Probing binding requirements of NAD kinase with modified substrate (NAD) analogues

Article abstract of DOI:10.1016/j.bmcl.2007.01.012

Synthesis of novel NAD⁺ analogues that cannot be phosphorylated by NAD kinase is reported. In these analogues the C2′ hydroxyl group of the adenosine moiety was replaced by fluorine in the ribo or arabino configuration (1 and 2, respectively) or was inverted into arabino configuration to give compound 3. Compounds 1 and 2 showed inhibition of human NAD kinase, whereas analogue 3 inhibited both the human and Mycobacterium tuberculosis NAD kinase. An uncharged benzamide adenine dinucleotide (BAD) was found to be the most potent competitive inhibitor (K_i = 90 μM) of the human enzyme reported so far.

Full text of DOI:10.1016/j.bmcl.2007.01.012

Bioorganic & Medicinal Chemistry Letters 17 (2007) 1512–1515
Probing binding requirements of NAD kinase with
modified substrate (NAD) analogues
Laurent Bonnac,a Liqiang Chen,a Rashmi Pathak,a Guangyao Gao,a
Qian Ming,a Eric Bennett,a Krzysztof Felczak,a Martin Kullberg,a Steven E. Patterson,a
Francesca Mazzola,b Giulio Magnib and Krzysztof W. Pankiewicza,*
aCenter for Drug Design, University of Minnesota, Minneapolis, MN 55455, USA
bInstituto di Biotecnologie Biochimiche, Universita Politecnica delle Marche, Via Ranieri, 60131 Ancona, Italy
Received 16 November 2006; revised 2 January 2007; accepted 3 January 2007
Available online 17 January 2007
Abstract—Synthesis of novel NAD+ analogues that cannot be phosphorylated by NAD kinase is reported. In these analogues the
C20 hydroxyl group of the adenosine moiety was replaced by fluorine in the ribo or arabino configuration (1 and 2, respectively) or
was inverted into arabino configuration to give compound 3. Compounds 1 and 2 showed inhibition of human NAD kinase, whereas
analogue 3 inhibited both the human and Mycobacterium tuberculosis NAD kinase. An uncharged benzamide adenine dinucleotide
(BAD) was found to be the most potent competitive inhibitor (Ki = 90 lM) of the human enzyme reported so far.
ꢀ 2007 Elsevier Ltd. All rights reserved.
NAD kinase catalyzes a magnesium-dependent phos-
HO OH
phorylation of the 20-hydroxyl group of the adenosine
ribose moiety of nicotinamide adenine dinucleotide
(NAD) using ATP or inorganic polyphosphates as phos-
phoryl donors to give NADP (Scheme 1).1 There are
two classes of the enzyme, one which is specific for
NAD+ and one which also phosphorylates NADH.
The Mycobacterium tuberculosis enzyme falls into the
latter category. No other pathway of NADP biosynthe-
sis has been found in prokaryotic or eukaryotic cells.
Although NAD kinase was discovered as early as in
1937, little was known about the structure, function,
and mode of action of this enzyme until very recently.2
NAD kinase plays a crucial role in controlling the cellu-
lar redox state by regulating the ratio of reduced coen-
zymes NADH/NADPH.2,3 More recently, it has
become clear that NADP-mediated signal transduction
affects numerous metabolic pathways and there is a
growing body of evidence that NADP serves as an
important component in the control of essential cellular
processes. For example, NADP is converted into nico-
tinic acid adenine dinucleotide phosphate (NAADP)
by the exchange of nicotinamide with nicotinic acid cat-
CONH2
NAD-kinase
O
O
+
O
N
O P O P O
HO
N
O
N
N
OH
ATP or poly(P)
N
HO OH
NAD
NH2
O
HO P O OH
HO
CONH2
O
O
+
O
N
N
O P O P O
HO
N
O
OH
N
N
NADP
HO OH
NH2
Scheme 1. NAD phosphorylation by NAD-kinase.
alyzed by CD38 and ADP-ribosyl cyclase.4–6 NAADP is
the most potent Ca2+ mobilizing agent in mammalian
cells. Interestingly, human NAD-kinase is highly selec-
tive and does not phosphorylate NAAD into NAADP.7
Keywords: NAD kinase; M. tuberculosis; Benzamide adenine dinucle-
otide (BAD); NAD analogues.
In contrast to NAD, which serves as a substrate for
a number of degradation enzymes such as ADP-
ribosyl transferases,8 poly-ADP-ribose polymerases,9
*
Corresponding author. Tel.: +1 612 625 7968; fax: +1 612 625
8252; e-mail: panki001@umn.edu
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.01.012

L. Bonnac et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1512–1515
1513
ADP-ribosyl cyclase,10 and NAD-dependent histone
OH
deacetylases,11 NADP is not known to be significantly
degraded. The affinities of ATP (Km = 3.3 mM) and
NAD (Km = 0.54 mM) for the human enzyme are low
and close to the physiological concentration of these
nucleotides.7
CONH2
+
N
OH
O
O
O
O P O P O
HO
N
O
N
N
OH
N
OH
OH
3
HO
HO
NH2
NAD-kinase is expressed in both prokaryotic and
eukaryotic cells. NADP is essential for the growth of
several microorganisms, including multi-drug resistant
M. tuberculosis, Bacillus subtillis, Escherichia coli, and
others.12–14 Bacterial NAD kinases showed substantial
differences from the human and mammalian enzymes,
and therefore they have been suggested as attractive tar-
gets for the development of novel antibiotics.12,15
HO OH
O
CONH2
O
O
O P O P O
HO OH
N
O
N
N
N
BAD
NH2
Figure 2. NAD with an inverted (arabino) configuration and benzam-
ide adnine dinucleoside (BAD).
The first crystal structure of NAD kinase was published
in 200412 describing the apo-enzyme from M. tuberculosis
followed by the structure of the same kinase in complex
with NAD.16 An X-ray structure of NAD kinase from
Archeoglobus fulgidus in complex with ATP, NAD, or
NADP17 revealed more details of these enzymes’ archi-
tecture. Studies of site-directed mutagenesis of the
M. tuberculosis enzyme15 and evidence of its allosteric
inhibition provided a wealth of potentially useful
information for drug design.
In addition, the activity of a known uncharged NAD
analogue, benzamide adenine dinucleotide (BAD or
1- deaza-NAD, Fig. 2)21 was also examined.
The 20-fluoro analogue 1 is structurally very similar to
NAD but cannot be phosphorylated by NAD kinase
to the corresponding NADP analogue. A fluorine atom
is a good mimic of a hydroxyl group in terms of size and
polarity, and it preserves the same C30-endo conforma-
tion of the adenosine ribose as that of NAD.22 We syn-
thesized analogue 1 by coupling of the commercially
available nicotinamide mononucleotide (NMN) with
50-monophosphate imidazolide of 20-deoxy-20-fluoroad-
enosine23 and found that at 1 mM concentration it
inhibited 43% of activity of the human but not the bac-
terial enzyme (Table 1).
In spite of this no inhibitors of the human or bacterial
enzyme have ever been reported and still little is known
about the mechanism of action of NAD kinase.17 No
crystal structure of the human NAD kinase has been
reported. Clearly further studies are needed to under-
stand the interactions of NAD kinase with ATP and
NAD as substrates, and other ligands as potential inhib-
itors. In this paper, we present our approach to drug
design through examination of interactions of modified
NAD analogues with NAD kinases to complement infor-
mation available from site-directed mutagenesis.15,18
Next, we investigated the importance of the ribo configu-
ration of the analogue 1 for the enzymatic activity of
NAD kinase. Using (20-deoxy-20-fluoro-1-b-D-arabino-
furanosyl) adenine22 and similar coupling with NMN
we prepared NAD analogue 2 in which the 20-fluoro atom
is in the arabino configuration. The adenine sugar of this
compound is in the C20-endo conformation, which is pre-
ferred in 20-deoxynucleosides rather than in ribonucleo-
sides. In spite of this different conformation, the NAD
analogue 2 showed a similar activity to that of 1. It inhib-
ited the human enzyme (37% inhibition at 1 mM) and was
inactive against M. tuberculosis NAD kinase.
We report herein the synthesis19 and NAD kinase inhib-
itory activity20 of novel analogues of NAD (1–3). These
analogues were prepared by replacement of the 20-hy-
droxyl group of the adenine ribose of NAD by fluorine
in the ‘down’ ribo or ‘up’ arabino configuration or by
inversion of the 20-OH into the arabino configuration
(Figs. 1 and 2).
F
OH
CONH2
+
N
O
O
Table 1. Effects of NAD analogues 1–3, BAD, and bromoacetyl
pyridines 11–13 on inhibition of human and M. tuberculosis NAD
kinase
O
O P O P
HO
O
N
O
N
N
OH
N
NAD analogue
concn 1 mM
Inhibition (%)
Ki (lM)
human enzyme
1
HO OH
NH2
Human enzyme M. tuberculosis
OH
CONH2
1
43
37
40
90
<10
<10
40
ND
ND
ND
90
+
F
O
O
2
O
N
O P O P
HO OH
O
3
BAD
N
O
N
N
<10
37
N
BAD (at 2 mM) 90
90
11
12
13
60
44
64
63
30
ND
ND
ND
HO OH
2
NH2
40
Figure 1. 20-fluoro ribo- and 20-fluoro arabino NAD.

1514
L. Bonnac et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1512–1515
O
Finally, we studied whether or not NAD analogues with
an inverted configuration of the 20-OH group or lacking
the positive charge of the nicotinamide ring would serve
as substrates.24 To answer these questions, we synthesized
NAD analogue 3 and prepared uncharged benzamide
adenine dinucleotide (BAD or 1-deaza-NAD, Fig. 2).
Interestingly, NAD analogue 3 inhibited both the
human and M. tuberculosis NAD kinase (40% inhibition
at 1 mM for both enzymes). Since the 20-OH group can
act as a donor or as an acceptor in hydrogen bonding,
whereas fluorine has only acceptor ability, this result
may indicate that the 20-substituent should act as a hydro-
gen donor in order to inhibit the bacterial enzyme. Such a
requirement is consistent with the X-ray structure of the
M. tuberculosis enzyme, in which the 20-OH hydrogen
bonds to Asp85; however, this residue is also conserved
in the human enzyme. In general, the inhibitory activity
of NAD analogues 1–3 was low; however, these com-
pounds may compete with NAD+ whose affinity to the en-
zyme is also in the low millimolar range.7
O
Br2
Br
N
N
8 2-pyridine
9 3-pyridine
10 4-pyridine
11 2-pyridine
12 3-pyridine
13 4-pyridine
Scheme 3. Bromoacetyl pyridines, irreversible NAD kinase inhibitors.
of the cyano group of nucleoside 5 into the carboxyam-
ido function of 6, and debenzylation afforded BR in 56%
overall yield. Using this new methodology we prepared
gram amounts of BR and converted it into BAD.21
As a positive control we synthesized three bromoacetyl
pyridines (11–13, Scheme 3) since one of them, 3-bromo-
acetyl pyridine 12, was reported as a potent irreversible
inhibitor of NAD kinase from pigeon liver.27 Apparent-
ly compound 12 alkylates a histidine residue present in
the active site of NAD kinase. It is a non-specific inhib-
itor which was reported to alkylate the active site of
other NAD-dependent enzymes.29,30 These compounds
were prepared in high yield (95–98%) from the corre-
sponding acetyl pyridines 8, 9, 10, by direct bromination
as previously reported.28 All three compounds showed
inhibitory activity against both the human and the
bacterial enzyme (Table 1).
In contrast to the charged NAD+ analogues 1–3, the
uncharged BAD was found to be a more potent
inhibitor of human NAD kinase (90% inhibition at
1 mM) than compounds 1–3. It showed competitive
inhibition of the human enzyme with Ki = 90 lM. In high-
er concentration (2 mM) it also inhibited the bacterial en-
zyme (36%). The TB enzyme does have NADH kinase
activity.18
In conclusion, we found that the replacement of the
20-OH group of the adenosine moiety of NAD+ by
fluorine (analogues 1 and 2) afforded only weak inhib-
itors of human NAD kinase, which were inactive
against the M. tuberculosis NAD kinase. In contrast,
inversion of the configuration of the OH (such as in
analogue 3) led to the inhibition of both the human
and the bacterial enzyme. BAD was found to be a
competitive inhibitor of the human enzyme with an
affinity 6-fold higher than that of NAD+ (Ki = 90 lM
vs Km = 540 lM, respectively). Apparently elimination
of the positive charge of the NAD+ increased the
inhibitory activity. In higher concentration BAD also
inhibited the bacterial enzyme. Our studies indicate
that the analogue of BAD with the inverted (arabino)
configuration of the 20-OH of adenosine moiety is
likely to show increased inhibition of M. tuberculosis
NAD kinase. The preparation of this and other
NAD analogues is now in progress.
BAD is an active metabolite of benzamide riboside (7,
BR, Scheme 2), the C-nucleoside which showed similar
anticancer activity to that of tiazofurin, an orphan drug
used clinically for treatment of patients with chronic
myelogenous leukemia.25 The laborious and inefficient
(10 step) synthesis of BR26 hampered research on bio-
logical activities of this compound in the last decade.
Recently, we developed a new, simple, 4-step procedure
for stereoselective synthesis of BR. Thus, Grignard reac-
tion of commercially available 3-iodobenzonitrile with
ribonolactone (4, Scheme 2) followed by ‘one pot’
removal of the anomeric hydroxyl group, conversion
CN
1) iPrMgCl,
BnO
BnO
3-iodobenzonitrile
O
O
O
2) BF3OEt2, Et3SiH
70%
BnO
OBn
BnO
BnO
OBn
Acknowledgment
4
5
These studies were funded by the Center for Drug
Design, University of Minnesota.
CONH2
BBr3
CONH
References and notes
HO
Me3SiOK
85%
O
O
1. Magni, G.; Orsomando, G.; Raffaelli, N. Mini-Rev. Med.
Chem. 2006, 6, 739.
94%
OBn
BnO
HO
OH
2. Berger, F.; Ramirez-Hernandez, M. H.; Ziegler, M. Trends
Biochem. Sci. 2004, 29, 111.
6
7
3. Grose, J. H.; Joss, L.; Velick, S. F.; Roth, J. R. Proc. Natl.
Acad. Sci. U.S.A. 2006, 103, 7601.
Scheme 2. A simple, stereoselective synthesis of benzamide riboside
(BR).

L. Bonnac et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1512–1515
1515
4. Yamasaki, M.; Churchill, G. C.; Galione, A. FEBS
J. 2005, 272, 4598.
(11 mg, sodium salt) was obtained as a white solid (yield
32%). H NMR (D2O, 600 MHz) d 9.35 (s, 1H), 9.21 (d,
1
5. Dowden, J.; Berridge, G.; Moreau, C.; Yamasaki, M.;
Churchill, G. C.; Potter, B. V.; Galione, A. Chem. Biol.
2006, 13, 659.
6. Graeff, R.; Liu, Q.; Kriksunov, I. A.; Hao, Q.; Lee, H. C.
J. Biol. Chem. 2006, E-pub ahead of print.
7. Lerner, F.; Niere, M.; Ludwig, A.; Ziegler, M. Biochem.
Biophys. Res. Commun. 2001, 288, 69.
8. Seman, M.; Adriouch, S.; Haag, F.; Koch-Nolte, F. Curr.
Med. Chem. 2004, 11, 857.
9. Gagne, J. P.; Hendzel, M. J.; Droit, A.; Poirier, G. G.
Curr. Opin. Cell Biol. 2006, 18, 145.
10. Potter, B. V.; Walseth, T. F. Curr. Mol. Med. 2004, 4, 303.
11. Sauve, A. A.; Schramm, V. L. Curr. Med. Chem. 2004, 11,
807.
12. Garavaglia, S.; Raffaelli, N.; Finaurini, L.; Magni, G.;
Rizzi, M. J. Biol. Chem. 2004, 279, 40980.
13. Garavaglia, S.; Galizzi, A.; Rizzi, M. J. Bacteriol. 2003,
185, 4844.
J = 6.0 Hz, 1H), 8.87 (d, J = 7.8 Hz, 1H), 8.38 (s, 1H),
8.27-8.22 (m, 1H), 8.19 (s, 1H), 6.32 (d, J = 6.6 Hz, 1H),
6.10 (d, J = 5.4 Hz, 1H), 4.60 (pseudo t, J = 6.6 Hz, 1H),
4.58–4.55 (m, 1H), 4.51 (pseudo t, J = 5.4 Hz, 1H),
4.48–4.44 (m, 2H), 4.40 (ddd, J = 11.8, 4.0, 2.6 Hz, 1H),
4.36 (ddd, J = 11.7, 5.1, 2.7 Hz, 1H), 4.31–4.25 (m, 2H),
4.19–4.15 (m, 1H). 31P NMR (D2O, 243 MHz) d ꢀ10.13
(d, J = 20.6 Hz), ꢀ10.56 (d, J = 21.1 Hz). MS ESI calcd
(MꢀH)ꢀ 662.1018, found 662.0819.
20. NAD kinase14,15 activity was determined by an HPLC
assay. The assay mixture (150 lL) consisted of 50 mM
Tris/HCl, pH 8.0, 20 mM MgCl2, 5 mM ATP, 1 mM
NAD, and an appropriate amount of inhibitor and
enzyme. After incubation for 30 min at 37 ꢁC, an aliquot
(70 lL) was withdrawn and added to ice-cold 1.2 M
HClO4 (35 lL) and kept for 10 min. The mixture was
centrifuged for 1 min at 16,000g. An aliquot (100 lL) of
the HClO4 supernatant was neutralized by addition of
0.8 M K2CO3 (28 lL), kept at ꢀ20 ꢁC for 15 min, and
centrifuged. An aliquot of the supernatant was loaded
onto a Supelcosil LC-18 reversed-phase column (250 ·
4.6 mm, 5 lm). The elution conditions: 9 min at 100%
buffer A (0.1 M potassium phosphate, pH 6.0), 6 min at up
to 12% buffer B (buffer A, containing 20% methanol),
2.5 min at up to 45% buffer B, 2.5 min at up to 100% buffer
B, and hold at 100% buffer B for 5.5 min; finally the
gradient returns to 100% buffer A in 5 min. The column
was flushed with buffer A for 4.5 min prior to the next run.
The flow rate was 1.3 ml/min, and the column was
thermostated at 25 ꢁC. The eluate absorbance was mon-
itored at 260 nm.
14. Zerez, C. R.; Moul, D. E.; Gomez, E. G.; Lopez, V. M.;
Andreoli, A. J. J. Bacteriol. 1987, 169, 184.
15. Raffaelli, N.; Finaurini, L.; Mazzola, F.; Pucci, L.; Sorci,
L.; Amici, A.; Magni, G. Biochemistry 2004, 43, 7610.
16. Mori, S.; Yamasaki, M.; Maruyama, Y.; Momma, K.;
Kawai, S.; Hashimoto, W.; Mikami, B.; Murata, K.
Biochem. Biophys. Res. Commun. 2005, 327, 500.
17. Liu, J.; Lou, Y.; Yokota, H.; Adams, P. D.; Kim, R.; Kim,
S. H. J. Mol. Biol. 2005, 354, 289.
18. Mori, S.; Kawai, S.; Shi, F.; Mikami, B.; Murata, K.
J. Biol. Chem. 2005, 280, 24104.
19. NAD analogues 1–3 were prepared by coupling of the
50-phosphoimidazolides of the corresponding adenine
sugars with nicotinamide mononucleotide (NMN) and
21. Zatorski, A.; Watanabe, K. A.; Carr, S. F.; Goldstein, B.
M.; Pankiewicz, K. W. J. Med. Chem. 1996, 39, 2422.
22. Pankiewicz, K. W. Carbohydr. Res. 2000, 327, 87.
23. Zatorski, A.; Goldstein, B. M.; Colby, T. D.; Jones, J. P.;
Pankiewicz, K. W. J. Med. Chem. 1995, 38, 1098.
24. Shi, F.; Kawai, S.; Mori, S.; Kono, E.; Murata, K. FEBS
J. 2005, 272, 3337.
25. Gharehbaghi, K.; Grunberger, W.; Jayaram, H. N. Curr.
Med. Chem. 2002, 9, 743.
26. Krohn, K.; Heins, H.; Wielckens, K. J. Med. Chem. 1992,
35, 511.
27. Apps, D. K. Eur. J. Biochem. 1971, 19, 301.
28. Taurins, A.; Blaga, A. J. Heterocycl. Chem. 1970, 7,
1137.
29. Ulmer, W.; Froeschle, M.; Jany, K. D. Eur. J. Biochem.
1983, 136, 183.
1
characterized by H, 31P NMR and MS. The synthesis of
P1-(nicotinamide-ribos-50-yl)-P2(adenine-arabinofuranos-
50-yl) pyrophosphate (3) serves as an example: NMN
(82 mg, 0.24 mmol) was added to a 0.2 M MnCl2/form-
amide solution (1.5 mL, 0.30 mmol), which was dried with
molecular sieves for several days before use. To this
mixture were added adenine-arabinofuranosyl 50-phos-
phoimidazolide (101 mg, 0.20 mmol) and anhydrous
MgSO4 (50 mg, 0.42 mmol). The reaction mixture was
stirred for 24 h at rt and the crude product was precip-
itated by adding CH3CN (6 mL). After washing with
CH3CN, the solid was dissolved in 1 M TEAB solution
(10 mL), filtered, and the filtrate was used directly for
purification by RP-HPLC. The resulting Et3N salt was
applied on a column of Dowex50WX8-200 (Na+ form) at
4 ꢁC and eluted with water to give the desired sodium salt.
From about one-fourth of the crude product compound 3
30. Woenckhaus, C.; Bieber, E.; Jeck, R. Prog. Clin. Biol. Res.
1987, 232, 53.