ARCHIVES

OF BIOCHEMISTRY

Vol. 293, No. 1, February

AND

BIOPHYSICS

14, pp. 25-31, 1992

Characterization of an Endopeptidase Trypanosoma brucei brucei M. J. Kornblatt,“”

of

G. W. N. Mpimbaza,? and J. D. Lonsdale-Ecclest

*The Enzyme Research G&up, Department of Chemistry and Biochemistry, Concordiu University, and TInternational Laboratory for Research on Animal Diseases (ILRAD), Nairobi, Kenya

Montreal,

Quebec, Canada;

Received July 1, 1991, and in revised form October 11, 1991

A soluble SO-kDa endopeptidaae has been isolated from Trypanosoma brucei brucei. The enzyme, which has a ~15.1, is optimally active at about pH 8.2 and has apparent pK, values of 6.0 and 210. It is inhibited by the serine protease inhibitor diisopropylfluorophosphate and by the serine protease mechanism-based inhibitor 3,4dichloroisocoumarin. Unexpectedly, the enzyme is inhibited by the cysteine protease inhibitor benzyloxycarbonyl-Leu-Lys-CHN, but not by the related diazomethane, butoxycarbonyl-Val-Leu-Gly-Lys-CHNz, nor by other cysteine protease specific compounds. Specificity studies with a variety of amidomethylcoumaryl (AMC) derivatives of small peptides show that the enzyme has a highly restricted trypsin-like specificity. The best substrate, based on the magnitude of k,, JK,,, , was benzyloxycarbonyl-Arg-Arg-AMC; other good substrates were benzyloxycarbonyl-Phe-Arg-AMC, benzoyl-Arg-AMC, and compounds with Arg at Pi and Ala or Gly at Pz . The hydrolysis of most substrates obeyed classical MichaelisMenton kinetics but several exhibited pronounced substrate inhibition. The enzyme did not activate plasminogen nor decrease blood clotting time; it was inhibited by aprotinin but not by chicken ovomucoid. We conclude that the enzyme is a trypsin-like serine endopeptidase with unusually restricted subsite specificities. o issz Academic

Press,

Inc.

Proteolytic enzymes have been identified in crude lysates and subcellular fractions of trypanosomes (1,2), but their physiological significance is unknown. Proteolysis could have a number of important functions in these organisms, including processing of trypanosomal proteins, hydrolysis of host proteins, and protein degradation in’ To whom correspondence should be addressed at Department of Chemistry and Biochemistry, Concordia University, 1455 de Maisonneuve Blvd. W., Montreal, Quebec, H3G lM8 Canada. Fax: 514-8483494. 0003.9861/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form

volved in the morphological changes occurring during the complex life cycle. We have observed endopeptidase activity against small, trypsin substrates in crude lysates of African trypanosomes (2). Both Bz-Arg-AMC? and Bz-Arg-pNA are hydrolyzed; this activity is inhibited by some cysteine as well as serine protease inhibitors. An alkaline peptidase with similar properties has also been identified in Trypanosoma cruzi and other trypanosomatids (3, 4). Since greater knowledge of the properties and function of these enzymes would not only increase our knowledge of the biochemistry of these organisms but might also lead to the design of compounds useful in the treatment of sleeping sickness, we have extended our studies to include subcellular location, changes in the levels of the proteolytic activities during the life cycle (5), and detailed studies of the mechanism and substrate specificity of the individual enzymes. In this paper we report the purification and characterization of the endopeptidase present in the soluble fraction of T. brucei. MATERIALS

AND

METHODS

Materials. AMC andp-nitroanilide substrates were purchased from Peninsula Laboratories Europe Ltd., St. Helens, U.K.; Protogen, Laufelfingen, Switzerland; Cambridge Research Biochemicals, Harston, U.K.; and Bachem Feinchemikalien AG, Bubendorf, Switzerland. Benzoyl-Arg-ethyl ester and 3-amidinobenzyl-phenyl ether were from Boehringer-Mannheim, Mannheim, Germany; glycine benzyl ester and DFP were from Sigma Chemical Co., Poole, U.K. The following inhibitors and inactivators were used: E-64 (Cambridge Research Biochemicals), p-amidinophenylmethylsulfonyl fluoride and 3,4dichloroisocoumarin (Protogen), 2,2’-dipyridyl disulfide (Aldrich Chemical Co., Gillingham, U.K.), plasminogen activator inhibitor active site hexapeptide (Peninsula

’ Abbreviations used: Boc, butoxycarbonyl; Bz, benzoyl; ‘2, benzyloxycarbonyl; AMC, 7-amido-4-methylcoumaryh pNA, p-nitroanilide; DFP, diisopropylfluorophosphate; Pipes, 1,4-piperazinediethanesulfonic acid; Ampso, 3-(N-cu,a-dimethylhydroxyethylamino)-2-hydrox~ropanesulfonic acid; E-64, L-tmn.s-epoxysuccinyl-leucyl-amido(4-guanido)butane; Glt, Glutaryl; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis. 25

Inc. reserved.

26

KORNBLATT,

MPIMBAZA,

Laboratories), urokinase inhibitor from placenta (Calbiochem, Lucerne, Switzerland), aprotinin, and type III-O trypsin inhibitor from chicken egg white (Sigma). The diazomethylketones were a gift from Dr. E. Shaw (Friedrich-Miescher Inst., Basel). Methods. The enzyme was purified from the high speed supernatant of approximately 10” trypanosomes (2’. brucei brwxi MITat 1.52). Growth and purification of trypanosomes and the preparation of the supernatant were performed as previously described (2), except that 2 mM E-64 was added before the cells were disrupted. The supernatant was applied to a column of DE-52 cellulose (Whatman, Maidstone, U.K.) equilibrated with 50 mM Tris-HCl, 0.1 mM dithiothreitol, pH 8.0, and the protease was eluted with a O-300 mM gradient of NaCl in the same buffer. Fractions containing activity were then applied to a column of Sephacryl200 (Pharmacia, Uppsala, Sweden) equilibrated with 100 mM Tris-HCl, 0.1 mM dithiothreitol, pH 8.0, and eluted with the same buffer. Further purification was achieved by isoelectric focusing using 1% ampholytes (pH 3-10, LKB, Uppsala, Sweden; or Serva, Heidelberg, Germany) in a 5-50% sucrose gradient. The activity focused between pH 5 and pH 5.2. After adjustment to pH 8.45, the enzyme was then applied to a column of immobilizedp-aminobenzamidine (Pierce Chemical Co., Rockford, IL) equilibrated with 50 mM Tris, 0.1 mM dithiothreitol, pH 8.45, and eluted with a gradient of O-240 mM NaCl in the same buffer. The enzyme was concentrated using a Centriprep 10 concentrator (Amicon, Lexington, KY), diluted 1:l with glycerol, and stored at -20°C. All purification steps were performed at 4°C. Protein concentration was measured using the Coomassie brilliant blue assay (Bio-Rad, Watford, U.K.) with bovine serum albumin as standard. After the last step in the purification, protein was measured by its absorbance at 215 and 225 nm (6). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed according to Laemmli (7). The enzyme activity was routinely measured at 37°C in 50 mM TrisHCl buffer, pH 8.1, containing 50 pM Bz-Arg-AMC. The release of AMC was followed in an Aminco SLM 8000 spectrofluorometer with the following settings: &,,,t, 380 nm; Xemiss,460 nm; all slits, 8 nm. The changes in fluorescence intensity were converted to micromoles of AMC by determining, under the same conditions, the fluorescence intensity of a standard solution of AMC. K,,, and V,.. were determined for a number of AMC substrates using the above assay. For 4-nitroanilide substrates, the release of I-nitroaniline was followed spectrophotometrically at 405 nm; under these assay conditions, e = 11,000 cm-i M-l. The enzyme was assayed in duplicate at 7-10 concentrations of substrate; data were analyzed using the ENZFITTER computer program (Elsevier-Biosoft, Cambridge, U.K.). In cases where substrate inhibition was apparent, the highest substrate concentration used in the analysis was 10 pM. The kinetic constants for these substrates were also analyzed by fitting data covering a wider range of concentrations (up to 100 pM) to the equation

u = Vm.x[SIIWm+ WI + lS12/K). The two fits gave similar values for K,,, and V-, with comparable errors. The data reported in Tables II and III are the results of the fit to the simple Michaelis-Menton equation (i.e., no substrate inhibition). Activity was measured as a function of pH at 25°C in buffers containing 25 mM each of citrate, Pipes, Tris, and Ampso (3-[N-cr,cr-dimethylhydroxyethylamino]-2-hydroxypropanesulfonic methylhydroxyethylamino]-2-hydroxypropanesul acid); the mixture NaO: was titrated to the desired pH with HCl or NaOH. Inactivation studies were performed in citrate, Tris, Pipes, Ampso buffer, bu each 22.5 mM, pH c6.9, n containing --..e”:-:m.r. 1~cx rl..a.r,,l lA.re.,-,. P IO..R” > incubated with the 10% glycerol. Enzyme, 8-10 nM, was compound of interest at room temperature for a minimum of 30 min and an aliquot was then assayed for activity. Other conditions are as described in Table V. In experiments with diisopropylfluorophosphate (DFP), samples were removed and assayed at varying times after addition of DFP: the data were fitted to a single decay curve using - exnonential the ENZFITTER program.

AND

LONSDALE-ECCLES

Active site titrations and identification of the DFP-sensitive enzyme were performed using [1,3-3H]DFP([3H]DFP) (New England Nuclear, Stevenage, U.K.). After elution from the benzamidine column and concentration, dithiothreitol(5 mM final concentration) and [3H]DFP (8.3 X 108 dpm/pmol; final concentration 0.23 mM) were added to the enzyme. The sample was incubated in the cold overnight; this resulted in >90% inactivation. The sample was then subjected to SDS-PAGE in a 10% gel. After staining with Coomassie Blue, the gel was impregnated with EnsHance (New England Nuclear) and the labeled band identified by fluorography. Alternatively, the band of interest was cut out and digested with perchloric acid and hydrogen peroxide (8), and the radioactivity was measured. For the determination of the molecular weight of the subunit by SDS-PAGE, the following [“C]methylated molecular weight markers (Amersham, Aylesbury, U.K.) were also run: myosin (200 kDa), phosphorylase b (92.5 kDa), bovine serum albumin (69 kDa), ovalbumin (46 kDa), carbonic anhydrase (30 kDa), and lysozyme (14.3 kDa). Values for the concentration of active sites are the averages of two determinations on the same preparation; the subunit molecular weight is the average of four determinations. The native molecular weight was determined by gel filtration on a Sephacryl300 (Pharmacia) column, equilibrated with 42 mM phosphate, 36 mM NaCl, 44 mM glucose, 20% glycerol, pH 8.0. (This column was being used for another project and had been calibrated using this buffer.) The column was calibrated with thyroglobulin (669 kDa), aldolase (158 kDa), bovine serum albumin (69 kDa), ovalbumin (46 kDa), and ribonuclease (13.7 kDa). The molecular weight standards were purchased from Pharmacia. The ability of the endopeptidase to activate plasminogen was tested by incubating plasminogen (Boehringer), 100 units/ml, with the enzyme (0.001 units) in standard assay buffer at room temperature and then measuring the activity using the assay for the T. brucei endopeptidase. The effects of the endopeptidase on blood clotting time were measured using normal titrated bovine plasma and thromboplastin prepared according to Magnusson (9). Citrated bovine plasma (0.2 ml) was mixed with 0.1 ml of 50 mM Tris-HCl, pH 8.2, and 0.002 units of the peptidase was added. The mixture was incubated at 37°C for varying lengths of time (up to 1.5 h) and then 0.1 ml of thromboplastin and 0.1 ml of the above Tris buffer were added. The incubations were continued for a further 5 min. The time to clotting was measured after the addition of 0.1 ml of 100 mM CaCl,.

RESULTS The endopeptidase present in the high speed supernatant was purified by a combination of column chromatography and isoelectric focusing. The results of one of our better preparations are shown in Table I. Although the protease activity was purified almost 700-fold, the material which eluted from the afhnity column in the final step still contained two or three major protein bands when analyzed by SDS-PAGE (Fig. 1, lane 1). Since we had earlier shown that the endopeptidase activity was inhibited by DFP (2), radioactive DFP was used both to identify this DFP-sensitive protein and to titrate the number of active sites. The results of SDS-PAGE ._-plus fluorography are shown in Fig. 1. Only one band, with a molecular mass of 80 f 3 kDa, is labeled by [3H]DFP (lane 2); reduction of the sample by dithiothreitol followed by subsequent alkylation with iodoacetamide nrior to SDS-PAGE did not significantly change the ___mobility of the labeled band (Fig 1, lanes-4 and-s). Based on the incorporation of [3H]DFP, the final fraction from the preparation shown in Table I contained 480 pmol of active site.

27

2’. brucei ENDOPEPTIDASE TABLE Purification

Fraction HSSb DE-52 s-200 IEF* AFF CHR*

Units (urn01 min-‘) 26.7 11.7 8.7 7.8 3.7

of

I

T. brucei Endopeptidase”

Yield %

Protein bd

Specific Activity

100 43.7 32.6 29.0 14.0

600 24 6 -c 0.12*

0.0445 0.488 1.45 30.8

a 1.6 X 10” trypanosomes used. * HSS, high speed supernatant; IEF, isoelectric focusing; AFF CHR. chromatography ’ Protein not determined due to interference by ampholytes. * Protein determined by absorbance at 215 and 225 nm.

The relative molecular mass of the enzyme, as determined by gel filtration, is 110 kDa. The molecular mass by SDS-PAGE is 80 kDa. The 80-kDa band accounts for about one-third the total amount of Coomasie bluestained protein. Since the protein is not 100% pure, we cannot state whether the protein is monomeric or whether it is composed of two polypeptides, one of 80 kDa and a second, smaller one. However, a 20- to 40-kDa band is not visible on the SDS-PAGE gels (Fig. 1). The purified, concentrated enzyme was stable for several months when stored in 50% glycerol at -20°C. Dilute

on immobilized

Enrichment 1 11 33 690

p-aminobenzamidine.

enzyme was unstable at 4°C unless diluted into buffer containing 5-10% glycerol. Dilution buffers normally contained 1 mM dithiothreitol because occasionally dithiothreitol was observed to improve the stability of the enzyme. The activity was not changed by the addition of 1 mM EDTA or of 20-300 mM NaCl to the assay. Isoelectric focusing indicates that the enzyme has a pI of 5.1. The pH vs activity profile for the endopeptidase (Fig. 2) is bell-shaped, with apparent pK, values of 6 and greater than 10. The kinetic constants for the enzymatic hydrolysis of a number of commercially available substrates are given in Table II. Most reactions conformed to MichaelisMenton kinetics. However, plots of velocity versus substrate concentration for several substrates were not hyperbolic but instead exhibited pronounced substrate inhibition; this inhibition was apparent at substrate concentrations of 50 PM or less. The data for one substrate, Boc-Glu(OBz)-Ala-Arg-AMC, are shown in the

40 a .z .E

_

2 20 -

FIG. 1. SDS-PAGE of the 2’. brucei endopeptidase labeled with [3H]DFP. SDS-PAGE, staining, and fluorography were performed as described under Materials and Methods. Lane 1, purified, labeled endopeptidase, Coomassie blue staining; the arrowhead marks the band that is labeled with [3H]DFP; lane 2, same gel as lane 1, fluorography; lane 3, ‘“C-labeled molecular weight markers; from top to bottom, the molecular weights are 200, 100,92.5,69,46, and 30 kDa; lanes 4 and 5, Auorography of labeled endopeptidase with (lane 4) and without (lane 5) reduction and alkylation. The same amount of radioactivity was not applied to lanes 4 and 5.

00 4

5

6

8

7

9

10

11

PH

FIG. 2. Materials second.

Activity vs pH. Activity was measured as described under and Methods; activity is plotted as fluorescence units per

28

KORNBLATT, TABLE

MPIMBAZA,

AND

LONSDALE-ECCLES

80 lb

II

1

Substrate Specificity

Substrate 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Cbz-Arg-Arg-AMC Z-Phe-Arg-AMC Z-Gly-Gly-Arg-AMC Bz-Arg-AMC Boc-Glu(OBz)-Gly-Arg-AMC Boc-Glu(OBz)-Ala-Arg-AMC Boc-Gln-Gly-Arg-AMC Boc-Leu-Gly-Arg-AMC Boc-Asp(OBz)-Pro-Arg-AMC Boc-Gln-Ala-Arg-AMC Bz-Arg-pNA Bz-Phe-Val-Arg-AMC Boc-Ala-Gly-Pro-Arg-AMC Tosyl-Gly-Pro-Arg-pNA Bz-Phe-Val-Arg-pNA Tosyl-Gly-Pro-Lys-pNA Glt-Gly-Arg-AMC Boc-Ile-Glu-Gly-Arg-AMC Arg-AMC Sue-Ala-Ala-Phe-AMC Leu-AMC Ac-Ala-Ala-Tyr-AMC Z-Gly-Gly-Leu-pNA Cbz-Gly-Pro-AMC

kcat/Km” (5-l

PM-‘)

809.6 114.9 64.1 46.5 47 37.3 29.3 24.6 21.1 11.4 11.0 10.9 10.4 3.7 2.0 1.9 1.3 0.7 0.37

a Standard errors of kcat were less than substrates 2 (4%), 14 (5%), 15 (7%), and K,,, were less than 6% except for substrates See Materials and Methods section for a when substrate inhibition was apparent. ’ Velocity at 20 @M substrate. ’ Velocity at 100 pM substrate. d Velocity at 50 pM substrate.

$)

72.9 83.9 157.8 145.2 103.4 67.5 116.3 72.1 92.3 88.2 108.9 42.7 99.1 98.4 37.0 48.7 63.8 37.1 35.9 3.77b 2.16’ 1.8gd 100 PM, is included for comparison. One possibility for the binding of two molecules of substrate is that the second molecule binds to the acyl-enzyme, occupying the leaving group site. Since this behavior was shown by the substrates containing Glu-0-benzyl at P3 [nomenclature of Schechter and Berger (ll)], but not Glu or Gln, we thought it possible that it was the 0-benzyl group that was required for this mode of binding. We therefore tested glycine benzyl ester as an inhibitor of the hydrolysis of Bz-Arg-AMC. No inhibition was observed at concentrations as high as 500 PM.

40

01 300

1

L>100 /’ d I H 200

01

0

40

so

120

160

[Sl, pm FIG. 3. Substrate inhibition by Boc-Glu(OBz)-Ala-Arg-AMC. Top, velocity (in fluorescence units per second) vs substrate concentration (PM). Bottom, additional assays were performed at high substrate concentrations and the data plotted as l/u vs [S].

The enzyme was also assayed for its ability to activate plasminogen and the blood clotting system. Incubation of the trypanosome endopeptidase with plasminogen did not result in any activation, whereas when the same amount of trypsin (based on activity toward Bz-ArgAMC) was used under the same conditions, activation of plasminogen was observed. The endopeptidase was incubated with titrated bovine plasma and the clotting time measured. Even after a 40-min preincubation with the endopeptidase, there was no difference in clotting times between the control and the endopeptidase-containing reaction mixtures. Various compounds were assayed for their ability to inhibit or inactivate the T. brucei endopeptidase; the results are shown in Tables IV and V. The inactivation by DFP was studied in more detail; at 0.88 mM DFP (1.67

TABLE

III

Substrate Inhibition

Substrate Bz-Arg-AMC Boc-Glu(OBz)-Gly-Arg-AMC Boc-Glu(OBz)-Ala-Arg-AMC Boc-Asp(OBz)-Pro-Arg-AMC Cbz-Arg-Arg-AMC Z-Phe-Arg-AMC

Ki for inhibition (PM) >600 (I 90 238 221 b

K,,, for hydrolysis (PM) 3.12 2.20 1.81 4.38 0.09 0.73

’ Ki not determined because plot of l/o vs [S] was not linear. b Substrate inhibition apparent at [S] = 15 pM; Ki not determined.

29

T. brucei ENDOPEPTIDASE TABLE IV

Possible Inhibitors of T. brucei Endopeptidase Compound A. Low

Result” molecularweight compounds Competitive inhibition (K; = 45 + 7 pMb) No inhibition Competitive inhibition (Ki = 36 PM') Competitive inhibition (Ki = 0.80 +_.04 mMb)

Bz-Arg ethyl ester Glycine benzyl ester Plasminogen activator inhibitor hexapeptide 3-Amidinobenzyl phenyl ether

B. Protein Urokinase inhibitor, 15 units (from human placenta) Chicken ovomucoid, 1 mg/ml Aprotinin, 10 pg/ml

inhibitors No inhibitiond No inhibitiond 55%d

a Inhibition was determined by measuringprotease activity in varying substrate concentrations in the presence and absence of fixed concentrations of inhibitor (glycine benzyl ester was tested at 250 and 500 PM). b Average of two determinations. ’ One determination only. d Activity in presence of inhibitor as percentage activity in its absence when the peptidase (5.7-7.1 PM) was assayed in 2 pM substrate.

tllP was between 10 and 11 min. The apparent first-order rate constant for the reaction, k,, , was 0.0654 * 0.0023 rnin~’ (average of four determinations); and the apparent second-order rate constant, kapp/I, was 74.3 nM enzyme),

M-l

rate (not shown) and E-64 and 2,2’-dipyridyl disulfide, both of which are active site titrants for cysteine proteases (l2), have little effect on the activity. Nor does the cysteine protease inhibitor Boc-Val-Leu-Gly-Lys-CHNz inhibit the enzyme. Thus the enzyme does not appear to be a cysteine protease. However, the enzyme is inactivated by Z-Leu-Lys-CHNz. Although peptidyl diazomethyl ketones were originally reported to be specific for cysteine proteases (13), it is now known that two serine proteases, kallikrein (14) and the post-proline cleaving enzyme (15, 16), are inactivated by diazomethylketones, while other serine proteases cleave and inactivate these compounds (14, 17). The T. brucei endopeptidase was inactivated by DFP; the DFP was covalently incorporated, based on the presence of 3H in a protein band following labeling with [3H]DFP and SDS-PAGE. The inactivation followed a single exponential decay rate, with a half-life of lo-11 min at 0.86 mM DFP; KJ[I] = 74 M-l min-‘. This reaction rate compares favorably with the inactivation rate of other serine proteases; it is slower than that of trypsin (300 M-’ min-‘) but is considerable faster than that of factor X, (1 M-’ min-‘) (18, 19). The T. brucei enzyme was inactivated by 3,4-dichloroisocoumarin, but was relatively insensitive compared to other serine proteases (20). Competitive inhibition was observed with 3-amidinobenzyphenyl ester (Ki = 800 PM). This compound inhibits thrombin and Factor Xa, with K:s of 60 and 7 /IM, respectively (21). p-Amidinophenylmethylsulfonyl fluoride, which inactivates serine pro-

min-‘. TABLE V

DISCUSSION

We have purified an endopeptidase from the soluble fraction of a trypanosome lysate. Although our final material is not 100% pure, all the activity toward Bz-ArgAMC can be inhibited by DFP and, when the enzyme is labeled with t3H]DFP, there is only one 3H-labeled band on the gel. Our final preparation, therefore, contains only one activity toward Bz-Arg-AMC. As further confirmation that we were dealing with only one activity, cleavage of Bz-Arg-AMC, Z-Gly-Gly-Arg-AMC, Arg-AMC, and ZArg-Arg-AMC were all measured in the presence and absence of aprotinin and with enzyme partially inactivated by DFP. Activity toward all four substrates was inhibited by aprotinin and equally inactivated by DFP (data not shown). Inhibitors and inactivators. The enzyme clearly requires one or more -SH groups for the maintenance of activity. The presence of dithiothreitol increases the stability of the enzyme (see legend to Table IV) and incubation with both iodoacetamide and N-ethylmaleimide leads to loss of activity. However, the presence of dithiothreitol in the assay has little or no effect on the reaction

Effects of Possible Inactivators on T. brucei Endopeptidase”

Reagent* DFP DCIC APMSF E-64 DPDS Iodoacetamide” N-Ethylmaleimide’ Boc-Val-Leu-Gly-Lys-CHN, Z-Leu-Lys-CHNz Z-Phe-Ala-CHNz

Concn. 0.9 mM 0.19 mM 8mM ImM 1mM 10 mM 8mM 90 /.LM 7 PM 90 PM

%Activity (rel. to control) 28 70 100 93 90 75 53 a7 18 94

a Enzyme, 2.5-3.6 nM, was incubated with the stated concentration of reagent in citrate-Pipes-Tris-Ampso buffer, pH 6.9, 1 mM dithiothreitol, for 30 min at room temperature and then an aliquot was assayed for enzymatic activity; results are the average of two experiments for each reagent. Control incubations had 91% activity relative to their activity at t = 0. * Abbreviations used: DFP, diisopropylfluorophosphate; DCIC, 3,4dichloroisocoumarin; APMSF, p-amidinophenylmethylsulfonyl fluoride; DPDS, 2,2’-dipyridyldisulfide. ’ No dithiothreitol present; after 30 min control incubations had 68% of their original activity.

30

KORNBLATT,

MPIMBAZA,

teases at l-2 pM (22), had no effect, even at 8 mM. These three effectors are all aromatic compounds. Because of the presence of the amidino group, two of these compounds are specific for trypsin-like enzymes, as opposed to chymotrypsin-like enzymes. The general lack of sensitivity of the T. brucei enzyme to these compounds suggests that the Lys/Arg-binding P1 subsite of the enzyme is smaller than that of trypsin and cannot accommodate the benzene ring of these compounds. Substrate specificity. Since only amide bonds with Arg or Lys at P1 were cleaved, the P, subsite specificity of the T. brucei endopeptidase is clearly trypsin-like. A comparison of the hydrolyses of tosyl-Gly-Pro-Arg-pNA and tosyl-Gly-Pro-Lys-pNA, which are identical except for Arg or Lys at Pi, indicates that Arg is preferred over Lys (Table II). Although the K,,, values for these two compounds are the same, the k,, is greater for substrate containing Arg than for that with Lys in the P1 position. The best substrate had Arg at both Pi and Pz. This preference is due to K,,, for this substrate being lower by a factor of 10 than that for any other substrate examined. Other good substrates had Phe, Gly, Ala, or Pro at Pz, (Table II, compounds 2-9). At P3 hydrophobic, neutral, or polar residues (Phe, Leu, Gln, Glu(OBz)) are all accepted by the enzyme but negatively charged residues are not (Glu, Glt). A comparison of the hydrolysis of Bz-Arg-AMC vs BzArg-pNA, of Bz-Phe-Val-Arg-AMC vs Bz-Phe-Val-ArgpNA and of Boc-Ala-Gly-Pro-Arg-AMC vs Tosyl-GlyPro-Arg-pNA (Table II) shows that the enzyme exhibits specificity for the leaving group. The p-nitroanilide substrates have Km values 3-4 times greater than that of the corresponding AMC substrates. The enzyme also has slightly lower ,& values (up to 25%) with the nitroanilide substrates than with the AMC substrates. The kinetic parameters for the hydrolysis of Bz-Arg-AMC (Km = 3 pM) and Bz-Arg-pNA (Km = 10 pM) can also be compared with the kinetic constants for the competitive inhibition observed with Bz-Arg-ethylester (Ki = 45 PM). [Under these conditions, the observed Ki is equivalent to Km (23).] The trend in K,,, values indicates that the enzyme prefers large, hydrophobic groups in the P’, subsite. Several substrates exhibited substrate inhibition. The substrates that showed pronounced substrate inhibition have the structures PI Pz P3

Arg Any amino acid Z, Glu(OBz), or Asp(OBz)

The fact that the diazomethyl compound that inactivates the enzyme also has this structure (Z-Leu-Lys-CHNJ suggests that the inactivation and substrate inhibition may occur via the same nonproductive mode of binding. Since both P1 and P2 will accept Arg (and, presumably, Lys), it is possible that, in the nonproductive mode, the Lys or Arg occupies Pz, rather than P1. This implies that

AND

LONSDALE-ECCLES

P4 is specific for large hydrophobic groups such as Z and Glu(OBz), a specificity that was not covered by the set of substrates we examined. In this model, Z-Leu-Lys-CHNz inactivates when bound in sites P2-P3-P4, with the benzyloxyl group in Pd. Boc-Val-Leu-Gly-Lys-CHN, is a poorer inactivator since, with Leu as the third residue, binding would be in the productive mode (which does not result in inactivation) rather than the nonproductive mode. Another possibility for nonproductive binding is that this large hydrophobic group occupies the leaving group site. This would also explain the difference in behavior of the two Lys-containing diazomethanes. Some of the substrates that are hydrolyzed by this endopeptidase have been reported to be suitable for the selective detection of specific enzymes that activate plasminogen (24) or components of the blood clotting system (25). Since we have shown that the T. brucei endopeptidase can be released from the parasite in vitro (26), we wished to learn whether the enzyme could activate some of the zymogens present in blood. The enzyme was inhibited by the active center hexapeptide (Ala-Arg-MetAla-Pro-Glu) from plasminogen activator inhibitor (27), which is consistent with its hydrolytic activity toward small substrates and its observed specificity for residues at P1 and Pz. The enzyme did not activate plasminogen, was not inhibited by the urokinase inhibitor from placenta, and had no effect on blood clotting time. The enzyme does, however, digest denatured fibrinogen (data not shown). The lack of generalized proteolytic activity in comparison to trypsin could indicate that we are looking at the intrinsic activity of a zymogen, rather than an enzyme (14,28-30). However, we have not found any other serine protease-like activity in our trypanosome preparations, nor have we observed any enhancement of activity in our preparations which could indicate we are dealing with a zymogen. In addition, the inhibition of this enzyme by aprotinin suggests that the enzyme can bind proteins at the active site. Thus we may be dealing with an enzyme whose specificity is tightly controlled by protein domains that are not present in smaller, trypsin-like molecules. Since the enzyme prefers dibasic residues, it may play a role in the processing of T. brucei proteins. Alternatively, since some T. brucei proteins, such as the glycolytic enzymes of the glycosome (31) have high isoelectric points, a protease that can cleave after both single and double basic residues may be necessary for normal protein degradation of such proteins. An alkaline peptidase with trypsin-like specificity has been purified from T. cruzi (3, 4). The pH optimum and substrate specificity are similar to those of the enzyme we have studied, in addition both are inhibited or inactivated by Z-Leu-Lys-CHNz and by DFP. However, the T. cruzi enzyme is also inhibited by Boc-Val-Leu-GlyLys-CHN, (which does not inhibit the T. brucei enzyme), has a larger molecular weight, and is inhibited by E-64.

T. brwei

31

ENDOPEPTIDASE

Ashall and co-workers (3,4) report that a peptidase with similar properties is present in other trypanosomes, including 2’. brucei. It is not clear to us what the relationship is between these two enzymes. They do not appear to be the same, and we have not seen any evidence for a second activity against Bz-Arg-AMC in our extracts. One possibility is that there are two activities but that the enzyme studied by Ashall (3) has been completely inactivated by the E-64 that we add before lysing the cells. The inactivation by DFP, plus the typical serine protease pH-activity profile lead us to conclude that the endopeptidase from T. brucei is a serine protease that does not exhibit the usual specificity toward inactivators. Its properties are similar to those of the postproline cleaving enzyme, which is inactivated by DFP (but not by phenylmethylsulfonyl fluoride), some sulfhydryl reagents, and by CBZ-Ala-Ala-Pro-CHNz (20,21). In order to confirm the identity of the enzyme as a serine protease, we intend to isolate and sequence the active site peptide. ACKNOWLEDGMENTS We thank S. Wasike for growing the trypanosomes. We also thank Dr. E. Shaw for the generous gift of the peptidyl diazomethylketones. This work was performed while M. J. Kornblatt was on sabbatical leave at the International Laboratories for Research on Animal Diseases. She thanks Concordia University for awarding the leave and ILRAD for its support and hospitality. This is ILRAD publication 954.

REFERENCES 1. North, M. J. (1982) Microbial. Reu. 46, 308-340. 2. Lonsdale-Eccles, J. D., and Grab, D. J. (1987) Eur. J. Biochem 169, 467-475. 3. Ashall, F. (1990) Mol. Biochem. Purositol. 38, 77-88. 4. Ashall, F., Harris, D., Roberts, H., Healy, N., and Shaw, E. (1990) Biochim. Biophys. Acta 1035, 293-299. 5. Mbawa, Z. R., Gumm, I. D., Fish, W. R., and Lonsdale-Eccles, J. D. (1991) Eur. J. Biochem 195, 183-190. 6. Waddell, W. J. (1956) J. Lob. Clin. Med. 48, 311-314. 7. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 8. Mahin, D. I., and Lofberg, R. T. (1966) Anal. B&hem. 16, 500-

509. 9. Magnusson, S. (1970) in Methods in Enzymology (Perlman, G. E., and Lorand, L., Eds.), Vol. 19, pp. 157-184, Academic Press, San Diego.

10. Dixon, M., and Webb, E. C. (1979) in Enzymes (Dixon, M., Webb, E. C., Thorne, C. J. R., and Tipton, K. F., Eds.), 3rd ed. pp. 126138, Elsevier Science, Amsterdam. 11. Schechter, I., and Berger, A. (1967) Biochem. Btiphys. Res. Commun.

27,157-162. 12. Brocklehurst, K., Willenbrock,

F., and Salih, E. (1987) in Hydrolytic Enzymes (Neuberger, A. and Brocklehurst, K., Eds.), pp. 39-158, Elsevier Science, Amsterdam. 13. Green, G. D. J., and Shaw, E. (1981) J. Biol. Chem. 256, 19231928. 14. Zumbrunn, A., Stone, S., and Shaw, E. (1988) Biochm. J. 250,

621-623. 15. Green, G. D. J., and Shaw, E. (1983) Arch. Biochem. Biophys. 225,

331-337. 16. Walter, R., Simmons, W. H., and Yoshimoto,

T. (1980) Mol. Cell. Biochem. 30, 111-127. 17. Watanabe, H., Green, G. D. J., and Shaw, E. (1979) Biochem. Biophys. Res. Commun. 89,1354-1360. 18. Morgan, P. H., Robinson, N. C., Walsh, K. A., and Neurath, H. (1972) Proc. N&l. Acod. Sci. USA 69, 3312-3316. 19. Kerr, M. A., Grahn, D. T., Walsh, K. A., and Neurath, H. (1978) Biochemistry 17,2645-2648. 20. Harper, J. W., Hemmi, K., and Powers, J. C. (1985) Biochemistry 24, 1831-1841. 21. Sturzebecher, J., Markwardt, F., and Walsmann, P. (1976) Z’hrombin Res. 9,637-646.

22. Laura, R., Robison, D. J., and Bing, D. H. (1980) Biochemistry 19, 4859-4864. 23. Segal, I. H. (1975) Enzyme Kinetics, pp. 113-120, Wiley, New York. 24. Zimmerman, M., Quigley, J. P., Ashe, B., Dorn, C., Goldfarb, R., and Troll,

W. (1978) Proc. N&l. Acad. Sci. USA 75, 750-753.

25. Kawabata, S., Miura, T., Morita, T., Kate, H., Fujikawa, K., Iwanaga, S., Takada, K., Kimura, T., and Sakakibara, S. (1988) Eur. J. Biochem. 172,17-25. 26. Lonsdale-Eccles, J. D., and Grab, D. J. (1987) J. Protozool. 34,405-

408. 27. Andreasen, P. A., Riccio, A., Welinder, K. G., Douglas, R., Sartorio, R., Nielsen, L. S., Oppenheimer, FEBS L&t. 209, 213-218.

C., Blasi, F., and Dano, K. (1986)

28. Gertler, A., Walsh, K. A., and Neurath, H. (1974) 1302-1310. 29. Lonsdale-Eccles, J. D., Neurath, H., and Walsh, chemistry 17,2805-2809. 30. Zur, M., and Nemerson, Y. (1978) J. Biol. Chem. 31. Marchand, M., Kooystra, U., Wierenga, R. K.,

Biochemistry

13,

K. A. (1978) Bio-

253, 2203-2209. Lambeir, A.-M., F. R., and Michels, P. A. M. (1989)

Beeumen, J. V., Opperdoes, Eur. J. Biochem. 184,455-464.