Productive viral infections in oligotrophic marine waters

Abstract Viral lysis of dominant bacterial groups is accepted as the main mechanism through which viruses impact marine microbial communities. However, the structure of virus-bacteria infection networks and the abundance of their members have yet to be empirically characterized in situ. Here, we tackled this problem by building the virus-bacteria infection network of an oligotrophic environment using proximity-ligation, prophage integration, and CRISPR spacers. This network of 1388 previously undescribed virus-host pairs was integrated with community rank-abundance curves using direct counts of bacteria and viruses to identify productive viral infections. The same dominant viruses remained abundant across samples, but productive infections were poorly explained by viral and host abundances. This suggests that slow turnover rates maintain the abundance top-ranked, while viruses with medium-to-low abundances infecting heterotrophs are more productive. We predict that these non-dominant viruses contribute significantly more to biogeochemical cycling and host ecology due to their high turnover rates.