Bacteria and viruses were collected from the littoral marine zone at Canoe Cove, Nahant, MA, USA, on August 22 [ordinal day 222], September 18 [261], and October 13 [286], 2010.

Bacteria were collected using a previously described size-fractionation method[1]. Bacterial strain naming convention is described using the example of 10N.286.54.E5: the first position (here “10N”) indicates the year (2010) and location (Nahant) of isolation, the second position (here “286”) indicates the ordinal day of isolation, the third position (here “54”) is a code representing the size-fraction of origin (0.2um: 45,46,47; 1um: 48,49,50; 5um: 51,52,53; 63um: 54,55,56), and the fourth position is the storage plate well identifier. Multiple codes within the size-fraction identifier reflect independent water samples for the 63um fraction, and independent water sample fractionation series for the other size classes (water sample A: 45,51,54; sample B: 46, 52, 55; sample C: 47, 53, 56).

Bacterial genome libraries were prepared for sequencing using a tagmentation-based approach and 1-2ng input DNA per isolate, as previously described[2]. Genomes were sequenced in multiplexed pools of 50-60 samples per Illumina HiSeq lane. Accession numbers for all bacterial genomes associated with this study are provided in Supplementary Table S1.

Bacterial phylogenetic relationships were determined by extracting ribosomal proteins from 278 genomes with hmmsearch[3] and aligning with MAFFT[4] as described in Hehemann et al. (2016)[5]. These strains were added to the Vibrionaceae ribosomal phylogeny used in Hehemann et al., 2016 and taxonomy was assigned using manual inspection. Full-length hsp60 sequences were also extracted from these genomes using hmmsearch with default parameters and the Cpn60 hmm (PF00118) from Pfam[6]. The hsp60 sequences were aligned using the mafft-fftnsi algorithm. Sanger-sequenced hsp60 fragments from 40 strains lacking genome sequences were added to this alignment using the mafft-fftnsi algorithm with the --addfragments option. The hsp60 alignment was concatenated to the ribosomal protein alignment and used to create a phylogeny using RAxML under a partitioned general time reversible (GTR) model (options: –q, -m GTRGAMMAX)[7]. SH-like supports were calculated using RAxML.

Viruses were collected using a previously described iron flocculation approach[8], using 4L sample volumes, 0.2um pre-filtration to remove bacteria, 0.2um filters for floc-capture, and oxalate solution for resuspension to maintain virus viability. Isolation of viruses was performed as follows. Iron-oxalate concentrate volumes equivalent to 15mL of seawater were mixed into agar overlays of 1,334 potential host Vibrio. The agar overlays were performed by combining 150uL overnight host culture and virus concentrate directly on a bottom agar (1% agar, 5% glycerol, 125mL/L of chitin supplement [40g/L coarsely ground chitin, autoclaved, 0.2um filtered] in 2216MB), directly pipetting 2mL of molten top agar (52 degrees C, 0.4% agar, 5% glycerol, in 2216MB) onto the bottom agar, and rapidly swirling to mix.  Plates were incubated for 2 weeks, and plaques were archived for later purification. Sequencing and genome analysis of viruses is described briefly, as follows. High titer lysates of serially purified viruses were concentrated using 30kD centrifugal filter units (Millipore, Amicon Ultra Centrifugal Filters, Ultracel 30K, UFC903024) and washed with 1:100 Marine Broth 2216 to reduce salts for nuclease treatment. Concentrates were brought to approximately 500uL using 1:100 diluted 2216MB and then treated with DNase I and RNase A for 65 min at 37 degrees C to digest unencapsidated nucleic acids. Nuclease treated viral lysates were extracted by addition of 1:10 final volume of SDS mix (0.25M EDTA, 0.5M Tris-HCl (pH9.0), 2.5% sodium dodecyl sulfate), 30 min incubation at 65C; addition of 0.125 volumes 8M potassium acetate, 60 min incubation on ice; addition of 0.5 volumes phenol-chloroform; recovery of nucleic acids from aqueous phase by isopropanol and ethanol precipitation. Genomes were sequenced in multiplexed pools using Illumina MiSeq and HiSeq technologies, assembled using CLC assembly cell, and manually curated to standardize genome start positions for the Caudovirales.

Viral strain naming convention is described using the example of 1.008.O._10N.286.54.E5, with specific identifiers separated by a period. The first position (here “1”) represents a unique identifier for each independent plaque isolated from a given host from the initial exposure of a given host to an environmental virus concentrate. The second position (here “008”) represents a unique working ID for a host strain. The third position (here “O”) indicates a unique sublineage generated from a single plaque during viral serial purification, for example due to the emergence of multiple plaque morphologies. Following the underscore is the full strain ID of the host of isolation, as described above.

Accession numbers for all viral genomes associated with this study are included under NCBI BioProject PRJNA328102.