Coumarin inhibitors of gyrase B with N-propargyloxy-carbamate as an effective pyrrole bioisostere

Article abstract of DOI:10.1016/S0960-894X(99)00654-X

The synthesis and biological profile in vitro of a series of coumarin inhibitors of gyrase B bearing a N-propargyloxycarbamate at C-3' of noviose is presented. Replacement of the 5-methylpyrrole-2-carboxylate of coumarin drugs with an N-propargyloxycarbamate bioisostere leads to analogues with improved antibacterial activity. Analysis of crystal structures of coumarin antibiotics with the 24 kDa N-terminal domain of the gyrase B protein provides a rational for the excellent inhibitory potency of C-3' N- alkoxycarbamates.

Full text of DOI:10.1016/S0960-894X(99)00654-X

Bioorganic & Medicinal Chemistry Letters 10 (2000) 161±165
Coumarin Inhibitors of Gyrase B with N-Propargyloxy-carbamate
as an E€ective Pyrrole Bioisostere
Anne-Marie Periers, a Patrick Laurin, a Didier Ferroud, a Jean-Luc Haesslein, a
Michel Klich, a Claudine Dupuis-Hamelin, b Pascale Mauvais, b Patrice Lassaigne, b
Alain Bonnefoy b and Branislav Musicki a,*
aMedicinal Chemistry, Hoechst Marion Roussel, 102 route de Noisy, 93235 Romainville Cedex, France
bInfectious Disease, Hoechst Marion Roussel, 102 route de Noisy, 93235 Romainville Cedex, France
Received 4 August 1999; accepted 18 November 1999
AbstractÐThe synthesis and biological pro®le in vitro of a series of coumarin inhibitors of gyrase B bearing a N-propargyloxy-
carbamate at C-30 of noviose is presented. Replacement of the 5-methylpyrrole-2-carboxylate of coumarin drugs with an N-pro-
pargyloxycarbamate bioisostere leads to analogues with improved antibacterial activity. Analysis of crystal structures of coumarin
antibiotics with the 24 kDa N-terminal domain of the gyrase B protein provides a rational for the excellent inhibitory potency of
C-30 N-alkoxycarbamates. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
higher antibacterial potency, it appears that 5-methy-
pyrrole-2-carboxylate contibutes to higher hydro-
phobicity and higher membrane permeability.
As part of our continuing structure±activity relationship
(SAR) study of coumarin antibiotics (1±3),1±5 we
describe our e€orts aimed at modi®cation of 5-methyl-
pyrrole-2-carboxylate moiety linked to the C-30 position
of the sugar noviose. Coumarin antibiotics are inhibi-
tors of the catalytic functions of DNA gyrase that are
dependent on ATP hydrolysis (i.e. DNA supercoiling
and decatenation), by competitive inhibition of ATP
binding6 (for a review on structure and function of
DNA gyrase, see refs 7±10). Family members of coumer-
mycin antibiotics (3±6) produced by Streptomyces
species di€er by the presence of 5-methylpyrrole-2-car-
boxylate or pyrrole-2-carboxylate moiety in the mole-
cule.11 The comparative studies of antibacterial activity
of this group of coumarin analogues indicated that the
superior antibacterial pro®le of coumermycin A1 could
be attributed to the 5-methyl group of the 5-methyl-
pyrrole-2-carboxylate. Similarly, Berger and Batcho12
pointed out that 5-methylpyrrole-2-carboxylate, and not
the replacement of the C-8 methyl group of the cou-
marin by a chlorine atom, is the principal reason for
higher antibacterial potency of clorobiocin 2 compared
to novobiocin 1. As clorobiocin displays 2-fold higher
inhibitory activity in negative supercoiling of DNA
gyrase with respect to novobiocin, whereas 10-fold
In order to develop e€ective bioisosteres of 5-methyl-
pyrrole-2-carboxylate13 it was necessary to ful®l two
criteria. Firstly, isosteric replacement of 5-methylpyr-
role-2-carboxylate should maintain similar, or provide
stronger inhibitory potency of negative supercoiling of
DNA gyrase to that of clorobiocin or novobiocin. Sec-
ondly, as DNA gyrase is an intracellular target, isosteric
replacement should confer membrane permeability to
the analogue. Since X-ray crystal structures of several
coumarin drugs with 24 kDa N-terminal domain of
gyrase B of E. coli have been solved,14±17 these results
were an obvious starting point for our research.
Examination of the binding site of the noviose portion
of clorobiocin bearing at C-30 5-methylpyrrole-2-car-
boxylate moiety revealed that the pyrrole ring is sur-
rounded by hydrophobic amino acid residues Val-43,
Val-71, Val-120, Val-167, and Ile-78 (hydrophobic
pocket; Fig. 1). Besides these hydrophobic contacts,
there is an important hydrogen-bonding interaction
between the N±H of the pyrrole and the side chain of
Asp-73, and the hydrogen bonding that links the car-
bonyl group of the pyrrole-2-ester, an ordered water
molecule and the side chain of Thr-165 and Gly-77.
As will be seen by the results of biological activity
(vide infra) of coumarin analogues, both hydrophobic
*Corresponding author. Fax: +33-1-4991-5087; e-mail: branislav.
musicki@hmrag.com
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(99)00654-X

162
A.-M. Periers et al. / Bioorg. Med. Chem. Lett. 10 (2000) 161±165
interactions as well as s-cis orientation of Ha±N and
CˆO groups around the amide bond are crucial in
achieving a bioisostere with good biological activity.
provided an a-glycoside 15 as the major product after
chromatographic separation (Scheme 3). Having pre-
pared diol 15, we investigated regioselective protection
of 30-OH with Et3Si, MEM (MeOCH2CH2OCH2-), and
BOM (PhCH2OCH2-) protecting groups. In all cases the
major regioisomers formed 16a±c were at 30-OH and
could be easily separated by chromatography from 20-
regioisomers 17a±c. In general, deprotection of silyl
groups is relatively easy and we decided to continue our
synthetic scheme with this protecting group. Dihy-
dropyranylation of 16a a€orded a diastereomeric mix-
ture of THP derivatives (ꢀ1:1) 18 that was smoothly
desilylated and hydrogenolysed to furnish intermediate
19. The free 30-OH of the noviose derivative 19 could be
selectively transformed under standard conditions
(RCOCl, DMAP, Py or EDAC, DMAP, CH2Cl2) into
30-noviose esters. Similarly, intermediate 19 could
be selectively converted by the reaction of its p-nitro-
phenylcarbonate activated form with amines into 30-
carbamates 20a, or, by the reaction with hydroxyl-
amines into 30-N-alkoxycarbamates 20b. Once the
potential pyrrole bioisosteres were introduced at C-30,
the intermediates were subjected to hydrolytic condi-
tions with TosOH cat. in MeOH to remove tetra-
hydropyranyl group and a€ord 30-carbamates 21a±d or
30-N-alkoxycarbamates 22a±g of the coumarin-3-car-
boxy series. Alternatively, the 3-ester group in 20a,b
could be ®rst exchanged by ammonia, amines or
hydroxylamines to provide coumarin-3-carboxamide
24a, N-substituted coumarin-3-carboxamides 24b±d or
coumarin-3-hydroxamate derivatives 25a±c, respectively,
after subsequent THP deprotectection.
Chemistry
The standard synthetic route that could be used to
introduce potential pyrrole isosters consists of opening
the conveniently a-glycosylated noviose 20,30-carbonate
7 with nucleophiles (Scheme 1).13,19,20 However, it is
known that the opening of the corresponding carbonate
under thermodynamic control leads to a mixture of 30-
substituted 8 and 20-substituted 9 derivatives. Although
thermodynamic equilibrium favours regioisomer
8
(from 4:1 to 1:1 depending on nucleophile), in many
cases separation of the regioisomers is troublesome. In
order to avoid testing the mixtures 8 and 9, we wanted
to develop a synthetic approach that would enable
regioselective introduction of potential bioisosteres at
the C-30 position of noviose.
Previously, we have described the short synthesis of a
valuable intermediate for the construction of coumarin
drugs: 7-hydroxy-8-methyl-4-benzhydryloxycoumarin
(10).3 The preparation of the equally useful coumarin
building block 13 starting with THP protected deriva-
tive 11 is depicted in Scheme 2.
As in the case of the Mitsunobu's coupling of com-
pound 10 with noviose 14 (Scheme 4), coupling of the
noviose 14 with coumarin 13 under same conditions
Figure 1. Interactions between the 24 kDa N-terminal protein fragment of gyrase B of E. coli with clorobiocin and novobiocin.16,18
Scheme 1.

A.-M. Periers et al. / Bioorg. Med. Chem. Lett. 10 (2000) 161±165
163
Scheme 2. Reagents and conditions: (a) ClCOOEt, DMAP, CH2Cl2, rt, 55%; (b) BnOH, PPh3, EtO2CN=NCO2Et, CH2Cl2, rt, 49%; (c) HCl (1
M), THF, rt, 89%.
Scheme 3. Reagents and conditions: (a) PPh3, EtO2CNˆNCO2Et, CH2Cl2, rt, 54%; (b) Et3SiCl, DIPEA, Im, CH2Cl2, rt, 66%; (c) MEMCl,
DIPEA, Im, CH2Cl2, rt, 60%; (d) BOMCl, DIPEA, Im, CH2Cl2, rt, 60%; (e) DHP, TosOH cat, CH2Cl2, rt, 79%; (f) H2, Pd-C/10%, THF, rt; (g)
Bu4NF, THF, rt, (81% from 16a); (h) (i) p-NO2C6H4OCOCl, DMAP, CH2Cl2, 0 ꢁC; (ii) R0NH2 or R0ONH2, DMAP, DMF, rt; (i) TosOH cat,
MeOH, rt, (40±60% from 19); (j) NH3, R0NH2, THF, rt; or R0ONH2, Py, rt, (k) TosOH cat, MeOH, rt. (30±60% from 19).
Scheme 4. Reagents and conditions: (a) 10, PPh3, iPrO2CNˆNCO2iPr, CH2Cl2, rt, 67%; (b) Et3SiCl, DIPEA, Im, CH2Cl2, rt, 70%; (c) DHP,
TosOH cat, CH2Cl2, rt; (d) H2, Pd-C/10%, THF, rt; (e) Ac2O, CH2Cl2, 0 ꢁC; (f) Bu4NF, THF, rt, (70% from 28); (g) (i) p-NO2C6H4OCOCl,
ꢁ
.
DMAP, CH2Cl2, 0 C; (ii) HCꢂCCH2ONH2 HCl, DMAP, DMF, rt; (h) TosOH cat, MeOH, rt, (67%, from 28); (i) RONH2HCl, KOAc, EtOH, rt,
80%-quant.
A similar synthetic approach to that described, but
starting with coumarin intermediate 10 provided access
to a 3-acetylcoumarin 31 and a series of corresponding
oxime analogues 32a±h bearing the a 30-N-propargyl-
oxycarbamate moiety (Scheme 4).
ethoxy inhibitors possessing 30-carbamates 21a±c or
30-N-alkoxycarbamates 22a±g and a series of diverse
coumarin analogues with 30-N-propargyloxycarbamate
on the sugar noviose. For the comparative assessment,
in the Table 1 results are also presented showing the
biological activity of some pyrrole counterparts (26, 27,
Scheme 3 and 33, 34, Scheme 4). A carbamoyl deriva-
tive 21a gave similar inhibition in negative supercoiling
of DNA gyrase as novobiocin. With N-alkyl substi-
tution, a diminution of the inhibitory potency was
observed (21b±d). In the series of N-alkoxycarbamates
22a±g, an inverse e€ect dominated. Increasing the
Biological Results
Table 1 shows the inhibition in the supercoiling and
antibacterial activity of E. coli or S. aureus DNA gyrase
by novobiocin, clorobiocin and the coumarin-3-carb-

164
A.-M. Periers et al. / Bioorg. Med. Chem. Lett. 10 (2000) 161±165
Table 1. In vitro activity of coumarin inhibitors against E. coli/S. aureus DNA gyrase supercoiling (IC50),a,b and selected in vitro antibacterial
activity (MIC)c
Compound
R1
Ratio
IC50nova;b
IC50comp
MIC (mg/mL)
S. aureus
011HT3
S. aureus
011GO64
O¯oOxaEry-R
S. aureus
011HT1
Nov-R
S. epidermidis
012GO42
Oxa-R
S. pyogenes
02A1UC1
E. faecium
02D31P2
VanTeiEry-R
Novobiocin
Clorobiocin
1a,b
1.7a
ꢃ0.04
ꢃ0.04
ꢃ0.04
10
0.6
ꢃ0.04
0.3
ꢃ0.04
0.6
ꢃ0.04
ND
ND
Scheme 3
21a
21b
21c
21d
22a
22c
22d
22e
22f
22g
24a
24b
-NH2
-NH-Allyl
-NH-Propargyl
-NHiPr
-NHOMe
-NHOEt
-NHOiPr
-NHO-Allyl
-NHO-Propargyl
-NHO-But-2-ynyl
-NH2
1.6a
>40
>4 0
40
>4 0
>4 0
20
>4 0
10
1.2
10
ND
>4 0
>40
>4 0
>4 0
>40
>4 0
20
2.5
40
ꢃ0.04
ꢃ0.04
0.08
>40
>4 0
>40
>4 0
>4 0
>40
>4 0
>40
>40
>40
0.3
ND
2 0
20
>4 0
>4 0
>40
>4 0
20
0.3
10
ꢃ0.04
ꢃ0.04
0.08
>40
>4 0
>40
>4 0
>4 0
40
>4 0
20
20
>40
0.6
>40
>4 0
>40
>4 0
>4 0
>40
>4 0
>40
10
0.38a
0.38a
0.025a
0.66a
2.0a
2.0a
2.0a
5.6a
0.5b
>40
0.08
0.6
1.0a
ꢃ0.04
ꢃ0.04
ꢃ0.04
-NH-3-Pyridyl
0.67b
2.6b
1.2
1.2
0.08
1.2
24c
2.5
24d
25a
25b
25c
-NHCH2CH2OH
-NHOMe
-NHOEt
2.6b
11.1b
5.6b
ꢃ0.04
0.08
0.08
0.6
0.08
ꢃ0.04
1.2
10
1.2
0.6
ꢃ0.04
0.08
0.6
0.6
0.6
0.15
0.08
1.2
2.5
0.6
0.3
ꢃ0.04
2.0b
ꢃ0.04
0.15
26
27
3.1a
1.67a
0.3
ꢃ0.04
ND
ND
>20
2.5
ND
ND
0.3
0.15
5
2.5
Scheme 4
31
32a
32b
32c
8.1b
12.5b
8.3b
ꢃ0.04
0.15
0.08
0.6
2.5
5
2.5
5
ꢃ0.04
0.08
0.15
0.3
0.6
0.6
1.2
1.2
2.5
1.2
20
-Me
-Et
0.15
0.15
0.6
4.0b
0.15
0.15
32d
32e
-CH2O(CH2)2OMe
4.0b
1.3b
0.08
ꢃ0.04
0.6
ꢃ0.04
10
2.5
1.2
ꢃ0.04
0.6
0.15
5
0.6
33
34
2a
3.3a
ꢃ0.04
ND
ND
5
40
ND
ND
0.3
0.3
2.5
10
0.15
aIC50 was determined for gyrase B of E. coli against novobiocin (0.25 mg/mL) as reference. For details see ref 3.
bSupercoiling assay using puri®ed DNA Gyrase from S. aureus: the enzyme was puri®ed from a crude extract of S. aureus E34159. The frozen cells
were suspended in TGED bu€er (pH 7.5) supplemented with 9 mM dithiothreitol, 20 mM EDTA, 0.4% Brij 58, 60 mg of lysostaphin per mL,
proteases inhibitors and incubated for 30 min at 20 ꢁC and 25 min at 37 ꢁC. After the addition of 0.4 M KCl, incubation was repeated 30 min at
20 ꢁC. The cell suspension were sonicated, centrifuged at 100,000 g for 45 min. The supernatant was treated with 4% streptomycin for the removal of
DNA. After centrifugation, the supernatant was supplemented with 65% (NH4)2SO4. The precipitate was dissolved and dialysed in TGED bu€er.
The dialysate was directly puri®ed on novobiocin±Sepharose anity column. The gyrase A was eluted with TGED bu€er containing 1 M KCl and
the gyrase B was eluted with TGED bu€er containing 1 M KCl and 5 M urea. After dialyse, the proteins were concentrated with PEG 2000. Relaxed
DNA was prepared from pBR322 plasmid with calf thymus topoisomerase I (GIBCO-BRL), 1 h at 37 ꢁC. The DNA concentration was determined
by spectrophotometric measurements. Supercoiling assay was performed on 40 mL assay mixture containing 40 mM Tris±HCl pH 7.3, 20 mM KCl,
4 mM MgCl2, 2 mM ATP, 2 mM spermidine-HCl, 0.8 mg tRNA, 50 ng relaxed DNA and 1 unit of DNA-gyrase. One unit of gyrase was de®ned as
the amount of activity that supercoiled 50 ng of relaxed pBR322 in 60 min at 37 ꢁC. The reaction was stopped by addition of protease K and
separated by electrophoresis. IC50 was determined for inhibitors against novobiocin (0.5 mg/mL) as reference.
cMIC, minimum inhibitory concentrations (mg/mL) were measured by using a 2-fold broth microdilution after 24 h incubation. Particular phenotype
of resistance (-R) of the tested bacterial strains were mentioned: O¯o for o¯oxacin, Oxa for oxacillin, Ery for erythromycin, Nov for novobiocin, Tei
for teicoplanin, Van for vancomycin. Otherwise, strains were fully susceptible.
length of the substituent (up to a 3-carbon substituent)
produced an increase in inhibitory potency. The poor
activity of the carbamate series (21b±d) could be a
re¯ection of their preferred s-trans conformation 35b
around the amide bond (Fig. 2).21,22 This conformation
is placing N±H and CˆO bonds in the opposite direc-
tion to that observed by the X-ray model in Figure 1.
In the case of N-alkoxycarbamates (22a±g), s-cis con-
former 36b prevails, directing at the same time the alkyl

A.-M. Periers et al. / Bioorg. Med. Chem. Lett. 10 (2000) 161±165
165
Figure 2. Calculated relative conformational energies of carbamate (35a,b) and alkoxycarbamate (36a,b) derivatives.
group to the hydrophobic pocket. However, this
hydrophobic domain is not large enough to accom-
modate the alkyl chains containing more than three
carbon atoms, and an important loss of activity was
observed with analogue 22g.
References and Notes
1. Klich, M.; Laurin, P.; Musicki, B.; Schio, L. WO 9747634,
1998; Chem. Abstr. 1998, 128, 75634.
2. Chartreaux, F.; Klich, M.; Schio, L. EP 894805, 1999;
Chem. Abstr. 1999, 130, 125344.
3. Laurin, P.; Ferroud, D.; Klich, M.; Dupuis-Hamelin, C.;
Mauvais, P.; Lassaigne, P.; Bonnefoy, A.; Musicki, B. Bioorg.
Med. Chem. Lett. 1999, 9, 2079.
4. Laurin, P.; Ferroud, D.; Schio, L.; Klich, M.; Dupuis-
Hamelin, C.; Mauvais, P.; Lassaigne, P.; Bonnefoy, A.;
Musicki, B. Bioorg. Med. Chem. Lett. 1999, 9, 2875.
5. Ferroud, D.; Collard, J.; Klich, M.; Dupuis-Hamelin, C.;
Mauvais, P.; Lassaigne, P.; Bonnefoy, A.; Musicki, B. Bioorg.
Med. Chem. Lett. 1999, 9, 2881.
6. Gellert, M.; O'dear, M. H.; Itoh, T.; Tomizawa, J. I. Proc.
Natl. Acad. Sci. USA 1976, 73, 4474.
7. Reece, R. J.; Maxwell, A. Crit. Rev. Biochem. Mol. Biol.
1991, 26, 335.
8. Levine, C.; Hiasa, H.; Marians, K. J. Biochim. Biophys.
Acta 1998, 1400, 29.
9. Berger, J. M. Biochim. Biophys. Acta 1998, 1400, 3.
10. Maxwell, A. Biochem. Soc. Trans. 1999, 27, 48.
11. Godfrey, J. C.; Price, K. E. Adv. Appl. Microbiol. 1972, 15,
231.
12. Berger, J.; Batcho, A. D. In Antibiotics: Isolation, Separation
and Puri®cation (J. Chromatography Library 15); Winstein, M.
J.; Wagmen, G. H. Eds.; Elsevier: London, 1979; pp 101±158.
13. Some e€orts in this direction has been made, without suc-
cess, however, by the group of Bristol±Mayers Squibb. Ueda,
Y.; Chuang, J. M.; Fung-Tomc, J.; Partyka, R. A. Bioorg.
Med. Chem. Lett. 1994, 4, 1623.
Comparison of the analogues possessing N-propargyl-
oxycarbamate moiety and 5-methylpyrrole-2-carboxy
group: 22f and 26 (Scheme 3), 24a and 27 (Scheme 3), 31
and 33 (Scheme 4), 32 and 34 (Scheme 4), clearly indi-
cates improved inhibitory potency of the ®rst. Thus, N-
propargyloxycarbamate has ful®lled the ®rst criteria
imposed for a good bioisostere of 5-methylpyrrole-2-
carboxylate: superior inhibitory potency to analogues
possessing 5-methylpyrrole-2-carboxylate. While three
N-alkoxycarbamates 22d±f display almost the same
inhibitory activity, a dramatic di€erence between the
three in MIC values is observed. While 22d is devoid of
antibacterial activity, 22f displays similar antibacterial
spectrum to pyrrole analogue 26. In this way, N-pro-
pargyloxycarbamate has ful®lled the second criteria of
the good bioisostere: it conferred the membrane perme-
ability that is prerequisite for a good antibacterial
activity.
Di€erent coumarin classes having 30-N-propargyloxy-
carbamate: coumarin-3-esters (22f), -3-amides (24a), -3-
O-alkylhydroxamates (25a), -3-ketones (31) and -3-
alkoxyimino (32a) derivatives, were highly active
against all the Gram-positive strains tested including
Enteroccus and multi-resistant staphylococci. In general,
improved anti-Enterococcal activity was observed with
the novel series compared to pyrrole analogues. Amide
and hydroxamate derivatives 24a and 25c, respectively,
displayed the best antibacterial activity, particularly
against novobiocin-resistant strains.
14. Wigley, D. B.; Davies, G. J.; Dodson, E. J.; Maxwell, A.;
Dodson, G. Nature 1991, 351, 624.
15. Lewis, R. J.; Singh, O. M. P.; Smith, C. V.; Maxwell, A.;
Skarzynsky, T.; Wonacott, A. J.; Wigley, D. B. J. Mol. Biol.
1994, 241, 128.
16. Lewis, R. J.; Singh, O. M. P.; Smith, C. V.; Skarzynski, T.;
Maxwell, A.; Wonacott, A. J.; Wigley, D. B. EMBO. J. 1996,
15, 1412.
17. Tsai, F. T. F.; Singh, O. M. P.; Skarzynski, T.; Wonacott,
J. A.; Weston, S.; Tucker, A.; Pauptit, R. A.; Breeze, A.;
Poyser, J. P.; O'Brien, R.; Ladbury, J. E.; Wigley, D. B. Pro-
teins: Struct. Funct. and Genet. 1997, 28, 41.
18. We successfully solved X-ray crystal structures of several
coumarin derivatives with 24 kDa N-terminal fragment of
gyrase B in collaboration with P. Oudet and D. Moras at the
University of Louis Pasteur, Illkirch, France (unpublished results).
19. Hinman, J. W.; Caron, E. L.; Hoeksema, H. J. Am. Chem.
Soc 1957, 79, 5321.
In conclusion, based on the results of crystallographic
determination of the active site of 24 kDa N-terminal
subdomain of gyrase B with coumarin antibiotics, we
succeeded in designing N-propargyloxycarbamate as an
e€ective 5-methylpyrrole-2-carboxylate bioisostere. This
leads to analogues with improved in vitro inhibitory
potency of the negative supercoiling of DNA gyrase as
well as improved antibacterial activity. Among the can-
didates, amido 24a±d and hydroxamate 25a±c deriva-
tives with the uniformly good antibacterial activity seem
to be the most promising ones.
20. Bell, W.; Block, M. H.; Cook, C.; Grant, A.; Timms, D. J.
Chem. Soc., Perkin Trans 1997, 1, 2789.
21. Oki, M.; Nakanishi, H. Bull. Chem. Soc. Japan 1971, 44,
3148.
22. The relative energies of the conformers 35a,b and 36a,b
were minimised with Insight/Discoverer software applying
CVFF force®eld and conjugate gradient method. Dinur, U.;
Hagler, A. T. Approaches to empirical force ®elds. In Reviews
of Computational Chemistry; Lipkowitz, K. B.; Boyd, D. B.
Eds.; VCH: New York; 1991; Vol. 2 (Chapter 4).
Acknowledgement
We would like to thank the Analytical Department
(HMR, Romainville) for performing the spectral analysis.