Purification, Enzymatic Characterization, and Inhibition of theZ-Farnesyl Diphosphate Synthase from Mycobacterium tuberculosis *

We have recently shown that open reading frame Rv1086 of the Mycobacterium tuberculosis H37Rv genome sequence encodes a unique isoprenyl diphosphate synthase. The product of this enzyme, ω,E,Z-farnesyl diphosphate, is an intermediate for the synthesis of decaprenyl phosphate, which has a central role in the biosynthesis of most features of the mycobacterial cell wall, including peptidoglycan, arabinan, linker unit galactan, and lipoarabinomannan. We have now purified Z-farnesyl diphosphate synthase to near homogeneity using a novel mycobacterial expression system.Z-Farnesyl diphosphate synthase catalyzed the addition of isopentenyl diphosphate to ω,E-geranyl diphosphate or ω,Z-neryl diphosphate yieldingω,E,Z-farnesyl diphosphate andω,Z,Z-farnesyl diphosphate, respectively. The enzyme has an absolute requirement for a divalent cation, an optimal pH range of 7–8, and K m values of 124 μm for isopentenyl diphosphate, 38 μm for geranyl diphosphate, and 16 μm for neryl diphosphate. Inhibitors of the Z-farnesyl diphosphate synthase were designed and chemically synthesized as stable analogs ofω,E-geranyl diphosphate in which the labile diphosphate moiety was replaced with stable moieties. Studies with these compounds revealed that the active site of Z-farnesyl diphosphate synthase differs substantially from E-farnesyl diphosphate synthase from pig brain (Sus scrofa).

Isoprenyl diphosphate synthases catalyze the condensation of an allylic diphosphate with isopentenyl diphosphate (IPP, 1 C 5 ) via an electrophilic alkylation reaction to produce longer allylic diphosphates (1,2). Chain elongation continues until a physiologically appropriate chain length is reached, at which time the molecule may undergo further modifications (dephosphorylation, cyclization, or head-to-head condensation reac-tions). Polyprenyl phosphate (Pol-P) is formed by dephosphorylation of an allylic prenyl diphosphate chain. The predominant form of prokaryotic Pol-P is ,diE,polyZ-undecaprenyl phosphate 2 (C 55 ); however, there are documented exceptions in Paracoccus denitrificans (3) and in Mycobacterium sp. (4 -8). M. smegmatis contains heptaprenyl diphosphate (8) (C 35 , four saturated, three Z double bonds) and decaprenyl diphosphate (5) (C 50 , , one E, and eight Z double bonds), whereas M. tuberculosis contains only decaprenyl phosphate (6). Although the stereochemistry of decaprenyl phosphate from M. tuberculosis has not been determined, our enzymatic studies suggest that it has similar stereochemistry to decaprenyl phosphate from M. smegmatis (9).
Pol-P is central to prokaryotic cell wall synthesis as a sugar carrier, and it has been reported that the levels of Pol-P may be rate-limiting for in vivo cell wall synthesis (10 -13). Our laboratory has shown that Pol-P is instrumental in the synthesis of each component of the covalently linked peptidoglycan-arabinogalactan-mycolic acid cell wall core of mycobacteria, and other noncovalently associated macromolecules such as lipomannan and lipoarabinomannan (5,14,15). The importance of Pol-P is also demonstrated in vivo by the fact that M. tuberculosis (16) (and other Mycobacterium sp.) 3 are sensitive to the antibiotic bacitracin, which specifically binds isoprenyl diphosphate (17) intermediates in Pol-P synthesis, inhibiting both chain elongation and dephosphorylation reactions.
Evolutionarily, there appear to be two independent families of isoprenyl diphosphate synthases, based on the type of stereochemistry (E or Z) introduced at the products' new double bond. E-Isoprenyl diphosphate synthases are capable of catalyzing the chain elongation of a range of substrates, the smallest one being dimethylallyl diphosphate (DMAPP, C 5 ) with IPP to form ,E-geranyl diphosphate (,E-GPP, C 10 ). Other shortchain isoprenyl diphosphates (,E,E-farnesyl diphosphate (,E,E-FPP, C 15 ) and ,E,E,E-geranylgeranyl diphosphate (,E,E,E-GGPP, C 20 )) are generated by a similar mechanism with additional molecules of IPP. Medium-chain E-isoprenyl diphosphate synthases use the short-chain products as allylic substrates to produce compounds that are C 30 to C 50 in length. Medium-chain E-isoprenyl diphosphate synthases are homologous to the short-chain E-isoprenyl diphosphate synthases, because both types contain two signature aspartate motifs (DD(XX) 1-2 D).
Thus far, only seven protein sequences have been biochemically correlated with Z-isoprenyl diphosphate synthase activity, including undecaprenyl diphosphate synthases from Micrococcus luteus, Escherichia coli, Hemophilus influenzae, Streptococcus pneu-moniae, the dolichol synthase from Saccharomyces cerevisiae, and the decaprenyl diphosphate and farnesyl diphosphate synthases from M. tuberculosis. Prior to the identification of the Z-farnesyl diphosphate (Z-FPP) synthase from M. tuberculosis (9), all known Z-isoprenyl diphosphate synthases utilized ,E,E-FPP or ,E,E,E-GGPP as the allylic substrate, added multiple units of IPP, and released long-chain (C 45 and greater) isoprenyl diphosphate molecules with mixed stereochemistry (2).
The crystal structure of the short-chain E-isoprenyl diphosphate synthase (avian FPP synthase) has been determined (18), and mutagenesis studies have been performed (19 -21) providing a solid understanding of how the active site determines the chain length of the product (22,23). However, little is known about Z-isoprenyl diphosphate synthases. We have purified and enzymatically characterized the short-chain Zisoprenyl diphosphate synthase from M. tuberculosis, which catalyzes the first committed step in the synthesis of decaprenyl diphosphate, a molecule whose role in cell wall synthesis is likely essential.

Synthesis
All oxygen-and water-sensitive reactions (Scheme 1) were performed under dry argon atmosphere. 1 H NMR spectra of all compounds were obtained with a Varian Gemini 200 operating at 200 MHz in approximately 2% solution of CDCl 3 or D 2 O, using Me 4 Si or Me 3 Si(CH 2 ) 3 SO 3 Na as the internal standard. Column chromatographies were performed using 230-to 400-mesh silica gel (Merck) or reverse phase silica gel (Macherey-Nagel Polygosil 60-4063 C 18 ). Mass spectra were recorded on a VG 70-25S mass spectrometer or an HP-5988 A spectrometer. Reagents and solvents were purchased from Aldrich or Fluka.

Purification of Z-FPP Synthase
Open reading frame Rv1086 (26) was initially cloned into a commercially available protein expression vector pTBY2 (New England Bio-Labs). However, due to the apparent toxic effects of the Rv1086 fusion protein to the E. coli cells harboring the plasmid, the coding sequence that encoded the complete insert/fusion was moved into a mycobacterial expression vector for expression in M. smegmatis. The following primers were designed to amplify open reading frame Rv1086 from M. tuberculosis H37Rv genomic DNA: 5Ј-GGTACATATGGAGATCATC-CCCCCGCG-3Ј and 5Ј-GTCCTGCGCTAGGGCCCCCTGCCGTAGCTG-3Ј. NdeI and SmaI restriction endonuclease sites were engineered into the above primer sequences respectively (underlined). The native stop codon was removed from the C-terminal primer to allow for an in-frame fusion with the intein/chitin binding domain coded in the vector pTBY2. PCR was performed on a PerkinElmer Life Sciences GeneAmp 2400 PCR system using Vent DNA Polymerase (New England BioLabs). The PCR products were digested with the restriction endonucleases NdeI and SmaI, then ligated into pTBY2, which had been previously digested with the same enzymes. The ligation mixture was electroporated into E. coli-competent cells (XL-1 Blue, Stratagene). Cells containing plasmid were selected on LB agar (EM Science) containing ampicillin at a concentration of 100 g/ml. Purified plasmid was subjected to restriction and sequence analysis. The resulting construct, pTBY2-Rv1086, was digested with the restriction endonucleases, NdeI and PstI (Life Technologies). The ϳ2200-bp fragment contained the entire coding sequence for the fusion/insert. This was ligated into pVV16 (a gift from Dr. Varalakshmi Vissa, Colorado State University) and the resulting construct, named pIMP-Rv1086, was electroporated into competent M. smegmatis at 2.5 V, 800 ohms, 25 microfarads. Cells were allowed to recover in LB broth for 90 min and were plated on LB agar with kanamycin (20 g/ml). A single colony was chosen to start a liquid culture in LB broth with kanamycin (20 g/ml). The recombinant M. smegmatis strain containing vector pIMP-Rv1086 was grown to mid-log phase in LB broth with 20 g/ml kanamycin. This genetic construct expressed the M. tuberculosis open reading frame Rv1086 with a Cterminal fusion to a chitin binding domain. Approximately 6 g (wet weight) of cells was harvested by centrifugation, washed with a 0.9% saline solution, and centrifuged again. The resulting pellet was resuspended in cell lysis and column buffer containing 20 mM Tris-HCl (pH 8.0), 500 mM NaCl, 0.1 mM EDTA, and 0.1% Triton X-100. The cells were disrupted by probe sonication on ice with a Sanyo Soniprep 150 sonicator (10 cycles of 60 s on and 90 s off), and the suspension was centrifuged at 20,000 ϫ g for 20 min. The pellet was discarded, and the supernatant was loaded onto a 0.8-ml column of chitin beads (New England BioLabs) that had been equilibrated with the same buffer. The expressed protein was eluted from the column essentially as described by the manufacturer, except that the cleavage buffer contained 20 mM Tris-HCl (pH 8.0), 0.1% Triton, 150 mM NaCl, 30 mM dithiothreitol, and 10% glycerol. The eluted fractions were assayed for Z-FPP synthase activity, and the protein concentration was estimated using a BCA protein assay kit (Pierce).
Enzymatic Assays and Product Characterization-In vitro Z-FPP synthase assays, the enzymatic treatment of reaction products, and the analysis of products were done as described previously (9,27). Assays were conducted under conditions that were linear for time and protein concentration. In the metal ion dependence studies, the majority of the endogenous divalent cations were removed by incubating the enzyme preparation with Bio-Rex 70 200 -400 mesh (sodium form, Bio-Rad) on ice for 20 min. The enzymatic activity was reduced to a basal level, but it was not completely abolished, indicating that residual divalent cations were present. To determine which divalent cations supported activity, CaCl 2, MgCl 2 , MnCl 2 , or ZnCl 2 were added to the assay mixtures at the indicated concentrations. In a separate experiment, 10 mM EDTA was added. To study the pH dependence of Z-FPP synthase activity, a broad-range buffer comprising 250 mM Tris-HCl, 125 mM MES, and 125 mM acetic acid was used. The pH was adjusted with tetraethylammonium hydroxide.
Preparation of Sus scrofa E-FPP Synthase-Pig brain cytosol was prepared by homogenizing pig gray matter in 10 mM HEPES (pH 7.4) and 0.25 M sucrose using 18 passes of a Dounce homogenizer. The homogenates were centrifuged at 9000 ϫ g for 15 min at 4°C. The supernatant was decanted and centrifuged at 142,000 ϫ g for 1 h. The supernatant was decanted, divided into 1-ml aliquots, stored at Ϫ70°C, and used as an enzyme source for the ,E,E-FPP synthase assays. The protein concentration was estimated with a BCA protein assay kit (Pierce).
Other Materials-The sources of all materials have been described previously (9) with some exceptions. Citronellyl diphosphate, ,E,Efarnesyl diphosphate, and ,E-geranyl diphosphate were synthesized as described by Davisson et al. (28). Authentic prenols and prenyl phosphates of various chain lengths were purchased from the Institute of Biochemistry and Biophysics, Polish Academy of Sciences (Warsaw, Poland). ,Z-Neryl diphosphate was a gift from Drs. J. S. Rush and C. J. Waechter (University of Kentucky).

RESULTS
Purification of Z-FPP Synthase-Z-FPP synthase was expressed in M. smegmatis as a fusion protein with a C-terminal chitin binding domain. Affinity chromatography on a chitin column permitted a simple one-step protein purification (New England BioLabs). After the cells were harvested and disrupted, the 20,000 ϫ g supernatant was applied to the chitin column and washed. Subsequent elution with a dithiothreitolcontaining buffer cleaved the Z-FPP synthase from the chitin binding domain via an internal intein site. The column eluate was collected and subjected to SDS-polyacrylamide gel electrophoresis analysis (Fig. 1). The eluate was also assayed for [ 14 C]IPP incorporation into butanol-extractable material with five different allylic primers, DMAPP (C 5 ), ,E-GPP (C 10 ), ,Zneryl diphosphate (,Z-NPP; C 10 ), ,E,E-FPP (C 15 ), and ,E,E,E-GGPP (C 20 ). ,E-GPP and ,Z-NPP were the only functional substrates (data not shown). Assays with the other allylic primers did not produce any detectable radioactive product. [ 14 C]IPP incorporation into butanol-extractable material in assays primed with ,E-GPP or ,Z-NPP was linear for at least 40 min (data not shown). The product of the ,E-GPP assay was analyzed for chain length and stereochemistry by TLC (Fig. 2), confirming that the protein fraction shown in Fig.  1 (lane 4) synthesized ,E,Z-FPP. The product of the ,Z-NPP assay was also analyzed by TLC, demonstrating the enzyme's ability to synthesize a C 15 molecule, presumably ,Z,Z-FPP (data not shown).
Z-FPP Synthase Reaction Requirements-Z-FPP synthase was absolutely dependent on the presence of divalent cation for activity, and addition of 10 mM EDTA abolished the enzymatic activity (Table I). Z-FPP synthase activity was supported by the addition of MgCl 2 or MnCl 2 . The optimal concentration of MgCl 2 and MnCl 2 fell between 0.01 and 1.0 mM. Higher concentrations of MgCl 2 (between 1.0 and 5.0 mM) reduced (with respect to optimal concentration) the activity, whereas high concentrations of MnCl 2 (5.0 mM) strongly inhibited it. The enzyme was not stimulated by ZnCl 2 and CaCl 2 at the concentrations tested (0.01-5 mM). Z-FPP synthase was also tested for optimal activity over a range of pH (5.5-9.5 in 0.5 increments). The enzyme had a broad peak of activity over pH 7-8 (data not shown). When the rate of Z-FPP synthesis was measured in the presence of saturating ,E-GPP and varying concentrations of [ 14 C]IPP (Fig. 3A), a K m value of 124 M was calculated for the The latter two structures represent the possible products of Z-FPP synthase depending on which substrate (,E-geranyl diphosphate or ,Z-neryl diphosphate) is available to the enzyme. Isoprenyl diphosphate synthase activity was assayed in mixtures containing 50 mM MOPS (pH 7.9), 10 mM sodium orthovanadate, 0.05 mM MgCl 2 , 2.5 mM dithiothreitol, 0.3% Triton, X-100, 100 M geranyl diphosphate, 30 M [ 14 C]isopentenyl diphosphate, and 60 g of protein in a final volume of 50 l. Reactions were incubated for 10 min, stopped by the addition of water saturated with NaCl, and extracted with n-butanol saturated with water. The extracted isoprenyl diphosphates were dephosphorylated with potato acid phosphatase and spotted on reverse phase TLC plates to determine chain length. The plates were developed in methanol:acetone (8:2, v/v). Products labeled with [ 14 C]IPP were visualized by a Bioscan System 200 Imaging Scanner (Bioscan Inc.). Standard polyprenols were located with anisaldehyde spray reagent. Migration of the nonradioactive standard (F, ,E,Efarnesol) is indicated by an arrow at the top of B. The material corresponding to farnesol was scraped from the reverse phase TLC plate and extracted as described previously (9). The recovered farnesol was spotted onto a silica gel TLC plate and developed in toluene:ethyl acetate isoprenyl donor by nonlinear regression (Table II). When the concentration of IPP was fixed and the concentration of ,E-GPP or ,Z-NPP was varied, Michaelis constants of 38 and 16 M were calculated ( Fig. 3 and Table II).
Inhibition of Z-FPP Synthase-Compounds 1-4 (Scheme 1) and citronellyl diphosphate (Cit-PP, Fig. 4A) were tested for the ability to inhibit the synthesis of ,E,Z-FPP by the Z-FPP synthase from M. tuberculosis and the synthesis of ,E,E-FPP by the E-FPP synthase from pig brain (S. scrofa). Z-FPP synthase was inhibited by compound 4 and Cit-PP (Fig. 4B). The respective IC 50 values were estimated to be 300 and 350 M in the presence of 100 M GPP. E-FPP synthase was inhibited only by Cit-PP with an IC 50 of 125 M under similar assay conditions (Fig. 4C). Compounds 1, 2, and 3 had no effect on either of the FPP synthases at the concentrations tested (data not shown).

DISCUSSION
Little is known about Z-isoprenyl diphosphate synthases. Some early partial purification and characterization work has been published for bacterial (29,30) and eukaryotic (31)(32)(33)(34)(35) Z-isoprenyl diphosphate synthases, but it was not until Shimizu et al. (36) purified undecaprenyl diphosphate synthase from M. luteus in 1998 that a gene and protein sequence were correlated with the biochemical data. The complete sequencing of the M. tuberculosis H37Rv genome (and other genomes) has since revealed a wealth of information. Apfel et al. (37) was able to find 28 homologs of M. luteus undecaprenyl diphosphate synthase using comparative genome queries. M. tuberculosis is unique among organisms that have had their genomes sequenced in that there are two genes with homology to Z-isoprenyl diphosphate synthases (Rv1086 and Rv2361c), as opposed to the single Z-isoprenyl diphosphate synthase homolog typically found in the genomes of other organisms (37). We previously identified the function of the enzymes encoded by Rv1086 (Z-FPP synthase) and Rv2361c (Z-decaprenyl diphosphate synthase) (9). In this study, we report the purification, characterization, and inhibition of Rv1086.
All isoprenyl diphosphate synthases (E and Z) studied thus far have a strict requirement for a divalent cation. However, the type of cation and the concentration required for optimal activity is unique to each enzyme. X-ray crystallography of avian E-FPP synthase revealed that Mg 2ϩ cations were positioned within the active site of the enzyme (bound to aspartate residues) while complexed with the diphosphate moiety of the substrate (18). Undecaprenyl diphosphate synthase from Lactobacillus plantarum and E. coli were both shown to require Mg 2ϩ ions to bind IPP and a radiolabeled photolabile analog of ,E,E-FPP (38). The mycobacterial Z-FPP synthase is no exception, because it is stimulated by both Mg 2ϩ and Mn 2ϩ . It is likely that the divalent cation is required for substrate binding in a similar mechanism to that seen in E-isoprenyl diphosphate synthases.
Studies have shown that many E-FPP synthases can accept both DMAPP and ,E-GPP as allylic substrates (39 -43). When  The results shown in A are from assays containing 100 M ,E-geranyl diphosphate and increasing amounts [ 14 C]isopentenyl diphosphate (structure inset). The assay mixtures for B contained 30 M isopentenyl diphosphate and increasing amounts of ,E-geranyl diphosphate (structure inset). The assay mixtures for C contained 30 M isopentenyl diphosphate and increasing amounts of ,Z-neryl diphosphate (structure inset). Following incubation at 37°C for 10 min, the reactions were stopped with 1 ml of water saturated with NaCl, and extracted with 1 ml of butanol saturated with water. An aliquot was taken for liquid scintillation spectrometry.  (3,44). If the Pol-P from M. tuberculosis is structurally similar to the Pol-P from M. smegmatis (,E,polyZ-decaprenyl phosphate), ,E-GPP is the natural substrate of Z-FPP synthase. It is possible that ,Z-NPP is a precursor to an, as yet, undescribed isoprenoid molecule in M. tuberculosis. Nevertheless, the Z-FPP synthase is a monofunctional enzyme, and a separate enzyme must exist in M. tuberculosis that synthesizes either ,E-GPP or ,Z-NPP from DMAPP and IPP. A comparison of the mycobacterial Z-FPP synthase with the E-FPP synthase from S. scrofa brain was of interest, because the two enzymes catalyze the addition of IPP to ,E-GPP but release products with opposite stereochemistry at the newly formed double bond. Substrate analogs such as Cit-PP have been previously shown to inhibit pig liver E-FPP synthase (45,46). Cit-PP and other substrate analogs are thought to bind the enzyme active site through nonspecific lipophilic forces and a diphosphate binding force (45). The enzyme activity is inhibited, because Cit-PP lacks the allylic double bond in the 2 position (Fig. 4A) and is not able to undergo the typical electrophilic alkylation reaction. As shown in Fig. 4 (B and C), Cit-PP was effective at inhibiting both the E-and the Z-FPP synthases.
Compounds 1-4 (Scheme 1) were designed as analogs of ,E-GPP in which the biologically labile diphosphate moiety was replaced by moieties that can act as stable isosters that possess different conformational and stereoelectronic characteristics such as an unsubstituted phosphonoacetamidoxy (compound 1), an ␣-ethyl-substituted phosphonoacetamidoxy (compound 2), a phosphonoacetamido group (compound 3), or an ␣,␣Ј-diethyl-substituted [methylen(hydroxy)phosphoryl-]methanphosphonic moiety (compound 4). These isosteric moieties were chosen bearing in mind previous studies on isosters of the diphosphate group of prenyl diphosphates as squalene synthase (25), protein:farnesyl transferase (47), and protein: geranylgeranyl transferase inhibitors (48,49). Compounds 1-3 had no effect on either the Z-FPP synthase or the E-FPP synthase (data not shown). Compound 4 specifically inhibited the Z-FPP synthase (Fig. 4B) but had no effect on the activity of the E-FPP synthase (Fig. 4C). Despite its relatively high IC 50 , the ␣,␣Ј-diethyl-substituted [methylen(hydroxy)phosphoryl]methanphosphonic moiety of compound 4 seems to possess the correct conformational and stereoelectronic features to selectively interact with the active site of the mycobacterial Z-FPP synthase, suggesting that the study of the active site of this previously uncharacterized enzyme could lead to the development of novel chemotherapeutic agents for the treatment of multiple drug-resistant tuberculosis.