Vol. 170, No. 8

Transformation in Bacillus subtilis: Involvement of the 17-KilodaltonDNA-Entry Nuclease and the Competence-Specific

18-Kilodalton ProteinBEN VOSMAN, GAUKO KUIKEN, JAN KOOISTRA, AND GERARD VENEMA*

Department of Genetics, University of Groningen, Kerklaan 30, 9751 NN Haren, The Netherlands

Received 28 December 1987/Accepted 18 May 1988

A protein complex, consisting of a 17-kilodalton (kDa) nuclease and an 18-kDa protein, is believed to beinvolved in the binding and entry of donor DNA during transformation of Bacillus subtilis (H. Smith, K.Wiersma, S. Bron, and G. Venema, J. Bacteriol. 156:101-108, 1983). In this paper, the nucleotide sequencesof the genes encoding both the nuclease and the 18-kDa protein are presented. The genes are encoded by a904-base-pair PstI-HindIll fragment. The open reading frames encoding both proteins are partly overlapping.A B. subtilis mutant was constructed by insertion of a Cmr marker into the gene encoding the nuclease. Thismutant lacked the competence-specific nuclease activity and the 18-kDa protein but retained 5% residualtransformation. The total DNA association of the mutant was higher than that of the wild-type cells, and DNAentry was reduced to 30% of the wild-type level. These results suggest that an alternative pathway exists forthe internalization of transforming DNA. A mutant, exclusively deficient for the 18-kDa protein, previouslysuggested to be involved in the binding of transforming DNA, was constructed by insertion of a kanamycinresistance gene into the coding sequence of the gene. Since the mutant showed wild-type DNA-binding activity,the 18-kDa protein is probably not involved in the binding of donor DNA to competent cells. The transformingactivity of the mutant was reduced to 25% of the wild-type level, indicating that the 18-kDa protein has afunction in the transformation process. In vitro experiments showed that the 18-kDa protein is capable ofinhibiting the activity of the competence-specific nuclease. Its possible role in transformation is discussed.

Bacillus subtilis competence is associated with the synthe-sis of a specific set of proteins. Some of these are assumed tobe involved in the binding, the entry, and the subsequentintegration of the donor DNA (1, 6, 17, 25).

Binding and entry of donor DNA were reported to begoverned by a 75-kilodalton (kDa) protein complex (26, 27).This complex consists of two subunits of the DNA-entrynuclease, with a molecular mass of 17 kDa, and two subunitsof an 18-kDa protein. Mutants which lacked the 18-kDaprotein were found to be DNA binding deficient, suggestingthat this protein is involved in DNA binding (25). In vitroexperiments showed that only the total complex was capableof binding single-stranded, as well as double-stranded, DNA(28). No DNA binding was observed with the purified18-kDa protein (28).Four competence-deficient mutants, mapping in different

chromosomal regions, have been characterized (7). Three ofthese appeared to be DNA binding deficient. Protein analysisshowed that one of the three DNA binding-deficient mutantsstill produced the 18-kDa protein, as well as the 17-kDanuclease (1), suggesting that, in addition to the 75-kDaprotein complex, other proteins are essential for DNAbinding and DNA entry. In addition, it has been reportedthat a 38.5-kDa competence-specific protein capable of bind-ing double-stranded DNA can be isolated from B. subtilis(8).

Nucleases involved in the entry of donor DNA intocompetent Streptococcus pneumoniae and B. subtilis cellswere identified previously (17, 22). Comparison of the nu-clease activities present in competent and noncompetentwild-type cells, as well as analysis of nuclease activitypresent in DNA entry-deficient mutants, led to the identifi-

* Corresponding author.

cation of at least three competence-specific nuclease activi-ties in B. subtilis, with apparent molecular masses of 14, 17,and 28 kDa (1, 16, 17). It has been suggested that the 14-kDanuclease is derived from the 17-kDa polypeptide (27), whichis part of the 75-kDa protein complex referred to above, andthat the 28-kDa nuclease is a dimer of the 14-kDa nuclease(27).

In a previous paper, we have reported on the cloning, inEscherichia coli, of a 700-base-pair (bp) PstI-EcoRI DNAfragment encoding 14- and 17-kDa nuclease activities (32). Inthe present paper, we report the nucleotide sequence of thegene encoding the DNA-entry nuclease, as well as that of the18-kDa protein. We also describe the construction andcharacterization of mutants deficient in the 14- and 17-kDanuclease activities, as well as the 18-kDa protein.

MATERIALS AND METHODSBacterial strains. The bacterial strains used are shown in

Table 1.Media. Minimal medium for B. subtilis consisted of Spizi-

zen minimal salts (29), supplemented as previously de-scribed (25). Starvation medium consisted of minimal saltsplus 0.5% glucose. Minimal agar consisted of minimal saltssupplemented with the required growth factors and 1.5%agar. TY medium and TY agar were prepared as previouslydescribed (2).

Chemicals. Chemicals used were of analytical grade andwere obtained from E. Merck AG (Darmstadt, FederalRepublic of Germany) or BDH (Poole, England). Restrictionenzymes, T4 DNA polymerase, and T4 DNA ligase wereused as recommended by the manufacturer (BoehringerGmbH, Mannheim, Federal Republic of Germany).

Purification of DNA. B. subtilis chromosomal DNA waspurified from strain OG1 as previously described (30). Radio-

3703

JOURNAL OF BACTERIOLOGY, Aug. 1988, p. 3703-37100021-9193/88/083703-08$02.00/0Copyright © 1988, American Society for Microbiology

on March 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

3704 VOSMAN ET AL.

TABLE 1. Strains used

Strain Relevant genotype Source or reference

B. subtilisOG1 Prototrophic 42G8 thy thyr-J Our laboratory collection8G5 trpC2 his met tyr-I 48G370 Cmr This work8G206 Kmr This work

E. coliJM83 A(Iac-proAB) +80 lacZAM15 31JM101 A(lac-proAB) (F' tra36 33

proAB+ lacIq ZAM15)

active DNA was obtained from strain 2G8 grown in thepresence of 1 mCi of [methyl-3H]thymidine (specific activity,20.7 mCi/mg; Radiochemical Centre, Amersham, UnitedKingdom) per 100 ml of minimal medium. The specificactivity of the DNA preparations was 1.5 x 105 cpm/Lg ofDNA. Plasmid DNA was purified by the method of Ish-Horowicz and Burke (11). EMBL465/1 DNA was purified as

previously described (9).Transformations. E. coli was made competent and trans-

formed as described previously (14). B. subtilis cells weremade competent essentially as described by Bron and Ve-nema (4).

Southern hybridization. DNA was transferred to Gene-Screen-Plus membrane as described by Maniatis et al. (15).Probes were nick translated with [a-32P]dCTP according toRigby et al. (19). Hybridization was performed as recom-mended by the supplier of the GeneScreen-Plus (NENResearch Products GmbH, Dreieich, Federal Republic ofGermany).

Preparation of lysates from E. coli. Lysates were preparedas previously described (32). Protein concentrations weredetermined as previously described (3), with bovine serumalbumin as a standard.

Isolation of membrane vesicles. Membrane vesicles wereisolated from the wild-type and mutant strains essentially as

described previously (12). Membranes were finally sus-

pended in 10 mM Tris hydrochloride (pH 7.5) and stored at-20°C until use. Protein concentrations were measured as

described by Bradford (3), with bovine serum albumin as a

standard.PAGE. Sodium dodecyl sulfate-polyacrylamide gel elec-

trophoresis (SDS-PAGE) was performed on slab gels ac-

cording to Laemmli (13).Two-dimensional gel electrophoresis. Proteins were sepa-

rated in the first dimension by isoelectric focusing, as

described by O'Farrell (18) under conditions described bySmith et al. (25). The second dimension (SDS-PAGE) was

performed on slab gels as described by Laemmli (13).Labeling of cellular proteins with [35S]methionine. Labeling

was performed under conditions described previously (25).Fluorography. The procedure used for fluorography was

previously described (24).Total association, entry, and breakdown of transforming

DNA. These quantities were essentially determined as de-scribed by Mulder and Venema (16).

Detection of nuclease activity after gel electrophoresis. Themethod used was essentially as described by Rosenthal andLacks (21). SDS-PAGE was performed on a 15% gel. In theseparation gel, 10 p,g of calf thymus DNA per ml was

incorporated. After electrophoresis, the gel was washed indistilled water for 1 h (one change), followed by 2 h of

B P

FIG. 1. Plasmid pGV290. Arrows indicate the direction of tran-scription of the kanamycin resistance (Kmr) and the nuclease (nuc)gene. ori, Origin of replication. C, H, HIl, B, P, Y, and E representClaI, HindIII, Hindll, BcIl, PstI, BalI, and EcoRI sites, respec-tively.

washing with 25 mM Tris hydrochloride (pH 7.4) buffer (onechange). The nuclease bands were developed overnight byincubation of the gel at 37°C in 25 mM Tris hydrochloridebuffer (pH 7.4), containing 5 mM MnCl2. Subsequently, thegels were stained with 5 p.g of ethidium bromide per ml.

Assay of nuclease activity in E. coli lysates. The methodused was adapted from Smith et al. (26). Assays wereperformed in a volume of 30 ,ul during 30 min at 37°C, afterwhich part of the sample was applied to a 0.8% agarose gel.Incubations were carried out in a 10 mM Tris hydrochloride(pH 7.4)-10 mM MnCI2 buffer to which 5 ,ug of E. coli lysateand 4 ,ug of plasmid DNA were added.

Sequence analysis. Fragments were sequenced in twodirections according to Sanger (23), with universal 17-merM13 primer. For internal priming, 17-mer primers were

synthesized on the basis of the nucleotide sequence deter-mined with the universal primers.

Construction of mutant 8G370. For the construction of thisB. subtilis mutant, deficient in the formation of the 14- and17-kDa nuclease activities, we used a derivative of pGV290.Plasmid pGV290 (Fig. 1) was obtained by cloning into pGV1of a HindIII-EcoRI DNA fragment of lambda phage EMBL465/1, which hybridized to pGV211 (32). The mutant wasconstructed in the following way. First, the pWVO1 repli-cation origin of plasmid pGV290 (carrying the nuclease gene[Fig. 1]) was fortuitously removed when pGV290 wascleaved with ClaI and ligated into AccI-cleaved pUC7 (31),and E. coli JM83 was transformed with the ligation mixture.The Kmr Apr transformants appeared to carry pUC7 con-

taining the ClaI-EcoRI part of pGV290, carrying the Kmrgene. This plasmid, designated pGV330, resulted in allprobability from recombination between the lacZ sequencespresent in pGV290 (to the left of the EcoRI site) and thesame sequence in pUC7, which should be present as a directrepeat flanking the EcoRI-ClaI fragment of pGV290, carry-ing the pWVO1 replicative origin, in the fused replicons.Because of the symmetry of the pUC7 multiple cloning site,the ClaI-EcoRI fragment of pGV290 (encoding kanamycinresistance and nuclease activity) thus became contained inan EcoRI fragment.

Next, the continuity of the nuclease gene on pGV330 (AprKmr) was interrupted by inserting the Cmr gene of pC194(10) in the BalI-cleaved pGV330 DNA. This results in a small

oripWV01 KmR

J. BACTERIOL.

on March 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TRANSFORMATION IN BACILLUS SUBTILIS 3705

7-bp deletion, because two BalI sites are present in thenuclease gene. To this purpose, pC194 was digested withDral (these sites flank the Cmr gene) and blunt end ligatedinto the BalI site of pGV330. Cmr Apr transformants of JM83were selected. One of the transformants contained a plasmidwhich carried a'1,300-bp insert conferring resistance tochloramphenicol. This plasmid, designated pGV370, wascleaved with EcoRI and was used to transform competentcells of B. subtilis 8G5. The plasmid was treated with therestriction enzyme to prevent integration by a Campbell-likemechanism. Cmr Kms transformants were obtained. Thesetransformants must have arisen from replacement recombi-nation in which the cross-over events took place between theplasmid-carried chromosomal DNA sequences flanking theCmr marker and the resident chromosome. The mutant wasdesignated 8G370.

Construction of plasmids used. The plasmids shown in Fig.2 were constructed in the following ways: to obtain plasmidpGV199, the previously described plasmid pGV191 (32) wascleaved with PstI and ClaI, and the 2,100-bp PstI-ClaIfragment, containing the genes encoding the DNA-entrynuclease and the 18-kDa protein, was ligated into PstI-Accl-cleaved pUC9 (31). This places the genes under transcrip-tional control of the lacZ promoter of pUC9. PlasmidpGV210 was previously described (32); it only encodes theDNA-entry nuclease which is expressed under control of thelacZ promoter. Plasmid pGV250 was obtained by ligation ofthe EcoRI fragment of pGV199 in the same relative orienta-tion as that of pGV199 into EcoRI-cleaved pUC9. Directlyadjacent to the ClaI-AccI fusion point, pGV199 contains theEcoRI site of the multiple cloning site of pUC9. PlasmidpGV282 was obtained by deleting the PstI-SacII fragmentfrom pGV199. For this purpose, PstI-SacII-cleaved pGV199was treated with T4 DNA polymerase prior to ligation.

RESULTS

Cloning in E. coli of the gene encoding the 18-kDa protein.We have previously reported on the cloning of a chromo-somal PstI-ClaI fragment of 2,100 bp in plasmid pHV60,resulting in plasmid pGV191. This plasmid contained a700-bp PstI-EcoRI fragment encoding the 14- and 17-kDaDNA-entry nucleases (32).The 2,100-bp PstI-ClaI fragment was inserted into plasmid

pUC9, resulting in plasmid pGV199 (Fig. 2). Furthermore,pUC9 derivatives were made containing the SacII-ClaIfragment'and the EcoRI-ClaI fragment from pGV191. Theseare the plasmids pGV282 and pGV250, respectively (Fig. 2).E. coli JM83 was transformed with these plasmids. Lysatesof these transformants were analyzed by SDS-PAGE. Inlysates prepared from cultures containing pGV199 andpGV282, an intense protein band at the 18-kDa position wasdetected, which was absent in lysates prepared from culturescontaining pUC9 and pGV250 (results not shown). Thissuggests that, in addition to the gene specifying the nucleaseactivity, the 2,100-bp fragment also contains the codingsequence of the 18-kDa protein, and that the EcoRI site islocated within the coding region of this protein. Furtheranalysis showed that both the nuclease activities and the18-kDa protein were encoded by the 900-bp PstI-HindIIIfragment (results not shown).

Nucleotide sequence analysis. To determine the nucleotidesequence of the DNA fragment encoding the 14- and 17-kDanucleases, as well as the 18-kDa protein, the 700-bp PstI-EcoRI fragment of plasmid pGV199 was cloned in M13mp18and M13mpl9 (33). This fragment encodes the 14- and

P S E HI I a 0

CpGV 199

pGV210-i pGV250-i pGV282

lkb

FIG. 2. Schematic representation of the relevant part of theplasmids pGV199, pGV210, pGV250, and pGV282. All fragmentswere cloned in the multiple cloning site of pUC9, as described inMaterials and Methods, in the same relative orientation with respectto the lacZ promoter, initiating transcription from the left to theright. P, S, E, H, and C represent PstI, SacII, EcoRI, HindIII andClaI sites, respectively. Plac is the promoter of the lacZ gene. kb,Kilobase.

17-kDa nucleases. Also cloned in M13 was the EcoRI-Hindlll fragment of pGV199, which is located downstreamof the PstI-EcoRI fragment. Finally the PstI-HindIII frag-ment of pGV199 was cloned in these vectors, to read acrossthe EcoRI site.The sequence derived from this analysis, together with

putative ribosome binding sites and the relevant restrictionsites, is shown in Fig. 3. From the PstI site to the HindIIlsite, it contains 904 bp. The sequence contains two partlyoverlapping open reading frames. One of these is betweenposition 81 to 512. This frame can encode a protein of 16,194Da, which is probably the 17-kDa DNA-entry nuclease.Although the reading frame that is supposed to encode the17-kDa nuclease is open to the 5' end of the DNA sequencesuggested, it is unlikely that this region of DNA contains17-kDa protein-encoding sequences, because this shouldhave been evident from the molecular weight of the DNA-entry nuclease.

In the same frame, there is a possibility for a secondarytranslational start at position 156 (GTG). When this startwould be used, a protein of 13,367 Da would -be produced,which might correspond to the 14-kDa nuclease.The second open reading frame starts at position 418 and

ends at position 813. This frame can- encode a protein of14,980 Da, which possibly represents the 18-kDa protein.Experiments supporting this view are presented in a subse-quent section of this paper.The ribosome-binding sites preceding the genes encoding

the nucleases are relatively weak as compared with severalother ribosome-binding sites in B. subtilis, which mightexplain the relatively weak expression of the genes. This isin contrast to the ribosome-binding site preceding the geneencoding the 18-kDa protein, which is synthesized in largeamounts in competent cells.

Construction and characterization of a B. subtilis deficientfor the 14- and 17-kDa nucleases. To examine the contribu-tion of the nucleases to transformation, a strain was con-structed in which the nuclease gene was interrupted. For theconstruction of this mutant, plasmid pGV370, in which thecoding sequence of the nuclease gene had been interruptedat the BalI site (Fig. 3) by insertion of the Cmr gene, wasused. The construction of the plasmid is described in Mate-rials and Methods. Transformation of B. subtilis with EcoRI-cleaved pGV370 yielded strain 8G370. Figure 4 shows therelevant part of the chromosome of the wild-type strain OG1and the expected genetic structure of the mutant. The

VOL. 170, 1988

on March 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

3706 VOSMAN ET AL.

IP 10 20 30 40 50 60

7080 90 100 110 120

MetA1a*.ProgerAprgTyrProGltutrAlaLysH130 140 150 160 170 180

isIletyslA1aIleAsnGllyHiserGlUValCysmrIleAs9 AG1yA

190 200

laGluGluArgArgGluGInSerIaul25s p 260A7W7WCCPAMI0CAR7GCAA

spGIuTrpPr0etA1aMfetCya310 S1 320

H P P Y EH COG1. I,

H P PQ R Q E H Ci . .

-. .

H P P Y E H KmR EH CKb

. . R *.

lKb

80370

80206

_ 240 FIG. 4. Relevant part of the chromosome of the wild type (OG1)tLysApValPro6er[ysLy MyTyrAq3ArgA and mutants 8G370 and 8G206. The arrows indicate the nuclease

270 280 290 300gene. H, B, P, Y, E, and C represent HindIII, BclI, PstI, Ball,

2706280 290 ~EcoRI, and ClaI sites, respectively. Q represents the BalI-DraIyG1yGtuG1yA1aSerV&lGlu rIleS fusion. Cmr and Kmr represent the chloramphenicol and kanamycin

resistance genes, respectively. The chromosomal DNA fragment330_ 340 350 _ 360 present in plasmid pGV290 (---), which was used as the probe in

erProlumr¶rAlaGluGInA1lauG1ySerGlylleGlyleuroIle'hrGInT370 380 390 400 410 420

hr/laGlnxrgPhtdyrSerG1fPhbrfrtysGn,tyrIleGluGlGuG1GnRisIlAMet

430 440 450 460 470 480

ntltGlu XlS_nd"silMtMWVlr~Ln490 500 S10 520 530 540

550 560 570 580 590 600

AllaAapGlyA1aI1eSerPhsGluA1aG1nAgul ltilnaPhlIl1IuPhe

610 620 630 640 650 660

ArgluAsnSerSerG1uThM lAMtryrGluLysZLYum zalPraPhonis

670 IE 680 690 700 710 720

ValThrGluAinGly!leRisleliusIle9erlgIarrmRtgptLu730 740 750 760 770 780

Pro yGlqfyrG1nL.tlIryWprpl!hrha1proA1a1uDetbrAsp(atgI790 800 810 820 830 840

Al trleIlleA*AaValSrV&1850 860 870 880 890 900m~~~~ 1mCAeA

IH

FIG. 3. Nucleotide sequence of the PstI-HindIII fragment en-coding the 14- and 17-kDa nucleases, as well as the 18-kDa protein.The sequence starts at the PstI site (base 1) and ends at the HindlIlsite (base 904). The predicted amino acid sequence based on thenucleotide sequence is shown below the open reading frames. Theputative transcription termination signal is indicated by arrows. Theoverlining above the sequence indicate the putative ribosome-binding sites. Relevant restriction sites are as follows; P, B, S, E,and H represent PstI, Ball, SacII, EcoRI, and HindIII sites,respectively.

the Southern hybridizations, is indicated. Kb, Kilobase.

Figure 6 shows that these activities were not detectable inthe mutant. However, when 100 ,ug of protein was applied tothe gel, after prolonged incubation (3 days), a nuclease bandwas visible just above the position of the 17-kDa nucleaseband. This nuclease migrated to a different position on atwo-dimensional nuclease detection gel, suggesting the pres-ence of a different nuclease (results not shown).We also examined mutant 8G370 for the presence of the

18-kDa protein. Two-dimensional gel electrophoresis (iso-electric focusing PAGE) was performed on total extracts of[35S]methionine-labeled competent cultures. The results(Fig. 7) show that the mutant also lacked the 18-kDa protein.This might suggest that the genes specifying the nucleasesand 18-kDa protein are part of one operon.Table 2 shows that the transformation frequency of the

mutant was approximately 5% of that of the wild type, thattotal DNA association was higher, and that DNA entry and

_M 320(C

,_0 _1900

-* -1800. -155(

U.4

construction was checked by Southern hybridization (Fig.5). As expected, the data indicate that the 700-bp PstI-EcoRIfragment, encoding the nuclease gene, was replaced by a2,000-bp PstI-EcoRI fragment, as a result of insertion of theCmr gene. Similarly, the HindlIl fragment of 1,900 bp in thewild type was enlarged to 3,200 bp in the mutant. Theseresults indicate that the Cmr gene had been integrated in thenuclease gene in the correct way.Mutant 8G370 was examined for the presence of the 14-

and 17-kDa nuclease activities on a nuclease detection gel.

45(.

FIG. 5. (A) Southern hybridization of PStI-EcoRl- and HindIII-cleaved chromosomal DNA of the wild type (OG1) and mutant8G370. Lanes: 1, OG1 cleaved with PstI and EcoRI; 2, OG1 cleavedwith HindIII; 3, 8G370 cleaved with PstI and EcoRI; 4, 8G370cleaved with HindIll. (B) Southern hybridization pattern of HindIll-cleaved OG1 (wild-type) DNA (lane 1) and mutant 8G206 DNA (lane2). Molecular weights (in base pairs) are indicated in the rightmargin. Plasmid pGV290 was used as the probe.

J. BACTERIOL.

on March 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TRANSFORMATION IN BACILLUS SUBTILIS 3707

1 2 ,.1PP-l .......'

...A

.-W

j0-dot.w A

*0"_

-17

- 14

_..B

0

I* -1 _I.. .- 4

o.eW t 4

- ._.-

-0Om .0.,

-: ~ y

FIG. 6. Nuclease activities present in membrane vesicles of thewild type and mutant 8G370 of B. subtilis. Nuclease activity wasvisualized as described in Materials and Methods. For each sample,15 jig of protein was applied to the gel. Lanes: 1, 8G5 (wild type); 2,8G370. The positions of the 14- and 17-kDa nucleases are indicated.The activity indicated by the arrow is caused by DNase I, used inthe preparation of the membrane vesicles.

the formation of acid-soluble products were less than thosein the wild type. The fact that the mutant was still endowedwith 5% residual transforming activity may suggest thepresence of a minor, alternative pathway for the internaliza-tion of transforming DNA.

Construction and characterization of an 18-kDa protein-deficient B. subtilis strain. To determine exclusively the effectof the 18-kDa protein on transformation, a mutant was

constructed in which the gene encoding the 18-kDa proteinwas disrupted. For this purpose, pGV191 (32) was cleavedwith EcoRI, and the EcoRI fragment, containing the Kmrgene from plasmid pPJ1 (B. P. H. Peeters, J. H. De Boer, S.Bron, and G. Venema, Mol. Gen. Genet., in press), was

ligated into it. The EcoRI site is shown in Fig. 3. The ligationmixture was transformed to E. coli JM83 and Tcr Kmrtransformants were selected. One of the plasmids obtained,designated pGV206, was linearized with ClaI and used totransform competent cells of B. subtilis 8G5. A Kmr Cmstransformant, designated 8G206, was selected and analyzedfurther.

Southern hybridization was performed to verify whetherthe Kmr gene was inserted in the correct way into thechromosome. Figure 4 shows the genetic structure of therelevant part of the wild type and that expected for themutant chromosome. The result of the Southern hybridiza-tion (Fig. SB) shows that the 1,900-bp HindIll fragment inthe wild type was replaced by two hybridizing fragments of1,800 and 1,550 bp, respectively, indicating that the Kmrgene had been inserted by replacement recombination. Hy-bridization of the 1,550-bp HindlIl fragment resulted fromthe presence of the Kmr gene, which was part of thehybridization probe.Two-dimensional gel electrophoresis was performed to

investigate the presence of the 18-kDa protein in the wild-type and the mutant strain. The protein was not detected instrain 8G206 (result not shown).

Analysis of nucleases present in membranes of mutant

FIG. 7. Two-dimensional gel electrophoresis of [35S]methionine-labeled proteins of the wild-type 8G5 (A) and mutant 8G370 (B).Both strains were subjected to the competence regimen, as de-scribed in Materials and Methods. The 18-kDa protein is marked bythe arrow. Isoelectric focusing was from the left (basic) to the right(acid).

cells subjected to the standard competence regimen showedthat the 14- and 17-kDa nucleases involved in the entry ofdonor DNA were present in amounts similar to those presentin the wild type (results not shown).The fact that insertion of the Kmr gene in the EcoRI site

(see Fig. 3) resulted in the absence of the 18-kDa protein isconsistent with the hypothesis that the second open readingframe specifies the 18-kDa protein.

TABLE 2. Transformation, total DNA association, DNA entry,and DNA breakdown of transforming DNA in an insertionally

inactivated nuclease-deficient mutant"

3H radioactivity (106 cpm/CFU)Strain Transformation

frequencyb Total Entryd Breakdowneassociationc

8G5 0.85 260 140 2558G370 0.041 410 52 75

a The transformation frequencies were derived from six independent exper-iments with trp+ DNA (1 ptg/ml). Total association, entry, and breakdowndata are from duplicate experiments.

b (Number of trp+ transformants/total count) x 100.' Total amount of radioactivity associated with cells both sensitive and

resistant to DNase I.d Amount of DNase I-resistant radioactivity associated with the cell.e Acid-soluble radioactivity in transformation mixture, derived from trans-

forming DNA.

VOL. 170, 1988

II :;,..1m.

n. 4b.-AO-

.Aw - ..

on March 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

3708 VOSMAN ET AL.

TABLE 3. Transformation, total DNA association, DNA entry,and breakdown of transforming DNA in an 18-kDa

protein-deficient mutanta

3H radioactivity (106 cpm/CFU)Strain Transformation

frequency' associationt Entryd Breakdowne

8G5 0.80 732 41 448G206 0.22 672 35 42

a The transformation frequency was calculated from four independentexperiments. Total association, entry, and breakdown data were from dupli-cate experiments.

b-e Footnotes as defined for Table 2.

The data presented in Table 3 show that, although themutant transformed less than the wild type (residual trans-forming activity of approximately 25%), the total DNAassociation, DNA entry, as well as breakdown of transform-ing DNA were hardly affected by the absence of the 18-kDaprotein. The differences between the data concerning thebinding, entry, and breakdown for the wild type in Tables 2and 3 originate from the use of different DNA preparations.

Effect of the 18-kDa protein on the nuclease activity. Sincethe mutant was only moderately reduced in transformation,the question can be raised what the function of the 18-kDaprotein could be. As shown above, it does not seem to beinvolved in the binding of donor DNA to competent cells.The protein is only found in the competent fraction of a B.subtilis culture, and has been shown to be associated withthe 17-kDa nuclease in membrane fractions (25, 26).

Smith et al. (27, 28) have shown that addition of thepurified 18-kDa protein to purified entry nuclease reducesthe activity of the nuclease, which led them to speculate thatthe 18-kDa protein might modulate the nuclease activity. Toinvestigate whether this inhibition occurred in E. coli carry-ing the genes encoding the nuclease and the 18-kDa protein,E. coli lysates were prepared from strains containing theplasmids shown in Fig. 2. For this purpose, plasmid DNAwas incubated for 30 min in the presence of crude lysates andsubsequently analyzed by agarose gel electrophoresis. Fig-ure 8A shows that in lysates prepared from JM83(pGV210),carrying only the nuclease gene, the added plasmid DNAwas completely degraded. Lysates prepared from JM83containing pGV199, which contained the intact nucleasegene in addition to the gene encoding the 18-kDa protein, didnot show any plasmid degradation. Analysis of the lysates bySDS-PAGE using a DNA-containing gel (21) clearly showedthe presence of nucleases in the lysates prepared from JM83(pGV199) and JM83(pGV210) (results not shown). Appar-ently, the plasmid-encoded nuclease activity was completelyinhibited in the lysate when E. coli also produced the 18-kDaprotein. Experiments in which lysates were mixed showedsimilar results. When the lysate from JM83(pGV210) wasmixed with lysates prepared from strains containing pGV199and pGV282, no degradation was observed (Fig. 8B). Asmentioned before, these lysates contained a prominent 18-kDa protein. Figure 8 also shows that the mobility of theplasmid was not influenced by the addition of the lysates.This suggests that the 18-kDa protein did not bind to theDNA.When, instead of a JM83(pGV210) lysate, DNase I was

used in a concentration just capable of degrading all theplasmid DNA in the assay mix, no inhibiting effect wasfound upon addition of lysates prepared from a strain con-taining pGV199 or pGV282 (results not shown). This sug-gests that the observed inhibition of the 14- and 17-kDa

FIG. 8. Plasmid degradation by E. coli lysates. Plasmid pGV191DNA was incubated with E. coli lysates and subsequently subjectedto agarose gel electrophoresis. The breakdown of the plasmid DNAindicates nucleolytic activity. Lane 1 in panels A and B showspGV191 incubated in buffer without lysate. (A) Nuclease activity inlysates prepared from E. coli JM83, containing the plasmids pUC9(lane 2), pGV210 (lane 3), pGV199 (lane 4), pGV250 (lane 5), andpGV282 (lane 6). (B) Nuclease activity in mixed lysates. Lanes 3 to6 show the effect on the nuclease activity when lysates preparedfrom JM83(pGV210) were preincubated for 10 min with the follow-ing lysates before pGV191 was added. Lane 2, no other lysate addedto the JM83(pGV210) lysate; lane 3, JM83(pUC9) added; lane 4,JM83(pGV199) added; lane 5, JM83(pGV250) added. lane 6,JM83(pGV282) added.

nucleases is the result of an interaction between the nucleaseand the 18-kDa protein, rather than being caused by generalbinding of the 18-kDa protein to the DNA, which also mightresult in protection against the nucleases.

DISCUSSION

The nucleotide sequence of the PstI-HindIII fragmentrevealed that there are two partially overlapping open read-ing frames on this fragment. The first one (from position 81to 512) is presumed to encode the 14- and 17-kDa nucleases.This conclusion was drawn from the observation that thePstI-EcoRI fragment encodes nuclease activity in E. coli(32), whereas the SacII-ClaI fragment of pGV282 does not(unpublished data), indicating that the SacII site (position313 [Fig. 3]) is located within the coding sequence. This openreading frame can encode a protein of 16,194 Da, which isclose to the previously determined molecular weight of theDNA-entry nuclease (17, 26). The presence of a second startcodon, at position 156, 13 bp downstream of a putativeribosome-binding site within the same open reading frame,offers the alternative that the 14-kDa nuclease is not derivedfrom the 17-kDa nuclease, as previously suggested (27), butis translated independently with the GTG as start codon.Whether this putative start is actually used remains to beestablished.The second open reading frame is partially overlapping

with the first one and starts at position 418. This openreading frame can encode a protein of 14,980 Da. In spite ofthe discrepancy between the size of the protein based onsequence analysis and the size determined by two-dimen-sional gel electrophoresis, we may conclude that the secondopen reading frame specifies the 18-kDa protein, because ofthe following observations. (a) The protein is produced in E.coli strains carrying the plasmids pGV199 and pGV282 but

J. BACTERIOL.

on March 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TRANSFORMATION IN BACILLUS SUBTILIS 3709

not in a strain harboring the plasmid pGV250, indicating thatthe EcoRI site (at position 674) is located within the codingsequence for this protein. (b) The protein was also producedin an E. coli strain containing pUC18 (33) carrying the clonedPstI-HindIII fragment of pGV199 (unpublished data), indi-cating that the coding region ended before the HindlIl site.(c) A B. subtilis strain carrying a Kmr marker in the EcoRIsite also lacked the 18-kDa protein. Although all of theseobservations strongly suggest that the 18-kDa protein isspecified by the sequence between position 418 and 816, itremains unclear why the apparent molecular weight of theprotein is larger than that predicted from the sequence.The strong polar effect of the introduction of the Cmr gene

in the coding sequence of the nuclease gene on the expres-sion of the 18-kDa protein as well as the absence of apromoterlike structure preceding the second open readingframe (Fig. 3) suggest that both genes are organized in oneoperon. However, the possibility that the 18-kDa protein isquickly degraded in the absence of the nuclease cannot beexcluded but seems unlikely, since in the absence of the17-kDa nuclease, the 18-kDa protein is stable in E. colilysates.

Just after the stop codon at position 816, an invertedrepeat can be identified which may act as a rho-independenttranscription termination signal (20).A mutation was created, by insertion of a Cmr gene into

the BalI site, which abolished the 14- and 17-kDa nucleaseactivities. Nevertheless, the mutant still showed 5% residualtransforming activity (Table 2). The total DNA associationwas somewhat higher than in the wild type, and the DNAentry was reduced to approximately 30% of the wild-typelevel. The fact that the mutant still showed considerableDNA uptake and that the transformation was reduced byonly a factor of 20 suggests that an alternative pathway mayexist which can internalize donor DNA. It would seem thatin this pathway also a nuclease is involved, since the mutantwas still capable of generating acid-soluble products fromtransforming DNA (Table 2). It is unclear whether, or towhat extent, this alternative pathway plays a role in trans-formation when the 14- and 17-kDa nucleases are present.The results presented in this paper show that it is unlikely

that the 18-kDa protein is involved in the binding of donorDNA to competent B. subtilis cells, as has been previouslysuggested (25). A mutant exclusively deficient for the 18-kDaprotein was perfectly capable of binding donor DNA (Table3). The transformation frequency in this mutant was reducedto approximately 25% of that of the wild-type level. Thissuggests that, although the protein is not responsible forbinding of DNA, it is to some extent involved in transfor-mation. This is in accordance with the view that both thenuclease and the 18-kDa protein genes seem to be organizedin an operon, and with the observation that the 18-kDaprotein is competence specific (25).A possible role of the 18-kDa protein in transformation

may be related to its capacity to inhibit the DNA-entrynuclease; it is conceivable that the uncontrolled action of thenuclease might degrade the DNA during internalization tosuch an extent that the resulting molecules are no longeroptimally integrated in the resident chromosome (5). Thismight explain the slightly reduced transformability of strain8G206.

ACKNOWLEDGMENTSWe are grateful to Henk Mulder for preparing the figures, and to

the Unilever Research Laboratory (Vlaardingen, The Netherlands)for synthesis of the 17-mer primers.

This work was supported by The Netherlands Organization forthe Advancement of Pure Research (The Hague, The Netherlands)under auspices of The Netherlands Foundation for Chemical Re-search.

LITERATURE CITED1. Barberio, C., R. Coppolecchia, G. Mastromei, and M. Polsinelli.

1985. Competence proteins in Bacillus subtilis commutants.Biochim. Biophys. Acta 842:184-188.

2. Biswal, N., A. K. Kleinschmidt, H. C. Spatz, and T. A. Trautner.1967. Physical properties of the DNA of bacteriophage SP50.Mol. Gen. Genet. 100:39-55.

3. Bradford, M. M. 1976. A rapid and sensitive method for thequantitation of microgram quantities of protein utilizing theprinciple of protein dye binding. Anal. Biochem. 72:248-254.

4. Bron, S., and G. Venema. 1972. Ultraviolet inactivation andexcision-repair in Bacillus subtilis. I. Construction and charac-terisation of a eightfold auxotrophic strain and two ultraviolet-sensitive derivatives. Mutat. Res. 15:1-10.

5. Cato, A., and W. R. Guild. 1968. Transformation and DNA size.I. Activity of fragments of defined size and a fit to a randomdouble cross-over model. J. Mol. Biol. 37:157-178.

6. De Vos, W. M., and G. Venema. 1981. Fate of plasmid DNA intransformation of Bacillus subtilis protoplasts. Mol. Gen.Genet. 182:39-43.

7. Fani, R., G. Mastromei, M. Polsinelli, and G. Venema. 1984.Isolation and characterization of Bacillus subtilis mutants al-tered in competence. J. Bacteriol. 157:152-157.

8. Finn, C. W., and 0. E. Landmann. 1985. Competence relatedproteins in the supernatant of competent cells of Bacillussubtilis. Mol. Gen. Genet. 198:329-335.

9. Frischauf, A.-M., H. Lehrach, A. Poustka, and N. Murray. 1983.Lambda replacement vectors carrying polylinker sequences. J.Mol. Biol. 170:827-842.

10. Iordanescu, S. 1975. Recombinant plasmid obtained from twodifferent, compatible staphylococcal plasmids. J. Bacteriol. 124:597-601.

11. Ish-Horowicz, D., and F. J. Burke. 1981. Rapid and efficientcosmid cloning. Nucleic Acids Res. 9:2989-2999.

12. Konings, W. N., A. Bisschop, M. Veenhuis, and C. A. Vermeu-len. 1973. New procedure for the isolation of membrane vesiclesof Bacillus subtilis and an electron microscopy study of theirultrastructure. J. Bacteriol. 116:1456-1465.

13. Laemmli, U. K. 1970. Cleavage of structural protein during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

14. Mandel, M., and A. Higa. 1970. Calcium-dependent bacterio-phage DNA infection. J. Mol. Biol. 53:159-162.

15. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

16. Mulder, J. A., and G. Venema. 1982. Isolation and partialcharacterization of Bacillus subtilis mutants impaired in DNAentry. J. Bacteriol. 150:260-268.

17. Mulder, J. A., and G. Venema. 1982. Transformation-deficientmutants of Bacillus subtilis impaired in competence-specificnuclease activities. J. Bacteriol. 152:166-174.

18. O'Farrell, P. H. 1975. High resolution two-dimensional electro-phoresis of proteins. J. Biol. Chem. 250:4007-4021.

19. Rigby, P. W. J., M. Dieckman, C. Rhodes, and P. Berg. 1977.Labeling deoxyribonucleic acid to high specific activity in vitroby nick translation with DNA polymerase I. J. Mol. Biol. 133:237-251.

20. Rosenberg, M., and D. Court. 1979. Regulatory sequencesinvolved in the promotion and termination of RNA transcrip-tion. Annu. Rev. Genet. 13:319-353.

21. Rosenthal, A. L., and S. A. Lacks. 1977. Nuclease detection inSDS-polyacrylamide gel electrophoresis. Anal. Biochem. 80:76-90.

22. Rosenthal, A. L., and S. A. Lacks. 1980. Complex structure ofthe membrane nuclease of Streptococcus pneumoniae revealedby two-dimensional electrophoresis. J. Mol. Biochem. 141:133-146.

VOL. 170, 1988

on March 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

3710 VOSMAN ET AL.

23. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequenc-

ing with chain-terminating inhibitors. Proc. Natl. Acad. Sci.USA 74:5463-5467.

24. Skinner, M. K., and M. D. Griswold. 1983. Fluorographicdetection of radioactivity in polyacrylamide gels with 2,5-diphenyloxazole in acetic acid and its comparison with existingprocedures. Biochem. J. 209:281-284.

25. Smith, H., W. M. De Vos, and S. Bron. 1983. Transformation inBacillus subtilis: properties of DNA-binding-deficient mutants.J. Bacteriol. 153:12-20.

26. Smith, H., K. Wiersma, S. Bron, and G. Venema. 1983. Trans-formation in Bacillus subtilis: purification and partial character-ization of a membrane-bound DNA-binding protein. J. Bacte-riol. 156:101-108.

27. Smith, H., K. Wiersma, G. Venema, and S. Bron. 1984. Trans-formation in Bacillus subtilis: a 75,000-dalton protein complex isinvolved in binding and entry of donor DNA. J. Bacteriol. 157:733-738.

28. Smith, H., K. Wiersma, G. Venema, and S. Bron. 1985. Trans-

formation in Bacillus subtilis: further characterization of a

75,000-dalton protein complex involved in binding and entry ofdonor DNA. J. Bacteriol. 164:201-206.

29. Spizizen, J. 1958. Transformation of biochemically deficientstrains of Bacillus subtilis by deoxyribonucleate. Proc. Natl.Acad. Sci. USA 44:1072-1078.

30. Venema, G., R. H. Pritchard, and T. Venema-Schroder. 1965.Fate of transforming deoxyribonucleic acid in Bacillus subtilis.J. Bacteriol. 89:1250-1255.

31. Vieira, J., and J. Messing. 1982. The pUC plasmids, an

M13mp7-derived system for insertion mutagenesis and sequenc-ing with synthetic universal primers. Gene 19:259-268.

32. Vosman, B., J. Kooistra, J. OliJve, and G. Venema. 1987.Cloning in Escherichia coli of the gene specifying the DNA-entry nuclease of Bacillus subtilis. Gene 52:175-183.

33. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. ImprovedM13 phage cloning vectors and host strains: nucleotide se-

quences of the M13mpl8 and pUC19 vectors. Gene 33:103-119.

J. BACTERIOL.

on March 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from