Genomic analysis of oceanic cyanobacterial myoviruses compared with T4-like myoviruses from diverse hosts and environments

Abstract

T4-like myoviruses are ubiquitous, and their genes are among the most abundant documented in ocean systems. Here we compare 26 T4-like genomes, including 10 from non-cyanobacterial myoviruses, and 16 from marine cyanobacterial myoviruses (cyanophages) isolated on diverse Prochlorococcus or Synechococcus hosts. A core genome of 38 virion construction and DNA replication genes was observed in all 26 genomes, with 32 and 25 additional genes shared among the non-cyanophage and cyanophage subsets, respectively. These hierarchical cores are highly syntenic across the genomes, and sampled to saturation. The 25 cyanophage core genes include six previously described genes with putative functions (psbA, mazG, phoH, hsp20, hli03, cobS), a hypothetical protein with a potential phytanoyl-CoA dioxygenase domain, two virion structural genes, and 16 hypothetical genes. Beyond previously described cyanophage-encoded photosynthesis and phosphate stress genes, we observed core genes that may play a role in nitrogen metabolism during infection through modulation of 2-oxoglutarate. Patterns among non-core genes that may drive niche diversification revealed that phosphorus-related gene content reflects source waters rather than host strain used for isolation, and that carbon metabolism genes appear associated with putative mobile elements. As well, phages isolated on Synechococcus had higher genome-wide %G+C and often contained different gene subsets (e.g. petE, zwf, gnd, prnA, cpeT) than those isolated on Prochlorococcus. However, no clear diagnostic genes emerged to distinguish these phage groups, suggesting blurred boundaries possibly due to cross-infection. Finally, genome-wide comparisons of both diverse and closely related, co-isolated genomes provide a locus-to-locus variability metric that will prove valuable for interpreting metagenomic data sets.

Fig. 1

Overview of 16 cyanophage genome annotations. Each drawn box represents a predicted open reading frame (ORF) with forward strand ORFs above and reverse strand ORFs below. ORFs are colour-coded as per the legend in the figure, while colour-coded lines on the genome represent experimentally determined structural proteins (see Experimental procedures). For spreadsheet version of these data, please see File S1.

Fig. 2

T4-like gene set relatedness representations. A. Venn diagram illustrating the hierarchical core gene sets among 26 T4-like genomes. B. T4-like phage presence/absence gene cluster network. T4 Gene Clusters (T4-GCs) were used to construct a network of phage genomes and gene clusters found in one or more of the 26 genomes. Genomes are represented as diamonds, with cyanophage genomes coloured blue and non-cyanophage coloured red. Non-core T4-GCs are represented as a light purple circle, core T4-GCs shared by all genomes are coloured dark purple. If a T4-GC is present in a phage genome, an edge (green line) is drawn between that genome and the associated T4-GC. Genomes sharing many T4-GCs are in close spatial proximity to each other in the network. A multifasta file with all ORFs examined in this study is provided to link specific ORFs, T4-GC assignments and functional annotation (File S2).

Fig. 3

The core and pan-genomes of the (A) cyanophage and (B) non-cyanophage groups, where the core and pan-genomes are represented by square and triangles respectively. The core and pan-genomes were analysed for k genomes from cyanophages (n = 16) or non-cyanophages (n = 10). Each possible variation is shown as a grey point, and the line is drawn through the average. The core genome is defined as genes that are present in the selected k genomes. The pan-genome is the total unique genes found in k genomes. All variations of n choose k: n!/k!(n − k)!.

Fig. 4

Whole-genome pairwise comparisons across the bounds of the cyano T4 phage genome diversity are examined here. In all three panels, two genomes are compared where lines between the genomes connect homologues, coloured ORFs indicate genes that are unique to one genome or the other, and the per cent identity of each ORF is plotted in the lower half of each panel. Pairwise genome comparisons are presented for (A) two co-isolated cyanophages, P-HM1 and P-HM2, as well as (B) the three closest non-co-isolated phages, P-RSM4, S-SSM5 and S-SM1, and (C) the three most distant non-co-isolated phages, P-SSM2, S-PM2, Syn9, among the 16 sequenced cyanophage genomes.

Fig. 5

Close-up genome representation of the phosphate genes cluster from cyanophages. Genomic features are as described in Fig. 1. To orient the reader to the genome location of the cluster being portrayed, a box is drawn in a reference genome for each or a group of similarly placed phage gene clusters.

Fig. 6

Proposed role of 2-oxoglutarate (2OG) during cyanophage infection. A. In uninfected cyanobacteria, nitrogen limitation causes 2OG to accumulate, leading to 2OG-dependent binding of NtcA to promoters of nitrogen-stress genes, resulting in their expression. B. Phage infection draws down cellular nitrogen causing N-stress and likely leading to 2OG accumulation. Several cyanophage-encoded enzymes (in bold) suggest that increased 2OG may facilitate phage infection. First, a putative phytanoyl-CoA dioxygenase may convert 2OG to succinate, a major electron donor to respiratory electron transport in cyanobacteria (Cooley and Vermaas, 2001) thus potentially generating energy for the infection process. Second, 2OG-dependent dioxygenase [2OG-Fe(II)] superfamily proteins may function in cyanophage DNA repair (Weigele et al., 2007). Third, cyanophage genomes have multiple NtcA promoters driving genes encoding diverse functions – possibly exploiting the host NtcA-driven N-stress response system.

Fig. 7

Close-up genome representation of the carbon metabolic gene cluster from cyanophage genomes. Genomic features are as described in Fig. 1, and genome location orientation is as described for Fig. 5.