Host genotype and genetic diversity shape the evolution of a novel bacterial infection

Emerging infectious diseases cause severe wildlife declines and can arise via spillover, host-jumps, novel pathogen traits, or invasion into new environments. In novel host-pathogen associations, virulence and replication can evolve, but how host genotype and population genetic diversity shape pathogen evolution early post-emergence is unclear. Prior work in established systems shows host genetic diversity, spatial structure, and gene flow can influence pathogen evolution, with homogeneous hosts potentially fostering pathogen specialization. However, at emergence most hosts may be susceptible, potentially altering diversity effects. This study asks whether host genotype and population-level genetic diversity drive evolutionary changes in virulence, infectivity, and host range of a newly introduced pathogen. The authors experimentally evolved Staphylococcus aureus across 10 passages in Caenorhabditis elegans populations that varied in genotype (24 wild isolates in monoculture) and diversity (polycultures of all 24). They predicted: (a) genotype-specific increases in virulence/infectivity on sympatric hosts, (b) constrained evolution of virulence/infectivity in diverse host polycultures, and (c) specialization with reduced performance on novel hosts. They also examined the molecular basis and its relationship to host genetic distance.

The introduction synthesizes prior findings that pathogen adaptation can shape emergence and outcomes, and that host genetic diversity can limit disease spread, virulence evolution, evolutionary rates, and parasite adaptation in established interactions. Pathogens may specialize in homogeneous host populations due to selection for mutations neutral or beneficial in the focal host but costly on others. Models often assume diverse populations include immune individuals, reducing spread; yet at emergence most hosts may be susceptible, and diversity could modestly increase transmission via individuals with high contribution to spread (super-spreaders). Empirical gaps remain regarding how host genotype and diversity influence pathogen evolution immediately following introduction to a novel host.

Hosts: Twenty-four C. elegans genotypes were used (23 wild isolates randomly sampled across a phylogeny, plus lab-adapted N2). Host populations were maintained at 20°C on NGM with E. coli OP50. Life stages were synchronized by bleach treatment (1:1 NaClO:5M NaOH), eggs hatched overnight at 20°C, and L4s were washed off, cleaned of surface bacteria with M9 + Triton X-100 and multiple centrifugation washes.

Pathogen: Staphylococcus aureus MSSA476 (human pathogen isolate) was cultured (TSB, 24 h, 30°C, 220 rpm). A single ancestral stock was prepared and frozen with glycerol.

Experimental evolution: Pathogens were passaged for 10 host generations in: (i) 24 monoculture host populations (each a different genotype; one replicate per genotype), (ii) six polyculture replicates containing all 24 genotypes, and (iii) six no-host control passaging replicates. Exposures used 24-well plates with 500 µL 1.1% viscous media (TSB + HPMC cellulose) plus 100 µL concentrated S. aureus to reduce avoidance/behavioral biases. Approximately 500 L4 worms per well were incubated at 25°C for 24 h. After exposure, all worms were collected, washed thoroughly, and 10% of the population was bead-beaten; homogenates were plated on MSA to select S. aureus (37°C, 24 h). Approximately 100 colonies per replicate were picked and grown in 5 mL TSB (overnight, 30°C, 220 rpm) to found the next passage.

Trait assays: For virulence (pathogen-induced host mortality) and infectivity (infection load), ~200 worms per replicate were exposed (25°C, 24 h) in viscous media with pathogen. Mortality was scored by touch-response; no death occurred in food controls, while infection treatments yielded ~1–4% mortality. Infection load was quantified by washing and crushing 10 randomly selected worms per replicate and plating 100-fold dilutions on MSA for cfu counts. Assays were repeated five times. Specialization was tested by exposing evolved pathogen populations (from five monoculture-selected genotypes and five polyculture replicates) to their sympatric host and to a novel host genotype (CB4857); three technical replicates per exposure, infection loads from five worms per replicate. Degree of specialization was calculated as the difference in pathogen performance between sympatric and novel hosts.

Genome sequencing: For the ancestor and passage 10, 40 clones per replicate were grown and pooled for DNA extraction (Qiagen DNeasy). Libraries were prepared and sequenced (Illumina HiSeq4000, 150 bp paired-end). Reads were quality-filtered (fastp), mapped to the S. aureus MSSA476 reference (bwa mem -M), duplicates and singletons removed (samtools). Variants were called with GATK HaplotypeCaller using sample ploidy 40 (average coverage ~323x, enabling ~2.5% SNP frequency detection per clone), hard-filtered, and variants present in the ancestor and no-host control were removed. Evolutionary distances were computed as Euclidean distances between variant frequencies. Host genetic distances among monocultures were obtained from the C. elegans Natural Diversity Resource fixed variants.

Statistics: Analyses were conducted in R 3.6.0. Normality (Shapiro) and variance equality (F-test/Levene) were checked; outliers by Dixon test. Multiple comparisons were FDR-corrected. Binomial GLMs and one-way ANOVA tested variation in mortality and infection load across genotypes. Kendall’s rank correlations compared ancestral vs evolved infection loads; Spearman’s correlations assessed associations between infectivity and virulence post-evolution. Split-plot ANOVAs compared monoculture vs polyculture across time points; two-sample t-tests were used where assumptions held. Specialization indices were compared to zero via Wilcoxon or one-sample t-tests; monoculture vs polyculture specialization compared by Wilcoxon. Mantel tests related host and pathogen genetic distances.

• At introduction, there was no significant variation among the 24 C. elegans genotypes in pathogen-induced host mortality (χ²=28.738, d.f.=23, p=0.19) or infection load (ANOVA F=0.38, d.f.=23, p=0.99). Infection load and virulence were not associated initially (Kendall z=0.036, τ=0.002, p=0.97).
• After 10 passages, pathogen virulence (host mortality) varied significantly across host genotypes (χ²=39.875, d.f.=23, p=0.016), indicating genotype-specific selection on virulence. GO-term analyses suggested that differences in host functions related to metal ion binding were enriched among host genotype pairs exhibiting divergent evolved pathogen virulence, implicating host metal ion acquisition/sequestration as a driver of increased host exploitation.
• Infection load did not evolve to vary significantly across host genotypes (ANOVA F=0.56, d.f.=23, p=0.94), but ancestral and evolved infection loads across sympatric hosts were positively associated (Kendall z=8.49, τ=0.53, p<0.001), indicating overall escalation of infectivity.
• Post-evolution, infectivity and virulence were not correlated across sympatric host genotypes (Spearman ρ=0.015, p=0.87), suggesting independent evolutionary trajectories of these traits.
• Diverse host polycultures selected for higher pathogen virulence, while constraining infectivity compared with monoculture-selected lines (as summarized in the abstract), indicating contrasting selection on pathogen traits in diverse vs homogeneous host populations.
• Pathogen genomic evolution reflected host genetic relatedness: pairwise genetic distances among evolved S. aureus lines were positively correlated with genetic distances among their corresponding host genotypes (Mantel r=0.41 after 10 passages), showing greater divergence when selected in distantly related hosts.
• Despite divergence, S. aureus maintained a broad host range, with no evidence of strong specialization reducing performance on a novel host overall (per abstract summary).

The findings demonstrate that host genotype constitutes a distinct selective environment capable of shaping pathogen virulence evolution soon after introduction, even when initial host susceptibility is broadly similar. Variation in host metal ion binding/sequestration pathways likely modulates nutrient availability for bacteria, influencing the evolution of host exploitation and disease severity. Host population genetic diversity had opposing effects on pathogen traits: it promoted higher virulence yet constrained infectivity, possibly because heterogeneous host assemblages include individuals that disproportionately facilitate transmission or select for elevated damage, while simultaneously presenting diverse barriers that limit infection load increases. Importantly, infectivity and virulence evolved independently, underscoring the multifaceted nature of pathogen adaptation. At the genomic level, pathogen divergence mirrored host genetic distances, indicating parallel ecological speciation-like pressures without loss of broad host range. These results have implications for emergence dynamics in wildlife and human systems, suggesting that managing host genetic structure could influence pathogen evolutionary trajectories during early emergence.

This study shows that host genotype and population genetic diversity rapidly shape the evolution of a novel bacterial infection. After just 10 passages, S. aureus evolved genotype-specific virulence differences, likely influenced by host metal ion acquisition pathways, while infectivity increased but remained constrained in diverse host populations. Pathogen genomic divergence scaled with host genetic distance, yet host range remained broad. These insights highlight the potential for host genetic context to direct pathogen evolution during emergence and inform management strategies in conservation and public health. Future work should dissect the mechanistic basis (e.g., specific host metal sequestration and bacterial uptake pathways), track longer-term coevolution, quantify broader host fitness impacts (e.g., reproduction, population growth), and examine how varying levels and compositions of host diversity modulate pathogen adaptation and specialization over extended timescales.

• Assay limitation: inability to account for pathogen phenotypic adaptation to laboratory protocols, which may influence measured traits.
• Low absolute mortality levels (~1–4%) across treatments in the assay conditions may limit detection of subtle virulence differences and generalizability to natural settings.
• Hosts were kept evolutionarily static; coevolutionary feedbacks were not assessed.
• Sequencing pooled 40 clones per replicate at passage 10, potentially masking within-population linkage and clonal structure.
• Only 10 passages were performed; longer evolutionary timescales may reveal different dynamics, including specialization costs.
• Trait measurements focused on mortality and infection load; other fitness components (e.g., reproduction, population growth) were not measured.
• Laboratory exposure conditions (viscous medium, high doses, equalized exposure) may not reflect natural transmission or behavioral avoidance.