Chlorella Virus Research Articles

Evidence for a novel Chlorella virus-encoded alginate lyase

To clone the genes encoding lysis protein from a Chlorella virus, water samples were collected from 13 aquatic environments located in the Kanto area of Japan. Eight water samples contained plaque-forming viruses on Chlorella sp. NC64A, but no virus was detected in the other five samples. A novel Chlorella virus, CVN1, was isolated from the Inba-numa marsh sample. CVN1 genomic DNA was partially digested and shotgun cloned into pUC118 to identify the genomic region responsible for the lytic phenotype on Chlorella sp. NC64A. A DNA fragment which encoded two ORFs, ORF1 and ORF2, was obtained by antialgal assay. The ORF2 gene product, CL2, consisted of 333 amino acids showing antialgal activity not only on the original host of Chlorella sp. NC64A, but also on the heterogeneous hosts of Chlorella vulgaris C-27 and C. vulgaris C-207. CL2 showed a weak homology (19.8% amino acid identity) to mannuronate lyase SP2 from Turbo cornutus. CL2 in Escherichia coli cells was purified using a nickel chelate column. Lyase activity of purified CL2 on alginic acid was observed in an enzyme assay. The specific activity of purified CL2 was 2.1×10 −2 U mg −1, the optimum pH for enzymatic activity was 10.5, and Ca 2+ was required for enzyme activity. This is the first report of a Chlorella virus protein with lyase activity.

Mutational analysis of Escherichia coli DNA ligase identifies amino acids required for nick-ligation in vitro and for in vivo complementation of the growth of yeast cells deleted for CDC9 and LIG4.

We report that the NAD-dependent Escherichia coli DNA ligase can support the growth of Saccharomyces cerevisiae strains deleted singly for CDC9 or doubly for CDC9 plus LIG4. Alanine-scanning mutagenesis of E.coli DNA ligase led to the identification of seven amino acids (Lys115, Asp117, Asp285, Lys314, Cys408, Cys411 and Cys432) that are essential for nick-joining in vitro and for in vivo complementation in yeast. The K314A mutation uniquely resulted in accumulation of the DNA-adenylate intermediate. Alanine substitutions at five other positions (Glu113, Tyr225, Gln318, Glu319 and Cys426) did not affect in vivo complementation and had either no effect or only a modest effect on nick-joining in vitro. The E113A and Y225A mutations increased the apparent K (m)for NAD (to 45 and 76 microM, respectively) over that of the wild-type E. coli ligase (3 microM). These results are discussed in light of available structural data on the adenylylation domains of ATP- and NAD-dependent ligases. We observed that yeast cells containing only the 298-amino acid Chlorella virus DNA ligase (a 'minimal' eukaryotic ATP-dependent ligase consisting only of the catalytic core domain) are relatively proficient in the repair of DNA damage induced by UV irradiation or treatment with MMS, whereas cells containing only E.coli ligase are defective in DNA repair. This suggests that the structural domains unique to yeast Cdc9p are not essential for mitotic growth, but may facilitate DNA repair.

Giant viruses infecting algae.

Paramecium bursaria chlorella virus (PBCV-1) is the prototype of a family of large, icosahedral, plaque-forming, double-stranded-DNA-containing viruses that replicate in certain unicellular, eukaryotic chlorella-like green algae. DNA sequence analysis of its 330, 742-bp genome leads to the prediction that this phycodnavirus has 376 protein-encoding genes and 10 transfer RNA genes. The predicted gene products of approximately 40% of these genes resemble proteins of known function. The chlorella viruses have other features that distinguish them from most viruses, in addition to their large genome size. These features include the following: (a) The viruses encode multiple DNA methyltransferases and DNA site-specific endonucleases; (b) PBCV-1 encodes at least part, if not the entire machinery to glycosylate its proteins; (c) PBCV-1 has at least two types of introns--a self-splicing intron in a transcription factor-like gene and a splicesomal processed type of intron in its DNA polymerase gene. Unlike the chlorella viruses, large double-stranded-DNA-containing viruses that infect marine, filamentous brown algae have a circular genome and a lysogenic phase in their life cycle.

Chlorella Virus PBCV-1 Encodes a Functional Homospermidine Synthase

Sequence analysis of the 330-kb genome of chlorella virus Paramecium bursaria chlorella virus 1 (PBCV-1) revealed an open reading frame, A237R, that encodes a protein with 34% amino acid identity to homospermidine synthase from Rhodopseudomonas viridis. Expression of the a237r gene product in Escherichia coli established that the recombinant enzyme catalyzes the NAD+-dependent formation of homospermidine from two molecules of putrescine. The a237r gene is expressed late in PBCV-1 infection. Both uninfected and PBCV-1-infected chlorella, as well as PBCV-1 virions, contain homospermidine, along with the more common polyamines putrescine, spermidine, and cadaverine. The total number of polyamine molecules per virion (∼539) is too small to significantly neutralize the virus double-stranded DNA (>660,000 nucleotides). Consequently, the biological significance of the homospermidine synthase gene is unknown. However, the gene is widespread among the chlorella viruses. To our knowledge, this is the first report of a virus encoding an enzyme involved in polyamine biosynthesis.

Aminoacylation of tRNAs Encoded by Chlorella Virus CVK2

Viruses that infect certain strains of the unicellular green alga, Chlorella, have a large, linear dsDNA genome that is 330–380 kb in size; this genomic size is the largest known among viruses and is equivalent to ∼60% of the smallest prokaryotic genome of Mycoplasma genitalium (580 kb). Besides many putative protein-coding genes, a cluster of 10–15 tRNA genes is present in these viral genomes. Some of these tRNA genes contain peculiar insertions. In infected host cells, the viral tRNAs of CVK2, a Chlorella virus isolate, have been demonstrated to be cotranscribed as a large precursor, ∼1.0 kb in size, that is precisely processed into individual mature tRNA species. Acidic Northern blot analysis of eight of these tRNAs has revealed that they are actually aminoacylated in vivo, indicating their involvement in viral protein synthesis. They may help the virus reach maximal replication potential by overcoming codon usage barriers that exist between the virus and its host. These results provide evidence that some components of the host protein synthesis machinery can be replaced by viral gene products. This is the first report of tRNA aminoacylation encoded by viruses of eukaryotes.

Characterization of Two Chitinase Genes and One Chitosanase Gene Encoded by Chlorella Virus PBCV-1

Chlorella virus PBCV-1 encodes two putative chitinase genes, a181/182r and a260r, and one chitosanase gene, a292l. The three genes were cloned and expressed in Escherichia coli. The recombinant A181/182R protein has endochitinase activity, recombinant A260R has both endochitinase and exochitinase activities, and recombinant A292L has chitosanase activity. Transcription of a181/182r, a260r, and a292l genes begins at 30, 60, and 60 min p.i., respectively; transcription of all three genes continues until the cells lyse. A181/182R, A260R, and A292L proteins are first detected by Western blots at 60, 90, and 120 min p.i., respectively. Therefore, a181/182r is an early gene and a260r and a292l are late genes. All three genes are widespread in chlorella viruses. Phylogenetic analyses indicate that the ancestral condition of the a181/182r gene arose from the most recent common ancestor of a gene found in tobacco, whereas the genealogical position of the a260r gene could not be unambiguously resolved.

Expression of a Chitinase Gene and Lysis of the Host Cell Wall during Chlorella Virus CVK2 Infection

A chitinase gene (vChti-1) encoded by the Chlorella virus CVK2 was cloned and characterized. The vChti-1 open reading frame consisted of 2508 bp corresponding to 836 amino acid residues. The predicted amino acid sequence contained two sets of a family 18 catalytic domain that is responsible for chitinase activity. Northern blot analysis revealed that the vChti-1 gene was expressed in virus-infected Chlorella cells late in infection, when a single transcript of about 2.5 kb appeared at 120 min postinfection. This result was confirmed by Western blotting with a specific anti-vChti-1 protein antibody; a protein of about 94 kDa was detected specifically beginning at 240 min postinfection and was present until cell lysis. The protein was not incorporated into viral particles but remained in the medium after cell lysis. The vChti-1 protein produced in virus-infected cells showed chitinase activity on zymogram assays.

Footprinting of Chlorella Virus DNA Ligase Bound at a Nick in Duplex DNA

The 298-amino acid ATP-dependent DNA ligase of Chlorella virus PBCV-1 is the smallest eukaryotic DNA ligase known. The enzyme has intrinsic specificity for binding to nicked duplex DNA. To delineate the ligase-DNA interface, we have footprinted the enzyme binding site on DNA and the DNA binding site on ligase. The size of the exonuclease III footprint of ligase bound a single nick in duplex DNA is 19-21 nucleotides. The footprint is asymmetric, extending 8-9 nucleotides on the 3'-OH side of the nick and 11-12 nucleotides on the 5'-phosphate side. The 5'-phosphate moiety is essential for the binding of Chlorella virus ligase to nicked DNA. Here we show that the 3'-OH moiety is not required for nick recognition. The Chlorella virus ligase binds to a nicked ligand containing 2',3'-dideoxy and 5'-phosphate termini, but cannot catalyze adenylation of the 5'-end. Hence, the 3'-OH is important for step 2 chemistry even though it is not itself chemically transformed during DNA-adenylate formation. A 2'-OH cannot substitute for the essential 3'-OH in adenylation at a nick or even in strand closure at a preadenylated nick. The protein side of the ligase-DNA interface was probed by limited proteolysis of ligase with trypsin and chymotrypsin in the presence and absence of nicked DNA. Protease accessible sites are clustered within a short segment from amino acids 210-225 located distal to conserved motif V. The ligase is protected from proteolysis by nicked DNA. Protease cleavage of the native enzyme prior to DNA addition results in loss of DNA binding. These results suggest a bipartite domain structure in which the interdomain segment either comprises part of the DNA binding site or undergoes a conformational change upon DNA binding. The domain structure of Chlorella virus ligase inferred from the solution experiments is consistent with the structure of T7 DNA ligase determined by x-ray crystallography.

Hyaluronan Synthesis in Virus PBCV-1-Infected Chlorella-like Green Algae

We previously reported that the chlorella virus PBCV-1 genome encodes an authentic, membrane-associated glycosyltransferase, hyaluronan synthase (HAS). Hyaluronan, a linear polysaccharide chain composed of alternating β1,4-glucuronic acid and β1,3-N-acetylglucosamine groups, is present in vertebrates as well as a few pathogenic bacteria. Studies of infected cells show that the transcription of the PBCV-1hasgene begins within 10 min of virus infection and ends at 60–90 min postinfection. The hyaluronan polysaccharide begins to accumulate as hyaluronan-lyase sensitive, hair-like fibers on the outside of the chlorella cell wall by 15–30 min postinfection; by 240 min postinfection, the infected cells are coated with a dense fibrous network. This hyaluronan slightly reduces attachment of a second chlorella virus to the infected algae. An analysis of 41 additional chlorella viruses indicates that many, but not all, produce hyaluronan during infection.

Several genes in Chlorella virus strain CVG-1 encode putative virion components.

We have started to characterize the capsid components of European Chlorella virus isolate CVG-1, a member of the Pbi subgroup of the Phycodnaviridae. The major coat protein, Vp49, was biochemically characterized and the amino acid sequence of the N terminus was determined. Subsequently, the corresponding gene was isolated from CVG-1 genomic DNA. Sequence data were compared to those available from PBCV-1 and other Chlorella virus isolates representing the NC64A subgroup of the Phycodnaviridae. The major coat proteins of all strains are homologous and similar in size, but apparently differ in their degree of glycosylation. Like PBCV-1, the major coat protein of CVG-1 is part of a gene family, as two open reading frames with high similarity to Vp49 were also isolated and characterized in this study. The predicted amino acid sequences of the CVG-1 and PBCV-1 virus genes examined show, with one exception, a divergence of about 25%. Taking into account that corresponding genes of NC64A viruses are almost identical, this divergence supports the original placement of the NC64A and Pbi viruses into separate subgroups of the Phycodnaviridae.

The Catalytic Mechanism of a Pyrimidine Dimer-specific Glycosylase (pdg)/Abasic Lyase, chlorella virus-pdg

The repair of UV light-induced cyclobutane pyrimidine dimers can proceed via the base excision repair pathway, in which the initial step is catalyzed by DNA glycosylase/abasic (AP) lyases. The prototypical enzyme studied for this pathway is endonuclease V from the bacteriophage T4 (T4 bacteriophage pyrimidine dimer glycosylase (T4-pdg)). The first homologue for T4-pdg has been found in a strain of Chlorella virus (strain Paramecium bursaria Chlorella virus-1), which contains a gene that predicts an amino acid sequence homology of 41% with T4-pdg. Because both the structure and critical catalytic residues are known for T4-pdg, homology modeling of the Chlorella virus pyrimidine dimer glycosylase (cv-pdg) predicted that a conserved glutamic acid residue (Glu-23) would be important for catalysis at pyrimidine dimers and abasic sites. Site-directed mutations were constructed at Glu-23 to assess the necessity of a negatively charged residue at that position (Gln-23) and the importance of the length of the negatively charged side chain (Asp-23). E23Q lost glycosylase activity completely but retained low levels of AP lyase activity. In contrast, E23D retained near wild type glycosylase and AP lyase activities on cis-syn dimers but completely lost its activity on the trans-syn II dimer, which is very efficiently cleaved by the wild type cv-pdg. As has been shown for other glyscosylases, the wild type cv-pdg catalyzes the cleavage at dimers or AP sites via formation of an imino intermediate, as evidenced by the ability of the enzyme to be covalently trapped on substrate DNA when the reactions are carried out in the presence of a strong reducing agent; in contrast, E23D was very poorly trapped on cis-syn dimers but was readily trapped on DNA containing AP sites. It is proposed that Glu-23 protonates the sugar ring, so that the imino intermediate can be formed.

Mutational analysis of Chlorella virus DNA ligase: catalytic roles of domain I and motif VI.

A conserved catalytic core of the ATP-dependent DNA ligases is composed of an N-terminal domain (domain 1, containing nucleotidyl transferase motifs I, III, IIIa and IV) and a C-terminal domain (domain 2, containing motif VI) with an intervening cleft. Motif V links the two structural domains. Deletion analysis of the 298 amino acid Chlorella virus DNA ligase indicates that motif VI plays a critical role in the reaction of ligase with ATP to form ligase-adenylate, but is dispensable for the two subsequent steps in the ligation pathway; DNA-adenylate formation and strand closure. We find that formation of a phosphodiester at a pre-adenylated nick is subject to a rate limiting step that does not apply during the sealing of nicked DNA by ligase-adenylate. This step, presumably conformational, is accelerated or circumvented by deleting five amino acids of motif VI. The motif I lysine nucleophile (Lys27) is not required for strand closure by wild-type ligase, but this residue enhances the closure rate by a factor of 16 when motif VI is truncated. We find that a more extensively truncated ligase consisting of only N-terminal domain 1 and motif V is inert in ligase--adenylate formation, but competent to catalyze strand closure at a pre-adenylated nick. These results suggest that different enzymic catalysts facilitate the three steps of the DNA ligase reaction.

Specificity and fidelity of strand joining by Chlorella virus DNA ligase.

Chlorella virus PBCV-1 DNA ligase seals nicked duplex DNA substrates consisting of a 5'-phosphate-terminated strand and a 3'-hydroxyl-terminated strand annealed to a bridging template strand, but cannot ligate a nicked duplex composed of two DNAs annealed on an RNA template. Whereas PBCV-1 ligase efficiently joins a 3'-OH RNA to a 5'-phosphate DNA, it is unable to join a 3'-OH DNA to a 5'-phosphate RNA. The ligase discriminates at the substrate binding step between nicked duplexes containing 5'-phosphate DNA versus 5'-phosphate RNA strands. PBCV-1 ligase readily seals a nicked duplex DNA containing a single ribonucleotide substitution at the reactive 5'-phosphate end. These results suggest a requirement for a B-form helical conformation of the polynucleotide on the 5'-phosphate side of the nick. Single base mismatches at the nick exert disparate effects on DNA ligation efficiency. PBCV-1 ligase tolerates mismatches involving the 5'-phosphate nucleotide, with the exception of 5'-A:G and 5'-G:A mispairs, which reduce ligase activity by two orders of magnitude. Inhibitory configurations at the 3'-OH nucleotide include 3'-G:A, 3'-G:T, 3'-T:T, 3'-A:G, 3'-G:G, 3'-A:C and 3'-C:C. Our findings indicate that Chlorella virus DNA ligase has the potential to affect genome integrity by embedding ribonucleotides in viral DNA and by sealing nicked molecules with mispaired ends, thereby generating missense mutations.

Cryo-Electron Microscopy and Image Reconstruction of PBCV-1, an ALGAL Virus with T=169, Icosahedral Lattice Symmetry

Abstract Paramecium bursaria Chlorella virus l(PBCV-l), the type species of the Pycodnaviridae family, is a large, plaque-forming virus that replicates in certain unicellular, exsymbiotic, Chlorella-like green algae. The virion (∼lxl09 daltons) has a 330,740 base pair dsDNA genome encapsulated within an icosahedral shell of ∼190nm diameter. Purified virions contain more than 50 different proteins, which account for 65% of the total virion mass and range in size from 10 to more than 200kDa. Vp54, the major capsid protein (40% of total virion protein) is a myristilated glycoprotein but it is not phosphorylated. Three other proteins are located on the virus surface. Virions contain 5-10% lipid, located inside the glycoprotein shell, which is required for virus infectivity. Little is known about the detailed ultrastructure of PBCV-1 virions, though some information has been obtained with sectioned, negatively-stained, and metal-shadowed specimens. We have used cryo-electron microscopy and three-dimensional (3D) image reconstruction methods to examine the native morphology of PBCV-1.

Characterization of a Novel cis-synandtrans-syn-IIPyrimidine Dimer Glycosylase/AP Lyase from a Eukaryotic Algal Virus, Paramecium bursaria chlorellaVirus-1

Endonuclease V from bacteriophage T4, is a cis-syn pyrimidine dimer-specific glycosylase. Recently, the first sequence homolog of T4 endonuclease V was identified from chlorella virus Paramecium bursaria chlorella virus-1 (PBCV-1). Here we present the biochemical characterization of the chlorella virus pyrimidine dimer glycosylase, cv-PDG. Interestingly, cv-PDG is specific not only for the cis-syn cyclobutane pyrimidine dimer, but also for the trans-syn-II isomer. This is the first trans-syn-II-specific glycosylase identified to date. Kinetic analysis demonstrates that DNAs containing both types of pyrimidine dimers are cleaved by the enzyme with similar catalytic efficiencies. Cleavage analysis and covalent trapping experiments demonstrate that the enzyme mechanism is consistent with the model proposed for glycosylase/AP lyase enzymes in which the glycosylase action is mediated via an imino intermediate between the C1' of the sugar and an amino group in the enzyme, followed by a beta-elimination reaction resulting in cleavage of the phosphodiester bond. cv-PDG exhibits processive cleavage kinetics which are diminished at salt concentrations greater than those determined for T4 endonuclease V, indicating a possibly stronger electrostatic attraction between enzyme and DNA. The identification of this new enzyme with broader pyrimidine dimer specificity raises the intriguing possibility that there may be other T4 endonuclease V-like enzymes with specificity toward other DNA photoproducts.

Chlorella virus DNA ligase: nick recognition and mutational analysis.

Chlorella virus PBCV-1 DNA ligase seals nicked DNA substrates consisting of a 5'-phosphate-terminated strand and a 3'-hydroxyl-terminated strand annealed to a bridging DNA template strand. The enzyme discriminates at the DNA binding step between substrates containing a 5'-phosphate versus a 5'-hydroxyl at the nick. Mutational analysis of the active site motif KxDGxR (residues 27-32) illuminates essential roles for the conserved Lys, Asp and Arg moieties at different steps of the ligase reaction. Mutant K27A is unable to form the covalent ligase-(Lys-straightepsilonN-P)-adenylate intermediate and hence cannot activate a nicked DNA substrate via formation of the DNA-adenylate intermediate. Nonetheless, K27A catalyzes phosphodiester bond formation at a pre-adenylated nick. This shows that the active site lysine is not required for the strand closure reaction. K27A binds to nicked DNA-adenylate, but not to a standard DNA nick. This suggests that occupancy of the AMP binding pocket of DNA ligase is important for nick recognition. Mutant D29A is active in enzyme-adenylate formation and binds readily to nicked DNA, but is inert in DNA-adenylate formation. R32A is unable to catalyze any of the three reactions of the ligation pathway and does not bind to nicked DNA.

Viruses infecting marine brown algae.

Viruses infecting algal hosts possess large double-stranded DNA as genomes. We have recently identified a family of viruses specific for filamentous brown algae. In contrast to the better known Chlorella viruses with their lytic infection cycle, marine brown algal viruses latently occur in their host cells and are induced to multiply in response to a variety of external stimuli such as change in light and temperature. Here, I summarize the known properties of this family of viruses and discuss their taxonomic classification.

Chlorella virus DNA ligase: Nick recognition and mutational analysis

Chlorella virus PBCV-1 DNA ligase seals nicked DNA substrates consisting of a 5′-phosphate-terminated strand and a 3′-hydroxyl-terminated strand annealed to a bridging DNA template strand. The enzyme discriminates at the DNA binding step between substrates containing a 5′-phosphate versus a 5′-hydroxyl at the nick. Mutational analysis of the active site motif KxDGxR (residues 27–32) illuminates essential roles for the conserved Lys, Asp and Arg moieties at different steps of the ligase reaction. Mutant K27A is unable to form the covalent ligase-(Lys-ɛN-P)-adenylate intermediate and hence cannot activate a nicked DNA substrate via formation of the DNA-adenylate intermediate. Nonetheless, K27A catalyzes phosphodiester bond formation at a pre-adenylated nick. This shows that the active site lysine is not required for the strand closure reaction. K27A binds to nicked DNA-adenylate, but not to a standard DNA nick. This suggests that occupancy of the AMP binding pocket of DNA ligase is important for nick recognition. Mutant D29A is active in enzyme-adenylate formation and binds readily to nicked DNA, but is inert in DNA-adenylate formation. R32A is unable to catalyze any of the three reactions of the ligation pathway and does not bind to nicked DNA.

Evolution of viral DNA-dependent DNA polymerases.

DNA viruses as their host cells require a DNA-dependent DNA polymerase (Pol) to faithfully replicate their genomic information. Large eukaryotic DNA viruses as well as bacterial viruses encode a specific Pol equipped with a proofreading 3'-5'-exonuclease, and other replication proteins. All known viral Pol belong to family A and family B Pol. Common to all viral Pol is the conservation of the 3'-5'-exonuclease domain manifested by the three sequence motifs Exo I, Exo II, and Exo III. The polymerase domain of family A and B Pol is clearly distinguishable. Family A Pol share 9 distinct consensus sequences, only two of them are convincingly homologous to sequence motif B of family B Pol. The putative sequence motifs A, B, and C of the polymerase domain are located near the C-terminus in family A Pol and more central in family B Pol. Thus, family A Pol show a significant greater spacing between the Exo III motif and the Pol motif A that is especially extended in the case of the mitochondrial Pol gamma. From each host and virus family whenever possible the consensus sequences of two distantly related polymerase species were aligned for assessment of phylogenetic trees, using both maximum parsimony and distance methods, and evaluated by bootstrap analysis. Three alternative methods yielded trees with identical major groupings. A subdivision of viral family B Pol was achieved resulting in a branch with Pol carrying out a protein-primed mechanism of DNA replication, including adenoviruses, bacteriophages and linear plasmids of plant and fungal origin. Archaebacterial Pol and cellular Pol epsilon were consistently found at the base of this branch. Another major branch comprised alpha- and delta-like viral Pol from mammalian herpesviruses, fish lymphocystis disease virus, insect ascovirus, and chlorella virus. Due to a lower branch integrity Pol of T-even bacteriophages, poxviruses, African swine fever virus, fish herpesvirus, and baculoviruses were not clearly resolved and placed in alternate groupings. A composite and rooted tree of family A and B Pol shows that viral Pol with a protein-priming requirement represent the oldest viral Pol species suggesting that the protein-primed mechanism is one of the earliest modes of viral DNA replication.

Chlorella Viruses Encode Multiple DNA Methyltransferases

The >320 kb dsDNA genomes of 16 viruses which infect Chlorella strain NC64A and 5 viruses infecting Chlorella strain Pbi were tested for their sensitivity/resistance to more than 80 DNA restriction endonucleases. From the known methylation sensitivities of these enzymes to site-specific 5-methylcytosine and N6-methyladenine DNA modifications, we deduce that the 16 NC64A viruses encode at least 13 different sequence-specific DNA methyltransferases and the 5 Pbi viruses encode at least 7 sequence-specific DNA methyltransferases. Each DNA methyltransferase has a 2 to 4 base pair DNA recognition sequence. Some individual viruses encode as many as ten different DNA methyltransferases, making these chlorella virus genomes among the most concentrated sources of DNA methyltransferase genes known.