In Situ Viral Dynamics off the Coast of Southern California using Transcriptomes from Automated High-Frequency Sampling

Marine viruses play an essential role in marine biogeochemistry, however little is known about their variation and activity over short time periods. Because viruses are active only through infection of cells, cellular metatranscriptomes provide a valuable approach to assess viral activities. This study used autonomous sampling to provide frequent, regular sampling with robust metadata for 1 L marine samples in the spring of 2014. The water was filtered through a 1µm filter, collecting the large/attached size fraction (1µm-300µm). Metatranscriptomes were prepared from these filters for 12 day and 6 night samples. Our analysis used FastViromeExplorer to track the abundance and expression patterns by recruitment of the transcript reads to two different viral databases. Previous work on the microbial communities showed the phytoplankton community in the large size fraction was usually dominated by Synechoccocus, and the non-phytoplankton was usually dominated by SAR11. We detected and quantified the expression of genes from ~125 viruses (>1000 reads each) based on a custom marine database of viral isolates and >1100 putative viruses using our marine database which includes viral contigs. The two most abundant and 11 of the top 20 viruses over the time series identified using the viral isolate database were Synechococcus viruses. Our contig database supported the dominance of Synechococcus viruses, composing 7 of the top 20 viruses. There was also a strong signal from Pelagibacter and eukaryotic viruses. The previous work on the cellular data showed strong diel patterns in several taxa including Synechocccus, however, the signal was not as strong in the viral transcripts, though a few strains showed diel cycling. Eukaryotic viruses were dominated by Ostreococcus viruses, and their abundance increased toward the end of the time series, which corresponded to an increase in their host abundance. The remaining eukaryotic viruses tended to follow similar abundance patterns to each other, where most increased and decreased together. This study suggests a stable coexistence between microbes and viruses, rather than dramatic boom-bust cycles, may be the most prevalent virus-host lifestyle during relatively stable periods, even on a daily time scale.