INFECTION AND IMMUNITY, Nov. 1994, P. 4938-49470019-9567/94/$04.00+0Copyright C 1994, American Society for Microbiology

An Endo-Acting Proline-Specific Oligopeptidase fromTreponema denticola ATCC 35405: Evidence of

Hydrolysis of Human Bioactive PeptidesPIRKKO-LIISA MAKINEN, KAUKO K. MAKINEN,* AND SALAM A. SYED

Department of Biologic and Materials Sciences, School of Dentistry,The University of Michigan, Ann Arbor, Michigan 48109

Received 2 May 1994/Returned for modification 7 July 1994/Accepted 19 August 1994

An endo-acting proline-specific oligopeptidase (prolyl oligopeptidase [POPase], EC 3.4.21.26) was purifiedto homogeneity from the Triton X-100 extracts of cells of Treponema denticola ATCC 35405 (a human oralspirochete) by a procedure that comprised five successive fast protein liquid chromatography steps. ThePOPase is a cell-associated 75- to 77-kDa protein with an isoelectric point of ca. 6.5. The enzyme hydrolyzed(optimum pH 6.5) the Pro-pNA bond in carbobenzoxy-Gly-Pro-p-nitroanilide (Z-Gly-Pro-pNA) and bonds atthe carboxyl side of proline in several human bioactive peptides, such as bradykinin, substance P, neurotensin,angiotensins, oxytocin, vasopressin, and human endothelin fragment 22-38. The minimum hydrolyzablepeptide size was tetrapeptide P3P2P,P',, while the maximum substrate size was ca. 3 kDa. An imino acidresidue in position P1 was absolutely necessary. The hydrolysis of Z-Gly-Pro-pNA was potently inhibited by thefollowing, with the Kj(8pp) (in micromolar) in parentheses: insulin B-chain (0.7), human endothelin-1 (0.5),neuropeptide Y (1.7), substance P (32.0), T-kinin (4.0), neurotensin (5.0), and bradykinin (16.0). Chemicalmodification and inhibition studies suggest that the POPase is a serine endopeptidase whose activity dependson the catalytic triad ofCOOH ... Ser ... His but not on a metal. The amino acid sequence around the putativeactive-site serine is Gly-Gly-Ser*-Asn-Pro-Gly. The enzyme is suggested to contain a reactive cysteinyl residuenear the active site. Amino acid residues 4 to 24 of the first 24 N-terminal residues showed a homology of 71%with the POPase precursor from Flavobacterium meningosepticum and considerable homology with theAeromonas hydrophila POPase. The ready hydrolysis of human bioactive peptides at bonds involving an iminoacid residue suggests that enzymes like POPase may contribute to the chronicity of periodontal infections byparticipating in the peptidolytic processing of those peptides.

Treponema denticola is one of the predominant members ofthe human periodontal flora (15, 20-24, 34, 45). Previousstudies suggest that T. denticola is associated with periodontalinfections by adhering to epithelial cells (36), gingival fibro-blasts (50), fibronectin (7), laminin, fibrinogen, gelatin, andtype I and type II collagens (12); by degrading basementmembrane collagen (48); by invading healthy tissue (9, 14, 21,32); by suppressing fibroblast proliferation (4); by causingmicroulceration of the sulcular epithelium (28, 30); by showingkeratinolytic activity (31); and by exhibiting mutual symbioticgrowth enhancement with another periodontal pathogen, Por-phyromonas gingivalis (10). Treponemal cells have been shownto migrate through the basement membrane (11, 48), and T.denticola proteases activate host latent procollagenase (46).Because of the complexity of the overall process, the detailedchemical mechanism of human treponemal infections is notknown, although the above studies suggest that proteases andpeptidases present in the outer cell envelope or in the periplas-mic space of the treponemes may play a crucial role. Ourresearch has focused on the cell-associated oligopeptidases oftreponemes and discovered in these cells a novel prolyl endo-peptidase with a strict specificity profile.

Prolyl endopeptidases (EC 3.4.21.26; previously called post-proline endopeptidases) have received attention because oftheir role in the metabolism of vasoactive peptides (29, 49, 52,

* Corresponding author. Mailing address: Department of Biologicand Materials Sciences, School of Dentistry, The University of Mich-igan, Ann Arbor, MI 48109-1078. Phone: (313) 763-6166. Fax: (313)747-3896.

53, 62). A member of this enzyme family, the prolyl oligopep-tidase (POPase), may participate in the processing of brainangiotensin (Ang) (52) and in the degradation of oxytocin (59).After the N-terminal heptapeptide of Ang-I was also demon-strated to possess biological activity through a pathway that isnot dependent on Ang-converting enzyme, POPase was de-fined as one of the putative Ang-I-processing enzymes and thuspart of the RAS cascade (51). It was further suggested thatsynthetic peptide inhibitors of POPase act as antiamnesticagents (60). A POPase from human brain (16) and pig muscle(38) has been characterized and may be involved in thematuration and degradation of hormones and neuropeptides(29, 49, 53).Mammalian POPases are sensitive to diisopropyl fluoro-

phosphate, although they are also inhibited by p-hydroxymer-curibenzoic acid (pHMB) (49, 53). These and other character-istics have given the POPases an imprint of an "obscure" groupof serine proteases (1). The substrate often used in the assay ofthe POPases is carbobenzoxyglycyl-L-prolyl-p-nitroanilide (Z-Gly-Pro-pNA), in which the enzyme hydrolyzes the Pro-pNAbond. The same enzyme is responsible for all the activitiespreviously attributed to postproline endopeptidase, endooli-gopeptidase B, TRH-deamidase, brain kinase B, and oxytocin-degrading enzyme (57). This enzyme was later termed POPaseand recently called prolyl oligopeptide hydrolase. This enzymefamily reflects a further distinct evolutionary line of serinepeptidases and differs in catalytic mechanism from the chymo-trypsin and subtilisin families (40). In the family of POPases,the order of catalytic residues (Asp ... Ser ... His) is differentfrom that of chymotrypsin and subtilisin. POPase is an endo-

4938

Vol. 62, No. 11

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

PROLYL OLIGOPEPTIDASE FROM A SPIROCHETE 4939

peptidase with restricted specificity for substrate size, making itan oligopeptidase.Our studies on T. denticola ATCC 35405 showed that this

organism contains a POPase which hydrolyzes in humanbioactive peptides (HBPs) the bond involving the carboxylgroup of proline. It is possible that the chemical mechanism oftreponemal infection is associated with the degradation ofbradykinin (Bk), substance P (SP), Ang, vasopressin, or otherHBPs used in this study. A potential mechanism for regulatingthe levels of HBPs in mammalian tissues could arise bysubstrate competition for interaction with the enzyme, so thatone peptide may serve to regulate the in vivo level of anotherpeptide (13). It is possible that the POPases from pathogenicorganisms interfere with such reactions. The objective of thisstudy was to investigate the nature of the prolyl oligopeptidasereaction catalyzed by cell extracts of T. denticola. The specificaims of this study included the development of a purificationprocedure for the enzyme, the study of the specificity profile ofthe enzyme and the enzyme's inhibition by peptide inhibitors,and the study of the chemical modification of the enzyme forits possible classification.

MATERLILS AND METHODS

Source and cultivation of the organism and treatment ofcells. Cells of T. denticola ATCC 35405 were grown anaerobi-cally for 48 h in a tryptone-yeast extract-heart infusion brothcontaining 10% heat-inactivated rabbit serum (35). For thepurpose of enzyme purification, 1.2-liter aliquots of the growthmedium in 1.5-liter screw-capped flasks were inoculated with100-ml aliquots of cultures (27) and incubated anaerobicallyfor 4 days at 37°C. The optical density at 660 nm (OD660) wasused to determine the number of cells; an OD660 of 0.2corresponded to 5 x 108 cells per ml (12). The cells wereharvested by centrifugation for 10 min at 16,300 x g. Thisprocedure and all subsequent steps of enzyme purificationwere carried out at 0 to 4°C, except for fast protein liquidchromatographic (FPLC) separations, which were carried outat 22°C.

Chemicals. Unless specifically mentioned, the chemicalsused were obtained from Sigma. The water used in this studywas prepared with a Millipore Milli-Q system and had aresistance of 18 megaohms cm-'.

Purification of the enzyme. The harvesting of the cells forenzyme purification was performed after 4 days of growth,because the cell mass reached a sufficiently high level by 4 days.The cells were washed with 20 mmol of phosphate buffer (pH6.8) per liter and subsequently suspended in the same buffercontaining 0.1 mmol of EDTA (15 ml/8 g of cells, wet weight)per liter. Small volumes of 10% Triton X-100 (Pierce) wereadded to a final concentration of 0.05%. Sixty minutes later,the suspension was centrifuged for 15 min at 27,000 x g. Alarge number of separate purifications were carried out bysubjecting suitable aliquots of the Triton X-100 extracts toFPLC as described below.

(i) Hydroxyapatite-FPLC. The enzyme (normally in 50-mlaliquots) resulting from the detergent extraction was subjectedto high-resolution hydroxyapatite-FPLC (Fig. 1A). The activefractions were combined, and the enzyme was concentratedusing Amicon Centriprep-30 membrane filters.

(ii) Phenyl-Sepharose FPLC. Solid NH4Cl was added to theenzyme from the previous step to a final concentration of 3.5mol/liter. The enzyme was then subjected to separation on aphenyl-Sepharose gel involving a descending NH4Cl gradient(Fig. 1B). The active fractions were combined, and the enzymewas dialyzed (Spectrapor; cutoff, 12 to 14 kDa) for 16 h against

A m- -

I

60 90 120 150 180 210

11

0

100

50

3u

0 20 40

60u

60 RETENTION TIME (Minutes)FIG. 1. Purification of the POPase by means of FPLC. (A) Sepa-

ration of the enzyme on a Calbiochem high-resolution hydroxyapatitecolumn (2.2 by 35 cm). The elution was carried out with phosphatebuffer (pH 6.8) 20 mmol per liter with EDTA (0.1 mmol/liter) (initialbuffer), using a phosphate gradient from 0.02 to 1.0 mol/liter (0 to 35%in 120 min and 35 to 100% in 40 min). The gradient was applied afterthe application of the sample was completed. The fraction volume was3 ml (flow rate, 3 mUmin). (B) Separation of the enzyme from theprevious step on a phenyl-Sepharose CL-4B column (1 by 10 cm). Theelution was carried out with the above buffer using a descendingNH4C1 gradient from 3.5 to 0 mol/liter in 120 min. The column wasfinally eluted with Milli-Q water. The fraction volume was 1 ml (flowrate, 1 ml/min). (C) Separation of the enzyme from the previous stepon Fractogel EMD TMAE-650 anion exchanger (40 to 90 ,um; 1 by 40cm). The elution was performed using for the first 15 min 50 mmol ofTris (pH 7.5) per liter at 2.0 ml/min and subsequently an NaCl gradientfrom 0 to 1.0 mol/liter (0 to 50% from 15 to 80 min and 50 to 100%from 80 to 100 min). (D) Final separation of the enzyme with twosuccessive runs on a Superose 12 column (10/30) using 0.25 mol ofNaCl per liter in 20 mmol of phosphate buffer (pH 6.8) per liter withEDTA (0.1 mmolJliter) (the second separation is shown). The flow ratewas 0.5 ml/min. The elution of the POPase is shown in each panel witha bracket. The scale for the salt gradients (--- ) and the protein at 280nm (the latter shown as FPLC/Pharmacia printouts with values of 0.1to 2.0 for absorption units full scale [AUFS]) is shown on the left (from0 to 100%).

50 volumes of 25 mmol of Tris (pH 7.5) plus 0.1 mmol ofEDTA per liter, changing the dialyzing solution once. Theresulting dialysate was concentrated as above.

(iii) Anion-exchange FPLC. The dialyzed enzyme was chro-matographed through a strong (Tentacle type) anion ex-changer (Fractogel) (Fig. 1C). Normally, less than 10-mlvolumes of the dialysate were applied to the column using aSuperloop 10 sample applicator (Pharmacia). The resultingenzyme was concentrated as above.

(iv) Gel permeation chromatography. The enzyme wasfinally subjected to two consecutive separations on a Superose12 column (Fig. 1D). The purified enzyme was stored in the

D

.

VOL. 62, 1994

5

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

4940 MAKINEN ET AL.

TABLE 1. Purification of the POPase from T. denticola ATCC 35405

Step Vol (ml) Protein Total protein Sp acta Total activity(mg/ml) (mg) (,umol/min/mg) (,umol/min)

1. Triton X-100 extract after centrifugation 226 5.86 1,324.4 0.089 117.92. After hydroxyapatite chromatography 96.7 2.58 249.5 0.38 94.53. After phenyl-Superose 11.0 2.02 22.2 2.63 58.44. After Fractogel/concentration 2.6 1.73 4.5 8.58 38.65. After Superose 12 3.0 n.d. n.d. n.d. n.d.6. After Superose 12 6.0 0.102 0.61 24.02 14.8

a Determined with Z-Gly-Pro-pNA under conditions described in Materials and Methods with MES (50 mmol/liter, pH 6.5)." n.d., not determined.

elution buffer at 4°C. The purification procedure is summa-rized in Table 1. The actual enzyme yield was 0.61 mg from 80g (wet weight) of cells. This yield is about 12.5% of theexpected (theoretical) yield of about 5.0 mg.Enzyme determinations. The POPase was discovered as a

contaminant in crude preparations of the FALGPA-peptidase(an enzyme hydrolyzing 2-furylacryloyl-L-leucylglycyl-L-prolyl-L-alanine) (27), which hydrolyzed Bk at the Pro-7-Phe-8 bondin addition to the Phe-5-Ser-6 bond hydrolyzed by the purifiedenzyme proper (27). Subsequent studies showed that it was thePOPase that hydrolyzed the Pro-7-Phe-8 bond and that thisenzyme could be conveniently assayed using Z-Gly-Pro-pNAas the substrate. The activity of POPase was determined in1.0-ml reaction mixtures containing 0.05 mol of MES (mor-pholine ethanesulfonic acid) (pH 6.5), 0.2 mmol of Z-Gly-Pro-pNA (dissolved in methanol) (3), and 1 to 10 ,ul of enzyme perliter at 30°C. After enzyme addition, the increase in absorptionat 405 nm was monitored for 5 min at 30°C using a ShimadzuUV-265 recording spectrophotometer and a thermostatedcuvette holder. The value of 8,800 M-1 cm-' was used for E405.

Determination of the cleavage site of peptides. The study ofthe hydrolysis of peptides was based on the separation of theproducts of hydrolysis by reversed-phase chromatography on aPharmacia PepRPC 5/5 column. The peptide substrates werefirst incubated for various periods of time (5 min to 2 h) in 50mmol of MES (pH 6.5) per liter with 2 ,ug of enzyme and asuitable quantity of the peptide (0.1 mmol/liter) at 30°C.Aliquots of the mixtures were withdrawn at desired reactiontimes, and the reactions were quenched by adding eluent A(see below) to the aliquot (1:1). The mixtures were immedi-ately subjected to reversed-phase chromatography on a Pep-RPC R 5/5 column, using the Pharmacia FPLC system. EluentA was 0.1% trifluoroacetic acid in water, and eluent B was0.1% trifluoroacetic acid in acetonitrile or 0.05% trifluoroace-tic acid in isopropanol. The increase in the percentage of B perminute depended on the peptide used. The absorption of thepeptide fragments was monitored at 214 nm. The fractionscontaining these fragments were combined, and the resultingsolution was evaporated to dryness using a SpeedVac evapo-rator. The dry residues were hydrolyzed for 4 h at 145°C in 6mol of HCl per liter, and the resulting hydrolysates wereevaporated to dryness. The final dry residues were dissolved ina Beckman System 6300 dilution buffer for compositionalamino acid analyses on a Beckman System 6300 High Perfor-mance Analyzer. The molar ratios of the individual aminoacids were used to determine the structure of the peptidefragments involved.

Determination of Ki values. Determination of KI(app) valuesin the POPase-catalyzed hydrolysis of Z-Gly-Pro-pNA wascarried out by following the reaction in the absence of theinhibitor to establish the uninhibited linear rate of substratehydrolysis (v0). Inhibitor (at least 20-fold molar excess over

enzyme) in no more than 2% of the total assay volume wasadded. The reaction was followed to establish the inhibitedrate (vi). Substrate concentration was kept constant by allowingno more than 5% hydrolysis. Under these conditions, Ki(app)(i.e., Ki in the presence of substrate) was given by vGlvi = 1 +[1]114(app) where [I] is the concentration of the inhibitor (43).The values of Ki proper was calculated according to Cornish-Bowden (6).

Protein determination. The protein concentration was de-termined spectrophotometrically at 220 nm (54).

Chemical modification of POPase. Modification of serylresidues was performed with diisopropyl fluorophosphate.Modification of histidyl residues by diethyl pyrocarbonate wasstudied according to Miles (33). Treatment of the POPase withN-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline (EEDQ) tostudy the involvement of carboxyl groups in enzyme activitywas performed at 25°C (39, 42). E-64 IL-trans-epoxysuccinyl-leucylamide-(4-guanidino)-butane], Hg +, and pHMB wereused to study the modification of active sulfhydryl groups ofthe enzyme (2). Unless otherwise mentioned, all modificationswere carried out in an iced water bath.

RESULTS

Production and localization of POPase in the cells. Themaximum yield of the enzyme was obtained after 2 to 4 days ofgrowth. The maximum production thus took place during thelogarithmic growth phase, declining thereafter, although thegrowth of the cells had not reached a stationary phase by thefourth day. No POPase activity was demonstrated in thegrowth medium. The enzyme activity was functional as acomponent of intact cells. However, the cells need not lyse forthe POPase to be functional. Accordingly, thoroughly washedwhole cells of T. denticola ATCC 35405 were very active,suggesting that the enzyme may be located in the outermembrane or in the periplasmic space. Washing of the cells didnot affect their microscopic morphology. Treatment of the cellswith 0.05% Triton X-100 resulted in a virtually instantaneousliberation of most POPase; even 0.01% detergent was effective.Concentrations higher than 0.1% inhibited the enzyme.

Purity of the enzyme. The purity of the POPase after step 6(Table 1) was studied by means of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (using Phast-Gel Gradient 8-25 and PhastGel SDS Buffer Strips) and FPLCon a protein reversed-phase ProRPC 5/10 column. The Phast-System was from Pharmacia LKB Biotechnology Inc. ThePOPase was homogeneous in SDS-PAGE (Fig. 2) and inreversed-phase FPLC. The purity of the enzyme was indepen-dently reconfirmed by means of microbore-HPLC at theUniversity of Michigan Medical School Protein Structure andSequencing Facility.

INFECT. IMMUN.

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

PROLYL OLIGOPEPTIDASE FROM A SPIROCHETE 4941

kDa

29 _- -

45 _ ..

66 _. _o97.4 _- -

116w --

205 _- "`

"It

:~ ~4

_W- ---75kDa

!I

l 2 3 4 5 6 7

FIG. 2. SDS-PAGE of the enzyme after various purification steps.The PhastGel gradient 8-25 gels and PhastGel SDS Buffer Strips wereused in the PhastSystem (Pharmacia). Lanes: 1, size markers (SigmaSDS 6H); 2, Triton X-100 extract of cells; 3, after hydroxyapatitechromatography; 4, after phenyl-Sepharose; 5, after Fractogel; 6, afterfirst gel permeation chromatography; 7, after second gel permeationchromatography.

Amino acid composition, molecular weight, and isoelectricpoint. The amino acid composition of the POPase is shown inTable 2. These data gave for the molecular weight a range of77,023 to 77,121 (mean, 77,072; SDS-PAGE gave a value of75,000; FPLC on Superose-12 and Superose-6 gave 75,000 and77,000, respectively). The estimated minimum length of thepeptide was 689 amino acid residues. The isoelectric point ofthe POPase was 6.5 (determined in free solution by means ofan LKB isoelectric focusing column and a pH gradient of 3.5 to10).

Substrate specificity. Substrate specificity studies are sum-marized in Table 3. In oligopeptides, the enzyme hydrolyzedinternal peptide bonds which involved the carboxyl group of a

TABLE 2. Amino acid composition of the POPasefrom T. denticola ATCC 35405

No. of Nearestresidues integer

Asparagine/aspartic acid 86.2 86Threonine 35.9 36Serine 44.4 44Glutamine/glutamic acid 68.8 69Proline 28.1 28Glycine 56.0 56Alanine 46.4 461/2Cystineb -3 3Valine 30.1 30Methionine 10.6 11Leucine 60.7 61Isoleucine 35.9 36Tyrosine 15.9 16Phenylalanine 42.6 43Histidine 14.4 14Lysine 83.2 83Tryptophanc 9.0 9Arginine 17.6 18Unknown 0 0

a The hydrolysis of the enzyme (13.28 pmol) was carried out for 60 min at200°C. The values were not corrected for possible loss of amino acids duringhydrolysis.bDetermined after hydrolyzing the enzyme for 22 h at 1i10C in HCI (6

mol/liter).c Determined after hydrolyzing the enzyme for 22.5 h at 110°C in mercapto-

ethanesulfonic acid (3 mol/liter).

prolyl residue and the amino group of another amino acidresidue that did not exhibit strict structural requirements.Several peptides (human big endothelin, neuropeptide Y, andinsulin B-chain; see below) did not serve as substrates of thePOPase, although each one of these molecules contained oneor more suitable peptide bonds and although proline-contain-ing fragments of big endothelin and insulin B-chain werereadily hydrolyzed. Accordingly, insulin B-chain with a molec-ular weight (mol. wt.) of 3495.9 and with a potentially scissilebond of Pro-28-Lys-29 was not hydrolyzed, but it was a potentinhibitor. However, insulin B-chain fragment 22-30 (mol. wt.1086.3) was readily hydrolyzed at Pro-28-Lys-29. Similarly,human big endothelin, with a mol. wt. of 4282.9 and with thepotentially scissile bonds of Pro-25-Glu-26, Pro-30-Tyr-31,and Pro-36-Arg-37, was not hydrolyzed, but it acted as a stronginhibitor. Fragment 22-38 (mol. wt. 1,809) of the same mole-cule was readily hydrolyzed at all above bonds. NeuropeptideY (mol. wt. 4271.7) also contains bonds that in theory shouldbe hydrolyzed by the POPase (Pro-5-Asp-6, Pro-8-Gly-9, andPro-13-Ala-14), but it is too large a molecule to act as asubstrate. These and other data (Table 3) showed that thepresent POPase did not hydrolyze peptides with a mol. wt.higher than ca. 3,000. The hydrolysis of SP by POPase atPro-4-Gln-5 still left the fragment SP[5-11] biologically active(the C-terminal heptapeptide of SP is biologically more potentthan SP itself). The POPase hydrolyzed Bk at Pro-7-Phe-8 butdid not attack the Pro-3-Gly-4 bond. The minimum hydrolyz-able peptide size was P3P2P1P'1.

Affinity and specificity constants. The values ofKm and Vm.nfor the hydrolysis of Z-Gly-Pro-pNA, Bk, SP, and Ang-I weredetermined in 0.02 mol of phosphate (pH 6.8) per liter at 30°C,using the Enzpack 3 program (Biosoft, Ferguson, Mo.). Thehydrolysis followed the normal Michaelis-Menten kinetics.Therefore, plots of Lineweaver-Burk, Hanes-Wolf, and Eadie-Hofstee as well as the direct linear and the Wilkinson methodsgave essentially similar results, and the values of Ki,, as well asthose of kcat and the specificity constant (kcatlKm), are shown inTable 4. The high affinity and specificity constants for SP andAng-I deserve attention. Among the peptides studied, SP wasthe best POPase substrate.

Effect of pH on enzyme reaction. The enzyme hydrolyzedZ-Gly-Pro-pNA most rapidly at pH 6.5 when tested in 50 mmolof MES per liter and Bis-Tris buffers (Fig. 3). Washed wholecells hydrolyzed Z-Gly-Pro-pNA, Bk, and SP rapidly near pH6.5 as well.

Elect of NaCl. The POPase was strongly activated by NaClat up to 1.5 mol/liter. The degree of activation increased withincreasing pH (Fig. 4). The effect of NaCl was studied becausethis salt exerts a selective effect on proline-specific peptidases,i.e., low and high concentrations have quite different andcharacteristic effects on the rate of hydrolysis of the favoredsubstrates of three peptidases of T. denticola ATCC 35405(Fig. 4). The activity of proline aminopeptidase (active onNot-L-prolyl-2-naphthylamine) (25) did not appreciably dependon NaCl at the concentrations used, while the FALGPA-peptidase (27) was inhibited by all NaCl concentrations tested.This differentiation between treponemal peptidases by specificNaCl effects may be important to the in vivo function of theenzymes, because their natural environment (human gingivalcrevice) may contain Cl- at relatively high levels, i.e., up to0.01 mol/iter (8).

Effect of temperature. The stability of the isolated enzymedecreased rapidly above 37°C. The enzyme lost about 50% ofits activity after 35 min at 37°C in MES at 50 mmol/liter, pH 6.5.Summary of chemical modification studies. The order of the

catalytic triad residues in the members of the POPase family is

VOL. 62, 1994

40

2

Om

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

4942 MAKINEN ET AL.

TABLE 3. Determination of the cleavage site of peptides hydrolyzed by the POPase from T denticola ATCC 35405

Peptides identifiedSubstrate Amino acid sequence after reaction with

POPasea

Z-Gly-Pro-pNA Z-GIP-pNA

Substance P R-P-K-PQ--Q-F-F-G-L-M-NH2 1-4; 5-11

Angiotensin I D-R-V-Y-I-H--P'F-H-L 1-7; 8-10

Angiotensin II D-R-V-Y-I-H-PIF 1-7;8

Neurotensin pE-L-Y-E-N-K-PAR-R-P4Y-I-L 1-7; 11-13

Neurotensin 8-13 R-R-P1Y-I-L 8-10

Bradykinin R-P-P-G-F-S-P1F-R 1-7; 8-9

Oxytocin C-Y-I-Q-N-C-PLO-H2 1-7

[Arg8J-Vasopressin C-Y-F-QN-CN-P'R-G-NH2 1-7

[Lys8J-Vasopressin C-Y-F-Q-N-C-PK-G-NH2 1-7

FBI-Peptideb RSD-P- K-P 1-5; 6-10; 1-6

Hnge region pepddec P-T-P NS-N2 1-3

Endothdelin fragment 22-38d V-N-T-P 22-30; 31-3622-25; 26-36

a The reactions were carried out for 60 min in 0.21-ml mixtures containing 2 ,ug of enzyme and 0.5 mg of the peptide in MES (50 mmol/liter, pH 6.5) at 30°C.Z-Gly-Pro-pNA was tested at 0.2 mmol/liter. The peptide fragments were separated with a PePRP 5/5 column, and the compositional amino acid analysis was performedafter hydrolysis of the fragments. The arrow shows the site of cleavage. All substrates were from Sigma (unless otherwise shown), and all amino acids were in the L-form(Z in Z-G-P-pNA stands for carbobenzoxy).

b Inhibits the binding of fibronectin to fibroblasts.'Analog of the hinge region of human IgA2.d Human big endothelin-1 fragment (Bachem, Philadelphia, Pa.).

suggested to be Asp ... Ser ... His (from N to C terminus) (1).Therefore, the modification studies focused on these aminoacid residues and also on the possible presence of a sulfhydrylgroup.

(i) Histidyl residues. When treated with 0.21 to 0.64 mmolof diethyl pyrocarbonate per liter in 20-mmollliter phosphatebuffer, pH 6.8 (containing 0.1 mmol of EDTA and 0.25 mol ofNaCl per liter), the enzyme was inactivated in a time- anddose-dependent manner. Addition of hydroxylamine (finalconcentration, 0.2 mol/liter) to the inactivated enzyme not onlytotally reversed the inactivation, but slightly (10 to 15%)activated the enzyme. Plot of log(1/t05) versus [I], where tO5stands for the time (in minutes) required for 50% inactivation

TABLE 4. Kinetic constants of the hydrolysis of oligopeptides bythe POPase from T. denticola ATCC 35405'

Substrate Km (M) kcat klatIKm(min-') (M-1 min-')

Z-Gly-Pro-pNA 8.3 x 10-4 6,767 8.15 x 106Bk 9.1 x 10-5 1,078 1.18 x 107SP 1.1 x 10-5 1,286 1.17 x 108Ang-I 2.0 x 10-5 545 2.72 x 107

a Each of the peptides was hydrolyzed at only one peptide bond (Table 3). Theenzyme reactions were performed in buffer containing 20 mmol of phosphatebuffer (pH 6.8) and 0.1 mmol of EDTA per liter at 30°C. The POPaseconcentration was 7.3 nmol/liter. The peptides were separated on a reversed-phase column (PePRP 5/5) and identified and quantitated by means of aminoacid analysis.

and [I] is the concentration of the modifier (19), resulted in astraight line with a slope of 1.1, indicating that an average of atleast one modifier molecule binds to POPase when inactivationoccurs. Because hydroxylamine reactivated the enzyme, itsinactivation most likely correlated with the carbethoxylation ofone histidyl residue.

(ii) Seryl residues. Treatment of the enzyme with 43 to 540,umol of diisopropyl fluorophosphate per liter in the abovebuffer resulted in 80 to 100% irreversible inactivation of theenzyme in 10 min. The above plot gave a straight line with aslope of 0.9, suggesting that an average of at least one inhibitormolecule had reacted with POPase. Because dithiothreitol (1.0mmol/liter) failed to reverse the inactivation (thus excludingthe possibility that diisopropyl fluorophosphate reacted with acysteinyl residue), it can be assumed that an active-site serineresidue is involved. Benzamidine and phenylmethyl sulfo-nylfluoride (both at 1.0 mmol/liter) caused only 14 to 15%inactivation, while aprotinin was without effect at 6 ,umol/liter.Diisopropyl fluorophosphate has been claimed to cause a rapidand irreversible inactivation of serine-dependent proteolyticenzymes, while phenylmethylsulfonyl fluoride is much lessreactive.

(iii) Carboxyl groups. The enzyme was irreversibly inacti-vated by 1.0 to 4.0 mmol ofEEDQ (dissolved in methanol) perliter for 60 min at 25°C (tested in the above buffer). The aboveplot gave a straight line with a slope of 1.1, suggesting that atleast one carboxyl group had reacted with the EEDQ. Aprotonated form of a carboxyl group is required for EEDQmodification. The modification of POPase was therefore also

INFEcr. IMMUN.

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

~~~~~~PROLYL OLIGOPEPTIDASE FROM A SPIROCHETE 4943

0.1

0.4

A40!

0.4

5 6 7 8 pH

FIG. 3. Effect of pH on POPase-catalyzed reactions. (A) Rate (in

A405) of the hydrolysis of 0.2 mmol of Z-Gly-Pro-pNA per liter by the

purified enzyme in different buffer systems (90 mmol/liter). (B)Hydrolysis of 0.1 mmol of Z-Gly-Pro-pNA (Z) per liter by washed

whole cells (curve 1) and by the supernatant fluid of sonicated cells

(curve 2) in MES (50 mmol/liter) and the hydrolysis of the Pro-4--Gln-5

bond of 0.1 mmol of SP per liter (curve 3), and that of the Pro-7-Phe-8

bond of 0.1 mmol of Bk per liter (curve 4), by washed whole cells in

MES (50 mmol/liter). The rate of the hydrolysis of SP and Bk was

determined from the size of PepRPC 5/5 peaks SP[1-4] and Bk[1-7],respectively, and adjusted to the A45 axis.

performed at pHs from 5 to 7 (the acid sensitivity of POPase

did not allow experiments below pH 5). When the effect of

EEDQ (0.75 mmollliter) was tested at 250C in MES (0.1mmol/liter) at pHs from 5 to 7, the rate of modification

increased with decreasing pH, and the plot of kapp versus pH

gave a curve with apK value below pH 5.5, suggesting that the

reagent had reacted with the protonated form of the active

(carboxyl) group.

(iv) Sulflhydryl reagents. The enzyme activity could be

effectively destroyed in the presence of typical thiol-s?ecific

reagents. For example, pHMB and Hg2+ even at 10- molt

liter, caused a strong inactivation of the enzyme, which could

be reversed by 2-mercaptoethanol. The reactions were per-

formed in 1.0-ml mixtures to which to 10 pxl of HgCl2 solution

was added. The reaction time was 5 min. After each 5-mmnreaction, resulting in the inhibition shown, addition of 2-mer-

captoethanol to a final concentration of 1.0 mmol/liter imme-

diately returned the enzyme activity to approximately 90% of

the original level. If left standing for several days at 40C, the

enzyme activity could not be returned. It can be seen in Fig. 5

(inset) thatpHMB inactivated the enzyme nearly stoichiomet-

rically. Titration with pHMB could thus be used for the

determination of the concentration of the purified POPase.

Iodoacetamide was much less effective. However, E-64, which

has been claimed to be absolutely specific for active-site

(catalytic) sulfhydryl groups, was without effect at concentra-

~~pH 6.8

~10O00 0.5 1.0

(D>

B

150

1004

2

50

3

0 1 23

NaCI Concentration (M)

FIG. 4. Importance of NaCi to peptidase activity. (A) Effect of

NaCi at two pH values on the rate of the hydrolysis of 0.2 mmol of

Z-Gly-Pro-pNA per liter in phosphate buffer (25 mmol/liter) at 300C.

(B) Effect of NaCi on the activity of three peptidases from T denticola

ATCC 35405. Curve 1, POPase; curve 2, proline iminopeptidase (25);curve 3, peptidase hydrolyzing FALGPA (27). The reactions were

performed at 300C in 0.1 mol of bis-Tris-propane (pH 7.5) per liter

with the iminopeptidase and in 0.1 mol of MES (pH 6.5) per liter with

other enzymes.

tions of 1 to 10 pLmol/liter (which are normally effective in the

study of cysteinyl proteases). Also, because no thiols needed to

be added to the enzyme to maintain full activity, it is possiblethat the above strongly inhibitory thiol reagents had reacted

with a cysteinyl residue that is not necessary for catalysis but is

located close enough to the active site to affect the hydrolysisof Z-Gly-Pro-pNA. Such "reactive cysteines" have been re-

ported to be necessary for substrate binding.Inhibition by peptides. Among the potentially inhibitory

peptides were Bk, Bk fragment 1-7, Bk potentiator B, SP and

some of its fragments or modification products, peptidesderived from TgA2, Ang, and vasopressin (Table 5). The most

potent inhibitors were, however, human big endothelin, neu-

rotensin, neuropeptide Y, and oxidized insulin B-chain. In

general, larger peptides were stronger inhibitors than shorter

peptides. With peptides for which the value of was deter-

mnined, the inhibition was of the competitive type (except for

insulin B-chain). Because of the strong inhibition of the

POPase by oxiddized insulin B-chain (in which the sulfhydrylgroups have been oxiddized), the effect of cysteic acid was also

tested. The value of 2.14 mM for Kiap reflected low inhibi-

tion.'(a

Elfect of chelators, their analogs, and other inhibitors.

1,10-Phenanthroline and its nonchelating homologs 1,7-phen-anthroline and 4,7-phenanthroline were equally inhibitory,indicating the involvement of unspecific effects. EDTA had a

stabilizing effect during purification of the enzyme and did not

inhibit the purified enzyme. EDTA, 8-hydroxyquinoline sulfo-

nic acid, and EGTA [ethylene glycol-bis(3-:aminoethyl ether)-

io-A~~~~MESBIS-TRIS

25-

~~~~~BIS-TRISPROPANE

5_B

~~~~~~Z(whole cells)

25-Z (sonicate)

1~~~~~SP(hlecls2 BK (whole cells)>2r~ K(hlecls

44I

VOL. 62, 1994

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

4944 MAKINEN ET AL.

250

co

0

FIG. 5. Reactiojcysteinyl residue oconcentration in t]Z-Gly-Pro-pNA per(Inset) Titration ofmmol of phosphateliter) plus NaCl (0.2pHMB per liter wereach step, 1.0 pmc(dashed line) gaveconcentration of en

N,N,N',N'-tetraa(zyme. These findATCC 35405 ismammalian POP;reagents (see abowhile Zn2+ at 0.2of 130 pLM.

Effect of sulfhypounds (dithiothvated (10 to 15%Amino acid sec

terminus. The N-residues is M-Q-'R-W-L. When tisequences of all I(31,808 sequencescore of 71% for Iterium meningoseiI n f- ')A Ah1Z

TABLE 5. Inhibition of the T. den

Peptide

I ~~~~~~BkK ~~~~~~~~Bk[2-7]50% Bk[1-7]

X Bk[1-6]an I0> Bk[1-5]cc [Lys']-Bk

0 2 4 6 Lys-BkMoles/L (108) pHMB Ile-Ser-Bk (T-kinin)

Met-Lys-Bkdes-Pro2-Bkdes-Arg9-Bkdes-Arg9, [Leu8]-Bk

1 2 3 4 5 Bk potentiator

Moles Hg2+/L (106) Bk potentiator BBk potentiator C

n between Hg2+ and pHMB with a "reactive" SPIf POPase. Residual activity plotted versus Hg2+ SP[1-9]he POPase-catalyzed hydrolysis of 0.2 mmol of SP[1-7]r liter in 0.1 mol of MES (pH 6.5) per liter at 30°C. SP[2-11]f 70 nmol of POPase per liter with pHMB in 20 SP[3-11]buffer per liter (pH 6.8) with EDTA (0.1 mmoll SP[4-11]

25 mol/liter) at 30°C. Aliquots (1 ,ul) of 1.0 ,umol of [D-Pro2, D-Trp7'9]-SPre added at 10-min intervals to a 0.1-ml mixture (at IgA2 hinge regionc (Pro-Thr-Pro-)l of pHMB was added). The extrapolated curve Ser-NH2)the value of 0.9 to [pHMB]/[E], where [E] is the FBI peptidedizyme. IgGl Fc regione (Leu-Pro-Pro-Ser-

Arg)Ang-I

cetic acid] de facto slightly activated the en- Ang-IIlings suggest that the POPase of T. denticola Neurotensin[8-13]not a metalloenzyme. Bacitracin inhibits ,-Casomorphin

ases. Under the conditions used for sulfhydryl [Lys8]-vasopressinwe), bacitracin gave a K( value of 14 uM, [Arg ]-vasopressin20 mmollliter inhibited by6P4%, with a Ki( ) Human big endothelin (38 residues;

mol. wt. 4282.9),dryl compounds. Both tested sulfhydryl com- Human big endothelin fragmentireitol and 2-mercaptoethanol) slightly acti- 19-38 (mol. wt. 2221.6)eHuman endothelin-1 (mol. wt.) the enzyme. 2492)eluence. The POPase showed an unblocked N Neuropeptide Y (36 residues; mol.terminal amino acid sequence for the first 24 wt. 4271.7)Y-K-K-S-D-V-S-D-N-Y-F-G-T-I-V-P-D-?-Y- Insulin A-chain (oxidized; 21his sequence was scanned for homology to residues; mol.-wt. 2531.6)proteins present in the database CDPROT26 Insulin B-chain (oxidized; 30s) (37), sequence 4 to 24 showed a relative residues; mol. wt. 3495.9)homology with sequence 32 to 52 ofFlavobac- Insuli B-chain fragment 22-30pticum POPase precursor (Fig. 6). Sequence Insulin (51 residues1 mol. wt. 5700)I o1 h 0Insulin(51residu,es;r mo. wQtf 7 0)IlU L0L Usnowea close nomoIogy wiln sequenceil 1oL 5 oi

Aeromonas hydrophila POPase. A 6.2-kDa fragment of theTreponema POPase, obtained with CnBr treatment, was alsochecked for homology. Sequence 2 to 41 of this fragmentshowed a 74% score with sequence 525 to 564 of the POPaseprecursor of F. meningosepticum and even greater homologywith the sequence 506 to 545 of the POPase of A. hydrophila.A score of 68.9% with sequence 2 to 38 of pig brain POPasewas obtained. Sequencing of the 6.2-kDa fragment of theTreponema POPase revealed the amino acid sequence aroundthe putative active seryl residue to be G-G-S*-N-P-G. Aputative active-site aspartic acid residue was located in thisfragment (Fig. 6).

DISCUSSION

The only bacterial POPases more thoroughly studied havebeen obtained from F. meningosepticum (58), a Xanthomonas

Iticola POPase by oligopeptidesa

Kiap Ki Action as a(lfrM) (105 M) substrate'

1.6 +16.8 -

1.7 -

7.4 -

19.0 -

2.3 +1.7 +0.4 +2.0 +2.5 +0.9 +1.4 +6.7 n.t.1.8 n.t.

10.1 n.t.3.2 4.0 (C) +2.9 3.5 (C) +5.5 10.0 (C) +7.5 8.3 (C) n.t.4.1 3.9 (C) n.t.2.9 2.2 (C) n.t.1.5 n.t.4.9 +

4.13.5

2.73.40.52.13.01.71.30.36

0.06

0.05

0.17

6.7

0.07

+

++++nt.++

3.4 (C)4.1 (C)

+

0.11 (NC)

1.2 +

0.93

a The reactions were carried out in MES (50 mmol/liter, pH 6.5) at 30°C using0.2 mmol of Z-Gly-Pro-pNA per liter as substrate (C, competitive; NC, noncom-petitive). The molecular weight of some larger peptides is indicated in paren-theses. Unless otherwise indicated, the peptides were from Sigma.

b Suitability as a substrate of the POPase is indicated, if known (+, acts as asubstrate; -, is not hydrolyzed; n.t., not tested as substrate).

' Analog of the hinge region of human IgA2; inhibits proteolysis of IgA byNeisseria gonorrheae protease Type I.

d Inhibits fibronectin binding to fibroblasts.From Novabiochem (La Jolla, Calif.)

sp. (47), and A. hydrophila (17). The POPases isolated fromorganisms associated with human diseases display an interest-ing substrate specificity profile: they hydrolyze certain proline-containing HBPs at a high rate. Although the role of POPasesin bacterial inflammations remains to be determined, thepossibility exists that microbial POPases have developed simul-taneously with the human POPases to carry out functions that

INFECT. IMMUN.

- N.,

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

PROLYL OLIGOPEPTIDASE FROM A SPIROCHETE 4945

A

29

9

24 T-POP

52 F-POP

32 A-POP

B

FIG. 6. Homology between POPases from different sources. Alignment of the N-terminal segment (A) and the segments around the active-siteareas (B) of T. denticola POPase (T-POP) with segments of POPase from F. meningosepticum (F-POP) (61), A. hydrophila (A-POP) (17), and pigbrain (P-POP) (41). The putative active-site aspartic acid residues have been circled. The active seryl residues have been marked with an asterisk(for the 6.2-kDa T-POP fragment, the involvement of the corresponding seryl residue in enzyme activity remains to be vindicated). The sequencingstudies of the 6.2-kDa fragment provided inconclusive results for residue 24 (which is either R or S), for residue 36 (which is most likely G), andfor residue 43 (which is most likely V). Residues identical to those of T-POP are shaded.

are necessary for the perpetuation of the pathogen in humanhabitats. Such a functional adaptation to the host environmentmay enable the enzyme to escape from being recognized by thehost's immune system as nonself (18) and consequently avoidinactivation. Such bacterial functions may be pathogenic to thehost. This study demonstrated the presence in the cells of T.denticola ATCC 35405 of a POPase which is active on proline-containing HBPs. On the criteria of purity presented above,the T. denticola POPase was judged to be homogeneous,although the enzyme yield was regarded as somewhat unsatis-factory. The lower yield should be considered against the factthat in purification steps involving hydrophobic-interactionFPLC, the interactions that maintain protein conformation arealso those that mediate retention. Therefore, the FPLC itselfcan cause partial or even complete unfolding of a protein. Onthe other hand, this FPLC step is sensitive enough to separatethe pure native protein from other forms (Fig. 1B). ThePOPase is an example of "proline-specific peptidases" (thisterm implies that the peptidase requires an imino acid residueat or near the scissile bond). A typical proline-specific enzymeis proline aminopeptidase (EC 3.4.11.5), which is remarkablyactive in cells of T. denticola (25), suggesting that it plays animportant role in the propagation of this organism. A specificfunction may also be ascribed to the FALGPA-peptidase,which acts on Bk and small collagen fragments (27) andrequires the presence of a proline residue at the P'2 position ofthe substrate, which is hydrolyzed. The treponemal proline-specific peptidases, including the POPase, seem to be cellassociated, but so located (in the outer membrane?) that theycan readily inactivate HBPs present at the site of inflammation.It is thus possible that the proline-specific peptidases areimportant in the chemical aggressiveness of the treponemes.Based on the evidence presented in this paper, the T.

denticola peptidase can be considered to belong to the POPasefamily of serine proteases (40). This proposition is primarilybased on the strict specificity of the peptidase for a prolineresidue in the P1 position of peptides ranging from tetrapep-tides to about 3-kDa oligopeptides. Second, the chemicalmodification studies suggested the involvement of an activeseryl residue, an active histidyl residue, and an active carboxylgroup, which form the catalytic triad in the POPase family; the

amino acid sequence around the active-site serine residueclosely resembles that of POPase from F. meningosepticum, A.hydrophila, and pig brain. Third, the mol. wt. and the pHoptimum of the T. denticola peptidase were similar to those ofPOPases. Furthermore, the sequence of N-terminal aminoacid residues 4 to 24 showed >70% homology with residues 35to 52 of the F. meningosepticum POPase precursor (61), andresidues 10 to 24 showed a close similarity with sequence 18 to32 of A. hydrophila POPase (17).The T. denticola enzyme contains at least three Cys residues,

of which one reacts stoichiometrically with pHMB. It ispossible that this reactive Cys residue is necessary to conferconformational changes upon substrate binding. Previous lit-erature suggests that the Flavobacterium (58) and the Xantho-monas (47) POPases are not inhibited by sulfhydryl reagents.Consequently, the Treponema enzyme may be considered toresemble in this respect more closely the mammalian POPaseswhich show susceptibility topHMB (52). The POPase gene ofeukaryotes may have evolved further, with the introduction ofan additional Cys residue near the active site.The biological role of the POPase must be closely associated

with the location of the enzyme in the cells. The T. denticolaPOPase may be regarded as a bacterial "ectopeptidase,"because quite low detergent levels liberated the enzyme virtu-ally entirely. Triton X-100 is known to remove the outermostlayer of the cells of T. denticola (5), this structure being thesource of several peptidases. The ability of intact, washed cellsto bring about the rapid hydrolysis of Z-Gly-Pro-pNA, SP, andBk with enzymatic characteristics similar to those of thepurified POPase (to be reported) must be emphasized. Theenzyme indeed displayed high affinity for Bk, SP, Ang-I, andothers. It is possible that the POPase interferes with thenormal balance of such proline-containing peptides, therebycontributing to the maintenance of the inflammatory conditionin periodontal infections. For example, fragment SP[5-11],which is produced by the POPase from intact SP, has beenfound to be biologically more active than SP itself (56).Furthermore, the POPase can also serve as an ancillaryAng-I-converting system, because the T. denticola peptidasealso liberated fragment 1-7 from the intact Ang-I. This frag-ment is biologically active (52). Therefore, although it is

2KQtMF~45 T-POP 6.2kDa

525 ITI W M 568 F-POP

506 ~~~~~~~~~ ~ ~ ~~ ~~~~~~~~~~~~~~~~~~~~~A~ K E YF T D RL AR QiGt,V # M TO 549 A-POP523 Q AAY KEGCTPK RLNT1N CNG ,V A T C ANWQ 566 P-POP

I

VOL. 62, 1994

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

4946 MAKINEN ET AL.

possible that the hydrolysis of proline-containing HBPs by thePOPase is incidental, with no pathological meaning, thePOPase is expressly an oligopeptidase with a strict specificity,not a protease. The enzyme showed too low affinity constantvalues for several proline-containing HBPs to be categoricallyignored; low values of those constants speak for high affinity insubstrate binding. The cells of T. denticola possess the POPasefor a purpose, and it is conceivable that the natural substratesof this enzyme include natural, proline-containing HBPs. Fu-ture research must elucidate the possibility of collagen frag-ments and salivary proline-rich peptides acting as potentialsubstrates of the POPase. It is noteworthy that peptides canreplace free amino acids as preferential growth substrates of Tdenticola (44, 55).The reason for larger peptides being inhibitors and not

substrates may be the strict spacial geometry of the active siteof the POPase. The active site may have to accommodate theentire scissile peptide (<about 3 kDa), or most of it, in orderto trigger the cleavage of the proline-containing peptide bond.With larger peptides, binding of only the potentially hydrolyz-able site may take place; the peptide may be too large to triggerhydrolysis. This results in a strong and specific inhibition of thePOPase.The proposition that mammalian POPases participate in the

normal processing of human proline-containing peptides de-serves attention. The proline residue may serve as a regulatorysignal determining the life-time and degradation rates ofproline-containing HBPs. The close structural and functionalsimilarity between the host's and the pathogen's POPasessuggests that the "pathogenic" POPases may interfere with thenormal peptidolytic processing of the host's oligopeptideswithout being effectively eliminated by the immunologic de-fense of the host. The POPase seems to complement the richpeptidase arsenal of the treponemal cell wall. Other pepti-dases, such as the FALGPA-peptidase (27), the PZ-PLGPA-peptidase (26), BAPNA-peptidase (35), and proline amino-peptidase (25), complete each others' specificities. The com-bined peptidolytic capacity may render possible an effectiveexploitation of the host's bioactive peptides for the benefit ofthe treponeme's propagation.

ACKNOWLEDGMENTSupported in part by Center grant NIH 5 P60 AR20557.

REFERENCES1. Barrett, A. J., and N. D. Rawlings. 1992. Oligopeptidases, and the

emergence of the prolyl oligopeptidase family. Biol. Chem.Hoppe-Seyler 373:353-360.

2. Beynon, R. J., and G. Salvesen. 1989. Commercially availableprotease inhibitors, p. 241-249. In R. J. Beynon and J. S. Bond(ed.), Proteolytic enzymes: a practical approach. IRL Press, Ox-ford.

3. Blumberg, S., V. I. Teichberg, J. L. Charli, L. B. Hersch, and J. F.McKelvy. 1980. Cleavage of substance P to an N-terminal tet-rapeptide and a C-terminal heptapeptide by a post-proline cleav-ing enzyme from bovine brain. Brain Res. 192:477-486.

4. Boehringer, H., N. S. Taichman, and B. J. Shenker. 1984. Sup-pression of fibroblast proliferation by oral spirochetes. Infect.Immun. 45:155-159.

5. Cockayne, A., R. Sanger, A. Ivic, R A. Strugnell, J. H. Macdougall,R R B. Russell, and C. W. Penn. 1989. Antigenic and structuralanalysis of Treponema denticola. J. Gen. Microbiol. 135:3209-3218.

6. Cornish-Bowden, A. 1974. A simple graphical method for deter-mining the inhibition constants of mixed, uncompetitive andnon-competitive inhibitors. Biochem. J. 137:143-144.

7. Dawson, J. R., and R P. Ellen. 1990. Tip-oriented adherence ofTreponema denticola to fibronectin. Infect. Immun. 58:3924-3928.

8. Ferris, G., T. E. Grow, S. B. Low, and R. T. Ferris. 1987.Measurement of gingival crevicular fluid conductivity, in vivo. J.Periodontol. 58:46-50.

9. Goldhaber, P., and D. B. Giddon. 1984. Present concepts concern-ing the etiology and treatment of acute necrotizing ulcerativegingivitis. Int. J. Dent. Res. 14:486-496.

10. Grenier, D. 1992. Nutritional interactions between two suspectedperiodontopathogens, Treponema denticola and Porphyromonasgingivalis. Infect. Immun. 60:5298-5301.

11. Grenier, D., V.-J. Uitto, and B. C. McBride. 1990. Cellular locationof a Treponema denticola chymotrypsinlike protease and impor-tance of the protease in migration through the basement mem-brane. Infect. Immun. 58:347-351.

12. Haapasalo, M., V. Singh, B. C. McBride, and V.-J. Uitto. 1991.Sulfydryl-dependent attachment of Treponema denticola to lami-nin and other proteins. Infect. Immun. 59:4230-4237.

13. Hersch, L. B. 1981. Immunological, physical, and chemical evi-dence for the identity of brain and kidney post-proline cleavingenzyme. J. Neurochem. 37:172-178.

14. Heylings, R T. 1967. Electron microscopy of acute ulcerativegingivitis (Vincent's type). Demonstration of the fusospirochaetalcomplex of bacteria within prenecrotic gingival epithelium. Br.Dent. J. 122:51-56.

15. Holt, S. C., and T. E. Bramanti. 1991. Factors in virulenceexpression and their role in periodontal disease pathogenesis. Crit.Rev. Biol. Med. 2:177-281.

16. Kalwant, S., and A. G. Porter. 1991. Purification and character-ization of human brain prolyl endopeptidase. Biochem. J. 276:237-244.

17. Kanatani, A., T. Yoshimoto, A. Kitazono, T. Kokubo, and D.Tsuru. 1993. Prolyl endopeptidase from Aeromonas hydrophila:cloning, sequencing, and expression of the enzyme gene, andcharacterization of the expressed enzyme. J. Biochem. 113:790-796.

18. Kumar, S. 1993. Selective modification and immune evasion: ahypothesis. Immunol. Cell. Biol. 71:141-143.

19. Levy, H. M., P. D. Leber, and E. M. Ryan. 1963. Inactivation ofmyosin by 2,4-dinitrophenol and protection by adenosine triphos-phate and other phosphate compounds. J. Biol. Chem. 238:3654-3659.

20. Listgarten, M. A., and L. Hellden. 1978. Relative distribution ofbacteria at clinically healthy and periodontally diseased sites inhumans. J. Clin. Periodontol. 5:115-132.

21. Loesche, W. J. 1976. Periodontal disease and the treponemes, p.261-275. In R. C. Johnson (ed.), The biology of the parasiticspirochetes. Academic Press, New York.

22. Loesche, W. J., and B. Laughon. 1982. Role of spirochetes inperiodontal disease, p. 62-75. In R. J. Genco and S. E. Mergen-hagen (ed.), Host-parasite interactions in periodontal disease.American Society for Microbiology, Washington, D.C.

23. Loesche, W. J., S. A. Syed, E. Schmidt, and E. C. Morrison. 1985.Bacterial profiles of subgingival plaques in periodontitis. J. Peri-odontol. 56:447-456.

24. Loesche, W. J., S. A. Syed, and J. Stoll. 1987. Trypsin-like activityin subgingival plaque: a diagnostic marker for spirochetes andperiodontal disease. J. Periodontol. 58:266-273.

25. Miakinen, K. K., S. A. Syed, P.-L. Muikinen, and W. J. Loesche.1987. Dominance of iminopeptidase activity in the human oralbacterium Treponema denticola ATCC 35405. Curr. Microbiol. 14:341-346.

26. Maikinen, K. K., S. A. Syed, S. L. Salvador, and P.-L. Mikinen.1990. Hydrolysis of the Leu-Gly bond of phenyalzobenzyloxycar-bonyl-L-Pro-L-Leu-Gly-L-Pro-D-Arg (a substrate of microbial col-lagenases) by treponemes from the subgingival plaque of peri-odontitis patients. Curr. Microbiol. 20:69-74.

27. Makinen, K. K., P.-L. Makinen, and S. A. Syed. 1992. Purificationand substrate specificity of an endopeptidase from the human oralspirochete Treponema denticola ATCC 35405, active on furylacryl-oyl-Leu-Gly-Pro-Ala and bradykinin. J. Biol. Chem. 267:14285-14293.

28. Maltha, J. D., F. H. M. Mikx, and F. J. Kuijpers. 1985. Necrotizingulcerative gingivitis in beagle dogs. III. Distribution of spirochetesin interdental gingival tissues. J. Periodont. Res. 20:522-531.

INFECT. IMMUN.

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

PROLYL OLIGOPEPTIDASE FROM A SPIROCHETE 4947

29. Mentlein, R 1988. Proline residues in the maturation and degra-dation of peptide hormones and neuropeptides. FEBS Lett.234:251-256.

30. Mikx, F. H. M., D. N. B. Ngassapa, F. M. J. ReQntjens, and J. C.Maitha. 1984. Effect of splint replacement on black-pigmentedBacteroides and spirochetes in the dental plaque of beagle dog. J.Dent. Res. 63:1284-1288.

31. Mikx, F. H. M., and M. H. De Jong. 1987. Keratinolytic activity ofcutaneous and oral bacteria. Infect. Immun. 55:621-625.

32. Mikx, F. H. M., J. C. Maltha, and G. J. Campen. 1990. Spirochetesin early lesions of necrotizing ulcerative gingivitis experimentallyinduced in beagles. Oral Microbiol. Immunol. 5:86-89.

33. Miles, E. W. 1977. Modification of histidyl residues in proteins bydiethylpyrocarbonate. Methods Enzymol. 47:431 442.

34. Moore, W. E. C., L V. Holdeman, E. P. Cato, R M. Smibert, J. A.Burmeister, K. G. Palcanis, and R R Ranney. 1985. Comparativebacteriology of juvenile periodontitis. Infect. Immun. 48:507-519.

35. Ohta, K., K. K. Makinen, and W. J. Loesche. 1986. Purificationand characterization of an enzyme produced by Treponema denti-cola capable of hydrolyzing synthetic trypsin substrates. Infect.Immun. 53:213-220.

36. Olsen, L 1984. Attachment of Treponema denticola to culturedhuman epithelial cells. Scand. J. Dent. Res. 92:55-63.

37. PC/GENE User and Reference Manual. 1991. Release 6.5, Febru-ary 1991. IntelliGenetics, Inc., Mountain View, Calif.

38. Polgar, L 1991. pH-dependent mechanism in the catalysis ofprolyl endopeptidase from pig muscle. Eur. J. Biochem. 197:441-447.

39. Pougeois, R, M. Satre, and P. V. Vignais. 1978. N-Ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline, a new inhibitor of the mitochon-drial F1-ATPase. Biochemistry 17:3018-3023.

40. Rawlings, N. D., and A. J. Barrett. 1993. Evolutionary families ofpeptidases. Biochem. J. 290:205-218.

41. Rawlings, N. D., L. Polgar, and A. J. Barrett. 1991. A new familyof serine-type peptidases related to prolyl oligopeptidase. Bio-chem. J. 279:907-911.

42. Saccomani, G., M. L Bercellona, and G. Sachs. 1981. Reactivity ofgastric (H' + K+)-ATPase to N-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline. J. Biol. Chem. 256:12405-12410.

43. 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.

44. Shah, H. N., S. E. Gharbia, and M. I. N. Zhang. 1993. Measure-ment of electrical bioimpedance for studying utilization of aminoacids and peptides by Porphyromonas gingivalis, Fusobacteriumnucleatum, and Treponema denticola. Clin. Infect. Dis. 16(Suppl.4):S404-S407.

45. Simonson, L G., C. H. Goodman, J. J. Bial, and H. E. Morton.1988. Quantitative relationship of Treponema denticola to severityof periodontal disease. Infect. Immun. 56:726-728.

46. Sorsa, T., T. Ingman, K. Suomalainen, M. Haapasalo, Y. T.Konttinen, 0. Lindby, H. Saari, and V.-J. Uitto. 1992. Identifica-tion of proteases from periodontopathogenic bacteria as activators

of latent human neutrophil and fibroblast-type interstitial colla-genases. Infect. Immun. 60:4491-4495.

47. Szwajcer, E., J. Rasmussen, M. Meldal, and K. Breddam. 1992.Proline-specific endopeptidases from microbial sources: isolationof an enzyme of a Xanthomonas sp. J. Bacteriol. 174:2454-2459.

48. Uitto, V.-J., M. Haapasalo, T. Laakso, and T. Salo. 1988. Degra-dation of basement membrane collagen by proteases from someanaerobic oral microorganism. Oral Microbiol. Immunol. 3:97-102.

49. Walter, R, W. H. Simmons, and T. Yoshimoto. 1980. Proline-specific endo- and exopeptidases. Mol. Cell. Biochem. 30:111-126.

50. Weinberg, A., and S. C. Holt. 1990. Interaction of Treponemadenticola TD-4, GM-1, and MS25 with human gingival fibroblasts.Infect. Immun. 58:1720-1729.

51. Welches, W. R, K. B. Brosnihan, and C. M. Ferrario. 1993. Acomparison of the properties and enzymatic activities of 3 angio-tensin-processing enzymes-angiotensin-converting enzyme, prolylendopeptidase and neutral endopeptidase 24.11. Life Sci. 52:1461-1480.

52. Welches, W. R, R A. S. Santos, M. C. Chappel, K. B. Brosnihan,L. J. Greene, and C. M. Ferrario. 1991. Evidence that prolylendopeptidase participates in the processing of brain angiotensin.J. Hypertension 9:631-638.

53. Wilk, S. 1983. Minireview: prolyl endopeptidase. Life Sci. 33:2149-2157.

54. Wrigley, C. W., and H. L. Webster. 1968. Spectrophotometricestimation of protein in presence of ultraviolet-absorbing impuri-ties. J. Chromatogr. 33:534-536.

55. Wyss, C. 1993. Aspartame as a source of essential phenylalaninefor the growth of oral anaerobes. FEMS Microbiol. Lett. 108:255-258.

56. Yajima, H., K. Kitagawa, and T. Segawa. 1973. Studies onpeptides. XXXVIII. Structure-activity correlations in substance P.Chem. Pharm. Bull. (Tokyo) 21:2500-2506.

57. Yaron, A., and F. Neider. 1993. Proline-dependent structural andbiological properties of peptides and proteins. Crit. Rev. Biochem.Mol. Biol. 28:31-81.

58. Yoshimoto, T., and D. Tsuru. 1978. Proline specific endopeptidasefrom Flavobactenium. Agric. Biol. Chem. 42:2417-2419.

59. Yoshimoto, T., W. H. Simmons, T. Kita, and D. Tsuru. 1981.Post-proline cleaving enzyme from lamb brain. J. Biochem. 90:325-334.

60. Yoshimoto, T., K. Kado, F. Matsubara, N. Koriyma, H. Kaneto,and D. Tsuru. 1987. Specific inhibitors for prolyl endopeptidaseand their anti-amnesic effect. J. Pharmacobio-Dyn. 10:730-735.

61. Yoshimoto, T., A. Kanatani, T. Shimoda, T. Inaoka, T. Kokubo,and D. Tsuru. 1991. Prolyl endopeptidase from Flavobacteriummeningosepticum: cloning and sequencing of the enzyme gene. J.Biochem. 110:873-878.

62. Zolfaghari, R, C. R F. Baker, P. C. Canizaro, M. Feola, A.Amirgholami, and F. J. Behal. 1986. Human lung post-prolineendopeptidase: purification and action on vasoactive peptides.Enzyme 36:165-178.

VOL. 62, 1994

on Novem

ber 21, 2020 by guesthttp://iai.asm

.org/D

ownloaded from