Confirmation of the connectivity of 4,8,12,16,20-pentamethylpentacosylphoshoryl β-D-mannopyranoside, an unusual β-mannosyl phosphoisoprenoid from Mycobacterium avium, through synthesis

Article abstract of DOI:10.1021/ja0123958

The synthesis of the title glycolipid is reported. Comparison of the electrospray and high-energy collision-induced dissociation mass spectra of the synthetic material with those reported for the isolate confirm the structure of this unusual antigenic substance with its modified isoprenoid chain.

Full text of DOI:10.1021/ja0123958

Published on Web 02/16/2002
Confirmation of the Connectivity of
4,8,12,16,20-Pentamethylpentacosylphoshoryl
â-D-Mannopyranoside, an Unusual â-Mannosyl
Phosphoisoprenoid from Mycobacterium avium, through
Synthesis
David Crich* and Vadim Dudkin
Contribution from the Department of Chemistry, UniVersity of Illinois at Chicago, 845 West
Taylor Street, Chicago, Illinois 60607-7061
Received October 19, 2001
Abstract: The synthesis of the title glycolipid is reported. Comparison of the electrospray and high-energy
collision-induced dissociation mass spectra of the synthetic material with those reported for the isolate
confirm the structure of this unusual antigenic substance with its modified isoprenoid chain.
Introduction
the synthesis of all but the simplest â-mannosyl phosphates had
previously eluded other workers in the field.6,7 Numerous factors
Two unusual â-mannosyl phosphoisoprenoids were recently
isolated from Mycobacterium aVium and Mycobacterium tu-
bercolosis on the basis of their recognition by a CD1 c-restricted,
mycobacterial specific T-cell line.1 They were assigned the
structures 1 and 2, respectively, on the basis of degradation and
mass spectrometric studies.1 It was suggested1 that 1 and 2
follow a general mechanism for lipid antigen presentation by
CD12 in which the isoprenoid chain is bound within a
hydrophobic groove of the CD1 protein3 and the exposed
hydrophilic component is recognized by specific T-cell recep-
tors. The T-cell response was shown to be inversely proportional
to the length of the phosphoisoprenoid chain, preferring a C35
derived system over one of C55 and not recognizing a C95
dolichol-based â-mannosyl phosphoisoprenoid.1 Additionally,
it was found that full saturation of the aliphatic chain was
required and that a glucose, as opposed to a mannose, headgroup
was not recognized.1 Biosynthetically, 1 and 2 are interesting
because of their modified isoprenoid chains, which deviate from
the isoprene rule by incorporation of additional methylene
groups, at both ends of the chain. Chemically, they are of interest
because of the considerable challenge presented by the synthesis
of the â-mannosyl phosphate, surely one of the more significant
obstacles presented by nature given the already considerable
difficulty of synthesizing â-mannosides themselves.4,5 Indeed,
therefore combined to make 1 and 2 viable and challenging
targets with which to test the â-mannosylation protocols
developed in this laboratory.8,9 Accomplishment of this endeavor
would also point the way toward syntheses of other biologically
relevant â-mannosyl phosphates such as the partly saturated
â-mannosyl heptaprenyl phosphate from Mycobacterium smeg-
matis,10 that serves as a carrier of mycolic acid, and the
â-mannosyl decaprenyl phosphate from the same organism,11
thought to be an intermediate in R-1f6 linked manno-
oligosaccharide biosynthesis. Here, we report full details of our
synthesis12 of a dihydrophytyl model for 1 and the successful
synthesis of 1 itself, including that of its unusual, modified
isoprenoid chain in the form of a stereorandom mixture of
isomers.
* To whom correspondence should be addressed.
(1) Moody, D. B.; Ulrichs, T.; Muhlecker, W.; Young, D. C.; Gurcha, S. S.;
Grant, E.; Rosat, J.-P.; Brenner, M. B.; Costello, C. E.; Besra, G. S.; Porcelli,
S. A. Nature 2000, 404, 884-888.
(2) Moody, D. B.; Desra, G. S.; Wilson, I. A.; Porcelli, S. A. Immunol. ReV.
1999, 172, 285-296.
Results and Discussion
(3) Zeng, Z.-H.; Castano, A. R.; Segelke, B. W.; Stura, E. A.; Peterson, P. A.;
Wilson, I. A. Science 1997, 277, 339-345.
The synthetic effort required for the preparation of the
isoprenoid chain of 1 and the obviously difficult nature of any
(4) Barresi, F.; Hindsgaul, O. In Modern Methods in Carbohydrate Synthesis;
Khan, S. H., O’Niell, R. A., Eds.; Harwood Academic Publishers:
Amsterdam, 1996; pp 251-276.
(5) Pozsgay, V. In Carbohydrates in Chemistry and Biology; Ernst, B., Hart,
G. W., Sinay¨, P., Eds.; Wiley-VCH: Weinheim, 2000; Vol. 1, pp 319-
343.
(6) Sabesan, S.; Neira, S. Carbohydr. Res. 1992, 223, 169-185.
(7) Schmidt, R. R.; Stumpp, M. Liebigs Ann. Chem. 1984, 680-691.
(8) Crich, D.; Sun, S. Tetrahedron 1998, 54, 8321-8348.
9
10.1021/ja0123958 CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. VOL. 124, NO. 10, 2002 2263

A R T I C L E S
Crich and Dudkin
Scheme 1
Scheme 2
underwent rapid isomerization to the R-anomer simply on
standing in deuteriochloroform and could not be isolated.7 We
reasoned that the poor diastereoselectivity observed was a
function of the low nucleophilicity of 8 and that this might be
countermanded by moving to a less polar solvent. This would
enhance the stability of the triflate 10 and so suppress R-selec-
tive, dissociative mechanisms in favor of â-selective, associative
ones. This argument holds whether the nucleophilic displace-
ment of triflate from 10 is a true SN2 reaction or, as is quite
possible, proceeds via attack on a contact ion pair that is in
dynamic equilibrium with 10. In the event, in toluene at -78
°C, the reaction was completely selective and afforded only 12
in 56% isolated yield (Scheme 2) with the mass balance
consisting mainly of 2,3-di-O-benzyl-4,6-O-benzylidene-D-
mannopyranose, that is, the hydrolysis product of 10. We first
attempted deprotection of 12 by hydrogenolysis but, under all
conditions assayed, noted substantial hydrolysis and the forma-
tion of D-mannopyranose. We therefore turned to the Birch
reduction and found that exposure of 12 to sodium in liquid
ammonia, followed by quenching with ammonium chloride,
minimized this problem and enabled the isolation of 3, as its
sodium salt, in 92% yield. Importantly, no anomerization was
observed during the deprotection process and 3 was isolated in
the form of a single diastereomer. Its anomeric configuration
was confirmed by the NOE correlation of its anomeric hydrogen
synthesis of â-mannosyl phosphates dictated that we begin our
study with a model investigation. Accordingly, the phytanyl
â-mannosyl phosphate 3 was selected as a first target. Toward
this end commercial phytol, a mixture of isomers, was reduced
over Adam’s catalyst to give phytanol 4 quantitatively. Benzyl
2-cyanoethyl N,N-diisopropylphosphoramidite 5 was prepared
in 94% yield from commercial 2-cyanoethyl N,N-diisopropyl-
chlorophosphoramidite and benzyl alcohol in the presence of
Hunig’s base and coupled with 4 in the presence of tetrazole
giving 6, which was immediately oxidized with tert-butyl
hydroperoxide to provide 7 in 96% overall yield from 4 (Scheme
1). Treatment of 7 with tetrabutylammonium hydroxide in a
dichloromethane/water biphasic system then afforded the salt
8 quantitatively (Scheme 1).
The R-mannosyl sulfoxide 9,8 a single diastereomer,13 was
converted at -78 °C to the triflate 10,14 by treatment of a
dichloromethane solution with triflic anhydride in the presence
of 2,6-di-tert-butyl-4-methylpyridine (DTBMP). Two equiva-
lents of salt 8 were added and the resulting mixture was stirred
for several hours before quenching, extraction, and chromato-
graphic separation. In this manner, the R- and â-mannosyl
phosphates 11 and 12 were isolated in 63 and 11% yields,
respectively (Scheme 2). Both anomers were approximately 1/1
mixtures of diastereomers at phosphorus and, in 12, these could
be separated even if the configuration was not assigned. The
anomeric configuration of 11 and 12 was assigned on the basis
of the chemical shifts of the mannose H5 signals, especially
that of the â-anomer 12 (δ 3.41 and 3.44 for the two
diastereomers at P), which is diagnostic of configuration in 4,6-
O-benzylidene protected mannopyranosides.8 Although the
diastereoselectivity of this process was disappointing, the
isolation of 12, following chromatography on silica gel, was
viewed as very encouraging given that Schmidt and co-workers
had reported that a related dibenzyl â-mannosyl phosphate
1
to both H3 and H5 as well as by its anomeric JCH coupling15
of 157.6 Hz.12
With the essential methodology established, we turned to the
preparation of the branched C30 alcohol 13 required for the
synthesis of 1. As the structure of 1 had been deduced purely
from mass spectral considerations, no information was available
regarding the relative or absolute stereochemistry of the five
stereogenic centers in the isoprenoid chain.1,16 In the absence
of any stereochemical information or of a rational hypothesis
predicting a particular stereoisomer, a stereorandom synthesis
was designed with a view to expediency and the verification of
(15) Bock, K.; Pedersen, C. J. Chem. Soc., Perkin Trans. 2 1974, 293-297.
(16) The original report1 on the isolation, structural elucidation, and activity of
1 and 2 did not consider, being based solely on mass spectrometric evidence,
the relative and absolute stereochemistry of the five stereogenic centers in
their isoprenoid chains. It was further reported1 that a semisynthetic fully
saturated C35 â-mannosyl phosphoisoprenoid, obtained by transfer of
mannose from GDP mannose to a pure, naturally derived saturated
polyprenol with a mannosyl transferase, had a comparable effect on the
proliferation of human CD1 c-restricted T-cell lines to that of 1 and 2.
Given that the exact location of methylation (the first methyl of this model
compound being one methylene unit closer to the phosphate than in 1 and
2) does not have a significant effect on activity, it is unlikely that the
absolute and relative stereochemistry of the chain in 1 and 2 has a major
influence on the activity of these compounds.
(9) Crich, D. In Glycochemistry. Principles, Synthesis, and Applications; Wang,
P. G., Bertozzi, C. R., Eds.; Dekker: New York, 2001, pp 53-75.
(10) Besra, G. S.; Sievert, T.; Lee, R. E.; Slayden, R. A.; Brennan, P. J.;
Takayama, K. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 12735-12739.
(11) Yokoyama, K.; Ballou, C. E. J. Biol. Chem. 1989, 264, 21621-21628.
(12) Crich, D.; Dudkin, V. Org. Lett. 2000, 2, 3941-3943.
(13) Crich, D.; Mataka, J.; Sun, S.; Lam, K.-C.; Rheingold, A. R.; Wink, D. J.
J. Chem. Soc., Chem. Commun. 1998, 2763-2764.
(14) Crich, D.; Sun, S. J. Am. Chem. Soc. 1997, 119, 11217-11223.
9
2264 J. AM. CHEM. SOC. VOL. 124, NO. 10, 2002

Confirmation of the Connectivity
A R T I C L E S
Scheme 3
Scheme 4
the excursions from the isoprene rule at both ends of the chain.
To this end, geranyl acetate was converted to the distal aldehydo
ester 14 by the regioselective allylic hydroxylation.17,18 Standard
Wittig olefination installed the butyl group and gave ester 15
in 76% yield. Saponification and then hydrogenation gave
alcohols 16 and then 17 in 96% and 98% yields, respectively.
Mitsunobu reaction19 of 17 with 1-phenyltetrazole-5-thiol 18
as nucleophile afforded the sulfide 19 in 93% yield which, on
treatment with mcpba, gave 98% of the sulfone 20. Coupling
of 20 with a second aliquot of aldehyde, following deprotonation
with lithium hexamethyldisilazide, resulted in the formation of
ester 21 in 90% yield. This modified Julia sequence20-22 was
far superior in our hands to any alternative Wittig sequence.
Saponification and hydrogenation subsequently gave 22 and then
23 in 97% overall yield. The conversion of 23 to the sulfone
25 via sulfide 24 was conducted analogously to the preparation
of 20 from 17 (Scheme 3).
Scheme 5
The final six-carbon segment of 13 was obtained by conver-
sion of 2-methyl-δ-valerolactone (26)23 to the Weinreb amide24
27, benzylation to give 28 and, finally, DIBAL reduction25
giving the aldehyde 29 (Scheme 4). Coupling of 29 with sulfone
25 gave the C30 ether 30, which on exposure to hydrogen over
palladium on carbon underwent hydrogenation and hydro-
genolysis to afford 13 (Scheme 4).
Completion of the synthesis of 1 closely mimicked the
protocol established for the model 3. Thus, alcohol 13 was
converted to the phosphate 32, via the phosphite 31. Treatment
with tetrabutylammonium hydroxide gave the salt 33 which was
coupled in a highly â-selective manner with the mannosyl triflate
10 in toluene at -78 °C to give 34 in 49% yield, with no
indication of the R-anomer (Scheme 5). As with the model, the
mass balance in this coupling was provided by the pyranose
arising from the hydrolysis of triflate 10. Also in common with
(17) Barton, D. H. R.; Crich, D. Tetrahedron 1985, 41, 4359-4364.
(18) Bhalerao, U. T.; Plattner, J. J.; Rapoport, H. J. Am. Chem. Soc. 1970, 92,
3429-3433.
(19) Smith, A. B.; Wan, Z. J. Org. Chem. 2000, 65, 3738-3753.
(20) Blakemore, P. R.; Cole, W. J.; Kocienski, P. J.; Morley, A. Synlett 1998,
26-28.
(21) Baudin, J. B.; Hareau, G.; Julia, S. A.; Ruel, O. Bull. Soc. Chim. Fr. 1993,
336-357.
(22) Baudin, J. B.; Hareau, G.; Julia, S. A.; Ruel, O. Tetrahedron Lett. 1991,
1175-1178.
the model, the stereochemistry of 34 was initially determined
from the chemical shift of the mannose H5 resonance and
subsequently confirmed by NOE measurements and by the
(23) Oikawa, M.; Oikawa, H.; Ichihara, A. Tetrahedron 1995, 51, 6237-6254.
(24) Hamada, Y.; Yokokawa, F.; Kabeya, M.; Hatano, K.; Kurano, Y.; Shioiri,
T. Tetrahedron 1996, 52, 8297-8306.
(25) Marshall, J. A.; Johns, B. A. J. Org. Chem. 2000, 65, 1501-1510.
9
J. AM. CHEM. SOC. VOL. 124, NO. 10, 2002 2265

A R T I C L E S
Crich and Dudkin
magnitude of the anomeric 1JCH coupling constant. Finally, Birch
reduction of 34 gave the target molecule 1 in 93% isolated yield
with no detectable anomerization (Scheme 5). The anomeric
stereochemistry of 1 was fully confirmed by the NOE correlation
between the anomeric hydrogen and H’s 3 and 5 of the sugar
moiety.
sterochemistry, of the isoprenoid chain and the anomeric
stereochemistry of this highly unusual glycolipid is therefore
confirmed.
The protected â-mannosyl phosphates 12 and 34 described
here are configurationally stable and may be isolated by silica
gel chromatography, whereas Schmidt found closely related per-
O-benzyl-â-mannosyl phopshates to be unstable and very rapidly
equilibrated to the R-anomers in deuteriochloroform solution.7
This dichotomy, we believe, is another demonstration of the
torsionally disarming power28 of the 4,6-O-benzylidene protect-
ing group. In â-mannosylation by the sulfoxide method, this
protecting group is essential and functions by destabilizing the
anomeric oxacarbenium ion with respect to the covalent glycosyl
triflate.14 As anomerization of the glycosyl phosphate proceeds
via the same oxacarbenium ion, the configurational stability of
12 and 34 is readily understood.
The anomeric configuration of 1 and 2 and the unusual nature
of their modified isoprenoid chains were both originally
characterized by mass spectral fragmentation patterns.1 In
particular, in the negative ion spectrum a cross-ring fragmenta-
tion of the carbohydrate moiety, known to be characteristic of
cis-1,2-glycosyl phosphates,26 established the â-mannoside
stereochemistry, whereas a high-energy collision-induced remote
fragmentation pattern27 located the methylation sites on the lipid.
Mass spectrometric investigation of synthetic 1, using the same
array of techniques, reproduced the published spectra with a
high degree of fidelity. Thus, two cross-ring fragmentations (m/z
559.6 in the low-energy ESI CID spectrum and m/z 546 in the
high-energy FAB CID spectrum) and a dehydration (m/z 661.3
in the low-energy ESI CID spectrum) originally used to assign
the â-mannosyl phosphate configuration were nicely reproduced.
More importantly, given the unusual modified nature of the
isoprenoid chain, the remote fragmentation pattern of the
isoprene chain in the high-energy FAB CID spectrum was a
good match with the original. A feature of this spectrum is the
series of consecutive losses of 14 mass units with a jump at
every fifth position, indicating a methyl group on every fourth
carbon in an otherwise unbranched aliphatic chain. The sequence
of ions at m/z 662, 649, 635, 621, and 607 distinguish the remote
end of the chain, with its unbranched five carbon chain, from
that of a regular isoprenoid. The sequence m/z 327, 299, and
285, with the skip between 327 and 299, distinguish the
functionalized end of the chain from that of a regular C30
isoprenyl ester for which at peak at m/z 313 with a skip to m/z
285 would have been expected. The connectivity, if not the
In conclusion, we have described a synthesis of the â-man-
nosyl phosphoisoprenoid 1 from Mycobacterium aVium and, by
comparison of its mass spectral fragmentation patterns with
those of the original isolate, confirm the gross structure and
connectivity of this substance.
Acknowledgment. We thank the NIH (GM 57335) for
support of this work, Drs. S. Scheppele, J. A. Anderson (UIC,
Research Resources Center), and R. Van Breemen (UIC,
Department of Medicinal Chemistry) for assistance with mass
spectrometry, and JEOL USA for recording the high energy
CID MS/MS spectrum.
Supporting Information Available: Full experimental details
together with copies of 1H and 13C NMR spectra for compounds
1, 3, 7, 8, 11-13, 15, 19-21, 24, 25, 29, 32-34 and the mass
spectra of synthetic 1 (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA0123958
(26) Wolucka, B. A.; Rush, J. S.; Waechter, C. J.; Shibaev, V. N.; DeHoffman,
E. Anal. Biochem. 1998, 255, 244-251.
(27) Cheng, C.; Gross, M. L. Mass Spectrom. ReV. 2000, 19, 398-420.
(28) Andrews, C. W.; Rodebaugh, R.; Fraser-Reid, B. J. Org. Chem. 1996, 61,
5280-5289.
9
2266 J. AM. CHEM. SOC. VOL. 124, NO. 10, 2002