MexAB-OprM specific efflux pump inhibitors in Pseudomonas aeruginosa. Part 4: Addressing the problem of poor stability due to photoisomerization of an acrylic acid moiety

Article abstract of DOI:10.1016/j.bmcl.2004.03.007

Exchange of the ethylene tether in a series of pyridopyrimidine-based MexAB-OprM specific efflux pump inhibitors to an amide bond stabilized the olefin of the acrylic acid moiety, preventing facile photoisomerization to the Z-isomer. Furthermore, the activity was drastically improved in the amide tether variants, providing extremely potent acrylic acid and vinyl tetrazole analogues.

Full text of DOI:10.1016/j.bmcl.2004.03.007

Bioorganic & Medicinal Chemistry Letters 14 (2004) 2493–2497
MexAB-OprM specific efflux pump inhibitors in Pseudomonas
aeruginosa. Part 4: Addressing the problem of poor stability due to
photoisomerization of an acrylic acid moiety
Kiyoshi Nakayama,a,* Noriko Kuru,a Masami Ohtsuka,a Yoshihiro Yokomizo,a
Atsunobu Sakamoto,a Haruko Kawato,a Ken-ichi Yoshida,a Toshiharu Ohta,a
Kazuki Hoshino,a Katsuya Akimoto,a Junko Itoh,a Hiroko Ishida,a Aesop Cho,b
Monica H. Palme,b Jason Z. Zhang,b Ving J. Leeb and William J. Watkinsb
aDaiichi Pharmaceutical Co., Ltd, 1-16-13, Kitakasai, Edogawa, Tokyo 134-8630, Japan
bEssential Therapeutics, Inc., 850 Maude Avenue, Mountain View, CA 94043, USA
Received 16 February 2004; revised 1 March 2004; accepted 1 March 2004
Abstract—Exchange of the ethylene tether in a series of pyridopyrimidine-based MexAB-OprM specific efflux pump inhibitors to an
amide bond stabilized the olefin of the acrylic acid moiety, preventing facile photoisomerization to the Z-isomer. Furthermore, the
activity was drastically improved in the amide tether variants, providing extremely potent acrylic acid and vinyl tetrazole ana-
logues.
Ó 2004 Elsevier Ltd. All rights reserved.
The emergence of bacterial resistance continues to
challenge effective chemotherapy, especially among the
immune compromised. Measures to tackle the problem
comprise the judicious use of antimicrobial agents
together with discovery of new drugs, including those
targeting the root cause of the resistance in question.
Peudomonas aeruginosa is a problematic opportunistic
pathogen that shows intrinsic resistance to a wide vari-
ety of antimicrobial agents (including quinolones,
b-lactams, aminoglycosides, macrolides and tetracy-
clines). This is largely attributable to the expression of
several efflux pumps, four of which have been identified
to date: MexAB-OprM, MexCD-OprJ, MexEF-OprN
and MexXY-OprM.1–4 We recently reported the dis-
covery and optimization5–7 of agents that are specific for
the inhibition of MexAB-OprM. Our initial SAR studies
included scaffolding methodology and pharmacophore-
model-based design to improve both efficacy and
physicochemical properties, furnishing pyridopyrim-
idine-based leads typified by 1. This compound pos-
sessed high in vitro potency coupled with relatively low
protein binding (for this series), and exhibited encour-
aging efficacy in vivo potency for the potentiation of the
activity of Levofloxacin (LVFX). However, the series
showed poor stability due to extremely facile photoiso-
merization of the olefin of the acrylic acid moiety, as
shown Figure 1. Surprisingly, exposure of solutions of
E-isomers such as 1 to ambient light for as little as 2 h
was sufficient to generate predominantly the Z-isomer
(2), which was found to be devoid of activity.
OH
OH
N
ambient light
in solution,
N
1
N
9
2
8
N
N
N
S
S
N
5
N
7
3
4
6
O
OH
O
HO
i.e. CHCl3, MeOH
1
O
O
2
Ratio: ca. 1:9 (after 2 hr exposure to ambient light in CDCl3)
Keywords: Efflux pump inhibitor; Pseudomonas aeruginosa; Photoiso-
merization.
Figure 1. Structure of the pyridopyrimidine-based MexAB-OprM
specific efflux pump inhibitor (1), and its isomerization to the Z-isomer
(2).
* Corresponding author. Tel.: +81-3-5696-9079; fax: +81-3-5696-8344;
e-mail: nakaygoi@daiichipharm.co.jp
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.03.007

2494
K. Nakayama et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2493–2497
Herein, we report the results of structure–stability rela-
tionship studies to address the problem of photoiso-
merization by variation of the olefin, and through the
incorporation of electron-withdrawing groups on the
pyridopyrimidine scaffold so as to ameliorate the effect
of the electron-rich substituent ortho to the acrylic acid
moiety, together with the potency of the derivatives.
Other linkers
O
O
N
H
amide tether
R1:
ether tether
R1
N
OH
R2:
N
N
R2
O
O
S
N
N
N
We envisioned that incorporation of small alkyl, alkoxy,
or fluorine substituents might alter the electronic char-
acteristics of the unsaturated olefin, leading to analogues
with greater stability towards photoisomerization.
Another strategy to circumvent the problem was based
on the hypothesis that the isomerization was facilitated
by the electron-donating 2-substituent in conjugation
with the acrylic acid moiety. Incorporation of further
substituents on the pyridopyrimidine scaffold might
counterbalance this effect. Alternatively, exchanging the
ethylene tether connecting the thiazole to the pyrido-
pyrimidine nucleus to an amide surrogate would render
the scaffold less electron rich. Conversely, an analogue
with an ether tether (29) would donate electron density
to the scaffold (Fig. 2).
O
OH
O
NH2
EW group
( F- )
O
Me-
MeO-
F-
Figure 2. Outline of strategies to prevent photoisomerization of the
olefin in the acrylic acid motif.
thiazole motif (R1) did not affect the stability (data not
shown); its selection was driven by the availability of
intermediates from earlier syntheses.5
The routes for incorporation of methyl, methoxy, or
fluorine substituents on the double bond of the acrylic
acid are shown in Schemes 1 and 2. In general, Vilsmeier
formylation of the parent pyrimidopyrimidine, followed
by Wittig reactions provided the a-substituted acrylic
Schemes 1–6 illustrate the routes by which these
derivatives were prepared. The alkyl substituent on the
R1
R1
1) TsCl, DMAP, THF, DMF
N
N
O
or
N
O
OH
N
O
N
S
S
(PhO)2P(O)Cl, iPr2NEt
N
N
2) Amine, DMF
3a (R1=cyc-Bu-)
3b (R1=iPr-)
R1
4a (R1=cyc-Bu-) 68%
4b (R1=iPr-) 75%
N
O
1) POCl3, DMF
2) Ph3PCHR2CO
THF
N
O
N
S
2 Me
N
OH
3) LiOH, H O,
EtOH
2
R2
O
5 (R1=cyc-Bu-, R2=-Me, 57% for 3 steps)
6 (R1=iPr-, R2=-OMe, 10% for 3 steps)
Scheme 1. Synthesis of methyl and methoxy a-substituted acrylic acid analogues.
N
N
1) POCl3, DMF
N
O
OH
N
O
OH
S
S
N
2) (EtO)2POCHFCO2Et,
n-BuLi, THF, -78 °C
N
OEt
3b
to rt
F
7
69% for 2 steps
O
N
O
1) TsCl, Et3N
2) morpholine
1) I2, CH2Cl2
N
O
N
S
2) LiOH, H2O, EtOH
68% for 4 steps
N
OEt
8
F
O
N
S
O
N
N
N
O
OH
F
9
O
Scheme 2. Synthesis of a-fluorinated acrylic acid analogue.

K. Nakayama et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2493–2497
2495
Me
Me
N
Me
N+
1) Ac2O
2) Pd-C, H2
N
Toluene,
reflux
F
O2N
BF4-
3) NaNO2, HBF
N
NHAc
NHAc
N
10
NH2
4
47%
12
11
83%
Me
1) n-BuLi, THF
2)
F
1) HCl, EtOH
2) Boc2O, t-BuOH
N
NHBoc
S
N
13
NHBoc
33%
53%
N
14
N
F
S
Br
N
TFA
92%
N
S
CH2(CO2TCP)2
xylene, reflux
84%
NH2
S
N
O
OH
N
F
N
15
16
O
F
O
H
H
1) TsCl, DMAP, THF,
DMF
2) 3-hydroxypiperidine
1) Ph3P=CHCO2tBu, THF,
reflux (84%)
2) NaOMe (65%)
N
N
O
N
O
S
3) POCl3 , DMF
69%
N
17
F
3) CCl3CONCO, AcOEt
4) HCO2Na, MeOH, H2O
OCONH2
5) HCl-dioxane
N
S
(57% for 3 steps)
N
O
N
Cl
N
F
TCP:
OH
O
Cl
18
Cl
Scheme 3. Incorporation of fluorine onto the pyridopyrimidine scaffold.
N
O
NHAc
HOOC
NH2
EDC, HOBt, DMAP
CH2Cl2
N
NHAc
S
N
H
N
S
N
19
20
21
90%
1) HCl, EtOH
2) CH2 (CO2TCP)2
xylene
1) (PhO)2P(O)Cl,iPr2NEt
2) (R)-3-hydroxypiperidine
N
O
N
O
OH
S
N
H
65%
62%
N
22
OH
1) POCl3, DMF
2) Ph3PCHCO2 Bu, THF
3) NaOMe, MeOH
N
O
t
N
N
S
N
H
N
47% for 3 steps
23
O
OH
1) Cl3CONCO, AcOEt
N
O
2) HCO2Na, MeOH, H2O
N
N
S
N
98% for 2 steps
H
N
O
OtBu
24
O
O
O
O
O
NH2
O
NH2
N
O
N
O
HCl-dioxane
quant.
N
O
N
N
N
S
N
H
S
N
H
N
N
OH
O
OtBu
25
26
O
Scheme 4. Synthesis of an amide tether variant.

2496
K. Nakayama et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2493–2497
1)
CH2(CO2TCP)2
xylene
N
N
2) POCl3, DMF
O
N
O
OH
3) Ph3PCHCO2tBu, THF
NH2
O
S
S
N
N
46%
27
28
OtBu
OH
O
N
1) TsCl, then, 3-hydroxypiperidine
2) HCl-dioxane
O
N
N
S
N
75%
O
OH
29
O
Scheme 5. Synthesis of an ether tether variant.
tional tosylate gave starting material 22 predominantly.
Vilsmeier formylation followed by Wittig reaction and
deprotection led to the acrylic acid derivative. Finally,
formation of the primary carbamate and deprotection
gave 26. A similar protocol was utilized for the ether
tether analogue 29 (Scheme 5), but in this case tosylate
was the preferred leaving group. The synthetic interme-
diate 27 was prepared via Curtius rearrangement.5
N
EtO
O
N
N
N
N
O
N
O
OH
OH
N
30
S
(COCl)2
N
H
OMe
N
DMF, CH2Cl2
piperidine, pyridine
58% for 2 steps
22
N
O
1) (PhO)2P(O)Cl, iPr NEt
2
N
PMB
S
2) (R)-3-hydroxypiperidine
N
H
We next examined the photostability of the various
analogues. The ratios of E and Z-isomers after exposure
to light from a white fluorescent lamp are illustrated in
Table 1.
N
N
N
O
N
O
31
N
O
O
NH2
1) CCl3CONCO
2) HCO2Na
The effect of a 6-fluoro substituent on the pyridopyr-
imidine framework was modest; the stability of analogue
N
O
N
S
N
H
H
N
N
3) TFA
N
N
40% for 5 steps
32
N
Table 1. The ratio (Z=E) of the photoisomerization of the olefin with
timea
Scheme 6. Synthesis of the vinyl tetrazole analogue.
O
O
O
NH2
acid derivatives. The reactivities of the Wittig reagents
varied significantly, and we found that reactions pro-
ceeded most satisfactorily when matched with appro-
priate substrates: the 2-morpholino intermediate for
Me- and MeO-containing analogues, and the 2-hydroxyl
intermediate for Wittig–Horner reaction to generate a
fluorine-containing analogue. In the latter case, a mix-
ture of E and Z-isomers was generated. Treatment with
iodine8 furnished the E-isomer exclusively.
N
O
N
N
O
N
N
O
N
S
S
N
N
F
OH
OH
Me
18
5
O
O
O
O
NH2
N
O
N
O
N
S
N
H
Incorporation of fluorine onto the pyridopyrimidine
scaffold was achieved via the Shieman reaction9 as the
key step, starting with 10 (Scheme 3). The resulting
product (13) was treated with excess base to generate a
dianion, which coupled satisfactorily with the requisite
2-halomethylthiazole.7 Subsequent steps were similar to
those utilized in nonfluorinated variants.7
N
OH
26
Time (min)
5b (Z/E)
18b (Z/E)
26 (Z/E)
0
60
4/96
6/94
7/93
10/90
3/97
0/100
1/99
1/99
1/99
24/76
31/69
35/65
120
240
Scheme 4 outlines the synthesis of amide tether deriva-
tives. Condensation of a 2-aminothiazole with a pyridine-
4-carboxylic acid derivative using the conventional EDC
protocol, removal of the acetyl protecting group, and
cyclization with an active ester of malonic acid provided
the 2-hydroxypyridopyrimidine 22. For this analogue,
diphenylphosphate was found to be a superior leaving
group for the subsequent nucleophilic displacement by
piperidine derivatives; by contrast, the more conven-
a The ratio was calculated by peak area using LC–MS. Each sample
was dissolved in H2O–MeOH, 9:1 or MeOH (concentration; 25–
50 lg/mL). Sample solutions were illuminated with a white fluores-
cent lamp (500 Lx). No peaks other than the E and Z isomers were
detected during the study.
b The initial sample contained a slight amount of Z-isomer, presum-
ably generated during the evaporation of the solvent after column
chromatography.

K. Nakayama et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2493–2497
2497
O
R
O
NH2
N
O
N
O
N
O
N
N
O
N
S
N
H
S
N
H
N
N
H
N
OH
N
32
O
N
N
O
O
NH2
O
NH2
O
R=
26
33
Figure 3. Advanced photostable analogues.
Table 2. In vitro potentiation activity of novel EPIs
enhancement in potency.7 Like 26, this compound was
not susceptible to photoisomerization (Fig. 3).
MPC8 (AZT) lg/mLa
MPC4
(LVFX)
lg/mLa
MPC8
(LVFX)
lg/mLa
Without
HSA
With
0.125%
HSA
In conclusion, replacement of the ethylene tether
between the thiazole moiety and the pyridopyrimidine
scaffold with an amide bond gave analogues whose
stability with respect to photoisomerization was greatly
enhanced. This change also unexpectedly led to signifi-
cant improvement in potency. However, these com-
pounds did not possess sufficient solubility for
intravenous use. Efforts to address this issue will be
reported in due course.
1
8
16
1
0.5
<0.25
<0.25
0.5
8
26
32
33
0.5
<0.25
1
1
2
0.5
2
4
a Values are for 100% growth inhibition.
18 (Table 1) was similar to that of the parent 1 (data not
shown).
Compound 5, bearing a methyl group on the acrylic acid
moiety, was found to be more stable, though the isom-
erization was not prevented completely. Other modifi-
cations of the olefin gave no improvement (compounds
6 and 9).
References and notes
1. Poole, K.; Krebes, K.; McNally, C.; Neshat, S. J.
Bacteriol. 1993, 17, 7363.
2. Poole, K.; Gotoh, N.; Tsujimoto, H.; Zhao, Q.; Wada, A.;
Yamasaki, T.; Neshat, S.; Yamagishi, J.; Li, X. Z.;
Nishino, T. Mol. Microbiol. 1996, 21, 713.
3. Koehler, T.; Michea-Hamzehpour, M.; Henze, U.; Gotoh,
N.; Curty, L. K.; Pechere, J. C. Mol. Microbiol. 1997, 23,
345.
4. Mine, T.; Morita, Y.; Kataoka, A.; Mizushima, T.;
Tsuchiya, T. Antimicrob. Agents Chemother. 1999, 43, 415.
5. Nakayama, K.; Ishida, Y.; Ohtsuka, M.; Kawato, H.;
Yoshida, K.; Yokomizo, Y.; Hosono, S.; Ohta, T.;
Hoshino, K.; Ishida, H.; Yoshida, K.; Renau, T. E.;
By contrast, introduction of the amide tether (26) gave a
dramatic increase in stability. Conversely, the instability
of the analogue bearing an ether tether (29) was such
that a significant quantity of Z-isomer was already
detected in solution after only a few minutes exposure to
house light (data not shown).
The activities of the photostable efflux pump inhibitors
were assessed using PAM1723,10 an experimental strain
of P. aeruginosa in which the MexAB-OprM pump is
overexpressed and MexCD-OprJ and MexEF-OprN are
genetically disrupted. Data for the potentiation of two
MexAB-OprM substrates, LVFX and aztreonam (AZT),
in the presence or absence of human serum albumin
(HSA) is displayed in Table 2. To quantify the activity of
the inhibitors, we defined the term MPCn as the minimum
concentration (lg/mL) of inhibitor required to reduce
(potentiate) the activity of antibacterial drug n-fold.
ꢀ
Leger, R.; Zhang, J. Z.; Lee, V. J.; Watkins, W. J. Bioorg.
Med. Chem. Lett. 2003, 13, 4201.
6. Nakayama, K.; Ishida, Y.; Ohtsuka, M.; Kawato, H.;
Yoshida, K.; Yokomizo, Y.; Ohta, T.; Hoshino, K.; Otani,
T.; Kurosaka, Y.; Yoshida, K.; Ishida, H.; Lee, V. J.;
Renau, T. E.; Watkins, W. J. Bioorg. Med. Chem. Lett.
2003, 13, 4205.
7. Nakayama, K.; Kawato, H.; Watanabe, J.; Ohtsuka, M.;
Yoshida, K.; Yokomizo, Y.; Sakamoto, A.; Kuru, N.;
Ohta, T.; Hoshino, K.; Yoshida, K.; Ishida, H.; Cho, A.;
Palme, M. H.; Zhang, J. Z.; Lee, V. J.; Watkins, W. J.
Bioorg. Med. Chem. Lett. 2004, 14, 475.
8. Lutz, R. E.; Scott, G. W. J. Org. Chem. 1948, 13, 284–
296.
9. Kyba, E. P.; Liu, S.-T.; Chockalingam, K.; Reddy, B. R.
J. Org. Chem. 1988, 53, 3513.
10. Lomovskaya, O.; Warren, M. S.; Lee, A.; Galazzo, J.;
Fronko, R.; Lee, M.; Blais, J.; Cho, D.; Chamberland, S.;
The derivatives possessing the amide tether, in particu-
lar, showed excellent potency. We also extended the
SAR at the 2-substituent7 by the finding that the 3-(R)-
carbamoyloxy-substituted piperazine (26) was more
potent than the (S) isomer (33). Finally, we showed that
the replacement of carboxylic acid by tetrazole (32,
synthesized as shown in Scheme 6 via Knoevenagel
reaction with tetrazole acetate 30) gave a further
ꢀ
Renau, T.; Leger, R.; Hecker, S.; Watkins, W.; Hoshino,
K.; Ishida, H.; Lee, V. Antimicrob. Agents Chemother.
2001, 45, 105.