Characterization of an exocellular serine-thiol proteinase activity in Paracoccidioides brasiliensis

209

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

Characterization of an exocellular serine-thiol proteinase activity in Paracoccidioides brasiliensis Adriana 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 in Paracoccidioides brasiliensis culture filtrates. Chromatographic analysis showed that the activity was eluted from an anionexchange Resource Q column at 0.08-0.1 M NaCl, and by gel filtration near ovalbumin elution, in a single peak. Purification of the proteinase, however, was hampered by the low protein yield, in contrast to the high peptidase activity. Numerous chromogenic peptidyl p-nitroanilide derivatives and internally quenched fluorescent peptides, flanked by Abz (O-aminobenzoyl) and EDDnp (ethylenediaminedinitrophenyl), were tested as substrates. Cleavage was observed with Abz-MKRLTL-EDDnp, Abz-FRLVR-EDDnp, and Abz-PLGLLGR-EDDnp at LeuThr, Leu-Val and Leu-Leu/Leu-Gly bonds respectively as determined by isolation of the corresponding fragments by HPLC.

Leucine at P1 seemed to be restrictive for the activity of the exocellular enzyme, but threonine (P'1) and leucine (P'2) in AbzMKRLTL-EDDnp apparently were not essential. Also, a pair of alanines could substitute for lysine (P3) and arginine (P2) in this substrate, with a decrease in the Km values. The exocellular peptidase activity of P. brasiliensis had an optimum pH of > 9.0 and was irreversibly inhibited by PMSF, mercuric acetate and p-hydroxymercuribenzoate. Inhibition of the mercuriate compounds could be partially reversed by Cys/EDTA. E-64 [transepoxysuccinyl-L-leucylamido-(4-guanido)butene] was a weak and reversible inhibitor, whereas EDTA and pepstatin were not inhibitory. These results suggest that P. brasiliensis exocellular enzyme belongs to the subfamily of SH-containing serine proteinases.

INTRODUCTION Paracoccidioides brasiliensis is a pathogenic fungus that causes paracoccidioidomycosis in man, a potentially lethal systemic fungal disease that is prevalent in Latin America. It is a dimorphic fungus which grows in the mycelial form at room temperature, and in the yeast form at 37 °C or in infected tissues. The factors involved in the host-parasite relationship leading to the multiplicity of clinical forms in paracoccidioidomycosis are still poorly understood. Structural a-1,3-glucans, the major components of the yeast phase cell wall [1], as well as gp43, recently implicated as a cell surface binding receptor for laminin [2], have been suggested as virulence factors. gp43, a major glycoprotein secreted in P. brasiliensis culture supernatants [3], is the main serum diagnostic antigen of paracoccidioidomycosis [4,5] and is also involved in the cellular immune reactivity [6]. Exocellular proteinases of micro-organisms have been correlated with the pathogenicity of parasites [7] and fungi [8,9]. Preliminary results suggested that the gp43 component might have a caseinolytic activity at pH 5.5-6.0 [10]. In the present work we describe a thiol-containing serine proteinase activity in P. brasiliensis culture filtrates which partially associates with other exocellular components of the fungal metabolism including the gp43 diagnostic antigen.

p-nitroanilides (peptidyl-pNa) were synthesized, purified and characterized as previously described [11]. Internally quenched fluorescent peptides were synthesized by the classical solution methods [12,13] using 2-aminobenzoic acid (Abz) as fluorescent group and ethylenediamino-2,4-dinitrophenyl (EDDnp) as fluorescence quencher, attached respectively to the N- and Cterminal groups of the peptide. Fluorescence measurements were carried out in a Hitachi 2000 fluorimeter. Casein, Hide Power Azure, insulin 8-chain, trans-epoxysuccinyl-L-leucylamido(4-guanido)butene (E-64), PMSF, sodium 7-hydroxymercuribenzoate (p-HMB) and glucose were from Sigma, St. Louis, MO, U.S.A. N-Ethylmaleimide (NEM), l-chloro-4-phenyl-3-Ltosylamidobutan-2-one (Tos-Phe-CH2Cl; 'TPCK'), 7-amino-lchloro-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 and protein standards for gel filtration were from Pharmacia LKB Biotechnology, Uppsala, Sweden, and Affi-Gel 10 was from BioRad Laboratories, Richmond, CA, U.S.A. Casein peptone was from Biobras, Montes Claros, MG, Brazil, ammonium sulphate from Merck, Darmstadt, Germany, and yeast extract from Difco Laboratories, Detroit, MI, U.S.A. All other chemicals were of reagent grade.

MATERIALS AND METHODS Materials P. brasiliensis strain 339 was kindly provided by Dr. Angela Restrepo-Moreno (Medellin, Colombia). Chromogenic peptidyl-

P. brasiliensis was cultivated for 7 days at 35 IC, with shaking, in red tomato juice and casein hydrolysate-enriched TOM medium

Enzyme preparation for characterization of the proteolytic activity [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-2one ('TPCK'); Z, benzyloxycarbonyl; E-64, trans-epoxysuccinyl-L-leucylamido-(4-guanido)butene. tTo whom correspondence should be addressed.

210

A. K. Carmona and others

Abz-MKRLTL-EDDnp) was paper-filtered, dialysed against PBS and depleted of gp43 by affinity chromatography in Affi-Gel 10 with anti-gp43 murine monoclonal antibody 17c [15]. Removal ofthis major exocellular component increased the specific activity of the proteinase. This preparation was referred to as DSP (depleted supernatant proteinase) and was used to determine substrate specificity, Km and inhibitor activity.

Chromatographic characterization of the proteolytic activity Culture supernatants were precipitated with ammonium sulphate at 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 precipitate obtained by centrifugation (16300 g, 30 min) was solubilized in 1 mM phosphate buffer containing 0.015 M NaCl and 10% ethylene glycol (50 ml), concentrated 10-fold under N2 in the Amicon cell with a PM10 Diaflo membrane, and the procedure repeated once in order to dialyse the ammonium sulphate. Desalted solutions were depleted of gp43 by affinity binding as described above and the unbound fraction was concentrated to 2-3 ml in Amicon. This enzyme preparation was fractionated by ion-exchange chromatography in a Resource Q column (FPLC system; 1 ml) equilibrated with 1 mM phosphate buffer containing 0.015 M NaCl and 10 % ethylene glycol, and eluted with a NaCl gradient (0.015-0.5 M) in starting buffer. The protein elution profile was monitored by UV absorbance (280 nm) and the fractions (1 ml) were assayed for peptidase activity against Abz-MKRLTL-EDDnp. Those fractions corresponding to the peak of proteinase activity were pooled, concentrated in Amicon and mixed with 2 mM p-HMB to reversibly inhibit proteinase activity and prevent self-degradation. The sample was then chromatographed in a calibrated Superose 12 column (FPLC system; 1.0 cm x 30 cm), equilibrated and eluted with PBS containing 10% ethylene glycol and 2 mM p-HMB. The protein elution profile was monitored by UV absorbance (280 nm). Fractions (0.5 ml) were assayed for peptidase activity in the presence of 2 mM cysteine and 1 mM EDTA, as described elsewhere [16]. The molecular mass of eluted components was determined 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 at 4 °C, except for FPLC chromatography which was carried out at room temperature.

Proteolytic activities Casein 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 (in 0.1 M Tris/HCl) by the method of Kunitz [17]. Hide Power Azure was assayed under the same conditions for 40 min, as described [18]. Insulin f-chain was assayed as described for casein for 14 h, and samples were analysed by HPLC using a Vydac C18 column (5 gcm; 4.6 mm x 250 mm; flow rate 1.0 ml/min), eluted with solvents A [trifluoroacetic acid (TFA)/water; 1: 1000, v/v] and B [TFA/acetonitrile (ACN); 1: 1000, v/v] using the following gradient: 5 min with 0 % B, 60 min with 0-60 % B, 5 min with 60-100% B and an additional 5 min with 100% B. Fractions were monitored by UV absorbance at 215 nm.

final volume). Fluorescence was measured continuously for 10 min at Aem = 420 nm and Aex = 320 nm. The curve slope was converted into #,smol based on a calibration curve obtained for each substrate from the corresponding synthetic fluorescent fragment released by hydrolysis. The proteolytic activity was expressed as the rate ofrelease of fluorescent fragment (umol/min per mg of protein). Chromogenic peptidyl-pNa substrates were assayed as previously described [11]. Stock solutions of the peptidyl-pNa were prepared as DMSO (50 %, v/v) and aliquots were 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. The absorbance at 410 nm was followed for at least 10 min.

K, determinations Kinetic parameters for the hydrolysis of internally quenched positive peptide substrates were obtained at 37 'C, pH 9.0, in 0.1 M Tris/HCl buffer as described [12,13]. At least six different substrate concentrations below 10 ,uM were used. Higher concentrations were not used, to avoid filter effects. The kinetics data were analysed using the GraFit program [19].

Proteinase inhibitors DSP was pre-incubated with the potential inhibitors for different periods of time, depending on each inhibitor, and then incubated with 10 #M Abz-MKRLTL-EDDnp as substrate. The results were recorded as the percentage of residual activity relative to control reactions run simultaneously in the absence of the inhibitor. For determination of irreversible inhibition, the residual enzymic activity was measured as described above, after a 5 min preincubation with various concentrations of inhibitor. The K, values were obtained by analysing the experimental data using the Morrison (Enzfitter; Leatherbarrow; Elsevier-Biosoft) non-linear regression.

Determination of substrate cleavage site The fragments resulting from hydrolysis of the internally quenched fluorescent peptides with DSP at 37 'C, pH 9.0, in 0.1 M Tris/HCl buffer were isolated by HPLC in an Ultrasphere Cl. 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 were monitored by UV absorbance at 220 nm and fluorescence at Aem = 420 nm and Aex. = 320 nm. The scissile bond(s) of hydrolysed peptides were determined by isolation of the fragments by HPLC, and their structures were deduced both from amino acid composition and by comparison of their HPLC profiles with those of the synthesized authentic fragments.

Determination of optimum pH The activity of DSP at various pH values was measured at 37 'C by the fluorimetric assay described above, using 10 #M AbzMKALTL-EDDnp as substrate. The buffers used were as follows: 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) and 0.1 M Tris/HCl (pH > 8.0).

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

Analytical methods Protein was assayed by Bradford's method [20], using BSA as standard.

Exocellular proteinase of Paracoccidioides brasiliensis

211

RESULTS Dialysed 7-day-old culture supernatants of P. brasiliensis grown in TOM medium [14] had a specific activity of 6.2 x 10-3 gmol/ min per mg using Abz-MKRLTL-EDDnp as substrate. Purified gp43 preparations and culture supernatants depleted of gp43 gave specific activities of 1.3 x 10-5 jmol/min per mg and 8.2 x 10-3 jmol/min per mg respectively, indicating that most of the exocellular peptidase activity for the internally quenched substrate remained in the gp43-depleted culture supernatant (DSP). Abz-MKRLTL-EDDnp was specifically hydrolysed only at the Leu-Thr bond by DSP (Figure 1). The first step in the fractionation of the proteolytic activity against Abz-MKRLTL-EDDnp involved ammonium sulphate precipitation of whole culture supernatant fluids (2 litres). At 40-50 % ammonium sulphate saturation, the peptidase activity was enriched in the precipitated material. This was subsequently depleted of the remaining gp43 in an anti-gp43 immunoaffinity column as described. The proteolytic activity concentrated in the gp43-free 40-50% ammonium sulphate precipitate was eluted from the Resource Q anion-exchange resin at 0.08-0.1 M NaCl (Figure 2a). Gel filtration of the concentrated active fractions resulted in elution as a single peak near ovalbumin in a calibrated Superose 12 column (Figure 2b). These results and the presence of a single scissile bond in the Abz-MKRLTL-EDDnp substrate are in accordance with a single exocellular proteinase activity in P. brasiliensis culture supernatants. It is noticeable in both chromatographic profiles that, although the proteolytic activity was very high, the corresponding protein content was negligible, indicating a proteinase of very high specific activity. SDS/PAGE analyses of the more highly purified fractions have so far been inconclusive in determining the correct size of the enzyme. Preliminary results have also shown that the proteo-

2000 -,

-,

*E 1000

0

0.05 U-

20

0

158

800

F(b)

67 43

40

25

13.7 kDa 0.020

,

600

0.015

400

0.010 CN4

200

0.005

E U-

O

U-

.52

cJ1

o

Fraction no.

Figure 2 Chromatographic analysis of the proteolytic activity

(a)

Abz-MKRLTL-EDDnp II

II

(a) Anion-exchange chromatography on Resource Q (FPLC system) of solubilized gp43-depleted ammonium sulphate precipitate of P. brasiliensis culture filtrate. The bound material was eluted with a linear gradient of 0.015-0.5 M NaCI (----) and assayed for peptidase activity at pH 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 fluorescence units (AFU)/min] against Abz-MKRLTL-EDDnp, assayed in the presence of 2 mM cysteine and 1 mM EDTA when eluted from Superose 12. The positions of the molecular mass markers are indicated. Further details are given in the Materials and methods section.

I

i I

I

lytic activity against Abz-MKRLTL-EDDnp was detected in aggregates with other components ofthe culture supernatant fluid following concentration. The addition of 10 % ethylene glycol to the enzyme samples throughout fractionation has apparently prevented aggregation. In addition, partially purified preparations eluted from the anion-exchange column were inhibited with mercuriates to prevent self-degradation.

Abz-MKRL

(b)

Proteolytic and peptidase activItles _______________________ ,_ ,, , ,_ _ 0

2

4

6

8

10

12

14

16

lime (min) Figure 1 Determination of substrate cleavage sfte HIPLC profile of Abz-MKRLTL-EDDnp before (----) and after ( ) hydrolysis by F ' brasiliensis DSP. (a) Absorbance at 280 nm; (b) fluorescence at A., = 320 nm arnd Aem = 420 nm. The peptides corresponding to each peak are indicated.

Casein and Hide Power Azure were weakly hydrolysed by the P. brasiliensis DSP at pH 5.0 and 8.0, whereas no hydrolysis of oxidized insulin fl-chain was observed at either pH, even after 14 h of incubation. In order to verify the specificity of the proteolytic activity of DSP, several peptides (Table 1) were tested as substrates at pH 5.0 and 8.0. The first group consisted of chromogenic aminoacyl-and peptidyl-pNas, which have been used as substrates for trypsin-like (peptides 2 and 3), elastase-like (peptides 1, 4

212

A. K. Carmona and others

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

Substrate

Peptide no.

Table 2 KIneIc parameters for the hydrolysis of Internally quenched fluorogenic peptides by DSP of P. brasillensis culture flltrates Hydrolysis conditions were 37 0C, 0.1 M Tris/HCI buffer (pH 9.0). Values are means+ S.D. (1M/min

Substrate

(/AM)

per mg)

(min

9 10 11 12 13 14 15 16

Abz-MKRLTL-EDDnp Abz-MRRLTL-EDDnp Abz-MKKLTL-EDDnp Abz-MKALTL-EDDnp Abz-RMKRLTL-EDDnp Abz-RMARLTL-EDDnp

11.4 + 0.9 11.0+2.0 3.5+0.5 4.7+0.3 8.7+0.6

Abz-RMMLTL-EDDnp

0.9 +0.1 8.9 +0.1

29.2 + 0.9 46.0+ 0.1 24.0+ 0.1 60.0+0.1 40.7+0.3 36.7+0.3 14.0+0.2 2.0+0.5

2.6 4.2 6.9 12.8 4.7 4.4 15.5 2.3

Peptide Chromogenic peptides 1 2 3 4 5 6 7 8 Internally quenched fluorogenic peptides 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Ac-A-pNa Ac-FR-pNa (D)VLK-pNa Suc-AAA-pNa

Z-ML-pNa Bz-KAL-pNa Bz-GHL-pNa

DRVYIHPF-pNa Abz-MKRLTL-EDDnp Abz-MRRLTL-EDDnp Abz-MKKLTL-EDDnp Abz-MKALTL-EDDnp Abz-RMKRLTL-EDDnp Abz-RMARLTL-EDDnp

Leu-Thr Leu-Thr Leu-Thr Leu-Thr

Abz-RMAALTL-EDDnp

Leu-Thr

Abz-FRLVR-EDDnp Abz-PLGLLGR-EDDnp Abz-GGFLRRV-EDDnp Abz-LMKRP-EDDnp Abz-FHLVIH-EDDnp Abz-RPFHLVIH-EDDnp Abz-YIHPFHLVIH-EDDnp

Leu-Val Leu-Leu, Leu-Gly

Q L.C

0

.E

4

6

Abz-FRLVR-EDDnp

8.4±0.4

mg)g

aii-Thr

Leu-Thr

E

2

l(xlKm

Km

no.

8

10

pH

FRgure 3 pH-dependence of proteolytlc actvlty P. brasiliensis DSP activity against Abz-MKALTL-EDDnp was deteri mined at various pH values by the fluorometric method described in the Materials and methoKds section.

and 5) and chymotrypsin-like (peptides 7 and 8 enzymes. None of these chromogenic peptides was hydrolysed tay DSP. The second group of substrates tested for hiydrolysis consisted of internally quenched fluorescent peptides, synthesized in our laboratory for different purposes (Table 1, pe ptides 9-22). Nine of them (peptides 9-17) were hydrolysed. The scissile bond(s) of hydrolysed peptides were determined as descrit)ed in the Materials and methods section for Abz-MKRLTL-E]DDnp (Figure 1). This peptide and its analogues (peptides 10-1.5) were hydrolysed at the Leu-Thr bond by the DSP. As Abz-tAKRLTL-EDDnp

was the first peptide substrate tested which was cleaved by the proteinase, analogues of it were synthesized in order to establish the importance of the basic pair at the P2 and P3 positions for the

enzyme activity. Abz-FRLVR-EDDnp (peptide 16), a sequence of rat kininogen, was cleaved at the Leu-Val bond, and AbzPLGLLGR-EDDnp (peptide 17), a substrate for collagenases of mammalian sources, was very poorly hydrolysed at the Leu-Leu and 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 of Abz-MKALTL-EDDnp by DSP. This substrate was used in this experiment in order to obtain measurable fluorescence values at the acidic pHs where the activity is usually low. There is a small peak of activity at pH 5.0-5.5, but the maximum values are obtained at pH values above 8.0. Considering that the activity was still increasing at pH 9.5 and that the fluorescence readings at higher pHs are not reliable, owing to fluorogenic peptide degradation, we could not determine the optimum pH of P. brasiliensis exocellular proteinase activity. Determination of the Km and K1 values, as well as testing of inhibitors, was carried out at pH 9.0. The kinetic parameters of hydrolysis were determined for the susceptible peptides with one cleavage size (Table 1, peptides 9-16). Hydrolysis was continuously monitored in a spectro-

fluorimeter and the data fitted to Michaelian kinetics. Table 2 shows the values of the kinetic constants. Taking Abz-MKRLTLEDDnp for comparison (peptide 9, Table 2), the presence of Arg-Arg at positions P2 and P3 in peptide 10 kept the Km value unchanged and barely increased the VmJK,,. On the other hand, the presence of a Lys-Lys basic pair in the same positions (peptide 11) reduced the Km value by one order of magnitude value, but the Vm,, also decreased. Elongation of peptide 9 with an additional arginine at the N-terminus (peptide 13) did not affect the peptide's susceptibility to hydrolysis. Alanine at P2 instead of arginine (peptides 12 and 15) resulted in higher Vm,' /Km values, and the resulting peptides were, therefore, the best substrates tested in this work. Peptide 14, in which alanine replaced lysine at P3., showed kinetic parameters very similar to those obtained with peptide 9. Replacement by alanine of both basic amino acids at positions P2 and P3 was only possible with the addition of arginine at the N-terminal position, to provide solubility. Abz-FRLVR-EDDnp (peptide 16) was also cleaved at the Leu carboxyl bond, but with lower catalytic efficiency.

Exocellular proteinase of Paracoccidioides brasiliensis Table 3 Inhibition of the peptddase activity of DSP from P. braslllensis culture flltrates by common proteinase InhIbitors Experimental conditions were 0.1 M Tris/HCI buffer, pH 9.0, 37 0C. The substrate used was Abz-MKRLTL-EDDnp. Fluorescence detection: Aem. = 420 nm; Aex = 320 nm. K; values are means + S.D. Inhibitor

Concentration

EDTA Pepstatin

10

Tos-Phe-CH2CI Tos-Lys-CH2CI

PMSF E-64 NEM lodoacetamide Mercuric acetate p- HMB

mM

1 ,tM 0.1 mM 0.1 mM 0.05 mM 0.1 mM 0.1 mM 0.1 mM 2 uM 2 ,M

Preincubation time

Residual activity (%)

5h 4h 17 h 17 h 5 min 17 h 17 h 17 h 5 min 5 min

100 100 52 100 0 72 60 79 0 0

K1 (esM)

-

-

-

6+2 -

0.15+0.01 2.3+0.1

Inhibition of the proteolytic activity Potential inhibitors of the P. brasiliensis exocellular proteolytic activity were assayed for their effects upon hydrolysis of AbzMKRLTL-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-proteinase inhibitor E-64 [21] were poor inhibitors, and the complex formed with E-64 was completely removed by addition of an excess of substrate. The most effective inhibitors were mercuric acetate, p-HMB and PMSF, which completely abolished the peptidase activity at low concentrations, and in less than 5 min. The inhibition constants (K1) for these compounds were obtained by determining the residual activity of P. brasiliensis exocellular proteinase at different concentrations of inhibitors (Table 3). The inhibition by mercuric acetate and p-HMB could be reversed by 40 % with 2.0 mM cysteine (higher concentrations inhibited the enzyme 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 the enzyme activity obtained by fractionation in Superose 12. TosLys-CH2Cl did not inhibit the exocellular peptidase activity, but Tos-Phe-CH2Cl lowered it by almost 50 % after a preincubation of 17h.

DISCUSSION The present work describes an exocellular serine-thiol proteinase activity in the culture supernatants of P. brasiliensis yeast phase. The activity eluted in a single peak after anion-exchange and gelfiltration chromatography, using as starting material solubilized gp43-depleted 40-50 % ammonium sulphate precipitates of fungal culture filtrates. The use of 10 % ethylene glycol together with reversible inhibitors was necessary to prevent both aggregation of the proteinase with other exocellular components and self-degradation of the enzyme during the fractionation process. Purification to homogeneity of this proteinase of high specific activity has not been possible so far due to the low protein content of the preparations. Previous data have shown that partially purified gp43 preparations were able to hydrolyse casein at pH 5.5-6.0 [10], and also to cleave the internally quenched fluorescent peptide AbzMKRLTL-EDDnp [15]. The specific activities measured in the gp43 preparations were, however, low in comparison with that in

213

culture filtrates of P. brasiliensis depleted of the gp43 antigen by immunoaffinity binding. The proteolytic activity associated with the gp43 apparently resulted from protein aggregation, with the aggregate containing very small amounts of a potent exocellular proteinase. The proteolytic activity of the culture filtrates depleted of gp43 (DSP) could not be detected on using as substrates insulin ,Jchain or eight different aminoacyl- and peptidyl-pNa derivatives. Even Z-AAL-pNa, a common synthetic substrate for subtilisins which is hydrolysed with high specificity by extracellular serine proteases from different micro-organisms [22], was not cleaved by the P. brasiliensis exocellular proteinase. Pretreatment with trypsin for a short period of time (1 min) and subsequent inhibition by soybean trypsin inhibitor did not increase the hydrolytic activity (results not shown), thus excluding secretion of a fungal pro-enzyme that could be activated by trypsin. Our results indicate that the fungal exocellular enzyme has a very restrictive substrate specificity, including a leucine at P1 as an important requirement. The internally quenched fluorescent peptide Abz-MKRLTL-EDDnp, and analogues thereof, were cleaved at the Leu-Thr bond, the Abz-FRLVR-EDDnp derivative at the Leu-Val bond and the substrate for mammalian collagenases Abz-PLGLLGR-EDDnp at the Leu-Leu and LeuGly bonds. The presence of threonine and leucine respectively at positions P'1 and P'2 in Abz-MKRLTL-EDDnp and its analogues is apparently not essential for the hydrolytic activity, since AbzFRLVR-EDDnp was still hydrolysed at a significant rate. Cleavage of the Leu-Val bond possibly depends on the neighbouring residues, since peptides Abz-FHLVIH-EDDnp, AbzRPFHLVIH-EDDnp and Abz-YIHPFHLVIH-EDDnp were not substrates for the fungal proteinase. The basic pair of amino acids present in Abz-MKRLTL-EDDnp was not essential for hydrolysis, since Abz-MKALTL-EDDnp and Abz-RMAALTLEDDnp were even better substrates. In addition, the presence of another arginine at P5 did not change the enzyme-substrate affinity of peptide Abz-RMKRLTL-EDDnp. Moreover, the endoproteolytic activity requires amino acid residues at the P' site of the cleavage bond, since Bz-KAL-pNa and Z-AAL-pNa were not hydrolysed, despite containing the same P3 and P1 residues as the best internally quenched fluorescent substrates (Abz-MKALTL-EDDnp and Abz-RMAALTL-EDDnp). The exocellular peptidase activity of P. brasiliensis was irreversibly inhibited by PMSF, mercuric acetate and p-HMB. Inhibition by the mercuriate compounds could, however, be reversed by Cys/EDTA. E-64 was a weak and reversible inhibitor, whereas EDTA and pepstatin were not inhibitory. Inhibition by PMSF and the basic optimum pH of > 9.0 are suggestive of an exocellular hydrolytic activity similar to serine proteinases. An SH-dependence, however, was demonstrated by the inhibition with mercuriates that was partially reversed by Cys/EDTA. These results suggest that P. brasiliensis exocellular enzyme could belong to the subfamily of SH-containing serine proteinases, such as humicolin of Thermomyces lanuginosus [23], the extracellular serine proteinase from Thermoactinomyces vulgaris [22,24] and proteinase K [25,26], which are also related to the subtilisins. All proteinases of this class, already described in the literature [22-24,27-29], depend for their activities on free thiol groups. Most of them contain only one cysteine residue, except for humicolin which contains five residues, and proteinase K, which has two disulphide bridges and one free cysteine positioned very 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 activity was 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 aggregate which could be active at acidic pH. All the substrates and inhibitors of the present work were assayed at pH 5.0 and 9.0 with similar results, and the cleavage sites were identical in both cases. In conclusion, we have characterized in P. brasiliensis an exocellular thiol-containing serine proteinase of high specific activity that may play a role in the pathogenicity of this fungus. Further purification and detailed evaluation of the cleavage specificity 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.

REFERENCES 1 San-Blas, G. (1982) Mycopathologia 79,159-184 2 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, 1465-1469 3 Stambuk, B. U., Puccia, R., Almeida, M. L. C., Travassos, L. R. and Schenkman, S. (1988) J. Med. Vet. Mycol. 26, 367-373 4 Puccia, R., Schenkman, S., Gorin, P. A. J. and Travassos, L. R. (1986) Infect. Immun. 53, 199-206 5 Puccia, R. and Travassos, L. R. (1991) J. Clin. Microbiol. 29, 1610-1615 6 Rodrigues, E. G. and Travassos, L. R. (1994) J. Med. Vet. Mycol. 32, 77-81 7 McKerrow, J. H. (1993) Cienc. Cult./J. Braz. Assoc. Adv. Sci. 45, 319-321 8 RUchel, R., Bernardis, F., Ray, T. L., Sullivan, P. A. and Cole, G. T. (1992) J. Med. Vet. Mycol. 30,123-132 9 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

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

10 11 12 13 14 15

16 17 18 19 20 21 22

23 24 25

26 27 28 29

Puccia, R. and Travassos, L. R. (1991) Arch. Biochem. Biophys. 289, 298-302 Juliano, M. A. and Juliano, L. (1985) Braz. J. Med. Biol. Res. 18, 435-445 Chagas, J. R., Juliano, L. and Prado, E. S. (1991) Anal. Biochem. 192, 419-425 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, 39-46 Puccia, R., Takaoka, D. T. and Travassos, L. R. (1990) J. Med. Vet. Mycol. 29, 57-60 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 Shenolikar, S. and Stevenson, K. J. (1982) Biochem. J. 205, 147-152 Kunitz, M. (1946) J. Gen. Physiol. 30, 291-310 Rinderknecht, H., Geokas, M. C., Silverman, P. and Haverback, B. J. (1969) Clin. Chim. Acta 21,197-203 Leatherbarrow, R. J. (1992) GraFit Version 3.0, Erithacus Software Ltd., Staines, U.K. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 Rich, D. L. (1986) in Proteinase Inhibitors (Barrett, A. J. and Salvesen, G., eds.), pp. 153-178, Elsevier, Amsterdam 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 Hasnain, S., Adeli, K. and Storer, A. C. (1992) Biochem. Cell Biol. 70, 117-122 Meloun, B., Baudys, M., Kotska, V., Hausdorf, G., Frommel, C. and Hohne, W. E. (1985) FEBS Left. 183, 195-200 Ebelin, W., Hennrich, N., Klockow, M., Metz, H., Orth, H. D. and Lang, H. (1974) Eur. J. Biochem. 47, 91-97 Betzel, C., Bellemann, M., Pal, G. P., Bajorath, J., Saenger, W. and Wilson, K. S. (1988) Protein Structure Function Genet. 4,157-164 Mizusawa, K. and Yoshida, F. (1973) J. Biol. Chem. 248, 4417-4423 Kundu, A. K. and Manna, S. (1975) Appl. Microbiol. 30, 507-513 Ong, P. S. and Gaucher, G. M. (1973) Can J. Microbiol. 19, 129-133