Improved synthesis of phytanyl α - D -cellobiosyldiphosphate as substrate for α - D -mannosyltransferase

Polyisoprenyl-pyrophosphate-linked cellobiose is the natural acceptor of the α -1,3-mannosyltransferase AceA from Acetobacter xylinum, which transfers mannose from GDP-mannose during the assembly of the heptasaccharide repeat unit of the exopolysaccharide acetan. Phytanyl α - D -cellobiosyldiphosphate 4 has been previously synthesized as an analogue acceptor by condensation of hepta-O -acetyl- α - D -cellobiosylphosphate 1 with phytanylphosphate 2 , but the procedure was briefly described. We report here a modified detailed synthesis of 4 . The complete NMR characterization of 4 is provided and also a selection of NMR signals of all the intermediate compounds which facilitate monitoring the synthesis.


Introduction
Glycosyltransferases are a group of enzymes catalyzing the most common and important reactions in nature.Yet, they are a group of enzymes that are relatively poorly understood.In bacteria, most glycosyltransferases are involved in the synthesis of the essential components of the cell envelope, e.g., glycolipid intermediates for lipo-oligosaccharides, peptidoglycan, and polysaccharides 1,2 and may be rational targets for drug development against bacterial pathogens. 3lycosyltransferases are classified as inverting or retaining enzymes according to the anomeric configuration of the sugar residue transferred.So far, two structural superfamilies have been described for nucleotide-diphospho-sugar-dependent glycosyltransferases, called GT-A and GT-B.These topologies are variations of Rossman-like domain. 4The GT-A members display a core structure of a 7-8 β-strands with a DXD motif, where D is aspartic acid and X is any amino acid.This acidic motif coordinates the ribose and divalent cation in the catalytic center.In contrast GT-B proteins do not bind metals and are composed of two well defined Rossmann domains with a deep cleft between them, in which binding of substrates and catalytic activity occur. 5espite the relative low structural variety, reproducible predictions of the reaction catalyzed by a proposed glycosyltransferase cannot yet be made.Currently, there are 92 glycosyltransferase families in the Carbohydrate-Active Enzyme (CAZy) database. 6,7However, in December 2007, less than 10% of all proteins in CAZy were enzymatically characterized. 7his is in part because most of these enzymes are integral membrane proteins, whose biochemical characterization is a difficult matter due to protein purification and stabilization, and in part because natural substrates for many of the glycosyltransferases are very difficult to produce and purify.These difficulties have slowed down the biochemical characterization of these enzymes.In addition, for those glycosyltransferases with solved x-ray crystal structure, the unavailability of substrates have delayed the identification of key amino acids involved in binding and catalytic mechanism.
2][13] Although the allylic nature of the lipid portion of the natural acceptor has been confirmed, 14 the complete structure is not known yet.Lellouch et  al., synthesized phytanyl α-D-cellobiosyl diphosphate (Glc2-PP-phytanyl, 4) as an analogue acceptor for the reaction catalyzed by AceA. 15The lipid portion of this synthetic substrate mimics the polyisoprene natural structure and its branched carbon backbone, but is relatively short (C20) and saturated.The shorter lipid chain makes the handling of the substrate easier and the saturated nature stabilizes it, as allylic pyrophosphates are unstable.They showed that AceA was able to catalyze the transfer of [ 14 C]Man from GDP-[ 14 C]Man to the synthetic acceptor. 15The synthesis of phytanyl α-D-cellobiosyldiphosphate (Glc2-PP-phytanyl, 4), by condensation of hepta-O-acetylα-D-cellobiosylphosphate 1 with phytanylphosphate 2 was briefly described (Scheme 1) but much of the NMR data of the compounds were not provided.We encountered several problems associated with scale-up and reproducibility and therefore we introduced some modifications on the synthesis previously reported.We describe here in detail the preparation, purification and characterization of substrate 4 and intermediate products.Complete NMR characterization is now reported and diagnostic signals were tabulated to assist monitoring the synthesis.The synthesis of 4 began with the phosphorylation of a derivative of cellobiose.Selective anomeric deacylation of octa-O-acetyl-α-D-cellobiose 5 was performed with acetic acid/ethylenediamine 16 to afford 6 (Scheme 2).This one pot procedure is easier and high yielding with respect to the procedure previously used, involving silver carbonate catalyzed hydrolysis of the acetylated glycosyl halide. 17The phosphate was introduced by treatment of 6 with diphenylchlorophosphate and 4-N,N-dimethylaminopyridine (DMAP) under the conditions established to favor α-anomers. 17The 1 H NMR spectrum of 7 showed the H-1 signal as a doublet of doublets at  5.99 with J1,2 3.5 Hz indicating the α configuration, and JH,P 6.5 Hz, confirming that the phosphorylation occurred.The 13 C NMR spectrum showed the resonance at 100.7 ppm for C-1' and doublets at  94.9 (5.4 Hz) and 69.7 (3.8 Hz) for C-1 and C-2, respectively, also indicative of the coupling with phosphorous (Table 1).These values were similar to those observed for the monosaccharidic analogue per-O-acetyl-α-D-glucosyldiphenylphosphate. 17leavage of the phenyl groups from the anomeric phosphotriester of 7 was achieved by catalytic hydrogenation over PtO2.Compound 1 was isolated as the triethylammonium salt and used directly without further purification.Also for this compound the 13 C NMR spectrum showed signals corresponding to C-1 and 2 as doublets due to JC,P (Table 1), confirming the presence of the phosphate group.The α-anomeric configuration was confirmed by the J1,2 value (3.3 Hz), indicative of a 1,2-cis relationship.The lipid phosphate 2 was previously prepared using the corresponding phosphoramidite chemistry, which involves three steps. 18It was stored as the corresponding t-butyl ester and deprotected immediately before the coupling with 1. 15 We found several problems using this strategy, so an alternative approach for the installation of this group was investigated.In view of the success in the phosphorylation of cellobiose with diphenylchlorophosphate, we tested the same method with phytanol and found it suitable.Thus, phytanol 8 was treated with diphenylchlorophosphate and DMAP under the same conditions described for the preparation of 7 (Scheme 3).Phosphate 9 was efficiently obtained in 85% yield and could be stored for several months at -20 C.Hydrogenation of 9 over PtO2 provided phosphate 2. The characteristic coupling constant 13 C-31 P observed for C-1 at the 13 C NMR spectra of 9 (JC,P 6.7, Table 1) and 2 (JC,P 5.4, Table 1) confirmed that the phosphatation was efficient.The signal corresponding to geminal H-1 and H-1' of compound 9 was a complex multiplet, while the analogue signal in compound 2 was a well resolved doublet of doublets (Table 1).With triethylammonium salts of 1 and 2 in hand the coupling was effected by activation with 1,1'-carbonyldiimidazole (Scheme 1).For the purification of condensed compound 3 HPLC was previously used. 15As this is not an amenable procedure to preparative scale-up, we introduced a purification step by silica gel column chromatography instead, which afforded pure compound 3 in 79% yield, considerably higher than the reported procedure.The 1 H NMR spectrum of 3 was in agreement with the previously reported, but we now report the 13 C NMR data which helps the confirmation of the coupling.Thus, doublets at  92.4 (JC,P 8.0 Hz, C-1) and 70.1 (JC,P 8.0 Hz, C-2) are indicative that the cellobiose residue is phosphated and the doublet at 64.7 ppm (JC,P 6.7 Hz) corresponds to phosphorylated phytanyl C-1.
By O-deacylation of 3 with NaMeO/MeOH in CH2Cl2 solution, compound 4 was afforded.A purification step by anionic interchange chromatography was introduced.Thus, the crude Odeacetylated compound was fractionated on a DEAE-cellulose column in the acetate form by an ammonium formate gradient.Compound 4 was isolated as the triethylammonium salt in an excellent yield of 80%.

Conclusions
The synthesis of Glc2-PP-phytanyl 4 was carried out, providing alternatives with respect to the synthesis and purification previously reported.The procedure introduced for the preparation of 2, by phosphorylation of phytanol with diphenyl chlorophosphate and subsequent hydrogenation is simple and reproducible.The phosphorylated compounds were characterized by 1 H, 13 C and 31 P NMR spectra and the diagnostic signals are depicted in Table 1, which allows monitoring the sequence reaction.Scale-up the synthesis of 4 is thus easier and this fact could remove a significant obstacle in the study of the AceA.

Experimental Section
General.Analytical thin layer chromatography (TLC) was performed on silica gel 60 F254 (Merck) aluminum supported plates (layer thickness 0.2 mm) with solvent systems given in the text.Visualization of the spots was effected by exposure to UV light and charring with a solution of 10% (v/v) sulfuric acid in EtOH, containing 0.5% p-anisaldehyde.Column chromatography were carried out with silica gel 60 (230-400 mesh, Merck).The 31 P NMR spectra were recorded with a Bruker AMX 500 spectrometer, and the 1 H and 13 C NMR spectra with such equipment or with a Bruker AC 200, as indicated in each case.The chemical shift reference for 31 P was that of external phosphoric acid (85%) in D2O set at 0.0 ppm.All the proton and carbon resonances for compound 4 were performed at 25 °C using a Bruker Avance 400 spectrometer and were assigned by 2D spectra: COSY, TOCSY (mixing time 80 ms) to identify the spin systems, 19 1 H-13 C HSQC, 20 and 1 H- 13 C HSQC-TOCSY (with a mixing time of 60 and 80 ms) 21 to assist with cross-peaks assignment.High resolution mass spectra (HRMS ESI) were recorded in a Bruker micrOTOF-Q II and a VG ZAB-HF mass spectrometers.