Locality and diel cycling of viral production revealed by a 24 h time course cross-omics analysis in a coastal region of Japan

Abstract

Viruses infecting microorganisms are ubiquitous and abundant in the ocean. However, it is unclear when and where the numerous viral particles we observe in the sea are produced and whether they are active. To address these questions, we performed time-series analyses of viral metagenomes and microbial metatranscriptomes collected over a period of 24 h at a Japanese coastal site. Through mapping the metatranscriptomic reads on three sets of viral genomes ((i) 878 contigs of Osaka Bay viromes (OBV), (ii) 1766 environmental viral genomes from marine viromes, and (iii) 2429 reference viral genomes), we revealed that all the local OBV contigs were transcribed in the host fraction. This indicates that the majority of viral populations detected in viromes are active, and suggests that virions are rapidly diluted as a result of diffusion, currents, and mixing. Our data further revealed a peak of cyanophage gene expression in the afternoon/dusk followed by an increase of genomes from their virions at night and less-coherent infectious patterns for viruses putatively infecting various groups of heterotrophs. This suggests that cyanophages drive the diel release of cyanobacteria-derived organic matter into the environment and viruses of heterotrophic bacteria might have adapted to the population-specific life cycles of hosts.

Fig. 1

Virome abundance and viral transcript abundance in Osaka Bay. Sequence reads obtained from nine viromes a, b and nine metatranscriptomes c, d were mapped on OBV contigs (red), EVGs (blue), and RVGs (black). y-axis represents log2-scaled average FPKM (fragments per kilobase per mapped million reads) a, c and % of the read mapped region b, d. For each panel, only sequences with ≥ 1 mapped reads are shown. The numbers of the read mapped sequences are indicated on the top of panels a, c, with the total numbers of sequences in parentheses. The boxes represent the first quartile, median, and third quartile. metaT: metatranscriptome

Fig. 2

Rank-abundance curves of viruses with transcriptomic abundance. a Illustration of the Bank model proposed by [26]. b Average abundance of the nine time points for 878 OBV viral populations (contigs). y-axis represents normalized virome and metatranscriptome FPKM (i.e., each FPKM is divided by the FPKM of the most abundant sequence). c Abundance at 12:00 (the first sample). d Abundance at 15:00. metaT: metatranscriptome

Fig. 3

Infection dependent rank increases in virome abundance. a For each pair of different time points, it was investigated whether rank increased viruses show higher transcriptional activity compared with rank decreased viruses. Significance was detected by the Mann–Whitney U-tests (asterisks). Colors of sampling times represent sample groups (blue: oceanic, orange: coastal, gray: mixed). The statistical tests were performed within oceanic/mixed samples (areas in blue), and within coastal/mixed samples (areas in orange). b–d The transcriptional activities (i.e., metatranscriptome FPKM divided by virome FPKM) at the first time point were plotted against the virome abundance rank increases from the first time point to the second time point. Predicted host groups are color coded as in the panels. Ratios of rank increased viruses to rank decreased viruses, and p-values of the tests for each predicted host group are shown in the panels. For a–d, statistical analyses and plots were performed on OBV viral genomes whose FPKM values in the virome at the first time point were more than 10. Inf: infinite. NA: not available

Fig. 4

Summary of infection dependent increases. Different groups of viruses show different timings in their infection dependent rank increases. The tails (indicated by circles) and heads (indicated by triangles) of arrows indicate the two time points in comparisons. Colors of sampling times represent sample groups (blue: oceanic, orange: coastal, gray: mixed). Orange and blue arrows correspond to coastal/mixed and oceanic/mixed waters. For each pair of different time points, it was investigated whether rank increased viruses show higher transcriptional activity compared with rank decreased viruses. Significance was detected by the Mann–Whitney U-tests for each predicted host group (asterisks). The statistical tests were performed on OBV viral genomes whose FPKM values in the virome at the first time point were >10