Ribosome-linked mRNA-rRNA chimeras reveal active novel virus-host associations.

Viruses of prokaryotes greatly outnumber their hosts1 and impact microbial processes across scales, including community assembly, evolution, and metabolism1. Metagenomic discovery of novel viruses has greatly expanded viral sequence databases, but only rarely can viral sequences be linked to specific hosts. Here, we adapt proximity ligation methods to ligate ribosomal RNA to transcripts, including viral ones, during translation. We sequenced the resulting chimeras, directly linking marine viral gene expression to specific hosts by transcript association with rRNA sequences. With a sample from the San Pedro Ocean Time-series (SPOT), we found viral-host links to Cyanobacteria, SAR11, SAR116, SAR86, OM75, and Rhodobacteracae hosts, some being the first viruses reported for these groups. We used the SPOT viral and cellular DNA database to track abundances of multiple virus-host pairs monthly over 5 years, e.g. with Roseovarius phages tracking the host. Because the vast majority of proximity ligations should occur between an organisms ribosomes and its own transcripts, we validated our method by looking for self- vs non-self mRNA-rRNA chimeras, by read recruitment to marine single amplified genomes; verifiable non-self chimeras, suggesting off-target linkages, were very rare, indicating host-virus hits were very unlikely to occur by mistake. This approach in practice could link any transcript and its associated processes to specific microorganisms.