Structure aided design of chimeric antibiotics

Article abstract of DOI:10.1016/j.bmcl.2012.02.019

The rise of antibiotic resistance is of great clinical concern. One approach to reducing the development of resistance is to co-administer two or more antibiotics with different modes of action. However, it can be difficult to control the distribution and

Full text of DOI:10.1016/j.bmcl.2012.02.019

Bioorganic & Medicinal Chemistry Letters 22 (2012) 2428–2433
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure aided design of chimeric antibiotics
Tomislav Karoli, Sreeman K. Mamidyala, Johannes Zuegg, Scott R. Fry, Ernest H.L. Tee, Tanya A. Bradford,
⇑
Praveen K. Madala, Johnny X. Huang, Soumya Ramu, Mark S. Butler, Matthew A. Cooper
Institute for Molecular Bioscience, University of Queensland, St Lucia, Queensland 4072, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
The rise of antibiotic resistance is of great clinical concern. One approach to reducing the development of
resistance is to co-administer two or more antibiotics with different modes of action. However, it can be
difficult to control the distribution and pharmacokinetics of two drugs to ensure both concentrations
remain within the range of therapeutic efficacy whilst avoiding adverse effects. Hybrid drugs, where
two drugs are linked together with a flexible linker, have been explored, but the resultant large, flexible
molecules can have poor bioavailability. We have developed a chimeric approach using click chemistry
where the pharmacophores of two drugs are overlapped into a single smaller, more drug-like molecule.
Design and selection of compounds were assisted by in silico structural docking. We prepared a series of
compounds that include candidates showing activity against the targets of both trimethoprim; dihydro-
folate reductase, and ciprofloxacin; DNA gyrase and topoisomerase IV. The resultant triazole containing
molecules show modest, but broad spectrum activities against drug sensitive and resistant Gram-nega-
tive and Gram-positive bacteria, with no observable cytotoxicity.
Received 19 December 2011
Revised 6 February 2012
Accepted 7 February 2012
Available online 22 February 2012
Keywords:
Chimeric antibiotics
Drug resistant bacteria
DNA gyrase
Dihydrofolate reductase
Click chemistry
Ó 2012 Elsevier Ltd. All rights reserved.
Drug resistant microorganisms and the paucity of new antibiot-
ics in development pose a serious and growing threat to human
health.1 Of particular concern are drug resistant Gram-negative bac-
teria that possess multiple mechanisms to acquire genetic material
encoding antibiotic resistance. The recent outbreak of carbapenem
resistant Acinetobacter baumannii, Klebsiella pneumoniae and
Escherichia coli expressing the plasmid-encoded metallo-b-lacta-
mase NDM-1 was a frightening example of this rapid global
horizontal gene transfer.2 One recent approach to combating resis-
tant pathogens is the development of hybrid antibiotic molecules in
which two drugs with different modes of action are joined together,
often using a flexible linker, in an attempt to improve on traditional
drug combination therapies. While this strategy combines two
modes of action into one chemical entity, simplifying optimisation
of the pharmacokinetics/pharmacodynamic (PK/PD) profile, effi-
cacy at both targets is usually compromised. The hybrids often have
a more flexible structure, thereby reducing target binding efficiency,
and have a higher molecular weight, which negatively impacts both
oral bioavailability and the ability to penetrate the outer membrane
of Gram-negative bacteria.3
further development has been halted include MCB 3681, an oxazo-
lidinone–quinolone hybrid that reached Phase I,5 CBR-2092,
a
rifampicin–quinolone hybrid that completed Phase I, and
TD-1792,6 a vancomycin–cephalosporin hybrid that completed
Phase II.7
We propose a progressive improvement of the hybrid antibiotic
concept with a chimeric strategy.5 Existing drugs are conceptually
broken down into fragments and then recombined into a single,
chimeric, drug-like molecule bearing the required pharmacophores
for the different bacterial targets. This modification maintains the
requisite molecular size and physicochemical properties that may
be suitable for an efficacious, orally available Gram-negative anti-
biotic. Reassembly of the fragments into active molecules is aided
by extensive molecular modelling, docking and structural studies.
To validate our approach, we have selected the benzyl pyrimi-
dine (e.g., TMP and iclaprim) and the fluoroquinolone (FQ, e.g., cip-
rofloxacin (CIP)) classes of antibiotics as (i) both show broad
spectrum activity against both Gram-negative and Gram-positive
bacteria, (ii) drug synergy studies show that they have a additive
and no antagonistic effect8 (iii) X-ray crystal structures of benzyl
pyrimidines, and FQs bound to their targets are available and,
(iv) extensive SAR information is known for both classes.
The development of linked hybrid antibiotic has been problem-
atic to date, with only cadazolid (ACT-179811), an oxazolidinone
and quinolone hybrid being actively developed in the clinic
where it is being evaluated in Phase II trials for the treatment
of Clostridium difficile infections.4 Other recent hybrids where
Benzyl pyrimidines inhibit DHFR, an essential enzyme which is
critical in the biosynthesis of purines and amino acids. DHFR catal-
yses the reduction of 7,8-dihydrofolate to 5,6,7,8-tetrahydrofolate
using nicotinamide adenine dinucleotide phosphate oxidase
(NADPH) as a co-factor. Numerous X-ray crystal structures of bacte-
rial DHFR with various inhibitors have been published with the
⇑
Corresponding author. Tel.: +61 7 3346 2044; fax: +61 7 3346 2090.
E-mail addresses: m.cooper@uq.edu.au, m.cooper@imb.uq.edu.au (M.A. Cooper).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.02.019

T. Karoli et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2428–2433
2429
Protein DataBank (PDB) containing 133 entries to date, but lacking
the DHFR of important pathogenic Gram-negative organisms.
Homology models of DHFR from K. pneumoniae, Pseudomonas aeru-
ginosa and A. baumannii were built using the crystal structure of
E. coli DHFR (Pdb: 2ANO13) as template (Fig. 1; see Supplementary
data for detail). Comparison of the DHFR active site residues from
these bacteria revealed that they are very well conserved, showing
the potential for broad spectrum activity against these pathogens.14
Most importantly, bacterial DHFRs show only <30% homology with
human DHFR versus 30–45% within bacterial species, providing a
selectivity window for the benzyl pyrimidines.
key pharmacophore activity sites of TMP and FQs showed promis-
ing activity against resistant strains (e.g., MP-1341, MP-1350, and
MP-1362, Fig. 2). The activity profiles were equivalent, or better
than equimolar mixtures of CIP with TMP.
Translating these results into our chimeric strategy, we sought
to design new compounds where the pharmacophores of both anti-
biotic compounds are combined into a single small molecule. The
initial designs (Fig. 3) were done by evaluating the published SARs
for TMP, iclaprim and FQs. In addition, the design varied the degree
of overlap of the pharmacophores, exploring the effect of molecular
weight, potentially improving their oral bioavailability and Gram-
negative activity.15
Fluoroquinolones are known to predominantly target two sim-
ilar and essential bacterial enzymes, DNA gyrase (topoisomerase II)
and topoisomerase IV. DNA gyrase and topoisomerase IV both con-
tain four subunits (2 GyrA and GyrB and 2 ParC and ParE) and are
both classed as type II topoisomerases. DNA gyrase introduces neg-
ative supercoils into DNA, while topoisomerase IV catalyses ATP-
dependent chromosome decatenation and relaxation.9 The primary
target of most quinolones in Gram-negative bacteria is DNA gyrase,
while the primary target in Gram-positive bacteria is a combina-
tion of topoisomerase IV and DNA gyrase.10 The exact nature of
the quinolone interaction was first reported in 2009 by a X-ray
crystal structures of Streptococcus pneumoniae ParC breakage-re-
union and ParE TOPRIM domains stabilized by two FQs, moxiflox-
acin and clinafloxacin.11 The cleaved DNA is thereby bound into the
topoisomerase IV with the FQs intercalated into a gap of the DNA
between the À1 and +1 nucleotide pairs. Due to this complex mode
of inhibition a structural comparison between the mode of inhibi-
tion of Gram-positive and Gram-negative gyrases are not meaning-
ful, hence no homology models have been attempted for the
gyrases.
The initial target compounds were synthesised via a napthyri-
dine common intermediate, as either the ethyl ester 1 or free acid
2, prepared by cyclisation of an enamino keto ester using the meth-
od of Bouzard18 and Chu19 or Shin20 (Scheme 1). The resulting chlo-
ronaphthyridine was treated with an appropriate amine to form
the desired chimeric antibacterial candidate.
Chimera 4 was prepared in three steps via methylamino piper-
azine 6 (Scheme 3).21 Reduction of nitrile 5 to the desired amine 6
proved to be challenging due to instability of the intermediate
imine, however, small quantities of the amine were obtained. With
amine 6 in hand, the target molecule 4 was prepared using the
published condensation method.12 After preparative HPLC purifica-
tion, only a small quantity of the desired product was obtained at
moderate purity and the material was subjected to biological as-
say. Unfortunately, chimera 4 was found to be inactive against
Gram-positive and Gram-negative bacteria (Table 1).
The lack of activity led us to believe that we had not included
enough of the required pharmacophore of the DHFR inhibitors,
and attention shifted to the less overlapped compound 3. This
compound was intended to represent the case where the two anti-
biotics are more representative of the respective pharmacophores.
Precedent for the activity of aminobenzyl pyrimidines against
DHFR was known.22 Preparation of the TMP portion of the mole-
cule was relatively straightforward (Scheme 2). The aldehyde
group of 5-nitrovanillin 7 was protected as the dioxolane before
catalytic transfer hydrogenation of the nitro group and cyclisation
of the resulting amine onto the adjacent phenol was accomplished
using dibromoethane.23
Published results12 of a series of TMP-FQ hybrid style molecules
synthesised using a linking strategy that did not interfere with the
Deprotection of the dioxolane protected aldehyde 10 and cycli-
sation of the diaminopyrimidine using the standard methodology
provided the desired amine 12.24 Surprisingly, attempts to form
the diaminopyrimidine in the presence of the nitro group were
unsuccessful. In contrast, condensation of amine 12 with napthyri-
dine 2 proved to be much more problematic. Under standard
conditions12 only traces of compounds with the desired mass spec-
trum were observed by LC–MS. Attempts to isolate these materials
via preparative HPLC were unsuccessful. The use of Buchwald-type
S. aureus
E. coli
A. baumannii P. aeruginosa K. pneumoniae
Leu5
Ile5
Val11
Ala13
Leu26
Leu34
Phe37
Lys38
Thr52
Met56
Leu60
Ile107
Ile8
Met6
Ala7
Ala7
Ala10
Leu23
Leu31
Phe34
Lys35
Thr49
Leu53
Leu57
Ile104
Gly106
Tyr110
Ala8
Leu20
Leu28
Val31
Lys32
Thr46
Ile50
Met20
Leu28
Phe31
Lys32
Thr46
Ile50
Ile21
Gln29
Phe32
Lys33
Thr47
Met51
Leu54
Ser97
Gly99
Tyr103
Leu54
Phe92
Gly94
Phe98
Leu54
Ile94
Gly96
Tyr100
Gly109
Phe113
Figure 1. Top: Homology models of DHFR enzymes from E. coli (red), A. baumannii
(blue, 33% identity/52% similarity to E. coli), P. aeruginosa (orange, 44%/61%), K.
pneumoniae (green, 32%/56%). The dihydrofolate substrate is shown in magenta
bound at the centre of the enzymes. Bottom: Comparison on DHFR active site
residues for pathogens of interest.
Figure 2. Published trimethoprim/fluoroquinolone hybrid structures.

2430
T. Karoli et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2428–2433
O
N
O
activity for this class of molecules, compound 12 was assayed
and found to be only weakly active against two of the Gram-nega-
tive strains tested.
NH2
N
F
OMe
OMe
OH
N
N
H2N
Examination of the difficulties encountered in the preparation
of compounds 3 and 4 showed that the vigorous conditions re-
quired for coupling the amino pyrimidine and napthyridine com-
ponents were causing decomposition of the products. To provide
a more gentle method of combining the fragments, ‘click’ triazole
preparation26 was explored. At this stage various in silico methods
were added to the design of new compounds, aimed to select com-
pounds able to fit in to the binding sites of both DHFR and DNA
gyrase, with the latter having a much more spatially restricted
binding pocket.
Click chemistry is widely used in drug research for the synthesis
of chemical libraries, and 1,2,3-trizoles containing drugs like tazo-
bactam have reached the market. Compared to the previously
published results12 of TMP-FQ hybrid, the triazole compounds
show slightly better drug-like properties, with a smaller molecular
weight (500 instead of 650), improved predicted solubility
(À3.7logS compared to À5.2), fewer rotatable bonds (8 compared
to 11), and reduced lipophilicity (logP of 0.9 compared to 2.9).
Analysis of published crystal structures of proteins with triazole
containing ligands shows mainly hydrophobic interactions with
the protein, suggesting little electrostatic contribution of the tria-
zole moiety to the overall binding energy.
HN
OMe
Trimethoprim
Ciprofloxacin
H2N
N
NH2
O
O
N
O
N
O
F
NH2
N
OH
F
OH
N
N
H
N
N
MeO
N
N
H2N
O
3
4
Figure 3. Initial target molecules, showing the template drugs.
O
N
O
O
O
F
OR
F
OEt
Cl
N
Cl
N
Cl
Virtual docking studies were undertaken using Gold v5.0.1,17
using standard settings for docking and re-scoring with Chem-
Score. Binding sites were defined as residues within 1 nm of the
corresponding inhibitor found in the crystal structure. While the
docking into E. coli DHFR (Pdb: 2ANO13) binding site was straight-
forward, the docking into the Staphylococcus aureus topoisomerase
II DNA-gyrase (Pdb: 2XCT16) required two additional settings in
order to generate comparable docking poses in the FQ binding site:
(i) scaffold constraint to the core of the ciprofloxacin inhibitor and
(ii) definition of flexible side chains of Asp437, Arg458, Asn476 and
Glu477. Compound poses not in the FQ binding site were not con-
sidered in the ranking. In both docking studies the inhibitors found
in the crystal structures were used as an internal control, with both
inhibitors ranking at the very top. The proposed compounds were
ranked according to both virtual docking studies, eliminating com-
pounds that did not fit, or that had low ranking in both docking
studies (Fig. 4).
1: R = Et
2: R = H
Scheme 1. Common intermediate synthesis.
palladium chemistry25 provided products in sufficient yield and
quality for isolation by HPLC.
Small amounts of two distinct stable compounds were isolated
(compounds 13 and 14) showing the correct mass for the desired
product; however, NMR analysis showed that neither was the de-
sired material, and instead were the products of coupling at the
pyrimidine NH2 groups rather than the expected morpholino nitro-
gen. These compounds were assayed in our antibacterial panel
(Table 1) but were found to be inactive. In terms of the antifolate
activity, this result is not unexpected given that the diaminopyrim-
idine portion of the molecule is deeply buried in the enzyme and is
the key pharmacophore of the drug class. Attempts to protect the
diaminopyrimidine portion of 12 to allow reaction at the morpho-
lino group were unsuccessful. To examine the probability of
For the synthesis of those compounds, azide and alkyne
analogues of the pyrimidine and napthyridine structures were re-
quired. The problematic imine stability observed during reduction
of compound 5 (Scheme 1) was exploited27 for preparation of the
Table 1
Assay results for chimeric compounds
Compound
E. colia s–sj
G Àve (MIC
l
g/mL)
G +ve (MIC
S. aureuses–s
l
g/mL)
Enzymatic assays (IC50
l
M)
K. pneumbs–r
P. aerugcs–r
E. colidr–r
S. epidermfs–r
DNA Gyrase
TopoIV
DHFR
g
CIPh
60.03
0.125–0.25
60.03
>64
8–16
>64
>64
16
>64
2
0.125
2
0.5
—
60.03
>64
0.25–0.5
—
>64
>64
>64
>64
>64
>64
>64
>64
>64
—
>64
>64
>64
>64
>64
>64
>64
>64
>64
0.5
1
0.06
>64
0.125–0.25
—
>64
>64
>64
16–32
>64
4
1.0
—
—
—
>250
—
0.4
—
—
—
>250
—
—
TMPi
0.1
—
TMP/CIP 1:1
4
0.5–1
>64
8
>64
>64
32
>64
16
4
—
12
13
14
22
24
26
28
32
34
32
0.3
—
>64
>64
>64
>64
64
16–32
8
>64
—
—
—
166
>250
17
48
10
>250
>250
>250
4.5
4.4
11
>250
>250
—
>250
>250
50
0.5–1
0.25
4
>64
8–16
64
4–8
0.5
8
0.25–0.5
64
160
Notes: aATCC 25922; bATCC 700603; cATCC 10145; dCIP and SXT resistant clinical isolate; eATCC 25923; fNRS 60; g— = not tested; hciprofloxacin; itrimethoprim; jsensitivity or
resistance to CIP–SXT. SXT = co-trimoxazole (1:5 trimethoprim/sulfamethoxazole).

T. Karoli et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2428–2433
2431
H2N
N
NH2
O
O
O
CHO
O
O
N
i
iii
iv
v, vi
MeO
NO2
MeO
NH
MeO
NH
MeO
ii
R
MeO
NH
OH
O
O
OH
O
8: R = NO2
9: R = NH2
viii
7
10
11
12
vii
O
N
O
F
NH2
O
O
3
OH
H
N
F
N
OH
HN
N
H
N
+
N
N
H
N
N
O
N
OMe
N
NH2
O
OMe
13
14
Scheme 2. Reagents and conditions: (i) ethylene glycol, pTSA, Dean–Stark; (ii) Pd/C, Et3SiH; (iii) K2CO3, dibromoethane, DMF; (iv) H3O+; (v) 3-morpholinopropanenitrile,
NaOMe; (vi) guanidineÁHCl, NaOMe; (vii) 2, Et3N, TMSCl; (viii) 1, Pd2(dba)3, DavePhos, Cs2CO3 then LiOH
NH
H2N
N
NH2
H2N
N
NH2
N
i
H2N
N
NH2
NH2
ii
iii
+
N
4
N
O
5
6
N
iv
H2N
N
NH2
N3
N
vii
H2N
N
NH2
R
19
N
viii
Figure 4. Left: Docked pose for TMP (grey), 12 (cyan) and 34 (magenta) into DHFR.
Right: Docked pose for CIP (grey), 26 (cyan), 28 (red), 32 (pink) and 34 (magenta)
into DNA gyrase.
15: R = CHO
16: R = CH2OH
17: R = CH2Br
v
vi
H2N
N
NH2
N
O
pyrimidine analogues where nitrile 5 was reduced to the imine and
hydrolysed to aldehyde 15. The aldehyde was then reduced with so-
dium borohydride to alcohol 16 which was converted to the unsta-
ble bromide 17 as the hydrobromide salt. This material was readily
converted to the azide 19 or propargyl ether 20 using sodium azide
or propargyl alcohol, respectively (Scheme 3). The napthyridine
analogues were readily prepared from compound 1 by heating with
propargyl amine or sodium azide to give alkyne 21 and azide 25,
respectively. A click reaction28 produced the desired triazoles 24
and 28 after hydrolysis of the ethyl esters (Schemes 4 and 5).
To extend the pharmacophore of the fluoroquinolone portion of
the molecule, an additional compound bearing a piperazine ring
between the napthyridine and pyrimidine moieties was desired.
Beginning with mono-Boc protected piperazine, alkyne 31 was
prepared and reacted with azide 19 to produce chimera 34 after es-
ter hydrolysis (Scheme 6).
Biological assessment of the compounds was performed in
three phases using standard assay procedures, (i) antibacterial
MIC assays against Gram-positive and Gram-negative organisms,
using sensitive and resistant strains, (ii) enzymatic assays against
topoisomerase IV, DNA gyrase and DHFR9b,c and (iii) cytotoxicity
assay against HepG2 and HEK293 cell lines (see Supplementary
data for assay detail).
20
Scheme 3. Reagents and conditions: (i) NaOEt, rt, 53%; (ii) H2/Raney nickel;21
(iii) 2, Et3N, TMSCl: (iv) Raney nickel, HCOOH, reflux, 3 h; (v) NaBH4, MeOH, rt, 5 h;
(vi) HBr/HOAc 60 °C, 2 h, giving HBr salt; (v) NaN3, DMF, 80 °C, 3 h; (vi) propargyl
alcohol.
HEK293 cells. Compound 21 showed a CC50 of around 80 lM for
both cell lines. In general, the ethyl esters (21, 23, 25, 27, 31, and
33) showed no activity in either the cell based MIC assays or in
the enzymatic inhibition assays. Only compound 23 showed a
weak (80%) inhibition of DHFR at the highest concentration tested
(250 lM), but no effect on the growth of the bacteria. On the other
hand, FQ like intermediates (22, 26 and 32) showed, as to be ex-
pected, some activity in the DNA gyrase and topoisomerase IV as-
says, and some activity in the cell based assays, but always less
than is observed for CIP. The piperazine containing intermediate
(32), which is the structurally closest to ciprofloxacin, showed
the best activity profile in enzyme and cell based assays. Similarly,
the TMP-like intermediate 12 showed activity against DHFR that
was translated into mild in vitro activity against sensitive
Gram-positive and Gram-negative strains. Interestingly, 12
showed activity against the SXT resistant K. pneumoniae strain.
From the designed compounds only 28 and 34 showed some
enzymatic and in vitro activity, whereas 24 did not show any
Of the compounds tested, only 21 showed cytotoxicity up to the
maximum tested concentration of 100 lM, against either HepG2 or

2432
T. Karoli et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2428–2433
O
N
O
O
N
O
F
F
OR
OR'
i
R
iii
Boc
i
N
1
N
N
H
N
N
N
N
HN
N
21: R=Et
31: R' = Et
29: R=Boc
30: R = H
ii
ii
iv
22: R = H
32: R' = H
H2N
N
O
N
O
O
N
O
N
F
iii
F
OR
OR'
N
N
N
H
N
H2N
N
N
v
N
N
N
N
N
NH2
N
23: R=Et
24: R = H
N
ii
H2N
33: R' = Et
34: R' = H
iv
Scheme 4. Reagents and conditions: (i) Propargyl amine, THF, Et3N, reflux 12 h; (ii)
aq 2 N NaOH, MeOH, rt, 3 h; (iii) 19, CuSO4, sodium ascorbate, BmPy2, rt, 12 h.
Scheme 6. Reagents and conditions: (i) Propargyl bromide; (ii) TFA; (iii) 1, THF,
Et3N, reflux, 12 h; (iv) aq 2 N NaOH; (v) 19, CuSO4, sodium ascorbate, BmPy2, rt, 12 h
O
N
O
close to Trp22 and Lys28, which the shorter 24 and 28 are not able
to reach.
F
OR
i
1
Docking experiments using the S. aureus topoisomerase II DNA-
gyrase better reflected the experimental results with the most ac-
tive compounds (28, 32 and 34) among the top scoring compounds,
while the non-active compounds (23 and 24) ranked lowest. Dock-
ing poses of the compound in the binding site indicated that active
compounds have the pyrimidine located in a pocket around Arg458
and Asp437. Compound 24 exhibited a conformation with the pyrim-
idine close to DNA with no protein interaction possible.
N3
N
25: R=Et
ii
26: R = H
H2N
The work presented here demonstrates the incorporation of the
pharmacophores from two active compounds with variant modes
of action into one molecule. Limitations in the synthetic tractabil-
ity of the strategy necessitated a shift from a pure chimeric design
towards a more hybrid approach by increasing the linker spacer
between the two pharmacophores. The in vitro activity relation-
ships observed perhaps reflect the structural constraints in binding
sites of the targets and the relative large size of the individual ori-
ginal starting pharmacophores, rendering the linker a crucial part
of the molecule implicated in activity. The compounds reported
here show modest in vitro activity, but they do posses more
drug-like chemical properties and show no cytotoxicity. The work
further illustrates the use of click chemistry and provides an exam-
ple of triazole containing compounds with antibacterial activity,
which opens the possibilities of using click chemistry more widely
in the discovery of new antibiotics. Compounds 28 and 34 show
potential as starting points for further optimisation studies, espe-
cially against antibiotic-resistant strains.
N
O
N
O
iii
N
F
OR
H2N
O
N
N
N
N
27: R=Et
28: R = H
ii
Scheme 5. Reagents and conditions: (i) NaN3, THF, reflux, 5 h; (ii) aq 2 N NaOH,
MeOH, rt, 3 h; (iii) 20, CuSO4, sodium ascorbate, BmPy2, rt, 12 h.
activity at all. Only compound 34 satisfied the design criteria of
showing inhibitory effect against DHFR, DNA gyrase and topoiso-
merase IV. Unfortunately this compound’s ability to inhibit the
growth of bacteria was not as effective as the mixture of ciproflox-
acin and trimethoprim, with modest to poor activity against E. coli,
SXT resistant P. aeruginosa and a glycopeptide resistant strain of
Staphylococcus epidermidis. On the other hand 28 exhibited better
ability to inhibit CIP sensitive bacteria, including sensitive E. coli,
SXT resistant K. pneumoniae, sensitive S. aureus and glycopeptide
resistant S. epidermidis. This selectivity profile for compound 28
matches the observed enzymatic assays, as 28 shows no inhibitory
effect on DHFR, but modest inhibition of DNA gyrase and topoiso-
merase IV.
The in silico docking experiments gave some insights, but the
scoring of individual compounds did not reflect the inhibitory
activity against DHFR well, even though the reference compound
TMP had the best score and overlayed with its crystal structure.
Nevertheless, the poses of the compounds in the binding site were
able to give some indications on possible reasons for their activity.
Compound 34 appears to make additional hydrophobic interac-
tions by placing the quinolone group into a hydrophobic groove,
Acknowledgments
Thanks to Matthew Hiedecker and Ben Yakimoff for their con-
tribution to the chemistry during summer research scholarships.
Financial support for this project was provided by the National
Health and Medical Research Council (NHMRC) Fellowship AF
511105.
Supplementary data
Supplementary data (key experimental details for the synthesis
of new compounds and biological analyses) associated with this
article can be found, in the online version, at doi:10.1016/
j.bmcl.2012.02.019.

T. Karoli et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2428–2433
2433
13. Summerfield, R. L.; Daigle, D. M.; Mayer, S.; Mallik, D.; Hughes, D. W.; Jackson,
S. G.; Sulek, M.; Organ, M. G.; Brown, E. D.; Junop, M. S. J. Med. Chem. 2006, 49,
6977.
14. Hecht, D.; Tran, J.; Fogel, G. B. Mol. Phylogenet. Evol. 2011, 61, 212.
15. O’Shea, R.; Moser, H. E. J. Med. Chem. 2008, 51, 2871.
16. Bax, B. D.; Chan, P. F.; Eggleston, D. S.; Fosberry, A.; Gentry, D. R.; Gorrec, F.;
Giordano, I.; Hann, M. M.; Hennessy, A.; Hibbs, M.; Huang, J.; Jones, E.; Jones, J.;
Brown, K. K.; Lewis, C. J.; May, E. W.; Saunders, M. R.; Singh, O.; Spitzfaden, C.
E.; Shen, C.; Shillings, A.; Theobald, A. J.; Wohlkonig, A.; Pearson, N. D.; Gwynn,
M. N. Nature 2010, 466, 935.
17. Verdonk, M. L.; Cole, J. C.; Hartshorn, M. J.; Murray, C. W.; Taylor, R. D. Proteins
2003, 52, 609.
18. Bouzard, D.; Di Cesare, P.; Essiz, M.; Jacquet, J. P.; Remuzon, P.; Weber, A.; Oki,
T.; Masuyoshi, M. J. Med. Chem. 1989, 32, 537.
19. (a) Chu, D. T. W. J. Heterocycl. Chem. 1985, 1033, 22; (b) Chu, D. T. W.;
Fernandes, P. B.; Claiborne, A. K.; Gracey, E. H.; Pernet, A. G. J. Med. Chem. 1986,
29, 2363; (c) Chu, D. T. W.; Fernandes, P. B.; Maleczka, R. E.; Nordeen, C. W.;
Pernet, A. G. J. Med. Chem. 1987, 30, 504.
20. H.-I. D. Shin, J.-H. D. Chang, K.-W. D. Lee. U.S. Patent US 7,576,214, 2009.
21. Huber, W. J. Am. Chem. Soc. 1943, 65, 2222.
References and notes
1. (a) Arias, C. A.; Murray, B. E. N. Eng. J. Med. 2009, 360, 439; (b) Boucher, H. W.;
Talbot, G. H.; Bradley, J. S.; Edwards, J. E., Jr.; Gilbert, D.; Rice, L. B.; Scheld, M.;
Spellberg, B.; Bartlett, J. Clin. Infect. Dis. 2009, 48, 1; (c) Cooper, M. A.; Shlaes, D.
Nature 2011, 472, 32.
2. (a) Poirel, L.; Hombrouck-Alet, C.; Freneaux, C.; Bernabeu, S.; Nordmann, P.
Lancet Infect. Dis. 2010, 10, 832; (b) Moellering, R. C. N. Eng. J. Med. 2010, 363,
2377.
3. (a) Nikaido, H. J. Antimicrob. Chemother. 1988, 22, 17; (b) Nikaido, H. Drug Resist.
Updat. 1998, 1, 93.
4. (a) WHO International Nonproprietary Names for Pharmaceutical Substances
(INN) List 104. http://www.who.int/medicines/publications/druginformation/
issues/PL-104.pdf (accessed 15.11.11).; (b) Actelion Pharmaceuticals webpage.
http://www1.actelion.com/sites/en/scientists/development-pipeline (accessed
15.11.11).
5. Brötz-Oesterhelt, H.; Brunner, N. A. Curr. Opin. Pharmacol. 2008, 8, 564.
6. Robertson, G. T.; Bonventre, E. J.; Doyle, T. B.; Du, Q.; Duncan, L.; Morris, T. W.;
Roche, E. D.; Yan, D.; Lynch, A. S. Antimicrob. Agents Chemother. 2008, 52, 2313.
7. (a) TD-1792 in Gram-positive Complicated Skin and Skin Structure Infection.
http://clinicaltrials.gov/ct2/show/NCT00442832 (accessed 29.03.11).; (b)
22. (a) Kompis, I.; Then, R.; Wick, A.; Montavon, M. Verlag Chem. 1980, 177; (b)
Roth, B.; Rauckman, B. S.; Ferone, R.; Baccanari, D. P.; Champness, J. N.; Hyde, R.
M. J. Med. Chem. 1987, 30, 348.
23. (a) Ismailov, V. M.; Mamedov, I. A.; Yusubov, N. N. Russ. J. Org. Chem. 2004, 40,
284; (b) Mizar, P.; Myrboh, B. Tetrahedron Lett. 2006, 47, 7823.
24. Otzen, T.; Wempe, E. G.; Kunz, B.; Bartels, R.; Lehwark-Yvetot, G.; Hansel, W.;
Schaper, K.-J.; Seydel, J. K. J. Med. Chem. 2004, 47, 240.
Theravance, Inc.
–
Annual Report. http://investor.theravance.com/
secfiling.cfm?filingID=1047469-11-1548 (accessed 29.03.11).
8. Yeh, P.; Tschumi, A. I.; Kishony, R. Nat. Genet. 2006, 38, 489.
9. (a) Cheng, J.; Thanassi, J. A.; Thoma, C. L.; Bradbury, B. J.; Deshpande, M.; Pucci, M.
J. Antimicrob. Agents Chemother. 2007, 51, 2445; (b) Drlica, K.; Hiasa, H.; Kerns, R.;
Malik, M.; Mustaev, A.; Zhao, X. Curr. Top. Med. Chem. (Sharjah, United Arab
Emirates) 2009, 9, 981; (c) Fisher, L. M.; Pan, X.-S. Methods Mol. Med. 2008, 142, 11.
10. Jacoby, G. A. Clin. Infect. Dis. 2005, 41, S120.
11. Laponogov, I.; Sohi, M. K.; Veselkov, D. A.; Pan, X.-S.; Sawhney, R.; Thompson,
A. W.; McAuley, K. E.; Fisher, L. M.; Sanderson, M. R. Nat. Struct. Mol. Biol. 2009,
16, 667.
25. Cui, W.; Loeppky, R. N. Tetrahedron 2001, 57, 2953.
26. Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int. Ed. 2001, 40, 2004.
27. Wyss, P. C.; Gerber, P.; Hartman, P. G.; Hubschwerlen, C.; Locher, H.; Marty, H.
P.; Stahl, M. J. Med. Chem. 2003, 46, 2304.
28. Presolski, S. I.; Hong, V.; Cho, S.-H.; Finn, M. G. J. Am. Chem. Soc. 2010, 132,
14570.
12. H. Labischinski, J. Cherian, C. Calanasan, R. Boyce. World Patent application WO
2010/025906 A2, 2009/09/02, 2010.