3.1. Evaluation of the effect of 2D nanosheets on SARS-CoV-2 spike protein
At the first stage of this study, the interaction of spike protein and 2D nanosheets that results in the deformation of spike protein and change in its transmissibility were examined (Fig. 2A). The total energy of the interactions, including van der Waals (vdW) and electrostatic, is shown in Fig. 2B. It is notable that in all cases, vdW interactions have a decisive role in total energy. Surface tuning can readily improve the interaction of nanosheets with spike protein, i.e., doping of graphene lattice and decoration of the surface with carboxylate groups boasted the electrostatic energy from − 5 kJ/mole to more than − 26 kJ/mole. It indicates the critical role of surface engineering. The overall interaction of spike protein is the result of the individual residues interactions. In this regard, the contribution of residues is plotted in Fig. 2C and in online resource Fig. S3. According to the results, twenty residues are involved in the binding of nanosheets to spike protein. Shifting from graphene and phosphorene to p-doped graphene alters the residues’ contribution to the overall binding energy. For instance, Histidine (His), Glutamine (GLN), Aspartic acid (ASP) have the strongest binding with graphene (-10.35, -15.24, -12.48 kJ/mole, respectively) [84], while in the case of phosphorene, they contribute to the total energy in the scale of other residues (approximately − 7.5 to 9.5 kJ/mole). The combination of phosphorous and carbon atoms to form the p-doped graphene allows the benefit from both properties in the attraction of spike protein. For example, the contribution of Serine (SER) residue for individual graphene and phosphorene nanosheets is -3.71 kJ/mole and − 8.57 kJ/mole, respectively, while the value drops to -11.25 kJ/mole for p-doped graphene.
The comparison of functionalized (-) p-doped and p-doped graphene demonstrates that the addition of carboxylate groups on the surface intensifies the interaction with negatively charged residues (e.g., Threonine (THR), Arginine (ARG), Valine (VAL)) that adds up to the total binding energy. Moreover, this modification not only increases the electrostatic energy but also makes vdW attractions between spike protein and the nanosheet stronger. The stronger the interaction, the more the spike protein is deformed. In the following section, for each case, a long time-scale simulation (3000 ns) was performed to test the effect of nanosheets on the deformed protein with the ACE2 receptor.
3.2. Evaluation of the interaction of SARS-CoV-2 spike proteins with ACE2 receptor
The study of structural changes in the spike protein is important for the diagnosis, treatment, and prevention of COVID-19. Therefore, an approach to analyze the nanosheets’ impact on this protein is to evaluate the binding affinity of the deformed spike proteins and ACE2. Figure 3A outlines the binding energy of pristine spike protein fragments (hereinafter called the control group) and the deformed protein fragments with different nanosheets. As it is clear, the lowest binding energy and consequently the most stable complex of spike protein and ACE2 is related to the pristine sample with ca. -346.7 kJ/mole. Considering the interaction of spike proteins deformed by phosphorene and graphene with ACE2, it is understood that the performance of all these nanosheets is similar; however, because of the lower biodegradability in vivo and higher cytotoxicity of graphene, phosphorene have priority in the biological environments [52]. Moreover, oxygen exerts dissociating effects on the phosphorene nanosheet [85]. Therefore, utilizing this nanomaterial in the facial masks and air filters that are exposed to air can weaken their antiviral effect since it can be degraded at a higher rate compared to graphene. To take advantage of both materials, a phosphor-doped graphene nanosheet was designed and considered in the simulations to monitor its impact on SARS-CoV-2 spike protein. Interestingly, docking analysis revealed that the nanosheet carries the advantages of both original nanosheets and deform the spike protein in the way that its affinity toward ACE2 is not significantly changed (E~ -230.5 kJ/mole). To improve the physicochemical properties of the resultant nanosheet, its surface was decorated with carboxylate groups to improve the electrostatic interactions of spike protein and nanosheet to result in more deformation of the spike protein. The binding energy of the affected spike protein with the functionalized (-) nanosheet revealed the interaction with ACE2 when compared with other spike proteins.
Structural and functional analysis of SARS-CoV-2 revealed that during the binding, spike proteins’ structure transforms from prefusion to postfusion, i.e., at the later stage, it forms a long needle-like structure [86, 87]. In this regard, to further quantify the binding of spike proteins (pristine and deformed) with ACE2 receptor, we examined the average radius of gyration and ΔRg (Rgfinal–Rginitial) for each nanosheet (Fig. 3B). The lower the ΔRg, the more advanced stage of binding the virus can go through. As it can be seen, the tightest structure is related to the control group, where the final Rg size is lower than its initial state as indicative of postfusion structure. Spike proteins distorted by functionalized (-) p-doped graphene and phosphorene nanosheets exhibit the most extended structures. The results confirm the inhibitory effects of the employed nano-objects.
MD simulation of the study cases reveals more detail on the interactions. Figure 3C and online resource Fig. S4 represent the Rg of spike proteins as a function of interaction energies for spike proteins with ACE2. For example, the minimum energy range locates at ca.-310 kJ/mole, 1.92 nm for the control group, while this value for the spike protein deformed with phosphorene and graphene shifts to ca. -185 kJ/mole, 3.55 nm, and ca. -210 kJ/mole, 2.1 nm, respectively. P-doping and functionalizing the graphene pushes the centers toward ca. -150 kJ/mole, 3.8 nm, and ca. -130 kJ/mole, 3.2 nm, respectively. It is obvious that surface engineering of the nanosheets provides higher energetic interaction (i.e., less stable) with ACE2 and the higher extended structure of spike protein. Altogether, these effects avoid spike protein to form a stable complex with ACE2 receptors.
Like other proteins, we can analyze the secondary structures of the spike protein of the SARS-CoV-2. The β-sheets and α-helices augment the stability of spike protein molecular structure, while coil, bend, and turn structures abate its stability. Figure 4A demonstrates the distribution of spike proteinʼs secondary structures after interaction with 2D materials. Snapshots of the structures are provided with the same color code as the diagram to provide better insight into the conformational analysis. The highest stability of the protein was observed in the case of the control group that is not deformed or distorted. Interestingly, the effect of p-doped graphene (26% of β-sheet and 5% of α-helix) was between graphene (28% of β-sheet and 5% of α-helix) and phosphorene (25% of β-sheet and 5% of α-helix). However, surface modification with carboxylate groups (20% of β-sheet and 3% of α-helix) augmented the instability within the structure. Therefore, it is understood that both the architecture of the nanosheet and its surface chemistry determines its inhibitory effect against the virus.
The mean solvent accessible surface area (SASA) of the spike protein-nanosheet and spike protein-ACE2 pairs are presented in Fig. 4B. Increasing the interaction between 2D structures and spike protein resulted in the adsorption of this protein on the surface of 2D structures and reduction in their contact area with aqueous media. According to the results of SASA analysis, spike protein had the lowest contact area with aqueous media after interaction with functionalized (-) p-doped graphene, indicating the stronger interaction of spike protein with the nanosheet, and thus, its greater effect on the deformation of the spike protein. Wang et al. [88] reported that hydrophobic interactions between spike protein and its receptor are determinants of the more transmissibility of SARS-CoV-2. Therefore, compounds that can abrupt the hydrophobic interactions between ACE2 and spike protein can be employed to combat COVID-19. In this regard, the effect of spike protein deformation by 2D structures on its interaction with ACE2 was investigated. After deformation and during the interaction with ACE2, spike protein is exposed toward aqueous media, which indicated the decreased hydrophobic interaction of the structures, both ACE2 and spike protein. As it is obvious, the largest SASA value is associated with the spike protein that is deformed by functionalized (-) p-doped graphene, indicating the exposure of spike protein to solvent molecules rather than the ACE2 receptor.
Hydrogen bonds (H-bonds) are also recognized as the main factor in the more contagion of SARS-CoV-2 as compared with previous coronavirus family members [88, 89]. Based on the results of a long time-scale simulation (400 ns), Ghorbani et al. [20] stated that SARS-CoV-2 forms H-bonds as double as the numbers that SARS-CoV (widespread in 2002) forms with ACE2. Therefore, it seems that prevention and reduction of H-bonds between ACE2 and SARS-CoV-2 can regulate its transmissibility and consequently harmful effect. Figure 4C shows the average H-bonds formed between the spike proteins and ACE2 in the course of 3000 ns interactions. Although all 2D structures reduce the hydrogen bonds between spike protein and ACE2 by changing the structure, the functionalized (-)-p-doped-graphene-deformed spike protein reduces the average number of H-bonds with ACE2 from ca. 36.5 (pristine spike protein) to 6 (deformed protein) where using graphene or phosphorene the number drops only to 18 and 12, respectively. Therefore, surface engineering of 2D materials can significantly improve their antiviral performance. To confirm the effect of 2D structures on the deformation of spike protein and consequently deterioration of its interaction with ACE2, the entropy of deformed spike proteins interaction with ACE2 was investigated using Schlitter’s entropy [85]. The higher difference between initial and final entropy displayed the stronger interaction of spike protein and ACE2. Figure 4D illustrates the mean entropy resulted from the interaction between the spike protein and ACE2 at different stages. The utilization of these 2D structures reduced the absolute energy and the difference between initial and final entropy. This results in the reduction of the absolute Gibbs free energy, showing the effectiveness of spike proteinʼs structure deformation by 2D nanostructures and consequently the hindrance of complex formation between spike protein and ACE2.
3.3. Translocation of 2D nanosheets through phospholipid membrane
As a result of the fact that the membrane supports the whole structure of the virus and especially keeps the spike proteins in place to fuse into ACE2 [90], flexing and rupture of the membrane can interfere with the function of other parts, such as spike proteins [91]. The membrane consists of hydrophilic, hydrophobic domains with carboxyl end groups that form ion channels [92]. It has been shown that lipid groups are the main components of the membranes [93, 94]. Hence, Eslami et al. [95] utilized the dipalmitoyl phosphatidylcholine (DPPC) lipid bilayer as the model for the SARS-CoV-2 membrane to investigate the effect of alcoholic disinfectants on the virus functions. Chen et al. [96] showed that modification of graphene surface to graphene oxide resulted in more attracted lipids toward the structure, which can be in favour of distorting the virus membrane. To view the interaction of nanosheets with SARS-CoV-2, the penetration of them through a phospholipid membrane was simulated. In line with previous research [95], we also have utilized DPPC bilayer in the microsecond simulations to study the nanosheetsʼ contacts with the virus through the phospholipid membrane. It worth mentioning that due to the formation of ion channels through the membrane, together with carboxylate groups, amine groups (-NH2) are also being added onto the surface of p-doped graphene to have an array of positive functional groups alongside negative groups. Further surface modification is designed to improve the interaction of nanosheets with DPPC.
Initial screening of results from the previous section showed that phosphorene, graphene, p-doped graphene, and functionalized p-doped graphene provide better interactions with spike protein. Therefore, these 2D nanosheets are considered for further assessments in microsecond-long simulations (Fig. 5A). As it can be seen, graphene and functionalized p-doped graphene can easily penetrate through the membrane, while phosphorene is stuck on the membrane surface at the end of 3000 ns. Strikingly, p-doped graphene cannot also translocate through the membrane during the long-time interaction. Consistent with previous results [97], the insertion of graphene in the membrane is assisted through the interactions of the membranesʼ hydrophobic segments with carbon nanostructure. Ion channels and the presence of negative and positive functional groups are the main reasons for the easy translocation of functionalized p-doped graphene.
The average energy of the interaction over the course of simulations is quantitatively calculated and represented in Fig. 5B for the 2D nanosheets. Consistent with the schematics of simulations, the most stable complex is corresponding to the functionalized (±) p-doped graphene. It can be attributed to the synergy of the hydrophobic and amplified electrostatic interactions, which cause easy and better penetration through the phospholipid membrane. Moreover, the H-bond formation between functional groups of nanosheets and various domains of phospholipid membrane assists the interactions. In this line, it can be observed (Fig. 5C) that the highest number of H-bonds are formed in the case of functionalized (±) p-doped graphene due to the presence of –COOH and NH2 groups.
Figure 5D shows the changes in the radius of gyration, defined as the initial radius of gyration minus the final radius of gyration. The greater these changes, the greater the nanostructure attraction energy to adsorb the phospholipid membrane. As shown, the largest change in the radius of gyration was related to functionalized (±) p-doped graphene. As a result of its strong electrostatic attraction, this structure can well absorb phospholipid membranes and shrinks membrane molecules. Overall, it is clear that the proposed nanosheets provide both successful interactions with spike protein and also their penetration through the phospholipid membrane bilayer seems promising, bringing the hope that they can be used to destabilize the whole SARS-CoV-2.
3.4 2D materials against SARS-CoV-2 protease
To characterize more inhibitory potentials of proposed 2D materials against SARS-CoV-2, we set additional coarse-grained long time-scale simulations to elucidate their impact on SARS-CoV-2 Mpro. Up to date, researchers have investigated various inhibitors against the Mpro, such as peptide-inhibitors [98–100], previously FDA-approved drugs [101, 102], and herbal components [103]. Our observations from the previous section confirmed that 2D engineered nanosheets are competent in blocking spike protein. These successful results encouraged us to consider their impact on the prevention of the infection propagation in a case that the virus can enter the body.
Since the binding site of the Mpro structure includes hydrophobic, hydrophilic, and charged residues (positive and negative), in this stage, we decorated the p-doped surface alongside negative functional groups (carboxylate) with positive functional groups (amine) to amplify its performance against Mpro. Figure 6A-i outlines various stages of simulations for only two cases (p-doped and functionalized (±) p-doped interactions with Mpro). In a recent study, Han and Kral designed and investigated four different peptides composed of components from the virus-binding domains of ACE2 against SARS-CoV-2. They suggested that peptide group named inhibitor 3 can be used as a therapeutic for COVID-19. In this part, we designed another functionalized (±) p-doped graphene with a conjugated inhibitor3 on one side (due to the larger size of peptide compared with nanosheet, only one inhibitor is designed on the surface). Figure 6A-ii represents three stages of simulation with inhibitor3-conjugated functionalized (±) p-doped graphene, and Fig. 6A-iii represents the inhibitor3 molecular structure. It should be mentioned that for the simulation of the interaction of inhibitor3-conjugated nanosheet, time was set at 5000 ns to ensure sufficient time for the process.
The results of the initial evaluations are reported in Table 1. The results of the average root-mean-square-distribution (RMSD) and the average root-mean-square-fluctuation (RMSF) depict details on the stability of the Mpro-nanosheet complexes. The RMSD results show the extent of the Mpro in an extended configuration, which indicates its activity. Clearly, the nanosheets capture the Mpro and reduce its activity. The lowest range of distribution is reported for the Mpro/inhibitor3 conjugated-functionalized (±) p-doped graphene complex. Similarly, the RMSF results represent that the inhibitor3 conjugated-functionalized (±) p-doped graphene narrowed down the broad range of fluctuations (8.34 nm for the control group) to less than half (1.29 nm). All of these findings confirm the excellent capability of the modified p-doped graphene in blocking Mpro. It is worthy of mentioning that in all results reported in the table, the performance of p-doped graphene is between phosphorene and graphene. It can be attributed to the properties of p-doped graphene that inherit from its constituents (phosphor and carbon). The lowered SASA values illustrate the prevented exposure toward aqueous media and blockage of active sites that inhibit the infection spread throughout the body.
To elucidate the inhibitory mechanism of the nanosheets, we considered more detail on the configuration and structure of Mpro after distortion phenomena. Figure 6B presents the average Rg and ΔRg observed in each case. Projections on XY and XZ planes are helpful to perceive the trend of change in the ΔRg and average Rg, respectively. The size of spheres exhibits a scale of the average Rg during the simulation, i.e., the smallest ball is related to the control group indicative of the packed structure of Mpro in comparison with its peers. Like spike protein, the reduced Rg size identifies the advanced stages of Mpro configurational changes during 3000 ns. Moreover, employing nanosheets prevents Mpro from advancing to further stages. The observations validate the previous findings on the inhibitory effect of nanosheets.
Table 1
The results of coarse grained simulations for the interaction of nanosheets with SARS-CoV-2 Mpro .
Nanosheet
|
Average RMSD (nm)
|
Average RMSF (nm)
|
Average SASA (nm2)
|
Control group
|
5.16
|
8.34
|
354.32
|
Bismuthene
|
4.93
|
7.26
|
350.21
|
Graphene
|
4.81
|
5.81
|
345.74
|
Phosphorene
|
4.54
|
4.45
|
341.89
|
P-doped graphene
|
4.62
|
5.36
|
342.11
|
Functionalized P-doped graphene
|
4.36
|
3.95
|
338.61
|
Functionalized (±) P-doped graphene with peptide inhibitor
|
1.15
|
1.29
|
296.62
|
To shed more light on the configurational changes during the simulations, the secondary structure of Mpro in each simulation is reported in Fig. 6.C. Apparently, the presence of nanosheets increases the coil configuration and reduces the β-sheet and α-helix percentage in the secondary structure of the protein, which leads to destabilization of Mpro. However, the α-helix percentage dropped from 11% to 4 % with surface modifications where the presence of peptide conjugate drops it to zero. It demonstrates that the formation of α-helix in the structure of Mpro is blocked through the addition of peptide inhibitors. The β-sheet percentage drops from 31% in the control group to 7% in the presence of the inhibitor-conjugated functionalized (±) p-doped graphene. All the reported data confirm the excellent capability of a designed nanosheet in the inactivation of the Mpro.
A more plausible explanation for the observations can be provided through energetic evaluations. Comparison of energy levels (Fig. 6D) gives more insight into the surface decoration with functional groups. It can be seen that surface engineering of p-doped nanosheet increases the electrostatic interaction energy from − 11.47 kJ/mole to ca. -67.63 kJ/mole. This value even reached more than − 200 kJ/mole with an inhibitor conjugated to the surface. It worth mentioning that with each surface engineering (addition of functional groups and peptide inhibitor), the electrostatic interaction had almost no significant contribution to the total energy. Moreover, surface modifications hinder the α-helix structure, meaning that the surface tuning has a critical role in the deactivation and destabilization of the virus. Moreover, H-bond formation was observed only between functionalized (±) nanosheet and protease, while the assessments showed no H-bond formation with other nanosheets. During the simulations, ca. 20 H-bonds form between Mpro and functionalized (±) p-doped graphene, where the number hits 64 bonds during the simulation. These observations are attributed to the presence of –NH2 and –COOH groups and a long peptide chain that all offer potential positions for H-bond formation.
From the energetic point of view, the performance of p-doped graphene is between graphene and phosphorene, showing that it possesses both properties. Additional surface tuning improved inhibitory effect of 2D nanomaterials through boasted electrostatic interactions and H-bond formation. Therefore, considering the total energy of the complex, functionalized p-doped graphene was considered as the most suitable nanosheet to attack the Mpro and destabilize the secondary structure of the protein.