A Solanesyl-diphosphate Synthase Localizes in Glycosomes of Trypanosoma cruzi*

We report the cloning of a Trypanosoma cruzi gene encoding a solanesyl-diphosphate synthase, TcSPPS. The amino acid sequence (molecular mass ∼ 39 kDa) is homologous to polyprenyl-diphosphate synthases from different organisms, showing the seven conserved motifs and the typical hydrophobic profile. TcSPPS preferred geranylgeranyl diphosphate as the allylic substrate. The final product, as determined by TLC, had nine isoprene units. This suggests that the parasite synthesizes mainly ubiquinone-9 (UQ-9), as described for Trypanosoma brucei and Leishmania major. In fact, that was the length of the ubiquinone extracted from epimastigotes, as determined by high-performance liquid chromatography. Expression of TcSPPS was able to complement an Escherichia coli ispB mutant. A punctuated pattern in the cytoplasm of the parasite was detected by immunofluorescence analysis with a specific polyclonal antibody against TcSPPS. An overlapping fluorescence pattern was observed using an antibody directed against the glycosomal marker pyruvate phosphate dikinase, suggesting that this step of the isoprenoid biosynthetic pathway is located in the glycosomes. Co-localization in glycosomes was confirmed by immunogold electron microscopy and subcellular fractionation. Because UQ has a central role in energy production and in reoxidation of reduction equivalents, TcSPPS is promising as a new chemotherapeutic target.

We report the cloning of a Trypanosoma cruzi gene encoding a solanesyl-diphosphate synthase, TcSPPS. The amino acid sequence (molecular mass ϳ 39 kDa) is homologous to polyprenyl-diphosphate synthases from different organisms, showing the seven conserved motifs and the typical hydrophobic profile. TcSPPS preferred geranylgeranyl diphosphate as the allylic substrate. The final product, as determined by TLC, had nine isoprene units. This suggests that the parasite synthesizes mainly ubiquinone-9 (UQ-9), as described for Trypanosoma brucei and Leishmania major. In fact, that was the length of the ubiquinone extracted from epimastigotes, as determined by high-performance liquid chromatography. Expression of TcSPPS was able to complement an Escherichia coli ispB mutant. A punctuated pattern in the cytoplasm of the parasite was detected by immunofluorescence analysis with a specific polyclonal antibody against TcSPPS. An overlapping fluorescence pattern was observed using an antibody directed against the glycosomal marker pyruvate phosphate dikinase, suggesting that this step of the isoprenoid biosynthetic pathway is located in the glycosomes. Co-localization in glycosomes was confirmed by immunogold electron microscopy and subcellular fractionation. Because UQ has a central role in energy production and in reoxidation of reduction equivalents, TcSPPS is promising as a new chemotherapeutic target.
Trypanosoma cruzi is the etiological agent of Chagas disease or American trypanosomiasis, which is the leading cause of cardiac death in endemic areas throughout Latin America. More than 18 million people are infected with the parasite, and some 40 million more are at risk (1).
Chemotherapy of Chagas disease is unsatisfactory because of toxicity and lack of efficacy of existing drugs, and it is important to identify enzymes and metabolic processes in T. cruzi that might be potential targets for drug development. One pathway that has been particularly useful for the identification of new targets is the isoprenoid pathway. Enzymes studied so far involved in the synthesis of sterols (2), farnesyl diphosphate (3), and protein prenylation (4) have been reported to be good drug targets against this parasite. The farnesyl-diphosphate synthase, for example, has been demonstrated to be the target of bisphosphonates that have activity in vitro and in vivo against T. cruzi (3,(5)(6)(7)(8)(9).
Polyprenyl-diphosphate synthases are responsible for chain elongation in isoprenoid biosynthesis and catalyze the sequential condensation of isopentenyl diphosphate (IPP, 3 C 5 ) with allylic prenyl diphosphates (10). These condensations are catalyzed by a family of prenyltransferases, which are classified into two groups according to the stereochemistry of the E or Z double bond that is formed (10). Z-Polyprenyl-diphosphate synthases are used for the synthesis of dolichols for N-linked glycoprotein biosynthesis, Z-polyprenols for peptidoglycan biosynthesis in bacteria, and natural rubber, whereas E-poly-prenyl-diphosphate synthases are used for the synthesis of a vast variety of important natural isoprenoids, such as steroids, cholesterol, sesquiterpenes, heme a, dolichols, farnesylated proteins, carotenoids, diterpenes, geranylgeranylated proteins, chlorophylls, and archaebacterial ether-linked lipids (10). Long E-polyprenyl-diphosphate synthases producing compounds with chain lengths from C 30 to C 50 are involved in respiratory quinone biosynthesis (10).
The localization of the trypanosomatid enzymes involved in isoprenoid metabolism has been little studied, although some of them, like the T. cruzi FPP synthase (5), bear predicted targeting signals for the glycosomes. Glycosomes are specialized peroxisomes that, like them, contain several enzymes in pathways of ether lipid synthesis, fatty acid ␤-oxidation, and peroxide metabolism, and, in addition, contain the Embden-Meyerhof segment of glycolysis (19).
In the present study, we report the cloning, sequencing, and heterologous expression of a T. cruzi gene designated TcSPPS that encodes a functional SPPS. The expressed TcSPPS gene could complement the function of the corresponding polyprenyl-diphosphate synthase of Escherichia coli, and the cells produced mainly UQ-9. The kinetic properties of the recombinant TcSPPS were analyzed, and the enzyme was shown to localize in the glycosomes, supporting the role of these organelles in isoprenoid synthesis.
Culture Methods and Cell Extraction-T. cruzi amastigotes and trypomastigotes (Y strain) were obtained from the culture medium of L 6 E 9 myoblasts as described previously (21). T. cruzi epimastigotes (Y strain) were grown at 28°C in liver infusion tryptose medium (22) supplemented with 10% newborn calf serum. T. cruzi epimastigotes (CL Brener clone) were grown as described before (23).
DNA Sequencing and Bioinformatics-Sequencing grade DNA was obtained using a Qiagen kit. Sequencing was performed on an ABI 377 using a BigDye Terminator Cycle Sequencing Kit (PerkinElmer Life Sciences), or on a MegaBACE 1000 using the DYEnamic ET dye terminator kit (Amersham Biosciences). Vector primers and the following sequencing primers were used: Fwd (antisense), 5Ј-CACGTGCCACCATGGCAAAC-3Ј; Fwd2 (antisense), 5Ј-CAATGCCTTCTGCCATGTC-3Ј. Chromatograms were analyzed using Bio Edit software (24). Homology searches were performed at the NCBI Blast server (25), and sequences were aligned using ClustalX VI 1.81. The theoretical molecular weight and isoelectric point were obtained from the ExPASy Server (cn.expasy.org). The superimposed hydrophobicity profiles were calculated using the Kyte-Doolittle hydropathy algorithm (26) at bioinformatics.weizmann.ac.il/hydroph. The presence of a signal peptide was assessed by the SignalP 3.0 software (www.cbs.dtu.dk/) (27).
Hybridization to Cosmid Filters-Cosmid filters from a CL Brener cosmid library were used (28). The whole coding sequence of the gene was generated by PCR, purified from agarose gels using DEAE membranes (29), and 30 ng was labeled with [␣-32 P]dCTP by random priming (Prime a Gene, Promega). Cosmid filters were prehybridized and hybridized as described (28), using a Micro 4 oven (Hybaid, UK). Two of the positive clones (20i8 and 69i5) were further studied.
Hybridization to Pulsed Field Gel Electrophoresis and Northern Filters-Chromosomes from the T. cruzi CL Brener clone were separated by pulsed field gel electrophoresis using different running conditions (30) and transferred to nylon filters (kindly provided by Mario Galindo, Instituto de Ciencias Biomédicas, Facultad de Medicina, Universidad de Chile). Schizosaccharomyces pombe and Saccharomyces cerevisiae chromosomes were used as markers (Bio-Rad). Total RNA from epimastigotes was isolated using an SV Prep Total RNA kit (Sigma), according to the manufacturer's instructions. For Northern blot analysis, epimastigotes total RNA was subjected to electrophoresis in 1% agarose gel containing 1ϫ MOPS buffer 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 RNA was transferred to a Hybond-N filter. A T. cruzi probe encoding the 19-kDa cyclophilin, TcCyP19, was used as a positive control.
RT-PCR-T. cruzi CL Brener epimastigote mRNA was isolated by using a QuickPrep Micro mRNA kit (GE Healthcare Bio-Sciences), and RT-PCR was performed with the Access RT-PCR System (Promega) using the following primers at 1 M final concentration: Miniexon (sense), 5Ј-AACGCTATTATT-GATACAGTTTCTGTACTATATTG-3Ј, Fwd2 (antisense). As an internal positive control TcCyP19 was amplified.
Southern Blot Analysis and Genome Organization-T. cruzi CL Brener genomic DNA (3 g) was digested by NcoI and AatII (Fermentas), separated on a 1% agarose gel and transferred to Hybond-N ϩ membrane (Amersham Biosciences). Efficient transfer was confirmed by methylene blue staining (Sigma). Probe generation and target detection was performed using the Gene Images AlkPhos Direct Labeling & Detection system (Amersham Biosciences) following the manufacturer's instructions. Blast searches of the T. cruzi genome (www.genedb.org/genedb/tcruzi/) were performed with TcSPPS nucleotide sequence. In silico restriction analysis was performed at The Sequence Manipulation Suite web site (bioinformatics.org/sms/).
Expression and Purification of TcSPPS from E. coli-For expression in E. coli, the entire coding sequence of the TcSPPS gene was amplified by PCR using primers (PS5, 5Ј-CCGGAT-CCATGCTGAAAACAGGCCTTT-3Ј; PS3, 5Ј-CCAAGCTT-CATACTTGTCGCGTTAAAA-3Ј) that introduced BamHI and HindIII restriction sites for convenient cloning into the expression vector pET-28a ϩ to yield pET-TcSPPS. The joining region was sequenced for confirmation. E. coli BL21(DE3) bacterial cells transformed with pET-TcSPPS were induced, and the recombinant protein was purified by nickel-nitrilotriacetic acid-agarose, following the standard Qiagen procedure. The eluted fraction was desalted with a PD-10 desalting column. Proteins were quantified by the Bradford method (31) with bovine serum albumin as a standard and the absence of protein contaminants was checked by SDS-PAGE.
Measurement of Activity and Product Analysis-Enzyme activity was measured by determination of the amount of [4-14 C]IPP incorporated into butanol-extractable polyprenyl diphosphates. Because removal of the polyhistidine tag resulted in complete loss of activity of other prenyltransferases from trypanosomatids (5, 11), this was not done. The standard assay mixture contained, in a total volume of 100 l, 100 mM Tris-HCl buffer (at physiological pH 7.4), 1 mM MgCl 2 , 1% (v/v) Triton X-100, 100 M [4-14 C]IPP (1 Ci/mol), allylic substrate (400 M DMAPP, 400 M GPP, 30 M FPP, or 50 M GGPP), and 0.5-3 g of the purified protein. The mixture was incubated at 37°C for 30 min, and the reaction was stopped by chilling quickly in an ice bath. The reaction products were then extracted with 1 ml of 1-butanol saturated with water. The 1-butanol layer was washed with water saturated with NaCl, and radioactivity in the butanol extract was determined with a liquid scintillation counter. One unit of enzyme activity was defined as the activity required to incorporate 1 nmol of [4-14 C]IPP into extracted product in 1 min. To identify the reaction products after the enzymatic reaction, the radioactive prenyl diphosphates in the mixture were hydrolyzed to the corresponding alcohols with potato acid phosphatase as described before (32). The alcohols were extracted with n-pentane and analyzed by TLC on a reversed-phase Adsorbosil HPTLC plate with a solvent system of acetone/water (12:1, v/v). The positions of authentic standards were visualized by iodine vapors. The radioactivity was visualized by autoradiography.
Glycosome Enrichment-T. cruzi CL Brener epimastigotes (ϳ10 9 cells) were centrifuged for 10 min at 2,000 ϫ g, and washed twice in TEDS buffer (25 mM Tris-HCl, pH 7.4, 1 mM EDTA, 250 mM sucrose, 1 mM dithiothreitol) containing protease inhibitors (P8340, Sigma). After freezing at Ϫ80°C for 20 min and thawing at 37°C, cells were centrifuged and resuspended in homogenization buffer (250 mM sucrose, 1 mM EDTA, 0.1% v/v ethanol, 5 mM MOPS, pH 7.2, and protease inhibitors). The parasites were grinded in a pre-chilled mortar with 1ϫ wet weight silicon carbide until no intact cells were observed under the light microscope. The lysate was centrifuged at 100 ϫ g for 10 min to remove the silicon carbide, which was washed in homogenization buffer, and both supernatants were combined (Fraction A). A centrifugation at 1,000 ϫ g for 15 min was performed to remove the nuclei, and the supernatant (Fraction B) was centrifuged at 33,000 ϫ g to enrich in glycosomes. The supernatant was Fraction C (cytoplasm) and the pellet (Fraction D) was the glycosomal enriched fraction. The whole procedure was performed twice. Protein concentration of each fractionation step was measured by a colorimetric assay (Protein Assay, Bio-Rad).
Western Blot Analysis-To investigate for protein expression in the different stages, total trypanosome proteins (30 g of protein/lane) were separated by SDS-polyacrylamide gel (10%) and transferred to nitrocellulose. Membranes were probed with 1:3,000 dilution of a rabbit anti-SPPS and then with horseradish peroxidase-conjugated anti-rabbit IgG antibody (1:10,000). Immunoblots were developed using the ECL TM chemiluminescent detection kit (Amersham Biosciences).
For Western blot analysis of the different subcellular fractions, the blots were sequentially probed with a rabbit anti-TbgGAPDH antibody as a marker for glycosomes at a dilution of 1:3,000, and after a stripping step, with rabbit anti-TcSPPS antibody.
Complementation Analysis-The TcSPPS was tested for its capacity to complement the ispB gene of E. coli. Strain KO229, whose essential ispB gene was disrupted and complemented by the ispB expression vector pKA3 (spectinomycin-resistant), was subjected to a plasmid-swapping experiment (33). Because the pET construct would not be inducible in strain KO229, the gene was subcloned into pQE30 vector (pQE-TcSPPS). After transformation with pQE-TcSPPS, the colonies were grown and passaged for several days in LB medium (1% tryptone, 0.5% yeast extract, 1% sodium chloride, pH 7.5) supplemented with ampicillin (to select pQE-TcSPPS-carrying colonies) and isopropyl 1-thio-␤-D-galactopyranoside (to induce expression of the His 6 -TcSPPS fusion protein). Isolated colonies were checked for ampicillin resistance and spectinomycin sensitivity.
Ubiquinone Extraction and Measurement-Ubiquinone was extracted as previously described (34). The crude extract of UQ was analyzed by normal-phase TLC with authentic standard UQ-10. Normal-phase TLC was carried out on a Kieselgel 60 F 254 plate (Merck) with benzene/acetone (93:7, v/v). The band containing UQ was collected from the TLC plate following UV visualiza-tion and extracted with chloroform/methanol (1:1, v/v). Samples were dried and re-dissolved in ethanol. The purified UQ was further analyzed by HPLC with ethanol as the solvent.
Fluorescence Microscopy-For co-localization with a glycosomal marker, T. cruzi Y strain epimastigotes slides were prepared as previously described (35). Antibody concentrations were as follows: affinity-purified rabbit anti-TcSPPS antibody at 1:4,000; supernatant from an anti-TbPPDK producing mouse hybridoma culture at 1:10; anti-mouse Alexa 488 at 1:1,000; anti-rabbit Alexa 546 at 1:1,000. For co-localization studies with MitoTracker and the vacuolar pyrophosphatase, epimastigotes were fixed for 30 min in 4% paraformaldehyde in 0.1 M cacodylate buffer, washed twice in Dulbecco's phosphatebuffered saline, pH 7.2, adhered to poly-L-lysine-coated coverslips, and permeabilized for 3 min with 0.3% Triton X-100. Cells were blocked for 30 min in 50 mM NH 4 Cl and 3% bovine serum albumin in phosphate-buffered saline, pH 8.0, and incubated for 1 h with polyclonal primary antibodies raised against T. cruzi SPPS (1:1000), and monoclonal antibodies raised against T. brucei VP1 (1:200). For mitochondrial staining, cells were previously incubated for 30 min in culture medium containing 100 nM MitoTracker before fixation. Cells were then washed in 3% bovine serum albumin, incubated with secondary antibodies anti-mouse Alexa 488 (1:1,000), anti-rabbit Alexa 488, and anti-rabbit Alexa 546 (1:1,000) and mounted in prolong Antifade. Cells were observed in a Deltavision fluorescence microscope. Images were recorded with a Photometrics CoolSnap HQ camera and deconvolved for 15 cycles using Softwarx deconvolution software.

Identification of T. cruzi SPPS-
We determined the complete sequence of the T. cruzi cDNA clone TENU4155 (accession number AW324852) (36), which showed similarities to polyprenyl synthases. The sequence surrounding the first ATG complied with the published rules for start codons in protozoa (37). To obtain further upstream sequences a cDNA probe was hybridized to high density cosmid filters, and the sequence obtained from two of the positive clones (20i8 and 69i5) with the forward primer displayed a stop codon in the same reading frame as the first putative ATG, confirming the protein coding region. This sequence has been submitted to the GenBank TM under the accession number AF282771.
Translation of the open reading frame of 1092 bp yielded a polypeptide of 363 amino acids with a predicted molecular mass of 39 kDa and an isoelectric point of 6.01. A small residue (Ala) is found at position Ϫ5 before the first aspartate-rich motif. This position is diagnostic, determining the final product length (for a review, see Ref. 38). Bulky amino acids do not allow nascent long chains to extend further inside the hydrophobic cavity of the enzyme. A BLAST search of the protein data base showed that the amino acid sequence from T. cruzi shared up to 38% identity and up to 61% similarity with other polyprenyl synthases. Considering specifically the human homologue (accession number NP_055132), the identity reached 33%.
The amino acid sequence from the T. cruzi enzyme was aligned with other representative polyprenyl synthases (Fig.  1A). All the conserved motifs involved in catalysis or binding (regions I-VII) identified in other polyprenyl synthases (39) are present in the T. cruzi enzyme. The functional residues conform also several motifs present in databases, pfam00348 among them, related to trans-isoprenyl-diphosphate synthases, as well as motif COG0142, IspA, related to farnesyl-diphosphate synthases. Hydrophobicity analysis of the protein showed the characteristic pattern of this family of enzymes (40), consisting of alternating hydrophobic regions, which in Fig. 1B is superimposed to the pattern of the human homologue for comparison.
Four features were observed when comparing the putative T. cruzi polyprenyl synthase to those from other species: a shorter N terminus, an insertion, a "correct length" and a "correct charge" of the C terminus (Fig. 1A). The N-terminal length variation does not seem relevant, because some species also have a longer N terminus (Homo sapiens, Capsicum annum, and both SPPS from Arabidopsis thaliana) (41,42). The observed 15-amino acid insertion, based on the proposed structure of polyprenyl synthases (40), could be located in loop 2, which seems not to be involved in binding to the substrate. This should minimize the effect of this difference in the overall structure and in the activity. The C terminus length seems to be significant, because it is claimed to form a flexible flap that seals the active site upon substrate binding (40). Regarding the C terminus charge, the majority of these proteins have positive side-chain amino acids in some of the last three positions (40), which is also the case for TcSPPS (Fig. 1A).
The genes encoding TcSPPS were located in homologue chromosomes of sizes 800 and 1100 kb, as assessed by hybridization to pulsed field gel electrophoresis membranes (data not shown). By analysis of the codon usage (bioinformatics.org/ sms/) a preference for A-and T-ending codons was detected. The gene could then be assigned to groups TC2 or TC3, conformed by genes using non-optimal codons, which are not highly expressed (43).
Southern Blot and Genome Analysis-Four sequences with homology with TcSPPS were present in the contigs generated by the Genome Project. Three of the contigs contained incomplete genes. Two of them were identical in the overlap, had some differences in comparison to the whole gene, and could represent one allele. The third contig was a mix of the two alleles and could arise from an assembly problem. The fourth and longest contig (GenBank TM accession number AAHK01002353) contained within its 6233 bp the complete TcSPPS coding region (locus tag Tc00.1047053509445.30), with identical sequence to TENU4155. The size of the fragments generated by complete digestion of total genomic DNA with NcoI (bands of ϳ1000 and 2500 bp, Fig. 2) agreed with the theoretical restriction map of this contig (data not shown). The

Solanesyl-diphosphate Synthase from T. cruzi
same was observed with the AatII digestion pattern (band of ϳ1000 bp, data not shown). These experiments support the idea that TcSPPS is a single copy gene.
Ubiquinone Detection in Complemented E. coli and in T. cruzi Epimastigotes-E. coli has an octaprenyl-diphosphate synthase (44). To test whether the final product of TcSPPS, UQ-9, was able to replace the essential functions of UQ-8, a plasmid-swapping test was carried out using the insertion mutant strain KO229. Several colonies that were ampicillinresistant (carrying the TcSPPS-expressing plasmid) and sensitive to spectinomycin (free of the ispB-expressing plasmid) were obtained, suggesting that TcSPPS was fully functional in bacteria. When the UQ of complemented colonies were isolated and their length determined, UQ-9 was the main product detected (Fig. 3a). This is in contrast to KO229 harboring pKA3, which produced mainly UQ-8 (Fig. 3b). To establish the length of the native molecule in the parasite, UQ was extracted from 1 g of T. cruzi epimastigotes and run in HPLC (Fig. 3c). The peak corresponded to UQ-9, as reported for other trypanosomatids (12)(13)(14)(15).
Functional Analysis of the Candidate Long-chain Prenyldiphosphate Synthase-For functional analysis, the protein was heterologously expressed in E. coli cells as a fusion protein with an N-terminal polyhistidine tag and purified by affinity chromatography (Fig. 4A). The purified protein runs close to its predicted molecular mass.
The enzymatic activity assay was performed in the presence of different concentrations of Mg 2ϩ and Mn 2ϩ , to determine their effect on the TcSPPS when the allylic substrate was GGPP. Mg 2ϩ and Mn 2ϩ were added to the reaction mixture at concentrations between 0.5 and 20 mM. As shown in Table 1, optimal levels of activity were obtained by the addition of 0.5-1 mM Mg 2ϩ . The addition of 10 mM EDTA abolished SPPS activity. Enzymatic activity was not detected when the divalent cation was Mn 2ϩ . The T. cruzi enzyme activity was also assayed between 0.5 and 5% (v/v) Triton X-100. Maximum activity was observed at 1% (v/v) Triton X-100 (using DMAPP, FPP, GPP, or GGPP as primer) (data not shown).
Four kinds of allylic diphosphates were tested as a primer substrate with [4-14 C]IPP as described under "Experimental Procedures." The enzyme utilized the four allylic diphosphates as a primer substrate, however, the enzymatic activity using DMAPP as substrate was one order of magnitude lower than the enzymatic activity using FPP, GPP, or GGPP ( Table 2). The reaction products were dephosphorylated and then analyzed by   reversed-phase TLC. When either FPP, GPP, or GGPP was used as the primer substrate, solanesol (C 45 ) was predominantly detected by TLC analysis indicating that the protein generated solanesyl diphosphate (C 45 ) as the major product (Fig. 4B). When the primer substrate was DMAPP the labeled products were almost undetectable by TLC.
Kinetic Analysis-Standard procedures were used to determine kinetic parameters. K m and V max values were obtained by a non-linear regression fit of the data to the Michaelis-Menten equation (SigmaPlot 2000 for Windows), and the results are summarized in Table 3. TcSPPS showed a similar K m value for FPP and GGPP, however the K m value for GPP was 7.7-fold higher than that for GGPP. Moreover, the k cat value for GGPP was 2.7-fold higher than that for FPP and slightly higher than that for GPP. Consequently the k cat /K m value for GGPP was 1.8-fold higher than that for FPP and 10.4-fold higher than that for GPP, indicating that TcSPPS prefers GGPP to FPP or GPP. Moreover, TcSPPS showed a similar K m value for IPP when the allylic substrate was FPP, GPP, or GGPP. However, when GGPP was used as primer substrate, the enzyme showed a 1.6-fold higher k cat value for IPP than that for IPP with GPP, and a 5.0-fold higher than that for IPP with FPP. Consequently, the k cat /K m values for IPP with GGPP or GPP were similar, although the k cat /K m value for IPP with GGPP was 6.3-fold higher than that for IPP with FPP. These results suggest the preference of TcSPPS for GGPP over FPP or GPP.
Presence of mRNA and Protein in T. cruzi-Several assays were used in an attempt to establish the presence of the specific mRNA in CL Brener clone epimastigotes. Neither Northern blot analysis nor RT-PCR amplification allowed the detection of even a negligible amount of mRNA. On the other hand, the mRNA for the positive control used (TcCyP19 gene) was detected in the Northern and RT-PCR assays. A search was then done on an expressed sequence tag clusters data base (TcruziDB Version 4.1). A cluster (99127) composed of one read (TcEST_NCBI_AW324852.1) was identified. Its sequence (316 bp) was identical to the 3Ј-end of the original expressed sequence tag clone, TENU4155 (1262 bp).
Immunoblot analysis with affinity-purified antibody against TcSPPS showed a band of ϳ36 -37 kDa present in all developmental stages of the Y strain (Fig. 5). The antibody also recognized a 85-kDa band in all three stages and more weakly a 40-kDa band. These proteins might share a few epitopes with SPPS in the motif regions.
To further characterize the anti-TcSPPS cross-reaction to other proteins in the parasite, we performed subcellular fractionation experiments to enrich for glycosomes. Fraction D, enriched for glycosomes, showed the strongest reaction by Western blot analysis with antibodies against TcSPPS (Fig. 6A) and TbgGAPDH, a glycosomal marker (Fig. 6B). In this enrichment experiment, the 40-and 85-kDa bands were not detected in    Solanesyl-diphosphate Synthase from T. cruzi DECEMBER 22, 2006 • VOLUME 281 • NUMBER 51 the pellet fraction (fraction D), which was enriched in glycosomes, but were present in the previous steps of the fractionation (fractions A-C) (Fig. 6A). These bands could be soluble proteins, probably cytoplasmic, that did not localize in glycosomes. The localization of TcSPPS was further investigated. It is worth mentioning that no import signals (either to endoplasmic reticulum or to glycosomes) were detected. Immunofluorescence microscopy of T. cruzi epimastigotes showed co-localization in glycosomes of TcSPPS (Fig. 7B) with pyruvate phosphate dikinase (Fig. 7C), a known marker of glycosomes (45), as detected with monoclonal antibodies prepared against the recombinant TbP-PDK. TcSPPS (Fig. 7F) did not co-localize with markers for acidocalcisomes (vacuolar pyrophosphatase) (Fig. 7G) or with mitochondrial markers (MitoTracker; Fig. 7, J and K).
Immunogold electron microscopy confirmed the co-localization of TcSPPS (Fig. 8, A-C, 18 nm gold particles) with pyruvate phosphate dikinase (Fig. 8, A-C, 12 nm gold particles). The density of TcSPPS gold particles in the glycosomes was significantly higher than in other compartments (Fig. 8D). No background staining was observed when secondary antibodies were used alone in immunofluorescence or immunogold electron microscopy assays (data not shown).

DISCUSSION
We report here the functional characteristics of the solanesyl-diphosphate synthase of T. cruzi (TcSPPS). Heterologous expression of the TcSPPS gene in E. coli resulted in the production of a recombinant enzyme that was similar to other SPPSs with respect to its Mg 2ϩ requirement but differed in having GGPP as preferred substrate. TcSPPS could complement the function of the corresponding polyprenyl-diphosphate synthase of E. coli, and the cells produced mainly UQ-9. The enzyme was shown to localize in the glycosomes, supporting the role of these organelles in isoprenoid synthesis. It is interesting to note that the protein carried neither a PTS1 nor a PTS2 signal, arguing for an alternative import mechanism (for a review, see Ref. 46). This is the first report of a gene encoding a functional SPPS in a trypanosomatid and on its localization in the glycosomes.
Peroxisomes harbor a number of enzymes involved in the isoprenoid pathway. They have been shown to contain acetoacetyl-CoA thiolase (47), 3-hydroxy-3-methylglutaryl (HMG)-CoA synthase (48), HMG-CoA reductase (49), mevalonate kinase (50), and FPPS (51). In addition, they also have a transprenyltransferase responsible for the generation of SPP and a nonaprenyl-4-hydroxybenzoate transferase involved in the first step of UQ-9 synthesis (52). The soluble HMG-CoA reductase from T. cruzi epimastigotes was initially shown to be associated with the glycosomes of this trypanosomatid (53). However, more recent immunogold labeling studies using ultrathin sections of T. cruzi epimastigotes and polyclonal/monoclonal antibodies against T. cruzi recombinant HMG-CoA reductase and digitonin solubilization experiments suggested that the enzyme is predominantly located inside the mitochondrial matrix (54). Our results indicate that SPPS is located in the glycosomes of T. cruzi. Because this enzyme catalyzes a PP i -producing reaction, the presence in glycosomes of the PP i -consuming pyruvate phosphate dikinase would make this reaction thermodynamically possible (55,56).
TcSPPS prefers GGPP over GPP and FPP as substrate. This is in contrast to rat trans-prenyltransferase that prefers GPP and IPP as substrates (57). As occurs with the rat enzyme (57), TcSPPS poorly utilized DMAPP. In fact, most SPPSs characterized to date (for example, the ones obtained from rat liver, spinach leaves, and Micrococcus lysodeikticus) (57)(58)(59) prefer GPP as allylic substrate, so TcSPPS together with A. thaliana SPPS1 are the only ones to preferentially use GGPP (42).  We succeeded in complementing E. coli deficient in polyprenyl-diphosphate synthase. A number of genes encoding polyprenyl-diphosphate synthases from bacteria have been successfully expressed in these E. coli mutants (33,60). The TcSPPS protein is assumed to work as a homodimer, because it is functional without any other genes, unlike the case of Bacillus subtilis heptaprenyl-diphosphate synthase (heterodimer) (61), fission yeast decaprenyl-diphosphate synthase (62), mouse solanesyl-diphosphate synthase (63), and human decaprenyldiphosphate synthase (heterotetramers) (63).
Ubiquinone is synthesized de novo in both prokaryotes and eukaryotes. The two parts of the molecule, the benzoquinone ring and the isoprene chain, are synthesized independently and assembled in a reaction catalyzed by a prenyl-4-hydroxybenzoate transferase (64). 4-Hydroxybenzoate originates from tyrosine or phenylalanine in eukaryotes or from acetate through the shikimate pathway in most prokaryotes (65). In trypanosomatids, ubiquinone biosynthesis was investigated in L. major and T. brucei (12,16) where it was found that the isoprenoid portion of the molecule is synthesized by the mevalonate pathway (both parasites) and that the ring is synthesized from acetate or aromatic amino acids (Leishmania).
Regarding T. cruzi UQ chain length, it seems clear, from our in vitro (TLC experiment with the purified enzyme in the presence of different substrates) and in vivo (HPLC of native UQ) results, that epimastigotes synthesize and retain mainly UQ-9 in their membranes. Valuable functions, even non-traditional ones (66), have been ascribed to this molecule, granting further studies on the importance of the TcSPPS as a new chemotherapeutic target.
Bisphosphonates have been shown to have activity in vitro and in vivo against T. cruzi (1). Some of these compounds (nitrogen-containing bisphosphonates) target the farnesyldiphosphate synthase (5). However, bisphosphonates can also inhibit other prenyltransferases, and it is expected that bisphosphonates with long side chains, which are known to be potent inhibitors of the intracellular form of T. cruzi (67), could target prenyltransferases like TcSPPS.