Stereospecific synthesis of 5-phospho-α-D-arabinosyl-C-phosphonophosphate (pACpp): A stable analogue of the putative mycobacterial cell wall biosynthetic intermediate 5-phospho-D-arabinosyl pyrophosphate (pApp)

Article abstract of DOI:10.1016/S0040-4039(01)00118-6

The stereospecific synthesis of 5-phospho-α-D-arabinosyl-C-phosphonophosphate (pACpp) from D-glucosamine is described. This compound was evaluated for its ability to serve as a stable analogue of the putative mycobacterial cell wall biosynthetic intermediate 5-phospho-D-arabinosyl pyrophosphate (pApp). The phosphonophosphate proved incapable of interfering with formation of the mycobacterial arabinan precursor decaprenylphospho-arabinose (DpA) in vitro.

Full text of DOI:10.1016/S0040-4039(01)00118-6

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 2231–2234
Stereospecific synthesis of
5-phospho-a- -arabinosyl-C-phosphonophosphate (pACpp):
D
a stable analogue of the putative mycobacterial cell wall
biosynthetic intermediate 5-phospho-
(pApp)
D-arabinosyl pyrophosphate
Philip McGurk,a Grace X. Chang,b Todd L. Lowary,b Michael McNeilc and Robert A. Fielda,*
aSchool of Chemistry, University of St Andrews, St Andrews KY16 9ST, UK
bDepartment of Chemistry, The Ohio State University, Columbus, OH 43210-1173, USA
cDepartment of Microbiology, Colorado State University, Fort Collins, CO 80523, USA
Received 4 December 2000; revised 8 January 2001; accepted 17 January 2001
Abstract—The stereospecific synthesis of 5-phospho-a-
described. This compound was evaluated for its ability to serve as a stable analogue of the putative mycobacterial cell wall
biosynthetic intermediate 5-phospho- -arabinosyl pyrophosphate (pApp). The phosphonophosphate proved incapable of interfer-
D-arabinosyl-C-phosphonophosphate (pACpp) from D-glucosamine is
D
ing with formation of the mycobacterial arabinan precursor decaprenylphospho-arabinose (DpA) in vitro. © 2001 Elsevier
Science Ltd. All rights reserved.
Since the mid 1980s there has been a disturbing resur-
gence in the incidence of tuberculosis world-wide,
resulting in the highest ever annual mortality rate from
the disease in 1997. There are 8 million new infections
and 3 million deaths from TB every year; it is expected
that in the next decade 90 million new cases will occur,
resulting in 30 million deaths.1 There is a real need to
identify new targets and develop new classes of drugs
for the treatment of TB. To date, several front-line
anti-tubercular agents have targeted steps in mycobac-
terial cell wall biosynthesis, including isoniazid, ethion-
amide and ethambutol.2 It is the complex combination
of mycolylated arabinogalactan, lipoarabinomannan
and other components that contribute to the impreg-
nable nature of the cell wall. It has been established
The biosynthesis of DpA originates from 5-phosphori-
bosyl pyrophosphate (pRpp), which is formed by the
pentose shunt pathway, and might occur via one of two
routes (Scheme 1).6 Route A would involve epimerisa-
tion of the C-2 position of pRpp to give the arabino-
configured pApp, followed by conversion to 5-phospho
decaprenylphospho-arabinose (5p-DpA) and dephos-
phorylation to DpA. Alternatively, attachment of the
lipid tail might precede C-2 epimerisation; in route B,
pRpp might first be converted to 5-phospho
decaprenylphospho-ribose (5p-DpR), followed by C-2
epimerisation to arabino-configured 5p-DpA and
dephosphorylation to give DpA. Intermediates 5p-DpA
and 5p-DpR have been identified in cell-free incuba-
tions by HPLC and by TLC/autoradiography, as have
products DpA and DpR. However, experiments using
radioactive 5p-DpR and Mycobacterium smegmatis cell
wall preparations failed to produce either 5p-DpA or
DpA, instead producing DpR as the sole product.6 This
suggests the absence of 5p-DpR and DpR 2-epimerase
activities in such preparations, hinting at a role for
pApp as a biosynthetic intermediate in DpA formation.
that most of the
D-arabinofuranosyl residues in cell
wall arabinan-containing structures are derived from
decaprenylphospho-arabinose (DpA);3–5 it follows that
inhibition of the biosynthesis of this molecule should
prove deleterious to the organism.
Keywords:
D-arabinose; 2,5-anhydro-D-mannitol; phosphonophos-
phate; mycobacteria; cell wall biosynthesis.
In order to assess whether or not pApp does indeed
serve as an intermediate in the biosynthesis of DpA, its
chemical synthesis was considered. However, in light of
* Corresponding author. Present address: School of Chemical Sci-
ences, University of East Anglia, Norwich NR4 7TJ, UK. Fax:
+44-1603-592003; e-mail: r.a.field@uea.ac.uk
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)00118-6

2232
P. McGurk et al. / Tetrahedron Letters 42 (2001) 2231–2234
O
O
-O P O
O-
-O P O
O
O
HO
O
O
O
O
O-
A
O P O P O-
O P O P O-
OH
OH
OH
O- O-
O- O-
pRpp
pApp
A
B
O
-O P O
O-
O
-O P O
O-
O
O
O P O-polyprenyl
O P O-polyprenyl
O
O
HO
O-
B
O-
OH
5p-DpR
OH
OH
5p-DpA
A, B
O
O
HO
HO
O P O-polyprenyl
O P O-polyprenyl
O
O
HO
O-
O-
OH
OH
DpR
OH
DpA
Scheme 1. Prospective biosynthetic routes (A and B) from 5-phosphoribosyl pyrophosphate (pRpp) to decaprenylphospho-
arabinose (DpA).
HO
HO
RO
O
O
O
HO
RO
OH
OH
OH
I
HO
OH
OR
NH2
2,5-Anhydro-D-mannitol
D-Glucosamine
O
O
P
O-
R'O
P O
-O
O
HO
O
O
O
RO
O
HO
O
O
P
RO
O
P
R'O
P
OH
P
O
O-
OR''
OR
OH
OR
OH
O- O-
OR''
pACpp
Scheme 2. Synthetic approach to 5-phospho- -arabinosyl-a-C-phosphonophosphate (pACpp).
D
nitol itself is readily accessible from
D-glucosamine via
difficulties encountered in the attempted synthesis and
deprotection of this compound,7 an alternative
approach involving synthesis of a stable analogue,
a well established stereospecific diazonium ion-medi-
ated ring contraction reaction10,11 (Scheme 2).
namely 5-phospho-a-D-arabinosyl-C-phosphonophos-
phate (pACpp), was investigated.8 An earlier synthesis
of pACpp by McClard and co-workers9 employed
addition of dimethylmethylphosphonate to a protected
Synthesis (Scheme 3): 2,5-Anhydro- -mannitol 1 was
D
prepared from glucosamine hydrochloride on a 20 g
scale using established literature procedures.10,11 Selec-
tive mono-O-tritylation, subsequent benzylation and
de-tritylation, without purification of the intermediate,
gave tri-O-benzylated primary alcohol 212 in good
overall yield. Introduction of the phosphonate func-
tionality was achieved in a straightforward manner,
using Arbuzov chemistry, from the corresponding pri-
mary iodide 3, which was prepared from primary alco-
hol 2 using PPh3, NIS and DMF.13 The large up field
shift of the 13C NMR signal attributed to C-1 (from
62.7ppm in 2 to 6.9ppm in 3) reflected the success of
the reaction.
D
-arabinofuranose,
giving
an
inseparable
diastereomeric mixture of gluco- and manno-
configured phosphonate intermediates. This prevented
detailed analysis of all but the final synthetic com-
pounds, which themselves proved difficult to separate.
We favoured a stereospecific route to pACpp, where
the potentially problematic formation of the required
manno-configured anhydrosugar was in place at an
early stage in the synthesis. This led us to consider a
suitably protected 2,5-anhydro-
as a synthetic intermediate, since 2,5-anhydro-
D
-mannitol derivative
-man-
D

P. McGurk et al. / Tetrahedron Letters 42 (2001) 2231–2234
2233
HO
O
HO
O
BnO
O
20g scale
OH
HO
BnO
i,ii
OH
OH
OH
HO
(refs 10,11)
OH
OBn
NH2
(2)
(1)
D-Glucosamine
iii
O
PhO P O
PhO
BnO
BnO
O
O
O
BnO
O
v-vii
BnO
O
BnO
iv
P OEt
OEt
P OEt
OEt
I
OBn
OBn
OBn
(4)
(5)
(3)
vii, ix
O
PhO P O
PhO
O
-O P O
O
O
BnO
O
O
HO
O
O
O-
x
P O P O-
O- O-
P O P OPh
O- O-
OH
OBn
(6)
pACpp
Scheme 3. Synthesis of 5-phospho-a-
D
-arabinosyl-C-phosphonophosphate (pACpp). (i) TrCl, pyr (41%); (ii) NaH, BnBr, DMF;
TsOH, CH2Cl2/MeOH, 5:1 (82%, two steps); (iii) NIS, Ph3P, DMF (76%); (iv) P(OEt)3, D (83%); (v) conc. H2SO4, Ac2O; Na,
MeOH (81%, two steps); (vi) diphenylphosphoryl chloride, pyr (98%); (vii) TMS–Br, CH2Cl2 (quant.); (viii) carbonyldiimidazole,
DMF, phenylphosphate (44%); (ix) H2, Pd–C, AcOH; H2, PtO2, Pt–C, AcOH (46%, two steps).
Unfortunately a first attempt at the Arbuzov reaction,
using refluxing trimethyl phosphite, gave no conversion
of iodide 3 to the corresponding dimethylphosphonate,
ruling out direct comparison of subsequent synthetic
intermediates with McClard’s data.8,9† On switching to
the higher boiling triethyl phosphite, the Arbuzov
chemistry proceeded in good yield. The identity of the
diethylphosphonate 4 was supported by the appearance
of a characteristic signal at 28.2 ppm in the 31P NMR
spectrum and the observed couplings of H-1 and H-1%
(18.9 and 18.4 Hz), C-1 (137.5 Hz) and C-2 (7.5 Hz)
with phosphorus. Subsequent steps in the synthesis
followed the route described by McClard.9 Selective
acetolysis of the 6-O-benzyl group of 4 followed by
deacetylation resulted in the unmasking of the primary
alcohol, which was phosphorylated with diphenylphos-
1
phoryl chloride to give 5. In addition to H and 13C
NMR data, the appearance of a new signal at −13.0
ppm in the 31P NMR spectrum, characteristic of a
phosphate ester,14 confirmed the identity of the
product. Dealkylation of phosphonate 5 with bro-
motrimethylsilane and subsequent formation of the
phosphonophosphate, via activation with carbonyldi-
imidazole and coupling to phenyl phosphate, also pro-
†
Selected data: 2,5-Anhydro-3,4-di-O-benzyl-1-deoxy-1-diethylphos-
phono-6-O-diphenylphosphono-D-mannitol (5).
Amorphous solid (calcd for C36H42O10P2: C, 62.07; H, 6.08; found
C, 61.80; H, 6.35); lH (CDCl3), 1.27 (6H, dt, J 7.06, J 1.54,
ceeded
satisfactorily.
The
resulting
protected
2×CH3
JH,P 18.4, J1%,2 7.85, H-1%), 3.97–4.61 (14H, m, 2×CH3CH2
2×PhCH2 H-2,3,4,5,6,6%), 7.12–7.38 (20H, Ar); lC 16.3
(CH3CH2OP), 16.4 (CH3CH2OP), 61.6 (CH3CH2OP, JC,P 6.54),
61.9 (CH3CH2OP, JC,P 6.54), 68.1 (C-6, JC,P 6.60), 71.6 (PhCH2),
71.8 (PhCH2), 79.0 (C-3 or 4), 81.7 (C-5, JC,P 8.40), 84.3 (C-3 or 4),
6
CH2OP), 2.15 (1H, dd, JH,P 18.9, J1,2 6.4, H-1), 2.20 (1H, dd,
6
OP,
phosphonophosphate 6 gave rise to doublets in its 31P
NMR spectrum at 13.4 and −16.1 ppm, both with a
24.5 Hz P–P coupling, confirming formation of the
phosphonophosphate moiety. The phenyl and benzyl
groups in 6 were removed by catalytic hydrogenation,
which for no obvious reason required the sequential use
of palladium and platinum catalysts for maximum
efficiency.
6
,
6
6
6
6
6
6
85.8 (C-2, JC,P 7.95), 120.0–129.7 (Ar), 137.3 (quat. Ar), 137.5
(quat. Ar), 150.4 (quat. Ar), 150.5 (quat. Ar); lP (decoupled) 26.1
(phosphonyl), −13.0 (phosphoryl). Lit. for dimethylphosphonate,9
lP 30.22, −11.63.
5-Phospho-a-D-arabinosyl-C-phosphonophosphate (pACpp)
The stereospecific synthesis of the pACpp was success-
lH (D2O) 2.25 (2H, dd, J1&1%,P 18.7, J1,1% 7.0, H-1,1%), 3.98 (2H, m,
H-6,6%), 4.00–4.21 (4H, m, H-2,3,4,5); lC 28.9 (C1), 60.9 (C6), 74.0
(C4), 75.6 (C2), 77.3 (C3 or C5), 78.1 (C3 or C5); lP (decoupled)
14.3 (phosphonyl), 0.7 (phosphoryl), −10.6 (phosphoryl anhydride).
NMR data in accord with literature values.9 ES-MS: calcd for
C6H15O13P3: 387.9725, found [M−H]− 386.9650.
fully completed in 11 steps from readily available 2,5-
anhydro-D-mannitol in 3.4% overall yield. For
comparison, McClard’s synthesis9 required only seven
steps, but necessitated a difficult separation of three
polar compounds at the final stage.

2234
P. McGurk et al. / Tetrahedron Letters 42 (2001) 2231–2234
Biological assays: Phosphonophosphate pACpp was
tested for its ability to inhibit incorporation of radiola-
bel from pRpp into organic soluble material (i.e. 5-P-
DpA/5-P-DpR and DpA/DpR) by membrane
preparations of both M. bovis (BCG strain) and M.
smegmatis using standard protocols.6 Results showed
less than 20% inhibition at 5 mM concentration, in
both the presence and absence of detergent.
References
1. Daffe, M.; Draper, P. Adv. Microb. Physiol. 1998, 39,
131.
2. Chatterjee, D. Curr. Opin. Chem. Biol. 1997, 1, 579.
3. McNeil, M. In Genetics of Bacterial Polysaccharides;
Goldberg, J. B., Ed.; CRC Press: Boca Raton, 1999; pp.
207–223.
4. Wolucka, B. A.; McNeil, M.; de Hoffmann, E.; Choj-
nacki, T.; Brennan, P. J. J. Biol. Chem. 1994, 269, 23328.
5. Lee, R. E.; Mikusova, K.; Brennan, P. J.; Besra, G. S. J.
Am. Chem. Soc. 1995, 117, 11829.
In summary, we have developed a stereospecific synthe-
sis of 5-phospho- -arabinosyl-a-C-phosphonophos-
D
phate (pACpp), a stable analogue of the putative myco-
bacterial cell wall biosynthetic intermediate 5-phospho-
-arabinosylpyrophosphate (pApp), and shown that it
6. Scherman, M. S.; Kalbebournonville, L.; Bush, D.; Deng,
L. Y.; McNeil, M. J. Biol. Chem. 1996, 271, 29652.
7. McGurk, P. Ph.D. thesis, University of St Andrews,
2000.
8. Where comparison could be made, compounds prepared
in this study possessed spectral characteristics consistent
with those reported by McClard for the corresponding
dimethylphosphonates.9 Selected data for key compounds
are also given in the footnote.
D
is incapable of interrupting key steps in mycobacterial
arabinan formation in vitro. One might choose to inter-
pret these results as indicating that pApp is not in fact
an intermediate in mycobacterial cell wall biosynthesis.
However, the acidity of a phosphonophosphate does
not necessarily accurately reflect that of the pyrophos-
phate,15 which might contribute to the lack of activity
observed. Alternatively, the anomeric oxygen atom of
pApp might be important for epimerase recognition.
The enzymes responsible for conversion of pRpp to
DpA are membrane bound;6 substrate channelling aris-
ing from protein complex formation might also account
for the inability of pACpp to interfere with DpA
biosynthesis.
9. McClard, R. W.; Witte, J. F. Bioorg. Chem. 1990, 18,
165.
10. Horton, D.; Philips, K. D. Methods Carbohydr. Chem.
1974, 7, 68.
11. Piper, I. M.; MacLean, D. B.; Kvarnstrom, I.; Szarek, W.
A. Can. J. Chem. 1984, 62, 2721.
12. Compound 2 has previously been synthesised by other
routes. See: Kaye, A.; Neidle, S.; Reese, C. B. Tetra-
hedron Lett. 1988, 29, 1841; Persky, R.; Albeck, A. J.
Org. Chem. 2000, 65, 5632 and references cited therein.
13. Palcic, M. M.; Heerze, L. D.; Srivastava, O. P.; Hinds-
gaul, O. J. Biol. Chem. 1989, 264, 17174.
Acknowledgements
This work was supported by US NIH/NIAID AIDS-
FIRCA grant R03 TW00938-01 and NIH grant AI
33706, the UK BBSRC (studentship to P.McG.) and
the Ohio State University. We thank the EPSRC Mass
Spectrometry Service at the University of Wales,
Swansea for invaluable support.
14. Gorenstein, D. G. Phosphorus-31 NMR: Principles and
Applications; Academic Press: San Diego, 1984.
15. For a phosphonate the second ionisation can often be
more than one pH unit higher than that of the corre-
sponding phosphate. See: Blackburn, G. M. Chem. Ind.
(London), 1981, 134.
.