Oxazolidine-2-thiones: A molecular modeling study

Article abstract of DOI:10.1016/j.tetlet.2004.06.090

Two oxazolidine-2-thiones, thio-analogs of linezolid, were synthesized and their antibacterial properties evaluated. Unlike oxazolidinones, the thio-analogs did not inhibit the growth of Gram positive bacteria. A molecular modeling study has been carried out to aid understanding of this unexpected finding. Two oxazolidine-2-thiones, thio-analogs of linezolid, were synthesized and their antibacterial properties evaluated. Unlike oxazolidinones, the thio-analogs did not inhibit the growth of Gram positive bacteria. A molecular modeling study has been carried out to aid understanding of this unexpected finding.

Full text of DOI:10.1016/j.tetlet.2004.06.090

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 6269–6272
Oxazolidine-2-thiones: a molecular modeling studyq
Neha Gandhi, Brijesh K. Srivastava, Vidya B. Lohray and Braj B. Lohray*
Department of Chemistry, Zydus Research Centre, Cadila Healthcare Ltd, Sarkhej–Bavla N. H.8A, Moraiya,
Ahmedabad 382210, India
Received 5 May 2004; revised 14 June 2004; accepted 22 June 2004
Abstract—Two oxazolidine-2-thiones, thio-analogs of linezolid, were synthesized and their antibacterial properties evaluated. Un-
like oxazolidinones, the thio-analogs did not inhibit the growth of Gram positive bacteria. A molecular modeling study has been
carried out to aid understanding of this unexpected finding.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
lower toxicity in the drug candidate. In the case of
linezolid itself, replacement of the acetamide group in
the side-chain by thioacetamide 2 led to a significant
improvement of in vitro MIC values.3 This encouraged
us to explore the effect of replacing the carbonyl group
of the oxazolidinone ring (X=O) by a thiocarbonyl
group (X=S).
Linezolid 1 has been in clinical use since 2000 for treat-
ing infection caused by terminal Gram +ve organisms.
However, toxicity as well as emergence of resistance in
some patients receiving prolonged treatment has been
reported.1 A number of attempts have been made by
various research groups to obtain potent and safer ana-
logs without much success.2 Replacement of oxygen by
sulfur in several pharmacophores has led to dramatic
improvements in efficacy, pharmacokinetics, and to a
Seneci et al. have reported replacement of the carbonyl
group of oxazolidinone by a thiocarbonyl group
in N-[3-(4-acetyl-phenyl)-2-oxo-oxazolidin-5-ylmethyl]-
acetamide (DuP 721).4 We decided to synthesize oxazol-
idine-2-thiones 3 and 4, as analogs of linezolid, to eval-
uate their antibacterial potential. Compounds 3 and 4
were synthesized according to the method outlined in
Scheme 1. The yield reported in Scheme 1 is for com-
pound 3.5
X
O
N
N
NH
Me
F
Y
Evaluation of thio-analogs 3 and 4 in an in vitro MIC
assay against a panel of Gram +ve bacteria (Enterococ-
cus faecium, E. faecalis, Bacillus cereus, B. subtilis, Sta-
phylococcus epidermidis, S. aureus, and B. pumilus) led
to surprising results. The compounds were found to be
completely ineffective against all the organisms includ-
ing the most sensitive, B. subtilis (MIC value >64lg/
mL using the standard assay as described in NCCLS
guidelines). In contrast, replacement of ÔOÕ by ÔSÕ in
the acetamide side-chain of linezolid 2 led to superior
antibacterial activity against all the organisms we stud-
ied.3 In order to understand this contrasting behavior,
we decided to carry out modeling studies of linezolid
1, thioacetamide 2 and thio-analog 3 in the potential
binding pocket.
=
O
N
; X=O ; Y=S : thioacetamide
; X=S ; Y=O
1
2
3
N
X=O ; Y=O : Linezolid
O
; X=S ; Y=O
4
N
N
=
N
OH
Keywords: Oxazolidine-2-thiones; Linezolid; Molecular modeling; 23S
rRNA; Antibacterial; Piecewise linear potential (PLP); Thioacetamide.
q ZRC Communication No. 120.
*
Corresponding author. Tel.: +91-2717-250801; fax: +91-2717-
250606; e-mail: braj.lohray@zyduscadila.com
0040-4039/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.06.090

6270
N. Gandhi et al. / Tetrahedron Letters 45 (2004) 6269–6272
S
R
a, b
O
n-Pr
O
+
O
R
NCS
N
F
O
OH
F
c - f
3
N
N
O
R=
S
R
R=
N-COCH2OH
O
4
N
F
NHAc
Scheme 1. Reagents and conditions: (a) LiBr, Ph3PO, xylene, 110–115°C, 1h; (b) NaOMe, MeOH, 26–28°C, 2h, overall yield for steps a and b is
100%; (c) MeSO2Cl, TEA, DCM, 0–5°C, 30min, 70%; (d) NaN3, DMF, 60–70°C, 1h, 81%; (e) Ph3P, 1,4-dioxane, MeOH, aq NH3, 26–28°C, 1h; (f)
Ac2O, pyridine, 0–5°C, overall yield steps e and f is 21%.
2. Molecular modeling studies with oxazolidinones and
oxazolidine-2-thiones
It has been reported that eperezolid binding to the 50S
ribosomal subunit is competitively inhibited by chlo-
ramphenicol.6 This suggests that oxazolidinones share
proximate binding sites on the 50S ribosome with chlo-
ramphenicol but the mode of action of oxazolidinones is
distinctly different from chloramphenicol. Although ox-
azolidinones inhibit bacterial protein synthesis, they
have no effect on peptidyl transferase unlike chloram-
phenicol or lincomycin.6 Colca et al.7 have shown that
oxazolidinone is cross-linked specifically to 23S rRNA,
tRNA, and two polypeptides and a specific oxazolidi-
none binding site is formed in the translating ribosome
in the immediate vicinity of the peptidyl transferase cen-
tre. Since the exact site and mechanism of action of ox-
azolidinones is highly debatable, we modeled the
oxazolidinone ring in a potential binding pocket on
the 23S rRNA using the X-ray co-ordinates from
1K018 and Discovery Studio SBD 1.2 (www.accel-
rys.com) using the Ligandfit module9 on Windows plat-
form. Before docking, the co-ordinates at a distance of
Figure 1. Chloramphenicol binding pocket (shown in yellow) in the
23S ribosomal structure. Crystal co-ordinates were extracted along a
˚
radius of 20A from the binding pocket.
4
Colca et al.7 have provided evidence that a binding site
for the oxazoldinone is formed near the peptidyl trans-
ferase centre in actively translating bacterial ribosomes.
Initiation of protein synthesis takes place when the 30S
ribosomal subunit, mRNA, fMet-tRNA and 50S com-
bine whilst oxazolidinones prevent initiation of protein
synthesis through binding to 23S rRNA of the 50S sub-
unit and inhibit fMet-tRNA from attaching to the ribo-
some.13 Various modifications in the ring system of
oxazolidinones and its structure–activity relationships
have been reported. We report here an unusual binding
pattern shown by the modified oxazolidinone ring
through molecular modeling. The docking results show
the 23S rRNA-ligand binding affinities (Table 1). The
˚
20A from the ligand chloramphenicol were extracted
as shown in Figure 1. The docking protocols were
validated by docking chloramphenicol in the binding
pocket, which generated hydrogen bonds with various
nucleotides of the peptidase transferase cavity.6 The
structures 1, 2 and 3 were docked using CFF and ran-
dom Monte–Carlo simulations in the binding pocket.
Docking analyses were observed on the basis of PLP10
(piecewise linear potential) scores and hydrogen
˚
bonds (at a distance of 2.5A) of the inhibitors with
the 23S ribosomal subunit.
3. Results and discussions
Table 1. Results of modeling of the 23S ribosome with linezolid 1,
thioacetamide 2 and thio-analog 3
Oxazolidinones act by inhibiting protein synthesis. Shi-
nabarger11 and Hutchinson2a have comprehensively
compiled the studies reported to elucidate the mecha-
nism of action of oxazolidinones. Ganoza and co-work-
ers12 have suggested that linezolid interferes with the
binding of fMet-tRNA to the ribosomal P site while
Compound
PLP110a
PLP210b
1
2
3
À59.94
À39.76
À17.19
À77.03
À25.55
À15.3

N. Gandhi et al. / Tetrahedron Letters 45 (2004) 6269–6272
6271
Figure 2. The hydrogen bonds are represented by black dotted lines and the binding nucleotides in stick representation. Nucleotide numbering is
according to the D. radiodurans sequence. Nucleotides A2430, U2485 and G2484 are not shown for clarity. (a) Energy minimized conformation of
linezolid 1 and thioacetamide analog 2. (b) Energy minimized conformation of linezolid 1 and thio-analog 3. (c) Binding pocket interaction of
linezolid 1 in 23S rRNA. (d) Binding pocket interaction of thio-analog 3 in 23S rRNA.
PLP scores are in arbitrary units of energy. Lower PLP
scores indicate better rRNA-ligand binding (higher pKi
values).
In contrast, and to our surprise, when oxazolidine-2-thi-
one 3 was docked using the same model, it adopted a
completely reversed orientation (Fig. 2b) in which the
side-chain of the acetamide of 3 lies over the morpholine
ring of 1 and the oxazolidine-2-thione moiety of 3 over-
laps the aromatic ring of 1. Thus, 1 and 3 have a 180°
opposite orientation (Fig. 2b). Docking 3 in the 23S
rRNA binding pocket did not show any of the interac-
tions that were observed with oxazolidinone 1 (Fig.
2d). Thio-analog 3 did not bind to the 23S rRNA in
the expected way. Thus, compound 3 may not be inhib-
iting protein synthesis, because it cannot form the de-
sired tertiary structure needed for inhibition of the
enzyme. Similar modeling was carried out with thio-ana-
log of eperezolid, 4 with no improvement in the results.
Thus, these modeling studies support our experimental
observations on the antibacterial activity of 3 and 4
and also suggest that thio-analogs simply do not fit into
the binding pocket of 23S ribosomes.
In general, the acetamide group in linezolid 1 is optimal
for antibacterial activity. However, the thioacetamide
derivative 2 is equally as good as linezolid in its antibac-
terial activities. From modeling studies of linezolid 1
and thioacetamide 2, it is clear that both orientate sim-
ilarly (Fig. 2a) when docked in the binding pocket (Fig.
2c), the oxygen atom of the morpholine ring hydrogen
bonds with U2564Dr in both cases (docking of linezolid
1 only is shown). The oxygen atom of the carbonyl of
the oxazolidinone ring in 1 and 2 is hydrogen bonded
to the nitrogen of G2044Dr (G2061Ec) of 23S rRNA.14
Similarly, the acetamide of 1 and thioacetamide of 2
form hydrogen bonds with the C2431Dr (C2452Ec) nu-
cleotide. These interactions result in the formation of a
tertiary complex, which inhibits protein synthesis (Fig.
2c). Nucleotides A2430Dr (A2451Ec), U2485Dr
(U2506Ec) and G2484Dr (G2502Ec) are not shown
for clarity. Nucleotide numbering is according to the
D. radiodurans sequence with E. coli nucleotides labeled
in parentheses. Interestingly, Rib-X Pharmaceuticals15
have reported the crystal structure of linezolid bound
to the 23S ribosome of various organisms including hu-
mans. The crystal bound linezolid appears to adopt a
similar configuration to that, which we obtained in the
energy minimization studies.
Acknowledgements
We thank Mr. Pankaj R. Patel, CMD and the manage-
ment of Zydus Group for encouragement and Dr. Sonja
Meddab, Accelrys for suggestions. We would like to
thank the referee for a number of excellent suggestions
and drawing our attention to several publications.

6272
N. Gandhi et al. / Tetrahedron Letters 45 (2004) 6269–6272
8. Schlunzen, F.; Zarivach, R.; Harms, J.; Bashan, A.; Tocilj,
References and notes
A.; Albrecht, R.; Yonath, A.; Franceschi, F. Nature 2001,
413, 814.
9. (a) Bo¨hm, H. J. J. Comput. Aided Mol. Des. 1994, 8, 243;
(b) Bo¨hm, H. J. J. Comput. Aided Mol. Des. 1998, 12,
309.
10. (a) Gehlhaar, D. K.; Verkhivker, G. M.; Rejto, P. A.;
Sherman, C. J.; Fogel, D. B.; Fogel, L. J.; Freer, S. T.
Chem. Biol. 1995, 2, 317; (b) Gehlhaar, D. K.; Bouzida,
D.; Rejto, P. A. Rational Drug Design: Novel Method-
ology and Practical Applications In ACS symposium series
719; Parrill, L., Rami Reddy, M., Eds.; American Chem-
ical Society: Washington, DC, 1999; pp 292.
1. (a) Prystowsky, J.; Siddiqui, F.; Chosay, J. Antimicrob.
Agents Chemother. 2001, 45, 2154; (b) Gonzales, R. D.;
Schreckenberger, P. C.; Graham, M. B. Lancet 2001, 357,
1179; (c) Tsiodras, S.; Gold, H. S.; Sakoulas, G. Lancet
2001, 358, 207.
2. (a) Hutchinson, D. Curr. Top. Med. Chem. 2003, 3, 1021,
and references cited therein; (b) Brickner, S. J. Curr.
Pharm. Design 1996, 2, 175.
3. (a) Lohray, V. B.; Lohray, B. B.; Srivastava B. K. Pure
Appl. Chem. 2004, communicated; (b) Lohray, V. B.;
Lohray, B. B.; Srivastava, B. K.; Gupta, S.; Solanki, M.;
Kapadnis, P.; Takale, V.; Pandya, P. Bioorg. Med. Chem.
Lett. 2004, 14, 3139.
11. Shinabarger, D. 1999, 8, 1195, pp 1195.
12. Hiroyuki, A.; Lizhu, K.; Susan, M. P.; Toni, J. P.;
Elizabeth, A. W.; Robert, C. G.; Richard, C. T.; Dean, L.
S.; Ganoza, C. Antimicrob. Agents Chemother. 2002, 46,
1080.
13. Zurenko, G. E.; Gibson, J. K.; Shinabarger, D. L.;
Aristoff, P. A.; Ford, C. W.; Tarpley, W. G. Curr. Opin.
Pharmacol. 2001, 1, 470.
14. The nucleotide numbers on the ribosome of Dinococcus
radiodurans are represented by ÔDrÕ, which is compared in
parentheses, with the nucleotide numbers of ribosomes
from E. coli ÔEcÕ. This is done to match with the nucleotide
numbers in E. coli for modeling purposes.
15. Steitz, T. A.; Moore, P. B.; Ban, N.; Nissen, P.; Hansen,
J.; Sutcliffe, J. A.; Oyelere, A. K.; Ippolito, J. A. EP
1308457 A1 (07/05/2003).
4. Seneci, P.; Caspani, M.; Ripamonti, F.; Ciabatti, R. J.
Chem. Soc., Perkin Trans. 1 1994, 2345.
5. The thio-analog of eperezolid was also synthesized in an
analogous manner using 4-(2-fluoro-4-isothiocyanophe-
nyl)-piperazine-1-carboxylic acid tert-butyl ester and
finally the Boc group was removed using TFA followed
by N-acylation using a-hydroxyacetic acid.
6. Lin, A.; Murray, R.; Vidmar, T.; Marotti, K. Antimicrob.
Agents Chemother. 1997, 41, 2127.
7. Colca, J. R.; McDonald, W. G.; Waldon, D. J.; Thom-
asco, L. J.; Gadwood, R. C.; Lund, E. T.; Cavey, G. S.;
Mathews, R.; Adams, L. D.; Cecil, E. T.; Pearson, J. D.;
Bock, J. H.; Mott, J. E.; Shinabarger, D.; Xiong, L.;
Mankin, A. S. J. Biol. Chem. 2003, 278, 21972.