Isopolar phosphonate analogue of adenosine diphosphate ribose

Article abstract of DOI:10.1021/ol0488993

(Chemical Equation Presented) The synthesis of the bisphosphonate ADP-ribose, in which acetylene has replaced the oxygen of the pyrophosphate linkage, is reported.

Full text of DOI:10.1021/ol0488993

ORGANIC
LETTERS
2004
Vol. 6, No. 20
3461-3464
Isopolar Phosphonate Analogue of
Adenosine Diphosphate Ribose
Kim Van derpoorten and Marie E. Migaud*
School of Chemistry, DaVid Keir Building, Queen’s UniVersity Belfast,
Stranmillis Road, Belfast, BT9 5AG, Northern Ireland
m.migaud@qub.ac.uk
Received June 11, 2004
ABSTRACT
The synthesis of the bisphosphonate ADP-ribose, in which acetylene has replaced the oxygen of the pyrophosphate linkage, is reported.
Adenosine diphosphate (ADP) ribose (1, Figure 1) is a
ubiquitous cellular metabolite. It is the turnover product in
metabolic pathways of key cellular components such as
protein-bound poly-(ADP-ribose), nicotinamide adenine di-
nucleotide (NAD+) and cyclic ADP-ribose. Through its
reducing ribosyl moiety, ADP-ribose can damage cells by
acting as a nonspecific protein-glycating agent with affinity
for the numerous cellular enzymes possessing nucleotide-
binding sites.1 Recently, ADP-ribose was shown to exhibit
intracellular signaling regulatory functions by mediating
cellular Ca2+-gating.2 This latest discovery has renewed
interest in the biology of ADP-ribose.3 Intracellular levels
of ADP-ribose are thought to be regulated by divalent cation-
dependent pyrophosphatases that cleave ADP-ribose to yield
adenosine monophosphate and 5-ribosyl phosphate.4 Pyro-
phosphatases that selectively hydrolyze ADP-ribose belong
to a vast and widely studied family of enzymes, called
NUDIX enzymes, capable of hydrolyzing numerous nucleo-
side pyrophosphate derivatives.4 Recent investigations of the
biological importance of ADP-ribose pyrophosphatases and
that of their mechanisms, based on crystallographic data
obtained from Escherichia coli,5 Mycobacterium tuberculo-
sis,6 and human7 enzymes, have provided a new momentum
in the study of this subfamily of NUDIX enzymes.
Our current research efforts are directed toward the
synthesis of novel NAD+ analogues to probe the binding
site of NAD(P)+-dependent enzymes8 that do not possess
the commonly encountered Rossmann fold.9 During the
* Fax: 44 28 9038 2117. Tel: 44 28 9027 4339.
(1) Jacobson, E. L.; CervantesLaurean, D.; Jacobson, M. K. ADP-
Ribosylation Anim. Tissues 1997, 419, 371.
(2) Perraud, A. L.; Fleig, A.; Dunn, C. A.; Bagley, L. A.; Launay, P.;
Schmitz, C.; Stokes, A. J.; Zhu, Q. Q.; Bessman, M. J.; Penner, R.; Kinet,
J. P.; Scharenberg, A. M. Nature 2001, 411, 595.
(3) Sano, Y.; Inamura, K.; Miyake, A.; Mochizuki, S.; Yokoi, H.;
Matsushime, H.; Furuichi, K. Science 2001, 293, 1327.
(4) Bessman, M. J.; Frick, D. N.; O’Handley, S. F. J. Biol. Chem. 1996,
271, 25059.
Figure 1. Adenosine diphosphate ribose 1 and its pyrophosphonate
mimic 2.
10.1021/ol0488993 CCC: $27.50
© 2004 American Chemical Society
Published on Web 09/03/2004

development of divergent syntheses of NAD(P) analogues,
it came to our attention that very few ADP-ribose derivatives
possessing modified pyrophosphate linkages had been re-
ported in the literature.10 In addition, Blackburn reported the
only pyrophosphonate derivatives of ADP and ATP, in which
the oxygen of a P-O-P bond had been replaced by an
alkynyl moiety.11-13 In the ATP analogue, the substitution
took place between the â- and the γ-phosphorus atoms and
the synthesis involved the condensation of AMP-morpholi-
date with acetylene-1,2-bisphosphonic acid obtained from
the Michaelis-Arbusov reaction between dichloroethyne and
triethyl phosphite. Such alkynyl-containing bisphosphonates
were shown to have lower pKas than the corresponding
pyrophosphates due to the lack of hydrogen bond between
the Pâ-O-H and the Pγ-O- caused by the linear acetylenic
linkage.11 It can be anticipated that at physiological pH, an
acetylenic-containing pyrophosphonate ADP-ribose analogue
could have equal capabilities for ionic interactions with
enzyme binding residues as its parent pyrophosphate. How-
ever, weaker complexes are to be expected between this type
of pyrophosphonate and Mg2+ present in the pyrophosphatase
binding pocket. This assumption is based on Blackburn’s
work with alkynyl-containing ATP analogues and Ca2+.11
Yet, long-range interactions between acetylene bisphospho-
nate-containing nucleotides and binding residues and metals
through enzyme bond-water molecules can also be antici-
pated.14,15
As a consequence, we decided to synthesize the ADP-
ribose analogue (2, Figure 1) in which the bridging oxygen
has been exchanged for an alkynyl moiety. This analogue
can be viewed as a mimic of the transition state occurring
during the enzymatic cleavage of the P-O-P bond as the
distance between the two phosphorus atoms has been
calculated to differ by 2 Å (calculated P-P distance; 2.8 Å
in P-O-P and 4.8 Å in P-CC-P).16 In addition to the
binding interactions anticipated to occur between enzyme
and analogue, it is hoped that the acetylenic pyrophosphonate
will participate in some new enzymatic chemistry. Indeed,
in the presence of a strong nucleophile, tetraethyl 1,2-ethynyl
bisphosphonate was shown to decompose via the rupture of
one of the P-C bonds.17 Such chemical behavior could lead
Scheme 1. Synthesis of R/â 1,2,3-Tri-O,O,O-acetyl
D-Ribose 5a
a Conditions: (a) CH3OH, H-Dowex resin, rt, 48 h. (b)
(CH3O)2C(CH3)2, H-Dowex resin, rt, 12 h. (c) CH3CH2CH2Br, BaO,
wet DMF, rt, 12 h, 92%. (d) CH3COOH glacial, (CH3CO)2O, H2SO4
cat., 0 °C, 12 h. (e) PdCl2, DCM/CH3OH (1/1), 0 °C to room
temperature, 4 h, 75%.
to an unprecedented mode of irreversible inhibition of
pyrophosphatases. Similarly, in the presence of an acidic
residue, addition of an enzyme nucleophilic residue across
the triple bond can be anticipated,17 once again, leading to
covalent bonding and irreversible inhibition. To the authors’
knowledge, no analogue for which an alkynyl moiety has
been introduced between the R- and â-phosphorus atoms of
a sugar nucleoside diphosphate has been synthesized. Here,
we report a versatile synthesis of the bisphosphonate
acetylene analogue of ADP-ribose (2, Figure 1).
The synthesis of the ADP-ribose analogue 2 was achieved
via the early introduction of an alkynyl-phosphonate moiety
(Scheme 1). Allyloxy diisopropylaminotrimethylsilylethynyl-
phosphine (3, Scheme 2) was prepared from trichlorophos-
phine and allyl alcohol in the presence of pyridine to yield
allyloxy dichlorophosphine (31P NMR 178.3 ppm), which
was then converted to the allyloxy chlorodiisopropylamino-
phosphine (31P NMR 182.0 ppm).18 Subsequent treatment
with trimethylsilyl acetylene in the presence of n-BuLi in
THF at -78 °C gave 3. The use of an allyl-protecting group
was essential for completing the synthesis of 2, as other
alkoxy groups were initially employed unsuccessfully. While
the methyl-protected diisopropylaminotrimethylsilylethynyl-
phosphine was a suitable reagent for the synthesis of the
Scheme 2. Synthesis of Intermediate 7a
(5) Gabelli, S. B.; Bianchet, M. A.; Ohnishi, Y.; Ichikawa, Y.; Bessman,
M. J.; Amzel, L. M. Biochemistry 2002, 41, 9279.
(6) Kang, L. W.; Gabelli, S. B.; Cunningham, J. E.; O’Handley, S. F.;
Amzel, L. M. Structure 2003, 11, 1015.
(7) Shen, B. W.; Perraud, A. L.; Scharenberg, A.; Stoddard, B. L. J.
Mol. Biol. 2003, 332, 385.
(8) Bellamacina, C. R. FASEB J. 1996, 10, 1257.
(9) Bottoms, C. A.; Smith, P. E.; Tanner, J. J. Protein Sci. 2002, 11,
2125.
(10) Pankiewicz, K. W.; Lesiak, K.; Watanabe, K. A. J. Am. Chem. Soc.
1997, 119, 3691.
(11) Blackburn, G. M.; Kent, D. E.; Kolkmann, F. J. Chem. Soc., Perkin
Trans. 1 1984, 1119.
(12) Spelta, V.; Mekhalfia, A.; Rejman, D.; Thompson, M.; Blackburn,
G. M.; North, R. A. Brit. J. Pharm. 2003, 140, 1027.
(13) Blackburn, G. M.; Forster, A. R.; Guo, M.-J.; Taylor, G. E. J. Chem.
Soc., Perkin Trans. 1 1991, 2867.
(14) Gabelli, S. B.; Bianchet, M. A.; Bessman, M. J.; Amzel, L. M.
Nature Struct. Biol. 2001, 8, 467.
(15) Clarke, P.; Lincoln, S. F.; Tiekink, E. R. T. Inorg. Chem. 1991, 30,
2747.
(16) Migaud, M. E.; Malone, J. F. Unpublished results of calculations
on Gaussian Package.
a Conditions: (f) 5, 2,4-DNP, CH3CN, rt, 2 h. (g) H2O2, DCM,
0 °C, 7 min. (h) 50% CsF, CH3CH2OH, rt, 1 h. (i) Allyloxychlo-
rodiisopropylaminophosphine, n-BuLi, THF, -78 °C, 1 h, 53%.
3462
Org. Lett., Vol. 6, No. 20, 2004

Scheme 3. Synthesis of 2a
a Conditions: (j) 2′,3′-O,O-isopropylidene adenosine, 2,4-DNP, CH3CN, rt, 12 h. (k) t-BuOOH, -40 °C, 7 min, 31%. (l) H2O/TFA, rt,
1 h. (m) Pd[PPh3]4, PPh3, TEA, HCOOH, THF, rt, 8 h, 85%. (n) NaOMe, methanol, rt, 30 min. (o) NH3, CH3OH, rt, 12 h, 68%.
fully protected ADP-ribose acetylene analogue, only partial
deprotection of the methyl groups could be achieved in the
final step of the synthesis. Upon more forceful conditions,
only decomposition products could be detected. Cyanoethyl
was not a suitable protecting group due to its lability under
basic conditions, while p-methoxybenzyloxy chlorodiiso-
propylaminophosphine was easily prepared but was inap-
propriate to complete the synthesis, as the PMB group was
readily oxidized in a subsequent step. Finally, p-chloro-
phenyloxy diisopropylaminotrimethylsilylethynylphosphine
proved to be a synthetic precursor difficult to obtain in rea-
sonable purity and yields. Consequently, our endeavor
focused on synthesizing 2 starting from ribose and allyloxy
diisopropylaminotrimethylsilylethynylphosphine 3.
The alkynylphosphonate 6 was obtained in 71% yield over
two steps by coupling 1,2,3-O,O,O-triacetyl-D-ribose 5,
prepared in five steps from D-ribose (Scheme 1), with 3 in
the presence of 2,4-dinitrophenol (2,4-DNP),19 followed by
careful phase transfer oxidation of the resulting phosphonite
and desilylation with CsF. Initial attempts to catalyze the
coupling of 5 with 3 with either tetrazole20 or imidazolium
triflate offered poor to moderate yields over the two steps
(<30%). The sugar-containing alkynylphosphonate 6 was
obtained as a mixture of diastereoisomers (31P NMR -6.47,
-6.69 ppm). The fully protected ADP-ribose acetylene-
containing analogue 8 (Scheme 3) was obtained in 31%
isolated yield over two steps from the 2,4-DNP-catalyzed
coupling of 2′,3′-O,O-isopropylidene adenosine with P1-
[allyloxy diisopropylamino]-P2-[(1′′,2′′,3′′-O,O,O-acetyl)-D-
ribofuranose-5-yl, allyl]-acetylene-(1-phosphine, 2-phospho-
nate) 7 (Scheme 2), followed by oxidation of the resulting
phosphinate to the phosphonate with t-BuOOH. The inter-
mediate 7 had been obtained by adding n-BuLi to a THF
solution of allyloxy chlorodiisopropylaminophosphine and
6, cooled to -78 °C. One should note that initial preparation
of the alkynyl anion and subsequent addition of the chloro-
phosphine resulted in complete decomposition of 6.
The acetonide protecting group of the fully protected ADP-
ribose analogue 8 was hydrolyzed by a mixture of H2O and
TFA (v/v; 1/1.9) to give 9 in 65% yield after preparative
thin-layer chromatography. Unreacted starting material was
recovered in 23% yield. The synthetic intermediate 9 was
then treated with a mixture of Pd[PPh3], PPh3, Et3N, and
formic acid21 to remove the allyl group. An alterna-
tive sequence, which offered similar overall yields, started
with the deallylation reaction immediately followed by the
acetonide removal. Finally, the acetate esters were hydrolyzed
by a solution of sodium methoxide in methanol to obtain
the ADP-ribose analogue 2 in 5% overall yield. It should be
(17) Maier, L. Phosphorus 1973, 2, 229-233.
(18) Wippo, H.; Reck, F.; Kudick, R.; Ramaseshan, M.; Ceulemans, H.;
Bolli, M.; Krishnamurthy, R.; Eschenmoser, A. Bioorg. Med. Chem. 2001,
9, 2411.
(21) Koizumi, M.; Hiratake, J.; Nakatsu, T.; Kato, H.; Oda, J. J. Am.
Chem. Soc. 1999, 121, 5799.
(22) Holy, A.; Koenig, J.; Vesely, J.; Dieter, C.; Votruba, I.; De Clercq,
E. Collect. Czech. Chem. Commun. 1987, 52, 1589-1608.
(23) Izumi, M.; Tsuruta, O.; Hashimoto, H. Carbohydr. Res. 1996, 280,
287-302.
(19) Amigues, E. J.; Migaud, M. E. Tetrahedron Lett. 2004, 45, 1001.
(20) Timmer, M. S. M.; Ovaa, H.; Filippov, D. V.; van der Marel, G.
A.; van Boom, J. H. Tetrahedron Lett. 2001, 42, 8231.
Org. Lett., Vol. 6, No. 20, 2004
3463

mentioned that 8 rapidly degraded to 10 and another
derivative thought to be 11, if first treated with a saturated
methanolic ammonia solution. This observation came as no
surprise, as Maier reported that the P-C bond of tetraethyl
1,2-ethynyl bisphosphonate was cleaved upon basic treat-
ment.17 In addition, when 9 was first deallylated then treated
with a saturated methanolic ammonia solution, a mixture of
three partially deacetylated products were obtained. They
were the 1′′-O-, 2′′-O-, or 3′′-O- monoacetylated derivatives
on allyloxy-chlorodiisopropylaminophosphine. While this
route is somewhat lengthier than the one described above, it
still provides facile access to ADP and ATP analogues that
incorporate an acetylenic group between the R- and the
â-phosphorus atoms. Further work on the deprotection
sequence to yield ADP analogues is currently in progress,
as well as work on the stereoselective reduction of the alkyne
pyrophosphonate derivative to their (E)- and (Z)-alkene
pyrophosphonate equivalents. The ADP-ribose analogues
reported herein are being evaluated for binding affinity to a
variety of ADPR-binding enzymes that include pyrophos-
phatases, dehydrogenases, and glycohydrolases, and their
biological activity will be reported in due course.
To summarize, we have developed a novel and potentially
versatile synthetic route for the preparation of a new type of
modified ADP-ribose in which the oxygen of the pyrophos-
phate linkage has been replaced by an acetylene moiety. This
bisphosphonate is stable under a broad range of pH condi-
tions and is expected to be relatively resistant to nonenzy-
matic cleavage under aqueous conditions. We expect that
this analogue will offer further information on the catalytic
abilities of ADP-ribose-pyrophosphatases, behave as a
potential inhibitor of these enzymes, and finally provide new
tools to investigate the biological importance of the enzyme-
catalyzed hydrolysis of ADP-ribose. In time, this route
developed for the synthesis of ADP-ribose analogues will
be applied to the synthesis of other sugar and nucleoside
derivatives to probe pyrophosphatase specificity.
1
of compound 2 (detected by LCIMS, H NMR, and RP-
HPLC). Attempts to remove the remaining acetate with a
KCN solution or an NH4OH/methanol solution were unsuc-
cessful, even after heating at 50 °C overnight. While
unsatisfactory from a synthetic perspective, these experiments
confirmed that hydrolysis of the pyrophosphonate diester
P-CC-P bond did not occur under these conditions.
Consequently, the stability to nucleophilic attack of the
phosphonate monoester P-CC-P bond is greater than that
of the P-CC-P bond of the phosphonate triester.
Due to our interest in probing enzymes that bind NAD-
(P)+, we wanted this synthetic route to be as flexible as
possible with regard to the introduction of other residues than
the terminal ribose of ADP-ribose. For this reason, we
investigated the feasibility of carrying out a similar sequence
starting instead from 2′,3′-O,O-isopropyledene adenosine to
the triacetyl ribose 5. While the synthetic route remains
broadly identical, some slight modifications had to be
introduced. For instance, neither tetrazole nor imidazolium
triflate was suitable to promote any reaction between allyloxy
diisopropylaminotrimethylsilyl-ethynylphosphine 3 and 2′,3′-
O,O-isopropylidene adenosine. For this synthetic sequence,
only 2,4-DNP was an appropriate activator for the phospho-
namidite coupling reaction. Furthermore, protection of the
N6-adenosine was required in order to form the P-CC-P
linkage in reasonable yields via the n-BuLi-catalyzed nu-
cleophilic displacement by the acetylene anion of the chlorine
Acknowledgment. We are grateful to the McClay Trust
for financial support of this work.
Supporting Information Available: Experimental de-
tails. This material is available free of charge via the Internet
at http://pubs.acs.org.
OL0488993
3464
Org. Lett., Vol. 6, No. 20, 2004