Galactose-derived phosphonate analogues as potential inhibitors of phosphatidylinositol biosynthesis in mycobacteria

Article abstract of DOI:10.1039/b616450a

Galactose-based phosphonate analogues of myo-inositol-1-phosphate and phosphatidylinositol have been synthesized from methyl β-d- galactopyranoside. Michaelis-Arbuzov reaction of isopropyl diphenyl phosphite or triisopropyl phosphite with a 6-iodo-3,4-iso

Full text of DOI:10.1039/b616450a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Galactose-derived phosphonate analogues as potential inhibitors of
phosphatidylinositol biosynthesis in mycobacteria†
Zoran Dinev,a,b Carlie T. Gannon,a,b Caroline Egan,a,b Jacinta A. Watt,a,b Malcolm J. McConvilleb,c and
Spencer J. Williams*a,b
Received 10th November 2006, Accepted 5th January 2007
First published as an Advance Article on the web 30th January 2007
DOI: 10.1039/b616450a
Galactose-based phosphonate analogues of myo-inositol-1-phosphate and phosphatidylinositol have
been synthesized from methyl b-D-galactopyranoside. Michaelis–Arbuzov reaction of isopropyl
diphenyl phosphite or triisopropyl phosphite with a 6-iodo-3,4-isopropylidene galactoside afforded the
corresponding phosphonates. Deprotection of the diphenyl phosphonate afforded methyl
b-D-galactoside 6-phosphonate, an analogue of myo-inositol-1-phosphate. The diisopropyl esters of the
diisopropyl phosphonate were selectively deprotected and the corresponding anion was coupled with
1,2-dipalmitoyl-sn-glycerol using dicyclohexylcarbodiimide. Deprotection afforded a methyl
b-D-galactoside-derived analogue of phosphatidylinositol. The galactose-derived analogues of
phosphatidylinositol and myo-inositol-1-phosphate were not substrates for mycobacterial
mannosyltransferases (at concentrations up to 1 mM) involved in phosphatidylinositol mannoside
biosynthesis in a cell-free extract of Mycobacterium smegmatis. The galactose-derived phosphonate
analogue of phosphatidylinositol was shown to be an inhibitor at 0.01 mM of PimA
mannosyltransferase involved in the synthesis of phosphatidylinositol mannoside from
phosphatidylinositol, and a weaker inhibitor of the next mannosyltransferase(s), which catalyzes the
mannosylation of phosphatidylinositol mannoside.
The successful exploitation of PimA as a drug target will require
the development of alternative substrates with which simpler
assays for PimA can be established, and the identification of in-
Introduction
myo-Inositol-1-phosphate (IP) and phosphatidylinositol (PI) are
important metabolites in their own right and are elaborated to
hibitors as lead compounds for further development. We therefore
a variety of products that play crucial signaling and structural
became interested in the synthesis of a phosphonate analogue of
roles in biology.1,2 In mycobacteria, a group of pathogenic
PI. Phosphonates are well-established as surrogates for phosphate
groups, yet are hydrolytically stable.9,10 To this end we noted the
surprising stereochemical congruence of D-galactose and myo-
inositol (Fig. 1).‡ The 6-phosphonate derivative 1 of methyl b-
D-galactoside is an analogue of IP, with the b-configuration of the
glycoside ensuring that the stereochemistry of the anomeric centre
matches that of the inositol. Elaboration of 1 to the corresponding
diacylglycerol derivative 2 would give an analogue of PI. Both 1
and 2 could be evaluated as potential substrates and inhibitors of
PIM biosynthesis.
bacteria that includes Mycobacterium tuberculosis, myo-inositol
is a precursor for PI, which leads to the structurally complex
phosphatidylinositol mannosides (PIMs) and lipoarabinomannan
(LAM), and other related molecules, all of which share a common
IP core (Fig. 1).3–5 The diacylglycerol lipid anchor of the PIMs
and LAM allows these conjugates to anchor into membranes
and their biosynthesis is compartmentalized within the plasma
membrane.6 Both PIMs and LAM are potent immunogens and
modulators of the interactions of the tubercle bacillus and the
host immune system.7 The early steps in the biosynthesis of PIMs
and LAM have now been described in considerable detail.7 In
particular, one of the first steps in PIM and LAM biosynthesis,
the transfer of a mannosyl group from GDP-mannose to the 2-
position of PI to give PIM1 catalyzed by the mannosyltransferase
PimA, has been shown to be essential for mycobacterial survival
(Fig. 1).8 Inhibition of this enzyme therefore could provide new
anti-tuberculosis drugs.
Results and discussion
Our synthetic route to the galactose phosphonate 1 envisioned
the introduction of a phosphonate moiety using either Michaelis–
Arbuzov or Michaelis–Becker reactions of a protected galacto-
side equipped with a leaving group at C6. Substitution at C6
of galactose derivatives is notoriously difficult owing to steric
obstruction of nucleophiles by the axial C4 substituent.11 The
poor reactivity of galactos-6-yl electrophiles can allow a variety of
side reactions to intervene including: intramolecular cyclizations
to 3,6-anhydro sugars, elimination reactions and, in the case
aSchool of Chemistry, University of Melbourne, Parkville, Australia 3010.
E-mail: sjwill@unimelb.edu.au; Fax: +61 3 9347 8124; Tel: +61 3 8344 2422
bBio21 Molecular Science and Biotechnology Institute, University of Mel-
bourne, Parkville, Australia
cDepartment of Biochemistry and Molecular Biology, University of Mel-
bourne, Parkville, Australia
‡ Vasella and co-workers have applied the stereochemical similitude of D-
galactose and myo-inositol to the synthesis of a phosphonate analogue of
myo-inositol 1,4,5-triphosphate.38
† Electronic supplementary information (ESI) available: NMR spectra of
1–3, 7, 8, 11–13; HRMS of 2. See DOI: 10.1039/b616450a
952 | Org. Biomol. Chem., 2007, 5, 952–959
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Fig. 1 Early steps in the biosynthesis of PIMs in mycobacteria. myo-Inositol is converted to phosphatidylinositol (PI), and then to phosphatidylinositol
mannoside (PIM1) and PIM2 by the GDP-mannose dependent mannosyltransferases PimA and an as-yet undefined mannosyltransferase(s), respectively.
R,Rꢀ = various acyl groups derived from palmitic, stearic, tuberculostearic and 12-methoxypropanoyloxystearic acids. 1 and 2 are D-galactose-derived
phosphonate analogues of myo-inositol-1-phosphate and PI, respectively.
of the Michaelis–Arbuzov reaction, rearrangement of trialkyl
phosphites to dialkyl alkylphosphonates. Indeed, preliminary
studies aimed at performing a Michaelis–Becker reaction of a
benzylated galactoside bearing an iodide or mesylate at C6, with
the anion of dibenzyl phosphonate, were not encouraging. Palcic
and co-workers have reported difficulties in the preparation of
iodide 3 from 4,12 observing only cyclization to the 3,6-anhydro
sugar 5.13 In our hands we found that treatment of the alcohol
4 with triphenylphosphine, iodine and imidazole in toluene at
in a phosphonate substitution reaction, whilst at the same time
mitigating side reactions.
A 3,4-O-isopropylidene protecting group has been widely used
in substitution reactions at the 6-position of galactose.11,14 The
cyclic acetal restricts ring inversion, preventing intramolecular
cyclization, and distorts the conformation of the carbohydrate
ring altering the position of the C4-axial substituent and thereby
opens up C6 to nucleophilic attack. However, even with an
3,4-O-isopropylidene group in place, Golding and co-workers
have reported that the Michaelis–Arbuzov reaction of 6-deoxy-
6-iodo-1,2:3,4-di-O-isopropylidene-a-D-galactose with trimethyl
phosphite occurs in poor yield, owing to the undesirable re-
arrangement of this reagent to dimethyl methylphosphonate.14
These authors have pioneered the use the modified Arbuzov
reagents, methyl or isopropyl diphenyl phosphite, to afford
improved yields of the diphenyl phosphonates. When methyl
diphenyl phosphite was used, the reaction proceeded best under
vacuum to remove the methyl iodide by-product, which is a
reactive electrophile that can consume the phosphite reagent.14
In seminal studies, Nesterov and Arbuzov have reported that
isopropyl diph◦enyl phosphite is unreactive towards isopropyl
iodide at 200 C,15 and Golding and co-workers have shown
this phosphite is a superior reagent for the Michaelis–Arbuzov
reaction with poorly electrophilic 6-iodo galactose derivatives.14
While Golding and co-workers have reported the synthesis
of isopropyl diphenyl phosphite from hexaethylphosphorous
triamide,14,16 we chose to use hexamethylphosphorous triamide
(HMPT) as a more economical starting material. Thus, phenol
and HMPT in dimethoxyethane at reflux afforded diphenyl N,N-
dimethylphosphoramidite, which was treated with isopropanol
and 1H-tetrazole in dimethoxyethane to afford isopropyl diphenyl
phosphite (Scheme 2a). While this method was effective for the
synthesis of moderate amounts of isopropyl diphenyl phosphite,
for larger scale preparations a simpler route was desirable that
avoided the need for large amounts of HMPT and 1H-tetrazole.
A direct one-pot procedure was developed by treating phosphorus
trichloride with triethylamine and isopropanol, and then phenol,
to afford isopropyl diphenyl phosphite, isolated by distillation in
◦
90 C afforded the iodide 3 in moderate yield (35%) (Scheme 1).
Also isolated was the same 3,6-anhydro sugar 5, arising from
intramolecular attack by the benzyl ether at O3, by way of a ring
inversion to the1C4 conformation encouraged bythe establishment
of a favourable anomeric effect in that conformer. Such an
occurrence forecast difficulties in the Michaelis–Becker reaction.
In the event, treatment of either the iodide 3 or methyl 2,3,4-tri-
O-benzyl-6-methylsulfonyl-b-D-galactoside13 with the sodium salt
of dibenzyl phosphorous acid in THF at reflux or with dibenzyl
phosphorous acid/caesium carbonate/Bu4NI at 120 ◦C in DMF
resulted in no reaction. Higher temperatures lead only to the 3,6-
anhydro sugar 5. With these dissatisfying results an alternative
approach was planned that required careful choice of protecting
groups to ensure sufficient reactivity of the sugar electrophile
Scheme 1 Reagents and conditions: (i) I2, imidazole, triphenylphosphine,
toluene, 90 ◦C, 35% of 3, 57% of 5.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 952–959 | 953
©

View Article Online
Scheme 2 Reagents and conditions: (i) phenol, 1,2-dimethoxyethane,
reflux, 95%; (ii) isopropanol, 1H-tetrazole, 1,2-dimethoxyethane, 75%; (iii)
PCl3, isopropanol, Et3N, Et2O, then phenol, 30%.
Scheme 4 Reagents and conditions: (i) 2-mercaptobenzothiazole, EtiPr2N,
DMF, 94%; (ii) octanol, DCC, pyridine; (iii) 1,2-dipalmitoyl-sn-glycerol,
DCC, pyridine.
sufficient purity (Scheme 2b). The yield of our one-step process
(30%) is lower than the two-step approach but, owing to its
simplicity and the inexpensive nature of the reagents, is superior
and can be easily scaled to 10–20 g quantities.
to effect efficient condensation with 1,2-dipalmitoyl-sn-glycerol
using any of a range of coupling reagents (BOP,24 PyBOP,25
HATU, DCC, EDC, triisopropylbenzenesulfonyl chloride26 or
trichloroacetonitrile27). This troublesome road block therefore
demanded an alternative approach. It was forecast that the
phosphonate 11 would act as a better electrophile than 9 in
condensations with the poorly nucleophilic 1,2-dipalmitoyl-sn-
glycerol. However, the selective debenzylation of the phosphonate
of 8 or, alternatively, the selective dephenylation of the diphenyl
phosphonates of 8, in the presence of a 2-O-benzyl ether, presented
a particular challenge.28 Therefore, we revisited our approach to
11 by way of the diisopropyl phosphonate 12, anticipating the
chemoselective cleavage of the isopropyl phosphonate esters using
TMSBr (Scheme 5).29–31 Treatment of the iodide 6 with triisopropyl
phosphite at 180 ◦C cleanly afforded the diisopropyl phosphonate
12 in excellent yield (87%). Cleavage of the isopropyl groups of 12
in the presence of the isopropylidene moiety proceeded without
incident by treatment of 12 with TMSBr and Et3N, affording the
anion 11. Confirming our predictions, 11 proved a much better
electrophile than the monoester 9: treatment of 11 with 1,2-
dipalmitoyl-sn-glycerol and DCC in pyridine gave the protected
galactose phosphonate 13. Finally, removal of the isopropylidene
group (TFA–H2O) and hydrogenolysis using Pd(OH)2 proceeded
without incident to afford the phosphonate 2 in 75% yield.
With ample supplies of isopropyl diphenyl phosphite in hand,
we next turned to the Michaelis–Arbuzov reaction, with the
immediate aim of preparing the IP analogue 1 (Scheme 3).
The iodide 6 was synthesized from methyl b-D-galactopyranoside
according to the literature.17–19 Treatment of 6 with 5 equivalents
◦
of isopropyl diphenyl phosphite at 180 C afforded the diphenyl
phosphonate 7 in excellent yield (81%). Successful formation of
the phosphonate was revealed by 31P NMR, with a characteristic
signal at dP 20.63 ppm. Deprotection of 7 to afford the analogue
1 was achieved in three steps. Base-catalyzed transesterification
of the phenyl ester of 7 with benzyl alcohol, using the method of
Billington and co-workers,20 afforded the dibenzyl phosphonate
8. Hydrolysis of the isopropylidene acetal of 8 with TFA–water,
and then debenzylation by hydrogenolysis (Pd/C), afforded the
galactose phosphonate IP analogue 1, isolated as the ammonium
salt after ion exchange chromatography (51%).
i
Scheme 3 Reagents and conditions: (i) see ref. 17–19; (ii) PrOP(OPh)2,
180 ◦C, 81%; (iii) BnOH, NaH, THF, 74%; (iv) a) TFA–H2O (9 : 1), b)
Pd/C, THF, H2O, 51%.
We next targeted the PI analogue 2. Our initial synthetic plan
considered a dehydrative coupling of a monobenzyl phosphonate
such as 9 with 1,2-dipalmitoyl-sn-glycerol (Scheme 4). While
dipalmitoyl PI has not been isolated from mycobacteria, authentic
mycobacterial PIs and their derivatives are modified with long
chain acyl groups including palmityl, stearyl, tuberculostearyl and
12-methoxypropanoyloxystearyl groups, suggesting that the PI-
modifying enzyme PimA can tolerate unnatural combinations of
long chain acyl groups. In addition, PimA can act on PI from alter-
native sources such as soy-bean.21 Thus, the dibenzyl phosphonate
8 was monodebenzylated to afford 9 by prolonged treatment
with 2-mercaptobenzothiazole22 or, preferably, by treatment with
DABCO at reflux in toluene.23 While 9 could be condensed with
simple alcohols such as 1-octanol under the agency of BOP or
DCC to give 10, even after extensive investigation we were unable
Scheme 5 Reagents and conditions: (i) triisopropyl phosphite, 180 ◦C,
87%; (ii) TMSBr, Et3N, CH2Cl2, 92%; (iii) 1,2-dipalmitoyl-sn-glycerol,
DCC, pyridine, 59%; (iv) a) TFA–H2O (9 : 1), b) Pd(OH)2, MeOH, EtOAc,
75%.
IP analogue 1 and PI analogue 2 were assessed for their ability to
act as substrates and/or inhibitors in a Mycobacterium smegmatis
cell-free system obtained by nitrogen cavitation of log phase cells.
This cell-free system is capable of the synthesis of PIMs, which
can be assessed using GDP-[3H]mannose as a mannosyl donor,
partitioning of lipophilic products into 1-butanol, and separation
of radiolabelled products by HPTLC (Fig. 2).6 The products seen
are radiolabelled by [3H]mannosyl-transfer in the cell-free system,
954 | Org. Biomol. Chem., 2007, 5, 952–959
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
phatidylinositol. Notable steps include the introduction of the
phosphonate moiety using modified Michaelis–Arbuzov reagents
and a 6-iodo-3,4-isopropylidene-protected galactoside; the selec-
tive deprotection of a diisopropyl phosphonate in the presence of
a 3,4-O-isopropylidene group; and the DCC-mediated conden-
sation of the phosphonate 11 with 1,2-dipalmitoyl-sn-glycerol.
The IP analogue 1 and the PI analogue 2 were investigated
as substrates and inhibitors of PIM mannosyltransferases in a
mycobacterial cell-free system. Neither compound acted as a
substrate, demonstrating that these analogues are not effective
bioisosteres for native PI, and that these structural modifications
are too dramatic for recognition by PimA. The IP analogue 1 was
shown to be an effective inhibitor of PimA, the mannosyltrans-
ferase that catalyzes the synthesis of PIM1 from PI, and a partial
inhibitor of the mannosyltransferase that converts PIM1 to PIM2.
The PI analogue 2 was shown to be a weak inhibitor of this second
mannosyltransferase. Further analysis of the inhibition of both
mannosyltransferases by 1 and 2 will require the development of
improved assays for these two enzymes.
Fig. 2 HPTLC analysis of IP analogue 1 and PI analogue 2 as sub-
strates/inhibitors of M. smegmatis mannosyltransferases. M. smegmatis
cell lysates were incubated with GDP-[3H]mannose in the absence of any
analogue (lane 0) or the presence of different concentrations of either
PI analogue 2 or IP analogue 1. S = authentic PIM standards derived
from in vivo [3H]mannose labelling; A = Ac2PIM1; B = PPM, Ac2PIM2,
AcPIM1; C = AcPIM2; D = PIM1; E = PIM2; F = AcPIM5, AcPIM5ꢀ;
G = AcPIM6.39 The abbreviations used are: AcxPIMy, PIM species with x
(1 or 2) fatty acyl chains, linked to either the core a-1,2-linked Man or the
myo-inositol head group, and y Man residues; PPM, polyprenol (C35/C50)
phosphomannose.
Experimental
Pyridine was distilled before use. Diethyl ether, 1,2-dimethoxy-
ethane and THF were dried over sodium before use. Reactions
were monitored using thin layer chromatography (TLC), per-
formed with Merck Silica Gel 60 F254. Detection was effected by
charring in a mixture of 5% sulfuric acid in methanol, vanillin stain
(6% vanillin, 1% H2SO4 in EtOH), 10% phosphomolybdic acid in
EtOH, and/or visualising with UV light. Flash chromatography
was performed according to the method of Still et al. using
Merck Silica Gel 60.34 Melting points were obtained using a
Reichert–Jung hotstage microscope (corrected). Optical rotations
were obtained using a JASCO DIP-1000 polarimeter (Melbourne,
Australia). [a]D values are given in 10−1 cm2 g−1. 1H, 13C and
31P NMR spectra were recorded using Unity Plus 400 or Inova
400 or 500 instruments (Melbourne, Australia). For 31P NMR
spectra, 85% orthophosphoric acid (31P NMR: d 0.00 ppm)
was used as a reference. Elemental analyses were conducted by
C.M.A.S. (Belmont, Victoria). Refractive indices were obtained
on a Bellingham & Stanley refractometer (Melbourne, Australia).
Low resolution mass spectrometry was performed by Sioe See
Volaric (Melbourne University, Australia). High resolution mass
spectrometry was performed by Mr Hadi Loie on a Finnigan
hybrid LTQ-FT mass spectrometer (Thermo Electron Corp.).
largely from an existing pool of the direct intermediates, rather
than by de novo synthesis from PI. In the absence of exogenously
added 1 or 2 (lane 0), a family of products are seen, which were
assigned as various PIMs and PPM based on previous studies (see
Fig. 2 legend).6 At concentrations up to 1 mM neither 1 nor 2
acted as a substrate, with no additional products observed even at
the highest concentration tested (1 mM).
The same analyses allowed assessment of the ability of 1 and 2
to inhibit the formation of each of the individual PIMs, as each
labelled compound is formed largely from their direct precursors.
Careful inspection of the resulting chromatogram reveals a clear
reduction in the labelling of PIM1 (band D) and PIM2 (band
E) in samples treated with IP analogue 1. That this result does
not arise from an indirect effect of the detergent-like nature of 2
was shown in control experiments where the detergent octyl a-D-
mannoside was included in the cell free assay and resulted in no
observable change in the appearance of the spectrum of products
formed even at 1 mM (see ESI†). The IP analogue 1 thus appears
to be an inhibitor of PimA, the first mannosyltransferase in the
pathway of PIM biosynthesis (Fig. 1). The decrease in labelling of
PIM2 could reflect reduced levels of de novo-synthesized PIM1,
or inhibition of the second mannosyltransferase in the PIM
biosynthesis pathway.32,33 This transfer, which may be catalyzed
by more than one mannosyltransferase, adds a mannosyl group
to the 6-position of myo-inositol in PIM1 and must therefore
recognize structural features of the IP core of PIM1, and the
IP analogue 1. In contrast, PI analogue 2 had less effect on
PIM1 biosynthesis, while partially inhibiting PIM2 biosynthesis
at concentrations up to 1 mM. Therefore it appears that PimA
distinguishes between endogenous PI and the PI analogue 2,
whereas the second mannosyltransferase does not.
Methyl 2,3,4-tri-O-benzyl-6-deoxy-6-iodo-b-D-galactoside 3 and
methyl 3,6-anhydro-2,4-di-O-benzyl-b-D-galactoside 5
A suspension of the alcohol 412 (233 mg, 501 lmol), imidazole
(93.8 mg, 1.38 mmol) and triphenylphosphine (214 mg, 817 lmol)
in toluene (5 mL) was azeotropically dried (Dean–Stark). The
reaction mixture was cooled to 90 ◦C, iodine (210 mg, 829 lmol)
was added and the mixture was stirred under N2. After 40 min,
TLC indicated the formation of two compounds of lower polarity
than the starting material. The reaction was quenched with MeOH
(5 mL) and then diluted to 100 mL with toluene, washed with water
(8 × 25 mL), Na2S2O3 (0.25 M, 2 × 10 mL), dried (MgSO4) and
evaporated to yield a dark orange oil. The crude oil was purified by
flash chromatography (10–20% EtOAc/pet. spirits) to yield first
Conclusions
In summary, we describe the synthesis of galactose-derived
phosphonate analogues of myo-inositol-1-phosphate and phos-
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 952–959 | 955
©

View Article Online
the iodide 3 as an orange syrup (100 mg, 35%). Next to elute was
the 3,6-anhydro derivative 5 as a yellow syrup (101 mg, 57%).
Data for 3: mp 60.5–62.0 ◦C (colourless needles from petroleum
spirit); [a]2D3 +3.46 (c 0.305, CHCl3) (Found: C, 58.58; H, 5.38.
C28H31O5I requires C, 58.55; H, 5.44); 1H NMR (400 MHz; CDCl3)
d 3.17 (1H, dd, J5,6 7.2, J6,6 9.8 Hz, H6), 3.29 (1H, dd, J5,6 6.8, J6,6
9.8 Hz, H6), 3.51 (1H, ddd, J4,5 0.8, J5,6 6.8, 7.2 Hz, H5), 3.52
(1H, dd, J2,3 9.6, J3,4 2.9 Hz, H3), 3.57 (3H, s, OCH3), 3.80 (1H,
dd, J1,2 7.7 Hz, H2), 4.01 (1H, dd, J3,4 2.9, J4,5 0.8 Hz, H4), 4.27
(1H, d, J1,2 7.7 Hz, H1), 4.75, 4.82, 4.91 (4H, 3 d, J 11.0, 11.8 Hz,
2 × PhCH2), 4.67, 5.04 (2H, 2 d, J 11.3 Hz, PhCH2), 7.23–7.42
(15H, m, Ph); 13C NMR (100 MHz; CDCl3) d 3.08 (C6), 57.11
(OCH3), 73.34, 73.86, 74.75, 75.02, 75.06, 79.04, 82.08 (C2,3,4,5,
3 × PhCH2), 104.67 (C1), 127.51–128.41, 138.15, 138.25, 138.61
(Ph).
mixture was stirred at room temperature for 1 h, then a solution
of recrystallised phenol (10.8 g, 115 mmol) in anhydrous ether
(50.0 mL) was added dropwise to the reaction mixture. The
resultant mixture was stirred at room temperature overnight, then
filtered (Celite) and concentrated to produce a yellow oil. The
crude material was filtered through a plug of silica and eluted with
petroleum spirits (5 × 100 mL). The eluant was concentrated and
purified by vacuum distillation (120–126 ◦C at 5 Torr) to yield the
phosphite as a colourless liquid (4.74 g, 30%).
Methyl 2-O-benzyl-6-deoxy-6-diphenoxyphosphinyl-3,4-O-
isopropylidene-b-D-galactoside 7
The iodide 617–19 (1.61 g, 3.71 mmol) and isopropyl diphenyl
phosphite (5.13 g, 18.6 mmol) was heated at 180 ◦C for 4 days.
Unreacted phosphite was recovered by diluting the crude reaction
mixture with acetonitrile (50 mL) and washing with petroleum
spirits (12 × 50 mL). The petroleum phase was back-extracted with
MeCN (2 × 50 mL), and the combined petroleum extracts were
concentrated in vacuo. The crude phosphite (3.69 g) was distilled
for reuse. The MeCN extracts were evaporated to dryness to afford
a colourless solid. The solid was purified by flash chromatography
(20–60% EtOAc–pet. spirits) to afford the diphenyl phosphonate
Data obtained for 5 were in accordance with the literature.35
N,Nꢀ-Dimethyl diphenylphosphoramidite
Hexamethylphosphorous triamide (10.0 mL, 55.0 mmol) was
added to a solution of phenol (10.4 g, 110 mmol) in dry 1,2-
dimethoxyethane (20 mL). The mixture was refluxed under a N2
atmosphere for 20 h, after which time TLC (1% KMnO4 in H2O)
indicated conversion of the starting materials to a compound of
lower polarity. The solvent was removed by rotary evaporation,
and the residue was purified by distillation to afford N,Nꢀ-dimethyl
diphenylphosphoramidite as a colourless liquid (13.8 g, 96%); nD20
1.5698 (lit.36 nD20 1.5660); 1H NMR (400 MHz; CDCl3) d 2.79 (6H,
d, JH,P 9.6 Hz, 2 × CH3), 7.05–7.35 (10H, m, Ph); 13C NMR
(100 MHz; CDCl3) d 34.83 (d, JCNP 82 Hz, 2 × CH3), 120.10 (d,
JC,P 34 Hz, Ph), 120.74 (d, JC,P 27 Hz, Ph), 123.02, 124.25 (2s,
Ph), 129.54, 129.70 (2s, Ph), 151.61 (d, JC,P 15 Hz, Ph), 153.69 (d,
JC,P 27 Hz, Ph); 31P (161.8 MHz; CDCl3) d 139.02 (lit.37 139.09 in
C6H6).
◦
7 as a colourless solid (1.62 g, 81%), mp 94–96 C (EtOAc–pet.
spirits); [a]2D5 22.7 (c 0.950, CHCl3) (Found: C, 64.37; H, 6.09.
1
C29H33O8P requires C, 64.44; H, 6.15%); H NMR (400 MHz;
CDCl3) d 1.30, 1.38 (6H, 2 s, 2 × CH3), 2.56 (1H, ddd, J5,6 5.6, J6,6
16.0, JH,P 19.0 Hz, H6), 2.70 (1H, ddd, J5,6 8.0, JH,P 18.0 Hz, H6),
3.37–3.45 (1H, m, H5), 3.47 (3H, s, OCH3), 4.19–4.32 (3H, m,),
4.27 (1H, d, J1,2 8.0 Hz, H1), 4.79 (2H, ABq, CH2Ph), 7.15–7.39
(15H, m, Ph); 13C NMR (100 MHz; CDCl3) d 26.34, 27.72 (2 ×
C(CH3)2), 28.44 (d, JC6,P 144.2 Hz, C6), 56.88 (OCH3), 68.04,
73.48, 75.28 (d, JC,P 10.6 Hz), 78.99, 79.28 (C2,3,4,5,CH2Ph),
103.99 (C1), 109.96 (C(CH3)2OCH3), 120.52–120.72 (2d, Ph),
125.26, 128.12, 128.24, 129.77, 129.84, 138.29, 150.25, 150.30,
150.33, 150.39 (Ph); 31P NMR (161.8 MHz; CDCl3) d 20.63 (1P).
Isopropyl diphenyl phosphite
Method A. 2-Propanol (1.88 mL, 24.6 mmol) was added drop-
wise to a solution of N,Nꢀ-dimethyl diphenylphosphoramidite
(6.43 g, 24.6 mmol) and 1H-tetrazole (3.35 g, 47.9 mmol) in 1,2-
dimethoxyethane (25 mL) at 0 ◦C. The mixture was stirred and
allowed to warm to r.t. After 7 h TLC (vanillin stain) indicated
conversion of the phosphoramidite to a material of lower polarity.
The solvent was evaporated and the solid residue was suspended
in a small amount of pet. spirits. The suspension was filtered
through a short plug of silica, and was washed with pet. spirits
(4 × 100 mL). Concentration of the eluant in vacuo afforded a
pale yellow liquid. The residue was distilled to afford isopropyl
diphenyl phosphite as a colourless liquid (5.13 g, 75%); n2D0 1.5369
Methyl 2-O-benzyl-6-dibenzyloxyphosphinyl-6-deoxy-3,4-O-
isopropylidene-b-D-galactoside 8
˚
Benzyl alcohol (distilled and stored over 3 A molecular sieves,
62.0 lL, 600 lmol) and sodium hydride (60% in oil, 25.2 mg,
630 lmol) were added to THF (2 mL) in a two-necked flask under
N2. This mixture was stirred until effervescence ceased, and then
the diphenyl phosphonate 7 (108 mg, 199 lmol) was added via
the second neck. After 5 min the reaction was quenched with
sat. aq. NH4Cl (2 mL) and the mixture was partitioned between
water and dichloromethane. The aqueous phase was separated and
re-extracted with dichloromethane (4 × 10 mL). The combined
organic extracts were then washed with H2O (2 × 25 mL) and
sat. aq. NaCl (25 mL), dried (MgSO4) and evaporated to afford
a crude brown syrup. The crude material was purified by flash
chromatography (40–80% EtOAc–pet. spirits) to yield the dibenzyl
phosphonate 8 as yellow crystals (83.3 mg, 74%). Recrystallisation
from EtOAc–pet. spirits afforded colourless needles, mp 73–74 ◦C;
[a]2D5 23.5 (c 0.495, CHCl3) (Found: C, 65.51; H, 6.63. C31H37O8P
requires C, 65.48; H, 6.56%); 1H NMR (500 MHz; CDCl3) d 1.21,
1.32 (6H, 2 s, 2 × CH3), 2.29–2.41 (2H, m, H6,6), 3.32 (1H, dd,
J1,2 8.0, J2,3 7.0 Hz, H2), 3.44 (3H, s, OCH3), 4.01–4.10 (3H, m,
1
(lit.15 nD21 1.5394); H NMR (500 MHz; CDCl3) d 1.34 (6H, d,
J 6.0 Hz, 2 × CH3), 4.93 (1H, m, CH(CH3)2), 7.07–7.33 (10H,
m, Ph); 13C NMR (125 MHz; CDCl3) d 24.61 (d, JC,P 2.7 Hz, 2 ×
CH3), 68.28 (d, JC,P 3.6 Hz, CH(CH3)2), 120.59, 120.67, 121.05,
121.12, 123.99, 124.64, 130.02, 130.09 (Ph), 152.79 (d, JC,P 7.4 Hz,
Ph); 31P (161.8 MHz; CDCl3) d 129.82.
Method B. Triethylamine (48.0 mL, 346 mmol) followed by
isopropanol (4.50 mL, 58.8 mmol) in anhydrous ether (25 mL)
was added dropwise to a chilled solution of phosphorus trichloride
(5.00 mL, 57.3 mmol) in anhydrous ether (500 mL). The reaction
956 | Org. Biomol. Chem., 2007, 5, 952–959
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
H3,4,5), 4.14 (1H, d, J1,2 8.0 Hz, H1), 4.76 (2H, dd, CH2Ph),
4.94–5.08 (4H, m, 2 × POCH2Ph), 7.30–7.37 (10H, m, Ph); 13C
NMR (125 MHz; CDCl3) d 26.20, 27.72 (2 × C(CH3)2), 27.65
(C6), 56.83 (OCH3), 67.20–67.50 (m, 2 × POCH2Ph), 67.96 (C3),
73.45 (CH2Ph), 75.14 (d, C5), 78.84, 79.19 (C2,4), 103.80 (C1),
109.73 (C(CH3)2), 127.54–128.61 (Ph), 136.08–136.22 (m, Ph),
138.23 (Ph); 31P NMR (161.8 MHz; CDCl3) d 28.25.
(C(CH3)2), 28.57 (d, J6,P 138.1 Hz, C6), 56.32 (OCH3), 66.02 (d,
J
C,P 2.0Hz, POCH2Ph), 68.33, 72.88, 78.54, 79.06(C2,3,4,CH2Ph),
75.02 (d, J5,P 6.8 Hz, C5), 103.32 (C1), 109.10 (C(CH3)2), 125.88,
127.11, 127.15, 127.51, 127.72, 127.81, 128.02, 137.10, 137.16,
137.99, 142.52, 143.19 (Ar); 31P NMR (161.8 MHz; CDCl3) d
30.64.
Methyl 2-O-benzyl-6-deoxy-6-diisopropoxyphosphinyl-3,4-O-
isopropylidene-b-D-galactoside 12
Methyl 6-deoxy-6-dihydroxyphosphinyl-b-D-galactopyranoside,
ammonium salt 1
A mixture of the iodide 6 (603 mg, 1.39 mmol) and triisopropy-
lphosphite (1.6 mL, 7.0 mmol) was stirred at 180 ◦C for 20 h,
whereupon TLC indicated complete conversion of the iodide to
a compound of higher polarity. The reaction mixture was cooled,
and the crude material purified by flash chromatography (80–
100% EtOAc–pet. spirits) to afford the phosphonate 12 as a
yellow syrup (569 mg, 87%), [a]2D6 25.5 (c 1.225, CHCl3); 1H NMR
(400 MHz; CDCl3) d 1.29–1.33 (18H, m, 6 × CH3), 2.16–2.34 (2H,
m, H6,6), 3.34 (1H, dd, J 6.0, 8.0 Hz, H3), 3.53 (3H, s, OCH3),
4.04–4.15 (3H, m, H2,3,5), 4.18 (1H, d, J1,2 8.0 Hz, H1), 4.66–
4.74 (2H, m, 2 × CH(CH3)2), 4.77 (2H, ABq, CH2Ph), 7.20–7.37
(5H, m, Ph); 13C NMR (100 MHz; CDCl3) d 24.23, 24.26, 24.31
(CH(CH3)2), 26.60, 27.98 (CH3), 29.65 (d, JC6,P 143.5 Hz, C6),
57.04 (OCH3), 70.46, 70.66 (2d, JC,P 6.9 Hz), 68.53, 73.64, 75.41
(d, JC,P 6.6 Hz), 79.14, 79.53 (C2,3,4,5, 2 × CH(CH3)2,CH2Ph),
104.07 (C1), 109.88 (CMe2), 127.71, 128.31, 128.40, 138.51 (Ph);
31P NMR (161.8 MHz; CDCl3) d 25.01; HRMS (ESI+, m/z) calc.
for C23H37NaO8P [M + Na]+ 495.2118, found 495.2114.
A solution of dibenzyl phosphonate 8 (193 mg, 0.339 lmol) in
trifluoroacetic acid–water (9 : 1, 1.0 mL) was stirred for 40 min.
The solvent was evaporated by co-evaporation with toluene (4 ×
2 mL) to afford a crude brown syrup. A suspension of the crude
syrup, Pd/C (10%, 40 mg) in THF–water (2 : 1, 30 mL) was
treated with H2 for 18 h. The mixture was filtered (Celite), pyridine
(87 lL, 1.06 mmol) was added to the filtrate, and the solution
was evaporated to dryness to afford the pyridinium salt as pale
yellow syrup. Cyclohexylamine (0.39 mL, 3.39 mmol) in toluene
(5 mL) were added to the crude pyridinium salt and the mixture
was evaporated to dryness. Excess cyclohexylamine was removed
by co-evaporation with toluene (4 × 4 mL), yielding a colourless
solid. The crude material was purified by flash chromatography (7 :
2 : 1 then 2 : 2 : 1 EtOAc–MeOH–H2O), followed by size-exclusion
chromatography (Bio-Gel P2, 250 mM NH4HCO3 buffer). The
buffer was removed by lyophilisation to afford a colourless fluffy
solid. This material was passed through an ion exchange column
+
(Dowex 50WX8-400, NH4 form) and freeze-dried to afford the
Methyl 2-O-benzyl-6-deoxy-6-dihydroxyphosphinyl-3,4-O-
isopropylidene-b-D-galactoside, triethylammonium salt 11
ammonium salt of 1 as a colourless, hygroscopic powder (54.8 mg,
51%); [a]2D5 −11.4 (c 0.71, H2O) (Found: C, 27.98; H, 6.72; N,
4.85. C7H18NO8P·1.5H2O requires C, 27.82; H, 7.00; N, 4.63%);
1H NMR (400 MHz; D2O) d 1.89–2.04 (2H, m, H6,6), 3.44 (1H,
dd, J1,2 8.0, J2,3 10 Hz, H2), 3.53 (3H, s, OCH3), 3.65 (1H, dd, J2,3
10, J3,4 3.2 Hz, H3), 3.85 (1H, m, H5), 3.92 (1H, d, J3,4 3.2 Hz.
H4), 4.29 (1H, d, J1,2 8.0 Hz, H1); 13C NMR (100 MHz; D2O) d
23.91, 24.41, 29.13, 30.45, 48.97, 50.47, 57.18, 70.64, 71.54, 73.03,
103.67 (C1); 31P NMR (161.8 MHz; D2O) d 22.03.
TMSBr (280 lL, 2.16 mmol) was added to a solution of the diiso-
propyl phosphonate 12 (254 mg, 0.538 mmol) and triethylamine
(1.00 mL, 7.17 mmol) in dry dichloromethane (9 mL) at 0 ◦C
and the resultant mixture was stirred for 21 h. Additional TMSBr
(280 lL, 2.16 mmol) was added and stirring was continued for a
further 6 h. The reaction mixture was then concentrated in vacuo.
The residue was purified by flash chromatography (7 : 2 : 1 then
5 : 2 : 1 EtOAc–MeOH–H2O) to give the salt 11 as a yellow oil
(263 mg, 92%), [a]2D3 +22.6 (c 1.20, CHCl3); 1H NMR (400 MHz;
D2O) d 1.07 (9H, t, J 7.2 Hz, CH3CH2N), 1.15, 1.17 (2 × 3H, 2s,
C(CH3)2), 1.84 (1H, dd, J5,6 5.6, J6,P 5.6 Hz, H6), 1.89 (1H, dd J5,6
5.6, J6,P 5.6 Hz, H6), 2.98 (6H, q, J 7.2 Hz, CH3CH2N), 3.17 (1H,
dd, J1,2 8.2, J2,3 8.2 Hz, H2), 3.37 (3H, s, OCH3), 3.95–4.01 (1H, m,
H5), 4.07 (1H, dd, J2,3 6.0, J3,4 6.0 Hz, H3), 4.17 (1H, d, J3,4 5.2 Hz,
H4), 4.22 (1H, d, J1,2 8.4 Hz, H1), 4.59 (2H, s, CH2Ph), 7.18–7.26
(5H, m, Ph); 13C NMR (100 MHz, CDCl3) d 8.91 (CH3CH2N),
25.94, 27.58 (2C, C(CH3)2), 30.65 (d, JC6,P 134 Hz, C6), 47.30
(CH3CH2N)), 57.54 (OMe), 69.74, 73.91, 78.91, 79.62 (C2,3,4,
CH2Ph), 76.62 (d, JC5,P 7.3 Hz, C5), 103.39 (C1), 111.11 (C(CH3)2),
129.07, 129.33, 129.60, 137.60 (4C, Ph); 31P NMR (161.8 MHz,
CDCl3) d 25.46 (1P); HRMS (ESI−, m/z) calcd for C17H25O8P
[M − H]− 387.1203, found 387.1201.
Methyl 2-O-benzyl-6-benzyloxyphosphinyl-6-deoxy-3,4-O-
isopropylidene-b-D-galactoside, pyridinium salt 9
A solution of dibenzyl phosphonate 8 (267 mg, 470 lmol, 94 mM),
2-mercaptobenzothiazole (1.26 g, 7.36 mmol, final concentration
1.5 M) and ethyldiisopropylamine (1.3 mL, 7.6 mmol, 1.5 M) in
DMF (5 mL) was stirred at 50 ◦C under N2 for 19 h. The solvent
was evaporated under reduced pressure, and the crude orange
solid was purified by flash chromatography (100% EtOAc then
7 : 2 : 1 EtOAc–MeOH–H2O) to afford a clear colourless glass
(248 mg). Pyridine (2 mL) was added to the glass, and the excess
solvent removed by rotary evaporation. The residue was dissolved
in MeOH and passed through an ion exchange column (Dowex
50W X8-400, pyridinium form) to afford the pyridinium salt 9 as
a clear yellow syrup (247 mg, 94%); 1H NMR (500 MHz; CDCl3)
d 1.23, 1.32 (2 × 3H, 2 × s, C(CH3)), 2.27–2.42 (2H, m, H6,6),
3.37 (1H, dd, J1,2 7.0, J2,3 7.0 Hz, H2), 3.45 (3H, s, OCH3), 4.08–
4.16 (3H, m, H3,4,5), 4.19 (1H, m, J1,2 7.0 Hz, H1), 4.74–4.80
(2H, ABq, CH2Ph), 7.22–7.40 (10H, m, Ph), 7.60–7.70, 8.08–8.18,
8.65–8.80 (3 × m, pyr); 13C NMR (125 MHz; CDCl3) d 25.92, 27.49
Methyl 2-O-benzyl-6-deoxy-6-[(2R)-2,3-dipalmitoyloxypropyl-
oxy]hydroxyphosphinyl-3,4-O-isopropylidene-b-D-galactoside 13
The phosphonate salt 11 (192 mg, 392 lmol) was dried by co-
evaporation with pyridine (3 × 5 mL). Dipalmitoyl-sn-glycerol
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 952–959 | 957
©

View Article Online
(223 mg, 392 lmol) was then added and the reaction mixture dried
again by co-evaporation with pyridine (2 × 5 mL). The mixture
was dissolved in pyridine (8 mL), treated with DCC (809 mg,
3.92 mmol), and heated to 40 ◦C for 24 h. The reaction mixture was
then quenched with H2O (0.5 mL) and extracted with chloroform
(5 × 30 mL). The combined organic extract was dried (MgSO4) and
concentrated in vacuo. The resulting residue was purified by flash
chromatography (100% EtOAc then 17 : 2 : 1 EtOAc–MeOH–
H2O) to give the protected PI analogue 13 as a colourless solid
(215 mg, 59%), mp 77–78 ◦C; [a]D23 +22.2 (c 0.62, CHCl3) (Found:
C, 62.42; H, 9.19. C52H90KO12P. H2O requires C, 62.75; H, 9.32%);
1H NMR (400 MHz; D2O) d 0.87 (6H, m, CH3 of palmitoyl),
1.18–1.34 (54H, m, palmitoyl sidechain, C(CH3)2), 1.55 (4H, m,
MgCl2 before incubation at 37 ◦C and addition of analogue
1 or 2 followed by GDP-[3H]mannose (6.3 Ci mmol−1, final
concentration 9 mCi mL−1) to initiate labelling. The labelling
reaction was terminated at the indicated time points by adding
chloroform–methanol (1 : 1, v/v, 6.7 vol.). After sonication, the
insoluble material was removed by centrifugation (15 000 g, 2 min)
and the supernatant was dried under N2. Labelled lipids were
recovered by biphasic partitioning between water and 1-butanol
(1 : 2, v/v) and analysed by HPTLC developed in chloroform–
methanol–1 M ammonium acetate–13 M ammonia–water (180 :
140 : 9 : 9 : 23 v/v). Radiolabelled lipids were detected by treating
the HPTLC sheets with En3Hance (NEN Life Science Products,
Boston, MA, USA) and exposing them to a BioMax MR film
(Kodak, Rochester, NY, USA) at −80 ◦C.
=
CH2), 1.98–2.30 (6H, m, CH2C O, H6,6), 3.32 (1H, m, H2), 3.52
(3H, s, OCH3), 4.00–4.39 (8H, m, H1,3,4,5, 2 × CH2 of glycerol),
4.75 (2H, s, CH2Ph), 5.22 (1H, br s, CH of glycerol), 7.20–7.40
(5H, m, Ph); 13C NMR (100.5 MHz; CDCl3) d 14.1, 22.67, 24.84,
24.91, 26.34, 27.82, 29.18, 29.21, 29.36, 29.41, 29.58, 29.60, 29.67,
29.72, 31.91, 34.16 (d, JC6,P 17.7 Hz, C6), 57.15 (OMe), 62.78,
69.15, 70.60, 73.29 (C2,3,4,CH2Ph), 79.31 (d, JC5,P 71.4 Hz, C5),
103.72 (C1), 109.51 (C(CH3)2), 127.42, 127.92, 128.14, 138.35 (Ph);
Acknowledgements
This work was supported by the Australian Research Council (Re-
search contract DP0449625). Prof. Bernard Golding is gratefully
acknowledged for providing access to unpublished work.16 MJM
is a National Health and Medical Research Council Principal
Research Fellow.
173.39, 173.67 (C O); 31P NMR (161.8 MHz; CDCl3) d 19.38 (P);
=
HRMS (ESI−, m/z) calcd for C52H91O12P [M − H]− 937.6164,
found 937.6171.
References
1 R. H. Michell, N. M. Perera and S. K. Dove, Biochem. Soc. Trans.,
2003, 31, 11–15.
2 H. Hilbi, Cell. Microbiol., 2006, 8, 1697–1706.
3 J. Nigou, L. G. Dover and G. S. Besra, Biochemistry, 2002, 41, 4392–
4398.
4 T. Parish, J. Liu, H. Nikaido and N. G. Stoker, J. Bacteriol., 1997, 179,
7827–7833.
5 M. Jackson, D. C. Crick and P. J. Brennan, J. Biol. Chem., 2000, 275,
30092–30099.
6 Y. S. Morita, R. Velasquez, E. Taig, R. F. Waller, J. H. Patterson, D.
Tull, S. J. Williams, H. Billman-Jacobe and M. J. McConville, J. Biol.
Chem., 2005, 280, 21645–21652.
7 V. Briken, S. A. Porcelli, G. S. Besra and L. Kremer, Mol. Microbiol.,
2004, 53, 391–403.
8 J. Kordula´kova´, M. Gilleron, K. Mikusova´, G. Puzo, P. J. Brennan, B.
Gicquel and M. Jackson, J. Biol. Chem., 2002, 277, 31335–31344.
9 F. Chre´tien, N. Moitessier, F. Roussel, J.-P. Mauger and Y. Chapleur,
Curr. Org. Chem., 2000, 4, 513–534.
Methyl 6-deoxy-6-[(2R)-2,3-dipalmitoyloxypropyloxy]-
hydroxyphosphinyl-b-D-galactopyranoside 2
The protected PI analogue 13 (37 mg, 39 lmol) was treated with
TFA–H2O (9 : 1, 1.0 mL) for 30 min then concentrated in vacuo
with co-evaporation with toluene to remove residual solvent. The
residue was dissolved in methanol–ethyl acetate (1 : 1, 8 mL),
treated with 10% Pd(OH)2 on carbon (40 mg) and shaken under
H2 at 350 kPa for 24 h. The residue was filtered through Celite,
concentrated in vacuo and purified by flash chromatography (17 :
2 : 1 EtOAc–MeOH–H2O) to give the PI analogue 2 as a yellow
oil (24 mg, 75%); 1H NMR (400 MHz; D2O) d 0.88 (6H, m,
CH3 of palmitoyl), 1.25 (48H, m, palmitoyl sidechain), 1.58 (m,
=
CH2), 1.95–2.40 (6H, m, CH2C O, H6), 3.50 (3H, s, OCH3),
3.55–4.42 (9H, m, H1,2,3,4,5, 2 × CH2 of glycerol), 5.23 (1H,
bs, CH of glycerol); 31P NMR (161.8 MHz; CDCl3) d 21.83 (P);
HRMS (ESI−, m/z) calc. for C42H80O12P [M–H]− 807.5382, found
807.5374.
10 C. S. Rye and J. B. Baell, Curr. Med. Chem., 2005, 12, 3127–3141.
11 A. Borba´s, A. Biro and A. Lipta´k, ACH Models Chem., 1994, 131,
455–465.
12 J.-F. Valdor and W. Mackie, J. Carbohydr. Chem., 1997, 16, 429–440.
13 H. Yuasa, M. M. Palcic and O. Hindsgaul, Can. J. Chem., 1995, 73,
2190–2195.
14 M. F. Wang, M. M. L. Crilley, B. T. Golding, T. Mcinally, D. H.
Robinson and A. Tinker, J. Chem. Soc., Chem. Commun., 1991, 667–
668.
15 A. E. Arbuzov and L. V. Nesterov, Bull. Acad. Sci. USSR Div. Chem.
Sci. (Engl. Transl.), 1954, 361–367, (Chem. Abstr., 1955, 9541).
16 M. F. Wang, A. C. Tinker and B. T. Golding, personal communication.
17 P. L. Barili, G. Berti, G. Catelani, F. Colonna and A. Marra, Tetrahedron
Lett., 1986, 27, 2307–2310.
18 G. Catelani, F. Colonna and A. Marra, Carbohydr. Res., 1988, 182, 297.
19 J. C. McAuliffe, R. V. Stick, D. M. G. Tilbrook and A. G. Watts,
Aust. J. Chem., 1998, 51, 91.
20 M. G. N. Russell, R. Baker and D. C. Billington, Carbohydr. Res., 1992,
234, 263–268.
21 Y. S. Morita and M. J. McConville, personal communication.
22 A. Andrus and S. L. Beaucage, Tetrahedron Lett., 1988, 29, 5479–5482.
23 M. Saady, L. Lebeau and C. Mioskowski, J. Org. Chem., 1995, 60,
2946–2947.
24 S. Gobec, I. Plantan, J. Mravljak, R. A. Wilson, G. S. Besra and D.
Kikelj, Bioorg. Med. Chem. Lett., 2004, 14, 3559–3562.
Biochemical analysis using M. smegmatis cell-free system
M. smegmatis mc2155 was grown to mid-exponential phase at
37 ◦C in 7H9 broth (Difco, Detroit, MI, USA), supplemented with
0.2% (w/v) glucose, 0.2% (v/v) glycerol and 15 mM NaCl. Bacteria
were harvested by centrifugation in mid-exponential growth and
washed twice with 50 mM Hepes/NaOH (pH 7.4). Pellets were
resuspended at 0.2 g of wet pellet per mL using a lysis buffer
containing 25 mM Hepes/NaOH (pH 7.4), 25% (w/v) sucrose,
2 mM EGTA and a cocktail of protease inhibitors (Roche). The
cells were subjected to three rounds of nitrogen cavitation (Kontes,
Vineland, NJ, USA) at 15 MPa with a 30 min equilibration
period before the release of pressure. Unlysed cells were removed
by centrifugation (2500 g, 10 min) and the supernatant was
stored at −80 ◦C. The lysates were supplemented with 5 mM
958 | Org. Biomol. Chem., 2007, 5, 952–959
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
25 J. M. Campagne, J. Coste and P. Jouin, Tetrahedron Lett., 1993, 34,
6743–6744.
26 C. E. Dreef, A. R. P. M. Valentijn, E. de Vroom, G. A. van der Marel
and J. H. van Boom, Tetrahedron Lett., 1988, 29, 1199–1202.
27 T. K. Vinod, O. H. Griffith and J. F. W. Keana, Tetrahedron Lett., 1994,
35, 7193–7196.
28 E. T. Kaiser and K. Kudo, J. Am. Chem. Soc., 1967, 89, 6725–6728.
29 J. L. Montchamp, L. T. Piehler and J. W. Frost, J. Am. Chem. Soc.,
1992, 114, 4453–4459.
33 R. V. Tatituri, P. A. Illarionov, L. G. Dover, J. Nigou, M. Gilleron, P. G.
Hitchen, K. Krumbach, H. R. Morris, N. Spencer, A. Dell, L. Eggeling
and G. S. Besra, J. Biol. Chem., 2007, 10.1074/jbc.M608695200.
34 W. C. Still, M. Kahn and A. M. Mitra, J. Org. Chem., 1978, 43, 2923–
2925.
35 H. M. Zuurmond, P. A. M. van der Klein, G. A. van der Marel and
J. H. van Boom, Tetrahedron, 1993, 49, 6501–6514.
36 V. P. Kukhar and M. V. Shevchenko, Zh. Obshch. Khim., 1982, 52,
562–566.
30 C. E. McKenna, M. T. Higa, N. H. Cheung and M. C. Mckenna,
Tetrahedron Lett., 1977, 155–158.
31 C. D. Sun and R. Bittman, J. Org. Chem., 2004, 69, 7694–7699.
32 M. L. Schaeffer, K. H. Khoo, G. S. Besra, D. Chatterjee, P. J. Brennan,
J. T. Belisle and J. M. Inamine, J. Biol. Chem., 1999, 274, 31625–
31631.
37 E. E. Nifantyev, E. N. Rasadkina, P. V. Slitikov and L. K. Vasyanina,
Phosphorus, Sulfur, Silicon, Relat. Elem., 2003, 178, 2465–2477.
38 L. Czollner, G. Baudin, B. Bernet and A. Vasella, Helv. Chim. Acta,
1993, 76, 1013–1025.
39 Y. S. Morita, J. H. Patterson, H. Billman-Jacobe and M. J. McConville,
Biochem. J., 2004, 378, 589–597.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 952–959 | 959
©