Báo cáo khoa học: Differences in substrate specificities between cysteine protease CPB isoforms of Leishmania mexicana are mediated by a few amino acid changes potx

Differences in substrate specificities between cysteine protease CPB
isoforms of
Leishmania mexicana
are mediated by a few amino acid
changes
Maria A. Juliano
1
, Darren R. Brooks
2
, Paul M. Selzer
3
, Hector L. Pandolfo
1
, Wagner A. S. Judice
1
,
Luiz Juliano
1
, Morten Meldal
4
, Sanya J. Sanderson
5
, Jeremy C. Mottram
2
and Graham H. Coombs
5
1
Department of Biophysics, Escola Paulista de Medicina, Universidade Federal de Sa
˜
o Paulo, Brazil;
2

Wellcome Centre for Molecular
Parasitology, The Anderson College, University of Glasgow, UK;
3
Akzo Nobel, Intervet Innovation GmbH, BioChemInformatics,
Schwabenheim, Germany;
4
Center for Solid-Phase Organic Combinatorial Chemistry, Department of Chemistry, Carlsberg
Laboratory, Valby, Denmark;
5
Division of Infection and Immunity, Institute of Biomedical and Life Sciences,
Joseph Black Building, University of Glasgow, UK
The CPB genes of the protozoan parasite Leishmania mex-
icana encode stage -regulated cathepsin L-like cysteine pro-
teases that are important virulence factors and are in a
tandem a rray o f 1 9 genes. In this study, w e h ave c ompared
the substrate preferences of two CPB isoforms, CPB2.8 and
CPB3, a nd a H 84Y muta nt of the latter e nzyme, to analyse
the roles played by the few amino acid differences between
the isoenzymes in determining substrate specificity. CPB3
differs from CPB2.8 at just three residues ( N60D, D61N and
D64S) in the mature domain. The H84Y mutation mimics
an additional change present in another isoenzyme, CPB18.
The active recombinant CPB isoenzymes and mutant were
produced using Escherichia coli and the S
1
-S
3
and S
1
¢-S

3
¢
subsite specificities determined using a series of fluorogenic
peptide derivatives in which substitutions were made on
positions P
3
to P
3
¢ by natural amino acids. Carboxydipep-
tidase act ivities of CPB3 and H 84Y were also observed u sing
the peptide Abz-FRAK(Dnp)-OH and some of its ana-
logues. The kinet ic parameters of hydrolysis by C PB3, H84Y
and CPB2.8 of the synthetic substrates indicates that the
specificity of S
3
to S
3
¢ subsites is influenced greatly by the
modifications at amino acids 60, 61, 64 and 84. Particularly
noteworthy was the large preference fo r Pro in the P
2
¢
position for the hydrolytic a ctivity of CPB3, which may be
relevant to a role in the activation mechanism of the
L. mexicana CPBs.
Keywords: carboxydipeptidase; cysteine protease; fluoro-
genic p eptides; Leishmania ;parasite.
Cysteine proteases (CPs) are p resent in almost all organisms
and are associated with numerous physiological and
pathological conditions [1,2]. Cysteine proteases of the

papain supe rfamily, designated Clan CA, family C1 [3], are
synthesized as zymogens that are activated by cleavage of
the pro-domain to generate mature enzymes located
predominantly within lysosomes. The m ature protease folds
into an ellipsoid conformation with the a ctive site cleft
located between two structural domains. One domain
consists predominantly o f b-barrel folds, while a prominent
central helix of th e second domain is adjacent t o, and helps
define, the opposite side of the active site cleft [ 4].
Leishmania mexicana possesses three CPs of the papain
superfamily, d esignated CPA and CPB, both of which are
cathepsin L -like, and CPC, w hich is cathepsin B -like [ 5]. The
CPB proteases exist as multiple isoenzymes, which are
encodedbyatandemarrayof19similarCPB genes l ocated
in a single locus [6,7]. L. mexicana CPB i soenzymes are
expressed as inactive zymogens c omprising an 18 amino
acid pre-region that is thought to be rapidly removed by a
signal peptidase upon transfer into the endoplasmic reticu-
lum,a106aminoacidpro-region,a218aminoacidmature
domain that includes the active site, and a C-terminal
domain of either 16 or 100 amino acids [7]. The first two
genes of t he array, CPB1 and CPB2, are atypical because
they encode enzymes with a C-terminal domain of just 16
amino a cids [7]. Furthermore, CPB1 and CPB2 are
expressed almost exclusively in the infective metacyclic
stage, whereas the remaining isogenes, namely CPB3–
CPB17 (which include CPB2.8 and CPB3)andCPB18 are
expressed p redominantly i n a mastigotes, w hereas CPB19 is
a pseudogene [8]. The role of the C-terminal domain
remains unc ertain. Roles in intracellular targeting to the

megasomes, immune evasion and modulation of the
Correspondence to G. H. Coombs, Division of Infection & Immunity,
Institute of Biomedical and Life Sciences, Joseph Black Building,
University of Glasgow, Glasgow, G12 8QQ, UK.
Fax: +44 141 330 3516, Tel.: +44 141 330 4777,
E-mail: G.Coombs @bio.gla.ac.uk
Abbreviations:Abz,ortho-amino-benzoyl; AMC, 7-amino-4-methyl-
coumarin; CTE, C-terminal extension; DMF, dimethylformamide;
EDDnp, N-[2,4-dinitrophenyl]-ethylenediamine; K(Dnp), (2,4 di-
nitrophenyl)-e-NH
2
-lysine; MCA, 4-methylcoumarin-7-amide;
rCPB2.8, recombinant Leishmania mexicana cysteine protease CPB2.8
lacking the C-terminal extension, originally designated CPB2.8DCTE;
Suc-LY-MCA, N-succinyl-Leu-Tyr-7-amido-4-methylcoumarin;
t-Boc, tert-butyloxycarbonyl.
(Received 2 5 May 2004, revised 12 July 2004, accepted 28 J uly 2004)
Eur. J. Biochem. 271, 3704–3714 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04311.x
enzyme’s activity have all been postulated, although defin-
itive data are lacking [5].
Information about the functions and importance of the
Leishmania enzymes in host–parasite interactions has been
obtained by the generation of mutants deficient in the
multicopy CPB gene array (Dcpb). L. mexicana Dcpb
mutants have reduced virulence w ith poor lesion growth
in BALB/c m ice and induce a pr otective Th1 response [ 6,9].
Reinsertion of the amastigote-specific CPB2.8 or meta-
cyclic-specific CPB2 into Dcpb mutantsfailedtorestore
either a Th2 response or sustained virulence, and only the
re-expression of multiple CPB genes from a cosmid

significantly restored virulence [10].
A recombinant form of the enzyme encoded by CPB2.8
but lacking the C-terminal extension, originally designated
CPB2.8DCTE but herein nam ed r CPB2.8 (Table 1 lists
nomenclature of all proteins analysed), was expressed [11],
and its substrate specificity h as been studied extensively
[12–15] and several peptide inhibitors have also been
reported for it [16–18]. The CPB3 gene, originally designa-
ted cDNA C PB as it was isolated from a cDNA library [19],
is another CPB gene from the central region of the array [7].
The corresponding protein, CPB3, when expressed in Dcpb
mutants was devoid of the gelatinase activity in nondenat-
urating g el electrophoresis that was o bserved f or CPB2.8 [7].
These two CPB isoforms d iffer from e ach other in the
mature enzyme domain in only t hree positions, CPB2.8 h as
Asn60, Asp61 and Asp64 whereas CPB3 has Asp60, A sn61
and Ser64. Interestingly, CPB18, which also has Asp60,
Asn61 and Ser64 but also Tyr84 and Asn18 instead of the
His84 and Asp18 in CPB2.8, is active towards gelatin but
differs f rom CP B2.8 i n i ts a ctivity towards s ome s hort
peptidyl-7-amido-4-methylcoumarin substrates [7]. Thus the
substrate preferences of some CPB isoenzymes seem to be
determined by just a few amino acids.
In order to explore the effects on substrate utilization of
the r estricted l ocal amino acid variations of the CPB is oen-
zymes of L. mexicana, the recombinant CPB3 ( rCPB3), and
arecombinantH84YmutantofCPB3thatwasgenerated,
were expressed in Escherichia coli and their S
1
-S

3
and S
1
¢-S
3
¢
subsite (based o n t he Schechter & Berger nomenclatu re [20])
specificities i nvestigated in a systematic w ay u sing intramole-
cularly quenched fluorescence substrates derived f rom Abz-
KLRFSKQ-EDDnp ( where Abz is ortho-amino-benzoyl
and EDDnp is N-[2,4-dinitrophenyl]-ethylenediamine),
which were previously used to study the specificity of
rCPB2.8 [13]. Moreover, t h e locations of the varyi ng residues
were analyzed via molecular modeling to gain insight into
how the c hanges may impinge upon enzyme activity.
TherecombinantcysteineproteasecruzainfromTrypan-
osoma cruzi and r CPB2.8 of L. mexicana are both c athepsin
L-like and characteristically endopeptidases. However, we
have shown that these enzymes have carboxydipeptidase
activities and have compared them with those of human
recombinant cathepsin B and cathepsin L [21]. Therefore we
also comparatively a nalyzed the carboxydipeptidase activit-
ies of rCPB3 and r H84Y using the internally quenched
fluorescent peptide Abz-FRFK(Dnp)-OH and some of its
analogues, where K(Dnp) is (2,4 dinitrophenyl)-e-NH
2
-
lysine, in order to characterize further the importance of the
varying amino acids.
Materials and methods

Parasites
Leishmania mexicana (MNYC/BZ/62/M379) promastigotes
were grown in modified Eagle’s medium (designated com-
plete H OMEM medium when supplemented with 10% (v/v)
heat-inactivated fetal bovine serum, pH 7 .5) at 25 °Cas
described p reviously [7]. The required a ntibiotics were added
as follows: hygromycin B (Sigma) at 50 lgÆmL
)1
, phleomy-
cin (Cayla
1
, Toulouse, France) at 1 0 lgÆmL
)1
, and neo mycin
(G418, Geneticin, Life T echnologies Inc.) at 25 lgÆmL
)1
.
Molecular modeling
A homology-based protein model o f a Leishmania majorcath-
epsinL-likecysteineprotease(GenBanklocusU43706,PDB
identification code 1bmj) was built using
INSIGHTII
[22] soft-
ware (Accelrys Inc, San Diego, CA, USA) and the crystal
structures of papain [23] and cruzain [24] as reference proteins.
The L. mexicana CPB isoforms w ere then modeled by s uper-
imposition using
MIDAS PLUS
(Computer Graphics L aboratory,
University of California San Francisco, CA, U SA) [25,26].

Mutagenesis, constructs, transfections and production
of recombinant enzymes
Mutations were incorporated into pGL27, a pBluescript
SK– plasmid containing the S WB1a CPB cDNA gene [19]
now designated CPB3, using the QuikChange Site-Directed
mutagenesis kit (Stratagene) and the following reverse-
phase-purified oligonucleotides (only the sense strand
Table 1. Plasmids, L. mexicana cell lines, and proteins used in this study. CTE, C-terminal extension ; r, recombinant.
Plasmid
Cell lines (plasmids
expressed in Dcpb) Abbreviation
Expressed cysteine
protease Abbreviation Comments
pGL37 Dcpb [pXCPB3] DcpbCPB3 CPB3 CPB3 native CPB3
pGL43 Dcpb [pXCPB3
D18N
] DcpbCPB3D18N CPB3(D18N) D18N CPB3 mutated D18N
pGL44 Dcpb [pXCPB3
H84Y
] DcpbCPB3H84Y CPB3(H84Y) H84Y CPB3 mutated H84Y
pGL45 Dcpb [pXCPB3
D60N, N61D, S64D
] DcpbCPB3M3 CPB3(D60N,
N61D, S64D)
M3 CPB mutated D60N,
N61D, S64D
pGL46 Dcpb [pXCPB2.8] DcpbCPB2.8 CPB2.8 CPB2.8 native CPB2.8
pGL180 not applicable CPB2.8DCTE rCPB2.8 rCPB2.8 lacking CTE [11]
6
pGL400 not applicable CPB3DCTE rCPB3 rCPB3 lacking CTE

pGL401 not applicable CPB3(H84Y)DCTE rH84Y rCPB3(H84Y) lacking CTE
Ó FEBS 2004 Cysteine protease isoforms of Leishmania (Eur. J. Biochem. 271) 3705
primers are shown and the mutated sites are given in
bold): OL416 to generate pGL40: 5¢-GACGCCGGTGA
AGAATCAGGGTGCGTG-3¢, OL418 to generate pGL41:
5¢-CGAACGGGCACCTGTACACGGAGGACAGC-3¢
and OL420 to generate pGL42: 5 ¢-GCTGCGATGACA
TGAAC GATGGTTGCGACGGCGGGCTGATGC-3¢.
These mutant constructs were verified by sequence
analysis using an ABI 373 a utomated D NA sequencer
(PerkinElmer). The native and mutant CPB3 genes were
excised from pBluescript SK– with XbaI–XhoI. Blunt ends
were created using Klenow fragment (NEB) and these were
ligated to the SmaI site of the pX episomal shuttle vector
[27] to generate the pGL plasmids detailed in Table 1. The
CPB2.8 gene [6] was excised from p Bluescript SK– ( pGL28)
as a 2.0 kb Eco RV fragment and ligated to the SmaIsiteof
pX to generate pGL46 (Table 1). All pX-based constructs
were then used to transfect Dcpb [6].
Transfection of L. mexicana promastigotes was as des-
cribed previously [6]. Briefly, pX-based constructs were
prepared using Qiagen Tip100 columns as outlined by the
manufacturer. Transfection u tilized 10 lg o f DNA and
4 · 10
7
late-log phase Dcpb promastigotes. Following
electroporation, cells were allowed t o recover in 10 mL
complete HOMEM medium for 24 h at 25 °Candthen
transfectants were selected in complete HOMEM medium
containing 25 lgG418ÆmL

)1
.
To generate recombinant L. mexic ana CPBs, t he 203 bp
KpnI–SacIfragmentofpGL180(pQE-30CPB2.8DCTE)
[11] was replaced with the corresponding fragments from
pGL27 and pGL41 to give e xpression constructs pGL400
(pQE-30 CPB3) and pGL401 (pQE-30 H84Y) (Table 1).
These e ncode proteins comprising the pro- and mature
domains of the enzymes, and which lack the pre- and
C-terminal domains. The recombinant enzymes were pro-
duced without the C-terminal extension (CTE) to aid
refolding f rom t he insoluble inclusion body phase. The
production of active, mature recombinant enzyme using
E. coli was essentially as described previously for isoenzyme
CPB2.8 [11]. The concentration of the enzyme stock
solutions ( 11 l
M
) were determined b y active site titration
with human cystatin C, which was a generous gift from
M. Abrahamson (University of Lund, Sweden), using
Z-FR-7-amido-4-methylcoumarin (Sigma) a s the substrate.
Activity analyses of cysteine proteases using gelatin
SDS/PAGE
Parasite cysteine protease activities were analysed using
substrate SDS/PAGE a s d escribed previously [7,28].
Parasite cell lysates (10
7
cells) were subjected to electro-
phoresis under nonreducing conditions using 12% (w/v)
acrylamide gels containing 0.2% (w/v) gelatin. Following

electrophoresis, the gel w as washed for 1 h w ith 2.5%
(v/v)TritonX-100andthenincubatedfor2hin0.1
M
sodium acetate, pH 5.5, containing 1 m
M
dithiothreitol.
Gelatin hydrolysis w as detected by staining with Coomas-
sie Blue R-250 (0.25% w/v). When analysing activities
towards N-succinyl-Leu-Tyr-7-amido-4-methylcoumarin
(Suc-LY-MCA;
2
Sigma), the peptidyl amidomethylcou-
marin fluorogenic substrate was added to 0.01 m
M
and,
following a 10 m in incubation, fluorescence was detected
by exposure of the gel to low intensity UV light [28].
Western blotting
Western blotting utilized polyclonal a nti-CPB serum
(1 : 2500) raised against rCPB2.8 expressed in a nd purified
from E. coli [11].
Synthesis of Abz-peptidyl-Q-EDDnp
All the intramolecularly quenched fluorogenic peptides
contain N-[2,4-dinitrophenyl]-ethylen ediamine ( EDDnp)
attached to glutamine. This is a necessary result o f the
solid-phase peptide synthesis strategy employed, t he details
of which were provided elsewhere [29]. An automated
bench-top simultaneous multiple solid-phase peptide syn-
thesizer (PSSM 8 system; Shimadzu, Tokyo, Japan) was
used for the solid-phase synthesis of all the peptides b y the

Fmoc-procedure. The final de-protected peptides were
purified by semipreparative HPLC using an Econosil C-18
column (10 lm, 22.5 · 250 mm) and a two-solvent system:
(A) trifluoroacetic acid/H
2
O (1 : 1000) and (B) trifluoro-
acetic acid/acetonitrile/H
2
O (1 : 900 : 100). The column
was eluted a t a flow rate of 5 mLÆmin
)1
with a 1 0 ( or 30))50
(or 6 0)% g radient of solvent B o ver 30 or 4 5 min. Analytical
HPLCwasperformedusingabinaryHPLCsystemfrom
Shimadzu with a SPD-10AV Shimadzu UV-vis detector
and a Shimadzu RF-535 fluorescence detector, coupled to
an Ultrasphere C-18 column (5 lm, 4.6 · 150 mm) which
was e luted w ith s olvent systems A1 ( H
3
PO
4
/H
2
O, 1 : 1000)
and B1 (acetonitrile/H
2
O/H
3
PO
4

,900:100:1)ataflow
rate of 1.7 mLÆmin
)1
and a 10–80% gradient of B1 over
15 min. The H PLC column eluates were monitored b y their
absorbance at 220 nm and by fluorescence emission a t
420 nm f ollowing ex citation at 320 nm. The molecula r m ass
and purity of synthesized peptides were checked by
MALDI-TOF mass spectrometry (TofSpec-E, Micromass,
Manchester, UK)
3
and or p eptide sequencing using a protein
sequencer PPSQ-23 (Shimadzu). The concentrations of the
solutions of the substrates were determined by colorimetric
determination of 2,4-dinitrophenyl group (extinction coef-
ficient at 365 nm is 17 300Æ
M
)1
Æcm
)1
)
4
.
Enzymatic hydrolysis of fluorescent quenched substrates
Hydrolysis of the fluorogenic peptide substrates by
rCPB2.8, rCPB3 and rH84Y were carried out in 0.1
M
sodium acetate, 2 m
M
EDTA, 200 m

M
NaCl, pH 5.5, at
37 °C. All kinetic analyses were carried out at 37 °Cwith
5 min enzyme preincubation in 2.5 m
M
dithiothreitol and
by measuring the fluorescence at 420 nm, following excita-
tion at 320 n m, using a Hitachi F-2500 spectrofluorometer
to follow the Abz-peptidyl-Q-E DDnp s ubstrate hydrolysis.
The k inetic param eters were calc ulated according Wilkinson
[30] as well as by using Eadie–Hofstee plots. The standard
derivations of K
m
and k
cat
determinations were in no case
higher than 5% of the obtained value.
HPLC analysis of the enzymatic hydrolysis products
of the synthetic fluorogenic substrates
The c leaved p eptide bonds in each substrate w ere i dentified
by isolation of the fragments by HPLC reverse-phase
chromatography on a C 18 column equilibrated in 10%
3706 M. A. Juliano et al.(Eur. J. Biochem. 271) Ó FEBS 2004
solvent B (90% acetonitrile, 0.1% t rifluoroacetic acid, v/v).
The column w as eluted at a flow r ate of 1 mLÆmin
)1
with 10–
80% gradient of solvent B o ver 28 min. The elution profile
was monitored by absorbance at 220 nm and by fluorescence
at 420 nm after excitation at 320 nm. The Abz-containing

fragments were compared with authentic synthetic sequences
and/or by amino acid sequencing and molecular mass
determination by M ALDI-TOF mass spectrometr y.
Results
Correlation between structure and activity of CPB
isoenzymes using substrate SDS/PAGE
The CPB locus of L. mexicana consists of 19 genes in a
tandem array [7]. The proteins encoded by three of these
genes differ in j ust a few a mino acids. The p resent study was
based on the finding that CPB2.8 and CPB18 are both
highly active towards gelatin as a substrate w hen a ssessed b y
in situ substrate SDS/PAGE, whereas CPB3 was inactive
[6,7]. CPB3 differs from CPB2.8 i n just t hree residues (60, 61
and 64) whereas t he only difference between CPB3 and
CPB18 are residues 18 and 84. Thus it was reasoned that
one or more of these c hanges must play a key role in
modulating the enzyme activity. There appeared two
possible explanations for the activity differences observed:
that the amino acid substitutions modulated enzyme activity
directly, or they had an indirect effect by influencing the
enzyme’s folding and stability. To investigate t he role of the
different amino acid residues we generated two mutants of
CPB3 in which residues 18 and 84 were changed to those
present in CPB18. We also mutated residues 60, 61 and 64
of CPB3 to those present in CBP2.8 – a s a positive control
for the procedure.
Incorporating mutations into the CPB3 gen e and intro-
ducing the mutated genes into Dcpb by transfection a llowed
an investigation into the functional roles of these amino
acids. The activity of the enzyme expre ssed in the parasite

was then analyzed towards gelatin and a small fluorogenic
peptidyl substrate, in both cases using in situ substrate SDS/
PAGE. Conversion of three variant residues in the mature
domain of the CPB3 isoenzyme (Asp60, Asn61 and Ser64)
to those present in CPB2.8 (Asn60, Glu61 and Glu64) to
give a p rotein designated M3, restored gelatinase activity to
the protease as expected (Fig. 1A, lane 6). This demon-
strates that one or more of these residues play an i mportant
role either in modulating the activity towards gelatin or in
enabling the enzyme to reactivate after the electrophoresis
procedure used. Mutation of His84 to T yr84 (to g ive H 84Y)
also restored gelatinase activity to the CPB3 i soenzyme
(Fig. 1A, lane 5), whereas mutation of Asp18 t o Asn18 (to
give D18N) did not (Fig. 1A, lane 4). The level of
re-expressed CPBs was to the same order in all of the cell
lines, as assessed by Western blotting (Fig. 1B, lanes 3–7),
although there were differences in expression levels that may
account in part for the differences in proteolytic activities
apparent (for example between lanes 5–7 of Fig. 1A).
Interestingly, the H84Y a ctivity appeared to be somewhat
greater than that of M3. However, such in situ gel assays,
although they a re useful in providing qualitative r esults,
need to be interpreted with caution with respect to
quantitative data.
The mutated forms of CPB3 were also assessed for
activity towards a small fluorogenic peptide using the gel-
based a ssay, which a lso is u seful for assessing activity but is
not very quantitative (Fig. 2). As observed for gelatin, the
mutants expressing the CPB3 or D18N were inactive
towards Suc-LY-MCA ( Fig. 2A, lanes 1 a nd 2). In contrast,

the H84Y isoenzyme hydrolyzed Suc-LY-MCA well
(Fig. 2A, lane 3). The M3 enzyme was a lso active towards
this fluorogenic compound (Fig. 2A, lane 4). This was to be
expected, a nd so served as a positive control, as the mutant
has the same mature domain as CPB2.8 (Fig. 2A, lane 5),
which is active towards this substrate [7]. The higher
molecular mass activities towards Suc-LY-MCA (approxi-
mately 35 kDa) correspond to activated p recursor forms of
the isoenzymes [11]. To confirm t hat similar amounts o f the
re-expressed proteases had been applied to these activity
gels, a Western blot was performed on duplicate samples
with the CPB-specific antiserum (Fig. 2B).
These results indicated that residues 60, 61, 64 and 84
influenced the activity of CPB, thus we produced as
recombinant enzymes CPB3 and also H84Y in order to
carry out a fuller analysis of their substrate specificities and
123456 7
123 4 5 6 7
A
B
30 -
22 -
42 -
30 -
22 -
kDa
kDa
Fig. 1. Gelatin SDS/PAGE and West ern blot analyse s of L. mexicana
CPB isoenzym es ex pressed in Dcpb. (A) Extracts from 10
7

stationary
phase promastigotes were used for gelatin SDS/PAGE. Wild type
parasites (lane 1), Dcpb (lane 2), Dcp bCPB3 (lane 3), DcpbCPB3D18N
(lane 4), DcpbCPB3H84Y (lane 5), DcpbCPB3M3 (lane 6) and
DcpbCPB2.8(lane7).Thehighermolecular mas s activities (approxi-
mately 35 kD a) evident with some samples resulted from the activation
in situ of precursor forms of the isoenzyme s [11]. (B) E xtracts from
5 · 10
6
stationary phase promastigotes were used for Western b lotting
with anti-CPB serum (1 : 2500 ). Samples were applied to lanes 1–7 as
denoted for (A). Molecular mass markers are shown in kDa. The anti-
CPB serum recogn ized CPB isoenzymes tha t migrated as a major band
with a molecular mass of 27 kDa and a m inor band o f molecular m ass
26 kDa in cell l ysates of stationary p hase p roma stigotes o f wild type
L. mexicana (lane 1). The specificity of the antiserum was confirmed by
absence of d etected proteins in Dcp b (lane 2).
Ó FEBS 2004 Cysteine protease isoforms of Leishmania (Eur. J. Biochem. 271) 3707
compare them with that of CPB2.8 [13]. All enzymes were
produced without the C -terminal extension as previously
[11].
S
1
subsite specificity characterization of CPB3 and H84Y
A series derived from the peptide Abz-KLRFSKQ-EDDnp
were synthesized with systematic variation of Arg at P
1
using all natural amino acids as previously reported for the
subsite specificity studies of rCPB2.8 [13]. Table 2 shows
the kinetic parameters for the hydrolysis of this series of

peptides by CPB3 and H84Y, and, for comparison, also the
k
cat
/K
m
and K
i
values of rCPB2.8.
The substrate i nhibition w ith t he peptides containing
hydrophobic and non-charged amino acids that occurred
with rCPB2.8 was not observed with rCPB3 and rH84Y,
and some hydrolysis o ccurred with a ll substrates. H owever,
the s pecificity c onstants ( k
cat
/K
m
) v alues f or the se proteases
were considerably lower than those obtained with r CPB2.8.
The higher k
cat
/K
m
values for rCPB2.8 were due to both
lower K
m
s and higher k
cat
s. For example, with X ¼ R, the
rCPB2.8 values were 0.04 l
M

(K
m
) and 2.60 s
)1
(k
cat
)
compared with the values of 0.3 l
M
(K
m
)and0.48s
)1
(k
cat
)
for rCPB3 (Table 2).
rCPB3 hydrolyzed with highest k
cat
/K
m
values the
peptides containing Ser and Thr, followed by those with
Phe, Arg, Lys and Tyr. These higher c atalytic efficiencies are
mainly due to the low K
m
value rather than the catalytic
component. Similar hydrolytic behaviour was observed
with rH84Y and the best substrates were those containing
Met, Ser and Thr.

30 -
22 -
42 -
A
B
kDa
21345
21345
30 -
22 -
42 -
kDa
Fig. 2. Fluorogenic SDS/PAGE and Western blot analyses of
L. mexicana CPBs expressed i n Dcpb. (A) Extracts from 10
7
stationary
phase promastigotes of Dcpb re-expressing the following proteases
were an alysed for activity t owards Suc-L Y-MCA
12
. DcpbCPB3 (lane 1),
DcpbCPB3D18N (lane 2), DcpbCPB3H 84Y (lane 3), DcpbCPB3M3
(lane 4) a nd DcpbCPB2.8 (lane 5). (B) Extracts from 5 · 10
6
stationary
phase promastigotes were analysed with a nti-CPB serum to show that
roughly equivalent protein loadings were applied for the fluo rogenic
SDS/PAGE analy ses: samples we re applied t o lanes 1– 5 as d enoted for
(A). Mole cular mass markers are shown i n kDa. The mature C PBs are
arrowed.
Table 2. Kinetic p arameters for hydrolysis, by CPB, of the p eptides derived from Abz-KLXFSKQ-EDDnpwithmodificationsinX(P

1
). Conditions of
hydrolysis: 100 m
M
NaOAc, 200 m
M
NaCl, 2 m
M
EDTA, pH 5.5 and 3 7 °C. The enzymes were p reactiva ted by 2.5 m
M
dithiothreitol
7
for 5 min.
The cleavage si te is indicated b y ÔflÕ, and the numbe rs following give
8
the percentage of hydro lysis at each peptide bond. All the other hydrolyses were
at the X–F bond. rCPB2.8 data from reference [13]. The units for K
i
values are n
M
,andk
cat
/K
m
values are i n (m
M
Æs)
)1
. X indicate s the residue varied.
Modification

rCPB2.8 rCPB3 rH84Y
k
cat
/K
m
(m
M
Æs)
)1
K
m
(l
M
)
k
cat
(s
)1
)
k
cat
/K
m
(m
M
Æs)
)1
K
m
(l

M
)
k
cat
(s
)1
)
k
cat
/K
m
(m
M
Æs)
)1
R 65000 0.3 0.48 1600 1.6 0.22 138
K 31395 0.1 0.12 1200 0.2 0.04 200
H 36667 0.5 0.40 800 0.5 0.09 180
D 871 0.2 0.03 150 0.1 0.01 100
E 7727 0.8 0.53 663 0.3 0.04 133
C 14060 0.6 0.47 783 4.3 0.26 61
W K
i
¼ 28 1.2 0.12 100 0.3 0.04 133
Y K
i
¼ 26 0.1 0.11 1100 0.2 0.06 300
F K
i
¼ 18 0.1 0.21 2100 0.3 0.08 267

L K
i
¼ 15 k
cat
/K
m
¼ 600 (XflF ¼ 55, FflS ¼ 45) k
cat
/K
m
¼ 700 (XflF ¼ 50, FflS ¼ 50)
I K
i
¼ 9 k
cat
/K
m
¼ 350 (XflF ¼ 15, FflS ¼ 85) k
cat
/K
m
¼ 100 (XflF ¼ 19, FflS ¼ 81)
V K
i
¼ 29 k
cat
/K
m
¼ 250 (XflF ¼ 24, FflS ¼ 76) k
cat

/K
m
¼ 300 (XflF ¼ 28, FflS ¼ 72)
M K
i
¼ 22 0.3 0.15 500 0.05 0.16 3200
A K
i
¼ 27 0.2 0.14 700 0.1 0.12 1200
P K
i
¼ 400 0.8 0.04 50 3.8 0.05 13
S K
i
¼ 24 0.1 0.28 2800 0.04 0.09 2250
T K
i
¼ 16 0.03 0.08 2667 0.04 0.09 2250
N K
i
¼ 32 k
cat
/K
m
¼ 300 (XflF ¼ 67, FflS ¼ 33) k
cat
/K
m
¼ 133 (XflF ¼ 67, FflS ¼ 33)
Q K

i
¼ 43 0.1 0.08 800 0.2 0.15 750
3708 M. A. Juliano et al.(Eur. J. Biochem. 271) Ó FEBS 2004
Most of the peptides were hydrolyzed by rCPB3 and
rH84Y at the X–F peptide bond, wher e X represents all the
substitutions of Arg. A second cleavage, a t the F–S bond,
was observed for the peptides with X ¼ Leu, Ile, Val and
Asn. Like cathepsin L, t he S
2
–P
2
interaction i s d eterminant
in defining the cleavage point and these cysteine proteases
prefer hydrophobic amino acids at P
2
position [31,32]. Thus
the cleavage of the F–S bond with Asn at the S
2
subsite by
rCPB3 and rH84Y is surprising, a lthough this cleavage
corresponded to only 33% of the total (Table 2).
S
2
and S
3
subsite specificity characterization of rCPB3
and rH84Y
The kinetic parameters for hydrolysis of the peptides
modified at positions P
2

and P
3
areshowninTable3.S
2
specificity is considered critical for the activity of clan CA
5
cysteine proteases [31]. The r CPB3 preferred L eu at P
2
position of the substrate, while Pro is the worst amino acid
of those t ested for activity. O n the other h and, rH84Y
hydrolyzed the peptide with Arg at P
2
with higher k
cat
/K
m
;
this was mainly due to k
cat
contribution because the K
m
was
relatively high. The extended binding site of rC PB3 and
rH84Y also included the S
3
subsite, as the m odifications at
P
3
position of the substrates r esulted in significant variations
on the k

cat
/K
m
values (Table 3). rCPB3 hydrolyzed with
better efficiency the peptides with Lys and Leu at P
3
,
whereas rH84Y was most efficient with the peptides
containing Leu and Ala.
S
1
¢ to S
3
¢ subsite specificity characterization of rCPB3
and rH84Y
The kinetic parameters for hydrolysis of the peptides
modified at positions P
1
¢ to P
3
¢ are shown in Table 4. The
trend of t he k
cat
/K
m
values obtained w ith t he substrates with
modifications at P
1
¢ position were similar between rCPB3
and rCPB2.8, however, t he latter enzyme h ydrolyzed all the

peptides of this series with k
cat
/K
m
values a t l east one order
of magnitude higher. In contrast to rCPB3 and rCPB2. 8, the
mutant enzyme rH84Y hydrolyzed the peptide with Phe
only poorly. This enzyme hydrolyzed best the peptide
containing Ala at P
1
¢, f ollowed b y those with L eu and Arg.
The peptide with Pro w as almos t resistant to all three
proteases, although a low r ate of hydrolysis w as observed at
the P–S bond.
The modification at the P
2
¢ position of the substrates
revealed the very significant effect of Pro, resulting in
considerably higher k
cat
/K
m
values for the h ydrolysis by
rCPB3 and rH84Y – even above those with rCPB2.8. These
high values mainly reflected a marked decrease in the K
m
value. The k
cat
/K
m

values for the hydrolysis of Abz-
KLRFPKQ-EDDnp by these two proteases were the
highest found. Such preference for Pro in the S
2
¢ subsite is
a peculiarity of cruzain (of T. cruzi)andLeishmania CPB.
These enzymes accept P ro in this position in synthetic
substrates very well [13,1 5,33,34] and a lso in t he auto-
processing of their pro-enzymes to active enzymes – Pro is
the second amino acid in the mature form of the proteases
[11,35]. The importance of S
2
¢–P
2
¢ interaction was further
evidenced by the variations in the k
cat
/K
m
values for the
hydrolysis of various substrates with modifications at P
2
¢
position by rH84Y – changing the Ser favoured by rCPB2.8
in each case resulted in increased activity.
Table 3. Kinetic parameters for hydrolysis, by CPB, of the peptides
derived from Abz-KLRFSKQ-EDDnp, containing modifications at the
Leu (P
2
) and Lys (P

3
)residues.Conditions of hydrolysis: 100 m
M
NaOAc, 200 m
M
NaCl, 2 m
M
EDTA, pH 5.5 and 37 °C. The enzymes
were preactivated by 2.5 m
M
dithiothreitol for 5 min. rCPB2.8 data
from reference [13].
X indicates the residue varied. The c leavage s ite is
indicated by ÔflÕ.
9
Substrate
rCPB2.8 rCPB3 rH84Y
k
cat
/K
m
(m
M
Æs)
)1
K
m
(l
M
)

k
cat
(s
)1
)
k
cat
/K
m
(m
M
Æs)
)1
K
m
(l
M
)
k
cat
(s
)1
)
k
cat
/K
m
(m
M
Æs)

)1
Abz-KXRflFSKQ-EDDnp
L 65000 0.3 0.48 1600 1.6 0.22 138
F 15517 0.9 0.13 144 0.8 0.17 212
A 2204 0.2 0.02 100 0.4 0.04 100
R 3295 0.3 0.04 133 1.4 0.66 471
P 471 0.5 0.01 20 4.8 0.39 81
Abz-
XLRflFSKQ-EDDnp
K 65000 0.3 0.48 1600 1.6 0.22 138
L 6167 0.05 0.07 1400 0.05 0.13 2600
A 16769 0.1 0.08 800 0.1 0.25 2500
H 6320 0.7 0.28 400 2.2 0.46 209
R 10421 0.4 0.12 300 0.5 0.29 630
Table 4. Kinetic parameters for hydrolysis, by CPB, of the peptides
derived from Abz-KLRFSKQ-EDDnp, containing modifications at the
Phe (P
1
¢), Ser (P
2
¢) and Lys (P
3
¢)residues.Conditions of hydrolysis:
100 m
M
NaOAc, 200 m
M
NaCl, 2 m
M
EDTA, p H 5 .5 and 3 7 °C. The

enzymes were preactivated by 2.5 m
M
dithiothreitol for 5 m in.
rCPB2.8 data from reference [13]. The units for K
i
values are n
M
.
X indicates the residue varie d.
10
The cleavage site is indicated by ‘ fl’.
11
Substrates
rCPB2.8 rCPB3 rH84Y
k
cat
/K
m
(m
M
Æs)
)1
K
m
(l
M
)
k
cat
(s

)1
)
k
cat
/K
m
(m
M
Æs)
)1
K
m
(l
M
)
k
cat
(s
)1
)
k
cat
/K
m
(m
M
Æs)
)1
Abz-KLRflXSKQ-EDDnp
F 65000 0.3 0.48 1600 1.6 0.22 138

L 14117 0.7 0.28 400 0.06 0.03 500
A 14882 0.2 0.09 450 0.3 0.28 933
R 5438 0.4 0.11 275 0.4 0.17 453
P
a
K
i
¼ 490 0.5 0.01 20 0.3 0.03 93
Abz-KLRflF
XKQ-EDDnp
S 65000 0.3 0.48 1600 1.6 0.22 138
F 27778 0.2 0.18 977 0.3 0.19 633
A 8909 0.1 0.10 774 0.06 0.11 1833
R 15833 0.3 0.32 944 0.3 0.50 1667
P 11875 0.01 0.18 18000 0.01 0.16 16000
Abz-KLRflFS
XQ-EDDnp
K 65000 0.3 0.48 1600 1.6 0.22 138
F K
i
¼ 40 0.5 0.04 80 0.1 0.23 2300
A 12727 0.6 0.20 333 0.6 0.12 200
R 24286 0.4 0.15 375 0.2 0.31 1550
P 11000 0.2 0.02 100 0.6 0.27 450
a
Cleavage at P–S bond.
Ó FEBS 2004 Cysteine protease isoforms of Leishmania (Eur. J. Biochem. 271) 3709
The S
3
¢ subsite a lso influenced the binding of rCPB3 and

rH84Y. Significant variations in the k
cat
/K
m
values were
observed with t he amino acid substitutions at P
3
¢ position of
the substrates (Table 4). T he peptide with Lys was hydro-
lyzed best by rCPB3 (it is also favoured by rCPB2.8),
whereas r H84Y was more efficient on the peptides contain-
ing Phe and Arg.
Carboxydipeptidase activity of rCPB3 and rH84Y
The kinetic parameters for the carboxydipeptidase activit-
ies of rCPB3 and rH84Y on the internally quenched
fluorescent peptide Abz-FRFK(Dnp)-OH and some of its
analogues are shown in T able 5. The k
cat
/K
m
values for
human recombinant c athepsin L and rCPB2.8 [ 21] are a lso
shown for comparison. The carboxydipeptidase activities
of rCPB3, and rH84Y were lower than those of cathepsin
L and rCPB2.8, although relative activities t owards the
different substrates were rather similar and in each c ase
Abz-FRAK(Dnp)-NH
2
was the best substrate. The sub-
strates with free C-terminal carboxyl group were hydro-

lyzed with K
m
values t hat are a n order of magnitude higher
than those presented in Tables 2–4. The unfavorable
effects of the C-terminal negatively charged carboxyl
group on the protease activity of rCPB3 and rH84Y is
demonstrated by the pH-profiles of the carboxydipeptidase
activities of these enzymes on the peptides Abz-
FRAK(Dnp)-OH and Abz-FRAK(Dnp)-NH
2
(Fig. 3 ).
The pH optima of t he carboxydipeptidase activities
towards Abz-FRAK(Dnp)-OH of rCPB3 and rH84Y are
displaced to 4.0–4.5 and the activity decreases greatly by
pH 5.5. This pH range corresponds to the pK of
carboxylate g roup formation from the substrate, indicating
that the protonated carboxyl group fits better to the
enzymes than its carboxylate form. This is confirmed by
the pH-profiles for the hydrolysis of Abz-FRAK(Dnp)-
NH
2
. I n this c ase, the pH optimum is in t he range 6–8. It is
noteworthy that the pH-profile of carboxydipeptidase
activity of rCPB2.8 on Abz-FRAK(Dnp)-OH contrasts
greatly with tho se of rCPB3 and rH84Y. It has optimal
activity at pH 6–8, showing that rCPB2.8 accommodates
much better the negatively charged carboxylate group.
Interestingly, the pH-profile o f h yd rolysis o f A bz-
FRAK(Dnp)-OH by rH84Y presents a second small but
significant peak of activity around pH 7.5. It is difficult to

assign the group responsible for this effect because there
are several groups that change their ionization status
around this pH.
Analysis of amino acid locations via modelling
The location of the amino acids that differ between the
CPB isoenzymes at positions 18, 60, 61, 64 and 84 was
determined by constructing a model of the L. mexicana
CPB isoenzymes by superimposing upon that obtained for
the enzymes’ homologue in L. major [22] (Fig. 4). The
crystal structure of papain [23] and of a recombinant
cysteine protease of T. cruzi, cruzain [ 24], w hich has a
high degree of sequence identity to CPB, was used as a
template for the leishmanial CPB model. The mature
regions of the L. mexicana and L. major enzymes used for
the modelling have an overall 80% amino acid sequence
identity, which reaches even higher values within the
structurally conserved regions and especially within the
active site (Table 6). Therefore, their protein structures
appear to be very similar as determined by secondary
structure alignments. The polypep tide backbone of these
cysteine proteases folds into a series of a-helices and
b-she ets and the active site cleft is located between two
structural domains. The C-terminal extension is not shown
on the model, as the structure of this d omain has not been
solved.
The p rotease activity of all pap ain-like cysteine protease s
is associated with the catalytic triad (L. mexicana residues
Cys25, His163 and Asn183; Fig. 4) but the substrate
specificity is defined by the bindin g affinities of the subsites.
The c atalytic residues and t he S

1
and S
1
¢ subsites are highly
conserved between the three parasite species, and key
residues a t the three major subsites are completely conserved
between the L. me xican a CPB isoenzymes and the L. major
homologue (Table 6). Consequently, differences in activitie s
between the L. mexicana CPB isoenzymes must be associ-
ated with amino acid var iations in more peripheral positions
of the molecule.
The model revealed the location of the amino acids that
differ between the CPB isoenzymes under study and so
mediate the ob served activity changes (the residues are
highlighted in Fig. 4B). Residues 60, 61 and 64 are located
above the a-helix that forms a wall of the active site cleft.
Amino acid 18 is relatively close to the active site cleft and
also near to one disulphide bridge (Cys22–Cys63), whereas
residue 84 is sited on a surface loop of one domain of the
Table 5. Kinetic constant parameters for ca rboxydipeptidase activities. Conditio ns o f hydrolysis: 100 m
M
NaOAc, 20 0 m
M
NaCl, 2 m
M
EDTA,
pH 5.5 and 37 °C. The enzymes were preactivated by 2.5 m
M
dithiothreitol for 5 min. Cath L a nd rCPB2.8 data a re from reference [21].
Substrates

Cath L rCPB2.8 rCPB3 rH84Y
k
cat
/K
m
(m
M
Æs)
)1
k
cat
/K
m
(m
M
Æs)
)1
K
m
(m
M
)
k
cat
(s
)1
)
k
cat
/K

m
(m
M
Æs)
)1
K
m
(m
M
)
k
cat
(s
)1
)
k
cat
/K
m
(m
M
Æs)
)1
Abz-FRFK(Dnp)-OH 256 306 3.3 0.23 70 3.9 0.51 130
Abz-RRFK(Dnp)-OH 2.2 4.5 K
i
¼ 4.3 l
M
Resistant
Abz-ARFK(Dnp)-OH 0.5 1.1 K

i
¼ 2.3 l
M
4.7 0.004 0.85
Abz-FRK(Dnp)W-OH 477 625 0.7 0.14 210 1.9 0.24 121
Abz-FRAK(Dnp)-OH 667 389 2.9 0.35 118 2.2 0.46 203
Abz-FRAK(Dnp)-NH
2
5739 4909 1.2 0.90 780 0.85 1.14 1340
3710 M. A. Juliano et al.(Eur. J. Biochem. 271) Ó FEBS 2004
enzyme but adjacent to another disulphide bridge (Cys56–
Cys101).
Discussion
The CPB cysteine proteases of L. mexicana have been
shown t o be v irulence factors [ 6], therefore it is
important to understand the relative contributions that
different isoenzymes play in the pathogenicity of the
parasite. Those isoen zymes that have been characterized
so far are highly conserved (98% identical) and yet some
activity differences were apparent [7]. The aim of this
study was to determine the extent to which these amino
acid variations generated activity differences. The CPB
sequences obtained to d ate all possess identical residues
aligning the substrate subsites so it seemed that other
residues outside of the active s ite cleft must affect
activity.
Abz-FRAK(Dnp)-NH
2
Abz-FRAK(Dnp)-OH
pH

345678
0
400
800
1200
1600
2000
2400
2800
109
pH
34567891
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0
pH
345678910
0
200
400
600

800
1000
pH
345678910
0
20
40
60
80
100
120
140
160
180
pH
345678910
0
40
80
120
160
200
240
pH
345678910
kcat/Km (mM
-1
.s
-1
)kcat/Km (mM

-1
.s
-1
)kcat/Km (mM
-1
.s
-1
)
kcat/Km (mM
-1
.s
-1
) kcat/Km (mM
-1
.s
-1
) kcat/Km (mM
-1
.s
-1
)
0
2000
4000
6000
8000
10000
12000
14000
16000

CPB3H84Y CPB2.8
Fig. 3. pH-profile a ctivity (k
cat
/K
m
) for the h ydrolysis of Abz-FRAK(Dnp)-OH a nd Abz- FRAK(Dnp)-NH
2
by rCPB2.8, rCPB3 and rH84Y. The
reactions were carried out in standard buffer containing 25 m
M
acetic acid, 25 m
M
Mes, 75 m
M
Tris base, 2 5 m
M
glycine, and 2 m
M
EDTA. The
pH range w as 3.5–10 a nd adjusted with 2
M
NaOH or HCl. The enzymes wer e preactivated by 2.5 m
M
dithiothreitol for 5 min a t 37 °C.
Ó FEBS 2004 Cysteine protease isoforms of Leishmania (Eur. J. Biochem. 271) 3711
The activity results presented show that the few amino
acid variations known to exist between some isoenzymes of
CPB of L. mexicana are i ndeed important in modifying the
substrate specificities of the CPB isoenzymes. Both rCPB3
and r H84Y have lower activity towards some substrates

than does rCPB2.8, but they are able to accommodate a
wider r ange of amino acids at P
1
(Table 2). The kinetic
parameters of hydrolysis by rCPB3, rH84Y and rCPB2.8 of
the substrates with variations at P
1
position indicated that
the s pecificity of S
1
subsite i s greatly influenced by the
modifications at 60, 61, 64 and 84. However, the enzymes
have an extended binding site that goes at least from S
3
to
S
3
¢ and importantly each CPB isoform also shows signifi-
cant differences in the specificity of each subsite (Tables 3
and 4). Thus, the overall conclusion is that the different
isoenzymes do indeed have different hydrolytic c apabilities
and presumably t his is important for the parasite’s survival.
This view is supported by the recent observation that the
expression of multiple CPB genes encoding cysteine
proteases, rather than just one, is required for L. mexicana
virulence in vivo [10].
A particularly noteworthy finding was t he effect of Pro in
the P
2
¢ position in decreasing considerably the K

m
value
with rCPB3 and rH84Y. Clearly these isoenzymes greatly
favour Pro at this s ite. This may reflect a n important role for
the enzymes in the activation o f CPB in the parasite, as Pro
isthesecondaminoacidofthematuredomaininallofthe
CPB isoenzymes for which the structure i s known.
The variation of a mino a cid residues 6 0, 61 and 64
between the CPB isoenzymes examined in this work involve
amino acids with charged side chains and this necessarily
results in significant modifications on the electrostatic
potential on the s urface of the e nzymes. The modeling
analysis (Fig. 4) shows clearly that residues 60, 61 and 64
are located near the c atalytic g roove of the e nzyme and so it
is likely that the localized charge variations resulting from
the introduction or removal of residues with charged side
groups to these positions could be the basis for the
differences observed with respect to the utilization of
substrates. The pH-profile differences for the carboxydi-
peptidase activities of rCPB3 and rCPB2.8 clearly suggest
that the latter isoform accommodates the substrate carb-
oxylate group better than the former isoform, perhaps
indicating that re sidue 60 (Asn i n CPB2.8 but Asp i n CPB3)
is particularly important for this binding. These findings
agree well w ith a study of the p H-activity profile of c ruzain,
a related cathepsin L -like c ysteine pro tease of T. cruzi,
which highlighted the importance of several ionizable
groups and suggested that Asn60 is potentially involved in
substrate recognition [36]. Thus the data obtained provide
further evidence for the role of electrostatic potential in

defining the substrate specificity of the CPB isoforms.
84
60
61
64
18
SS
63
22
25
163
183
SS
204
156
SS
101
56
A
B
Fig. 4. Homology-based protein model of the
mature domain of L. mexicana CPB. The
protein i s shown as a ribbon s tructure ( spi-
rals, a-helices ; arrows, b-sheets). (A) High-
lighted in yellow are the location of the active
site triad ( Cys25, His163 an d Asn183) and
disulphide bonds (Cys22–Cys63, Cys56–
Cys101 and Cys156–Cys204). (B) Highlighted
in red a re the residues that d iffer between
CPB2.8, CPB3 a nd CPB18.

Table 6. Key active site res idues of cathepsin
L
-likecysteineproteases.The major variations between the active sites of parasite CPs and papain occur
between the S subsit es. Additional S and SÕ subsites are n ot listed because they are so far not identified by the cocrystallizat ion of a peptide substrate
and the proteases o f the p arasites.
Protease S2 subsite S1 subsite Catalytic triad S1¢ subsite
Papain W69, S205, F207 Y67, P68, V133, A160 C25, H159, N175 Q19, W177
Cruzain N69, E208, S210 L67, M68, A138, G163 C25, H162, N182 Q19, W177
L. mexicana CPB L69, Y209, V211 L67, M68, A139, G164 C25, H163, N183 Q19, W185
L. major CPB L69, Y209, V211 L67, M68, A139, G164 C25, H163, N183 Q19, W185
3712 M. A. Juliano et al.(Eur. J. Biochem. 271) Ó FEBS 2004
Residue 84 is located near t he surface of the mature
domain s tructure and so it is less easy to understan d how it
plays a role in affecting t he binding of substrates and their
hydrolysis. However, it is positioned near to a disulphide
bridge (Cys56–Cys101) and it is conceivable that the
replacement of His with Tyr may impinge upon this
structure a nd so change the active site i s s ome w ay. Clearly
the findings on the activity of this mutant compared with
rCPB3 suggest that mutation had some effects, although
more minor than those resulting from the three changes
(N60D, D61N and D64S) between CPB2.8 and the other
isoenzymes.
The results show that rCPB3 has activity towards p eptide
substrates and yet the enzyme showed no activity in
substrate SDS/PAGE analysis, whereas both CPB2.8 and
CPB18 were both highly active in similar analyses (Figs 1
and 2). This suggests not only that one or more of amino
acid changes H60D, D61N and D64S plays a key role in
modulating the enzyme activity directly but may also be

able to do so indirectly by affecting the enzyme’s refolding
and/or stability under the conditions e mployed for the
gelatin SDS/PAGE. Moreover, H84Y but not D18N can
counteract this effect. It is too early to be able to interpret
the way is which this is achieved, but the results show the
complexity of the interactions that occur both within the
mature protein and in the acquisition of its tertiary
structure.
In conclusion, the data reported suggest that the set of
CPB isoenzymes with only a few sequence modifications
have the modifications at strategic positions such that the
enzyme’s substrate specificity i s changed and that t hese
variations between isoenzymes provide the parasite with an
array o f h ydrolytic activity that is needed for its interac tion
with the mammalian host, and ensure its survival and
success as a parasite.
Acknowledgements
This work was supported by F undac¸ a
˜
odeAmparoPesquisadoEstado
de Sa
˜
o Paulo (FAPESP), Conselho Nacional de Desenvolvimento
Cientı
´
fico e Tecnolo
´
gico (CNPq) and Human Frontiers for Science
Progress (RG 0 0043/2000-M). J.C.M. a nd G.H.C. are supported by the
Medical Research Council.

References
1. Bromme, D. & Kaleta, J. (2002) Thio l-dependent cathepsins:
pathophysiological implicatio ns and recent advances i n inhibitor
design. Curr. P harm. Des. 8, 163 9–1658.
2. Sajid, M. & McKerrow, J.H. (2002) Cysteine prote ases of parasitic
organisms. Mol. Biochem. Parasitol. 120, 1–21.
3. Barrett, A.J. & Rawlings, N .D. (2001) Evolutionary lines of
cysteine peptidases. Biol. C hem. 382, 7 27–733.
4. McGrath, M.E. (1999) The lysosomal cysteine proteases. Annu.
Rev. Bio phys. Biomol. Struc t. 28, 1 81–204.
5. Mottram, J.C., Brooks, D.R. & Coombs, G.H. (1998) Roles of
cysteine proteinases of trypanosomes and Leishmania in host–
parasite interactions. Curr. Opin. Microbiol. 1, 455–460.
6. Mottram, J .C., Souza, A.E., H utchison, J .E., Carter, R., Frame,
M.J. & Coombs, G.H. (1996) Evidence from disruption of the
lmcpb gene array of Le ishmania mexicana that cysteine proteinases
are virulence fac tors. Proc.NatlAcad.Sci.USA93, 6008–6013.
7. Mottram, J.C., Frame, M.J., Brooks, D.R., Tetley, L., Hutchison,
J.E., Sou za, A.E. & Coombs, G.H. (1997) The m ultiple cpb
cysteine proteinase genes of Leishmania mexicana encode iso-
enzymes which differ in their stage-regulation and substrate pre-
ferences. J. Biol. Chem. 272, 14285–14293.
8. Brooks, D.R., Denise, H., Westrop, G.D., Coombs, G .H. &
Mottram, J.C. (2001) The stage-regulated expression of Leish-
mania mexicana CPB cysteine proteases is mediated by an
intercistronic sequence element. J. Biol. Chem. 276, 47081–
47089.
9. Alexander, J., Coombs, G .H. & M ottram, J .C. ( 1998) Leishmania
mexicana cysteine proteinase-deficient mutants have attenuated
virulence for mice and potentiate a T Helper 1 (TH1) response.

J. Immunol. 161 , 6794–6801.
10. Denise, H., McNeil, K.S ., Brooks, D.R., Alexander, J ., Coombs,
G.H. & Mottram, J.C. (2003) The expression of multiple CPB
cysteine protease genes is r eq uired for Leishmania me xicana viru-
lence in vivo . Infect. Immun. 71 , 3190–3195.
11. Sanderson, S.J., Pollock, K.G., Hilley, J.D., Meldal, M ., St. Hil-
aire, P.M., Juliano, M.A., Juliano, L., Mot tram, J.C. & Coom bs,
G.H. (2000) Expression and characterisation of a recombinant
cysteine proteinase of Leishmania mexicana. Biochem. J. 347,
383–388.
12. Alves, L .C., Melo, R.L ., Sanderson, S. J., Mottram, J .C., Coombs,
G.H.,Caliendo,G.,Santagada,V., Juliano, L. & Juliano, M.A.
(2001) S
1
subsite specificity of a recombinant cysteine proteinase,
CPB, of Leishmania mexicana compared with cruzain, human
cathepsin L and papain using substrates containing non-natural
basic amino acids. Eur. J. Biochem. 268 , 1206–1212.
13. Alves, L.C., Judice, W .A.S., S t. Hi laire, P.M., Meldal, M.,
Sanderson, S.J., Mottram, J.C., Coombs, G.H., Juliano, L. &
Juliano, M.A. (2001) Substrate speci ficity of recombin ant cysteine
proteinase, CPB, of Leishmania mexicana. Mol. Biochem. Para-
sitol. 116, 1 –9.
14. Alves, L.C., Melo, R.L., Cezari, M.H., S anderson, S.J ., Mottram,
J.C., Coombs, G.H., Juliano, L. & Juliano, M.A. (2001) Analysis
of the S
2
subsite specificities of the recombinant cysteine protei-
nases CPB of Leishmania mexicana, a nd cruzain of Trypanosoma
cruzi, using fluorescent substrates containing non-natural basic

amino acids. Mol. Biochem. Parasitol. 117, 137–143.
15. St. Hilaire, P.M., Alves, L.C., Sanderson, S.J., Mottram, J.C.,
Juliano, M.A., Juliano, L., Coombs,G.H.&Meldal,M.(2000)
The substrate specificity of a recombinant cysteine protease from
Leishmania mexicana: application of a combinatorial peptide
library a pproach. Chembiochem. 1, 115–122.
16. Alves, L.C., St. Hilaire, P.M., M eldal, M., Sanderson, S.J., Mot-
tram, J . C., Coombs, G .H ., Juliano, L. & Juliano, M.A. (2001)
Identification o f peptides inhibitory to recombinant c ysteine pr o-
teinase, CPB, of Leishmania mexicana. Mol. Biochem. Parasitol.
114, 81–88.
17. St. Hilaire, P.M., Alves, L.C., Herrera, F ., Renil, M., S anderson,
S.J., Mottram, J.C., Coombs, G.H., Juliano, M.A., Juliano, L.,
Are
´
valo, J. & Meldal, M. (2002) Solid-Phase library synthesis,
screening and selection of tight-binding reduced peptide bond
inhibitors of a recombinant Leishmania mexicana cysteine pro-
tease B. J. Med. Chem. 45 , 1971–1982.
18. Graven, A., St. Hilaire, P.M., Sanderson, S.J., Mottram, J.C.,
Coombs, G.H. & Meldal, M. (2001) Combinatorial library of
peptide isosters based on D iels-Alder reac tions: Identifi cation of
novel inhibitors against a recombinant cysteine protease from
Leishmania mexicana. J. Com b. Chem. 3, 441–452.
19. Souza, A.E., Waugh, S ., Coombs, G.H. & Mottram, J .C. (1992)
Characterization of a multicopy ge ne fo r a major stage-specific
cysteine prote inase of Leishmania mexicana. FEBS Lett. 311,
124–127.
Ó FEBS 2004 Cysteine protease isoforms of Leishmania (Eur. J. Biochem. 271) 3713
20. Schechter, I. & Berger, A. (1967) On the size of the active site

in proteases. I. Papain. Biochem. Biophys. Res. Commun. 27 ,
157–162.
21. Judice, W.A., Puzer, L., Cotrin, S.S., Carmona, A.K., Coom bs,
G.H., Juliano, L. & Juliano, M.A. (2004) Carboxydipeptidase
activities of recombinant cysteine peptidases. Eur. J. Biochem. 27 1,
1046–1053.
22. Selzer, P.M., Chen, X.W., Chan, V.J., Cheng, M.S., Kenyon,
G.L., Kuntz, I .D., Sakanari, J.A., Coh en, F.E. & M cKerrow, J.H.
(1997) Leishmania major: Molecular modeling of cysteine pro-
teases and prediction of new nonp eptide inhibitors. Exp. Pa rasitol.
87, 212–221.
23. Kamphuis, I.G., Kalk, K.H., Swarte, M.A. & Drenth, J. (1984)
Structure of papain refined at 1.65 A
˚
resolution. J. Mol. Biol. 179,
233–256.
24. McGrath, M.E., Eakin, A.E., Engel, J.C., McKerrow, J.H., Craik,
C.S. & Fletterick, R.J. (1995) T he crystal structure of cruzain: a
therapeutic target for Chagas’ disease. J. Mol. Biol. 247, 251–259.
25. Ferrin, T.E., Huang, C.C., Jarvis, L.E. & Langridge, R. (1988)
The midas display system. J. Mol. Graphics 6, 13–27.
26. Huang, C.C., Pettersen, E.F., Klein, T.E., Ferrin, T.E. & Lan-
gridge, R . (1991) Conic – a fast renderer f or space-filling molecules
with shadows. J. Mol. Graphics 9, 230–236.
27. LeBowitz, J .H., Coburn, C.M., McMahon-Pratt, D. & Beverley,
S.M. (1990) Development of a stable Leishmania expression vector
and application to the study of parasite surface-antigen genes.
Proc.NatlAcad.Sci.USA87, 9 736–9740.
28. Robertson, C.D. & Coombs, G.H. (1990) Characterization o f 3
groups of cysteine proteinases in the amastigotes of Leishmania

mexicana mexicana. Mol. Bioc hem. Parasitol. 42 , 269–276.
29. Hirata, I.Y., Cezari, M .H.S., Nakaie, C., Boschcov, P., Ito, A.S.,
Juliano, M.A. & Juliano, L. (1994) Internally quenched
fluorogenic protease sub strates: solid -phase synthesis and fluor-
escence spectroscop y of p eptides c ontainin g orth o-aminobenzoyl/
dinitrophenyl groups as d on or-acceptor pairs. Lett. Peptide S ci. 1,
299–308.
30. Wilkinson, G.N. (1961) Statistical estimations in enzyme kinetics.
Biochem. J. 80, 324–332.
31. Nagler, D.K., Ta m, W., Storer, A.C.,Krupa,J.C.,Mort,J.S.&
Menard, R . (1999) Interdependency of sequence and positional
specificities fo r cysteine pro teases of the papain family. Biochem-
istry 38, 4868–4874.
32. Portaro, F.C.V., Santos, A.B.F., Cezari, M.H.S., Juliano, M.A.,
Juliano, L. & Carmona, E. (2000) Probing the specificity of
cysteine proteinases at subsites remote from the active site: analysis
of P
4
,P
3
,P
2
¢ and P
3
¢ variations in e xten ded substrates. Biochem. J.
347, 123–129.
33. Serveau, C., Lalmanach, G., Hirata, I., Scharfstein, J., Juliano,
M.A. & Gauthier, F. (1999) Discrimination of cruzipain, the
major cysteine proteinase of Trypanosoma cruzi, and mammalian
cathepsins B and L, by a pH-ind ucible fluorogenic sub strate of

trypanosomal cysteine pr oteinases. Eur. J. Biochem. 259, 275–280.
34. Serveau, C., Lalmanach, G., Juliano, M .A., Sc harfstein, J ., J uli-
ano, L. & G authier, F. (1996) Investigation of the substra te spe-
cificity of cruzipain, the major cysteine proteinase of Trypanosoma
cruzi, through the use of cystatin-derived substrates and inhibitors.
Biochem. J. 313, 951–956.
35. Eakin, A.E., Mills, A.A., Harth, G., McKerrow, J.H. & Craik,
C.S. (1992) The sequence, organization, and expression of the
major cysteine protease (cruzain) from Tr ypanosoma c ruzi. J. Biol.
Chem. 267, 7411 –7420.
36. Stoka, V., M cKerrow, J.H., Cazzulo,J.J.&Turk,V.(1998)
Substrate inhibition o f cruzipain is no t affected by th e C-terminal
domain. FEBS Lett. 429 , 129–133.
3714 M. A. Juliano et al.(Eur. J. Biochem. 271) Ó FEBS 2004