1,5-Diaryl-2-ethyl pyrrole derivatives as antimycobacterial agents: Design, synthesis, and microbiological evaluation

Article abstract of DOI:10.1016/j.ejmech.2009.06.005

During the search of novel antitubercular drugs related to BM 212, new diarylpyrroles were designed and synthesized on the basis of a structure-activity relationship analysis of many pyrroles previously described by us. Among them, 1-(4-fluorophenyl)-2-et

Full text of DOI:10.1016/j.ejmech.2009.06.005

European Journal of Medicinal Chemistry 44 (2009) 4734–4738
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Short communication
1,5-Diaryl-2-ethyl pyrrole derivatives as antimycobacterial agents: Design,
synthesis, and microbiological evaluation
,
a
a
b
b
Mariangela Biava a , Giulio C. Porretta , Giovanna Poce , Alessandro De Logu , Rita Meleddu ,
**
*
Edda De Rossi c, Fabrizio Manetti d , Maurizio Botta
,
d
a
`
Dipartimento di Chimica e Tecnologie del Farmaco, Universita ‘‘La Sapienza’’, piazzale Aldo Moro 5, I-00185 Roma, Italy
b Dipartimento di Scienze e Tecnologie Biomediche, Sezione di Microbiologia Medica, viale Sant’Ignazio 38, I-09123 Cagliari, Italy
c
`
Dipartimento di Genetica e Microbiologia, Universita degli Studi di Pavia, via Ferrata 1, I-27100 Pavia, Italy
Dipartimento Farmaco Chimico Tecnologico, Universita degli Studi di Siena, via Alcide de Gasperi 2, I-53100 Siena, Italy
d
`
a r t i c l e i n f o
a b s t r a c t
Article history:
During the search of novel antitubercular drugs related to BM 212, new diarylpyrroles were designed and
synthesized on the basis of a structure–activity relationship analysis of many pyrroles previously
described by us. Among them, 1-(4-fluorophenyl)-2-ethyl-3-(thiomorpholin-4-yl)methyl-5-(4-methyl-
phenyl)-1H-pyrrole (2b) proved to be particularly active, with a minimum inhibitory concentration (MIC,
Received 19 December 2008
Received in revised form
3 June 2009
Accepted 4 June 2009
Available online 13 June 2009
expressed as
mg/mL) and a protection index (PI) better than or comparable to those of reference
compounds. Also the remaining compounds were very active, although their MIC and PI were in general
lower than those of their parent 2-methyl analogues.
Keywords:
BM 212
Ó 2009 Elsevier Masson SAS. All rights reserved.
1,5-Diarylpyrroles
Mycobacterium tuberculosis
Pharmacophoric model
Antitubercular agents
1. Introduction
XDR-TB) [5], is causing serious concern in some countries, because
patients could become virtually untreatable using currently avail-
Mycobacterium tuberculosis (MTB), responsible for tuberculosis
(TB) in humans, is positioned as the leading bacterial infectious
agent [1,2]. Its synergy with the HIV-1 epidemic led TB to increase,
in particular in some parts of the world, such as the World Health
Organization (WHO) African Region, where a large part of new TB
cases were attributable to HIV-1 co-infection [3]. Furthermore, the
emergence of MTB strains resistant to all of the first-line drugs
(leading to the multi drug-resistant TB, MDR-TB) [4] and to isoni-
azid and rifampin plus resistant to any fluoroquinolone and at least
one of three injectable second-line drugs (i.e., amikacin, kana-
mycin, or capreomycin) (leading to the extensive drug-resistant TB,
able anti-TB drugs. In addition, there have been no new drugs to
treat TB in the last 40 years, with the exception of fluoroquinolones
recently introduced [6], reflecting the inherent difficulties in
discovery and clinical testing of new antitubercular agents and the
lack of pharmaceutical industry research in the area [7]. As
a consequence, the development of new drugs with potent activity
toward MDR- and XDR-TB (as well as toward latent TB and infec-
tions caused by Mycobacterium avium complex) is urgently needed.
In continued attempts to identify new potent antimycobacterial
compounds, we have previously described many pyrrole deriva-
tives endowed with
a high inhibitory activity toward MTB,
including intramacrophagic mycobacteria and strains resistant to
the antitubercular drugs currently used in therapy [8]. Among these
compounds, 1,5-(4-chlorophenyl)-2-methyl-3-(4-methylpiperazin-
1-yl)methyl-1H-pyrrole (BM 212, 1a, Chart 1) proved to be the most
active and it was considered the lead compound of this new series
of derivatives. Structure–activity relationship (SAR) studies,
combined with a pharmacophoric model for antimycobacterial
compounds [8e], allowed us to find derivatives of 1a (compounds
1b–h, Chart 1) in which the following substituents and substitution
pattern on the pyrrole ring were identified as responsible for the
Abbreviations: MTB, Mycobacterium tuberculosis; TB, tuberculosis; WHO, World
Health Organization; MDR, multi drug-resistant; XDR, extensively drug-resistant;
BM 212, 1,5-(4-chlorophenyl)-2-methyl-3-(4-methylpiperazin-1-yl)methyl-1H-
pyrrole; SAR, structure–activity relationship; MIC, minimum inhibitory concentra-
tion; MNTD, maximum non-toxic dose; PI, protection index; DMEM, Dulbecco’s
minimum essential medium.
* Corresponding author. Tel.: þ39 06 4991 3812; fax: þ39 06 4991 3133.
** Corresponding author. Tel.: þ39 0577 234330; fax: þ39 0577 234333.
E-mail addresses: mariangela.biava@uniroma1.it (M. Biava), manettif@unisi.it
(F. Manetti).
0223-5234/$ – see front matter Ó 2009 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2009.06.005

M. Biava et al. / European Journal of Medicinal Chemistry 44 (2009) 4734–4738
4735
NCH3
S
N
N
S
N
N
N
N
Cl
1a, BM 212
MIC =1 µg/mL
PI = 4
1b
1c
MIC = 1 µg/mL
MIC = 0.4 µg/mL
Cl
F
PI = 32
PI = 20
S
S
S
N
N
N
F
N
N
N
F
F
F
F
1d
1e
1f
MIC = 0.5 µg/mL
MIC = 1 µg/mL
MIC = 0.5 µg/mL
PI = 8
PI = 32
PI = 16
S
S
N
N
N
N
F
H3C
1g
1h
MIC = 1 µg/mL
MIC = 0.25 µg/mL
F
F
PI = 8
PI = 256
Chart 1. Schematic representation of 1,5-diarylpyrroles 1b–h analogues of BM 212 (1a).
activity: (i) a substituted phenyl ring at both the positions 1 and 5
(F, Cl, and CH3 were the best substituents) and (ii) an amino methyl
group at position 3 (a thiomorpholinomethyl side chain was the
optimal moiety) (Chart 1).
On this basis, to enlarge SAR considerations on such a class of
compounds, and in view to verify the importance of the combi-
nation of both structural and lipophilic factors for the activity, we
planned the synthesis of the new derivatives 2a–h bearing an
ethyl group at position 2 of the pyrrole nucleus, while keeping, on
both N1 and C5 phenyl rings, the same substituents that gave the
best results in terms of activity in previous 2-methyl derivatives
[8b–8e]. Moreover, a thiomorpholinomethyl side chain at position
3 of the pyrrole was maintained, with the exception of compound
2a (the corresponding ethyl derivative of the lead compound 1a,
in which the thiomorpholine was replaced by a N-methylpiper-
azine moiety). The new derivatives 2a–h were evaluated for their
activity toward various strains of MTB, in comparison to 1a and
current antitubercular drugs used as reference compounds.
Moreover, SAR studies also suggested that a relationship does
exist among MIC values and the lipophilic character of pyrrole
derivatives [8a, 8b], leading us to the hypothesis that an increase of
lipophilicity could improve antimycobacterial activity of such
a class of compounds (as also stated by literature reports) [9–11].
As examples, the most active compounds (bearing a propyl or
isopropyl substituent at one of the two phenyl rings of the diary-
lpyrrole scaffold) showed the highest calculated log P values among
our pyrrole derivatives (7.23 and 7.03, respectively) and an activity
higher with respect to that of their ethyl or methyl congeners [8a].
X
R2
N
c
a
b
O
O
+
N
R2
H
R2
O
N
R2
3a-e
O
6a-h
4
R1
5a-e
2a-h
R1
Compounds. 3a: R2 = 4-Cl; 3b: R2 = 4-CH3; 3c: R2 = H; 3d: R2 = 4-F; 3e: R2 = 2-F; 5a: R2 = 4-Cl; 5b: R2 = 4-CH3;
5c: R2 = H; 5d: R2 = 4-F; 5e: R2 = 2-F; 6a: R2 = 4-Cl; R1 = 4-Cl; 6b: R2 = 4-CH3; R1 = 4-F; 6c: R2 = H; R1 = H; 6d:
R2 = 4-F; R1 = H; 6e: R2 = H; R1 = 4-F; 6f: R2 = 4-F; R1 = 4-F; 13g: R2 = 4-F; R1 = 2-F; 6h: R2 = 2-F; R1 = 2-F; 2a:
R1 = 4-Cl; R2 = 4-Cl; X = N-CH3; 2b: R2 = 4-CH3; R1 = 4-F; X = S; 2c: R2 = H; R1 = H; X = S; 2d: R2 = 4-F; R1 =
H; X = S; 2e: R2 = H; R1 = 4-F; X = S; 2f: R2 = 4-F; R1 = 4-F; X = S; 2g: R2 = 4-F; R1 = 2-F; X = S; 2h: R2 = 2-F;
R1 = 2-F; X = S. Reagents. a: 1) 3-Ethyl-5-(2-hydroxyethyl)-4-methylthiazolium bromide, NEt3, mw, 15 min; 2) 2N
HCl. b: amine, p-toluensulfonic acid, mw, 30 min. c: 1) N-methylpiperazine (2a) or thiomorpholine (2b-h), CH3CN,
HCHO, CH3COOH, rt, 1 h; 2) NaOH 20 w/v.
Scheme 1.

4736
M. Biava et al. / European Journal of Medicinal Chemistry 44 (2009) 4734–4738
2. Chemistry
hand, the best compound 2b showed the same MIC value previ-
ously found for the corresponding 2-methyl derivative (0.25 g/
m
Synthesis of the target compounds 2a–h is shown in Scheme 1.
Reaction of a suitable benzaldehyde 3a–e with ethyl vinyl ketone 4,
according to the Stetter reaction, by employing the Discovery
Microwave System apparatus (150 W, 70 ꢀC, 170 psi) [8a], afforded
1,4-diketones 5a–e. In the presence of the appropriate amine,
following the Paal–Knoor condensation conditions for 30 min, by
employing the Discovery Microwave System apparatus (150 W,
170 ꢀC, 170 psi) [8a], intermediates 5a–e cyclized to yield the
expected 1,5-diarylpyrroles 6a–h. Construction of the side chain at
C3 was achieved in good yield by reaction of 6a–h with formal-
dehyde and N-methylpiperazine or thiomorpholine, according to
the Mannich reaction conditions, to give the expected derivatives
2a or 2b–h, respectively.
mL) [8b], but a significantly higher PI value (>512 versus 256) due
to the very low cytotoxicity of this new compound
(MNTD > 128
toward both MTB H37Rv and MTB rifampicin-resistant strain (both
activity values are 0.25 g/mL). Such results showed that higher
mg/mL). The same compound was also very active
m
log P values were not always able to determine better activity and
led us to revise our previous hypothesis that an increase of log P
was in general associated with an improvement of activity. In
particular, the fact that the increase of log P caused by the insertion
at the position 2 of an ethyl group (instead of a methyl one) was not
profitable to improve the activity, led to the suggestion that the
substituent at position 2 was generally less important for activity in
comparison to substituents at positions 1, 3, and 5. As a conse-
quence, the correlation between higher lipophilicity and anti-
mycobacterial activity may depend on structural modifications
mainly involving the phenyl rings at N1 and C5, already identified
as crucial keys in determining antimycobacterial activity, being also
able to match the hydrophobic features of the pharmacophoric
model (Fig. 1) [8a, 8b].
3. Results and discussion
Compounds 2a–h were preliminary evaluated for their in vitro
activity toward MTB CIP 103471 and M. avium CIP 103317. Only
compounds showing MIC values lower than 16 mg/mL were also
assayed toward MTB H37Rv and MTB strains resistant to rifampicin.
Cytotoxicity in VERO cells was determined (and expressed as
maximum non-toxic dose, MNTD) and PI was calculated as the ratio
between MNTD and MIC values. Rifampicin, streptomycin, and 1a
were used as reference compounds. Biological data are reported in
Table 1.
4. Conclusions
Among new 1,5-diarylpyrrole derivatives, the best compound
2b showed a very good biological profile, having the same MIC
value toward MTB previously found for the corresponding
Analysis of the activity data toward MTB CIP 103471 clearly
showed that compounds 2a, 2c–e, and 2h were all characterized by
MIC values two- to fourfold higher than those found for the cor-
responding 2-methyl derivatives [8c, 8e] (even though some of
them possessed log P values comparable to those of previous very
active compounds of this class), with a significant cytotoxicity thus
leading to low values of PI. Compound 2f was inactive. On the other
2-methyl analogue (0.25
improved (>512 versus 256) because of the very low cytotoxicity
(MNTD > 128 g/mL). Compound 2b proved also to be very active
against both MTB H37Rv and MTB rifampicin-resistant strains,
being 0.25 g/mL both the MIC values. Since the biological data of
mg/mL) and a PI value significantly
m
m
2b were very promising in terms of high activity and low cyto-
toxicity, they prompted us toward in vivo investigations. In parallel,
Table 1
Structure, in vitro antimycobacterial activity (toward Mycobacterium tuberculosis CIP 103471, M. tuberculosis H37Rv, rifampicin-resistant M. tuberculosis and M. avium CIP
103317), cytotoxicity, and protection index of the new pyrrole derivatives 2a–h and reference compounds.
N
X
R2
R
N
1a: X = N-CH3; R = CH3
2a: X = N-CH3; R = C2H5
R1
2b-h: X = S; R = C2H5
Compoundsa
R1
R2
MIC (
m
g/mL)
MNTD (
m
g/mL)
PIc
clog Pd
M. tuberculosis strain
M. avium 103317
103471
H37Rv
Rifa-Rb
1a, BM 212
4-Cl
4-Cl
4-F
H
4-Cl
4-Cl
4-CH3
H
4-F
H
4-F
4-F
2-F
1
2
0.25
2
1
2
16
2
4
0.50
0.25
1
4
0.25
4
1
4
ND
2
8
0.50
0.25
ND
4
0.5
4
16
8
4
8
4
4
6.02
6.69
6.53
5.83
6.04
6.04
6.25
6.25
6.25
2a
2b
2c
2d
2e
2f
0.25
>128
16
8
16
8
8
8
128
64
>512
4
1
4
8
8
8
0.5
4
2
H
8
4-F
4-F
2-F
2-F
16
16
8
16
8
ND
2g
2h
Se
2
8
0.50
256
256
Rf
>100
0.25
a
Compounds 2a–h have been submitted to a PCT patent (see Ref. [8d]).
Rifa-R: rifampicin-resistant.
PI: protection index, expressed as MNTD/MIC ratio.
Calculated according to the Alog P98 method (Ref. [4]).
S: streptomycin.
b
c
d
e
f
R: rifampicin.

M. Biava et al. / European Journal of Medicinal Chemistry 44 (2009) 4734–4738
4737
silica gel and petroleum ether/ethyl acetate (3:1 v/v). The eluates
were combined after TLC control and the solvent was removed to
give 2a–h as solids in satisfactory yield. Re-crystallization from
diethyl ether gave the required products. Structure of the final
compounds was confirmed by observing the disappearance of the
singlet of the proton at the position 3 of the pyrrole (6.09 ppm) and
the appearance of the singlet corresponding to the methylene
group at 3.50 ppm [8a]. The elemental analyses for compounds 2a–
h are reported in Table 2 (Supporting Information).
5.3. 2-Ethyl-1-(4-methylphenyl)–3-(thiomorpholin-4-yl)-methyl-
5-(4-fluorophenyl)-1H-pyrrole (2b)
Mp 165 ꢀC (yield 45%); 1H NMR (CDCl3) 7.26 (m, 4H), 7.16 (m,
2H), 7.00 (m, 2H), 6.27 (s, 1H), 3.46 (s, 2H), 2.78 (s broad, 4H), 2.71 (s
broad, 4H), 2.25 (q, 2H), 2.18 (m, 3H), 0.85 (t, 3H). 13C NMR (CDCl3)
160.18 (d, 1C), 139.25 (s, 1C), 135.90 (s, 1C), 132.17 (s, 1C), 129.53 (s,
1C), 129.49 (s, 1C), 128.99 (s, 2C), 128.73 (s, 2C), 127.64 (s, 1C), 114.91
(s, 2C), 114.69 (s, 2C), 110.58 (s, 1C), 55.44 (s, 2C), 54.89 (s, 1C), 28.17
(s, 2C), 18.02 (s, 2C), 14.87 (s, 1C). Anal. (C24H27FN2S) C, H, N, S, F.
5.4. Microbiology
5.4.1. Compounds
Compounds 2b–h and reference drugs were dissolved in DMSO
at
a concentration of 5 mg/mL and stored cold until used.
Compound 2a was dissolved in ethanol at 6 mg/mL.
5.4.2. Antimycobacterial activity
Compounds were preliminarily assayed according to protocol
already described [8b], following a standard twofold agar dilution
method in Middlebrook 7H11 agar medium (Difco) containing 10%
of OADC (oleic acid, albumin and dextrose complex) [12].
Compounds with better preliminary activity were tested toward
M. tuberculosis CIP 103471, M. tuberculosis H37Rv ATCC 27294,
rifampicin-resistant M. tuberculosis ATCC 35838, and atypical
mycobacteria, such as M. avium CIP 103317. In all cases, MIC value for
each compound was determined for each mycobacterial strain.
Further details are in Supporting Information.
Fig. 1. Graphical representation of the superposition mode of 2a (green) and 1a (red)
into the pharmacophoric model. Side chains at positions 2 and 3 of 2a are oriented in
opposite directions with respect to the pyrrole plane and are only able to partially
satisfy the hydrogen bond acceptor group (green spheres) and the directionality
constraint imposed by one of the aromatic ring features (orange spheres, bottom right
corner). The hydrophobic region (cyan sphere) is filled by the p-Cl group of both
compounds, as well as the other aromatic ring feature (orange spheres, middle left).
5.4.3. Cytotoxic activity assays
further efforts are planned for a structural optimization of this class
of compounds.
Cytotoxic activity assays were performed in Vero cells to
determine the maximum non-toxic dose defined as the highest
drug concentration showing no morphological changes as deter-
mined microscopically by the observation at 72 h of incubation.
5. Experimental section
5.1. Chemistry
5.5. Computational details
The general procedure for the preparation of hexane-1,4-diones
5a–e and 1,5-diarylpyrroles 6a–h is described in Supporting
Information.
The QSARþ module of Cerius2 [13] was used to calculate
Alog P98 values of the studied compounds. Such a descriptor is an
implementation of the atom type-based Alog P method using the
latest published set of parameters [14].
5.2. General procedure for the preparation of compounds 2a–h
Acknowledgements
Following the Mannich reaction, to a stirred solution of an
appropriate pyrrole 6 (5.6 mmol) in acetonitrile (20 mL), a mixture
of thiomorpholine (0.57 g, 5.6 mmol) or N-methylpiperazine
(0.56 g, 5.6 mmol), formaldehyde (0.18 g, 5.6 mmol) (40% in water)
and 5 mL of acetic acid, was added dropwise. After the addition was
complete, the mixture was stirred at room temperature for 1 h. The
mixture was then treated with a solution of sodium hydroxide
(20%, w/v) and extracted with ethyl acetate. The organic extracts
were combined, washed with water and dried. After removal of
solvent, the residue was purified by column chromatography, using
`
Financial support from the Italian Ministero dell’Universita e
della Ricerca (PRIN 2005037820) and CARIPLO (Rif. 2006.0880/
10.8485 Bando 2006) is gratefully acknowledged.
Appendix A. Supplementary information
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.ejmech.2009.06.005.

4738
M. Biava et al. / European Journal of Medicinal Chemistry 44 (2009) 4734–4738
[7] (a) A systematic review of delay in the diagnosis and treatment of tubercu-
References
losis, BMC Public Health 8 (2008) 15;
(b) J.C. Sacchettini, E.J. Rubin, J.S. Freundlich, Nat. Rev. Microbiol. 6 (2008)
41–52.
[1] W.W. Yew, C.C. Leung, Am. J. Respir. Crit. Care Med. 177 (2008) 479–485.
[2] Global Tuberculosis Control: Surveillance, Planning, Financing WHO Report
2008, World Health Organization, 2008, (WHO/HTM/TB/2008.393).
[3] (a) R.A. Breen, L. Swaden, J. Ballinger, M.C. Lipman, Drugs 66 (2006) 2299–2308;
(b) K.D. Phillips, J. Assoc. Nurses Aids Care 18 (2007) 75–78;
(c) V. Idemyor, J. Natl. Med. Assoc. 99 (2007) 1414–1419;
(d) D.J. Pepper, G.A. Meintjes, H. McIlleron, R.J. Wilkinson, Drug Discov. Today
12 (2007) 980–989.
[8] (a) M. Biava, G.C. Porretta, G. Poce, A. De Logu, M. Saddi, R. Meleddu,
F. Manetti, E. De Rossi, M. Botta, J. Med. Chem. 51 (2008) 3644–3648;
(b) M. Biava, G.C. Porretta, G. Poce, S. Supino, D. Deidda, R. Pompei, P. Molicotti,
F. Manetti, M. Botta, J. Med. Chem. 49 (2006) 4946–4952;
(c) M. Biava, G.C. Porretta, G. Poce, D. Deidda, R. Pompei, A. Tafi, F. Manetti,
Bioorg. Med. Chem. 13 (2005) 1221–1230;
(d) M. Biava, M. Botta, D. Deidda, F. Manetti, R. Pompei, G.C. Porretta.
WO2006092822, OC/ACT/PCT 92767. (e) M. Biava, G.C. Porretta, D. Deidda,
R. Pompei, A. Tafi, F. Manetti, Bioorg. Med. Chem. 11 (2003) 515–520.
[9] G. Navarrete-Vazquez, G.M. Molina-Salinas, Z.V. Duarte-Fajardo, J. Vargas-
Villarreal, S. Estrada-Soto, F. Gonzalez-Salazar, E. Hernandez-Nunez, S. Said-
Fernandez, Bioorg. Med. Chem. 15 (2007) 5502–5508.
[10] R. Ragno, G.R. Marshall, R. Di Santo, R. Costi, S. Massa, R. Pompei, M. Artico,
Bioorg. Med. Chem. 8 (2000) 1423–1432.
[11] Z.J. Zhu, O. Krasnykh, D. Pan, V. Petukhova, G. Yu, Y. Liu, H. Liu, S. Hong,
Y. Wang, B. Wan, W. Liang, S.G. Franzblau, Tuberculosis 88 (2008)
S49–S63.
[12] NCCLS, The National Committee for Clinical Laboratory Standards NCCLS
document M24-A, Susceptibility Testing of Mycobacteria, Nocardiae, and
Other Aerobic Actinomycetes; Approved Standard, vol. 23, NCCLS, Wayne, PE,
2003, Number 18.
[13] Cerius2 Version 4.8.1 Accelrys, Inc., Scranton Road, San Diego, CA.
[14] A.K. Ghose, V.N. Viswanadhan, J.J. Wendoloski, J. Phys. Chem. 102 (1998)
3762–3772.
[4] (a) M.A. Espinal, Tuberculosis 83 (2003) 44–51;
(b) L.P. Ormerod, Br. Med. Bull. 73–74 (2005) 17–24;
(c) J.A. Caminero, Int. J. Tuberc. Lung Dis. 10 (2006) 829–837;
(d) M. Biava, G.C. Porretta, D. Deidda, R. Pompei, Curr. Drug Targets Infect.
Disord. 6 (2006) 159–172;
(e) W.W. Yew, C.C. Leung, Respirology 13 (2008) 21–46.
[5] (a) Revised definition of extensively drug-resistant tuberculosis, MMWR
Morb. Mortal. Wkly. Rep. 55 (2006) 1176;
(b) S.E. Dorman, R.E. Chaisson, Nat. Med. 13 (2007) 295–298;
(c) R.C. Goldman, K.V. Plumley, B.E. Laughon, Infect. Disord. Drug Targets 7
(2007) 73–91.
[6] (a) H. Tomioka, Curr. Pharm. Des. 12 (2006) 4047–4070;
(b) L.E. Ziganshina, S.B. Squire, Cochrane Database Syst. Rev.
CD004795;
1 (2008)
(c) L. Ballell, R.A. Field, K. Duncan, R.J. Young, Antimicrob. Agents Chemother.
49 (2005) 2153–2163;
(d) M. Spigelman, S. Gillespie, Lancet 367 (2006) 945–947;
(e) Y.L. Janin, Bioorg. Med. Chem. 15 (2007) 2479–2513.