Biochem. J. (1995) 309, 209-214 (Printed in Great Britain)

Characterization of an exocellular serine-thiol proteinase activity inParacoccidioides brasiliensisAdriana K. CARMONA,* Rosana PUCCIA,t Maria C. F. OLIVEIRA,* Elaine G. RODRIGUES,t Luiz JULIANO*and Luiz R. TRAVASSOStt*Departmento de Biofisica and tDisciplina de Biologia Celular, Escola Paulista de Medicina, Rua Botucatu 862, 82 andar, 04023-062 Sao Paulo, SP Brazil

An exocellular proteinase activity has been characterized inParacoccidioides brasiliensis culture filtrates. Chromatographicanalysis showed that the activity was eluted from an anion-exchange Resource Q column at 0.08-0.1 M NaCl, and by gelfiltration near ovalbumin elution, in a single peak. Purification ofthe proteinase, however, was hampered by the low protein yield,in contrast to the high peptidase activity. Numerous chromogenicpeptidyl p-nitroanilide derivatives and internally quenchedfluorescent peptides, flanked by Abz (O-aminobenzoyl) andEDDnp (ethylenediaminedinitrophenyl), were tested as sub-strates. Cleavage was observed with Abz-MKRLTL-EDDnp,Abz-FRLVR-EDDnp, and Abz-PLGLLGR-EDDnp at Leu-Thr, Leu-Val and Leu-Leu/Leu-Gly bonds respectively as de-termined by isolation of the corresponding fragments by HPLC.

INTRODUCTIONParacoccidioides brasiliensis is a pathogenic fungus that causes

paracoccidioidomycosis in man, a potentially lethal systemicfungal disease that is prevalent in Latin America. It is a dimorphicfungus which grows in the mycelial form at room temperature,and in the yeast form at 37 °C or in infected tissues. The factorsinvolved in the host-parasite relationship leading to the mul-tiplicity of clinical forms in paracoccidioidomycosis are stillpoorly understood. Structural a-1,3-glucans, the major com-

ponents of the yeast phase cell wall [1], as well as gp43, recentlyimplicated as a cell surface binding receptor for laminin [2], havebeen suggested as virulence factors. gp43, a major glycoproteinsecreted in P. brasiliensis culture supernatants [3], is the mainserum diagnostic antigen of paracoccidioidomycosis [4,5] and isalso involved in the cellular immune reactivity [6].

Exocellular proteinases of micro-organisms have been corre-

lated with the pathogenicity of parasites [7] and fungi [8,9].Preliminary results suggested that the gp43 component mighthave a caseinolytic activity at pH 5.5-6.0 [10]. In the presentwork we describe a thiol-containing serine proteinase activity inP. brasiliensis culture filtrates which partially associates withother exocellular components of the fungal metabolism includingthe gp43 diagnostic antigen.

MATERIALS AND METHODS

MaterialsP. brasiliensis strain 339 was kindly provided by Dr. AngelaRestrepo-Moreno (Medellin, Colombia). Chromogenic peptidyl-

Leucine at P1 seemed to be restrictive for the activity of theexocellular enzyme, but threonine (P'1) and leucine (P'2) in Abz-MKRLTL-EDDnp apparently were not essential. Also, a pair ofalanines could substitute for lysine (P3) and arginine (P2) in thissubstrate, with a decrease in the Km values. The exocellularpeptidase activity of P. brasiliensis had an optimum pH of > 9.0and was irreversibly inhibited by PMSF, mercuric acetate andp-hydroxymercuribenzoate. Inhibition of the mercuriate com-pounds could be partially reversed by Cys/EDTA. E-64 [trans-epoxysuccinyl-L-leucylamido-(4-guanido)butene] was a weak andreversible inhibitor, whereas EDTA and pepstatin were notinhibitory. These results suggest that P. brasiliensis exocellularenzyme belongs to the subfamily of SH-containing serineproteinases.

p-nitroanilides (peptidyl-pNa) were synthesized, purified andcharacterized as previously described [11]. Internally quenchedfluorescent peptides were synthesized by the classical solutionmethods [12,13] using 2-aminobenzoic acid (Abz) as fluorescentgroup and ethylenediamino-2,4-dinitrophenyl (EDDnp) asfluorescence quencher, attached respectively to the N- and C-terminal groups of the peptide. Fluorescence measurements werecarried out in a Hitachi 2000 fluorimeter. Casein, Hide PowerAzure, insulin 8-chain, trans-epoxysuccinyl-L-leucylamido-(4-guanido)butene (E-64), PMSF, sodium 7-hydroxymercuri-benzoate (p-HMB) and glucose were from Sigma, St. Louis,MO, U.S.A. N-Ethylmaleimide (NEM), l-chloro-4-phenyl-3-L-tosylamidobutan-2-one (Tos-Phe-CH2Cl; 'TPCK'), 7-amino-l-chloro-3-L-tosylamidoheptan-2-one (Tos-Lys-CH2Cl; 'TLCK')and iodoacetamide were from Fluka Chemie AG, Buchs,Switzerland. Superose 12 (HR 30-10 column), Resource Q andprotein standards for gel filtration were from Pharmacia LKBBiotechnology, Uppsala, Sweden, and Affi-Gel 10 was from Bio-Rad Laboratories, Richmond, CA, U.S.A. Casein peptone wasfrom Biobras, Montes Claros, MG, Brazil, ammonium sulphatefrom Merck, Darmstadt, Germany, and yeast extract from DifcoLaboratories, Detroit, MI, U.S.A. All other chemicals were ofreagent grade.

Enzyme preparation for characterization of the proteolytic activityP. brasiliensis was cultivated for 7 days at 35 IC, with shaking, inred tomato juice and casein hydrolysate-enriched TOM medium[14]. The supernatant fluid (bearing peptidase activity against

Abbreviations used: Abz, O-aminobenzoyl; Ac, acetyl; ACN, acetonitrile; Bz, benzyl; DSP, depleted supernatant proteinase; EDDnp,ethylenediaminedinitrophenyl; p-HMB, sodium 7-hydroxymercuribenzoate; NEM, N-ethylmaleimide; pNa, p-nitroanilide; Suc, succinyl; TFA,trifluoroacetic acid; Tos-Lys-CH2C0, 7-amino-i-chloro-3-L-tosylamidoheptan-2-one (TLCK); Tos-Phe-CH2CI, 1-chloro-4-phenyl-3-L-tosylamidobutan-2-one ('TPCK'); Z, benzyloxycarbonyl; E-64, trans-epoxysuccinyl-L-leucylamido-(4-guanido)butene.tTo whom correspondence should be addressed.

209

210 A. K. Carmona and others

Abz-MKRLTL-EDDnp) was paper-filtered, dialysed againstPBS and depleted of gp43 by affinity chromatography in Affi-Gel10 with anti-gp43 murine monoclonal antibody 17c [15]. Removalofthis major exocellular component increased the specific activityof the proteinase. This preparation was referred to as DSP(depleted supernatant proteinase) and was used to determinesubstrate specificity, Km and inhibitor activity.

Chromatographic characterization of the proteolytic activityCulture supernatants were precipitated with ammonium sulphateat 40% final concentration (with stirring for 2 h) and centrifuged(16300 g, 30 min). The supernatant was reprecipitated with 50 %ammonium sulphate and stirred overnight. The precipitateobtained by centrifugation (16300 g, 30 min) was solubilized in1 mM phosphate buffer containing 0.015 M NaCl and 10%ethylene glycol (50 ml), concentrated 10-fold under N2 in theAmicon cell with a PM10 Diaflo membrane, and the procedurerepeated once in order to dialyse the ammonium sulphate.Desalted solutions were depleted of gp43 by affinity binding asdescribed above and the unbound fraction was concentrated to2-3 ml in Amicon. This enzyme preparation was fractionated byion-exchange chromatography in a Resource Q column (FPLCsystem; 1 ml) equilibrated with 1 mM phosphate buffer con-taining 0.015 M NaCl and 10% ethylene glycol, and eluted witha NaCl gradient (0.015-0.5 M) in starting buffer. The proteinelution profile was monitored by UV absorbance (280 nm) andthe fractions (1 ml) were assayed for peptidase activity againstAbz-MKRLTL-EDDnp. Those fractions corresponding to thepeak of proteinase activity were pooled, concentrated in Amiconand mixed with 2 mM p-HMB to reversibly inhibit proteinaseactivity and prevent self-degradation. The sample was thenchromatographed in a calibrated Superose 12 column (FPLCsystem; 1.0 cm x 30 cm), equilibrated and eluted with PBS con-taining 10% ethylene glycol and 2 mM p-HMB. The proteinelution profile was monitored by UV absorbance (280 nm).Fractions (0.5 ml) were assayed for peptidase activity in thepresence of 2 mM cysteine and 1 mM EDTA, as describedelsewhere [16]. The molecular mass of eluted components wasdetermined in comparison with those of a mixture of standards:aldolase (158 kDa), BSA (67 kDa), egg ovalbumin (43 kDa),chymotrypsinogen A (25 kDa) and ribonuclease A (13.7 kDa).Processing of the enzyme preparations was always carried out at4 °C, except for FPLC chromatography which was carried outat room temperature.

Proteolytic activitiesCasein was tested as substrate with DSP at 37 °C for 30 min,both at pH 5.0 (in 0.1 M sodium acetate) and at pH 8.0 (in0.1 M Tris/HCl) by the method of Kunitz [17]. Hide PowerAzure was assayed under the same conditions for 40 min, as des-cribed [18]. Insulin f-chain was assayed as described for casein for14 h, and samples were analysed by HPLC using a Vydac C18column (5 gcm; 4.6 mm x 250 mm; flow rate 1.0 ml/min), elutedwith solvents A [trifluoroacetic acid (TFA)/water; 1: 1000, v/v]and B [TFA/acetonitrile (ACN); 1: 1000, v/v] using the followinggradient: 5 min with 0% B, 60 min with 0-60% B, 5 min with60-100% B and an additional 5 min with 100% B. Fractionswere monitored by UV absorbance at 215 nm.

Pepildase activitiesQuenched fluorescent peptides (10 1sM) were tested as substrateswith DSP in a spectrofluorimeter cuvette at 37 °C in 0.1 Msodium acetate, pH 5.0, and in 0.1 M Tris/HCl, pH 8.0 (2.0 ml

final volume). Fluorescence was measured continuously for10 min at Aem = 420 nm and Aex = 320 nm. The curve slopewas converted into #,smol based on a calibration curve obtainedfor each substrate from the corresponding synthetic fluorescentfragment released by hydrolysis. The proteolytic activity wasexpressed as the rate ofrelease offluorescent fragment (umol/minper mg of protein). Chromogenic peptidyl-pNa substrates wereassayed as previously described [11]. Stock solutions of thepeptidyl-pNa were prepared as DMSO (50 %, v/v) and aliquotswere added to solutions containing buffer (0.1 M sodium acetate,pH 5.0, or 0.1 M Tris/HCl, pH 8.0) and DSP at 37 'C. Theabsorbance at 410 nm was followed for at least 10 min.

K, determinationsKinetic parameters for the hydrolysis of internally quenchedpositive peptide substrates were obtained at 37 'C, pH 9.0, in0.1 M Tris/HCl buffer as described [12,13]. At least six differentsubstrate concentrations below 10 ,uM were used. Higher concen-trations were not used, to avoid filter effects. The kinetics datawere analysed using the GraFit program [19].

Proteinase inhibitorsDSP was pre-incubated with the potential inhibitors for differentperiods of time, depending on each inhibitor, and then incubatedwith 10 #M Abz-MKRLTL-EDDnp as substrate. The resultswere recorded as the percentage of residual activity relative tocontrol reactions run simultaneously in the absence of theinhibitor. For determination of irreversible inhibition, the re-sidual enzymic activity was measured as described above, after a5 min preincubation with various concentrations of inhibitor.The K, values were obtained by analysing the experimental datausing the Morrison (Enzfitter; Leatherbarrow; Elsevier-Biosoft)non-linear regression.

Determination of substrate cleavage siteThe fragments resulting from hydrolysis of the internallyquenched fluorescent peptides with DSP at 37 'C, pH 9.0, in0.1 M Tris/HCl buffer were isolated by HPLC in an UltrasphereCl. column (5,um; 4.6 mm x 150 mm; flow rate 1.0 ml/min)eluted with solvents (as before) and C (TFA/ACN/water; 1: 900:100, by vol.) in a 10-80% gradient for 15 min. Fractions weremonitored by UV absorbance at 220 nm and fluorescence atAem = 420 nm and Aex. = 320 nm. The scissile bond(s) of hydro-lysed peptides were determined by isolation of the fragmentsby HPLC, and their structures were deduced both from aminoacid composition and by comparison of their HPLC profiles withthose of the synthesized authentic fragments.

Determination of optimum pHThe activity ofDSP at various pH values was measured at 37 'Cby the fluorimetric assay described above, using 10 #M Abz-MKALTL-EDDnp as substrate. The buffers used were asfollows: 0.1 M sodium formate (pH < 4.0), 0.1 M sodium acetate(4.0 < pH < 5.5),0.1 M sodium phosphate (5.5 < pH < 8.0) and0.1 M Tris/HCl (pH > 8.0).

Analytical methodsProtein was assayed by Bradford's method [20], using BSA asstandard.

Exocellular proteinase of Paracoccidioides brasiliensis

RESULTSDialysed 7-day-old culture supernatants of P. brasiliensis grownin TOM medium [14] had a specific activity of 6.2 x 10-3 gmol/min per mg using Abz-MKRLTL-EDDnp as substrate. Puri-fied gp43 preparations and culture supernatants depleted ofgp43 gave specific activities of 1.3 x 10-5 jmol/min per mg and8.2 x 10-3 jmol/min per mg respectively, indicating that most ofthe exocellular peptidase activity for the internally quenchedsubstrate remained in the gp43-depleted culture supernatant(DSP). Abz-MKRLTL-EDDnp was specifically hydrolysed onlyat the Leu-Thr bond by DSP (Figure 1).The first step in the fractionation of the proteolytic activity

against Abz-MKRLTL-EDDnp involved ammonium sulphateprecipitation of whole culture supernatant fluids (2 litres). At40-50% ammonium sulphate saturation, the peptidase activitywas enriched in the precipitated material. This was subsequentlydepleted of the remaining gp43 in an anti-gp43 immunoaffinitycolumn as described. The proteolytic activity concentrated in thegp43-free 40-50% ammonium sulphate precipitate was elutedfrom the Resource Q anion-exchange resin at 0.08-0.1 M NaCl(Figure 2a). Gel filtration of the concentrated active fractionsresulted in elution as a single peak near ovalbumin in a calibratedSuperose 12 column (Figure 2b). These results and the presence

of a single scissile bond in the Abz-MKRLTL-EDDnp substrateare in accordance with a single exocellular proteinase activity inP. brasiliensis culture supernatants.

It is noticeable in both chromatographic profiles that, althoughthe proteolytic activity was very high, the corresponding proteincontent was negligible, indicating a proteinase of very highspecific activity. SDS/PAGE analyses of the more highly purifiedfractions have so far been inconclusive in determining the correctsize of the enzyme.

Preliminary results have also shown that the proteo-

(a)

Abz-MKRLTL-EDDnp

(b) Abz-MKRL

2000 -,

*E 1000 -,U-

0 20 40

158 67 43 25 13.7 kDa

800 F(b) ,

E

U-

U-.52cJ1

600

400

200

0

0.05

0.020

0.015

O0.010 CN4

0.005

o

Fraction no.

Figure 2 Chromatographic analysis of the proteolytic activity

(a) Anion-exchange chromatography on Resource Q (FPLC system) of solubilized gp43-depletedammonium sulphate precipitate of P. brasiliensis culture filtrate. The bound material was elutedwith a linear gradient of 0.015-0.5 M NaCI (----) and assayed for peptidase activity atpH 9.0 as described in the Materials and methods section. Active fractions were pooled,concentrated in Amicon and chromatographed by gel filtration in Superose 12 (FPLC system)in the presence of 2 mM pHMB (b). 0, A280; *, peptidase activity [arbitrary fluorescenceunits (AFU)/min] against Abz-MKRLTL-EDDnp, assayed in the presence of 2 mM cysteine and1 mM EDTA when eluted from Superose 12. The positions of the molecular mass markers areindicated. Further details are given in the Materials and methods section.

lytic activity against Abz-MKRLTL-EDDnp was detected inaggregates with other components ofthe culture supernatant fluidfollowing concentration. The addition of 10% ethylene glycol tothe enzyme samples throughout fractionation has apparently pre-vented aggregation. In addition, partially purified preparationseluted from the anion-exchange column were inhibited withmercuriates to prevent self-degradation.

Proteolytic and peptidase activItles_______________________, , ,,_,_ _ Casein and Hide Power Azure were weakly hydrolysed by the0 2 4 6 8 10 12 14 16 P. brasiliensis DSP at pH 5.0 and 8.0, whereas no hydrolysis

lime (min) of oxidized insulin fl-chain was observed at either pH, even after14 h of incubation.

In order to verify the specificity of the proteolytic activityigure 1 Determination of substrate cleavage sfte of DSP, several peptides (Table 1) were tested as substrates

IPLC profile of Abz-MKRLTL-EDDnp before (----) and after ( ) hydrolysis by at pH 5.0 and 8.0. The first group consisted of chromogenic' brasiliensis DSP. (a) Absorbance at 280 nm; (b) fluorescence at A.,= 320 nm aminoacyl-and peptidyl-pNas, which have been used as substratesnd Aem = 420 nm. The peptides corresponding to each peak are indicated. for trypsin-like (peptides 2 and 3), elastase-like (peptides 1, 4

F

HFar

211

II

II

I

iII

212 A. K. Carmona and others

Table 1 SyntheIc peptldes tested as substrates for DSP fom P. brasilllnsisculture flItrates: determinaon of the cleavage bond In fluorogenic peptides

Peptide bondPeptide no. Substrate hydrolysed

Chromogenic peptides12345678

Internally quenchedfluorogenic peptides

910111213141516171819202122

EQL.C

0

.E

Ac-A-pNaAc-FR-pNa(D)VLK-pNaSuc-AAA-pNaZ-ML-pNaBz-KAL-pNaBz-GHL-pNaDRVYIHPF-pNa

Abz-MKRLTL-EDDnpAbz-MRRLTL-EDDnpAbz-MKKLTL-EDDnpAbz-MKALTL-EDDnpAbz-RMKRLTL-EDDnpAbz-RMARLTL-EDDnpAbz-RMAALTL-EDDnpAbz-FRLVR-EDDnpAbz-PLGLLGR-EDDnpAbz-GGFLRRV-EDDnpAbz-LMKRP-EDDnpAbz-FHLVIH-EDDnpAbz-RPFHLVIH-EDDnpAbz-YIHPFHLVIH-EDDnp

2 4 6pH

FRgure 3 pH-dependence of proteolytlc actvlty

P. brasiliensis DSP activity against Abz-MKALTL-EDDnp was deteriby the fluorometric method described in the Materials and methoK

and 5) and chymotrypsin-like (peptides 7 and 8these chromogenic peptides was hydrolysed tThe second group of substrates tested for hi

of internally quenched fluorescent peptides,laboratory for different purposes (Table 1, peof them (peptides 9-17) were hydrolysed. Thehydrolysed peptides were determined as descritand methods section for Abz-MKRLTL-E]This peptide and its analogues (peptides 10-1.at the Leu-Thr bond by the DSP. As Abz-t

Table 2 KIneIc parameters for the hydrolysis of Internally quenchedfluorogenic peptides by DSP of P. brasillensis culture flltratesHydrolysis conditions were 37 0C, 0.1 M Tris/HCI buffer (pH 9.0). Values are means+ S.D.

Peptide Km (1M/min l(xlKmno. Substrate (/AM) per mg) (min mg)g

9 Abz-MKRLTL-EDDnp 11.4 + 0.9 29.2 + 0.9 2.610 Abz-MRRLTL-EDDnp 11.0+2.0 46.0+ 0.1 4.211 Abz-MKKLTL-EDDnp 3.5+0.5 24.0+ 0.1 6.912 Abz-MKALTL-EDDnp 4.7+0.3 60.0+0.1 12.813 Abz-RMKRLTL-EDDnp 8.7+0.6 40.7+0.3 4.714 Abz-RMARLTL-EDDnp 8.4±0.4 36.7+0.3 4.415 Abz-RMMLTL-EDDnp 0.9 +0.1 14.0+0.2 15.516 Abz-FRLVR-EDDnp 8.9 +0.1 2.0+0.5 2.3

Leu-ThrLeu-ThrLeu-ThrLeu-Thraii-Thr

Leu-Thr was the first peptide substrate tested which was cleaved by theLeu-Thr proteinase, analogues of it were synthesized in order to establishLeu-Val the importance of the basic pair at the P2 and P3 positions for theLeu-Leu, Leu-Gly enzyme activity.

Abz-FRLVR-EDDnp (peptide 16), a sequence of ratkininogen, was cleaved at the Leu-Val bond, and Abz-PLGLLGR-EDDnp (peptide 17), a substrate for collagenases ofmammalian sources, was very poorly hydrolysed at the Leu-Leuand Leu-Gly bonds. Peptides related to endorphins (peptide 18),human kininogen (peptide 19) and human angiotensinogen(peptides 20-22) were not substrates for the P. brasiliensis DSP.

Figure 3 shows the effect of pH on the rate of hydrolysis ofAbz-MKALTL-EDDnp by DSP. This substrate was used in thisexperiment in order to obtain measurable fluorescence values atthe acidic pHs where the activity is usually low. There is a smallpeak of activity at pH 5.0-5.5, but the maximum values are

obtained at pH values above 8.0. Considering that the activitywas still increasing at pH 9.5 and that the fluorescence readingsat higher pHs are not reliable, owing to fluorogenic peptidedegradation, we could not determine the optimum pH of P.brasiliensis exocellular proteinase activity. Determination of theKm and K1 values, as well as testing of inhibitors, was carried outat pH 9.0.The kinetic parameters of hydrolysis were determined for the

susceptible peptides with one cleavage size (Table 1, peptides8 10 9-16). Hydrolysis was continuously monitored in a spectro-

fluorimeter and the data fitted to Michaelian kinetics. Table 2shows the values ofthe kinetic constants. TakingAbz-MKRLTL-EDDnp for comparison (peptide 9, Table 2), the presence ofArg-Arg at positions P2 and P3 in peptide 10 kept the Km value

mined at various pH values unchanged and barely increased the VmJK,,. On the other hand, theds section. presence of a Lys-Lys basic pair in the same positions (peptide

11) reduced the Km value by one order of magnitude value, butthe Vm,, also decreased. Elongation of peptide 9 with anadditional arginine at the N-terminus (peptide 13) did not affect

enzymes. None of the peptide's susceptibility to hydrolysis. Alanine at P2 instead ofay DSP. arginine (peptides 12 and 15) resulted in higher Vm,' /Km values,ydrolysis consisted and the resulting peptides were, therefore, the best substratessynthesized in our tested in this work. Peptide 14, in which alanine replaced lysineptides 9-22). Nine at P3., showed kinetic parameters very similar to those obtainedscissile bond(s) of with peptide 9. Replacement by alanine of both basic amino)ed in the Materials acids at positions P2 and P3 was only possible with the additionDDnp (Figure 1). of arginine at the N-terminal position, to provide solubility.5) were hydrolysed Abz-FRLVR-EDDnp (peptide 16) was also cleaved at the LeuAKRLTL-EDDnp carboxyl bond, but with lower catalytic efficiency.

Exocellular proteinase of Paracoccidioides brasiliensis 213

Table 3 Inhibition of the peptddase activity of DSP from P. braslllensisculture flltrates by common proteinase InhIbitorsExperimental conditions were 0.1 M Tris/HCI buffer, pH 9.0, 37 0C. The substrate used wasAbz-MKRLTL-EDDnp. Fluorescence detection: Aem. = 420 nm; Aex = 320 nm. K; values aremeans + S.D.

Inhibitor Concentration Preincubation Residual K1 (esM)time activity (%)

EDTAPepstatinTos-Phe-CH2CITos-Lys-CH2CIPMSFE-64NEMlodoacetamideMercuric acetatep- HMB

10 mM1 ,tM

0.1 mM0.1 mM0.05 mM0.1 mM0.1 mM0.1 mM2 uM2 ,M

5 h4 h17 h17 h5 min17 h17 h17 h5 min5 min

100 -

100 -

52100 -

0 6+272 -

60 -

790 0.15+0.010 2.3+0.1

Inhibition of the proteolytic activityPotential inhibitors of the P. brasiliensis exocellular proteolyticactivity were assayed for their effects upon hydrolysis of Abz-MKRLTL-EDDnp, and the results can be seen in Table 3.EDTA (10 mM) and pepstatin (1 ,uM) were not inhibitory.lodoacetamide, NEM and the very specific cysteine-proteinaseinhibitor E-64 [21] were poor inhibitors, and the complex formedwith E-64 was completely removed by addition of an excess ofsubstrate. The most effective inhibitors were mercuric acetate,p-HMB and PMSF, which completely abolished the peptidaseactivity at low concentrations, and in less than 5 min. Theinhibition constants (K1) for these compounds were obtained bydetermining the residual activity of P. brasiliensis exocellularproteinase at different concentrations of inhibitors (Table 3). Theinhibition by mercuric acetate and p-HMB could be reversed by40% with 2.0 mM cysteine (higher concentrations inhibited theenzyme activity), while the inhibition by PMSF was irreversible,even after a 4 h incubation in the presence of cysteine, ,-mercaptoethanol or dithiothreitol. This also applied to theenzyme activity obtained by fractionation in Superose 12. Tos-Lys-CH2Cl did not inhibit the exocellular peptidase activity, butTos-Phe-CH2Cl lowered it by almost 50% after a preincubationof 17h.

DISCUSSIONThe present work describes an exocellular serine-thiol proteinaseactivity in the culture supernatants of P. brasiliensis yeast phase.The activity eluted in a single peak after anion-exchange and gel-filtration chromatography, using as starting material solubilizedgp43-depleted 40-50% ammonium sulphate precipitates offungal culture filtrates. The use of 10% ethylene glycol togetherwith reversible inhibitors was necessary to prevent both ag-

gregation of the proteinase with other exocellular componentsand self-degradation of the enzyme during the fractionationprocess. Purification to homogeneity of this proteinase of highspecific activity has not been possible so far due to the lowprotein content of the preparations.

Previous data have shown that partially purified gp43 pre-

parations were able to hydrolyse casein at pH 5.5-6.0 [10], andalso to cleave the internally quenched fluorescent peptide Abz-MKRLTL-EDDnp [15]. The specific activities measured in thegp43 preparations were, however, low in comparison with that in

culture filtrates of P. brasiliensis depleted of the gp43 antigen byimmunoaffinity binding. The proteolytic activity associated withthe gp43 apparently resulted from protein aggregation, with theaggregate containing very small amounts of a potent exocellularproteinase.The proteolytic activity of the culture filtrates depleted of gp43

(DSP) could not be detected on using as substrates insulin ,J-chain or eight different aminoacyl- and peptidyl-pNa derivatives.Even Z-AAL-pNa, a common synthetic substrate for subtilisinswhich is hydrolysed with high specificity by extracellular serineproteases from different micro-organisms [22], was not cleavedby the P. brasiliensis exocellular proteinase. Pretreatment withtrypsin for a short period of time (1 min) and subsequentinhibition by soybean trypsin inhibitor did not increasethe hydrolytic activity (results not shown), thus excludingsecretion of a fungal pro-enzyme that could be activated bytrypsin.Our results indicate that the fungal exocellular enzyme has a

very restrictive substrate specificity, including a leucine at P1 asan important requirement. The internally quenched fluorescentpeptide Abz-MKRLTL-EDDnp, and analogues thereof, werecleaved at the Leu-Thr bond, the Abz-FRLVR-EDDnp de-rivative at the Leu-Val bond and the substrate for mammaliancollagenases Abz-PLGLLGR-EDDnp at the Leu-Leu and Leu-Gly bonds. The presence of threonine and leucine respectively atpositions P'1 and P'2 in Abz-MKRLTL-EDDnp and its analoguesis apparently not essential for the hydrolytic activity, since Abz-FRLVR-EDDnp was still hydrolysed at a significant rate.Cleavage of the Leu-Val bond possibly depends on the neigh-bouring residues, since peptides Abz-FHLVIH-EDDnp, Abz-RPFHLVIH-EDDnp and Abz-YIHPFHLVIH-EDDnp werenot substrates for the fungal proteinase. The basic pair of aminoacids present in Abz-MKRLTL-EDDnp was not essential forhydrolysis, since Abz-MKALTL-EDDnp and Abz-RMAALTL-EDDnp were even better substrates. In addition, the presence ofanother arginine at P5 did not change the enzyme-substrateaffinity of peptide Abz-RMKRLTL-EDDnp. Moreover, theendoproteolytic activity requires amino acid residues at the P'site of the cleavage bond, since Bz-KAL-pNa and Z-AAL-pNawere not hydrolysed, despite containing the same P3 and P1residues as the best internally quenched fluorescent substrates(Abz-MKALTL-EDDnp and Abz-RMAALTL-EDDnp).The exocellular peptidase activity of P. brasiliensis was irre-

versibly inhibited by PMSF, mercuric acetate and p-HMB.Inhibition by the mercuriate compounds could, however, bereversed by Cys/EDTA. E-64 was a weak and reversible inhibitor,whereas EDTA and pepstatin were not inhibitory. Inhibition byPMSF and the basic optimum pH of > 9.0 are suggestive of anexocellular hydrolytic activity similar to serine proteinases. AnSH-dependence, however, was demonstrated by the inhibitionwith mercuriates that was partially reversed by Cys/EDTA.These results suggest that P. brasiliensis exocellular enzymecould belong to the subfamily of SH-containing serine protein-ases, such as humicolin of Thermomyces lanuginosus [23], theextracellular serine proteinase from Thermoactinomyces vulgaris[22,24] and proteinase K [25,26], which are also related to thesubtilisins. All proteinases of this class, already described in theliterature [22-24,27-29], depend for their activities on free thiolgroups. Most of them contain only one cysteine residue, exceptfor humicolin which contains five residues, and proteinase K,which has two disulphide bridges and one free cysteine positionedvery close to the active site [26].The pH-dependence of the enzymic activity shows a small but

significant proteolysis around pH 5.0, whereas the major activitywas obtained at basic pH. This could be related to a modified

214 A. K. Carmona and others

form of the same proteinase or to the property of an aggregatewhich could be active at acidic pH. All the substrates andinhibitors of the present work were assayed at pH 5.0 and 9.0with similar results, and the cleavage sites were identical in bothcases.

In conclusion, we have characterized in P. brasiliensis anexocellular thiol-containing serine proteinase of high specificactivity that may play a role in the pathogenicity of this fungus.Further purification and detailed evaluation of the cleavagespecificity of this enzyme with natural substrates are in progress.

This work was supported by FAPESP and CNPq. We thank Elaine D. N. Santos,Daniel C. Pimenta and Igor C. Almeida for help with graphs and scans.

REFERENCES1 San-Blas, G. (1982) Mycopathologia 79,159-1842 Vicentini, A. P., Gesztesi, J. L., Franco, M. F., Souza, W., Moraes, J. Z., Travassos,

L. R. and Lopes, J. D. (1994) Infect. Immun. 62, 1465-14693 Stambuk, B. U., Puccia, R., Almeida, M. L. C., Travassos, L. R. and Schenkman, S.

(1988) J. Med. Vet. Mycol. 26, 367-3734 Puccia, R., Schenkman, S., Gorin, P. A. J. and Travassos, L. R. (1986) Infect. Immun.

53, 199-2065 Puccia, R. and Travassos, L. R. (1991) J. Clin. Microbiol. 29, 1610-16156 Rodrigues, E. G. and Travassos, L. R. (1994) J. Med. Vet. Mycol. 32, 77-817 McKerrow, J. H. (1993) Cienc. Cult./J. Braz. Assoc. Adv. Sci. 45, 319-3218 RUchel, R., Bernardis, F., Ray, T. L., Sullivan, P. A. and Cole, G. T. (1992) J. Med.

Vet. Mycol. 30,123-1329 Kolattukudy, P. E., Lee, J. D., Rogers, L. M., Zimmerman, P., Ceselski, S., Fox, B.,

Stein, B. and Copelan, E. A. (1993) Infect. Immun. 61, 2357-2368

10 Puccia, R. and Travassos, L. R. (1991) Arch. Biochem. Biophys. 289, 298-30211 Juliano, M. A. and Juliano, L. (1985) Braz. J. Med. Biol. Res. 18, 435-44512 Chagas, J. R., Juliano, L. and Prado, E. S. (1991) Anal. Biochem. 192, 419-42513 Oliveira, M. C. F., Hirata, I. Y., Chagas, J. R., Boschcov, P., Gomes, R. A. S.,

Figueiredo, A. F. S. and Juliano, L. (1992) Anal. Biochem. 203, 39-4614 Puccia, R., Takaoka, D. T. and Travassos, L. R. (1990) J. Med. Vet. Mycol. 29,

57-6015 Puccia, R., Travassos, L. R., Rodrigues, E. G., Carmona, A. K., Oliveira, M. C. F. and

Juliano, L. (1994) in Molecular Biology of Pathogenic Fungi: A Laboratory Manual(Maresca, B. and Kobayashi, G. S., eds.), pp. 507-515, Telos Press, New York

16 Shenolikar, S. and Stevenson, K. J. (1982) Biochem. J. 205, 147-15217 Kunitz, M. (1946) J. Gen. Physiol. 30, 291-31018 Rinderknecht, H., Geokas, M. C., Silverman, P. and Haverback, B. J. (1969) Clin.

Chim. Acta 21,197-20319 Leatherbarrow, R. J. (1992) GraFit Version 3.0, Erithacus Software Ltd., Staines, U.K.20 Bradford, M. M. (1976) Anal. Biochem. 72, 248-25421 Rich, D. L. (1986) in Proteinase Inhibitors (Barrett, A. J. and Salvesen, G., eds.),

pp. 153-178, Elsevier, Amsterdam22 Stepanov, V. M., Chestukhina, G. G., Rudenskaya, G. N., Epremyan, A. S., Osterman,

A. L., Khodova, 0. M. and Belyanova, L. P. (1981) Biochem. Biophys. Res. Commun.100, 1680-1687

23 Hasnain, S., Adeli, K. and Storer, A. C. (1992) Biochem. Cell Biol. 70, 117-12224 Meloun, B., Baudys, M., Kotska, V., Hausdorf, G., Frommel, C. and Hohne, W. E.

(1985) FEBS Left. 183, 195-20025 Ebelin, W., Hennrich, N., Klockow, M., Metz, H., Orth, H. D. and Lang, H. (1974)

Eur. J. Biochem. 47, 91-9726 Betzel, C., Bellemann, M., Pal, G. P., Bajorath, J., Saenger, W. and Wilson, K. S.

(1988) Protein Structure Function Genet. 4,157-16427 Mizusawa, K. and Yoshida, F. (1973) J. Biol. Chem. 248, 4417-442328 Kundu, A. K. and Manna, S. (1975) Appl. Microbiol. 30, 507-51329 Ong, P. S. and Gaucher, G. M. (1973) Can J. Microbiol. 19, 129-133

Received 6 September 1994/14 February 1995; accepted 9 March 1995