Novel inhibitors of bacterial protein synthesis: Structure-activity relationships for 1,8-naphthyridine derivatives incorporating position 3 and 4 variants

Article abstract of DOI:10.1016/j.bmcl.2004.03.091

Structure-activity relationships for a recently discovered novel ribosome inhibitor (NRI) class of antibacterials were investigated. Preliminary efforts to optimize protein synthesis inhibitory activity of the series through modification of positions 3 an

Full text of DOI:10.1016/j.bmcl.2004.03.091

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3299–3302
Novel inhibitors of bacterial protein synthesis: structure–activity
relationships for 1,8-naphthyridine derivatives incorporating
position 3 and 4 variants
Richard F. Clark,* Sanyi Wang, Zhenkun Ma, Moshe Weitzberg, Christopher Motter,
Michael Tufano, Rolf Wagner, Yu-Gui Gu, Peter J. Dandliker, Claude G. Lerner,
Linda E. Chovan, Yingna Cai, Candace L. Black-Schaefer, Linda Lynch, Douglas Kalvin,
Angela M. Nilius, Steve D. Pratt, Niru Soni, Tianyuan Zhang, Xiaolin Zhang and
Bruce A. Beutel
Global Pharmaceutical Research and Development, Abbott Laboratories, Abbott Park, IL 60064, USA
Received 21 January 2004; accepted 25 March 2004
Abstract—Structure–activity relationships for a recently discovered novel ribosome inhibitor (NRI) class of antibacterials were
investigated. Preliminary efforts to optimize protein synthesis inhibitory activity of the series through modification of positions 3 and
4 of the naphthyridone lead template resulted in the identification of several biochemically potent analogues. A lack of corre-
sponding whole cell antibacterial activity is thought to be a consequence of poor cellular penetration as evidenced by the
enhancement of activity observed for a lead analogue tested in the presence of a cell permeabilizing agent.
Ó 2004 Elsevier Ltd. All rights reserved.
The clinical effectiveness of currently available antibio-
tics is threatened by an increasing prevalence of resis-
tant pathogenic bacteria. Multidrug resistance among
Gram-positive organisms, including Streptococcus
pneumoniae, the most common respiratory pathogen, is
of particular concern. A recent worldwide surveillance
study found that resistance rates among S.pneumoniae
isolates have reached alarming levels for frontline ther-
apies such as penicillin (36%) and macrolides (31%).1
Emerging resistance to fluoroquinolones has also been
documented.2 To address this pressing healthcare issue,
new antibacterial drug classes unaffected by existing
resistance mechanisms are needed.3;4
rases, NRIs function in a mechanistically distinct man-
ner. Detailed mechanistic studies revealed that these
agents exert their antibacterial effects by inhibiting bac-
terial protein synthesis at the translation level through a
direct interaction with the bacterial ribosome.5
R4
O
F
R3
F
5
4
1
N
2
N 7
8
N
N
N
N
X
H
(1)
(2)
Here we report our preliminary investigation of the
structure–activity relationships (SAR) at positions 3 and
4 of the NRI series (1, 2) in the context of bacterial
protein synthesis inhibition. Position 3 in A-72310 was
viewed as an attractive starting place for our study given
the general requirement for a carboxylic acid moiety at
this position in classical quinolone SAR.6 In addition to
providing a point of attachment for a potentially supe-
rior ribosomal binding element, the productive incor-
poration of a non-quinolone substituent at this site
would further differentiate the NRI class. In the course
of these studies we found that removal of the terminal
Recently, we described the discovery and characteriza-
tion of A-72310 (1a, X ¼ NH2, R3 ¼ COOH), the par-
ent member of a novel ribosome inhibitor (NRI) class
that exhibits selective and broad-spectrum antibacterial
activity.5 Despite a close structural relationship to
quinolone antibacterials that inhibit type II topoisome-
* Corresponding author. Tel.: +1-847-938-2509; fax: +1-847-938-3403;
e-mail: rick.clark@abbott.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.03.091

3300
R.F.Clark et al./ Bioorg.Med.Chem.Lett.14 (2004) 3299–3302
amino group from the C-7 pyrrolidine in A-72310 re-
sulted in a compound (1a, X ¼ H, R3 ¼ COOH) with
retained protein synthesis inhibitory activity and im-
proved antibacterial potency. We therefore explored the
SAR at position 3 on this modified parent template.
Additionally, to complement our acid-replacement
studies, we synthesized a series of analogues wherein the
carboxylic acid at position 3 was removed from the
naphthyridone nucleus. This modification enabled us to
assess position 4 as an alternative site for optimization
on the simplified 1,8-naphthyridine core 2.
aldehyde (1c, R ¼ H) and ketone (1d, R ¼ Me) deriv-
atives by reduction with DIBAL and treatment with
MeMgBr, respectively. Subsequent conversion of alde-
hyde intermediates (1c) to alcohol (1f) and amine (1g)
targets proceeded smoothly using conventional meth-
odologies as outlined. Finally, position 4 variants (2b–
2l) were prepared from reactive intermediate 2a by
displacement of the 4-chloro substituent with nitrogen
nucleophiles. Accordingly, decarboxylation of 1a
(R1 ¼ DMB, X ¼ H) with quinoline and copper (I)
oxide at elevated temperatures and ensuing N-1 de-
protection yielded naphthyridone 1h which, when he-
ated with phosphoryl chloride, provided the requisite
precursor 2a.
The synthetic targets were accessed through modifica-
tions of the C-3 carboxylic acid group in common
intermediates 1a as depicted in Scheme 1. Compounds
1a were prepared by adapting known quinolone
chemistry.7 Treatment of enol ether 38 with 2,4-di-
methoxybenzylamine followed by base-promoted cycli-
zation of the resulting enamine afforded precursor 4.
Subsequent nucleophilic aromatic substitution of the 7-
chloro substituent with either 3-Boc-aminopyrrolidine
or pyrrolidine provided esters 1b, which, upon saponi-
fication, gave the desired naphthyridone intermediates
(1a). The 2,4-dimethoxybenzyl (DMB) and Boc pro-
tecting groups were removed in the presence of triflu-
oroacetic acid to unmask all final products including 1a
(R1 ¼ H). Elaboration of 1a (R1 ¼ DMB) to corre-
sponding C-3 amides (1e) was achieved using various
standard coupling methods. In turn, Weinreb amides 1e
(R0 ¼ OMe, R00 ¼ Me) were readily transformed to C-3
As previously described,5;9 analogues were evaluated for
protein synthesis inhibitory activity in a panel of cell-
free translation assays. Reported here are the IC50 values
measured in ribosome-containing S.pneumoniae cell
extracts translating a luciferase mRNA. Inhibitory
activity was quantified using a luminescence assay to
measure luciferase production in the presence of com-
pound. All observed inhibition (S.pneumoniae transla-
tion IC50 < 100 lM) was confirmed by conducting a
parallel assessment of protein synthesis inhibitory
activity using an Escherichia coli transcription/transla-
tion assay9 employing a b-galactosidase rather than
luciferase reporter to ensure that the observed activity
was not a consequence of direct luciferase inhibition.
Inhibitory activities obtained for the corresponding
O
O
O
O
O
O
F
F
OEt
1
F
R = DMB
OEt
OEt
N
N
N
X
a,b
c
Cl
N
N
d
R1
Cl
N
Cl
OEt
R1
1b, R1 = DMB
4, R1 = DMB
3h
e
R1 = H
O
O
e
O
O
O
O
F
F
F
R''
OH
R
N
R'
R' = OMe,R'' = Me
1
1
1c,R = H
1d R = Me
R = DMB
R = DMB
N
N
N
N
N
N
N
N
N
X
X
X
R1
R1
R1
g
f
1a, R1 = DMB
R1 = DMB
R1 = H
1e, R1 = DMB
e
e
R1 = H
R1 = H
1
1
h, e R = H, R = DMB
i, e R = DMB, X = H
R4
N
O
O
F
F
F
2a, R4 = Cl
Y
1f, Y = OH
1g, Y = NHR
j
k
2b-2k, R4 = NR'R''
N
N
N
H
N
N
N
H
N
N
X
1h
Scheme 1. Reagents and conditions: (a) 2,4-dimethoxybenzylamine, CH2Cl2, 25 °C, 3 h; (b) K2CO3, CH3CN, reflux, 16 h, 81%; (c) 3-(t-Butoxy-
carbonylamino)pyrrolidine or pyrrolidine, K2CO3, CH2Cl2, 25 °C, 16 h, 93–97%; (d) LiOH, EtOH–H2O, 60 °C, 91–94%; (e) TFA, 25–80 °C, 63–92%;
(f) Method A (Table 1, entry 4): (i) isobutyl chloroformate, NEt3, DMAP (cat.), CH2Cl2, 0 °C; (ii) NH3, 25 °C, 83%. Method B (Table 1, entries 5–7):
HOAT, HATU, Hunig’s base, DMA, 25 °C, 27–49%. Method C (Table 1, entry 15): (i) CDI, DMF, 100 °C, 2 h; (ii) NH2OH HCl, pyridine, 34%.
Method D (1e, R0 ¼ OMe, R00 ¼ Me): HNMeOMe HCl, TBTU, Hunig’s base, CHCl3, 25 °C, 79–88%; (g) 1c: DIBAL, CH2Cl2, )78 °C, 54–61%; 1d
(X ¼ H): MeMgBr, THF–CH2Cl2, 0 °C, 84%; (h) 1f (X ¼ H): NaBH4, EtOH–CH2Cl2, 0 °C, 72%; 1g: RNH2, NaBH3CN, AcOH, MeOH–CH2Cl2,
25 °C, 43–76%; (i) Cu2O, quinoline, 180 °C, 20 h, 88%; (j) POCl3, 100 °C, 1 h, 68%; (k) 1° or 2° amine or hydrazine, DMA, 115 °C, 37–91%.

R.F.Clark et al./ Bioorg.Med.Chem.Lett.14 (2004) 3299–3302
3301
Table 1. Bacterial protein synthesis inhibitory activity for com-
pounds 1
Table 2. Bacterial protein synthesis inhibitory activity for com-
pounds 2
Entry Compd
X
R3
S.pneumoniae
translation
IC50, lM
Entry Compd
R4
S.pneumoniae
translation IC50, lM
1
2
2b
2c
2d
2e
2f
NHCH2COOH
NHCH2CONH2
NH(CH2)2OMe
NH(CH2)2OH
NH(CH2)2Me
NH(CH2)2NH2
NHMe(CH2)2NMe2
1-Piperazinyl
>100
82
39
1
1a
1b
1d
1e
1e
1e
1e
1g
1g
1g
1g
1a
1b
1c
1e
1f
NH2 COOH
NH2 COOEt
5
20
2
3
3
NH2 COMe
NH2 CONH2
50
23
4
91
19
4
5
5
NH2 CONHCHMe2
NH2 CONH(CH2)2CHMe2
NH2 CONMe(CH2)2CHMe2 >100
>100
25
6
2g
2h
2i
11
18
6
7
7
8
>100
>100
89
8
NH2 CH2NHPh
54
25
9
10
2j
2k
NHCH(Me)Ph
NHNH2
9
NH2 CH2NHCH2Ph
NH2 CH2NH(CH2)3Ph
NH2 CH2NH(CH2)3NEt2
10
11
12
13
14
15
16
17
18
19
20
10
10
jecting amine (2g,h) or alkyl (2f) side chains displayed
the best activity across a range of functional groups
examined. Furthermore, amine 2g was among the most
potent NRIs identified in the current study. C-4 adducts
containing acid (2b), amide (2c), alcohol (2e) or hydra-
zine (2k) groups were less active. The relative inactivity
of piperazinyl analogue 2i compared to noncyclized
amine entries 2g and 2h reveals a preference for side
chain flexibility within this series.
H
H
H
H
H
H
H
H
H
COOH
COOEt
10
>100
>100
>100
>100
>100
>100
>100
>100
CHO
CONHOH
CH2OH
1g
1g
1g
1h
CH2NHPh
CH2NHCH2Ph
CH2NH(CH2)2NMe2
H
The data presented demonstrate that the position 3
carboxylic acid in lead compounds 1a is not required for
potent inhibition of bacterial protein synthesis. A vari-
ety of alternative side chain functionality extending from
positions 3 or 4 resulted in good biochemical potency.
However, while not required for biochemical potency,
the C-3 carboxyl appears to be important for whole cell
antibacterial activity. Although parent acids 1a exhib-
ited good antibacterial activity against S.pneumoniae
7257, a quinolone-resistant strain,10 and Haemophilus
influenzae GYR 1435, a representative Gram-negative
pathogen (Table 3), all position 3 and 4 variants were
devoid of activity against these strains.11;12 To prelimi-
narily assess whether this lack of activity might be
associated with poor cellular penetration relative to the
C-3 acids, 2g was tested for antibacterial activity in the
presence of ethylenediaminetetraacetic acid (EDTA, 1.1
mg/ml), a permeabilizer of the bacterial cell envelope.
Interestingly, a greater than 4-fold improvement in
assays were found to be in good agreement. Represen-
tative quinolone antibacterials including ciprofloxacin,
levofloxacin, and norfloxacin did not inhibit bacterial
protein synthesis in our assays.
In general C-7 aminopyrrolidinyl derivatives (Table 1,
entries 1–11) displayed a high tolerance for structural
variation at position 3 with regard to both nature and
size of substituents. Replacement of the carboxylic acid
group with ester (entry 2), amide (entries 4 and 6) or
amine-linked substitutions (entries 8–11) produced po-
tent inhibitors. Steric parameters of C-3 variants had
greater effects on inhibitory potency than did the type of
functional group incorporated. Although sterically
demanding substitution near the naphthyridone nucleus
was detrimental (entries 5 and 7), small adducts more
closely approximating the spatial environment of the
parent carboxyl were better tolerated as exemplified by
entries 2–4. Similarly, elongated tethers lacking branch
points proximal to the core resulted in good potency for
amide- or amine-linked derivatives projecting a variety
of functionality including aliphatic (entry 6), aryl (en-
tries 9 and 10), and amino groups (entry 11). Signifi-
cantly, the best inhibitors (entries 10 and 11) within this
series exhibited potency comparable to the parent acid
(entry 1) despite pronounced structural deviation at
position 3.
Table 3. Antibacterial activitya for NRI parent acids 1a
Compd
X
R3
S.pneumoniae
7257b
H.influenzae
GYR 1435
1a
1a
NH2
H
COOH 32
COOH
4
4
16
1
0.015
Levofloxacin
a MIC (minimum inhibitory concentration), lg/mL.
b Quinolone-resistant.
In contrast, any modification at position 3 in the cor-
responding C-7 pyrrolidinyl series (entries 12–20, Table
1) abolished inhibitory activity. These results were sur-
prising given the potency of the parent acid (entry 12)
and in light of the tolerance for variation observed in the
corresponding C-7 aminopyrrolidine series. Interest-
ingly, however, by installing a compensatory substituent
at position 4 on related core 2, potent inhibition was
achieved in the absence of the C-3 carboxylic acid
moiety (Table 2). Nitrogen-linked C-4 derivatives pro-
Table 4. Effect of permeabilizera on the antibacterial activityb of NRI
2g
Compd
R4
E.coli 7703
E.coli 7703
+ EDTA
2g
NH(CH2)2NH2
>128
32
a Ethylenetetraacetic acid (EDTA).
b MIC, lg/mL.

3302
R.F.Clark et al./ Bioorg.Med.Chem.Lett.14 (2004) 3299–3302
Ruan, X.; Saiki, A. Y. C.; Shen, L. L.; Soni, N. B.;
Wagner, R.; Weitzberg, M.; Xuei, X.; Zhong, P.; Beutel,
B. A. Antimicrob.Agents Chemother. 2003, 47, 3831.
6. Gootz, T. D.; Brighty, K. E. In The Quinolones; Andriole,
V. T., Ed.; Academic: San Diego, 1998; pp 29–80.
7. Radl, S.; Bouzard, D. Heterocycles 1991, 34, 2143.
8. Atarashi, S.; Imamura, M.; Kimura, Y.; Yoshida, A.;
Hayakawa, I. J.Med.Chem. 1993, 36, 3444.
9. Pratt, S. D.; David, C. A.; Black-Schaefer, C.; Dandliker,
P. J.; Xuei, X.; Warrior, U.; Burns, D. J.; Zhong, P.; Cao,
Z.; Saiki, A. Y. C.; Lerner, C. G.; Chovan, L. E.; Soni, N.
B.; Nilius, A. M.; Wagenaar, F. L.; Merta, P. J.; Trapha-
gen, L. M.; Beutel, B. A. J.Biomol.Screen. 2004, 9, 3.
10. Bruggemann, A. B.; Coffman, S. L.; Rhomberg, P.;
Hungh, H.; Almer, L.; Nilius, A. M.; Flamm, R.; Doern,
G. V. Antimicrob.Agents Chemother. 2002, 46, 680.
11. Antibacterial activity was assessed using the broth micro-
dilution method as described in the NCCLS guidelines.
2000. Methods for Dilution Antimicrobial Susceptibility
Testing.Approved Standard M7–A5. National Committee
for Clinical Laboratory Standards, Wayne, PA.
antibacterial activity was observed in the permeabilized
cells (Table 4). These findings suggest that the contri-
butions made by the C-3 carboxylic acid in NRI lead
structures 1a are physical rather than biochemical. It is
hoped that a more systematic study of NRI cores 1 and
2 aimed at identifying structural features that both
facilitate penetration into the bacterial cell and enhance
ribosomal inhibition will yield NRIs from the current
series with more desirable antibacterial profiles.
References and notes
1. Felmingham, D.; Reinert, R. R.; Hirakata, Y.; Rodloff, A.
J.Antimicrob.Chemother. 2002, 50(Suppl. S1), 25–37.
2. Chen, D. K.; McGeer, A.; de Azavedo, J. C.; Low, D. E.
N.Engl.J.Med. 1999, 341, 233.
3. Cassell, G. H.; Mekalanos, J. JAMA 2001, 285, 601.
4. Sutcliffe, J. A. Bioorg.Med.Chem.Lett. 2003, 13, 4159.
5. Dandliker, P. J.; Black-Schaefer, C.; Bui, M.; Cai, Y.;
Cao, Z.; Chovan, L. M.; Clark, R. F.; Daly, M. M.;
David, C. A.; Englund, E. E.; Hickman, R. K.; Kakavas,
S. J.; Lerner, C. G.; Merta, P.; Nilius, A. M.; Pratt, S. D.;
12. Ester 1b (Table 1, entry 2) exhibited modest antibacterial
activity against H.influenzae GYR 1435 (32 mg/mL) but
was inactive against a broader panel of both Gram-
negative and Gram-positive bacterial strains.