Characterization of the Minimal Replicon of a Cryptic ...Deinococcus promoter was shown to efﬁciently express LacZ. Ever since its discovery in 1956 (3), Deinococcus radiodurans

APPLIED AND ENVIRONMENTAL MICROBIOLOGY,0099-2240/00/$04.0010

Sept. 2000, p. 3856–3867 Vol. 66, No. 9

Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Characterization of the Minimal Replicon of a CrypticDeinococcus radiodurans SARK Plasmid and Development of

Versatile Escherichia coli-D. radiodurans Shuttle VectorsROB MEIMA1 AND MARY E. LIDSTROM1,2*

Departments of Chemical Engineering1 and Microbiology,2 University of Washington,Seattle, Washington 98195-1750

Received 10 April 2000/Accepted 8 June 2000

The nucleotide sequence of a 12-kb fragment of the cryptic Deinococcus radiodurans SARK plasmid pUE10was determined, in order to direct the development of small, versatile cloning systems for Deinococcus.Annotation of the sequence revealed 12 possible open reading frames. Among these are the repU and resUgenes, the predicted products of which share similarity with replication proteins and site-specific resolvases,respectively. The products of both genes were demonstrated using an overexpression system in Escherichia coli.RepU was found to be required for replication, and ResU was found to be required for stable maintenance ofpUE10 derivatives. Gel shift analysis using purified His-tagged RepU identified putative binding sites andsuggested that RepU may be involved in both replication initiation and autoregulation of repU expression. Inaddition, a gene encoding a possible antirestriction protein was found, which was shown to be required for hightransformation frequencies. The arrangement of the replication region and putative replication genes for thisplasmid from D. radiodurans strain SARK is similar to that for plasmids found in Thermus but not to that forthe 45.7-kb plasmid found in D. radiodurans strain R1. The minimal region required for autonomous replica-tion in D. radiodurans was determined by sequential deletion of segments from the 12-kb fragment. Theresulting minimal replicon, which consists of approximately 2.6 kb, was used for the construction of a shuttlevector for E. coli and D. radiodurans. This vector, pRAD1, is a convenient general-purpose cloning vector. Inaddition, pRAD1 was used to generate a promoter probe vector, and a plasmid containing lacZ and aDeinococcus promoter was shown to efficiently express LacZ.

Ever since its discovery in 1956 (3), Deinococcus radioduransand other members of the Deinococcaceae have become theparadigm of natural resistance to high-level ionizing and UVradiation. D. radiodurans cells are able to survive levels ofgamma radiation 1,000 times higher than the lethal dose forhumans and were shown to survive and accurately repair mas-sive DNA damage. At 1.5 megarads, up to 130 double-strandbreaks occur per cell, which are repaired without mutagenesisor loss of viability (13). In contrast, the presence of a mere twodouble-strand breaks is lethal to Escherichia coli (21). As aconsequence, this pink-pigmented, non-spore-forming bacte-rium has received considerable attention from the scientificcommunity, from both a fundamental and an applied point ofview. Because of their radioresistance, Deinococcus specieshold great potential for bioremediation of complex waste mix-tures containing organic solvents, heavy metals, and radioiso-topes. Recently, Lange et al. (23) have demonstrated toluenedioxygenase (TDO) activity in engineered strains of D. radio-durans R1. TDO is a broad-spectrum dioxygenase capable ofdegrading trichloroethylene (42), a common organopollutantat many Department of Energy waste sites (32). In thesestrains, TDO activity was not inhibited by high levels of radi-ation, demonstrating the potential of D. radiodurans as a toolfor remediation of mixed wastes.

Recent advances in molecular biology and genomics havegreatly facilitated the study of D. radiodurans. The completegenome sequence has been available since 1998, and the

annotated version has recently been published, revealing agenome consisting of two chromosomes, a megaplasmid, anda 45.7-kb plasmid (44). However, the lack of small, versatileplasmid cloning and expression systems has hampered the ex-ploitation of D. radiodurans to its full potential. An E. coli-D. radiodurans shuttle vector, pI3, and derivatives for promotercloning were constructed and characterized previously by Mas-ters and Minton (26). However, due to their large size and thelack of knowledge concerning the sequence and the minimalreplication functions, these plasmids have not been convenientfor genetic engineering of D. radiodurans. The availability ofa broader repertoire of more well-defined and convenientgenetic systems would significantly facilitate the genetic ame-nability of D. radiodurans, for both fundamental and appliedresearch. The present paper describes the characterization ofthe minimal replicon of pI3, which is a derivative of pUE10, acryptic plasmid from D. radiodurans SARK (25). We report (i)sequencing and functional characterization of this deinococcalreplicon and (ii) construction of second-generation vectors foruse in D. radiodurans and E. coli. We show that the resultinggeneral-purpose vector can be shuttled efficiently betweenthese organisms and that the lacZ reporter gene, fused to en-dogenous D. radiodurans promoters, is successfully expressedfrom this plasmid in D. radiodurans.

MATERIALS AND METHODS

Bacterial strains and plasmids. The bacterial strains and plasmids used in thisstudy are listed in Table 1.

Chemicals and enzymes. All chemicals used were of analytical grade and,unless indicated otherwise, were obtained from Baker Chemical Co. (Phillips-burg, N.J.) or Fisher Scientific (Fair Lawn, N.J.). 5-Bromo-4-chloro-3-indolyl-b-D-galactopyranoside (X-Gal) and o-nitrophenyl-b-D-galactopyranoside werefrom ISC Bioexpress (Kaysville, Utah) and Sigma Chemical Co. (St. Louis, Mo.),

* Corresponding author. Mailing address: Department of ChemicalEngineering, Box 351750, University of Washington, Seattle, WA 98195-1750. Phone: (206) 616-5282. Fax: (206) 616-5721. E-mail: [email protected].

3856

on Septem

ber 2, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

http://aem.asm.org/

respectively. Enzymes for molecular biology were purchased from BoehringerMannheim Corp. (Indianapolis, Ind.) and New England BioLabs (Beverly,Mass.) and were used as described by the supplier. Taq DNA polymerase wasobtained from Gibco BRL (Grand Island, N.Y.).

Media and growth conditions. LB broth for growth of E. coli consisted of (perliter) 10 g of tryptone (Difco Laboratories, Detroit, Mich.), 5 g of yeast extract(Difco), and 10 g of NaCl (pH 7.4). TGY broth for D. radiodurans contained (perliter) 5 g of tryptone, 3 g of yeast extract, and 1 g of glucose (28). Solid mediawere prepared by addition of 1.5% agar (Difco) to either LB or TGY broth.Where necessary, media were supplemented with the appropriate antibiotics, allof which were obtained from Sigma. Ampicillin was used at 50 mg/ml for E. coli.Chloramphenicol was added to a final concentration of 3 mg/ml for D. radio-durans. Kanamycin was routinely used at 50 mg/ml for E. coli and at 8 or 4 mg/mlfor D. radiodurans grown on solid and liquid media, respectively. Transforma-tions of E. coli were performed either using commercially available cells (JM109and TOP10 [Promega, Madison, Wis., and Invitrogen, Carlsbad, Calif., respec-tively]) or by the CaCl2 method (35). D. radiodurans cells were transformedessentially as described previously (40), with the exception that cells from expo-nentially growing cultures were collected by centrifugation (12,000 3 g, 1 min)and concentrated 10-fold in TGY supplemented with 30 mM CaCl2.

DNA manipulations. Miniscale plasmid DNA preparations of E. coli wereobtained as described by Sambrook et al. (35) and were resuspended in a totalvolume of 25 ml. Plasmid DNA from D. radiodurans was isolated using a variantof the alkaline lysis method. Cells were collected by centrifugation (16,000 3 g,2 min) and resuspended in 100 ml of solution I (25 mM Tris-HCl, 10 mM EDTA,50 mM glucose [pH 8.0]) supplemented with lysozyme (10 mg/ml; BoehringerGmbH, Mannheim, Germany) and proteinase K (5 mg/ml; Boehringer GmbH).The suspension was incubated at 50°C for 30 min, followed by 5 min at 0°C and1 min at 100°C. Subsequently, lysis was achieved by the addition of 200 ml ofsolution II (1% [wt/vol] sodium dodecyl sulfate [SDS], 0.2 N NaOH). Afterprecipitation of chromosomal DNA and proteins (150 ml of solution III [60 ml ofpotassium acetate, 11.5 ml of glacial acetic acid, 28.5 of ml H2O]), the aqueousphase was extracted twice with an equal volume of phenol-chloroform-isoamy-lalcohol (24:24:1) (Boehringer Mannheim Corp.), and DNA was precipitatedwith 2.5 vol of 96% (vol/vol) ethanol. Finally, the pelleted DNA was resuspendedin 25 ml of T10E1 (10 mM Tris-HCl, 1 mM EDTA [pH 8.0]) containing 2 mg of

RNase A (Boehringer GmbH) per ml. This procedure produced sufficientamounts of DNA for several restriction analyses from 1 ml of cultures grownovernight at 30°C. PCR products were purified using a Qiaquick PCR purifica-tion kit (Qiagen Inc., Valencia, Calif.). Primers for PCR amplification andsequencing purposes were 18-mers and were obtained from Gibco BRL (Fred-erick, Md.). Nonradioactive nucleotide sequencing was performed by the Uni-versity of Washington’s Department of Biochemistry DNA Sequencing Facility,using an ABI Prism 377 sequencer (PE Biosystems). Southern hybridizationanalyses using 32P-labeled probes were performed at 60°C (both hybridizationand wash steps) as described by Sambrook et al. (35); [a32P]dCTP (6000 Ci/mmol) was from NEN Life Science Products (Boston, Mass.).

pI3 sequencing strategy. The sequence of the pUE10 moiety was determinedby using a primer-walking strategy starting from four positions on the pI3 ge-nome. The primers used for the initial sequencing reactions were based on thesequence of a 0.5-kb PstI-NsiI fragment present on pTCP1 (Table 1), onpKK232-8 (GenBank accession no. U13859, position 5060), and on available pI1sequences (24) GenBank accession no. M94966).

Sequence comparisons and predictions. Computational analyses of DNA anddeduced amino acid sequences were performed using the following World WideWeb-based programs. Similarity searches were carried out using the BLASTalgorithms described by Altschul et al. (2) at the National Center for Biotech-nology Information website (http://www.ncbi.nlm.nih.gov/BLAST/). Multiplealignments were performed using CLUSTAL W, available at the EuropeanBioinformatics Institute website (http://www.ebi.ac.uk/clustalw/). The presenceof possible signal peptidase I cleavage sites was analyzed using the parametersdescribed by Nielsen et al. (29) at the Center for Biological Sequence Analysiswebsite (http://www.cbs.dtu.dk/services/SignalP/). Analyses of primary proteinstructure were performed using the ExPASy ProtParam tool, available at theSwiss Institute of Bioinformatics website (http://www.expasy.ch/cgi-bin/prot-param). Preliminary sequence data for D. radiodurans were obtained from TheInstitute for Genomic Research website (http://www.tigr.org).

Plasmid constructions. Derivatives of pI3 were constructed as follows (Fig. 1).First, a 3.1-kb fragment containing the putative ardU gene was removed by CelIIdigestion and self-ligation, resulting in pI5. Plasmid pI6 was constructed by NsiIdigestion of pI3, followed by T4 DNA polymerase treatment to remove theprotruding 39 ends and self-ligation. In a parallel experiment, a 6.3-kb KpnI

TABLE 1. Bacterial strains and plasmids

Strain orplasmid Relevant genotype Reference and/

or source

E. coliJM109 F9 traD36 lacIq D(lacZ)M15 proA1 B1/e142 (McrA2) D(lac-proAB) thi gyrA96 (Nalr) endA1 hsdR17

(rK2 mK

1) relA1 supE44 recA143; Promega

TOP10 F2 mcrA D(mrr-hsdRMS-mcrBC) f80lacZDM15 DlacX74 recA1 deoR araD139 D(ara-leu)7697 galUgalK rpsL (Strr) endA1 nupG

Invitrogen

M15 Nals Strs Rifs lac ara gal mtl F2 recA1 uvr1 lon1 Qiagen

D. radiodurans R1 Wild-type strain 3

PlasmidspI3 E. coli-D. radiodurans shuttle vector composed of pKK232-8 and pUE10 sequences; Apr Cmr, 17 kb 24pMD209 E. coli-D. radiodurans shuttle vector composed of pBR322 and pUE10 sequences; Tcr Kmr M. DalypMTL23 E. coli cloning vector; Apr, 2.5 kb 9pMUTIN2mcs pUC-based gene inactivation and promoter probe vector for B. subtilis; Apr Emr, 8.6 kb 45; http://www.tigr.orgpTC23 pMTL23 carrying the XbaI-NdeI Tcr fragment of pMD209 This studypTCP1 pTC23 carrying the NsiI-PstI fragment of pI3 This studypI5 pI3 carrying a CelII deletion; 13.7 kb This studypI6 pI3 from which the NsiI site was removed by digestion and T4 DNA polymerase treatment This studypI8 pI3 from which the 6.3-kb KpnI fragment was removed by partial digestion and self-ligation This studypI9 pI3 lacking both the 4.7- and 6.3-kb KpnI fragments This studypI10 pI8 carrying a HindIII-SplI deletion; 8.6 kb This studypI12 pI10 carrying an AocI deletion; 8.0 kb This studypRAD1 pMTL23 carrying the SalI-FspI fragment of pI12; 6.3 kb This studypRAD1AocI pMTL23 carrying the SalI-FspI fragment of pI10; 6.8 kb This studypRADZ1 pRAD1 carrying the lacZ fragment of pMUTIN2mcs; 10.0 kb This studypRADZ3 pRADZ1 carrying the putative R1 groESL promoter This studypRADZ30 Like pRADZ3; opposite-orientation groESL fragment This studypREP4 p15A-derived vector carrying the lacI gene for in trans repression of the lac promoter; Kmr, 3.7 kb QiagenpQE30–pQE32 pBR322-derived series of vectors carrying a His6 linker for overexpression and purification of

proteins in E. coli; Apr, 3.5 kbQiagen

pQRESU1 pQE30 carrying the pI3 resU gene (primers resU59 and resU39) This studypQRESU10 Like pQRESU10; opposite-orientation resU fragment This studypQREPU1Bd pQE30 carrying the pI3 repU gene containing a mutation in the TTG startcodon

(primers repU59Bd and repU39A)This study

pQREPU10Bd Like pQREPU1Bd; opposite-orientation repU fragment This study

VOL. 66, 2000 MOLECULAR TOOLS FOR D. RADIODURANS GENETIC ENGINEERING 3857

on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

fragment was removed by partial KpnI cleavage and subsequent ligation of themixture, yielding pI8. Next, the resU gene was removed by HindIII-SplI digestion,followed by T4 DNA polymerase treatment in the presence of 0.25 mM de-oxynucleoside triphosphates to fill in the overhanging ends and ligation. Theresulting plasmid, pI10, was subsequently digested with AocI and religated,producing pI12. Finally, the minimal replicon of pI12 was fused to pMTL23,creating pRAD1.

Overexpression of the plasmid-encoded ResU and RepU proteins in E. coliwas achieved by fusing PCR fragments containing the genes to an IPTG (iso-propyl-b-D-thiogalactopyranoside)-inducible promoter. These fragments wereobtained using the following primers, bearing 59 tags containing restriction sitesfor BamHI and BglII (underlined): resU59 (GCGAGATCT-ATGTCTGCACAGAATCA), resU39 (CGCGGATCC-TTTTGAAACCGACGCAG), repU59 A(GCCAGATCT-TCTTGAGACACAATCCA), repU39 A (CGCGGATCC-CTTCTCGGCCTTTCTGT), repU59 Ad (GCCAGATCT-TCGTCAGACACAATCCA), and repU59Bd (GCGAGATCT-TTCACGGTCCGACTCCT); the last twocontain mutations (boldface) that eliminate the predicted translational start sitesof repU. After PCR amplification, the fragments were first cloned in pCR2.1-TOPO (Invitrogen) and subsequently transferred to the QIAexpress type IVseries of His6 tag-based overexpression vectors (Qiagen), thus fusing the hexa-histide-encoding tag to the 59 termini of resU and repU, respectively.

Expression from a plasmid-borne reporter gene was studied using the follow-ing plasmids. pRADZ1 was constructed by insertion of a 3.2-kb PCR fragmentcarrying the pMUTIN2mcs-derived b-galactosidase (lacZ) gene of E. coli fused

to an optimized ribosome-binding site (RBS) flanked by suitable restriction sites(41; R. Meima and M. E. Lidstrom, unpublished data) in the BglII-XbaI sites ofpRAD1. Subsequently, the promoter region of the putative groESL operon of D.radiodurans R1 was inserted into the BglII site of pRADZ1 to generate pRADZ3and pRADZ30 (opposite orientation), respectively.

Protein analysis. SDS-polyacrylamide gel electrophoresis was performed asdescribed by Laemmli (22), using a Hoefer Mighty Small vertical electrophoresissystem (Pharmacia Biotech, San Francisco, Calif.). Protein was visualized bystaining with Coomassie brilliant blue (Sigma). Overexpression and purification,under nondenaturing conditions, of the His-tagged RepU and ResU proteinsusing Ni-nitrilotriacetic acid (Ni-NTA) column chromatography was performedexactly as indicated by the supplier (Qiagen); samples were stored at 0°C forprolonged periods of time without loss of activity in subsequent gel mobilityassays. Gel retardation studies using purified N-terminally His-tagged RepUprotein were performed as follows. Approximately 300 ng of protein was mixedwith 32P-labeled pRAD1AocI-derived MaeI or PCR fragments in a bindingbuffer that consisted of 20 mM Tris-HCl (pH 8.0), 100 mM KCl, 5 mM MgCl2,0.5 mM dithiothreitol, 10 mM EDTA, and 20% (vol/vol) glycerol. Bovine serumalbumin and poly(dI-dC) were used as noncompetitive protein and DNA at 100and 50 ng/ml, respectively. After 30 min of incubation at room temperature, thesamples were electrophoresed on a 4% (wt/vol) polyacrylamide gel in 13 TAE(40 mM Tris-acetic acid [pH 8.0], 2 mM EDTA).

Expression assays. Expression of the lacZ reporter gene in E. coli and D.radiodurans colonies was detected using X-Gal (40 mg/ml). Quantitative analyses

FIG. 1. Schematic representation of the 11.9-kb pUE10 EcoRI-HindIII fragment present on pI3 and of deletion derivatives. Restriction sites that were mappedpreviously are indicated (26); those used for the construction of the deletion derivatives are shown in boldface. Abbreviations for restriction enzymes: E, EcoRI; Sm,SmaI; A, AocI; K, KpnI; Cl, ClaI; S, SplI; St, StuI; C, CelII; P, PstI; N, NsiI; and H, HindIII. See text for details on the construction of the derivatives. The methylatedClaI site is marked with an asterisk (Cl*); note that the position of this site, as determined in our sequencing effort, differs substantially from that on the physical mappublished previously (26). The genotypes of the resulting plasmids and their ability to replicate in D. radiodurans R1, as measured by the occurrence of Cmr

transformants, are indicated on the left. repU*, truncated form of the repU gene as a result of the out-of-frame mutation introduced in pI6 by removal of theNsiI-generated protruding ends followed by self-ligation (N*). wt, wild type.

3858 MEIMA AND LIDSTROM APPL. ENVIRON. MICROBIOL.

on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

of lacZ expression were performed as described by Miller (27). Cell extracts ofD. radiodurans were obtained by passing concentrated cell suspensions througha French pressure cell at 1,000 lb/in2 using a J5-598A laboratory pressure cellpress (Aminco, Silver Spring, Md.). Alternatively, samples for b-galactosidaseassays were prepared by toluene permeabilization of cells. To correct for theabsorption caused by the cell suspension added to the reaction mixtures, the A420value obtained with R1(pRAD1) cells (lacZ mutant) was subtracted from thoseobtained with the other constructs.

Nucleotide sequence accession number. The complete sequence of the pUE10moiety of pI3 can be accessed through GenBank (accession no. AF206717).

RESULTS

Determination of the pI3 sequence. Since previous attemptsto define the minimal replicon of pI3 by shotgun approacheswere unsuccessful (26), we determined the nucleotide se-quence to allow a more directed approach. Based on the phys-ical map published by Masters and Minton (26) we were ableto subclone a 0.5-kb NsiI-PstI fragment on pTC23 (Table 1;Fig. 1) and to sequence this insert on both strands. The se-quence of the remaining portion of the 11.9-kb EcoRI-HindIIIfragment was established using a primer-walking approach.This single-stranded sequence was verified in part by severaladditional runs using a set of primers directed against theminimal replicon (see below), altogether generating approxi-mately 2.8 kb of double-stranded sequence. No mistakes werefound, suggesting that the overall sequence of the entire inserthas an error rate of less than 1 mistake per 2,800 bp.

The total size of the pUE10-derived HindIII-EcoRI insert inpI3 was determined to be 11,910 bp, part of which is shown inFig. 2A. The overall G1C content is 64.5% which is com-parable to the genomic percent G1C of D. radiodurans R1.Nevertheless, a region of strikingly high percent A1T wasfound downstream of the putative replication protein-encod-ing gene (see below). Annotation of the sequence revealed thepresence of 12 putative open reading frames (ORFs), all ofwhich are transcribed in the same direction (Fig. 1). Thesewere analyzed using the BLAST algorithms (2) (Table 2), andfive ORFs showed significant similarity at the amino acid levelto other database entries. The second ORF encodes a proteinthat is highly similar to a class of hypothetical proteins, repre-sentatives of which are found in both the archaeal (e.g., Meth-anococcus jannaschii protein MTH993; 55% overall similarity)and eubacterial (e.g., YxiE of Bacillus subtilis and DR2132 andDR2363 of D. radiodurans R1; $42% overall similarity) do-mains. Since the function of these proteins is unknown, therelevance of the gene, which was designated orfB, for pUE10 asyet remains unclear. The third ORF is very similar to periplas-mic serine protease genes and was hence designated htrU. Infact, the highest similarity (71% overall) was found with thetranslated product of a chromosomal htrA-like gene of D. ra-diodurans R1 (DR1756). In accordance with its expected lo-calization, the predicted protein contains a possible signal pep-tidase cleavage site (18TDP2TE22). The product of ORF8showed considerable similarity to an antirestriction protein,ArdA, of Yersinia pestis plasmid pMT-1 (24; P. Hu, J. Elliott,P. McCready, E. Skowronski, J. Garnes, A. Kobayashi, A. V.Carrano, R. Brubaker, and E. Garcia, GenBank accession no.AF053947 [http://www.ncbi.nlm.nih.gov/Entrez/index.html])and to those of pKM101 (5) and Col1bP-9 (1, 4). In addition,the predicted protein is highly acidic (calculated pI of 4.25),which is also strikingly similar to the case for other knownplasmid-encoded antirestriction proteins (for comparison,ArdA and ArdB of pKM101 have pIs of 4.05 and 4.84, respec-tively). Therefore, this gene was designated ardU. The productof ORF11 showed similarity to replication proteins of twoThermus plasmids (15, 43). Although the overall identity wasslightly higher with the RepA protein of the ATCC 27737

plasmid, the strongest conservation was found with an internalregion of the pTSp45s RepT protein (Fig. 2B). This gene wasdesignated repU. Two possible DnaA-binding sites (14) werefound within the coding region of repU (Fig. 2A and C) (seebelow). No typical direct repeats (iterons) (16, 20) were foundin the vicinity of repU, although one imperfect direct repeat (17bp in length, 13 identical) was present, with one half of therepeat encompassing the DnaA box at position 8716 (Fig. 2A).Finally, the product of ORF12 showed similarity to the XerDfamily of site-specific recombinases-resolvases, and this ORFwas designated resU. The remaining seven putative ORFsshared little or no similarities at the amino acid level with otherproteins and/or did not contain recognizable RBS sequences,and they were designated orfA, orfC, orfD, orfE, orfF, orfG, andorfH (Table 2). Of these, orfA is nevertheless of interest, sincea gene encoding a protein with 46% overall similarity wasfound adjacent to the chromosomal origin of replication ofMycobacterium smegmatis (GenBank accession no. X92503)(34). Its role in plasmid replication, however, remains un-known. The predicted product of orfF showed similarity withantifreeze glycoproteins found in arctic fish species such asBoreogadus saida (arctic cod) and Dissostichus mawsoni (noto-theniid fish) (11). Although the overall similarity appears to besignificant (Table 2), the high alanine content (640 to 50%)and the Ala-Ala-Thr repeats typical of these proteins were notobserved with the predicted product of orfH (17.1% Ala),suggesting that this similarity may not be significant.

Establishment of the minimal origin of replication. With theavailability of the entire sequence, specific combinations ofgenes could be tested to determine the minimal region re-quired for autonomous replication of pUE10. The results ofthese analyses are shown in Table 3 and Fig. 1. pI3 transformedD. radiodurans R1 at a frequency of approximately 2.45 3 105

transformants per mg of DNA, which is similar to the fre-quency described previously for pI3 (26). When part of theardU gene was removed from pI3 by deletion of a 3.1-kb CelIIfragment, transformation of D. radiodurans R1 with the result-ing plasmid, pI5, produced Cmr colonies, albeit at a 50- to100-fold-lower frequency than observed with pI3. Next, theputative repU gene of pI3 was inactivated by an out-of-framemutation, yielding pI6. This construct produced only a verysmall number of Cmr colonies (,0.2% of those produced bypI3). Using a second DrepU derivative, pI9, lacking most of theoriginal fragment present on pI3, no transformants were ob-tained in D. radiodurans R1. In contrast, transformants wereobtained with pI8 (repU1 ardU2), at a frequency similar to thatwith pI5 (ardU2). Together, these data demonstrate that theputative repU gene encodes a function necessary for replica-tion, most likely the replication initiation protein. The fewtransformants obtained with pI6 may be the result of chromo-somal integration, conceivably via htrU. This gene which is notpresent on pI9, is highly similar to the htrA homolog located onchromosome I (see above), which could provide the source ofhomology for such an event to occur. The highest nucleotidesimilarity observed was 86.5% in a continuous 52-bp sequence.

The minimal origin was further delineated by removal of theputative resolvase gene (resU), resulting in pI10. In a final step,a 0.5-kb fragment was removed from the Cmr-orfA intergenicregion of pI10. Both pI10 and the resulting plasmid, pI12, wereable to replicate in D. radiodurans R1, producing Cmr trans-formants at a frequency similar to that for the other ardU2

derivatives. Thus, the minimal region required for autonomousreplication resides within a 2,624-bp fragment containing theputative repU gene and the A1T-rich sequence downstream ofthis gene.

Next, these deletion derivatives were isolated from D. radio-


on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/


on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

durans R1, and their respective restriction patterns were com-pared to those extracted from E. coli. These patterns wereidentical (results not shown). Southern hybridization usingpI12 as a probe demonstrated that whereas pI3 and pI5 arepresent as monomers in D. radiodurans R1, pI10 and pI12(DresU) exist predominantly in a multimeric form, suggestingthe existence of an active resolution system on the resU1 plas-mids (Fig. 3) (see below). Cells harboring pI8 (resU1) alsoaccumulated multimeric forms of the plasmid, albeit at a muchlower level than pI10 and pI12. Although the copy numbers ofpI3 and its derivatives were not accurately determined, it ap-pears that these are all similar, and judged from the intensity ofbands in gel electrophoresis, they were estimated at 5 to 10copies per cell.

Construction of a shuttle vector for E. coli and D. radio-durans. The pUE10-derived promoter driving expression of thechloramphenicol resistance gene of pI3 was shown previouslyto be located within an 0.43-kb fragment (26) (GenBank ac-cession no. M94966). In the present work we have furtherlocalized this promoter by transformation of D. radioduransR1 with pI12, carrying a 532-bp AocI deletion. This constructproduced Cmr colonies at a frequency comparable to that forpI10 (see above) (Table 3). Thus, it appears that the promoterdriving expression of the resistance gene is located within 129bp between the AocI and SmaI sites of the pI3 insert. Based onthese data, the region containing the minimal replicon andCmr cassette present on pI12 was transferred to a smaller E.coli vector, pMTL23 (9), which contains an extended multiple

FIG. 2. Nucleotide sequence of the minimal replicon and resolvase gene of D. radiodurans SARK plasmid pUE10. (A) Nucleotide and amino acid sequence of therepU gene and its predicted product. The alternative start codon (TTG) at position 7966 for repU and the first six amino acids are indicated in boldface italics. Theputative DnaA-binding sites boxes are boxed; the solid arrows indicate the imperfect repeats (see text for details). Nucleotide numbers correspond to GenBankaccession no. AF206717. (B) Alignment of the replication proteins of pUE10 (pI3) and Thermus plasmid pTsp45s. Stretches of $3 identical residues or conservedreplacements are boxed. (C) Comparison of the two putative DnaA-binding sites present on pUE10, with the DnaA-binding sites of the chromosomal origins ofreplication of E. coli (Ec) and B. subtilis (Bs), and those of Thermus sp. strain YS45 plasmid pTsp45s.


on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

cloning site and has a much higher copy number in E. coli thanthe pBR322 origin of pI3. The resulting plasmid, pRAD1 (Fig.4), is 6.3 kb in size and contains an extended multiple cloningsite. It efficiently transformed D. radiodurans R1 to Cmr (Table3) and could be shuttled back to E. coli.

Expression of a plasmid-encoded reporter gene and gener-ation of a promoter probe vector. To assess the feasibility ofusing the newly constructed shuttle vector, a promoter probevector was generated using the E. coli lacZ gene. First, a BglII-XbaI fragment containing lacZ was ligated to BglII-XbaI-di-gested pRAD1, yielding pRADZ1. pRADZ1 has cloning sitesupstream of lacZ that make it potentially useful for promotercloning and analysis. To test this function, a putative promotersegment of the D. radiodurans R1 groESL genes (R. Meimaand M. Lidstrom, unpublished data) was inserted in both ori-entations in the unique BglII site of pRADZ1 that is upstreamof lacZ, and LacZ expression was analyzed. When the groESLpromoter was fused to lacZ in the proper orientation (pRADZ3),b-galactosidase activity was significant (219 nmol min21 OD600unit21) compared to that for pRADZ1 (50 nmol min21 OD600unit21) and the groESL fragment in the wrong orientation(pRADZ30) (19 nmol min21 OD600 unit21). These results dem-

onstrate that a heterologous gene can be expressed efficientlyfrom the shuttle vector and that the promoter probe vector canbe used to assess promoter activity.

Functional analysis of resU, the putative resolvase gene.Although the mechanisms underlying the stabilizing effect ofplasmid-encoded site-specific recombinases on plasmid main-tenance are not fully understood, their role has been suggestedin several studies (8, 19, 30, 33, 38). It has long been estab-lished that these so-called resolvases contribute to plasmidmaintenance by resolving plasmid multimers to monomers,thus increasing the number of segregation units to be distrib-uted during cell division (for a review, see reference 30). Fail-ure to resolve multimers prior to segregation can cause rapid

FIG. 3. Analysis of pI3 and its deletion derivatives. Agarose gel electrophore-sis (A) and Southern hybridization (B) of undigested plasmid DNA isolated fromD. radiodurans R1 are shown. Lanes: 1 and 7, molecular size marker (1-kbladder; Gibco BRL); 2 and 8, pI3; 3 and 9, pI5; 4 and 10, pI8; 5 and 11, pI10; 6and 12, pI12. In each lane, 2 ml of miniprep DNA was loaded. EaeI-digested pI12was used as a probe. The signal observed with the molecular size marker (ar-rowhead) is caused by hybridization of the vector containing the 1-kb ladder withthe pKK232-8 moiety of pI12. The asterisk indicates the position of multimericpI10 and pI12; solid arrows mark the mono- and dimeric forms of these twoconstructs. The additional signal observed in the ethidium bromide stain in panelA (open arrow) most likely represents a chromosomal DNA contamination.

TABLE 2. Putative ORFs present on the 11.9-kb pUE10 insert of pI3 and their similarities to database entries

Gene Coordinatesa RBS(spacing)

Predicted protein(amino acids; Da) Similaritiesb

orfA 105–593 ? 162; 17,759 M. smegmatis orf194 (38% identity in a 73-amino-acid overlap)orfB 659–1078 AGGAGG (5) 139; 15,085 Family of hypothetical proteinsc

htrU 1198–2436 GGAGGA (5) 412; 42,568 Family of periplasmic serine Do proteasesc

orfC 3127–3678 AGAAGG (8) 183; 19,211 Schizosaccharomyces pombe proline-rich hypothetical proteinorfD 3691–4038 ? 115; 12,022 NoneorfE 4090–4551 ? 153; 15,991 C termini of several proline-rich extensin-like proteins (e.g., Volvox carteri and

Arabidopsis thaliana; 35% identity in 148 amino acidsorfF 4678–5364 AGAGG (7) 228; 24,331 NoneardU 5422–5973 GGAGGA (9) 183; 20,139 ArdA of Y. pestis plasmid pMT1 (31% identity) and enterobacterial plasmids

pKM101 (28%) and ColIb-P9 (27%)orfG 6164–6445 AAGGAG (10) 93; 10,526 NoneorfH 6748–7749 GGAGGA (6) 333; 33,939 Antifreeze proteins (e.g., B. saida; 26% identity in 215 amino acids)repUd 7966–9159 GGAGG (8) 397; 43,311 RepA of a cryptic plasmid from Thermus sp. strain ATCC 27737 (26% identity)

and RepT of Thermus sp. strain YS45 plasmid pTSp45s (24%)7984–9159 ? 391; 42,547resU 10625–11590 AGGAGG (5) 321; 36,181 XerD family of site-specific recombinases (e.g., Proteus mirabilis;

30% identity in full length)

a Numbers correspond to the positions of the ORFs on the 11,910-bp EcoRI-HindIII pUE10-derived fragment of pI3.b Similarity searches were performed using the BLASTP algorithms (2).c See text for details.d Two possible translation initiation codons (TTG and ATG) for the putative repU gene were identified.

TABLE 3. Transformation frequencies of pI3 and its derivativesa

Plasmid Relevantgenotype

Transformation(CFU/ml/mg

of DNA)

Transformationrelative to that

of pI3 (%)

pI3 Wild type 2.45 3 105 100pI5 DardU 2.8 3 103 1.1pI6 repUb 4.5 3 102c ,0.2c

pI8 DardU 2.05 3 103 8.4pI9 DardU-repU 0 0pI10 DardU DresU 1.3 3 104 5.3pI12 DardU DresU 1.5 3 104 6.1pRAD1 DardU DresU 1.4 3 104 5.6

a Competent cells were incubated with approximately 0.2 mg of plasmid DNAon ice for 30 min and diluted 10-fold in TGY broth. Samples were plated after1 h of expression at 30°C on TGY agar supplemented with 3 mg of chloram-phenicol per ml.

b Out-of-frame mutation in the putative repU gene.c Colonies grew very slowly and had a mucoid appearance.


on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

accumulation of plasmid-free cells, even with multicopy plas-mids such as pBR322 (dimer catastrophe hypothesis [38]). Inaddition, these enzymes also contribute to the fidelity of plas-mid replication by a variety of mechanisms. The site-specificrecombinase of Streptococcus pyogenes plasmid pSM19035 wasshown to act not only as a resolvase (8) but also as an efficientDNA invertase ensuring faithful replication of both arms ofthe inverted repeats present on this plasmid (33). In anotherinc18 plasmid, pAMb1 of Enterococcus faecalis, Resb was pos-tulated to enhance faithful replication by catalyzing D-looparrest of DNA polymerase I, allowing its replacement by thehighly processive DNA polymerase III holoenzyme during theearly stages of replication (19).

The possible role of the resU-encoded resolvase homolog inthe maintenance of pI3 and its derivatives was analyzed bymeasuring the segregational stability of these plasmids undernonselective conditions (Fig. 5). The results of both direct- andreplica-plating experiments clearly showed that the DresU plas-mids tested (pI10 and pRAD1) are unstable compared to theresU-containing plasmids pI3, pI5, and pI8, which exhibited ahigh level of stability (100% Cmr after 75 generations). Fromthe data shown in Fig. 5, the appearances of Cms cells withpI10 and pRAD1 were calculated at 1.22% (regression coeffi-cient [R] 5 0.9825) and 1.32% per generation (R 5 0.9876),respectively. The structural integrity of the plasmids was veri-

fied by isolation of DNA from chloramphenicol-enriched cul-tures at the end of the assay and transformation of E. coli usingthese isolates (results not shown). The observation that theDresU constructs are unstable is in good agreement with thehigh level of multimers produced by these derivatives (Fig.3B). Together, these data suggest that the resU1 plasmidscontain both an active resolvase and its cognate resolutionsite(s) and that maintenance of pUE10 and its derivativesdepends, at least in part, on the resolvase system.

Biochemical analysis of the RepU protein. To analyze whetherthe putative replication protein is able to recognize a specificsequence within the minimal replicon, the corresponding genewas fused to a His6 tag, allowing for overexpression and pu-rification of the protein. The resulting plasmids, pQREPU1through -3, carrying the insert fused in three different readingframes, were maintained in an E. coli host containing a plas-mid-encoded copy of lacI (pREP4) to ensure full repression ofthe lac promoter. Expression was induced by addition of IPTGat various concentrations. Samples were taken at regular in-tervals and analyzed for protein content. A protein with anapparent molecular mass of about 38 kDa was observed uponinduction of cells carrying pQRESU1 (37,579 Da; resU fused tothe His6 tag in the proper frame), confirming that resU en-coded a protein of the correct size. However, RepU overex-pression could not be obtained with any of the fusions tested

FIG. 4. Physical map of the E. coli-D. radiodurans shuttle vector, pRAD1. The positions of the genes and antibiotic resistance markers as well as the uniquerestriction sites are indicated. The gray and black segments represent pI3- and pMTL23-derived sequences, respectively; the hatched box indicates the A1T-rich regiondownstream of the repU gene. See text for details on the construction of pRAD1.


on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

(Fig. 6A). The same was true when a copy of the repU genecontaining a mutation of the ATG start codon was used (prim-ers repU59Ad and repU39A). However, when E. coli M15(pREP4)was transformed with a plasmid carrying a copy of the repUgene in which the TTG start codon was mutated (pQREPU1Bd),a product with an apparent molecular mass of 45 kDa wasobserved upon induction with IPTG (predicted mass of His-RepU fusion, 44,092 Da) (Fig. 6A). These data suggest that theTTG may in fact be the true translation initiation codon (Fig.2A; Table 2). Using Ni-NTA column chromatography, the His-RepU protein was purified (Fig. 6B). The purified fusion pro-tein was subsequently applied in gel retardation experimentswith 32P-labeled MaeI restriction fragments of pRAD1AocIcovering the minimal replicon. These analyses indicated thatthe protein binds to a fragment containing the AT-rich regionand DnaA boxes, as well as to sequences upstream of the repUgene, but not to other tested regions (not shown). These initialobservations were confirmed by performing retardation assayswith (nested) PCR fragments in the regions for which bindingwas observed (Fig. 7). Under the conditions used, binding was

observed with fragments AT5 (containing a portion of theAT-rich region), DnaA1 (containing one of the putative DnaAboxes), and Prep1 (containing the region immediately upstreamof repU). As a negative control, a PCR fragment derived fromthe resU gene (present in pI8) was used. These data suggestthat the RepU protein has affinity for both the AT-rich stretchand the upstream region of the repU gene. This property couldprovide a dual mechanism for copy number control of pUE10,namely, at the levels of (i) replication initiation and (ii) auto-regulation of repU expression.

DISCUSSION

As a result of its remarkable resistance to ionizing radiationand other DNA-damaging agents, D. radiodurans has becomeone of the most promising tools for bioremediation of mixedwastes containing radionuclides. To further expand the rangeof molecular tools for genetic engineering of this bacterium, weset out to construct small, versatile shuttle vectors for E. coliand D. radiodurans. For this purpose, we determined the se-

FIG. 5. Maintenance of pI3 and a number of its derivatives in D. radiodurans R1. At the onset of the experiment, cultures were inoculated from overnight culturesgrown in TGY broth containing 3 mg of chloramphenicol per ml. Diluted samples were plated at regular intervals on selective and nonselective agar. The stability isshown as the percentage of Cmr colonies of the total viable count, as determined by replica-plating colonies from nonselective to selective agar (closed symbols). Pointsagreed within 610%. Plasmids tested were: pI3 (resU1) (F), pI5 resU1 (■), pI8 (resU1) (3), pI10 (resU2) (Œ), and pRAD1 (resU2) (}). For comparison, the opensymbols indicate data obtained from direct plating of samples on both selective and nonselective media; shown here are pI3, pI10, and pRAD1.


on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

quence of a fragment within pI3, which contains the origin ofreplication of pUE10, a cryptic plasmid from D. radioduransSARK (25, 26), a strain that is different from the one for whichthe genome sequence is available (strain R1).

The sequence of the pUE10-derived fragment of pI3 con-sists of 11,910 bp. The insert contains only three Sau3A sites,which may explain why previous attempts to isolate the mini-mal replicon of pUE10 using a shotgun approach met with nosuccess (26). The overall percent G1C content of the pUE10moiety of pI3 is 64.5, which is comparable to that of the D.radiodurans R1 chromosomes I and II and the megaplasmid(44). In contrast, the fourth genetic element found in D. radio-durans R1, a 45.7-kb plasmid, has a considerably lower G1Ccontent (56.1%) than the chromosomes and megaplasmid. Al-though pUE10 (37 kb) (25) and the 45.7-kb plasmid are ofsimilar size, this difference in percent G1C content suggeststhat these elements are of different origins, and given the highG1C content of pUE10, it is possible that this plasmid, unlikethe 45.7-kb plasmid, was already present prior to the branchingof Deinococcus and the closely related Thermus species. Anno-

tation of the nucleotide sequence revealed 12 putative ORFs,all of which are transcribed in the same direction. None ofthese showed significant similarity to ORFs on the 45.7-kb plas-mid of D. radiodurans R1, except for a XerD-type recombinase(44% overall similarity with ResU). Again, these data demon-strate that these Deinococcus plasmids are very different interms of sequence, structural organization, and, possibly, modeof replication (see below).

FIG. 6. Identification and overexpression of the RepU and ResU proteins inE. coli. The corresponding genes were fused to a His6 tag, and the resultingconstructs were maintained in strain M15(pREP4) to ensure repression of tran-scription. (A) Overexpression of His-ResU and His-RepU. Expression was in-duced by addition of 100 mM IPTG to exponentially growing cells. Samples wereprepared by boiling pelleted cells in 13 SDS sample buffer. The arrowheadsindicate the positions of the RepU (45-kDa) and ResU (38-kDa) proteins.Lanes: 1, prestained low-molecular-weight marker (Bio-Rad, Hercules, Calif.);2 and 3, pQRESU1; 4 and 5 pQRESU10; 6 and 7, pQREPU1Bd; 8 and 9,pQREPU10Bd; 10, pQE30. Even-numbered lanes represent uninduced cultures;odd-numbered lanes indicate lysates obtained from cultures induced with IPTGfor 2 h. (B) Purification of the His-RepU fusion protein by Ni-NTA chromatog-raphy. Total protein of M15(pREP4, pQREPU1Bd) was isolated from a 50 ml of100 mM IPTG-induced culture and subjected to chromatography as indicated inthe supplier’s manual. Lanes: 1, prestained low-molecular-weight marker (Bio-Rad); 2, uninduced culture at 2 h; 3, induced culture at 6 h; 4, total lysate; 5,flowthrough (ft) from Ni-NTA column; 6 and 7, wash (W) steps; 8 to 10, elution(e) steps.

FIG. 7. Binding of the His-RepU fusion protein to pI8- and pRAD1-derivedPCR fragments. (A) Purified His-RepU protein was mixed with purified 32P-labeled PCR fragment (64,000 cpm) and incubated at room temperature for 30min with the fragments shown in panel B (lanes 1 to 6); Cntrl (lane 7), PCRfragment derived from the resU region of pI8. The reaction mixtures weresubsequently electrophoresed on 4% polyacrylamide–TAE gels. (B) Schematicrepresentation of the repU region of pI3, showing the locations of the PCRfragments bound by the His-RepU fusion protein. (C) Specificity of the His-RepU interaction with the repU upstream region (32P-Prep1; see panel B). Thespecificity of this interaction was assayed by adding increasing amounts of non-competitive DNA (lanes 2 to 5) and unlabeled (Unlab.) competitive DNA (Prep1)(lanes 6 to 9); the ratio between 32P-labeled probe and poly[d(I-C)] or unlabeledprobe, respectively, is shown.


on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

Two important characteristics for plasmids used as genetictools are replication and stability. The two most interestingORFs with respect to these two characteristics are repU andresU, with products showing similarity to replication and re-solvase proteins, respectively. The RepU protein shares simi-larity with the replication initiation proteins of two differentThermus plasmids (15, 43) (Table 2). The 244-amino-acidproduct of an ORF present on the D. radiodurans R1 45.7-kbplasmid showed weak similarity with the Rep proteins of gram-positive plasmids pSM19035 (RepS) (7), pIP501 (RepR) (6),and pAMb1 (RepE) (36) but not with RepU. These observa-tions provide further support for our hypothesis that pUE10and the 45.7-kb plasmid do not have a common ancestry.Downstream of the repU gene, an A1T-rich region (42%G1C) of approximately 500 bp was found. A region similarlyrich in A and T was found downstream of the RepA gene ofthe Thermus sp. strain ATCC 27737 plasmid (15) but not inpTsp45s (43). The presence of an A1T-rich sequence isthought to facilitate strand separation for initiation of replica-tion by RepU and DnaA. DnaA was shown to be crucial forchromosome replication as well as for replication of severaltheta-replicating plasmids (20). In the latter process, the con-certed activity of the plasmid-encoded Rep protein and DnaAis thought to be involved in loading of the replication helicaseDnaB by stimulating unwinding of the plasmid origin. In ad-dition to the A1T-rich region and the DnaA-binding sites,classical oriA-containing plasmids such as P1, F, pSC101, R6K,and the broad-host-range plasmid RK2 are characterized bythe presence of a series of directly repeated sequences callediterons (for reviews, see references 16 and 20). These repeatsconstitute the primary binding site for the replication initiationproteins, thereby forming a nucleoprotein complex at the ori-gin of replication. No such repeats are present in the repUregion, suggesting that the replication initiation protein ofpUE10, as well as those of the Thermus plasmids, recognize anonrepeated sequence. However, pUE10 replication resem-bles that of the oriA-containing plasmids in terms of its inde-pendence of DNA polymerase I (18). Together, these datashow that the pUE10 origin shares some characteristics withclass A-type replicons, while the absence of a typical oriA andthe locations of the DnaA boxes are features unique to pUE10and the Thermus plasmids. DNA binding (helix-turn-helix [17])or other motifs typical of regulatory proteins were not detectedin the amino acid sequence of RepU. However, we were ableto demonstrate binding of this protein to two separate regionsof the minimal replicon, consistent with our hypothesis that theRepU protein may be involved not only in initiation of repli-cation, but also in copy number control through autoregulationof its own expression (for a review, see reference 10). Furtherwork will be required to test this hypothesis. Our presentsequence data did not reveal the presence of additional puta-tive copy number control elements in the vicinity of repU,although such functions may be present on the parental plas-mid.

Although not strictly required for replication, a plasmid-encoded resolvase (Resb) was shown to enhance the proces-sivity of DNA synthesis during initiation of pAMb1 replication(19, 31). In addition, the presence of an active resB gene greatlyenhanced the segregational stability of pAMb1 derivatives bypreventing the accumulation of plasmid multimers (39). Like-wise, the pSM19035-encoded site-specific recombinase is re-quired for both maintenance and accurate replication, by act-ing as both a resolvase and a DNA invertase (8, 33). Thestability of pI3 and its derivatives also appears to depend onthe plasmid-encoded site-specific resolvase system. Whereasthe resU1 plasmids were found to be very stable, the DresU

derivatives analyzed in our studies were lost at a rate of ap-proximately 1% per generation, suggesting that resolution ofthe multimers observed with these constructs (Fig. 3B) is cru-cial for their maintenance.

The presence of an ardA homolog is of particular interestsince thus far, these antirestriction systems were thought toexist exclusively on bacteriophages and representatives of theF, I, and P incompatibility complexes of enterobacterial plas-mids (12). Since these functions were shown to be essential forthe survival and establishment of plasmids during conjugation,this finding suggests that pUE10 might be a conjugative plas-mid. However, in an earlier study, pUE10 could not be conju-gated (37). It is not clear why ArdU appears to be importantfor high transformation frequencies in our studies, but it mayreflect a need to protect incoming transforming DNA fromdegradation.

Combining the data obtained previously by Masters andMinton (26) (GenBank accession no. M94966) with the se-quence presented here, we further narrowed down the frag-ment containing the promoter that drives expression of theCmr marker in pI3 and its derivatives. This promoter, the Cmr

marker, and the minimal replicon defined in the present workwere used to construct a versatile E. coli-D. radiodurans shuttlevector with a large number of unique restriction sites and smallsize (pRAD1). It efficiently transforms D. radiodurans to Cmr,and the lacZ reporter can be expressed at substantial levelsfrom this vector. Although pRAD1 is poorly maintained in theabsence of selective pressure, we have opted not to include theresolvase system in order to limit the size of the vector andincrease the number of available restriction sites for cloning.Our present experience with this vector is that instability is nota problem in the presence of antibiotic selection. Likewise, wehave not included the ArdU region, since the transformationfrequencies in the absence of ardU are sufficiently high forroutine work. Therefore, pRAD1 is a convenient and usefulgeneral-purpose cloning vector. In addition, pRADZ1, con-taining a lacZ reporter, is a useful vector for analyzing pro-moter activity. These vectors are present in the cell at approx-imately the same copy number as the chromosome, which ispresent at 7 to 10 copies per cell.

Together with the annotated genome sequence, the avail-ability of these molecular tools will significantly enhance thegenetic amenability of this intriguing and potentially usefulmicroorganism.

ACKNOWLEDGMENTS

We are indebted to Michael J. Daly for generously providing plas-mids pI3 and pMD209 and to John Battista for making availableD. radiodurans R1. We thank Marion Franke and Khue Quang Trinhfor technical assistance.

This work was funded by a grant from the DOE EMSP program(DEFG0797ER20294).

REFERENCES

1. Althorpe, N. J., P. M. Chilley, A. T. Thomas, W. J. Brammar, and B. M.Wilkins. 1999. Transient transcriptional activation of the Incl1 plasmid anti-restriction gene (ardA) and SOS inhibition gene (psiB) early in conjugatingrecipient bacteria. Mol. Microbiol. 31:133–142.

2. Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller,and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generationof protein database search programs. Nucleic Acids Res. 25:3389–3402.

3. Anderson, A. W., H. C. Nordan, R. F. Cain, G. Parrish, and D. Duggan. 1956.Studies on a radio-resistant micrococcus. I. Isolation, morphology, culturalcharacteristics, and resistance to gamma radiation. Food Technol. 10:575–577.

4. Belogurov, A. A., E. P. Delver, and O. V. Rodzevich. 1992. IncN plasmidpKM101 and IncI1 plasmid ColIb-P9 encode homologous antirestrictionproteins in their leading regions. J. Bacteriol. 174:5079–5085.

5. Belogurov, A. A., E. P. Delver, and O. V. Rodzevich. 1993. Plasmid pKM101


on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

encodes two nonhomologous antirestriction proteins (ArdA and ArdB)whose expression is controlled by homologous regulatory sequences. J. Bac-teriol. 175:4843–4850.

6. Brantl, S., D. Behnke, and J. C. Alonso. 1990. Molecular analysis of thereplication region of the conjugative Streptococcus agalactiae plasmid pIP501in Bacillus subtilis. Comparison with plasmids pAM beta 1 and pSM19035.Nucleic Acids Res. 18:4783–4790.

7. Brantl, S., A. Nowak, D. Behnke, and J. C. Alonso. 1989. Revision of thenucleotide sequence of the Streptococcus pyogenes plasmid pSM19035 repSgene. Nucleic Acids Res. 17:10110.

8. Ceglowski, P., A. Boitsov, S. Chai, and J. C. Alonso. 1993. Analysis of thestabilization system of pSM19035-derived plasmid pBT233 in Bacillus subti-lis. Gene 136:1–12.

9. Chambers, S. P., S. E. Prior, D. A. Barstow, and N. P. Minton. 1988. ThepMTL nic2 cloning vectors. I. Improved pUC polylinker regions to facilitatethe use of sonicated DNA for nucleotide sequencing. Gene 68:139–149.

10. Chattoraj, D. K., and T. D. Schneider. 1997. Replication control of plasmidP1 and its host chromosome: the common ground. Prog. Nucleic Acid Res.Mol. Biol. 57:145–186.

11. Chen, L., A. L. DeVries, and C. H. Cheng. 1997. Convergent evolution ofantifreeze glycoproteins in Antarctic notothenioid fish and Arctic cod. Proc.Natl. Acad. Sci. USA 94:3817–3822.

12. Chilley, P. M., and B. M. Wilkins. 1995. Distribution of the ardA family ofantirestriction genes on conjugative plasmids. Microbiology 141:2157–2164.

13. Daly, M. J., L. Ouyang, P. Fuchs, and K. W. Minton. 1994. In vivo damageand recA-dependent repair of plasmid and chromosomal DNA in the radi-ation-resistant bacterium Deinococcus radiodurans. J. Bacteriol. 176:3508–3517.

14. Datta, H. J., G. S. Khatri, and D. Bastia. 1999. Mechanism of recruitment ofDnaB helicase to the replication origin of the plasmid pSC101. Proc. Natl.Acad. Sci. USA 96:73–78.

15. de Grado, M., I. Lasa, and J. Berenguer. 1998. Characterization of a plasmidreplicative origin from an extreme thermophile. FEMS Microbiol. Lett. 165:51–57.

16. del Solar, G., R. Giraldo, M. J. Ruiz-Echevarrıa, M. Espinosa, and R.Dıaz-Orejas. 1998. Replication and control of circular bacterial plasmids.Microbiol. Mol. Biol. Rev. 62:434–464.

17. Dodd, I. B., and J. B. Egan. 1990. Improved detection of helix-turn-helixDNA-binding motifs in protein sequences. Nucleic Acids Res. 18:5019–5026.

18. Fuchs, P., H. Agostini, and K. W. Minton. 1994. Defective transformation ofchromosomal markers in DNA polymerase I mutants of the radioresistantbacterium Deinococcus radiodurans. Mutat. Res. 309:175–184.

19. Janniere, L., V. Bidnenko, S. McGovern, S. D. Ehrlich, and M. A. Petit. 1997.Replication terminus for DNA polymerase I during initiation of pAM beta1 replication: role of the plasmid-encoded resolution system. Mol. Microbiol.23:525–535.

20. Kornberg, A., and T. A. Baker. 1992. DNA replication, p. 637–682. W. H.Freeman and Co., New York, N.Y.

21. Krasin, F., and F. Hutchinson. 1977. Repair of double-strands breaks inEscherichia coli, which requires recA function and the presence of a duplicategenome. J. Mol. Biol. 116:81–98.

22. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4. Nature 227:680–685.

23. Lange, C. C., L. P. Wackett, K. W. Minton, and M. J. Daly. 1998. Engineer-ing a recombinant Deinococcus radiodurans for organopollutant degradationin radioactive mixed waste environments. Nat. Biotechnol. 16:929–933.

24. Lindler, L. E., G. V. Plano, V. Burland, G. F. Mayhew, and F. R. Blattner.1998. Complete DNA sequence and detailed analysis of the Yersinia pestisKIM5 plasmid encoding murine toxin and capsular antigen. Infect. Immun.66:5731–5742.

25. Mackay, M. W., G. H. Al-Bakri, and B. E. B. Mosely. 1985. The plasmids ofDeinococcus spp. and the cloning and restriction mapping of the D. radio-philus plasmid pUE1. Arch. Microbiol. 141:91–94.

26. Masters, C. I., and K. W. Minton. 1992. Promoter probe and shuttle plas-mids for Deinococcus radiodurans. Plasmid 28:258–261.

27. Miller, J. H. 1982. Experiments in molecular biology. Cold Spring HarborLaboratory Press, Cold Spring Harbor, N.Y.

28. Murray, R. G. E. 1992. The family Deinococcaceae, p. 3732–3744. In A.Balows, H. G. Truper, H. Dworkin, W. Harder, and K.-H. Schleifer (ed.),The prokaryotes, 2nd ed. Springer Verlag, New York, N.Y.

29. Nielsen, H., J. Engelbrecht, S. Brunak, and G. von Heijne. 1997. Identifica-tion of prokaryotic and eukaryotic signal peptides and prediction of theircleavage sites. Protein Eng. 10:1–6.

30. Nordstrom, K., and S. Austin. 1989. Mechanisms that contribute to thestable segregation of plasmids. Annu. Rev. Genet. 23:37–69.

31. Pujol, C., S. D. Ehrlich, and L. Janniere. 1994. The promiscuous plasmidspIP501 and pAM beta 1 from gram-positive bacteria encode complementaryresolution functions. Plasmid 31:100–105.

32. Riley, R. G., J. M. Zachara, and F. J. Wobber. 1992. Chemical contaminantson DOE lands and selection of contaminant mixtures for subsurface scienceresearch. Subsurface Science Program, U.S. Department of Energy, Wash-ington, D.C.

33. Rojo, F., and J. C. Alonso. 1994. A novel site-specific recombinase encodedby the Streptococcus pyogenes plasmid pSM19035. J. Mol. Biol. 238:159–172.

34. Salazar, L., H. Fsihi, E. de Rossi, G. Riccardi, C. Rios, S. T. Cole, and H. E.Takiff. 1996. Organization of the origins of replication of the chromosomesof Mycobacterium smegmatis, Mycobacterium leprae and Mycobacterium tu-berculosis and isolation of a functional origin from M. smegmatis. Mol.Microbiol. 20:283–293.

35. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: alaboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, N.Y.

36. Simon, D., and A. Chopin. 1988. Construction of a vector plasmid family andits use for molecular cloning in Streptococcus lactis. Biochimie 70:559–566.

37. Smith, M. D., R. Abrahamson, and K. W. Minton. 1989. Shuttle plasmidsconstructed by the transformation of an Escherichia coli cloning vector intotwo Deinococcus radiodurans plasmids. Plasmid 22:132–142.

38. Summers, D. K., C. W. H. Beton, and H. L. Withers. 1993. Multicopy plasmidinstability: the dimer catastrophe hypothesis. Mol. Microbiol. 8:1031–1038.

39. Swinfield, T. J., L. Janniere, S. D. Ehrlich, and N. P. Minton. 1991. Char-acterization of a region of the Enterococcus faecalis plasmid pAM beta 1which enhances the segregational stability of pAM beta 1-derived cloningvectors in Bacillus subtilis. Plasmid 26:209–221.

40. Udupa, K. S., P. A. O’Cain, V. Mattimore, and J. R. Battista. 1994. Novelionizing radiation-sensitive mutants of Deinococcus radiodurans. J. Bacteriol.176:7439–7446.

41. Vagner, V., E. Dervyn, and S. D. Ehrlich. 1998. A vector for systematic geneinactivation in Bacillus subtilis. Microbiology 144:3097–3104.

42. Wackett, L. P., and D. T. Gibson. 1988. Degradation of trichloroethylene bytoluene dioxygenase in whole-cell studies with Pseudomonas putida F1. Appl.Environ. Microbiol. 54:1703–1708.

43. Wayne, J., and S.-Y. Xu. 1997. Identification of a thermophilic plasmid originand its cloning within a new Thermus-E. coli shuttle vector. Gene 195:321–328.

44. White, O., J. A. Eisen, J. F. Heidelberg, E. K. Hickey, J. D. Peterson, R. J.Dodson, D. H. Haft, M. L. Gwinn, W. C. Nelson, D. L. Richardson, K. S.Moffat, H. Qin, L. Jiang, W. Pamphile, M. Crosby, M. Shen, J. J.Vamathevan, P. Lam, L. McDonald, T. Utterback, C. Zalewski, K. S.Makarova, L. Aravind, M. J. Daly, K. W. Minton, R. D. Fleischmann, K. A.Ketchum, K. E. Nelson, S. Salzberg, H. O. Smith, J. C. Venter, and C. M.Fraser. 1999. Genome sequence of the radioresistant bacterium Deinococcusradiodurans R1. Science 286:1571–1577.

45. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. Improved M13 phagecloning vectors and host strains: nucleotide sequences of the M13mp18 andpUC19 vectors. Gene 33:103–119.


on Septem


.asm.org/

Dow

nloaded from

http://aem.asm.org/

Documents

Characterization of the Minimal Replicon of a Cryptic ...Deinococcus promoter was shown to efﬁciently express LacZ. Ever since its discovery in 1956 (3), Deinococcus radiodurans