Vol. 153, No. 3JOURNAL OF BACTERIOLOGY, Mar. 1983, p. 1424-14310021-9193/83/031424-08$02.00/0Copyright C 1983, American Society for Microbiology

Cloning Structural Gene sacB, Which Codes for ExoenzymeLevansucrase of Bacillus subtilis: Expression of the Gene in

Escherichia coliPHILIPPE GAY,1 DOMINIQUE LE COQ,1 MICHEL STEINMETZ,' EUGENIO FERRARI,2t AND

JAMES A. HOCH2*Ldboratoire GMn&tique et Membranes, Institut Jacques Monod, Centre National de la Recherche ScientUfique

et Universite Paris VII, Tour 43, 75251 Paris Cedex 05, France,' and Department of Cellular Biology,Research Institute of Scripps Clinic, La Jolla, California 920372

Received 20 September 1982/Accepted 30 December 1982

A clone bearing the structural gene sacB, coding for the exoenzyme levansu-crase, was isolated from a library of Bacillus subtilis DNA that was cloned inphage A charon 4A on the basis of the transforming activity of the chimeric DNA.This A clone also was found to contain the sacR and smo loci. Subcloning thesacB-sacR region in plasmid pBR325 resulted in a clone which directed levansu-crase synthesis in Escherichia coli. The nucleotide sequence coding for thesecreted protein was localized on the physical map of the cloned DNA.

Levansucrase (sucrose:2,6-P-D-fructan 6-n-D-fructosyltransferase; E.C. 2.4.1.10) is an en-zyme secreted in culture medium by Bacillussubtilis after induction by sucrose. The enzymecatalyzes mainly the following transfructosyla-tion reaction: sucrose + acceptor -+ glucose +acceptor-fructose.

Since the enzyme was purified (17), it hasbeen extensively investigated. Physical proper-ties of the protein have been described byGonzy-Treboul et al. (20) and Berthou et al. (2).The enzyme consists of a single polypeptidechain of 50,000 daltons. The sequence of the 444amino acids of the secreted enzyme was recentlyelucidated by A. Delfour (unpublished data). Amodel of the three-dimensional structure at 3.8-A resolution was proposed by Le Brun and VanRappenbusch (27). Detailed enzymological stud-ies by Chambert and Gonzy-Treboul showedthat the enzyme obeys a ping-pong mechanismwhich involves a covalent fructosyl-enzyme in-termediate which has been isolated (10-12).The genetic analysis of sucrose metabolism in

B. subtilis by Lepesant and co-workers led tothe identification of several loci related to levan-sucrase synthesis (29). The stuctural gene (sacB)coding for the enzyme was defined by mutationsthat affected either the thermostability of levan-sucrase or its ability to synthesize levans (31). Aregulatory locus (sacR) was defined by muta-tions resulting in constitutive synthesis of theenzyme. The two loci, sacB and sacR, mapped

t Present address: Genentech, Inc., South San Francisco,CA 94080.

close to one another and were localized betweenthe hisA and smo markers (29).We took advantage of the sacB mutants isolat-

ed by Lepesant et al. (29, 31) to screen for thepresence of wild-type sacB alleles in the libraryof B. subtilis DNA cloned in A charon 4A(described in a previous paper [19]). A clonebearing the sacB, sacR, and smo genes wasisolated. This communication describes the iso-lation of this clone, the reconstitution of thesacB-sacR locus in a plasmid, and the identifica-tion of levansucrase synthesized by the E. colihost of the plasmid.

MATERIAU AND METHODSBacterhl strains and cloning vectors. The strains and

vectors used are listed in Table 1. Plasmid pH4 is aderivative of pHV33 (37) obtained by partial HhaIdigestion. It has lost the pC194 functions and theEcoRI site of the pBR322 moiety of pHV33. Thenomenclature concerning the various phenotypes re-lated to levansucrase synthesis was taken from Lepe-sant et al. (29). Lvs+/Lvs- stands for the ability/inabi-lity to synthesize levansucrase as characterized by itssaccharolytic activity. Lvsc/Lvsi stands for the consti-tutive/inducible synthesis of the enzyme. Lev+/Lev-stands for the ability/inability of levansucrase to syn-thesize levans on appropriate solid medium. For ex-ample, strain QB2060 (sacR2 sacB204 cysB3) constitu-tively synthesizes a levansucrase, the activity of whichcan hydrolyze sucrose, but it is poorly able to poly-merize fructose (31). Its phenotype is Lvs' Lvsc Lev-Cys-. The gene responsible for the smooth/roughphenotype was cloned in our experiments. In mostlaboratories, the reference strain 168M carries a smoallele, the origin of which is not clearly identified. It isnoteworthy that the wild-type phenotype is rough (21).

1424

on April 5, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

CLONING OF THE sacB GENE OF B. SUBTILIS 1425

TABLE 1. List of the strains and cloning vectors usedBacterial strain or cloning Genotype Relevant Origin

vector phenotypeB. subtilis168M trpC2 smo Trp- SmoQB1058 sacA321 sacB197 sacS32 trpC2 smo Trp- Lev- Smo IRBM"PG790 sacA321 sacB197 sacS32 smo Lev- Smo IRBMQB2060 sacA321 sacB204 sacR2 cysB3 Lvs+ Lvsc Lev- IRBM

Rou Cys-PG795 sacA321 sacB204 sacR2 Lvs+ Lvsc Lev- IRBM

RouE. coli cloning hostsDP50 supF F- dapD8 lacY A(gal-uvrB) AthyA F. R. Blattner

nalA hdsS suII suIJIHVC45 thrAl leu-6 thi-J lacYJ supE44 hsdR Amps Cms Tets S. D. Ehrlich

rpsL tonAICloning vectors

X charon 4A F. R. Blattner(4)

A charon 4A derivativesA LSC66 This paperA gutF38 This paperpBR325 Amp' Cm' Tetr J. Doly (see

reference 6)pBR325 derivativespLS1, pLS2, pLS3, Ampr Cms Tetr This paperpLS4

pLS8, pLS10 Ampr Cm' Tets This paperpH4 Ampr Tetr I. JonespH4 derivativespLS5, pLS8 AmpT Tets This papera IRBM, Institut de Recherches en Biologie Moleculaire, Centre National de la Recherche Scientifique.

Media. The multiplication of X charon phage wassupported by Escherichia coli strain DP50 (supF)grown in a medium containing: tryptone (Difco Labo-ratories), 20 g/liter; yeast extract (Difco), 10 g/liter;NaCl, 10 g/liter; MgC12, 10 mM; diaminopimelic acid(Sigma Chemical Co.), 100 mg/liter; thymidine (Sig-ma), 50 mg/liter. LB medium (tryptone, 10 g; yeastextract, 5 g; NaCl, 10 g; water, 1,000 ml) was used as asolid or liquid medium to support the growth of otherE. coli strains.

B. subtilis was grown on tryptose blood agar base(Difco) plates or in MM (1) mineral medium. SG standsfor MM mineral medium supplemented by sucrose (20g/liter) and glucose (1 g/liter).DNA isolation and purification. (i) Bacteriophage k

DNA. Phage lysates were concentrated by precipita-tion (PEG 6000 [J. T. Baker Chemical Co.], 100 g/liter;NaCl, 0.5 M). Phages were then purified by CsClgradient centrifugation. DNA was extracted twicewith phenol, and residual phenol was extracted withether. After ethanol precipitation, the DNA was solu-bilized in TE (Tris-hydrochloride, 10 mM [pH 7.5];EDTA, 1 mM).

(ii) Plasmid DNA. For large-scale isolation, strainscarrying plasmids were grown according to the methodof Bolivar et al. (7), and plasmid amplification wasobtained by the addition of either chloramphenicol(170 mg/liter) or spectinomycin (300 mg/liter) (6). DNAwas extracted according to the method of Humphreyset al. (24). Plasmid DNA was purified by CsCl bandingor hydroxyapatite chromatography according to themethod of Colman et al. (14). For small-scale isolation,

we used the rapid alkaline procedure of Birnboim andDoly (3).

B. subtilis transforming DNA was extracted accord-ing to the method of Marmur (32).DNA analysis and cloning. Enzymes BamHI, EcoRI,

Hindlll, and bacteriophage T4 DNA ligase were pre-pared at the Research Institute of Scripps Clinic byAlain Meier. Enzymes KpnI and PvuII were fromBethesda Research Laboratories, ClaI was from NewEngland Biolabs, and HpaII was from BoehringerMannheim Corp. They were used according to thedirections of the manufacturers.

Slab gel electrophoresis (Sigma agarose, catalog no.A 6877) was run in Tris-acetate buffer (40 mM Trisbase, 20 mM sodium acetate, 2 mM EDTA, adjusted topH 7.9 with acetic acid). Charon 4A DNA, doublydigested with EcoRl and BamHI, or pBR322 DNA,digested with HpaII, was used as the molecular weightstandard.

Genetic procedures. E. coli strains were transformedaccording to the method of Dagert and Erhlich (15).Cmr, Tetr, and Ampr colonies were selected on LBplates supplemented with the related antibiotic(s):chloramphenicol, 15 mg/liter; tetracycline, 15 mg/liter;and ampicillin, 200 mg/liter, respectively.

B. subtilis strains were transformed according to themethod of Anagnostopoulos and Spizizen (1), includ-ing the use of the MG I and MG II media of Borensteinand Ephrati-Elizur (8). Prototrophic transformantswere selected on MM plates lacking either tryptophanor cysteine. The Smo/Rou phenotype was scored onTBAB plates by replica plating of patched colonies.

VOL. 153, 1983

on April 5, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

1426 GAY ET AL.

Levansucrase-related phenotypes were scored accord-ing to the method of Lepesant et al. (29). The Lev'phenotype was identified by the synthesis of levan,which surrounded the colonies on SG plates. TheLvsc/Lvsi phenotype was identified on coloniespatched on MM plates supplemented with glycerolinstead of glucose. After 24 h of growth at 37°C, theplates were sprayed with a mixture of chloramphenicol(2 mg/ml) and sucrose (500 mg/ml), incubated at 37°Cfor an additional 90 min, and then sprayed with GODPerid reagent (Boehringer Mannheim). The constitu-tive phenotype results in a green halo surrounding thecolonies.

Bacterial extacts and levansucrase assays. E. colicells treated with EDTA (1 mM in Tris-hydrochloride,100 mM, pH 7.5) were sedimented and suspended in50 mM potassium phosphate buffer, pH 6.0. Lyso-zyme (1 mg/ml) was added, and the cells were incubat-ed at 0°C for 30 min. Lysis was achieved by freezingand thawing. Extracts were treated with DNase (50,ug/ml) to avoid excess viscosity.Levansucrase was estimated according to the meth-

od of Chambert et al. (12). Uniformly labeled ['4C]su-crose and ['4C]glucose were obtained from AmershamCorp. They were purified by paper chromatographybefore use. Transfructosylation on glucose (exchangereaction) was assayed in the following mixture:['4C]glucose, 200 mM (0.1 uCi/,Lmol); sucrose, 100mM; potassium phosphate buffer, 50 mM, pH 6.0.Transfructosylation on levan was characterized in thefollowing mixture: [(4C]sucrose, 240 mM; purifiedlevan (molecular weight, 10,000), 10 mg/ml; potassiumphosphate buffer, 50 mM, pH 6.0. In the absence ofacceptor levan, incubation of levansucrase in thismixture resulted mainly in sucrose hydrolysis.

In all cases, portions of the reaction mixtures were

subjected to paper chromatography (washed paper,Schleicher & Schuell catalog no. 2043a; solvent, n-

butanol-acetic acid-water, 4:1:1 [vol/vol/vol). The ra-

dioactive spots were localized by autoradiography.The radioactivity of each spot was estimated directlyfrom the paper immersed in scintillation fluid.

RESULTS

Isolation of a clone carrying the sacB gene froma X charon 4A library. The construction of alibrary ofB. subtilis DNA originating from strain168M (trpC smo) cloned in the EcoRI sites ofcharon 4A was described previously. When thislibrary was tested for transforming activity on

competent B. subtilis, 70% of the markers testedwere found in a sample of 1,710 plaques (19).The same sample was tested for the presence ofclones that carried the structural gene coding forthe exoenzyme levansucrase. Transformation ofsacB mutants to wild-type does not allow adirect selection of recombinants. However,wild-type Lev' colonies can easily be distin-guished from the Lev- background on SG medi-um since they are surrounded by levan. Toidentify Lev' colonies in the transformed popu-lation, transformants were selected by congres-sion with an unlinked auxotrophic marker.A competent culture of strain QB1058

(sacB197 trpC2 sacS2) was exposed to a subsa-turating amount of DNA from strain PG790(sacB197 sacS2) and plated on SG medium to getless than 5,000 Trp+ colonies per plate (theappropriate concentration of PG790 DNA waspreviously determined by a saturation curve).DNA originating from each of 19 subpools ofabout 90 plaques of X recombinant phages wasspotted on the lawn of competent cells. Amongthe Trp+ colonies thus selected, Lev+ colonieswere found in two spots, indicating that somerecombinant clones carrying sacB alleles werepresent in two subpools. Crude lysates wereprepared from each plaque of one subpool andwere used further as DNA sources in a transfor-mation experiment as described above. Thisallowed the identification and isolation of the XLSC66 plaque, the DNA of which transformsthe sacB197 allele. The clone was purified twiceand then was grown in large amounts for furtherstudies.

Genetic mapping of A LSC66 insert. Previouswork by Lepesant et al. (29) showed that thesacB gene mapped close to the regulatory locussacR, not far from the smo marker (position 305in the map compiled by Henner and Hoch [23]).hisA, sacB, sacR, smo, and cysB were furtherfound by PBS1 transduction to be linked in thisorder (M. Steinmetz, unpublished data). Thepresence of sacB, sacR, and smo markers in theinsert of A LSC66 was tested with strain QB2060(sacA321 sacB204 sacR2 cysB3) as the recipientand a mixture of DNA originating from strainPG795 (sacA321 sacB204 sacR2) and X LSC66as the donor DNA. The patched Cys+ recombi-nants were checked for unselected markers onappropriate media. Table 2 shows that the allelessacB204, sacR2, and smo were transformed.Moreover, the recombination frequencies be-tween the markers were consistent with theorder found in transduction experiments.

Subcloning the B. subtils DNA insert of ALSC66. A preliminary physical map of the B.subtilis DNA insert of X LSC66 was establishedwith regard to EcoRI and HindIII sites (Fig. 1).It showed that five EcoRI sites delimited fourpieces of 2.1, 3.4, 6.0, and 5.1 kilobases (kb),respectively, in this order. Isolation of thesefragments from agarose gels allowed us to local-ize the sacB transforming activity on the 2.1-kbpiece, localized at the left end of the insert (datanot shown).An EcoRI digest of X LSC66 DNA was ligated

in the EcoRI site of plasmid pBR325 (6), and theligation mixture was used to transform E. colistrain HVC45. Among the Tetr, Ampr, Cmscolonies, recombinant plasmids carrying eachEcoRI piece were found (Fig. 1). The plasmidDNAs were used to transform the QB2060 strainas described above. The results (Table 2)

J. BACTERIOL.

on April 5, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

CLONING OF THE sacB GENE OF B. SUBTILIS 1427

TABLE 2. Genetic characterization of the markers cloned in A LSC66 and the derived subclonesaNo. in Cys+ recombinant classes"

Relevant Lev' Lev-Cross donor Lvsi Lvsc Lvs' Lsvc Alleles detected

Smo Rou Smo Rou Smo Rou Smo Roud

1 A LSC66 27 17 2 3 8 5 37 105 sacB sacR smo2 X LSC66 47 40 6 5 NSe NS NS NS sacB sacR smo3 pLS1 0 0 0 16 0 0 0 192 sacB4 pLS2 0 0 0 0 0 22 0 186 sacR5 pLS4 0 0 0 0 0 0 35 173 smo6 pLS5 0 39 0 13 0 40 0 408 sacB sacRa Recipient strain, QB2060: sacA321 sacB204 sacR2 cysB3 Lsv+ Lsvc Lev- Rou Cys-.b The DNA cloned in X LSC66 and the derived subclones was from strain 168M (trpC2 smo).c The congression of the cysB marker was utilized only to select the competent cells (see text).d Phenotype of the recipient strain; no recombination detected in the locus investigated.' NS, Not scored.

showed that transforming activities for sacB,sacR, and smo were present, respectively, inpLS1, pLS2, and pLS4, thus giving an unambig-uous map of this region of the B. subtilis chro-mosome.

Expression of the sacB gene in A LSC66 andreconstitution of the sacB sacR loci in a plasmid.We looked for a potential expression of sacB inE. coli cells infected by A LSC66. The presenceof the highly specific exchange activity of levan-sucrase (fructosyl transfer from sucrose to glu-cose) was tested in either crude lysates or in E.coli cells harvested in the stationary phase thatprecedes lysis. Both crude lysates and infectedcells were found to contain approximately 0.015U of enzyme per ml of culture, whereas nodetectable activity (less than 0.001 U) was foundunder the same conditions in cells infected byanother clone, k gutF38 (a clone which containspart of the gut operon; P. Gay, unpublisheddata). This result suggested that K LSC66 con-tained a functional sacB gene.No levansucrase activity was found in E. coli

A LS 66

soc B sac R

EH H E H E

XC4Lend'12.1

pILSL 3.41

pLS 12

3.15

095 pLSS

IpLSIO 4.1

cells carrying plasmid pLS1, suggesting that the2.1-kb EcoRI fragment contained in pLS1 wasinsufficient to direct levansucrase synthesis. Toobtain a plasmid with an insert capable of codingfor the synthesis of levansucrase, HindIlI frag-ments of k LSC66 were subcloned. DNA of KLSC66 was digested by HindIII and ligated inthe HindlIl site of plasmid pH4. The selection ofAmpr Tets transformants in E. coli allowed theisolation of a clone carrying plasmid pLS5, theinsert of which overlaps both the right end ofpLS1 and the left end of pLS2 inserts (Fig. 1).Table 2 shows that DNA of pLS5 transformedboth sacB204 and sacR2 alleles. All the allelestested further, sacB91, sacB182, sacB194,sacB236, sacR37, sacR41, and sacR44 (29), werealso transformed by pLS5 DNA (data notshown). However pLS5 did not direct the syn-thesis of levansucrase activity in E. coli hosts.We assumed that pLS5 was lacking one end ofthe sacB gene. Thus, we attempted to constructa plasmid sharing inserts of both pLS1 andpLS5. To generate this plasmid, K LSC66 DNA

smo

E E

IXC4Rend

i6.0

pLS3 5-.

pLS4

pLS

FIG. 1. Organization of the B. subtilis DNA inserts of A LSC66 and of the derived subclones. E and H stand,respectively, for EcoRI and HindIIl restriction sites. Five additional HindIII sites present in pLS4 were notmapped.

VOL. 153, 1983

-T-

on April 5, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

1428 GAY ET AL.

FIG. 2. Physical map of the sacB-sacR region and schematic diagram of the reconstitution of a functionalsacB gene in pLS7 and pLS8 plasmids. The 1,332 base pairs (bp) encoding the known amino acid sequence of thesecreted levansucrase are indicated as follows: dotted area, origin of the gene; oblique hatching, middle part ofgene; horizontal hatching, terminus of the gene. C, E, H, K, and P stand, respectively, for Clal, EcoRl, HindIIl,KpnI, and PvuII restriction sites.

digested by EcoRI was ligated in the uniqueEcoRI site present in the insert of pLS5 (Fig. 2).The recombinant plasmid pLS7 was identifiedby systematic restriction analysis of the DNA ofAmpr transformants. The 5.5-kb insert of thisplasmid contains 4.1 kb of the left end of XLSC66, part of this material being duplicated.Levansucrase was found to be synthesized by

pLS7 hosts. However, we observed that Lvs+colonies of E. coli segregated Lvs- colonies at arelatively high rate. These Lvs- colonies werefound to carry a plasmid similar to pLS5, proba-bly originating from internal recombination be-tween the duplicated segments (data not shown).The availability of both precise physical map-

ping of the sacB-sacR region and the sequenceof the protein allowed a precise location of thesacB gene. A. Delfour (personal communica-tion) elucidated the complete sequence of the444 amino acids of the secreted levansucrase ofB. subtilis. The localization of 11 restriction siteson the physical map of the sacB region (6 ofthem indicated in Fig. 2) was found to be per-fectly consistent with the amino acid sequenceof the protein. This comparison provided anunambiguous positioning of the nucleotide se-quence coding for the secreted form of theenzyme. This sequence indicated that pLS1 andpLS5 plasmids were lacking, respectively, theorigin and the terminus of the gene, which wasconsistent with the absence of expression ofsacB in their hosts.

It appeared that a convenient way to clone ashorter insert that would allow sacB expressionwas to start from a partial HindlIl digestion ofeither A LSC66 or pLS7. Such a partial digestionof pLS7 DNA was cloned in the HindlIl site ofpBR325. The experiment allowed the isolationof plasmid pLS8 in which the tet gene of pBR325was inactivated (Fig. 2).

Identification of levansucrase in E. coli hosts ofpLS8. A crude extract of strain HVC45 carryingpLS8 was obtained as described above. Thefructosyl transferase activities of levansucrasesynthesized by this strain were characterized bythree different reactions (Fig. 3): (i) the ex-change reaction (fructosyl transfer from sucroseto radioactive glucose); (ii) the hydrolysis ofsucrose; and (iii) the fructosyl transfer to accep-tor levan. Since neither exchange reaction norlevan synthesis could be shown in E. coli strainscarrying vectors that did not contain sacB al-leles, these experiments identify the expressionof the sacB gene. The specific activity of theenzyme was found to be constant during expo-nential growth, 0.8 U per mg of bacterial pro-tein.

DISCUSSIONThe use of a library of B. subtilis DNA cloned

in X charon 4A (4) allowed us to isolate a DNAfragment which included the structural genesacB coding for the exoenzyme levansucraseand the regulatory locus sacR. It is worthy to

J. BACTERIOL.

on April 5, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

CLONING OF THE sacB GENE OF B. SUBTILIS 1429

7 2 3 4 5FIG. 3. Characterization of lev,

sized by the E. coli strain HVC45 capLS8. Bacterial extracts were prepain the relevant reaction mixtures (tolas described in the text. The abbrevif stand, respectively, for levan, sucifructose. Tracks 1 to 5, assays of tltion with extracts originating froiHVC45; lane 2 strain HVC45(pBRSstrain HVC45(pLS8). To determinereaction, samples were taken after.of incubation. Tracks 6 and 7, ahydrolysis and levan synthesis catalextract of strain HVC45(pLS8). A 4two tracks shows that the accumulaity in the levan spot is correlated wittivity of the fructose spot.

note the ease with which cloniphages allows a genetic screeinserting B. subtilis DNA, sinceforming activity of either the pu

chimera or the DNA present in cbe used to transform relevant althis transformation does not resiphenotype.The presence of the sacB ge

cloned in A LSC66 and cloned fupLS8 was identified on the basiria: (i) the transforming activitDNA tested on various mutantstion between the physical mapthe amino acid sequence of lei(iii) the expression of the sacB I

All the sacB and sacR allel

found to be transformed by X LSC66 DNA. Thesubcloning experiments allowed us to localize inthe pLS1 insert the wild-type allele of thesacB204 mutation, which is responsible for thesynthesis of modified enzyme (31), whereas thewild-type alleles of sacR mutations were found

: f in the pLS2 insert. Furthermore, the DNA ofplasmid pLS5 was found to transform all thesacB and sacR alleles tested. The localization ofsacR was not as precise as sacB since, up untilnow, few restriction sites convenient for cloningwere found in this area. This localization doesnot bring much more information concerningsacR function than is already available. It isconsistent with an operator role, as suggested byLepesant et al. (30), but does not exclude anoth-er function. It is known that transformation byDNA of plasmids that do not replicate in B.subtilis involves mostly integration of the com-plete plasmid sequence in the chromosome (18,22). Extensive studies of such integrations of LS

* le plasmids bearing a resistance marker which isexpressed in B. subtilis are in progress. They

6 7 should provide more information concerning theansucrase synthe- sacR region and its function.trrying the plasmid On the basis of (i) the genetic localization, (ii)red and incubated the expression of sacB in pLS8 and the lack oftal volume, 0.1 ml) expression of the gene in pLS1 and pLS5, andiations le, s, g, and (iii) the known size of the protein (2), the posi-rose, glucose, and tion of the sacB gene could be determined withinhe exchange reac- a 1,600-base pair segment, which included them: lane 1, strain HindIII-EcoRI segment shared by pLS1 and325); lanes 3 to 5, pLS5 (see Fig. 1). On the basis of the amino acid20, 40, and 60 min sequence of the secreted enzyme, the position of2ssays of sucrose the gene was exactly determined on the physicallyzed by the same map of the cloned DNA. The occurrence of acomparison of the signal peptide (5, 16) involved in the secretion ofttion of radioactiv- levansucrase is still questioned. Evidence forth the low radioac- such amino-terminus extensions in proteins se-

creted by Bacillus sp. was obtained by Neuge-bauer et al. (33) and Palva et al. (34), whoelucidated the nucleotide sequences of the genescoding, respectively, for the penicillinase of

ing in X charon Bacillus licheniformis and the a-amylase of Ba-mning of clones cillus amyloliquefaciens. Although some obser-- the high trans- vations by Petit-Glatron and Chambert (35) sug-rified DNA of A gest that the secretion of levansucrase mayrude lysates can involve a membrane-bound form of the enzyme,lleles even when the very small amount of material detected didult in a selective not allow structural analysis. Progress in se-

quencing the sacB gene will provide more pre-ne in the insert cise data about this problem.irther in plasmid Levansucrase synthesis was characterized inis of three crite- the E. coli hosts of either A LSC66 or pLS7 andy of the cloned pLS8. The specific activity of the enzyme in the(ii) the correla- crude extract of HVC45(pLS8) strain was ap-

of the DNA and proximately 0.8 U per mg of protein. This corre-vansucrase; and sponds to a rate of synthesis of approximatelygene in E. coli. 10-11 U of enzyme per cell per min. This rateles tested were obviously depends upon the number of sacB

VOL. 153, 1983

on April 5, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

1430 GAY ET AL.

gene copies in the cells. pLS8 was derived frompBR325, and it has a ColEl mode of replication.Clewell and Helinski (13) found that plasmidColEl was present in E. coli to the extent ofapproximately 24 copies per cell. From the yieldof pLS8 DNA extractions without spectinomy-cin amplification, we may roughly estimate forthis plasmid a number of copies of the sameorder of magnitude. In B. subtilis, the levansu-crase rate of synthesis measured in mineralmedium varies from 10-12 U per cell per min inthe reference strain 168M to 1.5 x 100 U in thehigh producer QB112, which differs from 168Mby the presence of the sacU32 mutation un-linked to sacB (26, 36). Therefore, the rate ofsynthesis of levansucrase in E. coli related to thegene copy number may be compared to the ratemeasured in the 168M strain of B. subtilis.

It is not unlikely that levansucrase synthesisin E. coli depends upon the sacB promotor itself.One may assume that the 2,000 base pairs thatprecede the origin of the gene in pLS8 includesthe promotor. The orientation of sacB in pLS8,opposite to the tet gene orientation, is consistentwith this assumption. In the case of the expres-sion of the cloned penP gene ofB. licheniformis,Brammar et al. (9) and Imanaka et al. (25)discussed the occurrence of Bacillus promotorrecognition by E. coli RNA polymerase andconcluded that the transcription of penP waspromoted by its own promotor. Moreover, it isknown that some Bacillus promotors have sig-nificant homologies with E. coli promotors (28).

Preliminary experiments did not allow us toconclude clearly whether E. coli secretes levan-sucrase in the periplasmic space. Osmoticshocks carried out with various E. coli hosts ofplasmid pLS8 showed that levansucrase waspartially liberated in the shock fluid. However,in some cases we found that some ,3-galacto-sidase, supposedly a strictly intracellular en-zyme and thus chosen as a control, was alsopresent in the shock fluid. A similar phenome-non was observed in some hosts of a cloned a-amylase gene (K. Willemot and M. Shwartz,personal communication). This suggests that thesynthesis of relatively large amounts of a foreignsecreted protein results in fragility of the E. colienvelopes. How such fragility may be related tothe secretion process is to be investigated.

ACKNOWLEDGMENTS

This work was supported by grants from Centre National dela Recherche Scientifique. We are indebted to A. Delfour forthe communication of the amino acid sequence of levansu-crase and to A. Nguyen and F. Ferrari for their kind help inDNA work.

Biohazards associated with the experiments described inthis publication have been examined previously by the FrenchNational Committee.

LITERATURE CITED

1. A o , C., and J. Spizzen. 1961. Requirementsfor transformation in Bacillus subtilis. J. Bacteriol.81:741-746.

2. Berthou, J., A. Laurent, E. LeBrun, and R. Van Rappen-busch. 1974. Cristalography of Bacillus subtilis levansu-crase. J. Mol. Biol. 82:111-113.

3. BIrnboIm, H. C., and J. Doly. 1979. A rapid alkalineprocedure for screening recombinant plasmid DNA. Nu-cleic Acids Res. 7:1513-1523.

4. Blattner, F. R., B. G. WillIams, A. E. Blechl, K. D.Thompon, H. E. Faber, L. A. Furlog, C. J. Grunwald,D. 0. Kiefer, D. D. Moore, J. N. Schum, E. L. Sheldon,and 0. Smithies. 1977. Charon phages, safer derivatives ofbacteriophage lambda for DNA cloning. Science 196:161-169.

5. Blobel, G., and B. Dobberstein. 1975. Transfer of proteinacross membranes: presence of proteolytically processedand unprocessed nascent immunoglobin light chains onmembrane bound ribosomes of murine myeloma. J. CellBiol. 67:835-851.

6. Bollvar, F. 1978. Construction and characterization ofnew cloning vehicles. III. Derivatives of plasmid pBR322carrying unique EcoRI sites for selection of EcoRI gener-ated recombinant DNA molecules. Gene 4:121-136.

7. Bollvar, F., R. L. Rodriguez, M. C. Betlach, and H. W.Boyer. 1977. Construction and characterization of newcloning vehicles. I. Ampicillin resistant derivatives of theplasmid pMB9. Gene 2:75-93.

8. Borenstein, S., and E. Ephrati-Elizur. 1969. Spontaneousrelease of DNA in sequential genetic order by Bacillussubtilis. J. Mol. Biol. 45:137-152.

9. Brammar, W. J., S. Muir, and A. McMorris. 1980. Molec-ular cloning of the gene for the 3-lactamase of B. licheni-formis and its expression in E. coli. Mol. Gen. Genet.178:217-224.

10. Chambert, R., and G. Gonzy-Treboul. 1976. Levansu-crase of Bacillus subtilis: kinetic and thermodynamicaspects of transfructosylation process. Eur. J. Biochem.62:55-64.

11. Chambert, R., and G. Gonzy-Treboul. 1976. Levansu-crase of Bacillus subtilis: characterization of a stabilizedfructosyl-enzyme complex and identification of an aspar-tyl residue as the bindings site of the fructosyl group. Eur.J. Biochem. 71:493-508.

12. Chambert, R., G. Gonzy-Treboul, and R. Dedonder. 1974.Kinetic studies of Bacillus subtilis levansucrase. Eur. J.Biochem. 41:285-300.

13. Cleweil, D. B., and D. R. Helinki. 1972. Effect of growthconditions on the formation of the relaxation complex ofsupercoiled ColEl deoxyribonucleic acid and protein inEscherichia coli. J. Bacteriol. 110:1135-1146.

14. Colnan, A., M. J. Byers, S. B. Primrose, and A. Lyons.1978. Rapid purification of plasmid DNA by hydroxyapa-tite chromatography. Eur. J. Biochem. 91:303-310.

15. Dagert, M., and S. D. Ehrilch. 1979. Prolonged incubationin calcium chloride improves the competence of theEscherichia coli cells. Gene 6:23-28.

16. Davs, D. B., and P. C. Tal. 1980. The mechanism ofprotein secretion across membranes. Nature (London)283:433-438.

17. Dedonder, R. 1966. Levansucrase from Bacillus subtilis.Methods Enzymol. 8:500-505.

18. Duncan, C. H., G. A. Wilson, and F. E. Young. 1978.Mechanism of integrating foreign DNA during transfor-mation of Bacillus subtilis. Proc. Natl. Acad. Sci. U.S.A.75:3665-3668.

19. Ferrari, E., D. J. Henner, and J. A. Hoch. 1981. Isolationof Bacillus subtilis genes from a charon 4A library. J.Bacteriol. 146:430-432.

20. Gonzy-Treboul, G., R. Chanbert, and R. Dedonder. 1975.Levansucrase of Bacillus subtilis: reexamination of somephysical and chemical properties. Biochimie 57:17-28.

21. Grant, G. F., and M. I. Simon. 1969. Synthesis of bacterial

J. BACTERIOL.

on April 5, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

CLONING OF THE sacB GENE OF B. SUBTILIS 1431

flagella. II. PBS1 transduction of flagella-specific markersin Bacillus subtilis. J. Bacteriol. 99:116-124.

22. Haldenwang, W. G., C. D. B. Banner, J. F. Ollington,R. Losick, J. A. Hoch, M. B. O'Connor, and A. L. Sonen-sheim. 1981. Mapping a gene under sporulation control bythe insertion of a drug resistance marker into the Bacillussubtilis chromosome. J. Bacteriol. 142:90-98.

23. Henner, D. J., and J. A. Hoch. 1980. The Bacillus subtilischromosome. Microbiol. Rev. 44:57-82.

24. Humphreys, G. O., G. A. Willshaw, and E. S. Anderson.1975. A simple method for the preparation of large quanti-ties of pure plasmid DNA. Biochem. Biophys. Acta383:457-463.

25. Imanaka, T., T. Tanaka, H. Tsunekawa, and S. Aiba.1981. Cloning of the genes for penicillinase, penP andpenI, of Bacillus licheniformis in some vector plasmidsand their expression in Escherichia coli, Bacillus subtilis,and Bacillus licheniformis. J. Bacteriol. 147:776-786.

26. Kunst, F., M. Pascal, J. Lepesant-Kejzlarova, J. A. Lepe-sant, A. Billault, and D. Dedonder. 1974. Pleiotropicmutations affecting sporulation conditions and the synthe-ses of extracellular enzymes in Bacillus subtilis 168.Biochimie 56:1481-1489.

27. Le Brun, E., and R. Van Rappenbusch. 1980. The struc-ture of Bacillus subtilis levansucrase at 3.8 A resolution.J. Biol. Chem. 255:12034-12036.

28. Lee, G., C. Talkington, and J. Pero. 1980. Nucleotidesequence of promotor recognized by Bacillus subtilisRNA polymerase. Mol. Gen. Genet. 180:57-65.

29. Lepesant, J. A., F. Kunst, J. Lepesant-Kejzlarova, andR. Dedonder. 1972. Chromosomal location of mutationsaffecting sucrose metabolism in Bacillus subtilis Marburg.Mol. Gen. Genet. 118:135-160.

30. Lepesant, J. A., F. Kunst, M. Pascal, J. Lepesant-Kejzlar-

ova, M. Steimetz, and R. Dedonder. 1976. Specific andpleiotropic regulatory mechanisms in the sucrose systemof Bacillus subtilis 168, p. 58-69. In D. Schlessinger (ed.),Microbiology-1976. American Society for Microbiology,Washington, D.C.

31. Lepesant, J. A., J. Lepesant-Kejzlarova, M. Pascal,A. Billault, and R. Dedonder. 1974. Identification of thestructural gene of levansucrase in Bacillus subtilis Mar-burg. Mol. Gen. Genet. 128:213-221.

32. Marmur, J. 1961. A procedure for the isolation of deoxyri-bonucleic acid from microorganisms. J. Mol. Biol. 3:208-218.

33. Neugebauer, K., R. Sprengel, and H. Schaller. 1981.Penicillinase from Bacillus licheniformis: nucleotide se-quence of the gene implication for the biosynthesis of asecretory protein in Gram-positive bacterium. NucleicAcids Res. 9:2577-2588.

34. Palva, I., R. F. Peterson, N. Kalkinen, P. Lehtovaara,M. Sarvas, H. Soderlund, C. Takkinen, and L. Kiiriiinen.1981. Nucleotide sequence of the promotor and NH,-terminal signal peptide region of the a-amylase gene fromBacillus amyloliquefaciens. Gene 15:43-51.

35. Petit-Glatron, M. F., and R. Chambert. 1980. Levansu-crase of Bacillus subtilis: characterization of a formisolated from phenol treated cells and activated by TritonX100. Eur. J. Biochem. 103:189-195.

36. Petit-Glatron, M. F., and R. Chambert. 1981. Levansu-crase of Bacillus subtilis: conclusive evidence that itsproduction and export are unrelated to fatty-acid synthe-sis but modulated by membrane modifying agents. Eur. J.Biochem. 119:603-611.

37. Primrose, S. B., and S. D. Ehrlich. 1981. Isolation ofplasmid deletion mutants and study of their instability.Plasmid 6:193-201.

VOL. 153, 1983

on April 5, 2018 by guest

http://jb.asm.org/

Dow

nloaded from