Identification of Structure-Stabilizing Interactions in Enzymes: A Novel Mechanism to Impact Enzyme Activity

Abstract

Cruzain, a cysteine protease in the cathepsin family, is pivotal to the life-cycle of Trypanosoma cruzi, the etiological agent in Chagas disease. Current inhibitors of cruzain suffer from drawbacks involving gastrointestinal and neurological side effects and as a result have spurred the search for alternative anti-trypanocidals. Through sequence alignment studies and intra-residue interaction analysis of the pro-protein of cruzain (pro-cruzain), we have identified a host of non-active site residues that are conserved among the cathepsins. We hypothesize that these conserved amino acids play a critical role in structure-stabilizing interactions among the cathepsins and are therefore crucial for eventually gaining protease activity. As predicted, mutation of selected conserved non-active site amino-acid candidates in cruzain resulted in a compromised structural stability and a corresponding loss in enzymatic activity relative to wild-type enzyme. By advancing the discovery of novel, non-active-site-based targets to arrest enzymatic activity our results potentially open the field of alternative inhibitor design. The advantages of defining such a non-active-site inhibitor design space is discussed.

Introduction

Human Chagas’ disease, also known as American Trypanosomiasis, is caused by the protozoan parasite Trypanosoma cruzi [1]. According to WHO, approximately 5.7 million people are infected with T. cruzi in 21 Latin American countries alone [2]. However, in the recent past, Chagas has been increasingly detected in other non-endemic countries in the Region of the Americas (Mexico, Canada and the United States of America) [3]. Chagas disease may go almost unnoticed in the first stage (acute phase), but after the second stage (chronic phase) heart problems start developing, eventually leading to death. Chagas remains the primary cause of infectious myocarditis in the world resulting in 50,000 deaths per year [4]. Chagas disease can also cause enlarging of the esophagus and the colon, which can also be fatal [5].

T. cruzi is transmitted via the feces of a triatomine insect (or “kissing” bug) which after drawing blood from the host, defecates; thereby discharging feces containing the parasite into the skin wound [5]. Other modes of transmission include transfusion of infected blood, oral transmission through contaminated food, vertical transmission, organ transplantation, and even in laboratory accidents [3]. Cruzain (CZN), a cysteine protease that plays a very important role in the life cycle of T. cruzi [5–7], has been proposed as an attractive drug target [7–9]. The three-disulfide-bond containing enzyme is a lysosomal proteinase and is critical to the parasites’ nutrition. Cruzain’s participation in blood breakdown makes it critical for parasites survival, replication and proliferation [10]. Recent studies showed that T. cruzi is a complex of genetically diverse isolates highly phylogenetically related to pathogenic protozoans T. cruzi-like species, Trypanosoma cruzi marinkellei and Trypanosoma dionisii. This corroborates cruzain as valuable target for drugs, vaccine, diagnostic and genotyping approaches [11].

Cruzain, formally known as cruzipain [12], is synthesized as a pro-protein (PCZN), which is the inactive form (zymogen) of the enzyme. PCZN possesses a pro-domain and a catalytic domain comprised of approximately 104 and 215 amino acid residues respectively [13]. Cruzain, like all cysteine proteases, contains a catalytic triad comprising Cys 25, His 159, and Asn 175 [7, 14]. The activation is achieved by proteolytic excision of the prodomain, an event believed to occur in vivo by a multistep process that may involve multiple endosomal⁄lysosomal peptidases or even proteases present in the extracellular environment [14].

Therapeutics against T. cruzi includes the use of drugs like benznidazole and nifurtimox which are the only available medications for acute phase Chagas’ disease, but are highly toxic and have poor efficacy in long-lasting chronic infections [15–18]. Current strategies designed to inhibit cruzain have been limited to inhibit cruzain’s active site [8, 19]. In this study we propose a bottom-up approach to destabilize the zymogen before it becomes functional. Sequence analysis was used to discover critical residue–residue interactions within the pro domain. To assess whether these selected residues have a consequence on PCZN activity, we generated a recombinant version of wild-type PCZN and mutants E25V, D82V, E87V proteins. Herein, using biochemical assays and circular dichroism studies, we have demonstrated structural and activation rate changes in all mutants when compared to the wild type (WT)-PCZN.

Materials and Methods

Protein Sequence Alignment

BLAST (Basic Local Alignment Search Tool) [15] analysis of PCZN amino acid sequence was performed in order to identify conserved residues among proteases related to PCZN. Then, sequences obtained were selected for further alignment using the following criteria: total score >350, query coverage >90%, and maximum identity >50%. Proteases with the highest scores were aligned with PCZN sequence for comparison using Clustalw Omega [16].

Homology Modeled Structure of PCZN and Protein Residue–Residue Interaction Analysis

Since no crystal structure for PCZN has been deciphered yet, a homology modeled of PCZN was constructed using I-Tasser [17]. Inter-protein ionic and hydrophobic interactions were further calculated by Protein Interaction Calculator (PIC) [18], where salt bridges were assigned when distance between the two atoms of opposite charge was less than 6 Å and side chain–side chain hydrophobic interactions were also calculated within 6 Å distance. The homology-modeled structure, inter-protein interactions, and residue substitutions were visualized using PyMol software [20].

PCZN-WT and Mutant Constructs

pQE30-PCZN-WT construct was a gift from Dr. Ana Paula Lima from Federal University of Rio de Janeiro, Brazil (Biophysics Institute Carlos Chagas Filho) and was obtained from Tulahein 2 strain T. cruzi. DNA sequence of pQE30-PCZN-WT was verified by using pQE vector primers: forward Type III/IV primer: 5′-CGGATAACAAT TTCACACAG-3′ and reverse primer: 5′-GTTCTGAGGTC ATTACTGG-3′ (Eurofins MWG Operon, AL, USA) on a 3130xl Genetic Analyzer (GeneAmp PCR System 9700, Life Technologies). PCZN-WT-pQE30 was then subjected to mutagenesis to generate mutants: E25V; D82V; E87V; C141A by Genscript, USA. See Supplementary Fig. 1A–B for pQE30-PCZN-WT and mutants map constructs. Residues were mutated to valine, as opposed to a smaller residue like glycine or alanine, to retain a residue of similar size and for a lack of charge be the only variant. Additionally, glycine can induce disturbance in the backbone and is normally avoided.

Protein Expression of Recombinant PCZN-WT and Mutants (E25V, D82V, E87V, C141A)

Plasmids pQE-30 containing 6x histidine-tagged PCZN (WT and mutants) at the N-terminus were introduced into chemically competent E. coli DH5α (Invitrogen) for cloning and into E. coli M15 for expression (Invitrogen). A modified version of a method previously described by Murray, et al. 2009 [21] was used to identify and obtain high-expressing colonies for clones containing the 6xHis-tagged protein. A 5 mL LB culture media with antibiotics (amp 100 μM/kan 50 μM) was inoculated with a freshly transformed colony and grown to an O.D.₆₀₀ of 0.7–0.9. An LB/amp/kan plate was inoculated with bacteria and grown overnight at 37 °C. Three colonies were selected and fresh LB/amp/kan media was inoculated separately and grown overnight at 37 °C. A small-scale expression, described below, was performed for each 5 mL culture. Before induction with Isopropyl-B-D-1-thiogalactopyranoside (IPTG), cells from each expression were inoculated into a LB/amp/kan plate. The plates were identified as “colony A”, “colony B” and “colony C”. Expression levels were checked by harvesting 250uL of cells suspension at 3700 rpm for 5mim. The pellet was re-suspended in 100 μL lysis buffer, then 20 μL of lysate was mixed with 20 μL of 4x SDS sample buffer and heated at 75 °C for 20 min. The sample was centrifuged and the supernatant was loaded into SDS-PAGE verify expression. High expression colonies were grown and stored as glycerol stocks.

Expression

A small-scale culture was grown by inoculating 100 mL of LB broth containing 100uM Amp and 50uM Kan with one colony of transformed BL21 cells. The culture was grown overnight at 37°C while shaking. The small-scale culture was diluted ten times the original volume in LB/amp/kan and grown at 37°C while carefully monitoring the absorbance at 600 nm. Once the optical density of the culture reached 0.5–0.7, expression was induced by adding 1 M IPTG to a final concentration of 1 mM and was incubated 4 h at 37°C [8]. The cells were harvested by centrifugation at 5000 rpm and 4 °C for 15 min using a Sorvall Centrifuge RC28S (Wilmington, DE, USA) with GSA rotor, and lysed in 100 mM NaH₂PO₄, 10 mM Tris-HCl, 10 mM imidazole, and 8 M urea (pH 8), under agitation for 30 min at 4°C. The cell lysate formed was further disrupted by sonication at 60 W for 3 × 30 s pulses (with 1 min delay between pulses) on ice bath, with a Branson Sonifer 450. The soluble material was recovered by centrifugation at 11000 rpm for 20 min using a Sorvall Centrifuge RC28S with F28⁄36 rotor.

Protein Purification of Recombinant Mutants and Wild-Type

Recombinant mutants and wild-type PCZN proteins were purified using affinity chromatography. The solubilized His-tagged recombinant proteins were mixed with preequilibrated Ni-nitrilotriacetic acid (NTA)-agarose resin (loaded with Ni²⁺) (ThermoFisher) for 15 min. Unbound protein was released from the column and the resin was washed with wash buffer (100 mM NaH₂PO₄, 10 mM Tris-HCl, and 8 M urea, pH 6.3). The recombinant proteins were then eluted an imidazole gradient (10, 30, 50, and 200 mM) and the purity was analyzed by 12% SDS-PAGE. The eluted samples were pooled and submitted to in vitro refolding by incubation with 5 mM dithiothreitol at 37 °C for 45 min, followed by 20-fold dilution in ice-cold refolding buffer containing 100 mM Tris-HCl (pH 8), 1 mM EDTA, 250 mM arginine, and 20% glycerol, and further incubation at 4°C for 24 h. The solution was concentrated in a mini-reservoir (RC800 Millipore) under nitrogen pressure to a final volume of 30 mL using 10 kDa molecular weight cutoff cellulose acetate filters (Millipore, Bedford, MA, USA) at 4°C. Protein concentration was measured with Nano Drop 1000 instrument (Thermo Scientific, software ND 100 v3.6.0) by using extinction coefficient of 63,473 /cm/ M and molecular weight of 35.403 kDa.

Activation Assays of Recombinant PCZN-WT and Mutants

After protein was refolded and further dialyzed against cold water, the protein concentration was adjusted to 0.3 mg/mL. Protease activity of recombinant mutants and PCZN-WT was measured using Z-Phe-Arg-7-amido-4-methyl-cou-marin-HCl (Z-Phe-Arg-AMC) as substrate, and performed as previously described; albeit, with minor modifications [13]. To determine reproducibility, assays were carried out twice on different days. Activation assays were performed in a DM45 Spectrofluorimeter (On-line Instrument Systems, Inc.) using Spectra Works software and were carried out in a quartz cuvette 1 cm path length (Fisher Scientific). Fluorescence was recorded as a function of time at 440 nm emission wavelength and 380 nm excitation wavelength, and set for 0.100 integration time and 600 PMT HV. Recombinant protein was added to a 1:1 activation buffer (200 mM NaCH₃COO, 1.8 M NaCl, 10 mM DTT^red and 10 mM EDTA, pH 5.2) to substrate buffer (40 μM Z = F-RAMC, 5 mM EDTA, 50 mM NaH₂PO_4, 100 mM NaCl, 5 mM DTTred) for a final protein concentration of 73.2 μg/mL and pH of 5.3. The fluorescence was recorded promptly after protein addition and incubated at 37 °C for a period of 50,000 s. When maximum activity was observed, the mature enzyme was inhibited with 10uM Leupeptin (Sigma) for 15 min at 37 °C to prevent auto-degradation of the protein. An aliquot of the active protein was lyophilized and run on 12% SDS-PAGE and gels were stained in Coommassie brilliant blue R-250 (BioRad).

Circular Dichroism Studies of Folded Recombinant Mutants and Wild- Type PCZN

Refolded protein was dialyzed against phosphate-buffered saline buffer pH 7.4 using 3 kDa MWCO dialysis tubes (GE Healthcare) and further concentrated using centrifugal filter tubes (Vivaspin2, 3 kDa MWCO). Resultant dialyzed protein was passed through PVDF 0.22 μm syringe filters (Fisherbrand) to remove any debris of precipitated protein. Protein concentration was measured using NanoDrop instrument (Thermo Scientific Nano Drop 1000, software ND 100 v3.6.0). Assays were performed in 0.1 cm path length quartz cuvette (Fisher Scientific). Jasco 1500 spectropolarimeter was used for CD measurements that has Peltier temperature controller. Data was documented at 4°C and 37°C and samples were incubated for 10 min prior recording of the data. Mean molar ellipticity [theta] was calculated using the formula theta (mDeg)/ncl; where, n (number of chiral centers), c (concentration of the protein) and l is the path length.

Results

Protein Sequence Alignment

BLAST analysis of PCZN was used to identify proteases from closely related pathogenic parasites (Table 1). Results displayed five high-scored cysteine proteases from two classes of parasites, Trypanosoma and Leishmania, to be closely related to the cysteine protease from T. cruzi. Trypanosomas brucei, rangeli and congolense showed the highest identity as expected with 66, 75, and 64 maximum identity percentage respectively, while Leishmania braziliensis and Leishmania mexicana came last with 57 and 55 maximum identity percentage, respectively [16, 17].

Table 1.

Results of BLAST analysis of pro-cruzain amino acid sequence (pro and catalytic domains)

Protein	Accession #	Total score	Query coverage (%)	Maximum identity (%)
Trypanosoma brucei (brucipain)	XP_845225.1	470	98	66
Cathepsin B-like protein
Trypanosoma rangeli	AFA34859.1	463	100	75
Cathepsin L-like protein
Trypanosoma congolense (congopain).	AAA18215.1	436	96	64
Cathepsin L-like protein
Leishmania braziliensis	XP_001562140.1	371	93	57
Cathepsin L-like protease
Leishmania mexicana	CBZ24131.1	362	91	55
Cathepsin L-like protease

A multi-sequence alignment of PCZN and the resulted cysteine proteases from the above-mentioned BLAST inquiry was performed using Clustalw [16, 17]. The amino acid sequence alignment of brucipain, congopain, and cathepsin L-like proteases (from T. rangeli, L. braziliensis and L. mexicana) shows several conserved residues throughout the pro-domain as shown in Fig. 1. Residues addressed by an arrow (E25, F32, F35, E49, F56, D82, E86, and E87) coincide with interactions found using molecular dynamics studies (un-published data).

Fig. 1.

Pro-region sequence alignment of PCZN-WT and cysteine proteases. Sequence alignment of proteases (pro-domain region only): T. cruzi, T. rangeli, T. brucei, T. congolense, L. braziliensis, and L. mexicana. Arrows indicate the residues selected for substitution: E25, F32, F35, E49, F56, D82, E86, and E87

Homology-Modeled Structure of PCZN and Protein Residue–Residue Interaction Analysis

A 3D homology-modeled structure of the zymogen form of cruzain was successfully created using I-Tasser (Fig. 2). This model was based on the available X-ray data from pro-catalytic complexes (zymogen proteins). For instance, the best scored model was constructed using another cysteine protease proenzyme (Procathepsin L, PDB ID entry 1CS8) as a template with homology identity of 37% with PCZN and query coverage of 97%.

Fig. 2.

PCZN structure homology-modeled by I-Tasser. Light gray represents the pro domain and dark gray represents the catalytic domain

As interactions within a protein are essential for stability and function of the protein we aimed to enumerate critical residue–residue non-covalent-interactions that can occur within the pro domain, primarily ionic and hydrophobic interactions, by using PIC server. Enlisted in Table 2 are some of the intra-protein residue interactions found within the pro domain of PCZN. Intra-protein interactions within the catalytic domain were also observed, but are not included in this study since our focus is the impact of the pro domain. The underlined residues correspond to some of the residues found in the amino acid sequence alignment in Fig. 1. In addition, some of the abovementioned ionic interactions were visualized using Pymol (Fig. 3). Results obtained from multi-sequence alignment and intra-protein interactions allowed us to select residues E25, F32, F35, E49, F56, D82, E86, and E87 as candidates for single amino acid substitution as they are involved in ionic and hydrophobic interactions. The selection of critical residues was also influenced by structural impact data from molecular dynamic studies of in-silico mutants (un-published data). Nevertheless, only E25V, D82V, and E87V mutants were expressed and analyzed in this study. Mutant C141A was included in this study as control due to its lack of protease activity by the removal of the catalytic cysteine. Note: mutational analysis of other identified conserved sites were omitted due to the limited scope of this study.

Table 2.

Inter-protein ionic and hydrophobic interactions found within pro-segment of PCZN

Ionic	Hydrophobic
Glu87-Arg91	Phe32-Phe35
Glu25-Arg57	Phe32-Phe56
Arg41-Asp82	Phe35-Phe56
Lys36-Glu49
Glu86-Arg89
Arg52-Glu87

Interactions done by PIC server and found within 6 Å. The underlined residues correspond to some of the residues found in the amino acid sequence alignment in Fig. 1

Fig. 3.

Ionic interaction within the pro domain of homology modeled PCZN. Interaction between a Glu25-Arg57; b Asp82-Arg41; c Glu87-Arg91. Atom-atom distance measurements are made in angstroms

Protein Expression of PCZN-WT and Mutants (E25V, D82V, E87V, C141A)

DNA sequence of all constructs was confirmed by performing DNA sequencing. The resultant amino-acid sequence for the wild-type protein is shown in Supplementary Fig. 2. Protein expression of PCZN-WT and mutants using the pQE30 vector expression system was successfully accomplished using a previously described protocol [21] with modifications (Fig. 4).

Fig. 4.

Recombinant PCZN WT and Mutants. 12 % SDS-PAGE gel showing refolded and concentrated WT and mutants of PCZN. a Line 1: PCZN-E25V, line 2: PCZN-E82V, line 3: PCZN-D87V, line 4: PCZNWT, and line 5: MW marker. b Line 1: PCZN-C141A, and line 2: MW marker

Protein Purification of PCZN and Mutants

Based on conditions described earlier, we generated 1 L of IPTG-induced culture bacteria. The E. coli M15 cells were harvested, and the bacterial pellet was re-suspended in lysis buffer. The soluble protein was collected by sonication followed by centrifugation for removal of cellular debris. Purification of recombinant proteins was carried out using Ni-NTA affinity chromatography with a concentration gradient of imidazole as the eluting agent to obtain high purity protein fractions. In Fig. 5a, lines 2–4, we can observe a thick band at approximately 40 kDa corresponding to PCZN-C141A protein along with impurities. These fractions were eluted with 50 mM imidazole elution buffer followed by elution with 100 mM imidazole elution buffer (lines 5–7). The purity of the protein increases with an increase in imidazole concentration with the purest being eluted at 200 mM imidazole. Nevertheless, to avoid early elution of his-tagged protein we utilized elution buffers with lower imidazole concentrations (10 and 30 mM). These were applied several times at the beginning for removal of nonspecific binding proteins. The rest of the recombinant mutant proteins and WT were purified following the latest elution process. The final yield of protein for 1 L of cell culture was around 1.6 mg/mL. Unfortunately, more than half of the protein precipitated out during the dialysis processes resulting in approximately 0.7 mg/mL of refolded protein. His-tag remained on all recombinant proteins.

Fig. 5.

Purification of recombinant PCZN WT and Mutants. a PCZNC141A protein purification. Line 1: FT (cell lysate proteins from flow through fraction after His-tag recombinant protein has been bounded to resin), lines 2–4: mix of PCZN-C129A protein with undesired protein (eluted with 50 mM imidazole elution buffer), lines 5–7: mix of PCZN-C141A protein with undesired protein (eluted with 100 mM imidazole elution buffer), line 8: shows band at approximately 36 kDa corresponding to the 6xHis-PCZN-C141A (fraction eluted with 200 mM imidazole), and line 9: MW marker (EZ-Run Rec Protein Ladder, Fisher BioReagents). b line 1: MW marker, lines 2–3: mix of PCZNC141A protein with undesired protein (eluted with 200 mM imidazole elution buffer), lines 4–9: purified PCZN-C141A protein fractions (eluted with 200 mM imidazole elution buffer). c PCZN-E82V purification. Line 1: FT (showing unbound PCZN-E82V protein), lines 2–6 and 8–9 purified PCZN-E82V protein fractions (eluted with 200 mM imidazole elution buffer), line 7 MW marker. d PCZN-WT purification. Line 1: FT, lines 2–6 and 8–9 purified PCZN-WT protein fractions (eluted with 200 mM imidazole elution buffer), line 7 MW marker. Figure e: PCZN-E87V purification. Line 1: FT, line 2: washed fraction, line 3: fraction eluted with 10 mM imidazole buffer, line 4: fraction eluted with 30 mM imidazole buffer, lines 5–7: fractions eluted with 200 mM imidazole buffer, line 8: MW marker. f PCZNE25V purification. Line 1–8: Purified fractions eluted with 200 mM imidazole buffer, line 9: MW marker (EZ-Run Pre-stained Rec Protein Ladder, Fisher BioReagents)

Activation Assays of Recombinant Mutants and Wild-Type PCZN

Enzymatic activity of recombinant cysteine proteases was accessed by the liberation of the fluorescent leaving group 7-amino-4-methyl-coumarin (AMC) from the peptide substrate Z-Phe-Arg-AMC, upon incubation with the enzyme at 37°C. The fluorescence was collected for a period of 50,000 s although only 3000 s are shown as plateau was observed for all the enzymes at that time (Fig. 6). The half life, V_max and K (rate constant) of the autoactivation of WT PCZN and mutants are shown in the table below:

Fig. 6.

Activation of recombinant PCZN-WT and mutants. The kinetics of autoactivation of recombinant PCZN was monitored via cleavage of AMC from the Z-Phe-Arg-AMC substrate. The y axis designates the fluorescence intensity in arbitrary units (a.u.), while the x axis shows the time in seconds. The figure shows the non-linear least square regression fit of the activity assay. The bar over each data point shows the standard deviation of the mean of experiments run in duplicates. Purple circles: PCZN-WT, dark blue triangles: E25V, red diamonds: PCZN-D82V, green inverted triangles: PCZN-E87V, fluorescent blue squares: PCZ-C141A (color figure online)

Kinetics	WT	(C141A)	(E25V)	(E87V)	(D82V)
Vmax	0.3782	0.06162	0.4293	0.3908	0.06237
T 1/2	1.336	6.384	5.111	2.929	6.62
K (rate constant)	0.5187	0.1086	0.1356	0.2366	0.1047
(ΔV=Vmax-V0)	0.3829	0.0596	0.4751	0.4379	0.05975

From the observed results, the D82V mutation was found to be the most deleterious for enzymatic activity compared to WT, followed by E87V and lastly E25V. The C141A mutant showed no activity as expected. Activation products were observed in a 12% SDS-PAGE gel (Fig. 7). The values were calculated after fitting the data using non-linear least square regression. The equation used for the fit is as follows: V = V_max + (V₀ − V_max)e^−k.x

Fig. 7.

Products from autocatalytic activation of PCZN-WT and mutants. 12% SDS-PAGE gel showing the products of autocatalytic activation of WT and mutants of PCZN. Line 1: Elution fractions from purification of WT-PCZN, pooled, and concentrated before activation, lines 2–5: activated proteins (2) WT-PCZN, (3) PCZN-D82V, (4) PCZN-E87V, (5) PCZN-E25V; (6) MW marker. Thirty microliters of protein was loaded into each line

Circular Dichroism Studies of Folded Recombinant Mutants and Wild-Type PCZN

To determine the effect of mutation on the overall structure of PCZN, we acquired the far-UV CD of the PCZN-mutants and compared it with the structure of PCZN-WT. Putative 3D homology modeling predicted predominant helical structure (30%) for the PCZN-WT (Fig. 8). WT far-UV CD data confirmed the predicted primarily helical structure of PCZN WT as evidenced by the strong minima at 222 nm and 208 nm. The native conformation of WT and mutant variants were assessed at 4 °C (Fig. 9). PCZN-WT has a stronger minima at 208 nm than at 222 nm signifying rigidity of the helix. Overall, all mutations were shown to decrease the total helicity of the protein. D82V, E25V and E87V mutations increased the hydrophobic content of the protein as they replaced negatively charged amino acid residues involved in hydrogen bonding. This may partially explain decrease in helical content compared to the PCZN-WT. Figure 9 also shows the helical wheel comparison between WT and the mutants. In all helices analyzed, except for E25V, no other helices have a long stretch of hydrophobic residues (five and above). The distribution of hydrophobic residues and hydrophilic mostly alternate each rather than forming a patch. The E25V mutation shows a hydrophobic stretch facing the outside. This mutation increases the hydrophobicity of the peptide stretch that was already rich in hydrophobic residues (Fig. 9c).

Fig. 8.

Predicted secondary structure of PCZN-WT. Secondary structure of PCZN-WT reproduced based on amino acid sequence using Phyre2 server. Regions not covered are shown along with the predicted secondary structure. Overall, α-helical content is the predominant secondary structure for procruzain

Fig. 9.

Circular dichroism spectra and helical wheels for PCZN-WT and mutants. CD data demonstrate the native structure of PCZN-WT at 4 °C and that is compared with the mutants at same temperature. WT was compared to: (a) D82V; (b) E87V; (c) E25V; (d) C141A. The helical wheel diagram shows the change in the helical structure post mutation. Arrows indicate the residue mutated. “N” indicates N-terminal and “C” C-terminal positions

Figure 10a–e shows the effect of physiological temperature on the overall structure of PCZN-WT and their mutants. Figure 10f depicts the ratio of [θ]₂₀₈ and [θ]₂₂₂ since they are the minima for α-helical signature conformation. As PCZN-WT had a stronger minima at 4 °C, i.e., the native conformation, we compared it with that at 37 °C, i.e., conformation at physiological temperatures. From the activity data, PCZN-WT had the highest initial activity as well as highest V_max in relation to the mutants at 37 °C therefore we compared the native conformational change in the mutants when the temperature was increased from 4 °C to 37 °C with the WT. The change in conformation at 37 °C denotes the effect of temperature on the structure of the enzyme. We compared the change in the ratio of [θ]₂₀₈ and [θ]₂₂₂ at 4 °C to that of 37 °C. The WT showed increase in the ratio, i.e., there was an overall increase in the helical content at the physiological temperature. This was also seen in E87V and E25V mutants and they do show high enzymatic activity. The initial activity (first 5 min) was less in both the mutants and had an extended lag phase. This could be attributed to the reduced helical content in their native conformation at 4 °C when compared to the WT. At 37 °C there is an increase in helical content in both the mutants and E25V shows highest increase in the ratio. D82V shows a reverse trend where the ratio decreases when switched to 37 °C. This was because there was an overall increase in [θ]₂₂₂ signifying relaxation of the helix. It can be speculated that the rigidity of the helix could be important for higher enzymatic activity. Interestingly, the C141A mutant showed little or no activity despite the increase in ratio of [θ]₂₀₈ and [θ]₂₂₂ (akin to the WT and enzymatically active mutants). This is due to the fact that the mutation directly affects the active site unlike other mutations that affect other parts of the protein that are not involved directly in substrate processing. The other mutations may affect substrate acquisition partially explaining the extended lag phase. Thus, conformational adjustment at physiological temperatures by PCZN plays a critical role in enhanced enzymatic activity.

Fig. 10.

Far-UV circular dichroism of PCZN-WT and compared with that of the mutants at 37 °C. Secondary structure determination at 37 °C of WT and compared with the PCZN mutants (a–e). Ratio of [θ]₂₀₈/[θ]₂₂₂ (minima for α-helical conformation) at 4 °C (native conformation) and compared with that at 37 °C (conformation at physiological temperatures) (f)

Discussion

Protein Sequence Identity and Multiple Sequence Alignments

PCZN domains possess significant similarities in the sequence homology in comparison to several other papain-like cysteine proteases when compared separately by running a BLAST analysis. The pro domain is somewhat variable in homogeneity (85–50%) to several other papain-like cysteine proteases, such as cysteine peptidases from T. rangeli and T. brucei, whereas the catalytic domain shares lower sequence identity (72–65%). However, although some pro parts may differ in length and sequence identity their overall fold appears to be similar, as it has been demonstrated by the resolved X-ray structures[22]. Nonetheless, when PCZN sequence is compared as a pro-protein complex (pro-cat domains together) it shows lower sequence identity as shown in Table 1. Nevertheless, several residues (E25, F32, F35, E49, F56, D82, E86, and E87) are highly conserved within the pro domain sequence of cysteine proteases from homologous parasites. This allowed us to identify several residues as possible candidates for single amino acid substitution. Interestingly, the importance of these residues were confirmed by intra-protein interactions that showed ionic and hydrophobic interactions within the pro domain. CD data confirmed the impact of the mutations on the overall structure of the protein as they all reduced the helical content of the enzyme.

A predicted secondary structure of PCZN (Fig. 8) showed to have more percentage of alpha-helical content (30%), followed by 18% of beta strands and 17% disordered, while the rest remained as coil conformation. From this predicted secondary structure we could determine the localization of the selected residues within the different motifs in the PCZN structure. Residues F32, and F35; E49, and F56; E86, and E87 belong to a stable helix spanning E26–H39; A46–A67; R85–Y92, respectively. Nevertheless, residues E25, E82 were not part of a motif in the secondary structure. The predicted secondary structure of the pro domain region is in accord with the secondary structure represented in a sequence alignment of Cathepsins propeptides, including cruzain and bruzain [23], except for the first alpha helix shown in our prediction. This disagreement may be due to the presence of the 6-his tag at the N-terminus retained in our sequence.

Activation Assays of Recombinant Mutants and Wild-Type PCZN

Although pro domains of several cysteine proteases have shown to be potent inhibitors of other papain-like enzymes [23], no studies have been done to identify critical residues of these pro domains with the final objective of protein early inactivation via non-classical (active-site) methods. In this study, enzymatic activity of recombinant cysteine proteases was assessed by the liberation of the fluorescent leaving group, AMC, from the peptide substrate Z-Phe-Arg-AMC (Fig. 6). All data points were normalized and background corrected using buffer control. Using post non-linear least square regression fitting we calculated the Vmax, half life, rate constant and change in velocity. Change in velocity was shown in the ascending order: C141A = D82V < WT < E87V < E25V. Other parameters such as V_max, half life and rate constant were compared in context of change in velocity of the enzyme. WT had the shortest half life and the rate constant followed by E87V and E25V. V_max was highest for E25V followed by E87V and then WT.

From the observed results, D82V mutation was found to be the most deleterious compared to WT, E25V and E87V. As it has been established that mutation of active site of subtilisin (S22C) does not affect the folding rate[24], we considered mutant C25A (C141A) as negative control for the studies for autoactivation of mutants. Therefore, C141A mutant showed no activity, as anticipated.

Although mutant E25V was proposed as important residue by sequence alignment studies and by folding characterization, it showed a decrement in autoactivation rate during the above-mentioned experimental analysis. Although neither of the mutants was completely inactive, residue mutations have reduced the proteinase activity by delaying the reach of the substrate. Previous studies have shown the effect of mutated pro domains of proteases. That is the case of subtilisin E, where mutations generated within the pro-peptide were incapable of producing an active subtilisin [25]; or in the case of cathepsin L that resulted in improper folding after deletions in its pro domain [26]. Nevertheless, not all mutations produced a negative effect on the activity of proteases as demonstrated in another study done by Ruan et al., 1999 [27] where mutated pro-peptide increases subtilisin’s folding rate when compared to WT pro-peptide. Indeed, our mutants caused a deleterious effect on the rate of PCZN’s autoactivation proving important information for the identification for structural destabilization sites within the pro domain.

Pro-domain mutations can indeed have critical effect on enzyme activity. The nearly complete absence of activity in D82V mutant signifies that it could act as a potential therapeutic target. Other mutations had decreased autoactivation time, but once they became activated, they reached similar V_max compared to the PCZN-WT. D82V activation mimicked the C141A active site mutant which had hardly any activity. The effect of D82V mutation on the enzyme leading to complete lack of activity needs further in-depth study. A multi-set of mutagenesis comprised of a combination of all residues mutated in this study in one plasmid could be additionally helpful for the evaluation of these sidechains in the structural impact of PCZN structure. In addition, mutagenesis of observed hydrophobic cluster within the pro domain compromised of phenilalanines 32, 35 and 56, may be valuable to determine its implication in protein destabilization. Finally, an in-depth characterization of pockets within the crucial residue–residue interactions described in this study via bioinformatics tools, could be very beneficial as contribution of the surrounding residues could be important for the development of anti-chagastics drugs.

Our work describes a novel mechanism by which one could potentially inhibit enzymes. Such an approach first involves the identification of structure-stabilizing interactions in the tertiary structure of the protein. By disrupting such interactions, the contribution of such interactions to the overall stability and enzymatic activity can be assessed. Eventually, the design of small molecules that either disrupt such structure-stabilizing interactions or induce fold arrest can be envisaged. Therefore, this novel mechanism can be described as a bottom-up approach to the design of inhibitors that can block enzyme activity rather than the canonical top-down approach. Our work, involving, the identification and testing of structure-stabilizing sites in cruzain is the first step in this process.