Synthesis of the nucleoside moiety of liposidomycins: Elucidation of the pharmacophore of this family of MraY inhibitors

Article abstract of DOI:10.1016/S0960-894X(00)00349-8

Tunicamycins (TCMs) and liposidomycins (LPMs) are naturally occurring inhibitors of the bacterial translocase (MraY). Based on structure-activity relationship (SAR) studies, a molecular model has been proposed for their inhibitory mechanism. This study points out the importance of the nucleoside moiety of liposidomycins in the inhibition of MraY. A simplified molecule (I) based on the liposidomycin core structure has been synthesised and tested on MraY. The compound displayed a moderate inhibitory activity (IC₅₀ = 50 μM). The validation of the molecular model was then performed by synthesising higher homologues of I, containing an additional stereocentre in the 5' position (XIV and XV). In agreement with the prediction, only the (S) isomer XV showed significant activity against MraY (IC₅₀ = 5 μM). (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00349-8

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1839±1843
Synthesis of the Nucleoside Moiety of Liposidomycins: Elucidation
of the Pharmacophore of this Family of MraY Inhibitors
C. Dini,a,* P. Collette,c N. Drochon,a J. C. Guillot,a G. Lemoine,b
P. Mauvaisc and J. Aszodia
aMedicinal Chemistry Department, Hoechst Marion Roussel/Aventis, 102 rte de Noisy, 93235 Romainville Cedex, France
bMolecular Modelling Department, Hoechst Marion Roussel/Aventis, 102 rte de Noisy, 93235 Romainville Cedex, France
cScreening Department, Hoechst Marion Roussel/Aventis, 102
rte de Noisy, 93235 Romainville Cedex, France
Received 4 April 2000; revised 16 June 2000; accepted 20 June 2000
AbstractÐTunicamycins (TCMs) and liposidomycins (LPMs) are naturally occurring inhibitors of the bacterial translocase
(MraY). Based on structure±activity relationship (SAR) studies, a molecular model has been proposed for their inhibitory
mechanism. This study points out the importance of the nucleoside moiety of liposidomycins in the inhibition of MraY. A simpli-
®ed molecule (I) based on the liposidomycin core structure has been synthesised and tested on MraY. The compound displayed a
moderate inhibitory activity (IC50=50 mM). The validation of the molecular model was then performed by synthesising higher
homologues of I, containing an additional stereocentre in the 50 position (XIV and XV). In agreement with the prediction, only the
(S) isomer XV showed signi®cant activity against MraY (IC50= 5 mM). # 2000 Elsevier Science Ltd. All rights reserved.
In the course of our investigations targeted at the dis-
covery of new antibacterial entities able to overcome the
problem of nosocomial bacterial resistance,1 we have
found translocase2 (MraY) of interest as a potential
target. This enzyme belongs to the bacterial peptidogly-
can biosynthetic pathway, which is speci®c to bacteria.3
In addition, it has been recently proved as essential for
the survival of bacteria,4 and displays the advantage of
having known naturally occurring inhibitors5 such as
tunicamycins (TCMs), liposidomycins (LPMs) and mur-
eidomycins (MRDs) (Fig. 1). TCMs are good inhibitors of
translocase but are not selective. LPMs and MRDs are
potent and selective inhibitors of MraY.6 8 The low level
or complete lack of activity against bacteria might be rela-
ted to their high hydrophilicity which would be incompa-
tible with a passive di€usion through the lipophilic
cytoplasmic membrane.
by a phospho-lipid unit (dodecaprenyl phosphate: C55-
P) to provide lipid I (Scheme 1).
The presence of uridine moieties in the inhibitor chemi-
cal structures suggests that they recognise the UMA5/
UMP binding site. Moreover, the presence of lipid
groups in TCMs and LPMs suggests their positioning in
the C55-P binding site.
Because of the high ¯exibility of their peptide side
chain, molecular modelling has predicted several low-
energy conformations for MRDs. Therefore, TCM and
LPM structures have been used for conformational
studies. The structure of TCMs has been fully char-
acterised. For LPMs, however, the con®guration of four
chiral centres remain four uncertain and chemical
synthesis has never been completed.9 16 Nevertheless,
NMR studies carried out by Isono's group suggest that
the con®gurations are S, S, R, S, for the 50, 200, 500 and 600
positions, respectively9,17 (Fig. 2). We used these results
for our model.
Inspection of their chemical structures suggests that
TCMs, LPMs and MRDs might be either substrates or
transition state analogues of the reaction catalysed by
MraY. This reaction consists of the replacement of a
UMP unit from UDP-MurNAc pentapeptide (UMA5)
By using the cv€ force ®eld18 and a conjugate gradient
method with Insight/Discover, the lowest energy con-
formations of both LPMs and TCMs could be super-
imposed on the nucleoside substrate (UMA5) of the
enzymatic reaction (Fig. 3).
*Corresponding author. Tel.: +33-1-49991-5480; fax: +33-1-4991-
5087; e-mail: christophe.dini@aventis.com
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00349-8

1840
C. Dini et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1839±1843
Figure 1.
Scheme 1.
Figure 2.
Information given by this model can be summarised as
follows: (a) Uridine moieties are superimposed for
TCMs, LPMs and UMA5. (b) The muramoyl residue of
UMA5 and the GlcNAc moiety of TCMs are overlaid
(rings A and B, respectively). (c) The ribosamine unit of
LPMs ®ts perfectly the N-acetyl galactosamine like
central unit of TCMs (rings D and C), even at the level
of stereochemistry, mimicking the diphospho unit (E) of
UMA5. (d) TCM and LPM lipid chains point in di€er-
ent directions. (e) The 50, 200 bond that bears the diaze-
panone ring (F) of LPMs points in the same direction as
the 50 hydroxyl of TCMs.
Figure 3.
available) with 2 using mercury(II) cyanide led to 4.
Methanolysis of 4 gave 5, which was reacted without
puri®cation with 2-methoxy propene in the presence of
PTSA to provide the bis-acetonide 6. Mesylation of the
remaining primary alcohol of 6 was performed with
methanesulfonyl chloride and TEA to provide 7. Sub-
stitution of the mesylate 7 with sodium azide in DMF at
70 ꢀC led to the azido compound 8. Reduction of the
azido group in the presence of PPh3 and H2O in THF
gave 9. Deprotection of both acetonide protecting
groups by using 70% aqueous tri¯uoroacetic acid, gave
rise to I.19 The inhibitory activity (IC50) of I was deter-
mined by using toluene-permeabilised cells of Escheri-
chia coli20,21 and the speci®c nucleoside substrate of the
translocase UMA5. Tunicamycin (from Sigma) and
mureidomycin-B were used as standards (Table 1).
Common features among these three families might
re¯ect the minimal structure responsible for the activity
displayed by both inhibitors. In order to con®rm this
hypothesis, we undertook the synthesis of I, a structure
that might correspond to the minimal part of LPMs
responsible for activity (Scheme 2).
Chlorination of 1 (commercially available) with thionyl
chloride gave 2, which was stored and used as a 1 M meth-
ylene chloride solution. Condensation of 3 (commercially

C. Dini et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1839±1843
1841
Scheme 2. (a) SOCl2, AcOH, CH2Cl2, 2 days, rt; (b) Hg(CN)2, CH2Cl2, 2 days, rt, 80%; (c) NH3, MeOH, rt, 2 days; (d) PTSA, acetone, 2-meth-
oxypropene, 3 h, rt (72% for steps (c) and (d); (e) MsCl, Et3N, 1.5 h, rt, 50%; (f) NaN3, DMF, 70 ꢀC, ꢁQ; (g) PPh3, H2O, THF, rt, 90%; (h)
TFA:H2O (7:3), 30 min, rt, ꢁQ.
Table 1.
of dibutyltin oxide and azeotropic removal of water in
toluene was followed by sequential addition of cesium
¯uoride in DMF and benzyl bromide, at room tem-
perature. 11S was prepared from 11R, by inversion of
con®guration using Mitsunobu conditions and p-nitro-
benzoic acid. Subsequent methanolysis of the resulting
crude gave 11S.
Tunicamycins
0.48
Mureidomycin B
0.065
I
IC50 (mM)
50
As I showed a reasonable level of activity (50 mM)
against translocase, we hypothesised that the introduc-
tion of a chiral centre between the two sugar units could
further improve the activity, by decreasing the con-
formational ¯exibility. XIV and XV as (R) and (S)
(Scheme 3) hydroxymethyl homologues, respectively,
were chosen to provide an unambiguous answer for the
con®guration of the chiral centre. XIV and XV were pre-
pared from intermediates 11R and 11S, corresponding
respectively to the d-allofuranose and l-talofuranose ser-
ies. These intermediates are accessible from a unique
compound 10.22
The same chemical strategy as that used to give I was then
applied to both stereoisomers (11R and 11S) to provide
respectively, compounds XIV and XV (Scheme 4).
Condensation of 11 with 2 under Koenigs±Knoor condi-
tions led to 12. Methanolysis of 12, followed by acetonide
protection gave 13. Mesylation of the remaining primary
alcohol was performed with methanesulfonyl chloride and
TEA to give 14. Subsequent hydrogenolysis of benzyl
groups a€orded the diol 15. Mesylate substitution with
sodium azide in DMF led to the azido compound 16.
Deprotection of both acetonide-protecting groups was
accomplished by using 60% aqueous acetic acid.
Acetylation of the crude reaction product gave 17. The
10 was subjected to monobenzylation by using the stan-
nylene acetal method modi®ed by Ohno et al.23 Addition
Scheme 3. (a) Bu2SnO, PhMe/azeotropic removal, then BnBr, CsF, DMF; (b) pNO2-PhCO2H, DEAD, PPh3, PhMe; (c) MeONa, MeOH.

1842
C. Dini et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1839±1843
Scheme 4. (a) 2, Hg(CN)2, CH2Cl2, MS 4 A, rt; (b) MeONa, MeOH, rt; (c) 2-methoxypropene, PTSA, acetone, rt; (d) MsCl, TEA, CH2Cl2, rt; (e)
Pd/C, H2, MeOH, rt; (f) NaN3, DMF, 70 ꢀC; (g) 60% AcOH aq. 70 ꢀC; (h) Ac2O, Pyr., rt; (i) O,O0-bis(trimethylsilyl)uracil,TMSOTf, CH3CN, rt; (j)
MeONa, MeOH, rt; (k) PPh3, H2O, THF, rt.
4. Boyle, D. S.; Donachie, W. D. J. Bacteriol. 1998, 180, 6429.
5. For a recent review see Lee, V. J.; Hecker, S. J. Med. Chem.
Table 2.
Res. Rev. 1999, 19, 521, and references cited therein.
6. Brandish, P. E.; Kimura, K.; Inukai, M.; Southgate, R.;
Lonsdale, J. T.; Bugg, T. D. H. Antimicrob. Agents Chemother.
1996, 40, 1640.
Compound no. Tunicamycins Mureidomycin B
IC50 (mM) 0.48 0.065
I
XIV XV
50 425
5
7. Kimura, K.; Ikeda, Y.; Kagami, S.; Yoshihama, M.;
Suzuki, K.; Osada, H.; Isono, K. J. Antibiot. 1998, 51, 1099.
8. Inukai, M.; Isono, F.; Takatsuki, A. Antimicrob. Agents
and Chemother. 1993, 37, 980.
9. Spada, M. R.; Ubukata, M.; Isono, K. Heterocycles 1992,
34, 1147.
10. Knapp, S.; Nandan, S.; Resnick, L. Tetrahedron Lett.
1992, 33, 5485.
11. Kim, K. S., Cho, I. H., Ahn, Y. H., Park, J. I. J. Chem.
Soc., Perkin Trans. 1 1995, 1783.
introduction of the uracil moiety was performed using
the standard Vorbruggen procedure.24 Subsequent
hydrolysis of the crude compound led to the nucleoside
18. The azido group was reduced in the presence of
PPh3 and H2O in THF to give the desired isomers XIV25
and XV.26
The inhibitory activity (IC50) of XIV and XV on trans-
locase was determined and ®nal results are summarised
in Table 2.
12. Kim, K. S.; Lim, J. W.; Joo, Y. H.; Kim, K. T.; Cho, I.
H.; Ahn, Y. H. J. Carbohydr. Chem. 1995, 14, 439.
13. Moore, W. J.; Luzzio, F. A. Tetrahedron Lett. 1995, 36,
6599.
14. Gravier-Pelletier, C.; Charvet, L.; Le Merrer, Y.; Depe-
zay, J. C. J. Carbohydr. Chem. 1997, 16, 129.
15. Le Merrer, Y.; Gravier-Pelletier, C.; Gerrouache, M.;
Depezay, J. C. Tetrahedron Lett. 1998, 39, 385.
16. Kim, K. S.; Ahn, Y. H. Tetrahedron: Asymmetry 1998, 9,
3601.
17. Ubukata, M.; Kimura, K.; Isono, K.; Nelson, C. C.;
Gregson, J. M.; Mc Closkey, J. A. J. Org. Chem. 1992, 57,
6392.
18. Dinur, U.; Hagler, A. T. Approaches to Empirical Force
Fields. In Reviews of Computational Chemistry; Lipkowitz,
K. B., Boyd, D. B., Eds.; VCH: New York, 1991; Vol. 2,
Chapter 4.
19. Analytical data for I (formic acid salt): 1H NMR
(300 MHz, D2O): 3.05 (dd, 1H, J=9, 13 Hz, H500a), 3.40 (dd,
1H, J=2.5, 13 Hz, H500b), 3.76 to 4.37 (m, 8H, H400, H300, H200,
H40), 5.12 (s, 1H, H100), 5.87 (d, 1H, J=3.5 Hz, H10), 5.90 (d,
1H, J=8 Hz, H5), 7.74 (d, 1H, J=8 Hz, H6), 8.46 (s, 1H,
formic acid salt); MS (FAB): 376+ (M+H+).
20. Vosberg, H. P.; Ho€mann-Berling, H. J. Mol. Biol. 1971,
58, 739.
Only the (S) isomer XV displays an ecient inhibitory
activity on the enzyme. This result is in agreement with
the molecular model we propose, and also with the
con®guration proposed by Isono's group for liposido-
mycin. Moreover, modi®cation of I to give XV resulted
in a 10-fold improvement of the inhibitory activity.
In conclusion, the model based on stable conformations
of two highly potent enzyme inhibitors (TCMs and
LPMs) with that of the natural substrate (UMA5) has
led to the synthesis of smaller molecules which still dis-
play useful enzyme inhibitory activity. The model also
supports Isono's work on the con®guration of the 50
chiral centre. The next step will be to elucidate the role
of each functional group in such molecules.
Acknowledgement
We are grateful to the Analytical Department (HMR,
Romainville) for performing spectral analysis.
21. Maas, D.; Pelzer, H. Arch. Microbiol. 1981, 130, 301.
22. Kishi, Y.; Fang, F.; Forsyth, C. J.; Scola, P. M.; Yoon, S.
K. Patent WO 9317690, 1993.
23. Nagashima, N.; Ohno, M. Chem. Pharm. Bull. 1991, 39,
1972.
References and Notes
24. Niedballa, U.; Vorbruggen, H. J. Org. Chem. 1974, 39,
3668.
1. Plattner, J. J. Annual Reports in Medicinal Chemistry 1997,
32, 111.
2. Struve, G. W.; Sinha, R. K.; Neuhaus, F. C. Biochemistry
1966, 5, 82.
3. Kandler, O. Cell wall structure and their phylogenis impli-
cations, Zbl. Bakt. Hyg., I Abt. Orig. 1982, C3, 149.
25. Analytical data for XIV (acetate salt): 1H NMR
(400 MHz, D2O): 1.91 (s, 3H, CH3 acetate salt), 3.09 (dd, 1H,
J=9, 13 Hz, H500a), 3.36 (dd, 1H, J=2.5, 13 Hz, H500b), 3.74
(dd, 1H, J=5, 12.5 Hz, H60a), 3.89 (dd, 1H, J=3.5, 12.5 Hz,
H60b), 4.06 (dt, 1H, J=3.5, 5 Hz, H50), 4.10 to 4.20 (m, 4H,

C. Dini et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1839±1843
1843
H400, H300, H200, H40), 4.39 (dd, 1H, J=5, 6 Hz, H30), 4.41 (dd,
1H, J=6, 4.5 Hz, H20), 5.29 (s, 1H, H100), 5.81 (d, 1H, J=4.5
Hz, H10), 5.90 (d, 1H, J=8 Hz, H5), 7.66 (d, 1H, J=8 Hz,
H6); MS (ESI): 406+ (M+H+).
26. Analytical data for XV (acetate salt): 1H NMR (300 MHz,
D2O): 1.92 (s, 3H, CH3 acetate salt), 3.19 (dd, 1H, J=8,
13.5 Hz, H500a), 3.41 (dd, 1H, J=3, 13.5 Hz, H500b), 3.80 (dd,
1H, J=6.5, 12 Hz, H60a), 3.87 (dd, 1H, J=4, 12 Hz, H60 b),
4.01 to 4.30 (m, 6H, H400, H300, H200, H40, H30, H50), 4.35 (dd,
1H, J=3, 5 Hz, H20), 5.20 (s, 1H, H100), 5.79 (d, 1H, J=3 Hz,
H10), 5.89 (d, 1H, J=8, H5), 7.82 (d, 1H, J=8 Hz, H6); MS
(FAB): 406+ (M+H+).