A new DNA gyrase inhibitor subclass of the cyclothialidine family based on a bicyclic dilactam-lactone scaffold. Synthesis and antibacterial properties

Article abstract of DOI:10.1021/jm1014023

The DNA gyrase inhibitor cyclothialidine had been shown to be a valuable lead structure for the discovery of new antibacterial classes able to overcome bacterial resistance to clinically used drugs. Bicyclic lactone derivatives containing in their 12-14-membered ring a thioamide functionality were reported previously to exhibit potent antibacterial activity against Gram-positive bacteria. Moderate in vivo efficacy, however, was demonstrated only for derivatives bearing hydrophilic substituents, which were found to have a favorable impact on pharmcokinetics, and to reduce metabolic degradation, in particular glucuronidation. The incorporation of an additional amide unit into the 14-membered monolactam-lactone scaffold of cyclothialidine analogues provided a new "dilactam" subclass of DNA gyrase inhibitors of inherently higher polarity. After adjusting their lipophilicity by methyl-halogen exchange at the benzene ring, compounds of this series did not require the thioamide functionality to exert a decent antibacterial potency and consequently exhibited improved pharmacokinetic properties resulting in a pronounced in vivo efficacy in a mouse septicaemia infection model.

Full text of DOI:10.1021/jm1014023

ARTICLE
pubs.acs.org/jmc
A New DNA Gyrase Inhibitor Subclass of the Cyclothialidine Family
Based on a Bicyclic Dilactam-Lactone Scaffold. Synthesis and
Antibacterial Properties†
Peter Angehrn, Erwin Goetschi,* Hans Gmuender,‡ Paul Hebeisen, Michael Hennig, Bernd Kuhn,
Thomas Luebbers, Peter Reindl, Fabienne Ricklin, and Anne Schmitt-Hoffmann§
F. Hoffmann-La Roche Ltd., Discovery Chemistry, CH-4070 Basel, Switzerland
S Supporting Information
b
ABSTRACT: The DNA gyrase inhibitor cyclothialidine had
been shown to be a valuable lead structure for the discovery of
new antibacterial classes able to overcome bacterial resistance to
clinically used drugs. Bicyclic lactone derivatives containing in
their 12-14-membered ring a thioamide functionality were
reported previously to exhibit potent antibacterial activity
against Gram-positive bacteria. Moderate in vivo efficacy, how-
ever, was demonstrated only for derivatives bearing hydrophilic
substituents, which were found to have a favorable impact on
pharmcokinetics, and to reduce metabolic degradation, in
particular glucuronidation. The incorporation of an additional amide unit into the 14-membered monolactam-lactone scaffold of
cyclothialidine analogues provided a new “dilactam” subclass of DNA gyrase inhibitors of inherently higher polarity. After adjusting
their lipophilicity by methyl-halogen exchange at the benzene ring, compounds of this series did not require the thioamide
functionality to exert a decent antibacterial potency and consequently exhibited improved pharmacokinetic properties resulting in a
pronounced in vivo efficacy in a mouse septicaemia infection model.
’ INTRODUCTION
knowledge of the structure-activity relationships (SAR) with
regard to the DNA gyrase inhibitory activity and shown the
importance of physicochemical properties for both antibacterial
in vitro activity and in vivo efficacy. Furthermore, the ambivalent
role of one of the phenol groups of 1 had been revealed: on one
hand, the hydroxy group ortho to the thiomethyl side chain of the
benzene ring was found to be essential for the binding to the
enzyme,11 and we have early on postulated its role as a hydrogen-
bond donor and acceptor. On the other side, this main anchor
group I of the cyclothialidine inhibitor family was also found to
be its Achille’s heel for being a target for glucuronidation, leading
to rapid inactivation and clearance from the plasma of promising
derivatives such as the bicyclic lactones 2 and 3. Phenol
conjugation as well as strong plasma protein binding were
considered by us as major factors preventing the translation of
the antibacterial potency into an adequate in vivo efficacy.
Following our working hypothesis suggesting more hydrophilic
derivatives to be less affected by these drawbacks, we eventually
found analogues, e.g. the hydroxymethyl derivatives 4 and 5, that
displayed both improved pharmacokinetic properties and in vivo
efficacy.10
The emergence of resistance to clinically used drugs was
recognized early on as a threat in the treatment of bacterial
infections.1 Bacterial resistance mechanisms caused by mutations
in the binding site can be directed to a specific therapeutic agent,
but cross-resistance to other members of the same family is quite
common and can also affect different chemical classes. The use of
a different mode of action, in particular the addressing of novel
bacterial targets, is an obvious way out of this problem.2
The success of the quinolones in the 1980s had reinforced the
interest for DNA gyrase as a bacterial target.3 By using target-
based screening methods, cyclothialidine (1) and congeners of it
were discovered independently by research groups at Roche4 and
at Glaxo5 (GR122222X = cyclothialidine C6) and were shown to
be potent inhibitors of the ATPase activity conferred by the B
subunit of DNA gyrase (Chart 1).7 This target had been already
validated by the coumarin class of antibiotics,8 e.g. novobiocin,
which however, since the 1980s, was hardly used clinically.
Therefore, new subunit B inhibitors of DNA gyrase were
considered to have a potential for the development of drugs
devoid of cross-resistance with established antibacterial agents.
Although lacking activity against intact bacterial cells, cy-
clothialidine was followed up by the Roche group in a medicinal
chemistry program which provided analogues of potent anti-
bacterial activity covering broadly the spectrum of Gram-positive
bacteria.9,10 This early, ligand-based work had established a broad
The difficulties to achieve in vivo efficacy with cyclothialidine
derivatives had raised interest in alternative subclasses. We had
discovered that also seco-cyclothialidines can be good DNA
Received: October 28, 2010
Published: March 09, 2011
r
2011 American Chemical Society
2207
dx.doi.org/10.1021/jm1014023 J. Med. Chem. 2011, 54, 2207–2224
|

Journal of Medicinal Chemistry
ARTICLE
Chart 1
’ CHEMISTRY
The preparation of new bicyclic lactones relied on the
synthetic pathways described previously.10 In general, a primary
amine (6, 14, 21, 34b,c, 35b,c, 82) was acylated with an
appropriate linker unit (e.g., 7, 15, 22, 44). If required, the
hydroxy function at the C-5 position of the linker moiety was
unmasked and other functional groups were optionally pro-
tected, whereupon the allyl ester was cleaved and the resulting
ω-hydroxy-carboxylic acid was lactonized. A reaction sequence
consisting mainly in optional thiation of the lactam group,
optional derivatization, and eventual cleavage of protecting
groups provided the target compounds.
In the preparation of 12 and 13 (Scheme 1), ketal 8 was
hydrolyzed and the primary alcohol was protected as trityl ether
(9). Palladium-assisted cleavage of the allyl ester of 9 also
caused partial cleavage of the allyl carbamate group, which had
to be reestablished by subsequent treatment of the crude
product with allyl chloroformate. After the lactonization, thia-
tion, and trityl ether cleavage steps, the resulting alcohol 10 was
converted to the target compounds. For the syntheses of the 13-
hydroxymethyl derivatives 19a,b and 20a,b, amine 1410 was
acylated with dioxolane 7.23 Upon hydrolysis of ketal 16 and
protection of the primary alcohol as trityl ether, the mixture of
diastereomeric allyl esters 17a/17b was converted into the
lactones 18a and 18b, which were separated by chromatogra-
phy. The pure isomers were converted to the target compounds
19a,b and 20a,b in the usual manner. The stereochemical
assignment for the 13-hydroxymethyl substituent was made
possible by an X-ray crystal structure of lactone 19b.24 Analo-
gously, the 9-oxa lactone 24 was assembled starting from amine
2110 and linker-acid 22.25
For the synthesis of the 1-chloro- and 1-bromo-lactones 36-
41 (Scheme 2), the required amino intermediates 34b,c and 35b,
c were prepared starting from methyl ester 25.26 Treatment of 25
with sulfuryl chloride in dichloromethane provided a mixture of
regioisomers from which the major component 26 could be
isolated in pure form by crystallization. Saponification of the ester
26 afforded pseudoacid 27b, which in its NMR spectrum
(DMSO-d6) also displayed signals of the corresponding o-formyl
benzoic acid (ca. 20%). For the preparation of the corresponding
bromo-derivative (27c), 25 was converted to pseudoacid 28,
which subsequently was brominated in dichloromethane with
high regioselectivity. The resulting bromide 27c exhibited the
same pseudoacid/acid ratio in DMSO as 27b. Treatment of 27b,
c with tetramethylguanidine and allyl bromide in DMF followed
by silylation of the phenol group afforded aldehydes 30b,c. As
reductive thiolation27 of 30b,c with the cysteine derivatives 32 or
3310 was accompanied by extensive aldehyde to methyl reduc-
tion and partial debromination in the case of 30c, the benzyl
iodides 31b,c were prepared and used for the alkylative coupling
with thiols 32 and 33.28 Intermediates 34b,c and 35b,c were
converted to the target compounds 36-41, respectively, by
subjecting them to the reaction sequence described previously
for the preparation of 98, 3, and 100 (Table 1) from amines 14
and 6, respectively.10
gyrase inhibitors, a finding that was also taken up by others,9b,12
but for this as well as for other phenol subseries of the general
structure I,13 it so far has not been possible to combine a useful
level of antibacterial activity with appropriate physicochemical
properties. Other targeted screening programs have provided
nonphenolic DNA gyrase subunit B inhibitors. Triazines14 and
alkyl-urea substituted N-heterocycles15,16 were discovered by
screening for ATPase inhibitors, pyrazoles by assaying the
inhibition of chromosome partitioning in Escherichia coli,17 and
pyrrole-2-carboxamides by using a NMR guided screen.18 Struc-
tural information on the binding mode of cyclothialidine,19,20
novobiocin,20 but also of the ATP analogue ADPNP,21 to the
ATP binding site has been used more and more to guide the
optimization of lead structures, and in addition, it allowed to
apply fragment-based in silico tools in the search for new classes
of DNA gyrase inhibitors. Thus, a biased in silico “needle”
screening for potential inhibitors featuring a generic hydrogen-
bond donor/acceptor motif II, followed up by an extended hit
validation process, provided several new nonphenolic DNA
gyrase inhibitor classes such as indazoles and 2-hydroxymethyl-
indoles.22 Among the various nonphenolic series of subunit B
inhibitors described in the past decade, many still represent only
potential starting points for an optimization program. Others,
however, such as ethyl-urea-benzimidazoles15 or pyrrole-2-
carboxamides16 have been developed to potent antibacterials
which show promising efficacy in vivo.
We have described for cyclothialidine analogues how well-
balanced physicochemical properties are required to allow
translation of potent enzyme inhibitory activity into significant
efficacy in the therapy of infections.10 Prototypical illustration of
this concept is the conversion of the in vivo inactive, lipophilic
lactones 2 and 3 into efficacious analogues 4 and 5, respectively,
the difference being made by an additional hydroxymethyl
substituent. Herein we report on a new subclass of the cyclothia-
lidine inhibitor family, which emerged from our continued efforts
to improve on the conversion of potent enzyme inhibitory
activity into significant in vivo efficacy within the 14-membered
lactone series. To this end, we investigated not only new
substitution patterns on the parent monolactam-lactones but
also the incorporation of polar elements into the bicyclic scaffold.
For the preparation of dilactams 47-50, amine 14 was
coupled with glycine derivative 44, and the resulting product
was converted to the bicyclic lactone 47 in 4 steps. Heating of
dilactam 46a with 0.8 equivalent of Lawesson’s reagent in
toluene provided three products which were separated by
chromatography and thereafter desilylated by treatment with
2208
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Scheme 1a
a Reagents and conditions: (a) (R)-4-(2,2-dimethyl-[1,3]dioxolan-4-yl)-butyric acid (7), 3-(2,2-dimethyl-[1,3]dioxan-5-yl)-propionic acid (15), or (2-
trityloxy-ethoxy)-acetic acid (22), for 8, 16, or 23, respectively, EDC, MeCN, 0 °C; (b) 5 N HCl-EtOH, rt; (c) tritylpyridinium trifluoroborate, MeCN,
rt; (d) (i) Pd(PPh3)4, 4-TMS-morpholine, AcOTMS, DCM, rt, (ii) ClCOOCH2CHCH2, NMM, DCM, 0 °C; (e) DEAD, PPh3, toluene, 0-20 °C; (f)
Lawesson's reagent, toluene, 80 °C; (g) pTsOH, MeOH, 20-60 °C; (h) Pd(PPh3)4, Me2N-TMS, TFA-TMS, DCM, 0 °C; (i) NH4F, MeOH, rt; (j)
NaNO2, 25% aq AcOH, 0 °C; (k) ClC(Ph)3, pyridine, rt; (l) Pd(PPh3)4, morpholine,THF, 0 °C.
NH4F in methanol, to give the two regioisomeric monothioxo
derivatives 48 and 49 as well as the 8,11-dithioxo product 50. The
structural assignment for the monothioxo products is based on
the marked low-field shift of the NH-proton of the thioamide
functionality. Using the same reaction sequence, the chloro- and
bromo-analogues 51, and 52-54, were prepared from the amines
34b and 34c, respectively.
Amine 34c also served for the syntheses of lactones 58-59,
63-66, and 71-78 (Scheme 3). For the 9-methyl derivatives
58-59, the linker was introduced stepwise, whereas for the other
syntheses the readily available linker entities 60a,29 60b, and 6730
were attached in one step. The cyclic amines 71 and 78 were
obtained from 9-Boc intermediate 69 by optional thiation and
cleavage of protecting groups. N-acylation of amine 70 afforded
derivatives 72-77.
For the preparation of tertiary amine 81 (Scheme 4), inter-
mediate 34c was heated with lactone 79, and the resulting
acylation product 80 was subjected to the usual reaction
sequence. To modify the substituent of the oxadiazole ring,
carboxylic acid 85 was synthesized starting from benzyl iodide
31c and L-cysteine methyl ester. Coupling of 85 with the amide-
oximes 86a,b followed by thermal cyclization provided analo-
gues 87 and 88. Further variations of the oxadiazole substitu-
tion were accomplished by modifying the aminomethyl
substituent of lactones 91 and 92 using procedures described
previously.10
2209
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Scheme 2a
a Reagents and conditions: (a) SO2Cl2, DCM, rt; (b) 2 N NaOH, rt; (c) Br2, DCM, 0-20 °C; (d) allyl bromide, 1,1,3,3-tetramethylguanidine, DMF, rt;
(e) thexyldimethylchlorosilane, NEt3, DMF, 0-20 °C; (f) tetramethyldisiloxane, TMSCl, NaI, MeCN, 0 °C; (g) 32 or 33, respectively, NEt3, DCM, rt;
(h) prepared in analogous manner as 99, 3, and 100 starting from 34b,c or 35b,c, respectively, see Experimental Section; (i) Gly-OEt, EDC, NMM,
MeCN, 0 °C; (j) KOH, MeOH, 60 °C; (k) 44, EDC, MeCN, 0 °C; (l) pTsOH, MeOH, rt; (m) Pd(PPh3)4, morpholine,THF, 0 °C; (n) DEAD, PPh3,
toluene, 0-20 °C; (o) NH4F, MeOH; (p) Lawesson's reagent, toluene, 80 °C.
Conformational Aspects of the 14-Membered Lactones.
During our previous work with 12-14-membered monolactams,
e.g. 2 or 3, the AX2 spin system of the protons of the cysteamine
entity was found to display in various solvents always a large and a
small vicinal coupling constant, suggesting a high population of a
conformation with a synclinal/antiperiplanar arrangement of
these protons. This conformational feature was also found for
1 when bound to the ATPase site of DNA gyrase.19,20 In the
NMR spectrum of the excellent dilactam inhibitors, however,
two small- to medium-sized vicinal coupling constants for the
cysteamine protons were observed, indicating a synclinal/syncl-
inal arrangement to be predominant. Interestingly, the X-ray
analyses of the 14-memberd lactones 3 and 52 display ring
conformations which do not correspond to the conformational
features observed in solution (Figure 1). The crystal structure of
monolactam 3 exhibits a synclinal/synclinal relation for the
cysteamine hydrogens with the oxadiazole ring in a pseudoaxial
position, whereas in that of dilactam 52, the cysteamine hydro-
gens are synclinal/antiperiplanar and the oxadiazole is in a
pseudoequatorial position. Both structures show an intramole-
cular hydrogen bond between the N(9/12) hydrogen and the
lactone carbonyl. This hydrogen bond is more pronounced in the
dilactam scaffold where it is part of a perfect β-turn motif. It is
noteworthy that 3D structure generation for 3 and 52 proposes
low-energy conformations, which are in agreement with the
observed coupling constants in the solution NMR spectra.
Whereas the calculation for 3 affords a conformation which
can easily be superimposed with the “binding conformation” of 1,
the conformation of 52 in its bound state is presumably different
from that proposed in Figure 1 (Figure 2).
’ BIOLOGICAL RESULTS AND DISCUSSION
As improving of in vivo efficacy within the class of the bicyclic
14-membered lactones represents a multidimensional problem,
the impact of structural modifications of the lead compounds, 98
and 3 (Table 1), on their physicochemical and enzyme inhibitory
properties, as well as on their antibacterial activity in vitro and
in vivo, will be discussed in parallel. We used calculated log P
(CLOGP) values for all compounds as a rough estimate of
lipophilicity, which we consider (besides the intrinsic enzyme
inhibitory activity) to be the major player in the optimization
process. Lipophilicity/hydrophilicity properties control cell per-
meation, thus antibacterial activity, as well as solubility, protein
binding, and susceptility for glucuronidation, hence pharmaco-
kinetic behavior and in vivo efficacy. Regarding both DNA gyrase
inhibitory activity, expressed as maximum noneffective concen-
tration (MNEC) in a supercoiling assay10 and antibacterial
activity in vitro, assessed as minimum inhibitory concentration
(MIC), parent lactam 98 was markedly surpassed by the
2210
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Table 1. DNA Gyrase Inhibitory Activity, in Vitro and in Vivo Antibacterial Activity, and CLOGP of Monolactams. Comparison
with Reference Compounds
compd
R1
R2
R3
R4
X
MNECa(μg/mL)
MICb (μg/mL)
ED50c (mg/kg)
CLOGPd
98f
3f
5f
99f
100f,g
12
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Cl
H
H
H
H
H
H
H
H
H
Me
Me
Me
O
S
0.01
0.25
0.06
0.5
>25
>25
12.5
25
1.70
2.37
1.41
1.06
1.05
0.09
0.10
0.23
0.23
1.62
1.62
1.76
2.58
1.27
1.92
2.73
1.42
0.004
0.005
0.001
0.002
0.005
0.01
CH2OH
S
H
CH2OH
CH2NH2
CH2NH2
CH2OH
Me
S
0.25
0.25
32
H
S
8.5
nde
nd
CH2OH
S
13
CH2OH
S
10
19a
19b
20a
20b
36
H
H
H
H
H
H
H
H
H
H
β-CH2OH
O
O
S
0.005
0.1
8
nd
R-CH2OH
Me
64
nd
β-CH2OH
Me
0.002
0.05
0.5
12.5
>25
>25
nd
R-CH2OH
Me
S
1
H
H
H
H
H
H
Me
O
S
0.005
0.002
0.002
0.005
0.005
0.05
0.25
0.25
0.12
0.12
0.12
0.5
37
38g
Cl
Me
Cl
CH2NH2
Me
S
6
39
Br
O
S
25
40
41g
Br
Me
>25
12.5
Br
CH2NH2
S
a Supercoiling assay, E. coli gyrase, MNEC = maximum noneffective concentration. b Agar dilution (Mueller-Hinton agar); inoculum 104 CFU/spot, MIC =
c
minimum inhibitory concentration, Staphylococcus aureus Smith. In vivo efficacy, septicaemia in mice (S. aureus Smith), iv administration. d CLOGP =
calculated log P (version 4.94, program BioByte Corp., Claremont, CA). e Not determined. f Monolactam reference compounds.10 g Hydrochloride salt.
10-thioxo analogue 3. As a matter of fact, from a large number of
compounds prepared, only thiolactams seemed to have the
potential for in vivo efficacy, which eventually was found for
polar analogues such as the hydroxymethyl derivatives 4, 5, and
99, or the amine 100.10
We started out by investigating further substitution patterns of
the parent lactones 98 and 3. Table 1 shows a compilation of
biological data of novel derivatives of this series in comparison
with that of key compounds (98, 3, 5, 99, 100) from our previous
work in the 14-membered lactone series. The combination of
two hydrophilic substituents in the same molecule provided the
potent enzyme inhibitors 12 and 13, which however displayed a
strongly diminished antibacterial activity, presumably due to
their pronounced hydrophilic character.
We probed systematically other positions of the core lactone
for their suitability to carry polar substituents. Exemplary for this
investigation are the two isomeric thiolactams 20a and 20b
bearing a hydroxymethyl group in the 13β- and 13R-positions,
respectively. We noticed a 25-fold stronger enzyme inhibiton for
the β-hydroxymethyl derivative 20a when compared to the R-
substituted analogue 20b. A similar differentiation had been
observed already for the epimeric 10-oxo analogues 19a and 19b.
Whereas, however, the superiority of the 13β-hydroxymethyl
substitution was also reflected in a better antibacterial activity of
19a, the MICs of 20a and 20b were similar. Nevertheless, only
the 13β-substituted isomer 20a was efficacious in vivo, although
not surpassing the efficacy of our best compound. The fact that
no lactam, i.e. no 10-oxo-analogue displaying in vivo efficacy, had
been identified up to then was explained by the inherently higher
polarity, giving rise to a weak antibacterial activity due to slow cell
penetration. However, the failure of lactam 98 in the in vivo assay
was not obvious. It displayed a good in vitro activity against S.
aureus with only a small “plasma shift” when assessed in the
presence of plasma proteins. We tried to moderately increase the
lipophilicity by replacing the 1-methyl group by a chloro or a
bromo atom. This modification had been shown previously to be
beneficial for the in vitro potency of seco-cyclothialidines.12a We
found that both halogen analogues of 98, i.e. 36 and 39, showed a
slight improvement of both enzyme inhibitory and in vitro
antibacterial activity, but most importantly, bromo analogue 39
displayed in vivo efficacy. The same modifications at thiolactam 3
led to a slightly decreased activity against S. aureus, and the
1-bromo analogue 40 was found inactive in vivo. For the more
polar amine 100, however, the combination of these lipophilic
elements was helpful. In particular, the chloro derivative 38 was
active in the mouse septicaemia model with an ED50 of 6 mg/kg.
Obviously, and in accordance with our working hypothesis, it
appears that halogenation is only helpful for polar compounds
like 98 or 100 but not for lipophilic ones like 3.
For tuning the lipophilicity in the 14-membered lactone series,
we had used until that time four structural elements: the
conservation of the amide functionality itself and polar substit-
uents such as hydroxymethyl or aminomethyl for increasing
hydrophilicity, and the thioamide group and halogenation of the
benzene ring for amplifying hydrophobicity. An alternative way
to enhance the hydrophilic character of target compounds
2211
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Scheme 3a
a Reagents and conditions: (a) for 55, 61a, 61b, or 68, Boc-N(Me)CH2COOH, 60a, 60b, or 67, respectively, EDC, MeCN, 0 °C; (b) TFA, 0 °C; (c)
(Ph3COCH2COOH, HOBT, EDC, MeCN, 0 °C; (d) pTsOH, MeOH, rt; (e) Pd(PPh3)4, morpholine, THF, 0 °C; (f) DEAD, PPh3, toluene, 0-20 °C;
(g) NH4F, MeOH, rt; (h) Lawesson's reagent, toluene, 80 °C; (i) Ac2O, pyridine, 60 °C; (j) cyclopropane-carbonyl chloride, 2,2-dimethyl-propionyl
chloride, benzoyl chloride, or methanesulfonyl chloride, respectively, NEt3, DCM, 0 °C, or L-Ala-OH or L-Leu-OH, respectively, EDC, MeCN, 0 °C.
consists in using a more polar scaffold. In an early attempt to
apply this strategy, it had been found that the cyclic ether 24, the
12-oxa analogue of 101 (atom numbering of structure 101),10
displayed a significantly reduced antibacterial potency, which we
had attributed to unfavorable binding properties as well as to the
increased hydrophilicity when compared to the parent 101
(Figure 3). In contrast, dilactam 47, containing a dipeptide
moiety in the 14-membered lactone ring, was found to maintain
the DNA gyrase inhibitory activity of the parent monolactam 98.
The antibacterial activity, however, was strongly affected by its
polar character, making it also inefficacious in the therapeutic
in vivo assay (Table 2). Therefore, we set out to increase the
lipophilicity of 47 by either applying thiation of the amide groups
or halogenation of the benzene ring. The 11-thioxo analogue 48
was a 10 times more potent DNA gyrase inhibitor, exhibited
excellent antibacterial activity, and unlike its monolactam analo-
gue 3, was found to be efficacious in vivo. The 8-thioxo
regioisomer 49 was inferior to 48. The bis-thioxo product 50,
although equipotent to 48, was not followed up.
exception having been the bromo-lactam 39. Taking into account
the improved properties of halogenated monolactam-lactones,
we also prepared the 4-chloro and the 4-bromo analogues of the
new dilactams. In comparison to 47, chloro-dilactam 51 dis-
played the same modest antibacterial activity but nonetheless was
found to be active in vivo, like 39. Bromo-dilactam 52, however,
being equipotent as enzyme inhibitor, was found to be 8 times
more active in vitro against S. aureus Smith. To our surprise, this
compound showed an excellent efficacy in vivo in the mouse
septicaemia model with an ED50 of 3 mg/kg when compared to
the ED50 values assessed for novobiocin (3 mg/kg) and vanco-
mycin (1.5 mg/kg) in the same assay. The 11-thioxo analogue 53,
despite its superior antibacterial activity, was less efficacious
in vivo than 52, and the even more lipophilic bis-thioxo derivative
54 was found to exhibit both lower enzyme inhibitory and lower
antibacterial activity.
The favorable biological properties of bromo-dilactam 52 let
us consider it as a new lead structure. In particular, the fact that
this compound was devoid of the potentially toxic and metabo-
lically instable thioamide entity was regarded as an advan-
tage.32 We therefore explored further possibilities to optimize
the dilactam scaffold of 52, investigating at first the impact of
substituents. It was found that N-methylation of the newly
The need to introduce a certain level of lipophilicity to warrant
a decent antibacterial activity in vitro and in vivo had been
satisfied up to then by keeping at least one thioamide function-
ality in the 12-14-membered bicyclic scaffolds, the noteworthy
2212
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Scheme 4a
a Reagents and conditions: (a) 79, 1-oxy-pyridin-2-ol, toluene, 80 °C; (b) Pd(PPh3)4, morpholine, THF, 0 °C; (c) DEAD, PPh3, toluene, 0-20 °C; (d)
NH4F, MeOH, rt; (e) L-Cys-OMe, NEt3, DCM, 0 °C; (f) (i) 44, EDC, MeCN, 0 °C, (ii) pTsOH, MeOH, 20 °C; (g) 0.1 N NaOH, aq THF, rt; (h) (i) N-
hydroxy-cyclopropanecarboxamidine (86a) or N-hydroxy-2-methoxy-acetamidine (86b), respectively, DCC, 1-oxy-pyridin-2-ol, DMF, 0 °C, (ii) tolu-
ene, 110 °C; (i) (i) Pd(PPh3)4, 4-TMS-morpholine, AcOTMS, DCM, rt, (ii) ClCOOCH2CHCH2, NMM, DCM, 0 °C; (j) DEAD, PPh3, THF, 0-20
°C; (k) Pd(PPh3)4, (Me)2N-TMS, TFA-TMS, DCM, rt; (l) NaNO2, 25% aq AcOH, 0 °C; (m) Lawesson's reagent, toluene, 80 °C; (n) acetone, NaBH4,
aq AcOH, NaOAc, 0 °C.
introduced amide function (58) only slightly affected MNEC
and MIC values, but that in vivo efficacy was lost. A similar effect
was noted for the introduction of a 7β-methyl substituent (63),
whereas a 7β-phenyl group (65) markedly reduced the enzyme
inhibitory activity. To get a more complete SAR picture, the 11-
thioxo analogues 59, 64, and 66, were also prepared. Displaying
in general a reduced enzyme inhibitory activity, they had slightly
increased antibacterial potency, which, however, was not trans-
lated into in vivo efficacy.
The basic amine 71, formally obtained by reduction of the
8-oxo carbonyl group of 52, represents the 12-aza analogue of 98
(atom numbering of structure 98). In contrast to the lowered
enzyme inhibitory activity of the 9-oxa derivative 24, when
compared to its carba analogue 101, amine 71 was found to be
equally active as both monolactam 98 and dilactam 52. Its
antibacterial properties were less pronounced like those of
dilactam 52, but unlike 52 it showed only borderline in vivo
efficacy. The corresponding thiolactam 78 was assayed inactive
in vivo.
N-acylation (72-77) mostly had a negative impact on the
antibacterial activity, an exception being the pivaloyl derivative
74, which however was inactive in mice.
Modification of the oxadiazole side chain of 52 along the line
exercised earlier for the monolactams10 provided similar results
(Table 4). With the exception of cyclopropyl analogue 87, being
structurally closest to 52, compounds 88, 93, 94, and 96 lost
their in vitro potency, probably due to their strong hydro-
philic character. This is supported by the finding that the 11-
thioxo analogues of 94 and 96, the equipotent inhibitors 95
and 97, regained antibacterial activity, and 97, but not 95, was
also curative in the in vivo assay although four times less than
dilactam 52.
The impact of the three major structural modifications
applied, i.e. polar substituents, methyl/halogen exchange, and
monolactam/dilactam switch, can be summarized as follows:
DNA Gyrase Inhibiton. In the monolactam series, the attach-
ment of 1 or 2 polar substituents was found to be tolerated in
several positions, e.g. 14β, 13β, and the 30-position of the
oxadiazole, but in the 13R-position, it caused a strong reduction
of activity. Both 1-chloro and 1-bromo substituents brought a
2-fold increase of activity for lactam 98, but only 1-chloro
Attempts to improve the biological properties of amine 71 by
further modifications at its amino function were unsuccessful
(Table 3). N-Methylation (82) led to loss of in vivo efficacy, and
2213
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Figure 3. Comparison of biological properties and CLOGP of 12-
carba- and 12-oxa-lactones 101 and 24 (atom numbering of struc-
ture 101).
observed before, lipophilicity strongly influences the antibacterial
potency and thiolactams were typically found to be 2-16 times
more active than their oxo-analogues. As a general rule for mono-
and dilactams, replacement of the methyl group at the phenyl
ring by chloro or bromo was paid off well, the beneficial effect
being, however, more pronounced for polar molecules, like 47 or
98. The incorporation of a second amide entity into the lactam
ring generally lowered the MIC by a factor of 8-32, with the
exception of that against Streptococcus pyogenes (factor 2).
Because of their higher lipophilicity, the original MIC values of
thiolactams (3, 40) were less affected by this added polarity than
those of the corresponding 11-oxo compounds (98, 39).
The antibacterial activity of the most interesting dilactam 52
was further characterized by testing it against a larger number of
Gram-positive and Gram-negative isolates (Table 6). Against a
set of highly resistant MRSA, 52 was found to be only slightly
weaker than vancomycin and 2-4 times less active than line-
zolide. The lead compounds 3 and 5 were best in this compar-
ison. Interestingly, the hydroxymethyl derivative 5 was
equipotent to the parent thiolactam 3, indicating that polar
compounds might be less affected by the alterations presented
by these resistant strains. MIC values of 52 against a set of
vancomycin-resistant strains of Enterococcus faecalis were 2-8
times lower than those of linelozide and imipenem, as well as
those of 3 and 5, and a similar superiority of 52 was found for
Entercoccus faecium strains resistant to vancomycin and imipe-
nem. In line with this finding is the observation that for all
enterococci tested, 5 was more active than the more lipophilic
parent compound 3. For penicillin-resistant Streptococcus pneu-
monia, Haemophilus influencae, and Moraxella catarrhalis, 52 was
found to be 2-8 times more active than the thiolactams 3 and 5
and the reference compound linelozide. In the testing against the
latter two species, MICs of 3 and 5 presumably were negatively
affected by the presence of plasma proteins in the test medium
(“plasma shift”).
Figure 1. X-ray structures and calculated low-energy conformations of
lactones 3 and 52.24 In the solid state, dilactam 52 adopts a β-turn-like
conformation which is stabilized by a strong intramolecular hydrogen
bond (d (O HN) = 2.94 Å), whereas the corresponding hydrogen
3 3 3
bond in thiolactam 3 is longer (d = 3.19 Å), indicating weaker
stabilization of this inherently more flexible ring system. The oxadiazole
substituent in the crystal structure occupies a pseudoaxial position in 3
but a pseudoequatorial positition in 52. Interestingly, CORINA 3D
structure generation31a proposes low-energy conformations where the
heterocycle is pseudoequatorial in 3 and pseudoaxial for 52, which is in
agreement with the coupling constants found in the NMR spectra. The
strong intramolecular hydrogen bridge is preserved in the calculated
conformation of 52 (d = 2.83 Å).
Figure 2. Overlay of 1 (black, structure truncated for clarity of view)
bound to the ATP binding site of DNA gyrase19 with low-energy
conformations of 3 (cyan) and 52 (green). The conformation of 52
with the heterocycle in the pseudoequatorial position was adapted from
the CORINA generated conformation of 3 (Figure 1) with subsequent
energy minimization using Moloc.31b This conformation is within 1
kcal/mol of that of the relaxed X-ray structure of 52.
improved thiolactam 3. The 11-oxo dilactams showed a potent
and robust inhibitory activity slightly better than that of cy-
clothialidine (MNEC 0.05 μg/mL), but they were 2-4 times
weaker compared to 98. Thiation of the 11-oxo group of
dilactams had a favorable effect only for 4-methyl- but not for
the 4-chloro- and the most interesting 4-bromo-derivatives.
Antibacterial Activity in Vitro. For the assessment of the
antibacterial potency we have considered as indicator strain S.
aureus Smith, which was also used in the mouse infection model.
The antibacterial spectrum of the “dilactams” follows closely that
of the “mono-lactams”. It covers broadly the spectrum of Gram-
positive bacteria, but only selected strains of Gram-negative
bacteria, such as Xanthomonas maltophilia, Haemophilus influen-
cae, Neisseria meningitidis, and Moraxella catarrhalis are
susceptible.10 Some general trends in the SAR can be seen from
the MIC values of selected compounds compiled in Table 5. As
In Vivo Efficacy. The in vivo performance of the compounds
displayed in Table 5 can be rationalized in the following way:
Low intrinsic activity (MIC 8 ug/mL) is probably the main
reason why the polar dilactam 47 was found inefficacious.
Although very potent, the lipophilic thiolactams 3 and 40 were
not active in vivo. In addition to unfavorable pharmacokinetic
behavior, their MIC value is strongly affected by binding to
plasma proteins as is evidenced by a “plasma shift” of 8. On the
other side, the polar 98 is not efficacious despite a low MIC and a
small “plasma shift” of 2, so that other parameters, e.g. high
clearance by glucuronidation, have to be brought into play as
explanation. The slightly more potent bromo analogue 39,
2214
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Table 2. DNA Gyrase Inhibitory Activity, in Vitro and in Vivo Antibacterial Activity, and CLOGP of Dilactams: Modifications at
the Core
compd
R1
R2
R3
X
Y
MNECa (μg/mL)
MICb (μg/mL)
ED50c (mg/kg)
CLOGPd
47
48
49
50
51
52
53
54
58
59
63
64
65
66
71f
78f
Me
Me
Me
Me
Cl
H
H
H
H
H
H
H
H
H
Me
Me
H
H
H
H
H
H
O
S
O
0.01
0.001
0.005
0.002
0.02
0.02
0.05
0.2
8
0.5
2
>25
25
0.57
1.00
0.77
1.20
0.63
0.78
1.21
1.41
1.43
1.86
1.30
1.73
2.52
2.94
1.28
1.71
H
O
S
H
O
S
>25
nde
25
H
S
0.5
8
H
O
O
S
O
O
O
S
Br
H
1
3
Br
H
0.12
1
12.5
nd
Br
H
S
Br
H
O
S
O
O
O
O
O
O
H,H
H,H
0.05
0.05
0.05
0.2
2
>12.5
nd
Br
H
0.5
2
Br
Me
Me
Ph
Ph
H
O
S
>12.5
>12.5
nd
Br
0.5
>64
0.5
1
Br
O
S
0.2
Br
1
>25
25
Br
O
0.01
Br
H
S
0.1
0.5
>25
a Supercoiling assay, E. coli gyrase, MNEC = maximum noneffective concentration. b Agar dilution (Mueller-Hinton agar); inoculum 104 CFU/spot,
c
MIC = minimum inhibitory concentration, Staphylococcus aureus Smith. In vivo efficacy, septicaemia in mice (S. aureus Smith), iv administration.
d CLOGP = calculated log P (version 4.94, program BioByte Corp., Claremont, CA). e Not determined. f Hydrochloride salt.
Table 3. Biological Properties and CLOGP of 9-Aza-lactones
compd
R
MNECa (μg/mL)
MICb (μg/mL)
ED50c (mg/kg)
CLOGPd
71f
81f
72
73
74
75
76f
77f
H
0.01
0.05
0.01
0.10
0.05
0.10
0.10
0.10
1
1
25
>25
nde
nd
1.28
1.79
1.45
1.72
2.37
1.45
1.20
2.66
Me
Ac
16
8
(CO)-cProp
(CO)-tBu
SO2Me
0.5
64
>64
16
>12
nd
Ala-H
nd
Leu-H
nd
a Supercoiling assay, E. coli gyrase, MNEC = maximum noneffective concentration. b Agar dilution (Mueller-Hinton agar); inoculum 104 CFU/spot,
c
MIC = minimum inhibitory concentration, Staphylococcus aureus Smith. In vivo efficacy, septicaemia in mice (S. aureus Smith), iv administration.
d CLOGP = calculated log P (version 4.94, program BioByte Corp., Claremont, CA). e Not determined. f Hydrochloride salt.
however, displaying only a modest “plasma shift” of 4, was
assayed efficacious. The chloro analogue 38 was in vivo slightly
better than amine 100 in accordance with its improved MIC
value. Dilactam 97, an N-isopropyl derivative of the amine 95,
was found active in vivo in contrast to the significantly more
polar 95. We have described earlier a reversed case where
N-isopropylation of 100 led to an inefficacious compound.10
Obviously, isopropylation of 95 adjusts lipophilicity appropri-
ately, whereas in the case of 100, it was increased too much.
The best compound 52 was the most polar one among those
found active in vivo. Its modest antibacterial activity in vitro
(MIC = 1 ug/mL against S. aureus Smith) was hardly affected by
2215
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Table 4. Biological Properties and CLOGP of Dilactams: Modification of the Oxadiazole Substituent
compd
R
X
MNECa (μg/mL)
MICb (μg/mL)
ED50c (mg/kg)
CLOGPd
52
87
88
93
94f
95f
96f
97f
Me
O
O
O
O
O
S
0.02
0.02
0.05
0.02
0.10
0.05
0.02
0.02
1
2
3
6
0.78
1.23
c-propyl
CH2OMe
8
nde
nd
nd
>12
nd
12
0.31
CH2OH
32
32
1
-0.53
-0.54
-0.01
0.72
CH2NH2
CH2NH2
CH2NHCHMe2
CH2NHCHMe2
O
S
16
0.5
1.25
a Supercoiling assay, E. coli gyrase, MNEC = maximum noneffective concentration. b Agar dilution (Mueller-Hinton agar); inoculum 104 CFU/spot,
c
MIC = minimum inhibitory concentration, Staphylococcus aureus Smith. In vivo efficacy, septicaemia in mice (S. aureus Smith), iv administration.
d CLOGP = calculated log P (version 4.94, program BioByte Corp., Claremont, CA). e Not determined. f Hydrochloride salt.
Table 5. DNA Gyrase Inhibitory Activity, in Vitro and in Vivo Antibacterial Activity, and Physicochemical Properties of Selected
Mono- and Dilactams: Comparison with Reference Compounds
MICb (μg/mL)
compd
MNECa(μg/mL) E. c. 25922 X. m. 1AC 739 S. a. 25923 S. a. Smith S. e. 16/2 E. f. 6 S. p.b15 ED50c(mg/kg) CLOGPd log De PSf
3g
0.004
0.005
0.01
0.05
0.002
0.002
0.02
0.01
0.005
0.01
0.02
0.1
>64
>64
>64
>64
64
0.5
1
0.12
<0.12
1
0.06
<0.12
0.5
0.01
<0.12
<0.12
<0.12
<0.12
<0.12
<0.12
0.12
0.25
0.12
0.5
0.25
0.12
0.5
>25
>25
25
2.52
2.73
1.10
1.31
1.05
1.27
1.25
1.70
1.91
0.57
0.78
3.09
3.23
2.32
2.61
1.85
1.89
8
8
40
48
53
100g,h
38g
97g
98g
4
0.25
4
0.25
0.5
0.5
1
<0.12
0.25
0.12
0.5
<0.12
0.5
0.12
0.5
12.5
8.5
6
4
64
4
<0.12
0.25
0.5
0.25
0.25
1
>64
>64
64
16
12
4
2
0.5
<0.12
32
0.25
<0.12
8
>25
25
>25
3
1.99
2.18
1.05
1.40
2
4
39
<0.12
1
0.12 <0.12
47
>64
>64
>64
>64
>64
>64
>64
>64
8
2
52
4
1
<0.12
0.25
0.5
4
<0.12
2
novobiocin
0.25
0.12
1
3
vancomycin
1
1
2
4
1
1.5
e1
2
a Supercoiling assay, E. coli gyrase, MNEC = maximum noneffective concentration. b Agar dilution (Mueller-Hinton medium), inoculum 104 CFU/
spot, MIC = minimum inhibitory concentration. Test organisms: Escherichia coli 25922, Xanthomonas maltophilia IAC 739, Staphylococcus aureus ATCC
25923, Staphylococcus aureus Smith, Staphylococcus epidermidis 16/2, Enterococcus faecalis 6, Streptococcus pyogenes b15. c In vivo efficacy, septicaemia in
mice (S. aureus Smith), iv administration. d Calculated log P, see Table 1. e Octanol/water pH 7.4 shake-flask method. f “Plasma shift” = ratio MIC (BHI
agar-mouse plasma (1:1), S. a. Smith)/MIC (BHI agar, S. a. Smith). g Reference compounds.10 h Hydrochloride salt.
protein binding, resulting in a “plasma shift” of only 2. However,
dilactam 52 obviously displays improved pharmacokinetic
properties, which was demonstrated in a pharmacokinetic study
in rats, where it achieved the highest plasma levels measured so
2216
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Table 6. In Vitro Antibacterial Activity of 52 and of Reference Compounds against Multiresistant Gram-Positive and Gram-
Negative Bacteriaa
S aureus (11 strains) E. faecalis (16 strains) E. faecium (23 strains) S. pneumoniae (21 strains) H. influenzae (21 strains) M. catarrhalis (20 strains)
compd
MIC50/90
MIC50/90
MIC50/90
MIC50/90
MIC50/90
MIC50/90
52
3b
5b
4/4
1/1
0.25/0.5
4/4
0.5/0.5
8/8
0.5/2
1/2
1/4
1/2
8/16
2/2
1/1
256 />256 c
1/1
1/2
2/4
2/8 d
8/8
32/>32 d
2/2
8/32 d
penicillinc,d
linelozide
1/2
2/2
1/2
1/2
8/16
8/16
imipenem
vancomycin
32/64
2/4
2/2
64/>64
256 />256
2/256
a Agar dilution: Mueller-Hinton agar (inoculum 104 CFU/spot) for S. aureus, E. faecium, E. faecalis; Haemophilus test medium (inoculum 105 CFU/
spot) for H. influenzae; Mueller-Hinton agar (inoculum 105 CFU/spot) and IsoSensitest (inoculum 104 CFU/spot) supplemented with sheep blood
(5% v/v), for M. catarrhalis and S. pneumoniae, respectively. MIC50/90 = minimum concentration (in μg/mL) that inhibits at least 50%, 90% of the strains
tested. b Monolactam reference compounds.10 c Data for methicillin. d Data for penicillin G.
Figure 5. Correlation between calculated log P (CLOGP) and experi-
mental log D values of two analogous series of monolactams and
dilactams (see Table 5). The ratio y/x between derivatives differing
only in the additional amide group is <1, indicating an overestimation of
this hydrophilicity increment by the CLOGP value. The dilactam
analogues are positioned along a steeper line (slope 2.2) as compared
to the monolactams (line slope 1.3). This results from the overestima-
tion in the CLOGP of the hydrophilic impact of the additional amide
functionality in comparison to the lipophilicity contributions of methyl/
bromine replacement (52), thiation (48), or both (53).
explains their reduced antibacterial effect. Being aware of the
deficiencies of the CLOGP value, Figure 6 demonstrates anyway
that a certain degree of lipophilicity is a prerequisite for a
pronounced antibacterial potency. However, the general obser-
vation that high lipophilicity leeds to increased protein binding,
assessed directly or deduced from the “plasma shift” of the
antibacterial in vitro test, reinforced the view that only com-
pounds of a certain hydrophilicity can exhibit efficacy in vivo.
Figure 7 shows the ED50 values determined for all compounds of
significant activity against S. aureus Smith plotted against their
CLOGP value. In vivo efficacy (ED50 e 25 mg) was only found
for derivatives with CLOGP between 0.78 and 1.92. Some
inefficacious compounds within this lipophilicity window were
possibly not tested high enough (ED50 > 12 mg/kg), but others
(ED50 >25 mg/kg) must be hampered by other unfavorable
properties such as metabolic degradation, e.g. glucuronidation.
We have focused on compounds occupying this favorable
lipophilicity range, which seems to offer the best compromise for
optimizing concomitantly cell permeation and pharmacokinetic
Figure 4. Total plasma concentration of 52 and of reference com-
pounds10 4, 100, and 3 in rats (n = 3-6) after a single intravenous dose
of 20 mg/kg (microsuspension in gelatin). Comparison with vancomy-
cin data from ref 35.
far for cyclothialidine analogues and approaching those mea-
sured for vancomycin (Figure 4).
The comparison of CLOGP values with experimental log D
values reveals a charscteristic distinction between the monolac-
tam and the dilactam series (Figure 5). The calculated log P value
obviously overestimates the hydrophilicity impact of the addi-
tional amide functionality. This can be explained by a compen-
satory reduction of the hydrophilicity by the stronger
intramolecular hydrogen bond of the dilactam scaffold (see
Figure 1). As a consequence, the 11-thioxo-dilactams 48 and
53 have a higher log D but a smaller CLOGP value than the
monolactams 98 and 39. Nevertheless, dilactams 47 and 52 are
significantly more polar than the monolactams 98 and 39, which
2217
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
Figure 6. Antibacterial activity in vitro of all new compounds plotted
against their CLOGP value. A certain degree of lipophilicity (CLOGP >
1) is required for a pronounced antibacterial potency.
Figure 8. Impact on polarity (log D) and biological properties (in vivo
efficacy, antibacterial activity) of chemical modifications on derivatives
of parent monolactam 98. The derivatives shown are interconnected by
conversions A, B, and C. Small/large circle diameters indicate small/
large MIC values.
to new series of interest, e.g. with regard to conformational
adaption to different environments. Thus, it seemed obvious that
new lactone subclasses of intrinsically improved biological prop-
erties had to be found empirically.
Further attempts to identify more favorable hydrophilic
substitution patterns for the 14-membered monolactams
provided only compounds of similar efficacies as those pre-
viously reported, but the use of a chloro- or bromo-substitute
for the 1-methyl group of monolactams brought some pro-
gress with regard to the antibacterial activity both in vitro and
in vivo.
A noteworthy step forward was accomplished by using a new
and more polar scaffold obtained by incorporating an additional
amide moiety into the lactone ring. Dilactam analogues were
found to be still potent DNA gyrase inhibitors. Compared to the
monolactams, they had a slightly reduced antibacterial potency
due to their increased hydrophilic character, a drawback which,
however, was compensated by more favorable physicochemical
properties. The dilactam inhibitors cover the same antibacterial
spectrum than the monolactams and were found to be particu-
larly active against multiresistant strains of Enterococcus faecalis,
Entercoccus faecium , and Streptococcus pneumonia. The 4-bromo-
dilactam 52 displayed a significantly improved pharmacokinetic
behavior, which resulted in a substantial in vivo efficacy with an
ED50 of 3 mg/kg in a mouse septicaemia model. Regarding its
antibacterial potency, polar character, as well as pharmacokinetic
behavior, dilactam 52 resembles the properties of vancomycin,
which, however, is addressing an extracellular target. The fact that
the “dilactams” do not require a thioamide functionality to
achieve in vivo efficacy makes them attractive.
Figure 7. In vivo efficacy (septicaemia in mice, S. aureus Smith) plotted
against CLOGP for all compounds with a MIC e 2 μg/mL (S. aureus
Smith) tested. For pharmacokinetic data of compounds 3, 4, 52, and
100, see Figure 4.
behavior. Figure 8 shows schematically how the stepwise applica-
tion of structural lipophilicity modulaters on the parent lactone
98, i.e. conversions A, B and C, influences hydrophilicity and
biological properties, i.e. log D, ED50 and MIC values, of its
derivatives.
’ CONCLUSIONS
Our continued effort to optimize bicyclic lactones derived
from the natural product cyclothialidine was guided by the
hypothesis that only a certain degree of hydrophilicity will allow
the translation of a good antibacterial potency into in vivo
efficacy. According to this hypothesis, and confirmed by the
experience collected so far, efficacious compounds can only be
found within a rather narrow lipophilicity window. The CLOGP
value appeared to be a useful guidance tool for the modification
program, although the conformational flexibility of the 14-
memberd lactone rings together with the possibility to form an
intramolecular hydrogen bond could provide specific properties
Unfortunately, we did not succeed in further improving on the
lead compound 52, and only structurally close analogues dis-
played a similar in vivo effect. Nevertheless, the new dilactam
subclass confirms that DNA gyrase inhibitors of the cyclothiali-
dine family can indeed combine interesting antibacterial activity
with satisfactory pharmacokinetic properties, and therefore,
might earn further investigation.
2218
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
’ EXPERIMENTAL SECTION
L), followed by cooling and addition of hexane (0.4 L) to give
pseudoacid 28 (51.6 g, 88%) as white solid, mp 153-153 °C. 1H
NMR (DMSO-d6) δ 3.79 (s, 3H), 6.56 (d, 1H, J = 8), 6.66 (s, 1H), 6.76
(br s, 1H), 7.82 (d, 1H, J = 8), 10.42 (br s, 1H).
General. Solvents and chemicals used for reactions were purchased
from commercial suppliers and used without further purification. Reac-
tions were carried out under N2 or Ar. For standard workup procedures,
organic solvents and aqueous (aq) solutions used are given in brackets,
and if not stated otherwise, aq solutions of NaHCO3 and Na2CO3 used
were saturated, aq layers were back-extracted with the organic solvent
used, organic solutions were dried with Na2SO4, and solvents were
evaporated under reduced pressure at 20-40 °C on a B€uchi rotary
evaporator. Thin layer chromatography (TLC) was carried out on silica
gel glass plates (Kieselgel 60 F254, Merck). Normal phase silica gel (Silica
Gel 60, 230-400 mesh, Merck) was used for preparative chromatogra-
phy, and the eluents used are indicated in brackets. Melting points (mp)
(RS)-7-Bromo-3,4-dihydroxy-6-methoxy-3H-isobenzofur-
an-1-one (27c). Bromine (12.8 mL, 0.25 mol) was added over 1 min
to a stirred suspension of 28 (49.0 g, 0.25 mol) in DCM (0.5 L)
precooled to 0 °C. The temperature rose to 12 °C and was allowed to
rise to 20 °C over 10 min. Stirring was continued for 2 h at 20 °C, and
thereafter, the yellow reaction mixture was evaporated completely. To a
suspension of the solid residue in H2O (0.5 L) was added at 0 °C 32% aq
NaOH (100 mL). The mixture was stirred for 15 min to produce a dark-
yellow solution of pH 12. By slowly adding 7 N HCl (ca. 100 mL) at
0 °C, the pH of the reaction mixture was lowered to 2.5. The precipitate
produced was collected by filtration, washed with H2O, and dried. The
crude product was triturated with hot acetone (0.25 L), and after cooling
to 0 °C and addition of hexane (0.25 L), 27c (60.2 g, 88%) was isolated
as white solid, mp >200 °C. 1H NMR (DMSO-d6) δ 3.88 (s, 3H), 6.46
(d, 1H, J = 8.5), 6.62 (s, 1H), 7.59 (d, 1H, J = 8.5), 10.64 (s, 1H). Signals
of corresponding acid (2-bromo-6-formyl-5-hydroxy-3-methoxy-ben-
zoic acid, ca. 20%) δ 3.93 (s, 3H), 6.67 (s, 1H), 10.05 (s, 1H), 11.42
(s, 1H), 13.45 (br s, 1H).
Allyl 2-Bromo-6-formyl-5-hydroxy-3-methoxy-benzoate
(29c). N,N,N0,N0-Tetramethylguanidine (34.6 mL, 0.275 mol) was added
at 20 °C to a solution of 27c (68.8 g, 0.25 mol) in DMF (0.5 L). After
stirring for 0.5 h, allyl bromide (27.5 mL, 0.325 Mol) was added to the
reaction mixture and stirring was continued at 20 °C for 18 h. The mixture
was evaporated, and Et2O (0.4 L) was added to the residual oil. The
mixture was extracted with 2% aq NaHCO3 (0.4 L). The aqueous extract
was set to pH 5 by addition of 7 N HCl and then extracted with EtOAc (1 L).
The organic layer was washed with brine, dried, and evaporated. The
remaining oil (57.2 g) was taken up in hot hexane (0.9 L), insoluble
material was removed by filtration, and the solution was stirred at 0 °C to
give 29c (53.1 g, 67%) as white solid, mp 45 °C. 1H NMR (CDCl3) δ
3.96 (s, 3H), 4.91 (d, 2H, J = 6), 5.46 (d, 1H, J = 10), 5.47 (d, 1H, J = 17),
5.95-6.16 (m, 1H), 6.51 (s, 1H), 9.68 (s, 1H), 12.01 (s, 1H).
In addition, the byproduct allyl 3-allyloxy-6-bromo-2-formyl-5-meth-
oxy-benzoate (8.0 g, 9%) was isolated from the Et2O extract, and starting
material 27c (8.7 g, 13%) was recovered from the aqueous extract at pH
2.5 (EtOAc).
Allyl 2-Bromo-5-[dimethyl(thexyl)silyloxy]-6-formyl-3-
methoxy-benzoate (30c). To a stirred mixture of 29c (37.8g, 0.12mol)
and (dimethyl)-(1,1,2-trimethyl-propyl)-chlorosilane (35.3 mL, 0.18 mol)
in DMF (240 mL) was added at 0 °C NEt3 (25.1 mL, 0.18 mol), a
precipitate being formed immediately. The reaction mixture was stirred at
20 °C for 15 h, and then worked up (EtOAc, 1 N HCl, 5% aq NaCl). The
resulting oil was crystallized from hexane to give 30c (51.3 g, 93%) as white
crystals, mp 88-89 °C. 1H NMR (CDCl3) δ 0.34 (s, 6H), 0.94 (d, 6H, J =
7), 0.98 (s, 6H), 1.68-1.84 (m, 1H), 3.92 (s, 3H), 4.91 (d, 2H, J = 6), 5.30
(d, 1H, J = 10), 5.42 (d, 1H, J = 16), 5.98-6.17 (m, 1H), 6.39 (s, 1H), 10.18
(s, 1H).
Allyl 2-Bromo-5-[dimethyl(thexyl)silyloxy]-6-iodomethyl-
3-methoxy-benzoate (31c). Sodium iodide (10.34 g, 69.0 mmol)
and trimethylchlorosilane (9.45 mL, 73.6 mmol) were added at 20 °C to
a solution of 30c (21.0 g, 46.0 mmol) in MeCN (23 mL), a brown
precipitate being formed. The mixture was stirred for 5 min, subse-
quently cooled to 0 °C, and 1,1,3,3-tetramethyl-disiloxane (8.7 mL, 48.5
mmol) was added over 5 min. The mixture was stirred at 20 °C for 2 h,
and then diluted with EtOAc, washed with H2O, dried, and evaporated.
On the basis of NMR analysis, the resulting brown oil (30 g) consisted of
31c (ca.19.2 g, 73%), and 7-bromo-4-[dimethyl-(thexyl)silyloxy]-6-
methoxy-3H-isobenzofuran-1-one (10.8 g) as the major byproduct.
NMR (CDCl3) of 31c: δ 0.31 (s, 6H), 0.87 (d, 6H, J = 7), 0.99 (s,
6H), 1.64-1.82 (m, 1H), 3.79 (s, 3H), 4.34 (s, 2H), 4.86 (d, 2H, J = 6),
1
were determined on a B€uchi SMP-20K and are uncorrected. H NMR
spectra were recorded on a Bruker-AC-250 or on a Bruker-ARX-400. If
not stated otherwise, spectra were recorded at 250 mHz. Chemical shifts
are reported as δ values (ppm) downfield from internal TMS in the
indicated solvent. Coupling constants (J) are given in Hz, s = singlet, d =
doublet, t = triplet, q = quartet, m = multiplet, br = broad. Low-resolution
ISP-MS and ISN-MS were recorded on a PE-Sciex API III. High-
resolution mass spectra (HRMS) were determined on an Agilent LC/
MS (series 1200, 6120 Q-TOF). Elemental analyses were carried out in
our laboratory or by Solvias AG, Basel, and results are within (0.4% of the
theoretical values. Alternatively, target compounds without elemental
analysis were analyzed by HRMS and their purity was assessed by HPLC
(A: RP-18 (reverse phase), 250 mm ꢀ 4 mm, detector 278 nm, 0.01 M aq
K,Na-phospate buffer (pH 7.0) containing tetrakis-(decyl)ammonium
bromide (0.008 mol/L)/acetonitrile (38:62), 1 mL/min; B: C-18
(reverse phase), 2.1 mm ꢀ 30 mm, detector with various wavelengths
265 nm (70%), 225 (10%), 215, 220, 230, 250, 280 nm (4% each), eluent
0.01% aq HCOOH/5-95% MeCN-MeOH (4:1); 1.2 mL/min) and
shown to be >95% if not stated otherwise.
Methyl 2-Chloro-6-formyl-5-hydroxy-3-methoxy-benzo-
ate (26). To a solution of methyl 2-formyl-3-hydroxy-5-methoxybenzo-
ate (25,26 25.4 g, 0.12 mol) in DCM (120 mL) was added over 15 min at
20 °C sulfuryl chloride (10.7 mL, 0.13 mol), causing a light evolution of
gas. The mixture was stirred for 21 h at 20 °C and then evaporated. The
residual material was crystallized from EtOAc to give a mixture of 26 and
the 4-chloro regioisomer (23.6 g, 72:28 mixture of regioisomers).
Repeated recrystallization from t-BuOMe provided pure 26 (8.33 g,
28%) as white crystals, mp 123-124 °C. 1H NMR (DMSO-d6) δ 3.82
and 3.93 (2 s, 2 ꢀ 3H), 6.72, 10.10, and 11.43 (3 s, 3 ꢀ 1H).
(RS)-7-Chloro-3,4-dihydroxy-6-methoxy-3H-isobenzofur-
an-1-one (27b). To 26 (0.65 g, 2.5 mmol) was added 2 N NaOH
(2.5 mL, 5 mmol), and the mixture was stirred for 30 min. The solution
was acidified with 3 N HCl at 0 °C, a white precipitate being formed. The
mixture was extracted with EtOAc. The organic layer was washed with
brine, dried, and evaporated to give 27b (0.57 g, 99%) as white solid, mp
>230 °C. 1H NMR (DMSO-d6) δ 3.87 (s, 3H), 6.48 (d, 1H, J = 8.7),
6.86 (s, 1H), 7.81 (d, 1H, J = 8.7), 10.62 (s, 1H). Signals of the
corresponding acid (2-chloro-6-formyl-5-hydroxy-3-methoxy-benzoic
acid, ca. 20%) δ 3.92 (s, 3H), 6.70 (s, 1H), 10.10 (s, 1H), 11.38 (s,
1H), 13.46 (br s, 1H). Anal. (C9H7ClO5) C, H, Cl.
(RS)-3,4-Dihydroxy-6-methoxy-3H-isobenzofuran-1-one
(28). To methyl 2-formyl-3-hydroxy-5-methoxybenzoate (25, 63.1 g,
0.30 mol) was added 2 N NaOH (375 mL, 0.75 mol), and the mixture
was stirred at 20 °C for 0.5 h. Water (50 mL) was added, and the mixture
was cooled in an ice-bath. Upon acidification to pH 2.5 by addition of 7
N HCl (ca. 70 mL), a precipitate was formed. The mixture was stirred for
0.5 h at 0 °C, and subsequently, the solid was collected by filtration and
washed with cold H2O (130 mL). The wet solid was dissolved in acetone
(0.3 L)/EtOAc (0.6 L) and the solution was dried and evaporated. The
resulting residue was recrystallized by dissolving it in hot acetone (0.11
2219
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
5.32 (d, 1H, J = 10), 5.48 (d, 1H, J = 16), 5.99-6.17 (m, 1H), 6.36 (s,
1H).
preparation of 100 from 6;10 white solid. 1H NMR (DMSO-d6) δ
1.60-2.10 (m, 4H), 2.34 (s, 3H), 2.58-2.74 and 2.86-3.00 (2 m, 2 ꢀ
1H), 3.08 (dd, 1H, J = 14, 11), 3.40 (dd, 1H, J = 14, 4), 3.69 (d, 1H, J =
11), 3.80 (s, 3H), 3.91 (d, 1H, J = 11), 4.06-4.20 (m, 1H), 4.31 (br s,
2H), 4.50-4.66 (m, 1H), 5.72 -5.88 (m, 1H), 6.72 (s, 1H), 8.67 (br s,
3H), 10.45 (s, 1H), 10.76 (d, 1H, J = 7). HRMS calcd for
(C19H23ClN4O5S2) 486.0798, found 486.0793.
By subjecting amine 34c and 5-trityloxypentanoic acid in analogous
manner to the reaction sequence described previously for the prepara-
tion of 98 and 3 from 14,10 compounds 39 and 40 were obtained:
(R)-1-Bromo-4-hydroxy-2-methoxy-8-(3-methyl-[1,2,4]-
oxadiazol-5-yl)-8,9,11,12,13,14-hexahydro-5H,7H-15-oxa-6-
thia-9-aza-benzocyclotetradecene-10,16-dione (39). White
solid. 1H NMR (DMSO-d6) δ 1.60-2.10 (m, 5H), 2.34 (s, 3H), 2.33-
2.48 (m, 1H), 2.88 (dd, 1H, J = 14, 11), 3.29 (dd, 1H, J = 14, 4), 3.68 (d,
1H, J = 11), 3.80 (s, 3H), 3.85 (d, 1H, J = 11), 4.04-4.16 (m, 1H),
4.54-4.65 (m, 1H), 5.13-5.25 (m, 1H); 6.64 (s, 1H), 8.66 (d, 1H, J =
8), 10.34 (s, 1H). MS (ISN) 498.0 [(M - 1)-]. Anal. (C19H22-
Allyl 2-[(R)-2-Amino-2-(3-methyl-[1,2,4]oxadiazol-5-yl)-
ethylsulfanylmethyl]-6-bromo-3-[dimethyl(thexyl)silyloxy]-
5-methoxy-benzoate (34c). NEt3(6.4 mL, 46 mmol) was added to
a solution of crude 31c (30 g containing ca. 19.2 g, 34 mmol, of 31c) and
(R)-2-mercapto-1-(3-methyl-1,2,4-oxadiazol-5-yl)-ethylcarbamic acid
tert-butyl ester (32,10 11.9 g, 46 mmol) in DCM (90 mL). The reaction
mixture was stirred at 20 °C for 2 h, and then worked up (EtOAc, 1 N
HCl, brine). The resulting material was stirred with TFA (90 mL) at 0 °C
for 0.5 h. The solution was evaporated, and the remaining oil was worked
up (EtOAc, aq Na2CO3, brine). The crude product was purified by
chromatography (EtOAc/hexane 1:3) to give 34c (13,5 g, 66%) as a
light-yellow foam. NMR (CDCl3) δ 0.30 (s, 6H), 0.93 (d, 6H, J =
7), 0.99 (s, 6H), 1.67-1.85 (m, 1H), 1.84 (br s, 2H), 2.38 (s, 3H),
2.84 (dd, 1H, J= 13, 8), 3.06 (dd, 1H, J= 13, 5), 3.69 and 3.74 (2 d, 2 ꢀ 1H,
J = 13), 3.84 (s, 3H), 4,12-4.20 (m, 1H), 4.86 (d, 2H, J = 6), 5.32 (d, 1H, J
= 11), 5.47 (d, 1H, J = 16), 5.97-6.15 (m, 1H), 6.45 (s, 1H).
BrN3O6S 0.2EtOAc) C, H, N.
Allyl 2-Chloro-5-[dimethyl(thexyl)silyloxy]-6-iodomethyl-
3-methoxy-benzoate (31b). Obtained by subjecting 27b in analo-
gous manner to the reaction sequence described above for the preparation
of 31c from 27c. Brown oil, which, based on NMR analysis, consisted of
31b (ca. 60%) and 7-chloro-4-[dimethyl-(thexyl)silyloxy]-6-methoxy-3H-
isobenzofuran-1-one (ca. 32%) as the major byproduct. NMR (CDCl3) of
31b: δ 0.36 (s, 6H), 0.96 (d, 6H, J = 7), 1.08 (s, 6H), 1.65-1.83 (m, 1H),
3.84 (s, 3H), 4.39 (s, 2H), 4.90 (d, 2H, J = 6), 5.32 (d, 1H, J = 10), 5.48 (d,
1H, J = 16), 6.01-6.18 (m, 1H), 6.42 (s, 1H).
Allyl 2-[(R)-2-Amino-2-(3-methyl-[1,2,4]oxadiazol-5-yl)-
ethylsulfanylmethyl]-6-chloro-3-[dimethyl(thexyl)silyloxy]-
5-methoxy-benzoate (34b). Obtained by subjecting crude 31b in
analogous manner to the procedure described above for the preparation
of 34c from 31c; light-yellow foam. NMR (CDCl3) δ 0.31 (s, 6H), 0.93
(d, 6H, J = 7), 0.99 (s, 6H), 1.70-1.82 (m, 1H) superimposed by 1.79
(br s, 2H), 2.38 (s, 3H), 2.86 (dd, 1H, J = 14 and 8), 3.05 (dd, 1H, J = 14
and 5), 3.70 and 3.74 (2 d, 2 ꢀ 1H, J = 11), 3.85 (s, 3H), 4.15-4.21 (m,
1H), 4.86, 5.31, and 5.45 (3 d, 3 ꢀ 2H, J = 6, 10 and 18), 5.96-6.17 (m,
1H), 6.47 (s, 1H).
Allyl (R)-2-[2-Amino-2-(3-allyloxycarbonylaminomethyl-
1,2,4-oxadiazol-5-yl)-ethylsulfanyl-methyl]-3-[dimethyl-
(thexyl)silyloxy]-6-chloro-5-methoxybenzoate (35b). Ob-
tained by subjecting 31b in analogous manner to the procedure
described above for the preparation of 34c but replacing 31c by 31b
and 32 by (R)-1-(3-allyloxycarbonylaminomethyl-1,2,4-oxadiazol-
5-yl)-2-mercapto-ethylcarbamic acid tert-butyl ester (33);10 light-
yellow foam. NMR (CDCl3) δ 0.31 (s, 6H), 0.93 (d, 6H, J = 7), 0.99
(s, 6H), 1.70-1.82 (m, 1H), 2.87 (dd, 1H, J = 13, 7), 3.05 (dd, 1H, J =
13, 4), 3.69 and 3.76 (2 d, 2 ꢀ 1H, J = 14), 3.85 (s, 3H), 4.15-4.24
(m, 1H), 4.52, 4.61, and 4.86 (3 d, 3 ꢀ 2H, J = 6), 5.20-5.50 (m, 4H),
5.84-6.15 (m, 2H), 6.47 (s, 1H).
Allyl (R)-2-[2-Amino-2-(3-allyloxycarbonylaminomethyl-
1,2,4-oxadiazol-5-yl)-ethylsulfanyl-methyl]-6-bromo-3-[di-
methyl(thexyl)silyloxy]-5-methoxybenzoate (35c). Obtained
by subjecting 31c in analogous manner to the procedure described above
for the preparation of 34c but replacing 32 by 33; light-yellow foam.
NMR (CDCl3) δ 0.31 (s, 6H), 0.93 (d, 6H, J = 7), 0.99 (s, 6H), 1.76-
1.80 (m, 1H) superimposed by 1.65-1.95 (br s, 2H), 2.87 (dd, 1H, J =
13, 7), 3.05 (dd, 1H, J = 13, 5), 3.67 and 3.72 (2 d, 2 ꢀ 1H, J = 13), 3.85
(s, 3H), 4.19 (dd, 1H, J = 7, 5), 4.52, 4.60, and 4,86 (3 d, 3 ꢀ 2H, J = 6),
5.18-5.55 (m, 5H), 5.84-6.15 (m, 2H), 6.45 (s, 2H).
3
(R)-1-Bromo-4-hydroxy-2-methoxy-8-(3-methyl-[1,2,4]-
oxadiazol-5-yl)-10-thioxo-7,8,9,10,11,12,13,14-octahydro-
5H-15-oxa-6-thia-9-aza-benzocyclotetradecen-16-one (40).
White solid. 1H NMR (DMSO-d6) δ 1.60-2.10 (m, 4H), 2.34 (s, 3H),
2.58-2.73 and 2.85-2.99 (2 m, 2 ꢀ 1H), 3.04 (dd, 1H, J = 14, 11), 3.40
(dd, 1H, J = 14, 4), 3.66 (d, 1H, J = 11), 3.80 (s, 3H), 3.88 (d, 1H, J = 11),
4.06-4.19 and 4.51-4.64 (2 m, 2 ꢀ 1H), 5.69-5.81 (m, 1H), 6.64 (s,
1H), 10.37 (s, 1H), 10.63 (d, 1H, J = 8). MS (ISN) 514.0 [(M - 1)-].
Anal. (C19H22BrN3O5S2) C, H, N, Br.
Ethyl (2-Trityloxy-acetylamino)-acetate (43). To a mixture of
2-trityloxy-acetic acid (42,10 63.6 g, 0.20 mol), glycine ethyl ester
hydrochloride (30.7 g, 0.22 mol), and 4-methyl-morpholine (24.2 mL,
0.22 mol) in MeCN (0.8 L) was added dropwise at 0-4 °C over 30 min
a solution of DCC (45.4 g, 0.22 mol) in MeCN (0.2 L). The mixture was
stirred at 0 °C for 3 h, a white precipitate being formed. The mixture was
evaporated and the remaining material was stirred with EtOAc (0.5 L) at
0 °C for 15 min. Insoluble material was removed by filtration, and the
clear solution was washed with 1 N HCl (0.5 L) and with brine. The or-
ganic layer was dried and evaporated to give 43 (82 g, “102%”) as a white
solid. An analytical sample was obtained by recrystallization (EtOAc/
hexane) as white solid, mp 115-116 °C. 1H NMR (DMSO-d6) δ 1.21
(t, 3H, J = 7), 3.45 (s, 2H), 3.92 (d, 2H, J = 6), 4.12 (q, 2H, J = 7), 7.25-
7.48 (m, 15H), 8.24 (t, 1H, J = 6). Anal. (C25H25NO4) C, H, N.
(2-Trityloxy-acetylamino)-acetic Acid (44). A mixture of 43
(82.0 g, max. 0.2 mol) and KOH (14.6 g, 0.26 mol) in MeOH (100 mL)
was stirred at 60 °C for 0.5 h. The mixture was cooled and evaporated,
and the remaining oil was dissolved in H2O (0.2 L). The solution was
washed with t-BuOMe/hexane (2:1, 0.6 L) and then acidified to pH 3 at
0 °C by the addition of 5 N HCl, a white precipitate being formed. The
mixture was extracted with EtOAc (0.4 L). Insoluble material in the
organic layer (product) was isolated by filtration, and the filtrate was
washed with brine, dried, and evaporated. The residue (25.3 g) and the
isolated insoluble material were triturated together with EtOAc/hexane
(1:1, 1 L) at 0 °C to give, after filtration and drying, 44 (60.0 g, 80%) as
white solid, mp 183-184 °C. 1H NMR (DMSO-d6) δ 3.44 (s, 2H), 3.84
(d, 2H, J = 6), 7.25-7.50 (m, 15H), 8.14 (t, 1H, J = 6), 12.66 (s, 1H).
Anal. (C23H21NO4) C, H, N.
3-[Dimethyl(thexyl)silyloxy]-2-[(R)-2-[2-(2-hydroxy-acetyl-
amino)-acetylamino]-2-(3-methyl-[1,2,4]oxadiazol-5-yl)-
ethylsulfanylmethyl]-5-methoxy-6-methyl-benzoic Acid (45a).
(i) A mixture of allyl 2-[(R)-2-amino-2-(3-methyl-[1,2,4]oxadiazol-5-yl)-
ethylsulfanylmethyl]-3-[dimethyl(thexyl)silyloxy]-5-methoxy-6-methyl-
benzoate (14,10 2.95 g, 5.5 mmol), 44 (2.44 g, 6.5 mmol), and EDC
(1.53 g, 8.0 mmol) in MeCN (55 mL) was stirred at 0 °C for 5 h.
Workup (EtOAc, 1 N HCl, aq NaHCO3, brine) gave a viscous oil (5.38
(R)-8-(3-Aminomethyl-[1,2,4]oxadiazol-5-yl)-1-chloro-4-
hydroxy-2-methoxy-10-thioxo-7,8,9,10,11,12,13,14-octahydro-
5H-15-oxa-6-thia-9-aza-benzocyclotetradecen-16-one Hy-
drochloride (38). Obtained by subjecting amine 35b in analogous
manner to the reaction sequence described previously for the
2220
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
g, ca. 100%). (ii) This oil (5.38 g) was stirred with pTsOH hydrate (1.0
g) in MeOH (30 mL) at 20 °C for 2 h. Workup (EtOAc, 5% aq
NaHCO3, brine) and chromatography (EtOAc/hexane 1:1) gave a
light-yellow foam (3.1 g). (iii) To a solution of this material (3.1 g,
4.8 mmol) in THF (50 mL) were added at 0 °C morpholine (4.1 mL, 47
mmol) and Pd(PPh3)4 (0.115 g, 0.1 mmol). The mixture was stirred for
2 h at 0 °C, and then worked up (EtOAc, 1 N HCl, brine) to give 45a
(1.95 g, 58%) as light-yellow foam. 1H NMR (CDCl3) δ 0.28 (s, 6H),
0.94 (d, 6H, J = 7), 0.98 (s, 6H), 1.70-1.82 (m, 1H), 2.14 and 2.35 (2 s, 2
ꢀ 3H), 3.04 (dd, 1H, J = 12, 5), 3.17 (dd, 1H, J = 12, 7), 3.76 (s, 3H),
3.78 and 3.82 (2 d, 2 ꢀ 1H, J = 12), 4.12 (dd, 1H, J = 14.7, 4), 4.22 (s,
2H), 4.34 (dd, 1H, 14.7, 6), 5.36-5.48 (m, 1H), 6.37 (s, 1H), 7.32 (s,
1H), 7.55-7.64 and 7.78-7.90 (2 m, 2 ꢀ 1H).
50 (6 mg, 10%, from 77 mg F1, after chromatography), white solid
(see SI).
(R)-4-Chloro-1-hydroxy-3-methoxy-13-(3-methyl-[1,2,4]-
oxadiazol-5-yl)-9,10,13,14-tetrahydro-12H,16H-6-oxa-15-
thia-9,12-diaza-benzocyclotetradecene-5,8,11-trione (51).
Obtained by subjecting amine 34b in analogous manner to the reaction
sequence described above for the preparation of 47 from 34a; white
solid. 1H NMR (DMSO-d6) δ 2.31 (s, 3H), 2.90 (dd, 1H, J = 14, 3), 3.24
(dd, 1H, J = 14, 7), 3.54 (d, 1H, J = 12), 3.81 (s, 3H), 3.86 (d, 1H, J = 12)
superimposed by 3.80-3.96 (m, 2H), 4.71 and 4.90 (2 d, 2 ꢀ 1H, J =
13), 4.49-5.62 (m, 1H), 6.69 (s, 1H), 7.44 (d, 1H, J = 9), 9.30 (t, 1H, J =
6), 10.45 (s, 1H). HRMS calcd for (C18H19ClN4O7S) 470.0663, found
470.0659.
6-Bromo-3-[dimethyl(thexyl)silyloxy]-2-[(R)-2-[2-(2-hydroxy-
acetylamino)-acetylamino]-2-(3-methyl-[1,2,4]oxadiazol-5-
yl)-ethylsulfanylmethyl]-5-methoxy-benzoic Acid (45c). Amine
34c (6.0 g, 10.0 mmol) was subjected in analogous manner to the reaction
sequence described above (45a (i-iii)) to give crude 45c (6.34, 94%) as
slightly yellow foam. 1H NMR (CDCl3) δ 0.30 (s, 6H), 0.94 (d, 6H, J = 7),
0.99 (s, 6H), 1.70-1.82 (m, 1H), 2.36 (s, 3H), 3.04 (dd, 1H, J = 12, 5),
3.19 (dd, 1H, J = 12, 7), 3.72 and 3.79 (2 d, 2 ꢀ 1H, J = 12), 3.84 (s, 3H),
4.10 (dd, 1H, J = 14, 4), 4.22 (s, 2H), 4.36 (dd, 1H, 14, 6), 5.42-5.52 (m,
1H), 6.41 (s, 1H), 7.42-7.60 and 7.77-7.88 (2 m, 2 ꢀ 1H).
(R)-1-[Dimethyl(thexyl)silyloxy]-3-methoxy-4-methyl-13-
(3-methyl-[1,2,4]oxadiazol-5-yl)-9,10,13,14-tetrahydro-12H,-
16H-6-oxa-15-thia-9,12-diaza-benzocyclotetradecene-5,8,11-
trione (46a). A solution of 45a (1.43 g, 2.34 mmol) and PPh3 (1.31 g,
5.0 mmol) in toluene (60 mL) was cooled to 0 °C. DEAD (0.78 mL, 5.0
mmol) was added, and the mixture was stirred for 1 h at 0 °C and for 15 h
at 20 °C. The solution was evaporated and the residue was chromato-
graphed (EtOAc/hexane 3:1) to give 46a (0.58 g, 42%) as off-white
foam. NMR (CDCl3) δ 0.27 and 0.30 (2 s, 2 ꢀ 3H), 0.94 (d, 6H, J = 7),
0.99 (s, 6H), 1.70-1.84 (m, 1H), 2.04 and 2.29 (2 s, 2 ꢀ 3H), 2.94 (dd,
1H, J = 14, 4), 3.39 (dd, 1H, J = 14, 6), 3.63 (d, 1H, J = 11), 3.76 (s, 3H),
3.89 (d, 1H, J = 11), 3.97 (dd, 1H, J = 15, 5), 4.14 (dd, 1H, J = 15, 6), 4.54
and 5.00 (2 d, 2 ꢀ 1H, J = 12), 5.58-5.70 (m, 1H), 6.40 (s, 1H), 7.36 (d,
1H, J = 8), 8.38 (t, 1H, J = 6).
(R)-4-Bromo-1-[dimethyl(thexyl)silyloxy]-3-methoxy-13-
(3-methyl-[1,2,4]oxadiazol-5-yl)-9,10,13,14-tetrahydro-12H,-
16H-6-oxa-15-thia-9,12-diaza-benzocyclotetradecene-5,8,-
11-trione (46c). Subjecting 45c (4.96 g, 7.4 mmol) to the lactoniza-
tion procedure described in the preparation of 46a afforded 46c (2.39 g,
49%) as off-white foam. NMR (CDCl3) δ 0.30 and 0.32 (2 s, 2 ꢀ 3H), 0.94
(d, 6H, J = 7), 0.99 (s, 6H), 1.70-1.85 (m, 1H), 2.32 (s, 3H), 2.96 (dd, 1H,
J = 14, 4), 3.38 (dd, 1H, J = 14, 7), 3.62 (d, 1H, J= 11), 3.83 (s, 3H), 3.87 (d,
1H, J = 11), 3.98-4.19 (m, 2H), 4.58 and 4.99 (2 d, 2 ꢀ 1H, J = 12), 5.98-
5.68 (m, 1H), 6.44 (s, 1H), 7.42 (d, 1H, J = 8), 7.76 (t, 1H, J = 6).
(R)-4-Bromo-1-hydroxy-3-methoxy-13-(3-methyl-[1,2,4]-
oxadiazol-5-yl)-9,10,13,14-tetrahydro-12H,16H-6-oxa-15-
thia-9,12-diaza-benzocyclotetradecene-5,8,11-trione (52). A
mixture of 46c (462 mg, 0.70 mmol) and NH4F(50mg, 1.35mol) inMeOH
(6 mL) was stirred at 20 °C for 1 h. The mixture was diluted with EtOAc
(50 mL) and washed with H2O. The organic layer was dried and evaporated,
and the residue was purified by chromatography (acetone/hexane 1:1) and
crystallized (MeOH/EtOAc/hexane) to give 52 (151 mg, 42%) as white
solid. 1H NMR (DMSO-d6) δ2.31 (s, 3H), 2.88 (dd, 1H, J=14,4),3.22(dd,
1H, J = 14, 7), 3.55 (d, 1H, J = 12), 3.80 (s, 3H), 3.86 (d, 1H, J = 12)
superimposed by 3.78-3.89 (m, 2H), 4.74 and 4.90 (2 d, 2 ꢀ 1H, J = 12),
5.48-5.58 (m, 1H), 6.67 (s, 1H), 7.42 (d, 1H, J=8),9.31(t,1H,J=6),10.44
(s, 1H). HRMS calcd for (C18H19BrN4O7S) 514.0158, found 514.0154.
(R)-4-Bromo-1-hydroxy-3-methoxy-13-(3-methyl-[1,2,4]-
oxadiazol-5-yl)-11-thioxo-9,10,11,12,13,14-hexahydro-16H-
6-oxa-15-thia-9,12-diaza-benzocyclotetradecene-5,8-dione
(53) and (R)-4-Bromo-1-hydroxy-3-methoxy-13-(3-methyl-
[1,2,4]oxadiazol-5-yl)-8,11-dithioxo-7,8,9,10,11,12,13,14-
octahydro-16H-6-oxa-15-thia-9,12-diaza-benzocyclotetra-
decen-5-one (54). A mixture of 46c (328 mg, 0.50 mmol) and
Lawesson’s reagent (161 mg, 0.4 mmol) in toluene (5 mL) was heated to
80 °C for 0.5 h. The mixture was cooled and evaporated, and the residue
was chromatographed (EtOAc/hexane 1:1) to give two major products:
F1 (Si*-54, 55 mg, 16%) and F2 (Si*-53, 75 mg, 22%), with Rf values
(EtOAc) of 0.66 and 0.43, respectively. The silylated products were
treated individually with NHF4/MeOH as described in the preparation
of 47 to give, after purification by chromatography (EtOAc/hexane) and
crystallization (EtOAc/hexane), the following products:
(R)-1-Hydroxy-3-methoxy-4-methyl-13-(3-methyl-[1,2,4]-
oxadiazol-5-yl)-9,10,13,14-tetrahydro-12H,16H-6-oxa-15-thia-
9,12-diaza-benzocyclotetradecene-5,8,11-trione (47). To a
solution of 46a (198 mg, 0.33 mmol) in MeOH (6 mL) was added
NH4F (20 mg, 0.1 mmol), and the mixture was stirred at 20 °C for 1 h.
Workup (EtOAc, H2O) and chromatography (EtOAc/hexane 1:1)
followed by recrystallization (MeOH/EtOAc/hexane) of the crude
1
product gave 47 (83 mg, 55%) as white solid. H NMR (DMSO-d6)
δ 1.96 and 2.31 (2 s, 2 ꢀ 3H), 2.83 (dd, 1H, J = 15, 3), 3.28 (dd, 1H, J =
15, 7), 3.54 (d, 1H, J = 12), 3.74 (s, 3H), 3.80 (d, 1H, J = 12), 3.85-3.93
(m, 2H), 4.68 and 4.90 (2 d, 2 ꢀ 1H, J = 12), 5.52-5.63 (m, 1H), 6.55
(s, 1H), 7.51 (d, 1H, J = 8), 9.25 (t, 1H, J = 6), 9.85 (s, 1H). MS (ISN)
449.3 [(M - 1)-]. Anal. (C19H22N4O7S 0.2MeOH) C, H, N, S.
3
(R)-1-Hydroxy-3-methoxy-4-methyl-13-(3-methyl-[1,2,4]-
oxadiazol-5-yl)-11-thioxo-9,10,11,12,13,14-hexahydro-16H-
6-oxa-15-thia-9,12-diaza-benzocyclotetradecene-5,8-dione
(48), (R)-1-Hydroxy-3-methoxy-4-methyl-13-(3-methyl-[1,2,4]-
oxadiazol-5-yl)-8-thioxo-7,8,9,10,13,14-hexahydro-12H,16H-6-
oxa-15-thia-9,12-diaza-benzocyclotetra-decene-5,11-dione
(49), and (R)-1-Hydroxy-3-methoxy-4-methyl-13-(3-methyl-
[1,2,4]oxa-diazol-5-yl)-8,11-dithioxo-7,8,9,10,11,12,13,14-
octahydro-16H-6-oxa-15-thia-9,12-diaza-benzocyclotetra-
decen-5-one (50). A mixture of 46a (295 mg, 0.50 mmol) and
Lawesson’s reagent (161 mg, 0.4 mmol) in toluene (5 mL) was heated
to 80 °C for 0.5 h. The mixture was cooled and evaporated, and the residue
was chromatographed (EtOAc/hexane 1:1) to give three products: F1 (Si*-
50, 77 mg, 25%), F2 (Si*-48, 90 mg, 29%), and F3 (Si*-49, 51 mg, 16%),
with Rf values (EtOAc/hexane 1:1) of 0.63, 0.45, and 0.40, respectively. The
silylated intermediates were treated individually with NHF4/MeOH as
described in the preparation of 47 to give the following products:
48 (61 mg, 88%, from 90 mg F2), white solid. 1H NMR (DMSO-d6)
δ 1.99 and 2.33 (2 s, 2 ꢀ 3H), 3.09 (dd, 1H, J = 15, 3), 3.31 (dd, 1H, J =
15, 7), 3.57 (d, 1H, J = 12), 3.74 (s, 3H), 3.81 (d, 1H, J = 12), 4.32 (d, 2H,
J = 6), 4.76 and 4.91 (2 d, 2 ꢀ 1H, J = 13), 6.03-6.15 (m, 1H), 6.56 (s,
1H), 9.20 (t, 1H, J = 5), 9.30 (d, 1H, J = 8), 9.87 (s, 1H). HRMS calcd for
(C19H22N4O6S2) 466.0981, found 466.0978.
53 (11 mg, 18%, from 75 mg F2), white solid. 1H NMR (DMSO-d6)
δ 2.33 (s, 3H), 3.11 (dd, 1H, J = 15, 3), 3.32 (dd, 1H, J = 15, 8), 3.59 (d,
1H, J = 12), 3.81 (s, 3H), 3.82 (d, 1H, J = 12), 4.33 (d, 2H, J = 6), 4.78
49 (33 mg, 85%, from 51 mg F3), white solid (see SI).
2221
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
and 4.92 (2 d, 2 ꢀ 1H, J = 13), 6.05-6.16 (m, 1H), 6.68 (s, 1H), 9.20-
9.32 (m, 2H), 10.46 (s, 1H). HRMS calcd for (C18H19BrN4O6S2)
529.9929, found 529.9927.
(m, 1H), 6.71 (s, 1H), 9.31 (t, 1H, J = 6), 9.35 (d, 1H, 8), 10.50 (s, 1H).
HRMS calcd for (C21H26BrN5O6S2) 587.0508, found 587.0509.
DNA Gyrase Inhibition. The MNEC (maximum noneffective
concentration) was determined in a supercoiling assay as described
previously.10 The MNEC, determined visually, was found to be 2-5
times lower than the IC50 of the supercoiling reaction.
In Vitro Antibacterial Activity. Minimum inhibitory concentra-
tions (MICs) for the test compounds against the test strains were
determined by an agar dilution method as described previously.10 Details
are indicated in the table legends.
In Vivo Antibacterial Efficacy. The in vivo efficacy was deter-
mined in a septicaemia mice model as described previously.10 Staphy-
lococcus aureus Smith was used as test organism. In experiments with 0-
1 survivals in a control group of 5-10 mice, the lowest dose affording
3-5 survivals out of a group of 5 infected mice was taken as the 50%
effective dose (ED50, in mg/kg).
54 (5 mg, 11%, from 55 mg F1), white solid (see SI).
Allyl 2-{(R)-2-[3-(Allyloxycarbonylamino-methyl)-[1,2,4]-
oxadiazol-5-yl]-2-[2-(2-hydroxy-acetylamino)-acetylamino]-
ethylsulfanylmethyl}-6-bromo-3-[dimethyl(thexyl)silyloxy]-
5-methoxy-benzoate (89). Reacting 35c (7.0 g, 10.0 mmol) with 44
(4.5 g, 12.0 mmol) and treating the resulting product with pTsOH/
MeOH, as described above in the preparation of 45a (i-ii), gave 89
(6.81 g, 84%) as colorless foam. NMR (CDCl3) δ 0.30 and 0.32 (2 ꢀ
3H), 0.93 (d, 6H, J = 7), 0.98 (s, 6H), 1.68-1.80 (m, 1H), 2.89-3.11
(m, 2H), 3.61 and 3.75 (2 d, 2 ꢀ 1H, J = 13), 3.85 (s, 3H), 3.97-4.16
(m, 4H), 4.48, 4.58, and 4,85 (3 d, 3 ꢀ 2H, J = 6), 5.15-5.37 (m, 4H),
5.47 (d, 1H, J = 17), 5.62 (t, 1H, J = 6), 5.80-6.14 (m, 2H), 6.46 (s, 1H),
7.31 (d, 1H, J = 8).
Allyl [5-((R)-4-Bromo-1-[dimethyl(thexyl)silyloxy]-3-meth-
oxy-5,8,11-trioxo-7,8,9,10,11,12,13,14-octahydro-5H,16H-
6-oxa-15-thia-9,12-diaza-benzocyclotetradecen-13-yl)-[1,2,4]-
oxadiazol-3-ylmethyl]-carbamate (90). Allyl ester 89 (2.04 g,
2.50 mmol) was subjected successively to the ester cleavage and the
lactonization procedures described above (45a (iii) and 46a) to give 90
(0.77 g, 41%) as white foam. NMR (CDCl3) δ 0.30 and 0.33 (2 s, 2 ꢀ
3H), 0.94 (d, 6H, J = 7), 1.00 (s, 6H), 1.68-1.84 (m, 1H), 2.97 (dd, 1H,
J = 15, 3), 3.31 (dd, 1H, J = 15, 6), 3.60 (d, 1H, J = 11), 3.84 (s, 3H), 3.86
(d, 1H, J = 11), 4.15-4.22 (m, 2H), 4.47 and 4.62 (2 d, 2 ꢀ 2H, J = 6),
4.62 and 5.01 (2 d, 2 ꢀ 1H, J = 12), 5.20 (d, 1H, J = 10), 5.29 (d, 1H, J =
17), 5.44 (t, 1H, J = 6), 5.55-5.68 and 5.82-5.99 (2 m, 2 ꢀ 1H), 6.46 (s,
1H), 7.44 (t, 1H, J = 6), 7.48 (d, 1H, J = 8).
’ ASSOCIATED CONTENT
S
Supporting Information. Additional synthetic proce-
b
dures and characterization of new compounds; X-ray crystal-
lographic information for compounds 3, 19b, and 52. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Accession Codes
†X-ray structures of compounds 3, 19b, and 52, are deposited
with Cambridge Crystallographic Data Centre under CCDCs
797879, 797880, and 797881, respectively.
(R)-13-(3-Aminomethyl-[1,2,4]oxadiazol-5-yl)-4-bromo-1-
[dimethyl(thexyl)silyloxy]-3-methoxy-11-thioxo-9,10,11,12,
13,14-hexahydro-16H-6-oxa-15-thia-9,12-diaza-benzocyclo-
tetradecene-5,8-dione (92). A mixture of 90 (0.36 g, 0.48 mmol)
and Lawesson’s reagent (0.70 g, 0.24 mmol) in toluene (5 mL) was
heated to 80 °C for 0.5 h. The mixture was cooled and evaporated, and
the residue was chromatographed (EtOAc/hexane 1:1) to give three
products with Rf values (EtOAc/hexane 2:1) of 0.60, 0.46, and 0.29,
respectively: To a solution of the product of the second fraction (Rf =
0.46, 77 mg) in DCM (1.4 mL) were added dimethylamino-trimethyl-
silane (0.10 mL, 0.6 mmol) and trimethylsilyl trifluoroacetate (0.10 mL,
0.6 mmol). The solution was stirred at 0 °C for 10 min, whereupon
Pd(PPh3)4 (7 mg, 0.006 mmol) was added and stirring was continued
for 2 h at 0 °C. The mixture was evaporated and the residual oil was
worked up (EtOAc, 5% aq NaHCO3, brine) to give, after chromatog-
raphy (EtOAc) of the crude product, 92 (63 mg, 91%) as light-yellow
foam. TLC: Rf = 0.11 (EtOAc).
’ AUTHOR INFORMATION
Corresponding Author
*Phone: þ41 61 711 7385. E-mail: goetschi@intergga.ch
Present Addresses
‡GeneData AG, PO Box, CH-4016 Basel, Switzerland.
§Basilea Pharmaceutica International Ltd., PO Box, CH-4005
Basel, Switzerland.
’ ACKNOWLEDGMENT
We thank Silvia Herzig, Serge Corpataux, and Corinne Lohri
for chemical syntheses, Karin Kuratli, Veronique Schirmer, and
Harry Straub for the biological evaluation, Christian Bartelmus
and Wolf Arnold for spectroscopical analyses, and Bj€orn Wagner
for the determination of physicochemical parameters.
(R)-4-Bromo-1-hydroxy-13-[3-(isopropylamino-methyl)-
[1,2,4]oxadiazol-5-yl]-3-methoxy-11-thioxo-9,10,11,12,13,
14-hexahydro-16H-6-oxa-15-thia-9,12-diaza-benzocyclote-
tradecene-5,8-dione Hydrochloride (97). (i) To a mixture of 92
(30 mg, 0.044 mmol), acetone (0.06 mL), NaOAc (10 mg), AcOH
(0.07 mL), and H2O (0.15 mL) in THF (0.3 mL) was added at 0 °C
portionwise over 30 min NaBH4 (10 mg, 0.26 mmol). Stirring was
continued for 10 min, and the mixture was then worked up (EtOAc, aq
NaHCO3, brine). The residue was chromatographed (EtOAc) to yield a
major product (17 mg). (ii) A solution of this material (17 mg) and
NH4F (10 mg) in MeOH (1 mL) was stirred at 20 °C for 0.5 h. After
workup (EtOAc, H2O), a solution of the resulting residue in EtOAc
(1 mL) was treated with 3 N HCl-Et2O (0.05 mL) and Et2O (1 mL).
The precipitate formed was isolated to give 97 (12 mg, 41%) as white
solid. 1H NMR (DMSO-d6) δ 1.23 (d, 6H, J = 6), 3.19 (dd, 1H, J = 15,
3), 3.29 (dd, 1H, J = 15, 8), 3.64 (d, 1H, J = 11), 3.81 (s, 3H), 3.89 (d, J =
11), 4.22-4.50 (m, 3H), 4.82 and 4,88 (2 d, 2 ꢀ 1H, J = 12), 6.12-6.23
’ ABBREVIATIONS USED
ATP, 50-adenosine triphosphate;ADPNP, 50-adenylyl-β,γ-imido-
diphosphate; MNEC, maximum noneffective concentration;
MIC, minimum inhibitory concentration;MRSA, methicillin-re-
sistant Staphylococcus aureus; CLOGP, calculated log P value.;
SiO2, silica gel;DCM, dichloromethane;DEAD, diethyl azodicar-
boxylate; DCC, N, N0-dicyclohexyl-carbodiimide; EDC, N-(di-
methylamino-propyl)-N0-ethyl-carbodiimide hydrochloride; Law-
esson’s reagent, [2,4-bis-(4-methoxyphenyl)-2,4-dithioxo-1,3,2,-
4-dithiaphosphetane]; MeCN, acetonitrile; TFA, trifluoroace-
tic acid; pTsOH, p-toluenesulfonic acid monohydrate; trityl,
triphenylmethyl;thexyl, 1, 1, 2-trimethyl-propyl; Si *, dimethyl-
(thexyl)silyl; Boc, tert-butoxycarbonyl;Rf, TLC retention time;SI,
Supporting Information
2222
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
’ ADDITIONAL NOTE
(12) (a) Hebeisen, P.; Angehrn, P.; Gmuender, H.; Goetschi, E.;
Link, H.; Luebbers, T.; Schneider, F. New DNA gyrase inhibitors related
to cyclothialidine: seco-analogues as potent antibacterial agents. Pro-
gramme and Abstracts of the 20th International Congress of Che-
motherapy, Sydney, Australia, June 29-July 3, 1997, Poster 042. (b)
Rudolph, J.; Theis, H.; Hanke, R.; Endermann, R.; Johannsen, L.;
Geschke, F.-U. seco-Cyclothialidines: new concise synthesis, inhibitory
activity toward bacterial and human DNA topoisomerases, and anti-
bacterial properties. J. Med. Chem. 2001, 44, 619–626.
(13) (a) Luebbers, T.; Angehrn, P.; Gmuender, H.; Herzig., S.
Design, synthesis, and structure-activity relationship studies of new
phenolic DNA gyrase inhibitors. Bioorg. Med. Chem. Lett. 2007,
17, 4708–4714. (b) Brvar, M.; Perdih, A.; Oblak, M.; Maꢀsic, L. P.;
Solmajer, T. In silico discovery of 2-amino-4-(2,4-dihydroxyphenyl)thia-
zoles as novel inhibitors of DNA gyrase B. Bioorg. Med. Chem. Lett. 2010,
20, 958–962.
Systematic IUPAC (International Union of Pure and Applied
Chemistry) chemical names were used for all new compounds.11
’ REFERENCES
(1) (a) Wood, M. J. Chemotherapy of Gram-positive nosocomial
sepsis. J. Chemother. 1999, 11, 446–452. (b) Chopra, I. Antibiotic
resistance in Staphylococcus aureus: concerns, causes and cures. Expert
Rev. Anti-Infect. Ther. 2003, 1, 45–55. (c) Barrett, C. T.; Barrett, J. F.
Antibacterials: are the new entries enough to deal with the emerging
resistance problems?. Curr. Opin. Biotechnol. 2003, 14, 621–626.
(2) Singh, S. B.; Barrett, J. F. Empirical antibacterial drug discovery—
foundation in natural products. Biochem. Pharmacol. 2006, 71, 1006–
1015.
(14) Poyser, J. P.; Telford, B.; Timms, D.; Block, M. H.; Hales, N. J.
Triazine derivatives and their use as antibacterials. WO Patent 99/
01442, 1999.
(3) Hooper, D. C.; Wolfson, J. S. Mode of action of new quinolones:
new data. Eur. J. Clin. Microbiol. Infect. Dis. 1991, 10, 223–231.
(4) (a) Kamiyama, T.; Shimma, N.; Ohtsuka, T.; Nakayama, N.;
Itezono, Y.; Nakada, N.; Watanabe, J.; Yokose, K. Cyclothialidine, a
novel DNA gyrase inhibitor. II. Isolation, characterization and structure
elucidation. J. Antibiot. 1994, 47, 37–45. (b) Nakada, N.; Shimada, H.;
Hirata, T.; Aoki, Y.; Kamiyama, T.; Watanabe, J.; Arisawa, M. Biological
characterization of cyclothialidine, a novel DNA gyrase inhibitor. Anti-
microb. Agents Chemother. 1993, 37, 2656–2661.
(5) Lewis, J. R.; Singh, O. M. P.; Smith, C. V.; Skarzynsky, T.;
Maxwell, A.; Wonacott, A. J.; Wigley, D. B. Crystallization of inhibitor
complexes of an N-terminal 24 kDa fragment of the DNA gyrase protein.
J. Mol. Biol. 1994, 241, 128–130.
(6) Yamaji, K.; Masubuchi, M.; Kawahara., F.; Nakamura, J.; Nishio,
A.; Matsukuma, S.; Watanaba, J.; Fujimori, M.; Nakada, N.; Kamiyama,
T. Cyclothialidine analogues, novel DNA gyrase inhibitors. J. Antibiot.
1997, 50, 402–411.
(7) Nakada, N.; Gmuender, H.; Hirata, T.; Arisawa, M. Mechanism
of inhibition of DNA gyrase by cyclothialidine, a novel DNA gyrase
inhibitor. Antimicrob. Agents Chemother. 1994, 38, 1966–1973.
(8) (a) Gellert, M.; O’Dea, M. H.; Itoh, C.; Tomizawa, J. Novobiocin
and coumermycin inhibit DNA supercoiling catalysed by DNA gyrase.
Proc. Natl. Acad. Sci. U.S.A. 1976, 73, 4474–4478. (b) Lambert, H. P.;
O’Grady, F. W. Antibiotics and Chemotherapy. In Coumarins, 6th ed.;
Lambert, H. P., O’Grady, F. W., Eds.; Churchill Livingstone: Edinburgh,
1992; pp 140-141.
(9) (a) Arisawa, M.; Goetschi, E.; Kamiyama, T.; Masciadri, R.;
Shimada, H.; Watanabe, J.; Hebeisen, P.; Link, H. PCT Int. Appl. WO
9218490, 1992; Chem. Abstr. 1993, 119, 117285. (b) Geiwiz, J.;
Goetschi, E.; Hebeisen, P.; Link, H.; Luebbers, T. European Patent
Appl. EP 675122, 1995; Chem. Abstr. 1996, 124, 146213. (c) Goetschi,
E.; Angehrn, P.; Gmuender, H.; Hebeisen, P.; Link, H.; Masciadri, R.;
Nielsen, J. Cyclothialidine and its congeners: a new class of DNA gyrase
inhibitors. Pharmacol. Ther. 1993, 60, 367–380. (d) Goetschi, E.;
Angehrn, P.; Gmuender, H.; Hebeisen, P.; Link, H.; Masciadri, R.;
Reindl, P.; Ricklin, F. The DNA gyrase inhibitor cyclothialidine:
progenitor of a new class of antibacterial agents. In Medicinal Chemistry:
Today and Tomorrow; Yamazaki, M., Ed.; Blackwell Science Ltd.:
Oxford, 1996; pp 263-270. (e) Goetschi, E.; Angehrn, P.; Gmuender,
H.; Hebeisen, P.; Kostrewa, D.; Link, H.; Luebbers, T.; Masciadri, R.;
Reindl, P.; Ricklin, F.; Theil, F. P. From the DNA gyrase inhibitor
cyclothialidine to a new class of antibacterial agents. In Natural Product
Chemistry at the Turn of the Century; Atta-ur-Rahman, Choudhary, M. I.,
Khan, K. M., Eds.; Print Arts: Karachi, 2002; pp 489-497.
(10) Angehrn, P.; Buchmann, S.; Funk, C.; Goetschi, E.; Gmuender,
H.; Hebeisen, P.; Kostrewa, D.; Link, H.; Luebbers, T.; Masciadri, R.;
Nielsen, J.; Reindl, P.; Ricklin, F.; Schmitt-Hoffmann, A.; Theil, F.-P.
New antibacterial agents derived from the DNA gyrase inhibitor
cyclothialidine. J. Med. Chem. 2004, 42, 1487–1513.
(11) In contrast to our earlier papers, we used IUPAC nomenclature,
which gives different numbering, e.g., 1- or 4-hydroxy for the essential
phenol group, dependent on the core scaffold.
(15) Charifson, P. S.; Grillot, A.-L.; Grossman, T. H.; Parsons, J. D.;
Badia, M.; Bellon, S.; Deininger, D. D.; Drumm, J. E.; Gross, C. H.;
LeTiran, A.; Liao, Y.; Mani, N.; Nicolau, D. P.; Perola, E.; Ronkin, S.;
Shannon, D.; Swenson, L. L.; Tang, Q.; Tessier, P. R.; Tian, S.-K.;
Trudeau, M.; Wang, T.; Wei, Y.; Zhang, H.; Stamos, D. Novel dual-
targeting benzimidazole urea inhibitors of DNA gyrase and topoisome-
rase IV possessing potent antibacterial activity: intelligent design and
evolution through the judicious use of structure-guided design and
stucture-activity relationships. J. Med. Chem. 2008, 51, 5243–5263.
(16) (a) Basarab, G. S..; Gravestock, M. B.; Hauck, S. I. Pyrrole
derivatives as dna gyrase and topoisomerase inhibitors. Patent
WO2006/092599, 2006. (b) Basarab, G. S.; Bist, S.; Manchester, J. I.
Patent WO2008/06470, 2008. (c) Bist, S.; Dangel, B.; Sherer, B.
Heterocyclic urea derivatives and methods of use thereof. Patent
WO2009/106885, 2009. (d) Basarab, G. S.; Bist, S.; Manchester, J. I.;
Sherer, B. Alkyl urea substituted pyridines. U.S. Patent US 7,674,801,
2007.
(17) Tanitame, A.; Oyamada, Y.; Ofuji, K.; Suzuki, K.; Ito, H.;
Kawasaki, M.; Wachi, M.; Yamagishi, J. Potent DNA gyrase inhibitors;
novel 5-vinylpyrazole analogues with Gram-positive antibacterial activ-
ity. Bioorg. Med. Chem. Lett. 2004, 14, 2863–2866.
(18) (a) Basarab, G. S. Antibacterial pyrrolecarboxamides.
WO2008/020227, 2008. (b) Gravestock, M. B.; Hull, K. G.; Fleming,
P. R. Antibacterial pyrrolecarboxamides. WO2008/020229, 2008. (c)
Illingworth, R. N.; Uria-Nickelsen, M. R.; Bryant, J.; Eakin, A. New
inhibitors of bacterial topoisomerase GyrA/Par subunits. 48th Annual
ICAAC Meeting, Washington, DC, October, 2008, Poster F1-2028.
(d) Uria-Nickelsen, M. R.; Blodgett, A.; Eakin, A. 48th Annual ICAAC
Meeting, Washington, DC, October, 2008Poster F1-2029.
(19) Kostrewa, D.; D’Arcy, A. Unpublished results.
(20) Lewis, J. R.; Singh, O. M. P.; Smith, C. V.; Skarzynsky, T.;
Maxwell, A.; Wonacott, A. J.; Wigley, D. B. The nature of inhibition of
DNA gyrase by the coumarins and the cyclothialidines revealed by X-ray
crystallography. EMBO J. 1996, 15, 1412–1420.
(21) Wigley, D. B.; Davies, G. J.; Dodson, E. J.; Maxwell, A.; Dodson,
G. Crystal structure of a N-terminal fragment of the DNA gyrase B
protein. Nature 1991, 351, 624–629.
(22) Boehm, H.-J.; Boehringer, M.; Bur, D.; Gmuender, H.; Huber,
W.; Klaus, W.; Kostrewa, D.; Kuehne, H.; Luebbers, T.; Meunier-Keller,
N.; Mueller, F. Novel inhibitors of DNA gyrase: 3D structure based
biased needle screening, hit validation by biophysical methods, and 3D
optimization. A promising alternative to random screening. J. Med.
Chem. 2000, 43, 2664–2674.
(23) Hori, K.; Hikage, N.; Inagaki, A.; Mori, S.; Nomura, K.; Yoshii,
E. Total synthesis of tetronomycin. J. Org. Chem. 1992, 57, 2888–2902.
(24) X-ray structures of compounds 3, 19b, and 52, are deposited
with Cambridge Crystallographic Data Centre under CCDCs 797879,
797880, and 797881, respectively. The oxadiazole ring in compound 52
is disordered in the X-ray structure.
2223
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224

Journal of Medicinal Chemistry
ARTICLE
(25) Knutsen, L. J. S.; Andersen, K. E.; Lau, J.; Lundt, B. F.; Henry,
R. F.; Morton, H. E.; Naerum, L.; Petersen, H.; Stephensen, H.; Suzdak,
P. D.; Swedberg, M. D. B.; Thomsen, C.; Soerensen, P. O. Synthesis of
novel GABA uptake inhibitors. 3. Diaryloxime and diarylvinyl ether
derivatives of nipecotic acid and guvacine as anticonvulsant agents.
J. Med. Chem. 1999, 42, 3447–3462.
(26) Broadhurst, M. J.; Hassall, C. D.; Thomas, G. J. Tetracycline
studies. Part 5. New syntheses of anthracenes and anthraquinones
through benzophenone carbanions. J. Chem. Soc., Perkin Trans. 1
1977, 2502–2512.
(27) Geiwiz, J.; Goetschi, E.; Hebeisen, P. Concise synthesis of
cyclothialidine analogues with ring sizes from 12 to 15: novel macro-
cyclization protocol involving reductive thiolation. Synthesis 2003
1699–1704.
(28) Aizpurua, J. M.; Lecea, B.; Palomo, C. Reduction of carbonyl
compounds promoted by silicon hydrides under the influence of
trimethylsilyl-based reagents. Can. J. Chem. 1986, 64, 2342–2347.
(29) Joerres, V.; Keul, H.; Hoecker, H. Aminolysis of R-hydroxy acid
esters with R-amino acid salts. First step in the synthesis of optically
active 2,5-morpholinediones. Macromol. Chem. Phys. 1998, 199, 825–
833.
(30) Lowe, G.; Vilaivan, T. Amino acids bearing nucleobases for the
synthesis of novel peptide nucleic acids. J. Chem. Soc., Perkin Trans. 1
1997, 539–546.
(31) (a) CORINA, version 3.46, 3D structure generator (www.
molecular-networks.com); (b) Moloc, molecular modeling package, e.g.
for generation of conformation libraries (www.moloc.ch).
(32) (a) Ruse, M. J.; Waring, R. H. The effect of methimazole on
thioamide bioactivation and toxicity. Toxicol. Lett. 1991, 58, 37–41. (b)
Ruse, M. J.; Waring, R. H. The metabolism of thionicotinamide in the
rat. Drug Metab. Drug Interact. 1991, 9, 123–137.
(33) Merzouk, A.; Guibꢁe, F.; Loffet, A. On the use of silylated
nucleophiles in the palladium catalyzed deprotection of allylic carbox-
ylates and carbamates. Tetrahedron Lett. 1992, 33, 477–480.
(34) Iwata, C.; Maezaki, N; Murakami, M.; Soejima, M.; Tanaka, T.;
Imanishi, T. Asymmetric functionalization of prochiral 1,3-diols based
on an efficient 1,6-chiral induction: the diastereoselective C-O bond
fission in chiral β-arylsulfinyl acetal via two types of chelation control.
J. Chem. Soc., Chem. Commun. 1992, 516–518.
(35) Ngeleka, M.; Auclair, P.; Tardif, D.; Beauchamp, D.; Bergeron,
M. G. Intrarenal distribution of vancomycin in endotoxemic rats.
Antimicrob. Agents Chemother. 1989, 33, 1575–1579.
2224
dx.doi.org/10.1021/jm1014023 |J. Med. Chem. 2011, 54, 2207–2224