Bacteriophage adenine methyltransferase: a life cycle regulator? Modelled using Vibrio harveyi myovirus like



Shaun Bochow, Microbiology and Immunology, School of Veterinary & Biomedical Sciences, James Cook University, 1 Solander Drive, Townsville, Qld 4811, Australia.



The adenine methyltransferase (DAM) gene methylates GATC sequences that have been demonstrated in various bacteria to be a powerful gene regulator functioning as an epigenetic switch, particularly with virulence gene regulation. However, overproduction of DAM can lead to mutations, giving rise to variability that may be important for adaptation to environmental change. While most bacterial hosts carry a DAM gene, not all bacteriophage carry this gene. Currently, there is no literature regarding the role DAM plays in life cycle regulation of bacteriophage. Vibrio campbellii strain 642 carries the bacteriophage Vibrio harveyi myovirus like (VHML) that has been proven to increase virulence. The complete genome sequence of VHML bacteriophage revealed a putative adenine methyltransferase gene. Using VHML, a new model of phage life cycle regulation, where DAM plays a central role between the lysogenic and lytic states, will be hypothesized. In short, DAM methylates the rha antirepressor gene and once methylation is removed, homologous CI repressor protein becomes repressed and non-functional leading to the switching to the lytic cycle. Greater understanding of life cycle regulation at the genetic level can, in the future, lead to the genesis of chimeric bacteriophage with greater control over their life cycle for their safe use as probiotics within the aquaculture industry.


The Vibrionaceae are a diverse group of gammaproteobacteria with members exhibiting wide environmental preferences from fresh water to marine. Vibrio harveyi can cause severe mortality in a number of hosts, particularly crustaceans (Pizzuto and Hirst 1995). This pathogen has also been observed to be antibiotic resistant (Karunasagar et al. 1994). Of importance to this review is Vibrio campbellii Australian Collection of Marine Microbiology (ACMM) strain 642 isolated from diseased Penaeus monodon larvae from northern Queensland (Muir 1987) (reported in the literature as V. harveyi 642, see taxonomy section). Further investigation revealed the production of exotoxins capable of causing 100% mortality in larval P. monodon (Harris and Owens 1999). The virulence was attributed to a temperate prophage (a bacteriophage or ‘phage’ that incorporates its genome into the host bacterial genome). This prophage was designated Vibrio harveyi myovirus like (VHML) and contained a putative ADP-ribosylating toxin (Oakey and Owens 2000; Oakey et al. 2002). VHML was extracted, purified and challenged avirulent strains of Vharveyi, which subsequently produced toxins (Munro et al. 2003) and exhibited phenotypic variation (Vidgen et al. 2006). Putatively identified on the VHML genome was a deoxyribonucleic acid (DNA) adenine methyltransferase (DAM) known to be a gene regulator and virulence determinant. Methyltransferases are in most cells and some phages (Jones and Takai 2001). The enzyme functions as a powerful epigenetic gene regulator switching genes on and off. This is achieved by the addition of a methyl group to a particular base within a defined short sequence of DNA. The methylation affects the binding of regulatory and polymerase proteins, which changes the transcription of the methylated genes (Low et al. 2001). Currently, there are no publications on the role that DAM plays in phage conversion (the incorporation of the phage genome into the host genome), but the role DAM plays in bacteria has been comprehensively reviewed (Low et al. 2001; Lobner-Olesen et al. 2005; Marinus and Casadesus 2009). This review will examine whether DAM has a direct role in gene regulation, resulting from the epigenetic switch of genes related to the phage life cycle. A new model of phage life cycle gene regulation where DAM plays a pivotal role will be considered.

Identification of Vibrio Species

The Harveyi clade includes Vharveyi, Vcampbellii, Vibrio alginolyticus, Vibrio rotiferianus, Vibrio parahaemolyticus, Vibrio mytili, Vibrio natriegens and the recently described Vibrio owensii (Cano-Gómez et al. 2010), Vibrio azureus (Yoshizawa et al. 2009) and Vibrio sagamiensis (Yoshizawa et al. 2010). The clade importantly shares a similar genetic composition. The core Harveyi group (Vharveyi, Vcampbellii, Vrotiferianus and Vowensii) shares approx. 99–100% 16S rRNA similarity (Cano-Gomez et al. 2011). Thus, traditional phenotypic and biochemical tests lack the resolving power to confidently differentiate between members of this clade (Cano-Gómez et al. 2010).

A set of biochemical keys developed by Noguerola and Blanch (2008) incorporated a vast set of growth conditions, nutrient preference and antibiotic sensitivity. The keys produce a 90% differentiation threshold, but this is not sufficiently accurate for all members of the almost genetically indistinguishable Harveyi clade (Gomez-Gil et al. 2004). Furthermore, Vidgen et al. (2006) showed that phenotypic changes occur when Vibro spp. are infected with prophage. This was reinforced by Lin et al. (2010), who used comparative genomic hybridization to show that V. harveyi strains BAA-1116 and HY01 were actually V. campbellii. This is interesting because BAA-1116 has been the model for quorum sensing in V. harveyi, but this model now applies to V. campbellii, raising the question of how varied quorum sensing actually is in V. harveyi.

Molecular Identification

Multilocus sequence analysis (MLSA) using conserved house-keeping genes rpoA, pyrH, topA, ftsZ and mreB can identify bacteria within the Harveyi clade with high resolving power (Thompson et al. 2005). For example, Vharveyi-like strains 47666-1, isolated from diseased larval P. monodon, and DY05T, isolated from the larval rearing system of Panulirus ornatus, were identified using 16S rRNA, MLSA and DNA–DNA hybridization to be a new species subsequently named V. owensii (Cano-Gomez et al. 2011) (see Gomez-Gil et al. (2004) and Cano-Gómez et al. (2010) for reviews of the molecular techniques used to positively identify Vibrio spp.). Using MLSA, Cano-Gomez et al. (2011) showed that the type strain of Vcommunis shared 98·6% locus sequence similarities with Vowensii and is therefore considered a junior synonym of the latter. This has been independently supported by Hoffmann et al. (2012). Vharveyi stain 642 was also reclassified as Vcampbellii strain 642. It is likely that early literature identifying V. harveyi actually described bacteria from other species within the Harveyi clade. It is possible that V. campbellii is a much more significant pathogen in aquaculture than first believed (Gomez-Gil et al. 2004). As the taxonomy of the Vibrio spp. is ever changing, this review will report the species as they appeared in the literature, but confirmation of taxonomy of isolates is necessary.

Vibrio Infections of Aquatic Animals

Vibriosis is a major disease affecting the aquaculture industry, with a wide range of species affected including cold-water species such as Salmo salar to the tropical crustacean Pmonodon. Relevant literature contains a multitude of reports of Vibrio spp. impacting the production and culture of aquaculture species (Table 1). The major virulence factors include extracellular enzymes that damage organs and the cuticle of infected animals.

Table 1. Reports of vibriosis within the aquaculture industry
Species impactedBacteriaIncidence reportedLocationReferences
P. vannameiVibrio harveyi strain CAIM 1792‘Bright red’ syndromeNorth western MexicoSoto-Rodriguez et al. (2010)
Penaeus monodon

V. harveyi

Vibrio alginolyticus

Vibrio vulnificus

Aeromonas fischeri

Vibrio parahaemolyticus

P. damsela

White spot syndrome

and shell disease

Peninsular IndiaManilal et al. (2010)


(review article)

VibrioVibriosisN/AAguirre-Guzmán et al. (2004)
Salmo salarA. salmonicidaCold-water VibriosisNorwayColquhoun and Sorum (2001)
Oncorhynchus mykiss and S. salar14 isolates of V. harveyiVibriosisExperiments carried out in United KingdomZhang and Austin (2000)
P. vannameiV. harveyi strain E2

Diseased postlarval


EcuadorMontero and Austin (1999)
P. monodon larvae

V. harveyi

Strain: 47666-1 and ACMM 642

Concentrated protein fractions caused 100% mortalityAustraliaHarris and Owens (1999)
P. monodon

V. harveyi

Strain: See paper

Extracellular productsTaiwanLiuxy et al. (1996)

P. monodon

P. japonicas

V. harveyi strain ValWhite spot syndromeTaiwanLee et al. (1996)

Acanthopagrus cuvieri

Epinephelus tauvina

V. harveyiMortalityKuwaitSaeed (1995)
P. monodon larvaeV. harveyiMass mortality (70–90%)IndiaKarunasagar et al. (1994)

Vibriosis of P. monodon is characterized by weakness, opaque colour, feeble and intermittent swimming movements and greenish luminescence (Harris and Owens 1999). Histology of infected P. vannamei revealed necrosis of lymphoid tubules, muscle packs, melanization and haemocytic nodules associated with microcolonies, while significant infection of V. harveyi led to systemic infection (Soto-Rodriguez et al. 2010). The main mode of virulence of Vibrio spp. is the production of extracellular products (ECP) including gelatinase, caseinase, phospholipase, haemolysin, elastase, chitinase (see Manilal et al. (2010) for enzymatic activity), alkaline proteases, cysteine proteases, alkaline metal chelator-sensitive proteases, serine proteases, metalloproteases (Aguirre-Guzmán et al. 2004) and siderophores (Owens et al. 1996; Colquhoun and Sorum 2001). These studies also indicated that high levels of ECP were needed for virulence; strains that produced more than one ECP were considered to have increased virulence, but one ECP did not stand out as being solely responsible for virulence (Manilal et al. 2010). Virulence has been associated with bacteriophage, as is the case with V. campbellii strain 642.

Bacteriophage Conversion and Life Cycle

The Myovirus have an icosahedron or elongated icosahedron nucleocapsid that contains DNA (Harper and Kutter 2008; Munn 2011). Initiation of infection occurs once a phage comes into contact with a bacterial host that has susceptible surface cell receptors. This susceptibility is generally highly specific, with a narrow host range, although some phages have a broad host range (Jensen et al. 1998). Once the nucleic acid has entered the host, there are two possible outcomes: the lytic life cycle or the lysogenic life cycle. The lytic life cycle occurs when the bacterial replication machinery is taken over and used to produce virons before the host cell is lysed releasing progeny phage. Lysogeny occurs when the phage genome is incorporated into the host genome becoming a prophage. There are two major factors that determine whether the phage can become lysogenic. (i) Genes: if the phage does not carry lysogeny genes, it cannot undergo the lysogenic life cycle and become a prophage. (ii) Bacterial physiology: bacteria grown in rich broth generally produce lytic phage. In the case of Lambda phage, this is a result of higher concentrations of host RNase III that favours lytic growth. Conversely, carbon-starved bacterial hosts produce less RNase III (Wilson et al. 2002).

Bacteriophage Therapy

The emergence of pathogenic bacterial strains resistant to antibiotics has caused a renaissance in phage research in search for an alternative treatment (Goodridge 2004). Pseudomonas aeruginosa is a multidrug-resistant Gram-negative pathogen of humans, with phage treatment being successful in placebo-controlled, double-blind clinical trials (Harper and Enright 2011). Phage endolysins have also shown promise in killing bacterial pathogens. The enzyme is purified and applied exogenously to bacteria causing lysis and cell death. These lysins used to prevent colonization of mucosa and systemic disease, are highly specific to target bacteria and will kill both growing and non-growing bacteria, unlike penicillin (Fenton et al. 2010). Phage can also be used to create bacterial ghosts, that is, lysed bacteria containing only the cell wall, used as a vaccine. Phages have also been used to deliver drugs directly to specific target sites, thus reducing systemic toxicity and side effects (Paukner et al. 2005; Petty et al. 2007).

There is much research into the potential use of lytic phage in combating pathogenic bacterial infections occurring in aquaculture (Harper and Kutter 2008; Sulakvelidze et al. 2001). The Australian crustacean aquaculture industry is generally banned from using unregistered antibiotics, which drives research to find alternative methods (Akinbowale et al. 2006). Vinod et al. (2006) and Karunasagar et al. (2007) isolated lytic phage that reduced mortality caused by V. harveyi within P. monodon hatcheries. Connerton et al. (2011) also described the successful use of phage to control Campylobacter jejuni in chickens, thus suggesting that phage therapy may be an alternative to antibiotic treatment. Although phages were identified early as having antibacterial properties, phage research was hampered owing to the lack of understanding of their life cycle and mechanisms of lysis and the discovery of antibiotics.

Bacteriophage Superinfection

Bacteria that contain a temperate prophage are in some instances able to resist the action of other invading phage, a process known as superinfection exclusion. Attachment and penetration may occur as normal; however, not all the nucleic acid is inserted. The remaining nucleic acid is injected into the periplasm and is degraded by phage endonuclease. This process is regulated by genes found on the prophage genome (Lu and Henning 1994). Likewise, Myung and Calendar (1995) demonstrated that an exonuclease of the temperate coliphage P2 genome found in Escherichia coli was able to interfere with the growth of temperate prophage Lambda (λ). Nesper et al. (1999) were able to show that genes carried on the phage K139 were able to prevent superinfection of Vibrio cholerae. It would seem that bacteria with a temperate prophage, such as V. campbellii carrying VHML, have the ability to resist superinfection of lytic phage.

The genome of VHML has been sequenced by Oakey et al. (2002) and was found to contain an old protein similar to that of P2 phage that has both DNase and RNase activities. It is hypothesized that this may prevent other myoviruses from infecting V. campbellii, possibly protecting the host from infection by other phage. This raises the question of whether and how pathogenic V. campbellii carrying VHML can be removed, so phage can be used in therapy in an aquaculture hatchery.

Vibrio harveyi Myovirus Like Phage

The VHML genome [open reading frame (ORF) 17] contains an active site for an ADP-ribosylating toxin along with a DAM gene (Oakey et al. 2002). DAM methylates the adenine residue in GATC sequences, thus affecting the binding of regulatory proteins and polymerases important to gene expression (Oshima et al. 2002; Lobner-Olesen et al. 2005). There is a large body of literature on the subject of DAM as a virulence factor. Low et al. (2001) suggested that DAM was a gene regulator in E. coli and affected its virulence. DAM-negative Salmonella typhimurium had attenuated virulence with a much higher dose needed to cause mortality in BALB/c mice (Garcia del-Portillo et al. 1999). Research indicates that DAM has an epigenetic effect and can up- or downregulate gene expression (Lobner-Olesen et al. 2005). DAM also plays a role in DNA mismatch repair. As the new strand is not fully methylated but hemimethylated, enzymes are able to differentiate between the parent strand and new strand and fix the mismatch base. Furthermore, DAM helps initiate DNA replication by distinguishing between old origins and new origins (Marinus and Casadesus 2009). However, there is a gap in the literature regarding DAM and its possible role in phage life cycles, perhaps as an epigenetic switch between lysogenic and lytic life cycles.

Vibrio campbellii Strain 642

Vibrio spp. have long been known to produce exotoxins that induce morbidity and mortality in crustaceans such as P. monodon larvae (Karunasagar et al. 1994). Harris and Owens (1999) isolated proteinaceous exotoxins from V. harveyi 47666-1 (now V. owensii) and V. campbellii 642 designated T1 and T2, respectively. These toxins caused 100% mortality in larval P. monodon and CBA mice. The LD50 of concentrated protein fractions was 2·1 and 3·1 μg g−1 for T1 and T2, respectively, via intraperitoneal injection of mice, while LD50 of concentrated protein fractions was 1·8 and 2·2 μg g−1 for T1 and T2, respectively, via intramuscular injection of P. monodon.

The possibility of a phage conferring virulence to V. campbellii was realized when a new phage VHML was found to infect Vcampbellii 642 (Oakey and Owens 2000). Mitomycin C was used to damage host DNA, causing VHML to become lytic so that it could be purified. Transmission electron microscopy visualized and confirmed the presence of VHML in the Vcampbellii 642. VHML also infected strains of V. harveyi and V. campbellii 45 and 645, respectively, that did not carry native prophage. Subsequent sequencing of VHML found a putative ADP-ribosylating toxin similar to that of V. cholerae (Oakey et al. 2002) and may thus be involved in the virulence of Vcampbellii 642.

The phage VHML was later shown to cause virulence in once avirulent strains of V. harveyi and V. campbellii. Larval P. monodon were bath challenged using various Vibrio strains experimentally infected with VHML. V. harveyi strains with VHML obtained from 642 caused significantly greater mortality compared to controls, with upregulation of haemolysin and production of toxic proteins recognized (Munro et al. 2003).

Conversion by VHML also changed the phenotypic profile of strains of V. harveyi; most notably, V. campbellii 642 was observed to have resistance to the vibriostatic compound O/129 (150 μg) (Vidgen et al. 2006). This highlights the importance of phage as carriers of virulence genes and the ability to increase antibiotic resistance.


Bacteria have evolved many mechanisms for regulating gene expression, particularly those related to virulence. This includes phase variation or the switching of genes on and off. This leads to bacteria and phage alike exhibiting heterogeneous gene expression (Haagmans and Van Der Woude 2000). Phase variation is influenced by the environment and is characterized by site-specific recombination, homologous recombination, slipped strand mispairing and reversible insertion of sequence, all of which change the DNA sequence (Henderson et al. 1999). Another mechanism of controlling gene expression is epigenetics by DNA methylation: specific sequences can be methylated, hemimethylated or unmethylated. DNA methylation is a powerful tool used by bacteria, plants and mammals to control the epigenetic switch of genes, by altering the timing and level of gene expression under methylation (Jones and Takai 2001). Methylation affects the affinity of regulatory proteins and polymerases by preventing binding.

Classes of Methyltransferase

There are several classes of DNA methyltransferase (MTase) including DNA cytosine MTase (Dcm), which methylates the C-5 position of cytosine in CC(A/T)GG sequences, DNA adenine methyltransferase, which methylates N-6 of adenine in GATC sequences, and cell cycle-regulated methylase (CcrM), which methylates the N-6 adenine of GAnTC (Low et al. 2001). These classes all play various roles within the host cell: Dcm plays multiple roles in gene expression in eukaryotic cells; DAM regulates certain genes within the cell; CcrM is involved in cell cycle regulation (refer to Low and Casadesus 2008 on the difference between CcrM and DAM).

These enzymes all donate a methyl group to their respective base sequence from S-adenosyl methionine. Only DAM and CcrM do not have associated cognate restriction enzymes, suggesting they are purely involved in gene regulation and not restriction modification systems (Ratel et al. 2006). Sequence data show many Vibrio spp. and a variety of other bacteria carry the DAM gene (Table 2). It should also be noted that phages also carry the DAM gene (Table 2), thus raising the question of the role of DAM in phage gene expression.

Table 2. Results of a search for adenine methyltransferase in NCBI gene database. Highlighted in grey are phage, while the rest are host sequences
Species:Gene descriptionFamilyHostPrimary Life cycleGene IDReferences
Vibrio phage VHMLPutative adenine methyltransferaseMyovirus likeVibrio campbellii strain 642Lysogenic (genome)956159Oakey et al. (2002)
Vibrio phage ICP1Putative adenine methyltransferase Myoviridae V. cholerae O1Lytic10228677Seed et al. (2011)
Burkholderia phage phi1026bPutative DNA adenine methylase Burkholderia spp. 2744117 
Vibrio phage VP882DNA adenine methylase Myoviridae Vibrio parahaemolyticus strain 03:K6Lysogenic (plasmid)5076208Lan et al. (2009)
Vibrio phage phiVC8 N/A Vibrio spp.UnpublishedGenBank: JF712866.1 
Vibrio phage VSK N/A Vibrio spp.UnpublishedGenBank: AF453500.3 
V. cholerae strain MJ-1236Hypothetical adenine-specific methylase yfcB   7856047Grim et al. (2010)
V. cholerae O1 biovar El Tor strain N16961DNA adenine methylase   2615643Bandyopadhyay and Das (1994)
V. cholerae strain O395DNA adenine methylase   5136322Julio et al. (2001)
V. harveyi ATCC BAA-1116 (now V. campbellii)DNA adenine methylase   5553986Lin et al. (2010)
V. vulnificus strain MO6-24/OMethyl-directed repair DNA adenine methylase   10165054Park et al. (2011)
V. parahaemolyticus strain RIMD 2210633DNA adenine methylase   1190292Nasu et al. (2000) and Makino et al. (2003)
V. splendidus strain LGP32DNA adenine methylase   7162354 
V. anguillarum strain 775Adenine-specific methyltransferase   10776434Naka et al. (2011)
V. fischeri strain ES114DNA adenine methylase   3279427Ruby et al. (2005)
Treponema succinifaciens strain DSM 2489Site-specific DNA mythyltransferase (adenine-specific)   10432289 
Deinococcus proteolyticus strain MRPD12 class N6 adenine-specific DNA methyltransferase   10257623 
Paludibacter propionicigenes strain WB4TSite-specific DNA methyltransferase (adenine-specific)   10000511Gronow et al. (2011)

DNA Adenine Methyltransferase

Binding of adenine methyltransferase and transfer of a methyl group

Methylation of the adenine occurs at specific GATC sites after the DNA has replicated. DAM catalyses the transfer of a methyl group from the donor S-adenosyl methionine, producing S-adenosyl-l-homocysteine and the base 6-methyladenine (Fig. 1). DAM binds to the template DNA, sliding along and methylating the GATC sequences in a single binding event. Horton et al. (2006) used DAM of E. coli and base substitution experiments to show that the first guanine of the GATC sequence interacts with side chains K9 and Y138 in DAM. Although affinity of the enzyme decreased when the guanine base was substituted, DAM was still able to recognize the ATC part of the sequence. Recognition of the third base (thymine) occurs through hydrophobic side chains L122 and P134, which forms a Van der Waals bond. The guanine in the fourth base on the complementary strand (C:G) interacts via its O6 and N7 atoms with the guanidino group of R124 in a bifurcated hydrogen-bonding pattern. The interaction between R124 and the guanine on the complementary strand is required to activate DAM for catalysis. Once DAM has attached, base flipping of the adenine occurs, in which the base is flipped out and bound to the active site of the enzyme. Here, in the presence of S-adenosyl methionine, the methyl group is added to the base before back-flipping into position (Horton et al. 2006). The 6-methyladenine subsequently prevents the binding of regulatory proteins and polymerase at the site; the DNA also becomes thermodynamically unstable and alters the curvature of the DNA (Low and Casadesus 2008; Marinus and Casadesus 2009).

Figure 1.

(a) DAM enzyme, donor S-adenosyl methionine (circle) and DNA sequence GATC that is methylated. (b) DAM recognizes the sequence and attaches. (c) DAM, donor and sequence form a complex. (d) Adenine is flipped out into the active site. (e) Adenine is methylated. (f) 6-methyladenine is flipped back into the sequence, the complex disassociates and the DAM moves onto the next GATC

The various roles of adenine methyltransferase

DAM has a fundamental role in mismatched repair, an event that fixes mutations arising from errors in DNA replication. Using E. coli as a model, Marinus and Casadesus (2009) proposed DAM as the regulatory factor for repairing mismatch bases. Post-replication produces hemimethylated DNA that is a parent strand that is fully methylated, while the new strand is unmethylated behind the replication fork. Methylation of the new strand by DAM before the MutS can bind affects chromosomal replication in a number of species such as Caulobacter crescentus (Low and Casadesus 2008), V. cholerae and Yersinia pseudotuberculosis (Julio et al. 2001). Base mismatch is recognized by the MutS protein that recruits MutL. MutH then complexes with these allowing MutH to cleave the unmethylated strand at the site of mismatch. UvrD helicase then displaces MutH and this unwinds the DNA so exonucleases can cut the mismatch before it is repaired. The parent strand is identified by DAM methylation and prevents the binding of MutH. Therefore, mismatch repair is confined to the newly synthesized strand that is unmethylated. Over- or underproduction of DAM will cause increased mutations as distinction between parental strand and new strand cannot be made (Marinus and Casadesus 2009). Thus, overproduction of phage-encoded DAM could lead to higher phage mutation rates for adaptation in an ever-changing marine environment.

Gene Regulation and Virulence

DAM has been extensively studied in E. coli where it regulates the epigenetic switch of a variety of genes. DAM was identified as a virulence determinant in uropathogenic E. coli where it plays a pivotal role in the epigenetic switch of the pyelonephritis-associated pili (encoded by the Pap operon). The distal GATC site is methylated at low levels of PapI, while leucine-responsive regulator protein (Lrp) binds to the proximal GATC site, blocking transcription of the papBA, silencing the pili gene. DNA replication generates a hemimethylated GATC distal site; methylation of the GATC proximal site and increased levels of PapI result in Lrp binding to the GATC distal site. These events allow the expression of pili, thus giving rise to uropathogenic E. coli (Weyand and Low 2000; Low et al. 2001). Likewise, DAM plays an important co-regulatory role along with Lrp in Salmonella enterica that carry the conjugative virulence plasmid pSLT. Lrp has been shown to coordinate cellular metabolism in response to nutrient availability, while DAM has the ability to ‘sense’ the state of the host genome. Camacho and Casadesús (2002) hypothesized that together, these factors co-regulate conjugation, which is energetically expensive. In V. cholerae and Y. pseudotuberculosis, introduction of a plasmid from E. coli carrying DAM showed that overproduction caused virulence to be significantly attenuated. This research revealed that DAM is essential for viability in these species (Julio et al. 2001).

The Lambda Bacteriophage Model

The λ phage can take either the lysogenic or the lytic pathway, depending on the state of host E. coli, environmental signals and the number of infecting phage per cell. The lytic cycle is the default, while lysogeny is undertaken as a protective measure during host decline (Dodd et al. 2005; Oppenheim et al. 2005). Once inside the host, a series of events lead to the expression of either the lytic or lysogenic life cycle (Fig. 2). The gene coding for Cro (a lytic gene) is the first to be transcribed following phage infection or prophage induction. Cro weakly represses CII on promoter pR and CI on pL, CI and CII are repressors of Cro (Ptashne 2006). The N protein is transcribed and promotes assembly of a transcription complex on the RNA polymerase. This complex modifies the RNA so that the RNA polymerase can overcome the tL1 and tR1 transcription terminators. This modification allows for the expression of the delayed early function proteins downstream from the terminators. Phage DNA replication can now take place via O and P genes along with the transcription of Q, a late gene regulator. Q protein subsequently builds up, modifying the RNA polymerase. The modified polymerase can now overcome transcriptional terminators downstream to pR′, thus allowing for the expression of proteins for phage morphogenesis and host cell lysis. During host DNA damage, the host SOS system is activated that includes RecA (Sweasy et al. 1990). RecA causes CI autoproteolysis allowing Cro to be expressed, leading to the lytic life cycle. It should be noted that a search of the λ phage genome found no methyltransferase gene.

Figure 2.

Genetic map and transcriptional units of the lambda phage regulatory region. See in text for gene function (Oppenheim et al. 2005).

The option of undergoing lysogeny manifests early during infection of the host by λ phage. There is an initial overshoot of CI expression (programmed by CII), CI favours lysogeny by repression of pR, which also insures O, P and Q are prevented from being transcribed. CII also activates pRM to allow continual production of CI, which is now self-regulating and silences the lytic genes (see Dodd et al. 2005; Oppenheim et al. 2005; Atsumi and Little 2006 and Ptashne 2006 for reviews.)

Proposed Vibro harveyi Myovirus Like Conversion Model

A hypothetical model for the conversion of temperate prophage VHML has been proposed by Oakey et al. (2002, 2005) using putatively identified genes. However, we propose a new model based on the following: unlike λ phage, VHML does not contain the gene Int (catalyses the insertion of phage DNA into host chromosome), instead it is thought that VHML uses a transposition mechanism and a site-specific recombination protein to integrate. Like the λ phage, the lytic genes are believed to be repressed by a putative CI gene. However, the VHML antirepressor (which has homologous function with λ phage antirepressor CII) has homology with rha antirepressor of phage phi 80.

As DAM has previously been shown to play a role in gene and virulence regulation, it is possible that DAM plays an important role in the switching between lytic and lysogenic life cycles in VHML. The lytic and lysogeny genes are dispersed throughout the VHML genome, unlike the genes encoding structural proteins that are found in blocks (Oakey et al. 2005). However, ORF six that contains the putative lytic repressor gene CI is at the start of the genome, while ORF 48 carrying the putative antirepressor gene (homologous to rha that has homology with the CII antirepressor) is at the end of the genome. ORF 17 carrying the DAM gene is found before the putative rha anitrepressor gene (Oakey et al. 2002). It may be that the transcription of the DAM gene first allows for the methylation of the rha antirepressor gene, which has three GATC sites in the ORF, the first occurring 13 bases in (Fig. 3). Methylation of the rha antirepressor could allow the putative lytic repressor gene CI to repress the lytic life cycle allowing for lysogeny to occur.

Figure 3.

rha antirepressor gene homologous with bacteriophage phi 80. Bold underlined bases indicate methylation sequences used for DAM recognition. NCBI Reference Sequence: NC_004456.1.

Based on the known activity of DAM, it is plausible to suggest that DAM methylates the rha antirepressor gene. Like λ phage, the lytic life cycle could be induced by host RecA removing methylation of the rha antirepressor, which would subsequently repress the putative lytic repressor gene CI allowing for the lytic cycle to occur, in response to DNA damage, just as mitomycin C was used to induce VHML prophage production (Oakey and Owens 2000) (Fig. 4). This view is also supported by Oppenheim et al. (2005) who noted that λ phages with a defective CII gene were exclusively lytic, while those with a defective Cro gene were unable to follow the lytic pathway. The role of DAM could be tested with site-directed mutagenesis using a suicide vector to knock out the VHML DAM gene; likewise, this technique could be used to knockout the putative repressor gene. Oakey et al. (2005) also identified an ADP-ribosylating toxin that was located on the same ORF as the DAM gene and suggested that DAM controlled the expression of this toxin.

Figure 4.

Schematic of DAMs hypothesized role in VHML life cycle regulation.

Holin and Lysin and Their Possible Role in Vibro harveyi Myovirus Like

There are two possible strategies that phage can take during the lytic mode to lyse their host. These include the inhibition of host cell murein synthesis or the degradation of murein using holin and lysin (see Wang et al. 2000). The putative genome of VHML did not contain any holin or lysin genes; however, there were several unidentified ORFs that may contain these genes. These genes have been reported as having a wide genetic diversity (Summer et al. 2004). Targeting of the unidentified ORFs may be needed to ascertain their function and possible location on the genome. It is also possible that holin and lysin genes are under the control of DAM regulation.

Consequences of Genetically Modified Bacteriophage

The practical application of understanding how the DAM gene operates is to utilize it in phage therapy. Genetic modification of phage to produce a more controllable and lethal predator is not a new area of science (see Moradpour and Ghasemian (2011) for review). The production of genetically modified phage (GMP) has allowed for specific targeting of pathogens and the removal of unwanted side effects such as endotoxin release resulting from rapid cell lysis (Matsuda et al. 2005). Lu and Collins (2007) inserted a DspB gene (codes for dispersin B) into T7 phage that hydrolyses β-1, 6-N-acetyl-d-glycosamine, a crucial adhesive needed for biofilm formation of bacteria. Upon lysis of the cell, dispersin B was released into the biofilm disrupting it, thus producing two pronged attacks against the bacteria. Engineered phages have also been used to target bacterial gene networks in combination with quinolones. The quinolone loxacin leads to protein, lipid and DNA damage, which subsequently induces the SOS response resulting in DNA repair. By engineering the non-lytic M13mp18 phage to overexpress lexA3, a repressor of the SOS system, Lu and Collins (2009) demonstrated that the bactericidal effect of loxacin is improved by 2·7 orders of magnitude compared to treatment with unmodified phage. However, the host range of such phage is still limited but new genetic engineering is changing this.

For phage to be a desirable therapeutic, they should be highly virulent to the host bacterium and have a broad, ideally, intraspecific host range. The specificity of phage is also both beneficial and problematic for treatment. Specificity allows for targeted destruction of a particular pathogen, leaving the commensal flora untouched. Treatment with a single phage for multiple pathogens, such as those of the Harveyi clade, will result in only one specific pathogen being removed, leaving other pathogenic strains untreated. The creation of chimeric phage can solve this problem. The broad host range phage IP008 of E. coli present in sewage effluent has been shown to infect 37% of environmental E. coli isolates. Mahichi et al. (2009) removed the tail fibre genes of phage T2 and replaced them with that of IP008. This resulted in a chimeric phage that combined the virulence of the T2 phage with the broad host range of the IP008 phage. The established proof of concept can be used to produce a similar chimeric phage effective against the Harveyi clade.

The use of broad host range phage against strains within the Harveyi clade has previously been reported (Crothers-Stomps et al. 2009). Phage VhCCS-06 was isolated from an aquaculture water sample in northern Australia. This phage was able to lyse 60% of strains tested, including V. harveyi, V. campbellii, V. rotiferianus and V. parahaemolyticus, all members of the Harveyi clade. Further work is needed to sequence the genome of this phage and identify its tail fibre gene for use in genetic engineering. By replacing the narrow host range tail fibres of VHML with those of the VhCCS-06 phage, silencing of the DAM to produce a lytic phage, and silencing of the putative ADP-ribosylating toxin carried by VHML, a GMP chimeric phage with the potential to be highly virulent to members of the Harveyi clade may be produced. This could be used to treat vibriosis and related disease caused by this group of pathogens.

DAM in Vibrio and Bacteriophage

The DAM gene is found in both bacteria and phage alike. A search of the National Centre for Biotechnology Information's (NCBI) Gene database shows a multitude of DAM genes found across a wide range of phage and bacteria. The phages are of particular interest, as these show varied life cycles. Table 2 shows results of an NCBI search of DAM in Vibrio spp. and their phage. Almost all the Vibrio spp. with complete genome sequences contain the DAM gene, while most phage associated with Vibrio spp. also contained the DAM gene. This raises the question of why phage and bacteria both need the DAM gene. It is possible that the phage use their DAM gene to methylate its own genome to evade the restriction enzymes of their host's immune system and that DAM plays an important role in gene regulation possibly relating to virulence. Studies showed that infection of VHML into avirulent strains of V. harveyi and V. campbellii causes increased virulence (Munro et al. 2003) and changed the host phenotype (Vidgen et al. 2006). It would therefore be interesting to see whether there is a correlation between DAM in phage and the dominant life cycle (e.g. lytic vs lysogenic) that it takes upon infection of its host. Likewise, a correlation analysis of the DAM in phage and the pathogenicity of the host may also reveal a relationship between these two factors.


The consensus within the literature is that DAM in bacteria is a regulatory enzyme that can affect the affinity of transcription and translation of genes. Specific examples include the critical role in the viability and pathogenicity of both V. cholerae and Y. pseudotuberculosis. However, it is clear that the literature is lacking when it comes to the role DAM plays within bacteriophage. Although phage genomes have been sequenced and DAM has been found, the specific role it plays in gene regulation of the phage is unclear. Experiments are needed to test the role of DAM in phage, to better understand its role as a life cycle regulator and virulence regulator. Does the DAM of the phage control host gene regulation or does the host DAM regulate the phage gene expression? Both V. harveyi ATCC BAA-1116 (now V. campbellii) and VHML carry the DAM gene. Targeted gene knockout experiments using a suicide plasmid on a temperate prophage such as VHML may yield results that help piece together the role DAM plays in phage.

Conflict of interest

There are no conflicts of interest.