Molecular dissection of the functional domains of a unique, tartrate-resistant, surface membrane acid phosphatase in the primitive human pathogen Leishmania donovani.

The primitive trypanosomatid pathogen of humans, Leishmania donovani, constitutively expresses a unique externally oriented, tartrate-resistant, acid phosphatase on its surface membrane. This is of interest because these organisms are obligate intracellular protozoan parasites that reside and multiply within the hydrolytic milieu of mammalian macrophage phago-lysosomes. Here we report the identification of the gene encoding this novel L. donovani enzyme. In addition, we characterized its structure, demonstrated its constitutive expression in both parasite developmental forms, and determined the cell surface membrane localization of its translated protein product. Further, we used a variety of green fluorescent protein chimeric constructs as reporters in a homologous leishmanial expression system to dissect the functional domains of this unique, tartrate-resistant, surface membrane enzyme.

Leishmania donovani is an important protozoan pathogen of humans that causes severe and most often fatal visceral disease in the tropics and subtropics worldwide (1). This organism has a digenetic life cycle that consists of two major developmental forms: 1) extracellular flagellated promastigotes that reside and multiply in the alimentary tract of their sandfly vectors and 2) obligate intracellular nonflagellated amastigotes that reside and multiply within the phago-lysosomal system of infected human macrophages.
Acid phosphatases (AcPs) 1 are phosphomonoesterases that hydrolyze substrates under low pH conditions and are gener-ally considered to be typical marker enzymes of lysosomes (2, 3). Previously, it was shown that L. donovani promastigotes (4,5) and tissue-derived amastigotes possess a unique, externally oriented, surface membrane, tartrate-resistant acid phosphatase. 2 The presence of this enzyme on the parasite cell surface is an interesting observation considering that amastigotes of all pathogenic leishmanial species reside and multiply within host cell phago-lysosomes. Moreover, to date, no other tartrateresistant surface membrane AcP from any source has been reported in the literature.
The tartrate-resistant surface membrane AcP of L. donovani (MAcP) has a broad substrate specificity hydrolyzing glycerol phosphates and mono-and di-phosphorylated sugars (5), inositol phosphates, and phosphorylated proteins (6). Although the biochemical properties of this enzyme have been partially characterized, its biological function(s), as with virtually all acid phosphatases (3), remain to be elucidated. Understanding and investigating the role(s) that this AcP plays in parasite growth and survival would be facilitated by the characterization of the gene(s) that encodes this unique enzyme. To date, however, no such leishmanial genes have been reported. Thus, in the current study we identified the gene encoding this novel L. donovani enzyme and characterized its structure, expression, and localization in both developmental forms of the parasite. Further, a variety of green fluorescent protein (GFP) chimeric constructs was used as reporters in a homologous leishmanial expression system to dissect the functional domains of this unique, tartrate-resistant, surface membrane enzyme.

EXPERIMENTAL PROCEDURES
Reagents-All chemicals used, unless otherwise noted, were of analytical grade and were purchased from Sigma. Similarly, enzymes and DNA molecular mass standards were purchased from Roche Molecular Biochemicals. Protein molecular mass standards were purchased from Amersham Biosciences.
Parasite Cultures-L. donovani promastigotes ([1S, clone 2 D ]from the 1S strain World Health Organization designation: (MHOM/SD/62/ 1S-CL2 D ) were grown at 26°C in chemically defined 199(ϩ) medium as described by McCarthy-Burke et al. (7). Axenic amastigotes forms of this L. donovani clone were grown at 37°C as described by Joshi et al. (8). All of the cultures were harvested at log phase (2-3 ϫ 10 7 cells ml Ϫ1 ) by centrifugation as described (9). The cell pellets were resuspended in the appropriate buffers for isolation of nucleic acids, for preparation of surface membrane fractions, or for transfection experiments. Tissuederived amastigotes of this L. donovani strain were isolated from spleens of infected hamsters (Mesocricetus auratus; LVG strain, Charles River Laboratories, Inc., Wilmington, MA) as described previously (10).

Preparation of Parasite Surface Membrane-enriched Fractions-Cell
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  lysates of both promastigotes and axenic amastigotes were prepared from washed cell pellets by the addition of lysis buffer (10 mM Tris-HCl, 2 mM EDTA, 25 g ml Ϫ1 leupeptin, pH 8.0) to a final concentration of 5 ϫ 10 8 cells ml Ϫ1 and disruption in a prechilled, tightly fitting Dounce homogenizer (5). The lysates were centrifuged at 8000 ϫ g for 30 min at 4°C to obtain parasite surface membrane-enriched fractions (8K pellets). The supernatant was removed, and the 8K pellets were washed twice with buffer (10 mM HEPES, 145 mM NaCl, pH, 7.4) by centrifugation as above. The washed 8K pellets were solubilized (at ϳ5 ϫ 10 8 cell equivalents ml Ϫ1 ) in the same buffer containing 1.0% Triton X-100 (Surfact-Amps® X-100; Pierce), 1.0% octyl-␤-D-glucopyranoside (Calbiochem, La Jolla, CA) by stirring at 4°C for ϳ12 h (11). The samples were centrifuged at 48,000 ϫ g for 30 min at 4°C, and the supernatants, which contained the detergent-solubilized parasite surface membrane components, were assayed for enzyme activity or used in immunoprecipitations and Western blot analyses. The protein concentrations were determined using the bicinchoninic acid (microBCA; Pierce) method (12). Isolation of Nucleic Acids-Total genomic DNA (gDNA) was isolated from promastigotes using the GNome DNA isolation kit from Bio 101 (La Jolla, CA). Total RNA was isolated from promastigotes, axenic amastigotes, and tissue-derived amastigotes using the RNeasy method (Qiagen, Chatsworth, CA) as described previously (13).
Probe Labeling-A probe common to the 5Ј-ends of the L. donovani AcP gene family (13) was generated by PCR amplification with Taq polymerase and digoxygenin-dUTP according to the manufacturer's instructions (Roche Molecular Biochemicals). The conditions for amplification were 95°C for 1 min, 42°C for 1 min, 72°C for 1 min (40 cycles), and 72°C for 6 min. This 300-bp probe was generated with oligo primers forward 1, 5Ј-CAGAACGACCGACATGC, and reverse 2, 5Ј-GACATACTTGAACCAGG using the previously described pAcP396 (13) as a template. The resulting digoxygenin-labeled fragment (DIG-300) was used in hybridization studies.
In addition, an oligo unique to MAcP, forward 111, 5Ј-CGCGAAAT-GCATGAAGGGCGCAATCGCGTCTTCGATATGG, was labeled with digoxygenin-dUTP by tailing with terminal transferase according to the Genius Labeling kit 6 manufacturer's instructions (Roche Molecular Biochemicals). The resulting MAcP specific oligo probe (DIG-111) was used in hybridization studies.
Southern Hybridization-gDNA was digested in independent reactions with several restriction endonucleases, and the restriction fragments were separated by 0.8% agarose gel electrophoresis. These gels were prepared for blotting as described previously (13). DNA was transferred under vacuum and UV cross-linked to nylon membranes (Hybond-N; Amersham Biosciences). The membranes were processed for hybridization with the DIG-300 or DIG-111 probes according to the Genius system user's guide (Roche Molecular Biochemicals), and the hybridized fragments were visualized using the Genius detection system (Roche Molecular Biochemicals) and fluorography.
Cosmid Library Screening-A cosmid library of L. donovani gDNA (kindly provided by Dr. B. Ullman, Oregon Health Sciences University, Portland, OR) was screened by hybridization using the DIG-300 probe. DNA isolated from positive cosmid clones (Magic Miniprep; Promega, Madison, WI) was analyzed by restriction endonuclease digestion and Southern hybridization analyses. Cosmids containing an ϳ3.0-kb PstI AcP insert (Cos AcP-101, -102, and -103) were subjected to nucleotide sequence analysis.
Nucleotide Sequencing and Analysis-DNA was sequenced using the fluorescent dideoxy chain terminator cycle sequencing method (14) at the Johns Hopkins University DNA Analysis Facility (Baltimore, MD) as described previously (9). Sequence data from both strands were analyzed using the Genetic Computer Group software package (15) running on a National Institutes of Health Unix system and Sequencer TM 3.0 software (Gene Codes Corp., Ann Arbor, MI). Signal sequence and protease cleavage sites were predicted using Analyze-Signalase 2.0.3 (16).
RT-PCR-Reverse transcription was carried out with total RNA from L. donovani promastigotes, axenic amastigotes, and tissue-derived amastigotes using Superscript II (Invitrogen Co., Carlsbad, CA) and oligo(dT) to generate cDNA according to the manufacturer's instructions. PCR amplification reactions contained the following oligo primer pair: primer pair 1, forward B, 5Ј-GCACCAGCTGCGTCCTCT, and reverse B, 5Ј-ATCACGCCAACTGCAGA. PCR amplification reactions contained primer pair 1, 2 l of the cDNA generated above, dNTPs (Promega) Taq polymerase, and the appropriate buffer in a final volume of 50 l. The conditions for amplification were 95°C for 1 min, 42°C for 1 min, 72°C for 1 min (40 cycles), and 72°C for 6 min. PCR products were analyzed by 3.0% NuSieve® agarose (FMC® Bioproducts, Rockland, ME) gel electrophoresis and ethidium bromide staining. Gelpurified PCR products (Sephaglas TM Bandprep kit; Amersham Biosciences) were subjected to nucleotide sequence analysis.
Antibodies-Two individual peptides unique to the MAcP-deduced protein were synthesized using 9-fluoromethyloxycarbonyl chemistry: peptide 1, Lys-Phe-Pro-Phe-Phe-Arg-Phe-Pro-Tyr-Arg-Arg-Arg-Asp-Cys-Ala-Leu-His (kindly provided by the Laboratory of Molecular Structure NIAID, National Institutes of Health), and peptide 2, Arg-Phe-Pro-Tyr-Arg-Arg-Arg-Asp-Cys-Ala-Leu (Genosys Biotechnologies Inc., The Woodlands, TX). These peptides were conjugated to maleimide-activated keyhole limpet hemocyanin (Imject-activated immunogen conjugation kit; Pierce). Anti-peptide antibodies (anti-MAcP peptide Ab) were produced in a New Zealand White rabbit (No. 1407) by Spring Valley Laboratories, Inc. (Woodbine, MD) by subscapular immunization with a mixture containing 1.25 mg of each of the conjugated peptides (2.5 mg total) as described previously (9). The resulting anti-MAcP peptide Ab was used in both Western blot analyses and immunoprecipitation activity assays. In addition, a rabbit monospecific (No. 172) Ab to the L. donovani SAcP (17) and the appropriate control sera were used as positive controls in these assays. An anti-GFP mouse monoclonal antibody (CLONTECH Laboratories Inc., Palo Alto, CA) was also used in Western blots and in indirect immunofluorescence assays to detect GFP chimeric proteins. A rabbit anti-Bip antibody was a kind gift from Dr. James Bangs (University of Wisconsin, Madison, WI). Further, a goat anti-rabbit rhodamine-conjugated antibody (Jackson Immunoresearch Laboratories, West Grove, PA) and a goat anti-mouse fluorescein isothiocyanate-conjugated antibody (Sigma) were used as secondary antibodies for fluorescence microscopy.
Immunoprecipitation of L. donovani MAcP Activity-Solubilized surface membrane extracts from L. donovani promastigotes and axenic amastigotes were reacted with anti-MAcP peptide Ab or NRS in a protein A-Sepharose 4B/CL (Amersham Biosciences) bead-based assay as described previously (9). Immunoprecipitates were subsequently assayed for AcP activity in the presence or absence of 5 mM sodium tartrate (L(ϩ)-tartaric acid; ICN Pharmaceuticals Inc., Costa Mesa, CA), as described previously (18). All of the samples were assayed in triplicate, and these assays were repeated on multiple samples. One unit of enzyme activity (19) reflects the hydrolysis of 1 nmol of pnitrophenyl phosphate (U.S. Biochemical Corp., Cleveland, OH) to pnitrophenol min Ϫ1 mg Ϫ1 of detergent-solubilized parasite surface membrane protein at 42°C. The percentage of activity immunoprecipitated was determined using the following formula: (activity bound/[activity bound ϩ activity unbound]) ϫ 100. The results obtained from immunoprecipitations were normalized by subtracting the values obtained with NRS from those obtained with the anti-MAcP peptide Ab.
Nomenclature-The designations used in this report for proteins, genes, and plasmids follows the nomenclature for Trypanosoma and Leishmania as suggested by Clayton et al. (20) GFP Constructs-pKSNEO was used as the leishmanial vector (21) to express GFP-MAcP chimeras for studies of the functional domains of the MAcP. In a multi-step process, the potential signal peptide (SP) of MAcP, GFP, and the putative transmembrane anchor domain (TM) of MAcP were PCR-amplified and subcloned to generate pKS NEO MAcPSP::GFP::TM.
First, PCR amplification of the TM domain of MAcP was performed using Cos AcP-101 as template with the primer pair, forward MAcPTM 5Ј-TGGTGGAGCGCTCCCGCCTTATATAAATTGATAGCTACGTGT (Eco47 restriction endonuclease site in bold type) and reverse MAcPTM 5Ј-TGGTGGACGCGTTGGACTAGTCCCTTAATACACGCGAAATGC-ATG (MluI and SpeI restriction endonuclease sites in bold type). The resulting PCR product was cloned into the pCR2.1vector (TA Cloning System; Invitrogen) to generate the plasmid pCR2.1 MAcPTM.
The first chimera that contained GFP and the putative TM domain of MAcP was constructed by subcloning the gel-purified Eco47-digested GFP fragment from pCR2.1 EGFP into the Eco47-linearized pCR2.1 MAcPTM construct, which resulted in the plasmid construct pCR2.1 GFP::TM. The orientation of the GFP fragment was verified using gel electrophoresis analysis with the appropriate restriction endonucleases.
The second chimera containing the MAcP SP, GFP, and the putative MAcP TM domain was produced by subcloning the gel-purified MluI fragment of pCR2.1 GFP::TM into the MluI-linearized pCR2.1 MAcPSP, which resulted in the plasmid construct pCR2.1 MAcPSP::GFP::TM. The orientation of the GFP::TM fragment was verified using gel electrophoresis analysis with the appropriate restriction endonucleases.
pCR2.1 MAcPSP::GFP::TM was digested with SpeI, and the gelpurified fragment was subcloned into the SpeI-linearized leishmanial expression vector pKS NEO, resulting in the plasmid construct pKS NEO MAcPSP::GFP::TM. Orientation of the subcloned fragment was verified using gel electrophoresis analysis with the appropriate restriction endonucleases. The final construct pKS NEO MAcPSP::GFP::TM was verified by sequence analysis as described above and was subsequently transfected into L. donovani promastigotes for functional domain analysis studies. Other MAcP-GFP chimeric expression plasmids including pKS NEO MAcPSP::GFP and pKS NEO GFP were constructed using the same multi-step PCR-based cloning strategy above as appropriate.
Transfection of GFP Chimeric Constructs into Leishmania-For transfection experiments, harvested promastigote cells were resuspended in electroporation buffer to 10 8 cells ml Ϫ1 as described previously (22). 500 l of cell suspension was added to a 2-mm gap electroporation cuvette (BTX Inc., San Diego, CA). Immediately prior to electroporation, 20 l of purified plasmid DNA (1 mg ml Ϫ1 in sterile 10 mm Tris, 2 mm EDTA (Quality Biological, Inc., Gaithersburg, MD), pH 8.0) was added to the cell suspension. The cells were electroporated using a BTX Inc. ECM-600 electroporation system. Electroporation conditions were a single pulse at 475 V, 800 microfarads, and 13 ohms. The electroporated cells were incubated on ice for 10 min and transferred into 5 ml of culture medium as described above and incubated at 26°C for 24 h. The transfected cells were subsequently harvested by centrifugation as above and resuspended in culture medium containing 15 g ml Ϫ1 G418 (Geneticin, Invitrogen). Transfected cells were selected for growth in increasing concentrations of G418 over a period of several weeks and maintained at 200 g ml Ϫ1 drug. The drug-resistant cells were used in subsequent experiments.
Microscopy-Transfected promastigotes were washed three times in phosphate-buffered saline by centrifugation as described above. Fluorescence images of such live cells were acquired using a Zeiss Axioplan microscope (Carl Zeiss, Inc., Thornwood, NY), which was equipped with epifluorescence, a cooled CCD camera (Photometrics, Tucson, AZ), and the appropriate fluorescein isothiocyanate excitation/barrier filters. Transfected parasites were also examined in indirect immunofluorescence assays using anti-Bip and anti-GFP antibodies essentially as described by Debrabant et al. (22). Such cells were examined by confocal microscopy using a Zeiss LSM 410 system with fluorescein isothiocyanate and rhodamine excitation/barrier filters. The fluorescent images obtained in these channels were collected separately. All of the captured images were processed using Adobe Photoshop 5.5 (Adobe Systems, San Jose, CA).

RESULTS
Identification of the MAcP ORF-Previously we identified and characterized the genes that encode the 110-and 130-kDa isoforms of the L. donovani histidine secretory AcPs (SAcP-1 and SAcP-2) (13). In the current study, results of Southern analysis of L. donovani gDNA digested with PstI revealed three restriction fragments (5.2, 3.9, and 3.0 kb) that hybridized with an L. donovani AcP gene probe (DIG-300) (Fig. 1). The two larger PstI fragments, of 3.9 and 5.2 kb, contained the SAcP-1 and SAcP-2 genes described above (13). The third PstI fragment, of 3.0 kb, was subcloned, sequenced, and found to contain a partial ORF, very similar but not identical in sequence to either SAcP-1 or SAcP-2. To obtain a full-length ORF corresponding to the 3.0-kb PstI fragment, an L. donovani gDNA cosmid library was screened with the DIG-300 probe, which is common to all three PstI restriction fragments. Seventeen positive clones were identified and analyzed by restriction endonuclease digestion with PstI followed by Southern hybridization with the DIG-300 probe. Three of the 17 cosmid clones (Cos AcP-101, -102, and -103) contained a 3.0-kb PstI restriction fragment. Sequence analysis revealed that these cosmids contained a full-length (948 bp) ORF corresponding to the 3.0-kb PstI genomic fragment.
Comparison of the L. donovani AcP-deduced Proteins-Sequence analysis showed that the 948-bp ORF encoded a deduced protein with a calculated molecular mass of 35,192 Da and a predicted isoelectric point of 8.23. Analysis of the deduced aa sequence showed that it contained five potential Nlinked glycosylation sites (Asn 44 , Asn 96 , Asn 135 , Asn 219 , and Asn 245 ) ( Fig. 2A) and one predicted myristoylation site (Gly 37 ). In addition, the deduced protein possessed five putative phosphorylation sites by several different mechanisms (i.e. casein kinase II: Ser 83 ,Ser 154 , and Ser 248 and protein kinase C: Thr 107 and Thr 247 ). Moreover, the most abundant residues of this deduced protein were alanine and leucine, which constitute 23% of its aa content.
The deduced protein can be divided into three structural domains. Region I consists of a 23-aa putative signal peptide (Met 1 -Ala 23 ) based on the Von Heijne algorithm (23) (Fig. 2). The sequence of this putative signal peptide including its peptidase cleavage site is identical to those present at the N termini of the SAcP-1-and SAcP-2-deduced proteins (13). Therefore, cleavage at this site would result in Arg 24 as the N-terminal aa of the mature protein. In addition, the two SAcP-deduced proteins of L. donovani (SAcP-1 and SAcP-2) and the 948-bp ORF-deduced protein above are identical in aa sequence for 251 residues beyond the signal peptide (Region II) (Fig. 2B). The presence of a conserved histidine AcP signature sequence (i.e. a catalytic site consensus sequence) within Region II (Val 27 -Arg 39 ) (Fig. 2) indicates that the 948-bp ORFdeduced protein is a member of this highly conserved family of enzymes. However, the second histidine AcP signature se-  (24), Region III of 948-bp ORF-deduced protein contains a stretch of 29 hydrophobic aa residues (Leu 274 -Tyr 302 ) that could function as a transmembrane anchor domain (Fig. 2). This is followed by a short 13-aa putative cytoplasmic tail (Arg 303 -Tyr 315 ). Cumulatively, these observations suggested that the 948-bp ORF represented a membraneanchored member of the histidine acid phosphatase family; thus, this gene was designated as the L. donovani MAcP.
Characterization of the MAcP Gene Locus-Comparison of both nucleotide and deduced aa sequences revealed that the AcPs from L. donovani belong to a highly conserved multi-gene family. To determine whether MAcP constituted a distinct single gene or a multi-gene locus, we designed a specific oligo probe, DIG-111 (Fig. 3A), that would only recognize MAcP and not SAcP-1 or SAcP-2. This oligo probe was used in Southern hybridization analysis of L. donovani gDNA. The results of these analyses with both single (AatII, NcoI, and NotI) and double restriction endonuclease (AatII and NcoI; AatII and NotI; and NcoI and NotI) digestions demonstrated that within an ϳ6-kb region of L. donovani gDNA, only single restriction fragments were observed to hybridize with the DIG-111 probe (Fig. 3B). One would expect that multiple restriction fragments would have hybridized with this probe, if the MAcP locus had been present in more than one copy. Thus, the hybridization results indicated that the 6 kb surrounding and including the MAcP locus were present only once within the L. donovani genome. These results were confirmed by Southern hybridization of the three cosmids clones, Cos AcP-101, -102, and -103, each of which was shown to contain a single MAcP ORF (data not shown).
Identification of the MAcP Transcript-Transcription of MAcP was assayed in a two-step process involving RT and PCR amplification (Fig. 4). First, cDNA of promastigote, axenic amastigote, and tissue-derived amastigote mature mRNA was generated with oligo(dT) and reverse transcriptase. Second, an aliquot of each of these cDNAs was used as template in PCR with gene-specific primers (Fig. 4A), and the expected 250-bp amplified product should contain an internal sequence specific to the MAcP.
The results of these experiments showed that only a single amplified product of 250 bp was obtained using aliquots of promastigote cDNA as template (Fig. 4B, Pro lane). Similarly, a 250-bp amplification product was obtained with aliquots of cDNA from both axenic amastigotes (Fig. 4B, Ax Am lane) and tissue-derived amastigotes (Fig. 4B, Tis Am lane). These RT-PCR results were confirmed using several different RNA-cDNA preparations. In control reactions, in which cDNAs were not generated prior to PCR, no amplified products were obtained (data not shown). Moreover, in control reactions in which forward or reverse primers were omitted from the reaction mix, no amplified products were obtained (data not shown). The specificity of the RT-PCR-amplified products was verified by sequencing the gel-purified PCR-amplified reaction products. Sequence analysis revealed that the primer pair specifically amplified a single product, which corresponded to the specific MAcP gene sequence. Together, these data demonstrated that MAcP is actively transcribed by L. donovani promastigotes, axenic amastigotes, and tissue-derived amastigotes. Thus, the MAcP appears to be constitutively transcribed throughout the parasite developmental life cycle.
Localization of MAcP by Western Analysis-SDS-PAGE and Western blot analysis were used to determine the cellular localization of the protein product encoded by the MAcP. L. donovani isolated surface membrane fractions from promastigotes and axenic amastigotes as well as whole cell lysates of tissue-derived amastigotes were probed in Western blots using the anti-MAcP peptide Ab. The results of these assays showed that the anti-MAcP peptide Ab reacted with a single ϳ53-kDa protein in the isolated promastigote surface membranes (Fig. 5,  lane 1). These results are consistent with those reported previously concerning the apparent molecular mass of a partially purified tartrate-resistant AcP from the surface membranes of L. donovani promastigotes (25). The MAcP peptide Ab also recognized a single ϳ64-kDa protein in surface membranes of axenic amastigotes (Fig. 5, lane 2) and an ϳ60-kDa protein in lysates of tissue-derived amastigotes (Fig. 5, lane 3). The ϳ4-kDa difference in apparent molecular mass between the MAcP protein present in isolated axenic amastigote surface membranes and in lysates of tissue-derived amastigotes might reflect differences in their post-translational modifications (e.g. type and/or amount of glycosylation, phosphorylation, etc.) because of the diverse environments in which these organisms were grown (i.e. in vitro cell culture versus hamster spleen macrophages in vivo). Alternatively, this difference could result from proteolytic degradation/ hydrolysis of the MAcP protein during the isolation and processing of tissue-derived amastigotes.
In parallel Western blots (data not shown) these proteins were also recognized by a rabbit polyclonal anti-L. donovani  Taken together, these Western blot results demonstrated that the MAcP protein is produced by tissue-derived amastigotes and that it is expressed in the surface membranes of both in vitro grown promastigotes and axenic amastigotes. Further, these results showed that the MAcP protein expressed by promastigotes and amastigotes differed in apparent molecular mass. The latter may reflect their developmental differences in post-translational processing of this surface membrane enzyme.
Immunoprecipitation of MAcP Activity-Surface membranes isolated from L. donovani promastigotes and axenic amastigotes were detergent-solubilized and assayed for total acid phosphatase activity using p-nitrophenyl phosphate as substrate in the presence or absence of 5 mM sodium tartrate. The results of these assays demonstrated that both parasite developmental forms possessed comparable levels of MAcP specific activity (Table I). Further, these assays showed that all of the measurable MAcP activity present in these solubilized surface membranes was resistant to inhibition by sodium tartrate. Aliquots of such solubilized surface membranes were reacted with the anti-MAcP peptide Ab (antibody raised against a portion of the unique C-terminal peptide sequence of MAcP (i.e. aa residues Lys 294 -His 310 )) or NRS in a protein A-Sepharose 4B/CL beadbased assay (9,18). Immunoprecipitates from these were assayed for AcP activity in the presence or absence of 5 mM sodium tartrate. The results from these assays (Table I) showed that the anti-MAcP peptide Ab immunoprecipitated ϳ20% of the total AcP activity present in the detergent-solubilized surface membranes of each parasite developmental form. Further, these results showed that all of the activity immunoprecipitated by the anti-MAcP peptide Ab was tartrate-resistant. The relatively low amount of enzyme activity immunoprecipitated in these assays might be due to insertion of the C-terminal hydrophobic domain of MAcP into detergent micelles, thus making this epitope only partially available for interaction and recognition by the anti-MAcP peptide Ab.
The results of these immunoprecipitation assays in conjunction with our Western blot results demonstrated that the anti-MAcP peptide Ab specifically reacted with a single tartrateresistant parasite surface membrane AcP present in both parasite developmental forms. Such expression is in agreement with our RT-PCR results, which indicated that the MAcP gene is actively transcribed by both L. donovani promastigotes and axenic amastigotes.
Episomal Expression of GFP Chimeras in L. donovani Promastigotes-Chimeric proteins containing specific domains of the MAcP fused with GFP as a reporter were episomally expressed in L. donovani promastigotes using the pKS NEO leishmanial expression vector. These chimeric constructs were used to demonstrate that the N-terminal region of the MAcP functions as a signal peptide and that its C-terminal region functions as a membrane anchor. All chimeric constructs and control plasmids were transfected into L. donovani promastigotes, which were subsequently grown in increasing concentrations of G418 (to a final concentration of 200 g ml Ϫ1 ). These transfectants were subsequently analyzed in Western blots and by epifluorescence microscopy.
The first plasmid construct, pKS NEO MAcPSP::GFP::TM, encoded a protein (Fig. 6A, map 1) containing the putative N-terminal signal peptide domain (Met 1 -Arg 23 ) of MAcP including its peptidase cleavage site (Arg 24 ), the full-length GFP, and the C-terminal domain of the MAcP (Pro 275 -Tyr 315 ). Thus, a Assayed using para-nitrophenyl phosphate as substrate and specific activity expressed as nmol p-nitrophenol liberated min Ϫ1 mg Ϫ1 of detergent solubilized isolated surface membrane protein. Identical values were obtained in the presence or absence of 5 mM sodium tartrate. The data shown represent the mean results of triplicate assays for each sample from three separate experiments.
b Activity immunoprecipitated using ␣-MAcP peptide antibody in a protein A-Sepharose 4B/CL bead-based assay using the formula: (Activity Bound/[Activity Bound ϩ Activity Unbound])ϫ 100. The results were normalized by subtracting values obtained with NRS (preimmune serum) from those obtained with the ␣-MAcP peptide antibody. Identical values were obtained in these assays in the presence or absence of 5 mM sodium tartrate. The data shown reflect the mean results of triplicate assays for each sample from three separate experiments. this expressed GFP chimeric protein contained the putative MAcP signal peptide at its N-terminal end and both the MAcP putative transmembrane (TM) anchor domain and cytoplasmic tail at its C terminus. This construct was used to determine whether the C-terminal MAcP sequence would function to target and anchor the GFP in the cell surface membrane of transfected parasites.
The second plasmid construct, pKS NEO MAcPSP::GFP (Fig.  6A, map 2), contained the N-terminal putative signal peptide sequence of the MAcP as above and the full-length GFP reporter (i.e. it lacked the entire C-terminal domain of the MAcP). This construct was used to determine whether the N-terminal MAcP sequence functioned as a signal peptide that would target the nascent protein into the endoplasmic reticulum.
The third plasmid construct (Fig. 6A, map 3), pKS NEO GFP, contained no sequences from the MAcP gene. This construct was used as a GFP reporter control for these transfection studies.
Surface membranes were isolated from both pKS NEO MAcPSP::GFP::TM and pKS NEO control transfectants. These were subjected to SDS-PAGE and Western blot analyses with the anti-GFP Ab and anti-MAcP peptide Ab. In such blots, the anti-GFP Ab and anti-MAcP peptide Ab each reacted very strongly with a single ϳ32-kDa protein that was only present in the surface membranes of pKS NEO MAcPSP::GFP::TM transfectants (Fig. 6B, lanes 2 and 2Ј, respectively) and not in those from pKS NEO control transfectants (Fig. 6B, lanes 1 and  1Ј, respectively). The apparent molecular mass of this ϳ32-kDa chimeric protein reflects the sum of its GFP and MAcP Cterminal domain (42 amino acid residues) components (i.e. ϳ28 and ϳ4.2 kDa, respectively). In addition, the anti-MAcP peptide Ab also reacted with the single ϳ53-kDa endogenous MAcP protein present in the surface membranes of both the pKS NEO MAcPSP::GFP::TM and pKS NEO control transfectants (data not shown). Further, preimmune rabbit serum, normal mouse serum, and ascites controls showed no reactivity in these Western blots. The results of these assays demonstrated that the pKS NEO MAcPSP::GFP::TM episomal plasmid was readily translated and expressed as a ϳ32-kDa MAcPSP::GFP::TM chimeric protein in these parasites.
L. donovani promastigotes transfected with the above expression plasmids were also examined by epifluorescence microscopy. Such observations revealed that promastigotes transfected with pKS NEO MAcPSP::GFP::TM demonstrated bright GFP cell surface fluorescence (Fig. 6C). In contrast, pKS NEO control transfectants showed no surface fluorescence. These results indicated that the MAcPSP::GFP::TM chimeric protein was in fact targeted to and expressed on the cell surface membrane of these parasites. Further, in conjunction with our Western blot data, these results demonstrated that the MAcP transmembrane domain was present on the GFP chimeric protein and that it functioned to anchor this protein in the parasite cell surface.
In contrast to the above information, promastigotes transfected with pKS NEO MAcPSP::GFP showed only diffuse intracellular fluorescence, reflecting the processing of GFP in the endoplasmic reticulum of these cells. The latter was confirmed by its colocalization (Fig. 6D, panel 3, Merge) with Bip (a resident endoplasmic reticulum protein) in indirect immunofluorescence assays using both anti-GFP (Fig. 6D, panel 1) and anti-Bip (Fig. 6D, panel 2) antibodies. Cell lysates and cell-free culture supernatants from promastigotes transfected with the pKS NEO MAcPSP::GFP plasmid or control pKS NEO plasmid were also subjected to SDS-PAGE and Western blot analyses. Such blots were probed with the anti-GFP antibody and matched control reagents. The results of these assays showed that the lysates of pKS NEO MAcPSP::GFP but not pKS NEO transfectants contained a single ϳ28-kDa protein that reacted with the anti-GFP Ab (Fig. 6E, lanes 2 and 1, respectively). Further, these results also showed that pKS NEO MAcPSP::GFP but not pKS NEO transfectants secreted/released a single, soluble, ϳ28-kDa GFP into their culture supernatants (Fig. 6E, lanes 2Ј and 1Ј, respectively). The ϳ28-kDa protein expressed by these transfectants presumably reflects the apparent molecular mass of the mature GFP alone (i.e. lacking the MAcP signal peptide domain). Cumulatively, these results demonstrated that the N-terminal end of the MAcP (Met 1 -Ala 23 ) functions as a signal peptide to translocate the nascent GFP into the endoplasmic reticulum. Further, in the absence of a C-terminal membrane anchor, the mature soluble ϳ28-kDa GFP is released from these cells into their culture supernatants, presumably via default into the secretory pathway.
Promastigotes transfected with the pKS NEO GFP plasmid (i.e. devoid of any MAcP sequences) were also examined in Western blots and by epifluorescence microscopy. Lysates of such cells showed that they contained a single ϳ28-kDa protein that reacted with the anti-GFP Ab in Western blots; however, no GFP was detected in the cell-free supernatants from these parasites (data not shown). Examination of these cells by epifluorescence microscopy revealed that GFP was diffusely distributed throughout their cytoplasm. These observations are in agreement with our conclusions above concerning the functions of the N-and C-terminal domains of the MAcP. DISCUSSION In the current report we identified a new member (MAcP) of the AcP multi-gene family in L. donovani by Southern hybridization of a gDNA cosmid library. The MAcP gene is present in single copy and encodes a distinct but highly conserved deduced protein of 315 aa. Sequence analysis of gene-specific RT-PCR using RNA isolated from both developmental forms of this parasite demonstrated that L. donovani promastigotes, axenic amastigotes, and tissue-derived amastigotes actively transcribed MAcP mRNA. In Western blots, the MAcP peptide Ab reacted with a single polypeptide in promastigote surface membranes, axenic amastigote surface membranes, and whole cell lysates of tissue-derived amastigotes. Further, this Ab immunoprecipitated the tartrate-resistant MAcP activity present in surface membranes of both L. donovani promastigotes and axenic amastigotes. Cumulatively, the results of these studies demonstrated that MAcP is constitutively transcribed and translated by both promastigotes and amastigotes into an active tartrate-resistant AcP protein, which is expressed in the surface membranes of both of these parasite developmental forms.
Previously, we identified the genes encoding the two major secretory isoforms (SAcP-1 and SAcP-2) of the histidine AcPs in L. donovani (13). In the current study we used GFP chimeras to show that the MAcP-deduced protein possesses a functional 23-aa signal peptide sequence at its N terminus. Further, we showed that this signal peptide is identical to those of SAcP-1 and SAcP-2. In addition, the N termini of these three leishmanial AcPs (SAcP-1, SAcP-2, and MAcP) all contained 251 conserved aa residues. Moreover, within this region, these proteins contain a signature sequence (Val 27 -Arg 39 ) that is characteristic of a catalytic domain present in all histidine AcPs (13,26). Downstream of this catalytic domain, however, these three leishmanial AcPs have diverged evolutionarily presumably by gene duplication and polymorphisms. As such, SAcP-1 and SAcP-2 contain regions rich in serine and threonine repeat units (13) that are absent in MAcP.
The MAcP-deduced protein possesses a unique 41-aa C-ter-minal domain that is absent from SAcP-1 and SAcP-2. Further, based on the Kyte-Doolittle algorithm (24), the MAcP contains a sequence of 29 hydrophobic aa residues (Leu 274 -Tyr 302 ). In the current study GFP chimeras were used to show that this C-terminal domain functions to anchor the MAcP protein into the surface membrane of these parasites. Recently, the mechanism by which tartrate inhibits the enzymatic activity of the tartrate-sensitive histidine AcPs was delineated by LaCount et al. (27). In that report, the authors showed that in addition to the conserved sequence motif Arg-His-Gly-Xaa-Arg-Xaa-Pro, aa residues corresponding to Ala 79 , His 257 , and Asp 258 of the human prostatic AcP are involved in the binding of tartrate to the active site of these enzymes. In that regard, several genes for tartrate-sensitive leishmanial AcPs have been identified. These include the two L. donovani secretory AcPs, SAcP-1 and SAcP-2 (13); the two released AcPs, SAP1 and SAP2 (28); and one membrane-bound AcP, MBAP (29) of Leishmania mexicana. The deduced proteins of all of these leishmanial AcPs contain residues, which correspond to those involved in tartrate binding (27). Thus, conservation of these residues accounts for these leishmanial AcP enzymatic activities being inhibited by tartrate. In the current report, we identified a constitutively expressed gene (MAcP) that encodes a unique tartrate-resistant surface membrane AcP of L. donovani. Sequence analysis of this MAcP demonstrated that it lacked the critical residues corresponding to His 257 and Asp 258 of the human prostatic AcP. Thus, the absence of these residues in this leishmanial MAcP is consistent with the tartrate resistance of this parasite surface membrane enzyme.
Acid phosphatases are generally considered to be typical marker enzymes of lysosomes. In fact, a family of tartrateresistant acid phosphatases (TRAPs) has been identified and shown to be conserved membrane components of mammalian lysosomes (2, 3). All TRAPs contain five conserved signature sequences (3); however, none of these signature sequences were present in the L. donovani MAcP, indicating that this parasite enzyme is not a TRAP per se. Thus, the cumulative results of this study have identified and characterized, for the first time, the gene encoding the unique tartrate-resistant surface membrane histidine acid phosphatase of L. donovani. To date, no other tartrate-resistant surface membrane acid phosphatase has been reported in the literature. Molecular dissection of the functional domains of this novel parasite enzyme demonstrated that both the N-terminal signal peptide and the C-terminal anchor domains were required to target the MAcP to the surface membrane of this pathogen. Because L. donovani amastigotes reside and multiply within the phago-lysosomal system of mammalian macrophages, the MAcP may afford this parasite a survival advantage in such hostile hydrolytic environments. The availability of the MAcP gene therefore should facilitate studies concerning the biological role of this enzyme in parasite survival.