Inhibitors Incorporating Zinc-Binding Groups Target the GlcNAc-PI de-N-acetylase in Trypanosoma brucei, the Causative Agent of African Sleeping Sickness

Disruption of glycosylphosphatidylinositol biosynthesis is genetically and chemically validated as a drug target against the protozoan parasite Trypanosoma brucei, the causative agent of African sleeping sickness. The N-acetylglucosamine-phosphatidylinositol de-N-acetylase (deNAc) is a zinc metalloenzyme responsible for the second step of glycosylphosphatidylinositol biosynthesis. We recently reported the synthesis of eight deoxy-2-C-branched monosaccharides containing carboxylic acid, hydroxamic acid, or N-hydroxyurea substituents at the C2 position that may act as zinc-binding groups. Here, we describe the synthesis of a glucocyclitol-phospholipid incorporating a hydroxamic acid moiety and report the biochemical evaluation of the monosaccharides and the glucocyclitol-phospholipid as inhibitors of the trypanosome deNAc in the cell-free system and against recombinant enzyme. Monosaccharides with carboxylic acid or hydroxamic acid substituents were found to be the inhibitors of the trypanosome deNAc with IC50 values 0.1–1.5 mm, and the glucocyclitol-phospholipid was found to be a dual inhibitor of the deNAc and the α1-4-mannose transferase with an apparent IC50 = 19 ± 0.5 μm.

ognition than those of the mammalian pathway, enabling substratebased species-specific inhibitors to be designed (16,18).
The T. brucei GlcNAc-PI de-N-acetylase (deNAc) has been genetically validated as a drug target through the generation of a conditional null mutant (3) and has been shown to be a zinc metalloenzyme (22). Owing to their role in the progression of various human diseases, zinc metalloenzymes have gained much interest as potential drug targets (23,24). Synthetic inhibitors of zinc metalloenzymes typically consist of a backbone and a zinc-binding group (ZBG). The backbone is typically a drug-like structure that interacts with the protein through non-covalent interactions and contributes to both the affinity and selectivity of the inhibitor for its target. The ZBG coordinates to the zinc divalent cation and primarily contributes to the binding affinity of the inhibitor-metalloenzyme complex.
We have postulated that ZBGs could act as inhibitors of the T. brucei GlcNAc-PI de-N-acetylase (22). As part of our efforts to test this hypothesis, we have previously reported the synthesis of a GlcNAc-PI analog incorporating an N-hydroxyurea zinc-binding moiety (1-D-6-O-[2-(N-hydroxyaminocarbonyl)amino-2-deoxy-a-D-glucopyranosyl]myo-inositol 1-(n-octadecyl phosphate) (25). Unfortunately, this compound proved to be unstable under the conditions employed in the activity assay and was therefore judged to be unsuitable for further study. We recently reported the synthesis of eight deoxy-2-Cbranched monosaccharides incorporating ZBGs, Figure 2 (26). Here, we describe the synthesis of a glucocyclitol-phospholipid incorporating a ZBG and report the biological evaluation of the monosaccharides and the glucocyclitol-phospholipid as inhibitors of the trypanosome GlcNAc-PI de-N-acetylase.

Materials and Methods
General methods 1 H, 13 C, 31 P NMR spectra were recorded on a Bruker AVANCE spectrometer using deuteriochloroform as a solvent and tetramethylsilane as the internal standard, unless otherwise indicated. All coupling constants (J) are given in Hz. High-resolution electrospray ionization mass spectra (HRESIMS) and liquid chromatography mass spectra were recorded with a Bruker microTof spectrometer. Melting points were determined on a Reichert hot-plate apparatus and are uncorrected. Optical rotations were measured with a Perkin-Elmer 343 polarimeter. Thin-layer chromatography (TLC) was performed on Kieselgel 60 F 254 (Merck, Nottingham, UK) with various solvent systems as developers, followed by detection under UV light or by charring using either sulfuric acid ⁄ water ⁄ ethanol (15:85:5), phosphomolybdic acid, orcinol, or ninhydrin spray reagents. Flash column chromatography (FCC) was performed on Kieselgel 60 (0.040)0.063 mm) (Merck, Notttingham, UK). Radial-band chromatography (RBC) was performed using a Chromatotron (model 7924T; TC Research, Norwich, UK) with silica gel F 254 TLC standard grade as the adsorbent. All reactions were carried out in commercially available dry solvents, unless otherwise stated.

Expression of recombinant T. brucei GlcNAc-PI de-N-acetylase
Escherichia coli BL21 (DE3) transformed with pETB-TbGPI12 were grown in Luria-Bertani medium with 50 lg ⁄ mL carbenicillin at 37°C until A 600 0.5, induced with 250 lM isopropyl b-D-thiogalactoside, and cultured for a further 16 h at 21°C. Cells were harvested by centrifugation at 4500 · g for 20 min at 4°C, resuspended in 10 mL buffer A (50 mM Tris-HCl pH 8.0, 200 mM NaCl, 0.06% n-octyl-b-D-glucopyranoside, 10% glycerol v ⁄ v) per liter, and incubated with lysozyme (1 mg ⁄ mL) on ice for 20 min. Cells were lysed at 30 000 psi (OneShot; Constant Cell Disruption Systems, Daventry, UK), and the lysate clarified by centrifugation at 30 000 · g for 30 min at 4°C. The supernatant was filtered through a 0.4 lM membrane, mixed with glutathione-sepharose beads (GE Healthcare, Chafont St. Giles, UK) for 2 h at 4°C, washed with buffer A, eluted with 20 mM glutathione in buffer A. Pooled fractions were concentrated and washed in buffer A in a 10 000 molecular weight cut-off spin-concentration device (Satorius, Aubagne, France) repeatedly to remove glutathione. The concentration of GST-TbGPI12 was determined by absorbance at 280 nm using a calculated e = 7.464 · 10 4 ⁄ cm ⁄ M.

Tryptic peptide mass fingerprinting of GST-TbGPI12
The protein was reductively alkylated prior to SDS-PAGE and staining with Sypro Orange, the band excised and digested in 0.1% noctylglucoside and 20 mM NaHCO 3 with 12.5 lg ⁄ mL trypsin before analysis by MALDI-TOF MS and MS-MS. The protein was confirmed as T. brucei GlcNAc-PI de-N-acetylase with a Mascot score of 1162, with 87% sequence coverage.

ES-MS ⁄ MS de-N-acetylase assay
The recombinant protein GST-TbGPI12 (500 ng ⁄ lL) was incubated in incorporation buffer (50 lL) with or without inhibitor for 5 min at RT. The solution was transferred to tubes containing dry GlcN-IPC 18 3 (0.5 nmol), briefly vortexed, sonicated for 5 seconds, and incubated at 37°C for 30 min. The mixture was diluted with 5% propan-1-ol (1 mL), the glycolipids were bound to C8 resin (100 mg Isolute cartridge), washed (5% propan-1-ol, 5 mM NH 4 OAc), eluted (60% propan-1-ol, 5 mM NH 4  to calculate the percentage of substrate conversion to product in a given sample. Inhibitor IC 50 values were calculated using a fourparameter fit of eight-point potency curves derived from three independent experiments, and are quoted with standard deviation.

Results and Discussion
Synthesis of the glucocyclitol-phospholipid The chemical synthesis of the glucocyclitol-phospholipid 11, Figure 3, began from the carboxylic acid 5 (26). An anomeric mixture of 5 was coupled to O-benzylhydroxylamine hydrochloride in the presence of N-(3-dimethylaminopropyl)-N¢-ethylcarbodiimide hydrochloride (EDAC) to afford, after purification, the benzyloxyamide 6 which was treated with di-tert-butyl dicarbonate (Boc anhydride) and a catalytic amount of DMAP to give the N-Boc-protected hydroxamate 7. The introduction of the Boc protecting group was necessary in order to circumvent unwanted side reactions, which were apparent in previous studies, when activating the anomeric position for coupling to the cyclohexanediol moiety. The aforementioned coupling was achieved by converting the thioglycoside 7 to a glycosyl triflate through treatment with phenylsulfenyl triflate (PST) (31,32). PST was generated in situ by the addition of a freshly prepared solution of 0.4 M benzenesulfenyl chloride (29) to a solution of silver triflate (AgOTf) at )78°C. After the addition of activated powdered 4 molecular sieves, a solution of the thioglycoside 7 and the proton scavenger DTBMP were added at )78°C, followed by 1R,2R-trans-cyclohexanediol in THF ) CH 2 Cl 2 (1:1). Diligent RBC provided the a-and b-anomers 8 and 9, at 13% and 60% yields, respectively.
salt (27) with the pseudodisaccharide 9 furnished a mixture of diastereoisomeric phosphonic diesters that were converted into the phosphoric diester 10 on oxidation in situ with iodine in wet pyridine (33). An identical approach using the a-anomer 8 was unsuccessful owing to inseparable contaminants in the reaction mixture after coupling and in situ oxidation.
The generation of the final b-glucohydroxamic acid-cyclitol-phospholipid analog 11 was initially planned to proceed through ZØmplen de-O-acetylation, followed by the removal of the Boc protecting group under acidic conditions and, finally, catalytic hydrogenolysis to remove the benzyl (Bn)-protecting group. De-O-acetylation was carried out using sodium methoxide in methanol and when TLC indicated the complete disappearance of 10, then the resulting reaction mixture was neutralized with Amberlite IR 120 (H + ) ionexchange resin, filtered, evaporated to dryness under reduced pressure, and then subjected to column chromatography. Surprisingly, the major fraction isolated, from the purification, corresponded to the completely deprotected analog 11: triethylammonium 1R,2R-1-O-[2-C-(carboxymethyl N-hydroxyamide)-2-deoxy-b-D-glucopyranosyl]cyclohexanediol 2-(n-octadecyl phosphate). Thus, de-O-acetylation, Boc removal, and benzyl deprotection occurred in a single deprotection step. The 13 C and 1 H NMR data indicated the loss of all the protecting groups, which was further confirmed by the mass spectral data. Under the conditions used, de-O-acetylation is indisputable, and although it is known from the literature that acid treatment is the common deprotection method used for the Boc group (34), it has also been documented that this group can be removed under basic conditions (35,36). The removal of the benzylprotecting group of N-OBn, on the other hand, is not so obvious. Possibly, after the neutralization of the reaction mixture with Amberlite IR 120 (H + ) resin followed by its evaporation to dryness, the conditions became sufficiently acidic to lead to the removal of the benzyl-protecting group. This statement is pure conjecture, and the N-OBn fi N-OH reaction pathway may never be understood under the conditions described.

Evaluation of inhibitors in the trypanosome cell-free system
The ability of the compounds 11-19 to inhibit the deNAc was initially assessed in vitro using the T. brucei cell-free system (cfs), i.e., washed trypanosome membranes that are competent in GPI biosynthesis. Because de-N-acetylation of GlcNAc-PI must precede the addition of the three mannose residues (13), the activity of the deNAc can be indirectly monitored by measuring production of mannosylated GPI biosynthetic intermediates. In the assay, the cfs is primed with GDP-[ 3 H]Man and synthetic GlcNAc-PI 3 with and without inhibitor, and the radiolabelled mannosylated products separated by hptlc, quantified radiometrically, and visualized by fluorography. The compounds 11-19 were tested at an initial concentration of 10 mM in the cell-free system (Figure 4). Compounds 13, 15, 17, and 19 produced <10% inhibition of the formation of radiolabelled mannosylated products compared to the DMSO control, while 11,12,15,16, and 18 all produced >80% inhibition. The potency of the latter compounds was then determined using eight-point potency curves in triplicate (Table 1). Notably, the a ⁄ b anomers 17 and 18 gave >100-fold difference in potency. These compounds contain a thiophenyl group, predicted to adopt an axial position in the a-anomer 17, and an equatorial position in the b-anomer 18, which could potentially lead to differences in the abilities of the two anomers to fit into the active site of the enzyme. This result is surprising, given that it has previously been observed that the deNAc is able to de-Nacetylate both the natural substrate GlcNAc-a-PI 1 and the unnatural GlcNAc-b-PI (17).
The indirect cfs assay is unable to distinguish between the inhibition of the deNAc and inhibition of the first mannosyltransferase (MT1), because either will lead to an overall reduction in mannosylated GPI species. Furthermore, to account for the observation that priming the cfs with GlcNAc-PI is significantly more efficient than priming with GlcN-PI, it has been postulated that substrate channeling occurs between the deNAc and MT1 (14,17). Given the structural similarity of the deNAc substrate GlcNAc-PI 1 and the MT1 substrate GlcN-PI 2, the analogs 11-19 may also interact with MT1. We assessed the specificity of the two most potent inhibitors 11 and 18 by measuring their ability to inhibit the cfs primed with GDP-[ 3 H]Man and either the deNAc substrate GlcNAc-PI 1 or the MT1 substrate GlcN-PI 2. Compound 18 inhibited the formation of mannosylated products when the cfs was primed with GlcNAc-PI 1, but not when primed with GlcN-PI 2 ( Figure 5A), suggesting that it inhibits only the deNAc. Compound 11 inhibited the formation of mannosylated products when the cfs was primed with either Glc- NAc-PI 1 or GlcN-PI 2 ( Figure 5A), suggesting either that it inhibits both the deNAc and MT1, or that it inhibits MT1 only. The apparent inhibition of MT1 by the glucocyclitol-phospholipid 11 may be due to substrate channeling between the deNAc and MT1, such that inhibition of the deNAc is able to prevent GlcN-PI from accessing the MT1 active site. However, the present data do not rule out direct inhibition of MT1 only.
Evaluation of inhibitors against recombinant T. brucei GPI de-N-acetylase We have developed a mass spectrometry-based assay to measure the activity of a recombinant truncated rat deNAc construct, where the first 23 residues corresponding to the transmembrane region are replaced with 6 residues from an orthologous E. coli protein (22). Cloning and expression of the equivalent T. brucei deNAc construct failed to produce a significant yield of soluble protein.
Instead, an alternative construct containing a GST tag (TbGPI12-GST) afforded an improved yield of soluble protein, with protein identity confirmed by tryptic mass fingerprinting (Mascot score 1162, 87% coverage). The activity of TbGPI12-GST was confirmed in the electrospray tandem mass spectrometry (ES-MS ⁄ MS) assay using the synthetic GlcNAc-PI analog GlcNAc-IPC 18 3, where the diacyl glycerol portion of PI is replaced by a C 18 alkyl chain without affecting enzyme recognition (19). The ability of 11 and 18 to inhibit TbGPI12-GST was assessed in the ES-MS ⁄ MS assay, and both were found to inhibit with an IC 50 = 600 € 300 lM ( Figure 5B) and 980 € 220 lM (data not shown), respectively. Both compounds show reduced potency against the soluble truncated protein in the ES-MS ⁄ MS assay compared to their potency against the intact protein in the indirect cfs assay. This may be due to the loss of interactions with the truncated portion of the protein and ⁄ or ER membrane, or  in the case of 11, the lack of the postulated substrate channeling effect.

Conclusions and Future Directions
Disruption of GPI biosynthesis has been genetically (2-4) and chemically (5) validated as a drug target against T. brucei, the causative agent of African sleeping sickness in humans and the related disease Nagana in cattle. African sleeping sickness is invariably fatal if untreated, and there is an urgent need for new therapeutic agents that is not being met by the pharmaceutical industry. We have previously shown that the GPI biosynthetic enzyme GlcNAc-PI de-N-acetylase is a zinc metalloenzyme and postulated that ZBGs could act as inhibitors (22). As part of our efforts to develop drugs that target GPI biosynthesis in T. brucei, we have synthesized small molecules to probe the mechanism of the GlcNAc-PI de-N-acetylase (25,26). Here, we report that small molecules incorporating carboxylic acid or hydroxamic acid can inhibit the T. brucei GlcNAc-PI de-N-acetylase, confirming our hypothesis that ZBGs can be used to target the enzyme. However, the current compounds are neither sufficiently potent nor drug-like to be useful therapeutics. Future synthetic efforts will be directed toward developing a more druglike inhibitor backbone, to increase potency, and to introduce species selectivity.