E-protein variability in Zika virus strains: A possible new O-glycosylation site and its implications

(B) Molecular dynamics simulations

Molecular dynamic simulations were performed using Desmond module (Schrodinger). Factors analysed included root mean square deviation (RMSD), root mean square fluctuation (RMSF), the radius of gyration (Rg), and solvent-accessible surface area (SASA). Each of the 7 ZIKV E-protein models was solvated in predefined water solvent model and enclosed within orthorhombic periodic boundary conditions. Furthermore, the system was neutralised by adding sodium and chlorine ions, and energy minimised under the OPLS3e force field.¹⁹ The NPT ensemble was used to run the simulation at 300 K temperature and 1 atm pressure via AMBER22 for relaxation for 100 ns.^20,21

(C) Residue interaction network generation (RING) 4.0

The relation between seven E protein mutated residue with nearby residues was explored using 3D structural analysis of the residue interaction network (RIN).²² The RING 4.0 ( https://ring.biocomputingup.it/ ) produced the result in two steps, (i) identified the closest residues on the E protein, (ii) possible interaction(s). The network was analysed for residues in the form of nodes, and their bonding interaction in the form of edges. The network analysis considered different edge types for all possible interactions-van der Waals contacts, H-bonds, overlaps, and main-chain and side-chain interactions.

ZIKV evolution - Phylogenetic analysis

The iToL generated phylogenetic tree (Supplementary Fig. 1) revealed the divergence (through branch distance values) of ZIKV isolates, with the reference Indian strain (ZIKV REF MK238035.1) highlighted in red. There are 2 main clades of the ZIKV– one represented by the ancestral Senegal strain, while the other hosting the remaining strains. The Indian strain is the closest match to the Senegal strain, though they belong to different clades. The Dominican Republic strain is the farthest from the Senegal and Indian strains and the Haiti strain is also distant from both the strains, resting above the Dominican Republic strain. This suggests that the strains circulating in Asia, North and South America may have a common ancestor.

Whole genome comparison

The BRIG circular genome comparison of ten ZIKV virus genomes (Supplementary Fig. 2) against the reference Indian genome (MK2380351), highlights the degree of sequence conservation and/or divergence among different strains. The ZIKV virus genome is highly conserved, as shown by the predominant red shading, suggesting low genetic variability among these strains. However, specific genomic regions display reduced identity in the non-structural proteins (NS2A, NS2B, NS3, NS4B) and envelope glycoproteins (E, M), suggesting mutations (insertions, or deletions). ZIKV strain from Senegal, South Africa exhibits significant divergence from the other strains, which corroborates with the phylogenetic analysis results.

(i) Multiple sequence alignment, prediction of ZIKV E protein structure and their possible effects

Clustal omega MSA of 8 E-proteins (Supplementary Fig. 3) reveals that this protein is highly conserved in the studied ZIKV strains, with no changes in their amino acid sequences. Total 6 mutations were identified across four strains (Senegal, Brazil, India, Thailand) via MSA analysis, are listed below with possible biological effects:

• Senegal strain: ZIKV DakAr41667 strain (no. ASR91936.1) showed 3 mutations: (i) atno. 120, non-polar, hydrophobic alanine is replaced with polar, uncharged hydrophilic threonine (Thr) residue, (ii) at no. 169, isoleucine is replaced by valine, and (iii) at no. 285, non-polar, hydrophobic, aromatic phenylalanine has replaced polar, hydrophilic serine residue.

• Possible biological effects of the mutations: (i) Addition of Threonine at residue 120 may disrupt the hydrophobic packing and unfold or misfold the E protein, which may destabilise its conformation and functionality. This may also initiate and/or enhance the post-translational O-glycosylation protein modifications, thereby modulating immune evasion and receptor interactions. (ii) The swap at residue 169 may not significantly affect protein conformation since both amino acids are non-polar and hydrophobic in nature. (iii) The mutation at residue 285 may lead to conformational changes due to modification in the physicochemical properties of the residues.

• Brazilian strain (no. AYV74937): At residue no. 118, lysine replaced arginine both are charged, polar, hydrophilic amino acids with similar physicochemical properties, therefore, the overall E protein conformation may remain same ( Table).

• Indian strain (no. AZS35407.1): Polar, uncharged hydrophilic threonine at residue no. 280 was replaced by non-polar, hydrophobic alanine. This change abolishes any scope of O-glycosylation post-translational modification, which may interrupt biological functions.

• Thailand strain (no. QOF88708): At residue no. 255, valine mutated to alanine, both are non-polar, hydrophobic amino acids; this mutation may decrease the structural stability but not necessarily disrupt the global fold of E-protein.

(ii) Molecular dynamics simulation studies

These demonstrated mostly stable structures of the 7 mutants (Supplementary Fig. 4). Details are listed below:

• Root mean square deviation (RMSD) analysis: The calculation across the 100 ns time course showed clear patterns of stability between the 7E-protein mutants. The V255A mutant (Thailand strain) showed oscillations between mostly ∼2.0 and 4.5 nm, reflecting moderate conformational variations from the starting structure, while S285F (Senegal strain-single mutation) had a larger (∼3.0–6.0 nm) range with occasional spikes greater than 6.0 nm, reflecting greater flexibility. The triple mutant Senegal strain (A120T_I169V_S285F) exhibited the most fluctuations (∼3.0–7.0 nm) with several sharp peaks, indicating vast conformational changes. I169V mutant E-protein of the Senegal strain was relatively stable (∼3.0–5.0 nm) with moderate fluctuations, whereas T280A (Indian strain) E-protein exhibited the smallest RMSD range (∼2.5–4.0 nm) and few sharp fluctuations, suggesting increased stability. K118R (Brazilian mutant) showed steady fluctuations (∼3.0–5.5 nm) without spikes, and A120T (Senegal strain- single mutation) showed large variations (∼3.0–6.5 nm) with spikes towards the end of simulation, implying greater flexibility in post-stage periods.

• Root mean square fluctuation (RMSF) analysis: It showed residue-specific flexibility patterns unique to each protein mutant along the simulation trajectory. The V255A Thailand mutant demonstrated moderate fluctuations along most of the protein structure with prominent peaks at positions 12–15, 90–105, 145–155, 250–255, and 365–370. The S285F Senegal single mutant showed high flexibility at residues 5–15, 95–105, 145–155, and 250–255, while the triple mutant Senegal strain E-protein (A120T_I169V_S285F) showed sharp peaks at 95–105, 145–155, and 250–255, with otherwise moderate oscillations. The I169V Senegal single mutant showed peaks at 90–105, 145–155, and 245–255, and T280A Indian mutant showed increased movement at 5–15, 90–105, 145–155, and 250–255. The K118R Brazil mutant displayed noticeable peaks at 90–105, 145–155, 250–255, and 365–370, whereas the A120T mutant experienced higher oscillations at 90–105, 145–155, 250–255, and 365–370, with the remaining regions being quite stable.

• Solvent-accessible surface area (SASA) analysis: All E-protein variants were consistent during the simulation, with small fluctuations seen along the trajectories. The V255A Thailand mutant had SASA values mostly in the range of ∼19,900 to 21,200 nm², with fleeting peaks during the initial simulation phase followed by stable behaviour. The S285F single mutation Senegal variant had SASA values between ∼19,800–21,000 nm² with intermittent brief spikes but stability in the long run. The SASA values of the Senegal triple mutant strain (A120T_I169V_S285F) ranged between ∼20,000 and 21,000 nm² with moderate oscillations along the trajectory. The I169V single Senegal mutant had SASA values ranging from ∼20,000–21,200 nm² with periodic increases in the middle and end phases of the simulation. The SASA values of the T280A Indian mutant were oscillating around ∼20,000–21,000 nm² with short-term higher peaks at some intervals. The K118R Brazilian mutant showed SASA values of ∼20,000 to 21,500 nm² with visible spikes at the end of the simulation. Lastly, the A120T single Senegal mutant showed SASA values between ∼20,000 and 21,100 nm² with infrequent peaks and consistent trends throughout the majority of the simulation time.

• Radius of gyration (Rg) analysis: With 100 ns simulation time, the V255A Thailand mutant exhibited overall Rg profile oscillating around 34–36 Å, which indicates that the protein structure remains compact. The S285F single Senegal mutant displayed an Rg value ranging near 34–36 Å for the entire simulation time, which is reflective of a stable structural size. The Senegal A120T_I169V_S285F triple mutant had an Rg pattern largely between 28–33 Å with occasional insignificant fluctuations in the trajectory. I169V Senegal single mutation variant had Rg values with fluctuation mostly between 28–32 Å with a relatively stable profile. K118R mutant had an Rg range around 33–35 Å with minor transient fluctuations. T280A Brazilian mutant had an Rg range between 30–33 Å and revealed moderate stability in the compactness of molecules. Finally, the A120T Senegal single mutation variant had Rg values largely within the range of 30–34 Å, indicating a relatively stable structural organisation throughout the simulation.

(iii) Residue interaction network (RIN)

The RIN analysis was performed to assess the potential effects of respective mutations in E-protein through contact information provided by the RING 4.0, which includes hydrogen bonds, π- π stack, π cation, ionic interaction, van der Waals, and π-H bond calculated by RING, with detailed information displayed directly on the structure, and different representations of the contacts with residues. RING 4.0 tool processed the Senegal file ASR91936_A120T_2 and displayed the mutated residue Threonine (exposed on surface) at position no. 120 interacting with Serine at position no. 64 with one van der Waal 2.91 Å distance and three H-bonds. The Serine residue at position 64 is also interacting with Valine at position 257 with van der Waal force at 2.42 Å distance (Figure). As displayed by RING 4.0, Serine and Threonine both are exposed on the surface of the E protein. A 2D residue-residue interaction map depicted the interplay between these three amino acid residues, with Threonine, Serine and Valine interconnected with three hydrogen and two van der Waal bonds (Supplementary Fig. 5).

Close

Figure. Structural depiction of amino acid interaction networks associated with the mutation at position 120 in the Zika virus envelope (E) protein of the Senegal strain (ASR91936_A120T_2). (A) Illustrates the wild-type residue threonine at position 120, while panel B shows the mutated alanine at the same position. In the threonine-containing structure, Ser64 establishes a stabilizing interaction with Thr120 through three hydrogen bonds, complemented by a non-covalent van der Waals interaction at a distance of 2.91 Å. Additionally, Ser64 interacts with Val257 via van der Waals forces at a distance of 2.42 Å, contributing to the local structural stabilization of the E protein. In contrast, substitution of threonine with alanine at position 120 (B) alters the local interaction landscape due to the loss of the polar side chain, potentially affecting hydrogen bonding capacity and local conformational stability. These interaction changes highlight the structural consequences of the A120T mutation and its potential implications for E protein stability, dynamics, and functional interactions relevant to viral infectivity and antigenicity (figure generated using PyMol vs 3.1.3).

Export to PPT