The catalytic significance of the proposed active site residues in Plasmodium falciparum histoaspartic protease Charity L. Parr 1 , Takuji Tanaka 2 , Huogen Xiao 1 and Rickey Y. Yada 1 1 Department of Food Science, University of Guelph, Canada 2 Department of Food and Bioproduct Sciences, University of Saskatchewan, Saskatoon, Canada The aspartic proteases, termed plasmepsins (PMs), produced by the Plasmodium parasite are currently considered attractive targets for new antimalarial drugs given their involvement in hemoglobin degradation [1,2]. The Plasmodium falciparum parasite is known to encode 10 PMs, PMI, -II, -IV–X and histoaspartic protease (HAP) [3], four of which (PMI, PMII, PMIV and HAP) reside within the food vacuole and are involved in human hemoglobin degradation [4]. PMI, -II and -IV are considered ‘classic’ aspartic proteases retaining a catalytic dyad of two aspartic acid residues. HAP, however, does not share this characteristic. Although HAP shares $ 60% sequence identity with PMI, -II and -IV, it is characterized by several substi- tutions to residues generally conserved in aspartic pro- teases [5]. Most notable, the active site Asp34 residue is replaced by His, which eliminates an aspartic acid residue involved in substrate cleavage. In addition, the normally conserved Tyr77 and Val ⁄ Gly78 residues are replaced by Ser and Lys, respectively, in the flap Keywords active site; histoaspartic protease; malaria; model; plasmepsin Correspondence R. Y. Yada, Department of Food Science, University of Guelph, Guelph, Ontario, Canada, N1G 2W1 Fax: +1 519 824 6631 Tel: +1 519 842 4120; ext 58915 E-mail: ryada@uoguelph.ca (Received 16 October 2007, revised 9 January 2008, accepted 7 February 2008) doi:10.1111/j.1742-4658.2008.06325.x Alanine mutations of the proposed catalytically essential residues in his- toaspartic protease (HAP) (H34A, S37A and D214A) were generated to investigate whether: (a) HAP is a serine protease with a catalytic triad of His34, Ser37 and Asp214 [Andreeva N, Bogdanovich P, Kashparov I, Popov M & Stengach M (2004) Proteins 55, 705–710]; or (b) HAP is a novel protease with Asp214 acting as both the acid and the base during substrate catalysis with His34 providing critical stabilization [Bjelic S & Aqvist J (2004) Biochemistry 43, 14521–14528]. Our results indicated that recombinant wild-type HAP, S37A and H34A were capable of autoactiva- tion, whereas D214A was not. The inability of D214A to autoactivate highlighted the importance of Asp214 for catalysis. H34A and S37A mutants hydrolyzed synthetic substrate indicating that neither His34 nor Ser37 was essential for substrate catalysis. Both mutants did, however, have reduced catalytic efficiency (P £ 0.05) compared with wild-type HAP, which was attributed to the stabilizing role of His34 and Ser37 during catalysis. The mature forms of wild-type HAP, H34A and S37A all exhib- ited high activity over a broad pH range of 5.0–8.5 with maximum activity occurring between pH 7.5 and 8.0. Inhibition studies indicated that wild- type HAP, H34A and S37A were strongly inhibited by the serine protease inhibitor phenylmethanesulfonyl fluoride, but only weakly inhibited by pep- statin A. The data, in concert with molecular modeling, suggest a novel mode of catalysis with a single aspartic acid residue performing both the acid and base roles. Abbreviations 2837b, internally quenched fluorescent synthetic peptide substrate EDANS-CO-CH 2 -CH 2 -CO-Ala-Leu-Glu-Arg-Met-Phe-Leu-ser-Phe-Pro-Dap- (DABCYL)-OH; EK, enterokinase; HAP, histoaspartic protease; PM, plasmepsin. 1698 FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS region of HAP [5]. Tyr77 is conserved in all known pepsin-like aspartic proteases and plays a key role in substrate positioning for substrate cleavage [6]. Although HAP contains the alterations of Asp34 to His and Tyr77 to Ser, it is still functional indicating that hydrolysis of substrate by HAP must function by an alternative mode of action [5]. Resolution of a crys- tal structure for HAP would provide considerable insight into the unique functionality of this enzyme. However, no such structure has been reported to date. In the absence of a crystal structure, molecular model- ing has been used to propose possible modes of action for HAP [7,8]. Using homology modeling and molecu- lar dynamics simulations, it has been suggested that HAP may be a serine protease with a catalytic triad of Ser37–His34–Asp214 and an oxyanion hole formed by Ser38 and Asn39 [8]. Alternatively, it has been pro- posed that HAP functions through the direct participa- tion of only Asp214, with His34 providing critical stabilization to the reaction [7]. In this report, we describe the biochemical characterization, supported by molecular modeling, of three mutant HAP proteins, H34A, S37A and D214A, in an effort to identify the importance of residues proposed to be catalytically essential in HAP. Results Autoactivation of histoaspartic protease The H34A, S37A and D214A mutants were investi- gated to determine the importance of proposed catalyt- ically essential residues in HAP. The mutations were chosen based on the proposed mechanisms of action, which suggested that HAP was either a serine protease with a catalytic triad of His34, Ser37 and Asp214 [8], or a novel protease with Asp214 acting alone in sub- strate catalysis [7]. Alanine was chosen as the substitu- tion residue because it eliminates the side chain beyond the b-carbon without imposing any major structural deviations through electrostatic or steric effects [9]. The H34A, S37A and D214A mutants were success- fully introduced into the truncated HAP gene sequence using site-directed mutagenesis. Wild-type HAP and all mutants were expressed in Rosetta Gami Escherichia coli cells. The recombinant wild-type HAP and mutants were partially purified using nickel-affinity and gel-filtration chromatography. At this stage, sam- ples contained predominately fusion protein and a small amount of zymogen protein as determined using N-terminal sequencing. After initial purification samples were tested for autoactivation. SDS ⁄ PAGE analysis revealed that wild-type HAP, H34A and S37A were processed to their expected mature size of 37 kDa [4] in the pH range 5.5–8.0, whereas D214A did not autoactivate (Fig. 1A). The complete activation of fusion wild-type HAP, H34A and S37A protein (confirmed by N-terminal sequencing), as indicated by band-shift, i.e. 62 to 37 kDa, was a relatively slow process, taking 48 h (Fig. 1B). The final cleavage between Lys119p (p denotes prosegment) and Ser120p, confirmed by Edman sequencing (data not shown), yielded the mature enzyme. Activity was assessed during autoacti- vation to ensure that the observed band-shift during activation was related to activity. Activity levels during A a kDa 117 85 f z m f z m 48 34 26 kDa 0 h 1 h 4.5 h 17 h 24 h 48 h 118 90 48 36 27 bc d e B Fig. 1. HAP autoactivation. (A) Coomassie Brilliant Blue-stained 15% SDS ⁄ PAGE gel demonstrating processing of wild-type HAP (lane a), H34A (lane b) and S37A (lane c) after incubation at 37 °C for 48 h at pH 7.0. D214A (lane d) did not exhibit any processing at 37 °C after 48 h. The preincubated sample is shown in lane e. (B) Coomassie Brilliant Blue-stained 15% SDS ⁄ PAGE gel demonstrat- ing the progress of autoactivation of wild-type HAP over 48 h. Fusion protein (10 lg) was incubated at 37 °C in pH 7.0 for 0, 1, 4.5, 17, 24 and 48 h. f, fusion protein; m, mature protein; z, zymo- gen. C. L. Parr et al. Catalytic residues in histoaspartic protease FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS 1699 activation increased with maximal activity being observed at 48 h. Incubation beyond 48 h, under auto- activation conditions, resulted in a loss of activity (Fig. 2). To test whether improper folding of D214A was responsible for its inability to autoactivate, D214A was incubated in the presence of PMII, which has been shown to process wild-type HAP to a mature size (37 kDa) (H. Xiao, unpublished results). Upon incuba- tion at 37 °C for 8 h (HAP to PMII ratio of 50 : 1) wild-type HAP was processed to its expected mature size. Similar results were observed for D214A indicat- ing that the D214A precursor was correctly folded in order for processing by PMII to occur [10]. However, further characterization of the D214A mutant was not conducted because of difficulties in removing PMII from the activation sample. Removal of PMII from the sample, despite being added in a relatively small quantity, was critical because of its ability to more effectively cleave the synthetic substrate than HAP [11,12]. pH of maximal activity, kinetic studies and inhibition To investigate activity characteristics, protein was acti- vated in the presence of enterokinase (EK) (for thiore- doxin tag removal) consistent with the methodology of Xiao et al. [12]. Samples autoactivated in the absence of EK also produced comparable kinetic data (data not shown). The pH profiles of activity for wild-type HAP, H34A and S37A were determined over the pH range 3.5–9.5 using the internally quenched fluorescent synthetic peptide substrate EDANS-CO-CH 2 -CH 2 - CO-Ala-Leu-Glu-Arg-Met-Phe-Leu-ser-Phe-Pro-Dap- (DABCYL)-OH (2837b), as previously described by Istvan and Goldberg [11]. High levels of activity were observed over a broad pH range (5.0–8.0) with maxi- mal activity between pH 7.5 and 8.0 (Fig. 3). Kinetic parameters for wild-type HAP, H34A and S37A were determined at pH 7.5 using 2837b. Analysis using the Michaelis–Menten model (Fig. 4) is sum- marized in Table 1. K m values of 3.42 ± 0.81, 2.40 ± 1.06 and 3.44 ± 1.00 lm for wild-type HAP, H34A and S37A, respectively, were not significantly different (P > 0.05). The measured k cat value for wild- type HAP (3.15 · 10 )3 ± 2.3 · 10 )4 Æs )1 ) was similar to that previously reported for recombinant HAP [12] and was higher (P £ 0.05) than that measured for H34A (9.03 · 10 )4 ± 1.45 · 10 )4 Æs )1 ) and S37A (8.04 · 10 )4 ± 7.6 · 10 )5 Æs )1 ). Inhibition of wild-type HAP, H34A, S37A by phen- ylmethanesulfonyl fluoride is shown in Fig. 5A. All forms of HAP were inhibited by phenylmethanesulfo- nyl fluoride in a dose-dependent manner. Complete inhibition of all forms of HAP was achieved at 1 mm phenylmethanesulfonyl fluoride (data not shown). Interestingly, all forms of HAP were only weakly inhibited by pepstatin A (Fig. 5B). Wild-type HAP and S37A were completely inhibited at 150 lm pepsta- tin A (data not shown), whereas H34A did not exhibit complete inhibition in the concentration range tested. Secondary structure determination The far-UV CD spectra of mature wild-type HAP, H34A and S37A were determined at pH 6.5 and 8.5. Fig. 2. Relative activity measured over the course of autoactiva- tion. Synthetic substrate 2837b was used to assay activity by acti- vation sample (50 n M), in 100 mM sodium acetate (pH 6.5). Each data point represents the mean of two replicates with three deter- minations each and standard deviation. Fig. 3. Determination of pH optimum. Effect of pH on wild-type HAP ( ), H34A (d) and S37A (h) activity in 100 mM sodium ace- tate (pH 3.0–6.5), 100 m M Tris ⁄ HCl (pH 7.0–8.5) and 100 mM sodium carbonate (pH 9.0–9.5). Assays were conducted with 50 n M enzyme and 4 lM peptide substrate 2835b. Each data point represents the mean of three determinations and standard devia- tion. Catalytic residues in histoaspartic protease C. L. Parr et al. 1700 FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS The overall far-UV CD spectra for wild-type HAP, H34A and S37A were similar. This observation was reflected in the predicted secondary structures (Table 2), which were not significantly different (P > 0.05) and indicated that HAP consists of $ 10% a helix, 40% b sheet, 30% turn and 20% random coil. Similar results were reported for the model of HAP (PDB 1QYJ) proposed by Andreeva et al. [8], which predicted a secondary structure content of $ 43% b sheet and $ 9% a helix. Energy minimization calculations To investigate the structural effects of the alanine mutations in the absence of crystallographic data, energy minimization models were generated from a modeled structure of HAP (PDB 1QYJ) (Fig. 6) [8]. Figure 6A shows the predicted orientation of active residues and water molecules. In the energy-mini- mized models of H34A (Fig. 6b), the orientations of Asp214 and Water2 are altered in a way that increases the distance between the two. A similar ori- entation changed was observed in the S37A mutant (Fig. 6C), as well as a change in the orientation of His34. Discussion Recombinant wild-type HAP, H34A, S37A and D214A proteins were successfully expressed as a 62 kDa thioredoxin fusion protein. In the pH range 5.5–8.0, wild-type HAP, H34A and S37A were all pro- cessed between Lys119p and Ser120p to their expected mature size. This cleavage site is four amino acids upstream of the native cleavage site [13]. A similar finding was reported for both PMI and PMII, where it was demonstrated that the autoactivation cleavage site was 7 or 12 amino acids upstream of the native cleav- age site [14,15]. This discrepancy in activation sites likely indicates that autoactivation is not the predomi- nant mode of activation in vivo. The latter was con- firmed by Banerjee et al. [13] who suggested that the plasmepsins were activated by an acidic convertase and not through autoactivation in vivo. The differences in cleavage site likely reflect the differences in specificity between HAP and the acidic convertase protein. Inter- estingly, D214A did not exhibit such processing. The ability of the H34A and S37A mutants to autoactivate suggests that neither His34 nor Ser37 is essential to HAP functionality. Thus, HAP is not a serine protease with a catalytic triad of His34, Ser37 and Asp214. Alternatively, the inability of the D214A to autoacti- vate suggests that Asp214 is required for HAP func- tionality, as suggested by Bjelic and Aqvist [7]. Because D214A was unable to be processed though autoactivation it could not be further kinetically char- acterized. Fig. 4. Kinetic characterization of wild-type HAP and mutants. Michaelis–Menten kinetic analysis for wild-type HAP (A), H34A (B) and S37A (C) using 2837b. Kinetic assays were completed over the substrate range 0.1–12 l M using 50 nM enzyme in 100 mM sodium acetate (pH 6.5) with 10% glycerol. (Inset) Hydrolysis of substrate 2837b. The assay was conducted with 50 n M enzyme and 0.1 lM peptide substrate 2837b in 100 mM sodium acetate (pH 7.5). C. L. Parr et al. Catalytic residues in histoaspartic protease FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS 1701 Kinetic characterization and molecular modeling of both the H34A and S37A mutants suggests that although His34 and Ser37 were not essential to HAP functionality they may play a stabilizing role in the catalytic process. The molecular model of H34A and S37A mutants were calculated based on a model of HAP (PDB 1QJY) [8]. Figure 6 shows the models around the active site. Figure 6A–C represents wild- type HAP, H34A and S37A, respectively. As seen in Fig. 6A, the position of the catalytic Asp214 residue is fixed with the interaction to (a) the e-amine of Lys78 and (b) the imidazole ring of His34. Asp214 also inter- acts with a water molecule (Water2) and this water further forms a hydrogen bond network to Lys78 via another water molecule (Water3). Water2 may act as a medium in the catalytic reaction. The position of Lys78 is anchored with an interaction between the main chain oxygen of Ala216 and the e-amine of Lys78. In H34A mutant (Fig. 6B), the interaction between the imidazole ring and Asp214 disappeared and the side chain of Asp214 is slightly relocated, moving Water3 away from Lys78. This displacement diminishes the interactions (His34–Asp214 and Asp214–Water2–Water3–Lys78) that fixed the position of Asp214 observed in the wild-type HAP model. A similar displacement was observed in S37A model (Fig. 6C). In this mutant, the substitution of Ser37 with Ala removed the interaction between residue 37 and His34. Loss of the Ser37–His34 interaction results in repositioning of His34 and, in turn, an altered posi- tion of Asp214. Because Water2 is speculated as the catalytic medium, these shifts could explain why the k cat values were smaller in the H34A and S37A mutants than in wild-type HAP (Table 1). It is important to note that the differences in the measured kinetic parameters for H34A and S37A com- pared with wild-type HAP cannot be attributed to an overall change in secondary structure content. Wild- type HAP, H34A and S37A had similar secondary structures, i.e. $ 10% a helix, 40% b sheet, 30% turn and 20% random coil (Table 2). These results are consistent with the reported secondary structures of pepsin-like aspartic proteases which are predominately b sheet [16]. On the basis of the kinetic behavior of the H34A and S37A mutants and the inability of D214A to auto- activate, Asp214 was identified as the only catalytically essential residue studied in this investigation. The iden- tification of Asp214 as a catalytically essential residue Table 1. Summary of measured kinetic parameters for wild-type HAP, H34A and S37A, otained using least-squares for best fit to the Micha- elis–Menten model. All values were obtained by averaging analysis of two replicates with three determinations per replicate (n = 6). Means sharing the same letter are not significantly different (P > 0.05). Protein K m (lM) k cat (s )1 ) k cat ⁄ K m (lM )1 Æs )1 ) Wild-type 3.42 ± 0.81a 3.15 · 10 )3 ± 2.30 · 10 )4 b 9.40 · 10 )3 ± 1.80 · 10 )3 d H34A 2.40 ± 1.06a 9.03 · 10 )4 ± 1.45 · 10 )4 c 4.20 · 10 )3 ± 1.40 · 10 )3 e S37A 3.44 ± 1.00a 8.04 · 10 )4 ± 7.60 · 10 )5 c 2.12 · 10 )3 ± 5.80 · 10 )4 e Fig. 5. Effect of protease activity on substrate degradation by wild- type HAP, H34A and S37A. (A) Inhibition of wild-type HAP ( ), H34A (.) and S37A ( ) by phenylmethanesulfonyl fluoride (PMSF). (B) Inhibition wild-type HAP ( ), H34A (.) and S37A ( ) by pepsta- tin A. Assays were carried out in 100 m M sodium acetate (pH 6.5), with 4 l M substrate and 50 nM enzyme. Each data point represents the mean of two replicates with three determinations each and standard deviation. Catalytic residues in histoaspartic protease C. L. Parr et al. 1702 FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS is consistent with the proposed molecular mechanism of Bjelic and Aqvist [7], who suggested that substrate cleavage by HAP was achieved by Asp214, which acted as both the acid and the base. Bjelic and Aquist [7], however, also suggested that the positive charge on His34 provided critical stabilization (by a factor of $ 10 000) to the nucleophilic hydroxyl and developing negative charge on the substrate during catalysis. The proposed critical significance of His34 is, however, not consistent with the observation that H34A exhibited only a fourfold reduction in turnover number com- pared with wild-type HAP, although a fourfold change is not trivial (Table 1). It is possible that the hydroxyl and the negative charge that develops on the substrate as a result of catalysis do not require the degree of sta- bilization as suggested by Bjelic and Aquist [7]. Alter- natively, it is conceivable that another positively charged amino acid, namely Lys78 of the flap, stabi- lizes the reaction through its contribution of a positive charge to the active site region. For most aspartic pro- Table 2. Predicted secondary structure composition of wild-type HAP, H34A and S37A. All values were obtained by averaging analysis of two replicates with four scans per replicate. Means sharing the same letter are not significantly different (P > 0.05). Protein Secondary structure element a helix (%) b sheet (%) Turn (%) Random coil (%) Wild-type HAP 8.5 ± 1.6a 37.6 ± 2.1b 31.8 ± 1.7d 20.5 ± 2.5c H34A 9.7 ± 2.4a 37.7 ± 3.7b 31.8 ± 1.5d 20.4 ± 2.0c S37A 8.3 ± 2.1a 36.1 ± 2.1b 31.0 ± 3.6d 20.2 ± 2.9c Fig. 6. Energy minimized models: (A) wild- type HAP, (B) H34A, (C) S37A. Models were rendered using SWISS-PDB VIEWER. C. L. Parr et al. Catalytic residues in histoaspartic protease FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS 1703 teases, substrate binding results in the flap moving to a closed position [17]. It has been proposed that such movements are conserved in HAP [8] placing Lys78 in a position that can compensate for the loss of His34 which allows for the interaction with the lytic hydroxyl in the active site. Interestingly, residue 78 in most aspartic proteases is a conserved Gly residue [5], therefore, it is possible that its alteration to Lys is of functional significance. Wild-type HAP, H34A and S37A exhibited a broad pH of maximal activity (5.0–8.5) (Fig. 3). Activity in the basic pH range would also suggest that the positive charge on His34 is not critical for HAP functionality [7] because this residue would likely be neutral at pH 8.0 given that free histidine has a free pK a of $ 6.5 [7]. The neutral state His34 would, therefore, not be able to contribute a positive charge to substrate catalysis. Based on our results, a mechanism for substrate catalysis is proposed in Fig. 7. With the substrate in position, a precisely positioned water molecule (Water2) that is oriented by a hydrogen bond network (His34–Asp214–Water2–Water3–Lys78) donates a pro- ton to Asp214. The free hydroxyl group is momentar- ily stabilized by the positive charges on His34 and ⁄ or Lys78 and then attacks the scissile bond of the sub- strate. Concomitantly, Asp214 donates a proton, ulti- mately breaking the peptide bond and returning the active site to its original state. Inhibition assays showed that HAP, H34A and S37A were effectively inhibited by the serine protease inhibitor phenylmethanesulfonyl fluoride (Fig. 5A). Recombinant HAP, H34A and S37A were only weakly inhibited by pepstatin A (Fig. 5B). It is unclear why HAP, unlike other aspartic proteases, is highly sensitive to phenylmethanesulfonyl fluoride [4]. It may be that phenylmethanesulfonyl fluoride modi- fies a second serine (Ser38) in the active site region. Such an interaction may form a phenylmethanesulfo- nyl fluoride-transition state analog which would steri- cally block access to the active site. It has also been proposed that phenylmethanesulfonyl fluoride may interact with the flap region of HAP [4]. Unlike other aspartic proteinases, a serine residue replaces the nor- mally conserved tyrosine residue on the tip of the flap. It is possible that phenylmethanesulfonyl fluoride interacts with this usual serine residue to prevent sub- strate access to the active site [4]. Alternatively, an interaction may result from a non-covalent interaction between phenylmethanesulfonyl fluoride and the HAP active site [4]. Multiple weak interactions may com- bine to generate a strong interaction between HAP and phenylmethanesulfonyl fluoride. The development of a crystal structure of HAP bound to the phenylmethanesulfonyl fluoride inhibitor will provide insight into this unusual interaction. It is also unclear why recombinant HAP and the mutants were only slightly inhibited by pepstatin A. The crystal structure of pepstatin A bound to PMII (PDB 1W6I) reveals that the first statyl of the inhibitor occupies an area generally occupied by the catalytic water. The inhibitor is stabilized by two key hydrogen bonds, one from each of the active aspartic acid resi- dues. It is possible that recombinant HAP is less sensi- tive to pepstatin A as a result of the replacement of the normally conserved aspartic acid residue with a histidine residue. The alteration may affect the inhibitor ⁄ enzyme interaction and reduce the stability 3 4 2 1 Fig. 7. Formula representation of the proposed four step catalytic reaction for wild-type HAP in the acidic range. W2 and W3 repre- sent water molecules, Water2 and Water3. Catalytic residues in histoaspartic protease C. L. Parr et al. 1704 FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS of pepstatin A binding. The inhibition of wild-type HAP was higher than those observed for both H34A and S37A (Fig. 5B). This may be supportive evidence that mutation of His34 and Ser37 to alanine disrupts a hydrogen-bonding network that is critical for proper positioning of the Asp214 residue. Any disruption of the position of Asp214 would likely alter an important interaction between Asp214 and pepstatin A. In conclusion, a study was conducted to investigate residues previously proposed by other research groups as essential to catalytic activity. The model reported by Bjelic and Aquist [7], in concert with the biochemical data and molecular models presented in this study, support an alternative mode of catalysis with a single aspartic acid residue performing both the acid and base roles. However, based on the kinetic parameters for H34A, His34 was not critical for stabilizing cataly- sis though its positive charge as suggested by Bjelic and Aquist [7]. The kinetic parameters of H34A and S37A suggest that His34 and Ser37 may alternatively both play a role in stabilizing the active site for sub- strate catalysis through an aspartic protease like hydrogen-bond network. Experimental procedures Materials The pET32b(+) plasmid and E. coli Rosetta-gami B (DE3)pLysS cells were purchased from Novagen (Mississa- uga, Canada). GenEluteÔ Plasmid Miniprep Kit was pur- chased from Sigma-Aldrich Co. (St Louis, MO, USA). Pfu DNA polymerase was obtained from Fermentas Life Sci- ences (Burlington, Canada) and mutagenic primers were synthesized by Sigma Genosys (Oakville, Canada). QIA- quick Ò PCR Purification Kit was purchased from Qiagen Sciences (Germantown, MD, USA). HIS Ò -Select 6.4 mL cartridges were obtained from Sigma-Aldrich. Two milliliter YM50 Centricon Centrifugal Filter Units were supplied by Millipore Corp. (Bedford, MA, USA). All chemicals and media were obtained from Fisher Scientific (Nepean, Can- ada) or Sigma-Aldrich. Mutagenesis The HAP gene with the 70 N-terminal most amino acids of the prosegment removed cloned into pET32b(+) (pET32btHAP) was obtained from our laboratory [12]. Mutants were generated using the Quick-Change Site Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA) according to the manufacturer’s instructions. The following primers were designed to introduce the single point muta- tions at residues His34, Ser37 or Asp214 (numbering according to the native mature form of HAP) of wild-type HAP in the pET32btHAP plasmid: H34A sense 5¢-CAAAA ATTTAATTTCTTATTCGCTACAGCTTCATCTAATG-3¢; H34A antisense 5¢-CATTAGATGAAGCTGTAGCGAATA AGAAATTAAATTTTTG-3¢; D214A sense 5¢-GCAAAC GTTATTTTAGCTAGTGCCACCAGTGTCATAACTG-3¢; D214A antisense 5¢-CAGTTATGACACTGGTGGCACTA GCTAAAATAACGTTTGC-3¢;S37AsenseFwd5¢-CAAA AATTTAATTTCTTATTCCATACAGCTGCATCTAATG- 3¢; S37A antisense 5¢-CATTAGATGCAGCTGTATGGAA TAAGAAATTAAATTTTTG-3¢. The sequences were con- firmed with the T7 promoter and T7 terminator primers and a gene specific primer at the Guelph Molecular Supercentre (Guelph, Canada) using dye terminator cycle sequencing on an ABI PRISM model (Applied Biosystems, Foster City, CA, USA). Expression of fusion wild-type and mutant HAP protein Expression of all recombinant HAP fusion proteins was conducted according to the method described by Xiao et al. [12]. The expression constructs were transformed into E. coli Rosetta gami B (DE3)pLysS cells. Cells were cul- tured in 1.0 L Luria–Bertani media containing 15 lgÆmL )1 kanamycin, 34 lgÆmL )1 chloramphenicol, 12.5 lgÆmL )1 tet- racycline and 50 lgÆmL )1 ampicillin to a A 600 of 1.0 and then induced with isopropyl b-d-thiogalactopyranoside. After expression, cells were collected by centrifugation (2500 g for 15 min). Purification of the fusion protein and activation Cell pellets were resuspended in 50 mL 1· BugBuster (Nov- agen, Madison, WI, USA), pH 7.5 and incubated at room temperature for 1 h with gentle shaking. The sample was then centrifuged at 16 000 g for 20 min at 4 °C. The super- natant was applied to a HIS Ò -Select Cartridge (Sigma- Aldrich, Oakville, Canada) on an AKTAÔ FPLC system (GE Healthcare, Chalfont St Giles, UK). The column was washed with 50 mm sodium phosphate ⁄ 0.3 m NaCl ⁄ 10 mm imidazole (pH 7.5) wash buffer and a gradient of 0–10% 50 mm sodium phosphate ⁄ 0.3 m NaCl ⁄ 250 mm imidazole (pH 7.5) buffer was applied to the column over eight col- umn volumes. Recombinant thioredoxin fusion HAP pro- tein was eluted with 50 mm sodium phosphate ⁄ 0.3 m NaCl ⁄ 250 mm imidazole (pH 7.5) buffer. The sample was concentrated in 50 mm sodium phosphate (pH 7.5) contain- ing 0.2% Chaps. The concentrated sample was then applied to a SuperoseÔ 12 10 ⁄ 300 GL column (GE Healthcare). After separation, HAP containing fractions were concen- trated in 50 mm Mes buffer (pH 6.5) containing 0.2% Chaps. Samples were incubated at 37 °C for 48 h with EK (Sigma-Aldrich, Oakville, Canada) (1 : 20 EK ⁄ HAP) to C. L. Parr et al. Catalytic residues in histoaspartic protease FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS 1705 allow activation. After activation the sample was applied to a SuperoseÔ 12 10 ⁄ 300 GL column (GE Healthcare) to recover pure HAP protein. Protein sample was finally washed in 50 mm Mes (pH 6.5) and stored at 4 °C. The identity of proteins present prior to activation and resulting from activation were N-terminally sequenced by Edman sequencing (The Hospital for Sick Children, Toronto, Canada). Protein concentration determinations Protein concentration was determined using the Bio-Rad protein assay (Hercules, CA, USA). Standard dilutions of commercial BSA were used to generate a standard curve. pH of maximal activity, enzyme kinetic assays and inhibition studies The pH of maximal activity for wild-type HAP and the mutants was determined using 2837b (AnaSpec Inc, San Jose, CA, USA) [10] over a pH range of 3.5–9.5 in buffers 100 mm sodium acetate (pH 3.5–6.5), 100 mm Tris ⁄ HCl (pH 7.0–8.5) and 100 mm sodium carbonate (pH 9.5). All assays were conducted with 50 nm of HAP and 4 lm sub- strate 2837b [12]. Inhibition assays were conducted using the aspartic protease inhibitor pepstatin A (20–150 lm) and the serine protease inhibitor phenylmethanesulfonyl fluoride (1– 1000 lm). The reaction was carried out in 100 mm sodium acetate (pH 6.5) using 50 nm HAP and 4 lm substrate 2837b. Kinetic parameters were determined using substrate 2837b [11]. Progress curves were followed over 30 min to ensure that initial velocity was measured. In addition, dif- ferent enzyme concentrations were initially evaluated to ensure that activity increased linearly with increasing enzyme concentration. The reactions were carried out using 50 nm HAP and 0.1–12 lm substrate in 100 mm Tris ⁄ HCl (pH 7.5). The assays were conducted using a Victor 2 1420 multilabel counter (Perkin–Elmer, Woodbridge, Canada) with excitation at 335 nm and emission at 500 nm. The measured fluorescence was converted to moles per second using a conversion factor (3 714 000) derived from a stan- dard curve for the complete digestion of the substrate by Saccharomyces cerevisiae proteinase A (Sigma-Aldrich, Oakville, Canada) [12,14]. To generate these data, a range of substrate concentrations (0.1, 1, 2 and 5 lm) were uti- lized to allow for hydrolysis to proceed to completion as indicated by a plateau on progress curve. Data were ana- lyzed through linear regression (R 2 = 0.996) to generate a standard curve. Nonlinear regression with Michaelis–Men- ten model was used to determine K m and k cat was calcu- lated from k cat = V max ⁄ [E]. Each sample was done in triplicate, with the experiment repeated twice (n = 6) [12]. Autoactivation of wild-type HAP and mutants Samples were incubated over the pH range 3.0–8.0 without the addition of exogenous protein. Aliquots were taken at predetermined time points and run on SDS ⁄ PAGE to visu- alize band shift. In addition, activity was assayed over 48 h using 50 nm enzyme and 4 lm substrate 2837b [12]. Far-UV CD spectropolarimetry Far-UV CD spectra were determined from 250 to 185 nm at room temperature using a 100 lL quartz cuvette with a 0.1 cm path length. A Jasco J-810 spectropolarimeter (Jas- co, Tokyo, Japan) was used to determine the spectra. CD scans were performed in triplicate with the average buffer spectra being subtracted from the sample spectra. The wild- type and mutant enzymes were scanned in 100 mm sodium phosphate (pH 6.5 and 8.5). All samples were filtered prior to measurement. Ellipticity values (mdeg) were recorded as a function of wavelength. Spectra were expressed as mean residue ellipticity (degrees cm 2 Ædmol )1 ) using the following equation: ½h MRWk ¼ðMRW  h k Þ=ð10  d  cÞ where MRW is mean residue weight (110), h k is the mea- sured ellipticity at a particular wavelength (mdeg), d is the pathlength (cm) and c is the concentration of enzyme (gÆmL )1 ). CD results were analyzed using Dichroweb (http:// www.cryst.bbk.ac.uk/cd web/html/ home.html), an online CD analysis tool [18,19]. Three analysis programs (sel- con3, contin and cdsstr) and two data sets (4 and 7) con- taining 43 and 48 proteins, respectively, chosen in accordance with the range of input data, were used to determine secondary structure. The average of four scans was used to input into the above programs. Molecular modeling Molecular models were calculated using namd software package [20] running on a Macintosh G4 computer. Initial mutant models were constructed based on the crystallo- graphic structure of HAP (PDB 1QYJ) [8]. The models for the mutants as well as wild-type were placed in a water sphere and energy-minimized using topology force field data provided with the namd package. The calculation was carried out with a 15 angstrom cut-off distance for 10 000 iterations. Statistical analysis graphpad (http://www.graphpad.com) ANOVA and Tukey post-test analysis were used to test the statistical signifi- cance of the data. Catalytic residues in histoaspartic protease C. L. Parr et al. 1706 FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS Acknowledgements The authors would like to thank Dr Eric Brown, Department of Biochemistry and Biomedical Sciences, McMaster University, for his critical reading of the manuscript. The financial support of the Natural Sci- ences and Engineering Research Council of Canada and the Canada Research Chairs Program is gratefully acknowledged. References 1 Bremen J (2001) The ears of the hippopotamus: mani- festations, determinants and estimates of the malaria burden. Am J Trop Med Hyg 64S, 1–11. 2 Olliaro P & Goldberg D (1995) The Plasmodium diges- tive vacuole: metabolic headquarters and choice drug target. Parasitol Today 11, 294–297. 3 Coombs GH, Goldberg DE, Klemba M, Berry C, Kay J & Mottram JC (2001) Aspartic protease of Plasmo- dium falciparum and other parasitic protozoa as drug targets. Trends Parasitol 17, 532–537. 4 Banerjee R, Liu J, Beatty W, Pelosof L, Klemba M & Goldberg DE (2002) Four plasmepsins are active in the Plasmodium falciparum food vacuole, including a prote- ase with an active-site histidine. Proc Natl Acad Sci USA 99, 990–995. 5 Berry C, Humphreys MJ, Matharu P, Granger R, Horrocks P, Moon RP, Certa U, Ridley RG, Bur D & Kay J (1999) A distinct member of the aspartic proteinase gene family from the human malaria parasite Plasmodium falciparum. FEBS Lett 447, 149–154. 6 Tang J (1971) Specific and irreversible inactivation of pepsin by substrate-like epoxides. J Biol Chem 246, 4510–4517. 7 Bjelic S & Aqvist J (2004) Computational prediction of structure, substrate binding mode, mechanism and rate for a malaria protease with a novel type of active site. Biochemistry 43, 14521–14528. 8 Andreeva N, Bogdanovich P, Kashparov I, Popov M & Stengach M (2004) Is histoaspartic protease a serine protease with a pepsin-like fold? Proteins 55, 705–710. 9 Cunningham BC & Wells JA (1989) High-resolution epitope mapping of hGH–receptor interactions by ala- nine-scanning mutagenesis. Science 244, 1081–1085. 10 van den Hazel HB, Kielland-Brandt MC & Winther JR (1992) Autoactivation of proteinase A initiates activa- tion of yeast vacuolar zymogens. Eur J Biochem 207, 277–283. 11 Istvan ES & Goldberg DE (2005) Distal substrate inter- actions enhance plasmepsin activity. J Biol Chem 280, 6890–6896. 12 Xiao H, Sinkovits AF, Bryksa BC, Ogawa M & Yada RY (2006) Recombinant expression and partial charac- terization of an active soluble histo-aspartic protease from Plasmodium falciparum. Protein Expr Purif 49, 88– 94. 13 Banerjee R, Francis SE & Goldberg DE (2003) Food vacuole plasmepsins are processed at a conserved site by an acidic convertase activity in Plasmodium falcipa- rum. Mol Biochem Parasitol 129, 157–165. 14 Xiao H, Tanaka T, Ogawa M & Yada RY (2007) Expression and enzymatic characterization of the solu- ble recombinant plasmepsin I from Plasmodium falcipa- rum. Protein Eng Des Sel 20, 625–633. 15 Tyas L, Gluzman I, Moon RP, Rupp K, Westling J, Ridley RG, Kaya J, Goldberg DE & Berry C (1999) Naturally-occurring and recombinant forms of the aspartic proteinases plasmepsins I and II from the human malaria parasite Plasmodium falciparum. FEBS Lett 454, 210–214. 16 Dunn B (2002) Structure and mechanism of the pepsin- like family of aspartic peptidases. Chem Rev 102, 4431– 4458. 17 Tang J & Koelsch G (1995) A possible function of the flaps of aspartic proteases: the capture of substrate side chains determines the specificity of cleavage positions. Protein Peptide Lett 2, 257. 18 Whitmore L & Wallace BA (2004) DICHROWEB, an online server for protein secondary structure analyses from circular dichroism spectroscopic data. Nucleic Acids Res 32, W668–W673. 19 Lobley A, Whitmore L & Wallace BA (2002) DI- CHROWEB: an interactive website for the analysis of protein secondary structure from circular dichroism spectra. Bioinformatics 18, 211–212. 20 Kale L, Skeel R, Bhandarkar M, Brunner R, Gursoy A, Krawetz N, Philips J, Shinozaki A, Varadarajan K & Schulten K (1999) NAMD2: greater scalability for parallel molecular dynamics. J Comp Physiol 151, 283– 312. C. L. Parr et al. Catalytic residues in histoaspartic protease FEBS Journal 275 (2008) 1698–1707 ª 2008 The Authors Journal compilation ª 2008 FEBS 1707 . The catalytic significance of the proposed active site residues in Plasmodium falciparum histoaspartic protease Charity L. Parr 1 ,. with an interaction between the main chain oxygen of Ala216 and the e-amine of Lys78. In H34A mutant (Fig. 6B), the interaction between the imidazole ring