Host genotype and sex shape influenza evolution and defective viral genomes

Abstract

Viral evolution during initial pandemic waves favors mutations that enhance replication and transmission over antigenic escape. Host genotype and sex strongly shape this early adaptation, yet their individual and combined effects remain unclear. We experimentally adapted influenza A virus to male and female BALB/c and C57BL/6 mice, generating 28 independent lineages, and employed a novel "rolling sphere" approach to identify mutational hotspots in three-dimensional protein structures. In BALB/c mice, adaptation favored nonsynonymous substitutions linked to increased virulence, including a hemagglutinin variant exclusively fixed in female lineages. It also revealed the first demonstration of sex-dependent selection shaping a viral protein interface. In female-adapted viruses, substitutions disrupting a key NS1 dimerization motif converged on a single residue, while in male-adapted viruses, they were dispersed across the same interface. Conversely, adaptation to C57BL/6 resulted in fewer substitutions but promoted defective viral genome formation, leading to reduced cytopathic effect and attenuated virulence. This provides the first in vivo evidence that host genotype alone can modulate defective viral genome formation. Our results offer critical insights into host-pathogen interactions and reveal that selective pressures imposed by specific genotype-sex combinations can increase virulence across host genotypes, enabling new epidemiological modeling and disease control strategies.

Fig. 1 –

Titer and virulence measurements. A) Log₁₀ TCID₅₀ values for viral infections with ancestral HK68 (black, n = 10/host) or mouse-adapted virus (colored, n = 12/host). BALB/c females – open circles; BALB/c males – open squares; C57BL/6 females – solid circles; C57BL/6 males - solid squares. The line at 10³ TCID₅₀ represents the limit of detection. B) Log₁₀ copy number per μL of viral RNA (M segment) (n = 7/host). C) Correlation between log₁₀ TCID₅₀ and log₁₀ vRNA copy number. Spearman’s correlation coefficients (r_s) were calculated for all lineages of each host-adapted virus (n = 7/host). D) Weight loss (percent of initial weight) and E) survival curves for the ancestral and host-adapted viruses in the respective hosts. Black – Combined ancestral HK68 (n = 29); dashed red line – BALBF infections (n = 22); dashed blue line – BALBM infections (n = 20); solid brown line – BL6F infections (n = 22). Weight loss was modeled until 7 days post infection. Error bars are SEM.

Fig. 2 –

Three-dimensional protein models for HA (PDB: 7QA4), polymerase (PDB: 6QNW) and NS1 (PDB: 4OPH) adapted to each host. The mean variant frequency across seven independent lineages is represented by the thickness (HA, NS1, polymerase) and color intensity (HA, NS1) of the simplified amino acid chains. Hotspot nexuses are marked with solid lines when statistically significant in a host and dotted lines when present at high frequency in at least one lineage but not statistically significant.

Fig. 3 –

A, B) PCAs for the sum of cumulative frequencies at each site for 6 Å (A) and 12 Å (B) for each host group. Ancestral HK68 is marked as a black X. C) Sum of nonsynonymous variants above 5% frequency for each viral lineage. D) Frequency-weighted genetic divergence per site for each evolved sample compared to HK68. n = 7 per host type for all analyses. Error bars are SEM.

Fig. 4 –

Genomic deletion measures by host. A-D) Genomic position (x-axis), Frequency (y-axis) and length (width) of deletion variants for the polymerase segments and hemagglutinin. Lineages within each host are combined. E-J) Spearman’s correlations between TCID₅₀ and deletion number (E), deletion length (F) and deletion frequency (G), and correlations between genome copy number and deletion number (H), deletion length (I) and deletion frequency (J). Samples with mean polymerase coverage <100 reads are excluded. n = 7 per host type.

Fig. 5 –

A) Weight loss (mean percent weight lost per day) and B) survival curves for familiar (solid) and unfamiliar (dashed) host infections with BALBF and BL6F virus. BALB/c female mice - open red circles/red lines; C57BL/6 female mice – solid brown circles/brown lines. Mice that died are represented as crosses. BALBF familiar (n = 22); BALBF unfamiliar (n = 22); BL6F familiar (n = 22); BL6F unfamiliar (n = 14). Error bars are SEM.