1Division of Infectious Diseases and Immunology, Queensland Institute of Medical Research, Queensland, and 2Department of Microbiology and Parasitology, 3Institute for Molecular Bioscience, University of Queensland, Brisbane, Queensland, Australia; 4Department of Tropical Medicine, Tulane University, New Orleans, Louisiana, 5Department of Microbiology and Tropical Medicine, George Washington University Medical Center, Washington, DC; 6School of Biosciences, Cardiff University, Cardiff, and 7Boots Science Institute, University of Nottingham, Nottingham, United Kingdom
Received 11 July 2002; revised 7 October 2002; electronically published 24 January 2003.
| Hookworms are voracious blood-feeders. The cloning and functional expression of an aspartic protease, Na-APR-2, from the human hookworm Necator americanus are described here. Na-APR-2 is more similar to a family of nematode-specific, aspartic proteases than it is to cathepsin D or pepsin, and the term "nemepsins" for members of this family of nematode-specific hydrolases is proposed. Na-apr-2 mRNA was detected in blood-feeding, developmental stages only of N. americanus, and the protease was expressed in the intestinal lumen, amphids, and excretory glands. Recombinant Na-APR-2 cleaved human hemoglobin (Hb) and serum proteins almost twice as efficiently as the orthologous substrates from the nonpermissive dog host. Moreover, only 25% of the Na-APR-2 cleavage sites within human Hb were shared with those generated by the related N. americanus cathepsin D, Na-APR-1. Antiserum against Na-APR-2 inhibited migration of 50% of third-stage N. americanus larvae through skin, which suggests that aspartic proteases might be effective vaccines against human hookworm disease.
|
Financial support: Australian Research Council; Ramaciotti Foundation; National Health and Medical Research Council of Australia (NHMRC; block grant to Queensland Institute of Medical Research; grants to D.P.F. and G.A.); Australian Postgraduate Award and University of Queensland Graduate School Research Travel Award (to A.L.W.); Burroughs Wellcome Fund Scholar Award in Molecular Parasitology (to P.J.B.); Wellcome Trust (to D.I.P.); UK Biotechnology and Biological Sciences Research Council (to K.G.); Royal Society Research Fellowship (to C.B.); Bill and Melinda Gates Foundation (to Sabin Vaccine Institute; grant to P.J.H.). A.L. was a NHMRC Howard Florey Research Fellow.
Previous genetic nomenclature for hookworm aspartic proteases and their cDNAs was inconsistent. A uniform system in adherence with standardized nematode genetic nomenclature has since been established. Na-apr-2 represents entry AJ245458, formerly referred to as necepsin I. Ac-apr-1 represents entry U34888, formerly referred to as AcASP. Na-apr-1 represents entry AJ245459, formerly referred to as necepsin II.
Reprints or correspondence: Dr. Alex Loukas, Dept. of Microbiology and Tropical Medicine, Ross Hall, Rm. 726, George Washington University Medical Center, 2300 Eye St. NW, Washington, DC 20037 .
Hookworms are highly host specific, obligate blood-feeders. Infective, third-stage larvae (L3) in soil penetrate the skin and migrate (and mature) through the circulatory system via the lungs to the alimentary canal, where they reside as adult worms in the small intestine. The 2 main species of hookworms that infect humans, Necator americanus and Ancylostoma duodenale, are among the most prevalent soil-transmitted nematodes, infecting 
1.2 billion people worldwide [1]. There are records of N. americanus successfully infecting and maturing in dogs, especially in young pups. However, although infective larvae of N. americanus can penetrate the skin of dogs and migrate to the lumen of the small intestine, almost none mature to adulthood and sexual maturity [2
4]. N. americanus can mature in golden hamsters if weaning pups are infected with L3. However, >99% of larvae do not reach adulthood, and of those that do, survival time in the host gut is only a couple of months [5], compared with years in human hosts [6]. Likewise, the canine hookworm Ancylostoma caninum can penetrate the skin and infect humans, and, occasionally, migrate to the gut, where they can cause zoonotic eosinophilic enteritis. However, in humans, A. caninum infections are short lived, the worms die within a few weeks, and human infections with A. caninum never reach patency [7, 8].
Since hookworms are hematophagous and depend on host hemoglobin (Hb) and serum proteins from ingested blood and mucosal tissues for nutrition, these enzyme-substrate relationships might have evolved to reflect host specificity at the molecular level. Aspartic proteases have been shown to play a crucial role in the degradation of Hb in malarial parasites [9, 10] and schistosomes [11]. We recently showed that gut-derived, cathepsin Dlike aspartic proteases from N. americanus and A. caninum cleaved Hb in a host-specific fashion, providing a molecular example of host-parasite coevolution and its manifestation in the form of restricted host range [12]. Here, we describe the cloning, expression, localization, and host-specific cleavage of Hb and of several serum proteins by Na-APR-2, an aspartic protease expressed by immature and mature adult hookworms that is more similar in primary structure to a distinct family of nematode aspartic proteases (designated here as the nemepsins) than it is to cathepsin D. Moreover, mouse antibodies to the protease partially inhibited infective larvae from penetrating skin in vitro, which suggests that this protease might be targeted as an efficacious antihookworm vaccine.
MATERIALS AND METHODS
Parasites and cDNAs. Adult N. americanus were collected at necropsy of infected hamsters [12] at the George Washington University Medical Center (GWU). Na-apr-2 cDNA ( accession no. AJ245458) was amplified from an adult N. americanus cDNA library (available as pBluescript plasmid in our laboratories) by use of a degenerate primer targeting the active site of eukaryotic aspartic proteases [13]. Soluble adult hookworm proteins (somatic extracts) and excretory/secretory (ES) products of N. americanus were prepared, as described elsewhere [12].
Reverse transcription and reverse-transcription polymerase chain reaction (RT-PCR). RNA was prepared from N. americanus L3, L4, and adult worms, as described elsewhere [14]. cDNA was prepared from 5 g of total RNA from each of the L3, L4, and adult worms of N. americanus with an oligo-dT primer using Moloney murine leukemia virus (MMLV) RT. After denaturation, the single-stranded cDNA samples were used as templates for PCR by use of the Na-apr-2specific primers 5
-CTCGAGAAAAGAGGATCCGGTGTATATAAAATCCCATTG and 5-CTCGAGGATCCTCAATGTTTTACAGCTGCAAA. Primers designed to amplify the transcript for the constitutively expressed calreticulin mRNA of N. americanus [14] and Na-apr-1 mRNA (K.G. and C.B, unpublished data) were included in a PCR as a positive control (data not shown). PCR-amplified products were separated by electrophoresis on a 1% (wt/vol) agarose gel, stained with ethidium bromide, and photographed under UV light.
Expression and analysis of recombinant Na-APR-2. The region spanning the prodomain to the C-terminus of Na-APR-2 was amplified by PCR from a pBluescript template and was cloned into the transfer vector pBacPak6 (Clontech) by incorporating NotI and AscI restriction sites [15]. Recombinant viral expression vectors were generated by cotransfection, which was aided by bacfectin (Clontech) of lepidopteran Spodoptera frugiperda Sf 9 cells with Bsu36 I digested BacPak6 viral DNA and recombinant, modified transfer vector. Recombinant virus was located by Western blotting, using a rabbit antiserum raised against Schistosoma japonicum cathepsin D [11], isolated, and amplified, and the resulting high-titer virus was stock stored at 4°C. Adherent Sf 9 cell cultures were used for the initial plaque purification of recombinant virus and small-scale well amplifications. High Five (Trichoplusia ni) cells were used for large-scale cultures in which the transformed cells were cultured in suspension flasks at 27°C in protein-free medium (JRH Biosciences). Supernatant containing Na-APR-2 was harvested 4 days later. Recombinant protease was affinity purified from culture supernatant on pepstatin-agarose (Pierce), which was autoactivated at low pH, as described elsewhere [11]. Enzymatic activity of recombinant, activated Na-APR-2 was monitored with the substrate peptide o-aminobenzoyl-Ile-Glu-Phe*nPhe-Arg-Leu-NH2 (* indicates the site of cleavage; nPhe, p-nitrophenylalanine), a shortened variant of a fluorogenic peptide substrate series for which human cathepsin D exhibits high affinity [11, 16]. In brief, the substrate was employed at 25 M in 0.1 M sodium formate (pH 2.05.0) or 0.1 M sodium phosphate (pH 5.58.7) in the presence of 1 g of recombinant Na-APR-2 in a final volume of 1.0 mL. Reactions proceeded at 37°C for 1 h, and hydrolysis of o-aminobenzoyl-Ile-Glu-Phe*nPhe-Arg-Leu-NH2 was measured in a fluorometer (Turner Designs TD-700), with excitation at 325 nm and emission at 420 nm. Pepstatin A, a general inhibitor of aspartic proteases, was added at a final concentration of 1.0 M to some experiments 10 min before the addition of the substrate. All assays were conducted in triplicate, and data are presented as the mean values.
Production of antiserum to recombinant Na-APR-2. Antiserum against recombinant Na-APR-2 was generated by immunizing female BALB/c mice with a total of 50 g of protease mixed with 75 g of saponin as adjuvant administered subcutaneously over 5 equal doses at 2 weekly intervals. Blood samples were collected before the immunization and 2 weeks after the final injection. IgG was affinity purified from antiserum and normal mouse serum using Protein G sepharose (Sigma). Various concentrations of purified IgG (5, 50, and 500 ng) were incubated with 1 g of active recombinant hookworm aspartic protease (in 0.1 M sodium formate [pH 5.0]) and were allowed to bind at room temperature for 1 h before assessing enzymatic activity with the fluorogenic substrate o-aminobenzoyl-Ile-Glu-Phe*nPhe-Arg-Leu-NH2 and with Hb (below). All reactions were conducted in triplicate, and data are presented as mean values.
Immunolocalization. Intact, live, adult N. americanus obtained from the small intestines of hamsters were washed in PBS (pH 7.2), incubated in 70% formalin and 0.4% Na2H2PO4 (pH 7.4) (phosphate buffered formalin; PBF), and subsequently irradiated in a microwave oven on high power for 40 s to permeabilize the cuticle and allow for immediate fixation and preservation of internal organs. The fixed hookworms were washed 2 times for 5 min in PBS, dehydrated through a graded series of ethanol (70%100%), cleared in methyl salicylate, and embedded in paraffin using standard preparative procedures [17]. Five micrometer sections were deparaffinized, rehydrated in graded alcohols, and reacted with mouse antiserum to Na-APR-2, followed by peroxidase-conjugated antimouse IgG. Antibody binding was visualized with 3,3-diaminobenzidine (DAB) substrate, as described elsewhere [18]. Sections then were washed, counterstained with hemotoxylin and Scott's Blue for 5 min, dehydrated in an ethanol series (70%100%), cleared in 100% toluene, and mounted in DPX for examination by light microscopy. Two hookworm cross-sections on each slide were exposed to PBS and normal mouse serum (NMS), instead of antiserum, which served as negative controls. In addition to light microscopic visualization, longitudinal sections of adult A. caninum hookworms were probed with mouse antiserum, followed by Cy-3conjugated antimouse IgG (Sigma). Mouse antiAPR-2 serum or NMS (diluted 1 : 100) were applied to each section and were incubated for 2 h at RT in a humidity chamber. Sections were washed 6 times for 5 min in PBS, then goatanti-mouse Cy3conjugated IgG (BD Biosciences; diluted 1 : 500) was applied to each section and incubated for 1 h at RT in a light-proof humidity chamber. Sections were washed 6 times for 5 min in PBS, then mounted in Crystal Mount medium (Biomedia), and baked at 65°C for 20 min. Sections then were viewed with an Olympus BX-60 fluorescent microscope, with excitation at 550 nm and emission at 565 nm.
Western blots. SDS-PAGE and subsequent Western blots were performed by use of standard methods. Samples (1.0 g of recombinant Na-APR-2) were electrophoresed through precast 4%20% gradient gels (Gradipore) under denaturing conditions. Separated proteins were electrotransferred to nitrocellulose membranes, and nonreactive sites were blocked with 5% skim milk in PBS overnight at 4°C. Membranes were incubated in primary serum (either rabbit antiS. japonicum cathepsin D or mouse antiNa-APR-2 diluted 1 : 1000 in PBS/Tween-20) for 1 h at RT, washed (3 times for 5 min) with PBS/Tween-20, followed by incubation with appropriate HRP-conjugated secondary antibodies (1 : 2000) for 1 h at RT. Membranes were washed again and were developed by use of enhanced chemiluminescence (ECL; Amersham Pharmacia) and autoradiography.
Cleavage of Hb by recombinant Na-APR-2. Hb was prepared by lysis of fresh, washed, normal human or dog erythrocytes in hypotonic medium [11]. Hydrolysates of 100 g of Hb (after digestion in the presence of 1 g of recombinant hookworm protease for 530 min) were examined using nondenaturing 15% SDS-polyacrylamide gels stained with Coomassie brilliant blue. Cleavage sites within the Hb molecules were determined by 1 of 2 methods: (1) digestion products were transferred to polyvinylidine difluoride membrane and N-terminal sequencing of cleavage products using automated Edman degradation, or (2) hydrolysates were separated by use of an RPhigh-performance liquid chromatography (HPLC), using 2.1 × 220 mm 5-m Spheri-5 RP-18 column (Brownlee) equilibrated in aqueous trifluoroacetic acid (TFA)/3% acetonitrile that was mounted on a SMART HPLC System (Amersham Pharmacia). Hb peptides were eluted using a linear gradient of acetonitrile from 3% to 60% in aqueous 0.085% TFA over 15-column bed volumes. Peptide-containing fractions were collected, and peptide masses were determined by means of a Perkin Elmer SCIEX Atmospheric Pressure Ionization QSTAR PULSAR Electrospray Quadruple Time-of-Flight mass spectrometer. Mass data, which were analyzed by use of the Biomultiview and FASTA FindPept software packages, were used to predict cleavage sites within the Hb molecules, as described elsewhere [12].
Quantitation of free amino groups of Hb, albumin, fibrinogen, and collagens generated by hydrolysis with Na-APR-2. Release of 
-amino nitrogen from the hydrolysis of Hb by recombinant Na-APR-2 at pH 5.0 was measured by use of ninhydrin, as described elsewhere [12, 19]. In brief, Hb was prepared as described elsewhere [12], and 100 g was incubated with 1.0 g of recombinant protease (with or without 1.0 M pepstatin A) in 0.1 M sodium formate (pH 5.0) in a final volume of 100 L for 530 min at 37°C. Fifty microliters from each assay then was mixed with 50 L of Ninhydrin Reaction Solution (Sigma) and was incubated at 100°C for 15 min. Subsequently, samples were placed on ice, and 250 L of 50% (vol/vol) ethanol was added before measuring absorbance at 570 nm. Quantities of free amino groups released by the proteolysis were determined by use of a standard curve of free L-leucine (0.010.5 nM). Human and dog collagens (I, III, IV, and V), fibrinogen, and serum albumin (10 g of each; Sigma) were incubated with 1.0 g of recombinant Na-APR-2 in 0.1 M sodium formate (pH 5.0) for up to 4 h at 37°C, and free amino termini were assessed as described above. All reactions were conducted in triplicate. To assess the statistical significance of the differences observed, the slopes and intercepts were estimated from a regression of the amount of free amino groups released at 3 time points after the commencement of the hydrolysis reactions. Data were initially tested for normality and homoscedasticity to satisfy the assumptions of linear regression and subsequent tests of inference. Five minutes after initiation of proteolytic digestion was considered as the starting point of the assay, reflecting that it can take several minutes for worms to pump substrate into their intestines, lyse host cells, and begin to digest cellular proteins. The regression parameters were estimated for each assay and were grouped by substrate, aided by the SAS release 8.2 software.
Effect of antiserum to Na-APR-2 on L3 migration. Hamster skin (obtained from a freshly killed hamster) was shaved, scrubbed with gauze, washed with PBS, and stretched over a glass migration tube filled with PBS so that the buffer was in contact with the underside of the skin (adapted from [20]). N. americanus and A. caninum L3 (300 L3/group) were incubated in 50 L of PBS (pH 7.2), undiluted normal mouse serum (NMS), or undiluted mouse antiserum to Na-APR-2 for 30 min at 37°C. Subsequently, each group of L3 was transferred to the hamster skin (in 1.0 mL of PBS) and allowed to migrate for 30 min at RT. L3 that migrated through the skin were collected from the media (PBS) in the collection tube and counted, as were L3 that remained on the outside surface of the skin. Each experiment was conducted in triplicate, and the results were presented as the means of these 3 replicates.
RESULTS
Hookworm pepsinogen-like aspartic proteases. Partial cDNA fragments encoding pepsinogen-like aspartic proteases were amplified by PCR from N. americanus and A. caninum cDNA libraries. These cDNAs were termed Na-apr-2 from N. americanus and Ac-apr-2 from A. caninum. The full-length transcript of Na-apr-2 was cloned; however, despite concerted efforts (using PCR, hybridization, and 5 Rapid Amplification of cDNA Ends strategies), only 421 nt of the 3 sequence of Ac-apr-2 was identified. Moreover, transcripts corresponding to Ac-apr-2 were not identified from the 7625 A. caninum L3 expressed sequence tags in Expressed Sequence Tag database by BLAST searches. Na-apr-2 cDNA was 1366 nt including a polyA tail and contained a putative polyadenylation signal at position 1284. The Na-apr-2 mRNA was expressed in fourth-stage larvae (L4) and adult worms but not in infective larvae (L3) of N. americanus . The predicted protein included 425-aa residues and consisted of an N-terminal signal peptide (16-aa residues) followed by a prodomain (38-aa residues) and the C-terminal mature enzyme. The predicted Na-APR-2 protein was most similar (45% identical) to Pep1 from the related strongylid nematode Haemonchus contortus ( accession no. 079402), 30% identical to the N. americanus cathepsin Dlike protease Na-APR-1 ( accession no. AJ245459) and 29% identical to chicken pepsin C ( protein JE0371; ). Na-APR-2 had 5 predicted motifs for N-linked glycosylation, on the basis of the numbering of the mature enzyme only: Asn-108, Asn-142, Asn-249, Asn-299, and Asn-310. None of the APR-2 glycosylation sites corresponded with predicted glycosylation sites in the closest homologous proteases including H. contortus Pep1. Glycosylation sites of Na-APR-2 were not conserved with the 2 cathepsin Dlike hookworm aspartic proteases [12], which supports previous work that showed variability in aspartic protease N-linked glycosylation sites [21]. The Na-APR-2 sequence, along with the H. contortus and Caenorhabditis elegans sequences , included an atypical cysteine-rich insert, of 35 residues, occurring 15 residues to the C-terminal side of the first catalytic active-site motif, DTG. The insertion sequence was flanked by conserved cysteine residues, Cys 45 and Cys 50, porcine pepsin numbering, that are disulphide-linked in other aspartic proteases. Similar disulphide bonding in Na-APR-2 might act as a point of anchorage and impose a spatial constraint on the energetically favorable N- and C-terminal positions of the inserted residues. The insert may simply loop out in a region of flexible structure (secondary structural predictions indicate 2 extended 
-strands not shown) or may form a rigid domain, with some or all of the 6 cysteines within the loop being involved in disulphide bond interactions or even, potentially, in the binding of metal ions. The N-terminal insertion on the molecular surface also would be highly accessible to the surrounding environment and may play a role in recognition, for example, by targeting the proteins to a particular subcellular location. It has been suggested that the insert in the H. contortus enzyme may function to target or bind the protease to the membrane, possibly via a membrane receptor molecule/membrane bound protein [22].
| fig.ommitted |
Figure 1. Expression of Na-apr-2 mRNA in fourth-stage larvae (L4) and adult worms (Ad) but not in third-stage infective larvae (L3) of Necator americanus. RNAs from L3, L4, and Ad were incubated with Moloney murine leukemia virus reverse transcriptase (RT) (+) or water (-) before amplification of Na-apr-2 by polymerase chain reaction. M DNA markers (sizes in base pairs indicated). X, blank lane; p, cloned cDNA plasmid as template for RT-PCR. |
| fig.ommitted |
Figure 2. CLUSTAL-formatted alignment comparing the mature protease sequences of Na-APR-2 with Haemonchus contortus pepsinogen (Hc-pep-1), C. elegans pepsinogen from cosmid T29692 (T29692), Necator americanus Na-APR-1 (Na-APR-1), and chicken pepsinogen C (chicken). Numbering (S1-S4 |
Na-APR-2 is expressed in the gut and other tissues. Na-APR-2 was expressed and secreted as a proenzyme in Sf 9 lepidopteran cells, eluted from pepstatin agarose affinity resin as a single band (detected by SDS-polyacrylamide gels; ) and used to immunize mice for antibody production. AntiNa-APR-2 serum recognized the native protease in soluble extracts of adult N. americanus, but like antiserum to Na-APR-1 [12], the serum did not bind to any proteins from ES products of adult worms (data not shown). Na-APR-2 and the cathepsin Dlike Na-APR-1 share 30% sequence identity, and antiAPR-2 serum bound to recombinant Na-APR-1 (data not shown). The antiserum to recombinant Na-APR-2 was used to localize expression of the protease to the gut and, to a lesser extent, the amphidial glands and excretory glands of adult N. americanus ( and ), as well as the intestine of A. caninum . The microvillar surface in the hookworm gut lumen stained positively (orange) and the underlying intestinal cells stained less intensely and not at all in some regions. The amphidial glands, excretory glands, and some regions of the cuticle also stained, but reproductive organs did not. A similar pattern of staining was observed for Na-APR-1 [12], and immunological cross-reactivity between a family of Necator aspartic proteases cannot be discounted. APR-2 was detected in somatic extracts but not in ES products of adult N. americanus in Western blots (data not shown).
| fig.ommitted |
Figure 3. Secretion of recombinant Na-APR-2 by insect cells. SDS-polyacrylamide gel stained with Coomassie brilliant blue showing expression of the recombinant secreted Na-APR-2 proenzyme in culture medium of Trichoplusia ni High Five cells (lane 1, note overexpressed band at |
| fig.ommitted |
Figure 4. Na-APR-2 is expressed in the intestinal microvilli, amphidial, and excretory glands (orange staining) of adult Necator americanus. Sections were probed with normal mouse serum (A) or antiserum raised against recombinant Na-APR-2 (B) and were developed with a peroxidase-conjugated secondary antibody and DAB substrate. Longitudinal sections of adult Ancylostoma caninum hookworms also were probed with anti |
Na-APR-2 cleaves an aspartic protease-specific fluorogenic substrate. After autoactivation, Na-APR-2 was proteolytically active, as measured by use of the fluorogenic substrate for cathepsin D, o-aminobenzoyl-Ile-Glu-Phe-nPhe-Arg-Leu-NH2, with a pH optimum of 5.0 . Points on the graph represent the means of triplicate reactions that were highly reproducible (<5% variation at all pH values). Pepstatin A (1.0 M) completely inhibited the proteolytic activity against this peptide.
| fig.ommitted |
Figure 5. Recombinant Na-APR-2 pH profile for cleavage of the fluorogenic peptide o-aminobenzoyl-Ile-Glu-Phe-nPhe-Arg-Leu-NH2 in the presence or absence of the inhibitor pepstatin A. Data are plotted in relative fluorescence units (rfu), and the baseline was set at 500 rfu. Each data point represents the mean value of highly reproducible triplicate reactions. |
Na-APR-2 cleaved human Hb more efficiently than dog Hb. The relative activity of the recombinant Na-APR-2 for Hb was assessed by incubating the protease with Hb from erythrocytes of human (permissive) and canine (nonpermissive) mammalian hosts. The enzyme cleaved both human and dog Hb, but at different rates. Na-APR-2 degraded human Hb 3.3 times as efficiently as dog Hb over 530 min, as determined by regression analysis P < .025). Degradation of human Hb was apparent within 5 min, and significant digestion was evident after 30 min at 37°C. By contrast, canine Hb was only partially degraded after 30 min incubation at 37°C. Na-APR-2 fully degraded dog and human Hb after 24 h. Pepstatin A (1.0 M) completely ablated cleavage of Hb by the protease (data not shown).
| fig.ommitted |
Figure 6. Na-APR-2 preferentially cleaves hemoglobin (Hb) from its permissive host. Recombinant Na-APR-2 generated more hydrolytic fragments from human Hb than from dog Hb at 15 and 30 min (A). Quantitation of free amino groups from dog and human Hb released after hydrolysis by recombinant Na-APR-2 at different time points by use of ninhydrin (B). Data are linear regressions of substrate cleavage over time. When pepstatin A was included in reactions, there was no detectable increase in free amino groups above background levels. |
| fig.ommitted | Table 1. Regression parameters describing the linear trend of digestion of various substrate proteins by Na-APR-2 over time. |
Mapping of the Hb cleavage sites for Na-APR-2. To further investigate the apparent host species-specificity of the hydrolysis of Hb, cleavage sites were determined by the following methods: (1) hydrolysates were resolved by SDS-PAGE, and the major peptides that stained with Coomassie brilliant blue were sequenced by Edman degradation; (2) hyrolysates were separated by RP-HPLC, and hydrolytic peptide masses were determined. Specific cleavage sites were mapped to the - and -chains of human and dog Hb The major cleavage sites (as determined by Edman sequencing of the major peptides resolved by SDS-PAGE) of human Hb were Ala13-Trp14, Asp47-Leu48, and Trp37-Thr38, and the major cleavage sites in dog Hb were Gly51-Ser52, Val93-Asp94, and Gly136-Val137. Like other aspartic proteases, including the cathepsin Dlike Na-APR-1, Na-APR-2 cleaved preferentially between hydrophobic P1 (Ala, Phe, Val, Leu, and Gly) and P1 residues (Ala and Leu), but the individual amino acids at each site differed. Of the major cleavage sites resulting in electrophoretically discernible bands, none was shared between human and canine Hb; furthermore, none of the 3 major cleavage sites (as determined by the Edman procedure) within human Hb was shared with the cathepsin Dlike enzyme from N. americanus, Na-APR-1 [12]. Na-APR-2 cleaved human Hb at 4 sites with a P1 Lys, whereas Na-APR-2 cleaved only 2 sites with a P1 Lys. In agreement with the amount of free Hb N-termini generated after cleavage with Na-APR-2, the enzyme cleaved human Hb at a total of 32 sites (13 in the - and 19 in the -chain) and dog Hb at only 20 sites (7 in the - and 13 in the -chain).
| fig.ommitted |
Figure 7. Hemoglobin (Hb) cleavage sites of Na-APR-2. Sites within human and dog Hb |
Host-specific cleavage of serum proteins but not collagen. Na-APR-2 degraded human serum albumin (HSA; albumin from the hookworm's correct definitive host) more efficiently than it degraded canine serum albumin (CSA; from the nonpermissive canine host; ). Furthermore, Na-APR-2 degraded human fibrinogen, also a plasma protein, more efficiently than canine fibrinogen Na-APR-2 degraded human and canine collagens (types I and IIIV) with equal efficiency (data not shown). Degradation of albumin, collagen and fibrinogen was completely inhibited by the addition of pepstatin A to 1.0 M (data not shown). As with Hb (described above), heterogeneity of the regression coefficients were observed, and the rates of cleavage of human albumin and fibrinogen were 1.6 times (not significant) and 4 times (P < .025) greater, respectively, than that of the corresponding canine substrate proteins. Although the digestion rates for albumin were not significantly different between the 2 host species, the intercept for the assay with human albumin was markedly greater than that of the assay for canine albumin. Therefore, mean differences were compared between the 2 groups of the albumin assay at each time point, and the mean quantity of free amino groups was significantly different between the 2 species of Hb at each time point sampled .
| fig.ommitted |
Figure 8. Host-specific cleavage of human serum proteins by Na-APR-2. Na-APR-2 cleaved human serum albumin (A) and fibrinogen (B) more efficiently than canine albumin and fibrinogen. Quantitation of free amino groups released from the substrates after hydrolysis by recombinant Na-APR-2 was determined by use of ninhydrin. Data are linear regressions of substrate cleavage over time. When pepstatin A was included in reactions, there was no detectable increase in free amino groups above background levels. |
Effect of anti-Na-APR-2 immunoglobulin G on protease activity and L3 migration. Mouse IgG (50 ng) purified from antiNa-APR-2 serum completely inhibited the enzymatic activity of the protease against the diagnostic cathepsin D substrate peptide. AntiNa-APR-2 IgG also inhibited the enzymatic activity of recombinant Na-APR-1 and Ac-APR-1 against the peptide by 50% and 40%, respectively. AntiNa-APR-1 and antiAc-APR-1 IgG inhibited the enzymatic activity of Na-APR-2 against the fluorogenic substrate peptide by 40% . Results are presented as the mean of highly reproducible triplicate experiments (<3% variation between experiments). Ninety-five percent (±6%) of L3 that were incubated in PBS or NMS successfully migrated through hamster skin (from the outside to the inside) within 30 min. When L3 were preincubated in antiNa-APR-2 serum, migration of N. americanus was inhibited by 50% (±5%) and A. caninum by 30% (±3%).
| fig.ommitted | Table 2. Inhibition of the enzymatic activity of hookworm aspartic proteases (1.0 mg) against the fluorogenic peptide o-aminobenzoyl-Ile-Glu-Phe-nPhe-Arg-Leu-NH2, using purified IgG (50 ng) from unimmunized mice or mice immunized with recombinant Na-APR-2. |
DISCUSSION
Two families of aspartic proteases have been reported from parasitic nematodes; cathepsin Dlike enzymes [12, 23] and a family of ostensibly nematode-specific enzymes [12], including C. eleganspredicted proteins, Pep1 from H. contortus [22], Strongyloides pepsin from Strongyloides stercoralis [24], and Na-APR-2 that is described here. This family of nematode-specific aspartic proteases is equally distant in sequence from cathepsin D (including the hookworm aspartic protease Na-APR-1) and pepsin-like aspartic proteases [12] but contains all the hallmark features of the class. In accordance with the literature describing other aspartic proteases, in particular the plasmepsins from Plasmodium falciparum [25], we suggest that this family of nematode-specific enzymes be designated the "nemepsins," for "nematode pepsin-like enzymes." This is the second report of active expression of any class of aspartic protease from a nematode parasite since we recently described the cloning and expression of hookworm cathepsins D [12].
Na-APR-2, like the Necator cathepsin D, Na-APR-1, was expressed in the gut lumen and amphidial glands of adult hookworms. The pepsinogen-like aspartic protease (Pep1) from the obligate blood-feeder, H. contortus (45% identical to Na-APR-2), a parasite of sheep and cattle, is expressed in the nematode gut, where it is thought to function in Hb proteolysis [22]. We reasoned that this family of nematode aspartic proteases is involved in a proteolytic degradation cascade of Hb and serum proteins in blood-feeding nematodes and that this was strongly supported by the rapid and host-specific degradation of human Hb and fibrinogen in vitro and by the immunolocalization of this enzyme to the surface of the gut of the hookworm. Furthermore, aspartic protease activity was not detected in ES products of adult hookworms [12], which suggests that this family of enzymes acts locally to degrade ingested substrate proteins rather than contributing to extracorporeal digestion outside the buccal capsule of the adult hookworm, at the site of attachment to the intestinal mucosa.
Differences between Na-APR-1 [12] and Na-APR-2 were evident in both the pH optima and cleavage sites within various substrate proteins, which suggests an ordered pathway for degradation of substrates, especially of those required for nutrition. The hookworm cathepsin D (Na-APR-1) was most active at pH 5.5 and might be responsible for initial cleavage of substrates in the gut; after acidification of the bloodmeal has proceeded further, Na-APR-2 (pH optimum 5.0) is then responsible for further processing of Hb at a lower pH. Initially, 1 protease may catalyze a defined cleavage, thereby exposing other sites to digestion by a cascade of proteases; however, recombinant Na-APR-2 was capable of digesting intact Hb in vitro. A similar ordered pathway of Hb degradation has been proposed to occur in schistosomes [26] and is known to occur in P. falciparuminfected erythrocytes, in which an aspartic protease located in the acidic digestive vacuole cleaves the -chain of Hb at the hinge region (Phe-33 and Leu-34), after which other proteolytic enzymes digest the Hb fragments into smaller peptides and dipeptides [9, 10, 27]. The order of cleavage events in the hemoglobinolytic cascade of hookworms is currently being explored in our laboratories.
Na-APR-2 cleaved dog and human Hb but at different rates, with Na-APR-2 degrading human Hb much more readily than dog Hb. Moreover, human Hb was cleaved at 32 distinct sites, whereas canine Hb was cleaved at 20 sites. In a similar host-specific fashion, Na-APR-1 cleaved human Hb at 34 sites and dog Hb at only 21 sites [12], but most of the sites cleaved were specific to each enzyme, which further implicates both proteases in a multienzyme hemoglobinolytic cascade. Na-APR-2 completely degraded dog and human Hb after 24 h (determined by SDS-PAGE analysis; not shown), but this is unlikely to be a biologically relevant interval of time since the majority of erythrocytes ingested by hookworms are processed in vivo within a matter of minutes [28]. Na-APR-2 shared several cleavage sites with Ac-APR-1, Na-APR-1 [12], and schistosome cathepsin D [11]. However, of the 52 cleavage sites found for Na-APR-2 (total from the 2 Hb species), >60% were unique to the protease, and only 5 were shared with human cathepsin D (data not shown). The enzyme also favored the -chain, with 60% of its cleavage sites occurring here. Na-APR-2 cleaved human Hb at nearly twice the number of sites compared with that of dog Hb, as well as sharing more common cleavage sites with Ac-APR-1 (13 sites) than Na-APR-1 (2 sites) [12]. Na-APR-2 did not cleave dog or human Hb at the hinge region (3334). However, Na-APR-1 cleaved at the hinge of both human and dog Hb [12], further incriminating Na-APR-1 as an upstream protease in the proteolytic cascade of Hb degradation, since cleavage at the hinge would unravel the Hb tetramer and thereby expose other sites to proteolysis by enzymes downstream in the cascade. Na-APR-2 cleaved human Hb at 32 distinct sites and dog Hb at 20 sites, reflecting a specific preference for Hb from the permissive definitive host. Although the differences are not significant when considered in isolation, the cumulative effect of host-specific cleavage by all the proteases involved in the digestive cascade (including other aspartic proteases, cysteine proteases, and metalloproteases) would be expected to have a dramatic effect on the ability of a given hookworm to digest Hb and other proteins from different hosts.
In similar fashion to the situation with the Hbs, Na-APR-2 readily degraded fibrinogen and serum albumin in a host-specific manner. Hydrolysis of fibrinogen around the attachment site lesion produced by a hookworm when feeding could be expected to retard clot formation. Adult hookworms secrete a battery of anticoagulant peptides [29], and, although APR-2 would be expected to interfere with fibrin clot formation in vitro, Na-APR-2 was not detected in ES products of adult worms incubated in vitro. Perhaps a more plausible explanation is that adult N. americanus feed on Hb and serum proteins, and the host-specific (and gut localized) cleavage of these proteins arose as a consequence of host-parasite coevolution and adaptive mutations in the parasite protease to maintain an intimate enzyme-substrate relationship.
Na-APR-2 mRNA was not expressed by L3 of N. americanus, further implicating this enzyme in digestion of host Hb and serum proteins. Unlike blood-derived proteins, Na-APR-2 degraded human and canine collagens at similar rates with no apparent host-specificity. Given that L3, the tissue migratory phase of the hookworm life cycle, did not express Na-apr-2 mRNA and that proteolytic digestion of collagen is likely to facilitate migration rather than feeding, it is not surprising that human and dog collagens were cleaved equally efficiently by Na-APR-2, since collagen is not a likely substrate for APR-2 in vivo. However, antiserum against recombinant Na-APR-2 partially inhibited penetration of skin and connective tissues by L3 in vitro, whereas normal serum did not. It is possible that antiserum to Na-APR-2 binds to epitopes on aspartic proteases of Necator L3, including Na-APR-1. Indeed, pepstatin A inhibited N. americanus L3 from penetrating hamster skin in vitro [30], which suggests that the aspartic proteases involved in skin penetration are immunologically similar to Na-APR-2.
Na-APR-2, like Na-APR-1, localized to the intestine and amphidial glands of adult worms, further supporting its role as a hemoglobinase in vivo. It was unexpected, however, to find the protease in the amphidial and excretory glands, since it was not detected in ES products in Western blots. However, the localization to the amphids also might be due to immunological cross-reactivity with other aspartic proteases. Since Pep1, an orthologous protease from H. contortus, is localized exclusively to the gut lumen [22], it is possible that Na-APR-2 is primarily produced in the same region in vivo.
Overall, significant differences were observed in the ability of Na-APR-2 to degrade molecules from permissive and nonpermissive host molecules necessary for parasite nutrition. Our findings presented here and previously reported elsewhere [12, 31] indicate that host-specific cleavage of substrate proteins by hookworm aspartic proteases may contribute to the inability of hookworms to reach adulthood in the gut of nonpermissive hosts. Moreover, antisera raised to hookworm aspartic proteases interferes with larval migration, which implies that the targeting of these enzymes offers a potentially useful strategy for the development of a vaccine to control blood-feeding helminth parasites.
Acknowledgments
We thank Alun Jones (Institute of Molecular Biosciences, University of Queensland [UQ]) for high-performance liquid chromatography separation of peptides and hydrolysates; Chris Wood (Department of Biochemistry, UQ) for Edman degradation; and Michael Smout, Lynette King, Malcolm Jones, and Dave Knox (Moredun Research Institute) for technical assistance and helpful discussions.
A Murine Model of Coccidioidal Meningitis
A Potential Virulence Gene, hylEfm Predominates in Enterococcus faecium of Clinical Origin
Activation of Platelet gpIIbIIIa by Phospholipase C from Clostridium perfringens Involves Store-Operated Calcium Entry
Antigen Burden Is a Major Determinant of Human Immunodeficiency VirusSpecific CD8+ T Cell Maturation State: Potential Implications for Therapeutic Immunization
Cell-to-Cell Contact of Human Monocytes with Infected Arterial Smooth-Muscle Cells Enhances Growth of Chlamydia pneumoniae
Chlamydia trachomatis Infection Alters Host Cell Transcription in Diverse Cellular Pathways
Combination Antiretroviral Therapy Results in a Rapid Increase in T Cell Receptor Variable Region Repertoire Diversity within CD45RA CD8 T Cells in Human Immunodeficiency VirusInfected Children
Common Surface-Antigen var Genes of Limited Diversity Expressed by Plasmodium falciparum Placental Isolates Separated by Time and Space
Does Antigen-Specific Cytokine Response Correlate with the Experience of Oculorespiratory Syndrome after Influenza Vaccine?
Effect of GB Virus C Coinfection on Response to Antiretroviral Treatment in Human Immunodeficiency VirusInfected Patients