MexAB-OprM-specific efflux pump inhibitors in Pseudomonas aeruginosa. Part 1: Discovery and early strategies for lead optimization

Article abstract of DOI:10.1016/j.bmcl.2003.07.024

The identification of a series of compounds that specifically inhibit efflux by the MexAB-OprM pump system in Pseudomonas aeruginosa is described. Synthesis and in vitro structure-activity relationships (SARs) are outlined. Early leads lacked activity in

Full text of DOI:10.1016/j.bmcl.2003.07.024

Bioorganic & Medicinal Chemistry Letters 13 (2003) 4201–4204
MexAB-OprM-Specific Efflux Pump Inhibitors in Pseudomonas
aeruginosa. Part 1: Discoveryand EarlyStrategies for
Lead Optimization
Kiyoshi Nakayama,a,* Yohei Ishida,a Masami Ohtsuka,a Haruko Kawato,a
Ken-ichi Yoshida,a Yoshihiro Yokomizo,a Shigeru Hosono,a Toshiharu Ohta,a
Kazuki Hoshino,b Hiroko Ishida,b Kumi Yoshida,b Thomas E. Renau,c Roger Leger,c
Jason Z. Zhang,c Ving J. Leec and William J. Watkinsc
aMedicinal Chemistry Research Laboratory, Daiichi Pharmaceutical Co., Ltd., 1-16-13, Kitakasai, Edogawa, Tokyo 134-8630, Japan
bNew Product Research Laboratories I, Daiichi Pharmaceutical Co., Ltd., 1-16-13, Kitakasai, Edogawa, Tokyo 134-8630, Japan
cEssential Therapeutics, Inc., 850 Maude Avenue, Mountain View, CA 94043, USA
Received 16 June 2003; accepted 29 July 2003
Abstract—The identification of a series of compounds that specifically inhibit efflux by the MexAB-OprM pump system in
Pseudomonas aeruginosa is described. Synthesis and in vitro structure–activity relationships (SARs) are outlined. Early leads lacked
activity in animal models, and efforts to improve solubility and reduce serum protein binding by the introduction of polar groups
are discussed.
# 2003 Elsevier Ltd. All rights reserved.
The demand for the discovery of new antimicrobial
drugs is continually fuelled by the emergence of resis-
tance. The more an antibiotic is used for the treatment
of infections, the greater the probability that resistance
to it will develop. Even antibacterial drugs with novel
mechanisms of action suffer the same fate, and a fun-
damental solution is required to break the vicious cir-
cle.1À5 Active efflux of antimicrobial agents is
increasingly being recognized as a major cause of anti-
microbial resistance among a diverse set of antibacterial
agents including b-lactams, macrolides, tetracyclines
and fluoroquinolones.6À10 A particularly problematic
pathogen in the clinical setting is Pseudomonas aerugi-
nosa, an opportunistic Gram-negative pathogen char-
acterized by intrinsic resistance to a wide variety of
antimicrobial agents.11À13 This property has been
attributed both to the impermeability of the outer
membrane as well as to the activity of several efflux
systems of the Resistance-Nodulation-Division (RND)
superfamily, four of which have been identified to date:
MexAB-OprM, MexCD-OprJ, MexEF-OprN, and
MexXY-OprM.14À17
We have previously disclosed the discovery and optimi-
zation of compounds that inhibit all RND pumps in P.
aeruginosa and thereby potentiate the activity of the
fluoroquinolone Levofloxacin (LVFX).18À20 Herein, we
describe the first example of a MexAB-OprM-specific
efflux pump inhibitor (EPI). Such an agent might be
used for the potentiation of anti-pseudomonas b-lac-
tams that are substrates of this pump (e.g., aztreonam
(AZT), piperacillin, meropenem), as well as quinolones
(Fig. 1). We also summarize in vitro SAR and initial
attempts to optimize properties for use in vivo.
Compound 1 (Fig. 2) was identified in a high-through-
put screen for LVFX potentiation using a MexAB-
OprM over-expressing strain21,22 of P. aeruginosa
(PAM1032).23
Inhibitors of RND pumps in P. aeruginosa are not
expected to display intrinsic antibacterial activity, as
functioning pumps are not essential for survival of the
organism in vitro. Neither 1 nor any analogues reported
*Corresponding author. Tel.: +81-3-5696-9079; fax: +81-3-5696-
8772; e-mail: nakaygoi@daiichipharm.co.jp
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.07.024

4202
K. Nakayama et al. / Bioorg. Med. Chem. Lett. 13 (2003) 4201–4204
Figure 1. Levofloxacin and anti-pseudomonas b-lactams.
Scheme 1. Synthesis of variants of Parts A and C.
Figure 2. Structure of compound 1.
herein were antibacterials (MIC >320 mg/mL). When
combined with antibiotics, the potentiation activity was
conveniently expressed as the lowest concentration of
EPI achieving an 8-fold reduction in antibiotic MIC
(Minimum Potentiation Concentration, or MPC8). By
this measure, the activity of 1 as a potentiator of LVFX
against PAM1032 and PAM1723 (a strain in which
MexAB-OprM is overexpressed and MexCD-OprJ and
MexEF-OprN are genetically disrupted)21,22 was
impressive (MPC8 ꢀ0.63 mg/mL). However, it was
devoid of potentiation activity against MexCD-OprJ,
MexEF-OprN and MexXY-OprM over-expressing
strains, suggesting that the action of inhibition is spe-
cific for the MexAB-OprM pump. Direct inhibition of a
fluorescent substrate of this pump was also measured
spectrophotometrically.
Scheme 2. Synthesis of variants of Part B.
While the microbiological profile of 1 was attractive, the
physicochemical properties were problematic. The aqu-
eous solubility was poor (2.85 mg/mL in pH 6.8 buffer
solution), the lipophilicity was high (log D>4), and the
affinity for serum albumin was also high (>98% bound,
Kd 0.28 mM). In addition, no significant potentiation of
LVFX was seen in a murine sepsis model using a
MexAB-OprM overexpressing strain of P. aeruginosa
(data not shown). We wished to reduce the serum pro-
tein binding so as to allow an assessment of free drug
clearance. In order to accomplish this we generated two
strategies, one based upon polar group substitution and
the other looking for alternative scaffolds. In this paper,
we report the scope and limitation of the first strategy;
the results of the search for alternative scaffolds will be
the subject of subsequent publications.
the outcome of the incorporation of polar groups into
each was examined. The syntheses of novel analogues
are depicted in Schemes 1–3, and final structures are
displayed in Figure 3. Thiazoles were synthesized in a
conventional manner by the condensation of haloke-
tones with the requisite thioamide.24À26 Variants of the
homophthalic acid moiety in 1 were generated by direct
acylation of the key intermediate aniline with a variety
of carboxylic acid derivatives.27À30
Analogues containing pyridyl surrogates for the central
phenyl moiety (part B) were synthesized by a Knoeve-
nagel-type condensation between a 2-methylthiazole
and
a
2-aminopyridine-4-carboxaldehyde deriva-
tive,25,26 or by a Horner–Emmons reaction between the
pyridine-aldehyde and a phosphonate (Scheme 2).24
Ether tethers were conveniently constructed via alkyl-
ation of pyridones. The ethylene tethers resulted from
For convenience, the structure of 1 was divided into
three parts (A, B and C) as exemplified in Figure 2, and

K. Nakayama et al. / Bioorg. Med. Chem. Lett. 13 (2003) 4201–4204
4203
the coupling of the dianion of 2-(t-butylox-
ycarbonylamino)-4-methylpyridine with the same thia-
zole-containing electrophile (Scheme 3).31
in order to evaluate the effects of protein binding.35
Entries 1–5 in Table 1 describe the modifications to the
thiazole moiety A. Introduction of a carboxamide (as in
2) reduced activity dramatically, although the influence
of serum was also diminished; other more polar variants
abolished activity completely. Replacement of the iso-
propyl substituent with pentafluoroethyl (3) gave some-
what reduced activity. t-Butyl or phenyl variants (4,5)
could be accommodated with no deleterious effect on
potency, but also no apparent impact on the effect of
protein binding.
New analogues were tested for the potentiation of
LVFX and AZT using PAM1723.32 Assays were con-
ducted in the presence and absence of 10% horse serum
Next, the modification of styrene portion (B) was
examined. Three pyridyl variants gave very different
results, with 8 dramatically less active than the other
two. The 3,5-disubstituted pyridine 7 gave good activity
that was not greatly affected by the addition of serum.
The LVFX potentiation of 6 was excellent, but its
extreme insolubility confounded attempts to generate
meaningful data with AZT. To address this problem,
alternatives to the vinyl moiety were then evaluated.
Molecules bearing methyloxy (9) and ethylene (10)
tethers displayed favorable activity profiles similar to 7.
Scheme 3. Synthesis of ether and ethylene tethers.
Figure 3. Structure of 1 and novel analogues.
Table 1. Potentiation activity of 1 and novel analoguesa
MPC8 (AZT) (mg/mL)
Ratio MPC8 (AZT)
with/without serum)
MPC8 (LVFX);
(mg/mL)
Without serum
With 10% horse serum
1
2
3
4
5
6
7
8
ꢀ0.63
20
10
40
80
20
40
10
10
80
5
10
40
20
20
160
40
>16
2
ꢀ0.63
80
2.5
0.63
1.25
0.31
2.5
>80
2.5
2.5
40
20
2.5
ꢀ0.63
1.25
>10b
5
20
2.5
5
20
5
2.5
20
32
>32
32
—
2
4
2
2
2
4
8
8
16
9
10
11
12
13
14
15
1.25
10
2.5
2.5
aAll compounds lacked intrinsic antibacterial activity.
bThe apparent lack of activity is attributable to precipitation of the compound.

4204
K. Nakayama et al. / Bioorg. Med. Chem. Lett. 13 (2003) 4201–4204
Based upon the known structure of serum albumin and
its interaction with acidic molecules,33,34 we envisaged
that the introduction of bulky or hydrophilic groups
adjacent to the carboxylic acid moiety might reduce
protein binding. Variants in part C were designed with
this notion in mind. One enantiomer (12) of the dioxo-
lane-containing pair showed more activity than the
other (11), and the serum effect was indeed reduced. The
reasonable potency of the pyridyl analogue (13) was
compromised by the addition of serum; similarly, the
incorporation of hydroxyl and methoxy groups in this
region gave no improvement in serum effect (14, 15),
although the activity of the alcohol was encouraging.
14. Poole, K.; Krebes, K.; McNally, C.; Neshat, S. J. Bacter-
iol. 1993, 175, 7363.
15. Poole, K.; Gotoh, N.; Tsujimoto, H.; Zhao, Q.; Wada, A.;
Yamasaki, T.; Neshat, S.; Yamagishi, J.; Li, X. Z. Nishino,
Mol. Microbiol. 1996, 21, 713.
16. Koehler, T.; Michea-Hamzehpour, M.; Henze, U.; Gotoh,
N.; Curty, L. K.; Pechere, J. C. Mol. Microbiol. 1997, 23, 345.
17. Mine, T.; Morita, Y.; Kataoka, A.; Mizushima, T.; Tsu-
chiya, T. Antimicrob. Agents Chemother. 1999, 43, 415.
18. Renau, T. E.; Leger, R.; Flamme, E. M.; Sangalang, J.;
She, M. W.; Yen, R.; Gannon, C. L.; Grith, D.; Chamberland,
S.; Lomovskaya, O.; Hecker, S. J.; Lee, V. J.; Ohta, T.;
Nakayama, K. J. Med. Chem. 1999, 42, 4928.
19. Renau, T. E.; Leger, R.; Flamme, E. M.; She, M. W.;
Gannon, C. L.; Mathias, K. M.; Lomovskaya, O.; Chamber-
land, S.; Lee, V. J.; Ohta, T.; Nakayama, K.; Ishiada, Y.
Bioorg. Med. Chem. Lett. 2001, 11, 663.
20. Renau, T. E.; Leger, R.; Yen, R.; She, M. W.; Flamme,
E. M.; Sangalang, J.; Gannon, C. L.; Chamberland, S.;
Lomovskaya, O.; Lee, V. J. Bioorg. Med. Chem. Lett. 2002,
12, 763.
21. Lomovskaya, O.; Warren, M. S.; Lee, A.; Galazzo, J.;
Fronko, R.; Lee, M.; Blais, J.; Cho, D.; Chamberland, S.;
Renau, T.; Leger, R.; Hecker, S.; Watkins, W.; Hoshino, K.;
Ishida, H.; Lee, V. Antimicrob. Agents Chemother. 2001, 45,
105.
22. Lomovskaya, O.; Lee, A.; Hoshino, K.; Ishida, H.; Mis-
try, A.; Warren, M. S.; Boyer, E.; Chamberland, S.; Lee, V. J.
Antimicrob. Agents Chemother. 1999, 43, 1340.
23. Hoshino, K.; Ishida, H.; Okumura, R.; Nakayama, K.;
Hayakawa, I; Sato, K. In 42nd Interscience Conference on
Antimicrobial Agents and Chemotherapy, San Diego, CA,
September 2002; Abstract F-728. Further details of the
microbiological and biochemical characterization and poten-
tial clinical utility of MexAB-OprM specific inhibitors will be
reported elsewhere in due course.
Compounds with reasonable activity and little influence
of serum in vitro were evaluated as potentiators of
LVFX in murine neutropenic thigh or sepsis models.
Marginal activity was seen for only one compound (10),
which exhibited a slight effect in the sepsis model (data
not shown). The protein binding of this compound, as
measured by conventional ultra-filtration techniques,
was 96.7% (Kd 12.3 mM), whereas for all the other ana-
logues tested it was >98%. We concluded that lipo-
philic acids such as these exhibited extremely high
affinities for serum albumin and that more radical
structure manipulation would be required to reduce the
binding significantly.
In summary, we synthesized several derivatives of 1, the
first known MexAB-OprM specific inhibitor of efflux in
P. aeruginosa, in an attempt to lower serum protein
binding. Pyridine variants with ether or ethylene tethers
replacing the styrene moiety exhibited the most promis-
ing profile, but were still not efficacious in vivo. Incor-
poration of more polar groups simply abolished the
activity. We therefore concluded that other strategies
would be required to overcome the problem, and a suc-
cessful alternative approach is reported in the following
paper.
24. Maehr, H.; Yang, R. Tetrahedron Lett. 1996, 37, 5445.
25. Hayashi, Y; Oguri, T.; Shinoda, M.; Tsutsui, M.; Taka-
hashi, K.; Miida, H.; Eur. Pat. 0219436.
26. Francis, G. F.; Raymond, J. V.; Joseph W. Z.; Eur. Pat.
0355353.
27. Tamura, Y.; Sasho, M.; Akai, S.; Kishimoto, H.; Seki-
hachi, J.; Kita, Y. Chem. Pharm. Bull. 1987, 35, 1405.
28. Tamura, Y.; Sasho, M.; Akai, S.; Kishimoto, H.; Seki-
hachi, J.; Kita, Y. Tetrahedron Lett. 1986, 27, 195.
29. Zahra, J.; Hennig, L.; Findeisen, M.; Welzel, P.; Mueller,
D.; Sheldrick, W. S. Tetrahedron Lett. 2001, 42, 2969.
30. Deady, L. W.; Rodemann, T. J. Heterocycl. Chem. 2001,
38, 1185.
References and Notes
1. Gold, H. S.; Moellering, R. C., Jr. New Engl. J. Med. 1996,
335, 1445.
2. Chu, D. T. W.; Plattner, J. J. J. Med. Chem. 1996, 39, 3853.
3. Domagala, J. M.; Sanchez, J. P. In Annual Reports in
Medicinal Chemistry; Bristol, J. A., Ed.; Academic: New
York, 1997; p 111.
31. Ihle, N. C.; Krause, A. E. J. Org. Chem. 1996, 61, 4810.
32. MICs were determined using Mueller–Hinton broth in 96-
well microtiter plates, following the broth microdilution
methodology outlined in NCCLS Document M7-A4, Vol. 17,
No. 2. The starting inoculum was 106 CFU/mL. MICs were
determined after 18-h incubation at 37 ꢁC. For AZT, MICs
were determined by visual observation of growth; for LVFX,
the determination was made at 70% inhibition of growth, as
measured by OD 595 nm.
33. Frostell-Karlsson, A.; Remaeus, A.; Roos, H.; Andersson,
K.; Borg, P.; Hamalainen, M.; Karlsson, R. J. Med. Chem.
2000, 43, 1986.
34. Colmenarejo, G.; Alvarez-Pedraglio, A.; Lavandera, J.-L.
J. Med. Chem. 2001, 44, 4370.
4. Perl, T. M. Am. J. Med. 1999, 106, 26 S.
5. Chopra, I. Curr. Opin. Microbiol 1998, 1, 495.
6. Lawrence, L. E.; Barrett, J. F. Exp. Opin. Invest. Drugs
1998, 7, 199.
7. Marshall, N. J.; Piddock, L. J. Microbiologia 1997, 13, 285.
8. Nikaido, H. Clin. Infect. Dis. 1998, 27, S32.
9. Lee, V. J.; Lomovskaya, O. CLEAR 1998, 1, 39.
10. Levy, S. B. Antimicrob. Agents Chemother. 1992, 36, 695.
11. Li, X. Z.; Livermore, D. M.; Nikaido, H. Antimicrob.
Agents Chemother. 1994, 38, 1732.
12. Li, X. Z.; Ma, D.; Livermore, D. M.; Nikaido, H. Anti-
microb. Agents Chemother. 1994, 38, 1742.
13. Hancock, R. E. Clin. Infect. Dis. 1998, 27 (Suppl. 1), S93.
35. The effects of the addition of 0.125% human serum albu-
min upon activity in vitro were comparable to those with 10%
horse serum for key compounds.