THEJOURNAL OF BIOLOGICAL CHEMISTRY Vol. 257, No. 5, Issue of March 10. pp. 2605-2612, 1982 Printed in U.S.A.

Escherichia coli dam Methylase PHYSICAL AND CATALYTIC PROPERTIES OF THE HOMOGENEOUS ENZYME* (Received for publication, August 19, 1981)

Gail E. Herman4 and Paul Modrich8 From the Department of Biochemistry, Duke University Medical Center, Durham, North Carolina 27710

The Escherichia colidam methylase has been purified sible functions of dam methylation have beensuggested. 3000-fold to a purity of 95% from a clone which over- Meselson et al. have provided evidence that methyl groups produces the enzyme 10- t o 20-fold. Physical properties introduced by the enzyme function in strand discrimination of enzyme purified fromthe overproducing clone were during posreplication mismatch repair (10). Further, identifiidentical with those of enzyme previously obtained cation of 18 dam recognition sites within the E . coli origin of from a non-overproducing E. coli strain (Geier, G. E., DNA replication (11,12) and conservation of 17 of these sites and Modrich, P. (1979) J. BioL C h m . 254, 1408-1413). in the Salmonella typhimurium origin (13) has suggested a The methylase is comprised of a single polypeptide rolefor methylation of this enzymeinorigin function or chain of M, = 31,000, has an s20,w of 2.8 S, a Stokes maintenance. radius of 24 A, and exists in solution as a monomer. Its Although some information has accumulated concerning aggregation state is not affected by the presence of S- the biology of DNA methylation (14-17), little is known about adenosyl-L-methionine. The simple kinetic behavior of the methylase indi- structures and mechanisms of enzymes involved.To date, the cates that it functions as a monomer. Initial rates of only DNA methylases that have been obtained in pure form have been several modification enzymes which conferresistmethyl transfer are first order in enzyme concentration, and Michaelis-Menten behavior is obeyed with ance to certain restriction endonucleases. These enzymes aprespect to bothsubstrates. A t 37 “C, in the presence of pear to fall into two classes (16, 17). Type I DNA methylases, saturating DNA, the enzyme has a turnover number of as typified by Eco B and Eco K enzymes,’ are large multi19 methyl transfers/min with a K, for S-adenosyl-L- subunit enzymes which have been isolated in two forms (17). Type I complexes comprisedof polypeptide products of hsds methionine of 12.2 p ~At. half-saturatingS-adenosyl-Lmethylase activity while methionine, the apparent K,,, for d(G-A-T-C) sites in and hsdm genes possess modification ColEl DNA is 3.0 m.The mechanism of methyl transfercomplexes also containing the hsdR product possess both is also consistent with the monomer being the func- modification and restriction activity (17,19-21). Methyl transtional form of the enzyme. Studies with G4 RFI DNA fer to unmodified DNA by Type I enzymes requires only (two d(G-A-T-C) sites) indicate that the methylase AdoMet; although this reaction is extraordinarily slow (19). transfers 1methyl group t o a recognition site and then The methyl transfer reaction is at least 70 times faster on dissociates fromthis DNA prior to subsequent cataly- DNA methylated on one strand of the duplex (hemimethsis. It appears that kinetic parameters for methyl trans- ylated DNA), and this rate is enhanced another 10-fold by fer to sites already modified on one DNA strand may ATP and M$+ (21). be slightly more favorable than those for transfer to In contrast, Type I1 modification methylases are relatively sites in whichboth strands are unmethylated. simple with respect to both cofactor requirements and structure. AdoMet is the only cofactor required for methyl transfer by Type I1 modifkation enzymes (16). Furthermore, the Eco The dam methylase is responsible for the majority of meth- RI and Hpa I1 enzymes, which have been obtained in pure ylated adenine residues in Escherichia coli DNA (1, 2). The form, have been shown to be comprised of single polypeptide enzyme recognizesthe symmetric tetranucleotide d(G-A-T-C) chains and to function as themonomer (22,23). In thecase of and introduces 2 methyl groups/duplex site, with the product Eco RI methylase it has also been demonstrated that the of methylation being 6-methylaminopurine (3). Mutants of E. enzyme transfers a single methyl group per DNAbinding coli deficient in dam methylase activity in vitro,and contain- event and that methyl transfer to hemimethylated DNA is ing markedly reduced levels of 6-methylaminopurine in their not more favorable than transfer to unmodified sites (22). In view of the biological role of the dam methylase and in DNA in uiuo, show a variety of phenotypic alterations. These include increased spontaneous mutability (2, 4-6), increased order to compare its mechanism of methyl transfer with those sensitivity to some chemical mutagens including the base of enzymes discussedabove, we have undertaken isolation and analogs 2-aminopurine and 5-bromouracil (5),increased levels characterization of the dam methylase of E. coli K12.We of genetic recombination (7, 8), and lethality in conjunction have previously purifieddam methylase from a cytosine methwith non-lethal recA or recBC mutations (2, 4, 9). Two pos- ylase-deficient (dcm-) strain of E . coli and have determined * This work was supported by Grant GM 23719 from the National the recognition sequence of the enzyme in duplex DNA (3). Institues of Health and Grant PCM 7823036 from the National However, analysis of the structure and mechanism of the Science Foundation. The costs of publication of this article were protein was hampered by the extremely low levels of enzyme defrayed in part by the payment of page charges. This article must present in wild type E. coli. In order to obtain larger quantities therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Supported by Training Grant GM 07184. 9 Recipient of Development Award CA 00495 from the National Cancer Institute.

’ Restriction endonucleases and modification methylases are designated as proposed by Smith and Nathans (18). The abbreviations used are: AdoMet, S-adenosyl-L-methionine: SDS, sodium dodecyl sulfate; KPO, potassium phosphate.

2605

’

2606

E. coli dam Methylase

ATP, 10 m~ dithiothreitol, 2 mMMgC12,25 pg/ml of each DNA, and 5 units/ml of T4 DNA ligase (31). DNA was heated at 37 "C for 15 min and quick chilled on ice prior to addition of ligase. In some experiments, after Pst I cleavage, the vector pBR322 wastreated with bacterial alkaline phosphatase (32,36) for 60 min at 60 "C in the presence of 0.1 m~ ZnCb. Phosphatase-treated DNA was phenolextracted and dialyzed versus three changes of 0.02 M Tris/HCl (pH 7.6), 0.05 M NaCl, 1 m~ EDTA prior to ligation. Ligase reactions were used directly to transform a d a m 4 strain EXPERIMENTAL PROCEDURES (SK1036), with selection for resistance to tetracycline or ampicillin (37). Transformants were screened for resistance to 2-aminopurine Materials Bacterial Strains and Bacteriophage-E. coli K12 strains SKI031 (51, and extracts of base analog-resistant clones were assayed for dam (F- mtl-1 xyl-7 argHl his-4 ilvD188 lacMS286 @3OdIIlacBK1 metE46 methylase activity (3). Plasmid DNA (pGG503) from one such clone thi spcRtsx-3 SU- dam+) and SKt036 (asSK1031 but containing the was isolated and transformed into E. coli K12 hosts GM31 (dcm-6) dum-4 allele of E. B. Konrad (7)) were obtained from S. Kushner, and JC4583 (endA). Enzyme Assays-The dam methylase assay has been described (3) University of Georgia, Athens. E. coli K12 strain GM31 (dcm-6) has been described (26).E. coli K12 strain RS5033 (Hfr Hayes metBl rel- and measures transfer of r3H]methyl groups from AdoMet to calf 1 str-100 azi7 lac MS286 ~OdIIZucBK1 thidam-4) was obtained thymus DNA. One unit of dam methylase is defined as that which from E. B. Konrad (7). E. coli K12 strain JC4583 (F- endA gal thi catalyzes conversion of 1 pmol of r3H]methyl groups into a form thy lac SUII)was obtained from Dr. A. J . Clarke, University of which binds to DE81 in 30 min at 37 "C. SI nuclease assays were California, Berkeley. E. coli C strain CGHl was prepared by P1 vir performed as recommended by the supplier. Nonspecific nuclease transduction of the dam-4 allele from RS5033 into strain C-10 (27). assays (100pl) contained 0.02 M Tris/HCl (pH 8.0), 0.1 M NaCl, 5 m~ MgC12,400 pg/ml of bovine serum albumin, 5 pg/ml of RNase (50 pg/ Transductants were selected for streptomycin resistance (28) and screened for 2-aminopurine sensitivity (5). E. coli K12 strain JA200 ml in crude extracts), and 20 pm of T7 r3H]DNA or 150 p~ pBR322 (F+ AtrpE5recA thr leu lac) containing a Clarke-Carbon hybrid DNA. RNase A was included in nuclease assays to eliminate any inhibition by contaminating tRNA since the major nuclease contamplasmid pLC13-42(24)was provided by J. Carbon, University of inant was endonuclease I (38). After incubation at 37 "C for 30 min, California, Santa Barbara. Bacteriophage G4 (29) was obtained from reactions with T7 DNA were terminated by addition of carrier DNA J . Scott, Stanford University, Stanford, CA. and trichloroacetic acid-soluble radioactivity determined (39). ReacEnzymes and Proteins-The dam methylase purified from the E. tions with pBR322 DNA were electrophoresed on 1% agarose gels coli dcm strain GM31 (26) wasthe preparation previously described using a Tris-borate buffer system (40) and visualized as described (3) and is referred to here as wild type enzyme. Dpn I endonuclease (30). (3),Eco RIendonuclease (30),and T4DNA ligase (31)were prepared Purification of the dam Methylase-dam methylase was purified as described. Bacterial alkaline phosphatase (Worthington BAPC) from the E.coli K12 strain GM31/pGG503. The method is a modifwas further purified according to themethod of Weiss et al. (32).Bum cation of that used previously (3).A summary of the purification from HI endonuclease, Sal I endonuclease, and SI nuclease were from 1.0 kg of cell paste is presented in Table I. AU steps were performed Bethesda Research Laboratories and Pst I endonuclease from New at 4 "C and centrifugation was at 12,000 X g for 10 to 30 min in a England Biolabs. Egg white lysozyme, catalase, ovalbumin, and chy- Sorvall GS3 rotor. motrypsinogen A were fromworthington. Horse heart myoglobin, After thawing overnight at 4 "C, cells were suspended in 3 liters of bovinehemoglobin, horse heart cytochrome c, and yeast alcohol 0.05 M Tris/HCl (pH 7.6), 10 m~ 2-mercaptoethanol, and disrupted dehydrogenase were from Sigma. RNase A (Worthington) was heated in 250- to 300-ml portions with five I-min pulses at 85 to 95 watts for 10 min at 80 "C before use. using a Branson sonifier. Temperature was maintained below 7 "C. Nucleic Acids-Bacteriophage T7 DNA, T7 r3H]DNA(10,000 After clarification by centrifugation, the extract was diluted with the cpm/nmol of nucleotide), and unmethylated ColEl DNA (>95% above buffer to yield an A ~ M of)280 (Fraction I, 4300 m l ) . covalently closed circles) were prepared as previously described (30). To fraction I were added 350 ml of a 5% (v/v) solution of Polymin Bacteriophage G4 RFI r3'P]DNA (29) (>95%covalently closed circles, P. After stirring for 15 min,the precipitate was removed by centrifu2600 cpm/nmol of nucleotide) lacking the dam modification was gation and discarded. An additional 560 ml of Polymin P were then isolated as above after infection of E. coli strain CGHl (33). Plasmid added to the supernatant, and after centrifugation, the pellet was pBR322(25)was obtained from D. Stafford, University of North extracted with 1600 ml of 0.02 M KPO4 (pH 7.4), 0.3 M KCl, 1 mM Carolina, Chapel Hill. Plasmid pLC13-42 DNA was isolated as above. EDTA, 10 m~ 2-mercaptoethanol, 5% (w/v) glycerol. This eluate was OtherMaterial~-(~~P)Inorganic phosphate (carrier-free), [methyl- clarified by centrifugation and diluted by addition of 3200 ml of 0.02 3H]AdoMet (>IO Ci/mmol), and Aquasol 2 were from New England M KPO, (pH 7.4), 10 m~ 2-mercaptoethanol, 5% glycerol. Alumina Nuclear. For kinetic analyses and G4 RFI methylation experiments, Cy gel (350ml, 4.7% solids) was added, the solution stirred for 20 min, AdoMet was purified on prewashed Whatman No. 3MM paper using and insoluble material collected by centrifugation. The pellet was I-butanol/t.O N HCl/ethanol (505020) or I-butanol/acetic acid/waextracted sequentially with 1600-ml portions of 0.2 M and 0.8 M KPO4 ter (601520) to >97% purity, and concentration determined by iso(pH 7.4), containing 10 m~ 2-mercaptoethanol and 5% glycerol, with tope dilution. E. coli tRNA (Boehringer Mannheim, phenol-ex- activity eluting in the latter buffer (Fraction 11, 1600 ml). tracted) was used to prepare tRNA-agarose (3 mgof tRNA/ml of Fraction I1 was dialyzed versus 2 changes of25 liters of0.02 M agarose) as described (34). N-Ethylmaleimide was from Sigma. Other KPO, (pH 7.4), 0.12 M KC1, 1 m~ EDTA, 10 m~ 2-mercaptoethanol, materials have been described (3). 10% glycerol (2 h/change). After removal of insoluble material by centrifugation, the supernatant (1960 m l ) was diluted with 680 ml of Methods 0.02 M KPO, (pH 7.4), 1 m~ EDTA, 10 mM 2-mercaptoethanol, 10% Growth of Cells-T plates (28) (1.5%, w/v, agar) spread with 0.1 glycerol, and applied at 750 ml/h to a phosphocellulose column (14 cm X 22 cm2) equilibrated with 0.02 M KPO4 (pH 7.4), 0.2 M KCl, 1 ml of a 1:10 dilution of crude colicin E l (35) were used to test for m~ EDTA, 10 m~ 2-mercaptoethanol, 5% glycerol. The column was resistance to this bacteriocin. Selective T plates contained 10 p g / d of tetracycline (Sigma), 50 pg/ml of ampicillin (Wyeth), or 400 pg/ml washed with 600 ml of this buffer and eluted with a 3.0-liter linear of 2-aminopurine (Calbiochem). Liquid cultures of transformants gradient of KC1 (0.2 to 1.0 M) in 0.02 M KPO4 (pH 7.4), 1 mM EDTA, 10 mM 2-mercaptoethanol, 5% glycerol. Fractions containing dam were grownin L broth (28) without added glucose and contained 5-10 pg/ml of tetracycline. Cells to be utilized for enzyme purification were methylase activity, which eluted at approximately 0.60 M KC1, were pooled (Fraction 111, 200 ml). grown as described (3). Fraction I11 was diluted with 440 ml of 0.02 M KPO4 (pH 7.4), 1 mM Recombinant DNA Methods-Restriction endonuclease hydrolysis of plasmid DNA with commercial enzymes was performed as recom- EDTA, 10 m~ 2-mercaptoethanol, 10% glycerol and applied at 150 mended by the supplier, while Eco RI endonuclease reactions were as ml/h to a blue dextran-Sepharose column (28 cm X 2.9 cm2) equilibrated with 0.02 M KPO, (pH 7.4), 0.25 M NaC1, 5 mM EDTA, 10 mM described previously (30). Plasmid pBR322 and pLC13-42DNA's were each cleaved with restriction endonuclease Eco RI, Pst I, Bum 2-mercaptoethanol, 10% glycerol. The column was washed with 2 HI, or Sal I, and reactions heated for 15 min at 75 "C. Restriction liters of this buffer, and eluted with a 750-ml linear gradient of NaCl products were ligated at 10 "C overnight in reactions (204) containing (0.25 to 1.1 M ) in 0.02 M KPO, (pH 7.4), 5 m~ EDTA, 10 mM 20.02 M Tris/HCl (pH 7.6),0.5 mg/ml of bovine serum albumin, 66 mercaptoethanol, 10% glycerol.Active fractions, which eluted at

of enzyme, an overproducing strain has been constructed by transfer of the dum locus from a Clarke-Carbon plasmid (24) into pBR322 (25). The enzyme has been isolated from one overproducing strain and physical and catalytic properties of purified enzyme from overproducing and wild type strains of E. coli examined.

pM

2607

E. coli dam Methylase approximately 0.69 M NaCI, were pooled (Fraction IV, 67 ml). Fraction IV was dialyzed uersus 2 liters of 0.02 M KPOJ (pH 7.4), 0.12 M NaCI. 5 mM EDTA, 10 mM 2-mercaptoethanol, 10% glycerol for 2 h (final volume,74 ml), diluted by addition of 34 ml of this buffer without NaCI, and applied to a tRNA-agarose column (11 cm X 0.75 cm') at 25 ml/h. The column was washed with 10 ml of 0.02 M KP0.t (pH 7.4), 5 mM EDTA, 2.5 mM dithiothreitol, 10%glycerol containing 0.15 M NaCl (spectralgrade glycerol wasemployed in thisand subsequent steps), and enzyme eluted with this buffer containing 0.4 M NaCl (Fraction V, 11 ml). Fraction V was diluted with an equal volume of glycerol prechilled to 0 "C. and dithiothreitol added to 2.5 mM. The enqyme was stored at -20 "C and lost no detectable activity ((10%) in 12 months. Gel Electrophoresis-SDS-polyacrylamide gel electrophoresis was performed by the method of Weber and Osborn (41). except samples were heated a t 100 "C for 20 min in the presence of2% (w/v) dithiothreitol, 1% SDS, and dialyzed versus 10 mM NaP0, (pH 7.1). 0.1% SDS, 1% (w/v) dithiothreitol, 10% (w/v) glycerol prior to electrophoresis. Nonreduced samples were prepared in the same manner except dithiothreitol was omitted. Molecular weight standards were ovalbumin (M, = 44,000), Eco RI endonuclease ( M , = 31,000), chymotrypsinogen (M, = 24,500), horse heart myoglobin (M, = 16,900). and lysozyme (M, = 14,300). Isoelectric focusing under native conditions was performed for 24 h at 4 "C on 7.5% polyacrylamide gels containing2% ampholytes (BioRad Biolytes 3/10) as recommended by the manufacturerexcept that gels contained 2.5 mM dithiothreitol. Gels (10 cm) to be assayed were cut into approximately 50 slices of 0.21 cm which were soaked for 24 h at 0 "C in 0.2 ml of 0.02 M KPO, (pH 7.4). 0.2 M NaCI, 0.2 mg/ml of bovine serum albumin, 2 mM dithiothreitol, 1 mM EDTA, 10% (w/v) glycerol. For pH measurements, slices from parallel gels were incubated in 0.4 ml of degassed 0.1 M KC1 for 7 h at 4 "C andpH determined a t 4 "C. Gels to be stained were soaked in 10% isopropyl alcohol-lO% acetic acid for several days to remove ampholytes and stained for 30 min in 0.25% Coomassie brilliant blue. Other Methods-Protein was estimated by the procedureof Lowry a t 260 nm was assumed in calculation of et al. (42). Anof200 DNA concentrations. A molecular weight of3.7 X 10" for G4 RFI DNA was employed for determining molar yields (29). Other procedures have been described previously (3).

TABLE I Purification of dam methylase from I kg of E. coli K12 GM.31/ pGG503 Specific activitv

Fraction Protein step and

extract I. Crude 11. Alumina Cy gel 111. Phosphocellulose IV. Blue dextran-Sepharose V. tRNA agarose

mg

units/mg protein

78.000 1,500 44 33.2 8.6

286 12,100 298.000 580.000

3.0

Recovery

82 23

900,000-1,000,000 12

RESULTS

Cloning of the damMethylase-Since the damlocus is 90% cotransdusible with trpS (26), aplasmid of the Clarke-Carbon collection (pLC13-42) (24) which bears trpS was tested for the presence of the dam locus. Strains bearing the dam-4 allele grow poorly on T plates containing 400 pg/ml of 2-aminopurine, while dam' strains grow normally (5). Introduction of pLC13-42 DNA intostrain SK1036 (dam-4)renderedthe strain resistant to 2-aminopurine indicating the presence of a wild type dam locus. C o n f i a t i o n of this conclusion was provided by in vitro assay of dam-mediated methyltransferase activity. SK1036/pLC13-42 had a specific activity of 45 units/ mg as compared with 0.3 unit/mg for the plasmid-free dam-4 strain. In an attempt to obtain cloneswhich overproduced the methylase, the dam locus was recloned into pBR322 using several restriction endonucleases. While 2-aminopurine-resistant clones were obtained after ligation of Eco RI or Pst I restriction products, only latter clones contained elevated in vitro levels of the enzyme. One such clone, SK1036/pGG503, overproduced the enzyme 10- to 20-fold relative to wild type (data not shown). Comparablelevels of overproduction were found in the dcm strain GM31 (Table I)." Pst I endonuclease hydrolysis of pGG503 revealed that it contained two fragments: the vector pBR322 and the largest Pst I fragment in theClarke-Carbon plasmid pLC13-42 (Fig. 1).Restriction

FIG. 1. Cleavage of plasmid DNA with Pst I endonuclease. Samples were pLC13-42 (wells I and Z ) , pGG.503 (wells 3 and 4 ) . and pBR322 (ujells 5 and 6). Within each set the left sample contained undigestedDNA while therightsample was treated with Pst I endonuclease. Electrophoresis was on a 1% agarose gel using a Trisborate buffer system (40) for 4 h at 100 V. DNA bands were visualized by staining for 1 h in 1 pg/ml of ethidium bromide. There are two additional Pst I fragments of pLC13-42 which have run off the gel shown. Material not entering the gel in well 1 was not observed in all preparations and may represent aggregated or nicked DNA.

'' The elevation of in vitro dam activity is approximately 50-fold ir strain JC4588/pGG503 (43) which is a rec A host.In all o t h e backgrounds (strains JC4583, SK1036, GM31) elevations of 10- to 20 fold were found. Similar levels of overproduction were found usin either calf thymus or dam-E. coli DNA as substrates.

enzyme hydrolysis with Eco RI and doubledigests with Pst I and Eco RI (not shown)were also consistent with incorporation of the large Pst I fragment (M, = approximately 12 X lo6) from pLC13-42 into the pBR322 vector.

E. coli dam Methylase

2608

Purification of the dam Methylase from a n Overproducing Clone-E. coli K12 dam methylasewas purified fromthe dcm strain GM31 containing the hybridplasmid pGG503 (“Methods’’). In this background, the enzyme is overproduced approximately 10- to 20-fold in crude extracts (Table I).Yields of methylase obtained form the overproducingclonewere consistent with the degreeof overproduction of enzyme (Table 11) indicating that the increased activityresulted from a physical overproduction in vivo. The use of tRNA-agarose as the terminal stepin purification (Table I) served to remove more than 99% of a contaminating endonuclease which cofractionated with dam methylase through much of the procedure, and reproducibility was improved over the previous method which used phenylalanylSepharose for this purpose (3). Nevertheless, Fraction V did contain residual endonuclease activity. Although the preparation did not detectably convert T7 [”HIDNA to an acidsoluble form,it did introduce single strand breaks into plasmid pBR322 DNA at a low rate (100 pmolmin”mg”). This endonuclease activity was nonspecific and inhibitedby tRNA suggesting that itwas endonucleaseI (38).This was confirmed by isolation of the methylase from JCt583/pGG503 (EndA dcm’). Although this preparation was contaminated at the 3%level by DNA cytosine methylase activity due to the dcm’ genotype of JC4583, it was free of contaminating endonuclease (