Inhibitors of bacterial enoyl acyl carrier protein reductase (FabI): 2,9-disubstituted 1,2,3,4-tetrahydropyrido[3,4-b]indoles as potential antibacterial agents

Article abstract of DOI:10.1016/S0960-894X(01)00405-X

An SAR study of a screening lead has led to the identification of 2,9-disubstituted 1,2,3,4-tetrahydropyrido[3,4-b]indoles as inhibitors of Staphylococcus aureus enoyl acyl carrier protein reductase (FabI).

Full text of DOI:10.1016/S0960-894X(01)00405-X

Bioorganic & Medicinal ChemistryLetters 11 (2001) 2241–2244
Inhibitors of Bacterial Enoyl Acyl Carrier Protein Reductase
(FabI): 2,9-Disubstituted 1,2,3,4-Tetrahydropyrido[3,4-b]indoles
as Potential Antibacterial Agents
Mark A. Seefeld,* William H. Miller, Kenneth A. Newlander, Walter J. Burgess,
David J. Payne, Stephen F. Rittenhouse, Terrance D. Moore, Walter E. DeWolf, Jr.,
Paul M. Keller, Xiayang Qiu, Cheryl A. Janson, Kalindi Vaidya, Andrew P. Fosberry,
Martin G. Smyth, Deborah D. Jaworski, Courtney Slater-Radosti
and William F. Huffman
GlaxoSmithKline Pharmaceuticals, Antimicrobial and Host Defense Division, 1250 S. Collegeville Road,
Collegeville, PA 19426, USA
Received 12 February2001; accepted 11 May2001
Abstract—An SAR study of a screening lead has led to the identification of 2,9-disubstituted 1,2,3,4-tetrahydropyrido[3,4-b]indoles
as inhibitors of Staphylococcus aureus enoyl acyl carrier protein reductase (FabI). # 2001 Elsevier Science Ltd. All rights reserved.
Introduction
stituted 1,2,3,4-tetrahydropyrido[3,4-b]indoles that
exhibit antibacterial activitybyinhibiting the FabI
enzyme. We also report compounds tested for in vitro
antibacterial activityagainst selected Gram-positive and
Gram-negative FabI-containing organisms.6 Triclosan,
a commercial antibacterial agent, has recentlybeen
shown to inhibit FabI7 and will be discussed in context
with these results.
The increasing resistance of clinicallyimportant patho-
gens to antibiotic treatment is of world-wide concern.1
New antibacterial agents that operate with distinctly
different mechanisms of action from current drug
therapies offer hope towards combating these multi-
drug-resistant organisms.2 In an effort to develop a
novel class of antibacterial compounds, we targeted the
inhibition of bacterial fattyacid synthesis (FAS-II). The
FAS-II elongation cycle utilizes several discrete mono-
functional enzymes with activity corresponding to indi-
vidual polypeptides effecting fatty acid chain elongation
Chemistry
The FabI inhibitor 3 was synthesized as shown in
Scheme 1. Carbobenzyloxy (Z) protection of commer-
3
and ultimatelycell membrane production. Enoyl acyl
carrier protein reductase (FabI) is the component of
FAS-II that catalyzes the final reaction in the enzymatic
sequence.4 In contrast, mammalian fattyacid synthesis
(FAS-I) employs a multifunctional enzyme complex in
which all enzymatic activities reside on a single poly-
peptide.5 Thus, there exists a potential for selective
inhibition of Gram-positive and Gram-negative bacterial
cell growth bythe inhibition of the FabI enzmy e.
Herein, we describe the SAR of a series of 2,9-disub-
ciallyavailable
1,2,3,4-tetrdarhoy-9
H-pyrido[3,4-
b]indole (1, Scheme 1) in quantitative yield followed by
N-9 alkylation using sodium hydride and 4-benzyloxy-
benzyl chloride in DMF gave 2 (92% yield). Hydro-
genolysis in methanol using Pd(OH)2 at 45 psi and
subsequent EDC-mediated acylation afforded com-
pound 3 in 82% overall yield. Compounds 6–15 and 19–
34 were likewise prepared with yields ranging from 72 to
90%. The regioisomeric derivative 16 was generated
from 2,3,4,5-tetrahydro-1H-pyrido[4,3-b]indole and ela-
borated in the manner described in Scheme 1. Indole 18
was similarlyconstructed from Z-protected 2-(methyl-
aminomethyl)indole. The racemic compound 17 was
*Corresponding author. Fax: +1-610-917-4206; e-mail: mark_a_
seefeld@sbphrd.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00405-X

2242
M. A. Seefeld et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2241–2244
the lowest concentration of compound that inhibited
visible growth.
Results and Discussion
Benzodiazepine 4 was identified as a weak lead com-
pound (IC50=7.8 mM) from high throughput screening
of our compound collection against S. aureus FabI. The
keypharmacophore (IC 50=2.7 mM) was determined to
be 2-benzoyl-1,2,3,4-tetrahydro-9H-pyrido[3,4-b]indole
(5).
Scheme 1. Preparation of 2,9-disubstituted 1,2,3,4-tetrahydro-
pyrido[3,4-b]indoles. Reagents: (a) Z-succinamide, Et3N, DMF; (b)
NaH, 4-benzyloxybenzyl chloride, rt; (c) Pd(OH)2, 1 M HCl in diox-
ane, H2 (3 atm), CH3OH, rt; (d) 4-hydroxybenzoic acid, EDC, HOBt,
DMF, rt.
synthesized in 89% isolated yield by the reaction of
tyramine and 4-hydroxybenzaldehyde in a Pictet–Spen-
gler condensation followed by acylation with 4-hydroxy-
benzoic acid.
Initial investigation focused on substitution at the N-9
position of the b-carboline subunit (Table 1). para-
Hydroxybenzoyl was chosen as the N-2 acyl component
because of its increased aqueous solubilityrelative to
the benzoyl moiety of 5. Increased inhibition of both S.
aureus and E. coli FabI was observed for 4-substituted
benzyl analogues, with the 4-hydroxybenzyl compound 3
exhibiting the greatest potency. Basic (12) and acidic (15)
substituents on the benzyl ring were less well tolerated.
Biology
The Escherichia coli and Staphylococcus aureus FabI’s
were cloned, overexpressed and purified as described
previously.8a FabI enzyme inhibition assays were car-
ried out in half-area, 96-well microtitre plates. Com-
pounds were evaluated in 150-mL assaymixtures
containing 100 mM NaADA, pH 6.5 (ADA=N-[2-
acetamido]-2-iminodiacetic acid), 4% glycerol, 0.25 mM
crotonoyl CoA, 50 mM NADH, and an appropriate
dilution of E. coli Fab I (usually60 nM). Inhibitors were
typically varied over the range of 0.01–10 mM. The con-
sumption of NADH was monitored for 20 min at 30 ꢀC
byfollowing the change in absorbance at 340 nm
(e=5.28 mMꢁ1). Initial velocities were estimated from
an exponential fit of the nonlinear progress curves
represented bythe slope of the tangent at t=0 min.
IC50’s were estimated from a fit of the initial velocities
to a standard, four-parameter model and are typically
reported as the meanꢂSD of duplicate determinations.
Triclosan, an inhibitor of Fab I, was included in all
assays as a positive control. IC50’s with S. aureus FabI
were determined similarlyexcept that NADPH was
used as substrate, and the concentration of enzyme was
typically 100 nM.
Further modifications to the most potent analogue 3
were subsequentlyinvestigated (Fig. 1). Regioisomer 16
and phenolic analogue 17 lost potencyrelative to 3.
Eliminating the conformational constraint afforded by
the 4-methylene position of the 1,2,3,4-tetrahydro-
pyrido[3,4-b]indole framework afforded compound 18.
Although this compound showed an improvement in
activityover 16 and 17, it was still significantlyless
active than the parent compound 3. Additional altera-
tions to the tricyclic structure were unsuccessful in
defining a more potent template.
Table 1. Modifications to R1
Compd
R1
S. aureus FabI
IC50 (mM)a,b
E. coli FabI
IC50 (mM) a,c
6
7
8
9
10
11
3
12
13
14
15
H
CH3
CH2CO2CH3
(CH2)4CO2CH3
C6H4CH2
1.4
1.4
9.8
4.6
18.0
11.0
34.9
26.8
27.0
8.8
4.2
21.4
4.2
Whole-cell antimicrobial activitywas determined by
broth microdilution. Test compounds were dissolved in
DMSO and diluted 1:10 in water to produce a 256 mg/
mL stock solution. The stock solution was seriallydilu-
ted using a Microlab AT Plus 2 (Hamilton Co., Reno,
NV, USA) into cation adjusted Mueller–Hinton broth
(Becton Dickenson, Cockeysville, MD, USA) on a 96-
well microtitre plate. After dilution, a 50 mL aliquot of
the test isolate (ꢃ1ꢄ106 cfu/mL) was added to each well
of the microtitre plate. The final test concentrations
ranged from 0.06 to 64 mg/mL. Inoculated plates were
incubated at 35 ꢀC in ambient air for 18–24 h. The mini-
mum inhibitoryconcentration (MIC) was determined as
1.3
3-(OH)C6H4CH2
4-(OH)C6H4CH2
4-(NH2)C6H4CH2
4-(F)C6H4CH2
4-(SO2CH3)C6H4CH2
4-(CO2H)C6H4CH2
0.42
0.11
0.67
0.25
0.33
8.7
4.5
>50
aValues are the averages of at least two experiments, standard devia-
tion is <50% in all cases.
bStaphylococcus aureus Oxford.
cEscherichia coli WT.

M. A. Seefeld et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2241–2244
2243
Figure 1. Effects of structural modifications to 3 on S. aureus FabI enzyme activity.
An investigation into substitution at N-2 of the b-car-
boline revealed that an aromatic acyl group at this
position was required for FabI enzyme activity (data
not shown). A simple one-dimensional, solution-phase
array consisting of 9-(4-hydroxybenzyl)-1,2,3,4-tetra-
hydropyrido[3,4-b]indole and various substituted ben-
zoic acids was prepared to rapidlydetermine the
optimum substitution pattern on the benzoyl ring
(Table 2).
The SAR revealed that maximum inhibition of S. aureus
and E. coli FabI generallyoccurred when a small lipo-
philic or hydroxyl group was located at the 4-position of
the benzamide ring. Compounds with two substituents
in this ring maintained activityagainst both enzymes
(30–33). Three or more substituents on the ring (34–36)
led to a loss of FabI activity, possibly due to unfavor-
able steric interactions. It was also observed that small
increases in substituent size decreased potency, espe-
ciallyin E. coli (20 vs 21 and 25 vs 26). A comparison of
the S. aureus enzyme inhibition of 3 with that of triclo-
san reveals that compound 3 is 10 times more active
against this specific enzyme. Conversely, triclosan is 10-
fold more potent than 3 against the E. coli enzyme.
These differences maybe attributed to the time-depen-
dent formation of a stable FabI-NAD+-triclosan tern-
arycomplex in E. coli8 as compared to the highly
reversible binding activitycharacterized bythe b-carbo-
line inhibitors. Differences in the binding site con-
formations of E. coli and S. aureus could also contribute
to these disparities in potency.
Table 2. Modifications to R2
Compd
R2
S. aureus FabI
IC50 (mM)a
E. coli FabI
IC50 (mM)a
19
20
21
22
23
24
25
26
27
28
3
29
30
31
32
33
H
4-CH3
4-C4H10
4-Cl
2,4-Cl
4-CO2CH3
4-NH2
4-NHCH3
2-OH
3-OH
0.49
0.24
0.51
0.18
1.8
2.0
1.0
2.1
2.7
5.2
>50
16.8
35.8
>50
4.0
38.6
4.3
>50
4.2
35.5
8.2
6.1
4.4
3.4
9.9
23.0
>50
0.43
0.72
3.4
The X-rayco-crystallization studyof 3 bound to E. coli
FabI/NAD+ assists in the interpretation of the b-car-
boline SAR (Fig. 2).9 Although the full inhibitor has not
been observed, the 4-hydroxybenzamide portion of the
molecule is well ordered, and the data shows that the
inhibitor is bound in the active site of the enzyme. The
benzamide moietyof the inhibitor is surrounded by
lipophilic residues in which Tyr 146 appears to partici-
pate in an orthogonal p-stacking interaction with the
aromatic ring. The 4-hydroxy substituent on the ring is
closelyflanked byMet 206 and Phe 203 which may
4-OH
4-OCH3
0.11
0.37
0.12
0.16
0.18
0.16
1.8
2-OH, 4-CH3
2-OH, 4-Cl
3-CH3, 4-OH
3-Cl, 4-OH
3,5-Cl, 2-OH
3,5-CH3, 4-OH
3,6-CH3, 2,4-OH
—
34
35
2.9
>50
1.1
36
Triclosan
aSee footnotes to Table 1.
Figure 2. Stereoview of the benzamide portion of inhibitor 3 bound in the active site of E. coli FabI/NAD+. For clarity, only the residues defining
the hydrophobic pocket of FabI have been illustrated. The benzamide moiety was revealed by an electron-density map.

2244
M. A. Seefeld et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2241–2244
Table 3. In vitro antibacterial activityof selected inhibitors with
<1 mM S. aureus FabI enzyme activity
Each compound in Table 3 was tested against a FabI
over-expressing strain of S. aureus. Thus, if FabI is an
antibacterial target of a compound, an increase in MIC
relative to the wild type strain would be expected. MIC
values for the compounds in Table 3 showed increases
of 4- to >16-fold, supporting that the mechanism of
action does indeed involve FabI inhibition.
Compd
MIC (mg/mL)a
S. aureus
M. catarrhalis
E. coli
20
22
27
3
30
31
32
0.5
4
4
0.5
1
1
1
1
0.03
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
4
8
8
4
2
8
4
0.06
Conclusion
We have identified potent inhibitors of S. aureus enoyl
acyl carrier protein reductase (FabI) that also demon-
strate whole-cell antimicrobial activityin Gram-positive
and Gram-negative organisms.
33
Triclosan
aMICs were repeated at least twice (nꢅ2) with data variabilityof one
dilution or less.
References and Notes
1. Neu, H. C. Science 1992, 257, 1073.
2. Chu, D. T. W.; Plattner, J. J.; Katz, L. J. Med. Chem. 1996,
39, 3853.
3. Fulco, A. J. Prog. Lipid Res. 1983, 22, 133.
4. (a) Bergler, H.; Fuchsbilder, S.; Hoegenauer, G.; Turn-
owsky, F. Eur. J. Biochem. 1996, 242, 689. (b) Heath, R. J.;
Rock, C. O. J. Biol. Chem. 1995, 270, 26538.
account for the steric effects observed in Table 2. The
carbonyl oxygen of inhibitor 3 is apparentlyinvolved in
a
hydroxyl of NAD+ and to the hydroxyl group of Tyr
156 (dashed orange lines). This particular active-site
interaction is analogous to that reported for Triclosan in
which a phenol hydroxyl functions in a similar capacity.8
hydrogen bonding interaction to the 20-ribose
5. Smith, S. FASEB J. 1994, 8, 1248.
6. (a) Heath, R. J.; Rock, C. O. Nature 2000, 406, 145. (b)
Heath, R. J.; Su, N.; Murphy, C. K.; Rock, C. O. J. Biol.
Chem. 2000, 275, 40128.
7. (a) Levy, C. W.; Roujeinikova, A.; Rice, D. W.; Rafferty,
J. B. Nature 1999, 398, 383. (b) Heath, R. J.; Yu, Y.-T.; Sha-
piro, M. A.; Olson, E.; Rock, C. O. J. Biol. Chem. 1998, 273,
30316.
8. (a) Qiu, X.; Janson, C. A.; Court, R. I.; Smyth, M. G.;
Payne, D. J.; Abdel-Meguid, S. S. Protein Sci. 1999, 8, 2529.
(b) Heath, R. J.; Rubins, J. R.; Holland, D. R.; Zhang, E.;
Snow, M. E.; Rock, C. O. J. Biol. Chem. 1999, 274, 11110.
Selected compounds with submicromolar S. aureus
enzyme activity were tested for in vitro antibacterial
activityagainst S. aureus Oxford, Moraxella catarrhalis
1502 and E. coli 7623 AcrABEFD+ organisms (Table
3). The MICs against S. aureus ranged from 0.5 to 4.0
(mg/mL) and corresponded well with FabI enzyme inhi-
bition. Antibacterial activityagainst the Gram-negative
M. catarrhalis strain was observed for the hydroxyl
containing compounds (3, 27, and 30–33). While anti-
bacterial activitywas not observed against E. coli, this
was expected due to the relativelypoor E. coli FabI
enzyme activity of these compounds. The MIC values of
triclosan were lower than we would have predicted
based on the corresponding IC50 levels. The high level
of in vitro potencyfor triclosan maybe the consequence
of having multiple cellular targets,10 and therefore MICs
maynot be exclusivelyattributable to FabI inhibition.
9. Crystallization conditions and structure determination
8a
procedures are similar to previouslypublished results. The
flipping loop of the enzyme is disordered. The crystal structure
˚
has been solved and refined to 2.4 A resolution (R=0.209,
Rf 0.277). The atomic coordinates have been deposited at the
Protein Data Bank (PDB) (accession number 1I30).
10. (a) McDonnell, G.; Pretzer, D. Actions and Targets of
Triclosan. ASM News 1998, 64, 670. (b) Russell, A. D. J. Appl.
Microbiol. 1997, 83, 155.