Characterization of the M32 metallocarboxypeptidase of Trypanosoma brucei: differences and similarities with its orthologue in Trypanosoma cruzi

Abstract

Metallocarboxypeptidases (MCP) of the M32 family of peptidases have been identified in a number of prokaryotic organisms but they are absent from eukaryotic genomes with the remarkable exception of those of trypanosomatids. The genome of Trypanosoma brucei, the causative agent of Sleeping Sickness, encodes one such MCP which displays 72% identity to the characterized TcMCP-1 from Trypanosoma cruzi. As its orthologue, TcMCP-1, Trypanosoma brucei MCP is a cytosolic enzyme expressed in both major stages of the parasite. Purified recombinant TbMCP-1 exhibits a significant hydrolytic activity against the carboxypeptidase B substrate FA (furylacryloil)-Ala-Lys at pH 7.0–7.8 resembling the T. cruzi enzyme. S everal divalent cations had little effect on TbMCP-1 activity but increasing amounts of Co²⁺ inhibited the enzyme. Despite having similar tertiary structure, both protozoan MCPs display different substrate specificity with respect to P1 position. Thus, TcMCP-1 enzyme cleaved Abz-FVK-(Dnp)-OH substrate (where Abz: o-aminobenzoic acid and Dnp: 2,4-dinitrophenyl) whereas TbMCP-1 had no activity on this substrate. Comparative homology models and sequence alignments using TcMCP-1 as a template led us to map several residues that could explain this difference. To verify this hypothesis, site-directed mutagenesis was undertaken replacing the TbMCP-1 residues by those present in TcMCP-1. We found that the substitution A414M led TbMCP-1 to gain activity on Abz-FVK-(Dnp)-OH, thus showing that this residue is involved in specificity determination, probably being part of the S1 sub-site. Moreover, the activity of both protozoan MCPs was explored on two vasoactive compounds such as bradykinin and angiotensin I resulting in two different hydrolysis patterns.

1. INTRODUCTION

Carboxypeptidases are enzymes that cleave the C-terminal amino acid residue from peptides and proteins. This hydrolysis step may result in substrate activation or degradation [1]. Regardless the outcome, this activity belongs to two major catalytic classes of proteases, namely metallocarboxypeptidases (MCPs) and serine carboxypeptidases [1]. MCPs are ubiquitous in nature and typically have a single zinc ion bound to the active site. Among carboxypeptidases, the family M14 is the most extensively studied group. This family includes several enzymes which participate in diverse processes such as blood coagulation and fibrinolysis, inflammation, innate immunity response, food digestion and pro-hormone and neuropeptide processing [2].

Previous studies on Thermus aquaticus, an extremely thermophilic bacterium, led to the identification of a new zinc-dependent MCP [3]. Contrary to archetypical M14 peptidases, Thermus aquaticus carboxypeptidase (TaqCP) was distinct in both size and sequence. TaqCP contained a HEXXH signature which is a conserved sequence found in the active site of neutral metalloendopeptidases, but not described previously in a carboxypeptidase. These facts led to the placement of TaqCP in the newly formed M32 family of MCPs [4]. Later studies performed on other members of this family also confirmed that this group of enzymes had also strong topological similarity with archaeal, bacterial and mammalian endopeptidases including angiotensin I-converting enzyme (ACE), neurolysin and thimet oligopeptidase [4, 5].

But not only the structural similarity to endopeptidases as well as their broad substrate specificity were the new features of these peptidases; their phylogenetic distribution was also quite peculiar. Sequence alignments have identified M32 family members in species from a very limited phylogenetic range. These include bacteria, archaea, some green algae and trypanosomatids, but not metazoa, suggesting that the genes have been acquired by horizontal gene transfer or retained for some special function that is no longer essential for higher organisms [6].

The Trypanosomatidae family includes several vector-borne protist parasites, causing human and veterinary diseases. The human parasites include a number of species in the genera Leishmania and Trypanosoma. In Trypanosoma, the two major human parasites are T. cruzi, the causative agent of Chagas' disease, and T. brucei, the causative agent of the Human African trypanosomiasis (HAT) or Sleeping Sickness. Both parasites are responsible for over 60.000 deaths per year in Latin America and Sub-Saharan Africa, respectively. For these infections no vaccines are available, and their treatment is dependent on a small number of drugs that have limited efficacy and can cause severe side effects [7]. These limitations make urgent the need to develop new drugs, more efficient, less toxic and more affordable [8]. In this context the study of peptidases has acquired considerable importance since several works have shown that many parasites deploy proteinases to accomplish both “housekeeping” tasks common to many eukaryotes as well as functions highly specific to the parasite life style. Host cell invasion and egress, encystation, excystation, catabolism of host proteins, differentiation, cell cycle progression, cytoadherence, and both stimulation and evasion of host immune responses are some processes in which this group of enzymes has been shown to be involved [9]. Thus, the possibility of developing selective inhibitors of key proteases of pathogenic parasites is currently explored as a novel chemotherapeutic strategy [9].

Herein we focus on T. brucei, an extracellular parasite, transmitted by the bite of the tsetse fly (Glossina spp.). The disease caused by this protozoan progresses from non-specific signs to typical symptoms such as wake–sleep cycle alteration and neurological dysfunction, leading to death in untreated patients [10]. In this work we report the cloning, purification and biochemical characterization of an MCP belonging to the M32 family (TbMCP-1) from T. brucei Lister 427. These results add to our knowledge of the peptidases present in T. brucei and may offer a new target for the development of a rational chemotherapy against trypanosomiasis.

2. MATERIALS AND METHODS

2.1 Materials

Peptide substrates were purchased from Sigma–Aldrich and Bachem Bioscience, except for those with the structure Abz-XXK(Dnp)-OH [Abz: o-aminobenzoic acid; K(Dnp): N´, 2–4-dinitrophenyllysine, and X different amino acid residues], which were synthesized as described in reference [11]. All other reagents were purchased from Sigma unless otherwise stated.

2.2 Trypanosome strains and culture conditions

The T. brucei bloodstream form (BSF) Lister 427 “single marker” trypanosome cell line (T7RNAPol TetR NEO) and the procyclic form (PCF) 29–13 cell line (T7RNAPol NEO TetR HYG) [12] were gifts from G. A. M. Cross (Rockefeller University). BSF cells were maintained in HMI-9 medium supplemented with 20% heat-inactivated fetal calf serum [13]. T. brucei PCF were grown at 28°C in medium SDM-79 [14] supplemented with haemin (7.5 mg/l) and 10% heat-inactivated fetal-calf serum.

2.3 Genomic DNA purification and molecular cloning

In order to clone the MCP gene from T. brucei Lister 427, two synthetic oligonucleotide primers were designed: ATG (5´-GGATCCatgaagggcatacaaagagctcg-3´) and STOP (5´-GAATTCtcagttggcatcgtcacggtag-3´). PCR amplification was carried out using genomic DNA as template. PCR conditions were as follows: initial denaturation (5 min at 94°C), denaturation (1 min at 94°C), annealing (45s at 62°C) and elongation (90s at 72°C) followed by a final extension step (10 min at 72°C). The PCR products were purified from a 1% agarose gel, using the QiaQuick protocol (Qiagen) and cloned into pGEM-T Easy vector (Promega). Sequencing of the products was performed using an ABI 377 DNA sequencer (PerkinElmer).

Site-directed mutations were introduced into the TbMCP-1 gene cloned in the plasmid pGEX 2T, using the QuickChange™ site-directed mutagenesis kit (Stratagene).

DNA was obtained from PCF of T. brucei by using the Proteinase K/phenol/chloroform method [15].

2.4 Expression and purification of TbMCP-1 fused to glutathione S-transferase (GST)

The T. brucei MCP-1 gene (TbMCP-1) was excised as BamHI/EcoRI fragment from pGEM-T Easy plasmid (Promega), gel purified and subcloned into the BamHI and EcoRI sites of the pGEX 2T expression vector. The resulting construct presented a GST tag at the N-terminus with a thrombin cleavage site. The construct was transformed into E. coli BL21 Codon Plus (DE3) cells. GST fusion protein was expressed by induction of exponential phase cultures (A600 =0.6) with 0.5 mM IPTG for 12 h at 18°C with vigorous (250 rpm) shaking. Bacteria were harvested by centrifugation at 5000 g for 30 min at 4°C, resuspended in 50 mM Tris-HCl pH 7.6, 150 mM NaCl, 0.1% Triton X-100, 1 mM PMSF and 1 mg/m l lysozyme, and centrifuged at 12000 g for 30 min at 4°C to obtain the bacterial crude extract.

The recombinant TbMCP-1 fusion protein was purified using a glutathione-agarose resin (GE) equilibrated with 50 mM Tris-HCl pH 7.6 containing 150 mM NaCl. The column was washed with 10 column vol. of the equilibration buffer and the sample was eluted with 50 mM Tris-HCl pH 7.6, 150 mM NaCl containing 10 mM reduced glutathione. Samples were desalted using a PD-10 column (GE) according to the manufacturer’s specifications. Protein purity was evaluated on a Coomassie Blue-stained SDS 10% polyacrylamide gel. A 6xHis-tagged version of TbMCP-1 was constructed as in the case of TcMCP-1 [6], and the recombinant enzyme was purified by IMAC-Ni²⁺. This recombinant TbMCP-1 was very unstable, but could be used to exclude the influence of the GST tag in enzyme dimerization and kinetics. Recombinant TcMCP-1 from T. cruzi was purified as described in [5].

2.5 Enzyme assays

Routinely, recombinant TbMCP-1 activity was assayed using furyl-acryloyl (FA)-Ala-Lys (200 µM) as substrate in 100 mM MOPS pH 7.3. Initial steady-state velocity was determined by continuous assay for a range of substrate concentrations at 340 nm with a Beckman DU 650 spectrophotometer. One unit of activity was defined as the amount of enzyme that released 1 µmol of the residue being cleaved per min at 25°C. Kinetic parameters for recombinant TbMCP-1 were also determined using FA-Ala-Lys in 100 mM MOPS buffer, pH 7.3.

The substrate preference with respect to P1´ position was determined using benzyloxycarbonyl-Ala-X (ZAX) substrates (Bachem and Sigma), using the ninhydrin method for detection [16]. The reaction mixture contained 6 mM ZAX substrate in 100 mM MOPS, pH 7.3. One enzyme unit was defined as the amount of enzyme that produced a ninhydrin-positive substance corresponding to 1 µmol of amino acid per min.

The substrate specificity with respect to P2 and P1 positions was determined using a series of fluorescence-quenched tripeptides [17]. Enzymatic activity was continuously followed with an AMINCO^®–Bowman Series 2 Spectrofluorimeter by measuring the fluorescence at 320 nm (λ_ex) and 420 nm (λ_em). The slope was converted into micromoles of substrate hydrolyzed per minute based on a calibration curve obtained from complete hydrolysis of each peptide.

2.6 Effects of pH, metal ions and general inhibitors

For the determination of the optimum pH, the enzyme activity was determined in the following buffers at 100 mM: MES (pH 5–6.7), MOPS (pH 6.5–7.9) and HEPES (pH 7–8.2). Metal-ion-dependence and inhibition pattern were investigated by assaying the purified recombinant peptidase activity after pre-incubation with different metal chlorides ranging from 1 µM to 1 mM or the chosen inhibitor. In all cases the reaction mixture contained 200 µM FA-Ala-Lys in 100 mM MOPS, pH 7.3.

2.7 Gel-filtration experiments

These experiments were conducted using a Superose 12HR 10/30 (Pharmacia Biotech) column in an Äkta Purifier System (GE). The sample was eluted with 50 mM Tris-HCl containing 200 mM NaCl. The enzyme activity was assayed with 200 µM FA-Ala-Lys in 100 mM MOPS pH 7.3. The molecular-mass markers used were vitamin B12 (1.3kDa), myoglobin (17kDa), ovalbumin (44kDa), BSA (66kDa), and γ-globulin (158kDa).

2.8 Antibody generation

Anti-TbMCP-1 polyclonal antibodies were obtained by immunizing mice with three doses (15 days apart) of the purified TbMCP-1 recombinant protein (subcutaneous injection). Each mouse received a first dose consisting of 20 µg of antigen with Freund's complete adjuvant (1:1) and other three doses, 5 µg each, with Freund's incomplete adjuvant. After the mice were bled, the serum was separated by centrifugation at 3,000×g and kept frozen at −20°C until use.

2.9 Western blot analysis

The expression pattern in the different stages of the life cycle of T. brucei was analyzed by Western blot. Parasites (20×10⁶ cells) were resuspended in cracking buffer, submitted to 10% SDS-PAGE and then transferred onto a nitrocellulose Hybond ECL membrane (GE). The membrane was blocked with 3% (w/v) non-fat milk and 2% glycine in TBS (50 mM TrisHCl pH 7.6 containing 150 mM NaCl) for 30 min and incubated with anti TbMCP-1 serum (1:100, 1hr.) diluted in blocking solution. It was then washed with TBS for 20 min, with 0.05% NP40 in TBS for 10 min, and with TBS for 10 min. After that, it was incubated with horseradish-peroxidase-conjugated goat anti mouse IgG (Sigma) (1:5000, 1h) in blocking solution. Blots were washed as before and developed with SuperSignal WestPico Chemiluminescent Substrate (Pierce).

2.10 Immunofluorescence experiments

To determine the subcellular localization of TbMCP-1 in BSF, 1×10⁶ parasites were washed with TD Buffer (0.025 M KCl, 0.4 M NaCl, 0.005 M MgSO₄, 0.1 M Na₂HPO₄, 0.01 M NaH₂PO₄, 0.1 M glucose), fixed with paraformaldehyde (PFA) 4% in PBS pH 7.6 for 1h at room temperature and then attached to poly-lysine coated glass coverslips (30 min). Parasites were permeabilized with 0.3% Triton-X-100 in PBS, pH 7.4 for 30s, washed 3 times with PBS pH 7.4 and incubated with blocking solution (50 mM NH₄Cl, 3% BSA, 5% goat serum, 1% fish gelatin in PBS pH 7.4) for 1hr. at room temperature. Incubation with primary antibody (1:150 in 1% BSA pH 8) was made for 1hr. Coverslips were washed three times with PBS for 1 min. Alexa Fluor^® 488-conjugated goat anti-mouse (Molecular Probes) (1:1000 for 1h) was used as secondary antibody. Coverslips were washed again three times with PBS and then mounted using 15 µl of Fluor Save (Calbiochem). Nucleus and kinetoplast were stained with DAPI (4´,6-diamidino-2-phenylindole) (5 mg.ml⁻¹).

PCF were air-dried on poly(L-lysine)-coated glass coverslips, fixed with 4% (w/v) PFA in PBS (20 min), incubated with 25 mM NH₄Cl (10 min) and permeabilized with 2% BSA, 0.1% saponin, 2% normal goat serum in PBS (30 min). The antibodies were diluted in blocking solution and incubation conditions were the same as with the BSF. Samples were examined with a Nikon Eclipse E600 microscope and photographed with a digital Sport RT Slider camera.

2.11 Activity of TbMCP-1 on vasoactive peptides

Bradykinin (BK, RPPGFSPFR, 1061Da) and angiotensin I (Ang I, DRVYIHPFHL, 1296Da) were used at 10 pmol/µl in buffer 10 mM MOPS pH 7.3 (TbMCP-1) or 10 mM MES, pH 6.2 (TcMCP-1). At different times, the reaction was stopped by mixing one volume of the sample with one volume of 0.1% α-ciano-4-hidroxycinnamic acid (in 50% acetonitrile). Samples were seeded in a Ground Steal plate (Bruker Daltonics) and analyzed by MALDI-TOF mass spectrometry (Ultraflex, Bruker Daltonics).

3. RESULTS

3.1 The genome of T. brucei encodes a single M32 MCP

Within the trypanosomatidae family, the peptidase family M32 seems to have followed different evolutive pathways. Thus, T. cruzi genome encodes two different M32 MCPs (TcMCP-1 and TcMCP-2), which are regulated throughout the parasite life cycle [6]. Leishmania spp, on the other hand, present multiple M32 paralogs and pseudogenes [6], all of them homologs of TcMCP-1. Most of them are species specific. In contrast, the related parasite T. brucei appears to have reduced to a minimum the M32 protein repertoire. Analysis of the genomic data of this organism led us to identify a single M32 homolog. This enzyme, named TbMCP-1, shares 72% identity to the characterized TcMCP-1 and displays the highest sequence identity (47–33%) with those M32 proteins of the proteobacteria group. Figure 1 shows the schematic representation of the conserved M32 peptidase motifs found in TbMCP-1 including the canonical HEXXH sequence that contains the active-site glutamic acid residue at position 268 [4], flanked by two histidine residues (positions 267 and 271) that coordinate the catalytic metal ion.

Figure 1.

A. In the alignment, sequences were sorted by similarity to TbMCP-1 using the CLUSTALW program. Similar residues according to the BLOSUM62 matrix are indicated by different tones of gray, according to the degree of similarity. Accession numbers with percent of identity to TbMCP-1 are as follows: TcMCP-1, CAG28317.1, 71%; TcMCP-2, CAG28318.1, 59%; Vibrio cholerae O1 biovar eltor str. N16961, NP_231057, 47%; Rickettsia rickettsii, ZP_00153287, 33%; Bacillus subtilis subsp. subtilis str. 168, NP_390090, 33%; Pyrococcus furiosus, 1KA4_A, 28% and Prochlorococcus marinus subsp. pastoris str. CCMP1986, NP_892611, 28%. B. Schematic representation of the conserved M32 family motifs.

3.2 Expression, purification and biochemical characterization of recombinant TbMCP-1

The M32 MCPs characterized to date differ in various aspects, including gene copy number, substrate specificity and pattern of expression. To address these aspects on the only M32 family member present in T. brucei, we analyzed its enzymatic properties. Thus, the full-length TbMCP-1 gene was cloned into the bacterial expression vector pGEX2T, and expressed in E. coli BL21 Codon Plus (DE3) cells as an N-terminally GST-tagged recombinant enzyme. TbMCP-1 was purified in one step using a glutathione–agarose resin with a yield of 9.6 mg/l of bacterial culture. Recombinant TbMCP-1 was evaluated for purity on Coomassie Blue stained 7.5 % polyacrylamide gel. The apparent subunit molecular mass of the purified enzyme was in good agreement with the calculated mass of the affinity-tagged product of the corresponding gen (approx. 77 kDa, Fig. 2A). As T. cruzi MCPs, TbMCP-1 eluted in a single peak from a Superose 12HR 10/30 column with an apparent molecular mass corresponding to 159 kDa, indicating that the enzyme associates into catalytically active homodimers (Figure 2B). A possible role of the GST tag in dimerization was excluded by repeating the experiment with a 6xHis-tagged recombinant enzyme, which gave an approximate value of 106 kDa.

Figure 2.

A. 7.5 % SDS-PAGE showing the different fractions of the TbMCP-1 purification stained with Coomassie blue. Lane 1, crude extract obtained from the disruption of induced E. coli cells; lane 2, glutathione-agarose eluate; lane 3, GST-agarose washes with 50 mM Tris-HCl pH 7.6, 150 mM NaCl; lane 4, purified TbMCP-1 eluted with 50 mM Tris-HCl pH 7.6, 150 mM NaCl, 10 mM reduced glutathione. B. The molecular mass of active TbMCP-1 was determined by molecular exclusion chromatography on a Superose 12HR 10/30 column: TbMCP-1 (▲), molecular mass markers (●) [thyroglobulin (670 kDa), γ-globulin (158 kDa), ovalbumin (44 kDa), myoglobin (17 kDa) and vitamin B12 (1.3 kDa)].

In order to analyze the expression pattern of TbMCP-1 along the parasite life cycle, we obtained polyclonal antibodies against the recombinant enzyme. Western blot analysis employing whole parasite extracts demonstrated that TbMCP-1 is expressed in both mammal and insect stages (figure 3A). This pattern of expression is similar to that of TcMCP-1, and contrasts with those of other MCPs found in T. cruzi and Leishmania major which were restricted to the insect stages [6, 18]. The subcellular localization of the T. brucei enzyme was studied by indirect immunofluorescence. Using the polyclonal antiserum raised against recombinant TbMCP-1 we obtained a diffuse uniform fluorescence throughout the cytosol of both procyclic and bloodstream forms of the parasite (Figure 3, B and C).

Figure 3.

A. Western blot showing expression of TbMCP-1 in vector (PCF) and host (BSF) stages of the parasite. The apparent molecular mass of the native enzyme (approx. 55 kDa) was in good agreement with the mass calculated from the corresponding gene (approx. 57.7 kDa). B. Immunofluorescence showing subcellular localization of the enzyme in the bloodstream form (green). C. Immunofluorecence showing subcellular localization of the enzyme in the procyclic stage of the parasite. Nucleus and kinetoplast were stained with DAPI (blue). Figures were prepared using the program ImageJ 1.46b.

The purified recombinant TbMCP-1 was used in enzymatic assays with N-blocked furylacryloyl dipeptides commonly employed for screening carboxypeptidases. These experiments revealed that the T. brucei enzyme exhibits a significant hydrolytic activity against the carboxypeptidase B (CPB, subfamily M14B) substrate FA-Ala-Lys at an optimal pH 7.0–7.8 (Supplementary Figure 1). The FA-Phe-Phe dipeptide, on the other hand, was not cleaved (Table 1).

Table 1.

Kinetic parameters of TbMCP-1.

	FAAlaLys			FAPhePhe
Enzyme	Km (mM)	kcat (s−1)	kcat/Km (s−1/mM)	Km (mM)	kcat (s−1)	kcat/Km (s−1/mM)
TbMCP-1	0.260	32	123	N.D.	-	-
TbMCP-1 M304R	0.240	8.6	36	0.033	2.5	76
TbMCP-1 E268A	N.D.	-	-	-	-	-
TbMCP-1 R348A	N.D.	-	-	-	-	-
TcMCP-1a	0.144	10	69	N.D.	-	-
TcMCP-1 M304Ra	N.D.	-	-	0.011	0.3	27
TcMCP-2a	N.D.	-	-	0.019	1.2	63

N.D. Not detected under these measurement conditions.

^aData from Reference [6]

Metal-ion chelators EDTA (10 mM) and o-phenanthroline (10 mM) were strong inhibitors, consistent with the TbMCP-1 being a metal-dependent enzyme (Table 2). Carboxypeptidase B-like inhibitors such as the arginine analog GEMSA or Plummer´s inhibitor were also effective whereas DL-benzylsuccinic acid, a potent inhibitor of carboxypeptidase A, was less effective. Although M32 MCPs have a similar fold to that of the angiotensin I-converting enzyme (ACE1), a dipeptidyl peptidase of the family M2, the specific ACE1 inhibitor captopril had no effect on TbMCP-1 (Table 2). On the other hand, DX600, a commonly used ACE2 inhibitor [19], presented a modest inhibition on T. brucei enzyme most likely due to a non optimal amino acid sequence. It is noteworthy that TcMCP-1 was similarly insensitive to captopril and partially inhibited (residual activity 22.5%) by the same concentration of DX600 (not shown).

Table 2.

Effect of different inhibitors on TbMCP-1 activity.

Inhibitor	Concentration	Type	Residual activity (%)
-			100 ± 9.9
EDTA	10 mM	Metallo-proteases, metal activated proteases	8.7 ± 1.2
o-phenanthroline	10 mM	Metallo-proteases, metal activated proteases	0
PMSF	1 µM	Serine proteases	93 ± 24
E64	5 µM	Cysteine proteases	85 ± 3.7
Pestatin A	1 µM	Some aspartic proteases	89 ± 4.1
Captopril	5 µM	M2 dipeptidilpeptidases (ACE1)	114 ± 39
DX600	12.5 µM	M2 carboxypeptidase (ACE2)	60.3 ± 6.7
GEMSA	5 µM	Carboxypeptidases B-like	9.1 ± 1.2
Plummer´s inhibitor	5 µM	Carboxypeptidases B-like	6.8 ± 3.8
DL Benzylsuccinic Acid	5 µM	Carboxypeptidases A-like	33 ± 16

As for TcMCP-1, the presence of Co²⁺ inhibited TbMCP-1 activity on FA-Ala-Lys substrate. The residual activity measured in the presence of 10 µM CoCl₂ was 16±7%. Mn²⁺ had also effect but at a higher concentration, the residual activity being 18±5 % with 100 µM MnCl₂. Ca²⁺ and Mg²⁺ ions also resulted inhibitory (72±3 and 65±3 were the residual activities, respectively, when measured in the presence of 100 µM of the corresponding metal chloride). Ni²⁺ and Zn²⁺, two strong inhibitors of TcMCP-1, had no effect on the T. brucei enzyme.

3.3 P1´ preference

To further characterize the substrate preference with respect to the P1´ position of the identified TbMCP-1, several commercially available ZAX (where X position was successively occupied with 1 of 17 amino acids) substrates were used. As shown in supplementary figure 2, the recombinant TbMCP-1 acted best on substrates having a basic residue at P1´ position. Maximum activity was found with ZAK (0.28 µmol/min.mg). ZAR and ZAM were also hydrolyzed, a lthough the relative activity on these peptides was lower (37.1% and 18.3%, respectively, when compared with the ZAK substrate). Other polar C-terminal residues including His, Tyr, Ser and Asn, and the aliphatic Leu, where cleaved at a lower extent.

3.4 Kinetic parameters

The kinetic parameters of TbMCP-1 were evaluated using the FA-Ala-Lys substrate in 100 mM MOPS, pH 7.3. A possible effect of the GST tag on the kinetic parameters was excluded by determining the K_m for this substrate with a 6xHis-tagged recombinant enzyme, which was 239 µM. As shown in table 1, TbMCP-1 has less affinity for this dipeptide than TcMCP-1, although its catalytic efficiency is higher. Replacement of Glu268 (active-site glutamic acid residue) or Arg348 (residue involved in COOH terminal binding) by an alanine residue resulted in complete loss of activity.

Previous studies suggested that Met at position 304 could favor B-type substrate utilization [5]. In order to test this hypothesis Met304 present in TbMCP-1 was replaced by an Arg (residue found in the carboxypeptidase A-like TcMCP-2). Table 1 shows that the inclusion of an Arg in that position confers TbMCP-1 activity towards the FA-Phe-Phe substrate, thus confirming a role of this residue in substrate specificity.

3.5 FRET substrates

Fluorescence Energy Resonance Transfer (FRET) peptides are an excellent alternative for enzyme kinetic studies and for analysis of the enzymatic activities in biological fluids, crude tissue extracts or on the surface of cells in culture [20]. As a consequence we decided to evaluate the activity of TbMCP-1 on a FRET peptide series containing a free C-terminal group [21]. Table 3 summarizes the activity of the recombinant T. brucei enzyme on these fluorogenic tripeptides. The fixed modified Lys residue at P1' position led us to evaluate the effect of the amino acid present at P1 and P2 positions which strongly influence the efficiency of the enzyme. For comparative reasons, TcMCP-1 was also included in the experiment. This enzyme showed a slightly different substrate preference as compared with TbMCP-1, although the hydrolytic activity of the enzyme was also affected by the amino acid residues present in P1 and P2 positions. Notably neither the T. brucei nor the T. cruzi enzyme could hydrolyze Abz-ARK(Dnp)-OH, probably due to the presence of an Ala residue at P2 position. The inclusion of a Val residue in P1 only abolished the activity of TbMCP-1, whereas TcMCP-1 showed activity against this substrate (Table 3).

Table 3.

Kinetic parameters of TbMCP-1 and TcMCP-1 on fluorescence-quenched tripeptide substrates.

	TbMCP-1			TcMCP-1
Substrate	Km (µM)	kcat (s−1)	kcat/Km (s−1/µM)	Km (µM)	kcat (s−1)	kcat/Km (s−1/µM)
Abz-ARK(Dnp)-OH	ND	-	-	ND	-	-
Abz-HRK(Dnp)-OH	1.8	0.503	0.279	>10	-	-
Abz-LRK(Dnp)-OH	2.7	0.012	0.004	4.4	0.156	0.036
Abz-ERK(Dnp)-OH	2.2	0.010	0.005	>10	-	-
Abz-FRK(Dnp)-OH	1.1	1.374	1.249	2.9	0.346	0.119
Abz-FSK(Dnp)-OH	0.3	0.013	0.043	5.4	0.032	0.006
Abz-FHK(Dnp)-OH	3.0	2.159	0.720	5.4	0.886	0.164
Abz-FVK(Dnp)-OH	ND	-	-	0.3	0.153	0.510
Abz-FPK(Dnp)-OH	0.3	0.004	0.013	1.9	0.004	0.002

N.D. Not detected under these measurement conditions.

3.6 Mapping TbMCP-1 substrate specificity determinants by site-directed mutagenesis

In order to explain the lack of activity of TbMCP-1 on Abz-FVK(Dnp)-OH we submitted the amino acid sequence of this enzyme to the SWISS-MODEL comparative protein modeling server [22]. Using TcMCP-1 crystal structure as template we determined which residues in the TbMCP-1 substrate binding channel differ from the corresponding residues in the T. cruzi enzyme. Only five changes were found: K91L, R288M, F411S, I408V and A414M (Figure 4). To investigate the hypothesis that residues at these positions might be responsible for the lack of activity, site-directed mutations were introduced into this enzyme. Of the five point mutants constructed, only the replacement of Ala414 with Met resulted in the appearance of activity of TbMCP-1 on Abz-FVK(Dnp)-OH. The affinity of the mutant enzyme (K_m 0.21 µM) on Abz-FVK(Dnp)-OH was similar to the one of the wild-type TcMCP-1, although the k_cat was lower (0.01 s⁻¹). These results suggest that this residue is part of the S1 subsite.

Figure 4.

Homology models of TbMCP-1 based on the crystal structure of TcMCP-1 (pdb id 3dwc) were utilized to map the positions of the potentially critical residues involved in substrate specificity. A. Superimposed C-α ribbon of a single TcMCP-1 subunit (green) and TbMCP-1 (gray). B. Surface diagram of TbMCP-1 subunit. C. Close-up view of protozoan MCP active-site cleft. Residues relevant for catalysis as well as residues that differ between both enzymes are shown as sticks. Colour-code; blue for nitrogens, yellow for sulphurs, and red for oxygens. The catalytic metal ion of the protozoan enzymes is depicted as a pink. Figures were prepared using the program PyMOL (The PyMOL Molecular Graphics System, Version 1.3, Schrödinger, LLC.).

3.7 Protozoan MCPs are active on bioactive peptides

Contrary to other previously characterized M32 peptidases, the protozoan enzymes seem to have a marked substrate preference at P1´ position. Moreover, the nature of the amino acid residue present at P1 or P2 also affects the hydrolytic efficiency of both enzymes as shown in section 3.5. To explore these effects on natural peptides we employed two vasoactive compounds, BK (RPPGFSPFR) and Ang I (DRVYIHPFHL). As shown in supplemental figure 3, after 30 minutes of digestion TcMCP-1 solely produces Des-Arg9-BK (RPPGFSPF↓R), a B-type agonist of B1 receptors, whereas the T. brucei enzyme generates a mixture of BK (1–7) (RPPGFSP↓F↓R) and Des-Arg9-BK, which is eventually fully converted to BK (1–7).

The final products of Ang I hydrolysis are also different for both MCPs, thus TbMCP-1 mainly produced Ang II (DRVYIHPF↓H↓L) through the sequential removal of the last two amino acid residues of the peptide, whereas the T. cruzi enzyme produces a mixture of Ang (1–9), Ang II and Ang (1–7). According to these results TbMCP-1 generates the same products that ACE would produce acting on these peptides.

4. DISCUSSION

We report herein the identification and biochemical characterization of the only M32 MCP present in the T. brucei genome. This enzyme is expressed in the cytosol of both mammal and insect stages of the parasite. Interestingly, this is the only M32 peptidase that has orthologs in the TriTryps genomes. TcMCP-1, the TbMCP-1 counterpart described in T. cruzi, also presents a similar pattern of expression and in vitro utilizes the CPB substrate FA-Ala-Lys [6]. These features contrast with those of species-specific M32 MCPs found in L. major and TcMCP-2, an enzyme of T. cruzi with CPA-like substrate preference, these enzymes being restricted to the insect stages [6, 18].

Previous studies have shown that M32 MCPs could hydrolyze a broad spectrum of amino acid residues at P1´ [3, 16, 18, 23]. On the other hand, the recombinant T. brucei enzyme described here presented a marked substrate preference for basic residues (Lys and Arg). This characteristic is more reminiscent to Bacillus subtilis CP, an M32 MCP reported to act over a few substrates including ZAK and ZAR [24]. The use of internally quenched fluorogenic tripeptides has also pointed out the effect of the amino acid residue located at P1 and P2 positions (and so the selectivity of the S1 and S2 subsites). These aspects were not previously analyzed for this group of enzymes. Our results suggest that trypanosomatid MCPs act with a different efficiency on the substrates of this series, these differences being more marked for TbMCP-1. An explanation of this could be found in the structure. Several crystallographic structures of M32 family members have been published [4, 5, 24]. However, none of them was determined in the presence of a substrate or inhibitor, and thus the putative substrate binding sites are still unknown. Moreover, the role that some residues play in catalysis has been suggested but has not been confirmed. In this direction, we had pointed out the relevance of Arg348. This residue, proposed to form a salt bridge with the C-terminal carboxylate group of the bound substrate [4, 5] is essential since its replacement by an Ala abolished the enzyme activity. Arg348 is structurally conserved in other monocarboxypeptidases of the clan MA, namely, ACE2. In particular, the crystal structure of ACE2, revealed that the guanidino group of Arg273 (the equivalent residue of Arg348) made a bidentate H-bond with the terminal carboxylate of the inhibitor MLN-4760 [25]. Further, its larger size, compared with Gln present in ACE, was suggested to eliminate the S2′ subsite in ACE2 [26]. Two major implications were derived from this single residue change. First, the switch in the peptidyl dipeptidase activity of ACE to the observed carboxypeptidase activity of ACE2 and second, why the potent ACE inhibitors such a s lisinopril, enalaprilat, and captopril are inactive against ACE2 [25]. Thus, for both TbMCP-1 and TcMCP-1, not only this essential Arg residue is conserved compared with ACE2, but also the response to some inhibitors. Two other residues, Met304 and Ala414, were found to influence the substrate specificity. In the case of Met304, previous work has suggested that this residue could favor B-type substrate utilization [5]. In the case of the T. brucei enzyme, its replacement by a positively charged residue resulted in the appearance of activity towards the FA-Phe-Phe substrate, with the concomitant reduction of the catalytic efficiency over FA-Ala-Lys. Although the effect of this substitution is not as dramatic as it has been reported for TcMCP-1 [5], which completely lost the activity on basic substrates when the same mutation was introduced, it is clear that this residue plays a major role in A/B-type substrate discrimination. Due to its orientation respect to the catalytic Glu, this residue may be part of the S1´ subsite. The second residue studied was Ala414. This residue located in the vicinity of the proposed S1 pocket is entirely responsible for the lack of activity of TbMCP-1 over the Abz-FVK(Dnp)-OH substrate since the introduction of a Met confers the T. brucei enzyme the ability to act on this peptide. Thus, in spite of having 72% identity, in MCP-1 like enzymes a few single amino acid residues changes can cause significant changes in substrate utilization.

No biological function has yet been assigned to M32 peptidases. On the bases of their biochemical properties and stage-specific expression, it has been suggested that protozoan MCPs might play a role in peptide catabolism [18]. In particular, Isaza and co-workers suggested that a species-specific M32 MCP found in L. major could facilitate the growth of the promastigote form of the parasite participating in the degradation of peptides and proteins to free amino acids [18]. Other possible functions came from the structural similarity of this family with mammalian metallopeptidases, including ACE and ACE2 [5]. Both enzymes a re zinc peptidases that play a central role in the renin-angiotensin system. Taking into account the marked substrate preference reported for some M32 MCPs a regulatory role in the metabolism of small peptides could not be excluded. In this direction, we have shown that TcMCP-1 could produce exclusively des-Arg bradykinin by hydrolysis of bradykinin in vitro. This peptide was reported to facilitate T. cruzi cell invasion through B1 receptors since addition of [Leu8]-des-Arg9-BK, a B1 receptor antagonist, is known to reduce the parasite infectivity [27]. In the case of T. brucei, the kallikrein-kinin system was reported to be severely dysregulated in trypanosome-infected hosts [28]. Thus, TbMCP-1 could be a good candidate to influence the activity of host vasoactive kinins. Moreover, there is recent evidence that suggests that M32 peptidases are secreted by trypanosomatids [29, 30] making the latter hypothesis feasible. In fact, it has been reported that a pyroglutamyl peptidase released by T. brucei could modulate plasma neuropeptide levels of infected hosts in vitro and in vivo [31]. Alternatively TbMCP-1 might be released in vivo by dead or dying trypanosomes lysed in the circulation by the various antimicrobial host defense mechanisms. This is probably the case of T. brucei oligopeptidase B, an enzyme whose activity is found in the plasma of infected rats and, as TbMCP-1, does not contain any known secretion sequences. The authors hypothesize that the last explanation is the most plausible one [32]. Thus, trypanosome-derived peptidases, including TbMCP-1, could play a role in modulating peptide hormone dynamics in the plasma of infected hosts, having significant implications for our understanding of the lesions observed in African trypanosomiasis.

5. CONCLUSIONS

T. brucei contains only one member of the M32 metallopeptidase family, TbMCP-1. This enzyme has similar properties when compared with its orthologue in T. cruzi, TcMCP-1, but also has significant differences in its response to divalent cations and in substrate specificity. The ability of TbMCP-1 to act on bradykinin and angiotensin I open up the possibility of a role of the enzyme in the pathogenesis of Sleeping Sickness, which might make it a suitable target for the development of new chemotherapeutic agents.

HIGHLIGHTS.

• -The genome of T. brucei encodes a single M32 MCP, orthologue of T. cruzi MCP-1.
• -TbMCP-1 was cloned, expressed in E. coli, purified and biochemically characterized.
• -TbMCP-1 displays different substrate specificity compared to its T. cruzi orthologue.
• -Site-directed mutagenesis identified a new residue determinant of specificity.
• -Both MCPs cleaved vasoactive peptides showing different hydrolysis patterns.

SUMMARY.

Despite having 72% identity to TcMCP-1, T. brucei MCP-1 displays different substrate utilization. Comparative homology models led us to map several residues that could explain this difference.

ABBREVIATIONS

BSA

bovine serum albumin

BSF

bloodstream form

CPB

carboxypeptidase B

E64

L-transepoxysuccinyl-leucylamido-[4-guanidino] butane

FRET

Fluorescence Energy Resonance Transfer

GEMSA

guanidinoethylmercaptosuccinic acid

IMAC-Ni²⁺

immobilized metal affinity chromatography using Ni²⁺

MOPS

3-(N-morpholino) propanesulfonic acid

NP40

nonyl phenoxypolyethoxylethanol

Plummer´s inhibitor

DL-2-mercaptomethyl-3-guanidinoethylthiopropanoic acid

PMSF

phenylmethylsulfonyl fluoride

PCF

procyclic form

TD Buffer

trypanosome dilution buffer