Hydrolytically stable arabinofuranoside analogs for the synthesis of arabinosyltransferase inhibitors

Article abstract of DOI:10.1016/j.tetlet.2005.12.029

The first members of two new families of arabinosyltransferase inhibitors, derived from previously reported hybrid compounds covalently associating an iminoalditol with an α-d-arabinofuranoside, have been prepared. In place of the arabinofuranoside moiety

Full text of DOI:10.1016/j.tetlet.2005.12.029

Tetrahedron
Letters
Tetrahedron Letters 47 (2006) 1113–1116
Hydrolytically stable arabinofuranoside analogs for the synthesis
of arabinosyltransferase inhibitors
*
´
Manon Chaumontet, Valerie Pons, Karine Marotte and Jacques Prandi
Institut de Pharmacologie et de Biologie Structurale du CNRS, UMR 5089, 205 route de Narbonne, 31077 Toulouse Cedex 04, France
Received 8 November 2005; revised 6 December 2005; accepted 6 December 2005
Available online 27 December 2005
Abstract—The first members of two new families of arabinosyltransferase inhibitors, derived from previously reported hybrid
compounds covalently associating an iminoalditol with an a-D-arabinofuranoside, have been prepared. In place of the arabinofurano-
side moiety, they incorporate in their structure a suitably substituted tetrahydrofuran (C-glycoside family) or a cyclopentane (carba-
sugar family) for mimicking the a-D-arabinofuranoside ring.
Ó 2005 Elsevier Ltd. All rights reserved.
1. Introduction
wall have been identified many years ago as good targets
for the development of new antimycobacterial drugs8
and the field has been active during recent years.9,10
Despite the availability of numerous antibiotics effective
against Mycobacterium tuberculosis, the causative agent
of human tuberculosis, the disease remains a major
world-wide health problem, taking millions of lives
annually. New therapeutic strategies for fighting the dis-
ease are urgently needed because of emergence and slow
increase of infections by multi-drug resistant strains of
the bacillus. Biosynthesis of the mycobacterial cell wall
is a validated target for the development of new thera-
peutic agents against tuberculosis.1 It is already the site
of action of three major antituberculosis drugs: isoniazid
and ethionamide block the mycolic acid synthesis2,3
while ethambutol interferes with the biosynthesis of
arabinogalactan, a major structural element of the cell
wall.4–7 Glycosyltransferases of the mycobacterial cell
We recently described a new family of hybrid molecules,
which showed good arabinosyltransferase inhibitory
properties11 and very promising in vitro activity against
M. tuberculosis.12 These molecules were built from a
1,4-dideoxy-1,4-anhydro-D-pentitol (five-membered aza-
sugar) covalently linked to a small oligo-a-D-arabino-
furanoside, acting as an addressor for the iminosugar
(see compound 1, Fig. 1).
A drawback of these compounds is the presence of
an acid-sensitive arabinofuranosidic linkage in their
structure, which shortens their half-life in biological
medium and impairs their potential use as antibiotics.
OH
HO
OH
HO
OH
O
O
OR
N
N
N
OR
OR
HO
OH
OH
OH
R = n-C16H33
1
2
3
C-glycoside
Carba-sugar
Figure 1. Arabinosyltransferase inhibitor 1 and its analogs: C-glycoside 2 and carba-sugar 3.
Keywords: Inhibitor; Arabinosyltransferase; Mycobacteria; C-Glycoside; Carba-sugar.
*
Corresponding author. Tel.: +33 (0)5 61 17 54 85; fax: +33 (0)5 61 17 59 94; e-mail: Jacques.Prandi@ipbs.fr
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.12.029

1114
M. Chaumontet et al. / Tetrahedron Letters 47 (2006) 1113–1116
O
O
O
OR
OR
OC16H33
RO
a
RO
BnO
HO
BnO
b
d
RO
OR
OBn
OBn
4 R = H
6 R = Ac
7 R = H
8
c
5 R = Bn
6'
OH
1
O
O
2'
6
OC16H33
O
e
f
OC16H33
N
1'
2
7
BnO
5'
OBn
RO
OR
10 R = Bn
9
g
2
R = H
Scheme 1. Reagents and conditions: (a) 6.5 equiv BnBr, 10.0 equiv NaH, DMF, 4 h, 77%; (b) TFA, Ac2O, room temp, 1.5 h; (c) MeONa, MeOH,
room temp, 4 h, 77%, two steps; (d) 1.0 equiv C16H33I, 1.3 equiv NaH, DMF, room temp, 18 h, 42%; (e) (COCl)2, DMSO, NEt3, ꢀ78 °C, 0.5 h; (f )
2.0 equiv prolinol, 6.0 equiv NaBH3CN, MeOH, AcOH, reflux, 1.5 h, 48%, two steps; (g) H2, Pd/C, 0.5 M HCl in MeOH, room temp, 4 h, 87%.
Replacement of this glycosidic bond with a more stable
ether bond should give inhibitors with improved stabil-
ity and, possibly, greater potency. Depending on which
oxygen atom of the glycosidic bond is replaced by a car-
bon atom, exo-cyclic or endo-cyclic oxygen, two families
of hydrolytically stable compounds are obtained, a
tetrahydrofuran family (or C-glycosides) and a cyclo-
pentane family (or carba-sugars). The conformational
analysis of these arabinofuranoside analogs had been
carried out13,14 and showed them to be of possible
potential as arabinofuranoside analogs. There has been
recent interest in the synthesis of oligo-C-arabinofura-
nosidic compounds and the first members of this series
have been prepared.15,16
tures below 0 °C. The reaction had to be run at room
temperature and alcohol 8 was isolated in only 40%
yield, together with the recovered starting material and
some dialkylated product. Swern oxidation of the alco-
hol, immediately followed by reductive amination of
aldehyde 9 with (S)-prolinol and NaBH3CN in acidic
methanol19,20 gave the coupling product 10 in 48% yield
(two steps). Catalytic hydrogenation of 10 (H2, Pd/C,
MeOH, HCl) afforded the targeted compound 2 in
87% yield, completing the synthesis of the first C-glyco-
side analog of 1.
3. Synthesis of cyclopentane 3
We now report the preparation of two hydrolytically
stable analogs of 1, a simple but representative example
of our previously described arabinosyltransferase inhibi-
tors: C-glycoside 2 and carba-arabinofuranoside 3,
where the 1,4-dideoxy-1,4-anhydro-D-pentitol moiety is
the commercially available (S)-2-pyrrolidinemethanol
((S)-prolinol).
The elaboration of the carbocyclic analog of the a-D-
arabinofuranoside ring relies on the cobalt-mediated
oxygenative radical cyclization of suitably protected
polyhydroxylated 6-iodohex-1-enitols-1-hexenitols to
cyclopentane methanols.21,22 As we wanted the configu-
rations of the various substituents of cyclopentane 3
(two hydroxyls, an alkoxy, and the aminomethyl group)
to be identical to those in a-D-arabinofuranoside, the
hex-1-enitol to be chosen for the radical cyclization
had to be derived from any D-glucopyranoside. Alkyl-
2. Synthesis of C-glycoside 2
The most simple C-glycoside of a-D-arabinofuranoside
is 2,5-anhydro-D-mannitol 4 (Scheme 1), which is readily
available from D-glucosamine by a two-step proce-
dure.17 So, 4 was benzylated in 73% yield with benzyl
bromide and sodium hydride in DMF to afford the tetra-
benzyl derivative 5 and the primary benzyl groups of 5
were then selectively acetolyzed with trifluoroacetic acid
in acetic anhydride to give diacetate 6.18 Crude 6 was
taken up in basic methanol (MeONa, MeOH) to remove
the ester groups and diol 7 was isolated in 77% yield
for the two steps. Despite the report15 of the successful
selective monobenzylation of diol 7 in 84% yield, mono-
alkylation of 7 with 1-iodohexadecane was found to be
difficult to control, mainly because of the very low solu-
bility of 1-iodohexadecane in DMF (or THF) at tempera-
20
Selected data for 2 (see Scheme 1 for numbering). ½aꢁD +18.5 (c 0.7,
methanol). 1H NMR (500 MHz);
d (CD3OD): 0.93 (t, 3H,
3J = 7.1 Hz, CH3 hexadecyl), 1.24–1.46 (m, 26H, CH2 hexadecyl),
1.60 (m, 2H, O–CH2–CH2 hexadecyl), 1.90 (m, 1H, 30-H), 2.10 (m,
1H, 40-H), 2.09 (m, 1H, 40-H), 2.20 (m, 1H, 30-H), 3.21–3.35 (m, 2H,
50-H, 6-H), 3.53 (m, 2H, O–CH2 hexadecyl), 3.57–3.67 (m, 4H, 7-H,
20-H, 7-H, 6-H), 3.72 (m, 1H, 50-H), 3.75 (dd, 1H, J6 ;6 ¼ 12:2 Hz,
2
0
0
J6 ;5 ¼ 5:6 Hz, 60-H), 2.85 (dd, 1H, J4,5 = 5.8 Hz, 3J4,3 = 5.1 Hz, 4-
3
3
0
0
2
3
0
ꢀ
0
0
0
0
H), 3.90 (dd, 1H, J6 ;6 ¼ 12:2 Hz, J6 ;5 ¼ 4:0; Hz, 6 -H), 3.98–4.05
3
3
(m, 2H, 2-H, 3-H), 4.19 (ddd, 1H, J5,6 = 9.9 Hz, J5,4 = 5.8 Hz,
3J5,6 = 2.1 Hz, 5-H) ppm. 13C NMR (125.8 MHz); d (CD3OD): 13.1
(CH3 hexadecyl), 22.3 (40-C), 25.6 (30-C), 55.8 (50-C), 56.8 (6-C), 59.7
(60-C), 69.0 (20-C), 70.8 (7-C), 71.4 (O–CH2 hexadecyl), 77.2 (2-C),
79.5 (4-C, 5-C), 83.2 (3-C) ppm. HRMS (FAB): C27H53NO5 calcd
472.4002, found 472.4006.

M. Chaumontet et al. / Tetrahedron Letters 47 (2006) 1113–1116
1115
ation of the stannylene acetal of methyl 4,6-O-benzyl-
idene-a-D-glucopyranoside 1123 with 1-iodohexadecane
and dibutyltin oxide24 was very difficult in toluene under
tetrabutylammonium bromide activation, only 16% of
the 2-O-alkylated product 12 were isolated after 24 h
at reflux temperature, together with an equal quantity
of the 3-O-alkylated isomer and considerable amounts
of recovered 11. However, it was found that alkylation
of the stannylene acetal of 11 could be smoothly
obtained at 60 °C in DMF under cesium fluoride activa-
tion.25 The conversion was complete after 14 h and 12
was obtained in 38% yield after chromatography. Once
again, no regioselectivity was observed for the alkyl-
ation and an equal amount of the 3-O-alkylated isomer
of 12 was also isolated from the reaction.
quantitative yield. Selective iodination of the primary
position of 14 was achieved under GarregÕs conditions26
with triphenylphosphine and 1.3 equiv of iodine in
refluxing toluene, a 74% yield of 15 was obtained. A
tert-butyldimethylsilyl ether was finally introduced
under classical conditions (TBDMSOTf, pyridine,
CH2Cl2) on the hydroxyl group of 15 and 16 was
isolated in a quantitative yield. The protection of this
hydroxyl group by the bulky silyl ether group was
chosen to decrease, and possibly abolish the nucleophi-
licity of this oxygen atom during iodination of hexenitol
18 (to 19, Scheme 2) and avoid competitive tetrahydro-
furan formation during this step. With all protecting
groups secured, 6-deoxy-6-iodo-glucopyranoside 16
was reductively opened with zinc powder in refluxing
ethanol27,28 and, after a rapid work-up of the reaction,
the sensitive intermediate aldehyde 17 was immediately
reduced to hexenitol 18 with methanolic sodium borohy-
dride in fair yield (47% for the two steps). Introduction
of the iodine atom, the radical precursor for the radical
The 3-hydroxyl group of 12 was then protected as a benz-
yl ether (BnBr, NaH, DMF/THF, 72%) before acidic
removal of the 4,6-O-benzylidene group (catalytic cam-
phorsulfonic acid in refluxing MeOH) to afford 14 in
X
O
O
Ph
Ph
O
O
a
O
O
O
c
RO
2
HO
R O
1
BnO
HO
OMe
R O
OMe
C16H33O
OMe
11
12 R1= C16H33, R2 = H
13 R1= C16H33, R2 = Bn
14 R = H, X = OH
15 R = H, X = I
d
e
b
16 R = TBDMS, X = I
HO
TBDMSO
BnO
TBDMSO
X
g
i
f
O
OC16H33
BnO
TBDMSO
OC16H33
OC16H33
OBn
20
17
18 X = OH
19 X = I
h
6'
OH
O
H
2'
6
5
j
k
N
4
1'
2
OC16H33
OC16H33
1
5'
R O
TBDMSO
1
OR
OBn
21
22 R1 = Bn, R2 = TBDMS
3 R1 = R2 = H
l
Scheme 2. Reagents and conditions: (a) 1.0 equiv Bu2SnO, toluene, reflux, then 1.0 equiv CsF, 1.4 equiv C16H33I, DMF, 60 °C, 14 h; (b) 1.5 equiv
BnBr, 2.0 equiv NaH, DMF/THF 4/1 v/v, room temp, 3 h, 72%; (c) 0.04 equiv CSA, MeOH, reflux, 1 h, quant; (d) 1.2 equiv PPh3, 3.0 equiv
imidazole, 1.3 equiv I2, toluene, reflux, 0.5 h, 74%; (e) 3.0 equiv TBDMSOTf, pyridine, CH2Cl2, room temp, 2.5 h, quant.; (f ) 10.0 equiv Zn,
10.0 equiv pyridine, EtOH, reflux, 4 h; (g) NaBH4, EtOH, room temp, 0.5 h, 47%, two steps; (h) 1.2 equiv PPh3, 3.0 equiv imidazole, 1.3 equiv I2,
toluene, reflux, 0.5 h, 89%; (i) 4.0 equiv NaBH4, 3.7 equiv NaOH, 0.08 equiv Co(salen), air, 95% EtOH, 40 °C, 3.5 h, 78%; (j) 3.0 equiv PCC,
˚
5.0 equiv AcONa, 3 A MS, CH2Cl2, room temp, 2 h; (k) 2.0 equiv (S)-prolinol, 6.0 equiv NaBH3CN, AcOH, MeOH, reflux, 2 h, 62%, two steps; (l)
H2, Pd(OH)2/C, 0.2 M HCl in MeOH, room temp, overnight, 82%.

1116
M. Chaumontet et al. / Tetrahedron Letters 47 (2006) 1113–1116
cyclization, on the primary position of 18 was readily
accomplished, again under GarregÕs conditions as above
and gave 6-iodo-hex-1-enitol 19 in high yield (89%).
Thanks to the silyl ether group on the 3-oxygen atom
of 18, and in contrast with previously reported examples
where a more nucleophilic benzyl ether was at this posi-
tion,22,29 only very minor amounts (<2%) of cyclized tet-
rahydrofuran-type material could be isolated from the
reaction mixture. Elaboration of the cyclopentanemeth-
anol was then easily carried out using the oxygenative
radical cyclization catalyzed by Co(salen) complex.
Compound 19 was dissolved in basic 95% ethanol con-
taining NaBH4 and the cobalt catalyst at 40 °C. Air
was continuously bubbled through the medium with a
small pump and 19 cyclized smoothly to a mixture of
cyclopentanemethanols in 2.5 h. Yield was 78% and
the observed diastereoselectivity for the formation of
the new C-4 asymmetric center of the cyclopentane ring
was found to be 5.5/1. This mixture of diastereoisomers
could be chromatographically resolved on silica and the
absolute configuration of the major isomer, 20, was
established to be 4R by comparison of the NMR data
of 20 with those of previously reported examples,22 this
configuration was expected from RajanBabuÕs stereo-
chemical rules for these radical cyclizations.30 The end
of the synthesis was straightforward, oxidation of alco-
hol 20 was done with pyridinium chlorochromate (PCC,
AcONa, CH2Cl2) and gave aldehyde 21, which was
immediately engaged in the reductive amination process
with (S)-prolinol, as described above for the synthesis of
10. Coupling product 22 was isolated in 62% yield for
the two steps from alcohol 20. Final deprotection was
carried out by hydrogenolysis of 22 (H2, Pd(OH)2/C)
in acidic methanol (HCl) at room temperature. This al-
lowed smooth removal of both benzyl and TBDMS
ether protecting groups in one step, and afforded a
82% yield of the final compound 3.à
ether bond: a C-glycoside and a carba-sugar, respec-
tively. The present synthetic routes are general and
allow the preparation of two complete sets of potential
new inhibitors with improved biological stability. Fur-
thermore, in the preliminary in vitro testing on cultures
of M. tuberculosis, C-glycoside 2 showed the same level
of microbicidal activity as the original inhibitor 1, show-
ing the importance of this family of compounds. Results
for 3 and other members of these families will be
reported in the very near future.
References and notes
1. Chatterjee, D. Curr. Opin. Chem. Biol. 1997, 1, 579–588.
2. Takayama, K. Ann. N.Y. Acad. Sci. 1974, 235, 426–438.
3. Zhang, Y.; Heym, B.; Allen, B.; Young, D.; Cole, S. T.
Nature 1992, 358, 591–593.
4. Takayama, K.; Kilburn, J. O. Antimicrob. Agents Chemo-
ther. 1989, 33, 1493–1499.
´
5. Mikusˇova, K.; Slayden, R. A.; Besra, G. S.; Brennan, P. J.
Antimicrob. Agents Chemother. 1995, 39, 2484–2489.
´
6. Lee, R. E.; Mikusˇova, K.; Brennan, P. J.; Besra, G. S. J.
Am. Chem. Soc. 1995, 117, 11829–11832.
7. Khoo, K. H.; Douglas, E.; Azadi, P.; Inamine, J. M.;
Besra, G. S.; Mikusova, K.; Brennan, P. J.; Chatterjee, D.
J. Biol. Chem. 1996, 271, 28682–28690.
8. Maddry, J. A.; Suling, W. J.; Reynolds, R. C. Res.
Microbiol. 1996, 147, 106–121.
9. Lowary, T. L. Mini Rev. Med. Chem. 2003, 3, 689–702.
10. Lee, R. E.; Protopopova, M.; Crooks, E.; Slayden, R. A.;
Terrot, M.; Barry, C. E., III. J. Comb. Chem. 2003, 5, 172–
187.
´
11. Marotte, K.; Ayad, T.; Genisson, Y.; Besra, G. S.; Baltas,
M.; Prandi, J. Eur. J. Org. Chem. 2003, 2557–2565.
12. Marotte, K.; Prandi, J., unpublished results.
13. OÕLeary, D. J.; Kishi, Y. J. Org. Chem. 1994, 59, 6629–
6636.
14. Callam, C. S.; Lowary, T. L. J. Org. Chem. 2001, 66,
8961–8972.
15. Gurjar, M. K.; Nagaprasad, R.; Ramana, C. V. Tetra-
hedron Lett. 2002, 43, 7557–7559.
16. Dondoni, A.; Marra, A. Tetrahedron Lett. 2003, 44, 4067–
4071.
4. Conclusion
The first members of two new families of potential ara-
binosyltransferase inhibitors have been prepared. They
incorporate two different a-D-arabinofuranoside ana-
logs in their structure, where the acido-labile glycosidic
bond had been replaced by an hydrolytically resistant
17. Horton, D.; Philips, K. D. Carbohydr. Res. 1973, 30, 367–
374.
18. Eby, R.; Sondheimer, S. J.; Schuerch, C. Carbohydr. Res.
1979, 73, 273–276.
19. Borch, R. F.; Bernstein, M. D.; Durst, H. D. J. Am. Chem.
Soc. 1971, 93, 2897–2904.
20. Lane, C. F. Synthesis 1975, 135–146.
´
´
21. Desire, J.; Prandi, J. Tetrahedron Lett. 1997, 38, 6189–
6192.
20
à Selected data for 3 (see Scheme 2 for numbering). ½aꢁD +12 (c 0.3,
´
´
22. Desire, J.; Prandi, J. Eur. J. Org. Chem. 2000, 3075–
methanol). 1H NMR (250 MHz); d (4/1 CDCl3/CD3OD): 0.90 (t, 3H,
3J = 7.1 Hz, CH3 hexadecyl), 1.22–1.33 (m, 26H, CH2 hexadecyl),
1.56 (m, 2H, O–CH2–CH2 hexadecyl), 1.73 (ddd, 1H, 2J5,5 = 13.5 Hz,
3084.
23. Evans, M. E. Carbohydr. Res. 1972, 21, 473–475.
24. David, S.; Thieffry, A.; Veyrieres, A. J. Chem. Soc., Perkin
`
3J5,4 = 10.5 Hz, J5,1 = 8.0 Hz, 5-H), 1.88 (ddd, 1H, J5,5 = 13.5 Hz,
3
2
Trans. 1 1981, 1796–1801.
3J5,1 = 8.5 Hz, J5,4 = 3.0 Hz, 5-H), 1.90–2.23 (m, 4H, 30-H, 40-H),
3
25. Nagashima, N.; Ohno, M. Chem. Lett. 1987, 141–144.
26. Garegg, P. J.; Samuelsson, B. J. Chem. Soc., Perkin Trans.
1 1980, 2866–2869.
27. Bernet, B.; Vasella, A. Helv. Chim. Acta 1979, 62, 1990–
2016.
2.37 (m, 1H, 4-H), 3.01–3.13 (m, 2H, 20H, 6-H), 3.43 (dt, 1H,
2J = 9.0 Hz, 3J = 6.5 Hz, O–CH2 hexadecyl), 3.46–3.60 (m, 3H, 50-H,
6-H, O–CH2 hexadecyl), 3.66 (m, 1H, 1-H), 3.70 (dd, 1H,
3J3,4 = 9.0 Hz, J3,2 = 7.5 Hz, 3-H), 3.80–3.89 (m, 3H, 2-H, 50-H,
3
60-H), 3.95 (dd, 1H, 2J6 ;6 ¼ 12:0 Hz, 3J6 ;2 ¼ 3:5 Hz, 60-H) ppm. 13C
NMR (62.9 MHz); d (4/1 CDCl3/CD3OD): 13.8 (CH3 hexadecyl),
22.4, 22.5, 25.9, 26.0, 29.1, 29.3, 29.4, 29.4, 29.5, 29.6, 30.5, 31.3, 31.7,
54.7 (50C), 59.4 (60-C), 69.4 (20-C), 80.5, 81.5, 82.8 (1-C, 2-C, 3-C)
ppm.
0
0
0
0
28. Bernet, B.; Vasella, A. Helv. Chim. Acta 1984, 67, 1328–
1347.
29. Martin, O. R.; Yang, F.; Xie, F. Tetrahedron Lett. 1995,
36, 47–50.
30. Rajanbabu, T. V. Acc. Chem. Res. 1991, 24, 139–145.