A functional aquaporin co-localizes with the vacuolar proton pyrophosphatase to acidocalcisomes and the contractile vacuole complex of Trypanosoma cruzi.

We cloned an aquaporin gene from Trypanosoma cruzi (TcAQP) that encodes a protein of 231 amino acids, which is highly hydrophobic. The protein has six putative transmembrane domains and the two signature motifs asparagine-proline-alanine (NPA) which have been shown, in other aquaporins, to be involved in the formation of an aqueous channel spanning the bilayer. TcAQP was sensitive to endo H treatment, suggesting that the protein is N-glycosylated. Oocytes of Xenopus laevis expressing TcAQP swelled under hyposmotic conditions indicating water permeability, which was abolished after preincubating oocytes with very low concentrations of the AQP inhibitors HgCl(2) and AgNO(3). glycerol transport was detected. No Immunofluorescence microscopy of T. cruzi expressing GFP-TcAQP showed co-localization of TcAQP with the vacuolar proton pyrophosphatase (V-H(+)-PPase), a marker of acidocalcisomes. This localization was confirmed by Western blotting and immunofluorescence staining using polyclonal antibodies against a C-terminal peptide of TcAQP. In addition, there was a strong anterior labeling in a vacuole, close to the flagellar pocket, that was distinct from the acidocalcisomes and that was identified by immunogold electron microscopy as the contractile vacuole complex. Taking together, the presence of an aquaporin in acidocalcisomes and the contractile vacuole complex of T. cruzi, provides support for the role of these organelles in osmotic adaptations of these parasites.

We cloned an aquaporin gene from Trypanosoma cruzi (TcAQP) that encodes a protein of 231 amino acids, which is highly hydrophobic. The protein has six putative transmembrane domains and the two signature motifs asparagine-proline-alanine (NPA) which have been shown, in other aquaporins, to be involved in the formation of an aqueous channel spanning the bilayer. TcAQP was sensitive to endo H treatment, suggesting that the protein is N-glycosylated. Oocytes of Xenopus laevis expressing TcAQP swelled under hyposmotic conditions indicating water permeability, which was abolished after preincubating oocytes with very low concentrations of the AQP inhibitors HgCl 2 and AgNO 3 . No glycerol transport was detected. Immunofluorescence microscopy of T. cruzi expressing GFP-TcAQP showed co-localization of TcAQP with the vacuolar proton pyrophosphatase (V-H ؉ -PPase), a marker of acidocalcisomes. This localization was confirmed by Western blotting and immunofluorescence staining using polyclonal antibodies against a C-terminal peptide of TcAQP. In addition, there was a strong anterior labeling in a vacuole, close to the flagellar pocket, that was distinct from the acidocalcisomes and that was identified by immunogold electron microscopy as the contractile vacuole complex. Taking together, the presence of an aquaporin in acidocalcisomes and the contractile vacuole complex of T. cruzi, provides support for the role of these organelles in osmotic adaptations of these parasites.
Trypanosoma cruzi, the etiologic agent of Chagas' disease or American trypanosomiasis, has been found to possess an acidic calcium-storage organelle that was named the acidocalcisome (1). Acidocalcisomes are also found in a diverse range of trypanosomatid and apicomplexan parasites (2), in the green algae Chlamydomonas reinhardtii (3), in the slime mold Dictyostelium discoideum (4), and in the bacterium Agrobacterium tumefaciens (5). They are characterized, in addition to their acidic nature, by their high density (both in weight and by electron microscopy), and high content of pyrophosphate (PP i ), polyphosphate (polyP), calcium, magnesium, and other elements. These organelles share several properties with the vacuoles of plants (tonoplasts), such as the presence of two proton pumps, a vacuolar-type H ϩ -ATPase and a vacuolar proton-translocating pyrophosphatase (V-H ϩ -PPase), and a vacuolar Ca 2ϩ -ATPase.
One of the potential functions of acidocalcisomes is in osmoregulation (2). A link between acidocalcisomes and osmoregulation was evidenced by the rapid hydrolysis or synthesis of acidocalcisome polyP when epimastigotes of T. cruzi were submitted to hyposmotic or hyperosmotic stress, respectively (6). A role for acidocalcisomes in the response of Leishmania major promastigotes to osmotic stress has also been proposed on the basis of their changes in sodium and chlorine content after hyposmotic stress (7). A functional link between acidocalcisomes and the contractile vacuole complex of C. reinhardtii (3) and D. discoideum (4), which is involved in water extrusion under hyposmotic stress, has also been proposed. In this regard, early observations (8) of epimastigotes of T. cruzi, through phase contrast microscopy, have detected the presence of a contractile vacuole complex as a group of small vacuoles, which fuse as they fill. The pulsation period was found to be between 1 min and 1 min and 15 s (8). Electron micrographs of the contractile vacuole and surrounding spongiome of other trypanosomatids have also been published (9,10).
Aquaporins (AQPs), 1 or water channels, are important molecules for osmoregulation in a number of cells. They were initially suspected by noting that a number of cell types are much more permeable to water than predicted by simple diffusion of water through the lipid bilayer (11,12). Aquaporins are composed of two groups; one is permeable only to water (orthodox aquaporins) and the second is permeable to water, glycerol, and other small, uncharged molecules (aquaglyceroporins) (12). At least three mammalian aquaporins (13)(14)(15) and two recently cloned aquaporins from Plasmodium falciparum (16) and Toxoplasma gondii (17), respectively, are aquaglyceroporins. Aquaporins are also abundant in the tonoplast and are also known as tonoplast intrinsic proteins (TIPs) (18,19). A high water permeability for the tonoplast, facilitated by TIPs, is important for osmoregulation in plants (19).
In the present study, we report the cloning and sequencing of a gene from T. cruzi (TcAQP) which encodes a protein with * This work was supported by National Institutes of Health Grant AI-23259 (to R. D.). 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  homology to plant aquaporins. The water transport properties of TcAQP were analyzed by expressing the gene in Xenopus laevis oocytes and performing oocyte swelling assays. TcAQP was shown to localize to the acidocalcisomes and to the contractile vacuole complex of the parasites. Together, these results provide evidence for the presence of a functional aquaporin in acidocalcisomes and contractile vacuoles, and support a role for these organelles in osmoregulation.

EXPERIMENTAL PROCEDURES
Culture Methods-Different stages of T. cruzi (Y strain) were obtained as described previously (20).
Chemicals and Reagents-Fetal bovine serum, Dulbecco's phosphate-buffered saline (PBS), peroxidase labeled concanavalin A, and protease inhibitor mixture were purchased from Sigma. Restriction enzymes, T4 DNA ligase, Taq polymerase, the Klenow fragment of DNA polymerase, SuperScript PCR buffer, and Superscript II reverse transcriptase, TRIzol reagent, geneticin, and 0.24 -9.5 kb RNA ladder were from Invitrogen. PolyAtract mRNA isolation system IV was obtained from Promega (Madison, WI). Pfu DNA polymerase was from Stratagene (La Jolla, CA Isolation of the T. cruzi AQP gene and DNA Sequencing-To screen for genes encoding AQP in T. cruzi, the amino acid sequence of human AQP-1 (NM_000385) was used to search the T. cruzi data base and Integrated T. cruzi Genome Resource using tBLASTn. This search yielded five sequences and the expressed sequence tag clone 24g3 (5Ј) (GenBank TM accession number AI057697). A fragment of 345 bp of the TcAQP gene was amplified using a forward oligonucleotide primer (5Ј-ATGAGTGAGATTGTG-3Ј) and a reverse oligonucleotide primer (5Ј-CACAAGATAGCCTGC-3Ј) derived from the T. cruzi expressed sequence tag. The PCR was performed with 35 cycles of 94°C for 1 min for denaturation, 40°C for 1 min for annealing, and 72°C for 1 min for extension, using 1.5 units of TaqDNA polymerase with 500 ng of T. cruzi genomic DNA, 25 pmol of each of the two oligonucleotide primers, 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 1.5 mM MgCl 2 , and 0.2 mM each deoxynucleoside trisphosphate in a total volume of 50 l. PCR products were cloned into the pCR2.1-TOPO and sequenced. The complete coding sequence was obtained from a library of T. cruzi genomic DNA constructed in GEM11 (21) and screened with the PCR probe as previously described (22). After the third screening, one clone was selected for restriction analysis. The isolated GEM11 DNA was digested with restriction endonucleases, and a 3.5-kb BamHI-HindIII fragment that hybridized to the PCR probe was ligated into pBSKS ϩ and transformed into Escherichia coli DH5␣. The BamHI-HindIII insert was sequenced in a 373A DNA Automatic Sequencer (PerkinElmer Life Sciences). Appropriate primers were synthesized by using as a model both DNA strands from the coding region. The predicted amino acid sequence of TcAQP was aligned with the sequences of other AQPs using the Biology Workbench 3.2 utility (available at workbench.sdsc.edu). Phylogenetic distances were calculated using the Neighbor Joining (NJ) method of Saitou and Nei (23). The sequence of TcAQP was deposited in the GenBank TM data base (accession no. AAM76680).
Southern and Northern Blot Analyses-Total genomic DNA from T. cruzi epimastigotes was isolated by phenol extraction (21), digested with different restriction enzymes, separated on a 1% agarose gel, and transferred to nylon membranes. The blot was probed with a [␣-32 P]dCTP-labeled TcAQP. After hybridization, the blot was washed three times in 2 ϫ SSC, 0.1% sodium dodecyl sulfate at 65°C (SSC is 0.15 M NaCl, 0.015 M sodium citrate). For the Northern blot analysis, total RNA was isolated from amastigotes, trypomastigotes, and epimastigotes using TRIzol reagent, according to the manufacturer's instructions. Polyadenylated RNA was obtained with the PolyATract mRNA isolation system. RNA samples were subjected to electrophoresis in 1% agarose gels containing 1ϫ MOPS buffer (20 mM MOPS, 0.08 M sodium acetate, pH 7.0, 1 mM EDTA) and 6.29% (v/v) formaldehyde after boiling for 10 min in 50% (v/v) formamide, 1ϫ MOPS buffer, and 5.9% (v/v) formaldehyde. The gels were transferred to a Hybond-N filter and hybridized with a probe containing the entire coding sequence of the TcAQP gene obtained by PCR. All Southern and Northern blots were visualized by autoradiography. The TcP0 (T. cruzi ribosomal protein 1) fragment used as a control in Northern blots was obtained as described before (20). Densitometric analyses of Northern blots were done with an ISI-1000 digital imaging system (Alpha Inotech Corp.). Comparison of levels of TcAQP transcription between the different stages was performed by taking as a reference the densitometric values obtained with the TcP0 transcripts and assuming a similar level of expression of this gene in all stages (20). Similar results were obtained when the densitometric values were compared by taking into account the amount of RNA added to each lane.
Antibody Generation Against T. cruzi AQP and Western Blot Analysis-A polyclonal antibody was raised against a synthetic C-terminal peptide of TcAQP with an added N-terminal cysteine (N-LDTHDRVA-PIELSGQVF-C). The peptide was cross-linked by its primary amine groups to the keyhole limpet hemocyanin and used to immunize a rabbit. Polyclonal antibodies were affinity purified using HiTrap-protein A HP columns. Samples of acidocalcisomes (1.5 g of protein) (24) and total membranes of T. cruzi epimastigotes (1.5 or 15 g of protein) (25) were mixed with sample buffer (125 mM Tris-HCl, pH 7, 10% w/v ␤-mercaptoethanol, 20% w/v glycerol, 4.0% w/v SDS, and 4.0% w/v bromphenol blue) and boiled for 5 min before application to SDSpolyacrylamide gels (12%). Electrophoresed proteins were transferred to nitrocellulose and the nitrocellulose was blocked in 5% nonfat dry milk in Tween-PBS (0.1% Tween 20, 80 mM Na 2 HPO 4 , 20 mM NaH 2 PO 4 , 100 mM NaCl, pH 7.5) overnight at 4°C. A 1:2,000 dilution of purified polyclonal antiserum in 3% bovine serum albumin in Tween-PBS was then applied at room temperature for 60 min. The nitrocellulose was washed three times for 15 min each with Tween-PBS. After 60 min of incubating with horseradish peroxidase-conjugated anti-rabbit IgG antibody (1:20,000) and washing three times for 15 min each with Tween-PBS, immunoblots were visualized on blue-sensitive x-ray film using the ECL™ chemiluminescence detection kit and following the instructions of the manufacturer (Amersham Biosciences).
Endoglycosidase H and Peroxidase-labeled Concanavalin A Treatment-Acidocalcisome fractions obtained as described before (24) were incubated in the absence or presence of endoglycosidase H according to the manufacturer's instructions. Control reactions containing 15 g of ovalbumin confirmed activity of the enzymes. After the incubation, the samples were precipitated for 8 h at Ϫ20°C with 750 l of ethanol and pelleted at 16,000 ϫ g for 30 min at 4°C. The dried pellets were resuspended in sample buffer and boiled for 5 min before SDS-polyacrylamide gel electrophoresis (20%). The proteins were transferred from the gels to nitrocellulose and probed with anti-TcAQP polyclonal antiserum and stripped. Membranes were reprobed with peroxidase labeled concanavalin A (10 g/ml) at room temperature for 60 min and washed three times for 15 min each with Tween-PBS. Finally, blotted membranes were visualized on blue-sensitive x-ray film using the ECL™ chemiluminescence system.

Construction of GFP-TcAQP Expression Plasmid and Transfections-
The coding region of green fluorescent protein (GFP) was amplified from pXG-GFPϩ2Ј vector by PCR using Pfu DNA polymerase with a forward oligonucleotide primer containing a HindIII site (5Ј-G-CCGGGAAGCTTATGGTGAGCAAGGGCGAGGAG-3Ј) and a reverse oligonucleotide primer harboring a XhoI site (5Ј-CCGCTCGAGTTACT-TGTACAGCTCGTCCAT-3Ј). After digestion with HindIII and XhoI, the PCR product was cloned in the pTEX expression vector. To fuse the TcAQP to the N-terminal of GFP, the entire coding sequence of the TcAQP gene was amplified by PCR using Pfu DNA polymerase and oligonucleotide primers to intoduce BamHI (5Ј-CGGGATCCATGACG-TTCTCTCCGGGTATG-3Ј) and HindIII (5Ј-CCCAAGCTTGAAAACCT-GACCTGAAAGTTC-3Ј) sites at the 5Ј-and 3Ј-ends, respectively. The resulting plasmid, pTEX-TcAQP-GFP, was sequenced to confirm that the correct reading frame was used. Transfections were carried out in a 2-mm gap cuvette with a Bio-Rad Gene Pulser II set at 1.5 kV and 50 microfarad. 40 ϫ 10 6 parasites were harvested and washed with HBS buffer (21 mM Hepes, pH 7.5, 137 mM NaCl, 5 mM KCl, 0.7 mM Na 2 H-PO 4 , 6 mM glucose), resuspended in 0.4 ml of HBS with 50 g of plasmid DNA. Parasites were recovered in 5 ml of LIT supplemented with 10% fetal bovine serum at 28°C, and after 24 in culture, geneticin was added to a final concentration of 250 g/ml. Cells were cloned by limiting dilution in 96-wells plates. GFP-AQP expressing epimastigotes were differentiated to mammalian forms as described before (26).
Immunofluorescence Microscopy-For immunofluorescence, cells were fixed with 4% paraformaldehyde, adhered to poly-l-lysine cover-slips, permeabilized for 5 min with PBS/0.3% Triton X-100, blocked for one hour with PBS/3% bovine serum albumin/5% goat serum/50 mM NH 4 Cl, and incubated with primary antibody for 1 h and then secondary antibody for 45 min. For visualization of the vacuolar H ϩ -pyrophosphatase, mouse polyclonal anti-T. cruzi pyrophosphatase was used at 1:200, followed by goat anti-mouse Alexa 488 or Alexa 546 at 1:1000. For visualization of aquaporin, rabbit polyclonal anti-T. cruzi aquaporin was used at 1:200, followed by goat anti-rabbit Alexa 546 at 1:1000.
For visualization of the flagellar pocket and/or cytostome, an adaptation of a previously described method was used (27). Briefly, cells were fixed in 4% formaldehyde, adhered to poly-L-lysine coverslips, and then incubated in buffer A (116 mM NaCl, 5.4 mM KCl, 0.8 MgSO 4 , 5.5 mM glucose, 50 mM Hepes, pH 7.2) containing 3% bovine serum albumin and 5 g/ml TRITC-ConA for 1 h before visualization.
Confocal images were collected with a Leica laser scanning confocal microscope (TCS SP2) using a 63ϫ Plan-Apo objective with NA 1.32. Single optical sections were recorded with an optimal pinhole of 293 nm according to Leica instructions.
Oocyte Expression of TcAQP and Osmotic Swelling Assay-The entire coding sequence of the TcAQP gene was amplified by PCR using Pfu DNA polymerase. Oligonucleotide primers for amplification of the AQP coding region, TcAQP5 (5Ј-GAAGATCTATGACGTTCTCTCCGGGT-3Ј) and TcAQP3 (5Ј-GAAGATCTTCAGAAAACCTGACCTGA-3Ј), were designed so that BglII restriction sites were introduced at the 5Ј-and 3Ј-ends for convenient cloning in the pSP64T oocyte expression vector, which contains the 5Ј-and 3Ј-untranslated regions of the ␤-globin gene of X. laevis. Capped cRNA was transcribed in vitro using SP6 mMES-SAGE mMACHINE™ transcription kit after linearization of the plasmid with SmaI. Stages V and VI X. laevis oocytes were defolliculated by incubating ovarian fragments in 0.15% collagenase type II for two hours at room temperature and were injected with 32 nl of water or 5-10 ng of cRNA in 32 nl of water. Injected oocytes were maintained for 3 days at 18°C in Barth's buffer (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO 3  assays were carried out at room temperature and were monitored with a Zeiss phase-contrast microscope equipped with a video camera using a ϫ2.5 objective. An oocyte image was stored at 15 s intervals for a total of 8 min or until the oocyte ruptured using an MCID Elite/Elite Work station 6.0. Oocytes were transferred from 200 mOsm either to 70 mOsm Barth's buffer diluted with deionized water or to isosmotic Barth's buffer, in which 88 mM NaCl was substituted by 130 mM glycerol. Osmolalities were confirmed using an Advanced Instruments 3D3 Osmometer (Norwood, MA). Other oocytes were incubated for 5 min in Barth's buffer containing 1, 5, 10, 300 M of HgCl 2 or 1, 5, 10 M of AgNO 3 followed by swelling in the presence of HgCl 2 or AgNO 3 . Still other oocytes were incubated for 5 min with 300 M of HgCl 2 , then removed and incubated for 15 min in 5 mM of ␤-mercaptoethanol prior to osmotic swelling in the presence of ␤-mercaptoethanol. The surface area of the sequential images was calculated using AnalyzeAVW version 5.0. The oocyte volume was calculated using the following formula: V ϭ (4/3) ϫ (area) ϫ (area/) 1/2 . The change in relative volume with time (d(V/V 0 )/dt, in s Ϫ1 ) was fitted by computer to a quadratic polynomial, using SigmaPlot version 6.0, and initial rates of swelling were calculated. Osmotic water permeability (Pf, m/s) were determined from osmotic swelling data, initial oocyte volume (V 0 ϭ 9 ϫ 10 Ϫ4 cm 3 ), oocyte surface area (S ϭ 0.045 cm 2 ), and the molecular ratio of water (Vw ϭ 18 cm 3 /mol) by the following formula:

Isolation of the T. cruzi AQP Gene and Sequence Analysis-
The complete coding region of TcAQP was established as described under "Experimental Procedures," and the translation of the open reading frame of 693 bp yielded a polypeptide of 231 residues with a predicted molecular mass of 24,681 Da. A BLAST search of the protein data base showed that the amino acid sequence from T. cruzi AQP has 26 -38% identity and 40 -55% similarity to other AQPs (Fig. 1A). The T. cruzi protein contains an internal repeat characteristic of all AQPs. The Nand C-terminal halves are sequence-related and each has the signature motif Asn-Pro-Ala (NPA) (28,29) that forms a single aqueous channel spanning the bilayer. The hydropathy plot for the TcAQP (Fig. 1B) is similar to that of other members of the AQP family, and is consistent with six transmembrane domains (1-6) and five connecting loops (A-E). Cys 189 in loop E of the AQP1 was shown to be the mercury-sensitive residue (30). TcAQP has a glycine (Gly 168 ) at the position equivalent to Cys 189 in AQP-1, however it has two cysteines nearby (Cys 183 and Cys 188 ). There is a third cysteine at residue 121, which is present in a presumed bilayer-spanning domain. Analysis of amino acid sequence of TcAQP using PROSITE (31) revealed a putative N-glycosylation site between residues 137 and 140, which is in loop D. The sequence analysis program iPSORT (31) predicted a 30-amino acid N-terminal signal peptide in TcAQP. In general, aquaporins are synthesized and localized to their final destination without cleavage of the N-terminal part. Since the N-terminal hydrophilic tail of TcAQP is very short, it is not possible to estimate whether it is cleaved. Sequence alignments and phylogenetic tree analyses show that TcAQP falls into the orthodox aquaporin branch between Arabidopsis thaliana ␥-Tip and the yeast aquaporins (Fig. 1C).
Genomic Organization of the TcAQP Gene-Southern blot analysis was performed with a PCR fragment that encompasses the entire coding region of TcAQP ( Fig. 2A). Digestions with enzymes that cut at sites not contained within the coding region of the gene gave single bands (BamHI, EcoRI, SalI), whereas two bands were obtained with enzymes with unique restriction sites in the open reading frame (HaeII, Sau3AI), suggesting that the gene is single copy per genome. In order to confirm the transcription of the TcAQP gene, we performed Northern blot analysis using poly(A) ϩ RNA from different forms of the parasite and the TcAQP gene as a probe. The presence of a single TcAQP transcript of ϳ1.35 kb after prolonged exposure was detected in each of the three life stages of T. cruzi (Fig. 2B). Analysis of the 1.35-kb band by densitometry indicated that the TcAQP gene was expressed at moderately higher levels in trypomastigote than in amastigote and epimastigote forms, while the transcription of a ribosomal protein gene (TcP0) was at comparable levels in all three stages of the parasite (Fig. 2B).
Functional Expression of TcAQP in Xenopus Oocytes-The basic transport characteristics of TcAQP were analyzed by oocyte swelling assay. Fig. 3A shows the osmotic swelling curves of control oocytes and of oocytes expressing TcAQP after subjecting them to hypotonic shock in diluted Barth's buffer (from 200 to 70 mOsm). Water-injected oocytes swelled minimally (Pf ϭ 8.35 Ϯ 3.2 m/s), whereas TcAQP-injected oocytes swelled and ruptured within 5 min (Pf ϭ 31.8 Ϯ 9.1 m/s). The coefficient of water permeability Pf was 3.8-fold higher than in water-injected oocytes (Fig. 3B).
Oocytes expressing the protein from T. cruzi were analyzed for glycerol permeability in an isosmotic swelling assay with 130 mM glycerol. Oocytes injected with TcAQP cRNA were not permeated by glycerol (Pf ϭ 2.27 Ϯ 0.01 m/s). On the contrary, shrinkage of these cells was observed in comparison with control oocytes (Fig. 3, A and B). TcAQP contains three cysteines of which one is found in a transmembrane domain, and two are part of a membrane-spanning domain (loop E). Omotic water permeability though water channels is inhibited by HgCl 2 , via covalent modification of a cysteine residue in the vicinity of the conserved motive NPA, and also by AgNO 3 (32). The increase in water permeability of TcAQP-expressing oocytes was inhibited by incubation in 300 M of HgCl 2 (Pf ϭ 9.7 Ϯ 4.5 m/s), the inhibition being partially reversed by subsequent incubation in 5 mM ␤-mercaptoethanol (Pf ϭ 20.5 Ϯ 5.6 m/s), as shown in Fig. 3, A and B. At low concentrations of HgCl 2 , ranging from 1 to 10 M, the water permeability was also inhibited in TcAQPinjected oocytes (Fig. 3, C and D). The effect of AgNO 3 was analyzed at concentrations between 1 and 10 M. As shown in Fig. 3, E and F, AgNO 3 was a potent inhibitor of TcAQP.
Localization of TcAQP in T. cruzi-To investigate the localization of TcAQP we generated fusion proteins containing the GFP at the C-terminal of TcAQP. Immunofluorescence microscopy of T. cruzi epimastigotes expressing this protein showed co-localization in acidocalcisomes of TcAQP-GFP (Fig. 4B) with TcPPase (Fig. 4C), a known marker of acidocalcisomes (24), as detected with polyclonal antibodies prepared against a conserved epitope in the T. cruzi enzyme. In addition, there was an anterior localization of TcAQP in a vacuole, close to the flagellar pocket, that was distinct from the acidocalcisomes (Fig. 4D,  arrowhead). Concanavalin A has been shown to bind to the flagellar pocket of Leishmania donovani (27) or to the cytos-tome of T. cruzi (33). The cytostome is a specialized structure formed by invagination of the plasma membrane in a region close to the flagellar pocket of epimastigote and amastigote forms, and is involved in endocytosis (34). When cells transfected with TcAQP-GFP (Fig. 4F) were reacted with a concanavalin A-TRITC conjugate (Fig. 4G) to label those struc-   D, H, and L, show  the overlap of B and C, F and G, and J, and K, respectively, and indicate co-localization of the aquaporin in the acidocalcisomes with the V-H ϩ -PPase (D and L), and lack of co-localization in the cytostome with concanavalin A (H). A, E, and I, are bright field images of the same parasites shown by immunofluorescence microscopy. Arrowheads in D, H, and L show the contractile vacuole labeling (see Fig.  7). Scale bar ϭ 10 m.
tures (27,33), there was no co-localization (Fig. 4H), suggesting that most TcAQP-GFP resided in a compartment close to but distinct from the flagellar pocket and/or cytostome. The same aquaporin distribution was observed in epimastigotes reacted with affinity-purified polyclonal antibody against TcAQP (Fig.  4K), while preimmune serum gave no reaction (not shown). These results also confirmed that the GFP-AQP pattern is not an artifact of protein overexpression and/or mistargeting. Labeling of acidocalcisomes with antibody against TcAQP (Fig.  4K) was more complete than with TcAQP-GFP (Fig. 4, B and F) because of the difficulties associated with high resolution imaging of GFP. However, some acidocalcisomes were labeled with antibodies against TcAQP but not with antibodies against TcPPase, and vice versa. Similar heterogeneity in labeling of acidocalcisomes with antibodies against a V-H ϩ -ATPase and a Ca 2ϩ -ATPase has been reported before (35) and could suggest the existence of different populations of acidocalcisomes. However, we cannot rule out the possibility that some aquaporin could be labeling another unidentified organelle. In this regard, Sarkar et al. (36) recently reported a punctate distribution of phosphoenolpyruvate mutase in T. cruzi that does not co-localize with a variety of markers. Similar results to those observed with epimastigotes were obtained with infective stages. Colocalization of TcAQP-GFP (Fig. 5, B and J) or antibodies against TcAQP (Fig. 5, G and O) with antibodies against TcP-Pase (Fig. 5, C, F, K, and N) was detected in amastigotes (Fig.  5, D and H) and in trypomastigotes (Fig. 5, L and P). Labeling of a distinct vacuole was also detected in both stages (arrows in Fig. 5, D, H, L, and P).

Presence of a Contractile Vacuole Complex in Different Life Cycle Stages of T. cruzi, Association with Acidocalcisomes, and
Labeling with Antibodies Against Aquaporin-Acidocalcisomes have been linked to the function of a contractile vacuole complex in C. reinhardtii (3) and D. discoideum (4) and the presence of a contractile vacuole complex has been reported in different trypanosomatids (10,11), including T. cruzi (9). However, there has been only occasional mention of this structure (37) even though extensive electron microscopy studies have been done in T. cruzi. Since in trypanosomatids the contractile vacuole complex is located close to the flagellar pocket we explored the possibility that aquaporin could be located in this structure in T. cruzi.
In all organisms examined to date, the contractile vacuole complex consists of a large central bladder surrounded by a diffuse radial network of tubules and vesicles known as the spongiome (38). In all three life cycle stages of T. cruzi, a large irregular-to-round vacuole that was associated with a loose network of tubules could be easily identified by transmission electron microscopy in close opposition to the flagellar pocket (Fig. 6, A-D). These structures closely matched the descriptions of the contractile vacuole fine structure in the trypanosomatids Leptomonas collosoma (10) and Bodo sp. (11), as well as descriptions from other non-trypanosomatid organisms (39 -41). Interestingly, many of the observed acidocalcisomes and con- tractile vacuole bladders contained remnants of electron dense material (closed arrowheads). Accumulation of this material is generally diagnostic for acidocalcisomes under transmission electron microscopy (2), and suggests communication between the two compartments. Occasional observation of acidocalcisomes apparently fusing with the contractile vacuole (open arrowhead in Fig. 6B) further served to reaffirm this hypothesis.
In order to analyze in more detail the structures labeled with the TcAQP-GFP, or with antibodies against TcAQP, immunoelectron microscopy was performed on thin sections of parasites embedded in the hydrophilic resin Unicryl using the polyclonal antibody against the C-terminal of TcAQP and monoclonal antibodies against TcPPase. TcAQP and TcPPase co-localized to acidocalcisomes (Fig. 7E). TcAQP was also highly concentrated in the contractile vacuole and the immediate surround- ing region (presumable the spongiome) (Fig. 7C), as well as the flagellar pocket (Fig. 7, A and B). Control experiments with TcAQP preimmune serum showed very little nonspecific labeling (Fig. 6, D and F). TcPPase was also found in the contractile vacuole (Fig. 7C). A TcAQP immunogold density histogram was constructed for the following subcellular compartments: contractile vacuole/spongiome, flagellar pocket, nucleus, mitochondrion, cytosol, and acidocalcisomes (Fig. 7G). The density of TcAQP gold particles in the contractile vacuole and acidocalcisomes was significantly higher than in other compartments. Additionally, the density of gold particles within the flagellar pocket was higher than that in the nucleus, mitochondrion, and cytosol, but lower than within the contractile vacuole. Labeling of the flagellar pocket by antibodies against TcAQP confirmed that concanavalin A labels specifically the cytostome and not the flagellar pocket of T. cruzi (33).
TcAQP Expression and N-Glycosylation in T. cruzi-Immunoblot analysis with the affinity-purified antibody against TcAQP showed a band of ϳ30.9 kDa present in all developmental stages of the parasite, with higher expression in epimastigotes (Fig. 8A). There was a 5.8-fold enrichment of this band in the acidocalcisome fraction as compared with the total membranes (Fig. 8B), thus confirming the localization studies (Figs. 4, 5, and 7).
Acidocalcisome fractions were analyzed to examine if the potential N-glycosylation site of TcAQP was utilized. As shown in Fig. 8C, the protein was sensitive to endoglycosidase H treatment, which removes high mannose type asparaginelinked oligosaccharides. To confirm the presence of oligosaccharides the gels were stripped and treated with the peroxidaselabeled lectin concanavalin A. While the control was labeled, the endo H-treated sample was not stained. These results suggest that the protein carries N-glycans. DISCUSSION We report here that a gene, TcAQP, present in the T. cruzi genome encodes a functional aquaporin. The open reading frame corresponding to TcAQP encodes a protein of 231 amino acids and a molecular mass of 24.7 kDa (Fig. 1A) that is expressed in all life stages of the parasite (Figs. 2B, 4, 5, and 8). The TcAQP is present as a single-copy gene ( Fig. 2A) and is a member of the orthodox (water-transporting) family of aquaporins (Fig. 1C).
TcAQP was expressed in X. laevis oocytes, where it was able to function as a water, but not a glycerol channel (Fig. 3). TcAQP was found to have similar sensitivity to AgNO 3 and HgCl 2 . This is in contrast to other aquaporins that are about 20 times more sensitive to AgNO 3 than to HgCl 2 (32). The mechanism of silver and mercury inhibition is most likely due to their ability to interact with sulfhydryl groups of proteins. In the case of TcAQP it is likely that these metals react with the sulfhydryl group of cysteines in the vicinity of the conserved NPA motif and thus effectively block the constriction region of the water channel. TcAQP has two cysteines in the vicinity of the NPA motif (Cys 183 and Cys 188 ) that could potentially react with these metals. The coefficient of water permeability (Pf) was enhanced by 3.8-fold by injecting mRNA of TcAQP into Xenopus oocytes. The result suggests low water channel activity of TcAQP. In this regard, it has been suggested (12) that the low water permeability exhibited by AQP0 and AQP6 may reflect the need for an activation step, and we cannot rule out that this might be the case with TcAQP.
Immunofluorescence microscopy using a fusion protein containing the green fluorescent protein at the C-terminal of TcAQP or antibodies against a C-terminal peptide of TcAQP (Figs. 4 and 5) revealed its localization in acidocalcisomes and in a vacuolar structure close to the flagellar pocket of the parasites. This vacuolar structure was not the cytostome, as demonstrated by the absence of co-localization of TcAQP with concanavalin A, a marker of the cytostome of T. cruzi (33) (Fig.  4H), and was identified by electron microscopy as the contractile vacuole complex (Figs. 6 and 7). Both TcAQP and TcPPase co-localized to the acidocalcisomes and contractile vacuole complex. Further work will be necessary to investigate the mechanism by which dual targeting of T. cruzi aquaporin occurs.
The adaptation of a number of protozoa to hyposmotic stress involves, in addition to the release of ions and osmolytes, as occurs in mammalian cells (42), the release of water by a contractile vacuole complex. Recent work has shown that most, if not all, contractile vacuole complexes are composed of a two-compartment system enclosed by two differentiated membranes (38,43). One membrane (spongiome), which is often divided into numerous vesicles and tubules, contains many proton-translocating V-H ϩ -ATPase enzymes that provide an electrochemical gradient of protons for water transport and which can fuse only with the membrane of the second compartment. The membrane of the second compartment (bladder) which lacks V-H ϩ -ATPase holoenzymes, although has a V-H ϩ -PPase (3, 4) expands into a reservoir for water storage, and is capable of fusing with the plasma membrane (38,43). It is this second compartment that periodically undergoes contraction, with the expulsion of water (38). How water is accumulated in the contractile vacuole complex was unknown. Although a water channel was postulated to be involved (38) it was never before identified. Our work provides the first evidence for the presence of an aquaporin in the contractile vacuole complex. Interestingly, other vacuoles, besides the contractile vacuole, have been observed to take up water when some protozoa are placed in hyposmotic media (44,45), and they have been suggested to also play a role in volume homeostasis (46). It is possible that those vacuoles correspond to the acidocalcisomes that we described in several protozoa (2)(3)(4).
As T. cruzi progresses through its life cycle, it encounters diverse, severe environmental stressors to which it must successfully adapt. Of particular interest is the parasite's ability to cope with extreme fluctuations in osmolarity that occur within the gut of the vector (47,48) and also as the parasite moves from the insect gut through the acidic phagolysosome to the cytosol of the host cell (49). The infective form of the parasite passes out of the vector in the highly concentrated excreta and rapidly encounters the interstitial fluid of the mammalian host with a much lower osmolarity. Clearly the parasite must have mechanisms that allow it to adapt both to hyperosmotic and hyposmotic stress. In this work we have identified the presence of an aquaporin in acidocalcisomes and contractile vacuole complex of T. cruzi, providing support for the role of these organelles in osmotic adaptations of these parasites.