JOURNAL OF BACTERIOLOGY,

JUlY 1991, p. 4107-4115

Vol. 173, No. 13

0021-9193/91/134107-09$02.00/0

A Highly Thermostable Neutral Protease from Bacillus caldolyticus: Cloning and Expression of the Gene in Bacillus subtilis and Characterization of the Gene Product BERTUS VAN DEN BURG,'* HANS G. ENEQUIST,2 MARJAN E. VAN DER HAAR,1 VINCENT G. H. EIJSINK,1 BEN K. STULP,1 AND GERARD VENEMA' Department of Genetics, Centre of Biological Sciences, Kerklaan 30, 9751 NN Haren,l and Department of Biochemistry, State University of Groningen, Nijenborgh 16, 9747 AG Groningen,' The Netherlands Received 25 March 1991/Accepted 24 April 1991

By using a gene library of Bacillus caldolyticus constructed in phage lambda EMBL12 and selecting for

proteolyticaily active phages on plates supplemented with 0.8% skim milk, chromosomal B. caldolyticus DNA fragments that specified proteolytic activity were obtained. Subcloning of one of these fragments in a protease-deficient Bacillus subtUis strain resulted in protease proficiency of the host. The nucleotide sequence of a 2-kb Hinfl-MluI fragment contained an open reading frame (ORF) that specified a protein of 544 amino acids. This ORF was denoted as the B. caldolyticus npr gene, because the nucleotide and amino acid sequences of the ORF were highly similar to that of the Bacillus stearothermophilus npr gene. Additionally, the size, pH optimum, and sensitivity to the specific Npr inhibitor phosphoramidon of the secreted enzyme indicated that the B. caldolyticus enzyme was a neutral protease. The B. stearothermophilus and B. caldolyticus enzymes differed at only three amino acid positions. Nevertheless, the thermostability and optimum temperature of the B. caldolyticus enzyme were 7 to 8°C higher than those of the B. stearothermophilus enzyme. In a three-dimensional model of the B. stearothermophilus Npr the three substitutions (Ala-4 to Thr, Thr-59 to Ala, and Thr-66 to Phe) were present at solvent-exposed positions. The role of these residues in thermostabiity was analyzed by using site-directed mutagenesis. It was shown that all three amino acid substitutions contributed to the observed difference in thermostability between the neutral proteases from B. stearothermophilus and B. caldolyticus.

Bacterial species belonging to the genus Bacillus secrete a variety of enzymes, some of which are of economical importance, in particular the ax-amylases and proteinases (5). The various Bacillus species show large variations in optimum growth temperatures, which are often, but not invariably, reflected in the thermostabilities of their extracellular enzymes (23). By comparing thermostable and thermolabile variants of homologous enzymes, information on the mechanisms involved in thermostability of enzymes can be obtained (22, 31, 37). We selected the neutral proteases from bacilli as a model system in such an approach. Neutral proteases (Npr) are metalloendoproteinases which show optimum activity at neutral pH. These enzymes contain one zinc atom per molecule and may be stabilized by calcium binding (17). Several npr genes have been cloned and expressed in Bacillus subtilis, including the genes from B. subtilis, Bacillus stearothermophilus and Bacillus amyloliquefaciens, and their nucleotide sequences have been determined (8, 12, 18, 33, 38, 40, 44). Recently, the npr gene from Bacillus brevis was also cloned and sequenced (la). Additionally, the primary and tertiary structures of the neutral proteases from Bacillus cereus and Bacillus thermoproteolyticus have been determined (21, 28, 34, 35). The neutral proteases from B. subtilis and B. amyloliquefaciens are rather thermolabile, whereas the enzymes from B. stearothermophilus and B. thermoproteolyticus show a considerably higher level of resistance towards thermal inactivation. The enzyme from B. cereus exhibits an intermediate thermostability (34). *

By comparison of the primary and tertiary structures of neutral proteases, several residues and structures which might be involved in thermostability in this class of enzymes have been identified. The role of some of these residues and structures was examined by means of site-directed mutagenesis (5a, 6). On the basis of the consideration that accurate predictions of residues involved in thermostability would benefit from the availability of an extended range of neutral proteases, we isolated and cloned the neutral protease gene from Bacillus caldolyticus, which is able to grow at temperatures as high as 85°C (11). This paper describes the isolation of the gene by using a technique which enabled the direct selection of DNA fragments encoding proteolytic enzymes. MATERIALS AND METHODS Bacterial strains, phages, plasmids, and media. The bacterial strains and plasmids used are listed in Table 1. B. caldolyticus was grown in TY broth (3) at 70°C. B. subtilis and Escherichia coli were grown in TY broth at 37°C. Phage was plated on Trypticase-peptone agarose prepared essentially as described by Frischauf et al. (7), with E. coli NM538 and NM539 as indicators. Skim milk (0.8%, wt/vol) was added to the plates to visualize protease activity, and agar was replaced by agarose to obtain brighter halos. When required, the following antibiotics were added (in micrograms per milliliter): for E. coli, ampicillin (100), chloramphenicol (25), and kanamycin (50); for B. subtilis, chloramphenicol (5) and kanamycin (10). DNA isolations. Total genomic DNA was extracted from B. caldolyticus grown in TY broth at 70°C essentially as

Corresponding author. 4107

VAN DEN BURG ET AL.

4108

J. BACTERIOL. TABLE 1. Bacterial strains, phages, and plasmids

Strain, plasmid, or

phage

Strains Bacillus caldolyticus Bacillus subtilis DB104 DB117 Escherichia coli NM538 NM539 JM101 WK6 WK6mutS Plasmids pHPS9 pTZ12

pGDV1 pGE501 pGS1 pGB301 pGB302 pUC19 pMa5-8 pMc5-8 Phages Lambda EMBL12 calII-9

Source or or Property Prpryo genotype eoyereference

Caldo-active Bacillus species; also called YT-P

11

his nprR2 nprE18 aprA3 Emr his nprR2 nprE18 aprA3, constructed by replacement recombination of the npr gene by the Emr gene (13), by the method of Niaudet et al. (25)

16 6

supFhsdR supF hsdR(p2cox-30) supE thr A(lac-proAB) (F' traD36 proAB+ laClq lacZMJ5) A(lac-proAB) galE strA F' lacIq ZAM15 proA+B+ A(lac-proAB) galE strA mutS215::TnJO F' laclq ZAM15 proA+B+

7 7 24 36 36

Cmr Emr, 5.7 kb, cal-86::lacZcx, pTA1060-pUC9 derivative, low copy number Cmr, 2.5 kb, high-copy-number plasmid from Corynebacterium sp. replicating in B. subtilis Cmr, 2.5 kb, pTZ12 carrying multiple cloning site mp18 Cmr, 4.8 kb, pGDV1 derivative harboring the npr gene from B. stearothermophilus CU21 (37) Cmr, 5.8 kb, pGDV1 derivative harboring the npr gene from B. subtilis 1A40

10

Cmr Emr, 12.6 kb, pHPS9 containing a 7-kb B. caldolyticus chromosomal DNA fragment Cmr, 7.6 kb, pGDV1 containing a 5-kb fragment from pGB301 Apr, 2.7 kb, E. coli replicon, high copy number Apr, 3.8 kb, E. coli replicon, Fl-ORI, cat-amb Cmr, 3.8 kb, E. coli replicon, Fl-ORI, bla-amb

Genomic cloning vector Recombinant EMBL12 derivative containing a 20-kb B. caldolyticus chromosomal DNA insert

described previously (41). Plasmid DNA was isolated by the method of Ish-Horowicz and Burke (15). Lambda EMBL12 DNA was purified by the method of Frischauf et al. (7). Other DNA manipulations were performed essentially as described by Maniatis et al. (20). Construction of a B. caldolyticus genomic library in EMBL12. Lambda EMBL12 DNA was digested to completion with BamHI. B. caldolyticus chromosomal DNA was partially digested with Sau3A, and DNA fragments ranging from 8 to 23 kb were isolated from agarose gels by using NA-45 DEAE membrane filters according to the instructions of the manufacturer (Schleicher & Schuell, Dassel, Germany). Ligations were performed at a DNA concentration of approximately 1 ,ug/,ul; the ratio of chromosomal DNA to EMBL12 arms was 1:5 (wt/wt). The ligation mixture was packaged by the method of Grosveld et al. (9). The library was plated and amplified on E. coli NM539 by the method of Maniatis et al. (20). The amplified phages were stored with chloroform at 4°C. Isolation of hybridization probes and screening of the B. caldolyticus library. A 609-bp BglII-EcoRI B. subtilis and an 888-bp SalI B. stearothermophilus npr gene fragment were isolated from plasmids pGS1 and pGE501, respectively. Restriction fragments were isolated from agarose gels by using NA-45 DEAE membrane filters and labeled with digoxigenin-dUTP by using the random-primed labeling procedure as described by the manufacturer (Boehringer Mann-

1

Laboratory collection Laboratory collection

Laboratory collection (originally obtained from Gist-Brocades, Delft, The Netherlands) This work This work 44 36 36 7 This work

heim, Mannheim, Germany). Plaques were transferred to GeneScreen Plus membranes (NEN-Dupont, Boston, Mass.) by the method of Maniatis et al. (20). Hybridization and signal detection were performed with the nonradioactive DNA Labeling and Detection Kit (Boehringer Mannheim) essentially as described by the manufacturer. However, the hybridizations were performed at 58°C and the amount of blocking reagent used was 1% (wt/vol). In addition, membranes were washed less stringently (2x SSC [lx SSC is 0.15 M NaCl plus 0.015 M sodium citrate] instead of 0.1X SSC). Hybridizing plaques were picked from the plates and amplified, and purity was tested by a second hybridization step. Gel electrophoresis and Southern blot analysis. DNA gel electrophoresis was performed by the method of Maniatis et al. (20). Southern transfer was carried out by diffusion blotting of the restriction fragments from agarose gels (0.8%, wt/vol) on GeneScreen Plus (NEN-Dupont), under alkaline conditions, by the method of Maniatis et al. (20). Hybridization and signal detection was performed with digoxigenindUTP-labeled probes as described for plaque hybridizations. DNA sequencing analysis. DNA restriction fragments were subcloned in pUC19 (45) and transformed into E. coli JM101. Nucleotide sequences were determined by the dideoxy nucleotide chain termination method (32), with T7 DNA polymerase, universal sequence primers, and ox-35S-dATP (2). Nucleotide and amino acid sequences were analyzed and

HIGHLY THERMOSTABLE B. CALDOLYTICUS NEUTRAL PROTEASE

VOL. 173, 1991

4109

TABLE 2. Oligonucleotidesa used for site-directed mutagenesis Oligonucleotide

Mutation

GA C CGA GGT CGT ACC GGC GAC3' (KpnI, +) 5'GCC ATC GGC CCA CAA GCT CCC GG3' (SmaI, +) 5'CGC GTC ATA GCT CGC GAA AAA TTG3' (NruI, +)

Ala-4.Thr....... 5'GCC Thr-59-Ala. Thr-66->Phe .......

a Wild-type sequence was according to Takagi et al. (38). Nucleotide substitutions are in boldface, and changed codons are overlined. Restriction sites, created to facilitate selection of mutant clones, are shown in parentheses and are underlined in the sequence.

aligned by using the Microgenie Sequence Analysis Program

(30).

Detection of protease-producing phages and colonies and assay of protease activity. Protease-producing recombinant phages and B. subtilis colonies were identified on the basis of halo formation on plates containing skim milk (0.8%, wt/vol [29]). Protease activity was determined on the basis of casein hydrolysis, essentially as described by Fujii et al. (8). The amount of released tyrosin was determined spectrophotometrically at 275 nm after trichloroacetic acid precipitation. The thermal stability of purified neutral proteases was determined by incubating purified enzymes at different temperatures for 30 min, followed by measurement of the residual activity in 50 mM Tris-HCl-5 mM CaCl2, pH 7.0, containing 0.7% casein (wt/vol). The temperature at which 50% residual activity was present was denoted Tm. Temperature optima were determined by incubating purified enzyme in 50 mM Tris-HCl-5 mM CaCl2, pH 7.0, containing 0.7% casein (wt/vol) for 60 min at different temperatures. Inhibition of casein hydrolysis by phosphoramidon was determined at 37°C by using different concentrations of the Npr-specific inhibitor. Purification of extracellular protease. Extracellular proteases produced by B. subtilis DB117 harboring plasmids with the npr genes from B. stearothermophilus and B. caldolyticus, pGE501 and pGB302, respectively, were purified with Bacitracin-silica as described previously (39). Peak fractions were analyzed by means of casein hydrolysis and tested for purity with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), essentially by the method of Laemmli (19) with the Minitan gel electrophoresis system (Bio-Rad Laboratories, Watford, Hertfordshire, United Kingdom). Protein concentrations were determined by using the BCA protein assay reagent (Pierce, Rockford, Ill.). Bovine serum albumin was used as a standard. N-terminal sequence determination. Purified B. caldolyticus Npr was subjected to SDS-PAGE and electroblotted on Immobilon membranes (Millipore, Bedford, Mass.). After staining with Coomassie brilliant blue R-250, protein migrating at the position of mature Npr was cut out and used for N-terminal sequence determination in an automated gasphase protein sequenator (Applied Biosystems, Foster City,

Calif.). Site-directed mutagenesis. Mutagenic oligonucleotides were synthesized with an Applied Biosystems DNA synthesizer. The primers used are listed in Table 2. Mutagenesis was performed by using the gapped-duplex method of Stanssen et al. (36). The mutation efficiency was between 5 and 85%. Clones were analyzed by restriction analysis, and fragments containing the proper restriction sites were sequenced by the method of Sanger et al. (32). DNA fragments containing the mutations were then transferred to a deletion variant of the production plasmid pGE501 and transformed to B. subtilis DB117. Proteolytically active B. subtilis colonies were identified on TY-agar plates (supplemented with 0.8% skim milk as a protease substrate) by their ability to

form halos. Site-directed mutagenesis was used to construct the three single mutations (Fig. 1). The double and triple mutants were constructed by the exchange of DNA restriction fragments between pGE501 and mutant clones constructed therein and pGB302 (Fig. 1 and 2). The double mutant containing an Ala-59 and a Phe-66 (pGE501-T59A/ T66F) was constructed by substituting the 888-bp Sall fragment of pGE501 with that of pGB302. The other two double mutants and the triple mutant were obtained by ligating the 1.8-kb SnaBI fragment of pGE501-A4T to the 3.0-kb SnaBI fragments of pGE501-T59A, pGE501-T66F, and pGE501T59A1T66F, yielding pGE501-A4T/T59A, pGE501-A4T/ T66F, and pGE501-A4T/T59A/T66F, respectively. The presence of the mutations was tested by restriction analysis and sequence determination. Mutations were constructed in the B. stearothermophilus npr gene, since production of the B. stearothermophilus enzyme in B. subtilis was several times higher than that of the B. caldolyticus enzyme (see below). Enzymes and chemicals. Restriction enzymes, T4 DNA ligase, and Klenow polymerase, obtained from Boehringer Mannheim, and T7 DNA polymerase, purchased from Pharmacia-LKB (Uppsala, Sweden), were used according to the instructions of the manufacturers. a-35S-dATP was purchased from Amersham International plc, Buckinghamshire, United Kingdom. All other chemicals were of analytical grade. Nucleotide sequence accession number. The GenBank/ EMBL accession number for the sequence shown in Fig. 3 is M63575. RESULTS Cloning of the B. caldolyticus npr gene. Southern hybridization of B. caldolyticus chromosomal DNA with probes from the B. subtilis and B. stearothermophilus npr genes showed the presence of hybridizing fragments (results not shown). The signal obtained with the B. stearothermophilus probe was stronger than that obtained with the probe of the B. subtilis npr gene. Although in culture supernatants of stationary-phase B. caldolyticus cultures hardly any proteolytic activity was detectable, the hybridization signals obtained suggested the presence of DNA sequences similar to both npr genes. Shotgun cloning of signal-positive restriction fragments in pHPS9 did not result in protease-proficient B. subtilis DB117 colonies after transformation of competent cells. Therefore, we decided to clone the npr gene from B. caldolyticus by using a genomic library. This library was screened genetically for npr gene-encoding inserts by using plaque hybridization and enzymatically by plating on Trypticase-peptone agarose supplemented with 0.8% skim milk (wt/vol). By using the screening procedure for proteolytically active phages, several positive phages were identified. These were amplified and additionally tested by plaque hybridization. Three of four proteolytically active recombinant phages tested also hybridized with the probes. The

VAN DEN BURG ET

4110

pGE5O1 (4.8 kb)

B

pGB3O2 (7.6 kb)

B

J. BACTERIOL. AL.J.BTEO. Alo4 Thr59 Thr66 [A 9

If

C

Ala59.

Thr4

A

B

Phe66

A

iAf

B

B

Thr4 * r,

pGE501 -A4T (4.8 kb)

A

'i

I

l

Ala59 *

pGE501 -T59A B (4.8 kb)

I

A

I

IA

I

a

A

B

-II

I

Phe66

pGE501 -T66F (4.8 kb) FIG.

1.

A

B I

1.8 kb

Plasmids used for the construction of double

mutations in the npr gene of B.

Npr,

as

and

pGE501-T66F.

and

triple

mutants

-and triple

(see Materials and Methods),

are

are

are

Site-directed

mutants.

shown. The are

1

mutagenesis was used to construct th,e single stearothermophilus and B. caidolyticus constructed in pGE501, yielding

single

mutants were

pGE501-A4T,

indicated by asterisks. The restriction sites used for the construction of

indicated. The

also

e

The amino acids differences between B.

The amino acid substitutions

subclones in E. coii(pMalc), Sail (A) and Sacll (C),

B

-Ii

303.0 kb

stearothermophiius(pGE501).

specified by pGE5O1 and pGB3O2 respectively,

pGE501-T59A,

A

~~~~~~i

I

positions

of the SnaBI sites (B), used for the construction of the double

given.

results of the

screening procedure are.documented in Fig. 4. of these phages (calll-9) was isolated and characterized by restriction analysis. Since a 7-kb XhoI fragment from recombinant phage calII-9 strongly hybridized to the probes employed and because it was sufficiently large to contain a complete npr 'gene, this fragment was isolated and inserted in Sail-digested pHPS9. The subcloning strategy is represented in Fig. 2. The plasmid thus obtained, designated pGB301, gav'e rise to extracellular proteolytic activity in B. su'btilis DB117. However, probably because of the low copy number of this cloning vehicle, protease production was low. On the basis of the previous observation that Npr production in pGDV1 yields superior results (6), a 5-kb EcoRI fragment wa's isolated and ligated in the high-copy-number plasmid pGDV1, yielding pGB302. Plasmid pGB3O2 indeed gave rise to increased production of extracellular protea'se by B. subtilis compared with that of pGB3O1 (results not shown). Nucleotide sequence of the B. caldolyticus apr gene. Hybridization experiments showed that the putative B. caidolyticus npr gene was located on a 2-kb Hinfl-MluI fragment contained in plasmid pGB3O2 (Fig. 2). The nucleotide sequence of this fragment was dete'rmined. This sequence (Fig. 3) revealed an open reading frame (ORF) of 1,632 tiucleotides, a protein of 544 amino acids. A possible riboso'me binding site commonly used in bacilli was present 8 nucleotides upstream of the putative translation start codon (AG= -23.2 kcal/mol). The GC content of this ORF was 55.9%. The nucleotide sequence of this ORF was compared with the nucleotide sequence of the B. stearothermophilus npr gene a,nd was found to be similar. Sequence similarity was highest in the 3' part of the ORFs contai ning the sequence specifying the mature enzyme (see below), whereas identity in the 5' part, containing the sequence specifying the pre-pro part of DNA from

one

encodin'g

the enzyme (see below),

similarity

was

much lower (Table 3). This

and the fact that the

Hinfl-MluI

DNA

fragment

XhoI

tAgh 6L% C.J-.A Tr-ft u-v (49WkDJ

a

left

XhoI

XhoI

arm

XhoI

right arm

BcdI

Nphe

Sol!

Dde!

BdlI

*--coRI

Av!

Sty!

d

pGB3O2 7.6 kb

MItil

pGOV1 2.7kb

snaBI

EcoRI

SnoBI

sai!

Hpr HI FIG. 2. Construction of plasmids pGB3O1 and pGB302. Plasmid pGB3O1 was constructed by replacing the 640-bp Sail fragment of pHPS9 with a 7-kb Xhol fragment from phage calll-9. pGB3O2 was constructed by

insertion of

a 5-kb EcoRI fragment from pGB3O1 in pGDV1. Thin lines indicate vector DNA; hatched lines indicate insert DNA. The location of the Npr-specifying ORF from B. caidoiyticus is indicated in pGB302. Sites used for the

the EcoRI

site of

construction of subclones in the gene

are

shown in

pUC19 for

pGB302.

sequence determination of

VOL. 173, 1991 10 CCTGMTATA 100

ATTA

20 M

HIGHLY THERMOSTABLE B. CALDOLYTICUS NEUTRAL PROTEASE 30

40

50

110

120

130

140

150

TGACGATGTCMCATTAAAACTAGAATA 190

200

210

220

290

300

310

160

170

GCTGTATAG 230 AC

ACGTCATAGCTCGCGTCAAACTTCT 280

70 90 s0 ACAACAAAGTCTOGCUAACTA

60

ATOC^AAGTATAGTTCAGCATATTAATl

320

240 250 CC CCTAGAMTTM

330

340

T

260

4111

B

A

180 GAC 270

ATOO>WTA 350

360

GTTCAiCTATAGATAATTTTG GWAAATAGU;Gl GCTG Gl;w-TTn Me0AafLyaA6A1d4tLod;1yA1aI1eGyL.u.h.G 370 380 390 400 GATTGA =IT=iLG= =TC M

410 S

420

430

GTGAAGAAAG

1yL*MtklaTrpProPheG1yk1&rA1 LaGlyLysS-rMtValTrpAsd;GlulnTrpLysThrProS 460

470

480

500

490

310

520

440 450 rGITCMTT rEb V&lS rGlyS

530

540

CT GT CJC XAA0TGAATTTA--A---7G&TCA&GA .TTACTTCCA0CT?0GTGAAGCTCCWmAAC *rLuLulyArgCyzGlnGluL*uValTyrAr&TyrL.uAspGlnGluLyAThrGlnLlyGlyGlnAlar&GluA

550 560 570 580 590 600 610 620 630 C GCCTGAGCTA ITAC0GCA?GWTTT1GAACAGOCTA W C&?cCTG7G? ruSlrLouI lGlyALyaLuGluLoulylisThrVaLStArhduGlilIlllaAlaS*rL*uCytGlyA 640

650

660

670

1-V&lL:uV>lalisVaLsps;lyGluuSrS 730

740

750

680 rL nuS

760

690

700

710

720

GlyThrLouIleProdlsuLuspLysArSThrL uLysTWr 770

760

790

900

910

luhlaklaIllrIllGbflnAldGluMatI lahlLysGlnspVlalaphsValThrLyrGluArSProklaAldlGluGlu 620

630

$40

950

860

lyLysProThrAzgL.uValIleTyrProAsIp1luGluThzProArgL 910

920

930

940

950

970

660

890

900

SaTyrGluVl V&AtrPhL.uThrProValProG 960

970

9N0

990

FIG. 4. Identification of B. caldolyticus npr gene containing phage. (A) E. coli NM539 cells were infected with recombinant phage calII-9 containing a B. caldolyticus chromosomal DNA fragment and plated on Trypticase-peptone agarose supplemented with 0.8% skim milk (wt/vol). (B) Phages were transferred to GeneScreen Plus membranes and hybridized to an 888-bp Sall fragment from the B. stearothermophilus npr gene. Hybridization was performed at

580C.

0CAACT060TCTAC&A5TGTMA lyeAnTrpIleTyz rAtIlakphlalpklyLysValLuA,sLysTrpsGldXtAspluU LysPrdGlyGlyAlaGlnP 1000 1010 1020 I _ r _ _InWn a_v _a

1030 1050 1040 1060 1070 1060 a CSa a I OTT wASTCAGAAATATATCA&TSAm?A?TATWCCT?AT0CT

roYrAldlyThrSG1ThrValGlyVYa1GyArzglyVa1LlyAsplnLys&yrIlAsnThrThrTyrS rS.rTyrTyrGlyT 1090 1100 £C&TA?OC*C6T

1110 1120 1130 1140 W0WWCTTTA!AGm0WA&

1150

1160

1170

I6GT7CCCOOCAOCTnG1GO0GCC

yrTyrTyrLej;Ln&sp&AsThrArGlyS9rGlyIle8eThrTyrAspGlyArgAsA,SThrValL.ufro0lyS.rLuT2rpJUa 1160

1190

1200

1210

1220

1230

1240

1250

1260

spOlyAshnCl^bSPoAlSrTyrAspkl&Alakl&V&lspil UsTyr yald1yV&lV&TyrA*pTyrTyrLysAV 1270

1260

1290

1300

1310

1320

1330

1340

1350

a1EiGlyAr6LeuSerTyrAsplySerAsvAlal-IllArzsSrThrValLT yGlyAz.GlyTyrAsAsMAlPh.Trp&sz 1360

1370

1380

1390

1400

1410

1420

1430

1440

lySerGIstVITyrGlyAs GlykapOlyGIuThrPb.LeuProFhb.SrGlyGlyllAspV1lValGlyUi GluL.uThrRU1a 1430

1460

1470

1460

1490

1500

1510

1520

1530

laValThrAspTyrThrAldlyLeuValTyzGlnsGluS.zGlyAlaIleAsGlu&ld4tS rAspIlPb.01lyThrL.uV-IG 1540

1550

1560

1570

1560

1590

1600

1610

1620

lubt.TyraAL lulSlrs GlypluuAIpleTyrSpyrThzr lyValAlaGlyAspA1aLmaAzgSrNtS.rA 1630

1640

1650

1660

1670

1660

1690

1700

1710

AAA=TCA

spProA1aLysTyzGlyAspProAspBisTyrSerLyaArTyrThzG1lyThGlGsptasGlyGlyValUsThrAsSrzGlyul.I 1720

1730

1740

1750

1760

1770

1780

1790

1600

leAnLyaA6l1ayzL.uLeuS.rG0Dly1lyV1yVlasTyG1yV ValTrVh1TlylyIlGly6xAepLyd4stGlyLysIleP 1910

1620

1830

1640

1650

1660

1670

1600

1690

beTyrArsalaLuVelTyrTyrLeuThzProThzSerAenlh.StGlnLuArOlazkgAysVall;71G llAlaepLeuqTyG 1956 1900 1910 1920 1930 1940 1960 1970 GO CWm90~TMG0 G90GA?M~AAT .T G1AanC

lySerThrSerGl GluValAsezValLy1lnAlaePbeAsAlaVaelGyValTyrhd

FIG. 3. Nucleotide sequence of the B. caldolyticus npr gene and primary structure of the neutral protease. The nucleotide sequence of a 1,974-nucleotide Hinfl-MluI fragment was determined. The ORF starts at nucleotide 318 and proceeds to nucleotide 1949. The potential ribosome binding site is underlined. The putative signal peptide cleavage site is indicated by the open arrow, and the maturation site is indicated by the filled arrow.

region, containing a stretch of hydrophobic amino acids (amino acids 5 to 20) and several potential processing sites, as calculated by the method of von Heijne (42). The one indicated in Fig. 3 was determined to have the highest probability of processing (P = 57.6). The N-terminal sequence determination of the extracellular enzyme revealed the sequence Val-Ala-Gly-Thr-Ser-Thr-Val. Therefore, the maturation site of the B. caldolyticus Npr (Fig. 3) is identical to the one used in B. stearothermophilus Npr. Alignment of the amino acid sequences of the Nprs from B. stearothermophilus and B. caldolyticus (Fig. 5) showed that their mature parts were nearly identical. They only differed at positions 4, 59, and 66 of the mature enzymes. The presumed pre-pro parts of the enzymes were much less similar (Table 3). Characterization of B. caldolyticus Npr secreted by B. subtilis. B. caldolyticus Npr expressed in B. subtilis was purified to homogeneity from the culture supernatant by using Bacitracin-silica, and thermostabilities and temperature optima were determined and compared with the almost identical B. stearothermophilus enzyme (Fig. 6). Figure 6A shows that B. caldolyticus Npr had a temperature optimum approximately 8°C in excess of that of B. stearothermophilus Npr. A similar difference was observed when the enzymes were incubated at elevated temperatures and residual activities were determined (ATm = 8.2°C [Fig. 6B]). Since only three

amino acids are different in the two enzymes, these must be responsible for the observed differences in thermo-

stability. The difference in amino acid composition did not affect the specific activity of the enzymes, which were nearly identical: 21 and 27 kU/mg for B. caldolyticus and B. stearothermophilus Npr, respectively.

containing the ORF specified a neutral protease (see below) indicated that the ORF represented the B. caldolyticus npr gene. Amino acid sequence of the B. caldolyticus Npr. The amino acid sequence of B. caldolyticus Npr is similar to those of other cloned neutral proteases (8, 12, 44) and is, therefore, probably also synthesized as a pre-pro protein. The putative pre structure of the B. caldolyticus Npr shows characteristics of a typical signal sequence (43); the NH2-terminal part (n region, amino acids 1 to 4) contains two positively charged amino acids (Lys-3 and Arg-4) and is followed by the h

TABLE 3. Sequence identity comparison of B. caldolyticus and B. stearothermophilus neutral proteases Sequence

Pre-pro Mature Total gene

% Sequence identity

Nucleotides

Amino acids

70.3 98.9 86.9

66.5 99 84.3

4112 1

VAN DEN BURG ET AL.

J. BACTERIOL.

NNIKIANLGAIOLAVSLNAVIPVASAKGKSNV 111111111

I1111111:1

111111

NIQNKT?PS

100

1:1111111111

NIKANLGAIGLAVOLLAARIGASAKOI3ITVNEQUKTPS 41 41 76

61

ILTSRTLDOQKETFQLSGQARKRLS I I T*SGSLLeGGIQALEKLVTQTVDRKEG?FRLO*RA3DRLA

FVSOSLLORCSQ I1II11II1

1 1:11:

D

LI@NKLDKLGNTVNIFVQAIAASLCNGAVLTANVNDOILS I I :: 1 I |I I I I I I1 ||I I I I I I1

V

I

1

1: 11

LIGKQ?DKLONTVN3FtQOEE

11

11 1

IRYTOTNLAANVKDSZLI

11

SLSGTLIP?LD

121

ALSGSLIPNLD@QPRLKKAK?VTVQQAIAIAIQDYTDTTT

157

K

K*RP?TT

107

WITYIDAADGKVLNKONQNDKAKPGGAQPVAGTSTVSVGR

IP?AA&3 Pt6RLVITPDZKTP?LATIvuvRLLTPV

I

I111111

:

O

IIII11111111111

0

ONGQ ?tLVITWTDOTAlLATzvNvuFLTrv

1:1: 1 11: 11

201

:I111:

11: 1

1

1111:111

VTTIIDATDGAILNKFUQID3IQ?00eQPYAGAsTYOTG

R

237

OTLGDQKTIINTTS3TTGTTYTLQDTRGSOIVTYDIR?RT

241

OTLODQKTINITTSSTYOTYTLQDNTlCGSIF?TDURN?T

277

VLPIOSLWADODUQVASTDAAAVDAETTAGOYTDTTKNVE

2$1

VLPOSLWTDGDNQVTASYDAAAVDANTT&9VVTDTTKNVN

317

ORLUTDGUSAAI&3tVETOESTOENAFWOSQNVTODGDGQ

31

O*LSTDOSSAAIR STVIYSISTUUAVWOSQNVTSDODOQ

gI

357

I

Ii I I1 1 111 111

111

11III

111

?FLP?156IDVTGU3L?EAVTDTTAOLVTQ ISGAIN

1

3.7

SD1

401

SDIVOTLVIVTAUENPDWKISZDIT?POBABDALPSNSDP

437

£KTGDFDrTSK1TT?TQDUGGVETNIOIINKAATLLQGG

441

£KTOD?DUTXKRtTt?tQDUSGYUTESIIINKAATLL3QGS

6tLTVITANlUPDWlIIBDIT GPOTAGDALl33NSDP

IGIDKNKI

FTIULYVTTLTPTIEFSQL3tAAC

441

TETOTSYTGIB3DKNGKI

TUALVTTLTPTSEVSQLlAAC

517

VQAAADLTSST3QIOVNSVQAVEAVGVT VQAAADLTOSTIOQ3IVKCQAFUAVGVT

321

TB

65

70

80

75

85

100

_AN

TVL?VSOSIDVVBE3LTUAVTDTTAGLVTQNI3GAIE3AN

INTOVIV

60

11

561

477

50

KRTLKT3AAISIQQAZNIxAKQDvADRvT

161

II1::I

>

FIG. 5. Comparison of the primary structures of the neutral stearothermophilus. The primary sequences of the B. caldolyticus (BcI) and B. stearothermophilus (Bst) Npr were aligned by using the Microgenie Sequence Analysis Program (30). The one-letter code is shown. The maturation site is indicated by the arrow. proteases from B. caldolyticus and B.

Inhibition experiments with the specific Npr inhibitor phosphoramidon showed similar sensitivities of both enzymes for this inhibitor (K, = 0.3 and 0.5 ,uM for the B. caldolyticus and B. stearothermophilus enzymes, respectively). The pH optimum for the B. caldolyticus enzyme was determined to be pH 7.0 (results not shown), which was identical to that of B. stearothermophilus Npr. Although both npr genes were cloned in the same highcopy-number vector (pGDV1), the amount of B. stearothermophilus Npr produced was approximately seven times higher than that of the B. caldolyticus enzyme (results not shown). Construction and characterization of mutant neutral proteases. A set of seven mutant Nprs was constructed by using site-directed mutagenesis and restriction fragment exchange, as described in Materials and Methods. The mutant Nprs were purified with Bacitracin-silica, and their purity was analyzed by SDS-PAGE. Yields and mobilities of the mutant enzymes were similar to those of the wild-type B. stearothermophilus Npr. The specific activities of the mutant Nprs were not af-

4-0

50

0' 65

71

77

83

TemperatLre FIG. 6. Temperature optima and thermostabilities of Npr from B. caldolyticus and B. stearothermophilus. Temperature optima (A) and thermostabilities (B) were determined as described in Materials and Methods. +, B. stearothermophilus Npr; A, B. caldolyticus Npr.

fected by the mutations compared with that of the wild-type enzyme when casein was used as a substrate (results not shown). Thermal stabilities were determined for the three single mutations (Ala-4->Thr, Thr-59--*Ala, and Thr66->Phe), and Table 4 shows that all three substitutions enhanced the thermal stability of the B. stearothermophilus Npr. The mutation Thr-66-*Phe had the largest effect on thermostability (+6.2°C). Mutations Ala-4--Thr and Thr56--Ala increased thermostability of the enzyme by 1.75 and 1.5°C, respectively (Table 4). Determination of the thermostabilities of the double mutants indicated that the effects of the single mutations were additive. However, the thermostability of the triple mutant was not as high as might have been expected from the effects of the single mutations (Table 4). Combination of the three single mutations should theoretically result in an increase in thermostability of 9.45°C, which is more than one degree in excess of the observed stabilization. Nevertheless, the thermostability of the triple mutant was identical to that of wild-type B. caldolyticus Npr (Table 4).

HIGHLY THERMOSTABLE B. CALDOLYTICUS NEUTRAL PROTEASE

VOL. 173, 1991

TABLE 4. Differences in thermostabilitya of wild-type and mutant neutral proteases ATmb (oC) Neutral protease 0 B. stearothermophilus wild type ...............................

Ala-4---Thr ................................+ 1.75

Thr-59--*Ala .......

.........................

Thr-66---Phe ................................

+ 1.50

+6.2

Ala-4---Thr Ala-4--Thr

+7.9

B. caldolyticus wild type ...............................

+8.2 +8.2

+ Thr-59-*Ala ................................ + Thr-66-*Phe ................................ Thr-59--Ala + Thr-66-*Phe ................................ Ala-4--*Thr + Thr-59-*Ala + Thr-66-*Phe .................. '

+3.25 +7.4

Thermal stabilities were determined as described in Materials and Meth-

ods.

b ATm represents the change in temperature corresponding to 50% residual activity of the enzymes after incubation for 30 min, compared with that of B. stearothermophilus wild-type Npr.

Inspection of Table 4 also shows that the effects of the incorporation of a third mutation on the thermostability of a double mutant were less than the contribution of that mutation to the thermostability of the wild-type Npr. DISCUSSION

The incorporation of skim milk in the plates on which the genomic B. caldolyticus gene library in phage EMBL12 was plated enabled an enzymatic screening of phages containing DNA fragments encoding proteolytic enzymes. The combination of this enzymatic screening and plaque hybridization with heterologous probes resulted in the rapid identification of the fragment of B. caldolyticus chromosomal DNA specifying neutral protease activity. The presence of proteolytically active phages which did not hybridize with the nprspecific probes suggests that this screening method may be also useful for the isolation of other bacterial protease genes. DNA fragments from phage calII-9 were transferred to strain B. subtilis DB117, deficient in neutral and alkaline protease activity, by subcloning in pHPS9 and pGDV1, and proteaseproficient clones were obtained. The nucleotide sequence of a 2-kb Hinfl-MluI fragment was determined and revealed an ORF of 1,632 nucleotides similar to the npr gene from B. stearothermophilus (38). The derived amino acid sequence was compared with the Npr of B. stearothermophilus and shown to be similar (Fig. 5; Table 3). These results, in combination with the observation that the enzyme secreted by B. subtilis(pGB302) showed optimum activity at neutral pH and was inhibited by the specific Npr inhibitor phosphoramidon, indicate that the B. caldolyticus chromosomal DNA fragment contained the npr gene from B. caldolyticus. The deduced amino acid sequence and its comparison with the Npr from B. stearothermophilus strongly suggest that the B. caldolyticus Npr is also synthesized as a pre-pro protein. The N-terminal region of the Npr contained a sequence typical for procaryotic signal sequences. This sequence was highly similar (21 of 25 amino acids) to the signal sequence of the B. stearothermophilus Npr. The putative pro peptide showed less sequence identity with that from B. stearothermophilus, whereas the mature parts of the Nprs from B. stearothermophilus and B. caldolyticus were nearly identical (Fig. 5; Table 3). The maturation sites used in both Nprs were identical, as was shown by N-terminal amino acid determination of the extracellular Npr. The expression signals present in the B. caldolyticus npr

4113

gene were less effectively recognized in B. subtilis than were those of the B. stearothermophilus npr gene. These differences in expression in B. subtilis may derive from differences in the promoter regions of both genes. In the B. stearothermophilus npr gene a probable promoter region preceded the ORF. The -10 region of this putative promoter showed sequence similarity with that recognized by the RNA polymerase 843. However, in the B. caldolyticus sequence no promoter sequence similar to any of the known B. subtilis RNA polymerases was present upstream of the ORF. The absence of a canonical promoter sequence in the B. caldolyticus gene may well explain the low activity of the gene in B. subtilis. The difference in the primary sequences between the two extracellular enzymes was limited to three substitutions, at positions 3, 59, and 66. Nevertheless, considerable differences existed between specific properties of the two enzymes; both the optimum temperature of activity and the thermal stability of the B. caldolyticus enzyme were 7 to 8'C higher than those of B. stearothermophilus (Fig. 6). One of these residues was previously shown to be involved in the thermostability of the B. stearothermophilus Npr by Imanaka et al. (14), who used site-directed mutagenesis to replace Thr-66 with a Ser. This mutation resulted in a large decrease in thermostability of the mutant compared with that in the wild-type enzyme. The presence of a Phe at this position in the B. caldolyticus enzyme which is also present at this position in thermolysin, the highly thermostable Npr from B. thermoproteolyticus, suggests an important role of this residue in thermostability. This role was confirmed by the site-directed mutagenesis experiments. The substitution of Thr-66 by Phe in the B. stearothermophilus Npr enhanced the thermostability by 6.2'C. To a lesser degree, substitution of Ala-4 by Thr and Thr-59 by Ala also increased the thermostability of the B. stearothermophilus Npr (Table 4). Examination of the three residues in a three-dimensional model of the B. stearothermophilus Npr, constructed on the basis of the well-defined thermolysin structure (6), showed that residues 4, 59, and 66 are located at the surface of the

protein (Fig. 7). In general, solvent-exposed residues are not supposed to be key residues in the stabilization of proteins (4, 26). The results presented here suggest that solvent-exposed residues may be quite important in determining the thermal stabilities of neutral proteases. Stabilization of proteins by substitution of solvent-exposed residues has also been observed by Kubo and Imanaka (18) and Pakula and Sauer (26, 27) for the neutral protease NprM from B. stearothermophilus MK232 and for the lambda Cro protein, respectively. NprM differed at two positions from thermolysin, such that the hydrophilic Asp-37 in thermolysin was replaced by the hydrophilic Asn in B. stearothermophilus NprM and the hydrophilic Glu-119 was replaced by the hydrophilic Gln. The amino acids at these positions are also solvent exposed. These substitutions rendered the NprM even more thermostable than thermolysin. This stabilization was attributed to increased hydrogen bonding and an improved electrostatic balance at the protein

surface. The results obtained with lambda Cro suggested a correlation between decreased hydrophobicity of exposed residues and increased thermostability. In the present case two of the solvent-exposed substitutions (Thr-59--->Ala, Thr66--Phe) concerned the replacement of hydrophilic by hydrophobic amino acids. Thus, it would seem that additional mechanisms involving solvent-exposed amino acids in determining the thermostability of enzymes exist.

4114

VAN DEN BURG ET AL.

tracing of B. stearothermophilus Npr. The model of stearothermophilus Npr was built from the known structure of thermolysin (21) as described by Eijsink et al. (6). Positions of the amino acid substitutions observed between B. caldolyticus and B. stearothermophilus Npr are shown by arrows. FIG. 7. C-a

B.

ACKNOWLEDGMENTS This work was supported by the Netherlands Committee for Industrial Biotechnology. We thank H. Mulder for preparation of the figures, W. J. Weijer (Eurosequence, Groningen, The Netherlands) for performing the N-terminal sequencing, and B. van der Vinne for technical assistance throughout this work.

REFERENCES

Aoki, T.,

N. Nochugi, M. Sasatsu, and M. Kono. 1987. Complete nucleotide sequence of pTZ12, a chloramphenicol-resistance plasmid of Bacillus subtilis. Gene 51:107-111. la.Avakov, A. S. Personal communication. 2. Biggin, M. D., T. J. Gibson, and G. F. Hong. 1983. Buffer gradient gels and 35S label as an aid to rapid DNA sequence determination. Proc. Natl. Acad. Sci. USA 80:3963-3965. 3. 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. 4. Creighton, T. E. 1990. Protein folding. Biochem. J. 270:1-16. 5. Debabov, V. G. 1982. The industrial use of bacilli, p. 331. In D. Dubnau (ed.), The molecular biology of bacilli, vol. I. Academic Press, Inc., New York. 5a.Eijsink, V. G. H., B. K. Stuip, B. van den Burg, G. Venema, and G. Vriend. 1989. Abstr. 2nd Int. Conf. Protein Eng., abstr. SII, p. 47. 6. Eisink, V. G. H., G. Vriend, B. van den Burg, G. Venema, and B. K. Stulp. 1990. Contribution of the C-terminal amino acid to the stability of Bacillus subtilis neutral protease. Protein Eng. 4:99-104. 7. Frischauf, M., H. Lehrach, A. Poutska, and N. Murray. 1983. Lambda replacement vectors carrying polylinker sequences. J. 1.

J. BACTERIOL. Mol. Biol. 170:827-842. 8. Fujii, M., M. Takagi, T. Imanaka, and S. Aiba. 1983. Molecular cloning of a thermostable neutral protease gene from Bacillus stearothermophilus in a vector plasmid and its expression in Bacillus stearothermophilus and Bacillus subtilis. J. Bacteriol. 154:831-837. 9. Grosveld, F. G., H.-H. M. Dahl, E. De Boer, and A. Flavell. 1981. Isolation of alpha-globin-related genes from a human cosmid library. Gene 13:227-237. 10. Haima, P., D. van Sinderen, H. Schotting, S. Bron, and G. Venema. 1990. Development of a ,-galactosidase ox-complementation system for molecular cloning in Bacillus subtilis. Gene 86:63-69. 11. Heinen, U. J., and W. Heinen. 1972. Characteristics and properties of a caldo-active bacterium producing extracellular enzymes and two related strains. Arch. Microbiol. 82:1-23. 12. Honjo, M., K. Manabe, H. Shimade, I. Mita, A. Nakayama, and Y. Furutani. 1984. Cloning and expression of the gene for neutral protease of Bacillus amyloliquefaciens in Bacillus subtilis. J. Biotechnol. 1:265-277. 13. Horinouchi, H., and B. Weisblum. 1982. Nucleotide sequence and functional map of pE194, a plasmid that specifies inducible resistance to macrolide, lincosamide, and streptograminin type B antibiotics. J. Bacteriol. 150:802-814. 14. Imanaka, T., M. Shibazaki, and M. Takagi. 1986. A new way of enhancing the thermostability of proteases. Nature (London) 324:695-697. 15. Ish-Horowicz, D., and J. F. Burke. 1981. Rapid and efficient cosmid cloning. Nucleic Acids Res. 9:2989-2998. 16. Kawamura, F., and R. H. Doi. 1984. Construction of a Bacillus subtilis double mutant deficient in extracellular alkaline and neutral proteases. J. Bacteriol. 160:442-444. 17. Keay, L. 1969. Neutral proteases of the genus Bacillus. Biochem. Biophys. Res. Commun. 36:257-265. 18. Kubo, T., and T. Imanaka. 1988. Cloning and nucleotide sequence of the highly thermostable neutral protease gene from Bacillus stearothermophilus. J. Gen. Microbiol. 134:1883-1892. 19. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680-685. 20. Maniatis, T., E. F. Fritsch, and F. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 21. Matthews, B. W., J. N. Jansonius, P. M. Colman, B. P. Schoenborn, and D. Dupourque. 1972. Three dimensional structure of thermolysin. Nature (London) New Biol. 238:37-41. 22. Menendez-Arias, L., and P. Argos. 1989. Engineering protein thermostability; sequence statistics points to residue substitutions in ox-helices. J. Mol. Biol. 206:397-406. 23. Merkler, D. J., G. K. Farrington, and F. C. Wedler. 1981. Protein thermostability: correlations between calculated macroscopic parameters and growth temperature for closely related thermophilic and mesophilic Bacilli. Int. J. Pept. Protein Res. 18:430-442. 24. Messing, J. 1979. A multipurpose cloning system based on the single stranded DNA bacteriophage M13. Recombinant DNA technical bulletin. National Institutes of Health publication no. 79-99, p. 43-48. National Institutes of Health, Bethesda, Md. 25. Niaudet, B., A. Gose, and S. D. Ehrlich. 1982. Insertional mutagenesis in Bacillus subtilis: mechanism and use in gene cloning. Gene 19:277-284. 26. Pakula, A. A., and R. T. Sauer. 1989. Amino acid substitutions that increase the thermal stability of the lambda Cro protein. Proteins Struct. Funct. Genet. 5:202-210. 27. Pakula, A. A., and R. T. Sauer. 1990. Reverse hydrophobic effects relieved by amino-acid substitutions at a protein surface. Nature (London) 344:363-364. 28. Pauptit, R. A., R. Karlsson, D. Picot, J. A. Jenkins, A.-S. Niklaus-Reimer, and J. N. Jansonius. 1988. Crystal structure of neutral protease from Bacillus cereus refined at 3.0 A resolution and comparison with the homologous but more thermostable enzyme thermolysin. J. Mol. Biol. 199:525-537. 29. Priest, J. 1977. Extracellular enzyme synthesis in the genus

VOL. 173, 1991

HIGHLY THERMOSTABLE B. CALDOLYTICUS NEUTRAL PROTEASE

Bacillus. Bacteriol. Rev. 41:711-753. 30. Queen, C., and L. J. Korn. 1984. A comprehensive sequence analysis program for the IBM personal computer. Nucleic Acids Res. 12:581-599. 31. Querol, E., and A. Parrilla. 1985. Tentative rules for increasing the thermostability of enzymes by protein engineering. Enzyme Microb. Technol. 9:238-244. 32. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467. 33. Shimnada, H., M. Honjo, I. Mita, A. Nakayama, A. Akaoka, K. Manabe, and Y. Furutani. 1985. The nucleotide sequence and some properties of the neutral protease gene of Bacillus amyloliquefaciens. J. Biotechnol. 2:75-85. 34. Sidler, W., B. Kumpf, B. Peterhans, and H. Zuber. 1986. A neutral protease produced by Bacillus cereus with high sequence homology to thermolysin: production, isolation and characterization. Appl. Microbiol. Biotechnol. 25:18-23. 35. Sidler, W., E. Niederer, F. Suter, and H. Zuber. 1986. The primary structure of Bacillus cereus neutral proteinase and comparison with thermolysin and Bacillus subtilis neutral proteinase. Biol. Chem. Hoppe-Seyler 367:643-657. 36. Stanssen, P., C. Opsomer, Y. M. McKeown, W. Kramer, M. Zabeau, and H.-J. Fritz. 1989. Efficient oligonucleotide directed construction of mutations in expression vectors by the gapped duplex DNA method using alternating selectable markers. Nucleic Acids Res. 17:4441-4454. 37. Suzuki, Y., K. Oishi, H. Nakano, and T. Nagayma. 1987. A strong correlation between the increase in number of proline

38. 39.

40.

41. 42. 43. 44.

45.

4115

residues and the rise in thermostability of five Bacillus oligo1,6-glucosidases. Appl. Microbiol. Biotechnol. 26:546-551. Takagi, M., T. Imanaka, and S. Aiba. 1985. Nucleotide sequence and promoter region for the neutral protease gene from Bacillus stearothermophilus. J. Bacteriol. 163:824-831. van den Burg, B., V. G. H. Ejsink, B. K. Stulp, and G. Venema. 1989. One-step affinity purification of Bacillus neutral proteases using Bacitracin-silica. J. Biochem. Biophys. Methods 18:209220. Vasantha, N., T. L. D. Thompson, C. Rhodes, C. Banner, J. Nagle, and D. Filpula. 1984. Genes for alkaline protease and neutral protease from Bacillus amyloliquefaciens contain a large open reading frame between the regions coding for signal sequence and mature protein. J. Bacteriol. 159:811-819. Venema, G., R. H. Pritchard, and T. Venema-Schroeder. 1965. Fate of transforming deoxyribonucleic acid in Bacillus subtilis. J. Bacteriol. 89:1250-1255. von Heijne, G. 1986. A new method for predicting signal sequence cleavage sites. Nucleic Acids Res. 14:4683-4690. Watson, M. E. E. 1984. Compilation of published signal sequences. Nucleic Acids Res. 12:5145-5164. Yang, M. Y., E. Ferrari, and D. J. Henner. 1984. Cloning of the neutral protease gene of Bacillus subtilis and the use of the cloned gene to create an in vitro-derived deletion mutation. J. Bacteriol. 160:15-21. Yannish-Perron, C., J. Vierra, and J. Messing. 1982. Improved M13 phage cloning vectors and host strains: nucleotide sequence of the M13mpl8 and pUC19 vectors. Gene 33:103-119.