Using Signature Genes as Tools To Assess Environmental Viral Ecology and Diversity

T4-like g20 portal protein in the family Myoviridae.

The application of the myovirus T4 portal protein gene, g20, as a signature gene originated from the identification of a conserved sequence in marine cyanophages that corresponded with g20, which led to the development of primer set CPS1-CPS2 (cyanophage specific) (Table 1) (9). Using additional CPS primer sets combined with denaturing gradient gel electrophoresis (DGGE) on an Atlantic Ocean north-south transect and depth profile, it was found that similar cyanophages were common along the transect but showed much greater diversity along the depth profile, reflecting nutrient and host bacterial abundance levels (10, 11). The same observations were made for cyanophage communities at the coast of British Columbia, Canada (12). A larger-scale Atlantic north-south transect revealed widely distributed similar cyanomyovirus signatures, but the variation present could not be linked to environmental factors or the host bacteria (13). g20 clone libraries of estuarine and open ocean environments revealed a high diversity in these two environments as well as high novelty in sequences (14).

Using g20 PCR restriction fragment length polymorphism (RFLP) and sequence analysis, temporal changes in marine cyanophage community composition and abundance were found in coastal waters near Rhode Island, and several RIM (Rhode Island myovirus) clusters were defined (15). Several other studies found well-established seasonal changes in marine cyanophage diversity, linked to host cell density, but low spatial variation (16–18).

The g20 marker gene has also been successfully used for analysis of freshwater environments. In the mesotrophic Lake Bourget (France), a high diversity of cyanophages was identified, with significant similarity to clades of marine cyanophages (19). In accordance with results from the marine studies, phage abundance was lowest in winter months. Water samples from Lake Erie (Canada and United States) and a ship's ballast tank showed novel sequences and sequences related to marine cyanophages, suggesting a possible marine origin for the freshwater cyanophages or, alternatively, that they infect a different host in the lake environment (20). A quantitative PCR approach to assess the abundance of cyanomyoviruses in this lake yielded 1.3 × 10⁵ to 4.3 × 10⁶ g20 copies per ml of water over the stations, depths, and seasons, showing a significantly higher viral density in summer (21). Analysis of floodwater and soil of rice paddy fields in Japan revealed that these communities differed significantly from each other, with the soil sequences grouping into more distinct clades whereas the floodwater community contained more sequences similar to marine, freshwater, and isolated cyanophages (22, 23).

In a study that spanned a wide range of environments, including marine (from tropical to polar and over 3,000 m in depth) and freshwater environments (lakes and ponds), highly similar sequences were found in highly variable environments (24).

When investigating the prevalence of the g20 gene in cyanophages with a confirmed myovirus morphology, it was found that only two-thirds of the population carried the gene, and it was concluded that the full diversity of cyanomyoviruses cannot be explored using this marker (25). In addition, the primer sets described for this marker (Table 1) are specifically targeted to cyanophages, making them relevant only to habitats where cyanobacteria make up the majority of the bacterial population.

Major capsid protein gene (g23 of T4) in the family Myoviridae.

The major capsid gene of phage T4, g23, was one of the conserved virion genes used to distinguish between T-even-, Schizo-T-even-, Pseudo-T-even-, and Exo-T-even-type phages (26). Based on this phylogenetic analysis, a new degenerate primer set for g23 (MZIA1bis and MZI6) was developed as an alternative for the g20 signature gene described above (Table 1) (27). In a wide range of marine habitats, this primer set could identify five new groups of previously unidentified T4-type phages, revealing that some of the groups were present from the Gulf of Mexico to the Arctic Ocean while others were more geographically distinct. Combining all isolated T4-type phages and known environmental g23 sequences up to that time with data from the Global Ocean Survey (GOS), 1,399 sequences were aligned, leading to the creation of the Far-T4 phylogenetic group, comprised almost exclusively of uncultured sequences, and the Cyano-T4 group, which includes all the cultured cyanomyoviruses and half of the environmental sequences described by Filée and colleagues in 2005 (28).

Recently, g23 TRFLP (terminal restriction fragment length polymorphism) analyses have been used to assess temporal variations in marine viral communities. A study conducted over 78 days at one location off the coast of California showed significant fluctuations of the bacterial and T4-type myovirus populations from days to weeks but a more resilient population structure over longer time periods (29). A 3-year study off the California coast (San Pedro Ocean Time-Series [SPOT]) found seasonable variability, with spring/summer and autumn/winter T4-like operational taxonomic units (OTUs) but also persistent OTUs throughout the 3 years, in contrast with the Bank model and the “Killing the Winner” hypotheses, which predict much more dynamic virus-host fluctuations (1, 30–32). At the same site and over the same time period, the interactions between protists, bacteria, and T4-like phages were charted, revealing a complex microbial network in which viruses are thought to be mainly controlled by host availability (33). A similar seasonability (winter/spring, summer, autumn) combined with persistent OTUs throughout the seasons was found in a fjord system in Norway, albeit at a lower diversity (34). Using g23 amplicon sequencing on samples from the Chesapeake and Delaware Bays, a phylogenetic analysis showed no associations between g23 polymorphism, phage genome size, sampling time, or location (35).

The g23 marker gene has also been extensively used to investigate phage diversity in Japanese and Chinese rice paddy field-associated niche environments. Phylogenetic analyses revealed distinct clades associated with paddy soil, rice straw, and paddy flood water (paddy groups I to IX), as well as a small number of sequences similar to environmental marine samples from previous studies (36–40). When comparing manganese nodules in Japanese paddy soil with plow soil and subsoil layers, no differences in phage community structures were found, nor were differences found when comparing soil depth profiles or soils in different regions of Japan or different rice plant parts (40–43). In an investigation of decomposing straw, a reduction in richness in T4-type phage signatures was found in the late stages of decomposition, the opposite of what was observed for the local bacterial population (44). During the decomposition of root callus cells, a proliferation of T4-type phages belonging to a limited number of paddy groups was observed in aerobic soils; no signal was found in anaerobic soils (45).

In a comparison of the g23 sequences from five different rice field soil types in northeast China, there was only a small difference between the soil types and the phylogenetic analysis showed most sequences clustered with the Japanese paddy groups and one novel China-specific group (46). In several dry upland black soils, 40 to 50% of the g23 clones grouped with the paddy groups in a phylogenetic analysis, while the rest formed separate groups or clustered separately, indicating a relatedness with rice paddy soils but a distinct community structure, which varied according to sampling location (47, 48). T4-type community structure in flooded paddy field soils and natural wetland soils in northeast China was found to be very diverse, and while the compositions were similar, they were significantly different from paddy soils in Japan and from marine, lake, and upland soil environments (49, 50).

In the freshwater Lake Baikal (Russia), g23 clones were most closely related to uncultured or cultured (ExoT-even) marine groups, with a few clones clustering with paddy group VII (51). The eutrophic Lake Kokotel, which is situated less than 5 km from the eastern shore of Lake Baikal, comprised a T4-like community that was different from that of Lake Baikal and more closely related to that of Lake Donghu, a lake with a similar trophic status (52, 53). The Antarctic Lake Limnopolar, which was dominated by eukaryotic viruses, still had a wide diversity of g23 signatures, mainly clustering in the cyano-T4 clade (54).

g23 clone analysis of water and sediment from cryoconite holes in an Arctic glacier (Svalbard, Norway) led to the creation of three unique T4-type clades, polar I to III, containing sequences from only Arctic or Antarctic regions, and three novel environmental clades (Env I to III) with sequences from a variety of habitats (55). An interesting finding in this study was a cluster of sequences in polar group III which shared up to 99% amino acid sequence identity with sequences from Lake Limnopolar (54).

The g23 signature gene appears to span a much greater diversity of bacteriophage groups within the Myoviridae family than the g20 gene, encompassing both cyanophages and noncyanophages and making it a more desirable marker gene for investigating myovirus diversity. Nevertheless, the prevalence of g23 in myoviruses has not yet been studied, leaving the question of how much of this family's diversity remains unstudied.

Alternative major capsid protein genes (mcp).

To investigate a specific group of non-T4-related cyanophages that infect freshwater filamentous cyanobacteria, a primer set was developed that amplified neither the mcp gene of T4-related marine cyanophage S-PM2, T4 itself (g23), nor that of several phycodnaviruses (56). The use of this novel primer pair revealed a heterologous group of previously uninvestigated cyanomyoviruses which were genetically very different from their marine counterparts that infect unicellular cyanobacteria. This study emphasized the limitation of surveying cyanophage diversity with the g20 and g23 marker genes.

The mcp gene of the heterogenous group of NCLDVs that infect unicellular eukaryotes was used as an alternative marker to the family B polymerase marker (see below) (57). The genus subdivisions of the Phycodnaviridae family were largely supported in phylogenetic analyses of the mcp marker gene. However, a close relationship with sequences of the family Mimiviridae was observed, suggesting some form of evolutionary relationship. For Emiliana huxleyi viruses, sequence analysis of environmental mcp amplicons significantly increased the known species richness of this viral group (58).

Very recently, the first single-stranded DNA (ssDNA) virus signature gene study was published, in which the major capsid protein of the subfamily Gokushovirinae in the Microviridae family was investigated (59). mcp signatures were found in all the environments tested (estuarine, freshwater, sediments, sewage) and were only distantly related to the cultured isolates of this subfamily. The findings of this paper suggested that the ssDNA viruses are much more cosmopolitan than previously believed, and when generalizing we can hypothesize that there are many more groups of uninvestigated viruses that may have such a global spread.

Tail sheath protein.

Typically, a tail sheath protein is only present in myoviruses, making this a potential marker for this specific viral family, but a general myovirus tail sheath primer set has not yet been developed. A real-time PCR of the tail sheath protein gene specifically for Microcystis aeruginosa Ma-Lmm01-type cyanophages in a freshwater lake allowed successful quantification of the virus particles (60). However, these quantitative values could not be reliably linked to host blooms, potentially because of inactivation of the virus particles. Additionally, a combination of this real-time PCR with a real-time PCR analysis of the host cells gave a negative correlation between phage quantity and host cell number, indicating that cyanomyoviruses are an important factor in M. aeruginosa population dynamics (61).

Photosynthesis-related genes psbA and psbD targeting cyanophages.

The use of the psbA-psbD signature genes originated from the discovery that the cyanophage S-PM2 carries the genes for the core proteins of photosystem II, D1 and D2 (genes psbA and psbD, respectively), which were later found to be widespread in cyamyoviruses and cyanopodoviruses (62, 63). Using the psbA gene, significant differences were found between viral (cultured, environmental, and prophage) and host gene phylogenies, possibly reflecting different evolutionary stresses (64). Phylogenetic analysis of psbA in marine and freshwaters revealed that this marker can distinguish between these two environments as well as between Synechococcus and Prochlorococcus hosts and myoviruses and podoviruses, but not between different geographical locations ranging from the Mediterranean Sea to the Arctic Ocean (65). Analysis of Japanese rice paddy field floodwaters showed novel viral psbA signatures (66). In a coastal environment in Hawaii where Prochlorococcus dominated, environmental psbA sequences clustered with either Prochlorococcus podoviruses or myoviruses and with two “unrepresented” clusters that were more distantly related, but not with Synechococcus phages, while the psbD sequences all grouped into one cluster with the cultured cyanophage P-SSM4 (67). The psbD gene sequence was used on a group of isolated phages from the Indian Ocean and revealed no distinct community structure at different depths (68).

The psbA marker has been found in isolated myoviruses and podoviruses infecting both Synechococcus and Prochlorococcus cyanobacterial hosts (67, 69), potentially making it a more complete marker than g20 for cyanophages. However, the full cyanophage richness remains unexplored, since the signature gene was found neither in siphoviruses nor in the full complement of myo- and podoviruses.

phoH.

The phoH gene, the expression of which was linked to phosphate starvation conditions but with no identified specific function, has been recently developed as a novel marker gene (70, 71). This gene was found in 40% of cultured marine phages but in only 4% of nonmarine phages, making it a good signature gene for assessment of marine phage diversity. Phylogenetic analysis showed that phages clustered separately from their hosts, that phages infecting heterotrophic bacteria were more diverse than cyanophages, that eukaryotic viruses formed a distinct cluster, and that the phage community composition differed with depth and geographical location (70).

phoH has some advantages as a marker gene compared with some of the other signature genes described in this review. The gene has been found in phages infecting autotrophs and heterotrophs, is not restricted to one phage family, and has also been detected in viruses of photosynthetic green algae (70). Therefore, despite only including about 40% of marine viruses, the use of this single signature gene will produce a more complete picture of marine phage diversity in a chosen habitat than any other single marker gene. On the other hand, caution is needed when using this marker, since several enterobacteriophages also contain a copy of the phoH gene, which could lead to an increased signal in habitats contaminated with human or animal waste.

T7-like DNA polymerase polA in the family Podoviridae.

The family A DNA polymerase of the T7-like “supergroup,” which comprises the genera belonging to the subfamily Autographivirinae and several cyanopodophages, contains several conserved regions which can serve as targets for degenerate primers in this subfamily (Table 1) (72, 73). Two partial polymerase sequences, named HECTOR and PARIS, have been found in biomes around the world (marine, freshwater, estuarine, extreme, terrestrial, and metazoan associated), and while PARIS was less abundant than HECTOR, both were more abundant in marine environments than other environments (72). With a different primer set for the DNA polymerase gene (Podo-F/Podo-R2) (Table 1), three environmental groups were found in marine samples clustering separately from the HECTOR and PARIS sequences and from previously cultured phage groups, such as the T7 group, the P60 group, and the SI-01 group (74). In Chesapeake Bay, seasonal changes of the podovirus community were found between winter and summer based on use of a 550- to 600-bp fragment of the polA gene (75). Compared with the estuarine Chesapeake Bay community, the open ocean podoviruses were less diverse, yet globally ubiquitous (76). A bioinformatics analysis of the GOS data revealed that those viruses that were present all fell into the preexisting clades (74), suggesting that the currently used primer sets (Table 1) span the full diversity of polA-containing podoviruses. However, the metagenomic data suggested that there are many cyanopodovirus groups in which the polA gene is not detected (76).

T4-like DNA polymerase g43 in the family Myoviridae.

To date, only one study has used the T4-like DNA polymerase gene g43 for diversity studies. Isolated cyanophages infecting Synechococcus sp. WH7803 and seawater viral clone libraries were investigated in a 6-year study (77). Clear spatial (southern New England coast [United States], Long Island coast [United States], and Bermuda inshore waters) and temporal community compositional variations were found, although the overall richness and evenness did not change over the seasons at one location.

The g43 signature gene targets the same group of phages (cyanomyoviruses) as the g23 marker, but no comparative studies to evaluate the performance of these signature genes have yet been reported.

Family B DNA polymerase polB of NCLDVs.

Primers for the family B DNA polymerase gene polB were first designed for viruses infecting microalgae and were shown to amplify DNA from environmental viral communities, particularly viruses belonging to the Phycodnaviridae family (78, 79). When comparing the polB marker sequences to polymerase sequences of other virus groups, it was shown that the algal viruses were related to the Herpesviridae and that the Herpesviridae, Poxviridae, Baculoviridae, and African swine fever virus each clustered separately (79). PCR-RFLP analysis of a natural viral community in the Gulf of Mexico revealed five different OTUs which, after sequencing, were all grouped into the Phycodnaviridae family, in which Micromonas pusilla viruses seemed the most abundant (80).

PCR-DGGE of a polB fragment revealed that the algal virus population in coastal waters was much more stable than the eukaryotic population and, at times, correlated with tide height, salinity, and chlorophyll a content (81, 82). Variance in community composition through the seasons was found in the subtropical coastal water of Hawaii, with phylotypes specific for winter or summer and only two phylotypes present throughout the year and at high abundance (83). In freshwater lakes in North America, polB phylogenetic analysis showed that the freshwater Phycodnaviridae sequences clustered together but were still related to the marine sequences (84). In another freshwater study, no clear temporal and spacial patterns of Phycodnaviridae were found but, based on analysis of the monophyletic groups, 13 different freshwater hosts were inferred (85). A comparison of marine and freshwater polB sequences, with addition of sequences from Amazon River systems, suggested that there is no significant gene flow of phycodnaviruses between freshwater and oceanic aquatic systems (86).

The AVS primer set does not amplify certain marine Phycodnaviridae, such as Emiliana huxleyi viruses and Herterosigma akashiwo viruses, and is suspected to preferentially amplify Micromonas pusilla virus isolates (84, 85), implying that abundance studies are skewed toward the latter and that the full phycodnavirus richness is not captured with these primers.

RNA-dependent RNA polymerase in the order Picornavirales.

The RNA-dependent RNA polymerase, present in all RNA viruses except retroviruses, was targeted to study the diversity of marine picorna-like viruses (87). This study found that this virus type is widespread and persistent in ocean environments with sequences related to viruses infecting algae, shrimp, and mammals, making this group extremely diverse. The use of this marker gene for investigation of marine RNA virus metagenomes revealed a distinct clade of marine picorna-like (Mpl) virus sequences (88). Extending the environmental range to subtropical waters increased the diversity within the Mpl clade, which was subsequently assumed to be a protist-infecting clade (89).