Journal of Science 113, 4035-4041 (2000) 4035 Printed in Great Britain © The Company of Biologists Limited 2000 JCS1618

Processing and trafficking of in Leishmania mexicana

Darren R. Brooks1, Laurence Tetley2, Graham H. Coombs2 and Jeremy C. Mottram1,* 1Wellcome Centre for Molecular Parasitology, University of Glasgow, The Anderson College, 56 Dumbarton Rd, Glasgow G11 6NU, UK 2Division of & Immunity, Institute of Biomedical and Life Sciences, University of Glasgow, Joseph Black Building, Glasgow G12 8QQ, UK *Author for correspondence: (e-mail: [email protected])

Accepted 29 August; published on WWW 31 October 2000

SUMMARY

Removal of the pro-domain of a cysteine is via the flagellar pocket. Moreover, they show that another essential for activation of the . We have engineered protease can process CPB in the absence of either CPA or a cysteine protease (CPB2.8) of the protozoan parasite CPB, albeit less efficiently. Abolition of the glycosylation Leishmania mexicana by site-directed mutagenesis to site in the mature domain of CPB did not affect enzyme remove the cysteine (to produce CPBC25G). When processing, targeting or in vitro activity towards gelatin. CPBC25G was expressed in a L. mexicana mutant lacking This indicates that glycosylation is not required for all CPB genes, the inactive pro-enzyme was processed to the trafficking. Together these findings provide evidence that mature and trafficked to the . These the major route of trafficking of Leishmania cysteine results show that auto-activation is not required for correct proteases to is via the flagellar pocket and processing of CPB in vivo. When CPBC25G was expressed therefore differs significantly from cysteine protease in a L. mexicana mutant lacking both CPA and CPB genes, trafficking in mammalian cells. the majority of the pro-enzyme remained unprocessed and accumulated in the flagellar pocket. These data reveal that Key words: Cysteine protease, Protease, L, CPA can directly or indirectly process CPBC25G and Trypanosomatid, Leishmania, Transfection, Enzyme processing, suggest that cysteine proteases are targeted to lysosomes Trafficking, Flagellar pocket

INTRODUCTION Eakin et al., 1993) and leishmanial (Sanderson et al., 2000) cysteine proteases have also suggested that maturation of these Cysteine proteases of the superfamily, designated Clan can occur auto-catalytically. However, definitive A, family C1 (Barrett and Rawlings, 1996), are synthesised as studies of in vivo processing were lacking and the participation that are activated by cleavage of the pro-domain to of other enzymes remained a possibility. It seems likely that generate mature enzymes located predominantly within the mechanisms of processing and trafficking operating in lysosomes. Trafficking of the mammalian proteases is lower eukaryotes such as trypanosomatids may differ mediated primarily through binding of mannose-6-phosphate significantly from those in mammals, and the lack of mannose- residues present on the surface of the enzymes to appropriate 6-phosphate on lysosomal cysteine proteases of Trypanosoma receptors (Kornfeld and Mellman, 1989). A mannose-6- cruzi (Cazzulo et al., 1990) supports this postulate. The current phosphate-independent mechanism based upon recognition of study of in vivo processing and trafficking of leishmanial membrane-associated receptors has also been described for cysteine proteases was undertaken to test this hypothesis. cathepsin D and (McIntyre and Erickson, 1993; Leishmania mexicana contain many cathepsin L-like Zhu and Conner, 1994). Cleavage of the pro-domain of the cysteine proteases that are classified as members of the papain mammalian proteases appears to occur either in late endosomal superfamily (Berti and Storer, 1995). The most abundant compartments or in the lysosomes themselves (Mach et al., enzymes occur within the large lysosomes, known as 1994). The pro-domains of many cysteine proteases contain a megasomes, which characteristically are present in the short conserved region centred on a heptapeptide. Mutational amastigote (mammalian) stage of the life-cycle (Pupkis et al., analysis of this motif of the plant cysteine protease papain has 1986; Duboise et al., 1994). They are encoded by the CPB indicated that cleavage of the pro-domain is triggered by genes, which are arranged as a tandem repeat of 19 copies changes in the electrostatic charge of the motif at acidic pH (Mottram et al., 1996; Mottram et al., 1997). The genes of the (Vernet et al., 1995) and that completion of the activation array that have been analysed to date are highly conserved process is entirely auto-catalytic (Vernet et al., 1991). In vitro (about 99% identical) in the pre-, pro- and mature domains. studies with prokaryotic (Burns et al., 1997), mammalian However, only gene copies CPB3-CPB18, which are expressed (Mason et al., 1987), trypanosomal (Hellman et al., 1991; predominantly in amastigotes, encode isoenzymes with 4036 D. R. Brooks and others unusual 100 C-terminal extensions (CTE) that are purified oligonucleotides (only the sense strand primer is shown and characteristic of the most abundant group of trypanosomatid the mutated sites are given in bold): cysteine proteases. The finding that the first two genes of the OL 422 to generate pGL38 (encoding CPBC25G), 5′-GGTGCG- CPB array encode isoenzymes with a truncated CTE and yet TGCGGGTCGGCTGGGCGTTCTCGG-3′; and OL 424 to generate pGL39 (encoding CPBN103S), 5′-GCCCGAGTGCTCGAGCAGCA- become located in lysosomes refutes the possibility that the ′ CTE plays a part in intracellular trafficking of the enzyme GTGAGCTCG-3 . These mutant constructs were verified by sequence analysis using (Mottram et al., 1997). Moreover, a second leishmanial an ABI 373 automated DNA sequencer (Perkin-Elmer). The native cathepsin L-like cysteine protease, CPA, lacks a CTE (Mottram and mutant CPB2.8 genes were excised from pBluescript SK− as 2.0 et al., 1992) but also occurs in lysosomes (Duboise et al., kb EcoRV fragments and ligated to the SmaI site of the pX episomal 1994). A recent study on the effects of cysteine protease shuttle vector (LeBowitz et al., 1990) to generate pGL46 (pXCPB2.8), inhibitors on Trypanosoma cruzi showed that there was an pGL47 (pXCPBC25G) and pGL48 (pXCPBN103S). The Trypanosoma accumulation of unprocessed pro-enzyme of the major cysteine brucei cysteine protease cDNA (Mottram et al., 1989) was cloned into protease in late-Golgi vesicles, which suggests that cleavage of the HindIII/EcoRI site of the pTEX shuttle vector (Kelly et al., 1992). the pro-domain is necessary for successful trafficking to Plasmids were transfected into the L. mexicana ∆CPB and ∆CPA/CPB lysosomes (Engel et al., 1998). Moreover, it has been shown mutants as described previously (Souza et al., 1994; Mottram et al., (Huete-Pérez et al., 1999) that the pro-domain is both sufficient 1996). Briefly, plasmids were prepared using Qiagen Tip100 columns as outlined by the manufacturer. Transfection used 10 µg of DNA and and necessary for targeting of both Leishmania and T. cruzi 4×107 late-log phase ∆CPB or ∆CPA/CPB promastigotes. Following cysteine proteases to lysosomes. Nevertheless, the mechanisms electroporation, cells were allowed to recover in 10 ml HOMEM of processing and trafficking of cysteine proteases in these medium with serum for 24 h at 25°C and then transformants were parasites remains uncertain (Engel et al., 2000). selected by transfer of cells into HOMEM medium with serum We have generated a series of cysteine protease-deficient containing 25 µg/ml G418. mutants of L. mexicana. ∆CPA lacks the single copy CPA gene (Souza et al., 1994), whereas ∆CPB lacks the entire CPB Analysis of cysteine proteases gene array (Mottram et al., 1996). Double mutants lacking Parasite cysteine protease activity was determined using substrate- ∆ SDS-PAGE as previously described (Robertson and Coombs, 1990; both CPA and CPB ( CPA/CPB) have also been generated 7 (Mottram et al., 1996). Re-expression of different copies of Mottram et al., 1996). Briefly, parasite cell lysates (10 cells) were ∆ subjected to electrophoresis under reducing conditions using 12% CPB in CPB demonstrated that the isogenes encode (w/v) acrylamide gels incorporating 0.2% (w/v) gelatin. Following enzymes with different substrate specificities (Mottram et al., electrophoresis, the gel was washed for 1 hour with 2.5% (v/v) Triton 1997). We have now exploited this ability to re-express CPB X-100 and then incubated in 0.1 M sodium acetate, pH 5.5, containing isoenzymes in the cysteine protease-deficient mutants to 1 mM DTT. of gelatin was detected by staining with investigate the mechanisms whereby the enzyme is processed 0.25% (w/v) Coomassie Blue R-250. Western blotting used polyclonal to the mature, active form and at the same time trafficked to anti-CPB antiserum (1 in 2500) raised against recombinant CPB2.8 lysosomes. To this end, we have engineered an inactive CPB, expressed in and purified from Escherichia coli as a hexahistidine by mutation of a cysteine residue in the active site that is fusion protein (Sanderson et al., 2000). essential for activity, and a CPB that lacks the only N-linked Immunoelectron microscopy glycosylation site on the mature region of the enzyme. The Axenic amastigotes were processed and immunostained as previously data generated from the re-expression of these mutated genes described (Mottram et al., 1997). Briefly, pelleted cells were fixed in provide new insights into the processing and trafficking of 2% (w/v) p-formaldehyde/0.1% glutaraldehyde for 30 min at 4°C, these leishmanial enzymes. processed cold to −20°C in increasing alcohol concentrations, and finally embedded in LR White Medium Resin (BioCell International, UK) at 20°C, polymerising under 365 nm light for 2 days. Sections MATERIALS AND METHODS were collected on nickel grids and immunostained with anti-CPB antiserum at 1 in 200, followed by goat anti-rabbit IgG/10 nm gold Parasites conjugate (Aurion, The Netherlands) at 1 in 20. Labelling was L. mexicana (MNYC/BZ/62/M379) promastigotes were grown in observed and recorded with a Zeiss 902 EFTEM at 80 kV, using zero- HOMEM medium supplemented with 10% (v/v) heat-inactivated fetal loss imaging to enhance contrast. calf serum, pH 7.5, at 25°C as described previously (Mottram et al., 1992). Amastigote-like forms (axenic amastigotes) were grown in vitro in Schneider’s Drosophila medium (Life Technologies) with 20% (v/v) heat-inactivated fetal calf serum, pH 5.5, at 32°C and in RESULTS the presence of 5% (v/v) CO2 (Bates et al., 1992). The required antibiotics were added to the cultures of the cysteine protease null In vivo processing of cysteine proteases mutants of L. mexicana as follows: hygromycin B (Sigma) at 50 The L. mexicana CPBs are similar to other cathepsin L-like µg/ml, phleomycin (Cayla, France) at 10 µg/ml, nourseothricin cysteine proteases in being produced as inactive precursors that (Hans-Knöll Institute, Germany) at 25 µg/ml, puromycin are processed to produce the mature, active enzyme (Sanderson (Calbiochem) at 10 µg/ml, and neomycin (G418, Geneticin, Life µ et al., 2000). Processing of the papain family enzymes is Technologies) at 25 g/ml. thought to occur via autocatalytic and predominantly intra- Mutagenesis, constructs and transfections molecular cleavages (Vernet et al., 1991). We addressed the Mutations of the CPB active site and glycosylation site were mode of processing of CPB by mutating the active site cysteine incorporated into pGL28, a pBluescript SK− plasmid containing the (C25G) of one CPB isoenzyme with a well-defined substrate CPB2.8 gene (Mottram et al., 1996), using the QuikChange site- specificity (CPB2.8; Mottram et al., 1997) and then expressing directed mutagenesis kit (Stratagene) and the following reverse phase- this inactive protease, from the pX episome, in the ∆CPB null Cysteine proteases in L. mexicana 4037

Fig. 1. Expression of cysteine proteases in ∆CPB. Cysteine protease activity was analysed by gelatin-SDS-PAGE (A,C) and protease expression confirmed by western blotting (B). Extracts from 107 stationary phase promastigotes were used for gelatin- SDS-PAGE analysis and 5×106 stationary phase promastigotes were used for western blotting with anti-CPB antiserum (1 in 2500). (A,B) Wild-type L. mexicana (lane 1), ∆CPB (lane 2), ∆CPB[pXCPB2.8] (lane 3) and ∆CPB[pXCPBC25G] (lane 4). Note that the lower mobility activities (>42 kDa) are not encoded by CPB genes. (C) ∆CPB (lane 1) and ∆CPB[pTEXtb] (lane 2). mutant (to give cell line ∆CPB [pXCPBC25G]). The lack of cysteine protease activity in the mutated protein CPBC25G was confirmed by substrate-SDS PAGE analysis (Fig. 1A). Stationary phase wild-type L. mexicana had two dominant cysteine protease activities towards gelatin (with apparent molecular masses of approximately 23 and 25 kDa) (lane 1) that were absent in ∆CPB (lane 2) and ∆CPB[pXCPBC25G] (lane 4). Analysis of ∆CPB[pXCPB2.8] (lane 3) demonstrated that when native CPB2.8 was re-expressed in the null mutant a dominant CP activity of ~23 kDa is produced. In addition, there were slower mobility activities (approx. 33-38 kDa). These correspond to intermediate processing forms of the that have been activated in situ (Sanderson et al., 2000). Western blotting (Fig. 1B) confirmed the presence of CPBC25G in ∆CPB (lane 4). Moreover, CPBC25G was processed to the same size, 28 kDa, as the mature active enzyme (Fig. 1B lanes 3 and 4), indicating that its maturation in vivo was due Fig. 2. CPA is able to process CPBC25G in vivo. (A) Western blot C25G to the action of one or more protease other than CPB. The analysis of CPB2.8 (lane 1) and CPB (lane 2) re-expressed in ∆ × 6 respective precursor forms of CPB2.8 and CPBC25G were CPA/CPB. Extracts from 5 10 stationary phase promastigotes were used for western blotting with anti-CPB antiserum (1 in 2500). detected at lower abundance than the mature , and the (B) Gelatin-SDS-PAGE analysis confirming that CPB is active in profile of intermediates appeared to differ slightly (Fig. 1B, ∆CPA. Extracts from 107 stationary phase promastigotes were used lanes 3 and 4, 38-48 kDa bands). The African trypanosome, T. for substrate gel analysis: wild-type L. mexicana (lane 1), ∆CPA brucei, also contains a multicopy array of cysteine protease (lane 2) and ∆CPB (lane 3). genes that encode enzymes with a CTE (Mottram et al., 1989). When expressed in ∆CPB (cell line ∆CPB[pTEXtb]) the active trypanosome cysteine protease was detected at approx. 32 kDa ∆CPA (Souza et al., 1994). This indicates that CPB zymogen (Fig. 1C, lane 2), which is the size found in procyclic processing is not dependent upon an intermolecular activation trypanosome cell extracts (not shown). This shows that this by CPA and is also suggestive of a degree of redundancy cysteine protease from a closely related trypanosomatid can be between CPA and CPB. The protease responsible for correctly processed and activated in Leishmania. processing the approx. 10% of CPBC25G in ∆CPA/CPB remains To investigate the possibility that the cysteine protease CPA to be characterised. was responsible for the processing of CPBC25G in ∆CPB, the ∆CPA/CPB double null mutant (Mottram et al., 1996) was Trafficking of cysteine proteases in L. mexicana transfected with pGL47, and cells expressing CPBC25G were In wild-type L. mexicana amastigotes, mature CPBs are analysed by western blotting (Fig. 2A, lane 2). A single localised to large lysosomes (Mottram et al., 1997). In dominant precursor form of CPBC25G of approx. 45 kDa was agreement with the detection of fully processed and active detected in addition to fully processed enzyme of 28 kDa, the protease, immunoelectron microscopy of ∆CPB and relative abundance being approximately 9:1. This indicates that ∆CPA/CPB re-expressing CPB2.8 showed that this enzyme CPA is capable of participating in the in vivo processing of was efficiently targeted to the lysosomes (Fig. 3a and 3b). CPBs in L. mexicana. Interestingly, however, when native There was little labelling of other cellular compartments (for CPB2.8 was expressed in ∆CPA/CPB, western blot analysis example, the Golgi complex), as might have occurred if there showed that the protease was processed to the size of the had been a low efficiency of targeting. The inactivated mature enzyme with no unprocessed protein being detected CPBC25G was also localised to the lysosomes of ∆CPB (Fig. (Fig. 2A, lane 1). Moreover, gelatin SDS-PAGE analysis (Fig. 3c), again with little labelling of other cellular compartments. 2B) showed that CPB2.8 was fully active when expressed in In contrast, expression of CPBC25G in ∆CPA/CPB resulted in 4038 D. R. Brooks and others the protease accumulating predominantly in the flagellar the major cysteine protease of T. cruzi (Eakin et al., 1993) and pocket and vesicles subtending this compartment (Fig. 3d and also recombinant L. mexicana CPB2.8 (Sanderson et al., 2000). 3e). A low level of labelling was apparent in the lysosomes, However, these findings do not show if autoactivation occurs, which correlates with correct targeting of the minor proportion or is the sole means of precursor processing, in the living cell. of fully processed CPBC25G (Fig. 2A, lane 2). Indeed, the initial cleavage of the recombinant T. cruzi cysteine Some of the native L. mexicana CPBs are known to be protease is thought to be mediated by an uncharacterised E. glycosylated (Robertson and Coombs, 1990) and a potential N- coli enzyme (Eakin et al., 1993). linked glycosylation site is present on a surface loop in the Indirect support for the involvement of another enzyme in mature domain of all the CPBs characterised to date. The the natural processing of cysteine proteases has been obtained CPB2.8 glycosylation site (N103) was mutated to a serine previously with plants. It was found that the N-terminal (N103S) in order to address any potential role that this sequence of in vitro processed recombinant papain was motif may have in trafficking of the enzymes. Following heterogeneous, whereas the N-terminus of the naturally transfection of ∆CPB with pGL48, immunolocalisation of CPBN103S clearly showed this enzyme to be correctly targeted to the lysosomes (Fig. 3f). Again, no labelling was detected in other cellular compartments. Consistent with a correctly targeted enzyme was the observation that removal of this glycosylation site had no effect on the processing of the precursor to the mature protein (Fig. 4B) or on the activity of the enzyme (Fig. 4A).

DISCUSSION

One aim of this study was to determine the mechanism whereby the cathepsin L-like cysteine proteases of L. mexicana that are encoded by the CPB gene array are processed in vivo to the active, mature form. Studies on similar enzymes of a number of other eukaryotes have demonstrated that in vitro the proteases are able to auto- activate under acidic conditions (Vernet et al., 1991; Eakin et al., 1993; Mason et al., 1987; Burns et al., 1997). The precise mechanism of this autocatalytic process is still unclear, but the finding that the pro-region binds to the active site cleft in the opposite orientation to that of a standard substrate suggested that cleavage might not be catalysed by the enzyme’s usual mechanism (Coulombe et al., 1996). Work with cathepsin L has suggested that in vitro autocatalysis is exclusively an intramolecular event (Nomura and Fujisawa, 1997). In contrast, Menard and co-workers concluded that the autocatalytic processing of recombinant human pro-cathepsin L Fig. 3. Immunolocalisation of CPB in L. mexicana. Ultrathin sections of L. mexicana axenic was effected via both intermolecular amastigotes were labelled with anti-CPB antiserum and goat anti-rabbit 10 nm gold conjugate: and intramolecular cleavages (Menard ∆CPB[pXCPB2.8] (a), ∆CPA/CPB[pXCPB2.8] (b), ∆CPB[pXCPBC25G] (c), et al., 1998). Some trypanosomatid ∆CPA/CPB[pXCPBC25G] (d,e) and ∆CPB[pXCPBN103S] (f). Note the absence of flagellar pocket cysteine proteases can also autoactivate labelling in a,b,c and f but heavy labelling in d, enlarged in e. fp, flagellar pocket; m, megasome in vitro – this was observed with both (large lysosome). Scale bars: 0.5 µm (a-d,f); 0.25 µm (e). Cysteine proteases in L. mexicana 4039

Fig. 4. CPBN103S is processed normally and active. Gelatin SDS- PAGE analysis (A) and western blotting (B) of CPB2.8 and CPBN103S expressed in ∆CPB. Extracts from 107 stationary phase promastigotes were used for gelatin SDS-PAGE analysis and 5×106 stationary phase promastigotes were used for western blotting with anti-CPB antiserum (1 in 2500): ∆CPB[pXCPB2.8] (lane 1) and ∆CPB[pXCPBN103S] (lane 2).

protease of T. cruzi, which has been shown to lack mannose- 6-phosphate (Cazzulo et al., 1990), and also more recently for L. amazonensis (Boukai et al., 2000). There is another potential glycosylation site some 30 amino acids from the start of the CTE of the CPB2.8 isoenzyme (N246). However, this motif is absent from some other CPBs that lack a long CTE and yet are still targeted to the lysosomes (Mottram et al., 1997), which suggests that the motif plays no role in trafficking. This conclusion, that neither glycosylation nor the CTE mediate occurring enzyme is homogeneous (Vernet et al., 1990). This CPB trafficking, is supported by the results of Huete-Pérez and implies that another protease may predominate during the colleagues (Huete-Pérez et al., 1999) who showed that the pro- maturation process in vivo. Heterologous proteases are also thought to be responsible for the activation of several other cysteine proteases. For instance, a second cysteine protease is required for in vitro activation of the barley cysteine protease aleurain (Holwerda et al., 1990). Furthermore, pro- protease IV, a highly abundant cysteine protease from Carica papaya, is also unable to process autocatalytically as a consequence of steric crowding of the active site cleft (Baker et al., 1996). Most recently, in vitro activation of human has been shown to require the addition of cathepsin L (Nagler et al., 1999). We have now shown that other proteases are able to mediate maturation of leishmanial CPBs in vivo and that autoactivation is not the only way in which these cysteine protease precursors can be processed. Our results from using L. mexicana ∆CPA/CPB clearly demonstrate that CPB can be activated in vivo by another cathepsin L-like cysteine protease, CPA. It is likely that CPA can cleave the pro-domain of CPB directly as the two cysteine proteases are closely related (Mottram et al., 1996), although the possibility that CPA could act indirectly through processing of an intermediate protease cannot be excluded at this stage. However, CPA is not essential for in vivo processing of CPB as CPB was still efficiently processed in the absence of CPA (Fig. 2B, lane 2). Moreover, the mature forms of both the mutated, inactive CPB and the native protease following re- expression in ∆CPB were shown by western blotting to be the same size. This suggests that the same final cleavage site is used by both CPA and CPB. However, the western blots revealed several higher molecular mass proteins, which were not identical in the different lines Fig. 5. Model of cysteine protease processing and trafficking, superimposed (Fig. 1B, lanes 3 and 4). This indicates that multiple on a transmission electron micrograph of a Leishmania amastigote. The cleavages occur in vivo, as has been shown for proposed major route of trafficking involves zymogen (blue dots) in recombinant CPBs in vitro (Sanderson et al., 2000), and vesicles produced from the Golgi apparatus (pink) moving to and fusing that some cleavage sites used by CPB and CPA differ. with the flagellar pocket. Following binding to a putative flagellar pocket The mechanism mediating trafficking appears not to be receptor (represented by black lines extending from flagellar pocket membrane into the pocket itself), the cysteine protease is activated to the glycosylation. Our finding that a mutant CPB with its only mature form (red dots) and then trafficked to the lysosome (yellow N-linked glycosylation site on the mature domain vesicles). A second hypothetical pathway (indicated with a question mark) abolished was localised in an identical fashion to the may involve direct trafficking of vesicles (blue) to the lysosome with native enzyme provides good evidence on this point. The activation of zymogen occurring en route. f, flagellum; i, inclusion body same conclusion was reached for the major cysteine (lipid); kp, kinetoplast; m, mitochondrion; n, nucleus. 4040 D. R. Brooks and others domain, which is not glycosylated, is both sufficient and et al., 1999) reported that cysteine protease inhibitors that necessary for targeting of both Leishmania and T. cruzi prevent activation of CPB in L. major caused parasite death in cysteine proteases to lysosomes. vitro and in vivo. One possible explanation for this difference The data showing the accumulation of the unprocessed pro- is that the L. major CPB homologues were not the sole target CPB in or near the flagellar pocket (Fig. 3d,e) suggest that a of the CP inhibitors used, and hence cell death was not due to route for the trafficking of the proteases in Leishmania to the the observed disturbances in trafficking but rather a lysosomes is via this compartment. It is not at present possible consequence of the inhibition of some unknown cysteine to entirely dismiss the possibility that the unprocessed cysteine protease(s). Furthermore, Selzer and co-workers postulated proteases accumulated in the flagellar pocket owing to a default that the inhibitor blocks parasite replication through inhibition mechanism whereby wrongly processed proteins are directed of both -like and cathepsin L-like cysteine to the compartment for subsequent excretion from the cell. proteases. The L. mexicana ∆CPA/CPB [pXCPBC25G] mutant However, this seems unlikely, as the pocket of amastigotes is still possesses functional cathepsin B-like protease activity and effectively a sealed environment (Overath et al., 1997) and this may be crucial for the survival and replication of the line. hence release from this location would not be possible. Additional factors that could account for the differences Moreover, no secretion of CPBC25G was detected from observed include variation in the levels of cysteine protease ∆CPA/CPB promastigotes (data not shown) even though the expression, pro-enzyme accumulation between the Leishmania flagellar pocket acts as the major site of endocytosis/exocytosis lines and possibly other species-specific characteristics. in this stage of the life-cycle (Overath et al., 1997). Indeed, The observed differences between the accumulation of rapid turnover of unprocessed CPB would be more likely to precursor cysteine proteases in Leishmania (this study and occur via direct targeting of CPBC25G to the lysosomes of Selzer et al., 1999) and T. cruzi (Engel et al., 1998; Engel et al., ∆CPA/CPB. 2000) point to species variations either in the site of cysteine As the stage-specific expression of CPA and CPB is protease processing or the trafficking route, or perhaps both. The generally similar, and it seems likely that both types are discovery that a lysosomal membrane glycoprotein, p67, of trafficked to lysosomes via the same route (Mottram et al., bloodstream forms of T. brucei is targeted to this organelle via 1997; Duboise et al., 1994), the processing of CPB could occur the flagellar pocket (Brickman and Balber, 1994) indicates that at any stage during trafficking. Addition of cysteine protease such a trafficking route does exist in at least this species of inhibitors to T. cruzi resulted in unprocessed zymogen trypanosome. The involvement of this cell compartment in accumulating within the late Golgi/early endosome network, cellular trafficking in these trypanosomatids suggests that the which suggested that the processing of the cysteine proteases molecules underpinning this unusual route for trafficking to of this parasite must normally occur at or before that stage lysosomes may represent attractive chemotherapeutic targets. (Engel et al., 1998). However, our results with the mutated and so inactive CPB in ∆CPA/CPB demonstrate a build up of This work was supported by the Medical Research Council (UK). unprocessed precursor cysteine protease in the flagellar pocket J.C.M. is an MRC Senior Research Fellow. and subtending vesicles of L. mexicana amastigotes. Treatment of L. major with cysteine protease inhibitors also leads to a REFERENCES build up of inactive cathepsin B-like cysteine protease in the flagellar pocket and immediate surrounding vesicles (Selzer et Baker, K. C., Taylor, M. J., Cummings, N. J., Tunon, M. A., Worboys, K. al., 1999). The finding that processed CPBC25G is targeted to A. and Connerton, I. F. (1996). Autocatalytic processing of pro-papaya the lysosomes (Fig. 3c) and yet unprocessed CPBC25G proteinase IV is prevented by crowding of the active site cleft. Protein Eng. accumulates in the flagellar pocket (Fig. 3d,e) provides further 9, 525-529. Barrett, A. J. and Rawlings, N. D. (1996). Families and clans of cysteine evidence that the pocket is also likely to be the site of cleavage peptidases. Perspect. Drug Dis. Design 6, 1-11. of the pro-domain to produce mature CPB enzyme. Bates, P. A., Robertson, C. D., Tetley, L. and Coombs, G. H. (1992). Axenic Consequently zymogen activation is likely to be a pre-requisite cultivation and characterization of Leishmania mexicana amastigote-like for the delivery of CPBs from the flagellar pocket to the forms. Parasitology 105, 193-202. lysosomes. The processed CPB is likely to complete its passage Boukai, L. K., Costa-Pinto, D., Soares, M. J., McMahon-Pratt, D. and Traub-Cseko, Y. M. (2000). Trafficking of cysteine proteinase to to lysosomes from the flagellar pocket via interaction with a Leishmania lysosomes: lack of involvement of glycosylation. Mol. Biochem. membrane-bound receptor, the result being internalisation of Parasitol. 107, 321-325. the enzyme into a vesicle destined for the lysosomes (see Fig. Berti, P. J. and Storer, A. C. (1995). Alignment/phylogeny of the papain 5 for a model of trafficking and processing). Such a receptor superfamily of cysteine proteases. J. Mol. Biol. 246, 273-283. Brickman, M. J. and Balber, A. E. (1994). Transport of a lysosomal may have evolutionary links to the mammalian receptors membrane glycoprotein from the Golgi to endosomes and lysosomes via the involved in mannose-6-phosphate independent lysosomal cell surface in African trypanosomes. J. Cell Sci. 107, 3191-3200. targeting (McIntyre and Erickson, 1993; Zhu and Conner, Burns, E. H., Marciel, A. M. and Musser, J. M. (1997). Structure-function 1994). Nevertheless, both this study on L. mexicana and that and pathogenesis studies of Streptococcus pyogenes extracellular cysteine on T. cruzi (Engel et al., 1998; Engel et al., 2000) clearly protease. Adv. Exp. Med. Biol. 418, 589-592. Cazzulo, J. J., Hellman, U., Couso, R. and Parodi, A. J. A. (1990). Amino demonstrate that cleavage of the pro-region is essential for acid and carbohydrate composition of a lysosomal cysteine proteinase from successful targeting. As studies on T. cruzi and L. major using Trypanosoma cruzi – absence of phosphorylated mannose residues. Mol. cysteine protease-specific inhibitors showed so convincingly Biochem. Parasitol. 38, 41-48. (Engel et al., 1998; Selzer et al., 1999), disruption of this event Coulombe, R., Grochulski, P., Sivaraman, J., Menard, R., Mort, J. S. and Cygler, M. (1996). Structure of human procathepsin L reveals the molecular has severe consequences on the parasite. It is interesting that basis of inhibition by the prosegment. EMBO J. 15, 5492-5503. the mutants expressing the CPB that was not processed survive Duboise, S. M., Vannier-Santos, M. A., Costa-Pinto, D., Rivas, L., Pan, A. and replicate in vitro, whereas Selzer and co-workers (Selzer A., Traub-Cseko, Y. M., de Souza, W. and McMahon-Pratt, D. (1994). Cysteine proteases in L. mexicana 4041

The biosynthesis, processing, and immunolocalization of Leishmania in their stage-regulation and substrate preferences. J. Biol. Chem. 272, pifanoi amastigote cysteine proteinases. Mol. Biochem. Parasitol. 68, 119- 14285-14293. 132. Mottram, J. C., North, M. J., Barry, J. D. and Coombs, G. H. (1989). A Eakin, A. E., McGrath, M. E., McKerrow, J. H., Fletterick, R. J. and cysteine proteinase cDNA from Trypanosoma brucei predicts an enzyme Craik, C. S. (1993). Production of crystallizable cruzain, the major cysteine with an unusual C-terminal extension. FEBS Lett. 258, 211-215. protease from Trypanosoma cruzi. J. Biol. Chem. 268, 6115-6118. Mottram, J. C., Robertson, C. D., Coombs, G. H. and Barry, J. D. (1992). Engel, J. C., Doyle, P. S., Palmer, J., Hsieh, I., Bainton, D. F. and A developmentally regulated cysteine proteinase gene of Leishmania McKerrow, J. H. (1998). Cysteine protease inhibitors alter Golgi complex mexicana. Mol. Microbiol. 6, 1925-1932. ultrastructure and function in Trypanosoma cruzi. J. Cell Sci. 111, 597-606. Mottram, J. C., Souza, A. E., Hutchison, J. E., Carter, R., Frame, M. J. Engel, J. C., Torres Garcia, C., Hsieh, I., Doyle, P. S. and McKerrow, J. and Coombs, G. H. (1996). Evidence from disruption of the lmcpb gene H. (2000). Upregulation of the secretory pathway in cysteine proteinase array of Leishmania mexicana that cysteine proteinases are virulence inhibitor-resistant Trypanosoma cruzi. J. Cell Sci. 113, 1345-1354. factors. Proc. Natl. Acad. Sci. USA 93, 6008-6013. Hellman, U., Wernstedt, C. and Cazzulo, J. J. (1991). Self- of Nagler, D. K., Zhang, R. L., Tam, W., Sulea, T., Purisima, E. O. and the cysteine proteinase, , from Trypanosoma cruzi gives a major Menard, R. (1999). Human cathepsin X: A cysteine protease with unique fragment corresponding to its carboxy-terminal domain. Mol. Biochem. activity. Biochemistry 38, 12648-12654. Parasitol. 44, 15-21. Nomura, T. and Fujisawa, Y. (1997). Processing properties of recombinant Holwerda, B. C., Galvin, N. J., Baranski, T. J. and Rogers, J. C. (1990). human procathepsin L. Biochem. Biophys. Res. Commun. 230, 143-146. In vitro processing of aleurain, a barley vacuolar protease. Plant Cell Overath, P., Stierhof, Y. D. and Wiese, M. (1997). Endocytosis and secretion 2, 1091-1106. in trypanosomatid parasites: tumultuous traffic in a pocket. Trends Cell Biol. Huete-Pérez, J. A., Engel, J. C., Brinen, L. S., Mottram, J. C. and 7, 27-33. McKerrow, J. H. (1999). Protease trafficking in two primitive eukaryotes Pupkis, M. F., Tetley, L. and Coombs, G. H. (1986). Leishmania mexicana: is mediated by a prodomain protein motif. J. Biol. Chem. 274, 16249-16256. amastigote in unusual lysosomes. Exp. Parasitol. 62, 29-39. Kelly, J. M., Ward, H. M., Miles, M. A. and Kendall, G. (1992). A shuttle Robertson, C. D. and Coombs, G. H. (1990). Characterization of three vector which facilitates the expression of transfected genes in Trypanosoma groups of cysteine proteinases in the amastigotes of Leishmania mexicana cruzi and Leishmania. Nucleic Acids Res. 20, 3963-3969. mexicana. Mol. Biochem. Parasitol. 42, 269-276. Kornfeld, S. and Mellman, I. (1989). The biogenesis of lysosomes. Ann. Rev. Sanderson, S. J., Pollock, K. G., Hilley, J. D., Meldal, M., Hilaire, P. S., Cell Biol. 5, 483-525. Juliano, M. A., Juliano, L., Mottram, J. C. and Coombs, G. H. (2000). LeBowitz, J. H., Coburn, C. M., McMahon-Pratt, D. and Beverley, S. M. Expression and characterisation of a recombinant cysteine proteinase of (1990). Development of a stable Leishmania expression vector and Leishmania mexicana. Biochem. J. 347, 383-388. application to the study of parasite surface-antigen genes. Proc. Natl. Acad. Selzer, P. M., Pingel, S., Hsieh, I., Ugele, B., Chan, V. J., Engel, J. C., Bogyo, Sci. USA 87, 9736-9740. M., Russell, D. G., Sakanari, J. A. and McKerrow, J. H. (1999). Cysteine Mach, L., Mort, J. S. and Glössl, J. (1994). Maturation of human protease inhibitors as chemotherapy: lessons from a parasite target. Proc. procathepsin B. Proenzyme activation and proteolytic processing of the Natl. Acad. Sci. USA 96, 11015-11022. precursor to the mature proteinase, in vitro, are primarily unimolecular Souza, A. E., Bates, P. A., Coombs, G. H. and Mottram, J. C. (1994). Null processes. J. Biol. Chem. 269, 13030-13035. mutants for the lmcpa cysteine proteinase gene in Leishmania mexicana. Mason, R. W., Gal, S. and Gottesman, M. M. (1987). The identification of Mol. Biochem. Parasitol. 63, 213-220. the major excreted protein (Mep) from a transformed mouse fibroblast cell- Vernet, T., Berti, P. J., Demontigny, C., Musil, R., Tessier, D. C., Menard, line as a catalytically active precursor form of cathepsin L. Biochem. J. 248, R., Magny, M. C., Storer, A. C. and Thomas, D. Y. (1995). The ionization 449-454. state of a conserved amino-acid motif within the pro region participates in McIntyre, G. and Erickson, A. H. (1993). The lysosomal proenzyme the regulation of intramolecular processing. J. Biol. Chem. 270, 10838- receptor that binds procathepsin L to microsomal membranes at pH 5 is a 10846. 43KDa integral membrane protein. Proc. Natl. Acad. Sci. USA 90, 10588- Vernet, T., Khouri, H. E., Laflamme, P., Tessier, D. C., Musil, R., 10592. Goursalin, B. J., Storer, A. C. and Thomas, D. Y. (1991). Processing of Menard, R., Carmona, E., Takebe, S., Dufour, E., Plouffe, C., Mason, P. the papain precursor – purification of the zymogen and characterization of and Mort, J. S. (1998). Autocatalytic processing of recombinant human its mechanism of processing. J. Biol. Chem. 266, 21451-21457. procathepsin L – contribution of both intermolecular and unimolecular Vernet, T., Tessier, D. C., Richardson, C., Laliberte, F., Khouri, H. E., Bell, events in the processing of procathepsin L in vitro. J. Biol. Chem. 273, 4478- A. W., Storer, A. C. and Thomas, D. Y. (1990). Secretion of functional 4484. papain precursor from insect cells – requirement for N-glycosylation of the Mottram, J. C., Frame, M. J., Brooks, D. R., Tetley, L., Hutchison, J. E., pro-region. J. Biol. Chem. 265, 16661-16666. Souza, A. E. and Coombs, G. H. (1997). The multiple cpb cysteine Zhu, Y. and Conner, G. E. (1994). Intermolecular association of lysosomal proteinase genes of Leishmania mexicana encode isoenzymes which differ protein precursors during biosynthesis. J. Biol. Chem. 269, 3846-3851.