Báo cáo khoa học: The propeptide of cruzipain ) a potent selective inhibitor of the trypanosomal enzymes cruzipain and brucipain, and of the human enzyme cathepsin F ppt

of the trypanosomal enzymes cruzipain and brucipain, andof the human enzyme cathepsin F 5Flavia C.. Here we report potent and selective inhibition of cruzipain and brucipain by the recom

of the trypanosomal enzymes cruzipain and brucipain, and

of the human enzyme cathepsin F

5Flavia C G Reis1, Tatiana F R Costa1, Traian Sulea2, Alessandra Mezzetti2, Julio Scharfstein1, Dieter Bro¨mme3, Robert Me´nard2and Ana Paula C A Lima1

1 Instituto de Biofı´sica Carlos Chagas Filho, CCS, Universidade Federal do Rio de Janeiro, RJ, Brazil

2 Biotechnology Research Institute, National Research Council of Canada, Montreal, Canada

3 Department of Oral and Biological Sciences, University of British Columbia, Vancouver, Canada

Pathogenic trypanosomes are the cause of major

para-sitic diseases, which pose a threat to public health in

many developing countries [1] Trypanosoma cruzi is

the etiologic agent of Chagas’ disease, a chronic illness

endemic in Central and South America, associated

with heart failure and⁄ or denervation of the digestive

system Trypanosoma brucei causes sleeping sickness, a

widespread disease in sub-Saharan Africa that is

char-acterized by neurologic degeneration, psychiatric

dis-orders, and inevitable death if left untreated Present

chemotherapy relies on compounds presenting low

efficacy and⁄ or high toxicity, and there is an urgent

need for the development of alternative drugs

In many pathogenic parasites, the activity of

papain-like cysteine proteases (CPs) seems to be crucial for

growth, development and tissue⁄ host cell penetration

[2] The use of synthetic irreversible CP inhibitors in animals experimentally infected with T cruzi [3] or with T brucei [4] has significantly reduced parasitemia and mortality, validating these enzymes as promising targets for the development of new drugs The main targets of these compounds are the well-characterized major CPs of T cruzi, cruzipain (or cruzain) [5–7], and the CP of T brucei, brucipain (or rhodesain) [8,9] The resolution of the X-ray structure of cruzain in com-plex with diazomethane peptidyl inhibitors provided the basis for structure-based inhibitor design [10] Although several studies have explored the inhibition

of cruzipain by peptidyl epoxy ketones [11], aldehydes [12], vinyl sulfones [13], thiosemicarbazones [14] and alpha-keto-based compounds [15], there are still ongo-ing efforts to find new mechanism-based small-molecule

Keywords

brucipain; cathepsin F; cruzipain; inhibition;

propeptide

Correspondence

A P C A Lima

3 , Instituto de Biofı´sica

Carlos Chagas Filho, Bloco G, Centro de

Cieˆncias da Sau´de, Universidade Federal do

Rio de Janeiro, Cidade Universita´ria,

21949-900, Rio de Janeiro, RJ, Brazil

Fax: +55 21 2280 8193

Tel: +55 21 2209 6591

E-mail: anapaula@biof.ufrj.br

(Received 25 September 2006, revised 30

November 2006, accepted 22 December

2006)

doi:10.1111/j.1742-4658.2007.05666.x

Papain-like cysteine proteases of pathogenic protozoa play important roles

in parasite growth, differentiation and host cell invasion The main cysteine proteases of Trypanosoma cruzi (cruzipain) and of Trypanosoma brucei (brucipain) are validated targets for the development of new chemothera-pies These proteases are synthesized as precursors and activated upon removal of the N-terminal prodomain Here we report potent and selective inhibition of cruzipain and brucipain by the recombinant full-length prodo-main of cruzipain The propeptide did not inhibit human cathepsins S, K

or B or papain at the tested concentrations, and moderately inhibited human cathepsin V Human cathepsin F was very efficiently inhibited (Ki

of 32 pm), an interesting finding indicating that cruzipain propeptide is able

to discriminate cathepsin F from other cathepsin L-like enzymes Compar-ative structural modeling and analysis identified the interaction between the b1p–a3p loop of the propeptide and the propeptide-binding loop of mature enzymes as a plausible cause of the observed inhibitory selectivity

Abbreviations

CP, cysteine protease

4 ; PBL, proregion-binding loop; PCZ, recombinant propeptide of cruzipain.

inhibitors and structural information that could lead

to the development of efficient and specific inhibitors

for both cruzipain and brucipain

CPs are synthesized as inactive precursors that

pos-sess a signal peptide, a prodomain, and a mature

domain that retains catalytic activity In trypanosomal

CPs, the mature enzyme is composed of a central

domain that corresponds to the mature domain of

mammalian CPs and an additional 130-residue

minal extension with unknown function [2] The

C-ter-minal extension is not essential for enzyme activity

Activation of papain-like CPs is achieved by

proteo-lytic excision of the prodomain, an event believed to

occur in vivo by a multistep process that may involve

multiple endosomal⁄ lysosomal peptidases or even

pro-teases present in the extracellular environment [16,17]

Diverse functions have been attributed to the

prodo-main present in CP precursors (proenzyme): (a)

inhibi-tion of enzyme activity through its interacinhibi-tion with the

active site; (b) folding assistance [18]; and (c) targeting

of the precursors to the endosomal–lysosomal system

[19] Determination of the three-dimensional structure

of zymogens revealed that enzyme inhibition by the

propeptide is accomplished by its binding to the

enzyme’s active site with the backbone in the reverse

direction, thus protecting it from hydrolysis [20] The

discovery that the propeptides of cathepsins B [21], L

[22], K and S [23–25] are potent inhibitors that display

a certain degree of selectivity for their parent enzymes

has paved the way for the design of new selective

inhibitory compounds Indeed, this approach was

suc-cessful in the production of short peptidyl noncovalent

inhibitors based on the mode of inhibition of cathepsin

L by its prodomain These synthetic inhibitors present

300-fold selectivity for cathepsin L over cathepsin K,

whereas the full-length propeptide of cathepsin L

pre-sented only two-fold selectivity for the parental enzyme

[26]

Although CPs are widely expressed in parasitic

organisms, very little information has been reported

on the inhibitory properties of their prodomains The

propeptide of the main cathepsin L-like CP from the

worm Fasciola hepatica was described as highly

select-ive for the parasite’s enzymes, being practically

inca-pable of inactivating mammalian cathepsins L, K and

B and papain [27] In another study, the inhibitory

potential of 23 overlapping synthetic peptides

span-ning the propeptide of a CP from the cattle parasite

Trypanosoma congolense

described [28] This study identified a few peptides

presenting a certain degree of selectivity for the

inhi-bition of congopain and cruzipain relative to

cathep-sins B and L [28]

In this study, the propeptide of cruzipain was pro-duced and tested as an inhibitor of trypanosomal CPs Selectivity with regard to human lysosomal CPs was also investigated We show that the recombinant pro-peptide of cruzipain is a potent inhibitor of the mature enzyme, as well as brucipain, and further demonstrate that the propeptide of cruzipain selectively inactivates human cathepsin F with high potency The molecular basis for selective inhibition and the implications of these findings are discussed

Results and Discussion

Expression and purification of the propeptide

of cruzipain The predicted full-length propeptide of cruzipain (Cys19–Gly122) was cloned by PCR and expressed in Escherichia coli as a fusion protein with a 6xHis tag at the N-terminus This type of N-terminal tag was previ-ously used for the expression of the propeptide of human cathepsin K, and it did not interfere with the inhibitory activity of the protein [24] Although some of the recombinant propeptide of cruzipain (PCZ)

pro-duced in its soluble form, a great amount was found in inclusion bodies In order to maximize production and

to optimize purification, the fusion protein was dena-tured in a buffer containing urea, and was subsequently refolded and used in the inhibition assays The purity of refolded PCZ was checked by SDS⁄ PAGE (Fig 1A), and its chemical integrity was verified by MS Analysis

by MALDI-TOF MS showed a major single peak with a molecular mass of 14 019 Da, and a minor peak with a molecular mass of 14 161 Da (Fig 1B) The major peak corresponded to the full-length construct containing: (a) residues MRGS (introduced by the plasmid multiple cloning site immediately before the His-tag); (b) the six histidines of the tag; (c) the full-length propeptide of cru-zipain (Cys19–Gly122); and (d) residues RGVDLQPS-LIS at the C-terminus, which resulted from the fusion of the propeptide with a small stretch of coding sequence within the multiple cloning site of the pQE30 plasmid before the stop codon The predicted molecular mass of the fusion protein is 14 075 Da (14 kDa), which is within 1% error of the mass corresponding to the major peak This indicates that the great majority of the puri-fied propeptide was homogeneous and had no detectable truncations

Inhibitory activity of the propeptide Next, we tested the inhibitory properties of PCZ towards the parent enzyme, cruzipain, purified from

T cruzi epimastigotes We observed time-dependent

inhibition of cruzipain (Fig 2A) by PCZ, as noted

pre-viously for the inhibition of mammalian CPs by their

propeptides, which follows slow-binding kinetics, in the

pH 5–6 range [21–25] The linear portions of the pro-gress curves where steady state was reached were used to calculate the rate constant of substrate hydrolysis in the presence of the inhibitor (vi) and to determine Ki The Dixon plots (1⁄ v versus [I]) [29] resulted in intersecting lines, showing competitive inhibition (Fig 2B) At

pH 6.5, PCZ inactivated cruzipain competitively, with

an inhibition constant (Ki) of 0.018 nm Remarkably, this was the lowest Kivalue described so far for the inhi-bition of a papain-like CP by its cognate propeptide The propeptide was also found to be a very good inhib-itor of brucipain, with a Ki value of 0.0163 nm at

pH 6.5 This finding is not surprising, as cruzipain and brucipain share a high degree of sequence similarity, both in the mature part and in their propeptides Along these lines, a previous study using 15-mer synthetic

14.3

20 26.6

Mass (m/z)

0

878.2

0

10

20

30

40

50

60

70

80

90

100 14015.00

14063.56

A

B

Fig 1 Expression of the full-length propeptide of cruzipain.

(A) A DNA fragment encoding the full-length propeptide of cruzipain

(Cys19–Gly122) was cloned in fusion with a polyhistidine tag in

pQE30 and expressed in E coli Denatured fusion protein (PCZ)

was purified in an Ni–nitrilotriacetic acid column and subsequently

renaturated, filtered and dialyzed SDS ⁄ PAGE of the purification of

PCZ: bacterial homogenate after induction (lane 1), denaturated

PCZ eluted from Ni–nitrilotriacetic acid (lane 2), PCZ after

renatura-tion and dialysis (lane 3) (B) MS analysis by MALDI-TOF MS.

Approximately 5 nmol of the purified propeptide was precipitated

with trichloroacetic acid and analyzed by MS The measured mass

for His-tagged PCZ was 14 019.01 Da, which is within 56 Da of the

average calculated mass.

0 200 400

seconds

0.1 0.08 0.06 0.04 0.02 0 –0.02

1

0.8 0.6 0.4 0.2

A

B

Fig 2 Inhibition of cruzipain by its propeptide (A) Progress curves for the inhibition of cruzipain by its full-length recombinant propep-tide The figure shows the generation of product with time in the absence or in the presence of PCZ at the concentrations indicated for each curve (B) Plots of 1 ⁄ v versus [I] showing competitive inhi-bition The substrate concentrations used in the kinetic measure-ments are indicated in the graph.

peptides based on the prodomain of the CP congopain

from the cattle trypanosome T congolense showed

cross-inhibition of congopain and cruzipain [28]

How-ever, the Kivalues of the inhibitory peptides were in the

micromolar range

The potency of PCZ as an inhibitor of cruzipain

and brucipain decreased moderately at the more acidic

pH value of 5 (Table 1), but remained in the

subnano-molar range The very low Kivalues at pH values

ran-ging from 5 to 7 suggest that if full-length propeptide

was generated upon procruzipain maturation, it would

probably inhibit the mature enzyme very efficiently

Notably, it was suggested that in T cruzi, the

autocat-alytic processing of cruzipain zymogens occurs at the

Golgi complex [30], a compartment that usually

pre-sents only a mildly acidic environment In this context,

it is possible that the liberated full-length propeptide

exerts a crucial regulatory function in keeping

cruzi-pain inactive until it reaches lysosome-like organelles

(reservosomes) The propeptide of cruzipain is a potent

inhibitor of the cognate enzyme, and it could be

expec-ted that PCZ would likewise efficiently inactivate the

CPs from other trypanosomatids that share high

sequence similarity with cruzipain Indeed, a

prelimin-ary test using the homogenates of trypanosomatids

such as T rangeli, T brucei and Leishmania donovani

revealed that PCZ inhibited approximately 90% of the

overall peptidase activity detected in L donovani and

50–70% of the activity detected in the homogenates

from the other protozoa (Fig 3) These results suggest

that PCZ would be a good candidate as a lead for the

design of potent peptidyl inhibitors aimed at the

inacti-vation of CPs from several pathogenic protozoa

The propeptide of cruzipain is a potent inhibitor

of cathepsin F

To evaluate the selectivity of PCZ as an inhibitor of

papain-like enzymes, the propeptide was incubated for

30 min with papain or human cathepsins B, L, S, K, V

and F (in the subnanomolar range), and the residual activity of the peptidases was measured using a syn-thetic fluorogenic substrate This analysis revealed rel-atively weak inhibition of cathepsins L and V, whereas neither papain nor cathepsins B, S and K were inacti-vated by PCZ when incubated with up to 25 nm of the propeptide (Table 2) We observed degradation of PCZ by cathepsins S, L and V, but not by cathepsins

B or K after 60 min of incubation (data not shown) During the invasion of mammalian cells by T cruzi, the host cell lysosomes are recruited to the point of contact between host cell and parasites, and this is followed by the delivery of the lysosomal content to

Table 1 Dissociation constants for the inhibition of cruzipain by its

propeptide Cruzipain purified from T cruzi epimastigotes was

assayed for inhibition by the propeptide as described in

Experimen-tal procedures The steady-state velocities were calculated by linear

regression, and the results were used to plot 1 ⁄ v versus [I] to

cal-culate the Kivalues [36].

5.0 0.2637 ± 0.0001

5.5 0.0565 ± 0.0005

0,1 1 10

100

T rangeli

T brucei

L donovani

T cruzi

V o

Fig 3 The propeptide of cruzipain inactivates endogenous CPs of trypanosomatids Parasites were cultivated as described in Experi-mental procedures, and lysates were incubated with control buffer (clear bars), with 5 n M propeptide (gray bars), or with 30 l M E-64 (dark bars) The peptidase activity was measured by addition

of 10 l M Z-Phe-Arg-MCA and monitored continuously The graphs represent initial velocities calculated by linear regression of the progress curves.

Table 2 Inhibition of papain-like enzymes by the propeptide of cru-zipain The enzymes were assayed for inhibition by the propeptide

of cruzipain as described in Experimental procedures The steady-state velocities were calculated by linear regression and the results were used to plot 1 ⁄ v versus [I] to calculate the K i values [29] ND, not detected.

Enzyme K i (n M ) at pH 6.5

19

Cruzipain 0.018 ± 0.002 Brucipain 0.0163 ± 0.0001

Cathepsin F 0.032 ± 0.003

Cathepsin La 2.05 ± 0.25

Cathepsin V 5.18 ± 0.54

a Kidetermination was performed at pH 5, due to enzyme instabil-ity at higher pH values Degradation of PCZ was observed when it was incubated with cathepsins S, K, L and V at [E] > 500 n M for over 60 min.

the extracellular environment [31] In view of our

observations, if cathepsins S or L are present at high

concentrations at the parasite–host cell point of

con-tact, they could mediate degradation of the propeptide

present therein

Interestingly, significant inhibition of cathepsin F by

PCZ was observed The inactivation constants for

inhi-bition of cathepsin F were subsequently determined

(Table 2), and showed that PCZ exhibits very high

potency towards cathepsin F, and is 150-fold less

effective in inhibiting cathepsin V Interestingly, the Ki

for cathepsin F was in the same range as that observed

with the cognate enzyme (cruzipain) and with

bruci-pain, although mature cathepsin F shares considerably

less sequence identity with mature cruzipain (49%)

than does mature brucipain (71%) The activity of

cathepsin L was minimally inhibited at the maximal

concentration of PCZ used (25 nm) The Ki could not

be determined at pH 6.5, because cathepsin L

inacti-vates over time at this pH However, when determined

at pH 5.0, the Kifor inhibition of cathepsin L by PCZ

was in the same range as observed for the inactivation

of cathepsin V This was expected, as cathepsin L is

very similar to cathepsin V, and its sequence identity

to cruzipain (43%) is also comparable to that of

cath-epsin V (41%)

In order to be able to inhibit endosomal⁄ lysosomal

CPs in living cells, PCZ would have to be, at least to

some extent, resistant to degradation by peptidases

that are present in these acidic compartments We

determined whether PCZ would be able to inhibit CPs

present in the pool of endogenous peptidases of

mam-malian cell homogenates The tissue distribution of

cathepsin L is quite broad, whereas cathepsins F and

V present a much more restricted expression pattern

It has been reported that both cathepsin V and

cathep-sin F are expressed in human monocyte-derived

macrophages [32,33] and that the latter is also

expressed in smooth muscle cells Therefore, we tested

the inactivation of endogenous CPs of human cells

with diverse origins by PCZ (Fig 4) The incubation

of whole cell homogenates with PCZ led to a 20–30%

drop in the total peptidase activity of

mono-cytes⁄ macrophages (U937 cell line), smooth muscle

cells and Hela cells, whereas it reduced the peptidase

activity of prostate epithelial cells (DU145) by 4%

Substrate hydrolysis by the cell homogenates was fully

inhibited by E-64 (data not shown), confirming that

we were measuring mainly the activity of papain-like

CP

8 s in our assays

Cathepsin F is a key enzyme responsible for the

deg-radation of the invariant chain (Ii) in macrophages,

playing a fundamental role in antigen presentation

[34] Cathepsin F also modifies low-density lipoprotein particles in vitro, is present in human atherosclerotic lesions [35], and partially degrades lipid free apoA-1 and high-density lipoprotein (HDL3) in macrophages, thus blocking cholesterol efflux from these cells [32] It would be interesting to know whether T cruzi can secrete the cruzipain propeptide in its intact form If

so, it is possible that antigen-presenting cells such as macrophages could internalize the propeptide, which, once in the endosomal⁄ lysosomal compartments, would be able to inhibit cathepsin F As cathepsin F participates in the degradation of Ii, indirectly modula-ting the levels of class II MHC molecules available for binding of antigenic peptides, one can speculate that inhibition of this peptidase by the propeptide could affect antigen presentation by such cells

Comparative structural analysis of papain-like CPs

Given the remarkable selectivity of PCZ as an inhib-itor of cathepsin F and of the two trypanosomal CPs tested, we set out to investigate the possible molecular basis of selectivity The sequence alignment of the propeptide domains of several papain-like enzymes (Fig 5A) revealed that the prodomains of cruzipain and of brucipain are highly similar The high sequence identity observed between the prodomains of the trypanosomal enzymes (60%) could account for the

Prostate epithelium

muscle

Monocytes/ macrophages 0

5 10 15 20 25 30

Fig 4 The propeptide of cruzipain inactivates endogenous CPs of mammalian cells Human cell lines were cultivated as described in Experimental procedures, and cell lysates (10 lgÆmL)1) were incu-bated with 5 n M propeptide or with control buffer, for 15 min at room temperature The residual peptidase activity was measured

by the hydrolysis of 5 l M Z-Phe-Arg-MCA The peptidase activity of all lysates was fully inhibited by 10 l M E-64 (not shown) The initial velocities were calculated by linear regression of the substrate hydrolysis curves, and the graph represents the percentage of inhi-bition, considering the peptidase activity of untreated lysates as 100% The graph is representative of three independent experi-ments.

potent cross-inhibition of brucipain by PCZ The

pro-domain of cruzipain shares the highest identity with

the C-terminal portion (residues Pro147–Leu251) of

the prodomain of cathepsin F (32%) and that of

cath-epsin W (33%) among the mammalian CPs, albeit at a

significantly lower level than with brucipain The PCZ

similarity to the prodomains of human capthepsins L,

V, S and K is significantly lower (18–25%), showing

that sequence conservation qualitatively correlates to a

certain degree with the overall inhibition pattern

shown in Table 2 PCZ is significantly different from

the prodomain of cathepsin B (not shown) and is not

expected to inhibit this carboxypeptidase, due to the

occluding loop in the primed side of the binding site

groove Although the prodomains of cathepsins F and

W share the same degree of similarity to the

prodo-main of cruzipain, the similarity of the mature doprodo-main

of cathepsin F to that of cruzipain is higher (48%)

than that between the mature domains of cruzipain

and cathepsin W (32%)

In order to understand the structural basis for the

PCZ inhibitory selectivity among endopeptidases, we

modeled the three-dimensional structure of cruzipain

propeptide in complex with the crystal structure of

mature cruzipain The modeled PCZ–cruzipain

com-plex was then compared with the experimental

struc-tural data available for the other enzymes, particularly

surveying the residues and structural features at the

PCZ–cruzipain contact interface that are conserved in

brucipain and cathepsin F but not in cathepsins V, L,

S and K The most significant difference is located in

the loop connecting strand b1p to helix a3p of the

proregion, and occurs concertedly with a structural

change in the interacting portion of the mature enzyme

known as the proregion-binding loop (PBL) As seen

in Fig 5A, the b1p–a3p loop of the cruzipain

prore-gion, which is two residues shorter than those of the

cathepsin L, K, V or S proregions (see also Fig 5A),

would result in a steric conflict with the PBL side

chain Phe145 of mature cathepsin L, as well as with

the corresponding side chains at this position in

mature cathepsins K (Phe144), V (Phe146) and S

(Leu147) The interfering side chain appears to be

con-formationally restricted by the PBL main chain in all

these enzymes The collision with PCZ is, however,

cir-cumvented in the case of mature cruzipain and

cathep-sin F, due to a two-residue deletion in the PBL

sequence, which changes the main chain local

confor-mation relative to the other cathepsins, and in turn

moves the corresponding interfering side chain (the

smaller Thr147 of cruzipain, and Phe141 of

cathep-sin F) out of the way On the basis of homology-based

sequence alignments of mature enzymes, the same

would hold true for PCZ binding to brucipain and cathepsin W, which are, however, omitted from Fig 5B, due to the unavailability of experimental structures for these enzymes It remains to be tested whether PCZ is indeed a potent inhibitor of cathep-sin W, as predicted by this analysis However, active cathepsin W has not been successfully expressed to date

Conversely, the two-residue insertion Asn55p-(Ala⁄ His)56p in the proregion loop b1p–a3p of cathepsins

L, K, V and S relative to the cruzipain, brucipain and cathepsin F and W proregions (Fig 5A), allows good steric complementarity with the PBL side chain at position 145 (cathepsin L numbering) of the corres-ponding mature enzymes (Fig 5B), in addition to a hydrogen bond observed between the conserved Asn55p side chain and the PBL main chain at position

144 (not shown) The predicted structural determinant

of PCZ specificity for cathepsin F can be verified experimentally by mutagenesis in several ways, either

on the b1–a3 loops of propeptides or in the interacting PBL segments of the mature enzymes For example, a recombinant PCZ variant with a two-residue-longer b1–a3 loop, as in the prosegments of cathepsins L, K,

V and S, should exhibit a tendency to level off or even reverse the cathepsin inhibition pattern of the wild-type recombinant PCZ

Other structural features may be invoked to fur-ther explain the observed inhibitory selectivity of cruzipain propeptide against cathepsins V, L, K and

S (Table 2) In general, it is expected that the fine modulation of binding affinity and specificity between mature enzymes and the  100-residue pro-peptides results from the cumulative effect of small contributions scattered along the extensive binding interface For example, the shorter b1p–a3p loop of PCZ appears to force a small displacement of its helix a3p and to allow the accommodation of an additional residue (His81p) at the end of this helix, close to the active site, in comparison to the struc-tures of procathepsins L and K These more subtle structural changes may have a differential impact on the affinity of binding to mature cathepsins V, L, K and S A movement of helix a3p perpendicular to its axis has been documented between the proregions of cathepsins L and K, and is speculated to affect the proregion inhibitory selectivity [36] Also, residues Asn82p and Gly83p from the putative active site-spanning segment YHNGAA of PCZ are conserved

in the proregions of cathepsins L and K, where they bind in the reverse substrate-binding mode into the

S1¢ and S1 subsites, respectively [20] Their flanking residues, however, are more conserved in brucipain

and cathepsin F than in the cathepsin V, L, K and

S proregions, and thus they may represent another

source of selectivity among these enzymes, also

depending on the physicochemical properties of their

substrate-binding site groove Given that CP

prore-gions adopt folded structures upon binding to the

mature enzymes, we cannot exclude that even the

PCZ regions located remotely from the interface and

active site could be implicated in the fine-tuning of

inhibitory selectivity

The selectivity of PCZ for cathepsin F over other mammalian cathepsins is a remarkable observation indicating that it could serve as a lead for the develop-ment of potent selective inhibitors of cathepsin F The knowledge generated by our study could contribute to the design and production of potent small synthetic inhibitors selective for either cruzipain or cathepsin F Such selective inhibitors would be of great interest for the study of the biological functions and pathologic states in which these enzymes are implicated

A

B

Fig 5 Structural basis of PCZ inhibitory selectivity (A) Sequence alignment of propeptides from papain-like CPs Residues identical in at least half of the sequences are highlighted on a black background; those similar in at least half of the sequences are highlighted on a gray background Secondary structure elements, as observed in the crystal structure of procathepsin L (Protein Data Bank accession number: 1CS8), are indicated below the alignment Amino acid identities (%) of the cruzipain proregion with other CP proregions are given on the right (B) Interactions with the PBLs of the mature enzymes as a basis for the PCZ inhibitory selectivity The PBL segments of mature cruzi-pain (Protein Data Bank accession number: 1ME3, modeled in complex with its propeptide) and mature cathepsins F, L, K, V and S (Protein Data Bank accession numbers: 1MD6, 1CS8, 1BY8, 1FH0 and 1NQC, respectively), displayed as Ca traces, correspond to the sequences aligned in the upper right corner The PBL segments of brucipain and cathepsin W, whose propeptides are aligned in (A), are not shown in (B), due to the unavailability of experimental structures for these mature enzymes The PBL-interacting propeptide segments of cruzipain (model) and procathepsins L and K (Protein Data Bank accession numbers: 1CS8 and 1BY8, respectively), rendered by secondary structure, correspond to the sequences aligned in the lower left corner The last four displayed residues (underlined) of these propeptide segments extend into subsites S1¢ to S3 of the mature enzymes Cruzipain is shown in green, cathepsin F in magenta, and the other cathepsins in white An arrow indicates the two-residue insertions (highlighted) in the b1p–a3p loops of cathepsin L and K propeptides relative to the cruzi-pain propeptide This length variability in the propeptide sequences correlates with the steric requirements for interaction imposed by the conformation adopted by a PBL residue, displayed as a ball-and-stick model and highlighted in the sequence alignment.

Experimental procedures

Expression of the propeptide of cruzipain

and Gly122

(5¢-CGCCCGGGGCCAACAACCTCCACG-TC-3¢), which span the full-length predicted prodomain of

the cruzipain gene of T cruzi Dm28c, were used as primers in

a PCR using the cruzipain gene as a template and Pfx

polymerase (Invitrogen, Grand Island, NY, USA) The

amplification conditions were as follows: 30 cycles of

a 0.8% agarose gel, digested with BamHI and SmaI, cloned

into the respective sites of pQE30 (Qiagen, Valencia, CA,

USA), and sequenced in an automated Applied Biosystems

(Foster City, CA, USA)

(Qiagen) was used as a host for the expression of the fusion

protein containing a 6xHis N-terminal tag and the full-length

predicted propeptide of cruzipain The cells were grown in

reached Recombinant protein expression was induced by

4 h The cells were harvested by centrifugation

using a Sorvall Centrifuge RC28S (Wilmington, DE, USA)

with GSA rotor, washed in 20 mm phosphate buffer (pH 7.2)

for 1 h at room temperature The soluble material was

recov-ered by centrifugation

propeptide was purified in an Ni–nitrilotriacetic acid–agarose

resin according to the manufacturer’s instructions The eluted

45 min, followed by 100-fold dilution in ice-cold 100 mm

filtered through 0.2 lm cellulose acetate filters (Millipore,

Bedford, MA, USA), and centrifuged

using a Sorvall Centrifuge RC28S with F28⁄ 36 rotor, and the

soluble fraction was concentrated and dialyzed by successive

dilutions in ice-cold Milli-Q water by ultrafiltration in 3 kDa

cut-off DIAFLO membranes (Amicon, Bedford, MA, USA)

The titer of the propeptide was determined by active site

titra-tion of cruzipain The activated enzyme (previously titrated

with E-64) was incubated with different volumes of the

(pH 6.5), and 2.5 mm dithiothreitol, for 20 min at room

tem-perature The residual activity of the enzyme was determined

by the hydrolysis of 5 lm Z-Phe-Arg-AMC (Sigma, St Louis,

(pH 6.5), 2.5 mm dithiothreitol, and 5% dimethylsulfoxide Substrate hydrolysis was monitored in a Hitachi

spectrofluorimeter (Tokyo, Japan) at 380 nm excitation and

440 nm emission, in continuous assays The initial velocities were calculated by linear regression of the hydrolysis curves The propeptide titer was calculated by linear regression of the titration curves as being equal to the X-value at which

Proteolytic enzymes Cruzipain was purified from T cruzi Dm28c epimastigotes

as previously described [37] Human cathepsin B and human cathepsin S were purchased from Sigma and Calbio-chem (La Jolla, CA, USA), respectively Human cathepsins

L, V, K and F were expressed in Pichia pastoris as previ-ously described [38,39] The detailed cloning and produc-tion of recombinant brucipain will be described elsewhere (T F R Costa, F C G Reis, L Juliano, J Scharfstein, &

A P C A Lima, unpublished results) Briefly, a DNA fragment spanning the predicted first residue of the prodo-main (Cys22) and the last residue of the mature doprodo-main (Val339) of brucipain was cloned in the BamHI and KpnI sites of pQE30 for expression in E coli M15 (pREP4) The purification of His-tagged recombinant probrucipain and

propeptide of cruzipain, with the exception of a 6-day

The proenzyme was activated by diluting in 50 mm NaOAc (pH 5.5), 200 mm NaCl, 5 mm EDTA, and 1 mm dithio-threitol Mature enzyme was subsequently affinity-purified

in thiopropyl sepharose 6B (Pharmacia-Amersham, Upsala, Sweden), according to the manufacturer’s instructions The

enzyme concentration was determined by active site titra-tion with E-64, as previously described [40]

MS analysis The refolded propeptide (5.4 nmol) was precipitated with trichloroacetic acid and resupended in de-ionized water

(a-cyano-4-hydroxycinnamic acid), applied on the steel plate appropriate for the introduction of samples in the Voyager DE Pro equipment, and allowed to dry at room temperature The mass profile was subsequently obtained

Biosys-tem, Foster City, CA, USA)

Kinetic measurements Kinetic measurements were carried out in continuous assays using a Hitachi

enzymes were stable at the different pH values during

the assays The kinetics of inhibition of cruzipain were

measured using the Z-Arg-Arg-MCA substrate at 1.5 lm

started with the addition of enzyme to the cuvette The

steady-state velocities in the presence of 0–0.1 nm

average of both values was considered as the final value All

measurements were performed in 100 mm sodium phosphate

(pH 6–7) or 100 mm sodium acetate (pH 5–5.5) buffers

con-taining 5 mm EDTA, 200 mm NaCl, 2.5 mm dithiothreitol

and 5% dimethylsulfoxide The kinetics of inhibition of

cathepsin V and cathepsin L were determined using the

The kinetics of inhibition of cathepsin F was determined

using the Z-Phe-Arg-MCA substrate at 0.4 lm and 0.8 lm

experi-mental conditions used, progress curves for the inhibition of

cruzipain and cathepsin F were linear, whereas the progress

curves for brucipain showed curvature and fitted a model of

inhibition through slow binding kinetics

Cell culture and homogenates

The human monocyte cell line U937 (ATCC) was cultivated

in RPMI (Sigma), supplemented with 10% fetal bovine

semiconfluence was reached The cells were incubated with

10 nm 4b-phorbol 12-myristate 13-acetate (Sigma) for 3 days

to induce differentiation to macrophages Hela cells and

and-rogen-independent DU145 prostate cancer cells derived from

brain metastasis of human prostate cancer were obtained

from the ATCC, and human primary cultures of smooth

muscle cells (Cell Bank of Rio de Janeiro) were cultivated in

DMEM supplemented with 10% fetal bovine serum, as

des-cribed above (<20 passages), until semiconfluence was

reached The cells were washed twice with Hank’s balanced

salt saline before being lysed by addition of 50 mm NaOAc,

200 mm NaCl, 5 mm EDTA (pH 5.5), and 1% NP-40 The

soluble fraction was recovered by centrifugation

strain) were cultivated in HMI9 medium supplemented with

10% fetal bovine serum and 10% serum plus (Invitrogen) at

was a gift from G Atella (Federal University of Rio de

Janeiro, Brazil) The parasites were collected at mid-log phase, washed three times with Hank’s balanced salt saline, and lysed as described above The inhibition of peptidases by the propeptide of cruzipain was evaluated by incubating the

NaCl, 5 mm EDTA (pH 6.0) and 2.5 mm dithiothreitol with

5 nm recombinant propeptide for 15 min at room tempera-ture Controls were performed by incubating the lysates with

EDTA, 200 mm KCl, and 4% glycerol The residual pepti-dase activity was measured by the hydrolysis of 5 lm of

EDTA (pH 6.0), and 2.5 mm dithiothreitol The activity detected was fully inhibited by E-64, indicating that it was due to papain-like CPs

Sequence alignment and homology modeling The sequences of cruzipain and human cathepsins L, S, F, V,

K and W were retrieved from GenBank (accession numbers: M84342, NM_001912, NM_004079, BC036451, AB001928,

sequence of brucipain was determined in the laboratory after the cloning of the brucipain gene Propeptide sequences were initially aligned with the clustalw program [41] and using the structure-based sequence alignment between the prore-gions of procathepsins L and K whose crystal structures are available [42,43] (Protein Data Bank accession numbers: 1CS8 and 1BY8, respectively) The multiple sequence align-ment was further refined manually to ensure: (a) positioning

of gaps and insertions between secondary structure elements; and (b) suitability of the overall three-dimensional model of the cruzipain proregion bound to mature cruzipain

A structural model of the cruzipain propeptide bound to cruzipain was derived from the crystal structure of the mature enzyme [44] (Protein Data Bank accession number: 1ME3) and a homology model of the prodomain Homo-logy modeling was done with the composer program [45] and the protein loops module implemented in the sybyl 6.9 software (Tripos, Inc., St Louis, MO), using as template the crystal structure of procathepsin L [42] (Protein Data Bank accession number: 1CS8) and following the sequence alignment of the propeptides Thirteen crystallographic water molecules buried in the mature enzyme were retained Hydrogen atoms were added explicitly, and the protonation state at pH 6.5 was adopted (i.e ionized forms for the Asp, Glu, Arg, Lys and His side chains, for the chain termini, and for the catalytic residue Cys25) The model of the pro-peptide–cruzipain complex was refined by conjugate gradi-ent energy minimization with the AMBER molecular mechanics force field [46], using a distance-dependent

structural relaxation was accomplished in a four-stage step-wise minimization protocol: (a) only the propeptide loops corresponding to insertions and deletions relative to the

template structure were relaxed; (b) the entire propeptide

was relaxed in the context of the rigid mature portion;

(c) the propeptide and the cruzipain side chains in contact

with the propeptide were free to relax while the remainder

of cruzipain was constrained to the crystallographic

constraints were lifted, and the complex was minimized up

Acknowledgements

We thank Angela Alves and Alda Fidelis for technical

assistance We are grateful to Luiz Juliano (Federal

University of Sa˜o Paulo, Sa˜o Paulo, Brazil) for the

synthesis of e-NH2-(Cap)L-(SBzl)C-MCA, to Dr

Amil-car Tanuri (Federal University of Rio de Janeiro, Rio

de Janeiro, Brazil) for access to the Perkin-Elmer ABI

PRISM 3100 automated sequencer, and to Denis L S

Dutra (Federal University of Rio de Janeiro, Rio de

Janeiro, Brazil) for assistance with MS This work was

supported in part by grants from the Conselho

Nac-ional de Desenvolvimento Cientı´fico e Tecnolo´gico

(CNPq), the Fundac¸a˜o Carlos Chagas Filho de

Amparo a` Pesquisa do Estado do Rio de Janeiro

(FAPERJ), the Howard Hughes Medical Institute

(Departmental sharing grant no 55003669), NIH (AR

46182) (to DB) and the CIHR (to DB) This is NRCC

publication number 47553

References

1 WHO–TDR Chagas’ Disease http://www.who.int/tdr/

diseases/chagas/

2 Sajid M & McKerrow JH (2002) Cysteine proteases of

parasitic organisms Mol Biochem Parasitol 120, 1–21

3 Engel JC, Doyle PS, Hsieh I & McKerrow JH (1998)

Cysteine protease inhibitors cure an experimental

4 Scory S, Caffrey CR, Stierhof YD, Ruppel A &

Steverd-ing D (1999) Trypanosoma brucei: killSteverd-ing of bloodstream

forms in vitro and in vivo by the cysteine proteinase

5 Cazzulo JJ, Stoka V & Turk V (2001) The major

cysteine proteinase of Trypanosoma cruzi: a valid target

for chemotherapy of Chagas disease Curr Pharm Des 7,

1143–1156

6 McKerrow JH, McGrath ME & Engel JC (1995) The

cysteine protease of Trypanosoma cruzi as a model for

antiparasite drug design Parasitol Today 81, 279–282

7 Scharfstein J (2003) Activation of Bradykinin receptors

by Trapanosoma cruzi: a role for cruzipain in

microvasc-ular pathology

Disease, Vol 1, pp 111–137 Landes Bioscience, Austin,

TX

8 Troeberg L, Pike RN, Morty RE, Berry RK, Coetzer

TH & Lonsdale-Eccles JD (1996) Proteases from Trypanosoma brucei brucei Purification, characterisation and interactions with host regulatory molecules Eur J Biochem 238, 728–736

9 Caffrey CR, Hansell E, Lucas KD, Brinen LS, Alvarez Hernandez A, Cheng J, Gwaltney SL 2nd, Roush WR, Stierhof YD, Bogyo M et al (2001) Active site mapping, biochemical properties and subcellular localization of rhodesain, the major cysteine protease of Trypanosoma brucei rhodesiense Mol Biochem Parasitol 118, 61–73

10 McGrath ME, Eakin AE, Engel JC, McKerrow JH, Craik CS & Fletterick RJ (1995) The crystal structure

of cruzain: a therapeutic target for Chagas’ disease

J Mol Biol 247, 251–259

11 Roush WR, Gonzalez FV, McKerrow JH & Hansell E (1998) Design and synthesis of dipeptidyl alpha¢,beta¢-epoxy ketones, potent irreversible inhibitors of the cys-teine protease cruzain Bioorg Med Chem 8, 2809–2812

12 Scheidt KA, Roush WR, McKerrow JH, Selzer PM, Hansell E & Rosenthal PJ (1998) Structure-based design, synthesis and evaluation of conformationally constrained cysteine protease inhibitors Bioorg Med Chem 6, 2477–2494

13 Palmer JT, Rasnick D, Klaus JL & Bromme D (1995) Vinyl sulfones as mechanism-based cysteine protease inhibitors J Med Chem 38, 3193–3196

14 Fujii N, Mallari JP, Hansell E, Mackey Z, Doyle P, Zhou

YM, Gut J, Rosenthal PJ, McKerrow JH & Guy RK (2005) Discovery of potent thiosemicarbazone inhibitors

of rhodesain and cruzain Bioorg Med Chem 15, 121–123

15 Choe Y, Brinen LS, Price MS, Engel JC, Lange M, Grisostomi C, Weston SG, Pallai PV, Cheng H, Hardy

LW et al (2005) Development of alpha-keto-based inhibitors of cruzain, a cysteine protease implicated in Chagas disease Bioorg Med Chem 13, 2141–2156

16 Dalet-Fumeron V, Guinec N & Pagano M (1993) In vitro activation of pro-cathepsin B by three serine proteinases: leucocyte elastase, cathepsin G, and the urokinase-type plasminogen activator FEBS Lett 332, 251–254

17 Ishidoh K & Kominami E (2002) Processing and activa-tion of lysosomal proteinases Biol Chem 383, 1827– 1831

18 Tao K, Stearns NA, Dong J, Wu QL & Sahagian GG (1994) The proregion of cathepsin L is required for proper folding, stability, and ER exit Arch Biochem Biophys 311, 19–27

19 McIntyre GF & Ericksson AH (1993) The lysosomal proenzyme receptor that binds procathepsin L to microsomal membranes at pH 5 is a 43-kDa integral membrane protein Proc Natl Acad Sci USA 90, 10588–10592

20 Cygler M & Mort JS (1997) Proregion structure of members of the papain superfamily Mode of inhibition

of enzymatic activity Biochimie 79, 645–652