Malaria is one of the world's most prevalent parasitic diseases, with over 200 million cases annually. Alarmingly, the spread of drug-resistant parasites threatens the effectiveness of current antimalarials and has made the development of novel therapeutic strategies a global health priority. Malaria parasites have a complicated lifecycle, involving an asymptomatic ‘liver stage’ and a symptomatic ‘blood stage’. During the blood stage, the parasites utilise a proteolytic cascade to digest host hemoglobin, which produces free amino acids absolutely necessary for parasite growth and reproduction. The enzymes required for hemoglobin digestion are therefore attractive therapeutic targets. The final step of the cascade is catalyzed by several metalloaminopeptidases, including aminopeptidase P (APP). We developed a novel platform to examine the substrate fingerprint of APP from Plasmodium falciparum (PfAPP) and to show that it can catalyze the removal of any residue immediately prior to a proline. Further, we have determined the crystal structure of PfAPP and present the first examination of the 3D structure of this essential malarial enzyme. Together, these analyses provide insights into potential mechanisms of inhibition that could be used to develop novel antimalarial therapeutics.

Introduction

In October 2015, the Nobel Prize for medicine was awarded to Youyou Tu for the development of the antimalarial drug, artemisinin. Artemisinin and its derivatives (collectively referred to as the ARTs) are currently the leading antimalarial treatment, replacing chloroquine and quinine, which, due to drug resistance, have become ineffective. The ARTs have undoubtedly saved millions of lives worldwide and in combination with improved diagnostic testing and vector control interventions, have reduced the global death toll by 60% since 2000, to 438 000 in 2015 [1]. However, resistance of Plasmodium falciparum (Pf) to the ARTs has emerged in the Greater Mekong subregion, including Cambodia, Thailand, Myanmar, Vietnam, and Laos [2] and is expected to spread. We therefore urgently require novel antimalarial agents that work via different targets.

Peptide recycling, the process by which cellular proteins are broken down to single amino acid residues, is critical to parasite survival. In blood-stage malaria parasites, two major processes are responsible for peptide turnover: proteasomal (within the cytosol) [36] and vacuolar (in the specialized digestive food vacuole) [7,8]. The vacuolar pathway is responsible for the digestion of 60–80% of host cell hemoglobin, which is imported into the digestive vacuole and degraded into free amino acids [811]. This process is absolutely necessary for parasite growth and development [12,13]. The final step of peptide turnover, the removal of N-terminal amino acids from short polypeptide chains, is catalyzed by a panel of aminopeptidases, which work in concert according to different substrate specificities, to complete protein digestion [1416]. Four metalloaminopeptidases (MAPs) have been investigated for their role in peptide turnover in Pf parasites: leucyl aminopeptidase (PfA-M17) [14,17], alanyl aminopeptidase (PfA-M1) [14,18,19], aspartyl aminopeptidase (PfM18AAP) [14,20,21], and aminopeptidase P (PfAPP) [14,22]. Parasites with a disrupted PfA-M1, PfA-M17, or PfAPP gene could not be isolated, suggesting that they are each essential to parasite survival, while disruption of the PfM18AAP gene results in a fitness cost [14]. The essential role of the MAPs in parasite survival makes them attractive targets for drug discovery. Indeed, both PfA-M1 and PfA-M17 have been validated as novel antimalarial targets [15], and determination of their crystal structures has allowed them to become the focus of a series of structure-guided drug discovery campaigns [2327].

X-Prolyl aminopeptidases (aminopeptidase P, APP) are MAPs that catalyze the removal of N-terminal residues from peptide substrates with a proline at the P1′ position [28]. They are present in a range of organisms, including mammals, bacteria, and protozoa. While mammals, including humans, possess three isoforms of APP [2932], only one, the essential PfAPP, has been identified in malaria [14,22]. Within parasites, PfAPP has been localized to the cytosol, where it likely plays a role in the turnover of ubiquitinated cellular proteins, as well as to the digestive food vacuole, where a key role in hemoglobin digestion has been proposed [14,22]. The mature PfAPP has been characterized as a 157 kDa homodimer, which is proposed to have been formed by the removal of a 120 residue N-terminal extension absent from mammalian forms of the enzyme [22]. Sequence alignments suggest that the 121–777 residue mature form adopts a three-domain arrangement similar to the human APP1 (hAPP1), XPNPEP1, as opposed to the two-domain organization common to all other known structures of X-prolyl aminopeptidases and prolidases [6,22].

Herein, we present the design and synthesis of a library composed of 20 fluorogenic substrates, which we used to determine the substrate fingerprint of mature PfAPP (PfAPP, residues 121–777). The profile shows that PfAPP has the capacity to catalyze the removal of any N-terminal amino acid residue from peptides with a P1′ proline, and that the other MAPs in P. falciparum are unable to perform this function. Further, we determined the X-ray crystal structure of PfAPP both unliganded and in complex with apstatin, a known APP inhibitor. The structure showed a homodimer associated by an extensive interface between the catalytic domains (domain III) of both monomers. The active site is internally located at the junction of the three domains and shows a di-metal coordination consistent with the presence of two catalytic manganese ions. The structure also allowed us to examine the active site characteristics that confer PfAPP's broad substrate tolerance and provides insights into potential of targeting PfAPP as a novel antimalarial therapeutic strategy.

Materials and methods

Protein expression and purification

DNA coding for the previously described [22] mature form of PfAPP (residues 121–777) with a C-terminal His6 tag was chemically synthesized by DNA2.0 using codons optimized for gene expression in Escherichia coli and provided in the pJ404 vector for expression. The construct was transformed into BL21(DE3) cells and grown at 37°C to an OD of 0.8 before protein expression was induced with 1 mM isopropyl β-d-thiogalactopyranoside for 17 h at 20°C. Cells were lysed by sonication in PBS, pH 8.0, 300 mM NaCl, 5% glycerol, 20 mM imidazole, 0.01% Triton X-100. Clarified lysates were bound to a Ni-NTA-agarose column in nickel-affinity buffer (PBS, pH 8.0, 300 mM NaCl, 5% glycerol, 20 mM imidazole) and eluted in nickel-affinity buffer supplemented with 500 mM imidazole. The protein was further purified by size-exclusion chromatography on a Superdex 200 16/60 using an AKTAxpress high-throughput chromatography system (http://proteinexpress.med.monash.edu.au/index.htm) in 50 mM Tris–HCl, pH 8.0, 50 mM NaCl. The major peak, containing dimeric PfAPP, was pooled and concentrated to 10 mg/ml.

A construct of a prolyl iminopeptidase from Xylella fastidiosa (Xf PIP) was kindly provided to us for use in our coupled enzyme assays. Xf PIP was transformed into BL21(DE3) cells, and expressed and purified by nickel-affinity chromatography, with the same method as described for PfAPP earlier. The final size-exclusion chromatography step was performed on a Superdex 200 16/60 using an AKTAxpress high-throughput chromatography system (http://proteinexpress.med.monash.edu.au/index.htm) in 50 mM HEPES, pH 8.0, 300 mM NaCl, 5% glycerol. The major peak, containing Xf PIP, was pooled and concentrated to 1 mg/ml.

Synthesis of X-Pro-ACC library

The synthesis of the library was performed on a solid support according to the solid phase peptide synthesis method described in the literature and depicted in Scheme 1 [33,34]. In the first step, fluorescent-leaving group Fmoc-ACC (7-amino-4-carbamoylmethylcoumarin)-OH was attached to the Rink Amide resin using coupling reagents (HOBt/DICI in DMF). Then, Fmoc-protecting group was removed with 20% piperidine in DMF. In the next step, Fmoc-Pro-OH was attached with HATU/collidine. After Fmoc group deprotection, the resin was washed with dichloromethane and methanol, and dried over P2O5. Then, dried resin was divided into 20 portions. To each portion of the NH-Pro-ACC resin, one of 19 natural amino acids and norleucine was attached using the same coupling reagents as for Fmoc-ACC-OH. The amino-protecting group (Fmoc) was cleaved with 20% piperidine in DMF. In the last step, the library was cleaved from the resin with TFA:TIPS:H2O (95%:2.5%:2.5%) and precipitated in diethyl ether. Each ACC-labeled substrate was purified using reverse-phase HPLC, lyophilized and dissolved in DMSO at a concentration of 10 mM and stored at −20°C until use. The chemical structures of each of the compounds in the library are provided as Supplementary Figure S1.

Scheme 1.

NH2-X-Pro-ACC library synthesis (X — 19 natural amino acids and norleucine).

Scheme 1.

NH2-X-Pro-ACC library synthesis (X — 19 natural amino acids and norleucine).

Enzymatic analysis

Aminopeptidase activity was determined by measuring the hydrolysis of the commercially available quenched fluorogenic substrate Lys(Abz)-Pro-Pro-NA (Bachem) [35]. Reactions were continuously monitored in 384-well microtitre plates, 50 μl total volume at 37°C using a spectrofluorimeter (BMG FLUOstar) with excitation at 355 nm and emission at 460 nm. Enzyme was first added to 100 mM Tris–HCl, pH 7.5, supplemented with 1 mM MnCl2 for 10 min prior to the addition of substrate. The assays were supplemented with manganese and performed at pH 7.5, based on previous reports of recombinant PfAPP activity [22]. Initial rates were obtained at 37°C over a range of substrate concentrations spanning Km (0.5–500 μM) and at fixed enzyme concentration (5 nM) in Tris–HCl, pH 7.5, supplemented with 1 mM MnCl2. Calculations (Km, kcat) were performed using GraphPad Prism.

The activity of PfAPP against the X-Pro-ACC substrate library was determined by a coupled assay using Xf PIP and measuring the release of the fluorogenic-leaving group, ACC, from the peptide substrate. The reactions were performed (n = 3) in 96-well microplates at 37°C, with a final assay volume of 100 µl. Assay conditions were carefully optimized to ensure that Xf PIP did not cleave the X-Pro bond (Supplementary Figure S2) and to ensure that XfPIP was not limiting in the reaction. In setting up the assay, we noted some residual activity of Xf PIP alone cleaving Trp-Pro-ACC and then established that omission of the metal ion from the assay buffer resulted in cleaving only Pro-ACC, and not X-Pro-ACC. The reactions were initiated by PfAPP (250 nM) and Xf PIP (500 nM) to the peptide substrate (final concentration 10 µM) and monitored with excitation at 355 nm and emission at 460 nm. The linear portion of the assay progress curves was used to calculate activity, setting the highest value of RFU/s (reflective fluorescent units per second) as 100% activity and correcting the other values accordingly. The percent activity values were averaged and are presented ± the standard deviation. For enzyme assays that replaced PfAPP with PfA-M1, PfA-M17, or Pf M18AAP (0–2000 nM), assay buffers were supplemented with 1 mM Co2+ for PfA-M17 and Pf M18AAP. No residual activity of Xf PIP alone was noted with the selected substrates. Enzyme velocity was obtained from the linear portion of the assay progress curves.

Protein crystallization

Initial crystallization conditions (5% polyethylene glycol 8000, 0.1 M sodium cacodylate, pH 6.5, 40% 2-methyl-2,4-pentanediol) were identified by screening PfAPP (8 mg/ml) against the JCSG+ Suite using 96-well sitting drop vapor diffusion plates. Final crystals, 0.05 × 0.1 × 0.15 mm in size, were obtained by hanging drop vapor diffusion method, with 2:1 (vol/vol) ratio of PfAPP (6 mg/ml) to mother liquor (0.5 ml reservoir volume) in 40–50% 2-methyl-2,4-pentanediol, 0.1 M sodium cacodylate, pH 5.6–6.2. Crystals appeared after 3 days and reached full size in 7 days. Crystals of the apstatin:PfAPP complex were obtained by soaking unliganded crystals in mother liquor supplemented with 2 mM apstatin for 1 h. Crystals of both unliganded PfAPP and apstatin:PfAPP were snap frozen in liquid nitrogen prior to data collection.

X-ray data collection, processing, and refinement

Data were collected at 100K using synchrotron radiation at the Australian Synchrotron using the micro crystallography MX2 beamline 3ID1. Crystal quality varied substantially and deteriorated rapidly after crystals reached full size. Final datasets were collected on crystals harvested less than 2 weeks after tray setup. Diffraction images were processed and integrated using iMosflm [36] and scaled using Aimless [37] to 2.35 Å (unliganded PfAPP) and 2.3 Å (apstatin:PfAPP). The structures were solved by molecular replacement using Phaser [38] as part of the CCP4 suite [39]. For the unliganded PfAPP structure, the molecular replacement search model was prepared from the structure of hAPP1 (PDB ID: 3CTZ) [6] using chainsaw [40] to prune non-conserved residues (maintained all atoms common to the target and model residues). The refined co-ordinates of the unliganded PfAPP structure were used as the molecular replacement search model to solve apstatin:PfAPP. The structures were refined using iterative cycles of PHENIX [41], with 5% of reflections set aside from refinement for calculation of Rfree. Between refinement cycles, the protein structure, solvent, and apstatin were visualized in COOT [42] and manually built into 2Fo − Fc and Fo − Fc electron density maps or density-modified maps (generated by parrot [43], ccp4 suite [39], during phase and electron density improvement). Modeling of domain I of chain B, for which the electron density was poor, was also guided by non-crystallographic symmetry maps and refined using secondary structure restraints. Apstatin restraint files were generated by the PRODRG2 server [44]. The co-ordinates and structure factors are available from the Protein Data Bank with PDB Accession codes 5JQK (unliganded) and 5JR6 (apstatin-bound), and a summary of data collection and refinement statistics is provided in Table 1.

Table 1
Crystallographic data collection, processing and refinement
Data collection PfAPP Apstatin:PfAPP 
Space group C 1 2 1 C 1 2 1 
Cell dimensions 
a, b, c (Å) 146.7, 100.1, 106.7 147.2, 99.9, 105.2 
α, β, γ (°) 90.0, 105.4, 90.0 90.0, 105.2, 90.0 
Resolution (Å) 70.75–2.35 Å 71.03–2.30 Å 
(2.41–2.35 Å) (2.36–2.30 Å) 
Total observations 229 848 (17 146) 237 546 (16 840) 
Unique observations 61 665 (4530) 65 055 (4547) 
Multiplicity 3.7 (3.8) 3.7 (3.7) 
Completeness (%) 99.5 (99.2) 99.6 (99.8) 
 8.2 (1.4) 8.7 (1.4) 
CC(1/2) 0.995 (0.553) 0.997 (0.602) 
Rpim (%) 6.9 (68.5) 5.5 (64.4) 
PDB 5JQK 5JR6 
Structure refinement 
 Non-hydrogen atoms 
  Protein 9602 8973 
  Solvent 525 203 
  Metal (Mn) 
  Ligand 33 
  Rfree (%) 22.6 24.3 
  Rcryst (%) 18.9 19.6 
 RMS deviations 
  Bond lengths (Å) 0.003 0.007 
  Angles (°) 0.851 1.09 
 Ramachandran plot 
  Favored (%) 98.7 97.6 
  Outliers (%) 
B factors (Å2
  Protein 64.0 68.2 
  Solvent 59.9 58.3 
  Ligand – 86.0 
 Molprobity (percentile) 97 92 
Data collection PfAPP Apstatin:PfAPP 
Space group C 1 2 1 C 1 2 1 
Cell dimensions 
a, b, c (Å) 146.7, 100.1, 106.7 147.2, 99.9, 105.2 
α, β, γ (°) 90.0, 105.4, 90.0 90.0, 105.2, 90.0 
Resolution (Å) 70.75–2.35 Å 71.03–2.30 Å 
(2.41–2.35 Å) (2.36–2.30 Å) 
Total observations 229 848 (17 146) 237 546 (16 840) 
Unique observations 61 665 (4530) 65 055 (4547) 
Multiplicity 3.7 (3.8) 3.7 (3.7) 
Completeness (%) 99.5 (99.2) 99.6 (99.8) 
 8.2 (1.4) 8.7 (1.4) 
CC(1/2) 0.995 (0.553) 0.997 (0.602) 
Rpim (%) 6.9 (68.5) 5.5 (64.4) 
PDB 5JQK 5JR6 
Structure refinement 
 Non-hydrogen atoms 
  Protein 9602 8973 
  Solvent 525 203 
  Metal (Mn) 
  Ligand 33 
  Rfree (%) 22.6 24.3 
  Rcryst (%) 18.9 19.6 
 RMS deviations 
  Bond lengths (Å) 0.003 0.007 
  Angles (°) 0.851 1.09 
 Ramachandran plot 
  Favored (%) 98.7 97.6 
  Outliers (%) 
B factors (Å2
  Protein 64.0 68.2 
  Solvent 59.9 58.3 
  Ligand – 86.0 
 Molprobity (percentile) 97 92 

Limited proteolysis

PfAPP (5 mg/ml) was treated with trypsin (0.5 mg/ml; Worthington Biochemical) in 50 mM Tris, pH 8.0, 50 mM NaCl at 25°C. Ten microliter aliquots were removed at time intervals of 1 min, 30 min, 60 min, and 20 h. Reactions were terminated by the addition of 5× gel loading buffer [0.5 M Tris, pH 6.8, 50% (v/v) glycerol, 5% (w/v) SDS, 0.05% (w/v) bromophenol blue, 5% (v/v) β-mercaptoethanol] and boiling at 95°C for 5 min. Samples were diluted 1:5 or 1:75, run on 15% SDS–PAGE gels, and analyzed by Coomassie blue stain or western blot analysis (using anti-His antibodies; His Tag horseradish peroxidase-conjugated antibody, R&D Systems), respectively. Trypsin cleavage sites were predicted by ExPASy PeptideCutter (>85% cleavage probability).

Results

Fluorogenic library and substrate fingerprint

The production of recombinant PfAPP (residues 129–777), its stability, and pH- and metal-dependence have been previously described [22]. Using a similar approach, we produced recombinant PfAPP (residues 121–777), which purified as an active dimer (kcat = 1480 s−1, Km = 190 μM, kcat/Km = 7.7 × 106 M−1 s−1). Ragheb et al. [22] also characterized the activity of PfAPP against three peptides with an X-Pro amino terminus: the nonapeptide bradykinin (RPPGFSPFR) and two pentapeptides from hemoglobin (FPHFD and YPWTQ). We sought to utilize a substrate profiling approach to determine whether PfAPP can catalyze the removal of any amino acid residue from X-Pro substrates. To this end, we synthesized a fluorogenic substrate library containing 20 amino acids (Scheme 1 and Supplementary Figure S1). The library was designed to screen the specific activity of PfAPP for 19 natural amino acids (to avoid oxidation artifacts, we omitted cysteine) and the unnatural amino acid norleucine. Recent studies investigating aminopeptidase substrate profiles attached an ACC fluorophore directly to the P1 residues [34]. However, since PfAPP is selective for a Pro at P1′, we chose to design the substrates as X-Pro-ACC. Release of the fluorogenic product therefore required two cleavage steps and necessitated the development of a coupled enzyme assay. To achieve this, we coupled the activity of PfAPP, which cleaves the X-Pro bond, with the prolyl iminopeptidase (post-prolyl, or Pro-X, aminopeptidase) from X. fastidiosa (Xf PIP), which cleaves the Pro-ACC bond to release the fluorogenic product (Figure 1A).

Coupled assay to detect the PfAPP substrate fingerprint.

Figure 1.
Coupled assay to detect the PfAPP substrate fingerprint.

(A) The coupled assay uses a prolyl iminopeptidase to release the free 7-amino-4-carbamoylmethylcoumarin. The substrates are shown in blue for the 19 natural amino acids and norleucine (X side-chain), proline in red and the 7-amino-4-carbamoylmethylcoumarin in gray, until it is released by proteolysis and fluoresces (ACC, yellow). (B) Percent activity exhibited by PfAPP for each of the 20 substrates tested. Amino acid residues at the P1 position are indicated on the X-axis, and ordered according to percent activity. (C) Activity of PfA-M1 (red), PfA-M17 (green), PfM18AAP (orange), and PfAPP (blue) for selected substrates as indicated on graph. Concentration of each PfMAP is indicated on X-axis and enzyme velocity on Y-axis, bars indicate standard error.

Figure 1.
Coupled assay to detect the PfAPP substrate fingerprint.

(A) The coupled assay uses a prolyl iminopeptidase to release the free 7-amino-4-carbamoylmethylcoumarin. The substrates are shown in blue for the 19 natural amino acids and norleucine (X side-chain), proline in red and the 7-amino-4-carbamoylmethylcoumarin in gray, until it is released by proteolysis and fluoresces (ACC, yellow). (B) Percent activity exhibited by PfAPP for each of the 20 substrates tested. Amino acid residues at the P1 position are indicated on the X-axis, and ordered according to percent activity. (C) Activity of PfA-M1 (red), PfA-M17 (green), PfM18AAP (orange), and PfAPP (blue) for selected substrates as indicated on graph. Concentration of each PfMAP is indicated on X-axis and enzyme velocity on Y-axis, bars indicate standard error.

Direct monitoring of the fluorescence enabled us to calculate PfAPP activity for each of the X-Pro-ACC substrates relative to the rest of the library. Initial screening identified the substrate Trp-Pro-ACC as having the highest activity in the library. We therefore chose to use Trp-Pro-ACC as our positive control, or 100% activity substrate, and normalized the activity of the remaining substrates against this compound. The results of the assay show that PfAPP is capable of catalyzing the removal of any amino acid from X-Pro peptides (Figure 1B), which is consistent with the presence of only one X-Pro aminopeptidase in Pf. While the enzyme has the capacity to remove any residue from the P1 position, it does demonstrate greater activity towards those with bulky and/or hydrophobic groups, with some exceptions (e.g. Ile). The most preferred substrates (>80% activity) were Trp, Lys, Nle, Tyr, Phe, Met, Val, Ala, Leu, and Arg. The only residue showing low cleavage by PfAPP was Asp, which was removed with only 5% of the activity of Trp.

Within P. falciparum, PfA-M1, PfA-M17, Pf M18AAP, and PfAPP, have been postulated to work in concert to complete the final stages of hemoglobin digestion in intra-erthyrocytic parasites [16]. Within human hemoglobin, there are 14 X-Pro sequences, 7 in each of the α- and β-subunits. We wanted to determine whether PfAPP is unique in its ability to cleave an X-Pro bond or whether the other three characterized MAPs (PfA-M1, PfA-M17, and Pf M18AAP) were capable of performing a similar function to that of PfAPP, and thereby providing redundancy within the hemoglobin digestion pathway. To investigate this, we performed our coupled enzyme assay with substrates containing a P1 residue specific for each MAP, but replaced PfAPP with PfA-M1, PfA-M17, or Pf M18AAP (Figure 1C). Negligible fluorescence was detected in these assays, confirming the need for a specialized APP in P. falciparum to catalyze the X-Pro bond.

PfAPP is a three-domain homodimer

To investigate how PfAPP might accommodate all residues within its S1 pocket, we determined its X-ray crystal structure. The structure of PfAPP was determined to 2.35 Å resolution in the C21 space group (see Data Collection and Refinement Statistics, Table 1). The phases were solved by molecular replacement using the co-ordinates of monomeric hAPP1 as the search model, which shares 32% sequence identity with PfAPP. The structure revealed the predicted homodimer quaternary structure, composed of two closely related monomers (PfAPPA and PfAPPB, RMSD = 0.7 Å) (Figure 2). Each monomer adopts a three-domain arrangement composed of the structurally similar domains I and II and the catalytic domain III (Figure 2A–C).

Crystal structure of PfAPP homodimer.

Figure 2.
Crystal structure of PfAPP homodimer.

The structure is shown in both cartoon and surface representation in three orientations. (A) Side view, (B) front view, and (C) top view. Structures are colored according to chain and domain. Chain A is shown colored according to domain, with domain I in blue, domain II in gray, and domain III in teal. Chain B is shown in magenta. (DF) Individual domains shown in cartoon representations in orientations that show the greatest clarity of structural features. (D) Domain I, with β-strand insertion shown in orange, (E) domain II, and (F) domain III.

Figure 2.
Crystal structure of PfAPP homodimer.

The structure is shown in both cartoon and surface representation in three orientations. (A) Side view, (B) front view, and (C) top view. Structures are colored according to chain and domain. Chain A is shown colored according to domain, with domain I in blue, domain II in gray, and domain III in teal. Chain B is shown in magenta. (DF) Individual domains shown in cartoon representations in orientations that show the greatest clarity of structural features. (D) Domain I, with β-strand insertion shown in orange, (E) domain II, and (F) domain III.

Domain I, composed of residues 121–304, has the greatest sequence divergence from hAPP1. Examination of the electron density of the PfAPP structure revealed that compared with domains II and III, the electron density of domain I is disordered, particularly for chain B. While it was of sufficient quality to model the polypeptide chain of chain A (with the exception of the N-terminal residues 121–128), regions on the solvent-exposed edges of chain B could either not be modeled at all, or be modeled as main chain and Cβ only. The disordered density did not improve with substantial refinement (including phase and density improvement software). Given that the quality of the data is good (see Table 1), the disordered density indicates that domain I may have a degree of flexibility within the crystal. Furthermore, the domain I–II linker region, composed of residues 299–306, is partially disordered in both chain A and B, further supporting the presence of domain I motions. Regardless of the partially disordered regions, clear secondary and quaternary structural features of domain I were observed. The domain possesses a seven-stranded β-sheet core surrounded by five helices (Figure 2D). Sequence alignments show that domain I of PfAPP contains a 20-residue insertion that is not observed in either hAPP1 or other apicomplexan APPs [22]. These residues, Tyr271–Val291, form a β-strand (β6), which is inserted into the central β-sheet (shown in orange in Figure 2D).

Domain II (residues 305–475) is structurally related to domain I, also possessing a β-sheet core surrounded by helices (RMSD = 2.5 Å) (Figure 2E). While insertions similar to that observed in domain I are commonly observed in domain II of other apicomplexan APPs, none are observed in domain II of PfAPP. The β-sheet core of domain II is therefore composed of only six β-strands, rather than the seven observed in domain I, though the five flanking helices are common to both domains. At the base of domain II, chains A and B are in close proximity (Figure 2B), and both chains possess a flexible loop (residues 428–436), for which density is not observed.

Domain III, the catalytic domain (residues 476–777), is composed of a ‘pita-bread’ fold. This fold is made up of two repeats of a ααβββ structure, which are arranged sheet to sheet to present an ∼2-fold symmetry axis (Figure 2F). Lying perpendicular to the symmetry axis of the ‘pita-bread’ fold are an additional two C-terminal α-helices, which cap the ‘top’ of the structure (and shown at the rear in Figure 2F). At the center of the domain on the symmetry axis, the junction of the two β-sheets creates a concave cavity, containing the active site of PfAPP. Somewhat breaking the dyad symmetry is a β-hairpin, which extends from one ααβββ repeat and, in combination with a curved helix, forms the dimerization interface (Figures 2F and 3A).

Dimerization interface

The dimerization interface of PfAPP lies between domain III of each of PfAPPA and PfAPPB. The interface, of 1080 Å2, is pseudosymmetrical, with the same residues from each chain involved in dimerization interactions (Figure 3A). Two key salt bridges in the center of the interface are observed (Arg608A/B –Glu629B/A) as well as hydrogen bonds at the edge of the interface (Tyr617A/B–Asp637B/A). However, the bulk of the surface is made up of hydrophobic interactions. On either end of the interface, three key Phe residues from each chain (Phe625A,B, Phe632A,B, and Phe635A,B), protrude into shallow pockets on the opposite face: Phe635A/B makes hydrophobic contacts with the corresponding Ile613, the aliphatic chains of Lys398 and Glu699, and the ring of Tyr706; Phe632A/B with Ala615, the main chain of Phe614, and the ring of Tyr706; and Phe625A/B contacts Ile626, Ala622, Leu623, Thr618, and Val612. The six Phe-pocket contacts are arranged so as to form an interaction interface that stretches the entire breadth of the molecule (Figure 3B,C).

PfAPP dimerization interface.

Figure 3.
PfAPP dimerization interface.

(A) Cartoon representation of the PfAPP dimerization interface colored according to chain. Domain IIIA shown in teal, domain IIIB shown in magenta. Key residues are shown in stick representation and labeled accordingly. Black dashed lines indicate salt bridges. (B) Surface representation of chain A oriented to show the dimerization interface face-on. Domain I shown in blue, domain II in gray, and domain III in teal. Dimerization interface colored according to role in interface: Phe residues (gray and ball-and-stick representation) from each chain protrude into pockets on the opposing chain (orange). (C) Close up of the interaction interface with Phe residues indicated. Colored according to (B).

Figure 3.
PfAPP dimerization interface.

(A) Cartoon representation of the PfAPP dimerization interface colored according to chain. Domain IIIA shown in teal, domain IIIB shown in magenta. Key residues are shown in stick representation and labeled accordingly. Black dashed lines indicate salt bridges. (B) Surface representation of chain A oriented to show the dimerization interface face-on. Domain I shown in blue, domain II in gray, and domain III in teal. Dimerization interface colored according to role in interface: Phe residues (gray and ball-and-stick representation) from each chain protrude into pockets on the opposing chain (orange). (C) Close up of the interaction interface with Phe residues indicated. Colored according to (B).

Active site arrangement

The PfAPP active site is located at the base of the catalytic domain, domain III, and lined by domains I and II (Figure 4A). Difference density indicated the presence of two metal ions co-ordinated tightly within the active site. Based on previous studies implicating manganese in the catalytic mechanism of PfAPP [22], we modeled two manganese ions (designated 1Mn and 2Mn) into the difference density. Difference density also indicated the presence of a tetragonal molecule within the active site, coordinating both 1Mn and 2Mn. Examination of the buffer and precipitant molecules that PfAPP is exposed to during purification and crystallization leads us to conclude that the density represents an ordered PO4 molecule, which co-purified with the enzyme (Supplementary Figure S3). The 1Mn ion is co-ordinated by both Oδ atoms of Asp570, Oδ1 of Asp581, Oε1 of Glu690, and one of the phosphate oxygens. 2Mn in turn is co-ordinated by Oε2 of Glu676, Nε2 of His644, Oδ2 of Asp581, Oε2 of Glu690, and the same phosphate oxygen that co-ordinates with 1Mn (Figure 4B).

PfAPP active site and apstatin-binding mechanism.

Figure 4.
PfAPP active site and apstatin-binding mechanism.

(A) Surface representation of PfAPP sliced to show position of the binding pocket. Chain A is shown colored according to domain, with domain I in blue, domain II in gray, and domain III in teal. Chain B is shown in magenta. (B) PfAPP active site with manganese ions shown as gray circles, and coordinating residues and phosphate molecule shown as sticks. Metal coordination interactions are shown as yellow dashed lines. (C) Apstatin (gray sticks) binding pose in the PfAPP pocket (teal sticks). Interactions between apstatin and the pocket residues are indicated by black dashes. (D) Apstatin binding pose (gray space filling representation) in the PfAPP pocket (surface representation, domain I colored blue, domain II colored teal).

Figure 4.
PfAPP active site and apstatin-binding mechanism.

(A) Surface representation of PfAPP sliced to show position of the binding pocket. Chain A is shown colored according to domain, with domain I in blue, domain II in gray, and domain III in teal. Chain B is shown in magenta. (B) PfAPP active site with manganese ions shown as gray circles, and coordinating residues and phosphate molecule shown as sticks. Metal coordination interactions are shown as yellow dashed lines. (C) Apstatin (gray sticks) binding pose in the PfAPP pocket (teal sticks). Interactions between apstatin and the pocket residues are indicated by black dashes. (D) Apstatin binding pose (gray space filling representation) in the PfAPP pocket (surface representation, domain I colored blue, domain II colored teal).

Examination of the protein surface showed that the active site has two openings to solvent. The first opening is to the interior of the dimer via a wide cavity, and the second, narrower opening is to the exterior of the molecule via a short channel ∼10 Å wide formed by the interface of domains I, II, and III.

Limited proteolysis

The crystal structure suggested substantial flexibility of domain I within the crystal. To investigate this apparent flexibility, we performed a limited trypsin digestion. PfAPP was treated with trypsin, and the state of digestion monitored over time by SDS–PAGE and western blot (using antibodies against the C-terminal His6 tag) analysis (Supplementary Figure S4). The position of unlabeled and His6-labeled bands indicates that domain I is efficiently digested by trypsin (evident after 1 min) and is completely removed after 20 h. After 20 h, the major digestion products correspond to domains II–III (∼57 kDa) and domain III (∼40 kDa), while small, unlabeled bands of size ∼12–17 kDa are most likely to contain mostly domain I fragments, plus some domain II. These observations support the idea that domain I is conformationally flexible and therefore more susceptible to digestion by trypsin in solution than the more rigid domains II and III.

Apstatin-bound structure

To provide insights into the mechanism by which substrates bind to PfAPP, we also examined the binding of apstatin to PfAPP. Apstatin is a non-hydrolysable peptide analog that selectively inhibits APP enzymes. We first determined whether apstatin inhibited PfAPP activity by utilizing the quenched fluorogenic substrate, Lys(Abz)-Pro-Pro-NA. We determined that apstatin inhibited PfAPP activity with an IC50 value of 20.2 ± 1.2 μM. We therefore soaked apstatin into unliganded PfAPP crystals and solved the structure of the apstatin:PfAPP complex by molecular replacement to 2.3 Å resolution (Table 1). Electron density for apstatin was visible in the binding pocket of PfAPPA (Supplementary Figure S5), while density in the PfAPPB-binding pocket was reminiscent of the ordered phosphate observed in the unliganded structure. We therefore modeled apstatin into chain A only. Electron density for the Phe (P1 position) and Pro (P1′ position) are clear, however that for the remainder of the molecule is somewhat disordered, and the atoms could only be modeled at half occupancy.

Overall, the architecture of the unliganded PfAPP and apstatin:PfAPP structures, including the active sites, is conserved (RMSD = 0.5 Å), suggesting that no conformational changes occur on binding of apstatin to PfAPP. However, since apstatin was soaked into PfAPP crystals, we cannot rule out the possibility that PfAPP undergoes apstatin-induced conformational changes in solution, but not within pre-formed crystals. As observed in the unliganded structure, the electron density for domain I is disordered; domain I of chain A was comparable to that in the unliganded structure, whereas domain I of chain B demonstrated weaker density and large stretches of backbone could not be modeled. Given that chain B is incomplete, we have restricted our analysis to chain A. The apstatin-binding position identifies the S1 pocket, which is occupied by the P1 Phe residue, and S1′ pocket, to which the P1′ Pro binds. At the junction of the subsites, the inhibitor co-ordinates both manganese ions. The apstatin hydroxyl group co-ordinates both 1Mn and 2Mn, while the carbonyl group co-ordinates 2Mn and forms a hydrogen bond with His653 (Figure 4C). The phenyl group occupies the S1 pocket, which is lined on one side by hydrophobic residues Phe537, Tyr636, Val652, Ile540, and Ile163, and His563, Glu162, and His 551 on the other side. The apstatin amine group forms cation–π interactions with the ring of Phe537 and orients the Phe moiety into the S1 pocket, where it forms hydrophobic contacts with the β-carbon of Phe537 and the side-chains of Ile163 and Val652. The S1′ pocket, is shaped by Glu676 and three His residues: Glu676 forms a hydrogen bond with the tertiary amine of the P1′ Pro, His640 caps the top of the pocket and forms edge face interactions with the P1′ Pro ring, and His653 interacts with the main chain carbonyl of apstatin and together with His551, introduces an acute bend between the S1 and S1′ pockets. Beyond the S1′ site, the pocket opens to solvent, with the P2′ Pro and remainder of the molecule extending into the internal solvent between the two monomers.

Comparison of APP from P. falciparum, E. coli, and Homo sapiens

The human APP, hAPP1, and PfAPP share only a modest 32% sequence identity; however, superposition of the two crystal structures shows a highly conserved 3D structure (RMSD of hAPPA and PfAPPA = 2.1 Å). Each of the three domains in hAPP1 and PfAPP are similar (Figure 5A), with the greatest divergence observed between domain I, within which PfAPP possesses an additional β-strand inserted into the central sheet structure. In contrast, domains II and III of the two enzymes overlay closely. Interestingly, domains II and III of PfAPP also overlay with the two-domain structure of E. coli APP (EcAPP) (RMSD = 2.9 Å) (Figure 5B).

Comparison of PfAPP to hAPP1 and EcAPP.

Figure 5.
Comparison of PfAPP to hAPP1 and EcAPP.

(A) Superposition of chain A of PfAPP and hAPP1. hAPP1 is shown in light blue, PfAPP is colored according to domains: domain I in dark blue (with β-strand insertion in orange), domain II in gray, and domain III in teal. (B) Superposition of chain A of the three-domain structure of PfAPP (colored according to A) and the two-domain structure of EcAPP (colored wheat).

Figure 5.
Comparison of PfAPP to hAPP1 and EcAPP.

(A) Superposition of chain A of PfAPP and hAPP1. hAPP1 is shown in light blue, PfAPP is colored according to domains: domain I in dark blue (with β-strand insertion in orange), domain II in gray, and domain III in teal. (B) Superposition of chain A of the three-domain structure of PfAPP (colored according to A) and the two-domain structure of EcAPP (colored wheat).

While the overall domain organization of PfAPP is similar to that of hAPP, their dimerization interfaces possess many key differences. The secondary structure of the interaction interface is similar, with both made up of the curved outer helix and β-hairpin of the ααβββ repeat; however, the specific interactions within the interface are not conserved. In the PfAPP interface, six key Phe-pocket interactions are observed. These are in place of hydrogen bonding interactions that characterize the hAPP1 interface, resulting in the PfAPP dimerization interface possessing a more hydrophobic nature than that of hAPP1.

The catalytic domains of PfAPP, EcAPP, and hAPP1 are conserved. All three enzymes possess an identical di-metal arrangement and have closely conserved S1 and S1′ sites. The crystal structure of EcAPP in complex with apstatin (apstatin:EcAPP) has also been determined [45]. Superposition of the catalytic domains of apstatin:PfAPP and apstatin:EcAPP demonstrates that apstatin binds to both EcAPP and PfAPP with an identical binding mode. Although the structure of hAPP1 with apstatin has not been determined, superposition of the catalytic domains of apstatin:PfAPP and unliganded hAPP1 gives an indication of the potential binding mode. The active site is located at the base of domain III and is composed almost entirely of domain III residues, with minor contributions from domain I residues (Figure 4D). When apstatin is placed in the hAPP1 active site, no major clashes are observed between the P1 Phe or P1′ Pro and the hAPP1 active site. However, differences are observed in the binding region for the P2′ of apstatin. In apstatin:PfAPP, P2′ and P3′ are exposed to solvent within the putative substrate entrance channel, the broad channel leading to the central cavity of the dimer. In this region, differences on the periphery of the dimerization interfaces between PfAPP and hAPP1 have an impact on the S2′ region of the sites. Specifically, in PfAPP an eight residue loop (660–667) is directed away from the active site and forms only long range, or water-mediated interactions with the opposing chain. In contrast, in hAPP, the equivalent loop (505–514) is two residues longer and makes more extensive, closer, dimerization interactions with the opposing chain. The result is that the larger loop in hAPP1 puckers and directs Lys507 towards the binding pocket. In this position, the lysine residue occupies the equivalent space of that inhabited by the P2′ proline residue in the apstatin:PfAPP structure.

Discussion

Many aminopeptidase enzymes demonstrate broad specificity and are capable of accepting a range of amino acid residues at the P1 and P1′ position [4649]. However, placement of a proline residue at either of these positions requires specific active site geometry; therefore, distinct enzymes capable of accepting proline residues are present in a range of organisms [28]. Such aminopeptidases include prolidases (X-Pro dipeptidase), iminopeptidases, and APPs. Within the genomes of Plasmodium species, the only enzymes capable of cleaving proline-containing substrates are an iminopeptidase, which cleaves the Pro-X bond, and an APP, which cleaves X-Pro [14,50,51]. While iminopeptidases from the Plasmodium species have not yet been characterized, parasites with a disrupted PfAPP gene could not be isolated, providing strong evidence that it is essential for parasite survival [14]. Further, PfAPP has been localized to both the food vacuole and cytosol of parasites and has demonstrated activity in environments representative of both [22]. This implicates PfAPP within both the hemoglobin digestion and proteasomal peptide recycling pathways [22] and presents it as an attractive target for the design of novel antimalarial therapeutics.

The Pf aminopeptidases PfA-M1, PfA-M17, and PfM18AAP have been demonstrated to exhibit varying selectivities for substrate P1 positions [49,52]. In contrast, little is known about the substrate fingerprint of PfAPP beyond the selectivity for proline at P1′. In the present study, we have shown that although PfAPP prefers larger, hydrophobic residues, it is capable of removing all 20 residues sampled in the library. Furthermore, PfA-M1, PfA-M17, or Pf M18AAP were not capable of replacing PfAPP in our coupled enzyme assay, indicating that, within Pf, APP is uniquely capable of hydrolyzing the X-Pro bond.

Given the presence of only a single APP enzyme in Plasmodium, the observation that PfAPP can remove any residue preceding a proline is not entirely unexpected. However, how can such a variety of chemical groups be accommodated within a single active site? To investigate, we determined the X-ray crystal structure of both PfAPP alone and in complex with the inhibitor, apstatin. The structure of PfAPP confirmed earlier work from Ragheb et al. [22], who used sequence alignments to propose that PfAPP shared the three-domain homodimeric organization of hAPP1, as opposed to the two-domain tetrameric organization of other members of the X-prolyl aminopeptidase family, such as EcAPP [53]. PfAPP is therefore only the second known member of the X-prolyl aminopeptidase family to adopt the three-domain homodimeric structure. Domains I and II of both hAPP1 and PfAPP are structurally similar, each being made up of a β-sheet core surrounded by α-helices. This similarity suggests that the three-domain architecture of mammalian (hAPP1) and apicomplexan (PfAPP) APPs may have evolved from the two-domain structure exhibited by other known APPs (such as EcAPP) by a partial internal duplication event [54,55]. Despite having no evident role within the active site, domain I of hAPP1 is important for function, as a truncated version of the enzyme, composed of only domains II and III, is not active [6]. The catalytic domain III of PfAPP is composed of a ‘pita-bread’ fold, which is shared by methionine aminopeptidases, prolidases, creatinases, and APPs. In the dimeric enzymes PfAPP and hAPP1, domain III additionally contains the dimersiation interface.

PfAPP is able to function at acidic pH, representative of the specialized digestive vacuole within the Plasmodium parasite [22]. Li et al. [6] demonstrated that disruption of hAPP1 dimerization resulted in an inactive protein, leading to the suggestion that hAPP1 dimerization is key to maintaining the correct fold and therefore catalytic activity. The hydrophobic nature of the PfAPP dimerization interface could potentially be responsible for the observed stability of the PfAPP dimer at acidic pH [22], thereby allowing it to maintain catalytic function within the digestive vacuole of the parasite.

We examined the remarkably broad specificity of PfAPP by investigating the mode of binding of apstatin. The apstatin:PfAPP structure showed that the PfAPP site is particularly structured to allow binding of any residue at P1, and only a proline residue at P1′. The S1 site possesses a dual nature; one side is lined with hydrophobic residues, while the other is made up of hydrophilic residues. In the apstatin:PfAPP structure, the Phe of apstatin lies up against the hydrophobic face, leaving the hydrophilic face exposed. This suggests that hydrophilic residues could align themselves with the opposing, hydrophilic face and explains how PfAPP is able to catalyze the removal of any P1 residue. However, the hydrophilic face lies deeper within the S1 pocket, which may account for the observation that Asp-Pro is the least favored substrate for PfAPP, being cleaved 10-fold less efficiently than the longer Glu-Pro analog. In contrast with the broad specificity of the S1 pocket, the S1′ site is highly selective for a proline residue. The apstatin:PfAPP structure shows that the geometry of S1′ site is tightly contoured to the P1′ proline. This is primarily achieved by the presence of three histidine residues (His551, His640, and His653), which are conserved in the APP enzymes of all five Plasmodium species. These histidine residues act to enclose the pocket, interact tightly with the P1′ proline and also introduce an acute bend into the active site which is discretely shaped for the unique geometry of an X-Pro peptide. This unique geometry presents the possibility to design potent, selective inhibitors for PfAPP.

Despite the overall structural differences within the APP family of enzymes, particularly with respect to the domain and quaternary arrangement, the active site architecture is highly conserved. Therefore, if PfAPP is to be considered a potential target for antimalarial drug discovery, we need to be confident that compounds could be designed to selectively bind and inhibit the Plasmodium form of the enzyme over the human form(s). Unlike PfAPP, mammalian APP2, a closely related membrane-bound homolog of APP1, cannot cleave substrates that possess a large P2′ residue [22,56]. This is consistent with our observation that PfAPP and hAPP1 differ in the S2′ region of their active sites. Our comparison of the structures has identified that while large P2′ groups can be accommodated in the solvent-exposed channel of PfAPP, they are predicted to clash with the spatially restricted S2′ region of the hAPP1 site. The differences in channel structures between PfAPP and hAPP1 suggest that inhibitors specific for the Pf form of the enzyme could be designed.

The structures of PfAPP and apstatin:PfAPP also inform on the PfAPP catalytic mechanism. The catalytic mechanism of EcAPP has been extensively studied [45,53,57], and the observation that PfAPP recognizes the substrate analog apstatin in the same way suggests that it utilizes a comparable reaction mechanism. Beyond the active site, the PfAPP structures provide insights on the mechanism of substrate/product processing. The PfAPP structure showed the presence of two solvent-filled channels that connect the active site to the inner cavity of the dimer and the exterior of the molecule. We propose that these channels serve for substrate entrance and product egress, respectively. The channel itself is formed by the intersection of domains I, II, and III. In hAPP1, truncation mutants that lack domain I are inactive [6]. This is curious given that domain I does not contribute substantially to substrate binding or to the architecture of the active site. The domain arrangement of PfAPP is near identical and therefore potentially operates in a comparable manner. Our crystal structure showed greater disorder in the electron density of domains I (chain A and B) compared with the rest of the molecule. In combination with our limited proteolysis results, this suggests that it possesses some flexibility within the crystal structure. Additionally, a flexible loop in domains II, at the junction of the dimer, is also observed. Domain motions, be they partial or full domain motions, have been implicated in substrate guidance for other aminopeptidases, particularly the M1 family of MAPs, the dynamics of which have been extensively studied [5862]. There is therefore precedent for the use of domain motions to moderate substrate access and product egress. If this were the case in PfAPP and hAPP1, it could potentially account for the importance of domain I, despite the lack of structural participation in the active site. Further, domain I motions could complement the role of dimerization in the catalytic activity of dimeric APPs; dimerization could potentially control substrate access to the inner cavity, while domain I motions could control substrate access to the active site. Elucidating the details of such a mechanism would provide insights into the function of PfAPP, and how it could be interfered with drug discovery.

Abbreviations

ACC, 7-amino-4-carbamoylmethylcoumarin; APP, aminopeptidase P; ART, artemisinin; DICI, N,N′-diisopropylcarbodiimide; Fmoc, 9-fluorenylmethyloxycarbonyl; hAPP1, human APP XPNPEP1; HATU, 2-(1H-7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyl uranium hexafluorophosphate methanaminium; HOBt, N-hydroxybenzotriazole; MAPs, metalloaminopeptidases; OD, optical density; Pf, Plasmodium falciparum; PfAPP, Plasmodium falciparum APP; TFA, trifluoroacetic acid; TIPS, triisopropylsilane; Xf PIP, Xylella fastidiosa prolyl iminopeptidase.

Author Contribution

N.D., M.D., and S.M. designed the research; N.D., K.K.S., R.S.B., W.R., K.M., N.B.V., P.J.S., M.D. and S.M. performed research; W.R., K.M., and M.D. contributed new reagents/analytic tools; N.D., R.S.B., W.R., K.M., N.B.V., M.D., and S.M. analyzed data; and N.D. and S.M. wrote the paper.

Funding

We thank the National Health and Medical Research Council (Project Grant [1063786] to S.M. and P.J.S.) for funding support. The work was also supported by a statutory activity subsidy from the Polish Ministry of Science and Higher Education for the Faculty of Chemistry at Wroclaw University of Technology.

Acknowledgments

The authors would like to thank Dr Francisco Javier Medrano from the Centro de Investigaciones Biologicas for providing us with the X. fastidiosa prolyl imino-aminopeptidase plasmid construct. We thank the Australian Synchrotron (MX-1 and MX-2) and the beamline scientists for beamtime (CAP8208) and for technical assistance. We thank the Monash Technology Research Platforms (Protein Production and Crystallization) for technical assistance.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
WHO
(
2015
)
World Malaria Report 2015.
1
244
p
2
Ashley
,
E.A.
,
Dhorda
,
M.
,
Fairhurst
,
R.M.
,
Amaratunga
,
C.
,
Lim
,
P.
,
Suon
,
S.
et al. 
(
2014
)
Spread of artemisinin resistance in Plasmodium falciparum malaria
.
N. Engl. J. Med.
371
,
411
423
doi:
3
Aminake
,
M.N.
,
Arndt
,
H.-D.
and
Pradel
,
G.
(
2012
)
The proteasome of malaria parasites: a multi-stage drug target for chemotherapeutic intervention?
Int. J. Parasitol. Drugs Drug Resist.
2
,
1
10
doi:
4
Gantt
,
S.M.
,
Myung
,
J.M.
,
Briones
,
M.R.
,
Li
,
W.D.
,
Corey
,
E.J.
,
Omura
,
S.
et al. 
(
1998
)
Proteasome inhibitors block development of Plasmodium spp
.
Antimicrob. Agents Chemother.
42
,
2731
2738
PMID:
[PubMed]
5
Paugam
,
A.
,
Bulteau
,
A.-L.
,
Dupouy-Camet
,
J.
,
Creuzet
,
C.
and
Friguet
,
B.
(
2003
)
Characterization and role of protozoan parasite proteasomes
.
Trends Parasitol.
19
,
55
59
doi:
6
Li
,
X.
,
Lou
,
Z.
,
Li
,
X.
,
Zhou
,
W.
,
Ma
,
M.
,
Cao
,
Y.
et al. 
(
2008
)
Structure of human cytosolic X-prolyl aminopeptidase: a double Mn(II)-dependent dimeric enzyme with a novel three-domain subunit
.
J. Biol. Chem.
283
,
22858
22866
doi:
7
Francis
,
S.E.
,
Sullivan
,
D.J.
and
Goldberg
,
D.E.
(
1997
)
Hemoglobin metabolism in the malaria parasite Plasmodium falciparum
.
Annu. Rev. Microbiol.
51
,
97
123
doi:
8
Goldberg
,
D.E.
(
2005
)
Hemoglobin degradation
.
Curr. Top Microbiol. Immunol.
295
,
275
291
doi:
9
Ball
,
E.G.
McKee
,
R.W.
,
Anfinsen
,
C.B.
,
Cruz
,
W.O.
and
Geiman
,
Q.M.
(
1948
)
Studies on malarial parasites: IX. Chemical and metabolic changes during growth and multiplication in vivo and in vitro
.
J. Biol. Chem.
175
,
547
571
10
Morrison
,
D.B.
and
Jeskey
,
H.A.
(
1948
)
Alterations in some constituents of the monkey erythrocyte infected with Plasmodium knowlesi as related to pigment formation
.
J. Natl Malar. Soc.
7
,
259
264
PMID:
[PubMed]
11
Orjih
,
A.U.
and
Fitch
,
C.D.
(
1993
)
Hemozoin production by Plasmodium falciparum: variation with strain and exposure to chloroquine
.
Biochim. Biophys. Acta, Gen. Subj.
1157
,
270
274
doi:
12
Liu
,
J.
,
Istvan
,
E.S.
,
Gluzman
,
I.Y.
,
Gross
,
J.
and
Goldberg
,
D.E.
(
2006
)
Plasmodium falciparum ensures its amino acid supply with multiple acquisition pathways and redundant proteolytic enzyme systems
.
Proc. Natl Acad. Sci. USA
103
,
8840
8845
doi:
13
Rosenthal
,
P.J.
(
2002
)
Hydrolysis of erythrocyte proteins by proteases of malaria parasites
.
Curr. Opin. Hematol.
9
,
140
145
doi:
14
Dalal
,
S.
and
Klemba
,
M.
(
2007
)
Roles for two aminopeptidases in vacuolar hemoglobin catabolism in Plasmodium falciparum
.
J. Biol. Chem.
282
,
35978
35987
doi:
15
Harbut
,
M.B.
,
Velmourougane
,
G.
,
Dalal
,
S.
,
Reiss
,
G.
,
Whisstock
,
J.C.
,
Onder
,
O.
et al. 
(
2011
)
Bestatin-based chemical biology strategy reveals distinct roles for malaria M1- and M17-family aminopeptidases
.
Proc. Natl Acad. Sci. USA
108
,
E526
E534
doi:
16
McGowan
,
S.
(
2013
)
Working in concert: the metalloaminopeptidases from Plasmodium falciparum
.
Curr. Opin. Struct. Biol.
23
,
828
835
doi:
17
Stack
,
C.M.
,
Lowther
,
J.
,
Cunningham
,
E.
,
Donnelly
,
S.
,
Gardiner
,
D.L.
,
Trenholme
,
K.R.
et al. 
(
2007
)
Characterization of the Plasmodium falciparum M17 leucyl aminopeptidase: a protease involved in amino acid regulation with potential for antimalarial drug development
.
J. Biol. Chem.
282
,
2069
2080
doi:
18
Allary
,
M.
,
Schrevel
,
J.
and
Florent
,
I.
(
2002
)
Properties, stage-dependent expression and localization of Plasmodium falciparum M1 family zinc-aminopeptidase
.
Parasitology
125
,
1
10
doi:
19
Azimzadeh
,
O.
,
Sow
,
C.
,
Gèze
,
M.
,
Nyalwidhe
,
J.
and
Florent
,
I.
(
2010
)
Plasmodium falciparum PfA-M1 aminopeptidase is trafficked via the parasitophorous vacuole and marginally delivered to the food vacuole
.
Malar. J.
9
,
189
doi:
20
Klemba
,
M.
,
Gluzman
,
I.
and
Goldberg
,
D.E.
(
2004
)
A Plasmodium falciparum dipeptidyl aminopeptidase I participates in vacuolar hemoglobin degradation
.
J. Biol. Chem.
279
,
43000
43007
doi:
21
Teuscher
,
F.
,
Lowther
,
J.
,
Skinner-Adams
,
T.S.
,
Spielmann
,
T.
,
Dixon
,
M.W.A.
,
Stack
,
C.M.
et al. 
(
2007
)
The m18 aspartyl aminopeptidase of the human malaria parasite Plasmodium falciparum
.
J. Biol. Chem.
282
,
30817
30826
doi:
22
Ragheb
,
D.
,
Bompiani
,
K.
,
Dalal
,
S.
and
Klemba
,
M.
(
2009
)
Evidence for catalytic roles for Plasmodium falciparum aminopeptidase P in the food vacuole and cytosol
.
J. Biol. Chem.
284
,
24806
24815
doi:
23
McGowan
,
S.
,
Oellig
,
C.A.
,
Birru
,
W.A.
,
Caradoc-Davies
,
T.T.
,
Stack
,
C.M.
,
Lowther
,
J.
et al. 
(
2010
)
Structure of the Plasmodium falciparum M17 aminopeptidase and significance for the design of drugs targeting the neutral exopeptidases
.
Proc. Natl Acad. Sci. USA
107
,
2449
2454
doi:
24
McGowan
,
S.
,
Porter
,
C.J.
,
Lowther
,
J.
,
Stack
,
C.M.
,
Golding
,
S.J.
,
Skinner-Adams
,
T.S.
et al. 
(
2009
)
Structural basis for the inhibition of the essential Plasmodium falciparum M1 neutral aminopeptidase
.
Proc. Natl Acad. Sci. USA
106
,
2537
2542
doi:
25
Mistry
,
S.N.
,
Drinkwater
,
N.
,
Ruggeri
,
C.
,
Sivaraman
,
K.K.
,
Loganathan
,
S.
,
Fletcher
,
S.
et al. 
(
2014
)
Two-pronged attack: dual inhibition of Plasmodium falciparum M1 and M17 metalloaminopeptidases by a novel series of hydroxamic acid-based inhibitors
.
J. Med. Chem.
57
,
9168
9183
doi:
26
Sivaraman
,
K.K.
,
Paiardini
,
A.
,
Sieńczyk
,
M.
,
Ruggeri
,
C.
,
Oellig
,
C.A.
,
Dalton
,
J.P.
et al. 
(
2013
)
Synthesis and structure–activity relationships of phosphonic arginine mimetics as inhibitors of the M1 and M17 aminopeptidases from Plasmodium falciparum
.
J. Med. Chem.
56
,
5213
5217
doi:
27
Drinkwater
,
N.
,
Vinh
,
N.B.
,
Mistry
,
S.N.
,
Bamert
,
R.S.
,
Ruggeri
,
C.
,
Holleran
,
J.P.
et al. 
(
2016
)
Potent dual inhibitors of Plasmodium falciparum M1 and M17 aminopeptidases through optimization of S1 pocket interactions
.
Eur. J. Med. Chem.
110
,
43
64
doi:
28
Cunningham
,
D.F.
and
O'Connor
,
B.
(
1997
)
Proline specific peptidases
.
Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol.
1343
,
160
186
doi:
29
Hooper
,
N.M.
,
Hryszko
,
J.
,
Oppong
,
S.Y.
and
Turner
,
A.J.
(
1992
)
Inhibition by converting enzyme inhibitors of pig kidney aminopeptidase P
.
Hypertension
19
,
281
285
doi:
30
Rusu
,
I.
and
Yaron
,
A.
(
1992
)
Aminopeptidase P from human leukocytes
.
Eur. J. Biochem.
210
,
93
100
doi:
31
Vanhoof
,
G.
,
de meester
,
I.
,
Goossens
,
F.
,
Hendriks
,
D.
,
Scharpé
,
S.
and
Yaron
,
A.
(
1992
)
Kininase activity in human platelets: cleavage of the Arg1-Pro2 bond of bradykinin by aminopeptidase P
.
Biochem. Pharmacol.
44
,
479
487
doi:
32
Erşahin
,
Ç.
,
Euler
,
D.E.
and
Simmons
,
W.H.
(
1999
)
Cardioprotective effects of the aminopeptidase P inhibitor apstatin: studies on ischemia/reperfusion injury in the isolated rat heart
.
J. Cardiovasc. Pharmacol.
34
,
604
611
doi:
33
Maly
,
D.J.
,
Leonetti
,
F.
,
Backes
,
B.J.
,
Dauber
,
D.S.
,
Harris
,
J.L.
,
Craik
,
C.S.
et al. 
(
2002
)
Expedient solid-phase synthesis of fluorogenic protease substrates using the 7-amino-4-carbamoylmethylcoumarin (ACC) fluorophore
.
J. Org. Chem.
67
,
910
915
doi:
34
Poreba
,
M.
,
Szalek
,
A.
,
Kasperkiewicz
,
P.
and
Drag
,
M.
(
2014
) Positional scanning substrate combinatorial library (PS-SCL) approach to define caspase substrate specificity. In
Caspases, Paracaspases, and Metacaspases: Methods and Protocols. Methods in Molecular Biology
(
Bozhkov
,
P.V.
and
Salvesen
,
G.
, eds), vol.
1133
, pp.
41
59
35
Stöckel-Maschek
,
A.
,
Stiebitz
,
B.
,
Koelsch
,
R.
and
Neubert
,
K.
(
2003
)
A continuous fluorimetric assay for aminopeptidase P detailed analysis of product inhibition
.
Anal. Biochem.
322
,
60
67
doi:
36
Battye
,
T.G.G.
,
Kontogiannis
,
L.
,
Johnson
,
O.
,
Powell
,
H.R.
and
Leslie
,
A.G.W.
(
2011
)
iMOSFLM: a new graphical interface for diffraction-image processing with MOSFLM
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
67
,
271
281
doi:
37
Evans
,
P.R.
and
Murshudov
,
G.N.
(
2013
)
How good are my data and what is the resolution?
Acta Crystallogr., Sect. D: Biol. Crystallogr.
69
,
1204
1214
doi:
38
McCoy
,
A.J.
,
Grosse-Kunstleve
,
R.W.
,
Adams
,
P.D.
,
Winn
,
M.D.
,
Storoni
,
L.C.
and
Read
,
R.J.
(
2007
)
Phaser crystallographic software
.
J. Appl. Crystallogr.
40
,
658
674
doi:
39
Winn
,
M.D.
,
Ballard
,
C.C.
,
Cowtan
,
K.D.
,
Dodson
,
E.J.
,
Emsley
,
P.
,
Evans
,
P.R.
et al. 
(
2011
)
Overview of the CCP4 suite and current developments
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
67
,
235
242
doi:
40
Stein
,
N.
(
2008
)
CHAINSAW: a program for mutating pdb files used as templates in molecular replacement
.
J. Appl. Crystallogr.
41
,
641
643
doi:
41
Adams
,
P.D.
,
Afonine
,
P.V.
,
Bunkóczi
,
G.
,
Chen
,
V.B.
,
Davis
,
I.W.
,
Echols
,
N.
et al. 
(
2010
)
PHENIX: a comprehensive Python-based system for macromolecular structure solution
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
66
,
213
221
doi:
42
Emsley
,
P.
,
Lohkamp
,
B.
,
Scott
,
W.G.
and
Cowtan
,
K.
(
2010
)
Features and development of Coot
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
66
,
486
501
doi:
43
Cowtan
,
K.
(
2010
)
Recent developments in classical density modification
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
66
,
470
478
doi:
44
Schuttelkopf
,
A.W.
and
van Aalten
,
D.M.F.
(
2004
)
PRODRG: a tool for high-throughput crystallography of protein–ligand complexes
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
60
,
1355
1363
doi:
45
Graham
,
S.C.
,
Maher
,
M.J.
,
Simmons
,
W.H.
,
Freeman
,
H.C.
and
Guss
,
J.M.
(
2004
)
Structure of Escherichia coli aminopeptidase P in complex with the inhibitor apstatin
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
60
,
1770
1779
doi:
46
Byzia
,
A.
,
Szeffler
,
A.
,
Kalinowski
,
L.
and
Drag
,
M.
(
2016
)
Activity profiling of aminopeptidases in cell lysates using a fluorogenic substrate library
.
Biochimie
122
,
31
37
doi:
47
Cunningham
,
E.
,
Drag
,
M.
,
Kafarski
,
P.
and
Bell
,
A.
(
2008
)
Chemical target validation studies of aminopeptidase in malaria parasites using α-aminoalkylphosphonate and phosphonopeptide inhibitors
.
Antimicrob. Agents Chemother.
52
,
3221
3228
doi:
48
Drag
,
M.
,
Bogyo
,
M.
,
Ellman
,
J.A.
and
Salvesen
,
G.S.
(
2010
)
Aminopeptidase fingerprints, an integrated approach for identification of good substrates and optimal inhibitors
.
J. Biol. Chem.
285
,
3310
3318
doi:
49
Poreba
,
M.
,
McGowan
,
S.
,
Skinner-Adams
,
T.S.
,
Trenholme
,
K.R.
,
Gardiner
,
D.L.
,
Whisstock
,
J.C.
et al. 
(
2012
)
Fingerprinting the substrate specificity of M1 and M17 aminopeptidases of human malaria, Plasmodium falciparum
.
PLoS ONE
7
,
e31938
doi:
50
Le Roch
,
K.G.
,
Johnson
,
J.R.
,
Florens
,
L.
,
Zhou
,
Y.Y.
,
Santrosyan
,
A.
,
Grainger
,
M.
et al. 
(
2004
)
Global analysis of transcript and protein levels across the Plasmodium falciparum life cycle
.
Genome Res.
14
,
2308
2318
doi:
51
Trenholme
,
K.R.
,
Brown
,
C.L.
,
Skinner-Adams
,
T.S.
,
Stack
,
C.
,
Lowther
,
J.
,
To
,
J.
et al. 
(
2010
)
Aminopeptidases of malaria parasites: new targets for chemotherapy
.
Infect. Disord.: Drug Targets
10
,
217
225
doi:
52
Sivaraman
,
K.K.
,
Oellig
,
C.A.
,
Huynh
,
K.
,
Atkinson
,
S.C.
,
Poreba
,
M.
,
Perugini
,
M.A.
et al. 
(
2012
)
X-ray crystal structure and specificity of the Plasmodium falciparum malaria aminopeptidase PfM18AAP
.
J. Mol. Biol.
422
,
495
507
doi:
53
Wilce
,
M.C.J.
,
Bond
,
C.S.
,
Dixon
,
N.E.
,
Freeman
,
H.C.
,
Guss
,
J.M.
,
Lilley
,
P.E.
et al. 
(
1998
)
Structure and mechanism of a proline-specific aminopeptidase from Escherichia coli
.
Proc. Natl Acad. Sci. USA
95
,
3472
3477
doi:
54
Nacher
,
J.C.
,
Hayashida
,
M.
and
Akutsu
,
T.
(
2010
)
The role of internal duplication in the evolution of multi-domain proteins
.
Biosystems
101
,
127
135
doi:
55
Zhang
,
J.Z.
(
2003
)
Evolution by gene duplication: an update
.
Trends Ecol. Evol.
18
,
292
298
doi:
56
Orawski
,
A.T.
and
Simmons
,
W.H.
(
1995
)
Purification and properties of membrane-bound aminopeptidase P from rat lung
.
Biochemistry
34
,
11227
11236
doi:
57
Graham
,
S.C.
,
Lilley
,
P.E.
,
Lee
,
M.
,
Schaeffer
,
P.M.
,
Kralicek
,
A.V.
,
Dixon
,
N.E.
et al. 
(
2006
)
Kinetic and crystallographic analysis of mutant Escherichia coli aminopeptidase P: insights into substrate recognition and the mechanism of catalysis
.
Biochemistry
45
,
964
975
doi:
58
Chen
,
L.
,
Lin
,
Y.-L.
,
Peng
,
G.
and
Li
,
F.
(
2012
)
Structural basis for multifunctional roles of mammalian aminopeptidase N
.
Proc. Natl Acad. Sci. USA
109
,
17966
17971
doi:
59
Jones
,
P.M.
,
Robinson
,
M.W.
,
Dalton
,
J.P.
and
George
,
A.M.
(
2011
)
The Plasmodium falciparum malaria M1 alanyl aminopeptidase (PfA-M1): insights of catalytic mechanism and function from MD simulations
.
PLoS ONE
6
,
e28589
doi:
60
Kochan
,
G.
,
Krojer
,
T.
,
Harvey
,
D.
,
Fischer
,
R.
,
Chen
,
L.
,
Vollmar
,
M.
et al. 
(
2011
)
Crystal structures of the endoplasmic reticulum aminopeptidase-1 (ERAP1) reveal the molecular basis for N-terminal peptide trimming
.
Proc. Natl Acad. Sci. USA
108
,
7745
7750
doi:
61
Kyrieleis
,
O.J.P.
,
Goettig
,
P.
,
Kiefersauer
,
R.
,
Huber
,
R.
and
Brandstetter
,
H.
(
2005
)
Crystal structures of the tricorn interacting factor F3 from Thermoplasma acidophilum, a zinc aminopeptidase in three different conformations
.
J. Mol. Biol.
349
,
787
800
doi:
62
Nguyen
,
T.T.
,
Chang
,
S.-C.
,
Evnouchidou
,
I.
,
York
,
I.A.
,
Zikos
,
C.
,
Rock
,
K.L.
et al. 
(
2011
)
Structural basis for antigenic peptide precursor processing by the endoplasmic reticulum aminopeptidase ERAP1
.
Nat. Struct. Mol. Biol.
18
,
604
613
doi:

Author notes

*

These authors contributed equally.

Supplementary data