The first chemical synthesis of the core structure of the benzoylhydrazine-NAD adduct, a competitive inhibitor of the Mycobacterium tuberculosis enoyl reductase

Article abstract of DOI:10.1021/jo051901z

An isoniazid-NAD adduct has been recently proposed as the ultimate metabolite responsible for the antituberculous activity of isoniazid (INH). Its structure results from binding of the isonicotinoyl radical at C4 position of the nicotinamide ring of NAD w

Full text of DOI:10.1021/jo051901z

The First Chemical Synthesis of the Core Structure of the
Benzoylhydrazine-NAD Adduct, a Competitive Inhibitor of the
Mycobacterium tuberculosis Enoyl Reductase
Sylvain Broussy,† Vania Bernardes-Ge´nisson,† Anna¨ık Que´mard,‡ Bernard Meunier,† and
Jean Bernadou*,†
Laboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne, 31077 Toulouse Cedex 4,
France, and Institut de Pharmacologie et de Biologie Structurale du CNRS, 205 route de Narbonne,
31077 Toulouse Cedex, France
bernadou@lcc-toulouse.fr
Received September 9, 2005
An isoniazid-NAD adduct has been recently proposed as the ultimate metabolite responsible for
the antituberculous activity of isoniazid (INH). Its structure results from binding of the isonicotinoyl
radical at C4 position of the nicotinamide ring of NAD with further possible and debated cyclization
to form a cyclic hemiamidal derivative. Replacing the pyridine cycle of INH in INH-NAD adduct
by a phenyl cycle (BH-NAD adduct) was shown previously to still retain the activity. On these
bases, the core structure (4-benzoyl-1,4-dihydronicotinamide ribonucleoside) of the BH-NAD adduct
and a series of analogues have been synthesized by using 3,4-pyridinedicarboximide as starting
material. Depending on the nature of the substituent (pyridine or aryl) and on the oxidized or the
reduced state of the nicotinamide nucleus, they were found either in a cyclized hemiamidal or an
opened form or were shown to exist in equilibrium under cyclized or opened forms. Although none
of these compounds could significantly inhibit activity of the InhA or MabA reductases (two possible
targets of isoniazid), they represent attractive targets to develop potential second-generation
inhibitors, including the total chemical synthesis of the bioactive BH-NAD adduct.
Introduction
mycobacterial catalase-peroxidase KatG to generate an
active form, most likely the isonicotinoyl radical, respon-
sible for the lethal effect on bacterial cells. This active
form binds covalently to the nicotinamide moiety of
NAD(H)6-8 and NADP(H),9 coenzymes of the reductases
InhA10,11 and MabA,9 respectively, two key enzymes
involved in the biosynthesis of mycolic acids.9,12,13 INH-
Isoniazid (isonicotinic acid hydrazide, or INH) is one
of the oldest synthetic antituberculous drugs which has
been and is still widely used in prophylaxis and treat-
ment of tuberculosis.1,2 It is a prodrug3-5 activated by the
* To whom correspondence should be addressed. Tel: 33 (0)5 61 33
31 17. Fax: 33 (0)5 61 55 30 03.
† Laboratoire de Chimie de Coordination du CNRS.
‡ Institut de Pharmacologie et de Biologie Structurale du CNRS.
(1) Bartmann, K.; Iwainski, H.; Kleeberg, H. H.; Mison, P.; Offe, H.
H.; Otten, H.; Tettenborn, D.; Trnka, L. In Antituberculosis Drugs;
Bartmann, K., Ed.; Springer-Verlag: Berlin, 1985.
(2) Blanchard, J. S. Annu. Rev. Biochem. 1996, 65, 215-239.
(3) Sinha,B. K. J. Biol. Chem. 1983, 258, 796-801.
(4) Zhang, Y.; Heym, B.; Allen, B.; Young, D.; Cole, S. Nature 1992,
358, 591-593.
(5) Johnsson, K.; Schultz, P. G. J. Am. Chem. Soc. 1994, 116, 7425-
7426. Johnsson, K.; King, D. S.; Schultz, P. G. J. Am. Chem. Soc. 1995,
117, 5009-5010. Nguyen, M.; Claparols, C.; Bernadou, J.; Meunier,
B. C. R. Acad. Sci. 2001, 5, 1-6.
(6) Rozwarski, D. A.; Grant, G. A.; Barton, D. H. R.; Jacobs, W. R.,
Jr.; Sacchettini, J. C. Science 1998, 279, 98-102.
(7) Wilming, M.; Johnsson, K. Angew. Chem., Int. Ed. 1999, 38,
2588-2590.
(8) Nguyen, M.; Claparols, C.; Bernadou, J.; Meunier, B. ChemBio-
Chem 2001, 2, 877-883.
(9) Ducasse-Cabanot, S.; Cohen-Gonsaud, M.; Marrakchi, H.; Nguy-
en, M.; Zerbib, D.; Bernadou, J.; Daffe´, M.; Labesse, G.; Que´mard, A.
Antimicrob. Agents Chemother. 2004, 48, 242-249.
(10) Banerjee, A.; Dubnau, E.; Que´mard, A.; Balasubramanian, V.;
Sun Um, K.; Wilson, T.; Collins, D.; de Lisle, G.; Jacobs, W. R., Jr.
Science 1994, 263, 227-230.
10.1021/jo051901z CCC: $30.25 © 2005 American Chemical Society
Published on Web 11/10/2005
10502
J. Org. Chem. 2005, 70, 10502-10510

Toward Isoniazid-Derived InhA Inhibitors
FIGURE 1. Structures of reduced (dihydropyridine) INH-NAD adducts (forms A and B), oxidized (pyridinium) INH-NAD adduct
(form C), and benzoylhydrazine-NAD adduct (form A′).
NAD and INH-NADP adducts have been shown to be
inhibitors of InhA and MabA.9,14 The crystal structure of
the INH-NAD adduct bound to InhA evidenced the
covalent addition of an isonicotinoyl substituent at posi-
tion 4 of the nicotinamide ring. It has been interpreted
as the form A with configuration 4S (Figure 1).6 However,
our recent studies have shown that INH-NAD adducts
prepared under biomimetic conditions (using manganese
pyrophosphate as a mimic of KatG protein8) are present
FIGURE 2. 4-Benzoyl-1,4-dihydronicotinamide ribonucleoside
in solution as a mixture of opened keto-amide (form A;
two epimers, both present as minor compounds) and
cyclized hemiamidal structures (form B; four diastere-
oisomers, two of them are the major compounds),8,15
raising the question about the effective active form(s) of
INH-NAD adduct. One oxidized adduct (derived from
oxidation of the dihydropyridine form to a pyridinium
entity), initially proposed by Johnsson and Wilming as
being a ketoamide structure,7 was recently shown to have
a hemiamidal structure (form C, two epimers).16 Such
an adduct has been found to be unable to inhibit InhA
activity.7,16
A substantial fraction (more than one-third) of all
clinical isolates that are resistant to INH results from
KatG mutations.17 Consequently, compounds able to
inhibit InhA, the molecular target of INH, without
requiring activation by KatG have tremendous promise
as novel drugs to fight multidrug resistant tuberculosis.
Recently, it was shown that use of benzoylhydrazine (BH)
instead of isoniazid (i.e., isonicotinoylhydrazine) leads
also to potent inhibitors of InhA, named BH-NAD
adduct(s), through the covalent attachment of the benzoyl
group in the place of the isonicotinoyl group, at position
4 of the nicotinamide ring of NAD (Figure 1, the opened
structure A′ has been proposed for such adduct).14 In this
work, we described the first chemical synthesis of the core
structure of the BH-NAD adduct, i.e., the 4-benzoyl-1,4-
1: the core structure of the benzoylhydrazine-NAD adduct
(form A′, Figure 1), a competitive inhibitor of the InhA
reductase.
FIGURE 3. Possible equilibrium between opened and cyclized
forms of 4-aroylpyridine-3-carboxamides.
dihydronicotinamide ribonucleoside 1 (Figure 2), and
some related compounds as potential inhibitors of InhA.
This approach could also give access to a total chemical
synthesis of the BH-NAD adduct, currently only ob-
tained by biochemical preparation in very low yield and
without full chemical characterization. Furthermore,
since the 4-aroylpyridine-3-carboxamide fragment can
exist under opened (keto-amide) or cyclized (hemiamidal)
debated forms (Figure 3), additional information on these
structures and on factors governing the cyclization
process are also presented throughout this work and will
be discussed.
(11) Que´mard, A.; Dessen, A.; Sugantino, M.; Jacobs, W. R., Jr.,
Sacchetini, J. C.; Blanchard, J. S. J. Am. Chem. Soc. 1996, 118, 1561-
1562.
Results and Discussion
Syntheses of Hemiamidal Precursors. In a previ-
ously reported synthesis of 4-benzoylnicotinamide (Figure
3, left, X ) CH) by an ortho-metalation-electrophilic
substitution sequence on N,N-diisopropylnicotinamide,
we have shown that the equilibrium between the opened
form is completely shifted toward the cyclized form
(Figure 3, X ) CH).18 This initial observation led us to
consider a one-step synthesis of this cyclized hemiamidal
from the commercially available 3,4-pyridine dicarbox-
imide 2 (Scheme 1). Addition of phenyllithium at -80
°C in THF proceeded in good yield furnishing a mixture
of para and meta adducts (3 and 3′, R1 ) C6H5, Scheme
(12) Takayama, K.; Wang, L.; David, H. L. Antimicrob. Agents
Chemother. 1972, 2, 29-35.
(13) Que´mard, A.; Lacave, C.; Lane´elle, G. Antimicrob. Agents
Chemother. 1991, 35, 1035-1039.
(14) Rawat, R.; Whitty, A.; Tonge, P. J. PNAS 2003, 100, 13881-
13886.
(15) Broussy, S.; Coppel, Y.; Nguyen, M.; Bernadou, J.; Meunier, B.
Chem. Eur. J. 2003, 9, 2034-2038.
(16) Broussy, S.; Bernardes-Ge´nisson, V.; Coppel, Y.; Bernadou, J.;
Meunier, B. Org. Biomol. Chem. 2005, 3, 670-673.
(17) Ramaswamy, S. V.; Reich, R.; Dou, S.-J.; Jasperse, L.; Pan, X.;
Wanger, A.; Quitugua, T.; Graviss, E. A. Antimicrob. Agents Chemo-
ther. 2003, 1241-1250. Heym, B.; Alzari, P. M.; Honore´ N.; Cole, S.
T.; Mol. Microbiol. 1995, 15, 235-245. Morris, S.; Bai, G. H.; Suffys,
P.; Portillo-Gomez, L.; Fairchok, M.; Rouse, D. J. Inf. Dis. 1995, 171,
954-960. Pretorius, G. S.; van Helden, P. D.; Sirgel, F.; Eisenach, K.
D.; Victor, T. C. Antimicrob. Agents Chemother. 1995, 39, 2276-2281.
Rouse, D. A.; Li, Z.; Bai, G. H.; Morris, S. L. Antimicrob. Agents
Chemother. 1995, 39, 2472-2477.
(18) Broussy, S.; Bernardes-Ge´nisson, V.; Gornitzka, H.; Bernadou,
J.; Meunier, B. Org. Biomol. Chem. 2005, 3, 666-669.
J. Org. Chem, Vol. 70, No. 25, 2005 10503

Broussy et al.
SCHEME 1. Synthesis of Hemiamidal Precursorsa
SCHEME 2. Alkylation of Hemiamidal
Compounds with Alkyl Bromides
a The pyridine carboxamide ring is numbered in the usual way;
IUPAC nomenclature is indicated for each prepared compound in
the Experimental Section.
ridine (Table 1).20 From the mixture of the regioisomers
meta and para, compounds 4-7 were purified by silica
gel column chromatography or, more easily, by recrys-
tallization from acetone or ethanol. The structures of the
para isomers 4-6 were attributed by analogy with the
NMR chemical shifts of H2 and H6 previously observed
for compounds 3 (H2 was more deshielded than H6)18 and
3′ (this work; H6 was more deshielded than H2). Surpris-
ingly, in the case of the pair of regioisomers 7/7′ (with
the phenyl linked to C7 through a methylene unit), H6
was more deshielded than H2 in the major product.
HMQC, HMBC, and Jmod NMR experiments allowed a
complete attribution of protons and carbons of this
molecule. A 3J correlation between C7 and H5 in the
TABLE 1. Various Hemiamidal Precursors (See General
Formula in Scheme 1)
4
HMBC spectrum (instead of an unlikely J correlation)
and the spatial proximity between H5 and protons of the
methylene linker (revealed by NOE experiments) showed
that the major isomer was the para regioisomer 7, as for
compounds 4-6.
Syntheses of Pyridinium Derivatives (Scheme 2
and Table 2). Alkylation of compounds 3-5 by various
alkyl bromides in THF and/or ether afforded the pyri-
dinium derivatives 9-17. Condensation of 3 with (2-
acetoxyethoxy)methyl bromide in THF gave instanta-
neously 9 as a white precipitate. Products 14 and 16 were
obtained in the same way. A similar reactivity was also
observed with 1-bromo-(2,3,5)-tri-O-acetyl-â-D-ribofura-
nose giving the compound 12. However, all of these
compounds were rather unstable in solution at ambient
temperature due to their possible reaction with nucleo-
philes, giving back in 24 h substantial amounts of the
starting products 3-5 through N-side-chain or N-glyco-
sydic bond cleavage. So, we tried to lower the electrophilic
nature of carbon C1′ of the chain by replacing the adja-
cent ether oxygen atom with an alkoxycarbonyl group.
The reactivity of ethyl bromoacetate and ethyl 2-bro-
mopropionate with the corresponding hemiamidals 3-5
in THF at reflux was sufficient to obtain the products
10, 11, 15, and 17 in good yield. The stability of these
last pyridinium derivatives is highly improved. In con-
trast, despite considerable efforts including variations of
solvents (DMF, CH3CN) and temperature conditions,
neither the 1-bromo-2,3-propanediol nor the protected
1-bromo-2,3-O-acetylpropanediol reacted with 3 to give
13 or its diacetylated derivative (Table 2).
* Isolated para compounds 3-8 containing less than 5% of the
meta regioisomer.
1 and Table 1; meta and para refer to the position of the
carbonyl group that undergoes nucleophilic attack). The
desired and major product 3 was separated from the
minor one 3′ by selective recrystallization from acetone.
The regioisomer adduct 3 was unambiguously identified
because its spectroscopic data were identical to the ones
described previously for the hemiamidal 3 obtained by
an unequivocal synthetic approach.18
The versatility and efficiency of this method allowed
us to prepare other derivatives by using different aro-
matic Grignard reagents or lithio-aromatic nucleophiles
(see the Experimental Section). The reaction generally
proceeded in good yield to the formation of a mixture of
para (4-7) and meta (4′-7′) compounds, in favor of the
first ones (the ratio para/meta was always more than 2;
Table 1). Unfortunately, in the case of the arylation with
4-lithiopyridine, prepared as previously described from
4-bromopyridine,19 no addition product (8 and/or 8′) could
be detected, probably due to the instability of 4-bromopy-
All these pyridinium compounds were shown to exist
only under a cyclized hemiamidal form on the basis of
13C NMR analyses since the chemical shift of C7 is
characteristic of a tetrahedral carbon (around 88 ppm).
(19) Wibaut, J. P.; Heeringa, L. G. Recl. Trav. Chim. Pays-Bas Belg.
1955, 74, 1003-1020.
(20) Karig, G.; Spencer, J. A.; Gallagher, T. Org. Lett. 2001, 3, 835-
838.
10504 J. Org. Chem., Vol. 70, No. 25, 2005

Toward Isoniazid-Derived InhA Inhibitors
TABLE 2. Pyridinium Compounds (See General
Formula in Scheme 2)
ditional chiral center in the chain). The most widely used
reagent for selective 1,4-reduction of pyridinium salts is
sodium dithionite24,25 but by treating 10 with this reduc-
tant in the presence of sodium carbonate and in hetero-
geneous water/dichloromethane solution, we could only
obtained an unstable product which was not the expected
1,4-dihydropyridine/hemiamidal derivative 21 (Scheme
3). Attempts to realize the reduction step by electrochem-
istry failed also (only unknown compounds were detected
by NMR) and reduction methods involving hydride
transfer from transition-metal hydride are not yet pre-
parative methods.26-28 As expected, the use of a hydride
reagent like NaBH4 provided in majority the 1,2-DHP
isomer 18 (85% yield, Scheme 3) which was characterized
by 13C NMR as an exclusively cyclized structure (C7
carbon signal at 87.6 ppm). Only traces of the target 1,4-
DHP compound could be obtained in these conditions.29
Thus, we tried two other reducing agents derived from
NaBH4: NaBH3CN and NaBH(OAc)3. The former led to
the formation of tetrahydropyridines (identified by mass
spectrometry and 1H NMR) as major products, even when
used in default. The latter, which is a weaker reducing
agent,30 led to the formation of a 80:20 mixture of 1,4-
DHP 21 and 1,2-DHP derivatives in 80% yield. The
regioselectivity can be improved up to more than 95% in
favor of the desired 1,4-DHP isomer versus the 1,2-DHP
isomer by adding NaBH(OAc)3 in solution in acetic acid.
In this way, other intermediate reducing species are
probably created, as proposed by Nutaitis.30 The 1,4-DHP
compound 21 formed initially had a cyclized structure
as the starting material but, interestingly, it was then
slowly and spontaneously converted into the open struc-
ture 23. This conversion can be efficiently base-catalyzed
1
(NaHCO3 or NH3). The following H and 13C NMR data
supported a cyclized structure for 21 and an opened one
for 23: H4 resonance is deshielded in 23 (δ ) 5.05 ppm)
with respect to 21 (δ ) 3.79 ppm), NH2 resonance is a
broad singlet (δ ) 6.67 ppm) in 23 whereas NH and OH
are two distinct singlets (δ ) 8.03 and 6.35 ppm,
respectively) in 21, and finally, the chemical shift of C7
is characteristic of a carbonyl group for 23 (δ ) 198.4
ppm) and of a tetrahedral carbon (δ ) 92.5 ppm) for 21.
In addition, only one set of signals was obtained in the
1H and 13C spectra of the cyclized structure 21. These
results indicate that the reduction of the pyridinium ring
was completely stereoselective, giving only one pair of
enantiomers. To confirm the stereoselectivity and to find
the relative configuration of the two carbons C4 and C7,
we have realized NOE coupling experiments, which
showed a spatial proximity between H4 and the hydroxy-
It is interesting to note that, as expected, only compounds
11 and 12 are present as two pairs of diastereoisomers
since they have additional asymmetric(s) carbon(s) in the
alkyl chain. This feature results in a splitting of 1H and
13C NMR signals. Such a splitting of NMR signals has
been previously observed in the case of oxidized INH-
NAD adduct7 and interpreted as a consequence of in-
tramolecular stacking of the two aromatic moieties
present in the opened form of the INH-NAD adduct. In
the present case, stacking interactions cannot explain the
NMR data, and only the cyclization process with creation
of a second chiral center at C7 and formation of two pairs
of diastereoisomers can explain the splitting of the NMR
signals for 11 and 12.16
Syntheses of Dihydropyridine Derivatives (DHP).
One main difficulty in the route to the BH-NAD adduct
of dihydropyridine-type appeared to be the regioselective
reduction of the pyridinium ring. Indeed, few reagents
are available to achieve 1,4-reduction, and moreover, 1,4-
dihydropyrine products are known to be rather unstable.
In particular, they are sensitive to air oxidation21 and to
protonation under acidic conditions.22,23
(21) Eisner, U.; Kuthan, J. Chem. Rev. 1972, 1.
(22) Hentall, P. L.; Flowers, N.; Bugg, T. D. H. Chem. Commun.
2001, 2098-2099.
(23) Wong, C. H.; Whitesides, G. M. J. Am. Chem. Soc. 1981, 103,
4890.
(24) Mauzerall, D.; Westheimer, F. H. J. Am. Chem. Soc. 1955, 70,
2261.
(25) Carelli, V.; Liberatore, F.; Scipione, L.; Musio, R.; Sciacovelli,
O. Tetrahedron Lett. 2000, 41, 1235-1240.
(26) Lo, H. C.; Buriez, O.; Kerr, J. B.; Fish, R. H. Angew. Chem.,
Int. Ed. 1999, 38, 1429-1432;
(27) Konno, H.; Sakamoto, K.; Ishitani, O. Angew. Chem., Int. Ed.
2000, 39, 4061-4063;
(28) Steckhan, E.; Herrmann, S.; Ruppert, R.; Dietz, E.; Frede, M.;
Spika, E. Organometallics 1991, 10, 1568-1577.
(29) Bu¨chi, G.; Coffen, D. L.; Kocsis, K.; Sonnet, P. E.; Ziegler, F. E.
J. Am. Chem. Soc. 1966, 88, 3099-3109 and references therein.
(30) Nutaitis, C. F. J. Chem. Educ. 1989, 66, 673-675.
We first chose to study the reduction of 10 which is
simple and quite easy for preparation (it has no ad-
J. Org. Chem, Vol. 70, No. 25, 2005 10505

Broussy et al.
SCHEME 3. Reduction of Pyridinium Derivatives 10 and 12 by Sodium Borohydride and Related
Reducing Agents
lic proton (7-OH) and not between H4 and the phenyl
protons. Thus, we concluded that the phenyl ring and
the dihydropyridine nucleus were on the same side of the
pyrrolidinone ring.
the purification process (column chromatography on silica
gel). The use of guanidinium/sodium methylate,33 a more
potent deprotecting system for acetates, allowed to
shorten the reaction time (less than 10 min) and thus to
improve the yield of 1 (25%, two epimers). In the case of
the 1,2-DHP derivative 19, the NH3/MeOH standard
method was employed to give the 1,2-DHP ribonucleoside
20 (two epimers) in an exclusively cyclized structure
This synthetic pathway was then employed to synthe-
size the compounds bearing a ribose substituent. The
introduction of the ribose moiety was based on a protocol
described for the synthesis of nicotinamide mononucleo-
side which goes through a brominated intermediate.31,32
It required the use of acetate protecting group for the
following reasons: this protecting group is readily remov-
able in the presence of the reactive pyridinium function,
it is stable to the bromination conditions and it is able
to provide neighboring group participation during the
glycosylation step for a high â-diastereoselectivity.31 We
first considered to deprotect the ribose of 12 prior to the
reduction step, because of the instability of the DHP
compounds. However, the NH3/MeOH conditions used for
this deprotection (3 equiv of NH3 1 M in MeOH, -3 to
-5 °C for 20 h) gave only traces of the fully deprotected
compound (the product was isolated and identified by
TLC and mass spectrometry: ESI, positive mode, m/z 359
(M+)) and a very large amount of the N-glycosyl bond
cleavage product 3. We then tried to avoid this side
reaction by reducing the pyridinium prior to deprotection.
The reduction step with NaBH(OAc)3 was complicated
by partial cleavage of the N-glycosidic bond (regeneration
of hemiamidal 3) and formation of undesired 1,2-DHP
isomer (less than 10%). However, the crude product could
be characterized by NMR as the cyclized structure 22,
but it could not be isolated in pure form since the
purification step by silica gel column chromatography
gave a substantial amount of the open product 24. Since
this one is the stable isomer, it was isolated and fully
characterized. Further deprotection in standard condi-
tions with NH3 in MeOH for 4 h gave the desired product
1 in poor yield (10%), due to the instability of the DHP
compound in solution at ambient temperature and during
1
(checked by H and 13C NMR).
From a structural point of view concerning the possible
cyclization involving the 4-aroyl and 3-carboxamide sub-
stituents of the pyridine nucleus to give cyclized hemi-
amidals, we can conclude that such process is depending
on several factors (Table 3): (i) 4-benzoylnicotinamide
and various derivatives substituted on the phenyl ring
(column 2) are stable in cyclized hemiamidal structure
(unfortunately, we did not succeed up to now to prepare
the 4-isonicotinoylnicotinamide to check the influence of
the replacement of the phenyl ring by a pyridine nucleus);
(ii) compounds of pyridinium type (column 3) exist as
unique cyclized hemiamidal structure both in the case
of 4-isonicotinoyl- and of 4-benzoyl-substituted com-
pounds; (iii) in the case of 1,2-dihydropyridine adducts
(column 4), the cyclized hemiamidal was the only struc-
ture observed, possibly by stabilization of the pyrrolinone
ring formed when a 3,4-double bond is present; (iv) in
the case of 1,4-dihydropyridine adducts (column 5), the
nature of the 4-aroyl substituent plays an important
role: in INH-NAD adducts (4-isonicotinoyl substituent),
a dynamic equilibrium exists between the opened form
(minor form) and the cyclized hemiamidal structure
(main form),8,15 whereas in BH-NAD models (4-benzoyl
substituent), the cyclized hemiamidal structure initially
present is progressively converted to 100% of opened
form.
InhA and MabA Activity Assays in the Presence
of the INH-NAD Analogues. Based on the scaffold of
the isoniazid-NAD adduct, we have synthesized 4-ben-
zoyl-1,4-dihydronicotinamide ribonucleoside and a series
(31) Lee, J.; Churchil, H.; Choi, W.-B.; Lynch, J. E.; Roberts, F. E.;
Volante, R. P.; Reider, P. J. Chem. Commun. 1999, 729-730.
(32) Mickhailopulo, I. A.; Pricota, T. I.; Timoshchuk, V. A.; Akhrem,
A. A. Synthesis 1981, 388-389.
(33) Adapted from Ellerwik, U.; Magnusson, G. Tetrahedron Lett.
1997, 38, 1627-1628.
10506 J. Org. Chem., Vol. 70, No. 25, 2005

Toward Isoniazid-Derived InhA Inhibitors
TABLE 3. Opened Structure or Cyclized Hemiamidal Structure Depending on Structural Features
of related compounds. Since the Kd value of NADH for
InhA is 2 µM34 while the Ki estimated (and debated)
values for the INH-NAD adduct vary from ∼1 nM14,35
to 100 nM,7 binding of the 4-isonicotinoyl fragment to
NAD, and probably also binding of the benzoyl frag-
ment (the adduct BH-NAD has also been reported to
give inhibition of InhA14) could be responsible for an
increase of affinity for InhA of about 20-2000-fold
between the natural cofactor NADH and the INH-NAD
inhibitor (and probably also the BH-NAD inhibitor). Our
hopes were that truncated analogues of these inhibitors
still retain sufficient affinity to inhibit the InhA or/and
MabA targets.
All of the synthesized analogues were tested as poten-
tial inhibitors of InhA and MabA enzymes using proce-
dures previously described.9,36 A pool of INH-NAD
adducts prepared by oxidation of INH with manganese
pyrophosphate in the presence of NAD+ was used as
reference36 and displayed high inhibition of InhA activity
(50% inhibition at 10 nM). Products 5 and 17 behave like
InhA activators (activity was increased by a factor around
2.5), but at the moment, no evident explanation could be
proposed. All the other BH-NAD models tested at 100
µM, including the 4-benzoyl-1,4-dihydronicotinamide ri-
bonucleoside 1, did not significantly affect InhA nor
MabA activities. These results showed that the lack of
ADP moiety found in the NAD adduct is a critical point
and that simple addition of the benzoyl radical on
truncated adducts is not sufficient to compensate for this
absence. So we must consider these compounds only as
the first of a series of compounds that need further
structural modulation to improve their interaction with
the target(s) and to give access to potential inhibitors.
As example, especially in the case of the simplest
molecule 3, the replacement of the phenyl group by a
pyridyl group (like in the INH-NAD inhibitor), or/and
the addition of a hydrophobic substituent able to interact
with the neighboring site of the enoyl acyl substrate in
InhA might considerably increase the affinity and con-
sequently the inhibiting activity. Such syntheses are
currently in progress in our group, as also the full
chemical synthesis of the biochemical adduct BH-NAD.
Conclusion
In the context of reemergence of tuberculosis during
the past two decades, there is an urgent need to develop
new and better treatments, particularly for tuberculosis
infections resistant to traditional antibiotics. Variations
of the existing TB drugs or design of drugs acting at new
targets would substantially improve the actual TB control
programs.36-43 In our search of potential inhibitors of
InhA and MabA derived from the isoniazid-NAD adduct,
we achieved synthesis of the 4-benzoyl-1,4-dihydronico-
tinamide ribonucleoside 1, in only four steps by a simple
and versatile method using 3,4-pyridinedicarboximide as
starting material. To our knowledge, no other synthetic
pathway towards this original structure and some other
related ones has been previously developed (without
using the expensive NAD+ or NADH as part of the
(37) Velezheva, V. S.; Brennan, P. J.; Marshakov, V. Y.; Gusev, D.
V.; Lisichkina, I. N.; Pergudov, A. S.; Tchernousova, L. N.; Smirnova,
T. G.; Andreevskaya, S. N.; Medvedev, A. E. J. Med. Chem. 2004, 47,
3455-3461.
(38) Pasqualoto, K. F. M.; Ferreira, E. I.; Santos-Filho, O. A.;
Hopfinger, A. J. J. Med. Chem. 2004, 47, 3755-3764.
(39) Sacchettini, J. InhA crystals and three-dimensional structure.
United States Patent 5,882,878, March 16, 1999.
(40) Staveski, M. M.; Sneddon, S. F.; Yee, C.; Janjigian, A. (Applicant
Genzyme corp.). InhA inhibitors and methods of use thereof. PCT/
US01/40045, WO 01/56974 A2, August 9, 2001.
(41) Yatvin, M. B.; Pederson, R. L. (Applicant Enzrel, Inc.). Anti-
mycobacterial compounds. PCT/US01/21640, WO 02/04478 A1, Janu-
ary 17, 2002.
(34) Que´mard, A.; Sacchettini, J. C.; Dessen, A.; Vilcheze, C.;
Bittman, R.; Jacobs, W. R., Jr.; Blanchard, J. S. Biochemistry 1995,
34, 8235-8241.
(42) Que´mard, A.; Daffe´, M.; Ducasse, S.; Marrakchi, H.; Labesse,
G.; Cohen-Gonsaud, M.; Doughet, D. (Applicants CNRS/University
Montpellier 1/INSERM). Use of the protein MabA (FABG1) of Myco-
bacterium tuberculosis for designing and screening antibiotics. FP 02
04018, March 29, 2002; WO 03/082911, October 9, 2003.
(43) Andries, K.; Verhasselt, P.; Guillemont, J. et al. Science 2005,
307, 223-227.
(35) Lei, B.; Wei, C.-J.; Tu, S.-C. J. Biol. Chem. 2000, 275, 2520-
2526.
(36) Nguyen, M.; Que´mard, A.; Broussy, S.; Bernadou, J.; Meunier,
B. Antimicrob. Agents Chemother. 2002, 46, 2137-2144.
J. Org. Chem, Vol. 70, No. 25, 2005 10507

Broussy et al.
1
1701, 1613, 1084, 703; H NMR (400 MHz, DMSO-d6) δ 9.17
(s, 1H), 8.76 (d, J ) 5.2 Hz, 1H), 8.60 (s, 1H), 7.64 (d, J ) 4.8
Hz, 1H), 7.01-7.11 (m, 5H), 6.70 (s, 1H), 3.41 (d, J ) 13.6 Hz,
1H), 3.22 (d, J ) 13.2 Hz, 1H); 13C NMR (100 MHz, DMSO-d6,
attributions from HMQC-GS and HMQC-GS LR experiments)
δ 167.1 (Cq, C9), 157.2 (C4), 153.1 (C2), 144.9 (C6), 135.8 (C11),
131.3 (2CH, C12 and C16), 128.4 (Cq, C13 and C15), 128.3
(C3), 127.3 (CH, C14), 119.0 (CH, C5), 88.6 (Cq, C7), 44.5 (CH2,
C10); MS (EI) m/z 222 (M+ - H2O).
molecule). Although none of these compounds showed
significant inhibition of InhA and MabA enzymes, the
core structure of BH-NAD adduct can be seen as an
attractive target to develop new synthetic analogues as
possible antitubercular agents and also as a key inter-
mediate in the total synthesis of the bioactive BH-NAD
adduct. We are currently working to improve binding
affinity and selectivity of these analogues of the iso-
niazid-NAD adduct.
Preparation of Compounds 9, 14, and 16. To a stirred
solution of hemiamidal 3, 4, or 5 (0.5 mmol) in dry THF (20-
100 mL, 3 and 4) or THF/ether (50 mL, 5) was added
2-(acetoxyethoxy)methyl bromide (2 equiv)15,44 was added
dropwise at rt under argon. After 30 min, the white precipitate
(cases of 9 and 14) was filtered, washed with THF and ether,
and dried under vacuum to give 9 in 80% yield or 14 in 49%
yield. In the case of 16, spontaneous precipitation did not occur,
and only after addition of 10 mL ether could the gel/precipitate
be filtered, washed with ether, centrifuged (ether), and quickly
desiccated under vacuum to give 16 in 45% yield.
5-[(2-Acetoxyethoxy)methyl]-1-hydroxy-3-oxo-1-phen-
yl-1,2-dihydro-3H-pyrrolo[3,4-c]pyridinium Bromide (9):
IR (KBr, νmax/cm-1) 3003 (broad), 1735, 1726, 1647, 1254, 1053;
1H NMR (300 MHz, DMSO-d6) δ 10.27 (s, 1H), 9.59 (s, 1H),
9.28 (d, J ) 6.0 Hz, 1H), 8.28 (d, J ) 6.0 Hz, 1H), 7.78 (s,
1H), 7.59 (m, 2H), 7.44 (m, 3H), 6.08 (s, 2H), 4.13 (m, 2H),
3.89 (m, 2H), 1.96 (s, 3H); 13C NMR (75 MHz, DMSO-d6) δ
171.1 (Cq), 167.0 (Cq), 164.0 (Cq), 148.3 (CH), 140.9 (CH),
139.0 (Cq), 130.8 (Cq), 130.1 (CH), 129.6 (2CH), 126.8 (2CH),
122.9 (CH), 89.3 (CH2), 88.5 (Cq), 69.5 (CH2), 63.3 (CH2), 21.5
(CH3); MS (ESI) m/z 343 (M+), 327, 227.
Experimental Section
Preparation of Compounds 3-7. To a solution of 3,4-
pyridinedicarboximide 2 (2.0 g, 13.5 mmol) in dry THF (100
mL), under argon at -80 °C, was added phenyllithium (2.0 M
in dibutyl ether, 2 equiv) dropwise, and the mixture was
stirred at -80 °C for 1 h and then allowed to warm to room
temperature over 1 h. Water was added (50 mL), and the
product was extracted with ethyl acetate (6 × 100 mL). The
combined organic phase was dried over Na2SO4, and the
solvent was evaporated to give a yellow solid, which was
recrystallized in acetone to give hemiamidal 3 as a white solid,
yield 47%. Changing phenyllithium by 4-phenoxyphenylmag-
nesium bromide (0.5 M in THF) or benzylmagnesium bromide
(20% in THF) allowed us to obtain 6 in 29% yield (as a white
solid after recrystallization in acetone) or 7 in 60% yield (as a
white solid 85:15 mixture of para and meta isomers after
purification by silica gel chromatography with dichloromethane/
MeOH 97:3 as eluent), respectively. Compound 4 or 5 was
prepared in the same way by addition to a dry THF solution
of 3,4-pyridinedicarboximide 2 (2.0 g, 13.5 mmol in 50 mL) of
a solution obtained by dropwise addition of butyllithium (2.0
M in cyclohexane, 2.5 equiv) to a solution of 1-bromo-3,4-
methylenedioxybenzene or 1-bromo-4-butylbenzene in dry THF
(50 mL), under argon at -80 °C, and the mixture was stirred
at -80 °C for 30 min. The following steps are the same as for
compounds 3, 6, and 7 and afforded 4 in 26% yield (as a yellow
solid after rectristallization in ethanol) and 5 in 42% yield (as
a white powder after purification by silica gel chromatography
with dichloromethane/MeOH 97:3 as eluent or recrystallization
in acetone).
3
3
5-[(Ethoxycarbonyl)methyl]-1-hydroxy-3-oxo-1-phen-
yl-1,2-dihydro-3H-pyrrolo[3,4-c]pyridinium Bromide (10).
This compound was prepared from hemiamidal 3 as previously
reported, and all analytical data obtained were identical to
those previously described.18
5-[1-(Ethoxycarbonyl)ethyl]-1-hydroxy-3-oxo-1-phen-
yl-1,2-dihydro-3H-pyrrolo[3,4-c]pyridinium Bromide (11).
To a refluxed solution of hemiamidal 3 (180 mg, 0.66 mmol)
in dry THF (30 mL) was added racemic ethyl 2-bromopropi-
onate (2.40 g, 133 mmol) by aliquots (6 × 0.4 g) every 20 h.
After 120 h of reaction, the white precipitate was filtered,
washed three times with THF and ether, and dried under
vacuum to give 200 mg of 11: yield 78% (mixture of two
racemic diastereoisomers); 1H NMR (300 MHz, DMSO-d6) δ
1-Hydroxy-1-phenyl-1,2-dihydro-3H-pyrrolo[3,4-c]py-
ridin-3-one (3). Analytical data obtained were identical to
those previously described.18
3
10.28 (s, 1H), 9.69 and 9.66 (2s, 1H), 9.39 and 9.35 (2d, J )
1-(Benzo-1,2-dioxol-4-yl)-1-hydroxy-1,2-dihydro-3H-pyr-
3
rolo[3,4-c]pyridin-3-one (4): IR (KBr, νmax/cm-1) 3314, 3086,
6 Hz, 1H), 8.32 and 8.30 (2d, J ) 6 Hz, 1H), 7.78 (2s, 1H),
3
1
7.60 (m, 2H), 7.44 (m, 3H), 6.14 (q, J ) 6 Hz, 1H), 4.22 (2q,
1689, 1607, 1450, 1256, 1040; H NMR (250 MHz, DMSO-d6)
3J ) 6 Hz, 2H), 1.95 and 1.93 (2d, 3J ) 6 Hz, 3H), 1.23 (2t, 3J
) 9 Hz, 3H); 13C NMR (75 MHz, DMSO-d6) δ 169.0 (Cq), 166.6
(Cq), 164.0 and 163.9 (Cq), 150.4 and 149.8 (CH), 143.1 and
142.5 (CH), 139.0 (Cq), 130.9 and 130.6 (Cq), 130.1 (CH), 129.6
(2CH), 126.8 (2CH), 122.8 and 122.6 (CH), 88.5 and 88.4 (Cq),
68.6 (CH), 63.5 (CH2), 18.4 (CH3), 14.7 (CH3); MS (ESI) m/z
327 (M+).
δ 9.46 (s, 1H), 8.85 (s, 1H), 8.71 (d, J ) 5.1 Hz, 1H), 7.42 (d, J
) 5.0 Hz, 1H), 7.16 (s, 1H), 7.03 (s, 1H), 6.90 (m, 2H), 6.01 (s,
2H); MS (EI) m/z 270 (M+), 253.
1-(4-Butylphenyl)-1-hydroxy-1,2-dihydro-3H-pyrrolo-
[3,4-c]pyridin-3-one (5): mp 177-178 °C; IR (KBr, νmax/cm-1
)
3300, 3051, 2951, 2922, 1694, 1606, 1432, 1293, 1083; 1H NMR
(300 MHz, DMSO-d6) δ 9.45 (s, 1H), 8.86 (s, 1H), 8.71 (d, J )
4.8 Hz, 1H), 7.39 (d, J ) 4.8 Hz, 1H; d, J ) 8.4 Hz, 2H), 7.18
(d, J ) 8.4 Hz, 2H), 7.11 (s, 1H), 2.55 (t, J ) 7.5 Hz, 2H), 1.51
(quintuplet, J ) 7.5 Hz, 2H), 1.27 (sextuplet, J ) 7.3 Hz, 2H),
0.87 (t, J ) 7.2 Hz, 3H); 13C NMR (75 MHz, DMSO d6) δ 167.9
(Cq), 159.5 (Cq), 153.8 (CH), 145.4 (CH), 143.3 (Cq), 138.7 (Cq),
129.2 (2CH), 127.0 (Cq), 126.3 (2CH), 118.7 (CH), 88.1 (Cq),
35.3 (CH2), 33.9 (CH2), 22.6 (CH2), 14.9 (CH3); MS (EI) m/z
282 (M+), 265 (M+ - OH), 161, 149.
5-(Tri-O-acetyl-â-D-ribofuranose)-1-hydroxy-3-oxo-1-
phenyl-1,2-dihydro-3H-pyrrolo[3,4-c]pyridinium Bro-
mide (12). To a solution of tetra-O-acetyl-â-D-ribofuranose
(1.13 g, 3.54 mmol) in dichloromethane (20 mL) at 0 °C was
added dropwise hydrogen bromide 30 wt % in acetic acid (1.12
mL, 5.31 mmol). After 10 min, the solvent was evaporated,
and the residual acetic acid was coevaporated with toluene (5
mL). The orange oily residue was dried under vaccum (1 h),
dissolved in dry acetonitrile (40 mL), and added to a suspen-
sion of hemiamidal 3 (400 mg, 1.77 mmol) in dry acetonitrile
(160 mL). The mixture was stirred for 4 h at room tempera-
ture, under argon and in the dark. Afterward, the solvent was
evaporated, and the oily residue was dissolved in dichlo-
romethane (low volume), precipitated with ether,and filtered
1-Hydroxy-1-(4-phenoxyphenyl)-1,2-dihydro-3H-pyr-
rolo[3,4-c]pyridin-3-one (6): IR (KBr, νmax/cm-1) 3306, 3165,
1
3068, 1697, 1609, 1591, 1254; H NMR (250 MHz, DMSO-d6)
δ 9.51 (s, 1H), 8.86 (d, J ) 0.8 Hz, 1H), 8.71 (d, J ) 5.0 Hz,
1H), 7.49 (m, 2H), 7.42 (dd, J ) 4.9 Hz and J ) 1.0 Hz, 1H),
7.39 (m, 2H), 7.20 (s, 1H), 7.15 (t, J ) 7.4 Hz, 1H), 6.99 (m,
4H); MS (EI) m/z 318 (M+), 301 (M+ - OH), 197.
1-Hydroxy-1-benzyl-1,2-dihydro-3H-pyrrolo[3,4-c]pyri-
(44) Robins, M. J.; Hatfield, P. W. Can. J. Chem. 1982, 60, 547-
553.
din-3-one (7): mp 233 °C; IR (KBr, νmax/cm-1) 3425 (broad),
10508 J. Org. Chem., Vol. 70, No. 25, 2005

Toward Isoniazid-Derived InhA Inhibitors
to give 900 mg of a white powder. This crude product could be
used directly for the next step or purified by column chroma-
tography on silica gel (dichloromethane/MeOH 98:2 to 95:5)
to give 500 mg of 12 as a white powder: yield 50% (two
epimers); IR (KBr, νmax/cm-1) 3392, 3135,1734, 1652, 1232,
1062; 1H NMR (300 MHz, DMSO-d6, a splitting of most of the
signals was observed because of the presence of two epimers)
δ 10.34 (s, 1H), 9.53 and 9.48 (s, 1H), 9.28 and 9.26 (d, J ) 6.0
Hz, 1H), 9.36 and 9.32 (d, J ) 6.0 Hz, 1H), 7.83 and 7.80 (s,
1H), 7.61 and 7.44 (m, 5H), 6.74 (s, 1H), 5.64 (m, 1H), 5.45
and 5.41 (d, J ) 6.0 Hz, 1H), 4.67 (m, 1H), 4.46 (m, 2H), 2.16
(m, 9H); 13C NMR (75 MHz, DMSO-d6) δ 170.9 (Cq), 170.4 (Cq),
170.1 (Cq), 167.5 (Cq), 163.9 (Cq), 153.8 (CH), 145.4 (CH),
139.0 (Cq), 131.1 (Cq), 130.1 (CH), 129.6 (2CH), 126.9 (2CH),
123.2 (CH), 98.4 (CH), 88.5 (Cq), 82.4 (CH), 76.0 (CH), 69.0
(CH), 62.8 (CH2), 21.4, 21.3, 21.1 (3CH3); MS (ESI) m/z 485
(M+); HRMS (FAB+) found 485.1557, calcd 485.1560.
5-[(2-Acetoxyethoxy)methyl]-1-(benzo-1,2-dioxol-4-yl)-
1-hydroxy-3-oxo-1,2-dihydro-3H-pyrrolo[3,4-c]pyri-
dinium bromide (14): IR (KBr, νmax/cm-1) 3393 (broad), 3138,
1725, 1648, 1250, 1034; 1H NMR (300 MHz, DMSO-d6) δ 10.21
(s, 1H), 9.55 (s, 1H), 9.24 (d, 3J ) 6.0 Hz, 1H), 8.29 (d, 3J ) 6.0
Hz, 1H), 7.73 (s, 1H), 7.08 (s, 1H), 7.05 and 6.95 (2d, J ) 6
Hz, 2H), 6.05 (2s, 4H), 4.15 (m, 2H), 3.87 (m, 2H), 1.97 (s, 3H);
13C NMR (75 MHz, DMSO-d6) 171.1 (Cq; weak signal), 167.0
(Cq), 164.0 (Cq), 148.9 (Cq), 148.5 (Cq), 148.2 (CH), 141.9 (CH),
132.7 (Cq), 130.8 (Cq), 122.8 (CH), 120.5 (CH), 109.0 (CH),
107.5 (CH), 102.4 (CH2), 89.3 (CH2), 88.3 (Cq), 69.5 (CH2), 63.3
(CH2), 21.4 (CH3); MS (DCI) m/z 387 (M+), 371, 369, 271.
136.2 (Cq), 130.5 (Cq), 129.6 (2CH), 126.7 (2CH), 122.6 (CH),
88.5 (Cq), 63.3 and 61.3 (2CH2), 35.3, 33.9 and 22.6 (3CH2),
14.8 and 14.6 (2CH3); MS (DCI) m/z 369 (M+), 353, 283
(hemiamidal + H+).
Ethyl (1-Hydroxy-3-oxo-1-phenyl-1,2,4,5-tetrahydro-
3H-pyrrolo[3,4-c]pyridin-5-yl)acetate (18). To a solution
of 10 (50 mg, 0.13 mmol) in acetonitrile (10 mL) was added
sodium borohydride (5 mg, 0.13 mmol) with cooling in an ice-
water bath. After 10 min at 0 °C, water was added, the product
was extracted with dichloromethane and dried over Na2SO4,
and the solvent was evaporated under vacuum. The product
was precipitated from dichoromethane/hexane (9:1) to give 35
mg of 18 as a yellow powder: yield 85%; IR (KBr, νmax/cm-1
)
3370, 3210, 1736, 1674, 1548, 1193; 1H NMR (300 MHz,
DMSO-d6) δ 8.23 (s, 1H), 7.30 (m, 5H), 6.41 (s, 1H), 6.39 (d, J
) 6.9 Hz, 1H), 4.45 (d, J ) 7.2 Hz, 1H), 4.22 (s, 2H), 4.13 (q,
J ) 7.1 Hz, 2H), 3.88 (s, 1H), 3.91 (s, 1H), 1.21 (t, J ) 7.1 Hz,
3H); 13C NMR (300 MHz, DMSO-d6) δ 171.6 (Cq), 170.3 (Cq),
156.6 (Cq), 142.0 (Cq), 110.5 (Cq), 145.6 (CH), 128.9 (2CH),
128.4 (CH), 126.3 (2CH), 87.7 (CH), 87.6 (Cq), 61.5 (CH2), 55.9
(CH2), 46.3 (CH2), 15.0 (CH3); MS (DCI/NH3) m/z 315 (M +
H+).
1-Hydroxy-3-oxo-1-phenyl-5-(2′,3′,5′-tri-O-acetyl-â-D-ri-
bofuranosyl)-1,2,4,5-tetrahydro-3H-pyrrolo[3,4-c]pyri-
dine (19). A solution of crude 12 (0.53 mmol, 300 mg) in EtOH
(20 mL) at 0 °C was incubated with sodium borohydride (0.53
mmol, 20 mg) for 10 min at 0 °C. Then water was added, the
product was extracted with dichloromethane and dried over
Na2SO4, and the solvent was evaporated under vacuum. The
product was purified by two successive column chromatogra-
phy on silica gel (dichloromethane/MeOH, 100:0 to 97:3 and
ethyl acetate) to give 129 mg of 19 as a bright yellow powder
(50% yield for two steps): 1H NMR (300 MHz, DMSO-d6; a
splitting of some signals was observed because of the presence
of two epimers) δ 8.38 (s, 1H), 7.38 (m, 5H), 6.54 (2s, 1H), 6.46
(d, J ) 6.3 Hz, 1H), 5.33 (m, 1H), 5.14 (m, 2H), 4.58 (2d, J )
6.9 Hz, 1H), 4.26 and 4.16 (2m, 5H), 2.07 (m, 9H); 13C NMR
(75 MHz, DMSO-d6), δ 171.3 (Cq), 170.4 (Cq), 170.3 (Cq), 170.2
(Cq), 156.0 (Cq), 143.7 (CH), 141.5 (Cq), 129.0 (CH), 128.5
(CH), 112.8 (Cq), 95.2 (CH), 90.1 (CH), 87.6 (Cq), 78.6 (CH),
71.3 (CH), 68.3 (CH), 64.2 (CH2), 39.8 (CH2), 21.4, 21.3 and
21.2 (3CH3); MS (ESI, positive mode) m/z 487 (M + H+), 509
(M + Na+), 525 (M + K+).
1-Hydroxy-3-oxo-1-phenyl-5-â-D-ribofuranosyl-1,2,4,5-
tetrahydro-3H-pyrrolo[3,4-c]pyridine (20). To 30 mg (0.053
mmol) of 19 was added NH3 in MeOH 7 M (3 mL). After 3 h
at rt, the solvent was concentrated in a vacuum and the
product was precipitated from MeOH/ether (9:1). The precipi-
tate was chromatographed on silica gel (dichloromethane/
MeOH, 90:0 to 85:20) to give 6 mg of 20 (31% yield). The
product can be precipitated from methanol/ethyl acetate to give
a bright yellow powder: IR (KBr, νmax/cm-1) 3376 (broad), 1669,
1600, 1198, 1046; UV (λmax, ꢀ) 398 nm, 4600 M-1 cm-1; 1H NMR
(300 MHz, DMSO-d6; two epimers) δ 8.29 and 8.28 (2s, 1H),
7.37 (m, 5H), 6.51 (d, J ) 7.2 Hz, 1H), 6.43 and 6.42 (2s, 1H),
5.10 (2d, J ) 6.0 Hz, 1H), 4.99 (2d, J ) 4.7 Hz, 1H), 4.85 (2d,
J ) 5.3 Hz, 1H), 4.67 (2d, J ) 6.4 Hz, 1H), 4.48 (2d, J ) 7.2
Hz, 1H), 4.24 (2s, 2H), 3.81 (m, 1H), 3.66 (m, 1H), 3.56 to 3.42
(m, 3H); 13C NMR (75 MHz, DMSO-d6) δ 171.6 (Cq), 156.9 and
157.0 (Cq), 144.6 and 144.7 (CH), 141.9 (Cq), 128.9 (2CH),
128.4 (CH), 126.4 (2CH), 111.6 and 111.5 (Cq), 97.6 (CH), 88.5
(CH), 87.5 (Cq), 84.4 (CH), 71.3 and 71.2 (CH), 69.5 and 69.4
(CH), 62.5 (CH2), 39.8 (CH2); MS (ESI, positive mode) m/z 361
(M + H+), 383 (M + Na+), 399 (M + K+).
1-(Benzo-1,2-dioxol-4-yl)-5-[(ethoxycarbonyl)methyl]-
1-hydroxy-3-oxo-1,2-dihydro-3H-pyrrolo[3,4-c]pyri-
dinium Bromide (15). To a refluxed solution of hemiamidal
4 (100 mg, 0.37 mmol) in THF (100 mL) was added ethyl
bromoacetate (164 µL, 1.48 mmol). After 78 h, the white
precipitate was filtered, washed three times with THF and
ether, and dried under vacuum to give 74 mg of 15: yield 46%;
IR (KBr, νmax/cm-1) 3145, 3009, 1755, 1719 (and shoulder on
1
the left (ester)), 1654, 1488, 1242, 1037; H NMR (250 MHz,
3
DMSO-d6) δ 10.24 (s, 1H), 9.53 (s, 1H), 9.16 (d, J ) 6.1 Hz,
3
1H), 8.33 (d, J ) 6.2 Hz, 1H), 7.76 (s, 1H), 7.14 (s, 1H), 7.04
and 6.95 (2d, J ) 8.2 Hz, 2H), 6.05 (s, 2H), 5.69 (s, 2H), 4.24
3
3
(q, J ) 7.0 Hz, 2H), 1.26 (t, J ) 7.0 Hz, 3H); 13C NMR (75
MHz, DMSO d6) δ 167.1 (Cq), 166.5 (Cq), 163.8 (Cq), 151.2
(CH), 148.9 (Cq), 148.6 (Cq), 143.7 (CH), 132.6 (Cq), 130.4 (Cq),
122.5 (CH), 120.4 (CH), 109.1 (CH), 107.4 (CH), 102.4 (CH2),
88.4 (Cq), 63.3 and 61.2 (2CH2), 14.8 (CH3); MS (ESI) m/z 357
(M+); HRMS (FAB+) found 357.1085, calcd 357.1087.
5-[(2-Acetoxyethoxy)methyl]-1-(4-butylphenyl)-1-hy-
droxy-3-oxo-1,2-dihydro-3H-pyrrolo[3,4-c]pyridinium bro-
mide (16): IR (KBr, νmax/cm-1) 3152, 2957, 2930, 1728, 1652,
1249, 1064; 1H NMR (250 MHz, DMSO-d6) δ 10.26 (s, 1H),
3
3
9.57 (s, 1H), 9.22 (d, J ) 6.3 Hz, 1H), 8.26 (d, J ) 6.2 Hz,
1H), 7.72 (bs, 1H), 7.49 and 7.25 (2d, J ) 8.2 Hz, 4H), 6.02 (s,
2H), 4.15 (m, 2H), 3.87 (m, 2H), 2.58 (t, J ) 7.5 Hz, 2H), 1.95
(s, 3H). 1.52 (quintuplet, J ) 7.5 Hz, 2H), 1.27 (sextuplet, J )
7.4 Hz, 2H), 0.88 (t, J ) 7.2 Hz, 3H); MS (DCI) m/z 283
(hemiamidal + H+), no M+ detected.
1-(4-Butylphenyl)-5-[(ethoxycarbonyl)methyl]-1-hy-
droxy-3-oxo-1,2-dihydro-3H-pyrrolo[3,4-c]pyridinium Bro-
mide (17). To a refluxed solution of hemiamidal 5 (100 mg,
0.35 mmol) in THF/ether (1/1, 10 mL) was added ethyl
bromoacetate (58 µL, 0.53 mmol). After 20 h, the white
precipitate was filtered, washed two times with ether, and
dried under vacuum to give 100 mg of 22: yield 70%; IR (KBr,
νmax/cm-1) 3385, 3040, 2958, 2931, 1748, 1724, 1654, 1374,
Ethyl (1-Hydroxy-3-oxo-1-phenyl-1,2,5,7a-tetrahydro-
3H-pyrrolo[3,4-c]pyridin-5-yl)acetate (21) and Ethyl (4-
Benzoyl-3-carboxamido-1,4-dihydropyridin-1-yl)ace-
tate (23). To a solution of 10 (100 mg, 0.25 mmol) in EtOH
(10 mL) at 0 °C was added sodium triacetoxyborohydride (80
mg, 0.38 mmol) in solution in 3 mL of acetic acid. After 10
min at 0 °C, water was added, and the product was extracted
with dichloromethane. This organic layer was washed with
1
1215; H NMR (300 MHz, DMSO-d6) δ 10.28 (s, 1H), 9.55 (s,
3
3
1H), 9.14 (d, J ) 6.3 Hz, 1H), 8.30 (d, J ) 6.2 Hz, 1H), 7.75
(s, 1H), 7.47 and 7.26 (2d, J ) 8.3 Hz, 4H), 5.68 (s, 2H), 4.24
(q, 3J ) 7.0 Hz, 2H), 2.58 (t, J ) 7.5 Hz, 2H), 1.53 (quintuplet,
3
J ) 7.5 Hz, 2H), 1.27 (m, 2H), 1.26 (t, J ) 7.0 Hz, 3H), 0.88
(t, J ) 7.3 Hz, 3H); 13C NMR (75 MHz, DMSO-d6) δ 167.1 (Cq),
166.7 (Cq), 163.8 (Cq), 151.2 (CH), 144.5 (Cq), 143.7 (CH),
J. Org. Chem, Vol. 70, No. 25, 2005 10509

Broussy et al.
saturated NaHCO3 solution and dried over Na2SO4, and the
solvent was evaporated under vacuum. The product could be
precipitated from dichoromethane/hexane (9:1) to give 65 mg
of a pale orange powder, Yield 80% of cyclized 21, which was
slowly and spontaneously converted in 23. NH3 in MeOH (1
M) could be used to catalyze this conversion. Compound 21:
IR (KBr, νmax/cm-1) 3247 (very broad), 1742, 1687, 1662, 1606,
1219, 1164, 1044; 1H NMR (300 MHz, DMSO-d6) δ 8.03 (s, 1H),
7.31 (m, 4H), 7.25 (m, 1H), 6.70 (s, 1H), 6.35 (s, 1H), 5.67 (d,
J ) 7.8 Hz, 1H), 4.08 (q, J ) 7.1 Hz, 2H), 4.08 (s, 2H), 3.91
(dd, J ) 7.8 Hz and J ) 1.5 Hz, 1H), 3.79 (d, J ) 1.5 Hz, 1H),
1.17 (t, J ) 7.2 Hz, 3H); 13C NMR (75 MHz, DMSO-d6) δ 170.7
(Cq), 170.2 (Cq), 142.7 (Cq), 132.6 (CH), 131.7 (CH), 128.4
(2CH), 128.0 (CH), 127.2 (2CH), 103.7 (Cq), 98.6 (CH), 92.5
(Cq), 61.4 (CH2), 54.0 (CH2), 47.9 (CH), 14.9 (CH3); MS (ESI,
positive mode) m/z 337 (M + Na+), 353 (M + K+). Compound
23: IR (KBr, νmax/cm-1) 3344, 1738, 1683, 1652, 1596, 1393,
500 wet mesh, carboxylic acid, hydrogen form) was added.
After filtration and concentration by evaporation, the residue
was chromatographed on silica gel (dichloromethane/MeOH,
90:10 to 85:15) to give 22 mg of 1 (25% yield) The product can
be precipitated from methanol/ethyl acetate to give a pale
orange powder: IR (KBr, νmax/cm-1) 3360 (broad), 1683, 1648,
1596, 1221, 1097; UV (λmax, ꢀ) 318 nm, 4600 L mol-1 cm-1; 1H
NMR (300 MHz, DMSO-d6; two epimers) δ 7.98 (m, 2H), 7.63
(m, 1H), 7.53 (m, 2H), 7.37 and 7.33 (2s, 1H), 6.77 (bs, 2H),
6.33 and 6.29 (2d, J ) 8.1 Hz, 1H), 5.25 (bs, 1H), 5.12 and
5.07 (2d, J ) 4.5 Hz, 1H), 4.90 (bs, 1H), 4.69 (m, 2H), 4.09 (bs,
1H), 3.99 to 3.86 (m, 2H), 3.71 (m, 1H), 3.48 (m, 2H); 13C NMR
(75 MHz, DMSO-d6) δ 199.2 (Cq), 170.0 (Cq), 137.2 (CH), 135.7
(Cq), 134.2 (CH), 129.7 (2CH), 129.5 (2CH), 127.4 (CH), 103.5
and 103.2 (Cq), 102.3 and 101.7 (CH), 95.8 (CH), 84.7 (CH),
72.3 and 70.8 (2CH), 62.2 (CH2), 42.9 (CH); MS (ESI, positive
mode) m/z 361 (M + H+), 383 (M + Na+), 399 (M + K+).
Activity Assays for InhA and MabA. For InhA, the
preincubation reactions were performed in 80 µL (total volume)
of 100 mM sodium phosphate buffer solution, pH 7.5, at 25 °C
containing 19 nM InhA and 100 µM ligand compounds 1-24
(or 58 µM or 10 nM of the pool of INH-NAD adducts). After
5 min of preincubation, the addition of 112 µM 2-trans-
decenoyl-CoA and 100 µM NADH initiated the reaction.
Control reactions were carried out under the same conditions
as those described above without ligands. For MabA, the
preincubation reactions were performed in 80 µL (total volume)
of 100 mM sodium phosphate buffer solution, pH 7.0, at 25 °C
containing 3 µM MabA and 100 µM ligand compounds 1-24
(or 58 µM or 10 nM of the pool of INH-NAD adducts). After
5 min of preincubation, the addition of 350 µM acetoacyl-
coenzyme A (acetoacyl-CoA) and 100 µM NADPH to a final
volume of 1 mL initiated the reaction. Control reactions were
carried out under the same conditions as those described above
without ligands. All activity assays were performed in dupli-
cate.
1
1216; H NMR (250 MHz, DMSO-d6) δ 8.00 (m, 2H), 7.62 (m,
1H), 7.52 (m, 2H), 7.20 (s, 1H), 6.67 (s, 2H), 6.05 (d, J ) 7.8
Hz, 1H), 5.05 (d, J ) 4.6 Hz, 1H), 4.64 (dd, J ) 7.8 Hz and J
) 4.6 Hz, 1H), 4.12 (s, 2H), 4.12 (q, J ) 7.2 Hz, 2H), 1.20 (t, J
) 7.2 Hz, 3H); 13C NMR (75 MHz, DMSO-d6) δ 198.4 (Cq),
170.5 (Cq), 169.6 (Cq), 138.4 (CH), 136.4 (Cq), 133.5 (CH),
131.6 (CH), 129.5 (4CH), 103.5 (Cq), 100.7 (CH), 61.6 (CH2),
54.0 (CH2), 42.4 (CH), 14.9 (CH3); MS (ESI, positive mode) m/z
337 (M + Na+), 353 (M + K+).
4-Benzoyl-3-carboxamido-1-(2′,3′,5′-tri-O-acetyl-â-D-ri-
bofuranosyl)-1,4-dihydropyridine (24). To a solution of
crude 12 (0.56 mmol, 317 mg) in EtOH (20 mL) at 0 °C was
added sodium triacetoxyborohydride (0.73 mmol, 155 mg) in
solution in 5 mL of acetic acid. After 10 min at 0 °C, water
was added, and the product was extracted with dichlo-
romethane. This organic layer was washed with saturated
NaHCO3 solution and dried over Na2SO4, and the solvent was
evaporated under vacuum. The crude product was character-
ized by NMR as the cyclized structure 22, but after purification
by two successive column chromatography on silica gel (dichlo-
romethane/MeOH, 100:0 to 97:3 and ethyl acetate), only pure
24 was obtained as a pale orange powder (90 mg; 35% yield):
1H NMR (300 MHz, DMSO-d6; two epimers) δ 7.99 (m, 2H),
7.63 (m, 1H), 7.52 (m, 2H), 7.41 and 7.38 (2s, 1H), 6.86 (bs,
2H), 6.33 (d, J ) 7.8 Hz, 1H), 5.27 to 5.09 (m, 3H), 4.78 (m,
1H), 4.30 to 4.14 (m, 4H), 2.07 (s, 9H); 13C NMR (75 MHz,
DMSO-d6) δ 198.5 and 198.4 (Cq), 171.0 (Cq), 170.3 (Cq), 170.2
(Cq), 169.1 and 169.3 (Cq), 136.2 (Cq), 135.2 (CH), 133.8 (CH),
129.6 (2CH), 129.5 (2CH), 127.5 (CH), 106.6 and 106.2 (Cq),
102.8 and 102.2 (CH), 93.4 (CH), 78.9 (CH), 71.1 and 71.0 (CH),
70.6 and 70.4 (CH), 64.2 and 64.1 (CH2), 47.2 (CH), 21.4, 21.2,
21.1 (3CH3); MS (ESI, positive mode) m/z 487 (M + H+), 509
(M + Na+), 525 (M + K+).
Enoyl reductase activity of InhA was assayed by monitoring
the oxidation of NADH at 340 nm (25 °C), and initial velocities
were determined. Results are expressed as a percentage of
remaining InhA activity determined on the basis of control
experiments. â-Ketoacyl reductase activity of MabA was
assayed in the same way, monitoring the oxidation of NADPH
at 340 nm.
Acknowledgment. This work was supported by the
CNRS and in part by a grant from the European
Community (Grant No. QLK2-CT-2000-01761).
Supporting Information Available: The general experi-
mental methods and a compound characterization checklist
4-Benzoyl-3-carboxamido-1-â-D-ribofuranosyl-1,4-di-
hydropyridine (1). To 113 mg of 24 (0.23 mmol) was added
14 mL of a solution of guanidine hydrochloride (50 mM) and
sodium methanolate (11 mM) in methanol/dichloromethane (9:
1). The mixture was allowed to stand at room temperature
for 15 min, and Amberlite IRP 64 ion-exchange resin (100-
1
are provided, and copies of a H NMR spectrum or a proton-
decoupled 13C NMR spectrum for each new compound are
included. This material is available free of charge via the
Internet at http://pubs.acs.org.
JO051901Z
10510 J. Org. Chem., Vol. 70, No. 25, 2005