Fig 1.
Computational pipeline for structural and evolutionary analysis of H5N1 neuraminidase (NA1) and subtype-specific positions.
This figure presents the multi-phase computational pipeline used in the study. Symbols represent data and processes: cylinders indicate databases (green for PDB structural data, red for GISAID/BV-BRC sequence data), parallelograms denote data inputs/outputs (grey for raw sequences and processed data), and rectangles represent processing steps colored by phase. Dashed boxes group elements by phase, and ovals show final outcomes. Key visual elements include 3D protein structures to illustrate alignment, colored bars for Family-Specific Positions and conserved residues and DNA helix structure to illustrate NA1 from multiple hosts.
Table 1.
Pairwise sequence identity of monomeric chains from nine avian influenza neuraminidase subtypes used for structural superimposition.
Fig 2.
Structural superimposition of nine avian influenza A neuraminidase (NA) subtypes (N1–N9) illustrating conserved and variable regions essential for identifying Family-Specific Positions (FSPs).
(A) Shows the overlay of NA subtypes as cartoons highlighting shared folds and subtype-specific conformational differences. Structures are color-coded for clarity and as follows; 2HTY (Beige/Tan), 1NN2 (Sky Blue), 4HZY (Salmon or Light Coral), 2HTV (Light Green), 3SAL (Golden Yellow/Amber), 4QN4 (Deep Pink/ Fuchsia), 4QN3 (Light grey), 2HT5 (Pale Magenta), and 7NN9 (deep blue). (B) Zooms in on the conserved active site and 150-loop residues (labeled by 2HTY numbering), demonstrating their critical and consistent roles across all subtypes, which supports functional analysis and FSP identification via the ZEBRA2 algorithm.
Fig 3.
Phylogenetic reconstruction of nine influenza A neuraminidase subtypes.
This unrooted phylogenetic tree reconstructs the evolutionary relationships among 645 neuraminidase homologs, representing the nine major influenza A virus (IAV) NA subtypes (N1-N9). Inferred using PhyML + SMS version 3.0 based on the comprehensive structure-guided alignment (as detailed in Fig 1) with SH-like aLRT branch support, the tree distinctly groups the NA subtypes into their nine established clades. These clades further cluster into two broad phylogenetic groups: Group 1 (N1, N4, N5, N8) and Group 2 (N2, N3, N6, N7, N9). Notably, a distinct node representing Neuraminidase from Influenza B virus is positioned between these two major Influenza A groups, reflecting its evolutionary divergence. Each NA subtype is visually differentiated by uniquely colored eclipse-shaped nodes.
Table 2.
Classification and distribution of neuraminidase homologs into four statistically significant subgroups by Zebra2.
Fig 4.
ZEBRA2-derived subfamily-specific position profile of neuraminidase homologs.
This figure presents a ZEBRA2-generated subfamily-specific position profile (or conservation-variability profile), highlighting the top 41 statistically significant amino acid positions crucial for NA functional divergence across subtypes. Each row corresponds to a specific residue position from the comprehensive structure-guided alignment, with its Z-score (statistical significance of specificity) listed in descending order on the left, directly corresponding to its alignment position. Residue conservation is annotated by letter case: capital letters (>70% conservation) denote highly conserved residues, while lowercase letters (30-70% conservation) represent partially conserved. Where applicable, the three most frequent amino acids are listed. The figure includes an explicit, color-coded key that defines the residue identity based on nine distinct structural and functional groups for immediate interpretation. These groups are: Unique Backbone (G, P: Grey); Aliphatic Hydrophobic (A, V, L, I: light Green); Polar (Hydroxyl) (S, T: Light Blue); Sulfur-Containing (C, M: Deep Yellow); Aromatic (F, Y, W: Green); Basic (Stable Cation) (K, R: Deep Blue); Acidic (D, E: Red); Polar (Amide) (N, Q: Faded Pink); and Histidine (pH-Responsive/Catalytic) (H: purple). This profile directly visualizes the FSPs identified by ZEBRA2, pinpointing variations correlating with functional specialization within the NA protein family.
Table 3.
Statistical Significance for RMSD, Rg, and SASA. P-values were calculated using the Mann-Whitney U test, comparing the full trajectory of the mutant system to the Wild-Type (WT) ensemble. The statistical significance threshold (α) was set at 0.05.
Fig 5.
Molecular Dynamics Equilibration Trajectories for Native and Mutant Neuraminidase 1 (NA1) Constructs.
This figure presents the equilibrium trajectories for the MD simulations of the Native NA1 structure (PDB ID: 2HTY) and the five investigated mutant constructs. Each panel plots the System Density (kg/m3) against the Simulation Time (ns) for the duration of the equilibration phase (1.6 ns). The stability of all trajectories, characterized by minimal fluctuations and the maintenance of a consistent average value over the 1600-time steps, confirms that all simulation systems achieved thermodynamic equilibrium prior to the production runs. This stability is a prerequisite for reliable statistical sampling of the conformational landscape and analysis of dynamic properties. The mutants shown are: K207I, K207H, K207W, E229S, and G324T.
Fig 6.
Cα RMSD Analysis: Time Evolution and Ensemble Distribution.
The RMSD was calculated for the Cα atoms of the backbone relative to the energy-minimized structure of the WT protein. (A) Plots the RMSD (nm) over the first of the trajectory, illustrating the overall structural stability and convergence of the backbone. (B) Displays the KDE of the entire ensemble (300 ns total simulation). The KDE represents the probability density of sampling specific conformations, where the vertical dashed lines indicate the Mean for each system, visually confirming the statistically significant shifts in the ensemble (Table 3).
Fig 7.
Root Mean Square Fluctuation (RMSF) Profiles.
The RMSF of Cα atoms was calculated over the 300ns molecular dynamics trajectories to quantify the local flexibility of each residue along the protein chain. (A) Overlays the profile for the WT and all five mutants, enabling a visual comparison of highly mobile regions (peaks) across the entire sequence (Residue Number on the X-axis). Other Panels (B): Top Sub-Panel: Compares the absolute profiles of the mutant (colored dashed line) against the WT (black solid line). Bottom Sub-Panel (ΔRMSF): Plots the differential fluctuation (RMSFMutant −RMSFWT). Positive peaks indicate segments of the chain that became more flexible due to the mutation, while negative valleys indicate regions that became stabilized or more rigid. Each differential plot includes the Wilcoxon Signed-Rank Test -value, comparing the overall distribution of the mutant’s RMSF values to the WT, with the conclusion confirming whether the global change in flexibility is statistically significant.
Fig 8.
Ensemble distributions of size (Rg) and solvent exposure (SASA).
This figure presents ensemble properties of variants analyzed from molecular dynamics trajectories. Left panels feature box plots, with the central line indicating the median, box edges representing the interquartile range (IQR), and red circles marking outliers. Right panels display KDEs for each metric, illustrating the full conformational range, statistical mean (vertical dashed lines), and distinct shapes of conformational ensembles. The Rg (A) and SASA (B) plots demonstrate statistically significant shifts in ensemble properties for all mutants compared to the WT, confirming altered conformational landscapes. The precise median values and the statistical significance (P-values) for all systems are quantified and presented (Table 4).
Table 4.
Genetic diversity and selective pressure metrics of H5N1 neuraminidase (NA1) genes isolated from diverse hosts (2023–2024).
Fig 9.
Conformational Free Energy Landscapes (FELs) of the NA1 WT and Variants.
The FELs were constructed by projecting the 300 ns molecular dynamics trajectories onto the first two principal components (PC1 and PC2), which capture the maximum variance in the overall backbone motion. (A) The Wild-Type (WT) shows a bi-modal landscape with two connected deep minima (dynamic switching). (B) The K207W mutant exhibits a single, deep, rigid minimum (structural lock). (C) The E229S mutant displays a fragmented, shallow landscape indicative of maximum conformational entropy (disordered collapse). (D-F) show the landscapes for K207H (dynamic preservation), K207I (expanded disorder), and G324T (single-axis restriction), respectively.
Fig 10.
Dynamical Cross-Correlation Maps (DCCMs) of wild-type and family-specific position-based mutant NA1 catalytic domains.
The DCCMs for the WT NA1 catalytic domain (specifically, Chain A) and its FSP-based mutant proteins, were derived from 100 ns of molecular dynamics simulation. These maps visualize correlated motions between residue pairs, with red indicating strong positive correlation (residues moving in the same direction) and blue indicating strong negative correlation (residues moving in opposite directions). White and black rectangular boxes highlight regions (e.g., residues 1-40, 140-180, 200-240, 300-380) where the most notable alterations in correlation patterns are observed in the mutants compared to WT. These maps provide atomic-level insights into how FSP mutations perturb the protein’s internal communication networks and conformational flexibility.
Table 5.
McDonald-Kreitman test results for H5N1 neuraminidase (NA1) gene sequences, indicating selection pressures.
Fig 11.
Host-associated phylogenetic analysis of H5N1 neuraminidase (NA1) gene sequences from six host species.
This figure presents a comprehensive set of representative rooted phylogenetic trees of the H5N1 neuraminidase (NA1) gene, illustrating distinct patterns of genetic diversity across different hosts and geographical regions. The trees are displayed as follows: Panel (11A) displays the tree for Chicken isolates, Panel (11B) shows the tree for Turkey isolates, Panel (11C) shows the tree for Swan isolates, Panel (11D) shows the tree for Geese isolates, Panel (11E) shows the tree for Human isolates, and Panel (11F) shows the tree for Cattle isolates. For all panels, branch and node colors indicate the country of isolation as defined in the accompanying color legend, with the exception of Panel (C) for Cattle isolates, where node colors indicate the specific US state of isolation.