Cyanomyoviruses and Photosynthesis
Cyanomyoviruses are unique among T4-like phages in that their hosts utilize light as their primary energy source; therefore it is not to surprising cyanomyoviruses carry genes that may alter the photosynthetic capability of their hosts. The most well studied of the photosynthetic phage genes are
psbA and
psbD, which encode for the proteins D1 and D2 respectively. The D1 and D2 proteins form a hetero-dimer at the core of photosystem II (PSII) where they bind pigments and other cofactors that ultimately result in the production of an oxidant that is strong enough to remove electrons from water. As an unavoidable consequence of photosynthesis there is photo-damage to D1 and to a lesser extent the D2 protein, therefore all oxygenic photosynthetic organisms have evolved a repair cycle for PSII [
59]. The repair cycle involves the degradation and removal of damage D1 peptides, and replacement with newly synthesized D1 peptides [
59]. If the rate of removal and repair is exceeded by the rate of damage then photoinhibiton occurs with a loss of photochemical efficiency in PSII [
60]. A common strategy of T4-like phages is to shutdown the expression of host genes after infection, but if this was to occur in cyanomyoviruses then there would be a reduction in the reduction efficiency of the PSII repair cycle and thus reduced photosynthetic efficiency of the host. This would be detrimental to the replication of phage and it has therefore been proposed that cyanomyoviruses carry their own copies of
psbA to maintain the D1 repair cycle [
52]. There is strong evidence to suggest that this is the case with Q-PCR data proving the
psbA gene is expressed during the infection cycle for the phage S-PM2 and that there is no loss in photosynthetic efficiency during the infection cycle [
56]. Further evidence for the function of these genes can be gained from P-SSP7 a podovirus that also express
psbA during infection with phage derived D1 peptides also being detected in infected cells [
61]. Although as yet phage mutants lacking these genes have yet to be constructed the results of modelling with in silico mutants suggests that
psbA is a non essential gene [
62] and that its fitness advantage is greater under higher irradiance levels [
62,
63]
The carriage of
psbD is assumed to be for the same reason in the maintenance of photosynthetic efficiency during infection, indeed it has been shown that
psbD is also expressed during the infection cycle [Millard et al unpublished data]. However, not all phage are known to carry both
psbD and psbA, in general that the broader the host range of the phage the more likely it is to carry both genes[
40,
49]. It has therefore been suggested that by carrying both of these genes that phage can ensure the formation of a fully functional phage D1:D2 heterodimer [
49].
Cyanomyoviruses may maintain the reaction centres of their host in additional and/or alternative ways to the replacement of D1 and D2 peptides. The reaction centre of PSII may also be stabilized by
speD a gene that has been found in S-PM2, P-SSM4 and S-RMS4.
speD encodes S-adenosylmethionine decarboxylase a key enzyme in the synthesis of the polyamines spermidine and spermine. With polyamines implicated in the stabilising the
psbA mRNA in the cyanobacterium
Synechocystis[
64], altering structure of PSII [
65] and restoring photosynthetic efficiency [
66], it has been proposed they also act to maintain the function of the host photosystem during infection [
11].
Whilst
psbA and
psbD are the most studied genes that may alter photosynthetic ability, they are certainly not the only genes. The carriage of
hli genes that encode high light inducible proteins (HLIP) are also thought to allow the phages host to maintain photosynthetic efficiency under different environmental conditions. HLIP proteins are related to the chlorophyll
a/b-binding proteins of plants and are known to be critical for allowing a freshwater cyanobacteria
Synechocysti s to adapt to high-light conditions [
67]. The exact function in cyanomyoviruses is still unknown, they probably provide the same function of as HLIPs in their hosts, although this function is still to be fully determined. It is apparent that the number of
hli genes in phage genome is linked to the host of the cyanomyovirus with phage that were isolated on
Prochlorococcus (P-SSM2 & P-SSM4) having double the number of
hli genes found on the those phage isolated on
Synechococcus (S-RSM4, Syn9, S-PM2) (Table
2). The phylogeny of these genes suggest that some of these
hli genes are
Prochlorococcus specific [
68], probably allowing adaptation to a specific host.
A further photosynthetic gene that may be advantageous to infection of a specific host is
cepT. S-PM2 was the first phage found to carry a
cepT gene [
5], it is also now found in Syn9 [
23], S-RSM4 and 10 other phages infecting
Synechococcus[
43], but is not found in the phage P-SSM2 and P-SSM4 which were isolated on
Prochlorococcus[
49].
cepT is thought to be involved in regulating the expression of phycoerythrin (PE) biosynthesis [
69], PE is a phycobiliprotein that forms part of the phycobilisome that is responsible for light-harvesting in cyanobacteria [
70], the phycobilisome complex allows adaptation to variable light conditions such as increased UV stress [
70]. Recently it has been shown that amount of PE and chlorophyll increases per cell when the phage S-PM2 infects its host
Synechococcus WH7803, with this increases in light harvesting capacity thought to be driven by the phage to provide enough energy for replication [
6] with phage
cpeT gene responsible for regulation of this increase [
71]. As
Prochlorococcus do not contain a phycobilisome complex that contains PE, which the
cpeT regulates expression of, it is possibly a gene advantageous to cyanomyoviruses infecting
Synechococcus.
Phage genes involved in bilin synthesis are not limited to
cepT, within P-SSM2 the bilin reductase genes
pebA and
pcyA have been found and are expressed during infection [
72]. The
pebA gene is functional
in vitro and catalyses a reaction that normally requires two host genes (
pebA &
pebB) and has since being renamed
pebS, this single gene has been suggested to provide the phage with short tern efficiency over long term flexibility of the two host genes [
72]. Despite evidence of expression and that the products are functional it is unclear how these genes are advantageous to cyanomyoviruses infecting
Prochlorococcus which do not contain standard phycobilisome complexes.
Alteration of host photosynthetic machinery appears to be of prime importance to cyanomyoviruses with a number of genes that may alter photosynthetic function. In addition to maintaining PSII centres and altering bilin synthesis, a further mechanism for diverting the flow of electrons during photosynthesis may occur. A plastoquinol terminal oxidase (PTOX)-encoding gene was first discovered in P-SMM4 [
25] and then in Syn9 [
23] and more recently has been found to be widespread in cyanomyoviruses infecting
Synechococcus. The role of PTOX in cyanobacteria, let alone cyanomyoviruses, is not completely understood, but it is thought to play a role in photo-protection. In
Synechococcus it has been found that under iron-limited conditions CO
2 fixation is saturated at low light intensities, yet the reaction centres of PSII remain open at far higher light intensities. This suggests an alternative flow of electrons to receptors other than CO
2 and the most likely candidate acceptor is PTOX [
73]. The alternative electron flow eases the excitation pressure on PSII by the reduction of oxygen and thus prevents damage by allowing an alternative flow of electrons from PSII [
73]. Further intrigue to this story in that PTOX encoding genes are not present in all cyanobacterial genomes and are far more common in
Prochlorococcus genomes than in
Synechococcus genomes. Therefore, phage may not only maintain the current
status quo of the cell as in the same manner
psbA is thought to, but may offer an alternative pathway of electron flow if its host does not carry its own PTOX genes. Although this is speculative it is already known that cyanomyoviruses that carry PTOX genes can infect and replicate in
Synechococcus WH7803 that does not have PTOX-encoding gene of its own.
All sequenced cyanomyoviruses have genes that may alter carbon metabolism in their hosts, although not all cyanomyoviruses have the same complement of genes [
5,
23,
25]. Syn9 [
23] and S-RSM4 have
zwf and
gnd genes encoding the enzymes glucose 6-phosphate dehydrogenase (G6PD) and 6-phosphogluconate dehydrogenase which are enzymes utilised in the oxidative stage of the pentose phosphate pathway (PPP). The rate-limiting step in the PPP is the conversion of glucose-6-phosphate, which is catalysed by G6PD. It could be advantageous for a phage to remove this rate-limiting step in order to increase the amount of NADPH or ribulose 5-phosphate it requires for replication. Whether the phage removes this rate limitation by encoding a G6PD that is more efficient than the host G6PD or simply producing more, is not known. Without experimental data the proposed advantages of these genes are speculative.
There are at least 5 modes in which the PPP can operate depending on the requirements of the cell [
74]. It might be assumed that for a phage the priority might be to produce enough DNA and protein for replication, thus use the mode of PPP that produces more ribulose 5- phosphate at the expense of NAPH. The production of ribulose 5-phosphate could then be used as the precursors for nucleotide synthesis. This mode of flux would result in the majority of glucose-6-phosphate being converted to fructose-6-phosphate and glyceraldehyde 3-phosphate. These molecules could then be converted to ribulose 5-phosphate by a transaldolase and transketolase.
Therefore, it is not surprising that
talC has been detected in four of the five sequenced cyanomyovirus genomes, in viral metagenomic libraries [
54], and in fragments of cyanomyovirus genomes S-BM4 [
53] and SWHM1 (this lab unpublished data).
talC encodes a transaldolase, an important enzyme in linking PPP and glycolysis, that if functional would catalyze the transfer of dihydroxyacetone from fructose 6-phospate to erythrose 4-phosphate, giving sedoheptulose 7-phosphate and glyceraldehyde 3-phosphate. However, currently this alteration of the PPP is speculation as other modes of flux are just as possible depending on the circumstances the phage find it self within its host with alternative modes leading to an increase in the production ATP and NADPH [
23].
It does appear that maintaining or altering carbon metabolism is important to cyanomyoviruses as the genes
trx is also found Syn9 and S-RSM4. The product of
trx is thioredoxin, an important regulatory protein that is essential in the co-ordination of the light-dark reactions of photosynthesis by the activation of a number of enzymes, one of the few enzymes that it suppresses is glucose-6-phosphate dehydrogenase [
75]. The reduced form of thioredoxin controls enzyme activity, with thioredoxin itself reduced by ferredoxin in a process catalysed by ferredoxin-thioredoxin reductase [
76]. Whilst no cyanomyovirus have been found to have ferredoxin-thioredoxin reductase, the cyanomyovirus S-RSM4 and P-SSM4 do have
petF, that encodes ferredoxin,. Ferredoxin acts as an electron transporter which is associated with PSI, whether the phage petF replaces host petF function is not known.
The function of another electron transporter is also unclear, some cyanophages (S-RSM4, Syn9, P-SSM2) have a homologue of
petE. Host
petE encodes plastocyanin, which transfers electrons from the cytochrome
b6f complex of photosystem II to P700
+ of photosystem I. It is known cyanobacterial
petE mutants show both a reduced photosynthetic capacity for electron transport and slower growth rate [
77]. Thus, it is possible that the phage
petE is beneficial by means of maintaining photosynthetic function.
Whilst there are a number of genes, trx, zwf, gnd, petE, petF that may alter host carbon metabolism, unravelling their function is not a trivial task, this is exemplified genes such as trx that can regulate enzymes in the Calvin cycle, PPP, and gluconeogenesis. This is further complicated by the fact that to date no two cyanomyovirus to date have exactly the same complement of genes that may alter carbon metabolism, with S-PM2 having none of the above mentioned and at the opposite end of the spectrum S-RSM4 has the full complement. However, the widespread distribution of these genes in cyanomyoviruses suggests their presence is not coincidental and they may be advantageous to cyanomyovirus under certain environmental conditions.
Gene transfer between the T4-likes and their hosts (impact on host genome evolution in the microbial world)
As discussed in the preceding sections there is clear evidence that cyanophages have acquired a plethora of genes from their bacterial hosts. These are recognisable either by being highly conserved such as
psbA which is conserved the amino acid level, or by the presence of a shared conserved domain with a known gene. Phages potentially have two methods of donating phage genes back to their hosts; through generalised or specialised transduction. Generalised transduction results from non-productive infections where phages accidently package a head full of host DNA during the stage when their heads are being packaged and they inject this into a second host cell during a non-fatal infection. Specialised transduction in comparison results from the accidental acquisition of a host gene resulting from imprecise excision from a host which would occur during lysogenic induction. Although this area has been poorly studied there is some evidence for both generalised and specialised transduction in cyanophages [
85].
Despite little direct evidence of lysogeny in marine cyanophages the relationship between host and phage genes can be established from phylogenetic analyses. When host genes are acquired by phages, they generally drift from having the GC composition of their hosts to that of the phage genome. This difference is much clearer in Synechococcus-phage relationships because Synechococcus genomes have a GC % of around 60% compared to the phages which have a GC% of around 40%. The GC of psbA in Synechococcus phages has drifted to a value between the average host and phage GC% so is around 50%. These differences are less clear in Prochlorococcus as it tends to have a similar CG% to the phages which infect it and thus phylogenetic analysis can be dominated by homoplasies (the same mutation happening independently).
All of the robust phylogenetic analyses that have been performed on metabolic phage genes that are shared between hosts and phages suggest that phages have generally picked up host genes on limited occasions and this has been followed by radiation has within the phage populations for example see Millard et al. 2005 [
53].
There is nothing known about the biology and molecular basis of lysogeny or pseudolysogeny in T4 type cyanomyoviruses. Indirect evidence for the abundance of lysogens was obtained from studies on inducing wild populations of cyanobacteria and quantifying the number of potential phages using epifluorescence. This work demonstrated that more temperate phages could be induced in winter when the number of cyanobacterial hosts was low and so conditions were hostile for phages in the lytic part of their life cycle. Other studies have suggested that the apparent resistance
Synechococcus shows to viral infection may be due to lysogenic infection [
3]. It is also clear that the phosphate status of cyanobacteria influences the dynamics of integration [
86]. During nutrient starvation cyanoviruses enter their hosts but do not lyse the cells, their genes are expressed during this period (Clokie et al., unpublished). The cells are lysed when phosphate is added back into the media. It is not known exactly how cyanophage DNA is integrated into the cell during this psuedolysogenic period but this may be a time in which genes may be donated and integrated from the phage genome to that of the host.
Despite a lack direct evidence for phage-mediated gene transfer, it is likely that transduction is a major driver in cyanobacterial evolution as the other methods of evolution are not available to them. In the open oceans DNA is present at such low levels (0.6 - 88 μg liter
-1) that it is probably too dilute for frequent transformation [
87]. Also both
Synechococcus and
Prochlorococcus appear to lack plasmids and transposons rendering conjugation an unlikely method for the acquisition of new genes. The large number of bacteriophages present in the oceans as well as the observation that phage-like particles appear to be induced from marine cyanobacteria, along with phage-like genes found in cyanobacterial genomes suggests that transduction is evident as a mechanism of evolution.
The genetic advantages that the T4-like cyanomyoviruses may confer to their hosts were listed in a recent review, but in brief they are: (1) prophages may function as transposons, essentially acting as foci for gene rearrangements, (2) they may interrupt genes through silencing non-essential gene functions, (3) they may confer resistance to infection from other phages, (4) they may excise and kill closely related strains, (5) they may cause increased fitness by the presence of physiologically important genes or (6) the phages may silence host genes.
In summary, it is difficult to pin down the exact contribution that T4-like cyanoviruses play in microbial evolution but their abundance, modes of infection and genetic content imply that they may be extremely important for cyanobacterial evolution. Their contribution will become clearer as more genomes are sequenced and as genetic systems are developed to experiment with model systems.
The impact of cyanomyoviruses on host populations
The two major biotic causes of bacterial mortality in the marine environment are phage-induced lysis and protistan grazing, currently efforts are being made to assess the relative impacts of these two processes on marine cyanobacterial communities. Accurate information is difficult to obtain for the oligotrophic oceans because of intrinsically slow rate processes [
88]. It must also be borne in mind that there are likely to be extensive interactions between the two processes e.g. phage-infected cells might less or more attractive to grazers, phage-infected cells might be less or more resistant to digestion in the food vacuole and phages themselves might be subject to grazing. Estimates of the relative effects of phage-induced lysis and grazing on marine cyanobacterial assemblages vary widely e.g. [
89‐
91] and this probably reflects the fact the two processes do vary widely on both temporal and spatial scales.
A number of methods have been developed to assess viral activity in aquatic systems, but all suffer from a variety of limitations such as extensive sample manipulation or poorly constrained assumptions [
92,
93]. The application of these approaches to studying cyanomyovirus impact on
Synechococcus populations has produced widely varying results. Waterbury and Valois [
3] calculated that between 0.005% (at the end of the spring bloom) and 3.2% (during a
Synechococcus peak in July) of the
Synechococcus population was infected on a daily basis. Another study [
94] indicated that as many as 33% of the
Synechococcus population would have to have been lysed daily at one of the sampling stations. A subsequent study using the same approach [
95] yielded figures for the proportion of the
Synechococcus community infected ranging from 1 - 8% for offshore waters, but in nearshore waters only 0.01 - 0.02% were lysed on a daily basis. Proctor and Fuhrman [
96] found that, depending on the sampling station, between 0.8% and 2.8% of cyanobacterial cells contained mature phage virions and making the questionable assumption that phage particles were only visible for 10% of the infection cycle, it was calculated that percentage of infected cells was actually ten-fold greater than the observed frequency.
An important consideration in attempting to establish the impact of cyanomyoviruses on their host populations is to ask at what point the infection rate becomes a significant selection pressure on a population, leading either to the succession of intrinsically resistant strains, or the appearance of resistant mutants. It has been calculated that the threshold would occur between 10
2 and 10
4 cells ml
-1[
10] and this is in agreement with data from natural
Synechococcus populations that suggest that a genetically homogeneous population would start to experience significant selection pressure when it reached a density of between 10
3 and 10
4 cells ml
-1[
97].
The community ecology of cyanomyovirus-host interactions is complicated by a number of factors including the genetic diversity of phages and hosts, protistan grazing and variations in abiotic factors (e.g. light, nutrients, temperature). Thus simple modelling of predator-prey dynamics is not possible. However, a "kill the winner" model [
92,
98] in which the best competitor will become subject to infection has gained widespread acceptance. Recently, marine phage metagenomic data have been used to test theoretical models of phage communities [
99] and the rank-abundance curve for marine phage communities is consistent with a power law distribution in which the dominant phage keeps changing and in which host ecotypes at very low numbers evade phage predation. A variety of studies have looked at spatio-temporal variations in cyanomyovirus populations. The earliest studies showed that cyanomyovirus abundance changed through an annual cycle [
3] and with distance from shore, season and depth [
94]. The ability to look at the diversity of cyanomyovirus population using
g20 primers revealed that maximum diversity in a stratified water column was correlated with maximum
Synechococcus population density [
30] and changes in phage clonal diversity were observed from the surface water down to the deep chlorophyll maximum in the open ocean [
28]. Marston and Sallee [
35] found temporal changes in both the abundance, overall composition of the cyanophage community and the relative abundance of specific
g20 genotypes In Rhode Island's coastal waters. Sandaa and Larsen [
34] also observed seasonal variations in the abundance of cyanophages and in cyanomyovirus community composition in Norwegian coastal waters. Cyanomyovirus abundance and depth distribution was monitored over an annual cycle in the Gulf of Aqaba [
40]. Cyanophages were found throughout the water column to a depth of 150 m, with a discrete maximum in the summer months and at a depth of 30 m. Whilst it is clear from all these studies that cyanomyovirus abundance and community composition changes on both a seasonal and spatial basis, little is know about short term variations. However, one study in the Indian Ocean showed that phage abundance peaked at around 0100 at a depth of 10 m, but the temporal variation was not as strong at greater depths [
84]. It may well be the case that infection by cyanomyoviruses is a diel phenomenon as phage adsorption to host is light-dependent for several marine cyanomyoviruses studied [
100]. A similar observation for the freshwater cyanomyovirus AS-1 [
101]. There is currently only one published study that describes attempts to look at the co-variation in the composition of
Synechococcus and cyanomyovirus communities to establish whether they were co-dependent [
102]. In the Gulf of Aqaba, Red Sea, a succession of
Synechococcus genotypes was observed over an annual cycle. There were large changes in the genetic diversity of
Synechococcus , as determined by RFLP analysis of a 403 bp
rpoC1 gene fragment, which was reduced to one dominant genotype in July. The abundance of co-occurring cyanophages capable of infecting marine
Synechococcus was determined by plaque assays and their genetic diversity was determined by denaturing gradient gel electrophoresis analysis of a 118 bp
g20 gene fragment. The results indicate that both abundance and genetic diversity of cyanophage covaried with that of
Synechococcus. Multivariate statistical analyses show a significant relationship between cyanophage assemblage structure and that of
Synechococcus. All these observations are consistent with cyanophage infection being a major controlling factor in cyanobacterial diversity and succession.
Analysis of the impact of cyanomyoviruses on host populations has been based on the assumption that they follow the conventional infection, replication and cell lysis life cycle, but there is some evidence to suggest that this may not always be the case. There is one particularly controversial area of phage biology and that is the topic of pseudolysogeny. There are in fact a variety of definitions of pseudolysogeny in the literature reflecting some quite different aspects of phage life history, but the one adopted here is "the presence of a temporarily non-replicating phage genome (a preprophage) within a poorly replicating bacterium" (S. Abedon - personal communication). The cyanobacterial hosts exist in an extremely oligotrophic environment posing constant nutritional stress and are exposed to additional environmental challenges such as light stress that may lead to rates of growth and replication that are far from maximal. There is evidence that obligately lytic
Synechococcus phages can enter such a pseudolysogenic state. When phage S-PM2 (a myovirus) was used to infect
Synechococcus sp. WH7803 cells grown in phosphate-replete or phosphate-deplete media there was no change in the adsorption rate constant, but there was an apparent 80% reduction in the burst size under phosphate-deplete conditions and similar observations were made with two other obligately lytic
Synechococcus myoviruses, S-WHM1 and S-BM1 [
86]. However, a more detailed analysis revealed this was due to a reduction in the proportion of cells lysing. 100% of the phosphate-replete cells lysed, compared to only 9% of the phosphate-deplete cells, suggesting that the majority of phosphate deplete cells were pseudolysogens.
From very early on in the study of marine cyanomyoviruses it was recognized that phage-resistance was likely to be an important feature of the dynamics of phage-host interactions. Waterbury and Valois [
3] found that coastal
Synechococcus strains were resistant to their co-occurring phages and suggested that the phage population was maintained by a small proportion of cells sensitive to infection. For well studied phage-host systems resistance is most commonly achieved by mutational loss of phage receptor on the surface of the cell, though there are other mechanisms of resistance to phage infection e.g. [
103]. Stoddard et al. [
104] used a combination of 32 genetically distinct cyanomyoviruses and four host strains to isolate phage-resistant mutants. Characterization of the mutants indicated that resistance was most likely due to loss or modification of receptor structures. Frequently, acquisition of resistance to one phage led to cross-resistance to one or more other phages. It is thought that mutation to phage resistance may frequently involve a fitness cost and this trade-off allows the coexistence of more competitive phage-sensitive and less competitive phage-resistant strains (for review see [
105]). The cost of phage resistance in marine cyanobacteria has been investigated by Lennon et al. [
106] using phylogenetically distinct
Synechococcus strains and phage-resistant mutants derived from them. Two approaches were used to assess the cost of resistance (COR); measurement of alterations in maximum growth rate and competition experiments. A COR was found in roughly 50% of cases and when detected resulted in a ~20% reduction in relative fitness. Competition experiments suggested that fitness costs were associated with the acquisition of resistance to particular phages. A COR might be expected to be more clearly observed when strains are growing in their natural oligotrophic environment. The acquisition of resistance to one particular cyanophage, S-PM2, is associated with a change in the structure of the lipopolysaccharide (LPS) (E. Spence - personal communication).
A variety of observations arising from genomic sequencing have emphasized the role of alterations in the cell envelope in the speciation
Prochlorococcus and
Synechococcus strains, presumably as a result of selection pressures arising from phage infection or protistan grazing. An analysis of 12
Prochlorococcus genomes [
107] revealed a number of highly variable genomic islands containing many of the strain-specific genes. Amongst these genes the greatest differentiator between the most closely related isolates were genes related to outer membrane synthesis such as acyltransferases. Similar genomic islands, containing the majority of strain-specific genes, were identified through an analysis of the genomes of 11
Synechococcus strains [
108]. Among the island genes with known function the predominant group were those encoding glycosyl transferases and glycoside hydrolases potentially involved in outer membrane/cell wall biogenesis. The cyanomyovirus P-SSM2 was found to contain 24 LPS genes that form two major clusters [
25]. It was suggested that these LPS genes might be involved in altering the cell surface composition of the infected host during pseudolysogeny to prevent infection by other phages. The same idea could apply to a normal lytic infection and could be extended to protection against protistan grazing. Similarly, cyanomyovirus S-PM2 encodes a protein with an S-layer homology domain. S-layers are quasi-crystalline layers on the bacterial cell surface and so this protein, known to be expressed in the infected cell as one of the earliest and most abundantly transcribed genes [
56], may have a protective function against infection or grazing.
The potential value of continuing research on the "eco-genomics" of cyanophages
Eco-genomics is defined as the application of molecular techniques to ecology whereby biodiversity is considered at the DNA level and this knowledge is then used to understand the ecology and evolutionary processes of ecosystems. Cyanophage genomes encode a huge body of unexplored biodiversity which needs to be understood to further extend our knowledge of cyanophage-cyanobacteria interactions and thus to fully appreciate the multiple roles that cyanophages play in influencing bacterial evolution, physiology and biogeochemical cycling.
As cyanophage genomes are stripped down versions of essential gene combinations an understanding of their genomics will assist in defining key host genes that are essential for phage reproduction. As many of the host genes encoded in phage genomes have an unknown function in their hosts, the study of phage genomes will impinge positively on our understanding of cyanobacterial genomes. The other major spin-off from researching the products encoded by phage genomes is the discovery of novel enzymes or alternative versions existing enzymes with novel substrate specificities. This is likely to be of major importance to the biotechnology and pharmaceutical industries.
As more phage genomes and metagenomes are sequenced, the core set of phage genes will be refined and the extent of phage encoded host metabolic and other accessory genes will be revealed. We would expect to find specific environments selecting particular types of genes. This research area is often referred to as 'fishing expeditions' especially by grant panels. However it is analogous to the great collections of plants and animals that occurred during the 19
th Century. These data were collected over a long period of time and it was only subsequently that scientists understood patterns of evolution, biogeography, variance and dispersal. This is an exciting time to be mining cyanophage genomes as metagenomic analysis of the viral fraction from marine ecosystems has suggested that there is little restriction to the types of genes that bacteriophages can carry [
109]. These data will likely provide the bedrock on which generations of scientists can interpret and make sense of.
To drive our understanding of cyanophage genomes forward however there needs to a concerted effort to capitalise on the sequence libraries that are being collected from both phage metagenomes and phage genomes. Sequencing even large data-sets is now comparatively easy and sequence information should be seen as the exciting starting point rather than the endpoint. To determine the function of the reservoir of genes will require extensive biochemical, chemical and molecular biological investigations as well as physiological experimentation.
Currently when new T4-like cyanophage genes are identified using bioinformatic approaches, they are compared to T4 and their function is deduced on the basis of known genes in the T4 genome. In order to really progress with understanding the role of the genes which have no homology (and to confirm the homology in genes where an identity can be hypothesised) a genetic system needs to be developed where cyanophages can be mutated. This will take extensive research effort and hopefully international research groups will come together in the way that researchers on the T phages in the 1960 s did to gradually piece together to determine the function of the genes that constitute largest reservoir of genetic diversity on earth.