Non-Human Primates Harbor Diverse Mammalian and Avian Astroviruses Including Those Associated with Human Infections

Abstract

Astroviruses (AstVs) are positive sense, single-stranded RNA viruses transmitted to a wide range of hosts via the fecal-oral route. The number of AstV-infected animal hosts has rapidly expanded in recent years with many more likely to be discovered because of the advances in viral surveillance and next generation sequencing. Yet no study to date has identified human AstV genotypes in animals, although diverse AstV genotypes similar to animal-origin viruses have been found in children with diarrhea and in one instance of encephalitis. Here we provide important new evidence that non-human primates (NHP) can harbor a wide variety of mammalian and avian AstV genotypes, including those only associated with human infection. Serological analyses confirmed that >25% of the NHP tested had antibodies to human AstVs. Further, we identified a recombinant AstV with parental relationships to known human AstVs. Phylogenetic analysis suggests AstVs in NHP are on average evolutionarily much closer to AstVs from other animals than are AstVs from bats, a frequently proposed reservoir. Our studies not only demonstrate that human astroviruses can be detected in NHP but also suggest that NHP are unique in their ability to support diverse AstV genotypes, further challenging the paradigm that astrovirus infection is species-specific.

Fig 1. NHPs harbor diverse AstV genotypes.

RdRp sequences were aligned using MAFFT v7.058b and phylogenetic trees were constructed and evolutionary history inferred using the Neighbor-Joining method in MEGA6. GenBank accession numbers for the reference strains are given before the strain name and assigned or putative (in italics) AstV genogroups listed in parenthesis. Human viruses are in red and NHP samples in blue.

Fig 2. NHPs harbor AstV genotypes associated with human infections.

(A) Magnified HAstV clade from Fig 1. (B-C) Clustal W alignments on ~300 nucleotides from the ORF2 capsid gene of NHP BG36 (B) or ~900 nucleotides of NHP BG31 (C) were performed using BioEdit and MEGA6. Phylogenetic trees were constructed and evolutionary history inferred using the Neighbor-Joining method. GenBank accession numbers for the reference strains are given before the strain name and assigned AstV genogroups listed. Human viruses are in red and NHP in blue.

Fig 3. Evidence of recombination.

A cBrother analysis established the recombinant relationship between the human AstV lineage represented by sequence N.L23513 and the NHP AstV lineage represented by the sequence from BG31. Both the top and bottom panels share a common X-axis representing the position within the trimmed alignment. The top panel represents the posterior probability of ancestral assignment for the corresponding ancestral line at a given position within the alignment. The bottom panel represents the number of crossover points which occur at a given alignment position out of the 1000 samples taken from the MCMC chain.

Fig 4. Diversity metrics illustrated graphically for a given phylogenetic tree and tip labeling.

The tips of the focus community are colored blue, and represent either the monkey or bat virus population. The uncolored or red tips represent the reference community. Branches in each case are colored according to whether and how the corresponding branch lengths would be counted for the given metric. The descriptions of each metric and how they are computed are presented in Results.

Fig 5. NHP and Bat AstV diversity are differently distributed with respect to the diversity of other canonical AstVs.

As described in Results, MMCC, Phylogenetic Diversity (PD) and UniFrac metrics are computed for each of two phylogenetic trees—one built with NHP AstV sequences, the other bat AstV sequences, both with a shared set of “reference community” astroviruses. The columns of plots represent different clustering thresholds meant to account for novelty bias in bat sequence submissions, while the x-axis represents subsampling to account for differences in sampling depth between the two populations. As we attempt to correct for the possibility of novelty bias, the phylogenetic diversity of the NHP AstVs begins to match that of the bat AstVs. The MMCC and UniFrac metrics indicate that across subsampling depths and novelty bias correction levels, bat AstVs are on average more distant from reference community viruses, and form more isolated clades. The error bars and central points are computed via the 50% confidence interval quantiles and medians from the random subsamples, and smoothed using R’s LOESS fitting.