APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Mar. 1993, p. 777-7850099-2240/93/030777-09$02.00/0Copyright © 1993, American Society for Microbiology

ScrFI Restriction-Modification System of Lactococcus lactissubsp. cremoris UC503: Cloning and Characterization of Two

ScrFI Methylase GenesRUTH DAVIS,' DANIEL VAN DER LELIE,2t ANNICK MERCENIER,2 CHARLES DALY,"3

AND GERALD F. FITZGERALD"*Department ofFood Microbiology' and National Food Biotechnology Centre,3 University College,

Cork, Ireland, and Transgene, S.A., 67082 Strasbourg, France2

Received 15 May 1992/Accepted 7 January 1993

Two genes from the total genomic DNA of dairy starter culture Lactococcus lactis subsp. crenoris UC503,encoding ScrFl modification enzymes, have been cloned and expressed in Escherichia coli. No homologybetween the two methylase genes was detected, and inverse polymerase chain reaction of flanking chromosomalDNA indicated that both were linked on the Lactococcus genome. Neither clone encoded the cognateendonuclease. The DNA sequence of one of the methylase genes (encoded by pCI931M) was determined andconsisted of an open reading frame 1,170 bp long, which could encode a protein of 389 amino acids (Mr, 44.5).The amino acid sequence contained the highly characteristic motifs of an m5C methylase. Extensive regions ofhomology were observed with the methylases of NIaX, EcoRII, and Dcm.

Members of the genus Lactococcus are commerciallyimportant bacteria, involved in the manufacture of a range offermented food products. In milk fermentations, the singlemost significant factor leading to poor starter performance isthe proliferation of lytic bacteriophage. Several mechanismsof phage defense have been identified among the lactococci.These include adsorption blocking, abortive infection, andrestriction-modification (R-M) (see references 24 and 38 forreviews). R-M systems, which are widely distributed amongthe lactococci, are generally plasmid encoded and tend to becharacterized on the basis of their effect on host-dependentphage replication (13). Multiple, independent R-M systemshave been observed to occur naturally in a number oflactococcal strains and can result in additive levels of phageresistance (7, 21).Type II R-M systems in bacteria consist of two enzymatic

activities: (i) an endonuclease (ENase) that recognizes andrestricts DNA at or very near a specific recognition sequenceand (ii) a modification activity (methylase [MTase]) thatmethylates the DNA at the same sequence, protecting itfrom restriction by the corresponding ENase (30). Recently,a third gene which is associated with some R-M systems andappears to encode a trans-acting regulatory protein has beenidentified (44).

Considerable effort has focused on cloning the genesinvolved in type II R-M systems, especially as they repre-sent a very promising model system for the study of DNA-protein interactions. A comparison of the active sites ofvarious systems at both the DNA level and the protein levelwill help determine the mode of action of these enzymes andpossibly elucidate some general principles of DNA se-

quence-specific protein recognition.ScrFI is the first type II restriction ENase to have been

isolated from Lactococcus spp. (15). It was identified inLactococcus lactis subsp. cremoris UC503 (originally desig-nated Streptococcus cremoris F [41]), an isolate from a

* Corresponding author.t Present address: VITO Laboratory of Genetics and Biotechnol-

ogy, 2400 Mol, Belgium.

commercial mixed-strain starter culture. It has an absoluterequirement for MgCl2 and recognizes the nucleotide se-

quence 5' CC C NGG 3', cleaving at the point indicated bythe arrow. Nelson and McClelland (32) have suggested thatthe modification component of ScrFI is a cytosine MTase.Evidence to date, based on plasmid curing and stabilityexperiments, indicates that the ScrFI R-M genes are chro-mosomally located. LlaI, the type II R-M system identifiedon pTR2030 in Lactococcus lactis subsp. lactis ME2 hasbeen cloned (20), and its MTase component has been se-

quenced (19), although characterization of the ENase itencodes has not yet been reported.

In this paper, the biological contribution of the ScrFI R-Msystem in protecting the host bacterium L. lactis subsp.cremoris UC503 against infection by lytic bacteriophage isdemonstrated. We also report the cloning and analysis oftwo MTase genes (encoded by pCI931M and pCI932M) fromUC503 in Escherichia coli that fully protect host DNA fromdigestion by ScrFI ENase. In addition, the nucleotide se-

quence of one of the MTases (encoded by pCI931M) was

determined. The deduced amino acid sequence was com-

pared with other cytosine MTases including those of NlaX,EcoRII, and Dcm. The latter two MTases modify a subset ofthe recognition sequence of ScrFI.

MATERIALS AND METHODS

Bacterial strains, plasmids, bacteriophages, and media. Thebacterial strains, plasmids, and bacteriophages used are

listed in Table 1. L. lactis strains were grown at 30°C in M17medium (46) containing glucose or lactose (0.5%) as re-

quired. E. coli was grown in Luria-Bertani broth at 37°C withcontinuous agitation. Where necessary, ampicillin (100 ,ug/ml) or tetracycline (12.5 ,ug/ml) was added to the relevantmedium. E. coli TG1 (8) was used as the host strain for allsubcloning and sequencing experiments involving M13tgl30or M13tgl31 bacteriophage (22). Recombinant M13 phagewere identified as colorless plaques when plated on H agar(10 g of Bacto Tryptone and 8 g of NaCl per liter), supple-mented with 0.004% 5-bromo-4-chloro-3-indolyl-3-D-galac-

777

Vol. 59, No. 3

APPL. ENvIRON. MICROBIOL.

TABLE 1. Bacterial strains, bacteriophages, and plasmids

Strain, phage, or plasmid Relevant genotype or phenotype Reference or source

Escherichia coliHB101 F- hsdSA20(rB- mB-) recA13 ara-14 proA2 lacYgalK2 rpsL20 (Smr) xyl-5 5

mtl-l supE44TG1 supE hsdA5 thi A(lac-proAB) F' [traA36 proAB+ lacIq lacZAM15] 8ED8739 supE supF met hsd mcrA mcrB 31aER1648 A(mrr-hsd RMS-mcrB) mcrAl272 serB28 50aGM31 thr-1 hisG4 leuB6 rpsL ara-14 supE44 lacYl tonA31 tsx-78 galK2 galE2 29a

xyl-S thi-1 mtl-l dcm-6Lactococcus lactis subsp.

cremorisUC503 Wild-type industrial starter strain UCC Culture

CollectionbUC563 UC503 cured of pCI528 12

L. lactis subsp. lactisC2 Wild-type strain UCC Culture

CollectionUC811 Wild-type strain UCC Culture

CollectionBacteriophages

c2 Homologous phage of L. lactis subsp. lactis C2 L. L. McKaycuc3811 Homologous phage of L. lactis subsp. lactis UC811 UCC Culture

CollectionXvir Wild-type lambda bacteriophage 17

PlasmidspBR322 4.3-kb E. coli cloning vector, Apr Tcr 4M13tgl3O and M13tg131 M13 cloning and sequencing vector, Apr LacZ 22apCI931M 3.4-kb EcoRI-Sau3A chromosomal fragment from L. lactis subsp. cremoris This work

UC503 cloned in pBR322 exhibiting m5C MTase activity, AprpCI932M 2.6-kb EcoRI-Sau3A chromosomal fragment from L. lactis subsp. cremoris This work

UC503 cloned in pBR322 exhibiting m5C MTase activity, AprM13 2(1) and 5(2) 2.2-kb EcoRI-HindIII fragment of pCI931M subcloned in M13tgl3O and This work

M13tgl31, respectivelyM13 3(1) and 6(1) 1.47-kb HindIII-SalI fragment of pCI931M subcloned in M13tgl3O and This work

M13tgl31, respectivelya Transgene, S.A., Strasbourg, France.b UCC, University College, Cork, Ireland.c University of Minnesota, St. Paul.

toside (X-Gal) and 0.5 mM isopropyl-13-D-thiogalactopyrano-side (IPTG).

Bacteriophage assays. Lactococcal phage propagation andplaque assays were carried out with M17 broth or agar.Luria-Bertani broth and agar were used for propagating andplaquing Xvir which was maintained and diluted in SM buffer(37). An initial stock of Xvir was prepared with E. coli GM31as host, as this strain is deficient in Dcm methylation (29),which partially protects against restriction by ScrFI. Theefficiency of plaquing (EOP) of bacteriophage on varioushosts was defined as the phage titer on the restrictive hostdivided by the phage titer on a nonrestrictive host (39).Modified bacteriophage were obtained by picking a singleplaque from cell lawns of the restrictive host and propagatingthem to a high titer on the same host (18).DNA preparation. Total cellular DNAwas isolated from L.

lactis subsp. cremoris UC503 according to the methodoutlined by van der Vossen et al. (47). Plasmid DNA wasprepared from E. coli by alkali lysis (37) and purified byCsCl-ethidium bromide ultracentrifugation. Plasmid mini-preparations were carried out by the alkaline-sodium dode-cyl sulfate lysis method (1).DNA restriction, molecular cloning, and gel electrophoresis.

All restriction enzymes (except ScrFI) and T4 DNA ligasewere purchased from Boehringer Corporation Ltd. (Dublin,Ireland) or Promega Corporation (Madison, Wis.) and usedaccording to the manufacturer's recommendations. ScrFI

was purchased from New England Biolabs (Beverly, Mass.).Molecular cloning protocols were performed according tothe methods described by Sambrook et al. (37). DNAdigested with restriction enzymes was analyzed on 0.8%horizontal agarose gels in TAE buffer (40 mM Tris-acetate, 2mM EDTA, pH 8.0) stained with ethidium bromide (0.03A±g/ml).Transformation. E. coli strains were transformed either by

CaCl2 treatment (10) or electrotransformation. Electrotrans-formations were performed with a Bio-Rad Gene Pulser(Bio-Rad Laboratories, Richmond, Calif.) according to themanufacturer's instructions, and the maximum volume ofDNA used was 1/10 that of the cells. In addition, ligationmixtures and restriction digests were dialyzed to removesalts with Millipore VS filters (0.025 ,um; Millipore S. A.,Molsheim, France) before electrotransformation.

Construction of genomic library and selection of plasmidsexpressing ScrFI MTase activity. The L. lactis subsp. cremo-is gene bank was constructed by partially digesting 40 p,g oftotal DNA with EcoRI (1 U/,g of DNA, 3 min) and Sau3A (1U/,ug of DNA, 5 min) and ligating the fragments into pBR322digested with EcoRI and BamHI. Transformants were se-lected on Luria-Bertani broth plus ampicillin, and the fre-quency of insertion was estimated by the percentage oftransformants exhibiting insertional inactivation of the tetra-cycline resistance gene of pBR322. As many of the deriva-tives of E. coli K-12 restrict DNA that is cytosine methylated

778 DAVIS ET AL.

TYPE II R-M SYSTEM IN L. LACTIS 779

(3), the gene bank was maintained in E. coli ED8739 (31),which is McrA- McrB- and is, therefore, tolerant of cy-

tosine methylation (35).The plasmid population of the library was isolated and

purified by CsCl-ethidium bromide density gradient centrif-ugation. The purified plasmid DNA (1 ,ug) was subsequentlydigested to completion with ScrFI (10 U) and used totransform E. coli ED8739. Surviving transformants were

assessed for tetracycline sensitivity, and the plasmid contentof tetracycline-sensitive clones was analyzed by restrictionenzyme digestion and agarose gel electrophoresis.

Southern hybridization. DNA was transferred from agar-

ose gels to Hybond N+ nylon membranes (Amersham, LittleChalfont, Buckinghamshire, United Kingdom) by the methodof Southern (42). DNA probes were labelled with peroxidaseand detected by enhanced chemiluminescence (ECL, Amer-sham). Alternatively, the DNA probes were labelled withdioxygenin-dUTP from the nonradioactive DNA Labellingand Detection kit supplied by Boehringer Corporation Ltd.

Analysis of flanking chromosomal DNA by inverse PCR. APerkin-Elmer Cetus DNA Thermal Cycler (Perkin-Elmer,Norwalk, Conn.) was used to carry out the polymerase chainreaction (PCR). TaqI polymerase was purchased from Cetus(Emeryville, Calif.). Ultrapure deoxynucleotides were ob-tained from Pharmacia Ltd. (Uppsala, Sweden). The relativepositions of the oligonucleotide primers (see Fig. 4) were

based on sequence data. The sequences are as follows: oligo1, 5' GCC-GTA-TAT-TTC-TTC-CTG-AAA-TAT-CAG-3',and oligo 2, 5' GAA-TTC-CAG-AAA-1TT-TAT-TAT-ACC-TG-3'. Total L. lactis subsp. cremoris DNA (1 ,ug) was

digested to completion with HindIII and ligated under con-

ditions facilitating the formation of circular monomers (<1,ug of DNA per ml of ligation). The circularized DNA was

precipitated and added to a 50-,ul PCR mixture containingprimers (0.1 ,uM each), deoxynucleotide triphosphates (200mM each), 50 mM KCI, 10 mM Tris-HCl (pH 8.3 at roomtemperature), 1.5 mM MgCl2, 0.01% gelatin, and 2.5 U ofTaqI polymerase. The reaction mixture was overlaid withparaffin. The DNA Thermal Cycler was programmed asfollows: 1 min at 94°C (denaturation), 2 min at 45°C (anneal-ing), and 3 min at 68°C (extension). This profile was repeatedfor 30 cycles. The amplified DNA fragment was analyzed byagarose gel electrophoresis and used without further purifi-cation as a DNA probe in Southern hybridization experi-ments.DNA sequencing. The nucleotide sequence of the M

ScrFI gene was determined by the dideoxy chain terminationmethod of Sanger et al. (40) with the Sequenase (Version 2.0)kit supplied by United States Biochemical Corporation(Cleveland, Ohio). Single-stranded DNA derived fromM13tgl30 and M13tgl31 subclones and alkali-denatured dou-ble-stranded DNA from pCI931M were both used as tem-plates for the reactions. The universal M13 primer was usedwhere appropriate, and subsequently, internal primers (18-mer) were prepared with an Applied Biosystems PCR-MATE DNA synthesizer (see Fig. 2B). a- 5S-dATP aS(New England Nuclear, Boston, Mass.) was the labellingdeoxynucleotide used. Products of the sequencing reactionswere separated on a Sequi-Gen Nucleic Acid SequencingCell (Bio-Rad) with 6% acrylamide gels. After fixation in10% acetic acid-10% methanol, the gels were transferredonto 3MM paper (Whatman, Maidstone, Kent, United King-dom) and dried under a vacuum at 80°C (Savant Slab GelDryer; Savant Instruments, Inc., Farmingdale, N.Y.). Au-toradiography was performed at room temperature for 24 to48 h with Hyperfilm-omax X-ray film (Amersham). Se-

TABLE 2. Demonstration of the biological activity of ScrFI R-M

Activity in host strain:

Phage subs.lactis L. lactis subsp. L. lactis subsp.Phage subsp. lactis lactis UC811 cremoris UC563

PFU/mI EOP PFU/ml EOP PFU/ml EOP

c2 9 x 108 1 1 1 x 104 1 x 10-5uc3811 4 x 109 1 130 3 x 10-8c25- 63a 4 x 109 1 2 x 108 0.05uc3811 .563a 5 x 109 1 1 x 107 2 x 10-3c2 * 563 C2 1 x 109 1 6 x 104 6 x 105uc3811 * 563 * 811 2 x 1010 1 50 2 x 109

a c2 * 563 and uc3811 563 designate c2 and uc3811, respectively, whichwere passaged through L. lactis subsp. cremonis UC563.

b c2 563 C2 and uc3811 563- 811 designate c2 563 and uc3811 * 563,respectively, which were passaged through their corresponding homologoushost strains.

quence data were compiled and analyzed with the Microge-nie sequence analysis software program (Beckman Instru-ments, Inc., Palo Alto, Calif.) and GeneJockey (AppleComputer, Inc., Cupertino, Calif.).

Nucleotide sequence accession number. The nucleotidesequence of M- ScrFI encoded by pCI931M is published inGenBank-EMBL under accession no. M87289.

RESULTS

Biological evidence for ScrFl R-M activity in L. lactis subsp.cremoris UC563. The plaque-forming abilities of bacterio-phages c2 and uc3811 on restricting and nonrestricting hostswere examined in order to determine whether the ScrFI R-Msystem had an in vivo effect on phage replication. Thehomologous hosts of these bacteriophages, L. lactis subsp.lactis C2 and UC811, respectively, were used as nonrestrict-ing strains. L. lactis subsp. cremoris UC503 is known tocontain a plasmid, pCI528, which encodes two bacterio-phage resistance mechanisms that mask the effect of theScrFI system in this strain (9, 12). To eliminate the effects ofpCI528, L. lactis subsp. cremoris UC563 (a variant ofUC503lacking pCI528 but still retaining the ScrFI activity) was usedas the host to demonstrate R-M activity.The data, presented in Table 2, showed that unmodified

bacteriophages c2 and uc3811 plaqued at very low efficiencyon the restricting host UC563 (EOPs of 1 x 10-5 and 3 x10-8, respectively). However, when plaque isolates fromthis host were propagated and retitered on both restrictingand nonrestricting hosts, the modified phage showed greatlyincreased plaquing abilities on UC563, a 3-log-cycle increasein titer in the case of phage c2 and a 5-log-cycle increase inthe case of uc3811. A single passage through their respectivenonrestricting hosts resulted in the loss of UC563 modifica-tion, and both phage c2 and phage uc3811 were restrictedagain by UC563.

Isolation of ScrFI MTase clones. pBR322 DNA contains 16ScrFI recognition sites, 10 of which remain susceptible torestriction when the vector is maintained in a Dcm-proficienthost strain. Selection of recombinant clones carrying theScrFI modification gene was based on the expectation thatexpression of the MTase gene would modify these sites,protecting the vector from subsequent digestion in vitro byScrFI ENase. Similar approaches have been used to isolateseveral individual MTase genes as well as a number ofcomplete R-M systems (28, 48).

VOL. 59, 1993

780 DAVIS ET AL.

1 2 34 5 6 7 89101112 A

RsaIEcoRI kde

pCI931M (7.4 kb)

EcoRV Hind, i I SallIPfal Rjal S41

pBR322

pCI932M (6.6 kb)

EcoRI Hindll Rsa SII VI Il~alI Si

UC503

M*SCRFI _

NdeI

RsaI

FIG. 1. Restriction digest analysis of the recombinant plasmidspCI931M and pCI932M encoding ScrFI MTase genes. Lane 1,molecular weight standards, X DNA digested with HindIII andEcoRI; lanes 2 to 4 and 6, recombinant plasmid DNAs isolated fromfour individual E. coli clones from which pCI931M (lane 2) was

chosen to be representative, digested with ScrFI; lanes 5 and 7,pCI932M and pBR322, respectively, both digested with ScrFI.Lanes 8 to 10 and 12, DNAs as described for lanes 2 to 4 and 6digested with EcoRI and SalI; lane 11, pCI932M digested withEcoRI and Sall.

A gene library of L. lactis subsp. cremoris UC503 totalDNA was constructed and consisted of 14,800 clones, 60%of which insertionally inactivated the tetracycline resistancegene of pBR322. The average size of the inserted DNA was3 kb. The size of the lactococcal genome is estimated to bebetween 2.3 and 2.6 megabases (26); therefore, approxi-mately 10 genomic equivalents were represented in thelibrary. Four hundred fifty ampicillin-resistant, tetracycline-susceptible transformants were recovered after digestion ofthe plasmid content with the ScrFI restriction ENase. Fortyof these were individually picked, and their plasmid contentwas analyzed by restriction enzyme digestion and agarose

gel electrophoresis. A total of 39 of the 40 plasmids contin-ued to exhibit resistance to ScrFI digestion and were judgedto encode a potential ScrFI modification gene. Thirty-eightof the plasmids appeared to carry a common 3.4-kb EcoRI-Sau3A fragment. One, designated pCI931M, was chosen forfurther study. The remaining ScrFI-resistant clone carried a

2.6-kb EcoRI-Sau3A fragment and was designated pCI932M(Fig. 1, lanes 5 and 11). As neither the BamHI nor Sau3Arestriction sites were regenerated during the cloning proce-

dure, the lactococcal DNA inserts of pCI931M and pCI932Mwere released from the pBR322 vector by digesting theplasmids with EcoRI and Sail, which excised an additional346 bp of vector with the inserts. The resultant insertfragment of pCI931M comigrated with the vector portion ofthis plasmid (Fig. 1, lanes 8 to 10 and 12). Both pCI931M andpCI932M were also shown to confer partial resistance toNciI, which restricts a subset of the ScrFI recognition sites(CCSGG) (data not shown).

Characterization of pCI931M and pCI932M. Cells thatrestrict are less susceptible to attack by unmodified phage

I kb _

FIG. 2. (A) Restriction maps of the two inserts of L. lactis subsp.cremoris UC503 chromosomal DNA that exhibit ScrFI methylaseactivity. (B) Sequencing strategy of M ScrFI from pCI931M. Thethree types of arrows indicate the sequence information obtainedfrom different strategies by the dideoxy chain termination method.Sequences of M13 subclones were initially obtained with the uni-versal primer (solid arrows). The remainder of the sequence was

determined by using synthetic oligodeoxynucleotides to prime reac-tions with the single-stranded M13 clones (checkered arrows) ordouble-stranded DNA templates of pCI931M (open arrows). Theposition of the open reading frame is indicated by the large shadedarrow above the sequence information.

and, therefore, less efficient at plaquing phage than are

nonrestricting cells. No decrease in EOP was observed,however, when E. coli strains harboring either pCI931M or

pCI932M were challenged with unmodified X phage, suggest-ing that the inserts did not encode the complete ENase geneor that, if present, it was not expressed in E. coli. This wasconfirmed when no in vitro ScrFI activity was detected inlysates of the E. coli strains (data not shown). Preliminaryrestriction maps were constructed for both the inserts whichmediated resistance to ScrFI restriction (Fig. 2A).

Hybridization analysis of pCI931M and pCI932M. In orderto confirm that the cloned fragments were derived from L.lactis subsp. cremois UC503, the EcoRI-HindIII fragmentof pCI931M (2.2 kb) was labelled and used to probe variousdigestions of UC503 total DNA. A single 3.6-kb EcoRI bandhybridized to the pCI931M-derived probe (Fig. 3A, lane 5).In addition, hybridizing fragments of approximately 20 kbwere detected in the PstI-PvuI, PvuII, and PstI digestions(Fig. 3A, lanes 2 to 4, respectively). Minor hybridizingfragments observed with the control plasmid pCI931M (Fig.3A, lane 6) arose from partially digested fragments whichwere detected by the sensitive ECL probe but were notvisible in the original agarose gel. When the same gel wasprobed with the EcoRI-HindIII (1-kb) fragment of pCI932M(Fig. 3B), the same-size hybridizing fragments were ob-served in lanes 2 to 4, while a much larger EcoRI fragmentthan that which hybridized to pCI931M hybridized topCI932M (Fig. 3B, lane 5). No homology between the twocloned inserts was observed (Fig. 3A, lane 8, and Fig. 3B,lanes 6 and 7).

Ability of pCI931M and pCI932M to become established in

UC503

RsaIINdeI Rsai EcoRI

Rsal

|NtdeIN&I

1 kb

pBR322

Rsal EcoRI

B

EcoRI-|-P

6-----"

APPL. ENVIRON. MICROBIOL.

TYPE II R-M SYSTEM IN L. LACTIS 781

A 1 2 3 4 5

, . A<a. .

HindM EcoRI6 7 8 HindM Sau3A

OLIGO ICHROMOSOMAL

- FRAOMGET CLONEDIN pCI93 IM

M,m FLANKING UCS03CHROMOSOMAL DNA

_ r pBR322

UC503chromosome

DIGEST UC503 CHROMOSOMALDNA WITH Hindm AND LIGATE(1 jg DNA/mi), TO FORM CIRCULARMONOMERS.

20-30 CYCLES OF PCR

_ FRAGMENT AMPLFED BY_------ - PCR AND WHICHHYBRIDIZED TO pCI932M

EcoRI t HindM SamuH]r CI9M1_

I PCI932M

RELATIVE POSITIONS OF CLONED INSERTSON UCS03 CHROMOSOME

Sau3A HindM EcoRI HindM Sau3A

FIG. 3. Hybridization analysis of pCI931M (A) and pCI932M (B)to determine the origin of the insert DNAs. (A) Hybridization withan ECL-labelled EcoRI-HindIII fragment of pCI931M (2.2 kb).Labelled DNA was included with the probe as molecular weightstandards. (B) Hybridization with an ECL-labelled EcoRI-HindIII

fragment of pCI932M (1 kb) only. Lanes 1, molecular weightmarker, K DNA digested with Hindlll; lanes 2 to 5, L. lactis subsp.cremoris UC503 total DNA digested with Pvul and Pstl, PvuII,

Pstl, and EcoRI, respectively; lane 6, pCI931M digested with EcoRI

and Hindlll; lane 7, pCI931M digested with EcoRI and Sall; lane 8,

pCI932M digested with EcoRI and HindIII.

certain E. coil hosts. It is well documented that many strains

ccli methylation-dependent restriction

systems that restrict DNA carrying foreign methylation such

as that found on plasmids encoding an MTase gene (2, 3, 33,

36). These systems, McrA, McrB, and Mrr, prevent the

establishment of such plasmids. It was decided, therefore, to

examine the ability of some commonly used laboratory E.

coli strains to tolerate the presence of pCI931M andpCI932M.

All the strains tested (HB101 [McrAsMcrBu1, GM31

[McrAn McrB)DcmH 1, TG1 [McrAn McrBe], and ER1648

[McrAg McrBIM)were transformed normally by nonmeth-

ylated pBR322 DNA (approximately 104 toi,L transfor-mants perUg of DNA). No transformants could be detectedwhen pCI931M was used to transform HB101, GM31, or

TG1. In the case of pCI932M, however, low numbers of

transformants were isolated with GM31 (70 transformants

per msg of DNA) and TG1 (120 transformants perh t g of

-<- INSERT OF pCI932M ON. -- INSERT OF pCI93IMM *.

FIG. 4. Schematic representation of the amplification of se-quences flanking the insert of pCI931M by inverse PCR (not drawnto scale).

DNA). None were isolated with HB101 as the transforminghost. E. coli ER1648 (50) was transformed by pCI931M andpCI932M with efficiencies comparable to those obtainedwith pBR322.

Identification of the flanking DNA sequence of pCI931M bymeans ofPCR. The position of the cloned MTases relative toeach other on the lactococcal chromosome was determinedthrough amplification of the flanking region of one of theinserts (pCI931M) by inverse PCR (45). The genome of L.lactis subsp. cremoris UC503 was digested to completionwith HindIII, resulting in junction fragments composed of aregion of the cloned DNA together with a region of theflanking chromosome (Fig. 4). The genomic DNA was thenligated to form a series of circular molecules. The primerswere constructed such that they directed DNA synthesisaway from one another. After the first cycle of synthesis, alinear form of the target DNA outside the original clonedarea was generated and PCR continued as normal. Theresultant band of DNA, 1 kb in size, was labelled and used toprobe digestions of UC503 total DNA and of pCI931M andpCI932M DNAs. Analysis of the hybridizing bands (Fig. 5)indicates that (i) the 1-kb fragment originated from L. lactissubsp. cremoris UC503 (lanes 1 to 4), (ii) the fragment didnot hybridize with the DNA of pCI931M (lanes 5 and 6), and(iii) the fragment did hybridize with the 1-kb EcoRI-HindIIIand the 2.9-kb EcoRI-SalI fragments of pCI932M (lanes 7and 8, respectively). These results may be explained if thetwo inserts are adjacent to one another on the UC503chromosome, linked by a common EcoRI site (Fig. 4). BothMTase-encoding fragments were also shown to hybridize toa common HindlIl fragment of the UC503 genomic DNA,which further supported this conclusion (data not shown).

B 1 2 3 4 5 6 7 8

_0 _W Om

VOL. 59, 1993

OLIGO 2

782 DAVIS ET AL. APPL. ENVIRON. MICROBIOL.

AGC

4 TACTGGAATAGTTTAATTAACGGATTGCCACCAATAAGAAATATGAGCATCTAATAAAAAATTAAAT71 TTTGGAACGATATTTATTCGTTCTTTTTTCTTTAATAAAAATAATC 3TCATTGAATTTAT

- 35

138 13TAGTGTCAATCAGAG GACAAAAAGTCCTTATAAAAACACTTTATGAGAAAGGATTTAT- 10 RBS

Met Thr Thr Ile Ser Arg Asn Thr Gly Thr Glu Ile Ser Ile Met Ile Lys205 ATG ACT ACT ATA TCA AGA AAT ACA GGA ACA GAA ATT TCT ATA ATG ATT AAA

Glu Lys Arg .Leu Arg Leu Asn Met Thr Gln Lys Glu Leu Ala Asp Ala Vai256 GAA AAG CGA CTA CGC CTA AAT ATG ACT CAA AAA GAG CTT GCA GAT GCA GTG

Gly Met Ser Lys Asn Gly Asp Arg Thr Ile Arg Arg Trp Glu Asn Gly Glu307 GGC ATG TCA AAA AAC GGA GAT AGA ACC ATA AGA AGA TGG GAA AAT GGT GAA

Thr Cys Pro Ser Gln Leu Glu Ile Ser Ala Ile Leu Arg Phe Pro Glu Ile358 ACT TGT CCA TCA CAG CTT GAA ATA TCT GCT ATT TTA AGA TTT CCA GAA ATT

FIG. 5. Hybridization analysis of DNA product obtained follow-

ing inverse PCR. Lanes 1 to 4, L. lactis subsp. cremoris UC503

chromosomal DNA digested with PvuI and Pstl, PvuII, PstI, and

EcoRI, respectively; lane 5, pCI931M digested with EcoRI and Sail;

lane 6, pCI931M digested with EcoRI and Sall; lane 7, pCI932M

digested HindIII; pCI932M digested

EcoRI and SailI; lane 9, molecular weight marker, DNA digestedwith Hindlll. The DNA was probed with the dioxygenin-dUTP-labelled product (1 kb) synthesized by inverse PCR.

DNA sequence analysis. The DNA fragment encoding the

MERScrFI of pCI931M was cloned in two pieces (EcoRI-HindIII [2.2 kb], and HindIII-SalI [1.47 kb]) into the appro-priate sites on Ml3tgl3O or M13tgl31. Attempts to clone theentire fragment into M13 failed. These clones were used to

provide single-stranded templates for sequencing with theuniversal primer or the synthetic 18-mer sequencing primers.

During the course of sequencing, deletions were observed in

the EcoRi-HindIIlM13 subclones. As a result, the remain-

ing sequence was determined by using pCI931M as a tem-

plate for double-stranded sequencing (Fig. 2B).

Computer analysis of the resultant sequence (1,612 bp)revealed an open reading frame 1,170 bp in length (Fig. 6)

with an ATG start codon. This open reading frame predicts

a protein of 389 amino acids with a calculated molecular

mass of 44.5 kDa. Located 5 bp immediately upstream of the

putative start codon is a possible ribosome binding site

(GAAAGGA). This sequence is homologous to the 3' end of

the lactococcal 16S RNA (27), with a free energy of binding

AG' of -16.2 kcal (-67,780.8 J)/mol. Putative -10

(TATAAT) and -35 (TTGCAA) promoter sequences were

located 58 bp upstream of the ATG codon. A total of 17 bp

separated the -10 region from the -35 region. As has been

found in most other lactococcal promoters, there is the

dinucleotide TG 1 bp before the -10 sequence and the region

upstream of the -35 sequence is very A+T rich (84% over

63 bp [14]). A search for a terminator structure showed a

Ala Pro Phe Glu Asn Arg Lys Thr Ala Lys Tyr Lys Met Ile Asp Leu Phe 85409 GCA CCA TTT GAA AAT AGA AAA ACA GCT AAA TAT AA ATG ATT GAC CTA TTT

Ala Gly Ile Gly Gly Thr Arg Leu Gly Phe His Gln Thr Glu Lys Val Lys 102460 GCA GGT ATT GGA GGG ACA CGC TTA GGT TTT CAT CAA ACA GAA AAA GTA AAG

Ser Val Phe Ser Ser Glu Ile Asp Lys Phe Ala Ile Lys Thr Tyr Lys Ala 119511 ,TCT GTA TTT TCA TCT GAA ATA GAT AAA TTT GCT ATT AAA ACT TAT AAA GCG

Asn Phe Gly Asp Glu Pro His Gly Asp Ile Thr Lys Ile Asp Glu Lys Asp 136562 AAT TTT GGA GAC GAA CCT CAT GIG GAT ATA ACA AAA ATT GAT GAG AAA GA

HI EcoRVIle Pro Asp His Asp Ile Leu Val Gly Gly Phe Pro Cys Gln Ala Phe Ser 153

613 AZ= CCA GAC.CAT GAT ATT CTT GTC GGT GGA TTT CCT TGT CAA GCA TTT AGTIV

Gln Ala Gly Lys Lys Leu Gly Phe Asp Asp Thr Arg Gly Thr Leu Phe Phe 170664 CM GCA GGA AAA AAM TTA GGA TTT GAT GAT,.ACA CGT GGA ACT TTA TTT TTT

V

Glu Ile Ala Arg Ile Ile Lys Glu Lys Arg Pro Lys Ala Phe Leu Leu Glu 187715 GAA ATA GCA AGA ATT ATA AAA GAA AAA PGC CCT AAA GCA TTT CTA CTC GM

Asn Val Lys Asn Leu Lys Thr His Asp Lys Gly Arg Thr Phe Lys Thr Ile 204766 AAT GTA AAA AAT TTA AAA ACC CAT GAT AAA GGA AGA ACT TTT AAA ACA ATA

VI

Leu Asn Thr Leu Glu Glu Leu Asp Tyr Glu Val His Thr Ala Leu Phe Lys 221817 TTA AAT ACT TTG GM GAA rTT OAT TAT GAA ITT CAC.ACA GCG CTA TTC AAA

VII

Ala Arg Asp Phe Gly Leu Pro Gln Asn Arg Glu Arg Ile Tyr Ile Val Gly 238868 GCA AGA GAT TTT GGT CTT CCA CAA AAT AGA GAA AGG ATT TAT ATT GTA GGT

Phe Asp Arg Lys Ser Ile Ser Asn Tyr Ser Asp Phe Gln Met Pro Thr Pro 255919 TTT GAC AGA AAA TCA ATT AGC AAC TAT TCT GAT TTT CAA ATG CCT ACA CCT

Leu Gln Glu Lys Thr Arg Val Gly Asn Ile Leu Glu Ser Val Val Asp Asp 272970 TTA CAA GAA AAA ACT CGT GTT GGA AAT ATT TTA GAA TCC GTT GTT GAT GAT

Lys Tyr Thr Ile Ser Asp Lys Leu Trp Asp Gly His Gln Arg Arg Lys Thr 2891021 AAA TAT ACA ATA TCA GAT AAA CTT TGG GAT GGG CAT CAA AGA AGA AAA ACT

Glu Asn Lys Lys Asn Gly Lys Gly Phe Gly Tyr Thr Leu Phe Asn Gln Asp 306

1072 GAA AAT AAA AAA AAT GGC AAA GGT TTT GGA TAT ACC CTC TTT AAT CAA GAT

Ser Glu Tyr Thr Asn Thr Leu Ser Ala Arg Tyr Tyr Lys Asp Gly Ser Glu 323

1123 AGT GAG TAT ACA AAC ACT TTA TCA GCT CGT TAT TAT AAA GAT GGT AGT GAA

Ile Leu Ile Glu Gln Lys Asn Lys Asn Pro Arg Lys Ile Thr Pro Arg Glu 3401174 ATA CTA ATT GAA CAG AAA AAT AAA AAT CCT CGA AAA ATT ACT CCT AGG GAA

Ix

Ala Ala Arg Leu Gln Gly Phe Pro Glu Asn Phe Ile Ile Pro Val Ser Asp 357

1225 GCT GCT CGA TTA CAA GGT TTT CCA GAA AAT TTT ATT ATA CCT GTT AGC GAT

Thr Gln Ala Tyr Lys Glu Phe Gly Asn Ser Val Ala Val Pro Thr Ile His 374

1276 ACC CA&A_GT TAT AAA GAA TTT GGA AAC TCT GTT GCA GTT CCC ACC ATT CAT

Ala Ile Ala Glu Lys Met Leu Glu Val Leu Glu Lys Ser Lys Lys Stop 389

1327 GCT ATT GCA GS AAG ATG TTG GAA GTT TTA GAG AAG TCA AAA AAA TAA CTA

1378 AAAAATCTCACGGgTAAATCTTqE2ATGAGAT?TTTTATATTCTCCAAATCTTCTTTTCTCTGMRsaI

1445 CTATTACGATATTAGACGTATTTTGAACAGAGGAATAGTTGATCTCCCTTATTTCCTCATCTGTTAG

1512 ATTATATACTTTGCTATCAAAAAGAATCATCTCACCATTCAATTTTATCTTTACCICACCAATACAABanI

1579 CGCTTAGGGGAAACAGGTAACATATAGTCAGCAT

FIG. 6. Nucleotide sequence of 1,612 bp of the EcoRI-SalI

fragment of pCI931M encoding one of the m scrFI genes. The

single large open reading frame present is translated into amino acid

sequence. The putative ribosome binding site (RBS) is denoted by

asterisks, and some significant restriction sites are underlined. The

-10 and -35 regions are boxed. The inverted repeat at the end of

the gene is indicated by facing arrows. The positions of the 10

conserved amino acid motifs present in m5C MTases (34) are

indicated by heavy lines beneath the nucleotide sequence and

numbered I to X.

17-bp imperfect inverted repeat beginning 1 bp after the stop

codon, TAA. The G+C content of the sequence is 31%,

which is slightly lower than the norm for Lactococcus spp.

(41). There is no recognition site present for ScrFI ENase

within the sequenced region.

Protein sequence analysis. The 10 conserved motifs (I to

X), characteristic of all known m5C MTases (34), were

clearly identified in the predicted protein sequence of

M * ScrFI and are indicated in Fig. 6. There is, however, one

notable deviation in block I (amino acids 81 to 100) in which

a normally invariant glycine residue is replaced by glutamic

acid at position 99. The 93-amino-acid-long internal variable

17

34

51

68

TYPE II R-M SYSTEM IN L. LACTIS 783

M.EcoR I H Q GIF T L R DII R F YrP P S F G ELF

M.Dcm R R D L N K A D F T L R D I E C FU TLA QL LM-NlaX L --D F R FT_ Q PI G Q A T - DOSI LM-ScrFI DK S IS QY PTDLFQEM -TLE- - G I I

M-EcoRII QP VV D|S K Y I L T P K L W EY L K K H A A KM.Dcm D P M T P i1L W KY L Y KKK ARM.NlaX IA YDEK Y T I SDIK L W Q GIYQ|R R K A ERNM.ScrFI S D K T I S K L W G R RKEN - - - N

MEcoR; G N G F Y F N 1EFIKE I - Rl T L S A R Y S

M-Dcm G N G F G Y - Y PN P S V R T L S A R Y Y K DAM.NlaX G K G F G Y L F N SI A Y NT[ S A R Y Y K D GSM- Scr]Fl G K G F G Y T L F N Q DS EI Y T1 NI T L S A R Y Y K D G SIM.EcoR E I L I D R G W D M A T G K DDP HM.Dcm E I L I D R G W D M A T G E T A NE EIAJM.NIaX E I L I P G KKM-ScrFl K N K

FIG. 7. Amino acid sequence alignment of the variable regions ofM * ScrFI, M N-aX, M Dcm, and M EcoRII. To optimize align-ments, gaps (dashes) have been introduced. Regions of homologyare boxed.

region, usually attributed to sequence recognition, is locatedbetween blocks VIII (amino acids 220 to 239) and IX (aminoacids 333 to 349). There is an additional variable N-terminalarm of 80 amino acids which may also contribute to se-

quence recognition. The dipeptide Pro-Cys (PC), which hasbeen implicated in the catalytic site of m5C MTases (51), ispresent in block IV at amino acid positions 148 and 149. Asexpected, M- ScrFI shows no homology with any knownENases in the GenBank data base. The protein showedsignificant homology to the following MTases: 0-3T (44.7%over 76 amino acids), p-lls (47% over 168 amino acids), SPR(44.6% over 186 amino acids), BsuRI (41.2% over 153 aminoacids), HhaI (43.9% over 305 amino acids), and NgoPII(41.1% over 168 amino acids). The highest levels of homol-ogy were observed between M ScrFI and the MTases ofMaX, EcoRII, and Dcm. The latter two MTases recognizethe sequence CCWGG, while the sequence specificity ofNiaX is unknown. M EcoRII was 59.3% homologous over

54 amino acids and 49.8% homologous over a further 213amino acids. This homology was increased to 70 and 63%,respectively, when conservative amino acid changes were

taken into account. M. Dcm, on the other hand, containednine short regions of between 62 and 88% identity with M-ScrFI. However, by far the greatest level of homology wasobserved between M ScrFI and M .NiaX, which exhibited67% identity over 307 amino acids composing almost theentire length of the deduced M NlaX protein sequence (313amino acids). Conservative amino acid changes increase thisto a remarkable 90% homology. We were especially inter-ested to note thatM * ScrFI and M. MaX were very similarthroughout the internal variable region. Furthermore, an

alignment of the variable regions of M MaX, M EcoRII,M. Dcm, and M ScrFI revealed the presence of a tetrapep-tide Gly-Phe-Gly-Tyr (GFGY) in all the sequences, followedby a second highly conserved segment of 17 amino acids(Fig. 7). Finally, there is a remarkable bias in codon usage.

Only 14% of the codons used have a C or G in the lastposition. This reflects the A/T richness of lactococcal DNAin general.

DISCUSSION

One of the principal functions of R-M systems in vivoappears to be the protection of the host strain against viralattack. Previous attempts to assess the biological role of

ScrFI R-M in L. lactis subsp. cremoris UC503 were ham-pered by the presence of a very effective phage resistanceplasmid, pCI528, in this strain. However, derivatives ofUC503 cured of pCI528 but retaining ScrFI activity exhib-ited between 3 and 5 log cycles of reduction in phageplaquing ability, confirming the defensive role of ScrFI inthis strain.

Selection by ENase digestion of the L. lactis subsp.cremoris UC503 gene bank resulted in the cloning andexpression of two MTase genes in E. coli whose geneproducts recognize and modify the recognition site of ScrFI,fully protecting the host DNA from digestion by this ENase.Both MTases also confer partial resistance to NciI restric-tion ENase. The two M. ScrFI genes show no homology toone another and originate from two distinct but adjacentEcoRI fragments on the UC503 chromosome.As expected, the E. coli strains encoding intact McrA

and/or McrB functions were poorly transformed bypCI931M and pCI932M. In particular, pCI931M could berecovered only in strains with an McrA- and McrB- phe-notype. By comparison, a low frequency of transformationwas observed for pCI932M DNA into E. coli TGI (McrA+McrB+) and GM31 (McrA+ McrB+ Dcm-). These transfor-mants expressed an active MTase and harbored plasmidDNA which appeared identical to that of pCI932M as judgedby restriction enzyme analysis. The L. lactis subsp. cremorisUC503 gene bank was maintained in E. coli ED8739, whichis McrA- and McrB- and as a result tolerates such methy-lation (35).The selection strategy employed to identify M ScrFI-

encoding plasmids frequently facilitates the cloning of largerfragments that carry both the MTase and ENase genes. In allthe instances described to date in which both genes havebeen cloned, the two have been linked (49). However,neither of the two M. ScrFI-carrying clones isolated in thisstudy exhibited any R ScrFI activity. It is possible that thesizes of fragments cloned were not sufficiently large toaccommodate both the ENase and MTase genes or that theENase was inactivated by the cloning procedure itself.Alternatively, it may be that these two genes are notadjacent to one another in UC503 or that the ENase genecannot be expressed in E. coli.As expected, no similarity between any type II restriction

ENases in GenBank and the M ScrFI protein was found.However, M- ScrFI contained all 10 predictive motifs nor-mally found in m5C MTases (34) with a variable region at theN terminus and between the conserved blocks VIII and IX.In particular, amino acid sequence comparison with otherm5C MTases revealed extensive regions of homology withthe MTases of NiaX, EcoRII, and Dcm. Especially signifi-cant were the tetrapeptide Gly-Phe-Gly-Tyr (GFGY) and the17-amino-acid peptide, which were highly conserved withinthe variable regions of all four proteins, perhaps suggestingsome common means by which these enzymes recognizetheir target sequence specificity (Fig. 7).The presence of two MTase genes with the same specific-

ity in one strain raises a number of interesting questionsregarding the origin and function of the genes. Systemsincorporating two MTase genes have been described previ-ously. The DpnII and HgaI R-M systems both consist of twoMTases linked to a single restriction ENase. In the case ofDpnII, the second MTase is a component of a complexcassette designed to allow genetic exchange within Diplo-coccus spp. while still maintaining protection against viralattack (6). HgaI, on the other hand, requires two MTases tomodify both sides of an asymmetric recognition site (43). In

VOL. 59, 1993

APPL. ENVIRON. MICROBIOL.

certain Bacillus strains, a number of MTases have beenfound to be prophage encoded (11, 16, 23). As temperatephages are also common in the lactococci, we were partic-ularly interested to determine whether one of the MTaseswas located on a prophage genome. However, attempts toinduce a temperate phage from UC503 with mitomycin C andUV failed (12), and neither pCI931M nor pCI932M hybrid-ized to a temperate phage induced from the closely related(though not identical) L. lactis subsp. cremoris UC509,which has also been observed to contain both M ScrFIgenes (data not shown).R-M systems active in commercially important lactococ-

cal strains are likely to play a significant role in the control ofbacteriophage proliferation during dairy fermentations (25).However, despite numerous reports describing R-M pheno-types in these bacteria, systematic searches within the genusfor classical type II ENase activities have met with very littlesuccess. It is interesting to note, therefore, that the twosystems that have been examined at a genetic level, LlaI andScrFI, both exhibit atypical features-LlaI appears to berelated to the less common type IIS subclass of enzymes(19), and there are two genetically linked though distinctMTase genes protecting against ScrFI digestion. Furtheranalyses of these and other lactococcal R-M systems arenecessary to determine the significance, if any, of theseobservations.DNA-protein interactions are a key feature of many im-

portant biological processes. The highly specific nature ofDNA recognition by MTases and the availability of a numberof different enzymes that recognize the same sequence makethese proteins attractive models for the study of such reac-tions. The cloning and sequencing of these two m scrFIgenes will facilitate comparison of the DNA and amino acidsequences and will help determine the functional domains ofthe proteins.

ACKNOWLEDGMENTS

This work was supported by the BAP (contract BAP-0008-IRL)and BRIDGE (contract BIOT-CT91-0263 [SSMA]) programs of theEuropean Community. In addition, we gratefully acknowledge theEuropean Molecular Biology Organization (EMBO), who supporteda short-term stay at the laboratories of Transgene, S.A., Strasbourg,France.

In this regard, we are indebted to A. Mercenier and the late J.-P.Lecocq, for welcoming R. Davis at their laboratory where some ofthese experiments were carried out and for their constant enthusi-asm and support. Thanks also to colleagues M. O'Regan and P. Slosfor many helpful discussions and to the other staff at Transgene.Special thanks to P. O'Reilly for technical assistance and to L.Burgess for the photography.

REFERENCES1. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extraction

procedure for screening recombinant plasmid DNA. NucleicAcids Res. 7:1513-1523.

2. Blumenthal, R. M. 1986. E. coli can restrict methylated DNAand may skew genomic libraries. Trends Biotechnol. 4:302-305.

3. Blumenthal, R. M. 1989. Cloning and restriction of methylatedDNA in Eschenchia coli. Focus 11:41-47.

4. Bolivar, F., R. L. Rodriquez, P. S. Greene, M. C. Betlach, H. L.Heynecker, H. W. Boyer, J. H. Crosa, and S. Falkow. 1977.Construction and characterization of new cloning vehicles. II. Amultipurpose cloning system. Gene 2:95-113.

5. Boyer, H. W., and D. Roulland-Dussoix. 1969. A complementa-tion analysis of the restriction and modification of DNA inEscherichia coli. J. Mol. Biol. 41:459-472.

6. Cerritelli, S., S. S. Springhorn, and S. A. Lacks. 1989. DpnA, amethylase for single-strand DNA in the DpnII restriction system

and its biological function. Proc. Natl. Acad. Sci. USA 86:9223-9227.

7. Chopin, A., M. C. Chopin, A. Moillo-Batt, and P. Langella.1984. Two plasmid-determined restriction and modification sys-tems in Streptococcus lactis. Plasmid 11:260-263.

8. Clark, P. 1987. Improved oligonucleotide directed mutagenesisusing M13 vectors. Methods Enzymol. 154:382-402.

9. Coffey, A. G., G. F. Fitzgerald, and C. Daly. 1991. Cloning andcharacterization of a plasmid encoding abortive infection ofbacteriophages from lactococcal plasmid pCI829. J. Gen. Mi-crobiol. 137:1355-1362.

10. Cohen, S. N., A. C. Y. Chang, and L. Hsu. 1972. Nonchromo-somal antibiotic resistance in bacteria: genetic transformation ofEscherichia coli by R-factor DNA. Proc. Natl. Acad. Sci. USA69:2110-2114.

11. Connaughton, J. F., W. D. Kalossk, P. G. Vanek, G. A.Nordone, and J. G. Chirikjuan. 1990. The complete sequence ofthe Bacillus amyloliquefaciens proviral H2, BamHI methylasegene. Nucleic Acids Res. 18:4002.

12. Costello, V. 1988. Characterization of bacteriophage-host inter-actions in Streptococcus cremoris UC503 and related lacticstreptococci. Ph.D. thesis. National University of Ireland (Uni-versity College, Cork), Cork, Ireland.

13. Daly, C., and G. F. Fitzgerald. 1987. Mechanisms of bacterioph-age insensitivity in the lactic streptococci, p. 259-268. In J. J.Ferretti and R. Curtiss III (ed.), Streptococcal genetics. Amer-ican Society for Microbiology, Washington, D.C.

14. de Vos, W. M. 1987. Gene cloning and expression in lactic acidbacteria. FEMS Microbiol. Rev. 46:281-295.

15. Fitzgerald, G. F., C. Daly, L. R. Brown, and T. R. Gingeras.1982. ScrFI: a new sequence-specific endonuclease from Strep-tococcus cremoris. Nucleic Acids Res. 10:8171-8179.

16. Gunthert, U., L. Reiners, and R. Lauster. 1986. Cloning andexpression of Bacillus subtilis phage DNA methyltransferasegenes in Escherichia coli and B. subtilis. Gene 41:261-270.

17. Hendrix, R. W., J. W. Roberts, F. W. Stahl, and R. A. Weisberg(ed.). 1983. Lambda II. Cold Spring Harbor Laboratory, ColdSpring Harbor, N.Y.

18. Higgins, D. L., R. B. Sanozky-Dawes, and T. R. Klaenhammer.1988. Restriction and modification activities from Streptococcuslactis ME2 are encoded by a self-transmissible plasmid, pTN20,that forms cointegrates during mobilization of lactose-ferment-ing ability. J. Bacteriol. 170:3435-3442.

19. Hill, C., L. A. Miller, and T. R. Klaenhammer. 1991. In vivogenetic exchange of a functional domain from a type II Amethylase between lactococcal plasmid pTR2030 and a virulentbacteriophage. J. Bacteriol. 173:4363-4370.

20. Hill, C., K. Pierce, and T. R. Klaenhammer. 1989. The conju-gative plasmid pTR2030 encodes two bacteriophage defensemechanisms in lactococci, restriction modification (R'/M') andabortive infection (Hsp+). Appl. Environ. Microbiol. 55:2416-2419.

21. Josephsen, J., and F. K. Vogensen. 1989. Identification of threedifferent plasmid-encoded restriction/modification systems inStreptococcus lactis subsp. cremoris W56. FEMS Microbiol.Lett. 59:161-166.

22. Kieny, M. P., R. Lathe, and J. P. Lecocq. 1983. New versatilecloning and sequencing vectors based on bacteriophage M13.Gene 26:91-99.

23. Kiss, A., G. P6sfai, C. C. Keller, P. Venetianer, and R. J.Roberts. 1985. Nucleotide sequence of the BsuRI restriction-modification system. Nucleic Acids Res. 13:6403-6421.

24. Klaenhammer, T. R. 1987. Plasmid-directed mechanisms forbacteriophage defense in lactic streptococci. FEMS Microbiol.Rev. 46:313-325.

25. Klaenhammer, T. R. 1989. Genetic characterization of multiplemechanisms of phage defense from a prototype phage-insensi-tive strain, Lactococcus lactis ME2. J. Dairy Sci. 72:3429-3443.

26. Le Bourgeois, P., M. Mata, and P. Ritzenthaler. 1991. Pulsed-field gel electrophoresis as a tool for studying the phylogeny andgenetic history of lactococcal strains, p. 140-145. In G. M.Dunny, P. P. Cleary, and L. L. McKay (ed.), Genetics andmolecular biology of streptococci, lactococci, and enterococci.

784 DAVIS ET AL.

TYPE II R-M SYSTEM IN L. LACTIS 785

American Society for Microbiology, Washington, D.C.27. Ludwig, W., E. Seewaldt, R. Kilpper-Balz, K. H. Schleifer, L.

Magrum, C. R. Woese, G. E. Fox, and E. Stakebrandt. 1985. Thephylogenetic position of Streptococcus and Enterococcus. J.Gen. Microbiol. 131:543-551.

28. Lunnen, K. D., J. M. Barsomian, R. R. Camp, C. 0. Card, S. Z.Chen, R. Croft, M. C. Looney, M. M. Meda, L. S. Moran, D. 0.Nwanko, B. E. Slatko, E. M. Van Cott, and G. G. Wilson. 1988.Cloning type II restriction and modification genes. Gene 74:25-32.

29. Marinus, M. G., and N. R. Morris. 1973. Isolation of deoxyri-bonucleic acid methylase mutants of Escherichia coli K-12. J.Bacteriol. 114:1143-1150.

30. Modrich, P., and R. J. Roberts. 1982. Type II restriction andmodification enzymes, p. 109-154. In S. Linn and R. J. Roberts(ed.), Nucleases. Cold Spring Harbor Laboratory, Cold SpringHarbor, N.Y.

31. Murray, N. E., W. J. Brammar, and K. Murray. 1977. Lamb-doid phages that simplify the recovery of in vitro recombinants.Mol. Gen. Genet. 150:53-61.

32. Nelson, M., and M. McClelland. 1989. Effect of site-specificmethylation on DNA modification methyltransferases and re-striction endonucleases. Nucleic Acids Res. 17(Suppl.):r389-r415.

33. Noyer-Weidner, M., R. Diaz, and L. Reiners. 1986. Cytosine-specific DNA modification interferes with plasmid establish-ment in Escherichia coli K12: involvement of rgtB. Mol. Gen.Genet. 205:469-475.

34. Posfai, J., A. S. Bhagwat, G. P6sfai, and R. J. Roberts. 1989.Predictive motifs derived from cytosine methyltransferases.Nucleic Acids Res. 17:2421-2435.

35. Raleigh, E. A., N. E. Murray, H. Revel, R. M. Blumenthal, D.Westaway, A. D. Reith, P. W. J. Rigby, J. Elhai, and D.Hanahan. 1988. McrA and McrB restriction phenotypes of someE. coli strains and implications for gene cloning. Nucleic AcidsRes. 16:1563-1575.

36. Raleigh, E. A., and G. Wilson. 1986. Escherichia coli K12restricts DNA containing 5-methyl cytosine. Proc. Natl. Acad.Sci. USA 83:9070-9074.

37. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring HarborLaboratory, Cold Spring Harbor, N.Y.

38. Sanders, M. E. 1988. Phage resistance in lactic acid bacteria.

Biochimie 70:411-421.39. Sanders, M. E., and T. R. Klaenhammer. 1980. Restriction and

modification in group N streptococci: effect of heat on develop-ment of modified lytic bacteriophage. Appl. Environ. Microbiol.40:500-506.

40. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequenc-ing with chain-terminating inhibitors. Proc. Natl. Acad. Sci.USA 74:5463-5467.

41. Schleifer, K. H., and R. Klipper-Bfilz. 1987. Molecular andchemotaxonomic approaches to the classification of strepto-cocci, enterococci and lactococci: a review. Syst. Appl. Micro-biol. 10:1-9.

42. Southern, E. M. 1975. Detection of specific sequences amongDNA fragments separated by gel electrophoresis. J. Mol. Biol.98:503-517.

43. Sugisaki, H., K. Yamamoto, and M. Takanami. 1991. The HgaIrestriction-modification system contains two cytosine meth-ylase genes responsible for modification of different DNAstrands. J. Biol. Chem. 266:13952-13957.

44. Tao, T., J. C. Bourne, and R. M. Blumenthal. 1991. A family ofregulatory genes associated with type II restriction-modificationsystems. J. Bacteriol. 173:1367-1375.

45. Taylor, G., and P. Quirke. 1989. Further applications of thepolymerase chain reaction. J. Pathol. 159:277-299.

46. Terzaghi, B. E., and W. E. Sandine. 1975. Improved medium forlactic streptococci and their bacteriophages. Appl. Microbiol.29:807-813.

47. van der Vossen, J. M. B. M., D. van der Lelie, and G. Venema.1987. Isolation and characterization of Streptococcus cremorisWg2-specific promoters. Appl. Environ. Microbiol. 53:2452-2457.

48. Wilson, G. G. 1988. Cloned restriction-modification systems-areview. Gene 74:281-289.

49. Wilson, G. G. 1991. Organization of restriction-modificationsystems. Nucleic Acids Res. 19:2539-2566.

50. Woodcock, D. M., P. J. Crowther, J. Doherty, S. Jefferson, E.DeCruz, N. Noyer-Weidner, S. S. Smith, M. Z. Michael, andM. W. Graham. 1989. Quantitative evaluation of Escherichiacoli host strains for tolerance to cytosine methylation in plasmidand phage recombinants. Nucleic Acids Res. 17:3469-3478.

51. Wu, J. C., and D. V. Santi. 1987. Kinetic and catalytic mecha-nism of HhaI methyltransferase. J. Biol. Chem. 262:4778-4786.

VOL. 59, 1993