Synthesis and evaluation of fluoroquinolone derivatives as substrate-based inhibitors of bacterial efflux pumps

Article abstract of DOI:10.1016/j.ejmech.2008.01.042

Bacterial efflux pump systems contribute to antimicrobial resistance in pathogenic bacteria. The co-administration of bacterial efflux pump inhibitors with antibiotics is being pursued to overcome efflux-mediated resistance to antibiotics. In this study we investigated a strategy for converting broad-spectrum efflux pump substrates, fluoroquinolone antibacterial agents which are inherently recognized by most efflux transporters, into expanded spectrum efflux pump inhibitors. Employing this strategy against organisms expressing efflux pumps from the MFS, MATE and RND classes of pump systems, we report here the identification of an ofloxacin-based EPI that is a potent inhibitor of MFS (NorA) and MATE (MepA) efflux pumps in Staphylococcus aureus. The methods described here outline a process that we feel will be broadly applicable to the systematic identification of bacterial efflux pump inhibitors.

Full text of DOI:10.1016/j.ejmech.2008.01.042

Available online at www.sciencedirect.com
European Journal of Medicinal Chemistry 43 (2008) 2453e2463
http://www.elsevier.com/locate/ejmech
Original article
Synthesis and evaluation of fluoroquinolone derivatives
as substrate-based inhibitors of bacterial efflux pumps
Nadezhda German a, Peng Wei a, Glenn W. Kaatz b, Robert J. Kerns a,
*
a Division of Medicinal and Natural Products Chemistry, University of Iowa, College of Pharmacy,
115 S. Grand Avenue, Iowa City, IA 52242, USA
b Department of Internal Medicine, Division of Infectious Diseases, School of Medicine,
Wayne State University and the John D. Dingell Department of Veterans Affairs Medical Center, Detroit, MI 48201, USA
Received 8 October 2007; accepted 25 January 2008
Available online 8 February 2008
Abstract
Bacterial efflux pump systems contribute to antimicrobial resistance in pathogenic bacteria. The co-administration of bacterial efflux pump
inhibitors with antibiotics is being pursued to overcome efflux-mediated resistance to antibiotics. In this study we investigated a strategy for
converting broad-spectrum efflux pump substrates, fluoroquinolone antibacterial agents which are inherently recognized by most efflux trans-
porters, into expanded spectrum efflux pump inhibitors. Employing this strategy against organisms expressing efflux pumps from the MFS,
MATE and RND classes of pump systems, we report here the identification of an ofloxacin-based EPI that is a potent inhibitor of MFS
(NorA) and MATE (MepA) efflux pumps in Staphylococcus aureus. The methods described here outline a process that we feel will be broadly
applicable to the systematic identification of bacterial efflux pump inhibitors.
Ó 2008 Elsevier Masson SAS. All rights reserved.
Keywords: Fluoroquinolones; Efflux; Efflux inhibitor; Staphylococcus aureus; Antibiotics; NorA
1. Introduction
phenotype [2,4]. As a direct resistance mechanism efflux
pumps expel antimicrobial agents once they enter the mem-
brane or cytoplasm, affording subtherapeutic intracellular con-
centrations. These diminished intracellular concentrations can
predispose organisms to the emergence of higher level, adap-
tive resistance mechanisms [6e11], and have been indicated in
reduced postantibiotic effect [12,13]. Intrinsic antimicrobial
resistance of certain bacterial species to classes of antibiotics
is the result of constitutive expression of drug efflux pumps,
thus contributing to the innate multidrug resistance (MDR) na-
ture of some organisms [4,14,15].
Several families of proteins capable of multiple drug extru-
sion have been described, some of which require ATP hydro-
lysis for drug transport (ATP binding cassette, or ABC pumps
[primary transporters]), and others which require a proton or
sodium gradient for drug efflux (major facilitator superfamily
[MFS]-, resistance-nodulation-division [RND]-, small multi-
drug resistance [SMR]-, and multiple drug and toxin extrusion
[MATE]-type pumps [secondary transporters]) [2,4,16e18].
Survival and growth of bacteria in the presence of antibi-
otics is the phenotypic expression of drug resistance. Mecha-
nisms of such resistance include the acquisition of target
mutations, enzymatic degradation of drug, and a reduction
of intracellular drug concentration which can be achieved
through decreased membrane permeability or the expression
of membrane-based efflux pump systems [1]. The action of
such efflux pumps contributes to antimicrobial resistance in
pathogenic bacteria [2e4], and can play a role in bacterial
pathogenicity [5]. These pump systems can be selective for
one class of antimicrobial agent, such as tetracyclines, or rel-
atively nonselective and transport a variety of structurally dis-
similar compounds and confer a multidrug resistant (MDR)
* Corresponding author. Tel.: þ1 319 335 8800; fax: þ1 319 335 8766.
E-mail address: robert-kerns@uiowa.edu (R.J. Kerns).
0223-5234/$ - see front matter Ó 2008 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2008.01.042

2454
N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
Many of these pumps are capable of extruding multiple struc-
turally unrelated compounds, resulting in an MDR phenotype.
Only recently have the structures of components of these
pump systems begun to be elucidated [19], affording valuable
information for understanding regulation and substrate recog-
nition toward further elucidating the mechanisms of drug rec-
ognition and transport [20e23].
C-7 position of FQs can accommodate great diversity of struc-
ture while often retaining antibacterial activity (Fig. 1)
[39,40]. The core structures of ciprofloxacin and ofloxacin
were thus chosen for the proof-of-concept study here because
properly integrating pump inhibitors and/or structural motifs
of pump inhibitors at the C-7 position of these FQs was antic-
ipated to also reveal a number of FQ-based EPIs that retain an-
tibacterial activity.
Fluoroquinolones (FQs) as antibiotics are extensively used
against Gram-positive and Gram-negative bacteria. Gram-
positive organisms express FQ efflux systems that utilize the
energy of the proton motive force and generally provide mod-
est resistance only to the quinolone-class antibiotics with little
or no recognition of other antibacterial agents [1,3,4]. Most
notable to the studies here is FQ resistance in Staphylococcus
aureus, which may be mediated by the norA-encoded and
mepA-encoded FQ efflux pump systems [2,24]. Unlike Gram-
positive MDR efflux pump systems that, with respect to anti-
microbial agents, almost exclusively recognize FQs, the
MDR efflux systems in Gram-negative organisms that efflux
FQs also recognize and transport many other antibiotics [2].
Most notable to this study is the MexAB-OprM tripartate
efflux system of Pseudomonas aeruginosa and the AcrAB
efflux system of Escherichia coli. These pump systems recog-
nize FQs and impart, depending on levels of expression,
efflux-mediated resistance to many FQs [2].
The co-administration of bacterial efflux pump inhibitors
(EPIs) with antibiotics is being pursued by a number of aca-
demic and commercial laboratories to overcome efflux-medi-
ated resistance to antibiotics [25e32]. These efforts have
revealed that inhibiting an efflux pump system responsible
for efflux of an antibiotic can reinstate or potentiate activity
of that antibiotic [33]. Inhibition of efflux pumps has also
been shown to suppress the emergence of higher level resis-
tance mechanisms such as target mutations that occur as a con-
sequence of upregulated efflux pumps [34]. During the last
decade dozens of structurally diverse compounds that inhibit
various efflux pump systems have been identified, including
those from natural sources [35], screening of compound librar-
ies [36], and secondary evaluation of current therapeutics [37].
However, individual EPIs trend toward inhibiting a limited
number of efflux systems. This limited spectrum of pump in-
hibition, presumably due to the number and diversity of pump
types, is an obstacle to pump inhibitors finding broad clinical
utility in antimicrobial therapy. Strategies to identify single
molecules that inhibit multiple classes of efflux systems are
needed to advance this field. Co-administration of EPIs that in-
hibit two different pump types have also shown remarkable
synergy when used in combination with certain substrates in
Gram-negative organisms [38], further supporting the impor-
tance of exploring strategies to identify broad, or at least
broader, spectrum pump inhibitors.
Many different EPIs have pharmacophore structures that
can be attached to, or integrated into, FQ structure. In this
study structural features of bisaryl urea-based NorA inhibitors
and peptide-based MexAB inhibitors were selected for incor-
poration at the C-7 position of FQ core structure (Fig. 2).
These EPIs were chosen for this initial proof-of-concept study
because they are representative of EPIs that target structurally
and mechanistically dissimilar pump systems, and because
many structural variations of these EPIs have been shown to
maintain EPI activity (Fig. 2).
2. Results and discussion
2.1. Synthesis of FQeEPIs
Conjugation of bisaryl urea-based NorA EPI structure with
ciprofloxacin was achieved by attachment of bisaryl urea to
the C-7 piperazine of ciprofloxacin (Scheme 1). Phenyl isocy-
anate (6) was coupled with 4-aminobenzylalcohol (5) to pro-
vide urea 7 [41]. Transformation of the hydroxymethyl
of 7 to a chloromethyl group afforded urea 8, which
group
was employed in direct alkylation of ciprofloxacin to provide
the ciprofloxacineurea conjugate (9) in excellent overall yield.
Conjugation of bisaryl urea-based NorA inhibitor with the
core of ofloxacin was achieved by first introducing C-7 amine
functionality into the ofloxacin core structure (Scheme 2) [42].
Unlike derivative 9 that retained the C-7 piperazine and thus
would be anticipated to retain antimicrobial activity the oflox-
acin derivative prepared here (14) has the aryl urea attached
directly to a C-7 amine group, thus affording very different
structure of FQeEPI conjugate likely at the expense of losing
antimicrobial activity. Displacement of the C-7 fluorine on
ofloxacin precursor 10 with 4-methoxybenzylamine (11) af-
forded intermediate 12. Removal of the methoxybenzyl moi-
ety under acidic conditions gave the C-7-amine derivative
ofloxacin (13) in good yield. Alkylation of amine 13 with
chloromethyl urea 8 provided novel ofloxacin-based bisaryl
In this study we investigated a strategy for the possible con-
version of broad-spectrum efflux substrates, FQ antibacterial
agents that are inherently recognized by most efflux trans-
porters, into expanded spectrum EPIs by incorporating or inte-
grating structural motifs of EPIs into the FQ structure.
Previous work by our group and others has shown that the
Fig. 1. Fluoroquinolone core structures employed in this work, where C-7
groups of the parent fluoroquinolones ciprofloxacin and ofloxacin (indicated
by arrow) are replaced or modified with EPI structural motif.

N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
2455
Fig. 2. Representative structures for the two classes of EPIs employed in this study. Structural variations in the aryl rings of bisaryl urea NorA inhibitors and the
dipeptide and aryl components of MexAB inhibitors are accommodated, where numerous derivatives retain inhibitory activity against the respective efflux pump
systems.
urea conjugate 14 in excellent overall yield (Scheme 2). It is
notable that numerous standard conditions to alkylate 13
with chloromethyl-substituted urea 8 unexpectedly failed or
gave poor yields, ultimately leading to identification of
HMPA as a critical solvent to achieve high yield in this
alkylation.
Novel ciprofloxacin derivatives 19 and 24, bearing dipep-
tide components of known RND-type efflux pump inhibitors,
were synthesized using standard amino acid coupling chemis-
try to modify the C-7 piperazine of ciprofloxacin (Schemes 3
and 4). Different order of amino acid addition was undertaken
because each dipeptide in 19 and 24 has been described in
pump inhibitors. Preparation of 19 was achieved by first cou-
pling N-Boc protected NHS-activated L-phenylalanine directly
to ciprofloxacin (Scheme 3). Removal of Boc protection to
give amine 16 was followed by coupling with pentafluoro-
phenyl-actived L-ornithine derivative 17 to give protected di-
peptide conjugate 18. Removal of Fmoc and Boc groups
using standard conditions afforded the fully deprotected dipep-
tide conjugate of ciprofloxacin (19).
13 were unsuccessful, as was direct formation of aryl ureas us-
ing aryl isocyanates. While unexpected, this is not surprising
because ring electronics that renders the C-7 fluorine readily
susceptible to nucleophilic displacement concomitantly dimin-
ishes nucleophilicity of the C-7 amine. In our hands this amine
group has been refractory to all attempts at acylation-type cou-
pling chemistry employing activated acids and isocyanates.
2.2. Susceptibility testing of FQeEPIs
Gram-negative organisms utilized in susceptibility testing
included E. coli AG100 and AG102, which represent parent
and acrAB overexpressing derivative, respectively [43], and
P. aeruginosa PAO1, OCR1, and K1536, which represent par-
ent, mexAB, and mexCD overexpressing derivatives, respec-
tively [44e46]. The S. aureus strains utilized all had the SA
K1758 background, which is S. aureus NCTC 8325-4 in which
norA has been deleted and replaced with an erythromycin
resistance gene [47]. This genetic background was used to
eliminate possible confounding effects of the presence of
a chromosomal copy of norA on subsequent data. Susceptibil-
ity testing and determination of MIC values against test organ-
isms was first undertaken to determine which, if any, of the
FQeEPI conjugates retained antimicrobial activity, and to de-
fine the concentration range at which the FQeEPI conjugates
could be studied for pump inhibitory activity without killing
the test organisms in pump inhibition studies.
Dipeptide derivative of ciprofloxacin 24 was achieved by
first forming amine protected dipeptide 22 (Scheme 4). Carbo-
diimide-mediated coupling of carboxylic acid 22 with cipro-
floxacin followed by amine deprotection steps afforded the
target ciprofloxacin-dipeptide derivative 24 in excellent yield.
Attempts to similarly couple the dipeptide EPI motif (22) di-
rectly to the arylamine of C-7-amino-ofloxacin intermediate
Scheme 1. Synthesis of ciprofloxacin derivative bearing a bisaryl urea EPI at the C-7 position.

2456
N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
Scheme 2. Synthesis of ofloxacin derivative bearing a bisaryl urea EPI at the C-7 position.
All FQeEPI compounds had modest (E. coli) to minimal
this proof-of-concept study only 9 retains the C-7 piperazine
ring with a basic terminal nitrogen, which is a common requi-
site feature for most quinolone-class antibiotics. The absence
of significant activity of 9 against P. aeruginosa and E. coli
is likely a consequence of reduced membrane permeability
due to increased hydrophobic character. The elimination of in-
nate antibacterial activity of 9 against test strains was neces-
sary to assess the potential efflux inhibitory activity of this
compound in general, and to specifically determine if equiva-
lent MICs against SA K1758-P and the mepA overexpressing
strain SA K1758-A were a consequence of inhibiting MepA
or simply not being recognized by the efflux pump. To this
end, ciprofloxacin-resistant derivatives of S. aureus test strains
(P. aeruginosa) antimicrobial activity against Gram-negative
test strains; in all cases significantly less activity than the par-
ent fluoroquinolones (Table 1). Unlike the parent compounds
few differences in MIC of FQeEPIs were observed between
parent and pump overexpressing strains, thus indicating that
these compounds are not substrates, or are much poorer sub-
strates, for the AcrB, MexB, or MexD efflux pumps.
With respect to S. aureus, antimicrobial activity of ofloxa-
cin-based EPI 14 and ciprofloxacin-based FQeEPIs 19 and 24
was poor, whereas bisaryl-containing ciprofloxacin conjugate
9 retained significant activity (Table 1). This result was not un-
expected because, of the four FQeEPI derivatives tested in
Scheme 3. Synthesis of ciprofloxacin derivative 19 bearing the L-ornithineeL-phenylalanine C-7 dipeptide.

N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
2457
Scheme 4. Synthesis of ciprofloxacin derivative 24 bearing the L-phenylalanineeL-ornithine C-7 dipeptide.
were created and not only had significantly raised ciprofloxa-
cin MICs (40e160 fold), but were also resistant to more than
100 mg/ml of 9.
the MepA and NorA pump systems. Evaluation of ciprofloxa-
cin-based derivative 9, which had significant antimicrobial ac-
tivity against test strains (Table 1), employed derivatives of
these strains made resistant to 9, containing vectors pCU1
and pALC2073 for norA and mepA expression, respectively.
Compound 9 showed no inhibition of ethidium efflux against
either the NorA or MepA pumps expressed in these strains.
This suggests that 9 being refractory to MepA efflux, and
less susceptible to NorA efflux, in susceptibility testing (see
Table 1) is likely due to 9 evading the efflux pumps rather
than actively inhibiting the pumps. In contrast, ofloxacin-
based FQeEPI 14 was found to be a potent inhibitor of the
NorA and MepA efflux pump systems in S. aureus. At
0.5 mM, 14 was shown to inhibit NorA-mediated and
MepA-mediated efflux of ethidium by 73.6% and 53.4%,
2.3. Inhibition of efflux pumps by FQeEPIs
Efflux inhibition assays for P. aeruginosa utilized strain
OCR1, which overexpressed mexAB [45]. Concentrations of
putative efflux inhibitors 9, 14, 19, and 24 ranged from 0 to
20 mM, representing no more than one third the MIC for any
compound. No FQeEPI achieved significant inhibition of
ethidium efflux by MexAB, and RND-type pump system, at
any tested concentration.
With respect to S. aureus, efflux inhibition studies were fo-
cused to evaluation of bisaryl urea conjugates 9 and 14 against
Table 1
Susceptibility data for FQeEPIs against study strains
Strain
Characteristics
Minimum inhibitory concentration (in mg/ml) of:
Ciprofloxacin (1)
ꢀ0.019
Ofloxacin (2)
9
14
19
10
24
10
E. coli AG100
Susceptible control
for AG102
0.08
5
10
E. coli AG102
P. aeruginosa PAO1
acrAB overexpresser
Susceptible control
for OCR1 and K1536
mexAB overexpresser
mexCD overexpresser
Susceptible control
for SA K2361
0.04
0.16
0.16
1.25
10
50
10
50
10
50
10
50
P. aeruginosa OCR1
P. aeruginosa K1536
SA K1902
0.78
0.78
0.16
6.25
6.25
0.63
50
50
50
100
50
50
50
50
10
0.31
>100
>10
SA K2361
SA K1758-P
norA overexpresser
Susceptible control
for SA K1758-A
2.5
0.16
2.5
0.63
1.25
0.63
100
>100
>10
>10
>10
10
SA K1758-A
mepA overexpresser
0.63
1.25
0.63
>100
>10
>10

2458
N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
Table 2
Inhibition of NorA and MepA pump systems by FQeEPI 14 and the known
EPI reserpine
3. Conclusions
In this study we tested the hypothesis that broad-spectrum
efflux pump substrates might be converted into expanded spec-
trum EPIs, whereby structural motifs of two different classes
of EPIs were incorporated into fluoroquinolone antibiotic
structure. From a design standpoint there are dozens of options
for linking and/or integrating these two classes of EPIs, as well
as other classes of EPIs, with fluoroquinolone core structures.
Employing just two classes of EPIs and two fluoroquinolone
cores in this proof-of-concept study we demonstrate both suc-
cessful and unsuccessful outcomes in identifying novel EPIs
derived from modification of efflux substrate. Indeed, dipep-
tide conjugates 19 and 24 did not display EPI activity. This re-
sult is perhaps not surprising because SAR of the parent RND
pump inhibitors indicates that optimal EPI activity requires the
internal amino acid of the dipeptide to be linked directly to an
arylamine core (see Fig. 2, structure 4). Coupling of the EPI
dipeptide moiety to C-7-amino quinolone structures, such as
the C-7 arylamine analog of the ofloxacin core (Scheme 1,
13), is one potential approach toward obtaining substrate-
based FQeEPIs based on this design strategy.
Incorporating a bisaryl urea EPI motif at the 7-position of
the ofloxacin core afforded the identification of FQeEPI con-
jugate 14 as a novel, potent inhibitor of both MFS and MATE
efflux pump systems in S. aureus. Indeed, FQeEPI conjugate
14 is one of the most potent inhibitors of the NorA and MepA
pump systems reported to date (Table 2), achieving greater
than 84% inhibition of efflux at 10 mM, which is a level of in-
hibition not previously observed with other EPIs against NorA
except for compounds such as carbonyl cyanide m-chloro-
phenyl hydrazone that deplete energy by collapsing the
Strain
Concentration of
FQeEPI 14 (mM)
Reserpine
(mM)
0.5
2.0
10
30
S. aureus K2361 (norA overexpresser) 73.6% 85.6% 96.4% 81%
S. aureus K1758-A (mepA overexpresser) 53.4% 65.9% 84.1% 32%
Data are presented as percent reduction in ethidium efflux.
respectively. Efflux inhibition was dose dependent, in that as
the concentration of 14 was increased, greater inhibition was
observed (Table 2).
Having identified the novel ofloxacin-based FQeEPI con-
jugate 14 as a new, potent inhibitor of NorA (MFS-type
pump) and MepA (MATE-type pump) efflux in S. aureus,
the structural features of 14 responsible for EPI activity
were investigated. Comparing EPI activity of 14 with C-7
truncated intermediates 12 and 13 revealed that the C-7 bisaryl
urea moiety is critical for pump inhibition (Fig. 3). Similarly,
the bisaryl urea alone (27), or modified ureas containing the
aminomethyl linking group (25, 26) [48], show only approxi-
mately 20% inhibition of efflux in the low concentration range
employed here. These results demonstrate that potent inhibi-
tion of NorA by FQeEPI 14 is not simply due to the presence
of the bisaryl urea attached to the ofloxacin core, but as a con-
sequence of the structure of FQeEPI 14 as a whole. These re-
sults indicate that further design and synthesis of analogs of 14
is warranted, where changes to both the urea moiety and the
FQ core structure will likely yield even more potent inhibitors
of NorA, and perhaps other S. aureus MFS pump-mediated
efflux.
Fig. 3. Inhibition of ethidium efflux from S. aureus K2361 (norA overexpresser) by FQeEPI 14 and truncated quinolone and urea components.

N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
2459
transmembrane proton motive force. Further investigation of
the structural features of both the quinolone and urea compo-
nents required for optimal activity of this novel type of EPI is
warranted. This study demonstrates that chemical modification
of substrates for bacterial efflux pumps is an untapped ap-
proach for identifying novel, potent inhibitors of bacterial ef-
flux pump systems, and because broad-spectrum substrates are
inherently recognized and gain access to multiple pump sys-
tems, may hold the key to future discovery of expanded spec-
trum EPIs.
50 mL of dichloromethane to give 8 (207 mg, 79% yield).
1H NMR was consistent with that previously reported [41].
4.2.2. 1-Cyclopropyl-6-fluoro-4-oxo-7-{4-[4-
(3-phenylureido)-benzyl]-piperazin-1-yl}-1,4-
dihydro-quinoline-3-carboxylic acid (9)
Previously prepared N-[4-(chloromethyl)phenyl]-N0-phe-
nyl-urea 8 (104 mg, 0.4 mmol) was added to a stirred solution
of ciprofloxacin hydrochloride (1) (66 mg, 0.2 mmol) and so-
dium bicarbonate (33.6 mg, 0.4 mmol) in HMPA (5 mL). The
resulting mixture was stirred at 130 ꢁC for 24 h, at which time
no ciprofloxacin remained as determined by TLC (7:1:0.5,
ethyl acetate/methanol/water). The reaction mixture was ex-
tracted with ethyl acetate, the organic layer was collected,
dried over magnesium sulfate, and solvent removed en vacuo.
The resulting residue was separated by flash chromatography
(7:1:0.5, ethyl acetate/methanol/water). Fractions containing
9 were combined, evaporated and dried to afford 9 (105 mg,
95% yield). (9) TLC (7:1:0.5 ethyl acetate/methanol/water)
4. Experimental
4.1. General synthetic methods
All reactions requiring anhydrous conditions were carried
out in oven-dried glassware under an argon atmosphere. Dry
solvents were prepared fresh by distillation. All reactants
and reagents were of reagent grade or employed as reported.
Ciprofloxacin hydrochloride (1) and 9,10-difluoro-2,3-dihy-
dro-3-methyl-7-oxo-7H-pyrido[1,2,3-de]-1,4-benzoxazine-6-
carboxylic acid (10) were from ICN Biomedicals, Inc. and
ACROS organics, respectively. Fmoc-Orn(Boc)-OPfp (17)
was from BACHEM, Boc-Phe-OSu (15) from Advanced
ChemTech and Boc-L-ornithine (21) from Chem-Impex Inter-
national. Reactions were monitored by thin layer chromatogra-
phy (TLC) carried out on 0.25 mm silica gel plates (60F254,
EM science). Visualization of the TLC plates was by one or
more of the following methods: (1) UV light (254 nm), (2)
staining solution (1.0 g ceric ammonium sulfate, 24.0 g am-
monium molybdate, 31 mL sulfuric acid and 470 mL water)
followed by heating, or (3) staining solution (0.3% ninhydrin
in n-butanol containing 3% acetic acid) followed by heating.
Flash chromatography was carried out on silica gel 60
1
Rf ¼ 0.5. H NMR (300 MHz, (CD3)2SO): d ¼ 8.65 (s, 3H),
7.89 (dd, J ¼ 3.8, 12 Hz, 1H), 7.56 (d, J ¼ 7.2 Hz, 1H), 7.43
(m, 2H), 7.25 (m, 4H), 6.97 (t, J ¼ 7.0 Hz, 1H), 3.81 (m,
1H), 3.50 (s, 2H), 2.59 (m, 2H), 1.30 (m, 2H), 1.08 (m, 2H).
13C NMR (CD3)2SO: d ¼ 7.64, 35.94, 49.55, 49.60, 52.19,
61.54, 106.52, 106.82, 110.84, 118.18, 121.86, 128.87,
129.53, 131.20, 138.73, 139.27, 139.82, 145.35, 148.07,
152.62, 166.04. ESIMS: calcd for (M þ Hþ) 556.2315, found
556.2380.
4.2.3. 9-Fluoro-2,3-dihydro-3-methyl-10-[[(4-methoxyphenyl)
methyl]amino]-7-oxo-7H-pyrido[1,2,3-de]-1,4-benzoxazine-6-
carboxylic acid (12)
A
solution of 9,10-difluoro-2,3-dihydro-3-methyl-7-
oxo-7H-pyrido[1,2,3-de]-1,4-benzoxazine-6-carboxylic acid
(10) (98 mg, 0.35 mmol) and 4-methoxybenzylamine (11)
(0.14 mL, 1.06 mmol) in pyridine (5 mL) was stirred at
100e110 ꢁC for 48 h. After cooling to room temperature
15 mL of cold water was added and the solution adjusted to
pH ¼ 4e5 by addition of acetic acid. The resulting precipitate
was collected by filtration and dried under vacuum to give 12
(136 mg, 98% yield). TLC (dichloromethane/methanol 10:1)
1
(230e400 mesh, Kieselgel, EM Science). H NMR spectra
were recorded using QE 300 (General Electric Co.) or Mer-
cury 400 (Varian) spectrometers. Chemical shifts are given
in ppm relative to solvent (CDCl3: 7.26, DMSO-d6: 2.49, ac-
etone-d6: 2.01, D2O: 4.80). Additional methods and data for
purity analysis of products using analytical HPLC are de-
scribed in supplementary data.
1
Rf ¼ 0.37. H NMR (300 MHz, (CD3)2SO): d ¼ 15.5 (s, 1H),
8.85 (s, 1H), 7.49 (d, J ¼ 12.0 Hz, 1H), 7.23 (d, J ¼ 8.4 Hz,
2H), 6.83 (d, J ¼ 8.4 Hz, 2H), 6.64 (m, 1H), 4.86 (m, 1H),
4.55 (d, J ¼ 9 Hz, 3H), 4.25 (d, J ¼ 11.1 Hz, 1H), 3.68 (s,
3H), 1.40 (d, J ¼ 7 Hz, 3H). 13C NMR (CD3)2SO: d ¼ 18.29,
47.68, 55.24, 55.42, 68.50, 103.66, 106.46, 114.12, 124.77,
128.64, 131.05, 133.00, 133.13, 146.20, 158.58, 166.77,
176.66. ESIMS: calcd for (M þ Hþ) 399.1312, found
399.1354.
4.2. Synthesis of FQeEPIs
4.2.1. N-[4-(Chloromethyl)phenyl]-N0-phenyl-urea (8)
4-Aminobenzylalcohol (5) (113 mg, 0.92 mmol) and phe-
nyl isocyanate (6) (0.1 mL, 0.92 mmol) were stirred in chloro-
form (5 mL) at room temperature for 15 min. The resulting
white precipitate was filtered, washed with chloroform
(15 mL) and dried to give N-[4-(hydroxymethyl)phenyl]-N0-
phenyl-urea
7
(216 mg, 97% yield). Thionyl chloride
4.2.4. 10-Amino-9-fluoro-2,3-dihydro-3-methyl-7-oxo-7H-
pyrido[1,2,3-de]-1,4-benzoxazine-6-carboxylic acid (13)
Previously prepared 12 (100 mg, 0.25 mmol) was added to
a stirred 30% solution of TFA in anhydrous dichloromethane
(5 mL) at 0 ꢁC. The reaction mixture was warmed to room
temperature and stirred for 5 h, at which time no 12 was
(0.13 mL, 1.8 mmol) was added dropwise at 30 ꢁC to a suspen-
sion of urea 7 (216 mg, 0.89 mmol) in dichloromethane
(5 mL). The reaction mixture was stirred at 40 ꢁC for 2.5 h.
After completion of the evolution of gas, the reaction was
cooled to room temperature, filtered, and solid washed with

2460
N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
detected by TLC (dichloromethane/methanol, 10:1). The re-
sulting solution was treated with water (10 mL) and adjusted
to pH ¼ 9 by the addition of 1 N solution of potassium hydrox-
ide, followed by extraction with dichloromethane
(2 ꢂ 50 mL). The aqueous layer was collected, and adjusted
to pH ¼ 4 by the addition of acetic acid. The resulting precip-
itate was collected by filtration and dried under vacuum to
yield 13 (62 mg, 89% yield). TLC (dichloromethane/methanol
anhydrous dichloromethane (5 mL) and treated with p-tolue-
nethiol (17 mg, 0.14 mmol) and trifluoroacetic acid (0.5 mL).
After 2 h, the reaction mixture was concentrated under re-
duced pressure. The resulting residue was washed with hexane
(2 ꢂ 50 mL), dissolved in minimum volume of methanol and
precipitated by addition to ice-cold diethyl ether (50 mL).
The resulting precipitate was centrifuged and dried under
high vacuum to give 16 (50 mg, 75% yield). TLC (7:1:0.5
1
1
10:1) Rf ¼ 0.29. H NMR (300 MHz, (CD3)2SO): d ¼ 15.6 (s,
ethyl acetate/methanol/water) Rf ¼ 0.1. H NMR (300 MHz,
1H), 8.87 (s, 1H), 7.53 (d, J ¼ 12.0 Hz, 1H), 6.25 (s, 2H), 4.88
(m, 1H), 4.55 (dd, J ¼ 2.0, 11.5 Hz, 1H), 4.33 (dd, J ¼ 2.0,
11.5 Hz, 1H), 1.44 (d, J ¼ 7.0 Hz, 3H). 13C NMR (CD3)2SO:
d ¼ 18.35, 55.31, 68.37, 102.97, 106.37, 113.57, 124.85,
131.0, 131.68, 131.91, 145.85, 149.34, 166.94, 176.80.
ESIMS: calcd for (M þ Hþ) 279.0736, found 279.0730.
(CD3OD)): d ¼ 8.7 (s, 1H), 7.9 (d, J ¼ 13.0 Hz, 1H), 7.2 (m,
8H), 4.6 (m, 1H), 3.2 (m, 4H), 3.3 (m, 2H), 3.0 (m, 7H), 2.3
(m, 1H), 1.2 (m, 2H), 1.0 (m, 2H). 13C NMR (CD3OD):
d ¼ 7.19, 17.91, 27.28, 35.54, 41.57, 45.10, 51.27, 96.29,
107.48, 126.62, 128.27, 129.34, 129.38, 136.83, 139.33,
148.09, 162.11, 168.20, 170.96. ESIMS: calcd for (M þ Hþ);
479.2050, found; 479.2081.
4.2.5. 10-(4-(-3-Phenylureido)-benzylamino)-9-fluoro-3,7-
dihydro-3-methyl-7-oxo-2H-[1,4]oxazino[2,3,4-
4.2.7. 1-Cyclopropyl-7-{4-[2-(2,5-diamino-pentanoy
ij]quinoline-6-carboxilic acid (14)
lamino)-3-phenyl-propionyl]-piperazin-1-yl}-6-fluoro-4-
oxo-1,4-dihydro-quinoline-3-carboxylic acid (19)
Previously prepared N-[4-(chloromethyl)phenyl]-N0-phenyl-
urea (8) (104 mg, 0.4 mmol) was added to a stirred solution
To a stirred solution of (16) (50 mg, 0.10 mmol) and Fmoc-
Orn(Boc)-Pfp (17) (68.3 mg, 0.11 mmol) in DMF (5 mL) was
added isopropylamine (28 mL. 0.2 mmol). After 0.5 h the re-
sulting mixture was diluted with ethyl acetate (50 mL), trans-
ferred to a separatory funnel, and washed with a 5% solution
of NaHCO3 (2 ꢂ 100 mL) and then brine. The organic layer
was collected, dried over Na2SO4, filtered and evaporated.
The resulting residue (18) was subjected to a deprotection
step without any further purification. To a solution of 18 in
DMF (3 mL) was added piperidine (0.75 mL), and the mixture
was stirred at room temperature for 15 h. The reaction mixture
was concentrated under the reduced pressure, dissolved in di-
chloromethane (6 mL), and treated with p-toluenethiol
(24.8 mg, 0.2 mmol) followed by trifluoroacetic acid
(0.6 mL). This mixture was stirred at room temperature for
2 h, diluted with ethyl acetate (5 mL) and evaporated to dry-
ness. The resulting solid was dissolved in a minimum volume
of methanol, precipitated by addition to hexanes (150 mL),
centrifuged and decanted. The collected precipitate was dis-
solved in methanol, precipitated with cold diethyl ether, centri-
of
10-amino-9-fluoro-2,3-dihydro-3-methyl-7-oxo-7H-pyr-
ido[1,2,3-de]-1,4-benzoxazine-6-carboxylic acid (13) (55 mg,
0.2 mmol) in HMPA (5 mL). The resulting mixture was stirred
at 130 ꢁC for 24 h, at which time 13 was not detected by TLC
(7:1:0.5, ethyl acetate/methanol/water). The reaction mixture
was extracted with ethyl acetate, the organic layer was col-
lected, dried over magnesium sulfate, filtered, and solvent re-
moved en vacuo. The resulting residue was separated by flash
chromatography (7:1:0.5, ethyl acetate/methanol/water). Frac-
tions containing 14 were combined, evaporated and dried to
afford 100 mg 14 (90% yield). TLC (7:1:0.5 ethyl acetate/
methanol/water) Rf ¼ 0.5. 1H NMR (300 MHz, (CD3)2SO):
d ¼ 8.59 (s, 1H), 8.53 (s, 1H), 7.2e7.4 (m, 7H), 7.2 (t,
J ¼ 7.5 Hz, 2H), 6.81 (t, J ¼ 7.5 Hz, 1H), 5.65 (br s, 2H),
5.04 (d, J ¼ 4.5, Hz 2H), 4.56 (m, 1H), 4.3 (dd, J ¼ 2.0,
11.0 Hz, 1 H), 4.1 (dd, J ¼ 2.5 Hz, 11.0, 1H), 1.25 (d,
J ¼ 7 Hz, 3H). 13C NMR (CD3)2SO: d ¼ 18.18, 54.29, 65.44,
68.49, 97.62, 103.31 (JCF ¼ 21 Hz), 108.702, 118.45, 118.55,
118.66, 122.28, 129.17 (JCF ¼ 21 Hz), 130.41, 139.83
(JCF ¼ 8 Hz), 146.16, 152.96, 165.30. ESIMS: calcd for
(M þ Naþ); 525.1550, found; 525.1572.
1
fuged, decanted and dried to give 19 in 80% overall yield. H
NMR (300 MHz, (CD3OD)): d ¼ 8.76 (s, 1H), 7.84 (d,
J ¼ 13.0 Hz, 1H), 7.38 (m, 6H), 5.13 (t, J ¼ 7.5 Hz, 1H),
3.44e4.05 (m, 7H), 2.90e3.15 (m, 6H), 2.48 (br s, 1H),
1.85 (m, 4H), 1.43 (m, 2H), 1.23 (m, 2H). 13C NMR
(CD3OD): d ¼ 8.69, 25.47, 29.74, 37.04, 39.26, 40.17, 43.19,
46.75, 51.37, 56.08, 107.7, 108.34, 112.33 (JCF ¼ 35 Hz),
116.51, 120.52 (JCF ¼ 12 Hz), 129.05, 130.98, 135.88,
140.91, 146.79 (JCF ¼ 14 Hz), 155.36 (JCF ¼ 243 Hz),
163.43, 169.62, 170.31, 178.31 (JCF ¼ 10 Hz). ESIMS: calcd
for (M þ Hþ); 593.2843, found; 593.2877.
4.2.6. 7-[4-(2-Amino-3-phenyl-propionyl)-piperazin-1-yl]-
1-cyclopropyl-6-fluoro-4-oxo-1,4-dihydro-quinoline-3-
carboxylic acid (16)
Diisopropylamine (56 mL, 0.4 mmol) was added to a stirred
solution of Boc-L-phenylalanine N-hydroxysuccinimide ester
(15) (79 mg, 0.22 mmol) and ciprofloxacin hydrochloride (1)
(66 mg, 0.2 mmol) in DMF (3 mL). The solution was stirred
for 4 h at room temperature, at which time no 1 was detected
by TLC and a single product had formed (TLC 7:1:0.5 ethyl
acetate/methanol/water, Rf ¼ 0.71). The mixture was concen-
trated under reduced pressure, dissolved in a minimum volume
of methanol and precipitated by addition to diethyl ether. Pre-
cipitate was collected by centrifuging and dried under high
vacuum. The resulting solid (82 mg) was dissolved in
4.2.8. 5-tert-Butoxycarbonylamino-2-
(2-tert-butoxycarbonylamino-3-phenyl-propionylamino)-
pentanoic acid (22)
Boc-protected L-ornithine (21) (362 mg, 1.0 mmol), Boc-
Phe-OSu (20) (232 mg, 1.0 mmol) and diisopropylamine

N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
2461
(140 mL, 1.0 mmol) were combined in DMF (10 mL) and
stirred overnight at room temperature. After removing solvent
on high vacuum the desired dipeptide (22) was obtained in
99% yield (475 mg). 1H NMR (300 MHz, (CD3OD)):
d ¼ 7.28 (m, 5H), 4.27 (dd, J ¼ 7.0 Hz, 7.5, 1H), 4.02 (dd,
J ¼ 5.0, 7.5 Hz, 1H), 3.16 (m, 2H), 3.03 (m, 1H), 2.84 (m,
1H), 2.68 (m, 2H), 1.49e1.80 (m, 4H), 1.45 (s, 9H), 1.38 (s,
10H). 13C NMR (CD3OD): d ¼ 24.89, 25.45, 27.28, 27.37,
38.53, 53.04, 79.05, 126.37, 128.05, 128.98, 129.32, 137.16,
156.19, 156.70, 173.67.
evaporation afforded 4-tert-butoxycarbonylaminomethyl ani-
line. A solution of 4-tert-butoxycarbonylaminomethyl aniline
(198 mg, 0.89 mmol) in chloroform (5 mL) was treated with
phenyl isocyanate (81.3 mL, 0.89 mmol, for 25) or 4-chloro-
phenyl isocyanate (113.8 mL, 0.89 mmol, for 26). The result-
ing precipitates were filtered, washed with chloroform and
dried to give the Boc-protected ureas in near quantitative yield.
Treatment of each with a 10% solution of TFA in anhydrous
dichloromethane, followed by separation as reported [48], af-
forded 25 and 26.
4.2.9. 7-{4-[5-Amino-2-(2-amino-3-phenyl-propionylamino)-
pentanoyl]-piperazin-1-yl}-1-cyclopropyl-6-fluoro-4-oxo-
1,4-dihydro-quinoline-3-carboxylic acid (23)
4.3. Susceptibility testing and evaluation of efflux
inhibition
Boc-protected dipeptide L-PheeL-Orn (22) (100 mg,
0.208 mmol) and N-hydroxysuccinimide (26.3 mg, 0.23 mmol)
were dissolved in DMF (2 mL). To the resulting solution
was added N,N0-dicyclohexylcarbodiimide (51.5 mg, 0.25
mmol) and the mixture stirred for 20 h at room temperature.
The resulting mixture of activated acid was treated with diiso-
propylamine (58.3 mL, 0.42 mmol) followed by addition of
ciprofloxacin hydrochloride (1) (68.8 mg, 0.208 mmol). After
18 h, 20 mL of ice-cold water was added to the reaction mix-
ture and the product extracted with excess ethyl acetate. The
organic layer was dried over Na2SO4, filtered, and evaporated
to give Boc-protected dipeptide conjugate (23) (135 mg, 85%
yield) TLC (ethyl acetate/methanol/water, 7:1:0.5) Rf ¼ 0.67.
Cleavage of Boc groups was achieved by treatment of conju-
gate 23 (100 mg, 0.12 mmol) with 10% TFA solution in anhydrous
dichloromethane (5 mL) in the presence of p-toluenethiol
(31 mg, 0.25 mmol). After TLC indicated complete reaction,
5 mL of ethyl acetate was added and solvent was removed
in vacuum. The resulting residue was washed with hexanes
(3 ꢂ 50 mL), dissolved in minimum volume of methanol, pre-
cipitated by addition to cold diethyl ether (30 mL), centri-
fuged, decanted and dried to give 24 as a TFA salt (64 mg,
90% yield). TLC (7:1:0.5 ethyl acetate/methanol/water)
4.3.1. Antimicrobial susceptibility testing
Antimicrobial activities of test compounds were deter-
mined using a standard broth microdilution approach follow-
ing CLSI guidelines [49]. MIC data were used to determine
appropriate concentrations to employ in efflux inhibition
assays to avoid possible confounding effects of inhibition of
bacterial growth.
4.3.2. Construction of plasmids
Plasmids were constructed to allow overexpression of norA
or the gene for the newly-described MATE family multidrug
transporter MepA in the SA K1758 background. Plasmid vec-
tors included pCU1 and pALC2073 for norA and mepA ex-
pression, respectively [50,51]. Using PCR norA and its
promoter were amplified and then cloned into pCU1, produc-
ing pK364. Likewise the mepA coding region was amplified
and then cloned into pALC2073, producing pK436 and plac-
ing the gene under the control of the xyl/tetO tetracycline in-
ducible promoter of pALC2073 [24]. Plasmids were
electroporated into SA K1758 producing SA K1902 (empty
pCU1), K2361 (pK364), SA K1758-P (empty pALC2073),
and SA K1758-A (pK436). Prior to MIC testing all S. aureus
strains were grown overnight in the presence of 10 mg/ml
chloramphenicol to maintain plasmids, and tetracycline
(50 ng/ml) was included in experiments involving strains con-
taining the pALC2073 backbone to induce high-level expres-
sion of mepA from pK436.
1
Rf ¼ 0.1. H NMR (300 MHz, (CD3OD)): d ¼ 8.73 (s, 1H),
7.85 (d, J ¼ 13.0, 1H), 7.56 (m, 1H), 7.3 (m, 5H), 4.53 (m,
1H), 4.03 (m, 2H), 3.77 (m, 4H), 3.47 (m, 3H), 3.2 (m, 4H),
1.8 (m, 2H), 1.6 (m, 2H), 1.4 (m, 2H), 1.2 (m, 2H). 13C
NMR (CD3OD): d ¼ 7.17, 23.86, 27.86, 35.56, 37.45, 38.52,
41.08, 45.14, 49.88, 54.46, 106.08 (JCF ¼ 2.8 Hz), 106.77,
111.32 (JCF ¼ 26 Hz), 114.90, 118.78, 119.52 (JCF ¼ 2.4 Hz),
127.43, 128.70, 129.05, 134.26, 139.29, 145.2 (JCF ¼ 9 Hz),
147.99, 153.61 (JCF ¼ 247 Hz), 167.23, 168.01, 168.52,
176.89. ESIMS: calcd for (M þ Hþ); 593.2843, found;
593.2898.
4.3.3. Creation of ciprofloxacin-resistant derivatives
of S. aureus
To accurately assess the efflux inhibitory potential of com-
pounds that have good antimicrobial activity such as 9, it was
necessary to separate the antibacterial and possible efflux
inhibitory effects. SA K1758 was passed repeatedly on cipro-
floxacin gradient plates to obtain a mutant with a significantly
raised MIC [52]. Plasmids pCU1, pALC2073, pK364, and
pK436 were electroporated into this background to produce
versions of SA K1902, SA K2361, SA K1758-P and SA
K1758-A having ciprofloxacin MICs increased to 6.25, 100,
6.25, and 100 mg/ml, respectively. Compared to derivatives
containing empty plasmids, the further 16-fold increase in ci-
profloxacin MIC observed for both the norA and mepA
4.2.10. General methods employed for preparing
bisaryl
urea
derivatives
N-(4-chlorophenyl)-N0-[4-
(aminomethyl)phenyl]urea (25) and N-phenyl-N0-[4-
(aminomethyl)phenyl]urea (26)
Briefly, aminomethylaniline (196.4 mg, 0.9 mmol) was
treated with di-tert-butyl dicarbonate (0.09 mL, 0.8 mmol) in
a 2:1 THF/water solution (1.5 mL). After 18 h water
(20 mL) was added and ethyl acetate extraction followed by

2462
N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
[6] P.N. Markham, Antimicrob. Agents Chemother. 43 (1999) 988e989.
[7] P.N. Markham, A. Neyfakh, Antimicrob. Agents Chemother. 40 (1996)
2673e2674.
overexpressing, ciprofloxacin-resistant strains is indicative of
avid efflux pump overexpression.
[8] O. Lomovskaya, A. Lee, K. Hoshino, H. Ishida, A. Mistry, M.S. Warren,
E. Boyer, S. Chamberland, V.J. Lee, Antimicrob. Agents Chemother. 43
(1999) 1340e1346.
4.3.4. Efflux inhibition assay
The loss of ethidium cation from S. aureus cells was deter-
mined fluorometrically as previously described [53]. Experi-
ments were performed in duplicate, and the results were
expressed as mean total efflux over a 5 min time course. The
effect of inclusion of increasing concentrations of inhibitors
was determined and total efflux under these circumstances
was compared to that in the absence of inhibitors to arrive at
the percent reduction in efflux.
[9] M. Oethinger, W.V. Kern, A.S. Jellen-Ritter, L.M. McMurry, S.B. Levy,
Antimicrob. Agents Chemother. 44 (2000) 10e13.
[10] A. Nakajima, Y. Sugimoto, H. Yoneyama, T. Nakae, Microbiol. Immu-
nol. 46 (2002) 391e395.
[11] D. Hocquet, C. Vogne, F. Garch, A. Vejux, N. Gotoh, A. Lee,
O. Lomovskaya, P. Plesiat, Antimicrob. Agents Chemother. 47 (2003)
1371e1375.
[12] W. Stubbings, J. Bostock, E. Ingham, I. Chopra, Antimicrob. Agents
Chemother. 49 (2005) 1206e1208.
For P. aeruginosa organisms were grown overnight in LB
broth, diluted 25-fold in fresh, pre-warmed LB broth and sub-
jected to vigorous agitation (300 rpm) at 37 ꢁC until the opti-
cal density at 600 nm (OD600) reached 1.2e1.5. Cells were
then pelleted, washed and resuspended in ice-cold uptake
buffer (100 mM NaCl, 50 mM sodium phosphate [pH 7.2],
0.1% glycerol) to an OD600 of approximately 1.2. The organ-
ism suspension was held on ice until use. At this point, organ-
isms were diluted with ice-cold uptake buffer to an OD600 of
0.2. Organisms then were warmed to 30 ꢁC for 10 min in
a shaking water bath (300 rpm), at which point ethidium
bromide was added (final concentration, 10 mM). Aliquots
(1 ml) were removed at frequent intervals for determination
of fluorescence (lex ¼ 530, lem ¼ 600). Intracellular ethidium
achieved steady state 10 min after its addition (data not
shown). At this point various concentrations of inhibitors
were added in some experiments and incubation continued
for an additional 10 min. The difference in fluorescence at
20 min between experiments without and with inhibitors was
determined to arrive at a percent reduction in efflux. Carbonyl
cyanide m-chlorophenyl hydrazone (250 mM) added at the
10 min time point in parallel experiments served as a positive
control.
[13] J.R. Aeschlimann, L.D. Dresser, G.W. Kaatz, M.J. Rybak, Antimicrob.
Agents Chemother. 43 (1999) 335e340.
´
[14] J. Vila, S. Martı, J. Sanchez-Cespedes, J. Antimicrob. Chemother. 59
´
´
(2007) 1210e1215.
[15] S. Yang, S.R. Clayton, E.L. Zechiedrich, J. Antimicrob. Chemother. 51
(2003) 545e556.
[16] K.P. Langton, P.J.F. Henderson, R.B. Herbert, Nat. Prod. Rep. 22 (2005)
439e451.
[17] C.F. Higgins, Nature 446 (2007) 749e757.
[18] A. Kumar, H.P. Schweizer, Adv. Drug Deliv. Rev. 57 (2005) 1486e1513.
[19] For representative examples, see (a) S. Murakami, R. Nakashima,
E. Yamashita, A. Yamaguchi, Nature 419 (2002) 587e593;
(b) R.J.P. Dawson, K.P. Locker, Nature 443 (2006) 180e185.
[20] O. Lomovskaya, H.I. Zgurskaya, M. Totrov, W.J. Watkins, Nat. Rev.
Drug Discov. 6 (2007) 56e65.
[21] E.W. Yu, J.R. Aires, H. Nikaido, J. Bacteriol. 185 (2003) 5657e5664.
[22] E. Yu, G. McDermott, H. Zgurskaya, H. Nikaido, D. Koshlander Jr., Sci-
ence 300 (2003) 976e980.
[23] K.S. McKeegan, M.I. Borges-Walmsley, A.R. Walmsley, Curr. Opin.
Pharmacol. 4 (2004) 479e486.
[24] G.W. Kaatz, F. Mcaleese, S.M. Seo, Antimicrob. Agents Chemother. 49
(2005) 1857e1864.
[25] G.W. Kaatz, Curr. Opin. Investig. Drugs 6 (2005) 191e198.
[26] J.E. Kristiansen, O. Hendricks, T. Delvin, T.S. Butterworth, L. Aagaard,
J.B. Christensen, V.C. Flores, H. Keyzer, J. Antimicrob. Chemother. 59
(2007) 1271e1279.
[27] A.S. Lynch, Biochem. Pharmacol. 71 (2006) 949e956.
[28] O. Lomovskaya, K.A. Bostian, Biochem. Pharmacol. 71 (2006) 910e918.
[29] G.W. Kaatz, Expert Opin. Emerg. Drugs 7 (2002) 223e233.
[30] O. Lomovskaya, W.J. Watkins, Curr. Med. Chem. 8 (2001) 1699e1711.
[31] T.E. Renau, R.C. Lemoine, Drugs Future 26 (2001) 1171e1178.
[32] J. Barrett, Curr. Opin. Investig. Drugs 2 (2001) 212e215.
[33] Dozens of reports describe lowering of antibiotic MIC against test
organisms by employing EPIs to block antibiotic efflux. For some exam-
ples and discussion see Refs. (reviews) [25,30,31].
Acknowledgements
This work was supported in-part by grants from the Univer-
sity of Iowa Biological Sciences Funding Program (RJK) and
VA Research Funds (GWK).
[34] For a representative study, see Ref. [7].
[35] For a recent review of EPIs from natural sources, see M. Stavri,
L.J.V. Piddock, S. Gibbons, J. Antimicrob. Chemother. 59 (2007)
1247e1260.
Appendix. Supplementary data
[36] For example, see P.N. Markham, E. Westhaus, K. Klyachko,
M.E. Johnson, A. Neyfakh, Antimicrob. Agents Chemother. 43 (1999)
2404e2408.
1H NMR spectra and purity analysis for FQeEPI conju-
gates evaluated in microbiological studies. Supplementary
data associated with this article can be found in the online ver-
sion, at doi:10.1016/j.ejmech.2008.01.042.
[37] For example, see P. Wei, G.W. Kaatz, R.J. Kerns, Bioorg. Med. Chem.
Lett. 14 (2004) 3093e3097.
[38] G. Tegos, F.R. Stermitz, O. Lomovskaya, K. Lewis, Antimicrob. Agents
Chemother. 46 (2002) 3133e3141.
References
[39] R.J. Kerns, M.J. Rybak, G.W. Kaatz, F. Vaka, R. Cha, R.G. Grucz,
V.U. Diwadkar, Bioorg. Med. Chem. Lett. 13 (2003) 2109e2112.
[40] R.J. Kerns, M.J. Rybak, G.W. Kaatz, F. Vaka, R. Cha, R.G. Grucz,
V.U. Diwadkar, T.D. Ward, Bioorg. Med. Chem. Lett. 13 (2003)
1745e1749.
[1] M.N. Alekshun, S.B. Levy, Cell 128 (2007) 1037e1050.
[2] X.-Z. Li, H. Nikaido, Drugs 64 (2004) 159e204.
[3] L.J. Piddock, Clin. Microbiol. Rev. 19 (2006) 382e402.
[4] K. Poole, J. Antimicrob. Chemother. 56 (2005) 20e51.
[5] L.J. Piddock, Nat. Rev. Microbiol. 4 (2006) 629e636.
[41] Synthesis of urea 8 starting from phenylisocyanate (6) and 4-amino ben-
zyl alcohol (5) was achieved with minor modifications of methods

N. German et al. / European Journal of Medicinal Chemistry 43 (2008) 2453e2463
2463
previously reported, see V. Wehner, H. Blum, M. Kurz, H.U. Stilz, Syn-
thesis 14 (2002) 2023e2036.
[48] Ureas 25e27 were prepared using previously reported methods, see (a)
Y.-S. Liu, C. Zhao, D.E. Bergbreiter, D. Romo, J. Org. Chem. 63
(1998) 3471e3473;
[42] Synthesis of the C-7 amine derivative of ofloxacin (13) was based on
similar methods previously employed to prepare putative metabolites
of grepafloxacin using C-8 bromo quinolone derivatives, see S. Morita,
K. Otsubo, J. Matsubara, T. Ohtani, Y. Kowano, K. Ohmori,
K. Ohguro, M. Uchida, Chem. Pharm. Bull. 43 (1995) 2246e2252.
[43] A.M. George, S.B. Levy, J. Bacteriol. 155 (1983) 531e540.
[44] B.W. Holloway, V. Krishnapillai, A.F. Morgan, Microbiol. Rev. 43 (1979)
73e102.
(b) M.D. Wendt, T.W. Rockway, A. Geyer, W. McClellan, M. Weitzberg,
X. Zhao, R. Mantei, V.L. Nienaber, K. Stewart, V. Klinghofer,
V.L. Giranda, J. Med. Chem. 47 (2004) 303e324.
[49] Clinical and Laboratory Standards Institute, Approved Standard M7-A7.
Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria
That Grow Aerobically, seventh ed. Clinical and Laboratory Standards
Institute, Wayne, PA, 2006.
[45] N. Masuda, S. Ohya, Antimicrob. Agents Chemother. 36 (1992) 1847e
1851.
[50] J. Augustin, R. Rosenstein, B. Wieland, U. Schneider, N. Schnell,
G. Engelke, K.D. Entian, F. Gotz, Eur. J. Biochem. 204 (1992) 1149e1154.
[51] B.T. Bateman, N.P. Donegan, T.M. Jarry, M. Palma, A.L. Cheung, Infect.
Immun. 69 (2001) 7851e7857.
[46] T.R. De Kievit, M.D. Parkins, R.J. Gillis, R. Srikumar, H. Ceri, K. Poole,
B.H. Iglewski, D.G. Storey, Antimicrob. Agents Chemother. 45 (2001)
1761e1770.
[52] V. Bryson, W. Szybalski, Science 115 (1952) 45e51.
[53] G.W. Kaatz, S.M. Seo, L. O’Brien, M. Wahiduzzaman, T.J. Foster, Anti-
microb. Agents Chemother. 44 (2000) 1404e1406.
[47] C.T.D. Price, G.W. Kaatz, J.E. Gustafson, Int. J. Antimicrob. Agents 20
(2002) 206e213.