Correspondence: Ruth-Anne Sandaa, Department of Biology, University of Bergen, PO Box 7800, N-5020 Bergen, Norway. Tel.: +47 5558 4646; fax: +47 5558 9671; e-mail: firstname.lastname@example.org
This study reports the diversity of uncultured environmental viruses harbouring photosynthetic genes (psbA and psbD) in samples from cold seawater (latitude above 60°). The viral community in coastal Norwegian waters was separated according to genome size using pulse field gel electrophoresis. Viral populations within a wide genome size range (31–380 kb) were investigated for the presence of the psbA and psbD genes using PCR, combined with cloning and sequencing. The results show the presence of photosynthetic genes in viral populations from all size ranges. Thus, valuable information could be obtained about the size class to which viral particles that encode photosynthesis genes belong. The wide genomic size range detected implies that a different cyanophage profile has been observed than has been reported previously. Thus, the method of phage gene detection applied here may represent a truer picture of phage diversity in general or that there is a larger range of size profile for viruses with psbA and psbD in higher latitudes than for the better-studied lower latitudes. Alternatively, a picture of diversity based on a different set of biases than that from either isolation-based research or from conventional metagenomic approaches may be observed.
The main prokaryotic component of the picophytoplankton in the photic zone of the world's oceans consist of marine unicellular cyanobacteria from the genera Synechococcus and Prochlorococcus. Together they contribute up to 89% of the primary production in the oligotrophic regions of the oceans (Li, 1995; Liu et al., 1997; Partensky et al., 1999). Cyanophages are viruses that infect cyanobacteria and are, like their hosts, ubiquitous in marine environments. They are found in concentrations up to 106 particles mL−1 in coastal waters during the summer period and are considered a significant factor in determining the dynamics of cyanobacterial populations (Suttle & Chan, 1993, 1994; Waterbury & Valois, 1993).
Prochlorococcus is essentially ubiquitous between 40°N and 40°S, but is not found in water where the temperature is <10 °C (Olson et al., 1990). This contrasts with Synechococcus, which is somewhat broader distributed in that it tolerates a wider range of temperature. Prochlorococcus is normally about 10-fold more abundant than Synechococcus in the oligotrophic regions of the open oceans (Vaulot et al., 1995), while Synechococcus populations are most abundant in both coastal and colder waters (reviewed in Partensky et al., 1999). Most Synechococcus phage research has been carried out in low-latitude temperate and tropical waters and there is a lack of knowledge about the diversity of these phages in waters of latitudes above 60°.
The objective of this study was to investigate the diversity of photosynthetic genes in uncultured cyanophages from Norwegian costal waters. All previous studies of photosynthetic viral genes have been either on isolated viruses (Lindell et al., 2004, 2005; Millard et al., 2004; Sullivan et al., 2005, 2006) or using whole viral fractions (Zeidner et al., 2005; Sullivan et al., 2006). The approach described here is unique as it allows one to investigate the presence of photosynthesis genes in noncultured viruses, but where the genome size of the virus is known. This type of information cannot be obtained from standard metagenomic data sets. To do this, viral PFGE bands within a wide genomic size range were investigated for the presence of the photosynthetic genes, psbA and psbD. PCR products with amplicons of these two genes were cloned, sequenced, and phylogenetic analyses were performed.
Materials and methods
Coastal water samples were collected at a station in Raunefjorden (60°16.2′N, 5°12.5′E), south of Bergen, Norway at nine different time points between 13 April and 2 November 2004 (Table 2). A total volume of 30 L was collected from a depth of 2 m. The samples were collected using a hand pump connected to a flask. Temperature was measured using a STD SAIV a/s SD 204 with a Sea Point fluorometer (SAIV A/S, Environmental Sensors & Systems, Bergen, Norway).
Table 2. Information about size range and sampling dates of investigated PFGE bands for the presence of the photosynthetic genes psbA and psbD, water temperature and number of cyanophages in the samples
Natural viral communities were concentrated from 25 L of seawater by ultrafiltration. The samples were filtrated through a 0.45-μm pore size low-protein-binding Durapore membrane filter 142 mm in diameter (Millipore) to remove zooplankton, phytoplankton and some bacteria. The filtered samples were then concentrated down to a final volume of around 150–250 mL, using a 30 000 MW cut-off spiral-wound Millipore ultrafiltration cartidge (Regenerated Cellulose, PLTK Prep/scale TFF 1 ft2, Millipore). One hundred and forty milliliters of this concentrate was concentrated further by ultracentrifugation (Beckman L8-M with SW-28 rotor, Beckman GmbH, Germany) for 2 h at 25 000 r.p.m. at 10 °C. The viral pellet was dissolved in 400 μL of SM buffer [0.1 M NaCl, 8 mM MgSO4·7H2O, 50 mM Tris-HCl (pH 8.0), 0.005% (w/v) glycerine]. Two hundred microliters was stored at −20 °C for quantitative real-time PCR analysis, while 200 μL was used for PFGE analysis.
Four virioplankton agarose plugs were made from the 200 μL concentrate. The samples were separated on a 1% w/v SeaKem GTG agarose (FMC, Rockland, ME) gel in 1 × TBE gel buffer using a Bio-Rad DR-II CHEF Cell (Bio-Rad, Richmond, CA) electrophoresis unit (Wommack et al., 1999). From each sample point three of the plugs were used, each at a different pulse ramp condition in order to separate the large range of viral genome sizes: (1) 1–5 s switch time with a 20-h run time for separation of small genome sizes (0–130 kb); (2) 8–30 s switch time with a 20-h run time for separation of medium genome sizes (130–300 kb); (3) 20–40 s switch time with a 22-h run time for separation of large genome sizes (300–600 kb). A molecular size standard [λ-ladder and 5-kb ladder (Bio-Rad)] was run on each side of the gel. Further details of the procedure are found in Larsen et al. (2001). The gels were visualized and saved as computer files using the Fujifilm imaging system, LAS-3000.
PCRs cloning and DNA sequencing
PFGE bands to be investigated for the presence of photosynthetic genes were excised from the gel and frozen at −20 °C. The DNA was extracted from the gel using the GeneClean Turbo kit (BIO101) for extraction of large DNA fragments from agarose gel, following the manufacturer's instructions, yielding c. 10 ng μL−1 of DNA (total 30 μL). The genomic DNA required for sequencing and PCR were produced by the GenomiPhi DNA amplification kit (Amersham Biosciences) according to the manufacturer's instructions yielding c. 1 μg μL−1 of DNA. This DNA was used in a PCR with both the primer sets targeting a section of the photosynthetic genes psbA and psbD (Table 1). PCRs were carried out in a total volume of 50 μL containing sterile distilled water, PCR buffer (10 × PCR buffer B, Promega, Madison, WI), dNTPs (each 200 nM), primers (each 0.5 μM), 1.5 mM MgCl, 2.5 U Taq DNA polymerase (Promega) and template amplicon (1–2 ng). Amplification conditions using the psbA primers were as follows: 94 °C for 5 min, 10 cycles of 94 °C for 30 s, 64 °C (−1 °C per cycle) for 30 s, and 72 °C for 1 min. There was then an extension of 2 min at 72 °C, followed by 25 cycles of 94 °C for 30 s, 56.5 °C for 30 s and 72 °C for 1 min. The final extension was at 72 °C for 10 min. Furthermore, the amplification conditions for the psbD primers were 94 °C for 5 min, 35 cycles of 94 °C for 1 min, 50 °C 1 min, and 72 °C for 1 min, and a final extension at 72 °C for 10 min. The PCR products were cloned with the TOPO PCR cloning kit (Invitrogen, Paisley, UK) following the manufacturer's description. The resulting reactions were used to transform competent Escherichia coli TOP10 (Invitrogen). Fifteen positive clones (white colonies) from each library were picked randomly and transferred by streaking onto agar plates. Positive clones were confirmed by PCR using the M13 primers according to the protocol (Invitrogen). Positive PCR products were purified using the DNA Clean & Concentrator-5 kit (Genetix Limited, New Milton, UK). Five positive PCR products from each cloning reaction were sequenced by cycle sequencing according to the protocol from Perkin-Elmer (Foster City) using the cloning primer M13f (Invitrogen) as the sequencing primer. Sequences were obtained on the ABI PRISM 3700 sequence analyser (Perkin-Elmer Applied Biosystem).
Table 1. Primers targeting the photosynthetic genes psbA and psbD
Between two and five clones resulted in sequences of good quality and were used in the phylogenetic analysis. Analysis of DNA sequences was carried out by alignment to the closest relative in the GenBank database using blastx (Altschul et al., 1990). Alignments were performed using clustalx (Thompson et al., 1997). Sequences were initially aligned based on protein sequences. The protein alignment was then used to align the corresponding DNA sequences. Maximum parsimony and neighbor-joining (NJ) analysis were conducted on nucleotide dataset using the test version of paup* 4.0 beta10 (Swofford, 2000). Supports for clades were estimated by means of bootstrap analysis, as implemented in paup* using 1000 replicates. The trees were viewed using the treeview program and rooted with either the psbA or psbD gene of Synechocystis PCC 6803. The nucleotide sequences reported in this paper have been submitted to GenBank and assigned the accession numbers psbD, DQ787206–DQ787235 and psbA, DQ787236–DQ787256.
Results and discussion
This is the first study to investigate the diversity of uncultured environmental viruses harbouring photosynthetic genes (psbA and psbD) in samples from cold seawater. The study was performed by investigating 19 PFGE bands in a broad size range (31–380 kb) from samples taken at different time periods (Fig. 1) based on information on the viral diversity and cyanophage seasonal dynamics in Raunefjorden (Sandaa & Larsen, 2006). The bands that were chosen represented the brightest bands in each of the different size classes. The viral community in Raunefjorden have shown a pronounced seasonal dynamic that correlates with changes in the abundance of possible hosts (Sandaa & Larsen, 2006). Most of the bands examined in the present study were from samples collected between late September and early November. This is at the same time when there is a bloom in the Synechococcus population in the fjord (Sandaa & Larsen, 2006). The bloom is followed by a peak in the cyanophage numbers, and is accompanied by a major change in the viral community structure (Sandaa & Larsen, 2006). The water temperature in Raunefjorden at the sampling times was between 5.3 and 13.3 °C (Table 2). Over a 10-month-period, the water temperature at the sampling depth in Raunefjorden may vary between 5.2 and 15.4 °C (Sandaa & Larsen, 2006). As Prochlorococcus has not been reported in water with temperatures below 10 °C (Olson et al., 1990) it is reasonable to conclude that the putative uncultured cyanophages presented in this study infect Synechococcus strains.
Eleven out of 19 investigated PFGE bands in this study contained detectable photosynthetic genes. The genes were detected in PFGE bands with genomic sizes from 31 to 380 kb (Table 2). Although the size range of genomes containing psbA and psbD is larger than has been described before, it may be even larger than the authorsapos; report, as bands smaller or larger than this size range were not investigated. Eight of the PFGE bands contained both the psbA and psbD genes, while three of the PFGE bands had only one of the genes (Table 2). The psbA and psbD genes are highly conserved (Lindell et al., 2004; Millard et al., 2004; Sullivan et al., 2006), suggesting that they encode functional proteins that may be involved in maintaining host photosynthesis during infections (Clokie et al., 2006a). Most cultured cyanophages carry both the psbA and psbD genes (Lindell et al., 2004; Millard et al., 2004; Sullivan et al., 2006); however, some only contain psbA (Millard et al., 2004; Sullivan et al., 2006). Two of the PFGE bands in this study contained only the psbA gene. However, in contrast to earlier published findings, only the psbD gene was amplified from one of the PFGE bands (Table 2). This is the first observation of cyanophages carrying the psbD gene, only. These observations, along with the fact that psbD and psbA are very far apart from each other in Synechococcus genomes, imply that the genes have been acquired independently from their hosts.
The present knowledge on the occurrence and diversity of the photosynthetic genes in cyanophages is mainly based on isolated phages, belonging to two viral families, Myoviridae and Podoviridae (Mann, 2003; Lindell et al., 2004; Millard et al., 2004; Sullivan et al., 2005, 2006) and from their abundance in metagenomic data sets (Zeidner et al., 2005; Sullivan et al., 2006). Metagenomic data sets have been very valuable in raising one's awareness of the widespread nature of phage encoded photosynthesis genes but do not provide further information about the phage biology. Although whole and partial genome sequencing of cultured cyanophages has revealed psbA and psbD genes in Synechococcus myoviruses, none have been shown to be in any Synechococcus podoviruses (Sullivan et al., 2006); however, it should be noted that very few Podoviridae infecting Synechococcus have been isolated or sequenced. The genome size in cultured cyanophages varies from c. 48 to 200 kb and the trend is for myoviral genomes (normally in the size range 100–200 kb) to be larger than podoviral genomes which normally range from 38 to 48 kb (Wichels et al., 1998; Mann, 2003). Forty-five per cent of the observed PFGE bands with photosynthetic genes had genome sizes in this range (95–200 kb); however, 55% fell outside this normal size range. Three of the populations (S9, S12 and S13) were in the genomic size range 31–67 kb, which according to size suggests they might belong to the Podo- or Siphoviridae. So far, there is no report of Synechococcus phages from Siphoviridae, and only few isolates reported belonging to Podoviridae (Waterbury & Valois, 1993; Fuller et al., 1998; Chen & Lu, 2002; Sullivan et al., 2005, 2006). On the other hand, if these viral populations belong to Myoviridae, this is also interesting as it suggests that the size range of this taxonomic group must be much broader that earlier reported. The other viral populations (3b, 5b and S1) were in the genomic size range of 240–380 kb. Although it is most likely that these are from Myoviridae, they are much bigger than any reported thus far. Another plausible explanation is that these photosynthetic genes are from viral populations infecting photosynthetic picoeukariotes, as viruses infecting algae do have genome sizes in this size range (Van Etten & Meints, 1999; Sandaa et al., 2001; Castberg et al., 2002). These psbA and psbD genes, however, exhibited highest similarity to other cyanophage and Synechococcus sequences in GenBank, not to photosynthetic genes in picoeukariotes. Therefore, it was believed that these genes are present in cyanophages infecting Synechococcus. It should though be stressed that these observations are based on genome size alone, and that physiological properties of the phage, e.g., viral morphology, essential for the correct definition of phage taxonomy.
Phylogenetic analysis of the psbA and psbD was carried out in paup* using both maximum parsimony and a neighbour-joining analysis (NJ). The two methods gave congruent results and those presented here are from the NJ analysis (Fig. 2a and b). Most of the clusters of the environmental sequences from this study lack sequences from cultured cyanophage or hosts, which suggests that these sequences may belong to phages that infects as-yet uncultured Synechococcus hosts. This is a common observation when using culture-independent approaches for investigation of environmental viruses. The psbA sequences fell into several distinct groups. From the top of the tree and downwards, one clade of seven psbA sequences has 87% bootstrap support and consists only of environmental sequences determined in this study. There is then an interesting group (with 76% bootstrap support) that contains both a new psbA sequence from this study, an environmental clone from the Monterey Bay, CA (Zeidner et al., 2003) and a known Synechococcus myophage S-RSM2. Two smaller clades again contain two and three new environmental psbA sequences, respectively. The two Synechococcus strains included in this analysis form a clade with an environmental psbA sequence from the Monterey Bay, California (Zeidner et al., 2003). There is one final large group that contains the remainder of the environmental psbA sequenced from Raunefjorden and which at its higher level contains two environmental sequences, one from the Red Sea (BAC9D04) and the other from the Mediterranean Sea (V141) (Zeidner et al., 2005). The sequence BAC9DO4 is from a putative podophage based on the fact that genes upstream of it are podophage like (Zeidner & Béjà, 2004), but the phage has not been isolated so the actual morphology has not been confirmed. Thirty-eight per cent of the psbA sequences from this study clustered in this large clade. If these viral populations are podophages, infecting Synechococcus, this is in contrast to what has been shown using different methods in other environments. Therefore, it may be that the method of phage gene detection employed here may represent a truer picture of phage diversity in general, or that phage diversity in high latitudes has a different profile to that from better-studied lower latitudes. It may alternatively be that the method used in this study simply has a different set of biases than isolation-based or strict metagenomic approaches. Although the phage psbA sequences isolated here cluster with the sequence in BAC9D4, it should be emphasized that there is significant variation at the nucleotide level, as there was a sequence difference of c. 25%.
The environmental psbD sequences obtained in this study fell into several distinct groups (Fig. 2b), showing greatest similarity to three Synechococcus myophages, syn1, syn10 and S-RSM2 (Millard et al., 2004; Sullivan et al., 2006). It is worth mentioning that the psbD gene has only been detected in cyanophage isolates belonging to Myoviridae (Sullivan et al., 2006). The isolate, S-RSM2 was isolated from the Gulf of Aquaba (Millard et al., 2004). PsbD appears to be a much better phylogenetic marker than psbA as all of these new psbD sequences were distinguishable from their host psbD genes, in contrast to the phage-encoded psbA genes that form a clade together with their host genes (Fig. 2a and b). Thus, compared with the phage-encoded psbA genes, these new psbD genes might have had a longer purifying selection time resulting in a clear divergence of the viral and host psbD genes. It may alternatively be the case that there is less evolutionary constraint on psbD than psbA genes and thus the phage-encoded versions can evolve more freely than phage-encoded psbA.
Interestingly, the psbD genes from the different clones partly cluster according to their original PFGE band. In contrast, the psbA genes did not show any such relationship. Thus, it seems that the level of diversity observed in the psbD gene is less than that observed in psbA where multiple very closely related genes were recovered from a single band. This might imply that the genes are moving or evolving (or both) independently, with different selection pressures that also would result in phylogenetic differences. Indeed, it may be that a specific phage ‘population’ may have multiple versions of psbA. Alternatively, differences between the trees may also be attributed to the fact that each PFGE band may consist of more than one viral population, with a similar genome size and so the clone library of one viral band can thus contain photosynthetic genes from different, but genetically similar cyanophages. Although, some of the photosynthetic genes in this study might originate from host genes, the fact that all the psbD sequences from Raunefjorden displayed greatest similarity to isolated cyanophages, and not to their hosts, is inconsistent with this assumption, confirming the reliability of both the viral concentration step and the PFGE analysis. In summary the psbD analysis shows a clear diversity of phage-encoded photosynthetic gene, while the psbA analysis appears to suffer from homoplasy; despite this, it is still useful as it has hinted at the family affiliation to which eight of the newly sequenced phage populations may belong.
Using PFGE for studies of the viral community, it is possible to detect between 105 and 106 viral particles mL−1 (Wommack et al., 1999; Steward, 2001). One important issue is the size of the viral genome that will influence the sensitivity of the method. For viruses with larger genome sizes, the detection limit will be lower, e.g., a viral genome of 300 kb might be detected down to 105 particles mL−1 (Steward, 2001). In this analysis 25 L of sample was collected, which was concentrated by a factor of c. 100. If a detection limit of 105 was assumed, 103 particles mL−1 of one population was needed to result in a signal on the PFGE gel.
In order to produce sufficient DNA from the different PFGE bands, a technique based on whole-genome amplification was applied, using the enzyme ϕ29 DNA polymerase. The ϕ29 DNA polymerase is an enzyme which is widely used for rolling-circle amplification of plasmids and circular DNA templates (Dean et al., 2002; Detter et al., 2002). The strand-displacing enzyme has proofreading activity, is extremely sensitive, and has been shown to amplify DNA of up to 70 kb (Blanco et al., 1989). This polymerase has also been used for environmental whole-genome amplification of bacterial and viral communities (Abulencia et al., 2006; Angly et al., 2006) and for a variety of genetical applications including sequencing and microsatellite marker and single nucleotide polymorphism (SNP) analysis of single individuals (Jiang et al., 2005). These results have demonstrated that the introduction of bias may be due to the sizes of DNA templates (possibly sheared during extraction and mixing), random primer availability, and stochastic effects of amplifying from very low concentrations of template (Jiang et al., 2005; Abulencia et al., 2006). Another bias may be the differential cloning efficiency of amplified DNA compared with unamplified DNA (Abulencia et al., 2006). Most of these biases will not be valid in the study as the multiple displacement amplification (MDA) approach was used to produce DNA that were used in a second PCR with specific primers. However, as this technique does apply several amplification steps, one should be aware of errors in the PCR product that might place clones from the same viral population at different positions in the tree.
The authors' objective with this study was to describe the distribution and diversity of photosynthetic genes in different genomic size ranges of viral populations, using a culture-independent approach such as PFGE in combination with cloning and sequencing. With PFGE it is possible to gain information about the dominant and presumably the most active viral populations in the samples, without the need for a cultureable host. So far, most knowledge on photosynthetic genes in phages is based on cultured cyanophages isolated using two cyanobacterial strains Synechococcus WH7803 and Prochlorococcus Med4 as hosts. Recent estimates have shown that more than 99% of marine bacteria (potential phage hosts) are unculturable (Rappé & Giovannoni, 2003), supporting the importance of a culture-independent approach for studies of genetical and functional viral diversity in environmental samples. Compared with other culture-independent approaches, e.g., metagenome cloning (Zeidner et al., 2005), PFGE also provides information about the genome size of the viruses. Such information can be used to suggest the likely family that these viruses belong to. Using this approach a much higher diversity in viral populations was discovered with photosynthetic genes than that reported previously. The results document the naturally occurring genetic diversity among uncultured environmental viruses in samples from cold seawater. Furthermore, the interesting findings that viral populations of different sizes harbour photosynthetic genes that were phylogenetically similar supports the idea of promiscuous horizontal gene transfer of gene modules within a common cyanophage gene pool. Thus, it might be hypothesized that marine cyanophages could have played a vital role in the evolution of photosynthetic gene diversity by providing an accessible pool of portable genes and facilitating the reshuffling, acquisition and exchange of such genetic material.
The authors are grateful to Thomas Sørlie for collecting the seawater samples and Mette Hordnes (UiB) for providing the water temperature data. This study was performed with financial support from Research Council of Norway project: Biodiversity patterns: blooms vs. stable coexistence in the lower part of marine food webs and EU project: Bacterial single-cell approaches to the relationship between diversity and function in the sea (BASIC), Contract number: EVK3-CT-2002-00078.