Copyright 0 1996 by the Genetics Society of America

DNA Methylase Mutants of Salmonella Eylbhimurium and a Novel Dam-Regulated Locus

Joaquin Torreblanca andJosep Casadesh Departamento de Genttica, Facultad de Biologia, Universidad de Sevilla, E-41080 Sevilla, Spain Manuscript received March 13, 1996 Accepted for publication June 6, 1996

ABSTRACT Mutants of Salmonella typhimurium lacking DNAadenine methylase wereisolated they include insertion and deletion alleles. The dam locus maps at 75 min between cysC and aroB, similar to the dam gene. Dm- mutants of S. typhimun’um resemblethose of E. coli inthe following phenotypes: (1) increased spontaneous mutations, (2) moderate SOS induction, (3) enhancement of duplication segregation, (4) inviability of dam recA and dam recB mutants, and (5) suppression of the inviability of the dam recA and dam recB combinations by mutations that eliminate mismatch repair. However, differ- ences between S. typhimun’um and E, coli dam mutants are also found: (1) S. typhimunum dam mutants do not show increasedUV sensitivity, suggesting thatmethyldirected mismatch repairdoes not participate in the repair of UV-induced DNA damage in Salmonella. (2) S. typhimunum dam recJmutants are viable, suggesting that the Salmonella RecJ function does not participate in the repair of DNA strand breaks formed in the absenceof Dam .We also describe a genetic screen for detecting novelgenes regulated by Dam methylation and a locus repressedby Dam methylation in the S. typhimuriumvirulence (or “cryptic”) plasmid.

HE biological functions of DNA adenine methyla- replication (MESSERet al. 1985; SMITHet al. 1985; B~YE T tion have been widely investigated in Eschm’chia and L0BNER-OLESEN 1990; CAMPBELL and KLECKNER coli and its phages (reviewed by HAITMAN 1981; MARL 1990; ABELES et al. 1993), segregation of the daughter NUS 1984, 1987a,b; MESSERand NOYER-WEIDNER 1988; chromosome molecules (OGDENet al. 1988; HERRICKet BARRASand MARINUS 1989; NOYER-WEIDNERand al. 1994), regulation of plasmid replication (RUSSELand TRAUTNER1993). TheDNA of E. coli contains -1.5 mol ZINDER 1987; GAMMIEand CROSA1991), and the activity of &methyl-adenine per 100 mol of adenine (HAITMAN of certain host and phage genes (HAITMAN 1982; MARI- 1981; MARINUS 1984). Formation of &methyl-adenine NUS 1985; ROBERTSet al. 1985; BRAUNand WRIGHT 1986; results from postreplicative modification of adenine res- K~CHERERet al. 1986; BLYNet al. 1990; CAMPBELL and idues in 5‘-GATC-3’ sites; methylation occurs in both KLECJSNER 1990; BRAATENet al. 1994). In addition, Dam strands of the palindromic target (LACKS and methylation is involved in the initiation of P1 DNArepli- GREENBERG1977). The methylation reaction is cata- cation (ABELES et al. 1993), packaging of bacteriophage lyzed by the DNA adenine , P1 DNA into virions (STERNBERGand COULBY1990) and often called “Dam methylase” (HERMANand MODRICH also affects functions associated with bacterial retropo- 1982). The dam gene is located at 74 min on the E. coli sons (HSUet al. 1990). This variety of processes affected genetic map and is cotransducible with cysG (MARINUS byDam methylation confirms earlier expectations on 1973; BACHMANN1990). The dam gene is part of an its relevance as a physiological signal (HAITMAN 1981; operon containing aroK, aroB, trpS and two additional MARINUS 1987a,b). genes involved in carbohydrate metabolism (LBBNER- A Salmonella typhimurium point mutant lacking DNA OLESENet al. 1992; LYNGSTADAASet al. 1995). adenine methylase was described a decade ago(RICHTIE In E. coli and its phages Mu and P1, many rolesof Dam et al. 1986). Given the pleiotropy of dam mutations of methylationhave been identified. Upon chromosome E. coli and the intricacies found in their manipulation replication, the existence of hemimethylated 5’-GATG (MCGRAWand MARINUS 1980; PETERSONet al. 1985; PE- 3’ sites directs mismatch repair toward the newly synthe- TERSON and MOUNT 1993), we judged that thestudy of sized, unmethylated strand (Lu et al. 1983; PUKKILAet al. Dam methylation in Salmonella might become easier 1983; RADMANand WAGNER 1986;MODRICH 1987). Dam if insertion and deletion alleles were available. Both methylation also controls the initiation of chromosome types of alleles are described in this paper. In addition, a detailed characterization of S. typhimun’um Dam- mu- tants shows that they share with those of E. coli many Corresponding author:Josep Casadesk, Departamento de Genktica, Facultad de Biologia, Universidad de Sevilla, Apartado 1095, Sevilla defects in recombination and DNA repair; however, dif- 41080, Spain.E-mail: [email protected] ferences are also found. The latter are intriguing and

Genetics 144: 15-26 (September, 1996) 16 J. Torreblanca and J. Casadesus

potentially insightful, like other examplesof divergence "locked-in" P22-Mu hybrid confemng chloramphenicol resis- between Escherichia and Salmonella (RILEY and Sm- tance (YOUDEW et al. 1988; BENSONand Gomm 1992). The insertion allele zu-6?05 ::Tn5[lexA :: carries a defective Tn5 DERSON 1990). h-4 derivative bearing a transcriptional fusion between the S. whzmu- The last part of this study describes a genetic screen- Gum /operator region and the E. coli lacZ gene ing for the detectionof genes regulatedby Dam methyl- ( GARRJGA1992). ation and the finding of a locus strongly repressed by Mediaand growth conditions: The E medium of VOCEL Dam methylation, located in the S. typhimurium viru- and BONNER (1956)was used as the standard minimal me- dium. NCE is E medium without citrate. Carbon sources were lence plasmid. either 0.2% glucose or 1% lactose. The rich medium was nutrient broth (8 g/l, Difco) with added NaCl (5 g/l). Mac- MATERIALSAND METHODS Conkey agar base was from Difco. Solid media contained Difco agar at 1.5% final concentration. Auxotrophic require- Bacterial strains, bacteriophagesand strain construction: ments and antibiotics were used at the final concentrations The S. typhimun'um strains used in this study are listed in Table described by WOY(1990). Green plates were prepared ac- 1. Transductional crosses using phage P22 HT 105/1 int201 cording to CW et al. (1972), except that methyl blue (Sigma) (SCHMIEGER1972; G. ROBERTS,unpublished data) were used substituted for aniline blue. For the selection of tetracycline- for strain construction operations involving chromosomal sensitive derivatives of Tet' strains, we used the medium of markers; the transducing phage will be henceforth referred as BOCHNERet nl. (1980) as modified by MAI>oY andNUNN "P22 HT". Phage sensitivity was tested by cross-streaking with (1981). the clear-plaque mutant P22 H5. To obtain phage-free isolates, Transposon substitutions: A lysate grown on strain TT10425 transductants were purified by streaking on "green" plates. was used for transposon replacement at the dam locus. This Strains SV3019-SV3030 arose from matings between the E. coli strain carries an F prime containing TnlOdKan (see the strain donors CClOl -CC106 (CUPPLESand MILLER1989) and either list). The lysatewas irradiated with UV light, using a 15 W of the S. typhimun'um recipients SV3016 or SV3017. In these Sylvania lamp at a distance of 30 cm for 30 sec; irradiation of crosses, F-prime transfer was scored on E plates, selecting Pro' the phage suspensions can be expected to increase recombina- Tet' or Pro' Cam' transconjugants; growth in the absence of tion inthe transductants (RuPP et al. 1971).Transductions proline selected for F-prime transfer, while the presence of the selecting the incoming marker (Kan') were carried out. Kan' antibiotic counterselected the E. colz donors. Strain UA1570, transductants were then scored for loss of the resident marker obtained from Xavier Garriga (Dep. Microbiologia, Universitat (Tet'). This procedure allowed us to obtain allele variants Autbnoma de Barcelona, Bellaterra, Spain), carries a Tn5 chro- tagged with different antibiotic resistance markers (e.g., Tet' mosomal insertion that has not been mapped; we have named and Kan'). Replacement of a lacZallele with the lacZ477::TnlO- this insertion zu-6305, following the nomenclatureof CHUMLEY net insertion was achieved using a lysate grown on strain et al. (1979) and using a z- allele number assigned to our TT16716. The lysate was W irradiated as above; Tet' transduc- laboratory by the Salmonella Genetic Stock Center (SGSC), tants were selected on lactose indicator plates supplemented University of Calgary, Alberta, Canada. The nature and the with tetracycline. origin of this Tn5 insertion are described in the following Measurement of spontaneous mutation rates using lac al- section. Using the same nomenclature system, a Mug insertion leles: Aliquots containing lo6 cells were added to six tubes in an unknown locus of the S. typhimun'um virulence plasmid of NCE liquid medium containing 0.2% glucose. The cultures has been named 2zu-6306::MudJ. were incubated at 37" until saturation (- lo9 cells/ml). Cells Plasmids and transposons:F'128 pro+ lac' nf1836::TnlO&an1 were then spread on NCE-lactose plates (10' cells per plate, is an E. coli episome carlying a defective TnlO element that four plates per culture). Revertants were scored after 48-hr confers cloramphenicol resistance (EILIO~and ROTH 1985). incubation at 37". Viable counts were carried out on NB pNK972 (Amp') is a pBR332 derivativecarrying the ISlOtranspo- plates. sax gene under the control of a tac promoter (BENDERand Mutagenesiswith TnZOdCam and isolation of insertions KLECKNER 1986). The E. coli episome F128 pro' lac+ .zf linked to the dum locus: We used the "nonhomologous trans- 1831 ::TnlOaet and the S. typhimurium episome F'152 nu& 2zf duction" procedure (ELLIOTTand ROTH 1985), with modifi- 18?3::TnlU&n were both obtained fromJ. R. ROTH, Depart- cations described elsewhere (FLOWSand CASADESUS1995). ment of Biology, University ofUtah, Salt LakeCity, Utah. Plasmid A lysate grown on strain SV2056 was used to transduce strain pTPl66, obtained from M. G. MARINUS (Department of Pharma- SV3002, selecting Cam' Amp' transductants.Transducing cology, University of Massachusetts, Worcester, Massachusetts), mixtures were made on NB plates and preincubated for 4-6 is a pBR322 derivative carrying the E. coli dam gene under the hr at 37" before replica-printing to NB supplemented with control of a tac promoter (MARINUS et al 1984). Repression of chloramphenicol and ampicillin. Cam'Amp' transductants the tux promoter was achieved by using the episome F pro' were replica-printed several (more than three) times to NB- lacP L8Z::TnIOA+HHl04, obtained from J. R ROTH. Plasmid Tet plates containing EGTA 10 mM; the latter was added to pGElO8 (Kan') is a ColEl derivative carlying a cea::lacZ fusion prevent reinfection (thus allowing the isolation of phage-free (SAI.IESet al. 1987). Plasmid pSE143 (Kan') is a pSClOl deriva- derivatives). Pools of 1000-2000 Cam' Tet' colonies were tive canying a umuDC::lacZfusion (ELLEDGEand WALKER 1983). made and lysed with phage P22 HT. The pools were then pIZ53 is a pUCl9 derivative canying the internal Hind111 fi-ag- used to transduce strain LT2; to detect cotransduction of the ment of Tn5; this fragment includes the kanamycin-resistance TnlOdCam and TnlOcEet elements, Tet' transductants were gene (MALDONADO et al. 1992). Mud is the specialized transduc- replica-printed to NB plates supplemented with chloram- ing phage Mudl (Amp Lac &62) originally constructed by CASA- phenicol. DABAN and COHEN(1979). Mud4 is a transpositiondefective Transposition of Mugand random formation of transcrip derivative of Mudl (HUGHESand ROTH 1984). Mud1734[Kan- tional lac fusions: A P22 HT lysate grown on strain TT10288 Lac] (C~TILHOet al. 1984) is a transpositiondeficient Mu deriva- (HUGHES andROTH 1988) was used to transduce TT1704, tive that generates operon fusions upon insertion; the element selecting kanamycin resistance. The recipient strain carried has been renamed Mug (HUGHESand ROTH 1988). MudQ is a the nontransducible deletion Ahis-95??. Transductants were Dam Methylation in S. typhirnuriurn 17

TABLE 1 Bacterial strains

Strain Genotype or phenotype Reference"

JC608 darn-201::TnlOflet rpsL1 JC609 dam-202:TnIOflet rpsLl JC610 darn-203::TnIOflet rpsLl SV80 Ahis-3050 SV1223 recB503:TnlO GARZONet al. (1996) SV1234 recF522:Tn5 A. GARZON SV1240 uz-6305Tn5 [lexA::lacz] A. GARZON SV1244 recAlrecB497::MudJ GARZONet al. (1995) SV2056 Ahis-3050/F' 128 lac+ pro' 2zf1836::TnlOdCam FI.ORES and CAsADEsirs (1995) SV3000 dam-201::TnlOflet SV300 1 dam-201::TnIOdKan SV3002 rlam-ZOI::TnIOflet/pNK972 SV3003 zzw6306::MudJ SV3004 DUP [tq2482 *Mudl-8* hisD99531 SV3005 DUP [trp2482 *Mudl-8* hisD99531 dam-2OI::TnlCMTet SV3006 Adam-204 SV3007 dam-201::TnIOflet wz-6305Tn5 [lexA::lacz] SV3008 Adam-204 m-6305Tn5 [lexA::hcZJ SV3009 LT2/pTP166 SV30 10 DUP [tq2482 *Mud** hisD99531 damA204 SV3011 Adam-204 mutL1II::TnlO SV3012 Adam-204 mutSl21::TnlO SV30 13 Adam-204 mutH101::Tn5 SV3014 Adam-Z04/pTP166 SV3015 DUP [trp-2482 "Mudl-8* hisD99531 dad204 mutHIOl::Tn5 SV30 16 poA692:MudQ SV3017 dam-Z01::Tn10flet proA692:MudQ SV30 19 proA692:MudQ/F' lac2101 proAF SV3020 poA692:MudQ/F' lacZl0.2 poAP SV302 1 proA692:MudQ/F' hcZ103 poAP SV3022 proA692:MudQ/Ft lac2104 proAB+ SV3023 proA692:MudQ/F' lacZI05 PoAB' SV3024 proA692:MudwF' lacZl06 proAF SV3025 dam-201::TnIOflet proA692:MudQ/F' lacZl01 poAF SV3026 dam-201::TnIOflet proA692:MudQ/F' lac2102 POALP SV3027 darn-201::TnIOflet proA692:MudQ/F' lacZl03 proAB+ SV3028 darn-201::TnlOflet proA692:MudWF' he2104*OAF SV3029 dam-201::TnIOflet proA692::MudQ/F' lacZ105 proAP SV3030 dam-201::TnIOflet proA692:MudQJF' lacZ106proALP SV3031 DUP ttlp-2482 *Mudl-8* hid99531 mutHlVl::Tn5 SV3069 uv-6306::MudJ dam-201::TnIOflet SV3070 zzv-6306::MudJ Ahis-3050 SV307 1 uw6306::MudJ Ahis-3050 dam-201::TnIOflet SV3072 uw6306::MudJ Ahis-3050 SV3073 2zw6306::MudJ Ahis-9533 dam-201::TnlOflet S\'3074 uv-6306:MudJ [lacZ::TnlOflet] TR5527 Ahis-712, cured of the virulence plasmid J. R. ROTH TR5667 ysG439 rpsLl J. R. ROTH TR5878 r(LT2)- m(LT2)- r(S)+ ilv-542 met A22 trpB2 Fels2- fZiA66 SGSC' rpsLl20 91-404 metE551 hspL56 hspS29 TT1704 Ahis-9533 J. R. ROTH TT2742 aroB542:Tn5 SGSC' TT9286 proAB47 led798 argI537 aru-9fol-I/F'128 proA+B+ argP J. R. ROTH lad9 Z+ Y+ A+ TT10425 nadA56/F'152 nad+ zf1833:TnlOdKan J. R. ROTH TT10'288 hisD9953:MudJ hisA944:Mudl HUGHESand ROTH (1988) TT10838 recA I J. R. ROTH TT11289 recAl srl-203:TnIOdCam J. R. ROTH TT15278 recJ504:MudJ MAHANet al. (1992) TT16716 his-644 pro-621/F' eo' lacZ477::TnIOflet SGSCb UA1570 uz-6305:Tn5 [le&.:lacZ] Rif X. GARRIGA Omitted for strains first described in this study. ' SGSC: Salmonella Genetic Stock Center, Department of Biological Sciences, University of Calgary, Calgary, Alberta, Canada T2N 1N4. 18 J. Torreblanca and J. Casadesk selected on lactose indicator media (MacConkey-lactose or X- MboI is blocked by Dam methylation, while DptI only cuts gal NB plates) supplemented with kanamycin. Rare Km' Ap' methylated DNA, cutting by Sau3AI is irrespective of the transductants generated by double-fragment transduction are methylation Status. easily identified because they are Kan' Amp', while transduc- Transformation of S. typhimurium: The transformable strain tants carrying MudJ "hops" are Kan' Amps (HUGHESand TU878 was used as the recipient of plasmids; preparation of ROTH 1988). Kan' transductants were made DhaPe-free bv I" competent cells and transformation followed the procedures replica-printing (more than two times) to plates containing Of LEDERBERGand COHEN (1974). Plasmids transformed into 10 mM EGTA. TU878 were transferred to suitable recipients by transduction Transduction and cotransductional mapping: Preparation with P22 HT. of phage P22 lysates and transductions were performed ac- Extraction of the S. typhimurium virulence plasmid One cording to GARZONet ul. (1995). The efficiency of transduc- milliliter of' an overnight culture in LBwas centrifuged at tion (EOT) of a given lysate is the ratio between the numbers 12,000 rpm for 2 min at 4". The pellet was resuspended in of transductants obtained on a pair of isogenic strains. For 150 /11 of E buffer; 300 p1 of lysis solution were then added. cotransductional mapping, transductants were replica-printed After incubating at 65" for 1 hr, the lysate was chilled on ice to suitable plates to score transduction of unselected markers. and shaken for 10 min (until a white precipitate was formed). Cotransduction frequencies are averages of more than four The preparation was then buffered by adding 150 pl of ice- independent crosses, scoring at least 300 transductants from cold 2 M Tris, shaken gently until it became transparent and each cross. The relative order of markers was determined centrifuged at 12,000 rpm for 20 min in the cold. The super- by three-factor crosses. Table 1 contains the original strains natant was transferred to a clean tube and mixed with one carrying markers cotransducible with the loci studied but not volume of nonsaturated pheno1:chloroform:isoarnylalcohol their derivatives constructed by adding oneof the alleles to be (25:24:1). After two to three extraction cycles, DNA was pre- mapped; such constructions were required toperform certain cipitated with 3 M sodium acetate and absolute ethanol. The three-factor crosses. pellet was rinsed with 70% ethanol and resuspended in 10 pI UV survival assays: Overnight cultures made in NB were of minimal TE. All preparations were treated with ribo- diluted 1/10 in the same medium. When the cultures reached nuclease (0.1 mg/ml, final concentration) before storage at an O.D.hoo= 0.5, the cells were harvested and z-euspended - 20". in E buffer (E medium without glucose). Five-milliliter ali- DNA hybridization: DNA hybridization followed the proce- quots were transferred to sterile, empty Petri dishes. Irradia- dures described by SAMBROOKet nl. (1989). DNAwas trans- tion was achieved by opening theplates under a 15 W Sylvania ferred to a nylon membrane using a vacuum blotting system UV lamp at a distance of 30 cm in the absence of daylight (TransVac TE80, Hoeffer Scientific Instruments) and cross- illumination. Cell suspensions were stirred during irradiation. linked by UV irradiation. The probes were labeled by random After serial dilution in foil-covered tubes, irradiated cultures priming with chemiluminescent digoxigenin-dUTP (Boeh- were plated on NB. ringer Mannheim); hybridization bands were visualized on an Analysis of duplication segregation: A single colony from X-ray film. a selective plate was used to inoculate 2 ml of nonselective NB broth, grown overnight, diluted and plated on nonselec- RESULTS tiveNB agar. Whencolonies appeared, they were replica- printed to plates selective for cells carrying the duplication Isolation and characterization of Dam- mutants of S. (NB supplemented with ampicillin). The percentage given refers to the fraction of colonies that lost the duplication. Eyphimuriwm: TnIUflet insertions in the S. typhimurium. Rapid mapping with Mud-P22 prophages: We followed the dum locus were identified following localized mutagenesis procedure of BENSONand GCKDMAN(1992), using a collec- of the region known to include this gene in 1'5 coli. In the tion of 67 "locked-in" Mud-P22 prophages (BKNSONand latter, dum is located between qsG and aroB at min 74 GOLUMAN1992; FI.ORESand CASADESUS 1995).This collection (MARINUS 1973; BACHMANN1990), a regon that corre- not included in the strain list. is sponds tomin 75 in S. typhimurium (SANDERSONet al. /%galactosidaseassays: Levels of P-galactosidase were as- sayed as described by MII.I.ER(1972), using the CHCl+odium 1995). In our search for TnIUflet insertions in dum, the dodecyl sulfate permeabilization procedure. qsG gene was used as the linked marker because muta- Isolation and purification of genomic DNA: A pellet from tions in the closer gene aroB often cause a leaky Dam- a 5 ml culture grown in NBwas resuspended in 1 mlTE- phenotype (data notshown) ; a similar effectof uroBmuta- glucose (25 mM Tris-HC1, 10 mM EDTA, 50 mM glucose pH tions has been described in E. coli (LOBNEK-OLESENet u1. 8.0) and 0.1 ml of a mixture containing lysozyme, proteinase K and ribonuclease A (final concentrations: 5 p,g/ml, 0.05 1992). Regional mutagenesiswith Tn IUctret was achieved pg/ml, and 0.5 pg/ml, respectively). After 30 min of incuba- in two steps: (1) nine pools of TnlOflet insertions were tion at 37", 0.1 ml SDS (20%) was added and incubation was generated in the wild-type strain LT2;each pool was made continued for another 30 min. DNA was sheared by passing from 2500-3000 independent Tet' colonies; (2) P22 HT once through a syringe and extracted once with phenol, twice phage grown on the pooled cells was used to transduce with phenol-chloroform and twice with chloroform. DNA was precipitated at -70" by adding 2.5 volumes of absolute etha- strain TR.5667, selecting tetracycline resistance. Prototro- nol and 1/10 (v:v) of 3 M sodium acetate.After centrifugation, phic (CysC') transductants were scored by replica-print- the DNA pellet was resuspended in 0.2 ml TE buffer (Tris- ing to minimal mediumcontaining tetracycline. Selection HCI 10 mM, EDTA 1 mM pH 7.6). for prototrophy eliminated potential candidates canying Discrimination of the methylation status of adenine resi- uroB mutations polar on dum. The total number of Tet' dues in genomic DNA: Genomic DNA preparations were di- Cy&+ transductants obtained was 889. These were then gested with restriction Sn&AI, DpnI and MboI (all from Boehringer Mannheim). All these enzymes recognize examined for slow growth and/or abnolmal colony mor- thesequence 5'-GATC-3'. The endonucleolytic activity of phology on green plates (Dan- mutants of E. coli show D am Methylation in Methylation Dam S. /y~~hirnuriurn 19

1 2 3 4 5 6 7 8 9 101112

FI(;I'KI.:2.-A mistulr ofcolonics IOrmctl by isogcnic I~IIII'. I and Dam- strainsof S. /y/~himurium(LT2 and SVBOOO, rcspcc- FIGURE1.-Digestion of genomic by restriction en- tively) on green plates. The wild type forms glossy convex zymeswith differentresponses LO themethylation state of colonies, while the colonies formed by the Dam strain are their targets (see the text for details). Lanes arefollows: as 1 - dull and flat. 4, Dam' (LT2); 5-8, Dam- (SVSOOO); 9-12, Dam-/pTPl(iG (SV3014). Lanes 1, 5 and 9 are controls of undigested DNA, scored for loss of tetracycline resistance and for the lanes 2, 6 and 10 contain Snu3AI digestions; lanes 3, 7 and maintenance of phenotypes characteristic of Dam- mu- 11 contain Mi101 digestions; lanes 4, 8 and 12 contain Q1n1 digestions. Lanes at both gel edges containDNA size markers tant.. (flat colony shape, distinct restriction pattern of (HindlII-digested lambda DNA). genomic DNA). One of the substitution isolates was the origin of strain SVSOO1. abnormal cellmorphology: see BARWS and MARINUS Isolation of TnlOdCam insertions linked to the dum 1989). locus: Pools of random TnlOKam insertions were Twenty candidates forming abnormal colonies were made on strain SVSOOO; insertions linked to the the physically analyzed by digestion of genomic DNA prepa- mutation dam-201 ::Tn IOdTetwere detected by cotrans- rations with DpnI, Sau%4I and MlmI. All duction, using strain LT2 as the recipient. Two indepen- these enzymes recognize the sequence 5'-GATGS', but dent insertions linked to the dam locus were obtained; Mho1 activity is blocked by Dam methylation, while DpnI the closest was z/lf-6304,>90% linked to dam (see Fig- only cuts methylated DNA, Sau3AIcuts both methylated ure 3). and umethylated DNA. Threeindependent isolates Isolation and characterization of deletion mutants: (each obtained from a different pool) proved to be Tetracycline-sensitive derivatives of strain SVSOOO were Dam-, as judged from their restriction patterns. These isolated on Bochner-Maloy plates ( BOCHNERPI al. 1980; isolates were the origin of strains JC608,JC609 and MALOYand NUNN1981). Although the selection agar JC610; their dam alleles were designated dam-201, dam- contains tryptone and yeast extract, aroB mutants are 202 and dam-203, respectively.The three strains looked unable to form colonies on this medium unless supple- identical in genetic and physical tests and their South- ern hybridization patterns against a TnlO probe were zhf-6304 identical (datanot shown); thus they were probably Y5G dam aro B generated by Tn 10fletinsertions in the same site. For 0.17 further work, the allele dam-201 ::TnIOof strain JC608 b was transduced to LT2, giving riseto strain SV3000 (see 0.92 its DNA restriction pattern in Figure 1, lanes 5-8). This 0.20- ." strain was used as the standard Dam- insertion mutant 0.93 0.23 of S. typhimun'um. 4 - 0.95 On green plates, Dam- mutants of S. typhimun'um 0.11 - form flat, dull colonies thatare easily distinguished 4 0.95 from the convex, glossy colonies of the wild type (Figure 0.04 2). This distinct colony morphology is a reliable trait b and can be used for strain construction. 0.09 For the generation of allele variants, a lysate grown - on strain TT10425 was UV-irradiated and used to trans- FIGURE%--Genetic map of the q.sGdam-nroB region of duce SV3000. Kan' transductants appeared at frequen- the S. typhirnurium chromosome, constructedby P'L2-mcdiated cies aroundper p.f.u.; Kan' transductants were transductional crosses. 20 andJ. Torreblanca J. Casadesus mented with “aromatic mix”. This selection permits nor (TT11289) was used to transduce isogenic Damf the isolation of Tets derivatives carrying only smalldele- and Dam- recipients. The srl and recA loci are 50% tions, because their extentis constrained by aroB at one linked (SANDERSONand ROTH 1983). Cotransduction side and by the essential gene trpS at the other. One of recA and srl was detected by scoring UV sensitivity. Tets derivative of strain SV3000 was the origin of strain The combination recA dam was judged inviable because SV3006; its deletion allele was named Adam-204. Geno- 300/300 Cam‘ (Srl-) transductants were W. Thus, as mic DNA from this strain was still able to hybridize in E. coli (PETERSONet al. 1985), the combination recA against a dam probe (anXbaI-PvuII fragment of plasmid dam is inviable in Salmonella. pTP166). However, aband shift was observed, sug- In E. coli the inviability of dam mutations in combina- gesting that the dam gene had suffered a partial dele- tion with certain recombination functions has been at- tion (data not shown). Furtherevidence that Adam-204 tributed to the needof recombination to repair double- is a deletion allele was obtained during the study of a strand breaks derived from MutHLScatalyzed incisions lac fusion repressed by Dam methylation (see below). (WmG and SMITH1986). An analogous picture can be Complementation of dam mutations of S. Eylbhjmu- drawn for S. typhimurium, because the presence of mutH, hum by the E. coli dam+ gene: Plasmid pTP166 was mutL or mutS mutations in the Dam- recipient permit- transferred to strain TR5878 by transformation, select- ted the isolation of RecB- transductants in P22-medi- ing ampicillin resistance. The resulting strain, TR5878/ ated crosses (Table 2). RecA- Dam- strains can be like- pTP166, was used as donor to transduce pTP166 to wise constructed if the recipient contains a mutH allele various recipients in crosses mediated by P22 HT. When (data not shown). plasmid pTP166was introduced in S. typhim~~z~mDam- Lack and overproductionof DNA adenine methylase mutants (e.g., SV3000 or SV3006), it complemented the cause hypermutability: Rates of spontaneous mutation dammutation: all the Amp‘ transductants regained both to rifampicin resistance were compared by plating early the wild-type colony shape (not shown) and a wild-type stationary cultures on NB plates supplemented with ri- pattern of digestion by endonucleases DpnI, Suu3AI and fampicin. The frequency ofRif ‘ mutants increased MboI (see lanes 10-12 in Figure 1). from in the wildtype to 9-15 X Dam-in Genetic mapping of dam mutations: Two- and three- strains. However, the highest mutation rates (>400-fold factor crosses werecarried out by P22 HT transduction; higher than the wild type) were observed when the E. linkage of unselected markers was scored by replica- coli dam gene was introduced into S. typhimurium on a printing. Preliminary mapping of the dam locus was multicopy plasmid. Thus lack of Dam methylase causes performed using the alleles dam-201 ::Tn10net, dam- a milder hypermutability than Dam overproduction, 202::TnlOaet and dam-203::TnIOflet. Further work, suggesting that faster DNA remethylation perturbs mis- including strain constructions for three-factor crosses, match repair moreseverely than the absence of methyl- was carried out using the alleles dam-201 ::TnlUnet ation. Similarresults have been reportedfor E. coli and Adam-204,as well as the linked markers ar- ( WNUSet al. 1984). oB542 ::Tn5, cysG439and zhf-6304: :Tn 1OdCam. The re- The mutation pattern of Dam- strains was examined sulting genetic map is shown in Figure 3. It must be using the lac2 allele collection constructed by CUPPLES noted thatlinkage between aroB and dam mutant alleles and MILLER(1989). Six F’ lac pro+ episomes, each con- must be established using the antibiotic-resistance of a taining a known base substitution in the lucZgene, were TnlOinsertion in dam, because the aroB542::Tn5 muta- introduced in isogenic Dam’ and Dam- strains of S. tion causes a leaky Dam- phenotype (data not shown). typhimurium. The six lacZ alleles cover all six possible Viability of dam mutations in Combination with other base substitutions (CUPPLES andMILLER 1989); the ge- mutations: Viability tests involvedtwo types of transduc- notypes of the strains used (SV3019-SV3030) are given tional crosses: in the strain list. The presence of a dum mutation in the (1) Insertionalleles (e.g., recD541 ::TnlOdCam, background increased around eightfold the reversion recF522::Tn5,recB503::Tn10, and recJ504::MudJ) rates of two alleles (lac2102 and lacZ106) but had little were transduced to an isogenic pair ofDam’ and or no effect on the other members of the lac2 allele Dam- recipients. A given combination was judged collection. These results indicate that the increase of inviable whenever the frequencyof transductants was Lac+ reversion observed ina Dam- host is clearly biased >lOOO-fold reduced.The results of theseexperi- toward transition mutations, as in E. coli (GLICKMAN ments, summarized in Table2, indicate thatdam mu- 1979). tations of S. typhimurium are inviable if combined Effect of dam mutations onSOS induction: SOS activ- with recB mutations, but not with recD,recF or red. ity was tested in lexA+/lexA::lacZ merodiploids (strains Notea relevant difference between S. typhimu~um SV1240,SV3007 and SV3008), constructed by using a and E. coli: in the latter, the combination dum recJis TnS-borne kd::lac fusion(GARRIGA 1992). Measure- inviable (PETERSONet al. 1985). ments of @-galactosidase activityindicated that dam muta- (2) A lysate grown on a recA 1 srl-203: :Tn I OdCam do- tions of Salmonella, like their E. coli counterparts (PE- Dam Methylation in S. typhimurium 21

TABLE 2 Viability of S. fyphimurium dam mutations in combination with other mutations

Transduced allele GenotypeRecipient of theEOT“ recipient recB497::Mug LT2 dam’ 1 SV3000 dam-201::TnlOaet < 10-5 SV3006 Adam-204 < 10-~ SV30 11 Adam-204 mutLl1l::TnlO 0.33 SV30 12 Adam-204 mutSI21::TnlO 0.45 recB503:Tn IO SV3013 Adam-204 mutHlOI::Tn5 1.2 recD541::Tnl OdCam LT2 dam” 1 SV3000 dam-201::TnlOaet 1.51 recF522:Tn5 LT2 dam+ 1 SV3000 dam-201::Tn106Iet 0.21 LT2 dam+ recJ504:MudCam LT2 1 SV3000 dam-201::Tn SV3000 1Oflet 0.49 EOT (efficiency of transduction) is the ratio between the numbers of transductants obtained on a pair of isogenic strains.

TERSON et al. 1985), undergo moderate (two- to threefold) derepression of Zed . Derepres- sion was not only observed for Zed; plate tests with plas- 1- mid-borne umuDC::lac and cea::lac fusions (ELLEDGE dam+ and WALKER 1983; SALLESet al. 1987) also showed in- 10-1 - dam-201::TnlOdTet creased expression in a Dam- background (datanot H shown). Thus S. typhimurium dam mutations cause mod- Id 10-2 - erate SOS induction, like those of E. coli (PETERSONet * recJ5W::MudJ al. 1985). ‘5 - W sensitivity assays: Dam- mutants of E. coli are m slightly UV-sensitive at low UV doses (GLICKMANet al. 1978), but theirUV sensitivity increases with respect to wildtype at higher UV doses (MARINUS and MORRIS 10-5 1974). We examined theUV sensitivity ofS. typhimurium 1 Dam- isolates and observed a difference with their E. 0 10 20 30 coli counterparts: Dam- mutants of Salmonella are UV- UV fluence (J / m2) resistant at both low and high UV doses (Figure 4). Given the differences in UV mutability between E. coli and Salmonella (reviewed by EISENSTADT1987) and the presence of umuDGlikegenes in the S. typhimuriumviru- lence (sometimes called “cryptic”) plasmid (NOHMIet al. 1991), we examined the possibility that the UV resis- tance of S. typhimurium Dam- mutants might involve plasmid-borne function(s). For this purpose, we con- dam+ structed a Dam- derivative ofa strain cured of the viru- dam-201::TnlOdTet lence plasmid and compared its UV sensitivity withthat of theparental strain, TR5527.Both UV sensitivity recB503::TnlOdTet curves were similar and roughly identical to those of Dam+ and Dam- LT2 derivatives shown in Figure 4. Thus Dam- strains of S. typhimurium may be intrinsically I, resistant to UV radiation. Or, if alternative functions 0 40 80 120 exist, they must lie in the chromosome and not in the UV fluence (J / m2) virulence plasmid. Effect of dam mutations on duplication segrega- FIGURE 4.-UV sensitivity curves of Damt and Dam- strains tion: Duplication segregation was examined using a of S. typhimurium (LT2 and SV3000, respectively). To illustrate the UV sensitivity levels,three isogenic strains carrying recom- test for homogenote formation formally similar to that bination mutations are also included (TT15278, TT10838 and of MARINUSand KONRAD(1976) ; the only difference is SVl223). 22 J. Torreblanca andJ. Casadesds

TABLE 3 Effect of dam mutations on the segregation of chromosomal duplications

Strain Genotype Strain % of segregants‘’ SV3004 DUP [ trp2482 *Mudl-8* hisD99531 67 SV3005 DUP [ trp-2482 *Mudl-8* hidl99531dam-201::Tn 1OdTet 97 SV30 10 DUP [trp2482 *Mud-8* hisD99531Adam-204 99 SV3031 DUP [ trp-2482 *Mudl-8* hisD99531 mutHlOl::Tn5 33 SV30 15 DUP [ trp-2482 *Mudl-8* hisD99531Adam-204 mutHlOl::Tn5 40 “Percentage of Amp’ colonies; average of three experiments. that we used chromosome merodiploids instead of F- is made Dam- by transduction of the insertion dam- prime heterogenotes.Isogenic Dami and Dam- pairs of 201 ::TnIOaet. This test intends to confirm that the strains carrying Mud-induced duplications with known change in color is caused by the dum mutation (and not endpoints were grown nonselectivelyin NB until satura- by another mutation carried by the Dam- recipient). tion. Absence of antibiotic selection permits segrega- (2) Reconstruction crosses, in which the fusion is tion of the Mud-held duplication (HUGHESand ROTH transduced to isogenic Dam+ and Dam- derivatives 1985). Colonies were isolated on NB plates and haploid (LT2 and SV3006, respectively). Reconstruction in re- segregants lacking the Mud-encoded antibiotic resis- cipients different from the pair of original strains in- tance (Amp”)were detected by replica-printing (FLOES tends to confirm that the fusion is regulated by Dam and CASADESUS1995). methylation irrespective of the backgroundof the recip- Table 3 shows that duplication segregation increases ient. in a Dam- background; the increase can be suppressed (3) Complementation with a cloned dum+ allele (car- by a mutH mutation. These results suggest that dum ried on plasmid pTP166). This test can confirm that mutations enhance duplication segregation via mis- the fusion is regulated by Dam methylation (and not match repair, probably because the MutH endonucle- by a gene located downstream of dam). This test is ase introduces nicks or double-stranded breaks in un- necessary whenever the dum alleles used are potentially methylated DNA. Because the activated form of MutH polar (insertions or out-of-frame deletions). can cleave both DNA strands at an unmethylated 5’- Geneticcharacterization of a fusion repressed by GATC-3’ site (Au et al. 1992), rare mismatches gener- Dam methylation: Among the candidates obtained with ated duringDNA replication can trigger MutHLS activ- thescreening described in theformer section, a lac ity. The free DNA ends generated can then be used to fusion repressed by Dam methylation was chosen for initiate homologous recombination, as proposed for E. further study. Damf isolates carrying this fusion form coli (MARINUSand KONRAD1976). white colonies on MacConkey lactose, while Dam- iso- A genetic screeningfor the detection of lac transcrip- lates form red colonies. In a Dam- host, the presence tional fusions regulated by DNA adeninemethyla- of plasmid pTP166 causes repression of the fusion to tion: Kan’ transductants generated byMudJ insertion levels similar to the wild type (data not shown). Histo- were classified according to their Lac phenotype; 100- grams of P-galactosidase activities are shown in Figure 500 insertions of the same class (Lac+ or Lac-) were 5. The activity of the fusion is 10-fold higher in a Dam- then pooled and lysed with P22 HT. The lysates were used to transduce aDam- recipient (SV3000),selecting Kan’ transductants on indicator plates. This procedure LTZdam+ can be expected to permit the detection of fusions in 0 LT2dam-201::TnlO genes whose transcription is regulated by DNA adenine n A his3050 dam+ methylation: (1) Lac- isolates thatturn Lac’ ina Dam- back- A his-3050 dam-201::TnlO ground carry fusions putatively repressed by Dam meth- A his-9533 dam+ ylation. 0A his-9533 dam-201::TnlO (2) Lacf isolates thatturn Lac- in a Dam- back- ground carry fusions putatively activated by Dam meth- ylation. FIGURE5.-P-galactosidase activities of isogenic pairs of To be classified as isolatescarrying a fusion regulated Dam+ and Dam- strains carrying the zzv-6?06 Mug fusion in different backgrounds: his+ dam+ (SV3003), his’ dam- by Dam methylation, candidates are required to pass (SV3069), Ahis-3050 dam’ (SV3070), Ahis-3050 dam- three additional tests: (SV3071), Ahis-9533 dam+ (SV3072) and Ahis-9533 dam- (1) A “backcross” in which the original Dam+ isolate (SV3073). Dam Methylation in S. t~~phimzrrium 23 background, a derepression level similar to that of the 123456 1234 5 6 IS10 transposase gene (ROBERTStt al. 1985) and higher than those of other Dam-repressed genes such as glnS, tV)R and SICLA(MARINUS 1985, 1987a,b; PETERSONtt al. 1985). The fusion is not inducible by DNA damage, as indicated by treatments with mitomycin C and nalidixic acid (data not shown). Isolates carrying the fusion in a Adam-204back- ground are stable and do not segregate Lac- colonies or sectors. In contrast, isolates carrying the fusion in a dam-201::Tn IOflet background yield colonies con- taining Lac- sectors at low frequency (< 1%). Sectored colonies are not foundif the plates contain tetracycline. I I I I These results suggest that thedam-201 ::Tn IOflet inser- FIGURE&-((A) Agarose gcl clectrophorcsis of virulencs low tion can undergo a but detectable rate of excision; plasmid DNA preparations from strains LT2 and SV300? in fact, transposase-independent Tn 10 excision has Lanes are as follows: (1) virulence plasmid from LT2, und been shown to be enhanced in Dam- mutants of S. gcstcd; (2) vinllence plasmid from LT2, I.;mRI-tligcstcd; (3 t@himurium (HAFNER andMACPHEE 1991). Most Lac- vinllence plasmid from LT2, HindIII-digested; (4) virulenc sectors seem to be formed by Dam+ revertants, as plasmid from SV3003, undigested; (.5) virdence plasmid fron SV3003, EcoRl-digested; (6) virulence plasmid from SV300? judged from the colony shape of purified Lac- isolates IlindIll-digested. (R) Southern hybridization of the DNA frap and from their ability to inherit a rtcB mutation (data menn shown in A, using the Hind111 fragment of plasmic notshown). These observations provide further evi- pIZ53 as a probe; this fragment contains the Tn5 kanamyci~ dence that theallele Adorn-204 cannot revert (and thus resistance gene (which is the same Kan" determinant carriec is a deletion allele). by the MdJ clement). The fact that the fusion had been isolated in strain TT1704, the usual recipient for transductional delivery obtained in transductional crosses: when the origina of Mug (HUG~IES ROTHand 1988), prompteda second, Mud element or its TnlOcontaining derivative werc unexpected observation: in the TTl704 background, transduced to the wild-type LT2,transductants were ob the ratio of /?-galactosidase activities found in Dam- tained at frequencies near IO-" per p.f.u. In contrast and Dam+ hosts is 26. This result suggests that TT1704 no transductants were obtained when the recipient wa carries a second mutation that activates the fusion in a strain cured of the virulence plasmid (TR5527). the absence of Dam methylation. This (hypothetical) Physical mapping confirmed that the Mug-generate( mutation has no effect in a Dam+ background (Figure fhion was located on the S. ly/)himurium virulence plar 5). The deletion carried by TT1704 (Altis-9533) isnon- mid. DNA preparations of the latter were digested wit1 transducible by P22; thus it must exceed the packaging EmRI or Hind111 and hybridized against the kanamycil capacity of the P22 capsid, 43 kb (CASIENSand HAYDEN gene of the Mud element. The results confirmed tha 1988). Smaller deletions such as Ahis-3050 do not in- the fusion is plasmid-borne: the Mud probe hybridizec crease the expression of the fusion (Figure 5). Thus with a -14kb LcoRI fragment (Figure 6B, lane 5), wit1 we hypothesize that a locus located near the histidine a -10-kb Hind111 fragment (Figure 6B, lane 6), anc operon may corepress the expression of the fusion. with undigested virulence plasmid DNA (Figure 6I? For mapping with the "locked-in" Mud-P22 proce- lane 4). The fusion has been named uv-6306::Mud. dure (BENSONand GOLDMAN 1992),a Tn IOdTet ele- ment was introduced by homologous recombination in DISCXJSSION the 1czcZ gene of the Mud-generated fusion. For this cross, a P22 lysate grown on strain TT16716 was UV- A S. typhimurium point mutant lacking DNA adenin1 irradiated and used to transduce SVSOOS, selecting tet- methylase was described a decade ago (RICHTIE et a, racycline resistance on X-gal plates. Four Tet' Lac- 1986) and specific aspects of Dam methylation in Sa' transductants were obtained. All carried a Tn 10 inser- monella have been since then investigated (RICHTIEt tion 100% linked to the resident Mug element. One al. 1988; HAFNER andMACPEKE 1991). Although thl of these isolates was the origin of strain SV3074. pioneering relevance of these studies must be emphz Attempts to map the Tet" insertion (and thus the sized, their use of point mutants can be viewed as Mug fusion) with the locked-in Mud-P22 procedure potential source of problems. Because dam mutation (BENSONand GOLDMAN1992) did not provide patches are deleterious and highly pleiotropic, Dam- mutant of Tef transductants, suggesting that the locus might are prone either to revert or to accumulate partial SUF not map on the chromosome.Evidence that the fusion pressors (MA(:GRAwand MARINUS1980); thus periodi mapped on the S. typhimurium virulence plasmid was cal reconstruction of Dam- mutants is highly advisable 24 andJ. Torreblanca J. Casadesfis

This can be readily achieved by the use of insertion (WmG and SMITH1986). This viewis supported by alleles such as the TnlOdTet insertions described in this the observation that Dam- Red- and Dam- RecB- work. In turn, certain operations of strain construction mutants of S. typhimurium are viable in the presence of are made easier by the use of deletion alleles that lack mutations thateliminate mismatch repair such as mutH, the Tn loencoded tetracycline resistance. Whenever mutL or mutS. Thus the RecA and RecBCD functions necessary, Dam- deletion strains can be reconstructed seem to be involved in the repair of the double-strand using nearby insertions such as zhf-6304 ::Tn 1OdCam. breaks typical of Dam- strains of Salmonella. For practi- For strain construction, one useful phenotype of Dam- cal purposes, the incompatibility between recB and dam mutants of s. typhimurium is their abnormalcolony mor- mutations permits a rapid identification test: any puta- phology ongreen plates. Like other phenotypes of tive Dam- construct can be identified as such if it fails Dam- strains, the unusualaspect of theircolonies tends to inherit arecB insertion allele in a P22-mediated trans- to disappear upon repeated subculture, but it is a reli- ductional cross (well-characterizedinsertions in recA are able trait to score for Dam- transductants. Confirma- not available in S. typhimurium). tion of the dam genotype may require either a genomic An important difference between Dam- mutants of DNA digestion test or a compatibility assay with recA or E. coli and Salmonella concerns RecJ function: Dam- recB mutations (see below). RecJ- mutants are viable in S. typhimurium but not in TnlOflet insertions in the dam locus of S. typhimu- E. coli (PETERSONet al. 1985). A recent study has sug- rium map at 75 min, as previously reported (RICHTIE et gested that repair of double-strand breaks in methyla- al. 1986).Cotransductional mapping shows thatthe tion-deficient strains of E. coli requires a RecBC-depen- dam locus is 10% linked to cysGand 95% linkedto aroB; dent pathway that includes RecJ (PETERSONand MOUNT these data suggest a position identical to that of the E. 1993). In contrast, our data indicate that RecJis not coli dam gene (MARINUS 1973; LYNGSTADAASet al. 1995). necessary for viability of Dam- strains of Salmonella, at The structure of the dam gene itself is likely to be con- least in an SbcB’ background. Although the difference served, because many (if not all) phenotypes of Salmo- is surprising, one should keep in mind that the RecJ nella Dam- mutants are efficiently complemented by a product may play slightly different roles in E. coli and plasmid containing the E. coli wild-type dam gene. A Salmonella (MAHAN et al. 1992). corollary is that the phenotypes that can be comple- Another unexpected difference between Dam- mu- mented are unequivocally caused by lack of DNA ade- tants of E. coli and Salmonella is UV sensitivity: unlike nine methylase (and not by a polar effect of the dam E. coli, S. typhimurium Dam- mutants are not UV-sensi- mutation on downstream genes). tive (not even at high UV doses: see MARINUS and MOR- Deletion alleles were obtained by selecting Tet’ deriv- RIS 1973; GLICKMANet al. 1978). This observation sug- ativesof TnlOflet insertions in dam. Deletions that gests that methyl-directed DNA repair does not play a remove genetic material from dam toward cysG are nec- significant role in the repair of W damage in Salmo- essarily small because of the presence of the essential nella. This observation is supported by the observation gene trpS (LYNGSTADAASet al. 1995;SANDERSON et al. that mut alleles do not cause UV sensitivity (SHANA- 1995). At the other side of dam, the extent of deletions BRUCH et al. 1981). However, we were surprised by the can be constrained by taking advantage of the presence difference between E. coli and Salmonella and consid- of the aroB gene: if Tets derivatives are obtained on ered the possibility that Salmonella might have alterna- plates lacking an aromatic mix supplement, only AroB+ tive repairfunctions. An obvious genetic entity that Tet’ derivatives are obtained. The advantage of these might harbor those hypothetical functions is the samAB constraints is the yield of deletions (such as Adunz-204 containingvirulence plasmid (NOHMIet al. 1991). How- and others) that remove little genetic material around ever, Dam- mutants proved to be still UV in the ab- dam and thus are sa€e to identify the consequences of sence of the virulence plasmid. Thus, if alternative func- loss of the Dam function (and not thecombined effects tions exist, they must map elsewhere. of multigenic losses). Aside from the differences discussed above, Dam- Attempts to construct double mutants containing a mutants of s. typhimurium share with those of E. coli dam mutation and anothermutation affectingrecombi- many relevant phenotypes: (1) increased incidence of nation and/or repair has indicated that the combina- spontaneousmutations, biased toward transitions tions red dam and recB dam are inviable, while the com- (GLICKMAN1979) ; (2) severe hypermutability (MARINUS binations recD dam, recF dam and recJ dam are viable. et al. 1984); (3) moderate SOS induction (PETERSONet These data indicate that viability of Dam- strains of S. al. 1985) and (4) enhanced duplication segregation, typhimurium requires certain recombination functions which can be suppressed by mutations that eliminate (RecA and RecBC) while others aredispensable (RecD, mismatch repair (MARINUS and KONRAD 1976).The bal- RecJ and RecF) . By analogy with E. coli, one may inter- ance of analogies and differences indicates that Dam pret this to mean that Dam- strains suffer a higher methylation performs similar roles in E. coli and Salmo- incidence of MutH-catalyzed doublestrand breaks nella. Moreover, the ability of the E. coli dam gene to Dam Methylation in S. typhimurium 25

BRAATEN,B. B., X. Nou, L. S. UTENBACHand D.A. Low, 1994 complement S. typhimurium dum mutations suggests that Methylation patterns inpap regulatory DNA control pyelonephri- Dam methyltransferase may be a function highly con- tis-associated pili phase variation in E. coli. Cell 76: 577-588. served among bothtaxa. On the other hand, thediffer- BRAUN,R. E., and A. WRIGHT,1986 DNA methylation differentially ences found areless surprising if one considers that the enhances one of the two E. coli dnaA promoters in uiuo and in vitro. Mol. Gen. Genet. 202: 246-250. genera Escherichia and Salmonella diverge in other CAMPBELL,J. L., and N. KLECKNER, 1990 E. coli onC and the dnaA aspects of recombination and repair (EISENSTADT gene promoter are sequestered from dam methyltransferase fol- 1987). lowing the passage of the chromosomal replication fork.Cell 62: 967-979. Genes regulated by Dam methylation have been tradi- CAS.DABAN,M. J,, and S. N. COHEN,1979 Lactose genes fused to tionally discovered by reverse genetics, after the pres- exogenous promoters in one stepusing a Mu-lac bacteriophage: ence of 5'-GATC-3' sites inor near the promoter(MARI- in uiuo probe for transcriptional control sequences. Proc. Natl. Acad. Sci. USA 76: 4530-4533. NUS 1987; NOYER-WEIDNERand TRAUTNER1993). In CGJENS,S., and M.HAYDEN, 1988 Analysis in vivo of the bacterio- contrast, classical genetic strategies like the lac fusion phage P22 headful nuclease. J. Mol. Biol. 199 467-474. screening described above can identify novel, uncharac- CAsTII.Ho, B. A,, P. OLFSONand M. J. CA~ADABAN,1984 Plasmid insertion mutagenesis and lacgene fusion with Mini-Mubacterio- terized loci regulated by Dam methylation. Our screen- phage transposons. J. Bacteriol. 158: 488-495. ing has two caveats: (1) it relies on the use of unmethyl- CHAN,R. K., D. BOTSTEIN,T. WATANABEandY. OGATA,1972 Special- ated DNA that, in general, is a nonphysiological ized transduction of tetracycline by phage P22 in Salmonella typhi- munum. 11. Properties of a high frequency transducing lysate. condition; thus it might give rise to artefacts whenever Virology 50: 883-898. hemimethylation and unmethylation are notequivalent CHUMLEY,F., R. MENZEI.and J. R.ROTH, 1979 Hfrformation di- signals; (2) small differences in transcriptional activity rected by TnlO. Genetics 91: 639-655. CUPPI.ES,C. G., and J. H. MILLER,1989 A set of lacZ mutations in may not be easily seen on plate tests. Despite these Escherichia coli that allow rapid detection of each of the six base potential limitations, the screening has already proved substitutions. Proc. Natl. Acad. Sci. USA 86: 5345-5349. useful to detect a locus repressed by Dam methylation, EISENSTADT,E., 1987 Analysis of mutagenesis, pp. 1016-1033 in Escherichia coli and Salmonella typhimurium: Cellular and Molecular located in the S. typhimurium virulence plasmid. Biology, edited by F. C. NEIDHARDT,J. L. INGRAHAM,K. B. LOW, This study was supported by grant PB93649 from the Direcci6n B. MAGASANIK,M. SCHAECHTER, and H. E. UMBARGER.American Society for Microbiology, Washington, DC. General de Investigaci6n Cientifica y TCcnica (DGIW) of the Gov- ELLEDGE,S. J., and G. C. WALKER,1983 The muc genes of pKMlOl ernment of Spain. ,J.T. was supported by a predoctoral fellowship are induced by DNA damage. J. Bacteriol. 155: 1306-1315. from the Regional Government of Andalusia (Junta de Andalucia, ELLIOTT,T., and J. R. ROTH, 1988 Characterization of TnlOd-Cam: Spain). We are gratefulto CARMENR. BEUZON,AND&S GARZON, a transposition-defective TnlO specifying chloramphenicol resis- AMANDO FLORESand MARTIN MARINUS for helpful discussions and to tance. Mol. Gen. Genet. 213: 332-338. RICHARD D'ARI for critical reading of the manuscript. Strains were FLORES,A,, and J. CASADESUS, 1995 Suppression of the pleiotropic kindly provided by MARTIN MARINUS,JOHN ROTH,GRAHAM WALKER effects of HisH and HisF overexpression identify four novel loci and by the Salmonella Genetic Stock Center, University of Calgary, on the Salmonella typhimunum chromosome: osmH, sjiW, sjiX, and Canada. Theassistance of ANA MORENO,GLORIA CHACON, JOSE COR- sjiY. J. Bacteriol. 177: 4841-4850. DOBA and LUISROMANCO is also acknowledged. GAMMIE,A. E., and J. H. CROSA,1991 Roles of DNA adenine methyl- ation in controlling replication of the REP1 replicon of plasmid pCOLV-K30. Mol. Microbiol. 5: 495-503. LITERATURECITED GARRIGA,X., 1992 Isolation and characterization of led genes from ABELES, A. L., T. BRENDLERand S. AUSTIN,1993 Evidence for two Salmonella typhimurium and lhinia carotouora. Ph.D. Thesis, Uni- levels of control of P1 oriR and host oriC replication origins by versitat Autbnoma de Barcelona, Barcelona, Spain. DNA adenine methylation. J. Bacteriol. 175: 7801-7807. GARZON,A., D. A CANU and J. CPSADESilS, 1995 Role of Erf recombt ARRAJ,J. A,, and M. G. MARINUS,1983 Phenotypic reversal in dam nase in P22-mediated plasmidtransduction. Genetics 140 427-434. mutants of Escherichia coli K-12 by a recombinant plasmid con- GARZON,A., C. R. BEUZON,M. J. MAHAN and J. C~s~uesirs,1996 recB taining the dam' gene. J. Bacteriol. 153: 562-565. recJmutants of Salmonella typhimunum are deficient in transduc- Au, K G., K. WELSHand P. MODRICH, 1992 Initiation of methyl- tionalrecombination, DNA repairand plasmid maintenance. drected mismatch repair. J. Biol. Chem. 267: 12142-12148. Mol. Gen. Genet. 250: 570-580. BACHMANN,B. J., 1990 Linkage map of Eschm'chia coli K-12, edition GI.ICKMAN,B. W., 1979 Spontaneous mutagenesis in Escherichiacoli 8. Microbiol. Rev. 54: 130-197. strains lacking 6methyladenine residues in their DNA. Mutat. BALE,A,, M. D'ALARCAOand M. G. MARINUS,1979 Characterization Res. 61: 153-162. of DNA adenine methylation mutam of Escherichiacoli K-12. GIKKMAN,B., P. VAN DER EISEN and M. RADMAv, 1978 Induced Mutat. Res. 59: 157-165. mutagenesis in dam mutants of Escherichia coli:a rolefor 6methyl- BARRAS,F., and M. G. MARINUS,1989 The greatGATC: DNA methyl- adenine residues in mutation avoidance. Mol. Gen. Genet. 163: ation in E. coli. Trends Genet. 5: 139-143. 307-312. BENDER,J., and N. KLECKNER,1986 Genetic evidence thatTnlO HAFNER,L. M., and D. G. MACPHEE, 1991 Precise excision of TnlO transposes by a nonreplicative mechanism. Cell 45: 801-815. in Salmon.ella typhimunum: effects of mutations in the polA, dam, BENSON,N. R., and B. S. GOLDMAN,1992 Rapid mapping in Salmo- mutH and mutB genes and of methionine or ethionine in the nella typhimunum with Mud-P22 prophages. J. Bacteriol. 174: plating medium. Mutat. Res. 263 179-184. 1673-1681. HATTMAN,S., 1981 DNA methylation, pp. 517-548 in The Enzymes, BLYN,L. B., B. A. BRAATENand D. A. LOW, 1990 Regulation of pap edited by P. D. BUYER.Academic Press, New York. pilin phase variation by a mechanism involving differential Dam HAm.W, S., 1982 DNA methyltmnsfedependent transcription of methylation states. EMBO J. 9: 4045-4054. the phage Mu m gene. Proc. Natl. Acad. Sci.USA 79: 5518-5521. BOCHNER,B. R., H. C. HUANG,G. L. SCHIEVENand B. N. AMES,1980 HERMANG. E., and P. MODRICH 1982 Esrherichia coli dam methylase. Positive selection for loss of tetracycline resistance. J. Bacteriol. J. Biol. Chem. 257: 2605-2612. 143: 926-933. HERRICK,J., R. KERN, S. GUHA,A. LANDOULSI, 0. FAYETet al., 1994 BOW, E., and A. LBBNER-OLESEN,1990 The role of dam methyltransfer- Parental strand recognition of the DNA replication origin by the ase in the control of DNA replication in E. coli. Cell 62: 981-989. outer membrane in E. coli. EMBO J. 13: 4695-4703, 26 J. Torreblanca and J. Casadesfis

HSU,M., M. INOWand S. INOLIYE, 1990 Retron forthe 67-base MODRICH,P., 1987 DNA mismatch correction. Annu. Rev. Biochem. multicopy single stranded DNA from Escherichia coli: a potentid 56: 435-466. transposable element encoding bothreveme transcriptase and Dam NOHMI,T., A. HAKURA,Y. NAKAI, M. WATANABE,S. Y. MURAYAMAet al., methylase functions. Proc. Natl. Acad. Sci. USA 87: 9454-9458. 1991 Salmonella typhimurium has two homologous but different HUGHES,K T., and J. R. ROTH,1984 Conditionally transpositiondefec- umuDC operons: cloning of a new umuDGlike operon (samAB) tive derivative of Mud (Amp Lac). J. Bacteriol. 159: 130-137. present in a 60-megadalton cryptic plasmid of S. typhimurium. J. HUGHES,K T., and J. R ROTH, 1985 Directed formation of deletions Bacteriol. 173 1051-1063. and duplications using Mud(Ap, lac). Genetics 109: 263-282. NOYER-WEIDNER,M., and T.A. TRAUTNER, 1993 Methylation of DNA HUGHES,K. T., and J. R. ROTH, 1988 Transitory cis complementa- in prokaryotes, pp. 39-108 in DNA Methylahm: Molecular Biology tion: a method to provide transposition functions to defective and Biological SigniJicance, edited by J. P. JOST and H. P. SAI.UZ. transposons. Genetics 119: 9-12. Birkhauser Verlag, Basel, Switzerland. PETERSON,K. and D.W. MOUNT,1993 Analysis of the genetic K~CHERER,C., H. LOTHER,R. KOLLING,M. A. SCHAUZUand W. MES R., requirements forviability of Escherichia coliDNA adenine methyl- SER, 1986 Regulation of transcription of the chromosomal dnaA ase (dam) mutants. J. Bacteriol. 175 7505-7508. gene of Escherichia coli. Mol. Gen. Genet. 205: 115-121. PETERSON,K. R., K. F. WERTMAN,D. W. MOUNTand M. G. MARINUS, LACKS,S., and B. GREENBERG,1977 Complementaly specificity of 1985 Viability of Escherichiacoli K-12 DNA adenine methylase restriction of Diplococcw pneumoniae with respect (dam) mutants requires increased expression of specific genes to DNA methylation. J. Mol. Biol. 114: 153-158. in the SOS regulon. Mol. Gen. Genet. 201: 14-19. LEDERBERG,E. M., and S. N. COHEN,1974 Transformation of Salmo- F’UKKIIA,P., J. PETERSON, G. HERMAN, P. MODRICHand M. MESELSON, nella typhimurium by plasmid deoxyribonucleic acid. J. Bacteriol. 1983 Effects of high levels of DNA adenine methylation on methyl 119: 1072-1074. directed mismatch repair in E. coli. Genetics 104: 571-582. L0BNERaLESEN,A., E. BOVE and M.G. IMARINUS, 1992 Expression RADMAN, M., and R. WAGNER,1986 Mismatch repair in Escherichia of the Escherichia coli dam gene. Mol. Microbiol. 6 1841-1851. coli. Annu. Rev. Genet. 20: 523-538. Lu, A. L., S. CLARKand P. MODRICH,1983 Methyldirected repair RICHTIE, L. J., R.M. HALL and D. M. PODGER,1986 Mutant of of DNA base pair mismatches in vitro. Proc. Natl. Acad. Sci. USA Salmonella typhimurium LT2 deficient in DNA adenine methyla- 80: 4639-4643. tion. J. Bacteriol. 167: 420-422. LYNGSTADAAS,A.,A. L0BNER-OLESENandE. BOYE, 1995 Characteriza- RICHTIE,L., D. M. PODGERand R. M. HALL, 1988 A mutation in the tion of three genes in the damontaining operon of Escherichia DNA adenime methylase gene (dam) of Salmonella typhimurium coli. Mol. Gen. Genet. 247: 546-554. decreases susceptibility to 4aminoacridine-induced frameshift mutagenesis. Mutat. Res. 194 131-141. MACGRAW, B.,and M. G. MARINUS, 1980 Isolation and characteriza- RILEY, M.,and K. E. SANDERSON,1990 Comparative genetics ofEsche- tion of Dam+ revertants and suppressor mutations that modify richia coli and Salmonella typhimurium, pp. 85-95 in TkBacterial secondary phenotypes of dam-3 strains of Escherichia coli K-12. Chromosome, edited by IC R. DRLICAand M. RILEY. American Soci- Mol. Gen. Genet. 178: 309-315. ety for Microbiology, Washington, DC. MAHAN,M. J., J. CAsm~sirsand J. R. ROTH, 1992 The Salmonella ROBERTS,D., B. C. HOOPES,W. R. MCCLUREand N. KLECKNER,1985 typhimurium RecJ function permits growth of P22 abc phage on IS10 transposition is regulated by DNA adenine methylation. Cell recBW+ hosts. Mol. Gen. Genet. 232: 470-478. 43 117-130. MALDONADO, R, A. GARZON,D. R DEANand J. CASADESUS, 1992 Gene RUPP,W. D., C. E. WILDE111, D. L. RENO and P. HOWARD-FLANDERS, dosage analysis in Rzotobacter vinelandii. Genetics 132 869-878. 1971 Exchanges between DNA strands in ultraviolet-irradiated MALOY,S. R., 1990 Experimatal Technquac in Bacterial Genetics. Jones Escherichia coli. J. Mol. Biol. 61: 25-44. and Bartlett Publishers, Boston. RUSSEL, D. W., and N. D. ZINDER,1987 Hemimethylation prevents MALOY, S. R., and W. D. NUNN,1981 Selection for loss of tetracycline DNA replication in E. coli. Cell 50: 1071-1079. resistance by Eschm.chia coli. J. Bacteriol. 145: 1110-1112; erra- SALLES, B.,J. M. WEISEMANNand G. M. WEINSTOCK,1987 Temporal tum 146 831. control of colicin El induction. J. Bacteriol. 169: 5028-5034. MARINUS, M. G., 1973 Location of DNA methylation genes on the SAMBROOK,J., E. F. FRITSCHand T. MANIATIS,1989 Molecular Clon- Escherichia coli K-12 genetic map. Mol. Gen. Genet. 127: 47-55. ing, Ed. 2. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. MARINUS, M. G., 1984 Methylation of procaryotic DNA, pp. 81-109 SANDERSON,K. E., and J. R. ROTH, 1983 Linkage map of Salmonella in DNA Methylation, edited byA. RAZIN,M. CEDARand A. D. typhimurium, edition Microbiol. Rev. 47: 410-453. RIGGS.Springer-Verlag, New York. VI. SANDERSON,K. E., A.HESSEI, and K. E. RUDD,1995 Genetic map of MARINUS,M. G., 1985 DNA methylation enhances trpR promoter Salmonella typhimurium, edition VIII. Microbiol. Rev. 52 485-532. activity in Escherichia coli K-12. Mol. Gen. Genet. 200: 185-186. SCHMIEGER,H., 1972 Phage P22 mutants with increased or de- MARINUS,M. G., 1987a DNA methylation in Escherichia coli. Annu. creased transducing abilities. Mol. Gen. Genet. 119 75-88. Rev. Genet. 21: 113-131. SHANABRUCH,W. G., I. BEHLAUand G. C. WALKER,1981 Spontane- MARINUS,M. G., 1987b Methylation of DNA, pp. 697-702 in Esche- ous mutators of Salmenella typhimurium LT2 generated by inser- richia coli and Salmonella typhimurium: Cellular and Molecular Biol- tion of transposable elements. J. Bacteriol. 147: 827-835. ogy, edited by F. C. NEIDHARDT,J. L. INGRAHAM,K. B. LOW, B. SMITH,D. W., A. M. GARLAND,G. HERMAN, R. E. ENNS,T. A. BAKER MAGASANIK, M. SCHAECHTER,and H. E. UMBARGER.American et al., 1985 Importance of state of methylation of oriC GATC Society for Microbiology, Washington, DC. sites in initiation of DNA replication in Escherichia coli. EMBO J. MARINUS, M. G., and E. B. KONRAD,1976 Hyper-recombinationin dam 4: 1319-1327. mutants of Eschenchiu coli K-12. Mol. Gen. Genet. 149: 273-277. STERNBERG,N., and J. COULBY,1990 Cleavage of the bacteriophage MARINUS, M. G., and N. R MORRIS,1973 Isolation of deoxyribonucleic P1 packaging site (pac) is regulated by adenine methylation. acid methylase mutants of E. coli. J. Bacteriol. 114: 1143-1150. Proc. Natl. Acad. Sci. USA 87: 8070-8074. VOGEL., H.,and D. BONNER,1956 Acetylornithase offichakhiu cdi: par- MARINUS,M. G., A. POTEETEand J. A. ARRAJ, 1984 Correlation of DNA adenine methylase activity with spontaneous mutability in tid purification and some properties. J. Biol. Chem. 218: 97-106. WANG,T. C. W., and K. C. SMITH,1986 Inviability of dum red and Escherichia coli K-12. Gene 28: 123-125. dam recB cells of Escherichia coli is correlated with their inability MESSER,W., and M. NOYER-WEIDNER,1988 Timing and targeting: to repair DNA double-strand breaks produced by mismatch re- the biological functions of Dam methylation in E. coli. Cell 54 pair. J. Bacteriol. 165 1023-1025. 735-737. YOUDERIAN,p., p. SUGIONO,K. L. BREWER,N. P. HISINS and T. MESSER, W.,U. BELLEKESand H. LOTHER,1985 Effect of dam methyl- E[,I.IOTT, 1988 Packaging specific segments of the Salmonella ation on the activity of the E. coli replication origin, WiC. EMBO chromosome with locked-in Mud-P22 prophages. Genetics 118: J. 4 1327-1332. 581-592. MILLER,J. H., 1972 Expmhentsin Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Communicating editor: D. BOTSTEIN