Vol. 173, No. 20JOURNAL OF BACTERIOLOGY, Oct. 1991, p. 6537-65460021-9193/91/206537-10$02.00/0Copyright C 1991, American Society for Microbiology

Erwinia carotovora subsp. carotovora Extracellular Protease:Characterization and Nucleotide Sequence of the Gene

SIRKKA R. M. KYOSTIO,t CAROLE L. CRAMER, AND GEORGE H. LACY*

Laboratory for Molecular Biology of Plant Stress, Department of Plant Pathology, Physiology, and Weed Science,Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061

Received 29 January 1991/Accepted 6 August 1991

The prtl gene encoding extracellular protease from Erwinia carotovora subsp. carotovora EC14 in cosmidpCA7 was subcloned to create plasmid pSK1. The partial nucleotide sequence of the insert in pSK1 (1,878 bp)revealed a 1,041-bp open reading frame (ORF1) that correlated with protease activity in deletion mutants.ORF1 encodes a polypeptide of 347 amino acids with a calculated molecular mass of 38,826 Da. Escherichia colitransformed with pSKI or pSK23, a subclone of pSK1, produces a protease (Prtl) intracellularly with amolecular mass of 38 kDa and a pI of 4.8. Prtl activity was inhibited by phenanthroline, suggesting that it isa metalloprotease. The prtl promoter was localized between 173 and 1,173 bp upstream of ORF1 byconstructing transcriptional lacZ fusions. Primer extension identified the prtl transcription start site 205 bpupstream of ORF1. The deduced amino acid sequence of ORF1 showed significant sequence identity tometalloproteases from Bacillus thermoproteolyticus (thermolysin), B. subtilis (neutral protease), Legionellapneumophila (metalloprotease), and Pseudomonas aeruginosa (elastase). It has less sequence similarity tometalloproteases from Serratia marcescens and Erwinia chrysanthemi. Locations for three zinc ligands and theactive site for E. carotovora subsp. carotovora protease were predicted from thermolysin.

Erwinia carotovora subsp. carotovora EC14 is a gram-negative bacterium which causes soft rot on many plantspecies (39). Soft rot of potato (Solanum tuberosum) tubersis associated with the production of several degradativeenzymes secreted by E. carotovora subsp. carotovora,including pectolytic enzymes, cellulases, proteases, andphospholipases (10, 57). Pectolytic enzymes probably playthe most important role in maceration by degrading pecticcomponents of the plant cell wall and middle lamella, result-ing in separation of the cells (5). Possible roles for otherdegradative enzymes have not been established.

Several erwinias and pseudomonads causing soft rot se-crete proteases. Among soft-rot pseudomonads, extracellu-lar protease production correlates more strongly with theability to macerate plant tissue than does pectolytic enzymeproduction (46). In E. carotovora subsp. carotovora, largeamounts of extracellular protease are produced when thebacterium is grown in rich broth, on bean (Phaseolus vul-garis) hypocotyls, or on sliced cucumber (Cucumis sativus)fruit (57), but the physiological role of the protease isunknown. Protease may aid in the degradation of plant cellwall components, cytoplasmic membranes, or cytosolic pro-teins. Purified E. carotovora subsp. carotovora proteasecauses limited cell death on cucumber disks and lysis ofcucumber protoplasts (57). The release of amino acids andsmall peptides by E. carotovora subsp. carotovora proteasemay increase the rate of bacterial growth, thereby increasingthe ability of the pathogen to colonize its host. The nutri-tional benefit derived from the proteolytic digestion of hostmacromolecules may contribute to greater virulence (in thesense of causing greater host damage) of protease-producinghuman pathogens, including Pseudomonas aeruginosa,Staphylococcus aureus, and Vibrio cholerae (61).

* Corresponding author.t Present address: Laboratory of Molecular and Cellular Biology,

National Institute of Diabetes and Digestive and Kidney Diseases,Bethesda, MD 20892.

A cosmid (pCA7) encoding extracellular protease (Prtl)was previously detected in an E. carotovora subsp. caroto-vora genomic library and complemented a TnS-induced,protease-deficient mutant, L-763, of E. carotovora subsp.carotovora (2). Here we report subcloning and sequencing ofthe protease gene (prtl) from pCA7. To clarify the functionof the protease encoded by prtl, we have further constructedan E. carotovora subsp. carotovora marker exchange mu-tant for prtl, since Southern analysis of protease-deficientmutant L-763 indicated that prtl was not interrupted by TnS(26). Our results show that the DNA sequence ofprtl is mostsimilar to Bacillus thermoproteolyticus thermolysin (55);prtl shows little sequence relatedness to metalloproteases ofE. chrysanthemi, a closely related soft-rot pathogen (12, 13).

MATERIALS AND METHODS

Bacterial strains, plasmids, media, and chemicals. E.carotovora subsp. carotovora EC14 and cosmid pCA7 havebeen described previously (2, 43). Escherichia coli DH5a(Bethesda Research Laboratories [BRL], Inc., Gaithers-burg, Md.) and CB806 (48) were used as plasmid hosts.Plasmids pSK- (Bluescript; Stratagene, La Jolla, Calif.),pBR322 (BRL), and pUC4-KIXX (Pharmacia, Inc., Piscat-away, N.J.) were used for cloning. Plasmid pCB267 containsa promoterless lacZ gene (48) useful for detecting promoteractivity. Bacteria were grown in LB broth (Difco, Detroit,Mich.), on LB agar (Difco) containing 100 ,ug of ampicillin or10 ,ug of tetracycline per ml in protein extraction medium(0.1% polygalacturonic acid, 0.5% tryptone, and 0.5% yeastextract) or on Davis minimal agar (28). Gelatin plates (19)were used to detect E. carotovora subsp. carotovora prote-ase production in E. coli transformed with plasmids. Chem-icals, antibiotics, and dyes were obtained from Sigma Chem-ical Co., St. Louis, Mo., unless stated otherwise.

IEF. Isoelectric focusing (IEF) gels were prepared and runby the method of Ried and Collmer (42) except that theampholyte concentration was modified; 0.4 ml of ampholytes

6537

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

6538 KYOSTIO ET AL.

pH 3 to 10 (Servalyte; Serva Fine Chemicals, Haake BuchlerInstruments, Inc., Saddle Brook, N.J.) and 0.9 ml of am-pholytes pH 4 to 6 (Servalyte) were used. Intra- and extra-cellular protein samples for IEF were prepared from 5-mlovernight cultures grown in LB at 37°C with orbital shaking(100 rpm). The turbidity at 600 nm was measured for eachculture. Cultures were centrifuged for 2 min at 1,000 x g topellet the cells; pellets and supernatants were saved. Eachculture supernatant (2.0 ml) was concentrated fivefold withmicroconcentrators (exclusion limit, 10 kDa; Centricon-10;Amicon Corp., Danvers, Mass.). For extracellular proteins,concentrated supernatants were standardized to the sameoriginal culture turbidity and applied (10 to 15 p.d) to the IEFgel. For intracellular proteins, cell pellets were washed twicewith 0.05 M Tris HCl (pH 8.0) and resuspended in 250 [1I ofthe same buffer, and 1 drop of toluene was added to lyse thecells. The mixture, in capped tubes, was vortexed for 30 sand then allowed to stand open to the atmosphere for 30 minto evaporate the toluene. Protein concentrations for celllysates were determined by the Bradford (9) method (Bio-Rad, Richmond, Calif.). Equal amounts of protein (4.2 pug)were applied to an IEF gel.

Protease activity in IEF gels was localized by using a0.35-mm gelatin overlay (4% gelatin and 1% agarose in 50mM Tris HCI [pH 8.0]). The overlay was cast on plasticsupport (Agarose Gelbond; FMC BioProducts, Rockland,Mass.) (42) and incubated in contact with the acrylamide gelat 37°C for 1 h to overnight. Protease activity was detectedby submerging the overlay in HgCl2 (12 g in 80 ml of waterand 16 ml of concentrated HCl), which causes a cloudy whitegelatin precipitation. Clear zones indicate areas where pro-tease has degraded the gelatin.

Protease inhibitor assays. 3,4-Dichloroisocoumarin,EDTA, iodoacetate, pepstatin, phenanthroline, phenyl-methanesulfonyl fluoride, and phosphoramidon (N-[a-L-rhamnopyranosyl-oxyhydroxyphosphinyl]-L-leucyl-L-tryp-tophan) were tested for ability to inhibit protease activity(45). Protease was preincubated with each inhibitor for 20min at 37°C before the assay for activity. Protease activitywas measured by azocasein degradation by the method ofReckelhoff et al. (41) with the following modifications: thereaction was started by adding 125 [L1 of prewarmed 2%azocasein (37°C) to 125 RI1 of enzyme preparation in 10 mMTris HCl (pH 8.0) containing 2 mM CaCl2 and then incubat-ing for 20 min at 37°C. The reaction was stopped by adding1 ml of5% trichloroacetic acid. The mixture was centrifugedfor 2 min at 1,000 x g at room temperature, and the A340 ofthe supernatant was measured. One unit of protease activitywas defined as an absorbance increase of 0.001 min-'.

Il-Galactosidase assays. Expression of ,-galactosidase wasdetected on LB plates plus ampicillin spread with 100 ,ul of a20-mg/ml solution of indolyl-,-galactoside (BRL). Enzymeassays for ,B-galactosidase activity were performed by themethod of Miller (34).DNA procedures. Procedures for agarose gel electropho-

resis, restriction enzyme analysis, and ligation were per-formed by using standard methods (30). Plasmid DNA usedfor cloning and sequencing was isolated from 5-ml LBcultures, using an alkaline lysis method (23). DNA fragmentsfor cloning were separated by electrophoresis in 0.7% (Sea-Plaque; FMC BioProducts) or 1.0% (NuSieve; FMC Bio-Products) low-melting-point agarose gels, and the DNA fromthe excised gel bands was used for ligation (52). Recombi-nant plasmids were transformed into E. coli by the method ofHanahan (18). Restriction enzymes, T4 ligase, exonucleasesIII and VII, and Klenow fragment of DNA polymerase I

were purchased from BRL. Alkaline phosphatase was pur-chased from Boehringer Mannheim Biochemicals, Indianap-olis, Ind.DNA sequencing and analysis. Unidirectional deletions of

pSK1 and pSK2 were prepared by cleaving the plasmidswith ApaI and Clal within the multiple cloning site of pSK-to create 3' and 5' overhangs, respectively, which wasfollowed by digestion of the 3' recessed strand with exonu-clease III (22). Single-stranded DNA was removed by diges-tion with exonuclease VII; the blunt ends were created bythe Klenow fragment, ligated, and transformed into E. coli.The approximate location ofprtl within pSK1 and pSK2 wasdetermined by screening deletion mutants for Prtl activityon gelatin plates. Deletion clones that had lost proteaseactivity and clones that flanked this region were sequencedby the Sanger dideoxy-chain termination method (47), usinga kit (Sequenase; United States Biochemicals Corp., Cleve-land, Ohio) and 35S-dATP (NEN Research Products, Du-Pont Co., Wilmington, Del.). Sequence data were analyzedby using the University of Wisconsin Genetics ComputerGroup 1989 software (14).

Primer extension. E. carotovora subsp. carotovora wasgrown in protein extraction medium at 30°C with orbitalshaking (100 rpm) to early and late stationary phases (tur-bidity at 600 nm = 0.9 and 1.4, respectively). Total cellularRNA for primer extension analysis was isolated by themethod of Szumanski and Boyle (53). The transcription startsite of prtl was determined by using a 22-nucleotide-longsynthetic oligonucleotide (5'-GGTGCCGTTTGCGATAATACGA-3') complementary to prtl synthesized by M. Leder-man (Department of Biology, Virginia Polytechnic Instituteand State University, Blacksburg). The primer extensionprotocol was obtained from R. Moore (Department of Patho-biology, Virginia-Maryland Regional College of VeterinaryMedicine, Virginia Polytechnic Institute and State Universi-ty). Approximately 20 to 40 ,ug of RNA and 40 ng of primerwere ethanol precipitated together overnight at 20°C. Aftercentrifugation for 45 min at 10,000 x g at 4°C, the pellet wasdissolved overnight in 12.5 ,ul of diethylpyrocarbonate-treated water. The primer was annealed by heating themixture at 100°C for 90 s and slowly cooling it to roomtemperature. The primer extension reaction was performedin reverse transcriptase assay buffer (50 mM Tris HCI [pH8.0], 6 mM MgCl2, 40 mM KCI, 10 mM dithiothreitol).dCTP, dGTP, and dTTP, 1 mM each; 1 ,ul of 35S-dATP(1,320 Ci/mmol; 10.4 mmol/ml); and 40 U of reverse tran-scriptase (Boehringer Mannheim) were added to the mixturein a total volume of 22 ,u1, and the mixture was incubated at42°C for 20 min. The reaction was chased by adding 1 mMconcentrations of all deoxynucleotides and incubating for anadditional 30 min. Three microliters of the mixture wasmixed with 1 pL. of stop dye from the Sequenase kit, and themixture was run on a 7% (wt/vol) acrylamide sequencing gel.A sequencing reaction performed with the same primer andpSK1 was used to determine the transcription start site.

Northern analysis. RNA was isolated from E. carotovorasubsp. carotovora grown in glycerol broth or in planta, andthe prtl mRNA expression was measured by Northern(RNA) analysis as described previously (65). A 32P-labelled,1-kb SmaI-EcoRI fragment of pSK1 was used as a prtlprobe. An RNA ladder (BRL) was used to determine themolecular size of the RNA transcript.

Bacterial mating, plasmid curing, and marker exchange.Plasmids with prtl inactivated by inserting a gene for kana-mycin resistance were mated into E. carotovora subsp.carotovora, and marker exchange for wild-type prtl was

J. BACTERIOL.

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 173, 1991

Plasmid VectorSi

pSKI pSK- placIN

Bm/B N

pSK2 pSK- lac !

4-prt

S EV ElIl I

N

prt

S/EV S EV/SpSK21 pCB267

pSK22 pCB267

pSK23 pSK-

EV/X

pSK16 pBR322 L

S EV/S

:: lacZ

EV/N S EV El N/EV

p lac I-i/Z pr\

N Sa/EVS EV/S S/EV El Nkn

prt

FIG. 1. Endonuclease restriction site maps for inserts of E. carotovora subsp. carotovora DNA in plasmids pSK1, pSK2, pSK21, pSK22,pSK23, and pSK16. The E. carotovora subsp. carotovora protease gene (prtl; thick line), directions of transcription (arrows: solid, functionalgene; dashed, nonfunctional), lacZ promoters in pSK- vector (P), P-galactosidase genes (lacZ), inserted kanamycin resistance gene (kan), anddigestion sites for endonucleases BamHI (Bm), BglII (B), EcoRI (El), EcoRV (EV), Hindlll (Hd), HpaI (H), NruI (N), Sall (Sa), SmaI (S),and XbaI (X) are indicated. Double restriction endonuclease sites are results of blunt-end ligations and are not functional. The constructionof the plasmids is described in Results.

forced by phosphate starvation. Triparental mating was

performed with helper plasmid pRK2013 (15). Bacterialstrains (E. carotovora subsp. carotovora, E. coli HB1O1/pRK2013, and E. coli DH5a/pSK16 containing inactivatedprtl) grown separately in LB medium with orbital shaking(100 rpm) at 30°C overnight were mixed together and im-pacted by filtration on membranes (0.45-,um pore size) andincubated on LB plates at 30°C overnight. Bacteria were

washed from membranes with sterile deionized water, di-luted serially, plated on Davis minimal agar containing 30 pugof kanamycin per ml, and incubated at 30°C for 48 h. Plasmidtransfer was confirmed for single Kanr colonies by plasmidisolation, electrophoresis in agarose, and visualization byethidium bromide fluorescence in UV light. To cure plasmidsand force marker exchange (44), bacteria were grown in A-Pphosphate starvation medium (56) containing 30 ,ug of kana-mycin per ml. Loss of plasmids and Prtl- phenotype were

confirmed by agarose gel electrophoresis and assay on

gelatin plates, respectively.Southern hybridization. Chromosomal DNA was isolated

by the method of Ausubel et al. (3) and digested with EcoRI,and fragments were separated electrophoretically on 0.7%agarose. DNA was transferred to a nylon membrane (Ny-tran; Schleicher & Schuell, Inc., Keene, N.H.) by themethod of Maniatis et al. (30). Probe labelling, hybridiza-tions, washings, and detection were carried out in accor-

dance with manufacturer's instructions (Genius System;Boehringer Mannheim).

Nucleotide sequence accession number. The nucleotidesequence for prtl was submitted to GenBank and assignedaccession number M36651.

RESULTS

Subcloning of prtl. BglII digested pCA7 (2) into six frag-ments; an 8.0-kb BglII fragment cloned into the BamHI siteof pSK- (pSK10) expressed protease activity (Prtl) in E. coliDH5a. Plasmid pSK10 was digested with HpaI and HindIIIto delete a 3.0-kb fragment. The remaining DNA, containinga 5.0-kb BglII-HpaI insert from E. carotovora subsp. caroto-vora and the pSK- vector, was purified from an agaroseelectrophoresis gel, blunt ends were made with Klenowfragment, and the DNA religated to create pSK2 (Fig. 1). Toclone the 5.0-kb BglII-HpaI fragment in the opposite orien-tation in pSK-, pCA7 was digested with BglII and HpaI, thefragments were treated with Klenow fragment, and thegel-purified BglII-HpaI fragment was ligated into SmaI-cleaved pSK- to produce pSK1 (Fig. 1). Prtl activity on

gelatin plates in E. coli transformed with pSKl or pSK2indicated that the 5.0-kb BglII-HpaI fragment contained theprtl promoter as well as the open reading frame (ORF).

E. carotovora subsp. carotovora and E. coli DH5aL con-

taining pCA7, pSK1, or pSK2 expressed a protease with a pIof 4.8 (Fig. 2). Occasionally, a second band, probably causedby protease degradation, was detected at pI 5.8. Most of thePrtl activity for E. coli DH5a/pSK1 or /pSK2 was detected

N El EV S N B/S

H/Hd

I_I J :: lacZ1 kb

Prtl 6539

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

6540 KYOSTIO ET AL.

Cr,,

FIG. 2. Protease activity for E. carotovora subsp. carotovora(lane 1) and E. coli DH5ot transformed with pCA7 (lane 2), pSK1(lane 3), pSK2 (lane 4), or pSK- (lane 5) determined by IEF followedby gelatin overlay. (A) Extracellular fractions. (B) Intracellularfractions. Loading sites at the bottom are visible in all lanes.

in the intracellular fraction. No protease activity was de-tected for E. coli DH5ot transformed with vector pSK-. E.coli DH5oa/pSK2 had approximately twofold more proteaseactivity than E. coli DH5o/pSK1 as estimated from the sizeof the clearing zone in the gelatin overlay. This increase wasprobably due to synergistic transcriptional fusion betweenprtl and the lacZ promoter in the pSK- vector.

Nucleotide sequence of prtl. The location of prtl in the

Hpa I

Nru I

1878 18681

EcoR I

1457

5.0-kb BglII-HpaI fragment of pSK1 was determined bytesting deletion mutants for Prtl activity (Fig. 3). A 1,878-bpregion of pSK1 was sequenced for both DNA strands, usingoverlapping deletion plasmids. A single 1,041-bp ORF,ORF1, was found between bases 663 and 1,703 (Fig. 4) thatcorrelated with Prtl activity of the deletion clones. Thededuced polypeptide of ORF1 contained 347 amino acidswith a calculated molecular mass of 38,826 Da. A putativeShine-Delgarno (ribosome binding) sequence, AGGAGA,was 7 to 12 bp upstream of the ATG (Met) initiation codon.A 24-bp palindromic sequence was located 14 to 39 basesdownstream from a termination codon (TGA). This se-

quence has the potential to form a stem-loop structure witha 9-nucleotide-long stem, with one mismatch, and an 8-nu-cleotide loop. Since the stem is not followed by a T-richsequence, it is not clear whether this structure is involved intranscription termination.

Localization of prtl promoter. Promoter activity of regionsupstream of ORF1 was tested by constructing transcriptional+(prtl'-lacZ) fusions to the promoterless lacZ gene in plas-mid pCB267. Plasmid pSK1 was digested with EcoRV or

SmaI and EcoRV. A gel-purified, 1.43-kb EcoRV restrictionfragment containing 1,173 bp upstream of ORFI was clonedinto the SmaI site of pCB267 to create pSK21 (Fig. 1).Plasmid pSK22 was created by cloning a 0.43-kb SmaI-EcoRV restriction fragment containing 173 bp upstream ofORF1 into the SmaI site of pCB267 (Fig. 1). E. coli CB806/pSK21 harvested during late logarithmic phase (turbidity at600 nm = 0.9) produced 41 U of ,-galactosidase; E. coliCB806/pSK22 expressed no 3-galactodisase activity. Thisconfirms that the prtl promoter and/or sequences affectingprtl transcription lie between 173 and 1,173 bp upstream ofORF1. E. coli CB806/pSK21 harvested at stationary phase(turbidity at 600 nm = 1.4) produced about ten-fold more

,B-galactosidase activity (500 U) than cells harvested in late

EcoRI I DlIIII I I pSKI

8a I

488Nru I

86 1

1703 p4prtORF

663

350

420

697

826

1763

PRT activity

1481

FIG. 3. Localization of the E. carotovora subsp. carotovora protease gene (prtl) by protease activity of deletion mutants. The ORFdetermined by sequencing (Fig. 4) is indicated by a thick bar with starting and ending base pairs indicated within the sequenced region (1,878bp). Regions contained in deletion plasmids are shown with thin bars. Numbers indicate the extent of the deletions to the nucleotide sequencein base pairs. Protease (PRT) activity (+/-) of the mutants was determined on gelatin plates. BglII, EcoRI, Hpal, Nrul, and SmaIendonuclease digestion sites are indicated.

J. BACTERIOL.

I

f' f,

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 173, 1991 prtl 6541

1 CGTTTGCAGCAGCTTAGTTATCAGATGCAAAGCGCAGTAGAACGTCAGTTGAATCAGAACAAGCAGAA

NruI'.69 GCTAGGTATCGCCTGTTCGCGACTGGAAGGCGTTAGCCGCTGGCGACGCTGGCACGCGGCTATAACGTC

138

207

276

345

414

483

552

621

690

ACCACCGCCCCGGACGGTAAAGTGCTGAAAAACGTCGCGCAAATTTCCCCTGGCGAAACGCTGAAAACC

CGTTTGCAGGACGGCTGGGTAGAAAGTCAGGTCACGACGCTGGTTCCGAATAAAAGTTCCGTGAAAAAG

CCGCGAAAACCCTCATCCTCGACGCCAAAATAACGCGAACGGCATAAAATGCATAAATGGGGATTTGTC

CTAAAACAACGGCGATCCCCCATTCTTACCCCCGGCATCTGAACTATACTCATTCCCAGCTCACTAACT

.-35 . -10 +1.TAATACCCGTTATACTTCAAGCTGCATGTGCGTTGGCTTTCCTCGCTCACCCCAGTCACTTACTGATGT

SmaIAAGCTCCCGGGGACTCCCTGCGTCGCCGCCTTCCTGCAAGTTGAATTATTTAGGGTATACATTCACCGT

TTTCACTGCATGGGTTACCAGCCTCCACGTGGCGAAATGAAGTGGATAAAGACAGTGAGCTTTGTTGAT

SDCGTCTGTGTTACCCCTGATGAGCGTTTTCAAGGAGATGAGGTATGAAGTCCAGACCGATTTGTAGTGTG

MetLysSerArgProIleCysSerValS

ATCCCCCCTTACATTTTGCATCGTATTATCGCAAACGGCACCGACGAGCAGCGCCACTGCGCGCAACAGIleProProTyrIleLeuHisArgIleIleAlaAsnGlyThrAspGluGlnArgHisCysAlaGlnGln10 20 30

759 ACGCTGATGCACGTTCAGTCATTAATGGTCAGCCACCATCCGCGCCCGGAACCCCATGAGAAATTACCCThrLeuMetHisValGlnSerLeuMetValSerHisHisProArgProGluProHisGluLysLeuPro

40 50

828 GCCGGGCAGGCAAATCGCAGCATTCATGATGCCGAACAGCAACAACAATTGCCCGGCAAGCTGGTGCGCAlaGlyGlnAlaAsnArgSerIleHisAspAlaGluGlnGlnGlnGlnLeuProGlyLysLeuValArg

60 70. EcoRV

GCTGAAGGTCAACCCAGCAACGGCGATATCGCCGTCGATGAGGCCTACAGCTACCTAGGCGTCACTTACAlaGluGlyGlnProSerAsnGlyAspIleAlaValAspGluAlaTyrSerTyrLeuGlyValThrTyr

80 90 100

966 GACTTCTTCTGGAAGATTTTTCAACGTAACTCACTGGACGCCGAAGGGCTGCCGCTGGCTGGCACAGTCAspPhePheTrpLysIlePheGlnArgAsnSerLeuAspAlaGluGlyLeuProLeuAlaGlyThrVal

110 120

1035 CATTACGGTCAGGATTATCAGAATGCCTTCTGGAACGGGCAGCAGATGGTGTTTGGAGATGGCGACGGCHisTyrGlyG1nAspTyrGlnAsnAlaPheTrpAsnGlyG1nGlnMetValPheGlyAspGlyAspGly

130 140

1104 AAAATCTTTAATCGCTTCACGATTGCGCTTGATGTGGTCGCACATGAACTCACTCACGGCATCACCGAALysIlePheAsnArgPheThrIleAlaLeuAspValValAlaHisGluLeuThrHisGlyIleThrGlu

150 160 170

1173 AACGAAGCGGGACTGATCTATTTCCGCCAGTCCGGTGCGCTAAATGAATCGCTGTCCGATGTCTTTGGCAsnGluAlaGlyLeuIleTyrPheArgGlnSerGlyAlaLeuAsnGluSerLeuSerAspValPheGly

180 190

1242 TCCATGGTCAAGCAGTATCATTTGGGGCAAACCACAGAGCAGGCCGATTGGCTTATCGGTGCCGAGCTTSerMetValLysGlnTyrHisLeuGlyGlnThrThrGluGlnAlaAspTrpLeuIleGlyAlaGluLeu

200 210

1311 CTGGCTGACGGTATTCACGGCATGGGGCTGCGGTCGATGTCACATCCGGGCACGGCGTATGATGATGAGLeuAlaAspGlyIleHisGlyMetGlyLeuArgSerMetSerHisProGlyThrAlaTyrAspAspGlu

220 230

1380 TTGCTCGGTATCGACCCCCAGCCCTCTCACATGAACGAGTATGTGAACACCCGTGAAGACAACGGCGGCLeuLeuGlyIleAspProGlnProSerHisMetAsnGluTyrVallAsnThrArgGluAspAsnGlyGly240 250 260

1449 GTACACTTGAATTCAGGCATCCCCAACCGGGCATTCTATCTGGCGGCCATCGCGCTAGGCGGCCATTCAValHisLeuAsnSerGlyIleProAsnArgAlaPheTyrLeuAlaAlaIleAlaLeuGlyGlyHisSer

270 280

1518 TGGGAAAAAGCGGGTCGCATCTGGTACGACACGCTGTGTGATAAAACGCTGCCGCAAAATGCGGATTTCTrpGluLysAlaGlyArgIleTrpTyrAspThrLeuCysAspLysThrLeuProGlr1nAsnAlaAspPhe

290 300

1587 GAAATTTTCGCGCGCCATACCATTCAACATGCCGCTAAGCGTTTTAACCACACGGTTGCTGACATTGTCGluIlePheAlaArgHisThrIleGlnHisAlaAlaLysArgPheAsnHisThrValAlaAspIleVal

310 320 330

1656 CAGCAGTCGTGGGAAACCGTGGGCGTGGAGGTTCGGCAGGAGTTCCTATGAAGACGCTGCCGGCGCTCAGlnGlnSerTrpGluThrValGlyValGluValArgGInGluPheLeuEnd

340 347

1725 ACGACGATGCCATCATTGAGCTAGCGCGTGAAGGGGGATTTGCCTTTATCCCTAAGCTGGCGGGGCCGC

1794 GACGCTTCGGCTCGCCAGCGTACCGCCATCCGAACGGAGCGTATTGTAACGCGATCCGTCATGCCTTCT

NruI1863 CAGGCTCGCGAACCGA

FIG. 4. Nucleotide and deduced amino acid sequences of the E.carotovora subsp. carotovora protease gene (prtl). Indicated arethose sequences most closely corresponding to the E. coli consensussequences for the Pribnow box (-10), the recognition site (-35),ribosome-binding site (SD), transcription initiation site (+1), theputative cleavage sites for signal peptidase (S), restriction endonu-clease sites (NruI, SmaI, and EcoRV), and the putative transcrip-tion terminator (horizontal arrows; break in arrow represents a

mismatch in the inverted repeat). Underlined within the ORF is thelocation of the sequence complementary to the primer used forprimer extension (Fig. 5).

FIG. 5. Determination of transcriptional start site for the E.carotovora subsp. carotovora protease gene (prtl) by primer exten-sion. The sequencing reactions (A, C, G, and T) and primerextensions were initiated from primer (Fig. 4) complementary toprtl. Primer extension reactions (lanes 1 to 4) contained the follow-ing: lane 1, 20 ,ug of RNA from the early stationary growth phase;lane 2, 40 jig of RNA from the early stationary phase; lane 3, 20 ,ugof RNA from the late stationary phase; lane 4, 40 ,ug of RNA fromthe late stationary phase. The base sequence on the left (comple-mentary to that shown in Fig. 4) indicates the first deoxynucleotide,cytosine (underlined), inserted in the prtl transcript.

logarithmic phase. No activity was observed in E. coliDH5ax/pSK22 cells harvested in the stationary phase.

Primer extension located the start of the prtl mRNA 205bp upstream of ORFi (Fig. 5). The transcript was detectedwhen RNA was isolated from E. carotovora subsp. caroto-vora during the late stationary growth phase (turbidity at 600nm = 1.4), but not at the early stationary phase (turbidity at600 nm = 0.9). The putative promoter sequences, TTTCCTand TTCAAG at -7 and -29, respectively, are separated by16 bp (Fig. 4). These sequences do not resemble strongly theE. coli -10 (TATAAT) and -35 (TTGACA) consensuspromoter sequences (20).The size of the prtl transcript, 1.3 kb, was estimated from

a Northern blot, using total RNA isolated from in plantagrown E. carotovora subsp. carotovora (Fig. 6). This size isconsistent with the length of the ORFi (1,041 bp) plus 205 bpupstream and 40 to 50 bp downstream.

Sequence comparisons with other proteases. The DNA andpredicted amino acid sequences for Prtl were used to searchthe GenEMBL and NBRF data bases in the University ofWisconsin Genetics Computer Group sequence analysissoftware (14). Prtl was found to be similar to severalbacterial Zn metalloproteases, including B. thermoprote-olyticus thermolysin (55), P. aeruginosa elastase (6, 17), B.subtilis neutral protease (64), B. cereus metalloprotease (50),B. amyloliquefaciens neutral protease (58), B. stearothermo-philus metalloprotease (54), Legionella pneumophila metal-loprotease (7), E. chrysanthemi protease B (13), and Serratiamarcescens neutral protease (37). Amino acid sequencesconserved among these metalloproteases were found in thecentral part of each enzyme. Two regions of Prtl, from

897/

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

6542 KYOSTIO ET AL.

.1

residue S z a z

ECC 157-173 L D V V A H E L T H G I[TJ N E A

BT 137-153 I D V V A H E L T HA V TRY T A

BS 359-375 L D V T A H E M T H G V T Q E T A

PA 333-349 L D V A A H E V S H G F T Q N S

LP 373-389 L G G H E V S H G F EQ H S

SM 187-203 R Q T F T H E I G H A L G L S H P

ECH 184-200 R H S F TIH E I G H A L G L S H P

FIG. 6. Expression of prtl mRNA of E. carotovora subsp.carotovora. Total bacterial RNA was isolated from glycerol broth-grown cells (lane 1) or 15 h after inoculation on sliced potato tubers(lane 2) at 30°C. Ten micrograms of RNA was separated electro-phoretically on a 1.2% agarose gel, transferred onto a Nytranmembrane, and hybridized to 32P-labelled prtl probe (1.0-kb EcoRl-SmaI fragment [Fig. 1]). The molecular size of the prtl transcript(1.3 kb) was estimated by comparison with RNA markers (RNAladder; BRL) and is indicated on the right.

ECC

BT

BS

PA

LP

SM

174-193

154-173

376-395

350-369

390-409

311-33 1

ECH 222-242

residues 157 to 193 and from residues 258 to 276, were 58 to71% and 42 to 74% identical, respectively, to proteases fromB. thermoproteolyticus, B. subtilis, L. pneumophila, and P.aeruginosa (Fig. 7). Proteases from E. chrysanthemi and S.marcescens had only 18 to 32% and 21 to 26% similarity,respectively.

Characterization of Prtl. The DNA sequence informationwas used to find a restriction fragment containing only theprtl ORF1; a 1.78-kb NruI fragment from pSK1 was clonedinto the EcoRV site of pSK- to produce pSK23 (Fig. 1). E.coli DH5a/pSK23 produced a 38-kDa protein with a pl of4.8, values which resemble those calculated for the deducedprotein from ORF1. Prtl activity in the intracellular fractionof E. coli/pSK23 was inhibited by phenanthroline and phos-phoramidon, but not by phenylmethanesulfonyl fluoride,iodoacetate, or pepstatin, confirming that it is a metallopro-tease (Table 1).Marker exchange mutagenesis of prtl. The prtl in plasmid

pSK2 was insertionally inactivated by cloning a 1.2-kb SmaIfragment containing the kan gene from pUC4-KIXX into theEcoRV site within ORF1 (Fig. 1). The resulting plasmid,pSK15, was transformed into E. coli DH5a, and its Prtl-phenotype was confirmed on gelatin plates. To ensure thatno vector-derived regions of homology remained, the pSK15insert was cloned into pBR322; pSK15 was digested withXbaI and Sall, and the gel-purified 6.2-kb XbaI-SalI frag-ment was filled in by using Klenow fragment and ligated intothe EcoRV site of pBR322. The resulting E. coli DH5attransformants with this 10.6-kb plasmid, pSK16, were Ampr,Kanr, Prt-, and Tets. Plasmid pSK16 was mobilized into E.carotovora subsp. carotovora by triparental mating, andthen marker exchange mutagenesis was promoted by growthin phosphate-limiting medium. Homologous recombinantsbetween prtl::kan of pSK16 and wild-type prtl in the E.carotovora subsp. carotovora genome were obtained byscreening the bacteria for Amps, Kanr, and Prtl-.

Several Kanr mutants of E. carotovora subsp. carotovorathat had reduced protease activity on gelatin plates weredetected. One, designated L-957, was selected for Southernanalysis, which indicated that kan was inserted into prtl,causing a 1.2-kb gain in mass for a 6.5-kb fragment from anEcoRI E. carotovora subsp. carotovora chromosomal DNAdigest (Fig. 8). Protease assay of the mutant on gelatin plates

ECC

BT

BS

PA

LP

SM

ECH

258-276

225-243

443-461

418-433

461-476

240-257

238-255

a z a

G L I Y F|R|SG A L N E - S|L S D V F GGLIYQNQSGALF]NE-SLSDVMIFGG L I Y Q N ESG AWN E -Al S DWF G

NSL I Y E N Q G A L N E -F S D V F GG L I Y R G Q S G G M N E -S F S D G

GL G Q S G G M N E -S F S D M A A

SGYTAN QRINLNE FSDVGGS R Q F S I M SY W E VE N T G K G

E

Q

G

I

L

G

G

a a

D N G G V H

D N G G V H

D G G V H

D V H

D - V H

D N G G - H

D--KG-V HD--VH

L N S G I P N R A F Y L

TIS G I P N A A Y L

H SS G N R A FY

Y S SGV YIY A A A P L L D D I A A

Y S ARjP L M D D I A A

FIG. 7. Comparison of portions of the deduced amino acidsequence of the E. carotovora subsp. carotovora (ECC) proteasegene (prtl) with similar regions from other bacterial metallopro-teases. Numbers refer to the location of the residues in the deducedamino acid sequence for each gene. Gaps (-) have been introducedto achieve optimal alignments. Residues involved in zinc binding (z),substrate binding (s), and the active site (a) are indicated. Aminoacids matching E. carotovora subsp. carotovora protease are boxed.BT, B. thermoproteolyticus thermolysin; BS, B. subtilis neutralprotease; PA, P. aeruginosa elastase; LP, L. pneumophila metallo-protease; SM, Serratia marcescens neutral protease; ECH, E.chrysanthemi metalloprotease B.

showed that protease activity was approximately 60 to 80%reduced compared with the wild type (Fig. 9). Weak hybrid-ization occurred between the probe for prtl and a 4.4-kbEcoRI fragment in the mutant and possibly in wild-type E.carotovora subsp. carotovora. This may represent a secondprotease gene and may account for the protease activityremaining in mutant L-957.

DISCUSSION

Little is known about the characteristics or role(s) ofproteases in soft rot caused by erwinias. We describe thenucleotide sequence of the prtl encoding a protease (Prtl)from E. carotovora subsp. carotovora EC14, partial charac-terization of Prtl, and construction of a Prtl-deficient mutantof E. carotovora subsp. carotovora.The deduced amino acid sequence of prtl showed a high

degree of sequence identity to several bacterial metallopro-teases. This is consistent with the inhibition of the cloned

J. BACTERIOL.

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 173, 1991

TABLE 1. Effects of metal ion chelators and protease inhibitorson E. carotovora subsp. carotovora protease (Prtl) activity fromprtl cloned into plasmid pSK23 and produced by E. coli DH5a

Inhibitor" Concn Remaining enzyme activity(mM) (%) + SDb

None None 100 ± 12.1 ac

EDTA 1 99+0.5a10 91 1.6a50 91 7.4 a

Phenanthroline 1 22 ± 0.9 b10 12+11.0b50 0 ± 0.0b

Phosphoramidon 1 76 ± 3.9 b5 47 ± 4.0b

PMSF 1 100 ± 9.4 a10 100 ± 12.3 a

3,4-DCI 1 100 ± 4.7 a

lodoacetate 1 100 ± 13.3 a

Pepstatin 0.1 95 ± 9.7 aa PMSF, phenylmethylsulfonyl fluoride; 3,4-DCI, 3,4-dichloroisocoumarin.b Incubation of protease with inhibitors and remaining protease activity

were measured as described in Materials and Methods. Each treatment wasreplicated twice.

c Means followed by a different letter were significantly different at the P <0.01 level as determined by Duncan's multiple-range test.

protease by phenanthroline, an inhibitor of zinc metallopro-teases with a high chelation affinity for zinc. The derivedamino acid sequence of prtl is more similar to the thermol-ysin from B. thermoproteolyticus than to the metallopro-teases from S. marcescens (37). Similarity to thermolysinwas also suggested by the inhibition of Prtl activity byphosphoramidon, an inhibitor that binds specifically to thethermolysin active site. In contrast, all identified extracellu-lar metalloproteases from E. chrysanthemi, a related soft-rotpathogen, have higher sequence identity to the S. marces-cens proteases (12, 13). This report is the first to characterizea metalloprotease from E. carotovora.Amino acid sequence comparison of Prtl with thermolysin

(31, 32) shows that all six active-site residues, all threezinc-binding residues, and five of seven substrate-bindingresidues are identical (Table 2). Similar conservation ofthese sites has been described for other thermolysin-typeproteases (6, 17, 64). In the predicted substrate-binding site,the two amino acid changes, from residues Leu-133 andVal-192 in thermolysin to Phe-153 and Leu-216 in Prtl,respectively, represent conservative changes and do notalter hydrophobicity. A helix (residues 137 to 150) connect-ing the two domains of the peptide and a second helix(residues 160 to 180) lining the hydrophobic substrate-bind-ing pocket of thermolysin (31) are conserved in Prtl (resi-dues 156 to 170 and 180 to 200, respectively). The thermol-ysin calcium-binding sites were not conserved in Prtl, whichmay suggest that Prtl does not require calcium for itsactivity. This is also supported by the slight inhibition of Prtlactivity by EDTA, a chelator with high affinity for calcium.The high degree of amino acid identity between Prtl and

thermolysin in the substrate-binding site suggests similaritiesin the substrate specificity. Thermolysin has been shown to

PRT K AN1 2 1 2

PB R1 2

8.0 -

6.0-

4.0 -

2.0 -

FIG. 8. Southern analyses of E. carotovora subsp. carotovora(lane 1) and prtl site replacement mutant L-957 (lane 2). The probes(prt, kan, and pBR322) used for hybridization are indicated on thetop, and DNA molecular size markers (in kilobases) are on the left.A 1.0-kb SmaI-EcoRI fragment from pSK1 (Fig. 1) was used as theprtl probe (PRT). A 1.2-kb SmaI fragment from pUC4-KIXX was aprobe for the kanamycin resistance gene (KAN). EcoRI-digestedpBR322 (4.4 kb) was used as the probe for the vector (PBR). Incontrol digestions (data not shown), PRT hybridized with 4.4-kb,EcoRI-linearized pSK23 (Fig. 1); KAN only hybridized to a 1.2-kbpUC4-KIXX; and PBR hybridized with the 4.4-kb, EcoRI-linearizedpSK23, 4.4-kb EcoRI-linearized pBR322, and the 2.8-kb SmaIfragment of pUC4-KIXX.

cleave plant proteins such as the heme-free horseradish(Armoracia rusticana) peroxidase (62) and hydroxyproline-rich glycoproteins (1), which are located in the plant cell walland serve a structural and defense function against microbialattack (49). Being highly positively charged molecules, theymay interact strongly with negatively charged Prtl. Colla-

FIG. 9. Assay of protease activity of E. carotovora subsp.carotovora (A), prtl site replacement mutant L-957 (B), E. colilpSK23 (C), and E. coli (D) on gelatin plates. Bacteria were grown onthe plate at 30°C for 24 h, after which gelatin was precipitated byflooding the plates with HgCl2. Clear zones in the white precipitateindicate areas of protease activity.

prtl 6543

."

.,.,s f:,.:.:,f -fl-,Mtf'

-Mq

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

6544 KYOSTIO ET AL. J. BACTERIOL.

TABLE 2. Comparison of amino acid residues fromE. carotovora subsp. carotovora protease with the residuesimplicated in B. thermoproteolyticus thermolysin function

E. carotovora subsp.carotovora

Active siteGlu-143 .................................... Glu-163Tyr-157 .................................... Tyr-177Asp-170..................................... Asp-190Arg-203 ..................................... Arg-227Asp-226.................................... Asp-259His-231 ..................................... His-264

Zinc-binding siteHis-143 ..................................... His-162His-146 .................................... His-166Glu-166 .................................... Glu-186

Substrate-binding sitePhe-130 .................................... Phe-150Leu-133 .................................... Phe-153Val-139 ..................................... Val-159Ile-188 .....................................Ile-212Gly-189 ..................................... Gly-213Val-192 .................................... Leu-216Leu-202 ..................................... Leu-226

gen, a major structural unit in animal connective tissue, likehydroxyproline-rich glycoproteins, is hydroxyproline richand is cleaved by thermolysinlike metalloproteases from L.pneumophila and P. aeruginosa (11, 21). Degradation ofpotato cell wall hydroxyproline-rich glycoproteins (solubleextensins) by an E. carotovora subsp. carotovora extracel-lular protease has been demonstrated (29).The NH2 terminus of the deduced polypeptide from the

prtl gene shows a typical E. coli signal sequence (38). This20-amino-acid sequence has two positively charged aminoacids (lysine and arginine) and then a hydrophobic core of 12amino acids as indicated by the hydropathy plot generatedby the Kyte-Doolittle algorithm (27) and a putative isoleu-cine-isoleucine-alanine signal cleavage site. Signal se-quences have been identified in several other extracellularmetalloproteases (6, 37, 54, 58, 64). In the E. chrysanthemimetalloprotease B sequence, a short (16-residue) NH2-ter-minal pro sequence was found, but no signal sequence wasobserved (13). The COOH-terminal region of E. carotovorasubsp. carotovora protease contains mainly polar andcharged residues (60%) and may be a-helical as indicated bya hydropathy plot. COOH-terminal processing involved inthe secretion of proteases has been reported (35, 40). Wehave not confirmed processing.

Partial loss of detectable protease activity by the markerexchange mutant on gelatin plates suggests that at least twoproteases may be produced by E. carotovora subsp. caroto-vora. Our preliminary analysis of extracellular proteins of E.carotovora subsp. carotovora grown in protein extractionmedium and run on sodium dodecyl sulfate-polyacrylamidegel electrophoresis showed the presence of a 45-kDa prote-ase, but not a 38-kDa protease as predicted by the prtl gene.The absence of the 38-kDa protease in E. carotovora subsp.carotovora grown in rich broth could suggest that Prtl is notinduced under these conditions, but is induced by gelatin orin planta. The 45-kDa protease is close to the size reportedpreviously for an extracellular protease in E. carotovorasubsp. carotovora EC14 (51). We are currently characteriz-

ing this second E. carotovora subsp. carotovora protease(Prt2) and comparing it with Prtl, the 38-kDa protease.Only one protease (pl 4.6 to 4.8) has been described from

E. carotovora subsp. carotovora 71 and SR394 (24, 63). In E.chrysanthemi, one to three extracellular proteases are pro-duced per strain. These have been identified as metallo-,serine, or unknown proteases and range in mass from 50 to55 kDa and in pl from 4.6 to 5.8 (4, 13, 59, 60). Differencesin the number and type of secreted proteases may vary inplant pathogenic erwinias similar to the variations observedfor pectolytic enzymes (25).The role(s) of bacterial metalloproteases in plant or animal

pathogens remains unclear. Several metalloproteases withcharacteristics similar to those of Prtl have been reported inhuman pathogens (6-8, 16, 17, 33, 36). In these pathogens,no single factor has been identified as being sufficient for theproduction of all disease symptoms. Secreted proteases mayenhance virulence by releasing nutrients from the hostand/or by degrading host defense proteins. So far, we do notknow the significance of Prtl in soft rot. However, thedetection of elevated levels of prtl mRNA from in plantagrown E. carotovora subsp. carotovora indicates that thisprotease is produced during potato maceration. We arestudying the prtl::kan marker exchange mutant, L-957, toelucidate the role of Prtl in plant pathogenesis.

ACKNOWLEDGMENTS

We thank Robert Moore for discussion and computer analyses, Z.Yang for Northern analysis, Judith S. Bond and Russ Wolz forhelpful discussions, Lyudmil Antonov for statistical analyses, andVerlyn K. Stromberg for critical review of the manuscript.

This work was supported by grants from the Finnish CultureSociety (Suomen Kulttuurirahasto), Kemira Oy, and Sigma Xi toS.R.M.K and by grants from the U.S. Department of Agriculture(85-CRCR-1-1776), U.S. Environmental Protection Agency (R-813805-02-2), and National Science Foundation (BSR-8705445) toG.H.L.

REFERENCES1. Adair, W. S., and H. Appel. 1989. Identification of a highly

conserved hydroxyproline-rich glycoprotein in the cell walls ofChlamydomonas reinhardtii and two other Volvocales. Planta179:381-386.

2. Alien, C., V. K. Stromberg, F. D. Smith, G. H. Lacy, and M. S.Mount. 1986. Complementation of an Erwinia carotovora subsp.carotovora protease mutant with a protease-encoding cosmid.Mol. Gen. Genet. 202:276-279.

3. Ausubel, F. M., R. Brent, R. E. Kingston, N. B. Moore, J. A.Smith, J. G. Seidman, and K. Struhl. 1989. Current protocols inmolecular biology. John Wiley & Sons, New York.

4. Barras, F., K. K. Thurn, and A. K. Chatterjee. 1986. Export ofErwinia chrysanthemi (EC16) protease by Escherichia coli.FEMS Microbiol. Lett. 34:343-348.

5. Bateman, D. F., and H. G. Basham. 1976. Degradation of plantcell wall and membranes by microbial enzymes, p. 316-355. InR. Heitefuss and P. H. Williams (ed.), Encyclopedia of plantphysiology, new ser. Physiological plant pathology, vol. 4.Springer-Verlag, New York.

6. Bever, R. A., and B. H. Iglewski. 1988. Molecular cloning andnucleotide sequence of the Pseudomonas aeruginosa elastasegene. J. Bacteriol. 170:4309-4314.

7. Black, W. J., F. D. Quinn, and L. S. Tompkins. 1990. Legionellapneumophila zinc metalloprotease is structurally and function-ally homologous to Pseudomonas aeruginosa elastase. J. Bac-teriol. 172:2608-2613.

8. Booth, B. A., M. Boesman-Finkelstein, and R. A. Finkelstein.1983. Vibrio cholerae soluble hemagglutinin/protease is a met-alloenzyme. Infect. Immun. 42:638-644.

9. Bradford, M. M. 1976. A rapid and sensitive method for the

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 173, 1991 prtl 6545

quantitation of microgram quantities of protein utilizing theprinciple of protein-dye binding. Anal. Biochem. 72:248-254.

10. Collmer, A., and N. T. Keen. 1986. The role of pectic enzymesin plant pathogenesis. Annu. Rev. Phytopathol. 24:383-409.

11. Conlan, J. W., A. Baskerville, and L. A. Ashworth. 1986.Separation of Legionella pneumophila proteases and purifica-tion of a protease which produces lesions like those of Legion-naire's disease in guinea pig lung. J. Gen. Microbiol. 132:1565-1574.

12. Dahler, G. S., F. Barras, and N. T. Keen. 1990. Cloning of genesencoding extracellular metalloproteases from Erwinia chrysan-themi EC16. J. Bacteriol. 172:5803-5815.

13. Delepelaire, P., and C. Wandersman. 1989. Protease secretionby Erwinia chrysanthemi. J. Biol. Chem. 264:9083-9089.

14. Devereux, J., P. Haeberli, and 0. Smithies. 1984. A comprehen-sive set of sequence analysis programs for the VAX. NucleicAcids Res. 12:387-395.

15. Ditta, G., S. Stanfield, D. Corbin, and D. R. Helinski. 1980.Broad host range DNA cloning system for Gram-negativebacteria: construction of a gene bank of Rhizobium meliloti.Genetics 77:7347-7351.

16. Dreyfus, L. A., and B. H. Iglewski. 1986. Purification andcharacterization of an extracellular protease of Legionella pneu-mophila. Infect. Immun. 51:736-743.

17. Fukushima, J., S. Yamamoto, K. Morihara, Y. Atsumi, H.Takeuchi, S. Kawamoto, and K. Okuda. 1989. Structural geneand complete amino acid sequence of Pseudomonas aeruginosaIFO 3455 elastase. J. Bacteriol. 171:1698-1704.

18. Hanahan, D. 1983. Studies on transformation of Escherichia coliwith plasmids. J. Mol. Biol. 166:557-580.

19. Hankin, L., and S. L. Anagnostakis. 1975. The use of solid mediafor detection of enzyme production by fungi. Mycologia 67:597-607.

20. Hawley, D. K., and W. R. McClure. 1983. Compilation andanalysis of Escherichia coli promoter DNA sequences. NucleicAcids Res. 11:2237-2255.

21. Heck, L. W., K. Morihara, W. B. McRae, and E. J. Miller. 1986.Specific cleavage of human type III and IV collagens byPseudomonas aeruginosa elastase. Infect. Immun. 51:115-118.

22. Henikoff, S. 1984. Unidirectional digestion with exonuclease IIIcreates targeted breakpoints from DNA sequencing. Gene 28:351-359.

23. Ish-Horowicz, D., and J. F. Burke. 1981. Rapid and efficientcosmid cloning. Nucleic Acids Res. 9:2989-2998.

24. Kelman, A. (North Carolina State University). 1990. Personalcommunication.

25. Kotoujansky, A. 1987. Molecular genetics of pathogenesis bysoft-rot erwinias. Annu. Rev. Phytopathol. 25:405-430.

26. Kyostio, S. R. M., and G. H. Lacy. Unpublished data.27. Kyte, J., and R. F. Doolittle. 1982. A simple method for

displaying the hydropathic character of a protein. J. Mol. Biol.157:105-132.

28. Lederberg, J. 1950. Isolation and characterization of biochemi-cal mutants of bacteria. Methods Med. Res. 3:5-22.

29. Lewosc, J., A. Kelman, aind L. Sequeira. 1989. Release ofhydroxyproline rich proteins from potato cell walls by enzymesfrom Erwinia carotovora subsp. carotovora. Phytopathology79:1181.

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

31. Matthews, B. W., P. M. Colman, J. N. Jasonius, K. Titani, K. A.Walsh, and H. Neurath. 1972. Structure of thermolysin. Nature(London) New Biol. 238:41-43.

32. Matthews, B. W., J. N. Jansonius, P. M. Colman, B. P. Schoen-born, and D. Dupourque. 1972. Three-dimensional structure ofthermolysin. Nature (London) New Biol. 238:37-41.

33. McKevitt, A. I., S. Bajaksouzian, J. D. Klinger, and D. E.Woods. 1989. Purification and characterization of an extracellu-lar protease from Pseudomonas cepacia. Infect. Immun. 57:771-778.

34. Miller, J. H. 1972. Experiments in molecular genetics. ColdSpring Harbor Laboratory, Cold Spring Harbor, N.Y.

35. Miyazaki, H., N. Yanagida, S. Horinouchi, and T. Beppu. 1989.Characterization of the precursor of Serratia marcescens serineprotease and COOH-terminal processing of the precursor duringits excretion through the outer membrane of Escherichia coli. J.Bacteriol. 171:6566-6572.

36. Morihara, K., and J. Y. Homma. 1985. Pseudomonas proteases,p. 41-79. In I. A. Holder (ed.), Bacterial enzymes and virulence.CRC Press, Boca Raton, Fla.

37. Nakahama, K., K. Yoshimura, R. Marumato, M. Kikuchi, I. S.Lee, T. Hase, and H. Matsubara. 1986. Cloning and sequencingof Serratia protease gene. Nucleic Acids Res. 14:5843-5855.

38. Oliver, D. 1985. Protein secretion in Escherichia coli. Annu.Rev. Microbiol. 39:615-648.

39. Perombelon, M. C. M., and A. Kelman. 1980. Ecology of soft roterwinias. Annu. Rev. Phytopathol. 18:361-387.

40. Pohlner, J., R. Halter, K. Beyreuther, and T. Meyer. 1987. Genestructure and extracellular secretion of Neisseria gonorrhoeaIgA protease. Nature (London) 325:458-462.

41. Reckelhoff, J. F., J. S. Bond, R. Beynon, S. Savarirayan, andC. S. David. 1985. Proximity of the Mep-1 gene to H-2D onchromosome 17 in mice; Immunogenetics 22:617-623.

42. Ried, J. L., and A. Collmer. 1985. An activity stain for rapidcharacterization of pectic enzymes in isoelectric focusing andsodium dodecyl sulfate-polyacrylamide gels. Appl. Environ.Microbiol. 50:615-622.

43. Roberts, D. P., P. M. Berman, C. Allen, G. H. Lacy, and M. S.Mount. 1986. Requirement for two or more Erwinia carotovorasubsp. carotovora pectolytic gene products for maceration ofpotato tuber tissue by Escherichia coli. J. Bacteriol. 164:279-284.

44. Roeder, D. L., and A. Collmer. 1985. Marker-exchange muta-genesis of a pectate lyase isogene in Erwinia chrysanthemi. J.Bacteriol. 164:51-56.

45. Salvesen, G., and H. Nagase. 1989. Inhibition of proteolyticenzymes, p. 83-104. In R. J. Beynon and J. S. Bond (ed.),Proteolytic enzymes: a practical approach. IRL Press, Oxford.

46. Sands, D. C., and L. Hankin. 1975. Ecology and physiology offluorescent pectolytic pseudomonads. Phytopathology 65:921-924.

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

48. Schneider, K., and C. F. Beck. 1987. New expression vectors foridentifying and testing structures for initiation and terminationof transcription. Methods Enzymol. 153:452-461.

49. Showalter, A. M., J. N. Bell, C. L. Cramer, J. A. Bailey, J. E.Varner, and C. J. Lamb. 1985. Accumulation of hydroxypro-line-rich glycoprotein mRNAs in response to fungal elicitor andinfection. Proc. Natl. Acad. Sci. USA 82:6551-6555.

50. Sidler, W., E. Niederer, F. Suter, and H. Zuber. 1986. Theprimary structure of Bacillus cereus neutral proteinase andcomparison with thermolysin and Bacillus subtilis neutral pro-teinase. Biol. Chem. Hoppe-Seyler 367:643-657.

51. Smith, F. D., P. M. Berman, and M. S. Mount. 1987. Charac-terization of an extracellular protease from Erwinia carotovorasubsp. carotovora strain EC14. Phytopathology 77:122.

52. Struhl, K. 1986. A rapid method for creating recombinant DNAmolecules. BioTechniques 3:452-453.

53. Szumanski, M. B. W., and S. M. Boyle. 1990. Analysis andsequence of the speB gene encoding agmatine ureohydrolase, aputrescine biosynthetic enzyme in Escherichia coli. J. Bacte-riol. 172:538-547.

54. Takagi, M., T. Imanaka, and S. Aiba. 1985. Nucleotide se-quence and promoter region for the neutral protease gene fromBacillus stearothermophilus. J. Bacteriol. 163:824-831.

55. Titani, K., M. A. Hermodson, L. H. Ericsson, K. A. Walsh, andH. Neurath. 1972. Amino acid sequence of thermolysin. Nature(London) New Biol. 238:35-37.

56. Torriani, A. 1960. Influence of inorganic phosphate in theformation of phosphatases by Escherichia coli. Biochim.Biophys. Acta 38:460-469.

57. Tseng, T.-M., and M. S. Mount. 1973. Toxicity of endopolyga-lacturonate trans-eliminase, phosphatidase and protease to po-

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

6546 KYOSTIO ET AL. J. BACTERIOL.

tato and cucumber tissue. Phytopathology 64:229-236.58. Vasantha, N, L. D. Thompson, G. Rhodes, C. Banner, J. Nagle,

nd D. FlIpula. 1984. Genes for alkaline protease and neutralprotease from Bacillus amyloliquefaciens contain a large openreading frame between the regions coding for signal sequenceand mature protein. J. Bacteriol. 159:811-819.

59. Wandersman, C., T. Andro, and Y. Bertheau. 1986. Extracellu-lar. proteases in Erwinia chrysanthemi. J. Gen. Microbiol.132:899-06.

60. Wandersman, C., P. Delepelaire, S. Letoffe, and M. Schwartz.1987. Characterization of Erwinia chrysanthemi extracellularprotease: cloning and expression of the protease genes inEscherichia coli. J. Bacteriol. 169:5046-5053.

61. Weinberg, E. D. 1985. Enzymes, nutrition, and virulence, p.1-16. In I. A. Holder (ed.), Bacterial enzymes and virulence.CRC Press, Boca Raton, Fla.

62. Welinder, K. G., and L. B. Smillie. 1971. Amino acid sequencestudies of horseradish peroxidase. II. Thermolytic peptides.Can. J. Biochem. 50:63-70.

63. Willis, J. W., and A. K. Chatterjee. 1987. Cloning and charac-terization of a gene for extracellular protease of Erwinia caroto-vora subsp. carotovora, p. 253. In E. L. Civerolo, A. Collmer,R. E. Davis, and A. G. Gillaspie (ed.), Proceedings of the 6thInternational Conference on Plant Pathogenic Bacteria, 2-7June 1985. Martinus Nijhoff Publishers, Boston.

64. Yang, M. Y., E. Ferrari, and D. J. Henner. 1984. Cloning of theneutral protease gene of Bacillus subtilis and the use of thecloned gene to create an in vitro-derived deletion mutant. J.Bacteriol. 160:15-21.

65. Yang, Z., C. L. Cramer, and G. H. Lacy. 1989. System forsimultaneous study for bacterial and plant genes in soft rot ofpotato. Mol. Plant Microbe Interact. 2:195-201.

on June 15, 2018 by guesthttp://jb.asm

.org/D

ownloaded from