Distribution and Biochemical Properties of an M1-family Aminopeptidase in Plasmodium falciparum Indicate a Role in Vacuolar Hemoglobin Catabolism*

Aminopeptidases catalyze N-terminal peptide bond hydrolysis and occupy many diverse roles across all domains of life. Here we present evidence that an M1-family aminopeptidase, PfA-M1, has been recruited to specialized roles in the human malaria parasite Plasmodium falciparum. PfA-M1 is abundant in two subcellular compartments in asexual intraerythrocytic parasites; that is, the food vacuole, where the catabolism of host hemoglobin takes place, and the nucleus. A unique N-terminal extension contributes to the observed dual targeting by providing a signal peptide and putative alternate translation initiation sites. PfA-M1 exists as two major isoforms, a nuclear 120-kDa species and a processed species consisting of a complex of 68- and 35-kDa fragments. PfA-M1 is both stable and active at the acidic pH of the food vacuole lumen. Determination of steady-state kinetic parameters for both aminoacyl-β-naphthylamide and unmodified dipeptide substrates over the pH range 5.0–8.5 reveals that kcat is relatively insensitive to pH, whereas Km increases at pH values below 6.5. At the pH of the food vacuole lumen (5.0–5.5), the catalytic efficiency of PfA-M1 remains high. Consistent with the kinetic data, the affinity of peptidic competitive inhibitors is diminished at acidic pH. Together, these results support a catalytic role for PfA-M1 in the food vacuole and indicate the importance of evaluating the potency of peptidic inhibitors at physiologically relevant pH values. They also suggest a second, distinct function for this enzyme in the parasite nucleus.

Human malaria is responsible for around one million deaths annually (1). Five species of the genus Plasmodium cause malaria in humans as they replicate within host erythrocytes. The cytoadherent properties of intraerythrocytic Plasmodium falciparum coupled with its ability to invade mature erythrocytes make it the most virulent of the species that infect humans. During its erythrocytic replication cycle, P. falciparum endocytoses and catabolizes over two-thirds of soluble erythrocyte proteins (2,3), the majority of which is hemoglobin. Hemoglobin catabolism provides amino acids for protein synthesis, general metabolism, and isoleucine import (4,5) and may also prevent premature hemolysis by reducing the colloid osmolarity of the erythrocyte (6). Blocking hemoglobin catabolism with protease inhibitors prevents parasite replication; therefore, enzymes that catalyze this process are attractive targets for the development of novel anti-malarial drugs (7).
Hemoglobin is extensively catabolized by the parasite within an acidic organelle called the food vacuole or digestive vacuole. In the vacuole, numerous types of endo-and exopeptidases act in a complementary and concerted manner to catalyze the hydrolysis of the ␣and ␤-globin chains of hemoglobin. Aspartic proteases (plasmepsin I, II, IV, and histo-aspartic protease) and cysteine endoproteases (falcipain-2, -2Ј, and -3) initiate cleavage of the globin chains and generate polypeptide fragments (8,9). The metallopeptidase falcilysin produces oligopeptides from these fragments (10), which are further reduced to dipeptides by the exopeptidase dipeptidyl aminopeptidase 1 (11,12). Peptides with a proline residue in the second position are trimmed by the aminopeptidase P homolog PfAPP (13).
What happens next is less clearly understood. The liberation of amino acids from globin peptides requires the action of broad specificity aminopeptidases or carboxypeptidases. As of yet no carboxypeptidases have been found to participate in globin peptide catabolism (14). Rather, two aminopeptidase have been implicated in this process: an M1-family aminopeptidase termed PfA-M1 (15)(16)(17)(18)(19) and the M17-family leucine aminopeptidase homolog PfLAP (19 -21).
One model for the generation of amino acids from globin diand oligopeptides that gained early traction is the "peptide export" model (for a recent elaboration of this model, see Ref. 18). According to this view, short peptides are transported out of the vacuole to the cytosol, where the final hydrolytic steps are catalyzed by aminopeptidases (14). Early support for this model included the apparent absence of aminopeptidase activity from enriched food vacuoles (14) and the neutral-to-basic pH optima of parasite aminopeptidase activities (22,23). The apparent localization of PfA-M1 and PfLAP to the cytosol appeared to provide further evidence for this model (15,18,21). More recently, a study in which these aminopeptidases were tagged with yellow fluorescent protein by allelic modification confirmed a cytosolic distribution for PfLAP but found that PfA-M1 exhibited a dual distribution in the food vacuole and the nucleus (24). This observation provided the first evidence that extensive amino acid production could occur in the acidic lumen of the food vacuole and was bolstered by the enrichment of PfA-M1 activity in food vacuole preparations (24).
Recently, two alternate interpretations of the food vacuole localization of Dalal and Klemba (24) have been advanced that do not invoke a role for PfA-M1 in the vacuole lumen. Whisstock et al. (25) have proposed that PfA-M1 is an integral membrane protein anchored in the food vacuole membrane with the catalytic domains in the cytosol, and Azimzadeh et al. (16) have suggested that PfA-M1 accumulates in cytosolic vesicles that reside proximal to the food vacuole. The idea that PfA-M1 plays a catalytic role in the vacuole has also been questioned on the basis of an apparent pH profile for PfA-M1 catalysis that precludes significant activity at pH values below 6 (18), such as that found in the P. falciparum food vacuole (26 -29). However, a rigorous evaluation of the stability and kinetic properties of PfA-M1 at acidic pH has not yet been reported, and its catalytic potential under these conditions remains unclear.
PfA-M1 is a member of the expansive M1 family of metalloaminopeptidases. Twelve members of this family have been identified in humans (30); however, PfA-M1 appears to be more closely related to prokaryotic M1-family enzymes such as Escherichia coli PepN (18). PfA-M1 possesses an ϳ190-amino acid sequence preceding the conserved M1-family domains that is not found in PepN (18). Crystal structures of PfA-M1 and PepN reveal that these aminopeptidases consist of four conserved domains, with a buried active site in the thermolysin-like domain II (18,31,32). A single metal ion, presumed to be zinc(II), is observed in the active site and is thought to activate a water molecule and coordinate the tetrahedral intermediate (32). The dipeptide mimetic bestatin exploits this active site configuration. A potent inhibitor of PfA-M1 (18), bestatin resembles a Phe-Leu dipeptide but contains an N-terminal ␣-hydroxy-␤-amino acid residue (33). In the bestatin-PfA-M1 co-crystal structure, bestatin occupies the active site with the ␣-hydroxy group displacing the catalytic water molecule (18).
In this study we tested the hypothesis that PfA-M1 has a catalytic role in globin degradation inside the food vacuole lumen. First, we localized untagged, native PfA-M1 to determine whether it could contribute to peptide catabolism within the vacuole. We then assessed the contribution of the N-terminal extension to the observed subcellular distribution. Next, we analyzed the stability and steady-state kinetic parameters of PfA-M1 over the pH range 5.0 -8.5. To further examine the role of pH in substrate binding, we determined its effect on inhibition of PfA-M1 by the substrate analogs bestatin and bestatin methyl ester.

EXPERIMENTAL PROCEDURES
Parasite Culture and Isolation-P. falciparum 3D7 parasites were cultured in human O ϩ erythrocytes (Interstate Blood Bank) in RPMI 1640 medium supplemented with 27 mM sodium bicarbonate, 11 mM glucose, 0.37 mM hypoxanthine, 10 g/ml gentamicin, and 5 g/liter Albumax I (Invitrogen). Cultures were synchronized by sorbitol treatment (34). In certain experiments, variants of 3D7 were employed in which the genomic PfA-M1 or PfLAP coding sequence had been modified to encode a C-terminal hemagglutinin (HA) tag (24).
Purification of Native PfA-M1-PfA-M1 was purified from a modified 3D7 parasite line that expressed PfLAP with a C-terminal HA tag (24). Trophozoite stage parasites were isolated from infected red blood cells with 0.15% (w/v) saponin in Dulbecco's phosphate-buffered saline (PBS). To remove any remaining hemoglobin, which co-purified with PfA-M1, parasite pellets were frozen at Ϫ80°C, thawed, and washed 3 times in PBS supplemented with 10 M N-(trans-epoxysuccinyl)-Lleucine 4-guanidinobutylamide, 10 M pepstatin, and 1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride. The pellet was then resuspended in 20 mM Bis-tris 2 ⅐HCl pH 6.0 with the same inhibitor concentrations, and cells were disrupted by sonication. Parasite lysate was centrifuged at 100,000 ϫ g for 1 h, and the supernatant was loaded onto a Mono Q 5/50 GL (GE Healthcare) column equilibrated with 20 mM Bis-tris⅐HCl, pH 6.0. The flow-through, which contained PfA-M1 activity, was adjusted to pH 7.5 by adding Tris base to 15 mM and injected onto a Mono Q 5/50 GL column equilibrated with 20 mM Tris-HCl, pH 7.5. Bound protein was eluted with a linear gradient from 0 to 1 M NaCl in 20 mM Tris-HCl, pH 7.5. PfA-M1 eluted at around 330 mM NaCl. Fractions with PfA-M1 activity were concentrated in an Ultra-4 centrifugal device (Amicon) and injected onto a Superdex 200 10/30 gel filtration column (GE Healthcare) equilibrated in 50 mM Tris-HCl, pH 7.5, 200 mM NaCl. Fractions were subjected to immunoblotting with anti-PfA-M1 serum to assess the distributions of the PfA-M1 species. Fractions enriched in p120 over p68/p35 and those enriched in p68/p35 over p120 were combined separately. Purity of the pools was assessed on a silver-stained polyacrylamide gel. For long term protein storage, Triton X-100 was added to 0.1% (v/v), and aliquots were snap-frozen in liquid nitrogen and stored at Ϫ80°C. Protein concentrations were calculated by quantitative immunoblotting as described in this section under "Immunoblotting and immunolocalization" using a standard curve generated with recombinant PfA-M1. Because the p68 fragment is two-thirds of the size of the recombinant PfA-M1 standard used to generate the polyclonal antibodies, it was assumed that the antibody signal per molecule of p68 was two-thirds that of the standard (i.e. antibodies binding to p35 were lost from the quantitation). To adjust for this, the amounts of p68 calculated from the standard curve were multiplied by a factor of 1.5.
Expression and Purification of Recombinant PfA-M1-DNA coding for residues 192-1085 of P. falciparum PfA-M1 (gene ID MAL13P1.56) was amplified by PCR from clone 3D7 genomic DNA using the forward primer, 5Ј-GCACGGGATC-CCGAAAACCTGTATTTTCAGAGCAAAAAGAACGAACC-AAAAATACATTATAGG, and the reverse primer, 5Ј-GCA-CGGCGGCCGCTTATAATTTATTTGTTAATCTTAATA-AATA. The PCR product was digested with BamH1 and NotI (underlined) and ligated into the same sites in the T7 expression vector pET45b (Novagen), which appended an N-terminal hexahistidine tag. The PfA-M1 coding sequence was immediately preceded by the primer-encoded tobacco etch virus (TEV) protease cleavage site ENLYFQS ( (35); see the italicized sequence in the forward primer). The sequence was confirmed by DNA sequencing.
The expression plasmid was transformed into E. coli BL21(DE3) Rosetta 2 (Novagen). Bacterial cultures were grown to an optical density of 0.8 at 600 nm, and protein expression was induced by the addition of 1 mM isopropyl-D-thiogalactopyranoside for 12 h at 25°C. Recombinant PfA-M1 was purified by immobilized metal affinity chromatography as described previously for P. falciparum aminopeptidase P (13). Fractions containing PfA-M1 were pooled and dialyzed at 4°C into 50 mM Tris-HCl, pH 7.5, 200 mM NaCl. To remove the N-terminal histidine tag, PfA-M1 was incubated with hexahistidine-tagged TEV protease (purified as described in Kapust et al. (36)) at a molar PfA-M1:TEV protease ratio of 10:1 in 50 mM Tris-HCl, pH 8.0, at 4°C overnight. Cleaved PfA-M1 was purified and concentrated as previously described (13) and dialyzed into 50 mM Tris-HCl, pH 7.5, 200 mM NaCl overnight at 4°C. The dialysate was injected onto a Superdex 200 10/30 gel filtration column (GE Healthcare) equilibrated in 50 mM Tris-HCl, pH 7.5, 200 mM NaCl. Fractions containing active PfA-M1 were pooled, dialyzed overnight at 4°C into 50 mM Tris-HCl, pH 7.5, containing 10 M ZnCl 2 , snap-frozen in liquid nitrogen, and stored at Ϫ80°C. Protein concentration was determined from absorbance at 280 nm using a calculated extinction coefficient of 1.150 ϫ 10 5 M Ϫ1 ⅐cm Ϫ1 .
Immunoblotting and Immunolocalization-Rabbit anti-PfA-M1 serum was produced using recombinant PfA-M1 as the immunogen (Cocalico Biologicals). An affinity column was generated by covalently coupling recombinant PfA-M1 to AminoLink Coupling Resin (Pierce) according to the manufacturer's instructions. Anti-PfA-M1 antibodies were bound, eluted at acidic pH, and rapidly neutralized according to an established procedure (37). The affinity-purified antibodies were dialyzed overnight at 4°C against PBS, pH 7.4, flash-frozen in liquid nitrogen, and stored at Ϫ80°C.
In immunoblotting experiments anti-PfA-M1 serum was used at a 1:10,000 dilution, or affinity-purified anti-PfA-M1 antibodies were used at a concentration of 0.13 g/ml. Other antibodies employed were rabbit anti-HA antibody (Invitrogen; 0.25 g/ml) and mouse anti-plasmepsin V monoclonal antibody 23.1.2 (1:400 (38)). These were followed by a horseradish peroxidase-conjugated anti-rabbit or anti-mouse secondary antibody. Chemiluminescent signal was developed with ECL Plus (GE Healthcare) and recorded on a Storm 840 imager. Band quantitation was performed with ImageQuant TL 7.0 software (GE Healthcare) using Rolling Ball background subtraction set to a radius of 100.
For localization of PfA-M1 by immunofluorescence, parasitized erythrocytes were fixed in 4% paraformaldehyde, 0.0075% glutaraldehyde in PBS, permeabilized with 0.1% Triton X-100 in PBS, and blocked as previously described (39). Fixed cells were labeled with affinity-purified anti-PfA-M1 antibodies (0.5 g/ml) followed by an Alexa 594-conjugated anti-rabbit IgG secondary antibody (2 g/ml; Invitrogen). Alternately, dried thin smears of parasitized erythrocytes were fixed with 1:1 methanol/ethanol at Ϫ20°C for 2 min, washed with PBS, blocked with 3% bovine serum albumin in PBS, and labeled as described for aldehyde-fixed parasites. In both cases samples were mounted in Prolong Gold (Invitrogen) containing 4Ј,6diamidino-2-phenylindole (DAPI). Images were collected on a Zeiss Axioimager M1 fluorescence microscope equipped with an Axiocam MRm digital camera and a 100ϫ/1.4NA objective lens. Contrast was adjusted using Adobe Photoshop.
Cryo-immunoelectron microscopy was carried out at the Molecular Microbiology Imaging Facility, Washington University, St. Louis, MO. Parasite-infected erythrocytes were enriched on a MACS magnetic LD column (40), fixed with 4% paraformaldehyde and 0.1% glutaraldehyde, and labeled with affinity-purified anti-PfA-M1 polyclonal antibody (0.4 g/ml) as previously described (11). Image contrast was adjusted with Adobe Photoshop.

Expression of PfA-M1 N-terminal Extension (NTE)-Yellow Fluorescent Protein (YFP) and NTE(M1L)-YFP Fusions in
Parasites-The YFP fusions were introduced into the genome of P. falciparum using the mycobacterium Bxb1 integrase system (41). The plasmid pLN-ENR-GFP (41) was modified to contain the NTE-YFP fusion with expression driven by the PfA-M1 promoter as follows. The green fluorescent protein sequence in pLN-ENR-GFP was replaced with the citrine allele of yellow fluorescent protein (24,42) by PCR amplification of citrine from plasmid pPM2CIT2 (24) with forward primer 5Ј-GTACGACTAGTCTCGAGATCGCCTAGGGAAAATTTA-TATTTTCAA (SpeI, XhoI, and AvrII sites are underlined) and reverse primer 5Ј-GTACGCTTAAGGCGGCCGCTTAACT-TCCTCCTAATCCTGCAT (AflII and NotI sites are underlined). The PCR product was digested with SpeI and AflII and ligated into the AvrII and AflII sites of pLN-ENR-GFP to generate pLN-YFP. DNA coding for residues 1-200 of PfA-M1 as well as the preceding 25 bases of the 5Ј-untranslated region was then amplified from genomic DNA by PCR with forward primer 5Ј-GTACGCTCGAGTATATATTTGTATATATAT-TACAAAATGAA and reverse primer 5Ј-GTACGCCTAGGA-TAATGTATTTTTGGTTCGTTCTTTTTAAC, digested with XhoI and AvrII (sites are underlined), and cloned into the same sites in pLN-YFP to generate pLN-NTE-YFP. The calmodulin promoter was replaced with the 5Ј-untranslated region of PfA-M1 (bases Ϫ868 to Ϫ22), which was PCR-amplified from genomic DNA with forward primer 5Ј-GTACGGGGCCC-AATAAATTATTTCTATTGATATAACAATAC and reverse primer 5Ј-GTACGCTCGAGTATATAAAAAAAAAAAAAT-TAAATAAAAATTA, digested with ApaI and XhoI (sites are underlined) and cloned into the same sites in pLN-NTE-YFP to generate pAPNp-NTE-YFP. A second plasmid in which the initial Met codon of the NTE was changed to the Leu codon TTG was generated by PCR amplifying the NTE with the forward primer 5Ј-GTACGCTCGAGTATATATTTGTATA-TATATTACAAATTGAAATTAACAAAAGGCTGTGCCT-ATAA (mutated codon in italics) and the reverse primer above. The product was cloned into the XhoI and AvrII sites of pAPNp-NTE-YFP to generate pAPNp-NTE(M1L)-YFP.
The expression plasmids were co-transfected with pINT into the parasite line 3D7 attB , which has an attB site integrated into the cg6 locus, as previously described (41). Parasites were selected with 2.5 g/ml blasticidin, and resistant parasites were cloned by limiting dilution; NTE-YFP clone D3 and NTE(M1L)-YFP clone E7 were used for this study. Integration of pAPNp-NTE-YFP at the genomic attB locus was confirmed by Southern blotting (supplemental Fig. S1). Microscopy was performed with live parasites using the Zeiss AxioImager described above. For NH 4 Cl treatment, parasites were transferred into RPMI medium supplemented with 50 mM NH 4 Cl (28) and maintained at 37°C in a 5% CO 2 incubator. The total time that parasites were exposed to NH 4 Cl, including imaging, was no longer than 75 min. DNA was visualized by adding 1 M Hoechst 33342 to parasite cultures immediately before imaging. Images were converted to TIF files and contrast was adjusted in Adobe Photoshop.
Parasite Fractionation-For analysis of membrane association, a saponin-treated trophozoite pellet was resuspended in PBS supplemented with 10 M N-(trans-epoxysuccinyl)-L-leucine 4-guanidinobutylamide, 10 M pepstatin, and 1 mM 4-(2aminoethyl)benzenesulfonyl fluoride and lysed by sonication. The lysate was split into two aliquots and centrifuged at 100,000 ϫ g for 1 h at 4°C. The supernatant and pellet fractions of one aliquot were separated and set aside. The pellet of the second aliquot was resuspended in 0.1 M sodium carbonate, pH 11, using a Dounce homogenizer and then centrifuged as above. Both sets of supernatant and pellet fractions were prepared in SDS-containing buffer and analyzed by immunoblotting. Amounts of sample loaded were normalized to that of the crude trophozoite lysate.
Stability Assays-Recombinant PfA-M1 (77 ng) was diluted from a stock solution into 100 mM buffer (Tris-HCl, pH 8.5 and 8.0, sodium HEPES, pH 7.5, sodium MES, pH 5.5, sodium acetate, pH 5.0) containing 100 mM NaCl and 0.1% Triton X-100 to a final volume of 200 l and incubated at 37°C. Immediately after the addition of enzyme and at 10-min intervals thereafter, 10-l aliquots were transferred to 140 l of ice-cold 100 mM HEPES, pH 7.5, containing 100 mM NaCl and 0.1% Triton X-100 and stored on ice. After all samples had been collected, they were warmed to 37°C and mixed with 50 l of 100 mM HEPES, pH 7.5, 100 mM NaCl, 0.1% Triton X-100 containing 300 M arginyl-␤-naphthylamide (Arg-␤NA), and changes in fluorescence were monitored as described in the next paragraph.
Enzyme Assays and Kinetic Analyses-To assess the effects of pH on steady-state kinetic parameters for the hydrolysis of fluorogenic substrates, recombinant PfA-M1 (116 ng) was added to solutions of 100 mM buffer (sodium acetate, pH 5.0; sodium MES, pH 5.5, 6.0, or 6.5; sodium HEPES, pH 7.0 or 7.5; Tris-HCl, pH 8.0 or 8.5) containing 0.1% Triton X-100, 4% dimethyl sulfoxide, sufficient NaCl to bring the total ionic strength to 150 mM, and Arg-␤NA or Ala-␤NA concentrations between ϳ0.2 K M and 5 K M . Assays were carried out in 200-l final volumes at 37°C. Changes in fluorescence were detected using a Victor 3 microplate fluorometer (PerkinElmer Life Sciences) with excitation and emission wavelengths of 340 and 410 nm, respectively, and converted to rates of ␤-naphthylamine production by reference to standard ␤-naphthylamine solutions prepared in the assay buffers to account for pH-dependent changes in fluorescence yield. Kinetic parameters were determined by non-linear regression fit to the Michaelis-Menten equation ϭ Vs/(K m ϩ s) using Kaleidagraph 4.1 (Synergy Software) where V is the limiting velocity and s is the substrate concentration. k cat was calculated from the relationship V ϭ k cat [E]. In some cases, substrate inhibition was observed; therefore, initial rates were fit to the equation for uncompetitive substrate inhibition, where ϭ Vs/(K m ϩ s ϩ s 2 /K si ), and K si is the inhibition constant. Rates were fit to the Michaelis-Menten equation as described in the above paragraph. Inhibition of PfA-M1 by bestatin and bestatin methyl ester was assayed at pH 5.5 and 7.5 with the substrate Arg-␤NA, and inhibition constants were calculated using the Dixon method (43).
Kinetic analysis of the hydrolysis of the dipeptide substrates Met-Phe and Leu-Leu (Bachem) was conducted by reverse phase ultra-performance liquid chromatography with precolumn derivatization of the amino acid products. Recombinant PfA-M1 (4 -115 ng) was preincubated in 100 mM buffer (sodium HEPES, pH 7.5; sodium MES, pH 5.5; sodium acetate, pH 5.0) and 110 mM NaCl for 10 min at 37°C. Reactions were initiated by the addition of substrate to give a final volume of 50 l. After 15 min at 37°C, reactions were stopped by the addition of 100 l of 500 mM sodium borate, pH 9.5. 15-l aliquots of the stopped reactions were mixed with 5 l of the primary amine derivatization reagent AccQ Tag Ultra (6-aminoquinolyl-N-hydroxysuccinimidyl carbamate; Waters) and injected onto a reverse phase Acquity C18 column (2.1 ϫ 100 mm, 1.7-m particles) equilibrated with 5% AccQ Tag Ultra Eluent A on an Acquity Ultra Performance Liquid Chromatography system. Derivitized substrate and products were resolved using a gradient of increasing AccQ Tag Ultra Eluent B solution, detected by absorbance at 216 nm, and quantified using Empower software (Waters). Product concentrations were determined from a standard curve of derivatized phenylalanine or leucine.
Mass Spectrometry-Purified native PfA-M1 preparations enriched in p120 or p68/p35 were resolved on a denaturing, reducing 7.5% polyacrylamide gel. Gel slices containing individual PfA-M1 polypeptides (p120, p68, or p35) were treated with 10 ng/l trypsin or endoproteinase Glu-C in 25 mM ammonium bicarbonate overnight at 37°C. The digests were analyzed on a 4800 MALDI TOF/TOF mass spectrometer (Applied Biosystems) with a matrix of 4 mg/ml ␣-cyano-4-hydroxy cinnamic acid in 50% (v/v) acetonitrile supplemented with 0.2% (v/v) trifluoroacetic acid and 20 mM ammonium citrate. A mass spectrum was obtained for each digest in reflectorpositive operating mode for the mass-to-charge range of 800 -4000, averaging data from ϳ1000 individual laser shots. Tandem mass spectra were then obtained for significant peaks observed in the MS spectrum utilizing the MS/MS 1-kV positive operating mode. Each tandem mass spectrum was typically the sum of ϳ1500 individual laser shots. A peak list for each digest was generated using 4000 Series Explorer software. Peak lists were submitted to a local Mascot Server Version 2.2 (Matrix Science Inc.) search engine using a data base containing PfA-M1. The tandem mass spectra were also validated manually by ensuring that the spectra contained at least four consecutive -y or -b ions matching the predicted amino acid sequence.

RESULTS
Cellular Distribution of PfA-M1-To assess the cellular distribution of endogenous, untagged PfA-M1 at high ultrastructural resolution, we localized the enzyme by cryo-immunoelectron microscopy using polyclonal affinity-purified anti-PfA-M1 antibodies. Labeled parasite sections clearly reveal that PfA-M1 is present throughout the lumen of the food vacuole and in the nucleus ( Fig. 1A and supplemental Fig. S2). The few gold particles observed outside of these two compartments might represent newly synthesized PfA-M1 before delivery to the food vacuole or nucleus. A similar distribution of PfA-M1 was observed in immunofluorescence assays with parasites fixed with either paraformaldehyde/glutaraldehyde (Fig. 1B) or 50% methanol/50% ethanol (supplemental Fig. S3). In contrast, fixation conditions that were close to those used previously to localize PfA-M1 (e.g. 25% methanol, 75% acetone, Ϫ20°C (15, 17)) yielded a diffuse distribution across the parasite (data not shown). This result suggests that the previously reported apparent cytosolic distribution (15,17) might stem from the use of fixation conditions that did not effectively preserve the relevant cellular compartments.
Role of the N-terminal Extension in Dual Targeting-PfA-M1 contains a unique, 194-amino acid NTE that precedes the conserved M1-family domains ( Fig. 2A, supplemental Fig. S4). This sequence is present in PfA-M1 orthologs from other Plasmodium species but, unlike the four M1 domains, is not highly conserved (supplemental Fig. S4). One feature that is conserved is a hydrophobic stretch of amino acids near the N terminus that could serve as a signal for translocation into the endoplasmic reticulum (Fig. 2B). Therefore, one possible role for the NTE could be to specify the dual targeting of PfA-M1 to the vacuole and the nucleus. To test this idea, we fused the N-terminal 200 amino acids of PfA-M1 to YFP and introduced the fusion into the parasite genome by Bxb mycobacteriophage integrase-mediated recombination (Ref. 41; supplemental Fig.  S1). Transgene expression was driven by the PfA-M1 promoter from a single copy of the expression cassette.
In microscopic images of live parasites, YFP was observed in two locations: the food vacuole and the cytosol (Fig. 2C). The accumulation of the NTE-YFP fusion in the cytosol rather than in the nucleus suggests that a nuclear localization signal resides outside of the NTE (i.e. in the four M1-family domains of PfA-M1). In some parasites it was difficult to discern whether YFP was present in the food vacuole due to the large size of the hemozoin crystal relative to the size of the vacuole. In addition, YFP fluorescence is reduced in the acidic food vacuole as the fluorescent form of YFP is in equilibrium with a protonated, non-fluorescent form with a pK a of 5.7 (42). We found that subjecting parasites to a short (45-75 min) treatment with culture medium containing 50 mM NH 4 Cl greatly facilitated the discrimination of vacuolar and extravacuolar fluorescence. As described previously (28), exposure to NH 4 Cl raises the pH of the food vacuole, thereby increasing the intensity of vacuolar YFP fluorescence. In addition, the vacuole increased in size and could be easily identified in phase contrast images of parasites. NH 4 Cl-treated parasites expressing the NTE-YFP fusion clearly exhibited vacuolar YFP fluorescence, which was readily distinguished from the cytosolic YFP fluorescence (Fig. 2D).
One possible mechanism for the dual targeting of PfA-M1 is the use of an alternate translation start site (44). Inspection of the NTE sequence indicates the presence of five potential alternate initiation (Met) codons that occur C-terminal to the signal peptide (Fig. 2B). Initiation at one of these internal Met residues could generate an NTE lacking the signal peptide. In the context of full-length PfA-M1, this species would initially be cytosolic but would be rapidly imported into the nucleus. To assess the plausibility of this model for dual targeting of PfA-M1, we mutated the initial Met codon of the NTE-YFP fusion to Leu and expressed the mutated fusion (designated NTE(M1L)) in parasites. The presence of YFP fluorescence in the cytosol (Fig.  2, C and D) is consistent with the use of an internal Met codon for translation initiation. Furthermore, YFP fluorescence appeared to be diminished or absent from the food vacuole (Fig.  2, C and D), an observation that is consistent with the expected absence of a signal peptide from the NTE(M1L)-YFP fusion. These results suggest that the NTE contributes to the dual targeting of PfA-M1 by enabling the translation of PfA-M1 variants with and without the N-terminal signal peptide.
Characterization of Endogenous PfA-M1 Polypeptides-Polyclonal antibodies against recombinant PfA-M1 recognized three polypeptides in SDS protein extracts of whole trophozoites (Fig. 3A). Two of these, named p120 and p68, have been described previously (15). A third species, designated p35, was detected with our antibody. To gain insight into the nature of the p120, p68, and p35 species, they were purified from saponin-treated parasites, resolved on an SDS-polyacrylamide gel, digested with trypsin or endoproteinase Glu-C, and subjected to tandem mass spectrometry. Results for all three species are presented in supplemental Fig. S5 and Table S1. Peptides from p120 mapped to all four conserved M1-family domains of PfA-M1. Peptides obtained from p68 corresponded to sequences in domains I, II, III, and the first 40 residues of domain IV but not to the C-terminal 290 amino acids of domain IV. Rather, all peptides originating from the p35 fragment were located in the C-terminal 290 residues of domain IV. These data indicate that the p68 and p35 polypeptides are N-and C-terminal fragments of PfA-M1, respectively. One peptide from the tryptic p35 digest appeared to derive from the N terminus of the p35 fragment, as the first residue of this peptide (Leu-796) followed an Asn residue rather than the Arg or Lys residue expected for trypsin cleavage (supplemental Table S1). In the crystal structure of recombinant PfA-M1, Leu-796 lies in a loop between two domain IV ␣-helices (18) and appears to be accessible to proteolysis. Cleavage at Leu-796 would produce a C-terminal polypeptide with a predicted mass of 34.2 kDa, which agrees well with the mobility of the p35 fragment on denaturing, reducing polyacrylamide gels (Fig. 3A). We propose that the p68 and p35 fragments are generated by proteolytic cleavage within the loop sequence containing Leu-796. No peptides from the NTE were recovered in our study; however, anti-peptide antibodies that were raised against the NTE (residues 111-123) recognize p120 but not p68 (15). Thus, the p120 species appears to retain at least some of the PfA-M1 NTE (Fig. 3C and Ref. 15).
We have previously generated a parasite line in which the genomic copy of the PfA-M1 coding sequence was modified to encode a C-terminal HA epitope tag (24). Immunofluorescence localization with anti-HA antibodies revealed the presence of PfA-M1-HA in the nucleus; however, the HA tag was not present in the food vacuole, presumably due to its cleavage and degradation by vacuolar proteases (24). To determine whether Met residues that could serve as alternate translation initiation sites are indicated in bold. The first residue of YFP is indicated by a star. C and D, live parasites expressing either NTE-YFP or NTE(M1L)-YFP were imaged without (C) or with (D) brief exposure to culture medium containing 50 mM NH 4 Cl. The two columns on the left side show fluorescence from YFP and Hoechst 33342 (DNA). In the YFP images the food vacuole is indicated with an arrowhead. In D, the two right-hand columns are identical phase contrast images; in the far right column, the food vacuole membrane is outlined in white. Scale bar, 2 m. p120 and/or p35 retained the HA tag (p68, being an N-terminal fragment, would not possess the tag), SDS extracts of parasites expressing PfA-M1-HA were subjected to immunoblotting with an anti-HA antibody. Only p120 possessed the HA tag (Fig.  3A). This finding strongly suggests that this species is the nuclear form of PfA-M1. We propose that p68/p35 is present in the proteolytic environment of the food vacuole based on the apparent loss of the NTE, the cleavage of the domain IV loop, and the loss of the HA tag from p35 derived from PfA-M1-HA; however, we have not yet been able to generate unambiguous experimental support for this idea.
Because pre-PfA-M1 has a hydrophobic sequence of amino acids close to the N terminus of the NTE (Fig. 2B, supplemental Fig. S4; Ref. 16), we examined whether p120 or p68/p35 associated with membranes. Immunoblot analysis of soluble and membrane fractions from a lysate of saponin-treated trophozoites revealed that most of the p120 and p68/35 resides in the soluble fraction (Fig. 3B). Much of that remaining in the membrane fraction was released by high pH carbonate treatment, which implies a peripheral interaction with membranes. As a control, the fractionation of plasmepsin V was analyzed, and this single-transmembrane helix integral membrane protein (38) was found to reside mainly in the carbonate-insoluble membrane fraction (Fig. 3B).
Comparison of Steady-state Kinetic Parameters of Recombinant PfA-M1 and Native PfA-M1 Isoforms-Native PfA-M1 was obtained from saponin-treated trophozoites that express an HA-tagged form of the M17-family leucine aminopeptidase PfLAP (24). PfLAP and PfA-M1 both catalyze the hydrolysis of fluorogenic aminopeptidase substrates (45); thus, the presence of PfLAP could potentially confound a kinetic analysis of PfA-M1. Purification of PfA-M1 to near homogeneity was achieved by ion exchange and size exclusion chromatography (supplemental Fig. S6). PfLAP was separated from PfA-M1 on the first ion exchange column (supplemental Fig. S6). The p68 and p35 polypeptides co-eluted on the size exclusion column, which suggests that these two species form a stable complex. The p120 isoform of PfA-M1 was partially resolved from p68/p35, which enabled the isolation of fractions highly enriched in either p120 or p68/p35 (supplemental Fig. S6). To generate recombinant PfA-M1, residues 192-1085 were fused to an N-terminal hexahistidine tag followed by a TEV protease site and expressed in E. coli. The PfA-M1 N-terminal extension was omitted from the recombinant protein to best represent the p68/p35 isoform, which has lost much of the NTE (15). After metal affinity chromatography, the hexahistidine tag was removed by TEV protease treatment, and the cleaved protein was further purified (supplemental Fig. S6).
The steady-state parameters for hydrolysis of the amide bond in the fluorogenic substrate Arg-␤NA at pH 7.5 and 37°C were determined for p68/p35, p120, and recombinant PfA-M1 ( Table 1). The p68/p35 and p120 isoforms had very similar values of k cat , K m , and k cat /K m , which indicates that neither the presence of an extended NTE in p120 nor the cleavage that takes place to generate the p68 and p35 fragments has a dramatic effect on catalysis. The catalytic efficiency of recombinant PfA-M1 at pH 7.5 was identical within experimental error to that of p68/p35. Because the quantities of purified native enzyme were very limited, analysis of the effects of pH on PfA-M1 activity was conducted with the recombinant enzyme. Effects of pH on PfA-M1 Stability, Catalysis, and Inhibition-Most experimental determinations of food vacuole pH fall in the range 5.2-5.7 (26 -29). To assess the feasibility of a catalytic role for PfA-M1 in the food vacuole, the stability of PfA-M1 activity at 37°C over the pH range 5.0 -8.5 was determined. PfA-M1 exhibited high stability within this range (Fig. 4A), with Ն90% of the starting activity remaining after 1 h at all pH values tested.
Steady-state parameters for cleavage of Arg-␤NA by PfA-M1 were determined over the pH range 5.5-8.5. The turnover number (k cat ) was not strongly affected by pH; this parameter changed little over the pH range 6.0 -8.0 and dropped by a factor of 2 at pH 5.5 (Fig. 4B). In contrast, the Michaelis constant (K m ) increased below pH 6.5 and was 7-fold higher at pH 5.5 than at pH 7.5 (Fig. 4C, Table 1). The net effect of these changes on the catalytic efficiency (k cat /K m ) was a 10-fold decrease at pH 5.5 versus 7.5; however, the catalytic efficiency at pH 5.5 remained high at ϳ10 4 M Ϫ1 ⅐s Ϫ1 . The catalytic efficiency at pH 5.0 (estimated by using a substrate concentration that is much lower than the K m value) was more than a magnitude lower than that at pH 5.5 but still greater than 10 2 M Ϫ1 ⅐s Ϫ1 (Fig. 4B). Because the substrate used in these analyses, Arg-␤NA, has a positively charged side chain, it was possible that the effect of pH on catalysis was due to interactions involving the substrate side chain. Thus, we examined the effect of pH on the hydrolysis of alanyl-␤-naphthylamide. The decrease in catalytic efficiency at pH 5.5 compared with 7.5 was a similar order of magnitude as that with arginyl-␤-naphthylamide (Table 1).
Because the ␤-naphthylamide substrates differ substantially in structure from physiological peptide substrates, we assessed the effect of pH on PfA-M1-catalyzed hydrolysis of two unmodified dipeptides, Met-Phe and Leu-Leu. Both substrates are found in the sequence of adult human hemoglobin (once and 4 times, respectively) and are, therefore, potential physiological substrates for PfA-M1. Michaelis-Menten parameters were determined at pH 7.5, 5.5, and 5.0. As observed with the fluorogenic substrates, the K m values increased as the pH decreased, whereas the k cat values were similar at pH 7.5 and 5.5. The catalytic efficiency of hydrolysis of Met-Phe at pH 5.5 was more than an order of magnitude higher than those for the fluorogenic substrates, which suggests that engagement of the S1Ј pocket by the substrate P1Ј side chain can help to drive the hydrolysis of peptides at acidic pH. At pH 5.0, the effects of pH on k cat were more complex; for Met-Phe, the value did not change appreciably from that at pH 5.5, whereas for Leu-Leu it dropped around 6-fold. Nevertheless, at pH 5.0 the catalytic efficiency of dipeptide hydrolysis remained in the range of 10 3 -10 4 M Ϫ1 ⅐s Ϫ1 .
To further probe the relationship between pH and substrateenzyme interaction, we evaluated the effect of pH on inhibition of PfA-M1 by the substrate analogs bestatin and bestatin methyl ester. The inhibition constant (K i ) for bestatin increased 30-fold at pH 5.5 compared with pH 7.5 ( Table 2). To determine whether the negative charge of the C-terminal carboxylate group of bestatin contributed to the pH effect, K i values were determined for a variant of bestatin in which the carboxylate group was methylated and uncharged. The potency of bestatin methyl ester was somewhat reduced relative to that of bestatin at both pH 5.5 and 7.5; however, the change in potency of bestatin methyl ester at pH 5.5 versus 7.5 was about the same as that for bestatin ( Table 2).

DISCUSSION
Aminopeptidases of the M1 family have evolved to fill a broad spectrum of biological roles in all three domains of life. In humans, M1 family members are key mediators of blood pressure regulation, neuropeptide levels, and antigen processing, to name just a few examples (46). Our studies reveal that the sole M1-family aminopeptidase in P. falciparum, PfA-M1, has been recruited to fulfill roles specific to the physiology of the intraerythrocytic parasite. We propose that one of these roles is the generation of amino acids from globin peptides in the food vacuole lumen. The accumulation of PfA-M1 in the nucleus implies a second function, the details of which remain to be elucidated. These roles differ from what has been proposed previously for this protein. In this section we interpret our results alongside those from prior studies.
Early immunofluorescence localization studies of PfA-M1 indicated a cytosolic distribution with possible accumulation around the food vacuole (15,17). In a previous study we tagged endogenous PfA-M1 with YFP and localized the fusion in live parasites (24). The YFP tag was observed to concentrate in two subcellular structures, the food vacuole, and the nucleus. To resolve the discrepancy between these and the earlier results, we undertook immunoelectron and immunofluorescence studies of native (untagged) PfA-M1 using affinity-purified polyclonal antibodies. These analyses support our earlier conclusion that PfA-M1 resides primarily in the food vacuole and the nucleus. Importantly, the immunoelectron microscopy and membrane fractionation studies reveal that PfA-M1 is a soluble enzyme distributed throughout the lumen of the food vacuole. Thus, it seems unlikely that PfA-M1 is anchored into the food vacuole membrane with the catalytic domains outside the vacuole or in vacuole-associated vesicles, as has been previously proposed (16,25). By examining a range of fixation conditions for indirect immunofluorescence, we also observed vacuolar and nuclear concentrations of PfA-M1 with this technique. However, some fixation conditions appeared not to preserve the relevant subcellular structures, giving rise to a diffuse distribution of PfA-M1 that could be interpreted as cytosolic.
One unique feature of PfA-M1 is the presence of a 194-residue N-terminal extension. The precise functions of the NTE are unknown; however, the presence of a hydrophobic putative signal sequence suggests that it plays a role in specifying the subcellular distribution of PfA-M1. To determine whether the NTE was sufficient to specify dual vacuolar/nuclear targeting, an NTE-YFP fusion was expressed in parasites under control of the PfA-M1 5Ј-untranslated region. YFP was present in the food vacuole of transfected parasites, which indicates that the NTE is sufficient for trafficking to this organelle. In a separate study the first 30 amino acids of the NTE were appended to the N terminus of green fluorescent protein. This fusion, which contained the putative signal peptide, was trapped in the endoplasmic reticulum (16). Thus, the signal peptide is sufficient to direct entry into the endoplasmic reticulum, whereas further transit to the food vacuole requires the presence of a longer NTE, which may encode a food vacuole targeting signal. In contrast, the NTE appeared not to be sufficient to direct the fusion to the nucleus; rather, it had a diffuse, apparently cytosolic distribution. Thus, it appears that the sequence that causes PfA-M1 to localize to the nucleus resides in the conserved M1 domains of PfA-M1 rather than in the NTE.
Given the presence of multiple Met residues in the NTE, we asked whether initiation of translation at an internal Met codon beyond the signal peptide could account for the presence of NTE-YFP in the cytosol. Mutation of Met1 of the NTE-YFP fusion to Leu appeared to diminish the amount of vacuolar fusion while preserving the cytosolic pool. These data are consistent with a mechanism for dual targeting that involves the use of alternate translation initiation sites to generate variants of PfA-M1 that are differentially targeted. According to this model, translation initiation at the first Met codon would generate a form of PfA-M1 that contains the signal peptide, enters the endomembrane system, and ultimately reaches the food vacuole. Initiation of translation at an internal Met codon would result in a form of PfA-M1 that lacks the signal peptide, thereby re-directing PfA-M1 from the endomembrane system to the cytosol. In the case of full-length PfA-M1, a nuclear localization signal would then lead to nuclear import. Whether the choice of initiation codon is dictated by ribosomal skipping (alternate translation) or the generation of alternate transcripts cannot be distinguished from our data. Future studies will attempt to better define the mechanism of initial codon selection and will attempt to locate the nuclear localization signal. Our results are consistent with a prior report of expression of an N-terminal-truncated form of PfA-M1 in P. falciparum in the cytosol but not the food vacuole (18).
Previously, three isoforms of PfA-M1 have been reported: p120, p96, and p68. We have identified a new species, p35, and demonstrate that p68 and p35 are N-and C-terminal fragments of PfA-M1, respectively. The p68 and p35 species, which are generated by proteolytic cleavage of a domain IV loop, remain associated during purification and exhibit catalytic activity that is very similar to that of the p120 species. A schematic diagram of the p120 and p68/p35 isoforms that is consistent with existing data is shown in Fig. 3C. We propose that the p120 form resides in the nucleus. The p68/p35 complex may inhabit the food vacuole; however, the absence of a unique tag on this PfA-M1 isoform makes this proposition difficult to test. The p96 isoform of PfA-M1 was initially thought to arise from artifactual proteolytic cleavage of p120 (15), but more recently it has been suggested that p96 is generated in the parasitophorous vacuole during trafficking to the food vacuole (16). We do not observe PfA-M1 in the parasitophorous vacuole in live cells expressing PfA-M1-YFP or in fixed parasites analyzed by immunofluorescence; however, it is possible that the sensitivity of these experiments is inadequate. We do sometimes observe the p96 species in immunoblots of parasite lysate and PfA-M1 purifications (supplemental Fig. S6, A and B), but the amounts are minor in comparison to p120 and p68/p35. It has been asserted that PfA-M1 does not have adequate activity at pH values below 6 to function as a catalytic aminopeptidase within the food vacuole (16,18,25). Reflecting this line of thinking, the enzyme has been referred to as a "neutral" aminopeptidase (45). To assess the validity of this proposal, we undertook a detailed kinetic characterization of the catalytic properties of PfA-M1 over the pH range 5.0 -8.5 using both fluorogenic aminoacyl-␤-naphthylamide and unmodified dipeptide substrates. The latter are present in the sequence of hemoglobin and are, therefore, potentially physiological substrates. The k cat values for PfA-M1-catalyzed hydrolysis of both sets of substrates were relatively insensitive to pH within the range 5.5-8.5, whereas the K m increased as the pH dropped below 6.5. Importantly, catalytic efficiencies (k cat /K m ) for hydrolysis of the two dipeptide substrates were Ͼ10 4 M Ϫ1 ⅐s Ϫ1 at pH 5.5 and Ͼ10 3 M Ϫ1 ⅐s Ϫ1 at pH 5.0. For comparison, the k cat /K m values for the hydrolysis of several peptide substrates by P. falciparum aminopeptidase P, an enzyme also located in the food vacuole, are around 10 4 M Ϫ1 ⅐s Ϫ1 at pH 5.5 (13). Given that most experimental measurements of food vacuole pH fall into the range of 5.2-5.7 (26 -29), it would appear that the catalytic efficiency of PfA-M1 is sufficient for a role in the production of amino acids from globin-derived peptides. The in vivo substrates of PfA-M1 are unknown but likely consist of a diverse pool of di-and oligopeptides generated by vacuolar enzymes including falcilysin, dipeptidyl aminopeptidase 1, and aminopeptidase P (10 -13). Our data support a model for hemoglobin catabolism in which amino acids are generated in the food vacuole lumen. We cannot exclude the possibility that some (di)peptides move from the vacuole lumen to the cytosol; however, it seems unlikely that peptide transport is an obligate step for amino acid production. Rather, our model leads to the prediction of one or more amino acid transporters in the food vacuole membrane.
So why has it been reported that PfA-M1 has negligible activity at pH values below 6 (18)? Prior analyses of the pH-activity relationship were conducted with a single concentration of a fluorogenic substrate, and the results were reported as fractional activity (16,18). These types of studies do not yield the catalytic efficiency of the enzyme and may lead to unwarranted conclusions (47). Our data are consistent with the proposal that PfA-M1 does not operate at its pH optimum in the acidic food vacuole lumen. The relevant question, however, is not whether activity is optimal but rather whether it is sufficient for the purpose at hand. In this regard, we believe that the observation that k cat does not change dramatically down to pH 5.5 is a key insight, as the pH-dependent increase in K m can be compensated for by high substrate concentrations. Indeed, the concentration of the hemoglobin tetramer in the erythrocyte cytosol is 5 mM (48), which likely leads to high micromolar/low millimolar concentrations of peptides in the vacuole. It is plausible that this high concentration drives substrate binding to PfA-M1 and leads to the efficient generation of amino acid products. PfA-M1 is not the first M1-family aminopeptidase for which a role in an acidic environment is proposed; the mammalian enzyme aminopeptidase B processes neuropeptides in acidic secretory vesicles and retains substantial activity at pH 5 (49).
Why does the substrate K m increase with pH? The pH dependence of inhibition by bestatin and bestatin methyl ester provides a clue. A co-crystal structure of PfA-M1 with bestatin reveals that the positively charged N terminus of the inhibitor interacts with a pair of glutamate side chains in the active site (18). One explanation for the increase in inhibitor K i and substrate K m at acidic pH could be the protonation of one or both of these glutamates, which would likely destabilize the enzymesubstrate or enzyme-inhibitor complex. These results have important implications for inhibitor discovery against this enzyme. Efforts to identify new inhibitor chemotypes (for example, by high throughput screening) should be carried out at pH ϳ 5.5 to best represent the electrostatic environment of the active site as it occurs in the food vacuole.
A lingering question is the role of PfA-M1 in the nucleus. At present, it is not possible to propose a specific function; however, there are interesting precedents for participation of M1-family aminopeptidases in nuclear processes. Mammalian puromycin-sensitive aminopeptidase has been localized to the nucleus and was found to associate with mitotic spindles (50). Interestingly, the puromycin-sensitive aminopeptidase inhibitors puromycin and bestatin inhibited DNA synthesis and blocked the cell cycle in cultured COS cells. Homologs of puromycin-sensitive aminopeptidase have also been implicated in homologous recombination during meiosis in plants (51) and in meiotic exit and anteroposterior axis formation in Caenorhabditis elegans (52). Investigation into the role of PfA-M1 in the P. falciparum nucleus is in progress.