Vol. 55, No. 11

Inducible Secretion of a Cellulase from Clostridium thermocellumin Bacillus subtilis

GWENNAEL JOLIFF, ALEX EDELMAN, ANDRE KLIER, AND GEORGES RAPOPORT*

Unite de Biochimie Microbienne, Departement des Biotechnologies, URA 1300 CNRS, Institut Pasteur,75724 Paris Cedex 15, France

Received 16 March 1989/Accepted 18 July 1989

A host-vector system for inducible secretion during the logarithmic growth phase in Bacillus sublilis has beendeveloped. The B. subtilis levansucrase gene promoter and the region encoding its signal sequence have beenused. The endoglucanase A of Clostridium thermocelum was used as a model protein to test the efficiency of thesystem. Effective inducible secretion of the endoglucanase A was observed when either the levansucrase signalsequence or its own signal sequence was used. Expression of the endoglucanase A in different geneticbackgrounds of B. subtilis showed that its regulation was similar to that of levansucrase, and high enzyme

activity was recovered from the culture supernatant of a hyperproducing B. subtilis sacU(Hy) strain. Themolecular weight of 46,000 estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis for thesecreted endoglucanase A is compatible with the calculated molecular weight of the mature polypeptide.

Various heterologous proteins have been secreted in Ba-cillus subtilis, and most of the secretion vectors developeduse the a-amylase secretory system (27, 28, 30), although,more recently, protease export systems have been used (13,14, 28, 35, 37). However, expression from these systemsoccurs during stationary phase, when most of the secretedenzymes of B. subtilis, including the major proteases, are

secreted. Here, we describe a secretion vector based on theB. subtilis levansucrase (Lvs) system, which is active duringthe logarithmic growth phase. Lvs is encoded by the sacBgene and expressed from a strong sucrose-inducible pro-

moter located in the closely linked sacR locus. The controlof its expression has been studied extensively (12, 33), andseveral regulatory up mutants are available: sacQ(Hy),sacS(Hy), and sacU(Hy). The sacQ(Hy), sacS(Hy), andsacU(Hy) mutants produce high levels of Lvs (19, 20). ThesacQ, sacS, and sacU gene products are transcriptionalregulators of the expression of sacB. The sacQ and sacUgenes act on a region located near or just upstream from thepromoter, and the sacS gene, which is involved in theinduction process, acts on the palindromic structure locatedbetween the promoter and the ribosome binding site (3, 16).These factors make it a suitable candidate for the construc-tion of secretion systems for heterologous proteins. Thissystem has already been used to overproduce an intracellu-lar protein, the cathechol 2,3-dioxygenase, encoded by thexylE gene of Pseudomonas putida (38). Fusions of the Lvsgene of B. subtilis and the a-amylase gene of Bacilluslicheniformis containing various lengths of the mature partsof both proteins have also been described (18). Furthermore,we recently reported the secretion of the TEM P-lactamaseof Escherichia coli using the sacB system (11).The extracellular endoglucanase A (EGA; 1,4-p-D-glucan

glucanohydrolase [EC 3.2.1.4]) of the thermophilic anaero-

bic bacterium Clostridium thermocellum encoded by thecelA gene has been used as a model protein to test theefficiency of the sacB secretion system in B. subtilis. Thisprotein was chosen for its convenient size (22), its successfulexpression in various microorganisms (10, 25, 32), and itsenzymatic activity, which is easily detectable directly on

* Corresponding author.

plates (10). Furthermore, EGA has been characterized andpurified (22, 31), and the DNA sequence (5) of its gene andthe sites of transcription initiation (7) are known. Thecalculated molecular weight of mature EGA is 49,125, but onsodium dodecyl sulfate-polyacrylamide gels of C. thermocel-lum culture supernatants, it appears as a 56,000-molecular-weight band believed to be due to glycosylation (5). Thesefactors facilitate the use of this protein as a model forheterologous secretion by B. subtilis.

Here, we show that B. subtilis can secrete enzymaticallyactive EGA into the culture medium after transformationwith hybrid plasmids containing precise fusions, after induc-tion by sucrose. We also compare the efficiency of the Lvsand EGA signal sequences when fused to the structural geneof EGA.

MATERIALS AND METHODS

Strains, bacteriophage, plasmids, and media. The strainsand plasmids used are listed in Table 1. The phageM13mpl8-am4 was used for the oligonucleotide site-directedmutagenesis. For growth of E. coli, Luria broth (21) was

used and supplemented with ampicillin (100 ,ug/ml), chlor-amphenicol (10 pug/ml), kanamycin (50 ,ug/ml), or tetracy-cline (10 ,ug/ml). For growth of B. subtilis, sporulation broth(29) was used and supplemented with chloramphenicol (5,ug/ml) or kanamycin (5 pxg/ml) for selection and 2% (wt/vol)sucrose for induction of the sacB promoter. E. coli was

transformed by the method of Cohen et al. (9), and B. subtiliswas transformed by the method of Anagnostopoulos andSpizizen (2).Enzyme assays. Carboxymethyl cellulose (CMC) (no. C-

4888, medium viscosity; Sigma Chemical Co.) was used as

the substrate for the EGA. Endoglucanase activity was

detected by the Congo red assay on plates by the method ofCornet et al. (10). All enzyme assays were performed in PCbuffer (50 mM K2HPO4, 12 mM citric acid, pH 6.3) at 60°C.Carboxymethyl cellulase (CMC-ase) activity was measuredby the method of Beguin et al. (6). One unit of activity wasdefined as the amount of enzyme that released 1 ,mol ofglucose equivalent per min. Cellulase activity was detectedin polyacrylamide gels by using Congo red-stained agarreplicas by the method of Bdguin (4).

2739

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Nov. 1989, p. 2739-27440099-2240/89/112739-06$02.00/0Copyright © 1989, American Society for Microbiology

on June 14, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

TABLE 1. Bacterial strains and plasmids

Strain or Genotype Relevant Referenceplasmid phenotype'

B. subtilis168 trpC2 Lvs+ Lepesant, thesisQB25 trpC2 sacS(Con)49 Lvsc This laboratoryQB136 leuA8 trpC2 sacU(Hy)32 Lvsh This laboratoryQB907 trpC2 sacA321 sacS(Hy)3 LvSh This laboratoryQB671 hisAl leuA8 thr-5 sacS(Hy)3 Lvsh This laboratory

E. coliTG1 K-12 A(lac-pro) supE thi hsdD5/F' traD36 rK- mK- repair' T. J. Gibson, Ph.D. thesis, University

proA+B+ lacIq lacZAM15 of Cambridge, 1984HB2154 K-12 ara A(lac-pro) thilF' proA+B+ lacIq rK+ mK+ repair- 8

lacZAM15 mutL::TnlOPlasmidspUC8 Apr 36pC194 Cmr 15pCT105 celA Apr Tcr 10pAE101 sacR AsacB Kmr Apr 11pAE119 AsacRB aphA3 Kmr This laboratorypAT21 aphA3 Kmr 34pBS413 sacR sacB Apr 1pBS430 sacR AsacB Apr Cmr 16

a Lvs, Extracellular levansucrase; Lvs+, wild-type production; Lvsc, constitutive production; Lvsh, hyperproduction. Apr, Cmr, Tcr, and Kmr denoteresistance to ampicillin, chloramphenicol, tetracycline, and kanamycin, respectively.

Oligonucleotide site-directed mutagenesis. The procedureused for oligonucleotide site-directed mutagenesis was es-sentially that described by Carter et al. (8). pGJ3 (see Fig. 4)was used to produce a single-stranded template which wasprimed with both the mutagenic primer DGF3 (5' AAAAGGCACACCTGCGGCAAACGCTTGAGT 3') and a selectionprimer, SELl (5' AAGGAGTCTGTCCATCAC 3') to re-move the selectable marker from the minus strand and wastransfected into HB2154 cells, which are repair deficient andselect against the marker; therefore, a selection for progenyphage derived from replication of the minus strand wasobtained. The 15 5'-nucleotides of the mutagenic primerDGF3 are complementary to the first 15 nucleotides of theEGA mature sequence, and the 15 3'-nucleotides are com-plementary to the last 15 nucleotides of the Lvs signalsequence contained in pAE101 (11). Southern blotting wasperformed by the method of Carter et al. (8) after a wash in6x SSC (lx SSC is 0.15 M NaCl plus 0.015 M sodiumcitrate) at room temperature and at 45°C, 55°C, and 65°C,successively; the mutants containing the in-frame fusionbetween the Lvs signal sequence and the EGA maturesequence hybridized strongly to DGF3 after the last wash.The mutation corresponding to a deletion of 19 nucleotidesof the EGA signal sequence was verified by dideoxy se-quencing (26) by using the synthetic oligoprimer DGF4 (5'TCGGCAATAGAAGTAGG 3'), which hybridizes down-stream of the deleted region.

Construction of pGJl, pGJ2, and pAE119. Plasmid pGJ1was constructed by a three-way ligation of (i) 50 ng of thevector pUC8 cut with BamHI and HindIII, (ii) 25 ng of the440-base-pair (bp) BamHI-RsaI fragment of the sacR regioncontaining the sacB promoter and its regulation region frompBS430 (3, 16), and (iii) 100 ng of the 2.2-kbp DraI-HindIIIfragment of the plasmid pCT105 (10) containing the putativeribosome binding site, the region encoding EGA (signalsequence and mature sequence) (5), and the putative termi-nation site of the gene (7) (Fig. 1). The 440-bp and 2.2-kbpfragments were isolated from a 1.5% agarose gel and werepurified by electroelution from the agarose gel. The desired

construct, pGJ1, was isolated as an Apr transformant of E.coli exhibiting CMC-ase activity (CMC+ phenotype) in theCongo red plate assay (10). To test the secretion of EGA inB. subtilis, pC194 (15), a plasmid of 2.91 kbp containing a B.subtilis functional origin of replication and an antibioticresistance marker (chloramphenicol), was ligated to pGJ1 atthe HindlIl site, resulting in plasmid pGJ2.The plasmid pAE119 was used for the deletion of the

sacRB region in B. subtilis, by recombination. pBS413 (1)contains the sacRB region on contiguous HindIII fragments.The plasmid was linearized with ClaI, which cuts in the sacBgene, and was treated with BAL 31 exonuclease to delete theentire sacRB region. After filling of ragged ends, the frag-ment was ligated to the aphA3 gene from Streptococcusfaecalis to result in pAE119, which is not replicative in B.subtilis. The aphA3 gene used is included in a 1.5-kbp ClaI

pUC8 digested with

DrI Nae I Hind II BamHl and HindML(Ava D celA AvP 1 AvpI

2.09 kbpN*

B=TR4 2.24 kbp 440bp

ligae, ufotm E&&li seect Ap

Ap' 0

4t -i.ni9v5Ia4ZkHbp 4I

AvaH~ ~ ~ AviBarn HI t

Ava

Ava I

FIG. 1. Construction of PGJ1 (see Materials and Methods fordetails). Solid bars represent DNA fragments containing the sacRregion or the celA gene. Open bars represent regions from pCT105flanking the celA gene. Arrows indicate the direction of transcriptionof the genes.

2740 JOLIFF ET AL.

on June 14, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

SECRETION OF A CELLULASE IN B. SUBTILIS 2741

. 1.0

0.9 'j.0.8

.0.7

.0.6 <

.0.53

0.4

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

TIME ( h)FIG. 2. Expression of celA gene from plasmid pGJ2 in different

B. subtilis strains after induction by sucrose. 0, Bacterial growth ofthe wild-type strain. The mutants strains gave the same growthcurve. The symbols correspond to CMC-ase activity per milliliter ofsupernatant where the host strain was: 0, sacU(Hy); A, sacS(Hy);O, sacS(Con); *, wild type. 0, CMC-ase activity per milliliter ofsupernatant in a noninduced sacU(Hy) strain.

fragment from pAT21 (34). After transformation of B. sub-tilis with pAE119, recombinants were Kmr and Lvs-, sug-gesting a deletion in the sacRB region by a double-crossoverevent.

RESULTS

Expression of the celA gene in B. subtilis under control of thesacR region. A construct for the expression of EGA wasmade by a three-way ligation, as mentioned in Materials andMethods. The resulting plasmid, pGJ2, was introduced intodifferent B. subtilis strains: 168 (wild type), sacS(Con),sacS(Hy), and sacU(Hy) (Table 1). Cmr transformants of B.subtilis carrying the C. thermocellum celA gene on themulticopy plasmid pC194 were identified on the basis ofenlarged zones of CMC hydrolysis when the bacteria wereinduced by sucrose. However, the halo size varied, depend-ing on the host strain. Endoglucanase assays of liquid culturesupernatants demonstrated substantial sucrose-inducible se-cretion of EGA (Fig. 2) (Table 2). These results suggestedthat the regulation of the expression of the celA gene was thesame as that of the sacB gene. No expression of the celAgene was detected in the absence of sucrose in the induciblestrains. The highest level of EGA secretion was obtainedwith the sacU(Hy) mutant, which also hypersecreted lev-ansucrase, producing about 100-fold more of it than wild-type cells produced. EGA secretion by this strain was onlyabout eightfold higher than that by the induced transformed

TABLE 2. CMC-ase activity in culture supernatants of differenttransformed B. subtilis strains

Relevant Induction CMC-aseStrain phenotype Plasmid by sucrose (U/mi)

168 Wild type pC194 + <0.01pGJ2 - <0.01pGJ2 + 0.10

QB25 SacS(Con) pC194 + <0.01pGJ2 - 0.11pGJ2 + 0.38

QB907 SacS(Hy) pC194 + 0.01pGJ2 - 0.01pGJ2 + 0.56

QB136 SacU(Hy) pC194 + 0.02pGJ2 - 0.02pGJ2 + 0.80

a The activity was measured in culture supernatants recovered at late logphase.

wild type, even though the gene was expressed from amulticopy plasmid. However, the expression level in asacU(Hy) strain of sacB on a multicopy plasmid is only 5- to10-fold higher than the level obtained in a sacU+ strain(unpublished results). This may indicate limitation factors inthe expression or secretion mechanisms. The expression ofcelA in a sacS(Con) mutant was hyperinducible, as it hasalready been observed with the expression of sacB in thesame background (J.-A. Lepesant, Ph.D. thesis, UniversitdParis VII, 1974). B. subtilis possesses an endogenous ,-1,4-endoglucanase activity (24). However, in strain B. sub-tilis sacU(Hy) (pGJ2), which presented a higher endogenousactivity than the wild-type strain, this accounts for onlyabout 2% of CMC-ase activity.

Endoglucanase assays showed that cell extracts of strainsacU(Hy) (pGJ2) contained about 20% of the total EGAactivity, indicating that most of the EGA (80%) was secretedin the medium (data not shown).

Secretion of EGA by the levansucrase signal sequence andregulation region sacR. The DNA region encoding the signalsequence of levansucrase and carrying sacR was fused to theregion encoding mature EGA by oligonucleotide site-di-rected mutagenesis in an M13 vector (8) (Fig. 3). The vectorM13mpl8-am4 (8), which has an amber mutation in anessential phage gene, was cut with BamHI and Hindlll and

DGF3 (30 mer)pGJ3 (strand+)

3' TGAGTTCGCAAACGG CGTCCACACGGAAAA 5'5'-GCGCAACTCAAGCGTTTGCC GCAGGTGTGCCTTTTAACAC-3'-----signal sequence 4+-mature sequence-----

Lvs G G EGAhaIM G C

.CGA AA CA TC GA TCTG

t19 nucleotides from thesignal sequence of EGA

FIG. 3. Fusion of the region encoding the mature part of EGA tothe region encoding the signal sequence of Lvs by site-directedmutagenesis. DGF3 is the mutagenic primer. The sequence deletedfrom pGJ3 (Fig. 4) is shown as a hairpin.

VOL. 55, 1989

on June 14, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

Nae I1,Ava 11 ccl A Av,a II

Hind IIIAva 11 I

M13mpl8-am4 digested withI mHI and HiallI

BamHI_ Nael_ 550 bp

lII

DGF 3

SELEI S

FIG. 4. Construction of pGJ3, pGJ4, and pGJ5 (see Results for details). The symbols used are as described in the legend to Fig. 1. DGF3corresponds to the mutagenic primer, and SELl corresponds to the selection primer. A, Sites of hybridization for the two primers.

ligated to (i) a BamHI-NaeI fragment of 0.55 kbp from theplasmid pAE101 (11) carrying sacR and the levansucrasesignal sequence, and to (ii) a NaeI-HindIII fragment of 2.14kbp from the plasmid pCT105 (10), containing the DNAencoding the mature EGA and 19 nucleotides from the signalsequence of EGA (Fig. 4). The resulting plasmid was calledpGJ3. The 19 nucleotides belonging to the EGA signalsequence were deleted from pGJ3, and constructs containingthe desired sequence were named pGJ4 (Fig. 4). However,the plasmids contained an insertion of only about 0.9 kbp inM13mp18, instead of the expected 2.7 kbp, corresponding toa deletion of about 1.8 kbp in the 3' region of the insert. Thiswas verified by sequencing with the universal primer. Thecomplete gene fusion was reconstructed. Three fragments,(i) the 0.58-kbp fragment BamHI-AvaII from pGJ4, contain-ing the in-frame fusion between the signal sequence of Lvsand the mature sequence of EGA; (ii) the 2.09-kbp partialAvaII-HindIII fragment from pCT105 (Fig. 1); and (iii) thevector pUC8, cut with BamHI-HindIII, were ligated to-gether. The resulting plasmid, containing the sacB-celAtranslational fusion, was named pGJ5 (Fig. 4).To test the secretion in B. subtilis of EGA resulting from

this fusion, the plasmid pC194 was ligated at the HindlIl siteof pGJ5. The resulting plasmid, pGJ6, was introduced into B.subtilis sacU(Hy) (QB136). All of the Cmr transformantstested possessed a sucrose-inducible CMC+ phenotype onthe basis of the Congo red plate assay. The supernatant ofone clone was assayed for its activity on CMC after induc-tion by sucrose, and the level of activity in the supernatantwas the same as that of QB136(pGJ2). This result seemed toindicate that the signal sequences of Lvs and EGA areequally efficient for secretion in B. subtilis.

Molecular weight and stability of the EGA secreted by B.

subtilis. Analysis of the supernatant of sucrose-induced B.subtilis QB136 harboring pGJ6 by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed a protein bandwith an estimated molecular weight of 46,000 which was notdetected in the controls (Fig. 5). The same profile wasobtained for QB136(pGJ2). This band was shown to possessCMC-ase activity by replication onto agarose gels containingCMC (4) (data not shown). The molecular weight of thisband is compatible with the calculated molecular weight ofmature EGA, 49,125, as deduced from the nucleotide se-quence (5). Even in a sacU(Hy) mutant, which is known tooverproduce proteases (17), no loss of CMC-ase activity wasobserved during an overnight incubation of cells grown tostationary phase, indicating that EGA was resistant to pro-teolytic degradation. These observations indicate that EGAis secreted into the culture medium in a complete and stableform. Strains QB136 and QB136(pGJ6) were induced withsucrose, and supernatants were assayed for Lvs activity.The untransformed strain produced three times more Lvsactivity than did the strain carrying pGJ6. This result was inagreement with densitometer tracings of Coomassie blue-stained polyacrylamide gels (Fig. 5) and suggests competi-tion between the sacB gene and the heterologous sacBlcelAgene for expression or secretion (or both).

Expression of the sacB-ceA fusion in a sacRB-deleted back-ground. To test whether sacB expression from the chromo-somal gene interferes with celA over-expression, the sacRBregion was deleted from the chromosome of host strains. Toachieve this, host strains were transformed with pAE119(see Materials and Methods) and Kmr transformants wereselected. In more than 98% of transformants, the sacRBregion was replaced by aphA3, as shown by the absence ofLvs activity (data not presented). Plasmid pGJ6 was intro-

i i

2742 JOLIFF ET AL.

on June 14, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

SECRETION OF A CELLULASE IN B. SUBTILIS 2743

0--.IC0

Rf

FIG. 5. Sodium dodecyl sulfate-polyacrylamide gel electropho-resis. Aliquots (25 Rl) of the supernatants from B. subtilis QB136late-log-phase cultures, concentrated 25 times on a Amicon mem-

brane, were applied to a 10% polyacrylamide gel. The figure showsdensitometer tracings of Coomassie blue-stained gels: (a) QB136; (b)QB136(pGJ6); (c) molecular size reference markers: phosphorylaseb (94 kilodaltons [kDa]), bovine serum albumin (67 kDa), levansu-crase (50 kDa), ovalbumin (43 kDa), carbonic anhydrase (30 kDa).

duced into B. subtilis QB671 in which the sacRB region hadbeen deleted as described above. QB671 is isogenic withQB907 (Lepesant, thesis), except that it is sacA+ (sucroseinhibits growth in sacA sacB strains; data not shown).Strains QB671(pGJ6) and QB671 AsacRB (pGJ6) were grown

in broth and induced with sucrose. The CMC-ase activitiesof the supernatants were assayed and were indistinguishablefrom each other.

DISCUSSIONThe heterologous protein EGA of C. thermocellum was

expressed and secreted from B. subtilis by using the sacBpromoter after induction by sucrose. EGA is secreted duringthe logarithmic growth phase and accumulates in the me-

dium. The concentration of EGA is highest during thestationary phase, suggesting that the protein is secreted in a

stable form. Efficient secretion was obtained by using eitherthe signal sequence of Lvs or its own signal sequence.Analysis of these two signal sequences showed no significantdifferences from typical signal sequences from gram-positivebacteria. Expression of the enzyme in different geneticbackgrounds of B. subtilis [sacS(Con), sacS(Hy), andsacU(Hy)] showed that its regulation was as that of the sacBgene, and high enzyme activity was recovered from theculture fluid of a sacU(Hy) strain of B. subtilis. The size ofthe protein (-46 kilodaltons) is compatible with the molec-ular weight deduced from the sequence of the mature pro-tein.

The level of Lvs secretion decreases when the celA gene isexpressed from the sacB promoter on the multicopy vectorpC194, suggesting competition between the homologous andheterologous system in expression or secretion (or both).However, in a AsacRB strain the level of secreted EGA isnot increased. Presumably, therefore, the presence of asingle copy of the sacB gene on the chromosome does notcompete significantly with the expression of genes from thesacB promoter on a multicopy plasmid.The EGA activity in the supernatant of a sacU(Hy) strain

was 0.8 U/ml. Because the specific activity of purified EGAis 140 U/mg (23), this corresponds to a protein concentrationof approximately 1 mg/liter per OD660 unit. This valuecompares favorably with those obtained for EGA productionin C. thermocellum and E. coli (10).

Analyses of the limiting factors of the system are inprogress, with a view to improve the expression of heterol-ogous genes. Secretion of an eucaryotic gene will also betested in order to assess possible biotechnological applica-tions of this system.

ACKNOWLEDGMENTS

We thank R. Dedonder, in whose laboratory this work was carriedout. We are grateful to J.-P. Aubert for providing plasmid pCT105.We acknowledge P. Roux for the densitometric analysis and E.Cracy for typing the manuscript.

This work was supported by research funds from the Ministere dela Recherche et de la Technologie, Institut Pasteur and CentreNational de la Recherche Scientifique. G. Joliff and A. Edelmanwere supported by a grant from the "Contrat Couples H6tes-Vecteurs Performants."

LITERATURE CITED1. Amory, A., F. Kunst, E. Aubert, A. Klier, and G. Rapoport.

1987. Characterization of the sacQ genes from Bacillus licheni-formis and Bacillus subtilis. J. Bacteriol. 169:324-333.

2. Anagnostopoulos, C., and J. Spizizen. 1961. Requirements fortransformation in Bacillus subtilis. J. Bacteriol. 81:741-746.

3. Aymerich, S., G. Gonzy-Treboul, and M. Steinmetz. 1986. 5'-Noncoding region sacR is the target of all identified regulationaffecting the levansucrase gene in Bacillus subtilis. J. Bacteriol.166:993-998.

4. Beguin, P. 1983. Detection of cellulase activity in polyacryl-amide gels using Congo red stained agar replicas. Anal. Bio-chem. 131:333-336.

5. Beguin, P., P. Cornet, and J.-P. Aubert. 1985. Sequence of a

cellulase gene of the thermophilic bacterium Clostridium ther-mocellum. J. Bacteriol. 162:102-105.

6. Beguin, P., P. Cornet, and J. Millet. 1983. Identification of theendoglucanase encoded by the celB gene of Clostridium ther-mocellum. Biochimie 65:495-500.

7. Beguin, P., M. Rocancourt, M.-C. Chebrou, and J.-P. Aubert.1986. Mapping of mRNA encoding endoglucanase A fromClostridium thermocellum. Mol. Gen. Genet. 202:251-254.

8. Carter, P., H. Bedouelle, and G. Winter. 1985. Improved oligo-nucleotide site-directed mutagenesis using M13 vectors. Nu-cleic Acids Res. 13:4431 4443.

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

10. Cornet, P., J. Millet, P. Beguin, and J.-P. Aubert. 1983. Char-acterization of two cel (cellulose degradation) genes of Clostrid-ium thermocellum coding for endoglucanases. Bio/Technology1:589-594.

11. Edelman, A., G. Joliff, A. Klier, and G. Rapoport. 1988. Asystem for the inducible secretion of proteins from Bacillussubtilis during logarithmic growth. FEMS Microbiol. Lett.52:117-120.

12. Fouet, A., M. Arnaud, A. Klier, and G. Rapoport. 1984. Char-

MGA

I (b)o

go:0

'.5030*0a ,

bSS-

VOL. 55, 1989

on June 14, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

acterization of the precursor form of the exocellular levansu-crase from Bacillus subtilis. Biochem. Biophys. Res. Commun.119:795-800.

13. Honjo, M., A. Akaoka, A. Nakayama, H. Shimada, I. Mita, K.Kawamura, and Y. Furutani. 1986. Construction of secretionvector and secretion of hIFN-P, p. 89-100. In A. T. Ganesanand J. A. Hoch (ed.), Bacillus molecular genetics and biotech-nology applications. Academic Press, Inc., New York.

14. Honjo, M., A. Nakayama, H. Shimada, A. Lio, I. Mita, K.Kawamura, and Y. Furutani. 1988. Construction of an efficientsecretion host-vector system in Bacillus subtilis, p. 365-369. InA. T. Ganesan and J. A. Hoch (ed.), Genetics and biotechnol-ogy of bacilli. Academic Press, Inc., New York.

15. Horinouchi, S., and B. Weisblum. 1982. Nucleotide sequenceand functional map of pC194, a plasmid that specifies induciblechloramphenicol resistance. J. Bacteriol. 150:815-825.

16. Klier, A., A. Fouet, M. Debarbouille, F. Kunst, and G. Rapo-port. 1987. Distinct control sites located upstream from thelevansucrase gene of Bacillus subtilis. Mol. Microbiol. 1:233-241.

17. Kunst, F., M. Pascal, J. Lepesant-Kejzlarova, J.-A. Lepesant, A.Billault, and R. Dedonder. 1974. Pleiotropic mutations affectingsporulation conditions and the syntheses of extracellular en-zymes in Bacillus subtilis 168. Biochimie 56:1481-1489.

18. Le Coq, D., P. Joyet, B. Le Reverend, and H. Heslot. 1986.Expression of levansucrase/a-amylase gene fusions in Esche-richia coli and Bacillus subtilis, part B, p. 457-463. In M.Alacevic, D. Hranueli, and Z. Toman (ed.), 5th InternationalSymposium on Genetics of Industrial Microorganisms, Split,Yugoslavia. Pliva, Zagreb, Yugoslavia.

19. Lepesant, J.-A., F. Kunst, J. Lepesant-Kejzlarova, and R. De-donder. 1972. Chromosomal location of mutations affectingsucrose metabolism in Bacillus subtilis Marburg. Mol. Gen.Genet. 118:135-160.

20. Lepesant, J.-A., F. Kunst, M. Pascal, J. Lepesant-Kejzlarova, M.Steinmetz, and R. Dedonder. 1976. Specific and pleiotropicregulatory mechanisms in the sucrose system ofBacillus subtilis168, p. 58-69. In D. Schlessinger (ed.), Microbiology-1976.American Society for Microbiology, Washington, D.C.

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

22. Petre, J., R. Longin, and J. Millet. 1981. Purification andproperties of an endo-P-1,4-glucanase from Clostridium thermo-cellum. Biochimie 63:629-639.

23. Petre, D., J. Millet, R. Longin, P. Beguin, H. Girard, and J.-P.Aubert. 1986. Purification and properties of the endoglucanaseC of Clostridium thermocellum produced in Escherichia coli.Biochimie 68:687-695.

24. Robson, L. M., and G. H. Chambliss. 1984. Characterization ofthe cellulolytic activity of a Bacillus isolate. Appl. Environ.

Microbiol. 47:1039-1046.25. Sacco, M., J. Millet, and J.-P. Aubert. 1984. Cloning and

expression in Saccharomyces cerevisiae of a cellulase genefrom Clostridium thermocellum. Ann. Microbiol. Inst. Pasteur135A:485-488.

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

27. Sarvas, M. 1986. Protein secretion in Bacilli. Curr. Top. Micro-biol. Immunol. 125:103-125.

28. Saunders, C. W., B. J. Schmidt, R. L. Mallonee, and M. S.Guyer. 1987. Secretion of human serum albumin from Bacillussubtilis. J. Bacteriol. 169:2917-2925.

29. Schaeffer, P., H. lonesco, A. Ryter, and G. Balassa. 1965. Lasporulation de Bacillus subtilis, dtude gdnetique et physi-ologique, p. 553-563. In Mecanismes de regulation des activitescellulaires chez les microorganismes. Centre National de laRecherche Scientifique, Paris.

30. Schein, C. H., K. Kashiwagi, A. Fujisawa, and C. Weissmann.1986. Secretion of mature IFN-a2 and accumulation of un-cleaved precursor by Bacillus subtilis transformed with a hybridoa-amylase signal sequence-IFN-a2 gene. Bio/Technology 4:719-725.

31. Schwarz, W. H., F. Grabnitz, and W. L. Staudenbauer. 1986.Properties of a Clostridium thermocellum endoglucanase pro-duced in Escherichia coli. Appl. Environ. Microbiol. 51:1293-1299.

32. Soutschek-Bauer, E., and W. L. Staudenbauer. 1987. Synthesisand secretion of a heat-stable carboxymethylcellulase fromClostridium thermocellum in Bacillus subtilis and Bacillus stear-othermophilus. Mol. Gen. Genet. 208:537-541.

33. Steinmetz, M., D. Le Coq, S. Aymerich, G. Gonzy-Treboul, andP. Gay. 1985. The DNA sequence of the gene for the secretedBacillus subtilis enzyme levansucrase and its genetic controlsites. Mol. Gen. Genet. 200:220-228.

34. Trieu-Cuot, P., and P. Courvalin. 1983. Nucleotide sequence ofthe Streptococcus faecalis plasmid gene encoding the 3'5"-aminoglycoside phosphotransferase type III. Gene 23:331-341.

35. Vasantha, N., and L. D. Thompson. 1986. Fusion of pro regionof subtilisin to staphylococcal protein A and its secretion byBacillus subtilis. Gene 49:23-28.

36. Vieira, J., and J. Messing. 1982. The pUC plasmid and M13mp7-derived system for insertion mutagenesis and sequencing withsynthetic universal primers. Gene 19:259-268.

37. Wang, L.-F., S.-L. Wong, S.-G. Lee, N. K. Kalyan, P. P. Hung,S. Hilliker, and R. H. Doi. 1988. Expression and secretion ofhuman atrial natriuretic a-factor in Bacillus subtilis using thesubtilisin signal peptide. Gene 69:39-47.

38. Zukowski, M. M., and L. Miller. 1986. Hyperproduction of anintracellular heterologous protein in a sacUh mutant of Bacillussubtilis. Gene 46:247-255.

2744 JOLIFF ET AL.

on June 14, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from