JOURNAL OF BACTERIOLOGY, Jan. 1994, p. 149-1560021-9193/94/$04.00+0Copyright © 1994, American Society for Microbiology

Purification and Characterization of a Clostridium perfringens120-Kilodalton Collagenase and Nucleotide Sequence

of the Corresponding GeneOSAMU MATSUSHITA, KUNIKO YOSHIHARA, SEI-ICHI KATAYAMA,

JUNZABURO MINAMI, AND AKINOBU OKABE*Department of Microbiology, Kagawa Medical School, Kagawa 761-07, Japan

Received 16 February 1993/Accepted 18 October 1993

Clostridium perfringens type C NCIB 10662 produced various gelatinolytic enzymes with molecular massesranging from -120 to -80 kDa. A 120-kDa gelatinolytic enzyme was present in the largest quantity in theculture supernatant, and this enzyme was purified to homogeneity on the basis of sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The purified enzyme was identified as the major collagenase of theorganism, and it cleaved typical collagenase substrates such as azocoll, a synthetic substrate (4-phenylazo-benzyloxy-carbonyl-Pro-Leu-Gly-Pro-D-Arg [Pz peptide]), and a type I collagen fibril. In addition, a gene (colA)encoding a 120-kDa collagenase was cloned in Escherichia coli. Nested deletions were used to define the codingregion of colA, and this region was sequenced; from the nucleotide sequence, this gene encodes a protein of1,104 amino acids (Mr, 125,966). Furthermore, from the N-terminal amino acid sequence of the purifiedenzyme which was found in this reading frame, the molecular mass of the mature enzyme was calculated to be116,339 Da. Analysis of the primary structure of the gene product showed that the enzyme was produced witha stretch of 86 amino acids containing a putative signal sequence. Within this stretch was found PLGP, theamino acid sequence constituting the Pz peptide. This sequence may be implicated in self-processing of thecollagenase. A consensus zinc-binding sequence (HEXXH) suggested for vertebrate Zn collagenases is presentin this bacterial collagenase. Vibrio alginolyticus collagenase and Achromobacter lyticus protease I showedsignificant homology with the 120-kDa collagenase of C. perfringens, suggesting that these three enzymes areevolutionarily related.

Certain bacteria produce collagenases, proteases whichcleave native collagen in its triple helical conformation (17, 18,21, 24, 34). Bacterial collagenases possess broad substratespecificities and degrade both native and denatured collagens(28), while vertebrate collagenases preferentially cleave thenative form at a specific site (3). Much of our knowledge ofbacterial collagenases has come from studies of the enzymesproduced by Clostridium histolyticum. Seven different polypep-tides with collagenolytic activity have been demonstrated in aculture of this organism, and these are divided into two groupson the basis of substrate specificity (6-8, 25, 35). Some otherclostridia can also produce collagenase, but little is knownregarding their biochemical properties.

Clostridium perfringens is one of the collagenase-producingclostridia (36). This organism is widely distributed in nature,particularly in soil and water contaminated with feces. It alsolives in the intestinal tracts of humans and animals andsometimes causes infection, representing the most commonpathogen of clostridial histotoxic infection (12). The collage-nase produced by C. perfringens appears to be involved intissue necrosis, along with other toxins such as phospholipaseC and thiol-activated hemolysin (12, 14, 29). It also seems toplay an important role in carcass putrefaction by the sapro-phytic life style of the organism, since collagen represents themost abundant protein in an animal body (12). A collagenasewith a molecular weight of 80,000 was purified from C.

* Corresponding author. Mailing address: Department of Microbi-ology, Kagawa Medical School, 1750-1 Miki-cho, Kita-gun, Kagawa761-07, Japan. Phone: +81 (878) 98-5111, ext. 2512. Fax: +81 (878)98-7109. Electronic mail address: a64235@sakura.kudpc.kyoto-u.ap.ip.

perfringens several years ago (14), but it has not yet been fullycharacterized despite its importance as a virulence factor andas an enzyme involved in geochemical cycles.

In order to characterize the biological properties and themolecular structure of the C. perfringens collagenase, we havechosen to study the enzyme from C. perfringens type C NCIB10662. While this strain produces collagenase in a moderatequantity, it expresses extremely low levels of alpha-toxin (16),which is often a contaminant in collagenase preparations.Therefore, this strain was judged to be useful for purificationof the enzyme. This paper describes our successful purificationand characterization of the 120-kDa collagenase as well as thenucleotide sequence of the corresponding gene (colA). Fur-thermore, since certain other bacterial proteases exhibit hydro-lytic activity against denatured collagen, we suspected that theamino acid sequence of the collagenase might be similar tothose of other bacterial proteases. We have therefore investi-gated similarities between these enzymes at the protein se-

quence level.

MATERIALS AND METHODS

Bacterial strains, plasmids, and bacteriophage. C. perfrin-gens type C NCIB 10662 (16) was used throughout this study.Escherichia coli LE392 (22) was used to grow bacteriophageX2001 (15). Recombinant X phage were grown on E. coli P2392(a P2 lysogen of LE392). A plasmid vector, pUC18 (41), andthe host strain E. coli DH5a (2, 11) were used for allrecombinant plasmid experiments.Media and culture conditions. C. perfringens was precul-

tured in cooked meat medium (Nissui, Tokyo, Japan) and was

grown in GAM broth (Nissui) unless otherwise stated.

149

Vol. 176, No. 1

on February 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

150 MATSUSHITA ET AL.

NZCYM medium (Gibco, Paisley, United Kingdom) (22) wasused to grow X2001 phage. Recombinant phage were plated onNZCYM agar plates containing 22 g of NZCYM broth powder(Gibco), 2 g of maltose, and 15 g of agar per liter. TransformedE. coli cells were selected on LB plates containing 20 g of LBbroth base (Gibco) and 15 g of agar per liter, supplementedwith 100 jig of ampicillin (Sigma Chemical Co., St. Louis, Mo.)per ml, 0.2 mM isopropyl-3-D-thiogalactopyranoside (IPTG)(Sigma), and 0.004% 5-bromo-4-chloro-3-indolyl-,-D-galacto-pyranoside (X-Gal) (Sigma).Assay for collagenase. Protein concentrations were deter-

mined by using the Pierce bicinchoninic acid protein assayreagent (32) with bovine serum albumin as a standard. Colla-genase activity was assayed with azocoll by the method ofKameyama and Akama (13) with slight modifications. Thestandard assay conditions were as follows. Azocoll (Sigma)powder was washed once with and resuspended in 0.2 M boricacid-borate (pH 7.2)-0.15 M NaCl to give a concentration of 5mg of azocoll per ml. Four hundred microliters of this assaymixture was taken into a 1.5-ml microcentrifuge tube, and 5 to10 pd of enzyme solution was added. The tube then was placedhorizontally and shaken at 37°C for 1 h. The supernatant wasobtained by centrifugation at 15,000 x g for 15 min. The A520of the supernatant was read. One unit of enzyme activity isdefined as the amount of enzyme which causes an increase of1.0A520 unit per min. In some experiments collagenase activitywas assayed by using 4-phenylazobenzyloxy-carbonyl-Pro-Leu-Gly-Pro-D-Arg (Pz peptide) as described by Wunsch andHeidrich (40).Zymography. Gelatin zymography was carried out essen-

tially by the method of Wilson et al. (37). GAM broth wasinoculated with a 5% volume of C. perfringens preculture, andthe cells were grown at 37°C without shaking. Portions of theculture were withdrawn after 2, 4, 6, and 8 h of incubation andchilled on ice, and the cells were removed by centrifugation.The samples were mixed with nonreducing loading dye andapplied to a 10% sodium dodecyl sulfate (SDS)-polyacryl-amide gel, which had been copolymerized with gelatin (WakoPure Chemicals, Osaka, Japan). After running, SDS wasremoved by washing the gel with 2.5% Triton X-100-100 mMNaCl, and then the gel was incubated overnight at 37°C in 50mM Tris-HCl (pH 7.5)-10 mM CaCl2 and stained with Coo-massie brilliant blue R-250.The activity of purified collagenase against collagen fibril

was examined by the method of Birkedal-Hansen and Tayler(4), using acid-soluble type I collagen from calf skin (Sigma).Four micrograms of the purified collagenase was loaded on a

conventional SDS-10% polyacrylamide gel under nonreducingconditions. After removal of SDS, the gel was briefly rinsedwith phosphate-buffered saline (pH 7.4) containing 1 mMCaCl2 and 1 ,uM ZnCl2, and then it was placed on a collagenfibril film and incubated overnight at 37°C. The film was thenstained with Coomassie brilliant blue R-250 in order to visu-alize an unstained band on a blue background.

Purification of collagenase. C. perfringens was grown incooked meat medium and then in GAM broth at 37°C for 12h. The final preculture was diluted 100-fold into 10 liters ofGAM broth and incubated overnight at 25°C. Cells wereremoved by centrifugation at 15,000 x g for 30 min at 4°C, andthe culture supernatant was concentrated to approximately 700ml by ultrafiltration using a Pellicon apparatus and a polysul-fone membrane (PTGC [5 ft2]; Millipore Co., Bedford, Mass.).Ammonium sulfate was then added to 60% saturation, and theprecipitate was collected by centrifugation at 20,000 x g for 60min at 4°C. The pellet was dissolved in 50 ml of 50 mMTris-HCl buffer (pH 7.5)-5 mM CaCl2, and the fraction was

dialyzed against 1 liter of the same buffer with two changes.The material was then applied to a DEAE-Sephadex A-25column (2.2 by 50.0 cm). Proteins were eluted with a lineargradient from 0 to 500 mM NaCl in the starting buffer (over 4bed volumes). Collagenase activity eluted in a single peak atapproximately 100 mM NaCl.The pooled collagenase fractions were concentrated by

ultrafiltration with a dialysis tubing to approximately 10 ml.After dialysis against 50 mM Tris-HCl (pH 7.5)-5 mM CaCl2,a 2-ml aliquot of the sample was applied to an ion-exchangefast protein liquid chromatography column (MonoQ; bedvolume, 1 ml; Pharmacia LKB Biotechnology, Uppsala, Swe-den) which was preequilibrated with 100 mM Tris-HCl (pH7.0)-5 mM CaCl2. Proteins were eluted by a 20-ml lineargradient from 0 to 0.6 M NaCl in 100 mM Tris-HCl (pH 7.0)-5mM CaCl2. Collagenase activity eluted at 210 mM NaCl. Thepooled collagenase fractions were concentrated by ultrafiltra-tion with a Centricon 30 (Amicon, Danvers, Mass.) to approx-imately 500 pl. A 100-,ul aliquot of the sample was applied toa size exclusion column (Superose 12; bed volume, 25 ml;Pharmacia). The column was run in 100 mM Tris-HCI (pH7.0)-5 mM CaCl2, and collagenase activity eluted at 0.44 bedvolume.

Determination of molecular weight and N-terminal aminoacid sequence. Estimations of the molecular weight of thecollagenase by gel filtration chromatography were made byusing the Superose 12 column with detection at 280 nm.Approximately 30 ,ug of purified enzyme in 50 pl of 100 mMTris-HCl (pH 7.0)-5 mM CaCl2 was run in the system, usingthe same buffer at a flow rate of 0.5 ml/min. The followingmolecular weight standards were used at concentrations of 2 to5 mg/ml; cytochrome c (Mr, 12,400), ovalbumin (Mr, 43,000),bovine albumin (Mr, 67,000), and bovine gamma globulin (Mr,150,000).

Estimation of the molecular weight of the purified enzymeby SDS-polyacrylamide gel electrophoresis (SDS-PAGE) wasperformed by the method of Laemmli (19). A mixture ofapproximately 0.5 ,ug of purified enzyme, 62.5 mM Tris-HCl(pH 6.8), 1% SDS, 1% 2-mercaptoethanol, 0.001% bromphe-nol blue, and 10% glycerol was heated to 100°C for 3 min andthen electrophoresed on a 10% polyacrylamide gel. The gelwas stained with 0.2% Coomassie brilliant blue R by theconventional method. Pharmacia high-molecular-weight stan-dards (Mrs, 53,000,76,000, 116,000, 170,000, and 212,000) wereused as standards.The N-terminal amino acid sequence of the purified enzyme

was determined on a protein sequencer (model 476A; AppliedBiosystems, Foster City, Calif.).

Effects of pH, temperature, and metal ions on enzymeactivity. The effect of pH was examined by using azocoll underthe standard assay conditions, except for the buffer. Thebuffers used were 0.2 M sodium phosphate (pH 4.5 to 8.5) or0.2 M sodium borate (pH 6.8 to 9) containing 0.15 M NaCl.The effect of temperature was also examined under thestandard assay conditions by varying the temperature (20 to57°C). For examination of metal ion specificity, divalent metalions present in the enzyme solution and the reaction mixturewere removed prior to the assay. To the purified enzyme in 100mM Tris-HCl (pH 7.0)-5 mM CaCl2 was added EDTA at afinal concentration of 10 mM (5 mM excess). Removal of themetal-chelator complex with Sephadex G-10 was carried out bythe spin-column method of Salvesen and Nagase (30). Divalentmetal ions which might be present in the Pz peptide assaymixture were removed by Na+-charged chelating Sepharose(Pharmacia) column chromatography. CaCl2, MgCl2, andZnCl2 were added to the metal ion-free assay mixture at

J. BAcrERIOL.

on February 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

C. PERFRINGENS COLLAGENASE GENE 151

various concentrations. The reaction was initiated by addingthe metal-free enzyme solution to the mixture.DNA manipulations. Restriction endonucleases were pur-

chased from Takara Shuzo (Kyoto, Japan), Toyobo (Osaka,Japan), and New England Biolabs (Beverly, Mass.). Klenowenzyme and the DNA ligation kit were products of TakaraShuzo. All recombinant DNA procedures were carried out asdescribed by Maniatis et al. (22).

Construction of genomic library of C. perfringens. C. perfrin-gens cells were harvested from a 1-liter culture by centrifuga-tion at 8,000 x g for 20 min at 4°C. The cell pellet wasresuspended in 60 ml of 100 mM Tris-HCl (pH 8.0)-10 mMEDTA. A 1/10 volume of a 2-mg/ml lysozyme solution wasadded; this was followed by incubation at 37°C for 5 min. A1/10 volume of 20% SDS was then added to the suspension,and cells were lysed by further incubation at 37°C for 5 min. Anequal volume of ice-cold phenol saturated with 100 mMTris-HCl (pH 8.0) was added to the sample, and the mixturewas gently shaken at 4°C for 30 min. The sample was extractedonce with phenol-chloroform at 37°C and once with chloro-form. The aqueous phase was recovered and treated withRNase A (10 ,ug/ml; Sigma) for 30 min at 37°C, and then theprotein was removed by phenol-chloroform extractions as

described above. The DNA was precipitated with 2 volumes ofethanol and dissolved in 10 mM Tris-HCl (pH 8.0)-i mMEDTA.Nine hundred micrograms of the above-described DNA was

digested with 60 U of MboI at 37°C for 2 min, and then theDNA was precipitated with ethanol and DNA fragments of 9to 20 kb were isolated by sucrose density gradient centrifuga-tion (22). The sticky ends of both the size-fractionated insertDNA and XhoI-digested X2001 were partially filled in (toprevent self-ligation) by treating each with the Klenow frag-ment in the presence of dATP and dGTP and of dCTP anddTTP, respectively. The DNAs were then ligated by T4 DNAligase and packaged in vitro.

Preparation of anticollagenase antiserum and cloning of thecolA gene. Collagenase was purified from the culture superna-tant of C. perfringens type A NCTC 8237 as described byKameyama and Akama (14). Rabbits were immunized withthis purified collagenase by a single subcutaneous injection incomplete adjuvant followed by a booster injection in incom-plete adjuvant over 42 days. The resulting antiserum wastreated with the E. coli extract of the ProtBlot kit (Promega,Cleveland, Ohio) to reduce the background on immunoscreen-ing.

Plaques of recombinant phage were grown on E. coli P2392host cells and lifted on nitrocellulose membranes (BA85;Schleicher & Schuell, Inc., Keene, N.H.). The filters wereprobed with the anticollagenase rabbit antiserum by using theProtBlot kit (Promega) according to the supplier's instructions.

Nucleotide sequencing. DNA was sequenced by the dideoxychain termination method (31), using modified T7 DNA poly-merase (Sequenase; United States Biochemicals, Cleveland,Ohio). Deoxyadenosine 5'-a-[35S]thiotriphosphate (> 1,000 Ci/mmol) was purchased from New England Nuclear (Wilming-ton, Del.). Plasmid DNAs were prepared as described byBirnboim and Doly (5), purified by CsCl density gradientcentrifugation with ethidium bromide (33), denatured, andneutralized for use as sequencing templates. Synthetic oligo-nucleotides for the sequencing primers were prepared on aDNA synthesizer (model 381; Applied Biosystems).

Southern hybridization. A nonradioactive DNA-labelingand detection kit (Genius; Boehringer Mannheim Biochemi-cals, Indianapolis, Ind.) was used for hybridization experi-ments. To prepare a DNA probe, pKY3135 plasmid DNA was

completely digested with XbaI and SmaI, and the 1.61-kbinsert (see Fig. 3) was isolated by agarose gel electrophoresisand electroelution. The DNA fragment was purified by usingthe GeneClean kit (Bio 101 Inc., La Jolla, Calif.), and approx-imately 1 ,ug was labeled with digoxigenin-11-dUTP by randompriming as described by the supplier (Boehringer Mannheim).C. perfringens chromosomal DNA was digested with BglII,HincII, or PvuII at 37°C overnight, and the digests (5 ,ug each)were applied to a 0.8% agarose gel and electrophoresed at 100V for 3 h. The gel was soaked in 0.5 M NaOH-1.5 M NaCl todenature the DNA and then neutralized with 0.5 M Tris-HCl(pH 7.0)-3 M NaCl. DNAs were transferred onto a sheet ofnylon membrane (Hybond-N; Amersham) with 20 x SSC (1 xSSC is 150 mM NaCl plus 15 mM sodium citrate) by a standardmethod (22). Hybridization and immunological detection wereperformed as described by the supplier, except that a chemi-luminescent dye (Lumi-Phos 530; Lumigen, Detroit, Mich.)was used as the substrate for alkaline phosphatase. Themembrane was sealed in a plastic bag and placed on an X-rayfilm for 1.5 min to record the results.Amino acid sequence similarity search. A data base search

for similar protein sequences was carried out by using theBLAST mail server (1) at the National Center for Biotechnol-ogy Information, National Library of Medicine, National In-stitutes of Health (Bethesda, Md.). The translated amino acidsequence of the entire open reading frame of the collagenasegene was used as the query sequence, and this was searched foragainst the nonredundant protein data base, including theSWISS-PROT, PIR, GenPept, and GenPept update databases. The default parameters of the program were used forthis similarity search.

Nucleotide sequence accession number. The DNA sequenceof the C. perfringens colA gene is available from GenBank/EMBLIDDBJ data bases under accession number D13791.

RESULTS

Gelatinolytic activities in the culture supernatant. Zymog-raphy revealed the presence of multiple gelatinases of differentsizes in the culture supernatants of C. perfringens (Fig. 1). Inthe mid-logarithmic phase (2 h), gelatinolytic activity wasdetected as one major band corresponding to a molecular massof -120 kDa, with three very faint bands at -110, -102, and-80 kDa. While the intensities of these three minor bandsincreased with incubation time, the 120-kDa band remainedthe major band. However, during stationary phase (8 h), theintensities of all four bands decreased.

Purification and characterization of the 120-kDa enzyme.The major 120-kDa enzyme was purified to homogeneity fromthe C. perfringens culture supernatant (Fig. 2A). The molecularmass of the collagenase was estimated to be 120 kDa bySDS-PAGE. The same value was obtained from analysis by gelfiltration on Superose 12, indicating that the 120-kDa collag-enase is in a monomeric form. From 10 liters of the culturesupernatant, 1.6 mg of purified enzyme was obtained. Toexamine the activity of this enzyme on a type I collagen fibril,zymography using this substrate was carried out. As shown inFig. 2B, the hydrolytic activity was visible as a single band, andthe location of this band coincided with that of the 120-kDaenzyme after staining of the polyacrylamide gel. From theseresults, we conclude that the type I collagen fibril was cleavedonly by the 120-kDa enzyme in the absence of other proteasesand that the purified enzyme was indeed a collagenase.The specific activity of the purified enzyme was 12.3 U/mg of

protein when assayed for azocoll hydrolytic activity under thestandard assay conditions and 2.25 ,umol/min/mg of protein

VOL. 176, 1994

on February 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

152 MATSUSHITA ET AL.

1 2 3 4pKY3134

---5kL: -

(XhoUIMbol) Xi Bg9 I(kDa)

94 -

67 -

43 -

FIG. 1. Zymogram of gelatinolytic activities in culture supernatantsof C. perfringens. Portions of the C. perfringens culture were removedafter 2 (lane 1), 4 (lane 2), 6 (lane 3), and 8 (lane 4) h. Fifteenmicroliters of each culture supernatant was applied to an SDS-polyacrylamide gel copolymerized with gelatin. Numbers on the leftare molecular masses (in kilodaltons) of the markers.

when assayed for Pz peptide. The effect of pH on azocollhydrolytic activity was assayed in the borate and phosphatebuffers. Enzyme activity was detectable over a broad range,from pH 4.5 to 9. The maximal enzyme activity in the boratebuffer was attained at pH 7.2 and that in the phosphate bufferwas at pH 7.0, the former being 1.7-fold higher than the latter.The optimal temperature for azocoll hydrolytic activity was42°C, and at higher temperatures the activity fell, probablybecause of thermal instability of the enzyme. Like the collag-enase of C. histolyticum, the 120-kDa collagenase appears torequire zinc for hydrolytic activity, since the addition of 1 mM1,10-phenanthroline to the assay mixture completely abolished

A B1 2 (kDa)

- 212- 170

- - -116-- 76

- - 53

FIG. 2. Analysis of the purified collagenase by SDS-PAGE andzymography. (A) SDS-PAGE. Lane 1, purified collagenase (0.5 ,ug);lane 2, molecular mass markers. Numbers on the right are molecularmasses (in kilodaltons) of the markers. (B) Zymography. Four micro-grams of the purified collagenase was loaded; this was followed byzymography using soluble collagen (type I) as the substrate.

nIWD/KY313 . ; I.

Azocoll hydrolysis;(-)H(+)

(MboVXhod)

Hincl Hin Pst1 kb

pKY361 1kb

pKY3135 _ir. * IpKY3136

FIG. 3. Partial restriction map of C perfringens DNA fragmentsencoding collagenase and subclones used for nucleotide sequencing. Apartial restriction map of the pKY313 insert is shown in the middle.The collagenase activities of two subclones (pKY3134 and pKY3132)are indicated by (+) and (-). Various deletions were introduced fromthe left end of the pKY3132 insert, resulting in loss of the enzymaticactivity when the deletion size exceeds approximately 2.5 kb (arrow).Three subclones were used to construct sequencing templates. Dele-tions of various sizes were introduced from the end of each insertindicated by the arrowheads.

the activity against azocoll, even in the presence of excessCaCl2 (10 mM). In order to confirm this, divalent metal ionspresent in the enzyme and substrate solutions were removed,and metal ions such as Ca2 , Mg2 , and Zn2+ were added tothe reaction mixture. No appreciable activity was observedwhen divalent metal ions were not added. The addition ofZn2+ (30 ,uM) resulted in reactivation of the enzyme activity tothe highest level. Ca2+ also reactivated it to a similar extent,but a much higher Ca2+ concentration (.1 mM) was requiredin order for activity to reach the maximal level, which corre-sponded to 90% of the maximal level attained by addition ofZn2+. The maximal reactivation attained by addition of Mg2"was only 30% of that attained by addition of Zn2+. Highconcentrations (.100 ,uM) of Zn2+ inhibited the enzymeactivity, which is consistent with the finding reported forzinc-dependent metalloproteases (30). These results clearlyindicate that the 120-kDa collagenase is a zinc protease. TheN-terminal amino acid sequence of the purified enzyme wasdetermined to be Ala-Arg-Asn-Asn-Lys-Ile-Tyr-Thr-Phe-Asp-Glu-Leu-Asn-Arg-Met-Asn-Tyr-Ser-Asp-Leu-Val-Glu-Leu-Ile-Lys-Thr-Ile by Edman degradation.

Cloning of the coU gene. A genomic library of C. perfringensDNA constructed in X2001 was immunologically screened fora clone expressing a collagenase gene. The antiserum raisedagainst collagenase purified from a C. perfringens type A strainwas shown to cross-react with the collagenase from the type Cstrain by Western blotting (immunoblotting) (data not shown),and was used for the screening. Approximately 4,000 recom-binant phage plaques were examined, and approximately 60plaques were reactive against the antiserum. Four of theseimmunologically positive plaques were randomly selected andpurified four times by using E. coli P2392 host cells, whichinhibits the growth of nonrecombinant phage. Lysates of thepurified recombinant phages were examined for collagenolyticactivity by using azocoll and Pz peptide. Three recombinantsshowed strong activities on both of the substrates, and oneshowed weak activities. Xcol3, one of the former clones, wasused for further analyses.

Restriction analysis showed that the Xcol3 recombinantphage contained a 12.9-kb MboI insert. The insert was excisedfrom the recombinant with XbaI and ligated into the XbaI siteof pUC18. Two different groups of recombinant plasmids wereobtained, one carrying the entire 12.9-kb insert (Fig. 3,

J. BAcrERIOL.

on February 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

C. PERFRINGENS COLLAGENASE GENE 153

10 20 30 40 50 60 70 80 90 100AACAATCCTAAATACC1 _ T_GTGA T

110 120 130 140 150 160 170 180 190 200

210 220 230 240 250 260 270 280 290 300

310 320 330 340 350 360 370 380 390 400A0=CI_TAAATATTAmAAGAT~~~AAG-~A -A T TTAA ATA

410 420 430 440 450 460 470 480 Aul 500ACAAAACTGAAGTTAGTTGAAAGATGGTCATC AAAAT K L K L V E R W S A T F T L A A F L F N S S F K V F A A D K K V510 520 530 540 550 560 570 580 590 600

TAGAGAATAGTAATAATGG;ACAGATTACTAGAGAGATrAATGCTGTAACAACGATATAGGTGTCGCxATGCE N S N N G 0 I T R E N AD I SKTELNNEVATDNN U610 620 630 640 OaiII 650 660 670 680 690 700

IIIIIIIIIIIIIIIENISITA_P710 720 730 740 750 760 770 780 790 800s r E N V P D L F N F NOD G SY T F F S N R D R V O A Y G

8sl 810 820 830 840 850 860 870 880 890 900LESR Y A DD DK GI PT L V E F L R A G Y Y L G F Y N

910 920 930 940 950 960 970 980 990 1000TAAACAATTATCATACCTAATACACCACGTrA AAAT AATGM TGTACCAGCTATGAGGAT AATATATA -1TIMAGATrATTK 0 L S Y L N T P 0 L K N E C L P A M K A 0 Y N S N F R L G T K1010 1020 1030 1040 1050 1060 1070 1080 1090 1100

AO0 D G V V E A L G R L I G N A S AOD P E V N N C Y V L S D F K1110 1120 1130 1140 1150 1160 1170 1180 1190 1200AAGATAATATAGATAAGTATGGTTC TATATCAACCTTATAATATTGAAT TTATACTAATAGTAATATACAD N D K Y G S NYSK NAVFNL K I DYYT N S V Y N1210 1220 1230 1240 1250 1260 1270 1280 1290 1300

T K G Y D A K N T E F Y N R I D P YHM E R L E S L C T G D K L N1310 1320 1330 1340 1350 1360 1370 1380 1390 1400ATGATTC19T0ATTCCTTATATTACAC80GTAGMTTGGG TAA01T1A68AACCTMTATCT I-I-1-1-GAA AGAN D N A W L V N N A L Y Y T G R V G K F R E D P S S 0 R A L E R A1410 1420 1430 1440 1450 1460 1470 1480 1490 1500KEY_ YLSYOYIEAmA NDLOL INiIF GTGGCAAATCACGGAATGTAA0VKEYPYL SY 0Y IE AA ND LD L N F G G K N S S G N D D1510 1520 1530 1540 1550 1560 1570 1580 1590 1600nTCMATMATMAMAATGCA.AG0GAMATATCr1CCAAAAACATATAC0IGTrrAT0AT0TAAA.GA1--08TAAAAGTGT006TMAAGAACFNKIKADAR EK YL PK TY TF DD G K F V V K A G D K V T1610 1620 1630 1640 1650 1660 1670 1680 1690 1700GAAGGAATAAAAGATTATATTGOOCTTGAGCCATAGGGATCAAGTAGmCE E K K R L Y W A S K E V K A 0 F M R V V Q N D K A L E E G N P D1710 1720 1730 1740 1750 1760 1770 1780 1790 1800ATGATATTMAACTGITTGTrAMATAACTCACCAGMA6AAGTATAAGTrTAATCGTATATAAATGGAIAGTACTGATAATGGTGGTTATATATTGAD L T V V Y N S P E E Y K L N R N G F S T D N G G Y E1810 1820 1830 1840 1850 1860 1870 1880 1890 1900

AAACATAGGAACMCTrACTAAAAATGAATATCTAGAATATTTCAGN G T F F T YERTPEESI YTLEELFRH E F T H Y L 0 G1910 1920 1930 1940 1950 1960 1970 1980 1990 2000RYVVPGV wG 0G EF YQ EG VL T W Y E E G T A E F F A G S T

2010 2020 2030 2040 2050 2060 2070 2080 2090 2100CTAGAACTGATGGAAT AAATCAGTTA AGCTTACGATAGAAAT MTAGAATGTCM ATATGGTGTATrACATGCTAAATAR T D G K P R K S V T 0 G L A Y D R N N R M S L Y G V L H A K Y

2110 2120 2130 2140 2150 2160 2170 2180 2190 2200TGGCTCATGGGAMCTATAATTATGGAMGCTCTACAATCTGTACAACAATAACATGGGAATGTTAATAAGATG WTACAAATCTAAAGAATG S W D F Y N Y G F A L S N Y M Y N N N M G M F N K M T N Y K N

2210 2220 2230 2240. 2250 2260 2270 2280 2290 2300AATGATGTATCTGGTTATAAAGATTATATTGCATCAATGAGTAGTGATTACGGATTAAATGATAAATATCAAGACTATATGG,ATTCMATTAAACAATAN D V S G Y K D Y A S M S S D Y G L N D K Y 0 D Y M D S L LNN

2310 2320 2330 2340 2350 2360 2370 2380 2390 2400TTGATAACTTAGATGT1CCTAG 1CA0ATGA0TAT0TAMT00ACAT0M0CTA68TATMAT0AMTAACTMT0ACATAATTTCAAD N L D V P LVSDEYVN HEAK NE T N D K E V S N

25101 2420 2430 2440 2450 2460 2470 2480 2490 2500TATAAAAGATCM CTAGTAATGTTGAAAAGTCTCAATTCMACrACGATATGA TATGTGATK D L S S N V E K S 0 F F T T Y D M R G T Y V G G R S 0 G E E N

2510 2520 2530 2540 2550 2560 2570 2580 2590 2600GACTGGAAAGATATGAATTCTAAGTTAAATGATAIAAGGTGAGTAAACGTCGAAmTACD W K D M N S K L N D VLKELSKKS NGYKTVTAY F V N H

2610 2620 2630 2640 2650 2660 2670 2680 2690 2700T011ATAAGAGATGGATTACTATMATGATGTrGTATrC1AATGA ATACAGATACAAATACTGATGTTCATGTTAATAAAGAGCCTAAGGCK V D E N G N Y V E F H G N T D T N T D V H V N K E P K A2710 2720 2730 2740 2750 2760 2770 2780 2790 2800

TGTTATAAAA TA.8AAATCTATAG0CTAAGAAACTGATGA CA4GAGTCAAAGGATATGAAGT AAATCAAAGCTTATGAATGGV K S D S S V V E E E N F D G T E S K D E D G E K A Y E W2810 2820., 2830 2840 2850 2860 2870 2880 2890 2900

GACMTGGAGATGGAGAAAAATCTAATGAGGCTAAAGCACTCATAAGTATAATAAAACTGGAGAATATGAAGTAAAATTAACAGTTACAGATAATAACGD F G D G E K S N E A K A T H K Y N K T G E Y E V K L T V T D N N G2910 2920 2930 2940 2950 2960 2970 2980 2990 3000

01I N T E S K K K V V EOD K P V E V N E S E P N N D F E K A N3010 3020 3030 3040 3050 3060 3070 3080 3090 3100

CCAAATAGCTAAATCTAA TAGTrTAGTAAGGGTACl ATCA IATAMTA UATG TAGC T MAIAAT00-1AQ A K S N M L V K G T L S E E D Y S D K Y Y F D V A K K G N V K3110 3120 3130 3140 3150 3160 3170 3180 3190 3200ATCACTCTTAATAAMAAATrCAGTAGGAATAA CTTGGACACMAT AAGGGGCTAAACAATrATGTM||ATATGCAACTGGAAATGATGGAAT L N N L N S V G T W T L Y K E G D L N N Y V L Y A T G N D G T3210 3220 3230 3240 3250 3260 3270 3280 3290 3300CAGAATTAAAGGGTGAAAAGACMAGAGCCTGGAAGATACTACTrAAGTGTATATACTrATGATAATCAATCAGGAGCTrACACAGTAAATGTAAAAGGE L K G E K T L E P G R Y Y L S V Y T Y D N 0 S G A Y T V N V K G3310 3320 3330 3340 3350 3360 3370 3380 3390 3400AAACCTrAAAAATCGAAGTTAAAGA SAAAAGTGCTATAAAAGA AGTTGAAMTAACAATGATM1 GATAAAGCTATGkAGGTAGACAGTAATAGC

N L K N E V K E T E K D A K E V E N N N D F D K A M K V D S N S3410 3420 3430 3440 3450 3460 3470 3480 3490 3500

K~~~~~~~~ ~I 0 N P SD L N V V E N L D N3510 3520 3P1uI 3540 3550 3560 3570 3580 3590 3600

ATATAAAAATGAACTGGTTATrATATTCAGCTGATGATTAAGTAACTATGTGGATTACGCTAATGCAGATGGAAATAAATTAAGTAACACTTGTAAGTTK M N W L L Y S A D D L S N Y V D Y A N A D G N K L S N T C K L3610 3620 3630 3640 3650 3660 3670 3680 3690 3700AAATCCAGGTAAATATrACTTATGTGMATCAAMGM.ACTCAGGTACTGAArCCGAATAAACATAAATGrAA

N P G K Y Y L C V Y 0 F E N S G T G N Y T V N L Q N K *

3710 3720 3730 3740 3750 3760 3770 3780 3790 3800CAAATATAMTAOAOOCTGTACAAmAOTAACAATATACM1TrATA AATAC IATMrlCTATGTIMAAGAGGGATGATACAGCCTrAAAnIIM C

3810 3820 3830 3840 3850 3860 3870 3880 3890 3900MAAAATAAGTM1 ATGT llllTMATATGACTTCAACGTTCAGGAACATGm TTTA G3910 3920 3930 3940 3950 3960 3970 3980 3990 4000TATAGCAATAAGGACAATCTMCTGTAGCTTCAAC1AACTTCTTC CTTAGTGAATrMAAGCTTGM--AMGTIMI-I-ATAACTATAA4010 4020 4030 4040 4050 4060 4070 4080 4090 4100TATAGAAAATGCT1/1 1 1GA 1A AAAATAGTCC TTAA GTAGTTl GAGCTTCTTTATGAGTTA4110 4120 4130 Hincl

AAT'GTTGTCCACTTAAGTATGTATAAGTTAGAGAAGTC

FIG. 4. DNA sequence of the C. perfringens collagenase gene (coLA). An inverted repeat close to the termination codon is indicated by arrows.Some of the restriction sites are also shown. A deduced amino acid sequence which is the same as that of a synthetic peptide substrate forcollagenases is indicated by a black background. The N-terminal amino acid sequence of the collagenase is shaded.

pKY313) and the other carrying a 2.7-kb XbaI fragment (Fig.3, pKY361). The latter fragment was placed between the XbaIsite of polylinker sequence of the vector and the XbaI sitewithin the 12.9-kb insert by partial physical mapping. Twosubclones carrying the leftmost 7.6-kb XbaI (within the vectorpolylinker site)-PstI fragment and the 4.9-kb XbaI (within the12.9-kb insert)-PstI fragment were constructed from pKY313and were designated pKY3132 and pKY3134, respectively.While E. coli transformed with pKY3132 showed hydrolyticactivity against azocoll, that transformed with pKY3134showed no activity. Nested deletions were introduced from theleft end of pKY3132 to localize the left end of the colA gene.When the deletion exceeded approximately 2.5 kb from the leftend of the insert, azocoll-hydrolyzing activity was lost, indicat-ing that the left end of the colA gene is located 2.5 kb awayfrom the left end of pKY3132. The right end of the colA genewas not determined but could be placed at about 0.2 kb to theleft of the HinclI site on the basis of a maximum nucleotidelength of the colA gene encoding the 120-kDa polypeptide.

Nucleotide sequencing of the coU gene. Three plasmids,pKY361, pKY3135, and pKY3136, and their derivatives pro-duced by nested deletions of the three plasmid (Fig. 3) weresequenced by using the strategy outlined in Fig. 3. Thenucleotide sequences of some regions were determined byusing synthetic oligonucleotides in order to close gaps in thesequence or to confirm ambiguous nucleotides. The sequences

thus obtained were aligned by their overlaps to form a singlecontig (Fig. 4). The open reading frame encoding the colAgene extends from 0.43 kb to the left of the XbaI site to 0.46 kbto the left of the HinclI site, as expected. This open readingframe encodes a protein with a molecular mass of 125,966 Da.The N-terminal amino acid sequence of the purified 120-kDacollagenase is the same as that predicted from the nucleotidesequence starting at nucleotide 629, and the open readingframe continues to nucleotide 3685, which is a TAA termina-tion codon (Fig. 4). The molecular mass of the proteinpredicted from this reading frame is 116,339 Da, which is ingood agreement with that determined by SDS-PAGE (120kDa) for the purified collagenase of C. perfringens. The ATGinitiation codon at nucleotide 371 is preceded by a consensusribosome binding sequence site (AGGAGG) with a 9-bp gap,which is complementary to the 3' end of C. perfringens 16SrRNA (10). Since no other ATG or GTG initiation codons arepresent in this region, it seems certain that the collagenase isinitiated at the ATG at nucleotide 371.

Southern hybridization using a coU gene probe. In order toexamine whether the coL4 gene is duplicated, a Southernhybridization was carried out with a 1.61-kb XbaI-BglII frag-ment as a probe (Fig. 3). Chromosomal DNA from C. perfrin-gens was digested with BglII, HincII, or PvuII and separated ona 0.8% agarose gel. For the BglII and PvuII digests, hybridizedbands were 1.8 and 3.0 kb long, respectively (Fig. 5), which

VOL. 176, 1994

on February 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

154 MATSUSHITA ET AL.

1 2 3 4

(kb)

23.1 -

9.4-6.6-4.4-

No

Q-

2.3 -2.0 -

FIG. 5. Southern hybridization of C. perfringens DNA digested withBglII, HincII, or PvuII. Lanes: 1, size markers; 2, BglIH digest (5 gig); 3,HincII digest (5 ,ug); 4, PvuII digest (5 ,ug). Numbers on the left aresizes (in kilobases) of the markers. The 1.61-kbXbal-SmaI fragment ofpKY3135 was used as a collagenase gene probe after labeling withdigoxigenin-dUTP by random priming.

coincided with a 1,766-bp BglII fragment (nucleotides 642 to2407) and a 3,043-bp PvuII fragment (nucleotides 486 to 3528).A 5.4-kb fragment hybridized in the HincIl digest, in goodagreement with the predicted size of the HincIl fragmentestimated from the restriction map. In all cases only one bandwas detected for a given digestion, suggesting that the 120-kDacollagenase is encoded by a single colA gene.

DISCUSSIONZymography using gelatin revealed the presence of a major

120-kDa gelatinolytic enzyme and other minor ones in theculture supernatant of the C. perfringens type C strain. Duringthe initial phases of this work, it became clear that a gelatino-lytic enzyme fraction partially purified by the method describedfor purification of C. perfringens collagenase (14) containedpredominantly enzymes with molecular masses ranging from110 to 80 kDa and only a trace amount of the 120-kDa enzyme.We therefore examined both the culture conditions and thepurification procedure with the goal of successfully obtainingsufficient quantities of the 120-kDa enzyme. We identified twomodifications which were critical to isolating the 120-kDaenzyme. First, we found that large amounts of the 120-kDaenzyme were present in the culture supernatants even afterprolonged incubation if the cells were grown at 25 rather than37°C. In addition, for certain Zn proteases, Ca2+ has beenshown to inhibit their autocatalysis (23), and the C. histolyti-cum collagenase has been successfully purified by the addition

of Ca2" to the buffers used (28). We therefore carried out allthe purification steps in the presence of 5 mM CaCl2. Thecombination of these improvements allowed purification of the120-kDa enzyme, which cleaved type I collagen fibril and wasidentified as a collagenase. The discrepancy between themolecular mass of 120 kDa that we obtained and the 80-kDavalue reported by other groups (14, 39) for the collagenase ofa C. perfringens type A strain does not seem to arise from adifference in the type of strain. In our hands, zymographyindicated the presence of both 120- and 80-kDa gelatinolyticenzymes in the culture supernatants of both the type A andtype C strains (data not shown). One possible explanation isthat the 120-kDa enzyme may be proteolyzed to give the80-kDa enzyme.

Inspection of the N-terminal amino acid sequence startingwith a potential translational initiation codon (ATG) reveals atypical signal peptide, which consists of a positively chargedstretch (MKKNLKRGELTKLKLVER), a hydrophobic core(WSATFTLAAFILF), and a polar region (NSSFKVFAA).However, Ala-87, the N-terminal amino acid of the matureenzyme, is separated by 46 amino acid residues from theputative signal peptide. Within these amino acid residuesoccurs a unique amino acid sequence (PLGP), the samesequence as that of a synthetic peptide substrate for bacterialcollagenases. The C. histolyticum collagenase is known tocleave a peptide bond between leucine and glycine residues inthe amino acid sequence indicated above, which is probablythe case for the 120-kDa C. perfiingens collagenase. This maysuggest that the 120-kDa enzyme can do self-processing,triggering subsequent removal of the following octapeptide byanother protease.While vertebrate collagenases cleave native type I collagen

only at a specific site and other proteases are necessary forcomplete degradation, bacterial collagenases can degrade col-lagen in the absence of other proteases. The latter enzymesseem to be related to protease, and one might expect sequenceand structural similarities between bacterial collagenases andproteases. We therefore investigated this possibility by search-ing for proteins with amino acid sequences similar to that ofthe C. perfringens collagenase, using the BLAST algorithm.This uses an ungapped alignment algorithm to compare aquery with all protein sequences deposited in the data bases.This search identified two proteins, Vibrio alginolyticus collag-enase (34) and Achromobacter lyticus protease I (26), whichshow sequence similarity to the C. perfiingens collagenase. Theformer showed higher similarity (Poisson P value, 2.1 x10-15) than the latter (Poisson P value, 1.8 x 10-8). To testtheir statistical significance, a Monte-Carlo test was performedby using the Lipman-Pearson alignment algorithm with theDayhoff similarity scoring (27). In this analysis, the bestalignment of the two original sequences was compared with acontrol obtained from the best alignments between 100 differ-ent randomized sequences. The alignment between the Valginolyticus and the C. perfringens collagenases showed a scorethat was 14.8 standard deviations higher than the control, andthe alignment between A. lyticus protease I and the C. perfrin-gens collagenase was 10.3 standard deviations higher than thecontrol. It should be noted that the V alginolyticus collagenaseand A. lyticus protease I showed the highest similarity (19.3standard deviations) among the three enzymes.The alignment of the three homologous enzymes is shown in

Fig. 6. Note that regions of sequence similarity are notlocalized but are scattered over the whole sequences. Aconsensus zinc-binding sequence (His-Glu-Xaa-Xaa-His) re-garded as the catalytic center of eucaryotic zinc collagenases(20) is also conserved in both the C. perfringens collagenase

J. BACTERIOL.

on February 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

C. PERFRINGENS COLLAGENASE GENE 155

EKNLKRIEITKLKLVERWSATFTLA FNSSFKIFIAIKKVENSNNGO T|E NADOELK LSVA IA--------TTL1TG SEIV ITEOHAHSAHTHGVEFNE-ER--I CISIL--------LLGLIIS AAIAIRPAIFI-------ISKTELNNEVATDNNRPIGPSI IRIRNNKIYTFDEfNIMNYSEVELMTfSYENED-VEYOP TPI OPSKTIJVQSLEILDESSTACDL IA VTESSNOQISE SOGATCIN------YINISSVDKVMETMPIVDVIKAKA----IDEI DKRGIIPRF DAIVDMT

NFIDUSYTFFSNFIRVOA IYGLEDSGRTEADDDKGIP*VEF_ LG KOISAE IOISVFSSIHMYNIAKHTTTL KG G GS ELE FLYU WE DNI

AW YTADGOFAVWRORVRSEK LSLNF FT M----MPAGGRLI

*LNIPOLKNECLPAMK3IENSN RLGTKAODGEEALGRLIGN SADPEVINNCIYVIIFIEWVTMVK--ESVDIFVNT YENSDRHGKILSEVI ITMD GLOHAYLPOVTOQVIPAKQAOGD--RGL S-EDNNNSAROLWTAIVPGAEAV EIVI PRDKVGEFKLFI

SDFK NIDKYISNYSKG NLMK IDYYTNS IYNTKGYIAKNTEFYNRIDPYMERIERWJAOHWYMRNAVAT ILFIGOWNEOFI0OI GNOTILAKA---------- MGKVIHDIVGFIPLARRLA A .----------SGEKGVSGSCN DVV------------CP

SLCTIGIKLNNDNWLVNM YY MIFREDP ISORA-JE AIKEIPYLSIQYIEIADFAL ASS G EDEFMAA GREL t A SVVKSOIS IFEOUE--MIGR DIVEGDGIRIIIRIVGIYSKSGJAC VNN --------D KRYFLT--AHCIMGTNDLDLNFGGKNSSGIDIDFNKIKADAREKYIPKEFDDIKFVVKAGDKV E KIERLIWWLA DTAS YADCSEFGICNFETELKGLVISOUCSPT--IRIL NMIQIOHAACSAST SI VVWNYOI.----------TCRAPNTPASGANI--DGSMETOSGSTV1UIAAIKEVKAcIMRVVONDKAIEEGNPIDI I VUYNI E-LNR I NGFKMGYEEGYIHQSLETGEOPVKDDHNT N FDISTEG FD GT|FTLLELNNAANPAFNIFWAGWI_RRI0-N A IHHPNVAEKRIS-

GiF RTPEE YLEE FR VVP WGOGEFYOEGVLTDPSOP PNUI A SYAN DHFVWN -FVODLY FS ----TEKI VI.---STSPTSFV WGGG GTiHNVOWOPS VT----GS

YE TIFFIGSTRTDG KPRKSVTOGLAYDRNNRMSLYGVLHAKYGSWDFYNYGFALSNWS YVI-

YMYNNNMGMFNKMTNY KNNDVSGYKDY ASMSSDYGLNDKYIDYMDSLINN DNLDVPL.------------------------------------AO NDNIAAIET D---------.------------------------------------U;KRVLGO HG--G -- ---.

VSDEYVNGHEIKD N IKE SNIKDLSSNVEKSOIFTTYDMRGTYVGGRSOGEEND-----.GTYTLSE F VFD DREYRWYL 4F ---------------------

-----.PSCSITGTURSDOYIRVFTSWTGGIA AISIL----------------

WKDMNSKLI DK LSKKS NG TVTAYIVIHKVDNGNYVYDVIFHIMNTDTNTDVHV-ENHKDD TROGN N [TLA *LYOFEGWQQTLISN-----------------SD LDPA GAE IDGLDI .G---GGI .

NKEK KSDSV 1E1E1NPDGTEKE KAYE EKEA KMNKT--A T M-KGKIG S ITSSEN PflV VL _ Q GSEI--TETPPV FTSTTSGLTATFTDS S A RS N N SK AAAElKN( GIESKKIKVVEDKPVEVINESEPNNDFEKANOIAKSNMLJKGTIS|SISE LAMA--------------------------------IH I AIGT G A K------------------------------- SITVIGGPEELSDKYYFDVAIKGNIKELINLNSVGE1LYKEIDLNNYVLYA DGTEIKGEKT|N LPOICAVOSIVSGGRL AGE CLAN00IIWLIA--------p EISNI-----HAQEYTNI-------TDAIIPDIAESPJV ---- RG--------R ATT-----

LEPGRYYLSVYTYDNOSGAYIVIVK NIVETEKDAIKEVENNNDFDKAMKVDSNSK-----------------AITIGIG LSNS------------------------------------------P -----OVT YHTSD.__________

6C463C

12C1058C

180165124

240223181

300271218

359329267

419387315

477447363

531502398

591512405

651526417

711560449

771608471

831666529

891694557

951741593

1011759606

C IVGTLSNDDLKDIYSIDIONPSDINIVIE IKMNWLLYEDDL VD ANA NKL 1071V ----------------GW "H WSD IG GE----C TL WGGVKV D-- 797A ----------------LKVLVAPDITY H RT----GGEH OTFTKDLS@EA-- 644

C SNTCKLNPGKYYLCVY SG-TGNYTVNLONK 1104V ----------------- AIVVDFDAQKCRO 814A -----------------AOREP-GSC-------G 653

FIG. 6. Amino acid sequence alignment of C. perfringens collage-nase (C), V alginolyticus collagenase (V) (34), and A. lyticus proteaseI (A) (26). Identical amino acids are indicated by a black background.The asterisks indicate the consensus sequence for zinc collagenases,HEXXH. The catalytic triad of the Achromobacter serine protease isshaded. The numbering is according to the amino acid sequence of thepremature proteins, and the mature proteins for C, V, and A start atamino acid positions 87, 77, and 206, respectively.

(positions 502 to 506) and the V. alginolyticus collagenase(positions 477 to 481). On the other hand, this putativezinc-binding sequence is absent in A. lyticus protease I, whichis known to be a serine protease. The catalytic triad (His-262,Asp-318, and Ser-399) of the serine protease is present in A.lyticus protease I but not in the other collagenases. Theseresults may imply that the three enzymes, which differ infunction and origin, are evolutionarily related. They mighthave evolved from a common ancestor protein by usingdifferent strategies so that each could increase its catalyticefficiency on the specific substrate in each bacterium. Thecollagenase of V alginolyticus, a gram-negative bacterium, ismore homologous to protease I of A. lyticus, also a gram-negative bacterium, than to the collagenase of C. perfringens, agram-positive bacterium. Furthermore, members of the genusClostridium are considered to be among the most ancientforms of life known (9, 38). The diversity of these enzymes iscorrelated with the phylogenetic relatedness of the threebacteria. Accumulation of more sequence data for bacterialcollagenases and proteases would enable a more preciseanalysis of their homology and provide valuable insight into themolecular evolution of these proteins.

ACKNOWLEDGMENTS

We thank Jon D. Stewart (Department of Chemistry, The Pennsyl-vania State University, University Park) for invaluable discussions andassistance in preparing the manuscript. We also thank Akihisa Taka-mizawa and Mitsuo Takagi (Kanonji Institute, The Research Founda-tion for Microbial Diseases of Osaka University) for preparing syn-thetic oligonucleotides and concentrating protein samples. We areindebted to Yoshihiro Shionoya (Applied Biosystems Japan, Inc.) forthe extensive protein sequencing.

This work was partly supported by a grant in aid for ScientificResearch from the Ministry of Education and Culture, Japan.

REFERENCES1. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman.

1990. Basic local alignment search tool. J. Mol. Biol. 215:403-410.2. Bethesda Research Laboratories. 1986. BRL pUC host: E. coli

DH5aLTM competent cells. Focus 8:9.3. Birkedal-Hansen, H. 1987. Catabolism and turnover of collagens:

collagenases. Methods Enzymol. 144:140-171.4. Birkedal-Hansen, H., and R. E. Tayler. 1982. Detergent-activation

of latent collagenase and resolution of its component molecules.Biochem. Biophys. Res. Commun. 107:1173-1178.

5. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extractionprocedure for screening recombinant plasmid DNA. NucleicAcids Res. 7:1513-1523.

6. Bond, M. D., and H. E. Van Wart. 1984. Characterization of theindividual collagenases from Clostridium histolyticum. Biochemis-try 23:3085-3091.

7. Bond, M. D., and H. E. Van Wart. 1984. Purification and separa-tion of individual collagenases of Clostridium histolyticum usingred dye ligand chromatography. Biochemistry 23:3077-3085.

8. Bond, M. D., and H. E. Van Wart. 1984. Relationship between theindividual collagenases of Clostridium histolyticum: evidence forevolution by gene duplication. Biochemistry 23:3092-3099.

9. Canard, B., B. Saint-Joanis, and S. T. Cole. 1992. Genomicdiversity and organization of virulence genes in the pathogenicanaerobe Clostridium perfringens. Mol. Microbiol. 6:1421-1429.

10. Garnier, T., B. Canard, and S. T. Cole. 1991. Cloning, mapping,and molecular characterization of the rRNA operons of Clostrid-ium perfringens. J. Bacteriol. 173:5431-5438.

11. Hanahan, D. 1985. Techniques for transformation of E. coli, p.109-135. In D. M. Glover (ed.), DNA cloning: a practical ap-proach, vol. 1. IRL Press, Oxford, England.

12. Hatheway, C. L. 1990. Toxigenic clostridia. Clin. Microbiol. Rev.3:66-98.

13. Kameyama, S., and K. Akama. 1970. Titration of kappa toxin ofClostridium perfringens and kappa antitoxin by a modified azocoll

cvA

CVA

CVA

CVA

CVA

CVA

CVA

CVA

CVA

CVA

CVA

CVA

CVA

CVA

CVA

CV

CVA

VOL. 176, 1994

on February 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

156 MATSUSHITA ET AL.

method. Jpn. J. Med. Sci. Biol. 23:31-46.14. Kameyama, S., and K. Akama. 1971. Purification and some

properties of kappa toxin of Clostridium perfringens. Jpn. J. Med.Sci. Biol. 24:9-23.

15. Karn, J., H. W. D. Matthes, M. J. Gait, and S. Brenner. 1984. Anew selective phage cloning vector, X2001, with sites for XbaI,BamHI, HindIll, EcoRI, SstI and XhoI. Gene 32:217-224.

16. Katayama, S.-I., 0. Matsushita, J. Minami, S. Mizobuchi, and A.Okabe. 1993. Comparison of the alpha-toxin genes of Clostridiumperfringens type A and C strains: evidence for extragenic regula-tion of transcription. Infect. Immun. 61:457-463.

17. Kato, T., N. Takahashi, and H. K. Kuramitsu. 1992. Sequenceanalysis and characterization of the Porphyromonas gingivalis prtCgene, which expresses a novel collagenase activity. J. Bacteriol.174:3889-3895.

18. Keil, B. 1979. Some newly characterized collagenases from pro-caryotes and lower eucaryotes. Mol. Cell. Biochem. 23:87-108.

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

20. Lepage, T., and C. Gache. 1990. Early expression of a collagenase-like hatching enzyme gene in the sea urchin embryo. EMBO J.9:3003-3012.

21. Lwebuga-Mukasa, J. S., E. Harper, and P. Taylor. 1976. Collage-nase enzymes from Clostridium: characterization of individualenzymes. Biochemistry 15:4736-4741.

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

23. Matsubara, H., and J. Feder. 1971. Other bacterial, mold, andyeast proteases, p. 721-795. In P. D. Boyer (ed.), The enzymes, vol.III. Hydrolysis: peptide bonds. Academic Press, New York.

24. Merkel, J. R., and J. H. Dreisbach. 1978. Purification and charac-terization of a marine bacterial collagenase. Biochemistry 17:2857-2863.

25. Mookhtiar, K. A., D. R. Steinbrink, and H. E. Van Wart. 1985.Mode of hydrolysis of collagen-like peptides by class I and class IIClostridium histolyticum collagenases: evidence for both endopep-tidase and tripeptidylcarboxypeptidase activities. Biochemistry 24:6527-6533.

26. Ohara, T., K. Makino, H. Shinagawa, A. Nakata, S. Norioka, andF. Sakiyama. 1989. Cloning, nucleotide sequence, and expressionof Achromobacter protease I gene. J. Biol. Chem. 264:20625-20631.

27. Pearson, W. R., and D. J. Lipman. 1988. Improved tools for

biological sequence analysis. Proc. Natl. Acad. Sci. USA 85:2444-2448.

28. Peterkofsky, B. 1982. Bacterial collagenase. Methods Enzymol.82:453-471.

29. Rood, J. I., and S. T. Cole. 1991. Molecular genetics and patho-genesis of Clostridium perfringens. Microbiol. Rev. 55:621-648.

30. 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.

31. Sanger, F., S. Nicklen, and A. K Coulson. 1977. DNA sequencingwith chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA74:5463-5467.

32. Smith, P. K., R. I. Krohn, G. T. Hermanson, A. K. Mallia, F. H.Gartner, M. D. Provenzano, E. K. Fujimoto, N. M. Goeke, B. J.Olson, and D. C. Klenk 1985. Measurement of protein usingbicinchoninic acid. Anal. Biochem. 150:76-85.

33. Tachihara, S. 1985. Rapid separation of plasmid DNA by stepgradient centrifugation. Hitachi Sci. Instr. News 28:16-18.

34. Takeuchi, H., Y. Shibano, K. Morihara, J. Fukushima, S. Inami,B. Keil, A.-M. Gilles, S. Kawamoto, and K. Okuda. 1992. Struc-tural gene and complete amino acid sequence of Vibrio alginolyti-cus collagenase. Biochem. J. 281:703-708.

35. Van Wart, H. E., and D. R Steinbrink 1985. Complementarysubstrate specificities of class I and class II collagenases fromClostridium histolyticum. Biochemistry 24:6520-6526.

36. Willis, A. T. 1983. Clostridium: the spore-bearing anaerobes, p.442-475. In G. Wilson, A. Miles, and M. T. Parker (ed.), Topleyand Wilson's principles of bacteriology, virology and immunity,vol. 2. Edward Arnold (Publishers) Ltd., London.

37. Wilson, M. J., M. Strasser, M. M. Vogel, and A. A. Sinha. 1991.Calcium-dependent and calcium-independent gelatinolytic pro-teinase activities of the rat ventral prostate and its secretion:characterization and effect of castration and testosterone treat-ment. Biol. Reprod. 44:776-785.

38. Woese, C. R 1987. Bacterial evolution. Microbiol. Rev. 51:221-271.

39. Wolf, U., D. Bauer, and W. H. Traub. 1991. Collagenase ofClostridium perfringens type A: degradation of human complementcomponent Clq. Int. J. Med. Microbiol. 276:27-35.

40. Wunsch, E., and H.-G. Heidrich. 1963. Zur quantitativen Bestim-mung der Kollagenase. Hoppe-Seyler's Z. Physiol. Chem. 333:149-151.

41. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. ImprovedM13 phage cloning vectors and host strains: nucleotide sequencesof the M13mpl8 and pUC19 vectors. Gene 33:103-119.

J. BACTERIOL.

on February 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from