Chemical synthesis of the second messenger nicotinic acid adenine dinucleotide phosphate by total synthesis of nicotinamide adenine dinucleotide phosphate

Article abstract of DOI:10.1002/anie.200460054

Signaling dinucleotides: The first single-isomer synthesis of nicotinamide adenine dinucleotide phosphate (NADP) is reported. Installation and maintenance of sensitive phosphate and pyridinium functionalities were key to success. Significantly, conversion

Full text of DOI:10.1002/anie.200460054

Angewandte
Chemie
Figure 1. Compounds 1 and 2.
Natural Products Synthesis
including cardiac,[4a] skeletal,[4b] Tcells, [4c] smooth muscle,[4d]
pancreatic acinar[4e] and pancreatic beta cells.[3a,4f,4g]
Chemical Synthesis of the Second Messenger
Nicotinic Acid Adenine Dinucleotide Phosphate
by Total Synthesis of Nicotinamide Adenine
Dinucleotide Phosphate**
Compound 1 also appears to play an important role in
calcium signaling by triggering and coordinating calcium
signals evoked by other calcium-mobilizing messengers.
Sensitive radio-receptor assays for tissue measurements of
this dinucleotide have been developed,[5] but it is clear that
further understanding of its biology will draw heavily on the
development of selective molecules to probe this novel
signaling pathway and characterize its protein components.
A chemo-enzymatic route has so far generated 1 and a small
number of analogues from 2 by enzyme-mediated exchange
of nicotinamide for nicotinic acid. Clearly, the chemistry that
can be performed on this sensitive molecule and compatibility
with the enzyme limits the range of available analogues. A
synthetic route would provide a wider range of tailored
derivatives, such as non-hydrolysable analogues, that could
not be achieved by established methods.
James Dowden,* Christelle Moreau, Richard S. Brown,
Georgina Berridge, Antony Galione, and
Barry V. L. Potter
Modulation of intracellular calcium concentration creates
signals that control a range of biological pathways. Orches-
tration of these signals by an array of small molecule second
messengers provides the necessary complexity to complement
multiple cellular responses, such as activation of the immune
system or fertilization.[1] The discovery that nicotinic acid
adenine dinucleotide phosphate (NAADP, 1; Figure 1)
potently induces intracellular calcium release in the eggs of
sea urchins[2] raised its status, from minor contaminant of
nicotinamide adenine dinucleotide phosphate (NADP, 2), to
an important novel second messenger.[3] Recent studies attest
to the higher potency of this molecule than the well-known
messengers for calcium release [i.e., inositol trisphosphate
and cyclic adenosine diphosphoribose (cADPR)] and many of
these extend its actions to a wide range of mammalian cells,
We are currently exploring a range of strategies to
generate chemical tools with which to interrogate this bio-
logical system, including the total synthesis of both dinucleo-
tides. Whitesides and co-workers developed a semisynthetic
route to b-nicotinamide adenine dinucleotide (NAD) by
using NAD pyrophosphorylase immobilized on polyacryla-
mide gel, but this relied on a battery of enzymes and reagents
to generate stoichiometric quantities of adenosine triphos-
phate (ATP).[6a] More recently, Lee et al.[6b] reported a
practical total chemical synthesis of nicotinamide adenine
dinucleotide (NAD) that built upon a much earlier synthesis
[*] Dr. J. Dowden, Dr. C. Moreau, R. S. Brown, Prof. B. V. L. Potter
Wolfson Laboratoryof Medicinal Chemistry
Department of Pharmacy& Pharmacology
Universityof Bath
[7]
by Todd and co-workers. Indeed, a number of researchers
Claverton Down, Bath BA2 7AY (UK)
Fax: (+44)1225-386-114
E-mail: j.dowden@bath.ac.uk
have developed chemical approaches to NAD derivatives, not
least for the preparation of cyclic adenosine 5’-diphosphate
ribose.[8] However, the presence of a sensitive 2’-phosphate on
a ribose makes the synthesis of 1 or 2 significantly more
complex. During early studies Todd and co-workers reported
an apparent preparation of 2 inferred from biological
evaluation of the complex product mixture but did not
successfully isolate the desired coenzyme from the mixture.[7b]
Given that this dinucleotide is readily available from com-
mercial sources, it is perhaps not surprising that further
developments toward its chemical synthesis are not apparent
G. Berridge, Prof. A. Galione
Department of Pharmacology
Universityof Oxford (UK)
[**] This research was funded bythe Biotechnologyand Biology
Sciences Research Council and an Engineering and Physical
Sciences Research Council (EPSRC) studentship. We acknowledge
the use of the EPSRC Chemical Database Service at Daresbury.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2004, 43, 4637 –4640
DOI: 10.1002/anie.200460054
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4637

Communications
in the general literature. To the best of our knowledge reactive phosphytilating intermediate that does not react with
therefore, we describe the first chemical synthesis of 2 as a
the purinyl amine, thus saving two steps. Carefully controlled
acidic conditions, as previously reported by Scott and co-
workers afforded exclusive protodesilylation at the less
hindered 5’-position in quantitative yield.[13]
single isomer, its enzymatic conversion to the potent second
messenger 1, and both chemical characterization and bio-
logical proof of the integrity of the latter dinucleotide.
The potentially fragile nature of 2 and its intermediates
encouraged a conservative approach that involves the late
construction of the pyrophosphate bond by activation of b-
nicotinamide mononucleotide (b-NMN, 4; Scheme 1) and
reaction with a suitably decorated adenosine building block.
It is known that the 2’-phosphate motif is critical for the
activity of 1[9] and may be involved in its degradation.[10]
Structural modifications are likely to be informative and
thus we designed a synthesis that was both robust and
sufficiently flexible to accommodate future modifications at
any position on the NAADP framework.
Compound 4 was generated from tetra-O-acetyl-b-d-
ribose.[6,11] Reaction between adenosine and 1,3-dichloro-
1,1,3,3,-tetraisopropyl disiloxane in pyridine proceeded with-
out complication and lead to simultaneous tethering of the 3’-
and 5’-hydroxyls of adenosine, thus permitting unambiguous
phosphorylation of the 2’-alcohol. Subsequent elaboration of
the 2’-alcohol 5 could be achieved with a range of phosphor-
amidites, including bis(benzyloxy)-N,N-diisopropylamino-
phosphane 6 (Scheme 1), to afford the respective phosphate
7 in high yield after oxidation of the intermediate phosphite
with m-chloroperoxybenzoic acid (mCPBA). Imidazolium
triflate[12] was used instead of tetrazole to produce a less
Generally, phosphoramidite chemistry did not proceed in
acceptable yields at the 5’-hydroxyl of the precursor, presum-
ably because this position is somewhat crowded. Resorting to
the reactive tetrazole-activated phosphoramidites required
amine protection and actually led to incorporation of an
unwelcome silanol phosphate. It was frustrating that a
multitude of protection strategies failed due to unexpected
incompatibility with the desired route. Although this chemis-
try might be optimized to provide acceptable yields of the
desired compound, we found that H-phosphonate chemistry
offered a very effective alternative. Selective 5’-protodesilyl-
ation of bis(benzyloxy)phosphate 7, then treatment of the
product with trisimidazolylphosphane, generated in situ from
PCl3 and imidazole, rapidly furnished 8 in quantitative yield
(Scheme 1). Oxidation was best achieved by using conditions
reported by Sekine and co-workers[14] in a process monitored
by 31P NMR spectroscopy, such that the H-phosphonate
triethylammonium salt (d(31P) = 5.5 ppm) was treated with
N,O-bis(trimethylsilyl)acetamide (BSA) to afford the bis(tri-
methysilylphosphonate) (d(31P) = 116 ppm) after approxi-
mately 1 h; the oxidant (1R)-(+)-(10-camphorsulfonyl)oxa-
ziridine (CSO) was added to this to generate bis(trimethylsi-
lyl)phosphate (d(31P) = À16 ppm) in about 20 min, which
Scheme 1. a) Nicotinamide, CH3CN, then TMSOTf, RT; 1.5 h, then MeOH, RT; 1 h; b) POCl3, PO(OMe)3, 08C, 4 h; c) 6, CH2Cl2, RT; 3 h, then
mCPBA, À788C, 0.5 h, 98%; d) TFA/H2O/THF 1:1:4, 08C, 3 h, 98%; e) PCl3, imidazole, Et3N, THF, 08C, 15 mins, then 1m TEAB aq. pH 7, RT,
15 mins, quant.; f) BSA, CHCl3, RT, 1 h, then CSO, RT, 0.5 h, then MeOH/D2O, RT, 15 mins, 72%; g) b-NMN, carbonyl diimidazole, Et3N, DMF,
RT, 3 h, then 9, DMF, RT, 16 h; h) 1m TBAF/THF, AcOH, 08C, 1.5 h, 22% i) 10% Pd/C, cyclohexadiene, MeOH/H2O, RT, 2 h, 77%; j) Nicotinic
acid, Aplysia ADP-ribosyl cyclase, 1m NaOAc aq. pH 4, RT, 5 h, 63%.
4638
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 4637 –4640

Angewandte
Chemie
slowly hydrolyzed to give the phosphate 9 (d(31P) = 1 ppm)
upon addition of D2O.
Successful repetition of the approach to NAD reported by
Lee and co-workers[6b] encouraged initial exploration of
morpholidate activation of the precursor 8, but attempted
conversions were not productive. Production and reaction of
the b-NMN imidazolide could be monitored by using
31P NMR spectroscopy (d(31P) = À10.6 ppm).
Contrary to other reports,[6] we found that bisbenzylphos-
phate precursor 9 appeared to react with good conversion
when monitored by 31P NMR spectroscopy and offered an
attractive protecting-group strategy (Scheme 1). Both benzyl
groups remained intact during this reaction and the resulting
product could not be purified by using our preferred ion-
exchange chromatography method. Instead, the crude mate-
rial was treated with 1m TBAF in THF/AcOH to effect
simultaneous cleavage of the silanol ether and one of the
benzyl protecting groups to yield monobenzyl NADP 10 after
ion-exchange chromatography. Although there was no evi-
dence of migration of the 2’-phosphate, the overall yield of
around 22% for these two steps is modest, but is comparable
to many literature reports as pyrophosphate bond formation
and isolation of the resulting product is generally difficult and
yields are highly variable. An alternative route that employs a
bis(2-cyanoethyl) protected precursor related to 9 gave yields
in the range of 60–80% for the pyrophosphate coupling and
we are currently exploring improvements of this route to
achieve useful conversion. Nonetheless, transfer hydrogena-
tion of 10 yielded material that was identical to commercial
preparations of b-NADP 2 in satisfactory (77%) yield.
With a wider objective in mind we chose to convert
synthetic NADP 2 into the target NAADP 1 to test its
behavior in a relevant biological system. Base exchange to
replace nicotinamide with nicotinic acid was achieved by
using an excess of the latter and crude Aplysia ADP-ribosyl
cyclase [enzyme commission number E.C.3.2.2.5] prepared
from Aplysia ovotestis[15] to provide NAADP 1 in 63% yield.
The ability of 1 to induce release of Ca2+ ions was tested
by using a cell-free system derived from the eggs of sea
urchins. Aliquots of homogenates of sea-urchin eggs comprise
vesicles derived from the intracellular stores that sequester
calcium when supplemented with an ATP-regenerating
system. Concentration-dependent calcium release is observed
when this mixture is challenged with second messengers, such
as 1, which can be measured by using cuvette-based fluorim-
etry and the calcium reporter dye fluo-3.[16] Synthetic
NAADP, carefully quantified by using total phosphate assay
prior to evaluation,[17] potently induced calcium release in an
identical manner to “authentic” NAADP prepared from
commercial NADP (Figure 2). Both preparations display an
important characteristic of this signaling pathway that occurs
at subthreshold concentrations of NAADP in homogenate of
sea-urchin eggs. They both potently deactivate the calcium
store in a time dependent manner, so that further challenge
with NAADP does not lead to significant calcium release
(Figure 3). This property allows competition binding experi-
ments that measure the extent that nonlabeled NAADP
displaces subthreshold concentrations of [32P]NAADP
(0.2 nm) from the putative receptor (Figure 4).
Figure 2. Ca2+-ion release from homogenate of sea-urchin eggs by
100 nm authentic NAADP (dotted arrows) or 100 nm synthetic
NAADP, 1 (solid arrows): Samples diluted to 2.5% in GluIM in the
presence of regenerating system and kept at 178C with agitation for
3 h to facilitate Ca2+-ion uptake into stores. Ca2+-ion release deter-
mined byan increase of Fluo-3 fluorescence at 526 nm; data
expressed as released [Ca2+], determined byfluorescence in arbritrary
units; n=>3 for each point. R. F. U.=relative fluorescence units.
Figure 3. Inhibition of NAADP-induced Ca2+-ion release byauthentic
NAADP (squares) or synthetic NAADP 1 (triangles) in the homogenate
of sea-urchin eggs: Samples diluted to 2.5% in GluIM in the presence
of a regenerating system and kept at 178C with agitation for 3 h to
facilitate Ca2+-ion uptake into stores. Ca2+-ion release determined by
an increase of Fluo-3 fluorescence at 526 nm; data expressed as
pmoles of Ca2+-ions released; n=>3 for each point.
In summary, the strategy of using 1,1,3,3-tetraisopropyl
disiloxane protecting groups offers a robust (70%, five steps
from adenosine) route to 2’,5’-adenosine diphosphate pre-
cursors and the potential to incorporate a range of function-
ality at these positions. Pyrophosphate formation requires
some further development, but affords benzyl protected
precursor 10 that, after transfer hydrogenation, leads to the
first total chemical synthesis of 2. This flexible route to 2’-
phosphate furnished dinucleotides will allow the develop-
ment of new chemical tools that probe biological systems at
the molecular level. Analogues of 1 are of immediate
importance for dissecting this novel and important second
Angew. Chem. Int. Ed. 2004, 43, 4637 –4640
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4639

Communications
(CA6), 147.0 (CHA2), 144.6 (CHN6), 142.7 (CHN4), 142.5(CHA8) 141.8
(CHN2), 133.8 (CN3), 128.6 (CHN5), 118.4 (CA5), 99.7 (CHN1’), 86.8
(CHA4’), 83.9 (CHA1’), 77.4 (CHN2’), 76.8 (CHA2’), 73.8 (CHN4’), 70.3
(CHN3’), 69.9 (CHA3’), 65.1 and 64.8 ppm (2 ” CH2); 31P NMR
(162 MHz, D2O): d = 1.47 and À9.40 to À9.80 ppm (br m); MS
[FAB+] 744.8 [M+ÀH]; found 745.0658 [M+ÀH C21H29N7O17P3
requires 745.0673.
Received: March 19, 2004
Keywords: biological activity· calcium · intracellular signaling ·
nucleotides · total synthesis
Figure 4. Competitive displacement of [32P]NAADP (0.2 nm) with
.
authentic NAADP (squares) or synthetic NAADP 1 (triangles) from
sea-urchin homogenate: Samples diluted in GluIM, then 0.2 nm
[32P]NAADP in the presence of increasing NAADP added and incu-
bated at room temperature for 20 mins. Samples filtered through
Whatman GF/B filters to separate bound and free [32P]NAADP ligand.
Nonspecific binding is defined byincubation of the homogenate in the
presence of 10 mm NAADP; n=3ÆSEM; data expressed as specific
cpm.
[1] M. J. Berridge, M. D. Bootman, H. L. Roderick, Nat. Rev. Mol.
Cell Biol. 2003, 4, 517 – 529.
[2] H. C. Lee, Curr. Biol. 2003, 13, R186-R188.
[3] a) R. Masgrau, G. C. Churchill, A. J. Morgan, S. J. H. Ashcroft,
A. Galione, Curr. Biol. 2003, 13, 247 – 251; b) G. C. Churchill,
J. S. OꢀNeill, R. Masgrau, S. Patel, J. M. Thomas, A. A.
Genazzani, A. Galione, Curr. Biol. 2003, 13, 125 – 128.
[4] a) J. Bak, R. A. Billington, G. Timar, A. C. Dutton, A. A.
Genazzani, Curr. Biol. 2001, 11, 987 – 990; b) M. Hohenegger, J.
Suko, R. Gscheidlinger, H. Drobny, A. Zidar, Biochem. J. 2002,
367, 423 – 431; c) A. H. Guse, Cyclic ADP-Ribose NAADP 2002,
301 – 318; d) F. X. Boittin, A. Galione, A. M. Evans, Circ. Res.
2002, 91, 1168 – 1175; e) J. M. Cancela, G. C. Churchill, A.
Galione, Nature 1999, 398, 74 – 76; f) J. D. Johnson, S. Misler,
Proc. Natl. Acad. Sci. USA 2002, 99, 14566 – 14571; g) K. J.
Mitchell, F. A. Lai, G. A. Rutter, J. Biol. Chem. 2003, 278,
11057 – 11064.
messenger pathway. Enzymatic conversion into 1 is confirmed
both by identical behavior when evaluated for Ca2+-release
properties against sea-urchin-egg homogenate and spectro-
scopic characterization. Expansion of the route to include
suitably protected beta-nicotinic acid mononucleotide (b-
NAMN) towards the first total chemical synthesis of 1 and
improved isolated yields of pyrophosphate will be reported in
due course.
[5] A. Galione, J. M. Cancela, G. C. Churchill, A. A. Genazzani, C.
Lad, J. M. Thomas, H. L. Wilson, D. A. Terrar in Methods in
Calcium Signalling (Ed.: J. W. J. Putney), CRC, Boca Raton, FL,
2000, pp. 249 – 296.
Experimental Section
[6] a) D. R. Walt, M. A. Findeis, V. M. Rios-Mercadillo, J. Auge,
G. M. Whitesides, J. Am. Chem. Soc. 1984, 106, 234 – 239; b) J.
Lee, H. Churchil, W. B. Choi, J. E. Lynch, F. E. Roberts, R. P.
Volante, P. J. Reider, Chem. Commun. 1999, 729 – 730.
[7] a) L. J. Haynes, N. A. Hughes, G. W. Kenner, A. Todd, J. Chem.
Soc. 1957, 3727 – 3732; b) N. A. Hughes, G. W. Kenner, A. Todd,
J. Chem. Soc. 1957, 3733 – 3738.
[8] V. C. Bailey, J. K. Sethi, S. M. Fortt, A. Galione, B. V. L. Potter,
Chem. Biol. 1997, 4, 51 – 61.
[9] H. C. Lee, R. Aarhus, J. Biol. Chem. 1997, 272, 20378 – 20383.
[10] G. Berridge, R. Cramer, A. Galione, S. Patel, Biochem. J. 2002,
365, 295 – 301.
2: Benzyl NADP 9 (15 mg, 0.018 mmol), 10% Pd/C (15 mg) and
cyclohexadiene (100 mL) in a degassed mixture of H2O:MeOH (3:1, v/
v, 2 mL) were stirred at room temperature under an argon atmos-
phere for 3 h, after which the palladium was filtered and the resulting
filtrate subject to ion-exchange chromatography (AG MP-1) by using
a 150 mm aqueous TFA gradient. Fractions were combined and
lyophilized to yield 2 as a white powder (10 mg, 77%): 1H NMR
(400 MHz, D2O): d = 9.29 (s, 1H; HN2), 9.11 (d, J = 6.3 Hz, 1H; HN6),
8.78 (d, J = 8.2 Hz, 1H; HN4), 8.41 (s, 1H; HA8), 8.23 (s, 1H; HA2),
8.13 (m, 1H; HN5), 6.11 (d, J = 5.9 Hz, 1H; HA1’), 6.01 (d, J = 5.1 Hz,
1H; HN1’), 4.96 (m, 1H; HA2’), 4.49 (m, 1H; HA3’), 4.44 (br s, 1H;
HN4’), 4.40 (m, 1H; HN2’), 4.31 (m, 1H; HN3’), 4.27 (m, 1H; HA4’),
4.26 (m, 1H; HA5’a), 4.12 ppm (m, 3H; HA5’b, HN5’); 31P NMR
(162 MHz, D2O): d = À0.73 and À11.42 to À11.84 ppm (br m); MS
[FAB + ]: m/z (%): 744.0 (50) [M+ÀH]; found 744.0842 [M+ÀH]
C21H29N7O17P3 requires 744.0833.
[11] S. Tanimori, T. Ohta, M. Kirihata, Bioorg. Med. Chem. Lett.
2002, 12, 1135 – 1137.
[12] Y. Hayakawa, M. Kataoka, J. Am. Chem. Soc. 1998, 120, 12395 –
12401.
[13] X. F. Zhu, H. J. Williams, A. I. Scott, Tetrahedron Lett. 2000, 41,
9541 – 9545.
[14] T. Wada, A. Mochizuki, Y. Sato, M. Sekine, Tetrahedron Lett.
1998, 39, 7123 – 7126.
[15] M. E. Migaud, R. L. Pederick, V. C. Bailey, B. V. L. Potter,
Biochemistry 1999, 38, 9105 – 9114.
[16] A.Minta, J. P. Y. Kao, R. Y. Tsien, J. Biol. Chem. 1989, 264,
8171 – 8178
1: NADP 2 (4 mg 5 mm) and 100 mm nicotinic acid (12 mg) in a
100 mm aqueous AcOH/NaOH (pH 4, 1 mL) were incubated with
5 mL of ADP-ribosyl cyclase at room temperature. After 5 h, HPLC
analysis (AG MP-1, aqueous TFA) showed complete consumption of
NADP and formation of NAADP (RT = 16.8 mins). The crude
mixture was purified on an ion-exchange resin (AG MP-1) by using
an aqueous TFA gradient (150 mm), the product eluted at 15% TFA.
Combined fractions were evaporated under vacuum at room temper-
ature and lyophilized overnight to afford 1 as a white powder (2.5 mg,
[17] D. Lampe, C. S. Liu, B. V. L. Potter, J. Med. Chem. 1994, 37, 907 –
912.
1
63%): H NMR (400 MHz, D2O): d = 9.32 (s, 1H; HN2), 9.17 (d, J =
5.9 Hz, 1H; HN6), 8.87 (d, J = 8.1 Hz, 1H; HN4), 8.45 (s, 1H; HA8),
8.28 (s, 1H; HA2), 8.14 (m, 1H; HN5), 6.14 (d, J = 5.5 Hz, 1H; HA1’),
6.03 (d, J = 5.1 Hz, 1H; HN1’), 4.97 (m, 1H; HA2’), 4.50 (m, 1H; HA3’),
4.44 (br, 1H; HN4’), 4.41 (m, 1H; HN2’), 4.30 (m, 1H; HN3’), 4.27–4.22
(m, 2H; HA4’ and HA5’a), 4.12–4.10 ppm (m, 3H; HA5’b, HN5’);
13C NMR (100 MHz, D2O): d = 165.4 (CO2H), 149.8 (CA4), 148.3
4640 ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 4637 –4640