Untangling lineage introductions, persistence and transmission drivers of HP-PRRSV sublineage 8.7

Abstract

Despite a rapid expansion of Porcine reproductive and respiratory syndrome virus (PRRSV) sublineage 8.7 over recent years, very little is known about the patterns of virus evolution, dispersal, and the factors influencing this dispersal. Relying on a national PRRSV surveillance project established over 20 years ago, we expand the available genomic data of sublineage 8.7 from China. We perform independent interlineage and intralineage recombination analyses for the entire study period, which showed a heterogeneous recombination pattern. A series of Bayesian phylogeographic analyses uncover the role of Guangdong as an important infection hub within Asia. The spatial spread of PRRSV is highly linked with a composite of human activities and the heterogeneous provincial distribution of the swine industry, largely propelled by the smaller-scale Chinese rural farming systems in the past years. We sequence all four available modified live vaccines (MLVs) and perform genomic analyses with publicly available data, of which our results suggest a key "leaky" period spanning 2011-2017 with two concurrent amino acid mutations in ORF1a 957 and ORF2 250. Overall, our study provides an in-depth overview of the evolution, transmission dynamics, and potential leaky status of HP-PRRS MLVs, providing critical insights into new MLV development.

Fig. 1. Global spatiotemporal distribution of lineage 8 sequences.

Global spatiotemporal proportion and lineage distribution of 3708 lineage 8 sequences.

Fig. 2. PRRSV sublineage 8.7 global phylogeographic reconstruction.

A Maximum-likelihood phylogeny of global sublineage 8.7 strains until 2022 with countries annotated on the ancestral nodes and branches. Lineage 8 transition periods were attached in the right panel with different genotypes (deep pink: HP-PRRSV isolates, pink: transition isolates from classical CH-1a cluster to HP-PRRSV cluster). B Phylogeographic reconstruction of PRRSV sublineage 8.7. The line thickness signifies the captured spatial transmission routes. The colors of the circle and relevant routes correspond to the color of the ancestral nodes. C Zoom-in map represents the detailed transmission routes in Southeast Asia. The map showed in Panels B and C were generated using a custom-built Nextstrain pipeline.

Fig. 3. PRRSV sublineage 8.7 phylogeographic reconstruction within China.

A Maximum clade credibility tree with ancestral nodes and branches colored according to estimated (province) location, depicting the spread of PRRSV within China. B Spatial spread of PRRSV in China based on the posterior expectations of Markov jumps. In this plot, the colors of migration link of each location correspond to the source location. The thickness of migration link correspond to the values of Markov jumps. C Sankey plots summarizing Markov jump estimates for the transition between provinces. The plots show the relative number of transitions between origin (top) and destination (bottom) locations. Note that locations may both be origin locations (in the left row) and destination locations (in the right row), and there is no temporal order for the transitions involved.

Fig. 4. The support and contribution of PRRSV diffusion predictors among 30 Chinese provinces.

Support for each predictor is represented by an inclusion probability that is estimated as the posterior expectation for the indicator variable associated with each predictor. Indicator expectations corresponding to Bayes factor support values of 3, 20, and 150 are represented by a dotted vertical line in this bar plot. Here we only showed the predictors which had BF values >3. The contribution of each predictor is represented by the mean and credible intervals of the GLM coefficients on a log scale conditional on the predictor being included in the model. The support and contribution of all predictors was included in Supplementary Fig. S10.

Fig. 5. Recombination landscape.

A Phylogenetic network of full-length genomes of lineage 8, using the SplitsTree5 software with the Kimura 2-parameter model. Isolates in the red shaded region corresponding to the intralineage recombination within lineage 8, the green shaded region corresponds to the interlineage recombination with lineage8, with a statistically significant difference using Phi Test (p < 0.0001). B Overview of interlineage recombination patterns. The relative size of linkages from the upper part to the lower part correlates to the recombination frequency of each lineage as minor parent specific to the recombined region. For example, lineage 1 is more likely recombined as minor parent in ORF4-ORF7. C Overview of intralineage recombination patterns. The relative thickness of curve in the upper part correlates to the recombination frequency of each region as minor parent. For example, most intra-lineage recombination events result in a new doner in non-structural region. D Cumulative number of interlineage recombination events per year with color corresponding to different lineages. E Cumulative number of interlineage recombination events per year with color corresponding to different regions. F Cumulative number of intralineage recombination events per year with color corresponding to different regions. G Cumulative proportion of interlineage recombination events relating to each lineage in specific region of ORF1a (top panel) and ORF1b (bottom). H Cumulative proportion of interlineage recombination events relating to each year in specific region of ORF1a (top panel) and ORF1b (bottom). I Cumulative proportion of intralineage recombination events relating to each year in ORF1a (top panel) and ORF1b (bottom).

Fig. 6. Vaccine homogeneous analysis.

A TCS haplotype network reconstruction, with nodes colored by MLV clusters. MLV vaccines were annotated with a red star, with a single Vietnam isolate annotated with a triangle, indicative of the potential implication of MLV-related isolate transmission in the South-east. B Temporal homogeneous analysis of field strains in each vaccine cluster, respectively. The shaded area represents the 95% confidence interval of the fitted values using Poisson parameterization estimation.