Cruzain structures: apocruzain and cruzain bound to S-methyl thiomethanesulfonate and implications for drug design

High-resolution structures of apocruzain and cruzain bound to S-methyl thiomethanesulfonate provide insights for the structure-based design of new cruzain inhibitors.

Abstract

Chagas disease, which is caused by Trypanosoma cruzi, affects more than six million people worldwide. Cruzain is the major cysteine protease involved in the survival of this parasite. Here, the expression, purification and crystallization of this enzyme are reported. The cruzain crystals diffracted to 1.2 Å resolution, yielding two novel cruzain structures: apocruzain and cruzain bound to the reversible covalent inhibitor S-methyl thiomethanesulfonate. Mass-spectrometric experiments confirmed the presence of a methylthiol group attached to the catalytic cysteine. Comparison of these structures with previously published structures indicates the rigidity of the cruzain structure. These results provide further structural information about the enzyme and may help in new in silico studies to identify or optimize novel prototypes of cruzain inhibitors.

1. Introduction  

Chagas disease, or American trypanosomiasis, is one of the leading causes of heart disease in Latin America (Longo & Bern, 2015 ▸). The illness, which is transmitted by the protozoan parasite Trypanosoma cruzi, affects more than six million people worldwide and causes more than seven thousand deaths per year (DNDial, 2018 ▸). The current therapies approved for this condition, nifurtimox and benznidazole, are suboptimal regarding efficacy and present numerous side effects that make treatment increasingly problematic. Both drugs are usually prescribed during the early acute stages of the disease and are less effective for the treatment of the chronic condition. Thus, the development of new drugs is an urgent priority (Longo & Bern, 2015 ▸; Thakare et al., 2018 ▸).

T. cruzi produces an array of potential target enzymes implicated in pathogenesis and host-cell invasion, including essential and closely related papain-family cysteine proteases. Cruzain (also known as cruzipain; gene product gp75151; McKerrow, 1999 ▸), which is the major cysteine protease in T. cruzi, is involved in diverse functions in the life cycle of the parasite. It is necessary for nutrition, host-cell invasion, differentiation and evasion of the host immune system (McKerrow, 1999 ▸; da Silva et al., 2016 ▸; Sajid & Mckerrow, 2002 ▸). Its active site contains two catalytic residues (Cys25 and His162), with a third residue (Asn182) contributing to proteolysis (Kerr et al., 2009 ▸). Similar to other cysteine proteases, cruzain is initially expressed as an inactive zymogen, which undergoes a pH-, temperature- and concentration-dependent self-activation to transform into the mature enzyme (Verma et al., 2016 ▸; Lee et al., 2012 ▸). The catalytically active domain is highly unstable, and its inhibition is necessary to avoid autoproteolysis. S-Methyl methane­thiosulfonate (MMTS) is extensively used to prevent this autoprocessing, since it is a specific and reversible cysteine inhibitor. This reagent forms mixed disulfides with the sulfhydryl group of cysteine, with the release of sulfonylmethane (Fig. 1 ▸). This modification is highly specific for sulfhydryl groups and is usually reversed by reducing agents such as β-mercapto­ethanol, tris(2-carboxyethyl)phosphine (TCEP) or dithiothreitol (DTT) (Karala & Ruddock, 2007 ▸; Lundblad, 2014 ▸).

Figure 1.

Schematic representation of the reaction of MMTS with a cysteine sulfhydryl group in the protein. The thiol group of the protein reacts with MMTS to form a dithiomethane adduct. The disulfide bond is reversible in the presence of reducing reagents such as DTT or TCEP.

The discovery or design of selective cysteine protease inhibitors has great therapeutic potential. Numerous strategies have been employed towards this end (Vieira et al., 2017 ▸). These include structure-based drug design, which uses structural information on molecular targets to identify and prioritize new ligands contributing to the development of drugs (Santos et al., 2018 ▸). 25 crystallographic structures of cruzain–inhibitor complexes have already been obtained and deposited in the Protein Data Bank (PDB). Most of the ligands in these complexes inhibit the enzyme by irreversibly binding to the thiol group of the catalytic Cys25. Despite these reports of cruzain complexes, apo structures of active cruzain are currently not available because of self-degradation. Here, we report an apo structure, which provides relevant data by allowing comparison with the complexes in order to understand the extent to which the cruzain structure is altered by induced fit upon ligand binding. Furthermore, we report the crystal structure of cruzain bound to MMTS. These results contribute further insights into the understanding of the structure–function relationships of cruzain. We observe that the apo and ligand-bound structures are very similar, with no significant ligand-induced conformational changes, information that is extremely relevant to molecular-docking studies. It is noteworthy that cruzain serves as a model for the cysteine cathepsin proteases, a large family of (patho-)physiologically important enzymes.

2. Methods  

2.1. Protein expression  

The plasmid containing the gene sequence encoding the C-terminally truncated procruzain (Δc; GenBank entry M84342.1) was kindly provided by Dr William Fernandes and Dr James McKerrow. The oligonucleotide encoding the His₆-tagged procruzain sequence was ligated into a pET-21a vector (Novagen/EMD) and transformed into ArcticExpress (DE3) RIL Competent Cells (Agilent). Recombinant cruzain was expressed and purified using a modified version of the method described by Lee et al. (2012 ▸).

The starter culture used to express His₆-tagged procruzain consisted of 5 ml Luria broth containing ampicillin (100 µg ml⁻¹) and gentamycin (20 µg ml⁻¹). After shaking for >15 h at 310 K, the overnight LB starter culture was added directly to 1 l of autoinduction medium N-5052 (Studier, 2005 ▸) containing ampicillin (100 µg ml⁻¹) and gentamycin (20 µg ml⁻¹) in baffled 2 l shaker flasks. After incubation at 293 K and 210 rev min⁻¹ for 24 h (the OD₆₀₀ was >0.5), the temperature was reduced to 289 K and the culture was incubated for an additional 48 h at 210 rev min⁻¹. The cells were collected by centrifugation (4000g, 2 × 30 min, 277 K) and lysed.

2.2. Protein purification and activation of procruzain  

The cell pellets were resuspended in 50 ml lysis buffer per litre of expression culture and lysed by sonication (20 s pulse, 120 s intervals for a 2 min total pulse time at 50 W power). The lysis buffer consisted of 50 mM Tris pH 10, 300 mM NaCl, 10 mM imidazole, 1 mM CaCl₂, 1 mM MgSO₄, a dash of DNase I (Sigma), 2 µM phenylmethanesulfonyl fluoride (PMSF) and 2 mM MMTS. The resulting lysate was clarified by two cycles of centrifugation (5697g, 30 min, 277 K) and the supernatant was passed through a 0.45 µm filter for FPLC purification. All subsequent purification steps were carried out at 277 K. Recombinant procruzain was purified from the supernatant using Ni–NTA Superflow resin (Qiagen, Hilden, Germany). The following buffers were used: wash/binding, 50 mM Tris pH 10, 300 mM NaCl, 10 mM imidazole; elution, 50 mM Tris pH 10, 300 mM NaCl, 500 mM imidazole. Elution took place with 25% elution buffer. Fractions containing His₆-tagged procruzain obtained upon elution were combined and MMTS (final concentration of 2 mM) was added. The eluates were concentrated using Amicon Ultra centrifugal filter units (3 kDa molecular-weight cutoff; Millipore). For zymogen activation, the buffer was changed to 0.1 M acetate pH 5.5, 0.9 M NaCl, 10 mM EDTA containing no reducing agents using NAP-5 desalting columns (GE Healthcare). For the autoactivation of procruzain, the pH was changed to 5.3, the procruzain concentration was adjusted to 1 mg ml⁻¹, and 5 mM dithiothreitol (DTT) was added and incubated at 310 K. Transformation of the inactive zymogen (∼37 kDa) into the catalytically active domain (cruzain, ∼23 kDa) was monitored by removing 50 µl aliquots of the solution at selected time points. The reaction was interrupted by placement of the sample tubes on ice. These samples were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis under denaturing conditions and by a fluorescence-based activity assay using the peptide substrate Z-Phe-Arg-AMC (Fonseca et al., 2015 ▸). After activation, DTT was removed using the activation buffer without reducing agents on NAP-5 desalting columns (GE Healthcare). The cruzain was then immediately inhibited with the covalent reversible inhibitor MMTS at a final concentration of 1 mM to prevent self-degradation of the protein. A fluorescence-based activity assay confirmed enzyme inhibition. The total reaction time for processing procruzain into active cruzain was generally 20–45 min, as indicated by a change in the aspect of the protein solution, which became clear. The final FPLC purification was performed via size-exclusion chromatography on a Superdex 75 10/300 GL column (GE Healthcare) pre-equilibrated in a similar buffer to that used for zymogen activation but without the reducing agents. MMTS (final concentration of 1 mM) was added to each of the fractions immediately upon elution from the column. Protein samples were concentrated before and after size-exclusion chromatography to a concentration of approximately 10 mg ml⁻¹ using Amicon Ultra centrifugal filter units (3 kDa molecular-weight cutoff; Millipore). The protein concentration was assayed by UV₂₈₀ using molar extinction coefficients ∊₂₈₀ of 68 910 and 60 430 M ⁻¹ cm⁻¹ for His₆-GS-procruzain and cruzain, respectively. The samples were stored at 193 K until use.

2.3. Cruzain crystallization  

Samples containing pure MMTS-inhibited cruzain were concentrated to 10 mg ml⁻¹ and buffer-exchanged into 2 mM bis-Tris pH 5.8 using an ultrafiltration unit with a 5 kDa cutoff membrane. To obtain cruzain crystals, crystallization plates were prepared using four previously described conditions: (i) 1.26 M ammonium sulfate, 0.2 M lithium sulfate pH 6 (PDB entry 3lxs; Chen et al., 2010 ▸), (ii) 0.1 M Tris pH 8.5, 2 M NH₄H₂PO₄ (PDB entries 3kku and 3i06; Ferreira et al., 2010 ▸; Mott et al., 2010 ▸), (iii) 20% PEG 3000, 0.1 M sodium acetate pH 4.5 (PDB entry 3iut; Brak et al., 2010 ▸) and (iv) 0.8 M sodium citrate pH 6 (PDB entries 1u9q, 1me3, 1me4, 1f29, 1f2a, 1f2b and 1f2c; Choe et al., 2005 ▸; Huang et al., 2003 ▸; Brinen et al., 2000 ▸). However, no cruzain crystals were obtained in these conditions. JBScreen Classic (Jena Bioscience, Jena, Germany) and Index (Hampton Research, Aliso Viejo, California, USA) screens were then set up to search for suitable crystallization conditions using 0.8 µl drops with a 1:1 protein:crystallization solution ratio in 96-well plates (Art Robbins Instruments, Sunnyvale, California, USA). The drops were equilibrated at 293 K. MMTS-inhibited cruzain crystals were obtained by the sitting-drop method in a condition consisting of 1.26 M NaH₂PO₄, 0.14 M K₂HPO₄ at pH 5.6 with no other salts and no buffer (from the Index screen). The crystals were cryoprotected in a 25% solution of ethylene glycol in mother liquor and were then flash-cooled in liquid nitrogen. To obtain the apocruzain structure, the crystals were soaked for 3 h in 40 mM DTT in the mother liquor before cryocooling.

Diffraction data sets were collected on beamline ID29 at the European Synchrotron Radiation Facility (ESRF), Grenoble, France using radiation of wavelength 0.97 Å. For apocruzain, 2760 frames were collected with a 0.08 s exposure time and a 0.1° oscillation range between frames, while for the MMTS-inhibited structure 2380 frames were collected with 0.08 s exposures and 0.1° oscillation between frames. The reflections were indexed and integrated using MOSFLM and scaled using SCALA (Battye et al., 2011 ▸). The structures of both cruzain bound to MMTS and apocruzain were solved by molecular replacement with Phaser (McCoy et al., 2007 ▸) using the existing structure of cruzain bound to a noncovalent and competitive inhibitor (PDB entry 3kku; Ferreira et al., 2010 ▸) as a template. Structural models were built using Coot (Emsley et al., 2010 ▸) and refined using PHENIX (Afonine et al., 2012 ▸). Coordinates and data sets were deposited in the Protein Data Bank (PDB entries 6n3s and 6o2x).

2.4. Mass spectrometry  

0.5 µg of protein was digested with chymotrypsin either with or without the prior reduction/alkylation of SH groups. The digestion buffer was 100 mM Tris–HCl, 10 mM CaCl₂ pH 8.0. The reduction buffer was dithiothreitol (Thermo Scientific; 30 min at 333 K). Alkylation was performed using iodo­acetamide (Thermo Scientific) for 30 min at room temperature. Digestion with chymotrypsin (Promega) took place for 20 h at 298 K. After desalting with ZipTip C18 (EMD Millipore), the peptides were loaded onto an Acclaim PepMap RSLC column (C18, 75 µm × 15 cm) and the column was developed with an acetonitrile gradient [solvent A, 0.1%(v/v) formic acid; solvent B, 0.1%(v/v) formic acid/90%(v/v) acetonitrile; 5–45% B in 60 min] at a flow rate of 300 nl min⁻¹ at 328 K. The HPLC (Dionex Ultimate 3000, Thermo Fisher Scientific) was directly coupled via nano-electrospray to a Q Exactive Orbitrap mass spectrometer (Thermo Fisher Scientific). The capillary voltage was 2 kV. For peptide identification, a top 12 method was used, with the normalized fragmentation energy at 27%. Peptide sequences and modifications were established using Peaks Studio X (Bioinformatics Solutions).

2.5. Analyses of cruzain structures  

All 25 cruzain structures available in the PDB were aligned against our apocruzain structure in PyMOL (Schrödinger) considering all atoms of the structures. A second analysis was also performed using only active-site residues. The active site of the enzyme was defined as the residues involved in catalytic activity and those observed to interact with ligands in structures deposited in the PDB. The differences among the aligned structures were analyzed using the root-mean-square deviation (r.m.s.d.). The PyMOL script ColorByRMSD (Shandilya et al., 2012 ▸) was used to color the structures of apocruzain and PDB entry 2aim (Gillmor et al., 1997 ▸) according to r.m.s.d., demonstrating the regions that were most different and flexible between them. A linear graph to compare the r.m.s.d. values between the apocruzain and PDB entry 2aim structures was prepared using the SUPERPOSE program (Krissinel & Henrick, 2004 ▸).

3. Results and discussion  

3.1. Expression of soluble cruzain and crystallization  

The cruzain zymogen was expressed as a soluble fraction using the autoinduction method described by Studier (2005 ▸). Recombinant cruzain was purified and activated following a modified version of a previously published protocol (Lee et al., 2012 ▸). The final yield of purified active cruzain was 20 mg per litre of autoinduction medium.

Crystallization attempts using four conditions that had previously been employed to obtain structures of cruzain–ligand complexes did not result in any crystals. Therefore, the JBScreen Classic (Jena Bioscience, Jena, Germany) and Index (Hampton Research, Aliso Viejo, California, USA) screens, containing 480 crystallization conditions, were set up to search for new conditions. Drops were equilibrated at 293 K and a novel cruzain crystallization condition was identified. MMTS-inhibited cruzain crystals were obtained by the sitting-drop method in 1.26 M NaH₂PO₄, 0.14 M K₂HPO₄ at pH 5.6 with no further salt or buffer components.

3.2. Crystal structure determination  

The cruzain–MMTS complex crystallized with two independent copies of the mature enzyme in the asymmetric unit. The complex was refined to a resolution of 1.2 Å, yielding an R _free of 17.7% and an R _work of 16.4%. Only a few cysteine protease structures complexed with MMTS are available, and they include only one structure of a cruzain–MMTS complex (PDB entry 3kku). However, the MMTS was modeled with a low occupancy of only 30% in this complex. Here, we have determined the structure of cruzain in complex with MMTS at atomic resolution (Figs. 2 ▸ a and 2 ▸ c, Table 1 ▸). The electron density confirmed the presence of a disulfide group (SCH₃) connected to the catalytic cysteine in both copies of cruzain present in the asymmetric unit. Interestingly, the SCH₃ group adopted two different conformations. Conformation 1 (Fig. 3 ▸, magenta sticks) is compatible with the active-site architecture usually observed in cruzain structures. However, conformation 2 of the SCH₃-modified Cys25 (Fig. 3 ▸, gray sticks) is in steric conflict with the catalytically productive orientation of His162. Therefore, conformation 2 is sterically feasible only if the imidazole ring of His162 adopts a different rotameric orientation (Fig. 3 ▸, gray sticks). However, we did not observe this alternative conformation in the density map. This finding suggests that the side chain of His162 is flexible to some extent if pushed out by an active-site ligand such as MMTS. Mass-spectrometric experiments were performed to further confirm the modification at Cys25. For this purpose, cruzain inhibited by MMTS was digested with chymotrypsin to generate peptide fragments harboring the catalytic Cys25. Peptide analysis was complicated by the close proximity of Cys25 to a disulfide bond (Cys22–Cys63), which could not be resolved by digestion with chymotrypsin. However, we still managed to identify some peptides and were able to confirm the modification of Cys25 by SCH₃ (Fig. 4 ▸ a).

Figure 2.

Close-up view of the active site of cruzain: active sites of the structures of the complex of cruzain with MMTS (a) and of apocruzain (b). The electron density (2F _obs − F _calc) defining the catalytic residues of cruzain bound to MMTS (c) and apocruzain (d) is contoured at 1σ over the mean. Cartoon representations of cruzain bound to MMTS and apocruzain are shown in white and pink, respectively. Selected residues of the active site are highlighted as sticks. O, N and S atoms are colored red, blue and yellow, respectively. This figure was prepared using PyMOL.

Table 1. Data-collection and refinement statistics.

Values in parentheses are for the highest resolution shell.

	Cruzain–MMTS	Apocruzain
Data collection
Space group	P1211	P1211
a, b, c (Å)	32.99, 72.37, 79.71	32.83, 72.21, 79.81
α, β, γ (°)	90, 91.80, 90	90, 91.53, 90
Resolution (Å)	53.57–6.53 (1.21–1.19)	36.11–6.53 (1.21–1.19)
R merge † (%)	5.09 (5.92)	2.7 (43.6)
Mean I/σ(I)	6.9 (2.4)	29.6 (2.5)
Completeness (%)	97.0 (94.5)	99.6 (97.9)
Multiplicity	4.1 (4.3)	3.9 (3.3)
Refinement
No. of unique reflections	115006 (5548)	115537 (5760)
R work‡/R free § (%)	16.4/17.7	16.6/13.7
R.m.s. deviations
Bond lengths (Å)	0.023	0.007
Bond angles (°)	1.36	0.93
B factors (Å2)
Protein	8.71	11.74
Water	20.16	24.31
Other ligands	19.82	12.24
Ramachandran plot
Most favored regions (%)	97.62	96.71
Additionally allowed regions (%)	2.38	3.06
Outliers (%)	0.00	0.23

^† R _merge = , where I_i(hkl) is the intensity of the ith measurement and 〈I(hkl)〉 is the mean intensity for that reflection.

^‡ R _work = , where |F _obs| and |F _calc| are the observed and calculated structure-factor amplitudes, respectively.

^§ R _free was calculated with 5.0% of the reflections in the test set.

Figure 3.

Inhibition of cruzain with MMTS leads to covalent modification of the catalytic Cys25 with an SCH₃ group. (a) Enlarged view of the active site of the cruzain–MMTS complex. The SCH₃-modified Cys25 (Sch25) adopts two conformations. Conformation 1 (magenta sticks) represents the classical cruzain active-site architecture. Conformation 2 of Sch25 (gray sticks) is in steric conflict with the typical orientation of His162 and thereby causes reorientation of the His162 imidazole ring (the alternative orientation of His162 was not observed in the density map and was therefore not included in the final PDB file). (b) The electron density (2F _obs – F _calc) surrounding the active-site residues is contoured at 1σ over the mean.

Figure 4.

Mass spectra confirming the reversible modification of the catalytic Cys25 by MMTS. (a) The addition of SCH₃ to Cys25 (monoisotopic mass Cys-SCH₃, 148.99) could be confirmed both on the y-series (y1–y2) and b-series (b11–b10) ions of the Ala15–Trp26 peptide. (b) After treating MMTS-modified cruzain with DTT and iodoacetamide, the SCH₃ group was removed and replaced by a carbamidomethyl group (monoisotopic mass Cys-carbamidomethyl, 160.03).

To obtain the apocruzain structure, crystals inhibited with MMTS were soaked in a solution containing the reducing agent DTT. Apocruzain was crystallized with two complete copies (residues 1–215) in the asymmetric unit, and its structure was determined to 1.19 Å resolution (Figs. 2 ▸ b and 2 ▸ d, Table 1 ▸). The model was refined to an R _free of 16.6% and an R _work of 13.7%. Mass-spectrometric experiments were also carried out to confirm the reversibility of the MMTS reaction after treatment with DTT. Cruzain inhibited by MMTS was reduced with DTT and then alkylated with iodoacetamide, and finally digested with chymotrypsin. After peptide analysis it was observed that the SCH₃ group bound to Cys25 was replaced by a carbamidomethyl group, further confirming the reversibility of MMTS inhibition (Fig. 4 ▸ b).

These structures contribute to the rational design of improved cruzain inhibitors by providing a better understanding of the flexibility and conformational changes that are characteristic of the active site of cruzain. Several structure–activity relationship (SAR) studies focusing on cruzain–inhibitor complexes have relied on X-ray crystallography. However, owing to concentration-dependent self-proteolysis, no apo structure of active cruzain was available. In silico docking and modeling studies in structure-based drug design usually consider the active site of cruzain to be rigid. This is supported by the high similarity amongst published structures; however, it was not known whether the binders cause similar conformational changes or whether the active site of the enzyme is mostly rigid.

To better understand the structural features of this T. cruzi target, the apocruzain structure was compared with the 25 crystallographic structures of cruzain that were already available. These include structures obtained in multiple crystal forms, with different crystal-packing contacts, in space groups such as C121, P6₅22, P4₃2₁2 and P3₂. Generally, the structures are very similar, showing only small fluctuations among them; the r.m.s.d. values ranged from 0.21 to 0.43 Å, which are considered to be minor differences when comparing protein structures (Fig. 5 ▸, Table 2 ▸). Global analyses showed that cruzain inhibited with benzoyl-tyrosine-alanine-fluoromethylketone (PDB entry 1aim; Gillmor et al., 1997 ▸) is the structure that is most different from the apo structure (r.m.s.d. of 0.43 Å). When the residues of the active site were analyzed, it was observed that overall the residue conformations were quite conserved throughout all of the structures, with the major differences being found for Ser142, Met145 and Glu208 (Fig. 5 ▸ c). Additionally, it was noticed that the structure inhibited by benzoyl-arginine-alanine-fluoromethylketone (Bz-Arg-Ala-FMK; PDB entry 2aim; Gillmor et al., 1997 ▸) presented the greatest differences in the active-site residues (r.m.s.d. of 0.35 Å).

Figure 5.

Alignment of apocruzain versus the 25 cruzain structures available in the PDB. (a) Cruzain structures are shown in cartoon representation. (b) C^α traces of cruzain structures. (c) Enlarged view of the active sites of all available cruzain crystal structures.

Table 2. Comparison of the overall structures and active sites of apocruzain and cruzain complexes available in the PDB.

Alignment and determination of the r.m.s.d. values of apocruzain (PDB entry 6n3s) versus cruzain structures deposited in the PDB were performed in PyMOL. R.m.s.d. values were calculated considering all atoms of the structures.

	R.m.s.d. values for apocruzain versus cruzain structures (Å)
Cruzain structure	Global structures	Active sites
1aim	0.430	0.305
1ewl	0.352	0.294
1ewm	0.360	0.291
1ewo	0.320	0.299
1ewp	0.339	0.256
1f2a	0.298	0.264
1f2b	0.306	0.292
1f2c	0.259	0.293
1f29	0.289	0.242
1me3	0.344	0.328
1me4	0.345	0.298
1u9q	0.286	0.237
2aim	0.365	0.350
2oz2	0.313	0.273
3hd3	0.211	0.217
3i06	0.263	0.275
3iut	0.307	0.315
3kku	0.243	0.179
3lxs	0.299	0.257
4klb	0.265	0.179
4pi3	0.310	0.242
4qh6	0.315	0.252
4w5b	0.255	0.178
4w5c	0.346	0.317
4xui	0.332	0.346
Cruzain–MMTS (chain A)	0.085	0.162
Cruzain–MMTS (chain B)	0.217	0.265

To highlight the principal differences between the apo­cruzain (PDB entry 6n3s) and Bz-Arg-Ala-FMK-inhibited cruzain (PDB entry 2aim) structures, we compared them using the ColorByRMSD (PyMOL) and SUPERPOSE (CCP4) functions. ColorByRMSD colors two structures according to the r.m.s.d. value, employing a color spectrum in which blue represents the minimum pairwise r.m.s.d. and red indicates the maximum. When we compare apocruzain and the Bz-Arg-Ala-FMK–cruzain complex (PDB entry 2aim), we observe that the structures are predominantly blue, presenting only three small loop regions in red (Fig. 6 ▸ a). In a complementary analysis, SUPERPOSE aligns and overlaps the two enzyme structures by matching graphs built on the secondary-structure elements of the protein, followed by an interactive three-dimensional alignment of protein backbone C^α atoms. The SUPERPOSE graph of apocruzain and Bz-Arg-Ala-FMK-complexed cruzain (PDB entry 2aim) presents r.m.s.d. values less than or equal to 2 Å for both the main chains and side chains for the majority of residues (Fig. 6 ▸ b). Therefore, both the SUPERPOSE graph and alignment using ColorByRMSD indicate only small differences in r.m.s.d. All of these results demonstrate that the cruzain structure is predominantly rigid (Fig. 6 ▸). We also investigated the residues that interact with the Bz-Arg-Ala-FMK inhibitor in PDB entry 2aim and evaluated how much their conformations differ from the conformations observed in the apoenzyme. Besides the covalent bond to Cys25, this ligand established polar interactions with Cys25 and Gly66 (Fig. 6 ▸ c). We also checked whether the nature of the ligand would influence the conformational changes. We then compared the apocruzain structure with a structure bound to a noncovalent and competitive benzimidazole inhibitor (PDB entry 3kku). This inhibitor forms hydrogen bonds to Gly66 and Asp161, and similarly to the structure of the Bz-Arg-Ala-FMK–cruzain complex (PDB entry 2aim) causes only minimum conformational changes in the active site of the enzyme (Fig. 6 ▸ d).

Figure 6.

Alignment of the apocruzain structure with the Bz-Arg-Ala-FMK–cruzain complex structure (PDB entry 2aim). (a) Alignment performed by the ColorByRMSD PyMOL script (Shandilya et al., 2012 ▸). (b) SUPERPOSE (Krissinel & Henrick, 2004 ▸) graph of each residue of the structure. MC_RMSD, root-mean-square deviation of the main chain. SC_RMSD, root-mean-square deviation of the side chain. (c) Comparison of the active sites of apocruzain and PDB entry 2aim. (d) Comparison of the active sites of apocruzain and PDB entry 3kku. The cruzain structures with PDB codes 2aim and 3kku and apocruzain are shown in yellow, white and pink, respectively. Selected residues from the active site are highlighted as sticks. O, N and S atoms are colored red, blue and yellow, respectively. Hydrogen-bond interactions are shown as yellow dashed lines. This figure was prepared using PyMOL.

The stability and rigidity of cruzain have been studied in silico through molecular-dynamics (MD) simulations. Five 20 ns MD studies were performed using the structure with PDB code 1me4. These simulations suggested that the active site of cruzain was predominantly rigid (Durrant et al., 2010 ▸). Another MD study carried out on a timescale of 100 ns using the structure with PDB code 4klb and two systems, apocruzain and cruzain bound to the NEQ176 inhibitor, evaluated the protein structure and dynamics by analyses of the secondary-structure (SS) elements, r.m.s.d. and root-mean-square fluctuation (r.m.s.f.). All of these parameters were consistent and corroborated the analysis of the experimental structures, with no significant variations during the simulation for SS elements, C^α-r.m.s.d. or C^α-r.m.s.f. Concerning the C^α-r.m.s.f., an exception was observed for two loops (loop 3, the region between Cys56 and Leu67, and loop 4, the region between Asp87 and Thr107), where the values increased to approximately 1 Å (Hoelz et al., 2016 ▸). The results presented here reinforce the idea that the cruzain structure is indeed quite rigid, as suggested by MD, without significant rearrangement upon ligand binding. These data will contribute to the understanding of the features of the active site of cruzain and can be utilized for in silico docking and modeling studies.

In addition to contributing to the analysis of cruzain rigidity, we provide a new protocol that can aid in the determination of new cruzain–inhibitor complexes. Previous complex structures were obtained by dedicating protein batches to co-crystallization with a given inhibitor. Here, we have established a protocol which allows the crystallization of the protein inhibited by MMTS followed by the removal of this inhibitor by soaking with a reducing agent. We have demonstrated that this protocol results in the apocruzain structure, yielding crystals that still diffract to atomic resolution. Therefore, we report a promising strategy that allows apocruzain crystals to be soaked with new inhibitors that may be useful for rational drug design for Chagas disease.

4. Conclusions  

Here, we report a new cruzain crystallization condition which yielded crystals that diffracted to atomic resolution. Two structures were obtained using this condition: a cruzain–MMTS complex and apocruzain. To obtain the latter, crystals of cruzain inhibited with MMTS were treated with a soaking solution containing DTT. Mass-spectrometric experiments confirmed the presence of a methylthiol group bound to the catalytic cysteine in the cruzain–MMTS structure and the absence of this modification in the apo structure. Comparison of apocruzain and cruzain complexes reinforced the overall rigidity of this enzyme, as previously predicted by molecular-dynamics studies.

Supplementary Material

PDB reference: apocruzain, 6n3s

PDB reference: cruzain bound to MMTS, 6o2x

Funding Statement

This work was funded by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior grant . Fundação de Amparo à Pesquisa do Estado de Minas Gerais grant . Conselho Nacional de Desenvolvimento Científico e Tecnológico grant . Austrian Science Fund grant W_01213 to Hans Brandstetter.