Chemoenzymatic synthesis of 7-deaza cyclic adenosine 5′-diphosphate ribose analogues, membrane-permeant modulators of intracellular calcium release

Article abstract of DOI:10.1021/jo071236p

(Chemical Equation Presented) An optimized synthetic route to 7-deaza-8-bromo-cyclic adenosine 5′-diphosphate ribose (7-deaza-8-bromo-cADPR 3), an established cell-permeant, hydrolysis-resistant cyclic adenosine 5′-diphosphate ribose (cADPR) antagonist, i

Full text of DOI:10.1021/jo071236p

Chemoenzymatic Synthesis of 7-Deaza Cyclic Adenosine
5′-Diphosphate Ribose Analogues, Membrane-Permeant Modulators
of Intracellular Calcium Release
Bo Zhang, Victoria C. Bailey, and Barry V. L. Potter*
Wolfson Laboratory of Medicinal Chemistry, Department of Pharmacy and Pharmacology, UniVersity of
Bath, ClaVerton Down, Bath, BA2 7AY, United Kingdom
b.V.l.potter@bath.ac.uk
ReceiVed June 13, 2007
An optimized synthetic route to 7-deaza-8-bromo-cyclic adenosine 5′-diphosphate ribose (7-deaza-8-
bromo-cADPR 3), an established cell-permeant, hydrolysis-resistant cyclic adenosine 5′-diphosphate ribose
(cADPR) antagonist, is presented. Using NMR analysis, we found that 3 adopted a C-2′ endo conformation
in the N9-linked ribose and a syn conformation about the N9-glycosyl linkage, which are similar to that
of cADPR. The synthetic route was also employed to produce 7-deaza-2′-deoxy-cADPR 4, a potential
cell-permeant cADPR analogue. 3 and 4 were more stable to chemical hydrolysis, consistent with the
observation that 7-deaza-cADPR analogues are more stable than their parent adenosine derivatives. 3
was also found to be stable to enzyme-mediated hydrolysis using CD38 ectoenzyme.
Introduction
Cyclic adenosine 5′-diphosphate ribose (cADPR 1) is a Ca2+
systems, cADPR is synthesized from nicotinamide adenine
dinucleotide (NAD+) by Aplysia ADP-ribosyl cyclase, an
enzyme known to have loose substrate selectivity, thus tolerating
structural modifications of its substrates.7,8 The crystal structure
of the enzyme has been studied, and a covalent bond linkage
between the enzyme and substrate has been proposed.9,10
Adenosine 5′-diphosphate ribose (ADPR) analogues have been
used to probe mechanistic questions for cyclization and hy-
drolysis.11 Numerous derivatives of the biologically important
molecule cADPR have now been synthesized, and their
-
mobilizing second messenger,1-3 releasing calcium from intra-
cellular stores in various cell types including sea urchin eggs,
pancreatic â cells, and smooth muscle and T-cells. cADPR-
induced calcium release is independent of the well-established
myo-inositol 1,4,5-trisphosphate pathway4,5 and is regulated by
gating the type 2 and/or type 3 ryanodine receptor.6 In biological
(1) Galione, A.; Lee, H. C.; Busa, W. Science 1991, 253, 1143-1146.
(2) Koshiyama, H.; Lee, H. C.; Tashjian, A. H. J. Biol. Chem. 1991,
266, 16985-16988.
(7) Lee, H. C.; Aarhus, R. Cell Regul. 1991, 2, 203-209.
(8) Walseth, T. F.; Lee, H. C. Biochem. Biophys. Acta 1993, 1178, 235-
242.
(9) Prasad, G. S.; McRee, D. E.; Stura, E. A.; Levitt, D. G.; Lee, H. C.;
Stout, C. D. Nat. Struct. Biol. 1996, 3, 957-963.
(10) Love, M. L.; Szebenyi, D. M. E.; Kriksunov, I. A.; Thiel, D. J.;
Munshi, C.; Graeff, R.; Lee. H. C.; Hao, Q. Structure 2004, 12, 477-486.
(3) Takasawa, S.; Nata, K.; Yonekura, H.; Okamoto, H. Science 1993,
259, 370-373.
(4) Clapper, D. L.; Walseth, T. F.; Dargie, P. J.; Lee, H. C. J. Biol. Chem.
1987, 262, 9561-9568.
(5) Dargie, P. J.; Agre, M. C.; Lee, H. C. Cell Regul. 1990, 279-290.
(6) Guse, A. H. FEBS 2005, 272, 4590-4597.
10.1021/jo071236p CCC: $40.75 © 2008 American Chemical Society
Published on Web 01/30/2008
J. Org. Chem. 2008, 73, 1693-1703
1693

Zhang et al.
FIGURE 2. (a) Mechanism of 6-amino group directed bromination
reaction at the pseudo-8-position (drawing based on the literature31).
(b) The 8-bromination was determined by gHMBC method, and a
coupling between C-1′ proton and the C-8 carbon was observed.
8-Bromo-7-deazaadenosine is proposed to adopt a syn conformation
as shown in (b).
FIGURE 1. Structures and numbering of cADPR and analogues. The
N9-linked ribose (also known as “southern”) and N1-linked ribose (also
known as “northern”) are distinguished by adopting prime (′) and double
prime (′′) symbols, respectively, for sugar carbons shown in this figure.
pharmacological activities have been investigated, although the
field is still in its infancy.12,13 A recent notable example is the
nonhydrolyzable cyclic inosine 5′-diphosphate ribose (cIDPR)14
and its 8-substituted analogues.15 As a stable analogue, cIDPR
was cocrystallized with the ectoenzyme ADP-ribosyl cyclase/
hydrolase CD38, providing insight into the binding and catalytic
process for the first time using this close analogue of the natural
ligand.16 Conservatively modified carbocyclic analogues of
cADPR (e.g., cADPcR) have also been synthesized and
investigated biologically.17-19 Analogues with more radical
modifications20 are also finding useful application.21
Among these analogues, deletion of the N7 nitrogen atom of
the purine of cADPR, inter alia, leads to a class of derivatives
that demonstrate hydrolysis-resistant properties, such as 7-deaza-
cADPR 2. 2 (Figure 1) was one of the first 7-deaza-cADPR
analogues ever synthesized, and its biological activities were
studied.22 As one of the closest possible analogues of the natural
ligand cADPR, 7-deaza-cADPR, however, was found to be a
partial agonist relative to cADPR, demonstrating that the N7
position is part of the recognition mechanism for effective
calcium release. It is also noteworthy that 7-deaza-cADPR is
more stable to heat-induced hydrolysis at the N1 ribosyl linkage
and is also a poor substrate for cADPR hydrolase.22 8-Position
modification is known to convert cADPR from an agonist to
an antagonist.12,13 Thus, 8-amino-cADPR 5 and 8-bromo-
cADPR 6 have been reported to demonstrate antagonistic effects
in both sea urchin eggs and Jurkat T-cells systems, and
moreover, 8-bromo-cADPR 6 was found to be cell permeant.23
With the partial agonistic property of 2 established, it was thus
of interest to study the combination of 8-substitution and N7
deletion, leading to the synthesis and biological evaluation of
7-deaza-8-bromo-cADPR 3.23,24 7-Deaza-8-bromo-cADPR ex-
hibited antagonistic activity in sea urchin homogenate, and its
potency was higher than that of 8-bromo-cADPR.23 It was also
active in T-lymphocytes and was employed in a seminal study
to investigate the signaling by cADPR in T-lymphocyte signal
transduction25 and also to establish a role for the cADPR
signaling pathway in long-term synaptic depression in the
hippocampus.26 In addition to antagonistic activity, 7-deaza-8-
bromo-cADPR 3 is also more resistant to both chemical and
enzymatic hydrolysis, as expected for 7-modified analogues,
and importantly is cell permeant. The cell membrane perme-
ability of 7-deaza-8-bromo-cADPR facilitates the administration
of this antagonist into intact cells, which usually relies on
techniques such as microinjection or patch clamping. Together
with these findings, 7-deaza-8-bromo-cADPR has thus become
one of the best-known pharmacological tools in this growing
area. A preliminary chemo-enzymatic synthetic route to 3 was
reported.24
The wide applicability of 3 prompted us to better optimize
its synthetic route, as the key intermediate 7-deaza-8-bromo-
NAD+ 12 in our earlier synthesis was prepared in a relatively
low yield of 28% via a dicyclohexylcarbodiimide (DCC)-
mediated condensation of 7-deaza-8-bromo-AMP 10 and
â-NMN+ (Scheme 1).24 Synthesis of cADPR derivatives in
general is a somewhat difficult process. The choice of routes is
limited to either a total synthetic methodology18,20,27 or a chemo-
enzymatic method relying on the cyclizing ability of certain
analogues of NAD+ by Aplysia ADP-ribosyl cyclase.28 Although
the total synthesis method in principle facilitates preparation
of more diverse cADPR compounds, this method has mainly
been used most efficiently for cyclic adenosine 5′-diphospho-
carbocyclic ribose (cADPcR) analogues, mimics of cADPR in
which a methylene group replaces the furan sugar oxygen atom
in the N1-linked ribose. Most of the published cADPR analogues
have not been synthesized by this method, partially because the
N1-ribose linkage is not stable enough to survive in chemical
treatments, although this is useful when the N1-linked ribose
(11) Zhang, B.; Muller-Steffner, H.; Schuber, F.; Potter, B. V. L.
Biochemistry 2007, 46, 4100-4109.
(12) Potter, B. V. L.; Walseth, T. F. Curr. Mol. Med. 2004, 4, 303-
311.
(13) Guse, A. H. Curr. Med. Chem. 2004, 11, 847-855.
(14) Wagner, G. K.; Guse, A. H.; Potter, B. V. L. J. Org. Chem. 2005,
70, 4810-4819.
(15) Moreau, C.; Wagner, G. K.; Weber, K.; Guse, A. H.; Potter, B. V.
L. J. Med. Chem. 2006, 49, 5162-5176.
(16) Liu, Q.; Kriksunov, I. A.; Moreau, C.; Graeff, R.; Potter, B. V. L.;
Lee, H. C.; Hao, Q. J. Biol. Chem. 2007, 282, 24825-24832.
(17) Shuto, S.; Fukuoka, M.; Kudoh, T.; Garnham, C.; Galione, A.; Potter,
B. V. L.; Matsuda, A. J. Med. Chem. 2003, 46, 4741-47490.
(18) Kudoh, T.; Fukuoka, M.; Ichikawa, S.; Murayama, T.; Ogawa, Y.;
Hashii, M.; Higashida, H.; Kunerth, S.; Weber, K.; Guse, A. H.; Potter, B.
V. L.; Matsuda, A.; Shuto, S. J. Am. Chem. Soc. 2005, 127, 8846-8855.
(19) Bailey, V. C.; Fortt, S. M.; Summerhill, R. J.; Galione, A.; Potter,
B. V. L. FEBS Lett. 1996, 379, 227-230.
(20) Gu, X. F.; Yang, Z.; Zhang, L.; Kunerth, S.; Fliegert, R.; Weber,
K.; Guse, A. H.; Zhang, L. J. Med. Chem. 2004, 47, 5674-5682.
(21) Guse, A. H.; Gu, X.; Zhang, L.; Weber, K.; Kramer, E.; Yang, Z.;
Jin, H.; Li, Q.; Carrier, L.; Zhang, L. J. Biol. Chem. 2005, 280, 15952-
15959.
(24) Bailey, V. C.; Sethi, J. K.; Galione, A.; Potter, B. V. L. Chem.
Commun. 1997, 695-696.
(25) Guse, A. H.; da Silva, C. P.; Berg, I.; Skapenko, A. L.; Weber, K.;
Heyer, P.; Hohenegger, M.; Ashamu, G. A.; Schulze-Koops, H.; Potter, B.
V. L.; Mayr, G. W. Nature 1999, 398, 70-73.
(26) Reyes-Harde, M.; Empson, R.; Potter, B. V. L.; Galione, A.; Stanton,
P. K. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 4061-40660.
(27) Shuto, S.; Matsuda, A. Curr. Med. Chem. 2004, 11, 827-845.
(28) Ashamu, G. A.; Galione, A.; Potter, B. V. L. J. Chem. Soc., Chem.
Commun. 1995, 1359-1360.
(22) Bailey, V. C.; Sethi, J. K.; Fortt, S. M.; Galione, A.; Potter, B. V.
L. Chem. Biol. 1997, 4, 51-61.
(23) Sethi, J. K.; Empson, R. M.; Bailey, V. C.; Potter, B. V. L.; Galione,
A. J. Biol. Chem. 1997, 272, 16358-16363.
1694 J. Org. Chem., Vol. 73, No. 5, 2008

Chemoenzymatic Synthesis of 7-Deaza-cADPR Analogues
SCHEME 1. Synthesis of 7-Deaza-8-Br-cADPRa
a Compounds 3 and 10-12 were isolated as triethylammonium salts.
has been deleted.20 Thus, we report here full details of an
optimized chemo-enzymatic approach to compound 3, in which
a morpholine-activated29 intermediate 11 was introduced, fol-
lowed by a Lewis acid catalyzed condensation with â-NMN+30
(Scheme 1) to give the desired intermediate dinucleotide 12 in
a high yield over two steps. We also report a synthesis of
compound 4, a potential cell-permeant, hydrolysis-resistant
agonist of cADPR-sensitive calcium release, by adaptation of
our modified synthetic route, demonstrating the route as a robust,
versatile approach for synthesizing cADPR analogues.
enosine 9 was assumed to adopt a syn conformation around
the N9-glycosyl bond to minimize the nonbonded repulsion
between the 8-bromo group and the ribose.33,34 The H-2′
chemical shift was explored as an indicator of the syn/anti
conformation. A significant downfield shift of H-2′ for the syn
conformer is normally expected, as seen for 8-bromo-7-
deazaadenosine 9, whose H-2′ chemical shift is at δ 5.12 (the
chemical shift for the anti 7-deazainosine is 4.30 ppm15),
indicating a syn conformation about the N9-glycosyl linkage.
An explanation was proposed to rationalize the 8-position
selectivity, in which the reaction was assumed to go through
an electrophilic substitution, as the free radical reaction is not
likely to happen under the conditions.31 Under the basic buffer
conditions, the 3,6-position nitrogens are not protonated and
therefore the lone pair on these nitrogens donates electron
density through the conjugated system, facilitating electrophilic
attack at C-8, and consequently producing the 8-bromotubercidin
as the predominant product, as shown in Figure 2.31
Dry 8-bromo-7-deazaadenosine was selectively phosphory-
lated at the 5′-hydroxyl group by adaptation of a published
method.35 Treatment of compound 9 in TEP with POCl3 and a
trace amount of water at 0 °C for 1 h resulted in the formation
of a sole product 10 as monitored by HPLC (for conditions,
see the Supporting Information). The desired 7-deaza-8-bromo
5′-adenosine monophosphate (7-deaza-8-bromo-AMP) 10 was
first purified using reverse-phase chromatography to remove
Results and Discussion
The synthetic route to compound 3 starts with bromination
of commercially available 7-deazaadenosine (tubercidin 8)
(Scheme 1).
Compound 8 was selectively brominated at the 8-position
using a combination of N-bromosuccinimide (NBS) and DMF,31
which had earlier been reported to be a mild combination for
electrophilic bromination.32 Purification of the product gave the
desired 9, showing only one NMR singlet at 6.74 ppm,
representing H-7, indicating that a bromination had occurred at
the 8-position. The occurrence of the 8-bromination was also
confirmed by gHMBC (Figure 2b). The 7-deaza-8-bromoad-
(29) Prabahar, K. J.; Cole, T. D.; Ferris, J. P. J. Am. Chem. Soc. 1994,
116, 10914-10920.
(30) Lee, J.; Churchill, H.; Choi, W. B.; Lynch, J. E.; Roberts, F. E.;
Volante, R. P.; Reider, P. J. Chem. Commun. 1999, 729-730.
(31) Bergstrom, D. E.; Brattesani, A. J. Nucleic Acids Res. 1980, 8,
6213-6219.
(33) Ikehara, M.; Uesugi, S.; Yoshida, K. Biochemistry 1972, 11, 830-
836.
(34) Sarma, R. H.; Lee, H. C.; Evans, F. E.; Yathindra, N.; Sundaral-
ingam, M. J. Am. Chem. Soc. 1974, 96, 7337-7348.
(35) Ikemoto, T.; Haze, A.; Hatano, H.; Kitamoto, Y.; Ishida, M.; Nara,
K. Chem. Pharm. Bull. 1995, 43, 210-215.
(32) Mitchell, R. H.; Lai, Y. H.; Williams, R. V. J. Org. Chem. 1979,
44, 4733-4735.
J. Org. Chem, Vol. 73, No. 5, 2008 1695

Zhang et al.
1
TABLE 1. Selected H-13C Connectivities from the gHMBC
any residual inorganic phosphate, which had been otherwise
normally eliminated by a yield-reducing charcoal treatment.
After further purification by an ion-exchange chromatography
to remove remaining starting material, compound 10 was
obtained as a triethylammonium salt in 77% yield and in good
purity. With 7-deaza-8-bromo-AMP in hand, we now proceeded
to synthesize the key intermediate 7-deaza-8-bromo-NAD+ 12.
7-Deaza-8-bromo-NAD+ was first synthesized by us24 using a
DCC-mediated condensation earlier published by Todd et al.36
A solution of 7-deaza-8-bromo-AMP 10 and nicotinamide 5′-
mononucleotide (â-NMN+) in pyridine water was treated with
DCC for 7 days. Purification of the product using ion-exchange
chromatography gave the desired dinucleotide 12 as its triethy-
lammonium salt in 28% yield. Although 12 was successfully
produced, the yield was relatively low and symmetrical products,
for example formed by the coupling of two â-NMN+ molecules,
were also generated. Clearly, as limited by its reaction mech-
anism, this DCC coupling reaction provides no coupling
selectivity and cannot be further optimized to increase the yield
of the desired NAD+ analogue and reduce the occurrence of
symmetrical products and is therefore not ideal for synthesizing
pyrophosphate 12. NAD+ analogues have also been produced
from the corresponding ATP analogues using NAD+ pyrophos-
phorylase (NADPP).37 These ATP analogues are often synthe-
sized enzymatically from the corresponding AMP analogues
using the combined actions of adenylate kinase and creatine
kinase. This, however, is not applicable to all AMP analogues;
for example, 8-substituted AMP analogues are not good
substrates for adenylate kinase.38 Clearly, in our case the
8-bromo-7-deaza-AMP is not suitable for this enzymatic reac-
tion. In our earlier report using the Michelson procedure,
diphenyl phosphorochloridate was used as the coupling reagent
and 7-deaza-NAD+ was successfully synthesized.22 However,
the coupling yield by this approach was relatively low and a
deprotection step was required, which reduced the overall yield
and also increased the risk of degrading the pyrophosphate. We
required a better route that gives the pyrophosphate in high yield
and requires minimum protecting steps.
An approach for synthesizing pyrophosphates was reported
by Moffatt and Khorana.39 This method involved the coupling
of a sugar phosphate and a nucleotide phosphoromorpholidate
and has been widely used for the synthesis of a large variety of
sugar nucleotide derivatives. Later improved,30 this method was
successfully explored for producing NAD+ analogues in a
relatively high yield.14,15 Compound 12 was thus successfully
synthesized by an adaptation of the method of Lee et al.30 As
shown in Scheme 1, the nucleoside monophosphate 10 was first
activated using a combination of triphenylphosphine/dipyridyl
disulfide and morpholine and the phosphoromopholidate 11 was
generated as the sole product in 87% isolated yield. The
phosphoromopholidate 11 was then condensed with â-NMN+
in the presence of a Lewis acid (MnCl2/formamide) to give the
desired pyrophosphate 12, which was then purified by reverse-
phase chromatography in 79% yield. The overall yield of this
two-step synthesis was 69%. An excess of 1.7 equiv of â-NMN+
was used in our experiment to completely consume the starting
Spectrum of 3
H-1′
H-C-N-C-8
H-C-N-C-4
H-1′′
n.d.a with C-2
n.d.a with C-6
a Not detected.
morpholidate. Residual â-NMN+ was conveniently removed by
the reverse-phase column as it had a higher polarity and shorter
elution time than the desired compound 12. The final compound
7-deaza-8-bromo-cADPR 3 was afforded by incubating com-
pound 12, obtained by either of the two routes, with Aplysia
cyclase. HPLC analysis indicated efficient cyclization of 12, at
a rate similar to that of NAD+, demonstrating 12 to be
presumably a good substrate for Aplysia cyclase. The resulting
7-deaza-8-bromo-cADPR 3 was purified by ion-exchange
chromatography to give its triethylammonium salt in 72% yield.
It was noteworthy that by HPLC analysis 7-deaza-8-bromo-
NAD+ was clearly over 90% converted into the cyclic analogue,
but some material was somehow unavoidably lost during the
subsequent purification process. gHMBC data for the cyclic
compound 3 were collected and are listed in Table 1. From this
2D spectrum, the connectivity between the C-1′ proton and a
quaternary carbon was clearly observed, indicating a substitution
at the 8-position. However, the C-1′′ proton and C-2/C-6
connectivity was not identified, possibly reflecting a weaker
bond linkage between C-1′′ and the purine N1. Indeed, as was
demonstrated in the X-ray structure of cADPR,40 the C-1′′-
N1 bond (1.54 Å) is significantly weaker than the C-1′-N9
bond (1.44 Å). The cyclic structure of compound 3 was firmly
identified by the 1D 1H NMR spectrum, where the nicotinamide
signals were clearly lost and two characteristic doublets were
observed at 6.12 and 6.05 ppm for both anomeric protons. Thus,
a robust optimized route to 3 has now been achieved, and it
should allow large amounts of the compound to be synthesized
for biological applications.
Synthesis of 7-Deaza-2′-deoxy-cADPR. We established
earlier that the 8-position unsubstituted compound 7-deaza-
cADPR is, interestingly, a partial agonist biologically and is
more stable than the parent.22 By adaptation of the synthetic
route for compound 3, we also synthesized the novel 7-deaza-
2′-deoxy-cADPR 4 (Scheme 2) as a potentially more cell-
permeant, hydrolysis-resistant partial agonist. With the lack of
a 2′-hydroxyl group on the N9-linked ribose, compound 4 should
be more lipophilic than its parent and has more potential to be
cell permeant. Also, it is likely that this ribose modification in
compound 4 should not radically affect any agonist activity for
intracellular Ca2+ release compared to its ribose analogue 2, as
we showed earlier that 2′-hydroxyl deletion of cADPR in the
N9-linked ribose does not decrease activity.41 An additional
hydrolysis-resistant property should also be added to compound
4 by virtue of the 7-deaza modification, and indeed, we
demonstrate herein that 7-deaza-2′-deoxy-cADPR 4 is more
stable than other 2′-deoxy compounds.
The synthesis of 4 employed compound 15 as the initial
building block, which could in principle conveniently be reacted
with a range of nucleophilic amines to introduce a variety of
N-6-substituted adenosine analogues, including compound 16
(36) Hughes, N. H.; Kenner, G. W.; Todd, J. Chem. Soc. 1957, 3733-
3736.
(37) Wang, L.; Aarhus, R.; Lee, H. C.; Walseth, T. F. Biochim. Biophys.
Acta 1999, 1472, 555-564.
(38) Walseth, T. F.; Aarhus, R.; Gurnack, M. E.; Wong, L.; Breitinger,
H. G.; Gee, K. R.; Lee, H. C. Methods Enzymol. 1997, 280, 294-305.
(39) Moffatt, J. G.; Khorana, H. G. J. Am. Chem. Soc. 1961, 83, 649-
658.
(40) Lee, H. C.; Aarhus, R.; Levitt, D. Nat. Struct. Biol. 1994, 1, 143-
144.
(41) Ashamu, G. A.; Sethi, J. K.; Galione, A.; Potter, B. V. L.
Biochemistry 1997, 36, 9509-9517.
1696 J. Org. Chem., Vol. 73, No. 5, 2008

Chemoenzymatic Synthesis of 7-Deaza-cADPR Analogues
SCHEME 2. Synthesis of 7-Deaza-2′-deoxy-cADPRa
a Compounds 4 and 17-19 were isolated as triethylammonium salts.
as currently desired, formed by a reaction with ammonia. The
noncommercial building block 15 was synthesized following a
published method,42 in which 13 (4-chloropyrrolo-[2,3-d]-
pyrimidine)43 was condensed with the R-halogenose 1444 under
phase-transfer conditions using tris-[2-(2-methoxyethoxy)ethyl]-
amine (TDA-1) as catalyst. Purification of product using flash
column chromatography gave compound 15 solely as the â
conformer in 79% yield, showing a characteristic anomeric H-1′
double doublet resonance in the 1H NMR spectrum at 6.80 ppm.
7-Deaza-2′-deoxyadenosine 16 was synthesized by treatment of
15 with methanolic ammonia and was isolated in 53% yield. A
similar yield was observed by Seela et al.42 The introduction of
the 5′-monophosphate group was also accomplished using the
POCl3/TEP combination, but under anhydrous conditions, as
2′-deoxy compounds were found to be particularly labile to acid-
induced hydrolysis of the N9-glycosyl linkage. It was found
that phosphorylation of compound 16 proceeded at a rate similar
to that of adenosine but that a mixture of products identified as
the 5′-monophosphate 17a and 3′,5′-bisphosphate 17b was
produced, as suggested by the 1H NMR and 31P NMR spectra.
HPLC analysis of the reaction mixture, however, deceptively
only showed a single peak for products. It is possible that
compound 17b might have an elution time beyond our monitor-
ing time (normally in 20 min). This, however, is highly unlikely
as no second peak was notable even after 45 min (for HPLC
conditions, see the Supporting Information). It is thus likely
that 17a and 17b may have the same retention time, since we
found that ATP also elutes at the same time as AMP under our
HPLC conditions (result not shown). It is not totally clear why
phosphorylation of 7-deaza-2′-deoxyadenosine produced two
products while phosphorylation of adenosine produced only one
product in high yield,35 but in light of the assumption of Ikemoto
et al. that the N7 nitrogen atom might be involved in the
formation of a TEP-adenosine complex to activate the 5′-
position hydroxyl group, thus facilitating the selective 5′-
phosphorylation,35 we reasoned that the unexpected product 17b
might be produced because of the formation of an unstable
7-deaza-2′-deoxyadenosine-TEP complex. Anhydrous condi-
tions might also contribute to the formation of 17b as water
might be crucial to modulate the reactivity of the phosphory-
lating reagent, thus maintaining a 5′-selectivity.
In the activation step for coupling, compounds 17a and 17b
were not separated, since this was deemed to be too difficult,
but were used directly to react with the combination of
triphenylphosphine, dipyridyl disulfide, and morpholine. Two
morpholidates (18a/18b) were produced, which were now
clearly separated by ion-pair HPLC (for conditions, see the
Supporting Information), and the retention times of these two
molecules were found to be 2.6 and 7.3 min, respectively.
Purification using ion-exchange chromatography produced the
desired 5′-morpholidate 18a in 83% isolated yield based on 17a
as its triethylammonium salt. The 5′-morpholidate 18a was
distinguished from the 3′,5′-bisphosphoromorpholidate 18b by
its 31P NMR spectrum in which only one peak was observed at
(42) Seela, F.; Muth, H.-P.; Bindig, U. Synthesis 1988, 670-674.
(43) Tsytovich, A. V.; Shamshin, D. V.; Burkovskii, V. B.; Shvets, V.
I. Russ. J. Bioorg. Chem. 1995, 21, 756-761.
(44) Hoffer, M. Chem. Ber. 1960, 93, 2777-2780.
J. Org. Chem, Vol. 73, No. 5, 2008 1697

Zhang et al.
1
TABLE 2. Selected Coupling Constants from 1D H NMR of 3 (pH 6.0)
N9-linked ribose
N1-linked ribose
J
1′,2′ 5.9 Hz
J2′,3′ 5.6 Hz
J2′′,3′′ 4.6 Hz
J3′,4′ 3.5 Hz
J3′′,4′′ 2.3 Hz
J4′,5′a 7.0 Hz
J4′,5′b n.d.a
J
1′′,2′′ 4.1 Hz
n.d.a
n.d.a
a Not detected.
FIGURE 4. Presumed different forms of 7-deaza-cADPR derivatives.
At acidic pH, cADPR adopts mainly the amino form, while at basic
pH, cADPR adopts the imino form.
FIGURE 3. C2′-endo/C3′-endo equilibrium for N9-linked ribose of
8-bromo-7-deaza-cADPR. Figure is based on the coupling constants
1
obtained from 1D H NMR and gCOSY data.
(J1′,2′ + J3′,4′)] × 100.45,46 Using 1D 1H NMR, 2D gCOSY, and
HMQC, we clearly identified the chemical shift of each proton
of compound 3 and calculated the coupling constants of these
protons. The coupling constants of protons in the N1-linked
ribose and N9-linked ribose are listed in Table 2.
8.22 ppm. It was also notable that in the 13C NMR spectrum
the C-5′ signal but not C-3′ was split by the phosphate,
indicating that the morpholidate is at C-5′ not C-3′. The minor
product, the 3′,5′-bisphosphoromorpholidate 18b, was also
isolated (ca 13% based on 16), the structure being confirmed
by the following spectroscopic evidence: in the 31P NMR
spectrum, two singlets were observed at 8.05 and 7.77 ppm,
indicating the existence of two morpholidates; morpholidates
normally have a chemical shift at around 8 ppm; in the 1H NMR
spectrum, multiplets at 3.66, 3.45, 3.08, and 2.85 ppm were
assigned for the two morpholine moieties. The structure was
further confirmed by an LC-MS study where the molecular
ion of 574.4 was identified. Evidence was also found in the
13C NMR spectrum, in which the C-3′ resonance was split into
a doublet suggesting the direct connection of a phosphate at
this position. Phosphoromorpholidate 18a was then condensed
with â-NMN+ in the presence of Lewis acid (MnCl2/forma-
mide), producing the desired 7-deaza-2′-deoxy-NAD+ 19, which
was isolated by reverse-phase chromatography in 60% yield.
The final compound 7-deaza-2′-deoxy-cADPR 4 was synthe-
sized by incubation of compound 19 with Aplysia cyclase and
after ion-exchange chromatography, the product was eluted to
give compound 4 in 67% yield as its triethylammonium salt.
Preliminary biological evaluation indicated that 4 may have
partial agonistic properties in permeabilized cells. Biological
data of 4 will be reported elsewhere.
Among these coupling constants, the important J3′,4′ coupling
for the N9-linked ribose could not be easily extracted from the
1D 1H NMR spectrum, since the resonance for H-3′ was largely
overlapped by those for H-2′′ and H-4′′ and therefore the
splitting of the signal by H-2′ and H-4′ was not identifiable.
Fortunately, H-4′ was clearly separated from all other signals,
offering us a good chance to calculate the desired H-3′/4′
coupling constant (J3′,4′). After decoupling of one of the two
H-5′ protons, H-4′ clearly presented a double doublet pattern
and the coupling constants (J3′,4′ and J4′,5′a) of this double doublet
were found to be 3.5 and 7.0 Hz. Further decoupling of the
H-3′ multiplet (signal overlaps with H-4′′ and H-2′′) made H-4′
also a double doublet from which two coupling constants (J4′,5′a
and J4′,5′b) were extracted. One of the coupling constants was
found to be close to 7.0 Hz (6.2 Hz), and the other was rather
small and was not clearly calculated. Altogether, clearly the
large coupling constant (7.0 Hz) was derived from the coupling
of H-4′ and H-5′a, while the smaller one (3.5 Hz) was produced
by the coupling of H-3′ and H-4′. The assignment of J3′,4′ was
confirmed by the finding that the sum of J1′,2′ and J3′,4′ is close
to 10 Hz.46 It is noted from Table 2 that the sum of J1′′,2′′ and
J3′′,4′′ for the N1-linked ribose is only 6.4, which is smaller than
the one for N9-linked ribose. It is known that nucleotides with
Conformational Study of Compound 3. To date, an
important issue that still needs to be answered is why 8-sub-
stituted cADPR analogues are antagonists, while cADPR itself
is an agonist. Shuto et al. reported that 2′′,3′′-dideoxydihydro-
cADPcR, an inactive compound, adopted a major C3′-endo and
high anti conformation in aqueous solution,18 indicating that
the ribose and glycosyl linkage conformations are of crucial
importance for the calcium release activities of cADPR ana-
logues. We thus assumed that substitutions such as a bromo
group at the 8-position might divert the N9-linked ribose into
a C3′-endo conformation, and this in turn may produce the
antagonistic effect. To determine the active conformation of
compound 3 in aqueous solution, we calculated its ribose C2′-
endo/C3′-endo ratio and glycosyl bond conformations. It has
been well-established that the C2′-endo/C3′-endo ratio of
a charged base (for example, NMN+) compared to those with
an uncharged base normally have smaller J1′,2′ and an unchanged
47
J3′,4′
.
Thus, the smaller sum for the N1-linked ribose in our
case probably suggests that this ribose is linked with a charged
base, and it indicates that under the NMR conditions at pH 6.0
the base is mostly in its “amino form” as would be anticipated
(Figure 4).
From the coupling constants listed in Table 2, we find that,
in aqueous solution (ca. pH 6.0), 7-deaza-8-bromo-cADPR
adopts 62% of the C2′-endo conformation in the N9-linked
(45) Hotoda, H.; Murayama, K.; Miyamoto, S.; Iwata, Y.; Takahashi,
M.; Kawase, Y.; Tanzawa, K.; Kaneko, M. Biochemistry 1999, 38, 9234-
9241.
(46) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1973, 95, 2333-
2344.
(47) Birdsall, B.; Birdsall, N. J. M.; Feeney, J.; Thornton, J. J. Am. Chem.
Soc. 1975, 97, 2845-2850.
1
nucleotide can be mathematically calculated from H NMR
spectrum by adaptation of the equation [C2′-endo] (%) ) [J1′,2′
/
1698 J. Org. Chem., Vol. 73, No. 5, 2008

Chemoenzymatic Synthesis of 7-Deaza-cADPR Analogues
ribose, which is consistent with the observation that, in the
crystal structure, the major puckering mode of cADPR is
observed to be C2′-endo.40 The conformation of the N1-linked
ribose was also calculated as 64% C2′′-endo. In agreement with
our observation, Graham et al.48 studied three cyclic compounds
(cADPR, 2′-OMe-cADPR, and 3′-OMe-cADPR), and all of
these compounds demonstrated a dominant C2′′-endo conforma-
tion in their N1-linked ribose. The conformations of the N9-
linked ribose are shown in Figure 3.
Clearly, the C2′-endo/C3′-endo ratio of 7-deaza-8-bromo-
cADPR (N9-linked ribose 62% C2′-endo) is close to that of
cADPR (N9-linked ribose 64% C2′-endo reported by Shuto et
al.18). In addition, 7-deaza-8-bromo-cADPR adopts a syn
conformation around the N9-glycosyl linkage due to a repulsion
of the bulky bromo group by the ribose moiety.33,34 cADPR
also presents a syn conformation around the glycosyl bond as
indicated in its crystal structure.40 The syn conformation of
cADPR is characterized by a large downfield shift of the H-2′
proton (+0.59 ppm relative to that of AMP).49 Compound 3
that has a +1.03 ppm downfield shift relative to that of 7-deaza-
AMP also indicates a syn conformation. On the basis of these
findings, we therefore assume that, at least for 7-deaza-8-bromo-
cADPR 3, the antagonistic activity is not caused by the C2′-
endo/C3′-endo conversion or the syn/anti conformation changes.
The issue, however, still remains open since we still do not
know what is the biologically active conformation of 3 or other
8-substituted antagonists when they actually bind with their
receptor. The reason 7-deaza-8-bromo-cADPR behaves as a
antagonist is still not clear, but it is conceivable that the
formation of a hydrogen bond by the 8-substitutent with the
receptor, or some other perturbing interaction at that site, may
be central to the antagonistic effect, consistent with the
observation that 8-amino-cADPR is more potent as an antagonist
than 8-bromo-cADPR. The amino group is both a hydrogen
bond donor and acceptor, whereas the bromo group is only an
acceptor and this could also be a key issue.
Hydrolysis Study. 7-Deaza-based cADPR analogues seem
to have the benefit of being more hydrolysis resistant chemically.
To further clarify this, hydrolysis of both 7-deaza-8-bromo-
cADPR 3 and 7-deaza-2′-deoxy cADPR 4 was investigated and
compared with 8-bromo-cADPR 6 and the other 2′-deoxy
analogue 8-amino-2′-deoxy-cADPR 750 at different pH values
and temperatures. These cADPR analogues were thus dissolved
in 50 mM phosphate buffer (pH 7.4), 0.2 M HCl solution (pH
0.7), and 0.01 M NaOH solution (pH 12) and were incubated
either at room temperature (rt) or at 70 °C. The hydrolysis
process was monitored by ion-pair HPLC (for conditions, see
the Supporting Information). At room temperature, at pH 7.4
and at pH 12, all cADPR analogues tested were found to be
perfectly stable for at least 24 h; under acidic conditions (pH
0.7), 7-deaza-8-bromo-cADPR 3, 7-deaza-2′-deoxy cADPR 4,
and 8-bromo-cADPR 6 remained largely as cyclic structures
for 24 h, while 8-amino-2′-deoxy-cADPR 7 was completely
degraded in 2 h. The degradation product of compound 7 was
analyzed by HPLC, and a major product was eluted at 2.56
min. It is assumed that compound 7 was degraded at the N9-
FIGURE 5. Proposed mechanism and products of acid-induced
hydrolysis.
TABLE 3. Stability of Cyclic Compounds 3 and 6 after Incubation
at 70 °C (pH 7.4)
cyclic form
retained after
1 h (%)
cyclic form
retained after
4 h (%)
compound
7-deaza-8-Br-cADPR 3
8-Br-cADPR 6
92
30
70
0
TABLE 4. Stability of Cyclic Compounds 4 and 7 after Incubation
at 70 °C (pH 7.4)
cyclic form
retained after
1 h (%)
cyclic form
retained after
3 h (%)
compound
7-deaza-2′-deoxy-cADPR 4
8-NH2-2′-deoxy-cADPR 7
97
78
88
49
glycosyl linkage, as we know that 2′-deoxy-cADPR analogues
are more acid labile at the N9-ribosyl linkage than their ribose
counterparts.51 The fact that 7-deaza-2′-deoxy-cADPR 4 exhibits
much better acid resistance demonstrates that N7 has a crucial
influence on stability for the N9-glycosyl linkage, with hy-
drolysis presumably proceeding by N7 being first protonated
in the normal series. Indeed, the N9-glycosyl bond of 7-deaza-
cADPR analogues was shown to be perfectly stable in our later
studies.
The hydrolysis of these cADPR analogues was also studied
at elevated temperatures. The cADPR analogues (3, 4, 6, and
7) were incubated at pH 7.4 and 70 °C, and the amount of cyclic
structure retained for each compound was monitored by HPLC
(for conditions, see the Supporting Information) and is shown
in Tables 3 and 4. From Tables 3 and 4, 7-deaza-cADPR
analogues (both 3 and 4) were found to be more resistant to
heat-induced hydrolysis at neutral pH (pH 7.4) compared to
those adenine-based cADPR analogues 6 and 7.
The degradation product of 8-bromo-cADPR 6 was isolated
1
and was identified to be 8-bromo-N9-ADPR by H NMR and
gHMBC (result not shown). Overnight incubation of 7-deaza
compound 3 at 70 °C (pH 7.4) resulted in a complete
degradation of this compound, producing mainly a new product
that had an HPLC retention time of ca. 15 min. We assume the
(48) Graham, S. M.; Macaya, D. J.; Sengupta, R. N.; Turner, K. B. Org.
Lett. 2004, 6, 233-236.
(49) Wada, T.; Inageda, K.; Aritomo, K.; Tokita, K.; Nishina, H.;
Takahashi, K.; Katada, T.; Sekine, M. Nucleosides Nucleotides 1995, 14,
1301-1314.
(50) Zhang, B.; Wagner, G.; Weber, K.; Garnham, C.; Morgan, A. J.;
Galione, A.; Guse, A. H.; Potter, B. V. L. J. Med. Chem. 2008, in press.
(51) Garrett, E. R.; Mehta, P. J. J. Am. Chem. Soc. 1972, 94, 8532-
8540.
J. Org. Chem, Vol. 73, No. 5, 2008 1699

Zhang et al.
new product to be 7-deaza-8-bromo-N9-ADPR 20 (Figure 5)
formed by breaking the N1-glycosyl linkage as we have seen
for 8-bromo-cADPR. Enzymic CD38 ectoenzyme-mediated
hydrolysis of 3 also gave 20 as the sole product, as indicated
by HPLC (data not shown). It has been reported by Moreau et
al.52 that, at acidic pH and elevated temperature, the hydrolysis-
resistant 8-substituted cIDPR was degraded at both the N9-
glycosyl bond and pyrophosphate linkages, producing a novel
nucleotide with ribose connected at the N1 position. In light of
this, we thus studied the hydrolysis behavior of these 7-deaza-
cADPR analogues in both acidic and basic conditions at 70 °C.
It was found that these 7-deaza-cADPR analogues degraded
much more quickly at both acidic and basic conditions than
that at neutral pH. It is assumed that under acidic conditions,
cADPR analogues may adopt mainly the “amino form” (Figure
4), thus forming a better leaving group at the N1 position and
then facilitating a quick degradation of the molecules. However,
this could not explain why the analogues also degraded quickly
under basic conditions as they are supposed to adopt the “imino
form” (Figure 4). It was proposed that, under basic conditions,
an ionized diol oxyanion on the nicotinamide-bearing ribose of
NAD+, by noncovalent stabilization of the oxocarbocation
intermediate, may be responsible for the large hydrolysis rate
increases observed.53 Such a mechanism might also stabilize
the corresponding oxocarbocation intermediate formed by basic
cleavage of the N1-linked ribose linkage of cADPR or an
analogue, thus accelerating the hydrolysis process. The final
detectable product of both acid- and base-mediated degradation
(see the following paragraph) suggested that the pyrophosphate
bond was also cleaved under either acidic or basic conditions.
It is not surprising that the anhydride pyrophosphate linkage
can be hydrolyzed in both acidic and basic conditions and under
basic conditions the hydrolysis possibly occurs because of the
intramolecular attack of the ionized 3′′-OH at the pyrophosphate
bond. The pyrophosphate backbone of ADPR was found to be
more labile than, for example, reduced diphosphopyridine
nucleotide, under basic conditions.54 The reason for the different
reactivity between pyrophosphate bonds towards base-induced
hydrolysis is so far not clear.
After being heated at pH 0.7 for 4 h, 7-deaza-8-bromo-
cADPR 3 was more than 90% degraded, producing two major
products with HPLC elution times of ca. 8.5 and 15 min,
respectively. The newly formed product at ca. 15 min is
consistent with 7-deaza-8-bromo-N9-ADPR 20, which is formed
via heat-induced hydrolysis (pH 7.4). The product at ca. 8.5
min, however, is most likely to be derived from compound 20
since, after incubating the mixture for a longer period, the peak
at ca. 15 min was almost completely consumed and the peak at
8.5 min was found to be significantly increased as the only
product. This product is likely to be a monophosphate, as
indicated by its elution time under our HPLC conditions
(normally 6-10 min for a monophosphate). The 31P NMR
spectrum of the hydrolysis mixture showed that the pyrophos-
phate bond was completely cleaved (the peak at ca. -10 ppm
disappeared) and a new peak at 0.02 ppm appeared, indicating
the formation of a monophosphate (Supporting Information).
A 1H NMR spectrum of the crude hydrolysis mixture revealed
FIGURE 6. CD38-mediated hydrolysis of cADPR analogues. 7-Deaza-
8-bromo-cADPR 3 was found to be more stable than 8-bromo-cADPR
6 and cADPR 1.
a rather downfield H-2′ signal at ca. 5.1 ppm (data not shown).
This indicated that the N9-ribosyl linkage was not broken and
the downfield shift of H-2′ was caused because the molecule
was in its syn comformation due to the nonbonded repulsion
by 8-bromo substitutent.33,34,49 On the basis of this evidence,
we thus confirmed that the hydrolysis product found at 8.5 min
is indeed 7-deaza-8-bromo-N9-AMP 10. A mass spectrometry
study further confirmed the result, and the expected molecular
ions for 10 at ca. 425 and 427 mass units were observed
(Supporting Information). Further evidence for the final product
being the N9-monophosphate came from the acid-induced
hydrolysis of 7-deaza-2′-deoxy-cADPR 4. After being incubated
at 70 °C and pH 0.7 for 1 h, 4 was completely decomposed,
resulting in an HPLC profile similar to that earlier described
for the acid hydrolysis of 3. The final product at 6.40 min was
identified to be 7-deaza-2′-deoxy-N9-AMP by coeluting with
the authentic standard compound 17a (HPLC trace in Supporting
Information). With the above studies, we thus propose that,
under acidic conditions, 7-deaza-cADPR derivatives decompose
first at the N1-glycosyl bond producing the 7-deaza-N9-ADPR
analogues, which in turn are further degraded by pyrophosphate
bond cleavage, forming 7-deaza-N9-nucleotides as the final
compounds. The proposed order for the acid-induced hydrolysis
of 7-deaza-cADPR analogues is shown in Figure 5. Under basic
conditions at 70 °C, analogues 3 and 4 also quickly degraded
to the same N9-nucleotides as observed earlier. However, no
clear order of bond cleavage could be established from the
HPLC study in this case. The N9-glycosyl bond was found to
be perfectly stable under both acidic and basic conditions.
In addition to chemical stability toward heat-induced hy-
drolysis, cADPR hydrolase-mediated hydrolysis of 7-deaza-
cADPR analogue 3 was also carefully investigated. Ectoenzyme
CD38 is a very effective hydrolase.55 In Figure 6, CD38-
catalyzed hydrolysis of 7-deaza-8-bromo-cADPR 3 is directly
compared with that of 8-bromo-cADPR 6 and cADPR 1,
showing that 3 retains largely the cyclic form (over 60%) after
incubation with CD38 for 20 h in pH 7.5 Tris-HCl buffer, while
1 and 6 are almost totally hydrolyzed. Together with the above
chemical stability studies, we can conclude that 7-deaza-cADPR
derivatives are indeed more resistant to both chemical and
enzymatic hydrolysis and these are attractive features for
biological studies. The N9-glycosyl bonds of 7-deaza-cADPR
analogues were found to be stable throughout the study.
(52) Moreau, C.; Woodman, T. J.; Potter, B. V. L. Chem. Commum.
2006, 1127-1129.
(53) Johnson, R. W.; Marschner, T. M.; Oppenheimer, N. J. J. Am. Chem.
Soc. 1988, 110, 2257-2263.
(54) Kaplan, N. O.; Colowick, S. P.; Barnes, C. C. J. Biol. Chem. 1951,
191, 461-472.
(55) Graeff, R.; Munshi, C.; Aarhus, R.; John, M.; Lee, H. C. J. Biol.
Chem. 2001, 276, 12169-12173.
1700 J. Org. Chem., Vol. 73, No. 5, 2008

Chemoenzymatic Synthesis of 7-Deaza-cADPR Analogues
1H, H-2), 6.38 (s, 1H, H-7), 5.99 (d, J1′2′ 5.4 Hz, 1H, H-1′), 5.03
(m, 1H, H-2′), 4.53 (m, 1H, H-3′), 4.06 (m, 3H, H-4′ and H-5′);
13C NMR (D2O, 68 MHz) δ 153.2 (C-6), 148.9 (C-4), 148.2 (C-2),
109.9 (C-8), 104.4 (C-7), 103.2 (C-5), 89.2 (C-1′), 82.8 (d, JCP 7.5
Hz, C-4′), 71.4 (C-2′), 69.5 (C-3′), 64.4 (C-5′); 31P NMR (D2O,
109 MHz) δ 2.35 (s); m/z (FAB+) 425.0, 427.0 (M + H)+.
HRMS: Calcd for [M + H]+ C11H1579BrN4O7P+ (ES+), 424.9862;
found, 424.9860.
Conclusion
In summary, we present a high-yielding synthesis of the
biologically important compound 7-deaza-8-bromo-cADPR 3,
a hydrolysis-resistant, membrane-permeable antagonist of cADPR-
induced Ca2+ release from intracellular stores. The synthetic
route was also extended to facilitate the synthesis of 7-deaza-
2′-deoxy-cADPR 4, a candidate hydrolysis-resistant agonist/
partial agonist that may be membrane permeant. The hydrolysis
resistance of 7-deaza-cADPR analogues was carefully investi-
gated chemically, and the hydrolysis product is confirmed as a
7-deaza-N9-nucleotide, characteristic of the stable N9-glycosyl
linkage of these 7-deaza compounds. 7-Deaza-8-bromo-cADPR
3 was also found to be more stable to enzymatic CD38-mediated
hydrolysis compared with cADPR and 8-bromo-cADPR. 7-Deaza-
8-bromo-cADPR thus possesses generally attractive properties
that should continue to make it an attractive tool with which to
explore cADPR-mediated signaling processes.
7-Deaza-8-bromoadenosine 5′-Monophosphate Morpholidate
(11). A solution of 7-deaza-8-Br-AMP 10 triethylammonium salt
(65 mg, 112 µmol) in anhydrous DMSO (0.5 mL) was coevaporated
with anhydrous DMF (3 × 1 mL). The resulting residue was
redissolved in anhydrous DMSO (0.5 mL) to give a yellow solution,
to which was added in sequence triphenylphosphine (100 mg, 381
µmol), morpholine (65 µL, 744 µmol), and dipyridyl disulfide (84
mg, 381 µmol). After being stirred at rt for three more hours, the
reaction was quenched by addition of a solution of NaI in acetone
(0.1 M, 10 mL) and the precipitated sodium salt was filtered, washed
with acetone, and dried under reduced pressure. Crude 11 was
further purified by ion-exchange chromatography, eluting with a
gradient of 0-50% 1 M TEAB buffer against MilliQ water.
Appropriate fractions were collected and evaporated in vacuo.
Excess TEAB was removed by evaporating with methanol (3×) to
give compound 11 as a glassy solid in the triethylammomium form.
Yield (97 µmol, 87%). HPLC: 3.2 min at 254 nm; 1H NMR (D2O,
270 MHz) δ 7.93 (s, 1H, H-2), 6.48 (s, 1H, H-7), 6.02 (d, J1′,2′ 4.7
Hz, 1H, H-1′), 5.21 (m, 1H, H-2′), 4.61 (m, 1H, H-3′), 4.00 (m,
3H, H-4′ and H-5′), 3.36 (m, 4H, 2 × CH2O), 2.76 (m, 4H, 2 ×
CH2N); 13C NMR (D2O, 68 MHz) δ 155.4 (C-6), 150.9 (C-2), 149.9
(C-4), 109.6 (C-8), 104.2 (C-7), 103.9 (C-5), 89.5 (C-1′), 82.3 (d,
Experimental Section
CD38-Mediated Hydrolysis of cADPR Analogues. 7-Deaza-
8-bromo-cADPR 3, cADPR 1, and 8-bromo-cADPR 6 (100 µM)
were incubated with CD3855 (2 µg/mL) in 40 mM Tris-HCl buffer
(pH 7.5, total volume 1 mL). The hydrolysis progress was monitored
by HPLC at 0, 0.5, 2, 4, and 20 h (for HPLC system, see the
Supporting Information).
7-Deaza-8-bromoadenosine (9). To a stirred solution of dry
7-deazaadenosine (tubercidin 8, 125 mg, 0.47 mmol) and potassium
acetate (100 mg, 1.02 mmol) in anhydrous DMF (1.6 mL) was
added dropwise a solution of freshly recrystallized NBS (180 mg,
1.02 mmol) in anhydrous DMF (1.2 mL). The reaction mixture
immediately turned dark red and was stirred at rt for a further 15
min. The DMF was removed by evaporation in vacuo, and the
residue was purified by column chromatography, eluting with a
gradient of 0-10% methanol/DCM to give the title compound as
a yellow oil (71 mg, 44%). NMR assignment is based on gCOSY,
HMQC, and HMBC. 1H NMR (CD3OD, 400 MHz) δ 8.03 (s, 1H,
H-2), 6.74 (s, 1H, H-7), 6.08 (d, J1′,2′ 7.4 Hz, 1H, H-1′), 5.12 (dd,
J
CP 8.7 Hz, C-4′), 71.3 (C-2′), 69.4 (C-3′), 66.9, 66.8 (2 × CH2O),
64.1 (d, JCP 4.6 Hz, C-5′), 44.7 (2 × CH2N); 31P NMR (D2O, 109
MHz) δ 8.33 (s); m/z (ES-) 492.05, 494.19 (M - H)-; HRMS
calcd for [M - H]- C15H2079BrN5O7P- (ES-), 492.0284; found,
492.0282.
7-Deaza-8-bromo nicotinamide Adenine Dinucleotide (7-
Deaza-8-bromo-NAD+, 12). Method 1. 12 was first synthesized
by a Todd carbodiimide method24 (Supporting information).
Method 2. A suspension of 7-deaza-8-Br-AMP morpholidate
TEA salt 11 (52 mg, 80 µmol), â-NMN+ (zwitterionic form, 45
mg, 135 µmol), and MgSO4 (30 mg, 250 µmol) in MnCl2/
formamide solution (0.2 M, 0.95 mL) was stirred at rt for 48 h.
HPLC analysis indicated that nearly all the morpholidate had
reacted. To the resulting clear yellow solution, MeCN (2 mL) was
added dropwise with constant stirring and the precipitated yellow
solid was filtered off and washed with acetone. The solid was
dissolved in a small amount of MilliQ water and treated with Chelex
resin (Na+ form) for 1 h. The resin was filtered off and washed
with MilliQ water, and the filtrated was evaporated. The solid
residue was again dissolved in a small amount of 0.05 M TEAB
buffer (5-10 mL) and was then purified by a reverse-phase system.
Combined fractions were evaporated under vacuum, and the excess
TEAB was removed by coevaporating with methanol (3×) to give
the desired compound in its triethylammonium form (63 µmol,
J2′,1′ 7.4 Hz, J2′,3′ 5.3 Hz, 1H, H-2′), 4.35 (dd, J3′,2′ 5.3 Hz, J3′,4′ 1.7
Hz, 1H, H-3′), 4.16 (m, 1H, H-4′), 3.87 (dd, J5′a,4′ 2.4 Hz, J5′a,5′b
12.6 Hz, 1H, H-5′a), 3.73 (dd, J5′b,4′ 2.4 Hz, J5′b,5′a 12.6 Hz, 1H,
H-5′b); 13C NMR (CD3OD, 100.5 MHz) δ 158.1 (C-6), 151.9 (C-
2), 150.5 (C-4), 111.5 (C-8), 105.9 (C-5), 103.9 (C-7), 92.2 (C-1′),
88.3 (C-4′), 73.7 (C-2′), 73.1 (C-3′), 64.3 (C-5′); m/z (ES+) 345.11,
347.12 (M + H)+; HRMS Calcd for [M + H]+ C11H1479BrN4O4
+
(ES+), 345.0198; found, 345.0190.
7-Deaza-8-bromoadenosine 5′-Monophosphate (10). A suspen-
sion of 7-deaza-8-bromoadenosine 9 (50 mg, 145 µmol) in
triethylphosphate (0.8 mL) was heated strongly using a heating gun
for 10 min, and the resulting clear solution was cooled to 0 °C in
an ice bath. H2O (4 µL) was then added followed by dropwise
addition of POCl3 (70 µL, 750 µmmol). After being stirred for
another hour at 0 °C, the reaction was quenched with ice (10 mL)
and the excess triethylphosphate was removed by partitioning
between aqueous solution and cold ethyl acetate (5 × 20 mL). The
aqueous layer was neutralized using 5 M NaOH and then loaded
on top of a reverse-phase column, which was eluted with a 0-30%
gradient of MeCN against 0.05 M TEAB buffer. Appropriate
fractions were combined, and the solvent was evaporated under
reduced pressure. Excess TEAB buffer was removed by coevapo-
rating with methanol (3×), and the resulting glassy residue was
further purified by ion exchange chromatography using 0-50%
gradient of 1 M TEAB/MilliQ water as the eluent. After the
appropriate fractions were combined and evaporated, the resulting
white solid was coevaporated with methanol (3×) to produce the
desired compound 10 as its triethylammonium salt (112 µmol, 77%).
1
79%). HPLC: 3.9 min at 254 nm; H NMR (D2O, 400 MHz) δ
9.16 (s, 1H, HN-2), 9.00 (d, J6,5 6.4 Hz, 1H, HN-6), 8.63 (d, J4,5 8.0
Hz, 1H, HN-4,), 8.05 (dd, J5,4 8.0 Hz, J5,6 6.4 Hz, 1H, HN-5), 7.87
(s, 1H, H-2), 6.36 (s, 1H, H-7), 5.96 (d, J1′,2′ 5.6 Hz, 1H, H-1′),
5.91 (d, J1′′,2′′ 4.3 Hz, 1H, H-1′′), 5.06 (m, 1H, H-2′), 4.48 (m, 1H,
H-3′), 4.42 (m, 1H, H-4′′), 4.33-4.11 (m, 7H, ribose-H); 13C NMR
(D2O, 100 MHz) δ 164.8 (CO), 155.2 (C-6), 151.0 (C-2), 149.6
(C-4), 145.2 (CN-4), 142.1 (CN-6), 139.5 (CN-2), 133.2 (CN-3), 128.5
(CN-5), 108.8 (C-8), 103.8 (C-7), 103.1 (C-5), 99.9 (C-1′′), 88.7
(C-1′), 86.7 (d, JCP 9.2 Hz, C-4′′), 82.2 (d, JCP 8.4 Hz, C-4′), 77.5
(C-2′′), 70.9 (C-2′), 70.3 (C-3′′), 69.0 (C-3′), 65.6 (d, JCP 5.4 Hz,
C-5′′), 64.7 (d, JCP 5.4 Hz, C-5′); 31P NMR (D2O, 109 MHz) δ
-10.57 (d, JPP 21.1 Hz), -11.02 (d, JPP 21.1 Hz); m/z (ES-) 739.31,
+
741.20 (M - H)-; HRMS Calcd for [M + H]+ C22H2879BrN6O14P2
1
HPLC: 9.2 min at 254 nm; H NMR (D2O, 270 MHz) δ 7.87 (s,
(ES+), 741.0322; found, 741.0325.
J. Org. Chem, Vol. 73, No. 5, 2008 1701

Zhang et al.
Cyclic 7-Deaza-8-bromoadenosine 5′-Diphosphate Ribose (7-
Deaza-8-bromo-cADPR, 3). To a solution of 7-deaza-8-bromo-
NAD+ (12, 20 mg, 22 µmol) in HEPES buffer (25 mM, pH 7.4,
52 mL) was added Aplysia californica ADP-ribosyl cyclase (100
µL). The resulting clear solution was allowed to stir at rt until HPLC
analysis indicated that all the starting pyrophosphate 12 had been
consumed. The reaction solution was then diluted with MilliQ water
until its conductivity reached an appropriate value (<200 µS/cm)
and product was purified by ion-exchange chromatography, eluting
with a gradient of 1 M TEAB buffer against MilliQ water.
Appropriate fractions were collected, evaporated in vacuo, and
coevaporated with methanol (3×) to give the final compound as a
glassy solid. Yield (16 µmol, 72%). HPLC: 5.2 min at 254 nm;
UV (H2O) λmax 277 nm (ꢀ/dm3 mol-1 cm-1 10 850); 1H NMR (D2O,
400 MHz) δ 8.84 (s, 1H, H-2), 6.97 (s, 1H, H-7), 6.12 (d, J1′,2′ 5.9
Hz, 1H, H-1′), 6.05 (d, J1′′,2′′ 4.1 Hz, 1H, H-1′′), 5.52 (dd, J2′,1′ 5.9
Hz, J2′,3′ 5.7 Hz, 1H, H-2′), 4.66 (m, 2H, H-3′ and H-4′′), 4.65 (m,
1H, H-2′′), 4.42 (m, 2H, H-3′′ and H-5′a), 4.34 (m, 1H, H-5′′a),
(d, J7,8 3.7, 1H, H-7), 6.42 (dd, J1′,2b′ 6.2 Hz, J1′,2a′ 5.9 Hz, 1H,
H-1′), 4.33 (m, 1H, H-3′), 3.83 (m, 1H, H-4′), 3.53 (m, 2H, H-5′),
2.48 (m, 1H, H-2′a), 2.18 (m, 1H, H-2′b); 13C NMR (DMSO6, 68
MHz) δ 158.1 (C-6), 152.2 (C-2), 150.2 (C-4), 122.2 (C-8), 103.5
(C-5), 100.2 (C-7), 87.3 (C-1′), 83.7 (C-4′), 71.6 (C-3′), 62.6 (C-
5′), C-2′ is under DMSO peak.; m/z (ES+) 251.4 (M + H)+; HRMS
+
Calcd for [M + H]+ C11H15N4O3 (ES+), 251.1144; found,
251.1144.
7-Deaza-2′-deoxyadenosine 5′-Monophosphate (17a). A sus-
pension of 7-deaza-2′-deoxyadenosine (16, 60 mg, 240 µmol, dried
in vacuo at 60 °C overnight) in triethylphosphate (0.7 mL) was
heated with a heating gun until formation of clear solution, which
was then cooled in an ice-salt bath and followed by a dropwise
addition of POCl3 (70 µL, 750 µmol). After being stirred in the
ice-salt bath for 2 h, the reaction was quenched by the addition of
ice (10 mL) and was immediately neutralized to pH 7 with a NaOH
solution (5 M). The reaction solution was extracted with cold ethyl
acetate (5 × 20 mL), and the aqueous layer was further purified
on a reverse column, eluting with 0-30% gradient MeCN against
0.05 M TEAB. Appropriate fractions were evaporated, and the
residual white solid was coevaporated with MeOH (3×) to give
the crude title compound as a glassy solid (96 mg, 100%). Crude
compound contained 17a (135 µmol, 56%) and 17b (57 µmol,
4.29 (m, 1H, H-4′), 4.12 (m, 1H, H-5′′b), 4.04 (m, 1H, H-5′b); 13
C
NMR (D2O, 100 MHz) δ 149.4 (C-6), 147.0 (C-4), 139.7 (C-2),
114.2 (C-8), 104.6 (C-7), 104.0 (C-5), 95.0 (C-1′′), 90.3 (C-1′),
88.1 (d, JCP 10.7 Hz, C-4′′), 84.4 (d, JCP 10.7 Hz, C-4′), 76.8 (C-
2′′), 72.1 (C-2′), 71.8 (C-3′′), 70.2 (C-3′), 64.7 (d, JCP 4.8 Hz, C-5′′),
64.0 (JCP 4.7 Hz, C-5′); 31P NMR (D2O, 109 MHz) δ -10.00 (brs),
-10.71 (brs); m/z (ES-) 617.25, 619.20 (M - H)-; HRMS Calcd
1
23.8%). Yield was calculated based on 1D H NMR integration.
1
HPLC: 6.8 min at 254 nm; H NMR (D2O, 270 MHz) δ 7.86 (s,
-
for [M - H]- C16H2079BrN4O13P2 (ES-), 616.9685; found,
1H), 7.32 (d, J8,7 3.7 Hz, 1H), 6.43 (d, J7,8 3.7 Hz 1H), 6.39 (m,
1H), 4.53 (m, 1H), 4.01 (m, 1H), 3.74 (m, 2H), 2.57 (m, 1H), 2.31
(m, 1H); 31P NMR (D2O, 109 MHz) 4.48 (s); 17b 1H NMR (D2O,
270 MHz) δ 7.91 (s, 1H), 7.47 (d, J8,7 3.7 Hz, 1H), 6.44 (d, J7,8 3.7
Hz, 1H), 6.43 (m, 1H), 4.69 (m, 1H), other ribose-H overlaps with
17a; 31P NMR δ (D2O, 109 MHz) 4.48 (s), 3.95 (s).
616.9699.
4-Chloro-7-(2-deoxy-3,5-di-O-(p-toluoyl)-â-D-erythropento-
furanosyl)-7H-pyrrolo[2,3-d]pyrimidine (15). To a fast stirred
suspension of powdered potassium hydroxide (81 mg, 1.44 mmol)
in dry MeCN (4.9 mL) was added TDA-1 (8.1 µL) under an argon
atmosphere. This mixture was stirred at rt for 5 min. Compound
13 (100 mg, 0.66 mmol) was added, and the mixture was again
stirred for 5 min. R-Halogenose 14 (259 mg, 0.67 mmol) was then
added, and the reaction mixture was stirred for another 20 min until
TLC (DCM/ethyl acetate, 12:1) suggested that all the starting
material was consumed and a fast moving spot (Rf 0.46) was given.
The crude product was evaporated with silica gel, loaded on top of
a silica column, and eluted with DCM/ethyl acetate, 20:1. The
appropriate fractions were collected and concentrated to dryness
under reduced pressure to produce the title compound as a white
7-Deaza-2′-deoxyadenosine 5′-Monophosphate Morpholidate
(18a). A solution of 7-deaza-2′-deoxy AMP triethylammonium salt
17a/17b (where 17a 70 mg, 98 µmol) in anhydrous DMSO (0.4
mL) was coevaporated with anhydrous DMF (3 × 1 mL). The
yellow residue was dissolved in anhydrous DMSO (0.4 mL), and
to the resulting solution was added dipyridyl disulfide (146 mg,
668 µmol), morpholine (150 µL, 1.72 mmol), and powdered
triphenylphosphine (176 mg, 668 µmol) in sequence. The reaction
progress was monitored by HPLC, and after 3 h at rt, the starting
material had been completely consumed and two new peaks were
produced (2.6 min for 18a and 7.3 min for 3′,5′-bisphosphate
morpholidate 18b). The reaction was then quenched by dropwise
addition of a solution of NaI in acetone (0.1 M, 10 mL) under
vigorous stirring, and the sodium salt precipitate was filtered,
washed with acetone, and dried in vacuo. After further purification
by ion-exchange chromatography, eluting with a gradient of 0-50%
1 M TEAB buffer against MQ water, the title compound 18a was
obtained as a white solid (81 µmol, 83%). A small amount of
7-deaza-2′-deoxyadenosine 3′,5′- bisphosphate morpholidate 18b
(10 mg, ca. 13% based on 16) was also isolated. HPLC: 2.6 min
1
foam (262 mg, 79%). H NMR (CDCl3, 270 MHz) δ 8.62 (s, 1H,
H-2), 7.97 (d, JArH,ArH 8.2 Hz, 2H, 2 × ArCH), 7.91 (d, JArH,ArH
8.4 Hz, 2H, 2 × ArCH), 7.41 (d, J8,7 3.7 Hz, 1H, H-8), 7.28 (d,
JArH,ArH 8.2 Hz, 2H, 2 × ArCH), 7.25 (d, JArH,ArH 8.4 Hz, 2H, 2 ×
ArCH), 6.80 (dd, J1′,2b′ 8.4 Hz, J1′,2a′ 5.9 Hz, 1H, H-1′), 6.59 (d,
J7,8 3.7 Hz, 1H, H-7), 5.76 (m, 1H, H-3′), 4.64 (m, 3H, H-4′ and
H-5′), 2.82 (m, 2H, H-2′), 2.43 (s, 3H, CH3), 2.39 (s, 3H, CH3);
13C NMR (CDCl3, 68 MHz) δ 166.2 (2CO), 151.1 (C-6), 151.1
(C-2), 151.1 (C-4), 144.9 (ArC), 144.4 (ArC), 130.0, 129.7, 129.4
(ArCH), 126.78, 126.5 (ArC), 126.1 (C-8), 118.4, (C-5), 101.0 (C-
7), 84.7 (C-1′), 82.6 (C-4′), 75.2 (C-3′), 63.9 (C-5′), 38.1 (C-2′),
21.9 (2CH3); m/z (ES+) 506.45, 508.47 (M + H)+; HRMS Calcd
for [M + H]+ C27H2535ClN3O5+ (ES+), 506.1483; found, 506.1474.
7-Deaza-2′-deoxyadenosine (16). Ammonia gas was bubbled
into a suspension of compound 15 (260 mg, 0.52 mmol) in methanol
(20 mL) at 0 °C for 40 min. The resulting clear saturated methanolic
ammonia solution was warmed to rt and then heated at 60 °C in a
pressure tube for 4 days. TLC of the reaction mixture showed that
a more polar product was formed (Rf 0.23, DCM/methanol, 5:1).
After removal of the solvent, the crude product was purified by
column chromatography, eluting with DCM/methanol, 6:1, to give
the title compound as a colorless oil (70 mg, 53%). 1H NMR
(DMSO6, 270 MHz) δ 8.04 (s, 1H, H-2), 7.35 (d, J8,7 3.7 Hz, 1H,
H-8), 7.05 (s, 2H, NH2), 6.57 (d, J7,8 3.7 Hz, 1H, H-7), 6.48 (dd,
1
at 254 nm; UV (H2O) λmax 272.3 nm; H NMR (D2O, 270 MHz)
δ 7.96 (s, 1H, H-2), 7.31 (d, J8,7 3.7 Hz, 1H, H-8), 6.47 (d, J7,8 3.7
Hz, 1H, H-7), 6.45 (dd, 1H, overlap with H-7, J not given, H-1′),
4.63 (m, 1H, H-3′), 4.09 (m, 1H, H-4′), 3.84 (m, 2H, H-5′), 3.43
(m, 4H, 2 × CH2O), 2.83 (m, 4H, 2 × CH2N), 2.67 (m, 1H, H-2′a),
2.45 (m, 1H, H-2′b); 13C NMR (D2O, 68 MHz) δ 156.6 (C-6), 150.4
(C-2), 148.9 (C-4), 122.1 (C-8), 103.2 (C-5), 100.6 (C-7), 85.1 (d,
J
CP 9.3 Hz, C-4′), 82.9 (C-1′), 71.3 (C-3′), 66.9, 66.8 (2 × CH2O),
64.5 (d, JCP 5.6 Hz, C-5′), 44.6 (2 × CH2N), 38.5 (C-2′); 31P NMR
(D2O, 109 MHz) δ 8.22 (s); m/z (ES-) 398.42 (M - H)-; HRMS
Calcd for [M - H]- C15H21N5O6P- (ES-), 398.1229; found,
398.1226.
7-Deaza-2′-deoxyadenosine 3′,5′-Bisphosphate Morpholidate
1
(18b): HPLC: 7.3 min at 254 nm; UV (H2O) λmax 271.1 nm; H
J1′,2b′ 6.2 Hz, J1′,2a′ 5.9 Hz, 1H, H-1′), 5.24 (m, 2H, 3′-OH, 5′-OH),
NMR (D2O, 270 MHz) δ 8.09 (s, 1H, H-2), 7.47 (d, J8,7 3.7 Hz,
1H, H-8), 6.63 (d, J7,8 3.7 Hz, 1H, H-7), 6.59 (dd, 1H, overlap
with H-7, J not given, H-1′), 4.85 (m, 1H, H-3′), 4.28 (m, 1H,
H-4′), 3.89 (m, 2H, H-5′), 3.66 (m, 4H, 5′ morpholidate 2 × CH2O),
4.34 (m, 1H, H-3′), 3.81 (m, 1H, H-4′), 3.52 (m, 2H, H-5′), 2.49
(m, 1H, H-2a′), 2.16 (m, 1H, H-2b′); 1H NMR (DMSO6, D2O shake,
270 MHz) δ 8.02 (s, 1H, H-2), 7.32 (d, J8,7 3.7 Hz, 1H, H-8), 6.57
1702 J. Org. Chem., Vol. 73, No. 5, 2008

Chemoenzymatic Synthesis of 7-Deaza-cADPR Analogues
3.45 (m, 4H, 3′ morpholidate 2 × CH2O), 3.08 (m, 4H, 5′
morpholidate 2 × CH2N), 2.85 (m, 4H, 3′ morpholidate 2 × CH2N),
2.71 (m, 1H, H-2′a), 2.59 (m, 1H, H-2′b); 13C NMR (D2O, 68 MHz)
δ 155.4 (C-6), 149.0 (C-4), 148.5 (C-2), 123.0 (C-8), 103.0 (C-5),
101.3 (C-7), 83.4 (dd, JCP 11.5 Hz, JCP 5.6 Hz, C-4′), 83.4 (C-1′),
75.1 (d, JCP 5.6 Hz, C-3′), 67.0 (4 × CH2O), 64.7 (C-5′), 44.8 (4
× CH2N), 38.3 (C-2′); 31P NMR (D2O, 109 MHz) δ 8.05 (s), 7.77
(s); m/z (ES-) 547.44 [(M - H)-, 40%]; HRMS Calcd for (M +
Cyclic 7-Deaza-2′-deoxyadenosine 5′-Diphosphate Ribose (4).
To a solution of 7-deaza-2′-deoxy-NAD+ (19, 5 mg, 6 µmol) in
HEPES buffer (25 mM, pH 7.4, 16 mL) was added Aplysia cyclase
(23 µL). The reaction mixture was stirred at rt, and progress was
monitored by HPLC. After 1 h, the reaction solution was diluted
with MilliQ water (conductivity < 200 µS/cm) and was purified
using an ion-exchange system. Evaporation with MeOH gave the
final compound as a glassy solid (4 µmol, 67%). HPLC: 4.3 min
at 254 nm; UV (H2O, pH 4.9) λmax 275.8 nm (ꢀ/dm3 mol-1 cm-1
+
H)+ C19H31N6O9P2 (ES+), 549.1628; found, 549.1630.
7-Deaza-2′-deoxynicotinamide Adenine Dinucleotide (19). A
suspension of 7-deaza-2′-deoxy-AMP morpholidate TEA salt (18a,
54 µmol), â-NMN+ (zwitterionic form, 31 mg, 93 µmol), and
MgSO4 (20 mg, 167 µmol) in MnCl2/formamide (0.2 M, 0.68 mL)
was stirred at rt for 48 h in an argon atmosphere to give a near
clear yellow solution. HPLC analysis indicated that all the starting
18a was reacted. The reaction was quenched by addition of MeCN
(2 mL), and the resulting yellow solid was filtered, washed with
acetone, and dissolved in a small amount of MilliQ water. The
yellow solution was treated with Chelex (sodium form) ion-
exchange resin for 1 h to remove any residual Mn2+. The resin
was filtered off and washed with MilliQ water (10 mL), and the
washings evaporated in vacuo. The residue, containing crude
7-deaza-2′-deoxy-NAD+ sodium salt, was dissolved in 0.05 M
TEAB buffer (5 mL) and was purified using a reverse-phase system,
eluting with a gradient of 0-30% MeCN against 0.05 M TEAB
buffer. Fractions containing the desired compound were collected
and concentrated under reduced pressure. The white residue was
coevaporated with MeOH (3×) to give compound 19 as its
triethylammonium salt (32 µmol, 60%). HPLC: 3.3 min at 254
1
10 900); H NMR (D2O, 400 MHz) δ 8.69 (s, 1H, H-2), 7.22 (d,
J8,7 3.5 Hz, 1H, H-8), 6.63 (dd, J7,8 3.5 Hz, 1H, H-7), 6.15 (t, J1′,2′a
) J1′,2′b 7.0 Hz, 1H, H-1′), 5.91 (d, J1′′,2′′ 3.9 Hz, 1H, H-1′′), 4.71
(m, 1H, H-3′), 4.56 (m, 1H, H-4′′), 4.51 (m, 1H, H-2′′), 4.40 (m,
1H, H-5′a), 4.30 (m, 1H, H-3′′), 4.29 (m, 1H, H-5′′a), 4.22 (m,
1H, H-4′), 4.12 (m, 1H, H-5′′b), 4.00 (m, 1H, H-5′b), 3.23 (m, 1H,
H-2′a), 2.36 (m, 1H, H-2′b); 13C NMR (D2O, 68 MHz) δ 150.5
(C-6), 146.0 (C-4), 139.2 (C-2), 129.8 (C-8), 104.3 (C-5), 100.7
(C-7), 94.8 (C-1′′), 89.5 (C-1′), 87.8 (d, JCP 10.1 Hz, C-4′′), 86.4
(d, JCP 10.8 Hz, C-4′), 76.8 (C-2′′), 71.6 (C-3′, C-3′′), 65.4 (C-5′′),
64.6 (C-5′), 37.6 (C-2′); 31P NMR (D2O, 109 MHz) δ -10.05 (brs),
-10.60 (brs); HRMS Calcd for [M - H]- C16H21N4O12P2- (ES-),
523.0631; found, 523.0632.
Acknowledgment. We acknowledge award of an MRC
studentship (V.C.B.). We thank Dr. C. Moreau for practical
advice on this work, Dr. T. J. Woodman for collecting NMR
data, and Ms. Alison Smith for LC-MS data for most of the
compounds. We acknowledge the EPSRC National Mass
Spectrometry Service at Swansea for support.
1
nm; H NMR (D2O, 400 MHz) δ 9.25 (s, 1H, HN-2), 9.05 (brs,
1H, HN-6), 8.67 (d, J4,5 7.4 Hz, 1H, HN-4), 8.00 (m, 1H, HN-5),
7.90 (s, 1H, H-2), 7.36 (brs, 1H, H-8), 6.43 (m, 2H, H-1′ and H-7),
6.06 (d, J1′,2′ 4.7 Hz, 1H, H-1′′), 4.64 (m, 1H, H-3′), 4.52 (m, 1H,
H-4′′), 4.11 (m, 7H, ribose-H), 2.60 (m, 1H, H-2′a), 2.38 (m, 1H,
H-2′b); 13C NMR (D2O, 68 MHz) δ 154.9 (C-6), 148.6 (C-2), 148.1
(C-4), 145.4 (CN-4), 142.2 (CN-6), 140.1 (CN-2), 132.0 (CN-3), 128.5
(C-8), 122.8 (CN-5), 102.5 (C-5), 101.0 (C-7), 100.1 (C-1′′), 87.0
(C-4′′), 85.2 (C-4′), 83.0 (C-1′), 77.7 (C-2′′), 71.6 (C-3′′), 70.6 (C-
3′), 66.1 (C-5′′), 65.0 (C-5′), 38.8 (C-2′); 31P NMR (D2O, 109 MHz)
δ -11.44 (brs); m/z (ES-) 645.07; HMRS Calcd for (M - H)-
Supporting Information Available: Todd carbodiimide syn-
thesis of compound 12; 1H NMR spectra for compounds 3, 4, 9-12,
and 17-19; 31P NMR and mass spectra for the hydrolysis reaction;
HPLC traces for compounds 3, 4, 12, and 19; and some HPLC
evidence for the heat (chemical)-induced degradation of 3 and 4.
This material is available free of charge via the Internet at
http://pubs.acs.org.
-
C22H27N6O13P2 (ES-), 645.1111; found, 645.1111.
JO071236P
J. Org. Chem, Vol. 73, No. 5, 2008 1703