Jorvik: A membrane-containing phage that will likely found a new family within Vinavirales

Summary Although membrane-containing dsDNA bacterial viruses are some of the most prevalent predators in aquatic environments, we know little about how they function due to their intractability in the laboratory. Here, we have identified and thoroughly characterized a new type of membrane-containing bacteriophage, Jorvik, that infects the freshwater mixotrophic model bacterium Rhodobacter capsulatus. Jorvik is extremely virulent, can persist in the host integrated into the RuBisCo operon and encodes two experimentally verified cell wall hydrolases. Jorvik-like prophages are abundant in the genomes of Alphaproteobacteria, are distantly related to known viruses of the class Tectiliviricetes, and we propose they should be classified as a new family. Crucially, we demonstrate how widely used phage manipulation methods should be adjusted to prevent loss of virus infectivity. Our thorough characterization of environmental phage Jorvik provides important experimental insights about phage diversity and interactions in microbial communities that are often unexplored in common metagenomic analyses.


INTRODUCTION
5][6][7] Nevertheless, recent microscopic, metagenomic and bioinformatic studies have revealed that non-tailed viruses are abundant in the environment, which emphasizes their potential importance and thus the necessity for further research. 5,8,91][12] Bacteria-infecting viruses of this kingdom belong to the class Tectiliviricetes and include some of the most prevalent predators of aquatic environments. 5,8The protein shell of Tectiliviricetes phages is reinforced by its anchoring into an internal membrane, which is likely to provide extra stability against mechanical stress. 13Based on the genome type, gene synteny and capsid characteristics, phages of this class are split into three families: Tectiviridae, Autolykiviridae, and Corticoviridae. 14ut of the three families, Tectiviridae is the best-studied.Phages belonging to the Tectiviridae contain a linear dsDNA genome and create a transient membrane protrusion with a characteristic tubular shape for genome delivery. 15Linear dsDNA phages belonging to the Autolykiviridae and Tectiviridae differ in their respective virion proteins, with the former being structurally similar to those of Corticoviridae phages.Unlike other phages of the lineage, Autolykiviridae phages often possess a broad host range, dominating the viral infection network of Vibrionaceae species. 5In contrast to these two families, phages belonging to the Corticoviridae contain a circular dsDNA genome.Members of this family have been shown to infect marine bacteria 16 but only one, pseudoalteromonal phage PM2, has been studied in detail. 1,4,7,17,18nterestingly, double jelly-roll fold capsids may combine even with ssDNA replication modules, 19,20 and much of the lineage diversity remains to be explored.
In this study, we identified a novel membrane-containing circular dsDNA phage, Jorvik -named after the Viking designation of the city of York where this phage was described.Jorvik is highly virulent and homologous prophages are abundant in proteobacterial genomes.We performed experimental validation for several phage components, critical for understanding how this group of phages function.Phylogenetic analysis suggests that Jorvik-like phages form a new family-level taxon within the Tectiliviricetes that shares a common ancestor with the Autolykiviridae and Corticoviridae families as well as recently described fNo16-like phages. 21We propose that all these phages should be classified in the same order, Vinavirales, and we refer to them as such in the rest of the manuscript.When working with an active virus stock, our experimental process ll OPEN ACCESS focused on the constant propagation of Jorvik with minimal culture manipulation.We suggest that this approach can be applied to the cultivation and characterization of other membrane-containing dsDNA viruses, allowing their hitherto unseen diversity to be revealed.

Origin of phage jorvik and its ability to spontaneously form more virulent variants
When cultivated in a defined RCV liquid medium, we observed that the Rhodobacter capsulatus B10 strain spontaneously produced bacteriophage particles capable of infecting the R. capsulatus SB1003, YW1, DE442 and St. Louis strains when they were grown in a rich YPS medium (Table 1).However, even for a relatively short time, the phage particles were unstable in liquid media, as demonstrated by a 90% decline in infective titer after just 24 h of incubation in YPS (Data S1.1).The phage was thus routinely propagated on YPS agar plates with a soft agar overlay, under aerobic conditions.For long-term storage, pieces of soft agar were taken from confluently lysed plates by a sterile inoculation loop, transferred to an Eppendorf tube, frozen and kept at À80 C.
The phage did not form plaques on cells growing under anaerobic conditions nor on aerobic RCV plates.The optimal temperature range for plaque formation on YPS agar plates was 20 C-34 C (Figure S1A).The plaques were 0.1-1.2mm in diameter and had a turbid halo with a less turbid center (Figure 1A).After several rounds of passage with R. capsulatus YW1, a plaque with a turbid halo and a completely clear center appeared (Figure 1B).We designated the phage Jorvik1, for the original turbid plaque variant, and Jorvik2, for the clear plaque variant.It is notable that spontaneous induction of two phages from R. capsulatus B10 that were capable of infecting St. Louis was previously observed by Wall et al. 26 These phages were not characterized in any detail; therefore, we were not able to make any comparison to phage Jorvik.
The titer of the phage was completely lost when the lysate was treated with chloroform, which is indicative of membrane-containing phages.Surprisingly, the titer declined substantially during repeated rounds of centrifugation and was negatively affected by filtration (Data S1.2).Cryo-EM imaging showed that phage particles extracted from confluently lysed plates were present in tightly packed formations rather than as individual virions (Figure 1C), with most of the grid being free of particles.Tight arrangements have been observed before when viral particles are at high density, 27 however, here Jorvik only reached relatively low titers in the range of 10 6 to 10 7 pfu mL À1 .These results suggest that in YPS media, Jorvik virions may behave as larger aggregates.

The Jorvik genome and virion morphology are characteristic of the Tectiliviricetes
The Jorvik virion is an icosahedron with approximately 59 nm diameter, decorated with short spikes and containing a spherical density most likely corresponding to an inner membrane (Figures 1D and 1E).The genome of Jorvik1 consists of 8,762 bp circular dsDNA (Figure 1F).The See also Table S3.a same results were observed for variants Jorvik1 and Jorvik2.b tested by positive PCR amplification of genes Slt and M15, spontaneous induction of YW1-infecting phage particles when cultivated in RCV; +, positive result; -, negative result; NA, data not available.
GC content is 62.4%, which is similar to the 66.6% GC of the host strain R. capsulatus SB1003. 22We identified 20 ORFs in the phage genome and predicted their function (Tables S1 and S2, Figure S2, and Data S2).The ORFs are encoded in three distinct operons, two on the forward strand and one on the reverse strand.The 8,889 bp genome of Jorvik2 differs from that of Jorvik1 in two loci.The first is a duplication of a non-coding region (position 60-186) located between the phage integration attachment site, attP, and a gene predicted to encode a helix-turn-helix (HTH) domain-containing protein, Gp1.The second difference is a single nucleotide change inside the gene encoding Gp1 that results in a W56L single amino acid substitution.Since both differences were mapped within or near gp1, we conclude that the product of this gene plays a role in the lysislysogeny decision.
We did not detect tubulation morphologies resembling those reported for Tectiliviridae and Autolykiviridae phages, either for Jorvik particles in solution or attached to cell membranes. 5,15Instead, we observed particles with ruptured virions from which DNA was escaping and cell membrane-attached particles that seem to lack part of the capsid at the attachment site (Figures S1B-S1G).It would be speculative to conclude that this rupturing is of biological significance, but nevertheless, these observations agree with the previously proposed hypothesis that the genome delivery mechanism of membrane-containing circular dsDNA phages differs from that of membrane-containing linear dsDNA phages. 28

Phage Jorvik1 integrates into the RuBisCO operon
The turbid zone of phage plaques is often caused by the integration of phage DNA into the genome of the host bacterial cells, resulting in lysogenic immunity against superinfection. 29,30To test if this was also the case for Jorvik1, the phage was plated on R. capsulatus St. Louis, which produced the most turbid plaques.The turbid zone of a single plaque was picked with a sterile tip and passaged on fresh plates to propagate phage-resistant mutants.Three individual colonies were subsequently tested for spontaneous induction of the phage, for resistance to Jorvik infection, and for the presence of Jorvik genes gp3 (M15) and gp15 (slt) by PCR (Data S3.1).One of the three colonies tested, designated St. Louis C1 hereafter, produced positive results for all three tests.After three passages, 10/10 St. Louis C1 colonies remained positive for the presence of the virus, suggesting stable maintenance in the host.Interestingly, we were unsuccessful in isolating a stable lysogen of the Jorvik2 variant when employing the same approach, testing 20 colonies from different susceptible R. capsulatus strains.
Genomic DNA of both the parental R. capsulatus B10 strain and the lysogenic St. Louis C1 strain was isolated and sequenced to identify the integration site of the phage.In both St. Louis C1 and B10, the phage was integrated into the bacterial genome within the RuBisCO cbbII operon between the genes encoding NAD(P)H-dependent quinone oxidoreductase (rcc01836) and phosphoglucomutase (rcc01837).Jorvik does not possess a predicted integrase gene but the attP and attB sequences show similarity to dif motifs recognized by a cellular XerCD recombinase system, 31,32 with the motif being conserved among Jorvik-like phages (Figure S3A).The actual location of the attB in the genomes of bacteria containing Jorvik-like prophages seems variable.It was recently hypothesized that the host XerCD is utilized for the integration of gammaproteobacterial membrane-containing dsDNA phages. 21Our results suggest the same mechanism for Jorvik-like phages.
Phage Jorvik has a narrow host range and requires the pleiotropic regulator CtrA for infection Phages Jorvik1 and Jorvik2 had identical host ranges for the R. capsulatus strains tested here (Table 1).Jorvik infected closely related strains, all of which had intact attB sites in their genome (Table 1).Apart from the original B10 source strain and the St. Louis C1 lysogen, PCR amplification of the M15 and slt genes did not produce a product for any other strain tested (Data S3.1); this suggests that there are no other lysogens of Jorvik among these strains.Bioinformatic analysis revealed the presence of CRISPR-Cas spacers matching the phage in diverse R. capsulatus isolates.These are strain A12 isolated in China, A52 and B41 isolated in Turkey and B6 isolated in the USA.Here, the B6 strain was shown to be resistant to infection by Jorvik (Table 1), suggesting that the system is functional.
To identify host genes that are required for successful phage infection, six available R. capsulatus SB1003 gene knockout strains were tested against both Jorvik1 and Jorvik2 (Table S3). 33Of the six strains, only one carrying a deletion of the pleiotropic regulator ctrA was consistently resistant to infection.The involvement of CtrA was verified by successful restoration of infection by in trans complementation with a ctrA plasmid (Table S3).The correct phosphorylation state of CtrA seems to be important for the phage infection because the phospho-null (D51A) and phospho-mimetic (D51E) forms do not fully complement the wildtype product; a similar effect was observed for R. capsulatus gene transfer agent production. 34We identified one putative CtrA-binding site, located in the intergenic region between the structural and packaging operon near a putative regulatory repeat (Figure S3B).It has been demonstrated that CtrA binds to the regulatory regions of several temperate-tailed alphaproteobacterial phages. 35Our results support the hypothesis that CtrA is widely involved in the regulation of lysogenic phages in the Alphaproteobacteria.

Growth characteristics show that phage Jorvik2 is highly virulent
Phage Jorvik adsorption kinetics followed a linear trend until approximately 60% of particles had adsorbed; the adsorption rate for this 10 min post-infection period was equal to 4.9 3 10 À10 mL min À1 (Data S1.3).This adsorption rate suggests that the phage recognizes 10-100 receptors per host cell. 36Over a more prolonged adsorption period, a subfraction of slower-adsorbing phage was identified.The adsorption efficiency of Jorvik reached 72% by the 20 min post-infection time point (Figure 2A).Inefficient adsorption of a sub-population is common in phages 37 and was reflected in the relatively broad rise period observed in the phage growth curve (Figure 2B).The latent period of the phage is 80-95 min and the titer peaks at $130 min post-infection, with the inflexion point of the curve at around 110 min.These times are longer in comparison to Alphatectivirus phages infecting faster-growing E. coli (inflexion point $45 min) 38 or the Corticovirus phage PM2 that infects Pseudoalteromonas ($70 min), 4 but shorter than that of tailed Roseobacter phage RDJL phi1 ($140 min). 39To our knowledge, apart from Jorvik, RDJLphi1 is the only Rhodobacteraceae-infecting phage with an estimated growth curve and more microbiological data are required to assess what variability in propagation speed exists among these phages.For the growth conditions tested, the burst size of Jorvik was estimated as 45 G 27 phage particles per infected cell (Data S1.4).
Jorvik phage can effectively suppress growth of all three R. capsulatus propagation strains -YW1, SB1003 and St. Louis (DE442 was not tested as it is an SB1003 derivative).With an initial multiplicity of infection (MOI) in the range of 0.0005-0.005(Data S1.5), the bacterial density starts to decline 5-6 h post-infection (Figures 2C-2E).Based on the phage growth curve, we estimate this as the end of the third lytic cycle of the phage.After >20 h of incubation, phage-resistant cell growth is noticeable in all three strains (Figures 2C-2E).Since the estimated MOI required for the clearance of the cultures was strikingly low, we tested the effect of cell growth state and the choice of titer method on the efficiency of phage plating (Data S1.2).The data showed that both variables affected the titer estimation by around 2-fold, adjusting our MOI estimate to 0.001-0.01.

Phage virions survive repeated freeze-thaw and pH shifts
The stability of the Jorvik virions in response to temperature, freezing-thawing, pH and salinity was tested.The phage was resistant to 30 min incubation at temperatures up to 45 C, but the titer decreased rapidly when incubated at 47.5 C and 50 C (Figure 3A).In the freeze-thaw experiment the titer dropped to an average of 32% after a single cycle, compared to a control incubated at room temperature (Figure 3B).Additional freeze-thaw cycles had only a minor additional influence on the phage titer.Incubation for 24 h in YPS media at pH 5 to 9 did not affect phage stability compared to the pH 7 control.There was a slightly larger decline in infectivity during incubation at pH = 10 and a complete loss of titer in pH = 4 media, even after 90 min of incubation (Figure 3C).When diluted in media of increasing salinity, the phage titer dropped to 33, 18, and 21% on average for YPS media with 0.1 M, 0.3 M, and 0.6 M NaCl, respectively, compared to YPS media with 0 M NaCl.After the initial osmotic shock, prolonged incubation in high salinity media did not lead to a larger decline of the titer compared to the decline in salt-free media (Figure 3D).

Proteomic analysis confirmed an abundance of predicted structural proteins in the virion-enriched sample
To assess which gene products are part of the Jorvik virion and indirectly confirm their function, an LC-MS/MS analysis was performed on a partially purified sample of Jorvik2 and the relative abundance of the proteins was estimated (Data S4).The partially purified sample was prepared by differential centrifugation as described in the STAR Methods section.A thorough purification was not possible due to the problems associated with different purification procedures and prolonged storage described above.The structural operon of phage Jorvik shares equivalent synteny and structural homology with the model phage PM2, thus we apply the established phage PM2 nomenclature for the Jorvik gene products.
Gp9, which corresponds to the putative major capsid protein P2, was the most abundant viral protein followed by the putative receptorbinding spike P1 (Gp14) and the putative structural membrane protein P3 (Gp10).Structural proteins P1, P2 and P3 are also the three most abundant proteins in phage PM2 virions. 7,18Another 13 phage gene products were identified in the sample, and all were present in comparable amounts to background host bacterial proteins (Data S4).All 13 products were encoded in the structural and packaging module of the genome.Four putative phage proteins were not detected by the MS analysis.These include the HTH-domain protein Gp1, replicase Gp2, putative holin Gp20 and, unexpectedly, a homologue of phage PM2 structural membrane protein P8 (Gp11).The presence of the very short Gp4 in the sample could not be estimated, as the theoretically acquired peptides after trypsin cleavage would be too short to analyze.

The phage encodes two lytic enzymes
Two gene products have predicted similarity to peptidoglycan hydrolases.One designated as Slt is 231 amino acids (AA)-long, located in the structural module, and has a C-terminal domain similar to the lytic transglycosylase P7 from phage PRD1 (HHpred; P27380, E-value = 9.7e-8, Score = 71.57;region 102-225).The second designated as M15 is 184 AA-long, located on the opposite strand packaging locus and is similar to peptidases from the M15 family (HHpred; PF08291.14, E-value = 1.4e-21,Score = 137; region 24-129).To confirm their function in R. capsulatus cell wall digestion, the genes encoding these products were cloned, overexpressed and the proteins purified using his-tag affinity chromatography.Both purified proteins showed R. capsulatus cell wall degradation activity in zymogram and plate lysis assays (Figure 4, Data S3.2).

Phylogenetic analysis classifies Jorvik-like phages as a novel family-level group within the Vinavirales order
Phylogenetic analysis of two conserved proteins, the packaging ATPase P9 (Gp5) and the major capsid protein P2 (Gp9) (Data S5), as well as Genome-BLAST Distance Phylogeny, 40 showed that phages Jorvik and Marinomonas phage YY form a unique family-level group within the Vinavirales (Figure 5).This is consistent with the level of differences observed during the manual inspection of genome synteny among the phages of Vinavirales (Figure 6).The major distinction between Jorvik and other phages of Vinavirales lies in a different set of regulatory genes, the replicase gene and lytic genes.The synteny of structural and packaging loci is more conserved, with several Jorvik proteins identified as homologous phage PM2 counterparts using HHpred 41 (Table S2).
To estimate the prevalence of prophages similar to Jorvik integrated into sequenced bacteria genomes, PSI-BLAST bioinformatic analysis of the two most conserved proteins, P9 and P2, was used to identify sequences with both products encoded within a 10 kb window.In total, 849 prophage hits were identified, belonging to 737 unique bacterial strains and 436 unique bacterial species (Data S6.1).These prophages were found predominantly in aquatic Alpha, Beta and Gammaproteobacteria species (Table S4).We manually inspected contig sequences from twelve selected representatives spread throughout the taxonomic lineages, and confirmed that they all encoded additional PM2-like proteins in the P9/P2 locus (Table S4).Clustering of P9 and P2 sequences showed the existence of putative prophage groups different from the known phages of Vinavirales 42 (Data S6.2-S6.4).Five prophages integrated in the genomes of Alphaproteobacteria clustered closely with Jorvik and Marinomonas phage YY, whose host is likely misannotated (Data S6.2-S6.4,Table S1).

Jorvik-like phages represent a new family-level group of membrane-containing circular dsDNA phages
Phage Jorvik is a unique double jelly-roll virus.This double jelly-roll classification is supported by the capsid diameter of around 59 nm, the presence of an inner membrane, the decoration of pentons with spikes (Figures 1D and 1E) and bioinformatic comparison with the archetypal Corticovirus phage PM2 (Figure 6; Table S2).To our knowledge, it is only the second member of the order Vinavirales that has been characterized in detail, after phage PM2. 1,6,7,18Moreover, it is the first characterized freshwater phage of the group.This does not come as a shock since our data demonstrate that established laboratory methods used for phage propagation and storage are unsuitable for work with these types of phages.Indeed, if a surreptitious lysogenic strain of phage Jorvik had not been available, the phage is likely to have been lost a long time ago.The modifications to existing phage isolation and manipulation methodologies, e.g., immediate processing of the isolation material without storage, reduced centrifugation and filtration steps, propagation of the phage via soft agar overlaying on a daily basis, long term storage of the phage in the form of frozen agar stabs, can be applied to the discovery and study of other membrane-containing dsDNA phages.
The most striking finding was that repeated rounds of centrifugation as well as filtration led to a rapid decline of the phage titer.Since centrifugation and filtration of samples is a common procedure when isolating and handling environmental phages, 43,44 many membrane-containing dsDNA phages could easily be lost during purification.Also, due to practical considerations, environmental samples are likely to be stored for various lengths of time prior to analysis.As demonstrated here for Jorvik, storage at 4 C for even one day leads to an order of magnitude loss of titer.Thus, the few membrane-containing dsDNA phages isolated so far are likely to be a gross underestimate of their true abundance.Interestingly, even though the survival of the phage in laboratory conditions was low, the virions were quite robust against pH shifts and freeze-thawing cycles.This contrasts with phage PM2 where one cycle of freeze-thawing led to >99% loss of titer. 7The virions also seem to favor low salt conditions reflecting the freshwater origin of this phage.
A good estimate of the environmental abundance of phages can be inferred from the presence of homologous prophages residing in bacterial genome sequences.In the case of phage Jorvik, more than 800 similar putative prophage sequences were found throughout the Proteobacteria (Pseudomonadota) phylum mainly in aquatic isolates.These included previously characterized phages belonging to Vinavirales, which group outside of Jorvik-like phage cluster (Figure 5, Data S5 and S6.2), suggesting Jorvik-like phages form a separate taxon at the same level as current Corticoviridae and Autolykivirdae families.Similar observations were recently made for fNo16-like phages infecting Vibrio species, 21 further suggesting that a major redefinition of taxonomy within the order Vinavirales is required.
The phage Jorvik2 is capable of clearing liquid cultures of Rhodobacter strains in several hours even when present at an MOI as low as 0.001 (Figures 2C-2E).This makes Jorvik2 highly virulent even when compared to potentially therapeutic phages, 45,46 and thus likely to be ecologically relevant.While we corrected the MOI based on our estimate of the efficiency of plating, one needs to consider that other unknown conditions may influence the real efficiency of plating and the measured MOI might still be underestimated.Interestingly, the fast-growing R. capsulatus strain St. Louis was capable of escaping clearance by the phage in one biological replicate (Figure 2E).This suggests that cells entering later growth phases become less susceptible to infection, which was supported by the observation that plaque production was hampered when the phage was plated with late log-phase to stationary phase cells compared to early log-phase cells (Data S1.2).No plaque formation nor clearing of the culture was observed when cells were infected in RCV defined media or during anaerobic cultivation.In all of these conditions, R. capsulatus cells tend to produce more extracellular polysaccharide, 47,48 which may hinder the phage's ability to infect.

Jorvik structural operon
The phage Jorvik genome is 8,762 bp, which is the smallest genome of the known membrane-containing dsDNA phages, and contains three distinct operons.The largest operon is a structural operon that is organized with equivalent synteny to other members of Vinavirales (Figure 6).The major protein products are P1 (receptor-binding spike), P2 (major capsid protein), P3, P7 and P8 (inner membrane protein), and P10 (unknown function).Interestingly, the putative inner membrane protein P8 was the only PM2 homologue that was not identified as part of the virion by mass spectrometry analysis.P8 is short and hydrophobic and thus hard to detect using MS methods, therefore its absence could be a false negative and its presence in the sample cannot be ruled out.Another interesting protein is a putative N-acetyltransferase designated Gp12, which is encoded between P8 and P10.The protein is conserved among the phages of Vinavirales, with the notable exception of phage PM2 (Figure 6). 42MS analysis detected only a very low amount of Gp12 in the phage-enriched sample, suggesting that it is not part of the virion but rather it is necessary for the acetylation of other structural components, similar to the function of Gp14 from Salmonella phage P22. 49omologues of phage PM2 conserved inner membrane proteins P5 and P6 could not be identified in Jorvik phages.In the genome, they are substituted with Slt and a hypervariable region encoding several short hydrophobic proteins Gp16-19 (Table S1).Since Gp16-19 are present in a structural module and were detected in the virion-enriched sample, they could be structural virion membrane proteins.Their low sequence conservation might be a result of co-evolution with bacterial host factors, for example, they could be involved in virus-host membrane fusion.A putative holin gene, gp20, is located at the very end of the structural operon.A homologue of this gene product from Marinomonas phage YY has predicted similarity to the R. capsulatus gene transfer agent holin 50 (HHpred; PF11351.11,E-value = 0.03, Score = 37.17; region 4-109 out of 118 AA).

Jorvik replication operon
The replication operon encodes two gene products.The first is the HTH domain-containing protein Gp1, which we propose is involved in the lysis-lysogeny decision of the phage.
The Gp1 protein of virulent phage variant Jorvik2 carries a mutation in a key structural residue and a duplication of the putative upstream regulatory region.It is likely that these changes promote the lytic replication, which explains why Jorvik2 forms clear plaques and why we could not isolate a stable lysogen of this variant.
The second protein encoded in the locus is the putative replication protein Gp2.Gp2 substantially differs from the replicases of characterized phages of the Tectiliviricetes (Figures 6 and S2D), having only 212 amino acid residues and two distinctive domains.In comparison, both PM2 and PRD1 replicases have more than 500 residues.The N-terminal domain of Jorvik replicase (34-116 AA) is similar to the lytic replication protein of coliphage P1 (HHpred; P19654, E-value = 2.1e-5, Score = 55.9); while the C-terminal domain does not resemble any characterized proteins.Due to the short length of Gp2, we speculate that additional host factors are required for Jorvik genome replication and that the C-terminal domain might be involved in the interaction with those factors.A recent metagenomic analysis reported that the major distinguishing factor of Vinavirales phages is the type of replicase protein, 42 which appears to coevolve with a specific host to allow for optimal function.

Jorvik packaging operon
The reverse strand packaging module encodes the packaging ATPase P9, putative endolysin M15, P15-like repressor Gp7 and two hypothetical proteins, Gp4 and Gp6.The function of Gp7 is predicted from the fact that its N-terminal domain shows high similarity to known repressors including P15 of phage PM2 (HHpred; Q9XJS8, E-value = 3.8e-6, Score = 59; region 12-61 out of 149 AA) and cI repressor of phage lambda (HHpred; P03034, E-value = 4.9e-5, Score = 51; region 5-58 out of 149 AA).The C-terminal domain is rich in prolines, and its function is unknown.Upstream of this gene lies an intergenic region (sequence 3263-3448) containing a putative CtrA-binding site (3265-3276), a noncoding repeat sequence identified by blastn (3337-3384), and two promoters for expression of two oppositely facing operons (Figure 1F).In phage PM2, the P15 repressor affects expression from two oppositely facing promoters 51 and Gp7 might work similarly.We believe CtrA, Gp7 and this intergenic region play a major regulatory role of the phage life cycle, but the exact molecular mechanism of the regulation remains to be determined.The pseudo-bootstrap support values from 100 replications are shown next to nodes, with an average support equal to 90%.The scale bar defines the branch length, scaled in terms of the D6 distance formula.The legend on the right defines different taxonomy rankings of the phages with the color coding differentiating between different taxon groups that were assigned within a rank.The color of the name depicts approved or officially proposed family ranks -Autolykiviridae (teal), Jorvik-like (magenta), Corticoviridae (green), fNo16-like (orange) and Tectiviridae (blue).See also Table S4 and Data S5 and S6.
The function of protein Gp6 may lie in virion membrane assembly.The gp6 gene precedes the packaging ATPase ORF, which is an equivalent location to the membrane assembly factor gp-h of other Vinavirales phages.It also possesses a weak similarity to protein P10, the gp-h homologue in phage PRD1 42 (Figure S2C).Moreover, the protein was abundant in the partially purified phage sample.This can be explained by the incorporation of the protein in mature virions or virion assembly intermediates.The conservation of a virion membrane assembly factor across genome packaging modules of Tectiliviricetes phages is striking and more research is required to explain if and how the factor is involved in the packaging process.The existence and potential function of the last hypothetical product of the operon, Gp4, remains unknown.

Jorvik lytic enzymes
Phage Jorvik encodes two lytic enzymes.Slt (Gp15) is encoded in the structural operon while M15 (Gp3) is encoded in the packaging operon.Both Slt and M15 were identified in the phage-enriched sample by MS, with Slt being more abundant.Many phages encode two PG hydrolases, a virion-associated hydrolase and an endolysin. 52,53The former is required for creating a pore in the cell wall during the injection of DNA into the host.The latter is required for the release of infectious virions after their assembly.In the phages of Vinavirales, functional homologues of Slt in the same genome position can be identified (Figure 6). 42In contrast, M15 homologues were found only among Jorvik-like phages infecting Alphaproteobacteria and as such are unique.Based on the genome position and MS data, we hypothesize Slt is a virion-associated hydrolase while M15 is an endolysin, analogous to the P7 and P15 hydrolases of phage PRD1. 54he isolation and detailed characterization of phage Jorvik prove that membrane-containing dsDNA phages of Vinavirales are widespread predators of aquatic niches.Modifications to classical laboratory methodologies and approaches described here can be applied in the cultivation of other membrane-containing dsDNA bacterial viruses and may drive a revolution in their study.

Limitations of the study
The instability of the phage in laboratory conditions might have affected the outcomes of some microbiological assays such as phage growth curve and burst size determination, leading to an underestimation of the real values.Also, due to this instability only a partially purified sample of the phage could be analyzed by mass spectrometry, thus the virion association of individual components needs to be interpreted carefully.The study did not target the molecular mechanisms of phage integration and operon regulation, the mechanisms proposed were based on similarity with homologous proteins and the behavior of mutant phenotypes.Due to very low sequence similarity among proteins of Tectiliviricetes, the support of several phylogenetic tree branches is low.For this reason, we did not provide any extensive interpretation of the topology of trees.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:  S2.The scale bar at the bottom shows the length of the genome in base pairs.Phages are grouped according to the replicase gene type.Genes are colored according to their putative function as explained in the legend.

Bacterial and virus strains
See Tables 1 and S3

EXPERIMENTAL MODEL AND SUBJECT DETAILS Bacterial strains
The bacterial strains used in this study are summarized in Tables 1 and S3.In addition, several Escherichia coli strains used for cloning are specified in the description of the corresponding methods.Rhodobacter strains were grown either in minimal media RCV 71 or rich media YPS, 26 E. coli strains were grown in LB.Rhodobacter strains complemented for the production of CtrA were cultivated in the presence of 10 mg.ml -1 of kanamycin.Unless otherwise stated, Rhodobacter strains were incubated aerobically at 30 C and E. coli strains at 37 C.For anaerobic cultivation, bacteria were incubated in sealed 15 ml screw-cap glass tubes filled with YPS up to the rim or YPS plates in Anaerocultâ A jars (Merck Millipore).The samples were incubated at 30 C in an illumination cabinet, placed 30 cm from three 40 W light-emitting tubes (Panasonic FL40SS・W/37c).

Bacteriophage Jorvik
The phage was propagated using YPS soft agar overlaying.The overnight culture of the propagation strain was mixed with phage lysate, then 3 ml of soft agar (YPS + 0.7% w/v agar) were added and overlaid on a YPS agar (1.5% w/v) plate.The plate was incubated at 30 C overnight.
The soft agar containing phage particles was harvested using a bacterial loop and resuspended in YPS media.The tube was vortexed vigorously and centrifuged at 6000 RCF for 4 min to separate the agar from the phage solution.Repeated or high-speed centrifugation was avoided as we observed that these procedures led to substantial loss of plaque forming units.Filtration of phage lysates was not performed as this also led to a large reduction of phage titre; instead, the supernatant was transferred to a clean tube using a 30G gauge needle, stored in the fridge and used for any experiments within 8 h.For pH and salt stability assays, soft agar containing the phages was resuspended in 5-10 times of excess of YPS by passing through a 30G gauge needle and directly used for experiments.This led to one order of magnitude higher titres, but with the presence of residual agar and cell debris in a lysate.Unless otherwise stated, the phage titre was estimated by a drop assay.Here, 10 ml of diluted samples were dropped onto freshly solidified soft agar containing early log phase culture of Rhodobacter capsulatus YW1 (OD 600nm = 0.1-0.25)and left with an open cover for 7-10 minutes in a flow hood to dry.For estimating the titre from soft agar overlay, the host cell culture was mixed with diluted phage samples before the addition of soft agar and subsequently plated.For the estimation of efficiency of plating for bacterial cultures at different growth phases, OD 600nm of log and stationary phase cultures was measured, cells diluted to OD 600nm =0.10 in YPS and 250 ml of the culture was used for phage plating either by soft agar overlay or drop assay as stated.

Standard laboratory handling stability assay
For the chloroform treatment, 5 % v/v of chloroform were added to the lysate.The sample was vortexed, incubated for 10 minutes on a bench, centrifuged at 6000 RCF for 4 min, and 100 ul of the supernatant from the top of the tube was transferred to a new tube and titred.For the phage filtering, Merck MillexâGP PES Membrane 0.22 mm (Merck Millipore), Minisart NML Plus cellulose acetate membrane 0.45 mm (Sartorius), yellow syringe filters PES membrane 0.45 mm (Starlab), Nylon 66 membrane 0.2 mm (Supelco) and Duraporeâ PVDF membrane 0.22 mm (Merck Millipore) filters were used to filter 2 ml of phage lysate.For centrifugation experiments, 500 ul of phage lysate were spun at 6000 RCF for 4 min, and 100 ul of the supernatant from the top of the tube were subsequently transferred to a fresh tube and titred.

Isolation of lysogenic strain
Phage lysate was mixed with the Rhodobacter capsulatus St. Louis overnight culture and plated using a soft agar overlay.After the overnight incubation, several plaques were picked with a sterile tip and streaked onto fresh plates and incubated for two days to get individual phageresistant colonies.Single colonies were restreaked onto a fresh plate and subsequently inoculated into RCV.After two days of incubation in RCV, the cells were spun, and the supernatant was dropped onto St. Louis WT strain to discover phage-inducing lysogens.Colonies positive for phage production were subsequently tested for the presence of the phage genes gp15 (slt) and gp3 (M15) by PCR.To verify the colonies do not come from the parental B10 strain, after obtaining the lysogen sequence, the average nucleotide identity between St. Louis C1, B10 and St. Louis WT strains was calculated using OrthoANIu web tool. 70For St. Louis C1 and B10, contig sequences obtained from Illumina sequencing were used as an input.For St. Louis WT, contig sequences deposited under GenBank project NZ_VIBE00000000.1 were used.This resulted in 99.95 and 99.47 % average nucleotide identity when St. Louis C1 was compared to St. Louis WT and B10 respectively.

Growth characteristics
The growth of different strains in presence of phage was determined using OD 600nm measurement on Spectrostar Nano Microplate Reader (BMG LabTech).The strains were grown to OD 600nm 0.2 to 0.5 and diluted to a final OD 600nm =0.1 in 200 ml of media containing between 2 and 7 3 10 4 phage particles per well (Data S1.5).The plate was incubated at 30 C, and the OD 600nm was measured at 30 min intervals for 20 h.The plates were shaken orbitally at 100 rpm for 10 s before each measurement.The range of phage titre rather than specific titre value was used for growth experiments because of the inherent laboratory instability, a fresh phage lysate was used for each experiment without previous knowledge of its precise titre.
The growth curve of phage was determined using an approach similar to Kropinski, 2018. 72YW1 strain was used as a host.Log phase cultures (OD 600nm = 0.2-0.5)diluted to OD 600nm = 0.1 and 2 3 10 4 to 7 3 10 4 phages were added to 3 ml of final culture volume.The mixture was then incubated at 30 C and gently vortexed before sampling each time point.Ten and hundred-fold dilutions of the infected cells were prepared one hour post-infection, which was still in the lag phase of the phage.For sampling, 200 ml aliquots were spun at 12 000 RCF for 2 min and plated with log-phase YW1 strain using soft agar overlaying.Plaques were counted the next day.The burst size was calculated as the difference between the averages of plaque-forming units before and after the lytic burst.
For the adsorption assay, an approach similar to Kropinski, 2009 36 was employed.In brief, 100 ml of between 1 3 10 5 and 9 3 10 5 phage was added to 900 ml of the log-phase grown culture, with a final OD 600nm of 0.1.The mixture was incubated at 30 C, shaking at 90 rpm.The aliquots were diluted 20 times in ice-cold YPS, spun at 12 000 RCF for 2 min and 200 ml were plated with a 100 ml of log-phase YW1 culture using the soft agar overlay.Plaques were counted the next day.To estimate the adsorption rate, the equation from Kropinski, 2009 36 was used.

Stability characteristics
For the heat stability assay, 250 ml of phage lysate was incubated in a heat block set at a specific temperature for 30 min.The titre of the phage was then estimated using a soft agar drop assay.For the freeze-thawing experiment, 250 ml of phage lysate with the titre between 1 3 10 5 and 1 3 10 7 PFU.ml - was frozen at -80 C for 30 min and then thawed at room temperature for another 30 min.After the last thawing cycle, the phage titre of all samples was estimated by a soft agar drop assay.For the pH and salinity experiment, YPS media was modified to different salt concentrations and pH by the addition of NaCl and HCl/NaOH respectively.Precise pH was determined using pH electrode InLab Easy BNC (Mettler Toledo) and adjusted as necessary.The phage lysate with the titre between 1 3 10 5 and 1 3 10 7 PFU.ml - was diluted 1:19 with the modified YPS media and incubated for 1.5 h at room temperature.After plating the phage, the lysates were transferred to the fridge and the titre was estimated again after an additional 22.5 h (24 h in total) of incubation in the modified YPS media.All the growth and stability experiments were done in three independent biological replicates, utilising different overnight cultures and different phage lysates.

DNA isolation and sequencing
The genome of Jorvik1 was isolated from 18 ml of lysate with a titre of 1 3 10 7 PFU.ml - .The phage was pelleted by the addition of PEG8000 to a final concentration of 10% (w/v), incubation at room temperature for 10 min and centrifugation at 10 000 RCF for 10 min.The supernatant was discarded, and the pellet was resuspended in 500 ml phage buffer (50 mM NaCl, 10 mM Tris pH7, 10 mM MgSO 4 ).Cellular nucleic acids were removed by digestion with 0.2 mg.ml -1 DNaseI and 0.02 mg.ml -1 RnaseA and incubation at 37 C for 1 h.EDTA was added to a final concentration of 50 mM, then encapsidated DNA was extracted using the ''Purification of Nucleic Acids by Extraction with Phenol:Chloroform protocol''. 73Purified DNA was linearized with blunt ends using EcoRV (New England Biolabs) and inserted into the ZeroBlunt Topo vector (Thermo Fisher Scientific).The genome of Jorvik2 was isolated from 6 ml of a lysate with the titre 4 3 10 6 in a similar manner.The integrity and amount were sufficient for direct sequencing.The genomic DNA of strains B10 and St. Louis C1 were isolated from 1.5 ml of stationary aerobic culture grown in YPS according to the published manual. 73All the samples were sequenced using Illumina sequencers (HiSeq/ NovaSeq) and a 250 bp paired-end protocol.Genome sequencing was provided by MicrobesNG (http://www.microbesng.com).

Estimation of phage integration site in St. Louis C1
The raw reads of St. Louis C1 were searched for the integration site by filtering for reads containing the first and last 17 nt of the NODE_41 contig sequence, which contained the circular sequence of the phage, using the ''grep'' command.This resulted in 596 reads, including the reverse complements.The sequence upstream to the start of the NODE_14 contig sequence and downstream to the end of the sequence was analysed in Vim text editor, resulting in two populations, one (479 reads) corresponding to the phage circular sequence and the other (91 reads) corresponding to the bacterial integration site in cbbII operon.The upstream/downstream sequence of the remaining 26 reads was shorter than 6 nt and these were not assessed.

R. capsulatus genes knockouts
Knockouts in R. capsulatus were created by gene transfer agent mediated gene replacement. 34pCM66T plasmid constructs were created with a gentamicin resistance cassette flanked by 500-1000 bp of DNA from either side of the target gene (Table S5).Assembly was achieved by a one-step, four-component NEBuilder reaction and transformation into NEB 10-beta cells (New England Biolabs).All oligonucleotides were obtained from IDT and designed with an optimal annealing temperature of 60 C when used with Q5 DNA Polymerase (New England Biolabs).All cloning reactions were carried out with NEBuilder according to the manufacturer's guidelines (New England Biolabs).
Deletion constructs were introduced into the E. coli conjugation strain S17-1 and then transferred to the R. capsulatus gene transfer agent hyperproducer strain DE442 by solid phase conjugal mating.One millilitre aliquots of overnight cultures of the E. coli S17-1 donor and Rhodobacter recipient strains were centrifuged at 5000 RCF for 1 min, washed with 1 ml YPS medium, centrifuged again and resuspended in 200 ml YPS.Ten microlitres of concentrated donor and recipient cells were mixed and spotted onto YPS agar or spotted individually as negative controls.Plates were incubated overnight at 30 C. Spots were scraped, suspended in 100 ml YPS broth and plated on YPS + 100 mg ml -1 rifampicin (counter-selection against E. coli) + 10 mg ml -1 kanamycin (plasmid selection).Plates were incubated overnight at 30 C then restreaked onto fresh agar to obtain single colonies.
A standard gene transfer agent bio-assay 74 was carried out to replace the intact chromosomal gene with the deleted version using 10 mg.ml -1 gentamicin for selection.Successful knockouts were confirmed by Sanger sequencing (Eurofins Genomics).

Bioinformatics and data visualization
ORFs were identified using ab initio prediction of Prokka 1.14.5 56 and GeneMarkS 4.28 57 with the ''phage'' algorithm.The annotation and circular maps were done in Artemis. 58The predicted function of protein was based on primary sequence similarity using blastp, 59 the comparison of profile hidden Markov models using HHpred 41 and predicted tertiary structure similarity using a combination of AlphaFold2 60,61 and DALI search. 62The terminators were predicted using the Genome2D web tool [http://genome2d.molgenrug.nl/g2d_pepper_transterm.php].
Psi-blast search against bacterial and viral taxids using matrix BLOSUM=45 was performed using the less conserved major capsid protein P2 sequence until convergence. 59,63Then, a list of nucleotide seqIDs encoding proteins found in all iterations was determined using NCBI batch Entrez search against identical protein groups database [https://www.ncbi.nlm.nih.gov/sites/batchentrez]. Unique Gis encoded by the longest nucleotide sequence were filtered and a list of proteins encoded within the sequences was generated using NCBI e-utilities. 63his list served as a subject for psi-blast with the packaging ATPase P9 until convergence using matrix BLOSUM=45.The position in the nucleotide sequence of these new hits was retrieved using a batch Entrez search and compared to the original major capsid protein P2 hits.Hits originating from the same nucleotide sequence with less than 10000 nt from each other were filtered using awk script and considered putative prophage sequences.The taxonomy information of the hits was retrieved using the NCBI tax identifier [https://www.ncbi.nlm.nih.gov/Taxonomy/TaxIdentifier/tax_identifier.cgi].The putative CtrA-binding site and dif motifs were identified using consensus sequences and regular expression search tool [https://regex101.com/].

Phylogeny analysis
Phylogeny analysis was performed according to Grybchuk et al. 75 In brief, a list of homologous proteins to those from Jorvik was generated by PSI-BLAST.The search was performed using the matrix BLOSUM=45 and the default settings until converging.The taxonomy database used for the search was Tectiliviricetes in the case of the major capsid protein P2, Alpha-proteobacteria (filtered out after the convergence) and Tectiliviricetes in the case of the packaging ATPase P9.The complete sequences were then aligned using MAFFT v7.487 e-ins-I algorithm. 64he uninformative columns were cropped using trimAl v1.4 algorithm automated1 65 and the phylogenetic tree was computed by maximum likelihood method using software iqtree2 v2.0.6 66 with a bootstrap test.A similar approach was performed using P9 and P2 sequences from Data S6.1 as an input, with the final tree computed using fast bootstrap test.The trees were visualized using FigTree v.1.4.4.[http://tree.bio.ed.ac.uk/software/figtree/].
For the Genome-BLAST Distance Phylogeny, genomic sequences classified to Autolykiviridae, Corticoviridae were retrieved from GenBank and together with the sequence of Jorvik used as a nucleotide input to VICTOR web server [https://ggdc.dsmz.de/victor.php]. 40he phylogeny was performed on a predicted amino acid sequence under settings recommended for prokaryotic viruses, using a recommended greed-with-trimming algorithm of formula D6.

Proteomic analysis
The phage sample for proteomic analysis was prepared in the same way as for the genome isolation.After pelleting and resuspending in B buffer, extra polishing was performed.Here, most remaining bacteria were separated by 6 000 RCF/4 min/8 C spin.Phage particle aggregates were enriched from the supernatant by an additional 17 000 RCF/25 min/8 C spin, resuspended in B buffer to final A 280nm $ 2 and used for LC-MS/MS analysis performed by The Centre of Excellence in Mass Spectrometry, University of York.Here, the phage-enriched sample was run 1 cm into a NuPAGE 10% BisTris gel (Invitrogen) and stained with SafeBLUE stain (NBS Biologicals).The stained gel segment was excised, destained, reduced with DTE and alkylated with iodoacetamide.Protein was in-gel digested using Promega sequencing grade trypsin and incubation overnight at 37 C.The resulting peptides were analysed by LC-MS/MS over a 1 h acquisition with elution from a 50 cm EN C18 PepMap column (Thermo Fisher Scientific) driven by an mClass UPLC (Waters) onto an Orbitrap Fusion Tribrid mass spectrometer (Thermo Fisher Scientific) operated in DDA mode.MS1 spectra were acquired at high resolution in the Orbitrap mass analyser.MS2 spectra were acquired in TopSpeed mode in the linear ion trap after HCD fragmentation.
LC-MS data in .rawformat were imported into Progenesis QI for peak picking with tandem mass lists exported in .mgfformat for database searching using Mascot.Data were searched against the combined Jorvik and NCBI R. capsulatus YW1 databases appended with common proteomic contaminants.Peptide matches were filtered to 1% FDR using the Percolator algorithm and as determined empirically by comparison to a reverse database search.Accepted peptide identifications were imported into QI and associated with the LC-MS chromatographic data.Results were further filtered to require a minimum of two unique peptides per protein identification.As a metric for relative abundance between proteins the Top3 approach has been used, whereby the best three responding peptide precursor ion areas for each protein are extracted and compared.

Electron microscopy
For cryo-EM grids, a confluently lysed plate of phage was washed with 2 ml of RCV using a pipette and the media was spun at 4000 RCF for 6 min.The pellet, containing aggregates of virus particles as well as particles attached on host membranes, was resuspended in the residual amount of liquid, mixed with 6 nm colloid gold fiducials (Auriol), applied onto glow-discharged 200 mesh copper Quantifoilâ grids and plunge frozen in liquid ethane.The grids were imaged using Glacios TEM (Thermo Fisher Scientific) operated at 200 kV at the cryo-EM facility of the University of York.

FFigure 1 .
Figure 1.Morphological and genomic characteristics of phage Jorvik (A and B) Morphology of Jorvik1 (A) and Jorvik2 (B) plaques on the lawn of R. capsulatus, strain YW1.The scale bar is 2 mm.(C) An aggregate of the phage obtained from a lysed plate wash, imaged by cryo-EM.The black dots are gold fiducials, scale bar represents 200 nm.(D and E) Native capsids of the phage imaged using cryo-EM (D) and cryo-ET reconstruction (E) with head spikes depicted by blue arrows, scale bar is 50 nm.(F) Genome map of the phage variant Jorvik1.The duplicated region of Jorvik2 adjacent to gp1 is highlighted in black and the single nucleotide change in gp1 is depicted by black dot.Genes encoding structural proteins (medium blue), acetyltransferase (light blue), hydrolases (light green), unknown function (yellow), holin (dark green), DNA-binding protein (magenta), replicase (orange), ATPase (dark blue) and P15-like repressor (dark magenta) are depicted.Genes shown in bold have sequence homologues in the database, white dots within ORF arrows depict predicted transmembrane regions.Putative regulatory repeat (gray), terminators (loop symbol) and attP site (X symbol) are highlighted.Operons are depicted in shades of orange in the center of the map.See also Figures S1-S3 and Data S2, S4, and S7.

Figure 2 .
Figure 2. Growth characteristics of phage Jorvik2 (A) Adsorption curve on the propagating strain YW1 (red) and phage-resistant strain P12F (black).On the y axis, values of plaque-forming units (PFUs) for specific time points relative to PFUs at time point 0 are shown.(B) Growth curve of the phage on the propagating strain YW1 (red) and sterile media control (black).On the y axis, PFUs for specific time points relative to PFUs at time point 0 are shown.(C-E) Growth curves of strains YW1 (C), SB1003 (D) and St. Louis (E) with (red) and without the phage (black) are shown.The multiplicity of infection was below 0.01.The raw data are present in Data S1.3-S1.5.

Figure 3 .
Figure 3. Phage survivability under different conditions Individual measurements of biological replicates are plotted as points, their averages are plotted as Bars.(A) Heating of the phage lysate for 30 min.(B) Freeze-thawing of the phage lysate.(C) Incubation of the phage in the medium with different pH.(D) Incubation of the phage in the medium with different NaCl concentrations.The raw data are present in Data S1.1 and S1.6-S1.8.

Figure 4 .
Figure 4. Peptidoglycan hydrolyzing activity of phage-encoded enzymes (A) SDS-PAGE gel of the lytic protein samples -lysozyme positive control (Lys), putative peptidase M15 (M15), putative soluble lytic transglycosylase (Slt) and negative controls protease 3C (3C) and bovine serum albumin (BSA).(B) Zymogram with R. capsulatus SB1003 cell walls embedded in the gel matrix.The raw gels are present in Data S3.2.(C) Plate lysis assay of protein samples dropped onto a layer of water agar with embedded cell walls.

Figure 5 .
Figure 5. Phylogenetic tree of phage Jorvik generated by Genome-BLAST Distance Phylogeny of the predicted amino acid sequence using the web server VICTOR The pseudo-bootstrap support values from 100 replications are shown next to nodes, with an average support equal to 90%.The scale bar defines the branch length, scaled in terms of the D6 distance formula.The legend on the right defines different taxonomy rankings of the phages with the color coding differentiating between different taxon groups that were assigned within a rank.The color of the name depicts approved or officially proposed family ranks -Autolykiviridae (teal), Jorvik-like (magenta), Corticoviridae (green), fNo16-like (orange) and Tectiviridae (blue).See also TableS4and Data S5 and S6.

Figure 6 .
Figure 6.Genome synteny of membrane-containing dsDNA bacteriophages Names of the phages with GenBank accession numbers are shown on the left.The color bar on the right shows HHpred score for the pairwise alignment between corresponding homologues of Jorvik and PM2 phage, with raw values present in TableS2.The scale bar at the bottom shows the length of the genome in base pairs.Phages are grouped according to the replicase gene type.Genes are colored according to their putative function as explained in the legend.

Table 1 .
Interaction between phage Jorvik and the tested host strain for a complete list of