Synthesis of 4-phenoxybenzamide adenine dinucleotide as NAD analogue with inhibitory activity against enoyl-ACP reductase (InhA) of Mycobacterium tuberculosis

Article abstract of DOI:10.1016/j.bmcl.2007.05.084

The chemical synthesis of 4-phenoxybenzamide adenine dinucleotide (3), a NAD analogue which mimics isoniazid-NAD adduct and inhibits Mycobacterium tuberculosis NAD-dependent enoyl-ACP reductase (InhA), is reported. The 4-phenoxy benzamide riboside (1) has been prepared as a key intermediate, converted into its 5′-mononucleotide (2), and coupled with AMP imidazolide to give the desired NAD analogue 3. It inhibits InhA with IC₅₀ = 27 μM.

Full text of DOI:10.1016/j.bmcl.2007.05.084

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4588–4591
Synthesis of 4-phenoxybenzamide adenine dinucleotide
as NAD analogue with inhibitory activity against
enoyl-ACP reductase (InhA) of Mycobacterium tuberculosis
Laurent Bonnac,a Guang-Yao Gao,a Liqiang Chen,a Krzysztof Felczak,a Eric M. Bennett,a
Hua Xu,b TaeSoo Kim,b Nina Liu,b HyeWon Oh,b Peter J. Tongeb
and Krzysztof W. Pankiewicza,*
aCenter for Drug Design, University of Minnesota, Minneapolis, MN 55455, USA
bDepartment of Chemistry, Stony Brook University, Stony Brook, NY 11794-3400, USA
Received 11 April 2007; revised 23 May 2007; accepted 25 May 2007
Available online 31 May 2007
Abstract—The chemical synthesis of 4-phenoxybenzamide adenine dinucleotide (3), a NAD analogue which mimics isoniazid-NAD
adduct and inhibits Mycobacterium tuberculosis NAD-dependent enoyl-ACP reductase (InhA), is reported. The 4-phenoxy benzam-
ide riboside (1) has been prepared as a key intermediate, converted into its 50-mononucleotide (2), and coupled with AMP imidazo-
lide to give the desired NAD analogue 3. It inhibits InhA with IC50 = 27 lM.
Ó 2007 Elsevier Ltd. All rights reserved.
X
Tuberculosis (TB) is a chronic infectious disease caused
O
O
NHNH2
by mycobacteria of the ‘tuberculosis complex’, including
Mycobacterium bovis, Mycobacterium africanum, and
mainly Mycobacterium tuberculosis. TB now kills more
adults than all other infectious diseases combined. Ac-
tive TB is usually treated with isoniazid in association
with one or more other anti-TB drugs but multi-drug
resistant TB (MDR-TB) and very recently extensively
drug resistant TB (XDR-TB) has become a serious
and unsolved public health problem.1–6
O
H
4
NH2
KatG
N
X
NAD(P)
ADPR(P)
X = N or CH, NADH adduct
(
) - inhibits InhA
4S
X = N, isoniazid (INH)
X = N,
X = N,
NADP adduct (
NADP adduct (
)
4S
4R
- inhibits MabA
- inhibits DHFR
X = CH, benzoic acid hydrazide (BH)
)
Scheme 1. Inhibition of M. tuberculosis reductases.
Isoniazid (isonicotinic acid hydrazide, INH) discovered
in 1952 is still the most important drug for treatment
of TB. It is a prodrug which requires metabolic oxida-
tion by the M. tuberculosis enzyme, catalase-peroxidase
katG,4,7,8 to an isonicotinoyl radical which binds cova-
lently to the position 4 of NAD(P) cofactor (Scheme
1).8–10 The INH-NAD(P) adducts inhibit two enzymes
involved in the fatty acid biosynthetic pathway of
M. tuberculosis,11–14 NAD-dependent enoyl-acyl carrier
protein reductase (enoyl-ACP reductase, InhA) and
NAD(P)-dependent b-keto-ACP reductase (mycolic acid
biosynthesis A, MabA).15 Addition of the isonicotinoyl
radical to the C4 of the nicotinamide ring can result in
two stereoisomers, however, only 4(S) isomers of the
INH-NAD and INH-NAD(P) are potent inhibitors of
InhA (Ki = 0.75 nM) and MabA (Ki = 2.2 lM). In con-
trast, the 4(R) isomer INH-NAD(P) inhibits M. tubercu-
losis dihydrofolate reductase (DHFR), which is essential
for nucleic acid synthesis (Scheme 1).16 Interestingly, it
was found that benzoic acid hydrazide is also metabo-
lized by M. tuberculosis in a similar manner as INH
and shows a potent inhibition of InhA.
Keywords: M. tuberculosis; Enoyl-ACP reductase (InhA); NAD ana-
logues; Isoniazid; Fragment-based drug discovery.
*
A high prevalence of resistance to INH was observed,
mainly due to emerging KatG mutants that do not acti-
Corresponding author. Tel.: +1 612 625 7968; fax: +1 612 625
8252; e-mail: panki001@umn.edu
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.05.084

L. Bonnac et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4588–4591
4589
vate or poorly activate INH. Therefore, it has been sug-
gested that compounds that inhibit final targets of INH
but do not require activation by KatG have tremendous
promise as novel drugs for combating MDR-TB.8,17
Preformed INH-NAD would be such a compound,
but as dinucleotide it is useless as a potential drug due
to metabolic instability and lack of cell permeability.
Nevertheless, several reports have been recently pub-
lished on chemical formation of the INH-NAD adduct
(X = N), its benzoyl analogue (X = CH), and on synthe-
sis of fragments of these important activated metabo-
lites.15,18–22 The studies of chemical activation of a
mixture of INH and NAD revealed a rather complicated
activation pathway in which two diastereomers [the
4(S)-an inhibitor and the 4(R)-inactive isomer)] were
formed together with four cyclic diastereoisomers
(hemiamidals).19,20
Chemical synthesis of fragments of the NAD adducts
(Scheme 2) supplemented earlier observation on poor
stability of INH-NAD adduct and formation of hemi-
amidals. However, none of the synthesized 4-substituted
nicotinamide fragments showed activity against
InhA.21,22
Figure 1. Binding of NAD and NAD adducts to InhA. (A) Cocrystal
structure with NAD.25 (B) Docked INH-NAD adduct (colored by
atom) superimposed with the INH-NAD X-ray structure (blue) 9. (C)
Docked (4R) INH-NAD superimposed with the (4S) X-ray structure
(blue). (D) Docked 4-phenoxy NAD superimposed with the INH-
NAD X-ray structure (blue). For all panels, a cutaway of the protein
surface is shown in green, except for the surface of Phe149, which
moved to accommodate the adducts and is shown in red. Docking
results are the lowest energy conformations found.
Chemical activation of a mixture of INH and NAD re-
sulted in the formation of INH-NAD adduct but the
cleavage of the unstable glycosidic bond of nicotinamide
riboside (NR) moiety of NAD was observed resulting in
release of ADP ribose.18 In addition, NAD+ is a sub-
strate in numerous biochemical processes23 in which
the glycosidic bond is cleaved. Thus, the key issue in
designing a drug, based on inhibition of InhA, is the
construction of a tight binding NAD mimic, which will
be metabolically stable.
In contrast, a 4-phenoxy substitution of NAD, where
the position 4 is sp2-hybridized, resulted in a global
˚
minimum with a low 0.69 A RMSd compared to the
INH-NADH adduct (Fig. 1D).
To address glycosidic bond instability, we also replaced
the nicotinamide ring nitrogen with carbon, a change
which by itself gives benzamide adenine dinucleotide
(BAD or 1-deaza-NAD).26,27 Thus, we focused on syn-
thesis of 4-phenoxy-substituted BAD analogue 3.
To address undesirable hemiamidal formation, we rede-
signed the attachment at the position 4, and began by
investigating the geometry of the position 4 using molec-
ular modeling.24 Docking reproduced the NAD binding
mode observed in the X-ray structure (Fig. 1A) with an
The synthesis of 3 (Scheme 3) was accomplished by
phosphorylation of 1 with P(O)Cl3 (path a) and cou-
pling of the 50-monophosphate 2 with adenosine-50-
monophosphate imidazolide (path b).28
˚
RMS error of 0.34 A and the INH-NADH adduct with
˚
an RMS error of 0.80 A (Fig. 1B). With NADH
adducts, the stereochemistry at position 4 is critical.
Modeling shows that changing from S to R causes the
INH ring to overlap with the backbone of residues
193–194. In the docking simulation, these bad contacts
forced the nicotinamide portion of the molecule out of
its normal binding pocket for all conformations found.
The key intermediate for the above synthesis is a novel
C-nucleoside 1, which we synthesized via stereo-specific
coupling of the Grignard reagent (prepared from com-
mercially available 2-fluoro-5-iodobenzo-nitrile) with
the protected ribonolactone 4 (Scheme 4) followed by
reduction. The fluorinated carbon in the activated ortho
position of the nitrile was easily substituted by a pheno-
late anion generated from phenol under basic condi-
tion.29 Then, the nitrile group was hydrolyzed to the
carboxyamido group with potassium trimethylsilanolate
to give 6.30 The protective groups of the ribose are final-
ly removed with boron tribromide to give the desired 1.
The structures of all new compounds 1-3 were confirmed
by 1H, 31P NMR and HRMS, and their purity was
established by HPLC.31
˚
The global minimum had an RMSd of 6.6 A from the
X-ray structure, despite the adenosine half of the
molecule remaining in the correct position (Fig. 1C).
X
X
OH
NH
O
O
H
H
O
NH2
N
N
Alk
Alk
X=N or CH
Compounds 1–3 were evaluated against InhA. The inhi-
bition was monitored by following the inactivation of
Scheme 2. Stereochemistry and equilibrium of fragments of INH-
NAD adducts.

4590
L. Bonnac et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4588–4591
HO OH
O
O
O
O
O
NH2
NH2
O
O
b
N
RO
O
P
O
P
O
N
N
O
O
HO
OH
N
HO
OH
1, R = H
2, R =PO(OH)2
HO
OH
NH2
3
a
Scheme 3. Synthesis of 4-phenoxybenzamide adenine dinucleotide (3).
Finally, in contrast to other reports we found that much
smaller fragments of our target molecule 3, such as
nucleoside 1 and mono-nucleotide 2, showed 25% and
12% of inhibition of InhA at concentration of 700 lM
and 100 lM, respectively. This inspired us to synthesize
a number 4-substituted aromatic C-nucleoside and car-
bocyclic nucleoside analogues. Indeed, some of these
compounds (at the nucleoside level) showed 50% or more
inhibition of InhA at 100 lM. According to the princi-
ples of fragmented-based drug design we expect to boost
their activity by converting them into the corresponding
NAD analogues.
O
O
F
CN
NH2
BnO
RO
BnO
a
b
c
O
O
O
O
d
BnO
OBn
RO
OR
BnO
OBn
4
5
6
1
, R = Bn
, R = H
e
Scheme 4. Efficient, stereoselective synthesis of a key C-nucleoside 1.
Reagents: (a) i-PrMgCl, 2-fluoro-5-iodobenzonitrile, THF; (b)
BF3OEt2, Et3SiH; (c) Phenol, K2CO3; (d) Me3SiOK, THF; (e) BBr3,
CH2Cl2.
Acknowledgments
InhA in reaction mixtures containing NADH, dodece-
noyl CoA, the inhibitor, and InhA.32
These studies were funded by the Center for Drug De-
sign, University of Minnesota. We thank the Minnesota
Supercomputing Institute for computer time.
The C-nucleoside 1 showed weak activity at 700 lM.
However, the monophosphate 2 showed 12% inhibition
at concentration of 100 lM. The BAD analogue 3 inhib-
ited InhA with an IC50 of 27 lM (Table 1), confirming
our initial hypothesis that the simple aromatic mimics
of the INH-NAD adduct would show inhibitory activity
against InhA. Although the BAD analogue 3 is a micro-
molar inhibitor it does not require metabolic activation
and did not form a complicated mixture of active and
inactive stereo-isomers. If phosphorylated to the corre-
sponding NAD(P) analogues, 3 could inhibit two other
M. tuberculosis enzymes, MabA and DHFR. Interest-
ingly, we recently found that benzamide adenine dinu-
cleotide (BAD) is a substrate for bacterial NAD-
kinase.28
References and notes
1. Aziz, M. A.; Wright, A.; Laszlo, A.; De Muynck, A.;
Portaels, F.; Van Deun, A.; Wells, C.; Nunn, P.; Blanc, L.;
Raviglione, M. Lancet 2006, 368, 2142.
2. Raviglione, M. C.; Smith, I. M. N. Engl. J. Med. 2007,
356, 656.
3. Pasqualoto, K. F.; Ferreira, E. I.; Santos-Filho, O. A.;
Hopfinger, A. J. J. Med. Chem. 2004, 47, 3755.
4. Wei, C. J.; Lei, B.; Musser, J. M.; Tu, S. C. Antimicrob.
Agents Chemother. 2003, 47, 670.
5. Morlock, G. P.; Metchock, B.; Sikes, D.; Crawford, J. T.;
Cooksey, R. C. Antimicrob. Agents Chemother. 2003, 47,
3799.
6. Ormerod, L. P. Br. Med. Bull. 2005, 73, 17.
7. Zhao, X.; Yu, S.; Magliozzo, R. S. Biochemistry 2007, 46,
3161.
8. Rawat, R.; Whitty, A.; Tonge, P. J. Proc. Natl. Acad. Sci.
USA 2003, 100, 13881.
9. Rozwarski, D. A.; Grant, G. A.; Barton, D. H.; Jacobs,
W. R., Jr.; Sacchettini, J. C. Science 1998, 279, 98.
10. Lei, B.; Wei, C. J.; Tu, S. C. J. Biol. Chem. 2000, 275,
2520.
11. Schaeffer, M. L.; Agnihotri, G.; Volker, C.; Kallender, H.;
Brennan, P. J.; Lonsdale, J. T. J. Biol. Chem. 2001, 276,
47029.
The results indicate that further searching for more po-
tent inhibitors of InhA based on a similar approach is
justified. Once a nanomolar inhibitor with aromatic
structure similar to that of 3 is found, a simple replace-
ment of the pyrophosphate oxygen by the CH2 group
will afford the corresponding bis(phosphonate) analogue
of INH-NAD with expected improved metabolic stabil-
ity and drug-like properties.33,34
Table 1. Inhibition of InhA by compounds 1–3
12. Kuo, M. R.; Morbidoni, H. R.; Alland, D.; Sneddon, S.
F.; Gourlie, B. B.; Staveski, M. M.; Leonard, M.;
Gregory, J. S.; Janjigian, A. D.; Yee, C.; Kreiswirth, B.;
Iwamoto, H.; Perozzo, R.; Jacobs, W. R., Jr.; Sacchettini,
J. C.; Fidock, D. A. J. Biol. Chem. 2003, 26, 26.
Compounds
InhA inhibition
4-PhenoxyBR (1)
4-PhenoxyBRMP (2)
4-PhenoxyBAD (3)
25% Inhibition at 700 lM
12% Inhibition at 100 lM
IC50 = 27 lM

L. Bonnac et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4588–4591
4591
13. Choi, K. H.; Kremer, L.; Besra, G. S.; Rock, C. O. J. Biol.
Chem. 2000, 275, 28201.
14. Kremer, L.; Douglas, J. D.; Baulard, A. R.; Guy, M. R.;
Morehouse, C.; Alland, D.; Dover, L. G.; Lakey, J. H.;
Jacobs, W. R., Jr.; Brennan, P. J.; Minnikin, D. E.; Besra,
G. S. J. Biol. Chem. 2000, 275, 16857.
15. Ducasse-Cabanot, S.; Cohen-Gonsaud, M.; Marrakchi,
H.; Nguyen, M.; Zerbib, D.; Bernadou, J.; Daffe, M.;
Labesse, G.; Quemard, A. Antimicrob. Agents Chemother.
2004, 48, 242.
16. Argyrou, A.; Vetting, M. W.; Aladegbami, B.; Blanchard,
J. S. Nat. Struct. Mol. Biol. 2006, 13, 408.
was determined without considering the carbonyl oxygen
of the INH/NAD adduct, since this atom has no counter-
part in 3.
25. Dessen, A.; Quemard, A.; Blanchard, J. S.; Jacobs, W. R.,
Jr.; Sacchettini, J. C. Science 1995, 267, 1638.
26. Zatorski, A.; Watanabe, K. A.; Carr, S. F.; Goldstein, B.
M.; Pankiewicz, K. W. J. Med. Chem. 1996, 39, 2422.
27. Pankiewicz, K. W.; Lesiak, K.; Zatorski, A.; Goldstein, B.
M.; Carr, S. F.; Sochacki, M.; Majumdar, A.; Seidman,
M.; Watanabe, K. A. J. Med. Chem. 1997, 40, 1287.
28. Bonnac, L.; Chen, L.; Pathak, R.; Gao, G.; Ming, Q.;
Bennett, E.; Felczak, K.; Kullberg, M.; Patterson, S. E.;
Mazzola, F.; Magni, G.; Pankiewicz, K. W. Bioorg. Med.
Chem. Lett. 2007, 17, 1512.
17. Basso, L. A.; Zheng, R.; Musser, J. M.; Jacobs, W. R., Jr.;
Blanchard, J. S. J. Infect. Dis. 1998, 178, 769.
18. Nguyen, M.; Claparols, C.; Bernadou, J.; Meunier, B.
Chembiochem 2001, 2, 877.
29. Kristensen, J. L.; Vedso, P.; Begtrup, M. J. Org. Chem.
2003, 68, 4091.
19. Nguyen, M.; Quemard, A.; Broussy, S.; Bernadou, J.;
Meunier, B. Antimicrob. Agents Chemother. 2002, 46,
2137.
20. Broussy, S.; Coppel, Y.; Nguyen, M.; Bernadou, J.;
Meunier, B. Chemistry 2003, 9, 2034.
21. Broussy, S.; Bernardes-Genisson, V.; Quemard, A.; Meu-
nier, B.; Bernadou, J. J. Org. Chem. 2005, 70, 10502.
22. Delaine, T.; Bernardes-Genisson, V.; Meunier, B.; Berna-
dou, J. J. Org. Chem. 2007, 72, 675.
23. Such as mono- and poly-ADP-ribosylation of protein, cell
signaling ((cyclic ADP-ribose), DNA repair and recombi-
nation, and histone deacetylation catalyzed by the Sir2
family of NAD-dependent deacetylases.
30. Merchant, K. J. Tetrahedron Lett. 2000, 41, 3747.
31. Compound 1, 1H NMR (d, CD3OD), 7.99 (H2, s, 1H), 7.56
(H4, d, J = 8.4 Hz, 1H), 7.39 [H3 (OPh), d, J = 7.8 Hz, 2H],
7.18 [H4 (OPh), t, J = 7.8, 1H], 7.04 [H2 (OPh), d,
J = 7.8 Hz, 2H], 6.97 (H5, d, J = 8.4 Hz, 1H), 4.71 (H10, d,
J = 6.6 Hz, 1H), 4.0 (H20, m, 1H), 3.98 (H40, m, 1H), 3.85
(H30, m, 1H), 3.75 (H50500, m, 2H). HRMS calcd for
C18H20NO6 [M+H]+ 346.1212 found 346.1192.; Compound
3, 1H NMR (d, D2O), 8.31 (H8, Ade, s, 1H), 8.21 (H2, Ade,
s, 1H), 8.04 (H2, s, 1H), 7.30 (H4, d, J = 8.4 Hz, 1H), 7.25
(H3(OPh), d, J = 7.8 Hz, 2H), 7.05 (H4(OPh), t, J = 7.8,
1H), 6.85 (H2 (OPh), d, J = 7.8 Hz, 2H), 6.66 (H5, d,
J = 8.4 Hz, 1H), 5.87 (H10, d, J = 5.4 Hz, 1H), 4.57 (H10, d,
J = 7.2 Hz, 1H), 4.44 (ribose, m, 1H), 4.29 (ribose, m, 1H),
4.22–3.88 (ribose, m, 8H). 31P NMR (D2O) ꢀ10.18, ꢀ9.98
(d, J= 21.14). HRMS calcd for C28H31N6O15P2 [M-H]ꢀ
753.1401, found 753.1388.
32. All kinetic experiments were carried out on a Cary 300 Bio
(Varian) spectrophotometer at 25 °C. A typical reaction
mixture contained 30 mM PIPES, 150 mM NaCl, pH 8.0,
100 nM InhA, 25 lM dodecenoyl CoA, 250 lM NADH,
and a certain concentration of the inhibitor. The reaction
was initiated by the addition of InhA, and then monitored
at 340 nm. The initial velocities were measured from the
linear period of the assays (usually the first 1/2 min). The
inhibitions of InhA by compounds 1 and 2 were tested at
the specified concentration. For compound 3, inhibition
assays were carried out at different inhibitor concentration
ranging from 0 to 200 lM, and then the data were fitted
into the following equation to obtain the IC50 value:
v = v0/(1 + [I]/IC50).
24. Modeling was based on the X-ray structure of the InhA/
NAD complex (PDB entry 1ENY). The OPLS-2005 force
field and the GB/SA solvation model, as implemented in
MacroModel 9.1 (Schrodinger, Inc., 2006), were used for
all simulations. Hydrogens were added and minimized,
then the entire protein was minimized subject to a 10 kcal/
2
˚
mol-A restraint. This model was used for validation
docking of NAD. When binding to INH-NADH adducts,
Phe149 changes position to accommodate the adduct. The
protein was frozen except for residues 149, 150, and 158,
and a torsion search was performed to identify a low-
energy conformation of Phe149 similar to that seen in the
adduct X-ray structure (PDB entry 1ZID). This model was
used for docking adducts, with Phe149 and the adduct
unrestrained and the remainder of the protein frozen.
MCMM/LMCS conformational searching of adducts was
performed in blocks of 10,000 steps until (a) no new
structures within 5 kcal/mol of the global minimum were
found and (b) the rate of finding new structures within
10 kcal/mol of the global minimum decreased at least 10-
fold relative to the first block of 10,000 steps. All searches
converged after 40,000 or fewer steps. Minimizations were
33. Pankiewicz, K. W.; Lesiak-Watanabe, K. B.; Watanabe,
K. A.; Patterson, S. E.; Jayaram, H. N.; Yalowitz, J. A.;
Miller, M. D.; Seidman, M.; Majumdar, A.; Prehna, G.;
Goldstein, B. M. J. Med. Chem. 2002, 45, 703.
34. Pankiewicz, K. W.; Patterson, S. E.; Black, P. L.;
Jayaram, H. N.; Risal, D.; Goldstein, B. M.; Stuyver, L.
J.; Schinazi, R. F. Curr. Med. Chem. 2004, 11, 887.
˚
performed to a gradient of 0.05 kJ/mol-A. RMS errors for
the adducts were determined after superimposing PDB
entry 1ZID on the minimized model. The RMS error for 3