2,3,4,5-Tetrahydro-1H-pyrido[2,3-b and e][1,4]diazepines as inhibitors of the bacterial enoyl ACP reductase, FabI

Article abstract of DOI:10.1016/j.bmcl.2009.07.094

In the search for new antibacterial agents, the enzyme FabI has been identified as an attractive target. Employing a structure guided approach, the previously reported ene-amide series of FabI inhibitors were expanded to include 2,3,4,5-tetrahydro-1H-pyri

Full text of DOI:10.1016/j.bmcl.2009.07.094

Bioorganic & Medicinal Chemistry Letters 19 (2009) 5359–5362
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
2,3,4,5-Tetrahydro-1H-pyrido[2,3-b and e][1,4]diazepines as inhibitors
of the bacterial enoyl ACP reductase, FabI
b
a
a
a
b
Jailall Ramnauth a, , Mathew D. Surman , Peter B. Sampson , Bryan Forrest , Jeff Wilson , Emily Freeman ,
*
David D. Manning b, Fernando Martin a, Andras Toro a, Megan Domagala a, Donald E. Awrey a,
Elias Bardouniotis a, Nachum Kaplan a, Judd Berman a, Henry W. Pauls a,
*
a Affinium Pharmaceuticals Inc., 1243 Islington Avenue, Suite 600, Toronto, ON, Canada M8X 1Y9
b AMRI, 26 Corporate Circle, PO Box 15098, Albany, NY 12212-5098, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
In the search for new antibacterial agents, the enzyme FabI has been identified as an attractive target.
Employing a structure guided approach, the previously reported ene-amide series of FabI inhibitors were
expanded to include 2,3,4,5-tetrahydro-1H-pyrido[2,3-b and e][1,4]diazepines. These novel series incor-
porate additional H-bonding functions and can be more water soluble than their naphthyridinone pro-
genitors; diazepine 16c is shown to be efficacious in a mouse infection model.
Received 11 June 2009
Accepted 20 July 2009
Available online 23 July 2009
Keywords:
FabI inhibitors
Ó 2009 Elsevier Ltd. All rights reserved.
2,3,4,5-Tetrahydro-1H-pyrido
[2,3-e][1,4]diazepineAntibacterial
Enoyl ACP reductase
MRSA
Thigh infection model
There is an urgent demand for new antibiotics due to an in-
crease in drug resistant pathogenic bacteria.1 One approach to
combat antibiotic resistance is to target novel mechanisms of
action,2–4 and the enzymes of the bacterial fatty acid biosynthesis
cycle (FASII) cycle are among the attractive targets.5–7
addition of exogenous fatty acids. The availability of additional
tools, as represented by potent FabI inhibitors, would be helpful
in understanding the role of the FASII pathway in pathogenic
bacteria.
In bacteria, fatty acid biosynthesis is carried out by a series of
sequential enzymatic steps.7 In certain pathogenic bacteria (e.g.,
Escherichia coli and the genus Staphylococcus), an enoyl acyl carrier
protein reductase, designated FabI, is responsible for the terminal
step in the synthesis, and its corresponding gene is essential.8 In
contrast, similar transformations in humans are carried out by an
unrelated multifunctional enzyme designated FASI.9 This has led
to the pursuit of specific FabI inhibitors as novel antibacterial
agents10 which would not be expected to interfere with compara-
ble human biochemical processes.
O
O
H
N
A
N
Ar
N
B
X
X
N
N
H
O
N
N
H
1, A = S, B = C
I
2, A = CH, B = N
O
N
O
NH
NR
Ar
Y Ar
Y
N
N
N
HN
X
N
H
II
X
III
O
A recent publication has challenged the validity of targeting
FASII in general, and FabI in particular, for antibiotic therapy.11 This
work showed that inhibition of this pathway by the relatively weak
FabI inhibitors triclosan and cerulenin can be bypassed by the
Novel structures, namely the naphthyridinyl-ene-amides, were
identified as potent FabI inhibitors by GSK.10 Representatives of
this class, benzothiophene 1 and indole 2 were shown to possess
activity against Staphylococci and selected E. coli strains. For the
purposes of our anti-infectives program, we required compounds
suitable for both oral and intravenous dosing. However, inhibitor
1, the more potent congener, could not be readily formulated for
iv dosing. Indole 2, although more aqueous soluble, achieved only
modest blood levels upon oral dosing in rats.
* Corresponding authors. Current addresses: NeurAxon Inc., 480 University Ave.,
Toronto, ON, Canada M5G 1V2. Tel.: +1 416 673 8461 (J.R.). Campbell Family
Institute, MaRS Centre, Toronto, ON, Canada M5G 1L7. Tel.: +1 416 581 7620
(H.W.P.).
E-mail addresses: jramnauth@nrxn.com (J. Ramnauth), hpauls@uhnresearch.ca
(H.W. Pauls).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.07.094

5360
J. Ramnauth et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5359–5362
Scheme 1 details the synthesis of 1,5-diazepin-2-ones 11a–d
and 1,5-diazepines 9a,b. The key step for the latter is a Heck cou-
pling of acrylamide 8 with bromo-diazepine 6. Diazepinone 5
was reduced to yield 6; the acrylamide coupling partner 8 was pre-
pared by acylation of amine 7. Iodobromoaminopyridine 3 was
selectively N-arylated with azetidinone under palladium-catalyzed
conditions to furnish 4 which was subsequently converted to 5 by
a ring expansion.16 Heck coupling of 5 with tert-butyl acrylate, fol-
lowed by cleavage gave acid 10. The diazepinones 11a-d were ob-
tained by coupling 10 with N-methylated benzyl amines.
The syntheses of 1,4-diazepin-2-ones 16a,c, 18 and 1,4-diaze-
pine 20a,b are shown in Scheme 2. Compound 1210 was reacted
with substituted glycine methyl esters 13a,b to give the corre-
sponding amino esters which were cyclized with NaH in DMSO
to provide the diazepinones 14a,b. The ene-amides were con-
structed by Heck coupling. For example, 14a,b was coupled with
tert-butyl acrylate followed by deprotection and then amidation
to provide 16a,b. Compound 16c was prepared by subsequent re-
moval of the methoxybenzyl group. The preparation of 18 was sim-
ilar. In this instance, the methoxybenzyl group was removed prior
to the Heck reaction; bromide 17 was coupled with acrylamide 29
to give 18. Reduction of 17 followed by Boc protection gave bro-
mide 19 which was subjected to Heck coupling as usual to obtain
the diazepine 20.
The diazepinone 27 and diazocinone 28 were obtained accord-
ing to Scheme 3. Reaction of 2-chloro-3-cyanopyridine with ethyl
glycine ester gave compound 21 while palladium-catalyzed N-ary-
lation with azetidinone gave 22. These nitriles were reduced by
hydrogenation with palladium on activated carbon to give the cy-
clized products in one pot. Filtration followed by bromination
yielded compounds 23 and 24. Heck coupling with tert-butyl acry-
late, followed by tert-butyl cleavage with trifluoroacetic acid gave
the acids 25 and 26. The final amides were prepared by standard
amide couplings. Reduction of 24 with LAH, followed by protection
of the secondary amine with a Boc group gave 30. Heck coupling
with 29 and deprotection of the Boc group furnished the desired
diazocine 31.
Figure 1. Model of compound 16c docked in the S. aureus FabI active site.13 H-
bonds shown as yellow dashes.
We reasoned that ring expanded analogs incorporating basic
nitrogen atoms, as represented by compounds I, II, and III would
have modified physicochemical properties vis a vis the naphthyrid-
inones. The calculated aqueous solubility is significantly enhanced
by such modifications, especially if the carbonyl of the saturated
ring is absent, that is, X,X or Y,Y = H,H.12
Examination of the putative binding modes of these analogs
(Fig. 1), as deduced from X-ray structures of representative naph-
thyridinones,13 suggests an additional benefit. Compound 16c
was subjected to a molecular dynamics simulation using AMBER.14
The simulation revealed favorable free energies of binding com-
pared to that for compound 1. Compound 16c had a low desolva-
tion penalty and an additional H-bond between the amine NH
and the Lys199 carbonyl of FabI. Similar interactions are predicted
for compounds I, II, and III. Herein we present results on a series of
seven- and eight-membered heterocyclic FabI inhibitors.15
The left-hand side heterocyclic arylmethyl amines used in
Schemes 1–3 were generally prepared from aldehydes, by reduc-
tive amination, as previously described.10
The data for compounds prepared in this study are presented in
Table 1. The 1,5 diazepinones 11a-d were effective FabI inhibitors,
however, all things being equal (11b vs 2), no in vitro benefit over
O
H
F
N
Ar
N
S
N
N
H
O
N
N
H
HCl
O
11d
11a
11b
11c
h
O
H
N
H
N
Br
I
HO2C
Br
f, g
b
Br
N
a
N
NH2
N
N
N
H
N
H
N
NH2
10
O
5
O
4
3
c
O
H
H
N
N
R
Br
N
N
e
Z
Z
9a
+
N
N
N
H
N
H
6
7, R = H, Z = S, O
9b
, Z = S; , Z = O
d
8
, R = COCH=CH2, Z = S, O
Scheme 1. Reagents and conditions: (a) azetidin-2-one, Pd2(dba)3, xantphos, Cs2CO3, toluene, 90 °C; (b) Ti(OiPr)4, toluene, 110 °C; (c) LiAlH4, THF; (d) acryloyl chloride, Et3N,
THF; (e) tert-butyl acrylate, Pd2(dba)3, P(t-Bu)3, (i-Pr)2EtN, DMF, 100 °C; (f) tert-butyl acrylate, Pd(OAc)2, P(o-tolyl)3, (i-Pr)2EtN, EtCN, DMF, 100 °C; (g) (i) TFA, CH2Cl2; (ii) 4 M
HCl/dioxane; (h) ArCH2NHCH3, EDC, HOBt, DMF.

J. Ramnauth et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5359–5362
5361
O
R
R
O
16a, R = Me
HO2C
, R = Me
N
N
N
16b, R = 4-(MeO)Bn
e
15a
15b
f
N
, R = 4-(MeO)Bn
16c, R = H
N
N
N
H
N
H
O
c, d
R
O
Br
NH
O
Br
Br
N
a, b
f
O
Br
H
N
N
• HCl
R
OMe
N
H
N
NH•HBr
N
N
H
12
13a
14a
17
, R = Me
, R = Me
13b, R = 4-(MeO)Bn
14b, R = 4-(MeO)Bn
g
h, i,
O
N
O
Br
Z
NH
g, j
O
NH
NBoc
N
N
N
N
N
N
H
N
H
O
H
19
18
20a, Z = S; 20b, Z = O
Scheme 2. Reagents and conditions: (a) Et3N, DMF; (b) NaH, DMSO; (c) tert-butyl acrylate, Pd(OAc)2, P(o-tol)3, (i-Pr)2EtN, EtCN, DMF 100 °C; (d) TFA, CH2Cl2; (e) ArCH2NHCH3,
EDC, HOBt, DMF; (f) ACE-Cl, DCE; (g) 8 or 29, Pd(OAc)2, P(o-tol)3, (i-Pr)2EtN, EtCN, DMF; (h) LiAlH4, THF; (i) (Boc)2O, Et3N, CH2Cl2; (j) (i) TFA, CH2Cl2; (ii) 4 M HCl/dioxane.
and above the naphthyridinones is apparent. The benzofuran ana-
log 11d possesses superior activity against Staphylococcus aureus
and E. coli FabI. In addition, excellent antibacterial activities were
obtained against wild type S. aureus and, more importantly, two
resistant strains: MRSA and TRSA. Water solubility is somewhat
enhanced but not pH sensitive.
At acidic pH the 1,4 diazepinones are more soluble than their
1,5 isomers. Qualitatively this is understood by the chemical mod-
ification involved, that is, compounds 16a,c and 18 incorporate a
basic amine. The enzyme assay indicates that the most potent
1,4 diazepinones (16c, 18) have a free NH available for interaction
with the carbonyl of Lys 199. Congener 18 showed excellent anti-
bacterial activity. However, in head to head comparisons, the 1,5
diazepinones are more effective than the 1,4 isomers against FabI;
models indicate a tighter H-bond to Lys199.
erate an eight-membered ring analog. The diazocinone 28 is equi-
potent to its seven-membered ring analog 27.
To further improve solubility, we prepared analogs in which the
carbonyl groups were absent. As expected, solubilities of com-
pounds 9b, 20b, and 31 are improved in relation to their compar-
ators 11d, 18, and 28, respectively, especially at acidic pH. Against
S. aureus FabI, activity of 9b and 31 has decreased significantly.
This suggests that the carbonyl group might be involved in an
important hydrogen bond or is affecting the hydrogen bonding
capabilities of the proton on the adjacent nitrogen. It has been
shown that the naphthyridine moiety of the right-hand side forms
two H-bonds with a backbone alanine residue of S. aureus FabI and
this donor–acceptor network is required for optimal potency.13
Note, however, that against E. coli FabI a reduction in activity is
not seen. In fact, entries 9b, 20a, and 31 are 5–10-fold better than
that measured for S. aureus FabI. This indicates subtle differences in
the binding of these inhibitors; further results to be reported
separately.
In compound 27 the carbonyl is migrated from the two to the
three position of the diazepinone ring; a marginally less potent
compound results. It is interesting that FabI active site can also tol-
O
CN
Cl
CN
a
NH
HO
O
N
N
N
H
CO2Et
n
N
N
H
21
25 n = 1
26 n = 2
HCl
b
c, d
NH
g
e, f
CN
N
O
N
Br
c, d
NH
n
O
N
O
n
N
N
O
22
23
2
7 n = 1
n = 1
H
HBr
N
N
O
24 n = 2
28 n = 2
H
h, i
O
O
N
NBoc
Br
NH
j
N
O
O
N
HN
N
29
30
31
HN
Scheme 3. Reagents and conditions: (a) glycine ethyl ester, DMF, K2CO3, 100 °C; (b) azetidin-2-one, Pd(dba), xantphos, Cs2CO3, toluene, 90 °C; (c) Pd/C, H2, HOAc; (d) Br2; (e)
tert-butyl acrylate, Pd(OAc)2, P(o-tol)3, (i-Pr)2EtN, DMF, 100 °C; (f) (i) TFA, CH2Cl2; (ii) 4 M HCl/dioxane; (g) 3-methylbenzofuran-2-CH2NHCH3, EDC, HOBt, DMF; (h) LiAlH4,
THF; (i) (Boc)2O, Et3N, CH2Cl2; (j) (i) Pd(OAc)2, P(o-tol)3, (i-Pr)2EtN, EtCN, DMF; (ii) TFA, CH2Cl2.

5362
J. Ramnauth et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5359–5362
Table 1
In vitro activity and physicochemical data for compounds used in this study
MIC18
MRSAb
(lg/mL)
TRSAc
pKae
Solubilityf (
pH 4.0
l
g/mL)
pH 7.4
17
Compd
FabI IC50 (nM)
S. aureus
E. coli
S. aureusa
E. colid
1
2
22
100
31
130
200
7
26
67
510
130
43
64
150
25
136
840
25
14
12
NT
330
29
9
0.125
60.016
60.016
0.25
0.031
60.016
60.016
60.016
4
0.125
60.063
60.016
0.063
60.063
60.016
0.25
60.063
0.031
60.016
0.125
0.063
60.016
60.016
0.031
0.5
0.125
60.063
0.031
0.063
60.063
0.031
0.5
4
0.5
0.125
1
4
3.6
3.6
4.9
4.9
4.9
4.9
8.4
8.4
4.6
4.8
4.8
7.7
7.7
4.2
4.8
8.3
0.2
6.2
31
25
39
1.5
30
>100
73
>100
170
100
>650
NT
0.3
5.6
38
33
37
1.7
1.1
29
9.7
82
20
33
440
NT
17
60.063
1
1
11a
11b
11c
11d
9a
2
4
60.016
60.016
0.5
8
1
0.125
2
4
4
0.5
8
60.063
60.063
60.063
>32
9b
16a
16c
18
20a
20b
27
1
60.063
60.063
60.063
60.063
60.063
8
57
48
60
61
14
39
30
28
31
46
1400
170
16
210
The following strains were from the Affinium bacterial collection or the American Type Culture Collection (Manassas, VA): awild type S. aureus 29213. bMethicillin resistant S.
aureus 43300. cTriclosan resistant S. aureus 934335. dE. coli efflux pump mutant. eAdvanced Chemistry Development software calculated pKa. fSolubility measured by Shake
Flask or Millipore MultiScreen Solubility Filter Plate Method. NT not tested.
9. Chirala, S. S.; Huang, W. Y.; Jayakumar, A.; Sakai, K.; Wakil, S. J. Proc. Natl. Acad.
Sci. U.S.A. 1997, 94, 5588.
10. See for example the preceding paper in this issue and Seefeld, M. A.; Miller, W.
Dose Response with FabI Inhibitors
API Inhibitor 16c
API Inhibitor 16c
H.; Newlander, K. A.; Burgess, W. J.; DeWolf, W. E., Jr.; Elkins, P. A.; Head, M. S.;
Jakas, D. R.; Janson, C. A.; Keller, P. M.; Manley, P. J.; Moore, T. D.; Payne, D. J.;
Pearson, S.; Polizzi, B. J.; Qui, X.; Rittenhouse, S. F.; Uzinskas, I. N.; Wallis, N. G.;
Huffman, W. F. J. Med. Chem. 2003, 46, 1627.
2.50
2.00
FabI Inhibitor 2
FabI Inhibitor 2
Linezolid
Linezolid
1.50
1.00
0.50
0.00
11. Brinster, S.; Lamberet, G.; Stael, B.; Trieu-Cuot, P.; Gruss, A.; Poyart, C. Nature
2009, 458, 83.
12. For
a common left-hand side (i.e., benzothiophene) the ACD calculated
solubility at pH 7.4 for the series 1, 9a and 20a is 0.12, 0.91 and 2.3 mg/ml,
respectively.
13. Manning, D. D.; Xie, D.; Duong, T. N.; Decornez, H. Y.; Dorsey, M.; Awrey, D. E.;
Kaplan, N.; Bardouniotis, E.; Clarke, T.; Berman, J.; Pauls, H. 44th Interscience
Conference on Antimicrobial Agents and Chemotherapy, Washington, DC,
October, 2004; F-318.
14. Case, D. A.; Pearlman, D. A.; Caldwell, J. W.; Cheatham, T. E., III; Wang, J.; Ross,
W. S.; Simmerling, C. L.; Darden, T. A.; Merz, K. M.;. Stanton, R. V.; Cheng, A. L.;
Vincent, J. J.; Crowley, M.; Tsui, V.; Gohlke, H.; Radmer, R. J.; Duan, Y.; Pitera, J.;
Massova, I.; Seibel, G. L.; Singh, U. C.; Weiner P. K.; Kollman, P.A., AMBER 7,
University of California, San Francisco15, 2002.
15. Pauls, H. W.; Manning, D. D., Xie, D.; Duong, T. N.; Decornez, H. Y.; Dorsey, M.;
Awrey, D. E.; Kaplan, N.; Bardouniotis, E.; Clarke, T.; Berman, J.; In Structure-
Guided Design, Synthesis and In Vivo Characterization of Inhibitors of Bacterial
Fatty Acid Biosynthesis, 230th ACS National Meeting, Washington, DC, Aug. 28–
Sept. 1, 2005; COMP 286.
-0.50
-1.00
0
3
10
30
100
Figure 2. Neutropenic murine thigh infection model.19
Selected inhibitors were tested in a neutropenic mouse thigh
infection model. Figure 2 shows that the diazepine 16c was compa-
rable to Linezolid at 100 mg/kg orally and superior to earlier mem-
bers of this series namely indole 2. This is despite the fact that
indole 2 possesses superior in vitro numbers, which underscores
the importance of ancillary properties in the optimization of
in vivo efficacy.
In conclusion, we described a series of seven- and eight-mem-
bered heterocyclic inhibitors of S. aureus and E. coli FabI. These
compounds are potent enzyme inhibitors with excellent antibacte-
rial activity. These inhibitors were designed to exhibit modified
physiochemical properties; improved efficacy results. These data,
underscored by the in vivo efficacy, tend to support FabI as a target
for antibacterial therapy.
16. Klapars, A.; Parris, S.; Anderson, K. W.; Buchwald, S. L. J. Am. Chem. Soc. 2004,
126, 3529.
17. S. aureus FabI inhibition assays were carried out in half area 96-well plates,
with an enzyme concentration of 3 nM enzyme, as described in Payne, D. J.;
Miller, W. H.; Berry, V.; Brosky, J.; Burgess, W. J.; Chen, E.; DeWolf, W. E., Jr.;
Fosberry, A. P.; Greenwood, R.; Head, M. S.; Heerding, D.; Janson, C. A.;
Jaworski, D. D.; Keller, P. M.; Manley, P. J.; Moore, T. D.; Newlander, K. A.;
Pearson, S.; Polizzi, B. J.; Qiu,X.; Rittenhouse, S. F.; Slater-Radosti, C.; Salyers, K.
L.; Seefeld, M. A.; Smyth, M. G.; Takata, D. T.; Uzinskas, I. N.; Vaidya, K.; Wallis,
N. G.; Winram, S. B.; Yuan, C. C. K.; Huffman, W. F. Antimicrob Agents Chemother.
2002, 46, 3118.E. coli FabI inhibition assays were carried out in half area 96-
well plates containing 150
lL of 100 mM MOPS pH 7.2, 50 mM ammonium
acetate, 3% glycerol, 25 M crotonyl-ACP, 50
l
lM NADH, and 0.5 nM enzyme.
Consumption of NADH was monitored at 340 nm and a titration of standard
compound was included on each dose response plate as a positive control.
18. MIC testing was performed according to CLSI guidelines using the
microdilution method in 96-well plates: National Committee for Clinical
Laboratory Standards. 2003. Methods for Dilution Antimicrobial Susceptibility
Tests for Bacteria that Grow Aerobically, 5th ed., Approved Standard M7-A6.
NCCLS, Wayne, PA, USA.
19. By a modification of the published method (vida infra), mice (6 per group) were
rendered neutropenic, and inoculated in each thigh with ꢀ105 CFU of S. aureus
29213. Two hours post infection single oral doses of compounds were
administered (time 0). Efficacy was determined 24 hours post-dosing and
expressed as the average change in log CFU per thigh compared to time 0:
Maglio D, Banevicius MA, Sutherland C, Babalola C, Nightingale CH, Nicolau DP.
2005. Pharmacodynamic profile of ertapenem against Klebsiella pneumoniae
and Escherichia coli in a murine thigh model. Antimicrob Agents Chemother. 49,
276–280.
References and notes
1. Yoneyama, H.; Katsumata, R. Biosci. Biotechnol., Biochem. 2006, 5, 1060.
2. Leef, M. Nature 2004, 431, 892.
3. Walsh, C. Nat. Rev. Microbiol. 2003, 1, 65.
4. Projan, S. J. Curr. Opin. Pharmacol. 2002, 2, 513.
5. Campbell, J. W.; Cronan, J. E., Jr. Annu. Rev. Microbiol. 2001, 55, 305.
6. Payne, D. J.; Warren, P. V.; Holmes, D. J.; Ji, Y. O.; Longsdale, J. T. Drug Discovery
Today 2001, 6, 537.
7. Heath, R. J.; White, S. W.; Rock, C. O. Prog. Lipid Res. 2001, 40, 467.
8. Moir, D. T. Curr. Drug Targets Infect. Disord. 2005, 5, 297.