Molecular Cloning and Characterization of a Protein Farnesyltransferase from the Enteric Protozoan Parasite Entamoeba histolytica*

Genes encoding α- and β-subunits of a putative protein farnesyltransferase (FT) from the enteric protozoan parasite Entamoeba histolytica were obtained and their biochemical properties were characterized. Deduced amino acid sequences of the α- and β-subunit of E. histolytica FT (EhFT) were 298- and 375-residues long with a molecular mass of 35.6 and 42.6 kDa, and a pI of 5.43 and 5.65, respectively. They showed 24% to 36% identity to and shared common signature domains and repeats with those from other organisms. Recombinant α- and β-subunits, co-expressed in Escherichia coli, formed a heterodimer and showed activity to transfer farnesyl using farnesylpyrophosphate as a donor to human H-Ras possessing a C-terminal CVLS, but not a mutant H-Ras possessing CVLL. Among a number of small GTPases that belong to the Ras superfamily from this parasite, we identified EhRas4, which possesses CVVA at the C terminus, as a sole farnesyl acceptor for EhFT. This is in contrast to mammalian FT, which utilizes a variety of small GTPases that possess a C-terminal CaaX motif, where X is serine, methionine, glutamine, cysteine, or alanine. EhFT also showed remarkable resistance against a variety of known inhibitors of mammalian FT. These results suggest that remarkable biochemical differences in binding to substrates and inhibitors exist between amebic and mammalian FTs, which highlights this enzyme as a novel target for the development of new chemotherapeutics against amebiasis.

Ras small GTPases function as a molecular switch of signal transduction in cell proliferation and differentiation (1). Ras small GTPases require a post-translational lipid modification called protein farnesylation in order to become membraneassociated and functional (2). Protein farnesylation, catalyzed by protein farnesyltransferase (FT) 1 (3), which is comprised of two heterologous ␣-(FT␣) and ␤-(FT␤) subunits, is a major post-translational lipid modification, together with protein geranylgeranylation (3). FT and protein geranylgeranyltransferase type I (GGT-I) catalyze the transfer of the farnesyl and geranylgeranyl group from farnesyl pyrophosphate (FPP) and geranylgeranyl pyrophosphate, respectively, to the cysteine residue of a C-terminal CaaX of small GTPases including Ras, Rac, and Rho, where C is cysteine, a is usually an aliphatic amino acid, and X is any amino acid. Marked differences in substrate specificity have been shown between FT and GGT-I, i.e. FT mainly utilizes, as substrates, small GTPases possessing the terminal CaaX motif, when X is serine, methionine, glutamine, cysteine, or alanine (4), whereas GGT-I prefers proteins with the C-terminal CaaL or CaaF motif (4). Among well characterized Ras proteins that terminate with a CaaX motif, human H-Ras-CIMF, N-Ras-CVVM, K-RasA-CIIM, and Rap2-CNIQ are known to be farnesylated by FT, while Rap1A-CLLL, as well as Rho family proteins are geranylgeranylated by GGT-I. It has also been shown that K-RasB-CVIM can be either farnesylated by FT or geranylgeranylated by GGT-I (5). Since constitutively active mutations of Ras proteins have been shown to induce carcinogenesis (6 -8), which is suppressed by the inhibition of farnesylation, FT has attracted attention as a target of cancer chemotherapy (9). In addition, several compounds targeting FT have proven promising against African sleeping sickness caused by Trypanosoma brucei and Malaria caused by Plasmodia species (10).
Entamoeba histolytica is an intestinal protozoan parasite, which causes amebic dysentery, colitis, and liver abscess in humans, and is responsible for an estimated 50 million cases of amebiasis and 40 -100 thousand deaths annually (11). A number of small GTPases have been studied including Ras/Rap (12,13), Rho/Rac (14 -18), and Rab (19 -21). The completion of the E. histolytica genome data base will help us to comprehensively understand the presence and complexity of these small GTPases in the ameba. The molecular and cellular functions of some of these small GTPases have begun to be unveiled (12,17,18,22). However, the molecular basis of the lipid modification of these small GTPases remains largely unknown in this parasite.
In this report, we describe the molecular and biochemical characterization of the ␣and ␤subunits of FT of E. histolytica (EhFT) using recombinant proteins co-expressed in Escherichia coli. We also show that only one amebic Ras protein among the many small GTPases tested is farnesylated by EhFT. In addition, we show that the amebic FT exhibits marked resistance to a variety of compounds that are known to inhibit mammalian FT, indicating that the amebic FT possesses distinct biochemical properties from the mammalian FT.

EXPERIMENTAL PROCEDURES
Parasite-Trophozoites of E. histolytica strain HM-1:IMSS cl6 (23) were cultured axenically in BI-S-33 medium at 35.5°C (24). Restriction endonucleases and modifying enzymes were purchased from Takara Biochemical (Tokyo, Japan). The other chemicals and reagents used were from either Sigma-Aldrich Fine Chemicals (Tokyo, Japan) or Wako Pure Chemical Industries (Osaka, Japan) unless otherwise mentioned and of the highest purity available.
cDNA Library of E. histolytica-A trophozoite cDNA library of E. histolytica was constructed using the poly(A) ϩ RNA and ZAP II phage (Stratagene, La Jolla, CA) as described previously (25).
Identification and Cloning of FT␣ and FT␤ of E. histolytica-We designed oligonucleotide primers to amplify FT␣ and FT␤ proteincoding regions from E. histolytica by PCR based on a homology search using yeast and mammalian FT in the E. histolytica genome data base available at The Institute for Genomic Research (www.tigr.org/tdb/). The sense primer for EhFT␣ was 5Ј-ATGGAAGAAGACGAAGAAATC-ACATTTG-3Ј. Sense and antisense primers for EhFT␤ were 5Ј-ATGG-AAATTGAAGAAGTAGAAGTAGAAACTGTTAC-3Ј and 5Ј-TTAGAGC-GAACGGAAATACTCACAACGCTTATC-3Ј, respectively. Since we did not find a sequence containing the C terminus of FT␣ in the data base, we used T7 primer, located downstream of the cloning site of our cDNA library, to amplify the FT␣-coding region. PCR was performed using a one-hundredth volume of the cDNA phage lysate as template with the following parameters. An initial step of denaturation at 94°C for 3 min was followed by 30 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 2 min, and extension at 72°C for 2 min. A final step at 72°C for 10 min was used to complete the extension. The amplified DNA fragments were electrophoresed, purified using a Geneclean II kit (BIO101, La Jolla, CA), and cloned into the SmaI site of pUC18 using a SureClone Ligation Kit (Amersham Biosciences). Nucleotide sequences were confirmed with an ABI Prism BigDye terminator cycle sequencing ready reaction kit (PE Applied Biosystems, Foster City, CA) on an ABI Prism 310 Genetic Analyzer.
Identification and Cloning of Ras Small G Proteins of E. histolytica-To identify substrates for the amebic FT, we searched for putative Ras homologues in the E. histolytica Genome data base via a TBLASTN search using amebic (EhRas1and EhRas2) and mammalian Ras as inquiry sequences. We identified 8 previously uncharacterized putative full-length Ras genes. The C terminus of these Ras proteins ended with Phe (four genes), Leu (two), Met (one), or Ala (one). The two latter genes encoding putative Ras proteins containing the CSVM-or CVVA-C terminus (identical to EH02830 and EH01021 in the E. histolytica genome data base) were assumed to be good candidates to be farnesylated by the amebic FT, designated EhRas3 and EhRas4, respectively, and characterized further. A protein coding region of EhRas1-4, and EhRacC were amplified by PCR using cDNA as template and appropriate primers based on the sequences in the genome data base.
Sequence Analysis-FT␣ and FT␤ protein sequences from E. histolytica and 9 other organisms, and 20 putative Ras and Ras-related proteins of E. histolytica were retrieved from the TIGR and the National Center for Biotechnology Information data bases (www.ncbi.nih.gov/) using the BLASTP and TBLASTN algorithms. The protein alignment and phylogenetic analyses were performed with ClustalW version 1.81 (26) using the neighbor-joining (NJ) method (27) with the Blosum matrix created using the ClustalW program (26). Unrooted NJ trees were drawn with TreeView ver.1.6.0 (28). Branch lengths and bootstrap values (1000 replicates) (29) were derived from the NJ analysis.
Construction of a Plasmid to Express Recombinant EhFT-A plasmid containing the protein-coding regions of FT␣ (without the stop codon), FT␤ (with the stop codon), and the ribosome binding sequence (GAG-GAGTTTTAACTT) between them were constructed by three rounds of PCR using the recombinant approach (30,31). Briefly, two sets of initial PCRs were conducted to amplify the FT␣ and FT␤ protein-coding region using a sense primer, 5Ј-ATGGAAGAAGACGAAGAAATCACATTTG-3Ј and an antisense primer, 5Ј-ATGATTAGTAATTTTTGTTAAATACCA-ATCCC-3Ј (for FT␣); a sense primer, 5Ј-ATGGAAATTGAAGAAGTAG-AAGTAGAAACTGTTAC-3Ј and an antisense primer, 5Ј-TAAGAGCG-AACGGAAATACTCACAAGCCTTATC-3Ј (for FT␤). Two sets of second PCRs were conducted using the respective product of the first reaction as a template. To amplify the FT␣ protein-coding region (excluding the stop codon), flanked by a BamHI site (italicized) and the ribosome binding site (underlined), a sense primer, 5Ј-GGAGGATCCCATGGAA-GAAGACGAAGAAATCACATTTG-3Ј (primer 1) and an antisense primer, 5Ј-AAGTTAAAACTCCTCATGATTAGTAATTTTTGTTAAATA-CCAATCCC-3Ј, were used. To amplify the FT␤ sequence including the stop codon, flanked by the ribosome binding site (underlined) and a HindIII site (italicized), sense, 5Ј-GAGGAGTTTTAACTTATGGAAAT-TGAAGAAGTAGAAGTAGAAACTGTTAC-3Ј and antisense, 5Ј-CCAA-AGCTTTAAGAGCGAACGGAAATACTCACAAGCCTTATC-3Ј (primer 2) were used. The third round of PCR was conduced using a mixture of the products of the second round, and primers 1 and 2. The resulting 2.1-kb PCR product was digested with BamHI and HindIII and ligated into BamHI-and HindIII-double digested pQE31 to construct pE-hFT␣␤. In pEhFT␣␤, the FT␣ and FT␤ protein-coding regions placed in tandem were presumably translationally coupled, facilitating co-expression of these two subunits at similar levels. An N-terminal histidine tag was also engineered in EhFT␣ to facilitate purification.
Expression and Purification of Recombinant Proteins-Plasmids constructed as described above were introduced into E. coli M15 cells. A 12-ml seed culture was grown overnight at 37°C in LB medium containing 100 g/ml of ampicillin and 25 g/ml of kanamycin. The overnight culture was then inoculated into 250 ml of fresh medium containing the antibiotics. The bacteria were grown for 1 h, and then another 4 h after the addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside to induce protein expression. The bacteria were harvested by centrifugation at 4,000 ϫ g for 20 min, and the pellet was stored at Ϫ20°C until purification. The recombinant proteins were purified according to the manufacturer's instructions. Briefly, the bacterial cells were resus-pended in cold lysis buffer, phosphate-buffered saline, pH 8.0, containing 10 mM imidazole and 1% lysozyme, sonicated, and centrifuged at 10,000 ϫ g for 20 min. The supernatant was applied to a Ni-NTAagarose column (Qiagen, Hilden, Germany), washed extensively with the wash buffer containing 20 mM imidazole, and eluted with the lysis buffer containing 250 mM imidazole. The recombinant FT proteins were further purified with Q Sepharose Fast Flow (Amersham Biosciences) at a flow rate of 0.5 ml/min as described (32). The purified recombinant FT and Ras proteins were then dialyzed against the enzyme assay buffer described below and 40 mM Tris-HCl, pH 8, containing 90 mM NaCl, 10 mM MgCl 2 , and 2 mM dithiothreitol and stored with 20 and 50% glycerol, respectively, at Ϫ80°C until use. Protein concentrations were determined by the method of Bradford (33) using Protein Assay CBB solution (Nacalai Tesque, Kyoto, Japan). Bovine serum albumin was used as the protein standard.
Enzyme Assays-The enzymatic activity of recombinant FT and the whole lysate of E. histolytica trophozoites were assayed for incorporation of [ 3 H]farnesyl pyrophosphate or [ 3 H]geranylgeranyl pyrophosphate into the recombinant small GTPases prepared as described above, human H-Ras-CVLS, or H-Ras-CVLL. The assay was performed essentially as described previously (35)  . The filters were washed with 4 ml of 100% ethanol and then subjected to scintillation counting (LS 6,000IC, Beckman Coulter, Fullerton, CA). The K m value was calculated from Lineweaver-Burk plots. FT assays were also conducted in the presence of known FT inhibitors: farnesylpyrophosphate analogues (FPT inhibitor-I, FPT inhibitor-II, Gliotoxin, ␣-hydroxyl farnesylphosphonic acid) and a peptidomimetic inhibitor (FTI-276) under the condition described above.
Guanine Nucleotide Binding Assays-GTP binding activity was measured using [ 35 S]GTP␥S (Amersham Biosciences) and a nitrocellulose filter (Millipore HA filter, Millipore Corporation) as previously described (36).

Features of FT␣ and FT␤ of E. histolytica-Nucleotide sequences of EhFT␣ and
EhFT␤ obtained by PCR were identical to sequences available from the genome data base (EH02829 and EH04188, respectively). The putative protein-coding region of EhFT␣ and EhFT␤, consisting of 894 and 1,125 bps, encodes proteins of 298 and 375 amino acids with a calculated molecular mass of 35.6 and 42.6 kDa and a pI of 5.4 and 5.7, respectively. A search for previously identified domains and motifs (37) using the NCBI Conserved Domain Search revealed that both EhFT␣ and EhFT␤ contained well conserved signature domains shared by other organisms. EhFT␣ contained one BET4 domain and five "protein prenyltransferase ␣-subunit repeats"; EhFT␤ possessed one CAL1 domain and five "prenyltransferase and squalene oxidase repeats" (Fig. 1). The deduced protein sequences of EhFT␣ and EhFT␤ were aligned with those from other organisms using the ClustalW program (Fig. 1). Both EhFT␣ and EhFT␤ were the smallest in size, and lack any recognizable secretory signal sequence, an organelle targeting signal, and any domain implicated in membrane association including the transmembrane domain and myristylation signal. EhFT␣ and EhFT␤ also lack the N-terminal extension of 15-58 amino acids found in these subunits from other organisms. EhFT␣ showed 29, 30, 27, and 24% positional identity with FT␣ of human, Arabidopsis thaliana, Saccharomyces cerevisiae, and T. brucei, respectively; EhFT␤ revealed 36, 35, 28, and 31% positional identity with the FT␤ of these organisms, respectively. All the residues implicated to be essential for catalysis (Fig. 1, A and B) (39 -41) are conserved in both EhFT␣ and EhFT␤.
Phylogenetic Analysis of EhFT␣ and FT␤-Phylogenetic trees of EhFT␣ and EhFT␤ were constructed (Fig. 2). Neither the ␣nor ␤-subunit of the amebic FT showed statistically significant (i.e. supported with high bootstrap values) affinity to those from other organisms while trypanosomal, mammalian, and plant proteins encoding ␣and ␤-subunits formed well supported independent clades, representing distinct groups. These results were compatible with the notion that both subunits of EhFT developed independently from other eukaryotes, suggestive of the presence of unique biochemical properties of EhFT (see below).
Expression of EhFT in E. coli-A complex consisting of EhFT␣ and EhFT␤ was expressed and purified as described under "Experimental Procedures." Purified proteins revealed two major proteins with an equal intensity of an apparent molecular mass of 38 and 43 kDa (Fig. 3), which likely correspond to EhFT␣ and EhFT␤, respectively. The apparent molecular mass of ␣and ␤-subunits agreed well with a theoretical value of 37.6 kDa with the N-and C-terminal addition of MRGSHHHHHHTDP and EEF, respectively, for the recombinant EhFT␣ and 42.6 kDa for EhFT␤. When histidine-tagged EhFT␣ or EhFT␤ were expressed separately, the apparent molecular mass of these proteins agreed well with their identity (data not shown). Densitometric quantitation of these two bands also supported that they contain an equal number of protein molecules (data not shown). Gel filtration chromatography of the recombinant EhFT using Sephacryl S-300 revealed a molecular mass of about 80 kDa (data not shown). Thus, the recombinant EhFT␣ and EhFT␤ were present as a stable dimer during the process of purification by Ni-NTA agarose and Q Sepharose Fast Flow chromatography (Fig. 3). After the purification, recombinant EhFT was estimated to be Ͼ95% pure by densitometric quantitation (data not shown), and further utilized for enzymatic assays.

Demonstration of FT Activity of the Recombinant EhFT against Human Ras Proteins-When assayed for incorporation of [ 3 H]farnesyl pyrophosphate, the recombinant
EhFT showed FT activity [1.03 Ϯ 0.012 nmol of FPP/mg of protein (mean Ϯ S.E.)] against human recombinant wild-type H-Ras-CVLS, whereas it showed ϳ20-fold less activity against mutant H-Ras-CVLL (0.05 Ϯ 0.01 nmol of FPP/mg of protein) (Fig. 4), which was previously shown to be predominantly geranylgeranylated by human GGT-I. The addition of EDTA (10 mM) to the reaction mixtures completely abolished the enzymatic activity of the recombinant EhFT (data not shown), suggesting, as shown for mammalian and yeast FT, that EhFT also requires Zn 2ϩ and Mg 2ϩ for its activity.
Identification and Phylogenetic Analysis of the Novel Ras Proteins in E. histolytica-To identify the protein substrates of EhFT in the parasite, we searched for putative Ras proteins in the genome data base based on homology to the EhRas1 and human H-Ras. In addition to EhRas1, EhRas2, EhRap1, and EhRap2, which have already been reported (12), we found 8 additional putative ras proteins previously uncharacterized, with the C-terminal CaaX motif. Of these 8 proteins, 6 possess Phe or Leu (4 with Phe, 2 with Leu) at the C terminus, while 1 each has Met or Ala. We tentatively designated proteins possessing Met or Ala as EhRas3 or EhRas4, and the other proteins as EhRas5-10. The nucleotide sequence of the EhRas1 cDNA we cloned was identical to that previously reported; the nucleotide sequence of our EhRas2 cDNA differed at one nucleotide (A368G) from the sequence previously reported (12), resulting in a Y123C substitution. EhRas3 and 4 consisted of 210 and 182 amino acids with a calculated molecular mass of 23.9 and 20.6 kDa and a pI of 5.5 and 5.8, respectively. The ClustalW multiple alignment showed that EhRas3 and 4, together with the previously identified EhRas1, EhRas2, EhRap1, and EhRap2, share conserved GTP binding consensus sequences (42) and also, at a moderate level, the effector binding region (42) (Fig. 5A).
Percent identity among the EhRas1-4 and EhRap1-2 ( Fig.  5B) also indicates that EhRas3 is, together with EhRas5 and EhRas6 (EhRas5 versus EhRas1-2, 62-66%; EhRas6 versus EhRas1-2, 39 -41%; EhRas5 and EhRas6 were not studied further in the present work) closely associated with EhRas1 The regions corresponding to the motif PXNYXXWYR (37), previously found in the turns connecting two helices of the coiled-coil, in FT␣ and a glycine-rich motif GGFXGXXP (37), corresponding to the loop regions that connect the C termini of the peripheral helices with the N termini of the core helices in the barrel (38) in FT␤ are boxed. All amino acids implicated in catalysis by crystallographic and mutational studies of mammalian FTs (38 -41) are shaded. Aromatic amino acids located in the hydrophobic cleft at the center of the ␣-␣ barrel implicated in the interaction with the farnesyl residue are marked with open circles (38). Arg 202 implicated in the binding of the essential C-terminal carboxylate of the protein substrate is marked with a filled circle (39). Amino acids implicated in the coordination of zinc are marked with open squares (40). Amino acids implicated in the binding and utilization of FPP, shown by a mutational analysis (39) are marked with filled squares. N-terminal extensions absent in EhFT␣ and EhFT␤ are also boxed. DDBJ/EMBL/GenBank TM accession numbers are given in Fig. 2. and EhRas2, showing 48 -51% identities, whereas EhRas4 showed only 26 -30% identities to EhRas1, EhRas2, EhRap1, and EhRap2. Phylogenetic reconstructions (Fig. 6) confirmed the results of the protein alignment: both EhRas3 and EhRas4, together with EhRas5 and EhRas6, represent new members of the Ras/Rap family. Rac and Rab proteins were categorized to independent clades, whose association was well supported by moderate to high bootstrap values (only representative Rac and Rab proteins were included in this analysis).
Identification of EhRas4 as a Substrate of Eh FT-We tested substrate specificity of EhFT toward EhRas1-4. We chose EhRas3-4, together with EhRas1-2, as possible candidates for EhFT substrates because it was previously shown that mammalian and yeast small GTPases with a C-terminal Ser, Met, Gln, Cys, or Ala have a tendency to be farnesylated whereas those containing Phe or Leu at the C-terminal end tend to be geranylgeranylated (4). The recombinant EhFT showed farnesyltransferase activity toward EhRas4 (1.03 Ϯ 0.005 nmol of FPP/mg of protein), which was comparable to the activity toward human H-Ras-CVLS (Fig. 4). In contrast, EhFT showed no detectable or only minimal activity toward EhRas1-3. We also tested if EhRab, EhRac, or EhRap are farnesylated by EhFT. The recombinant EhFT did not transfer farnesyl to EhRab7, EhRacC, or EhRap2 (data not shown). Furthermore, the recombinant EhFT did not utilize geranylgeranyl pyrophosphate as a isoprenyl donor to transfer the geranylgeranyl residue to EhRas1, EhRas2, EhRas3, EhRas4, EhRap2, EhRacC, or EhRab7 (data not shown). To confirm that EhRas3 and 4 are capable of binding GTP, a GTP binding assay was conducted. Both EhRas3 and 4 showed comparable GTP-binding activity to EhRas2 and EhRab7 (data not shown). We also assayed for the FT activity in the whole lysate of the E. histolytica trophozoites. Among the 2 human H-Ras and 4 E. histolytica Ras homologues described above, FT activity was detected only against human H-Ras-CVLS and EhRas4 in the whole lysate (data not shown), which excluded the possibility that some other EhFT protein (or proteins) exists to farnesylate these small GTPase in the amebic lysate.
Kinetic Properties of EhFT-Lineweaver-Burk plots showed the K m of recombinant EhFT for EhRas4 to be 5.13 Ϯ 0.02 M (plots not shown), significantly higher than that of bovine FT, the K m of which for Ras-CVLS and Ras-CVIM is 0.63 Ϯ 0.05 Sensitivity of Recombinant EhFT to Human FT Inhibitors-We examined the sensitivity of EhFT to known FT inhibitors. As shown in Table I, EhFT was virtually insensitive at up to 30 M except to FPT inhibitor-II and FPT 276 when recombinant EhRas4 was used as substrate. The lack of sensitivity of EhFT to FT inhibitors was not dependent on the substrate; EhFT was also insensitive to these inhibitors when H-Ras-CVLS was used as a substrate.

DISCUSSION
In this study, we have demonstrated for the first time the molecular identity of FT in enteric parasitic protozoa. Although all major subgroups of small GTPases, i.e. Ras, Rap, Rho, Rac, Rab, Arf, and Ran, have been identified (12, 14 -21) in E. histolytica and some of their functions studied (17,18), this is the first demonstration and characterization of an isoprenylation enzyme essential for correct membrane topology and organelle targeting of these small GTPases. We identified common features of FT hitherto recognized in other eukaryotes: both FT␣ and FT␤ contained well-conserved signature domains such as BET4 and CAL1 domains, and the repeats, i.e. protein prenyltransferase ␣-subunit repeats and prenyltransferase/ squalene oxidase repeats (Fig. 1). EhFT forms a heterodimeric Sequences over the alignment are GTP binding consensus sequences. Asterisks (*) and dots (.) indicate identical amino acids and conserved amino acid substitutions, respectively. GXXXXGKS and DXT depict GXXXXGK(S/T) and D(X) n T consensus sequences, respectively. DDBJ/ EMBL/GenBank TM accession numbers are given in Fig. 6 complex with a ratio of 1:1 between ␣and ␤-subunits, similar to the case in other organisms, as shown by co-purification (Fig.  3). Phylogenetic analyses indicate that both EhFT␣ and EhFT␤ are equally distant from homologues from other organisms. This may partially explain some of the unique biochemical characteristics of the amebic FT not shared by the mammalian counterpart. It is also worth noting that trees of both ␣and ␤-subunits are similar (Fig. 2), suggesting that the FT subunits co-evolved independently at a comparable rate in these organisms.
We identified EhRas4-CVVA as one of the intrinsic substrates of FT in E. histolytica. Although it was not possible to test all small GTPases as substrates for EhFT, we showed that EhFT exclusively utilized EhRas4 as a farnesyl acceptor. In contrast, recombinant EhRas1-3, Rap2, RacC, and Rab7 were not farnesylated by recombinant EhFT (Fig. 5) or the whole cell extract. The fact that the amebic lysate contained the activity to transfer the farnesyl residue to EhRas4, but not other EhRas isotypes, reinforces the notion that the FT characterized in the present study is the sole FT in this organism and also specific for this Ras protein. We also tentatively concluded that FTmediated farnesylation is not a major lipidation of Ras protein in this organism. It was unexpected that EhFT did not utilize EhRas3, which terminates with CSVM, as a farnesyl acceptor, because mammalian and yeast small GTPases containing a C-terminal Ser, Met, Gln, Cys, or Ala were shown to prefer to be farnesylated (4). An unexpected substrate specificity was also previously reported for FT from another protozoan parasite T. brucei, which farnesylates Ras protein with CVIM, but not CVLS (43). The fact that EhFT prefers smaller amino acids at the C terminus of EhRas (CVVA and CVLS) indicates that the amebic FT may possess a smaller binding cleft for the Ras C terminus.
Among newly found putative Ras-like proteins, EhRas3-6 were the only ones that contained a terminal CaaX and also showed a closer kinship to EhRas1 and EhRas2 than to other small GTPases (i.e. Rap, Rac/Rho, and Rab) (Fig. 6). This observation, together with the lack of farnesylation by EhFT of EhRas1-3, Rap2, RacC, and Rab7, indicates that EhRas4 protein is the sole Ras protein farnesylated by EhFT. It is also conceivable that EhRas1-3 proteins with the C-terminal Phe, Leu, or Met, respectively, are farnesylated by GGT-I, as shown for rat RhoB-CKVL (44). This is also the case for EhRas2-CELL, which has been shown to be farnesylated by the recombinant E. histolytica GGT-I in our preliminary experiment (data not shown). Although the C terminus of the previously identified amebic Ras/Rap (i.e. EhRas1-2 and EhRap1-2) (12) was presumed to be geranylgeranylated, a study using rabbit reticulocytes lysates (as a source of enzyme) and recombinant EhRas1 and EhRap2 showed that these proteins were not geranylgeranylated, but farnesylated (13). Considering that recombinant EhFT neither farnesylates nor geranylgeranylates EhRas1 and EhRap2, we have to conclude that the results of the previous report (13) are likely a consequence of artifactual farnesylation by heterologous prenylase(s), as observed for EhRas2-CELL, which was farnesylated by the rat GGT-I (data not shown). Alternatively, it is conceivable that the farnesylation of these small GTPases by GGT-I may require an unidentified accessory factor, like Rab escort protein for GGT-II (3), in E. histolytica. Altogether, these results suggest that the substrate specificity of prenyltransferases varies widely among organisms. Further studies, including the cloning and enzymatic characterization of GGT-I of E. histolytica to determine if EhRas proteins are geranylgeranylated by the amebic GGT-I, are now underway.
Although we did not show a specific role for EhRas4, this protein shares all the conserved domains characteristic of Ras (42) except for incomplete DXAG and D(X) n T consensus sequences, and showed a close kinship to other EhRas proteins in the phylogenetic reconstruction (Fig. 6). We demonstrated that EhRas4 was capable of binding GTP (data not shown), verifying its identity as a small GTP-binding protein. EhRas1-4 lack a cysteine residue located 5-8 amino acids upstream of the C terminus to be palmitoylated in H-and N-Ras (45), which was shown to be essential for membrane association. In addition, EhRas4, in contrast to EhRas1-3, also lacks the so-called polybasic region (Fig. 5A), which was found in K-RasB and attributed to membrane association (46). The polybasic region was also implicated in interaction with a negatively charged patch on the surface of FT␤, which is located in close proximity to the region responsible for the binding to the Ras C terminus (38). Interestingly, EhFT␤, which shows low affinity to EhRas1-3 with the polybasic region and high affinity toward EhRas4 without it, possesses a number of substitutions of negatively charged with either positively-charged or neutral amino acids particularly in helixes 3-5 (38) (e.g. D91S, E94M, E112R, D115G, E116Y, E131R, E166V, E167N, and D170Q, corresponding to rat FT␤). It is conceivable that these substitutions compensate for the repulsive force that interferes with proper binding, which would partially explain the observed Ras specificity of the amebic FT.
In addition to its unique (i.e. EhRas4-specific) acceptor specificity, the amebic FT revealed notable differences in sensitivity against compounds known to inhibit human FT by distinct mechanisms (Table I). Marked differences in sensitivity to FPP analogues were unexpected since all the aromatic amino acids (Trp 102 , Tyr 105 , Trp 106 , Tyr 154 , Tyr 205 , Phe 253 , Phe 302 , Trp 303 , Tyr 361 , and Tyr 365 of rat FT␤) that were shown to be located in the hydrophobic cleft at the center of the ␣-␣ barrel and implicated to be essential for the interaction with FPP within the FPP-binding pocket (41) were conserved. FPT inhibitors I, II, and ␣-hydroxyfarnesyl phosphonic acid share the common farnesyl (C15) portion (36), which interacts with these aromatic residues lined on this hydrophobic cleft (38). Therefore, the lack of sensitivity of EhFT against these FPP analogues suggests that the binding specificity of these compounds does not depend on the structure of the FPP-binding pocket per se, but on the neighboring spacial and electrostatical environment. The fact that EhFT is Ͼ10-fold more resistant to FPT inhibitor I and ␣-hydroxyfarnesyl phosphonic acid than FPT inihibitor II whereas human FT is equally sensitive to these inhibitors agrees well with the notion that EhFT has higher affinity to FPT inhibitor II. Considering the major structural differences between FPT inhibitors I and II: the presence of the O-ester linkage and the absence of the C-terminal residue in FPT inhibitor II, the observed differences in sensitivity may be partially explained by the substitutions of negative with neutral/positive amino acids found in the amebic FT described above. It is conceivable that EhFT is not sensitive to the CaaM peptidomimetic FTI-276 (Ͼ1000-fold less than human FT) since EhRas3-CSVM was not a substrate of EhFT. Finally, exploitation of critical differences in the affinity toward substrates and inhibitors between the mammalian and amebic FT should enable us to discover or design novel inhibitors selective for EhFT, leading to the development of new chemotherapeutics against amebiasis.