Evolutionary rate of SARS-CoV-2 increases during zoonotic infection of farmed mink

Abstract

To investigate genetic signatures of adaptation to the mink host, we characterised the evolutionary rate heterogeneity in mink-associated severe acute respiratory syndrome coronaviruses (SARS-CoV-2). In 2020, the first detected anthropozoonotic spillover event of SARS-CoV-2 occurred in mink farms throughout Europe and North America. Both spill-back of mink-associated lineages into the human population and the spread into the surrounding wildlife were reported, highlighting the potential formation of a zoonotic reservoir. Our findings suggest that the evolutionary rate of SARS-CoV-2 underwent an episodic increase upon introduction into the mink host before returning to the normal range observed in humans. Furthermore, SARS-CoV-2 lineages could have circulated in the mink population for a month before detection, and during this period, evolutionary rate estimates were between 3 × 10^-3 and 1.05 × 10^-2 (95 per cent HPD, with a mean rate of 6.59 × 10^-3) a four- to thirteen-fold increase compared to that in humans. As there is evidence for unique mutational patterns within mink-associated lineages, we explored the emergence of four mink-specific Spike protein amino acid substitutions Y453F, S1147L, F486L, and Q314K. We found that mutation Y453F emerged early in multiple mink outbreaks and that mutations F486L and Q314K may co-occur. We suggest that SARS-CoV-2 undergoes a brief, but considerable, increase in evolutionary rate in response to greater selective pressures during species jumps, which may lead to the occurrence of mink-specific mutations. These findings emphasise the necessity of ongoing surveillance of zoonotic SARS-CoV-2 infections in the future.

Keywords: SARS-CoV-2; evolutionary rate; mink; molecular clock; spike gene; spillover.

Figure 1.

The illustration of the anthropozoonotic dynamics of mink-associated SARS-CoV-2, including human-to-mink, mink-to-mink, and mink-to-human transmission networks. To highlight the risk of viral reservoirs, we have shown a potential scenario of spill-back of mink-adapted lineages into the human population from infected wildlife.

Figure 2.

The phylogenetic analysis of mink-associated SARS-CoV-2 genomes, focusing on two geographical outbreaks from the Netherlands and Denmark. (A) Maximum likelihood tree of SARS-CoV-2 sequences (n = 269). The two mink-associated clades are highlighted in bubbles: the Netherlands clade (n = 29) is highlighted in the upper left corner and Denmark clade (n = 40) in the lower right corner, with the remaining tips representing human isolates (n = 200). The tree is rooted with SARS-CoV-2 reference sequence Wuhan/IVDC-HB-04/EPI_ISL_402120. Bootstrap replicates (n = 1000) >70 per cent are marked with an asterisk. The scale bar represents substitutions per site (subs/site). (B) A root-to-tip regression plot of the genetic distance vs time (in decimal years) of the 269 SARS-CoV-2 genomes used in this analysis. Sequences in the Netherlands group (n = 29) are represented by green dots, and Denmark sequences (n = 40) are represented by purple dots, with the remaining human isolates represented by black diamonds (n = 200).

Figure 3.

Violin plots of posterior density for the evolutionary rates (substitutions/site/year). (A) The mean evolutionary rates of each clock model, from left to right: strict clock, uncorrelated gamma distributed (UCGD) relaxed clock, and FLC models: clade only, stem only, and stem and clade. (B) Evolutionary rates estimated from FLC models using a conservative prior on clock rate. From left to right, estimates are shown for the foreground branches for the Netherlands, Denmark, and shared mink models: clade only, stem only, and stem and clade. Dashed lines represent the mean evolutionary rate (and 95 per cent HPD intervals) from the strict clock.

Figure 4.

Maximum clade credibility phylogenies highlighting the two mink-associated SARS-CoV-2 clades. Mink-associated clades are shown in green and purple, for outbreaks in the Netherlands and Denmark, respectively, and for each site of interest, tips with the mutation present are shown in red. Mutations (A) Y453F, (B) S1147L, (C) F486L, and (D) Q314K are visualised. Each tree is mid-point rooted for clarity, and a scale bar represents substitutions/site/year.