A conserved genetic module that encodes the major virion components in both the coliphage T4 and the marine cyanophage S-PM2

Abstract

Sequence analysis of a 10-kb region of the genome of the marine cyanomyovirus S-PM2 reveals a homology to coliphage T4 that extends as a contiguous block from gene (g)18 to g23. The order of the S-PM2 genes in this region is similar to that of T4, but there are insertions and deletions of small ORFs of unknown function. In T4, g18 codes for the tail sheath, g19, the tail tube, g20, the head portal protein, g21, the prohead core protein, g22, a scaffolding protein, and g23, the major capsid protein. Thus, the entire module that determines the structural components of the phage head and contractile tail is conserved between T4 and this cyanophage. The significant differences in the morphology of these phages must reflect the considerable divergence of the amino acid sequence of their homologous virion proteins, which uniformly exceeds 50%. We suggest that their enormous diversity in the sea could be a result of genetic shuffling between disparate phages mediated by such commonly shared modules. These conserved sequences could facilitate genetic exchange by providing partially homologous substrates for recombination between otherwise divergent phage genomes. Such a mechanism would thus expand the pool of phage genes accessible by recombination to all those phages that share common modules.

In the marine environment, prokaryotes are estimated to account for between 70% (1) and 90% (2) of organic matter. Marine autotrophic prokaryotes, which require only simple inorganic substances to fulfill their nutritional requirements, are responsible for much of the inorganic carbon fixation in the sea. Among these organisms are the unicellular cyanobacteria. It has been estimated that cyanobacteria are so numerous that they are responsible for half of the oxygen produced on earth (2). In certain nutrient-poor marine habitats, cyanobacteria of the genera Synechococcus and Prochlorococcus are believed to be responsible for the major part of this oxygen production (3, 4).

Since an initial report in 1989 (5), convincing evidence has emerged for the remarkable abundance of viruses in the marine environment. Recent reviews have focused on the role of these viruses in the biogeochemistry and nutrient cycling in the sea (6, 7), as well as their ecology and diversity (8). Cyanophages are the viruses that infect cyanobacteria and, in the sea, they are typically an order of magnitude more abundant than their hosts (9–11). Suttle and Chan reported cyanophage titers in the Gulf of Mexico as high as 10⁸ per liter (10). These high titers imply that 5–10% of cyanobacteria in surface waters could be destroyed daily by phage infection (12, 13). As a result, the amount of fixed organic carbon that passes up the food chain is reduced because cyanophage-mediated lysis diverts host nutrients sideways by making them available for uptake by heterotrophic bacteria (14–16). Thus, cyanophages, through their impact on host abundance, have the potential to play a key role in the productivity and biogeochemistry of the oceans. Furthermore, there is evidence (9, 17) that cyanophages may affect the genetic diversity of the host population, and molecular studies have shown that the clonal diversity of bacterio- and phytoplankton is an important determinant of the virus community (18). This idea is supported by modeling of bacteriophage–host interactions, which predicts a reciprocal relationship between bacterial diversity and virus abundance (19). However, little is known still about these phages or their impact on marine ecosystems.

A large number of cyanophages have been isolated on strains of the marine phycoerythrin-containing cyanobacterium Synechococcus. The majority of these cyanophages appear to be myoviruses with contractile tail structures similar to the well-known coliphage T4 (9, 11, 20). The ubiquity of myoviruses in the marine environment is probably a consequence of a virion design that can be easily adapted to different conditions (21). The myoviruses are physically robust and remarkably versatile, and only minor modifications are required to infect bacteria whose habitats range from the animal gut to the open ocean (21, 22). S-PM2 is a cyanomyovirus isolated from the English Channel and infecting Synechococcus sp. WH7803 (20). It has previously been observed (23) that the S-PM2 genome contains a sequence with a homology to the T4 gene that encodes the virion head portal protein (g20). Here we report further molecular studies of S-PM2 that clarify its phylogenetic relation to the large and diverse group of T4-type phages.

Materials and Methods

Growth and Maintenance of Bacterial and Bacteriophage Strains.

The cyanobacterium Synechococcus sp. WH7803 and cyanophage strains, S-PM2 (23) and S-PWM3 (11), used in this study were propagated as previously described (20). The T4 phage was from the Toulouse collection. This isolate was originally obtained from R. H. Epstein of the University of Geneva in Switzerland. Phage T4 stocks were prepared on Escherichia coli B^E by using standard techniques for phage propagation (24).

Transmission Electron Microscopy (TEM) of Phage Particles.

Cyanophage morphology was determined by TEM. Phage particles were negatively stained before microscopy. Fresh phage lysate was adsorbed onto a carbon-stabilized formvar support on a 200-mesh copper grid and stained with 2% uranyl acetate. The grids were examined with a JEOL 1200 EX microscope at 100 kV accelerating voltage and photographed at a magnification of up to ×120,000. Photographic images were subsequently scanned (binuscan IPM, binuscan, Monaco) and digitally measured (nih image). The calibration of the magnification was confirmed with the use of the T4 virion as a standard.

Pulse-Field Gel Electrophoresis (PFGE) of Phage Genomes.

Cyanophage genome sizes were determined by PFGE. A suspension of cyanophage containing 10⁵-10⁶ plaque-forming units was added to an equal volume of 2% agarose in 0.5× TBE buffer (90 mM Tris/90 mM boric acid/0.45 mM EDTA, pH 8.0) and allowed to set. The blocks were incubated overnight at 55°C in a lysis buffer containing 100 mM EDTA, 100 mM Tris⋅HCl, pH 9, 1% SDS, and 0.5 mg⋅ml⁻¹ proteinase K. These blocks were then dialyzed three times for 1 h against 10 ml of TE medium (10 mM Tris·HCl/1 mM EDTA, pH 8.0). One-third of the block was analyzed by electrophoresis for 14 h in 1% agarose (0.5× TBE) at 275 V with a pulse time of 8 sec on the Pulsaphor Plus system (Amersham Pharmacia & Upjohn).

Cyanophage DNA Extraction.

A clonal cyanophage suspension was added to exponentially growing host cells at a multiplicity of infection of ≈4 × 10⁻³ and incubated until the culture had cleared (about 1 week). Host cell debris was cleared by centrifugation at 4,000 × g for 10 min. The supernatant was incubated with lysis buffer (1% SDS and 0.05 M EDTA, pH 8.0) for 15 min. DNA was isolated by one round of phenol extraction and three rounds of phenol/chloroform, 1:1, followed by precipitation through addition of 10% (vol/vol) 3.3 M NaOAc, pH 5.2, and 75% (vol/vol) isopropyl alcohol. After centrifugation at 16,000 × g for 30 min, the resulting pellet was resuspended in TE medium.

DNA Sequencing.

The oligonucleotides used to PCR amplify a ≈550-bp segment of gene (g)23 of S-PM2 were oMZIA1Bis (5′- GATATTTGIGGIGTTCAGCCIATGA-3′) and oMZIA6 (5′-CGCGGTTGATTTCCAGCATGATTTC-3′). Their sequences are based on the consensus of the g23 sequences in a number of T4 type phages (21, 24, 25). The oSPM23.2 (5′-AAGACCATGAATTGCCTTGAGGTC-3′) has a g23 sequence that was obtained from sequencing this PCR fragment. SPM23.2 was used in conjunction with an S-PM2 g20 oligo (SPM20.2, 5′-ATACGAACTCATTCGTAGATATCGT-3′) to amplify a ≈4.3-kb fragment of the phage genome that was completely sequenced by primer walking. To extend this sequence into g18, a standard PCR amplification was performed by using primers in the sequence of S-PM2 g20 (SPM20.1, 5′-TTGTGGTAAGTGATCTGCCATCA-3′) and in the g18 sequence of the pseudo T-even phage RB49 (FT18N2, 5′-GTAAATTCCAATGGGGTCCAGCTT-3′). These two primers amplified a ≈1.5-kb fragment of the S-PM2 genome. Analysis of this band by sequencing from FT18N2 indicated that it encoded a protein homologous to T4 gene product (gp)18. On the basis of this initial sequence, an oligo, SPM18.2 (5′-TCTCGGAGGTTCAGGCACGCA-3′), was designed and then used with SPM20.1 to cleanly amplify a slightly smaller fragment that was sequenced by primer walking. An Expand Long Template PCR System (Roche Diagnostics) with the same primers (FT18N2 and SPM20.1) also gave two significantly larger bands. One of these, a ≈4.5-kb fragment, was purified and then sequenced by primer walking in each direction from the g18 sequence.

The S-PM2 genome analysis was extended in the direction upstream of the g23 sequence by a PCR protocol that took advantage of the known S-PM2 sequence to provide a nested series of S-PM2-specific primers that were used in combination with a random primer. Initially, an S-PM2 primer internal to the g23 sequence was used with the random primer in a PCR reaction that involved a large number of cycles and thus generated a number of bands. The products of this reaction were used as template for a second PCR reaction with the same random primer but a more distal g23 primer. This reaction then gave a ≈1.6-kb band that had a sequence contiguous to g23.

The oligonucleotides used to PCR amplify a ≈1,300-bp segment of S-PWM3 across g22 and g23 were g22F1 (5′-TGGCTCRMWGAAAAYMAABTWGC-3′) and g23R2 (5′-TSRAKKTCRCGGTTVAKYTCMA-3′). Their sequences are based on a consensus of the cyanophage S-PM2, the vibriophages KVP20 and KVP40 and T4. The resulting PCR product was cloned (TA Cloning Kit, Invitrogen) and sequenced.

Computer Analysis of DNA and Protein Sequences.

ORFs in the S-PM2 nucleotide sequence were identified by using the ORF Finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html), and assessment of their likelihood of being genuine coding sequences was carried out by using genemark.hmm and genemark (26) in combination with National Center for Biotechnology Information blast similarity searches. Base composition analysis was carried out by using the commercial software package lasergene (DNAstar, Madison, WI). The segment pair overlap component of macaw (Multiple Alignment Construction and Analysis Workbench) (27) was used to detect homologous domains between proteins and to schematically display such features. Multiple sequence alignment was carried out by using clustalx (28) and phylogenetic analyses used treepuzzle 5.0 (http://www.tree-puzzle.de/).

Results and Discussion

The TEM image of cyanophage strain S-PM2 (Fig. 1) shows that it has a long contractile tail and an icosahedral head. The head is joined to the contractile tail by a tapered collar structure. The distal end of the contractile tail is terminated by a stubby baseplate, from which long tail fibers radiate. This cyanophage has only limited morphological resemblance to T4, the archetype of the large and diverse family of phages with contractile tails. The head of S-PM2 is an isometric icosahedral of ≈85 nm, and its tail is ≈180 nm long. Thus, the virion morphology is clearly distinguishable from T4, which has a prolate icosahedral head (78 by 111 nm) and a significantly shorter tail (113 nm). The S-PM2 genome size was estimated by pulse-field gel electrophoresis to be 194 kb (Fig. 2), 15% larger than the 169 kb of T4 genome. The relative size of S-PM2 and T4 genomes is consistent with what would be predicted, assuming the genome DNA entirely fills the volume available in the head of each phage, but that the density of the packing of the DNA it slightly higher in SPM-2 because of the absence of T4-like nucleotide modifications and glucosylation and/or because of a reduction in the content of internal proteins within the head (21).

Figure 1.

TEM of cyanophage strain S-PM2 stained with uranyl acetate. (Bar = 100 nm.)

Figure 2.

Pulse-field gel electrophoresis analysis of different cyanophage genomes compared with coliphage T4 (lane A). Cyanophage strains S-BP3 (lane B); S-WHM1 (lane C); S-PM2 (lane D) that have estimated genome sizes of 45, 170, and 194 kb, respectively. Size markers are labeled in kilobases (lane M). S-WHM1 is morphologically similar to S-PM2, and its genome contains a T4 g20 homologue (20). S-PB3 is a cyanopodovirus (20).

The hosts of S-PM2 and T4, Synechococcus spp. and E. coli, are extremely distant phylogenetically and occupy different ecological niches. Synechococcus spp. are abundant in the world's oceans, whereas E. coli is a ubiquitous inhabitant of the mammalian intestine. Despite their phylogenetic and ecological separation, previous work (23) has shown that the marine cyanomyoviruses S-PM2, as well as S-WHM1 and S-BnM1, all contain a sequence homologous to T4 g20, which codes for the head portal protein, a minor but important constituent of the virion capsid (29). Such a result argues that all of these phages must have had access to a common g20-like sequence at some point in their evolutionary history. It is widely believed that double-stranded DNA tailed phages were created by the sequential assembly of genetic modules, each encoding a group of related functions (30–32). Thus, it was of interest to determine whether the sequence homologous to T4 g20 was part of a larger T4-like module in the S-PM2 genome and, if so, to characterize this module. The first step in this effort was to determine whether the S-PM2 genome also contained a homologue of g23, the gene encoding the T4 major capsid protein (MCP), that in conjunction with gp20 and the prohead core proteins (gp21 and gp22) initiates T4 capsid assembly. The sequence of the MCP of a large number of phages with T4-like morphology has recently been established (21). When these diverse g23 homologues were aligned, two highly conserved regions of MCP sequence became evident. These conserved motifs in the g23 sequence were used to design degenerate PCR primers that could amplify the central third of the g23 in all of the currently characterized T4-like phages. These g23 PCR primers (oMZIA1Bis and oMZIA6) were used on a S-PM2 DNA template and, after the standard conditions of the PCR reaction were slightly modified, these oligonucleotides amplified a product of approximately the size expected for a g23 homologue (data not shown). This product was sequenced and revealed to be very distantly related to T4 g23. The presence in S-PM2 of sequences homologous to both T4 g20 and g23 gave strong support to the notion that this phage probably carried an entire T4-like head assembly module. To test this hypothesis, the nucleotide sequence of the region of the S-PM2 genome in the vicinity of these T4-like genes was determined.

The point of departure for this sequencing effort (Fig. 3) was the small central portion of the S-PM2 g23 sequence that we obtained by using the conserved g23 PCR primers. After this S-PM2 g23 sequence was determined, it was used as the basis for the synthesis of additional PCR primers. These S-PM2 g23 primers were used in conjunction with an appropriately oriented S-PM2 g20 primer to amplify the genome segment located between them. A PCR fragment of about 3 kb was obtained and was then sequenced by a primer walking procedure (33). Analysis of this sequence by the National Center for Biotechnology Information blast program revealed that it contained homologues of T4 g21 and g22 in addition to the expected g20 and g23 sequences. The order of these virion structural components was the same as in phage T4. This sequence was then extended beyond g20. We had established the sequence of the g18 homologues in four diverse T4-like phages (21). A variety of oligonucleotides PCR primers were designed on the basis of the common g18 sequences among these phages. These primers were then used in conjunction with appropriately oriented primers located in the N-terminal segment of S-PM2 g20. This approach allowed us to amplify and sequence the segment of the S-PM2 genome between g18 and g20 (see Materials and Methods).

Figure 3.

Genetic organization of the head and tail assembly region of bacteriophages S-PM2 (10,095 bp) and T4. The percentage identities (and similarities) at the amino acid sequence level, as determined by blast, between putative S-PM2 genes and their T4 homologues are shown between the S-PM2 and T4 maps. The PCR primer pairs involved in the S-PM2 sequencing, and the approximate size and position of the PCR products are indicated.

Analysis of this 10-kb S-PM2 sequence revealed 10 ORFs that are likely to be coding sequences (Fig. 3). The sequence actually appears to be composed of two distinct regions. At the 5′ end of the segment, the first three ORFs (ORFs 1, 2, and 3) are encoded on the opposite strand from the next seven ORFs. The polypeptides encoded by six of the next seven ORFs have significant sequence similarity to the virion proteins of the T4-type myoviruses. The extent of identity (and similarity) between the T4 and S-PM2 genes is indicated in Fig. 3. The first four of these ORFs correspond to T4 g18 (encoding a contractile protein of the tail sheath); g19 (encoding a tail tube protein), g20 (encoding a portal vertex protein), and g21 (encoding a prohead protease) are contiguous and are in the same order as in T4 (Fig. 3). In T4, g20 and g21 are separated by two small genes, g67 and g68, but both of these genes are absent from the S-PM2 sequence. However, in S-PM2, g21 and g22 (encoding the prohead core) are separated by ORF 4, a small polypeptide with no significant similarity to any proteins in the database. The g22 sequence is followed by the homologue of T4 g23 that encodes the major capsid protein of the T4 virion.

A further comparison of the T4 and the S-PM2 genomes was made by aligning the nucleotide sequence encoding the T4 g17 and g18 region with the sequence encoding S-PM2 ORF 3 and g18. The two phage sequences are completely divergent until well within in the T4 g18 sequence. The N-terminal 97 amino acids of the S-PM2 gp18 have no homology with any of the known T4-type g18 sequences. This result supports a suggestion by Tétart et al. (21) that g18 may have a modular organization similar to that of the tail fiber genes. This possibility is made more plausible by the analysis in Fig. 4 that uses the National Center for Biotechnology Information macaw program (27) to align the gp18-like sequences of a wide variety of phages with contractile tails. This analysis revealed a significant sequence similarity in the C-terminal half of the T4-type gp18 sequences with the tail sheath sequences in an unrelated group of phages (186, P2, PS17). Furthermore, the same conserved sequence is also found in Er, a phage-tail-like bacteriocin of Erwinia carotovora (34). This segment of all of these protein sequences seems to share a common, if very distant, origin. Even considering just the T4-type phages (KVP40, RB49, 42, nt-1, T4, and S-PM2), the N-terminal domain of the S-PM2 protein clearly differs from all of the others. These comparisons of the gp18-like sequences strongly support the view that they are composed of separate N- and C-terminal elements, and that recombination can shuffle the various possible versions of these modules during evolution.

Figure 4.

Modular organization of tail tube proteins (gp18) from a variety of bacteriophages and phage-related sequences. Boxed regions of the same size and shading are segments of similar sequence as detected by macaw by using the segment pair overlap method. The degree of similarity is indicated by the intensity of shading, with darker shading indicating regions of greater similarity. The sequences are as follows, with accession numbers in parentheses: 186, enterobacteria bacteriophage 186 (NP_052269); P2, E. coli bacteriophage P2 (NP_046778); PS 17, tail sheath protein from Pseudomonas aeruginosa bacteriophage PS17 (JC4865), Xfast, Xylella fastidiosa phage-related contractile tail sheath protein (C82769); Pecto, Pectobacterium carotovorum carotovoricin Er with similarity to tail sheath protein (BAA76521); KVP40, tail sheath protein from Vibrio parahaemolyticus bacteriophage KVP40 (BAA77375), nt-1, tail sheath protein from Vibrio natriegens bacteriophage nt-1 (AAG02027); T4, E. coli bacteriophage T4 tail sheath protein (225888), bp42, tail sheath protein from Burkholderia cepacia bacteriophage 42 (AAG02028); RB49, E. coli bacteriophage RB49 tail sheath protein (CAB96858); S-PM2 putative tail sheath protein from cyanomyovirus S-PM2 (this study).

Perhaps the most interesting of the T4-type proteins with sequence similarity to S-PM2 is the major capsid protein (gp23), because its sequence has been determined in such a large number of phages. A previous phylogenetic analysis of the major head and tail of a wide range of T4-type bacteriophages has supported their distribution into three subgroups (21). There is a cluster of phages termed the T-evens that are all closely related to T4, the type member of the Myoviridae. The second subgroup, which includes phages such as RB49 and -42, have previously been termed the pseudoT-evens (24). The third cluster includes Aeromonas phages and vibriophages such as nt-1, KVP20, and KVP40. These phages all have heads that are more elongated than those of T-evens and the pseudoT-evens and thus are called the schizoT-evens (21). A comparison was made, by using macaw, of the 16 currently publicly available gp23 sequences, together with the S-PM2 gp23 and a partial gp23 sequence from another T4-like marine cyanomyovirus S-PWM3, isolated from the Gulf of Mexico (11). This analysis revealed that there are only two sequence features of 17 and 59 amino acids, which are highly conserved in all 18 major capsid proteins. These conserved features correspond to residues 115–131 and 241–299 in the T4 major capsid protein. These two sequence features, which lacked gaps, were concatenated to give a 76-residue sequence, which was used for phylogenetic analysis by using a maximum likelihood approach. The highly conserved nature of the sequence means that it is unable to resolve closely related viruses and consequently does not completely recreate the groups defined by Tétart et al. (21), but it is useful in establishing the evolutionary distance of S-PM2 and S-PWM3 from all of the phages already characterized. An unrooted phenogram (Fig. 5) derived from these concatenated gp23 sequences reveals that S-PM2 and S-PWM3 are quite distinct from the other T4-like phages so far characterized and form a discrete group, which we suggest be called the exoT-even group. These are phages of the T4-type that have diverged the most from T4. Beyond the fact that they have a contractile tail, these phages have little morphological resemblance to the T4-type phages.

Figure 5.

An unrooted phenogram based on a 76-aa concatenated sequence from the T4-related bacteriophage major capsid proteins derived by maximum likelihood by using treepuzzle 5.0 with default settings. The numerical values of the quartet puzzling tree search estimates for each internal branch are given and may be interpreted in much the same way as bootstrap values. The sequences are as follows, with accession numbers in parentheses: T4 (P04535); 1 (AAF61700); 42 (AAF61692); RB43 (AAF61698); 44RR2.8t (AAF61693); RB49 (CAB01539); Aeh1, (AAF61695); 65, (AAF61694); nt-1 (AAF61697); KVP20 (BAA25880); KVP40 (BAA25567); S-PM2 and S-PWM3 (this study). Cyanophages S-PM2 and S-PWM3 are representative of what we suggest be called the exoT-even group.

In terms of its overall sequence, the gp23 of S-PM2 is most similar to those of the marine vibriophages nt-1, KVP20, and KVP40, which is consistent with the idea that in the marine environment, even very distantly related phages have had access to a common gene pool. Recombination between different phages can occur most easily if they both can infect the same cell. This coinfection may occur if their hosts are the same or phylogenetically relatively close to each other. Because members of the genus Vibrio and E. coli are both members of subgroup 3 of the γ-purple bacteria, these hosts are not actually so phylogenetically distant. However, the hosts of the cyanophages, the cyanobacteria, belong to a division of oxygenic phototrophs, a much more distant group. Nevertheless, it is possible that the diverse T4-like phages could infect some common intermediate bacterial hosts. In this regard, it is interesting that some other bacteriophages (35) have been shown to have a remarkably broad host range. It is also possible that these T4-type phages could exchange genetic modules indirectly by means of one or more intermediates among the T4-like phage that serve as a genetic link between them.

It has been suggested that phage genomes are really pastiches composed of combinations of genetic modules that have previously been used successfully in other contexts (31). The complete genomic sequence of the marine phage Roseophage S101 has revealed that components of its DNA replication apparatus have homologues in the coliphages T3 and T7 (36). This evidence supports the concept of the mosaic nature of phage genomes and indicates links between marine and nonmarine phages. The enormous advantage of a genome organization that facilitates the incorporation and utilization of exogeneous functional genetic elements, rather than relying on their evolution de novo, is self-evident. The work presented here and in other recent publications (21, 37–39) strongly supports the suggestion that at least some phage genes may themselves be patchworks or mosaics, composed of sequence motifs originating from diverse sources. A patchwork design for phage genes may be a much more common feature than had been previously imagined. A patchwork structure has been best characterized in the phage tail fiber genes, where this organization is believed to facilitate the swapping of the adhesin domains between distantly related phage and thus mediate rapid changes in the host range (40). Our current results indicate that the more general advantages of such a gene design merit being investigated further.

In conclusion, although the T4-type phage provide an excellent system to study the fundamental aspects of viral evolution, the ultimate interest may lie beyond that. Phages probably play an important role in the evolution of their hosts (41), because abortive infection provides an abundant source of genetic diversity that can be easily incorporated into the bacterial genome. Because the transfecting DNA can come from a previous host or from the phage genome itself, the host could eventually subvert many successful viral innovations. The large population size and the promiscuous recombination within a phage family create a situation where enormous diversity can be generated, tested, and exported at low biological cost. The two characteristics of phages that are most likely responsible for their rapid evolution, high recombination and broad host range, may be the same properties that make them efficient vectors to promote their host's evolution. Because biological diversity can be understood only in terms of the evolutionary processes that generate it, the role of extended virus families, such as the T4-type phages, as a potential major source of the biosphere's evolutionary innovation will need to be assessed.

Abbreviations

TEM

transmission electron microscopy

g

gene

gp

gene product