Aplysia californica mediated cyclisation of novel 3′-modified NAD+ analogues: A role for hydrogen bonding in the recognition of cyclic adenosine 5′-diphosphate ribose

Article abstract of DOI:10.1016/j.bmc.2003.10.012

Cyclic ADP-ribose mobilizes intracellular Ca²⁺ in a variety of cells. To elucidate the nature of the interaction between the C3′ substituent of cADP-ribose and the cADPR receptor, three analogues of NAD ⁺ modified in the adenosine ri

Full text of DOI:10.1016/j.bmc.2003.10.012

Bioorganic& Medicinal Chemistry 12 (2004) 475–487
Aplysia californica mediated cyclisation of novel 30-modified
NAD+ analogues: a role for hydrogen bonding in the recognition
of cyclic adenosine 50-diphosphate ribose
Christopher J. W. Mort,a Marie E. Migaud,a Antony Galioneb and Barry V. L. Pottera,*
aWolfson Laboratory of Medicinal Chemistry, Department of Pharmacy and Pharmacology, University of Bath,
Claverton Down, Bath BA2 7AY, UK
bUniversity Department of Pharmacology, Oxford University, Mansfield Road, Oxford OX1 3QT, UK
Received 21 August 2003; accepted 10 October 2003
Abstract—Cyclic ADP-ribose mobilizes intracellular Ca2+ in a variety of cells. To elucidate the nature of the interaction between
the C30 substituent of cADP-ribose and the cADPR receptor, three analogues of NAD+ modified in the adenosine ribase (xyloNAD+
30F-xyloNAD+ and 30F-NAD+ were chemically synthesised from d-xylose and adenine starting materials. 30F-NAD+ was readily
converted to cyclic 30F-ADP ribose by the action of the cyclase enzyme derived from the mollusc Aplysia californica. XyloNAD+
and 30F-xyloNAD+ were cyclised only reluctantly and in poor yield to afford unstable cyclic products. Biological evaluation of
cyclic 30F-ADP ribose for calcium release in sea urchin egg homogenate gave an EC50 of 1.5Æ0.5 mM. This high value suggests that
the ability of the C30 substituent to donate a hydrogen bond is crucial for agonism.
# 2003 Elsevier Ltd. All rights reserved.
1.Introduction
and sugar-modified cADPR analogues.17 ADP-ribosyl
cyclase catalyses the synthesis of cADPR from NAD+
by removing the nicotinamide moiety and forming a b-
glycosidic link between the N1 of adenine and the C100
of the nicotinamide ribose.18 The mechanism has been
investigated by co-crystallisation of the enzyme with
nicotinamide,12,19,20 mutagenesis of the enzyme,21 mass-
Recently, the calcium-releasing nucleotide cyclic adeno-
sine diphosphate ribose (cADPR 1, Fig. 1)1 has been
identified as a second messenger2 in a wide variety of
cell types including mammalian smooth muscle3 cells
and neurons.4 Initiated by activation of the cADPR
receptor, cyclic ADP ribose mediated Ca2+ release
occurs independently of both the inositol 1,4,5-
trisphosphate (IP3) and the nicotinic acid adenosine di-
phosphate (NAADP) Ca2+ release pathways.5À7 In
mammalian cells, the biosynthesis of cADPR from
NAD+ is catalysed by the ubiquitous membrane-bound
glycoprotein CD38,8,9 implicated in the endogenous
spectroscopy of an enzyme-inhibitor intermediate22 and
23À25
by kineticand inhibition studies.
Through their
altered biological properties, synthetic analogues of
cADPR have provided extensive information on the
structure–activity relationship (SAR) of the cADPR
receptor. Modification of the purine moiety has pro-
duced cADPR receptor antagonists,26 membrane-per-
meant antagonists,27,28 partial agonists,29 and
agonists.17,30 The Aplysia cyclase has also been found to
tolerate modifications to the adenine ribose component
of the NAD+ substrate.17,31À33 Among the cADPR
analogues chemo-enzymatically synthesised were 30-
production of cADPR10,11 and responsible also for cat-
alysing the hydrolyticdegradation of NAD
+
and
cADPR to ADPR.12À15 In contrast, it has been known
since 1991 that the ADP ribosyl cyclase enzyme derived
from the sea mollusc Aplysia californica possesses very
little hydrolase activity.16 Together with the loose sub-
strate specificity inherent in the A. californica cyclase,
this has permitted the synthesis of numerous purine-
deoxy-cADPR 2, 30-OMe-cADPR 3 and 30-OPO23À
-
cADPR 4 (Fig. 1).34 Sea-urchin egg homogenate (SUH)
calcium-release assays performed on these three analo-
gues defined a crucial role for the 30-hydroxyl group in
biological activity.33 The introduction of the 30-phos-
phate group (4) resulted in a complete loss of binding
affinity for the cADPR receptor, while complete
* Corresponding author. Tel.: +44-1225-386639; fax: +44-1225-
386114; e-mail: b.v.l.potter@bath.ac.uk
0968-0896/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2003.10.012

476
C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
zoyladenine proceeded smoothly to yield one main
product,37 the required N6-benzoyl-9-(20,30-di-O-ben-
zoyl-b-d-xylofuranosyl)-adenine 11. The relative
hydrophobicity of the partially protected nucleoside
permitted facile purification via conventional silica
gel chromatography. Application of Yoshikawa’s
methodology38 cleanly converted the partially pro-
tected nucleoside into the 50-phosphate ester, as
monitored by anion-exchange HPLC. 31P and 1H
NMR spectroscopy identified this material as the
phosphorylated nucleoside 12 contaminated only
with triethylphosphate. Removal of the benzoyl pro-
tecting groups was accomplished with methanolic
ammonia. Chromatographicpurifiaction of the resi-
due on AG-MP1 anion-exchange resin yielded the
requisite 9-(b-d-xylofuranosyl)-adenine-50-monophos-
phate 13. Conventional DCC-mediated phosphate
condensation39,40 of the modified adenine nucleotide
with nicotinamide mononucleotide to yield 14 was
expected to be low yielding since DCC might also
facilitate intramolecular condensation of 13 to yield
cyclic xylo-30,50-AMP.41 Hence, the alternative
DPPC-activation methodology42 was employed to
promote nucleophilic attack of the adenine nucleo-
side phosphate ester at the activated NMN pyro-
phosphate centre. Saponification of the 200,300-O-Ac
groups of the nicotinamide moiety with methanolic
ammonia followed by anion-exchange purification
yielded the free acid of xyloNAD+ 14 in modest
yield. No cyclic xyloAMP or symmetrical dinucleo-
tide by-products were detected.
Figure 1. Previously synthesised (2–4) and proposed (5–7) 30-modified
analogues of cyclic ADP ribose 1.
removal of the 30-hydroxyl (2) caused a 100-fold
decrease in agonism. Methylation of the C30 (3) yielded
an antagonist, thus far the only carbohydrate-modified
cADPR antagonist. These observations underline the
importance of the C30-hydroxyl group of cADPR for
binding to the receptor and initiating Ca2+ release. To
further elucidate the nature of the interaction between
the C30 substituent and the receptor, the synthesis of
three novel cADPR analogues (5–7) was undertaken.
Biological evaluation of these analogues would reveal
whether the stereochemistry of the C30-hydroxyl is
important for Ca2+ release, and clarify the hydrogen-
bonding requirements at the C30 for agonism and
antagonism. We present here the synthesis of adenosine
ribose-modified 30-modified analogues of NAD+ and
examine their behaviour to incubation with A. cali-
fornica cyclase. In particular, we have in 5 replaced the
30-hydroxyl group of the adenosine motif of cADPR by
the widely recognised surrogate fluorine to explore H-
bond donor/acceptor status with respect to the cADPR
receptor. The Ca2+ release profile of 5 is presented in
SUH.
Scheme 2 illustrates the synthesis of the fluorinated
analogue 30F-xyloNAD+. Careful kineticocntrol per-
mitted the regioselective benzoylation of 15,43 to pro-
ceed in high yield. Inversion of the 30-OH was
performed by a two-step oxidation and reduction
sequence to yield ribofuranose 17. The use of aqueous
THF in the reduction step obviated the transesterifica-
tion as witnessed by Imbach et al.43 Fluorine was intro-
duced with expected inversion of C30 stereochemistry by
DAST and confirmed by the large geminal splitting (J3-F
1
50.4 Hz) of the H NMR H-3 doublet and by the pre-
sence of a ‘ddd’ signal at around À200 ppm in the 19F
NMR spectrum of the product. The 1,2-O-iso-
propylidene group of 18 was hydrolysed using aqueous
acetic acid and the resultant diol 19 benzoylated and
chromatographed to yield an inseparable mixture (1:1 a/
b) of 1,2,5-tri-O-benzoyl-3-deoxy-3-fluoro-d-xylofura-
nose 20. Glycosidation and phosphorylation were per-
formed in a manner analogous to the synthesis of
xyloAMP to yield the fluorinated nucleotide 23. DPPC
activation of acetylated NMN followed by condensa-
tion with 23 yielded the required dinucleotide 24 after
deprotection of the acetate protecting groups.
2.Chemistry
In order to explore the potential effect of replacing the
adenine ribose by a xylose sugar xyloNAD+ 14 was
synthesised as shown in Scheme 1. 5-O-Benzyl-1,2-O-
isopropylidene-a-d-xylofuranose
8
was synthesised
from commercially available 1,2-O-isopropylidene-a-d-
xylofuranose 15 via the dialkylstannylene inter-
mediate.35 Hydrolysis of the unusually stable 1,2-O-
acetonide group was accomplished in good yield using
aqueous TFA at 0 ꢀC to avoid concomitant deprotec-
tion of the benzyl ether. The resultant diol was benzoy-
lated to yield 9. Catalytictransfer hydrogenation using
Pearlman’s reagent36 rapidly and quantitatively depro-
tected the primary benzyl ether affording an inseparable
anomericmixture of 10. The reaction of 1,2,3-tri-O-
benzoyl-d-xylofuranose 10 with persilylated N6-ben-
The synthesis of the prime fluorinated target 30F-
NAD+ 29 is outlined in Scheme 3. 1-O-Acetyl-2,5-di-O-
benzoyl-3-deoxy-3-fluoro-d-ribofuranose 25 was pre-
pared from d-xylose by the reliable 12-step procedure
of Mikhailopulo et al.44 Reaction of 25 under Vorbrug-
gen conditions37 afforded the C10-N7 b-nucleoside
which was then deprotected using methanolic ammonia

C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
477
Scheme 1. Synthesis of xyloNAD+ 14: (a) TFA/H2O (9:1 v/v), 0 ꢀC; (b) BzCl, py, 0 ꢀC!rt; (c) Pd(OH)2/C, EtOH/cyclohexene (2:1 v/v), reflux;
(d) N6-Bz-adenine-(TMS)2, TMS-OTf, DCE, reflux; (e) POCl3, PO(OEt)3, 0 ꢀC!rt; (f) NH3, MeOH, 50 ꢀC; (g) activated NMN, py/DMF (2:1 v/v).
Scheme 2. Synthesis of 30F-xyloNAD+ 24: (a) BzCl, py, À20 ꢀC; (b) CrO3, py, Ac2O, 0 ꢀC; (c) NaBH4, THF/H2O (5:4 v/v), 0 ꢀC; (d) DAST, py,
DCM, À10 ꢀC!rt; (e) AcOH/H2O (4:1 v/v), 80 ꢀC; (f) BzCl, py, DMAP; (g) N6-Bz-adenine-(TMS)2, TMS-OTf, DCE, reflux; (h) NH3, MeOH,
50 ꢀC; (i) POCl3, PO(OEt)3, 0 ꢀC!rt; (j) activated NMN, py/DMF (2:1 v/v).
to give the fluorinated nucleoside 27. Yoshikawa phos-
phorylation38 followed by ion-exchange purification
yielded 30-deoxy-30-fluoroadenosine 50-monophosphate
28 in good yield. The coupling of NMN to the adenine
nucleotide 28 was again performed using the DPPC
coupling to afford 30F-NAD+ 29. All three NAD+
analogues were then treated with ADP-ribosyl cyclase
and their substrate properties examined.
mal hydrolase activity.16 The commercial availability
and loose substrate specificity of the enzyme has
enabled the synthesis of a large number of cADPR
analogues. When modifications were made to the purine
ring of the substrate, the main kineticeffetc observed
was a slower turnover.45 Exchange of the 20- or 30-
hydroxyl groups of the adenosine ribose moiety for
charged groups such as phosphate31 or more hydro-
phobic moieties such as methoxy was also of little con-
sequence to the cyclase.33 Upon incubation of
xyloNAD+ with the Aplysia cyclase, three new entities
could be detected by HPLC. As expected, the first peak
corresponded to nicotinamide, the second possessed a
retention time very similar to that of cADPR. It was
3.Results and discussion
ADP-Ribosyl cyclase isolated from the mollusc A. cali-
fornica is known to have high cyclase activity and mini-

478
C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
Scheme 3. Synthesis of cyclic-30F-ADP ribose 29: (a) N6-Bz-adenine-(TMS)2, TMS-OTf, DCE, reflux; (b) NH3, MeOH, 50 ꢀC; (c) POCl3,
PO(OEt)3, 0 ꢀC!rt; (d) activated NMN, py/DMF (2:1 v/v).
therefore inferred that this latter product was the
desired cyclic nucleotide. However, after 60 min a large
proportion of the xyloNAD+ 14 starting material
remained unconverted. In an attempt to drive the reac-
tion to completion, a further quantity of cyclase was
added. No increase in the quantity of cyclised product
was detected, but an increase in the amount of a third
product was observed. This third product displayed a
retention time analogous to that of ADPR, the linear
product of NAD+ or cADPR hydrolysis. These obser-
vations seemed to indicate that although cyclisation was
occurring, the cyclic product was rapidly hydrolysed.
Indeed, purification of the reaction mixture by anion-
exchange chromatography failed to isolate the cyclic
product, which readily decomposed to xyloADPR over
the course of the purification. The attempted cyclisation
of 30F-xyloNAD+ 24 followed a similar pattern and no
stable cyclic product was isolated. The reluctance of
xyloNAD+ to cyclise may be rationalised by the fol-
lowing arguments. In purine nucleosides, a correlation
exists between the syn and anti equilibrium around the
glycosidic bond and the C30-endo-C20-endo conforma-
tions of the carbohydrate fragment.46 Analysis of the
solution conformation of several purine nucleosides by
13C NMR spectroscopy has demonstrated that popula-
tion of the C30-endo conformation of the furanose ring
of xylo-adenosine should result in repulsion between
the C30 substituent on the upper face of the sugar and
the base moiety. This repulsion is maximal when the
base lies in the syn conformation. Thus, it can be envi-
saged that the C30-endo conformation changes the syn/
anti equilibrium toward the anti form to a greater
degree in xylo-adenosine than in the parent adenosine.
sugar ring conformation with an anti conformation
around the glycosidic C10–N7 bond. In most natural
ribonucleosides, the C20-endo syn base conformation is
more energetically favourable than the C30-endo syn
base conformation (Fig. 2).49 However, in contrast to
xylo-adenosine, the barrier to pseudorotation in adeno-
sine between C30-endo and C20-endo is low (ꢁ3 kcal/
mol) (i.e., fairly equal population of both N and S
states) permitting fairly rapid interconversion. A low
energy pathway for anti/syn interconversion is impor-
tant in the cyclisation of NAD+ to yield cADPR since
the purine must be syn before ribosylation can occur at
the N1. X-ray crystallography18 and NMR spectro-
scopy of cADPR50 have established the conformation of
the adenosine furanose ring as C20-endo, also observed
in the adenosine furanose ring of NAD+.51 Enzymatic
cyclisation of xyloNAD+ 14 may therefore be impeded
by the strong tendency of the xylofuranose ring to adopt
a C30-endo conformation with the concomitant anti
orientation of the purine moiety. The instability of the
proposed cyclised products observed by HPLC might
also be explained by the tendency of the adenine
nucleoside moiety to adopt an anti conformation, thus
weakening the newly-formed N1–C100 bond to yield the
linear unstrained ADPR analogue.
Enzyme kinetic23À25 and mass spectroscopy experi-
ments22 have shown that Aplysia cyclase initially binds
the nicotinamide ribose moiety of the NAD+ substrate.
The enzyme is then believed to bind the second aromatic
purine terminus. It therefore seems plausible that the
cyclase first binds the nicotinamide riboside terminus of
xyloNAD+ and cleaves the C10–N1 glycosidic link to
form a covalent bond22 between the enzyme and the C10
of the nicotinamide ribose. The inflexible C30-endo anti
conformation of the xylo nucleoside unit may prevent
the enzyme from binding the purine terminus of the
dinucleotide. Alternatively, successful recognition may
1
Indeed, this has been demonstrated by H-coupled 13C
NMR experiments.47 Using X-ray crystallographic data
to refine previous observations,48 Birnbaum et al. con-
firmed that for xylo-adenosine, a strong preference
(population ꢁ80–87%) exists for an N-type (C30-endo)

C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
479
Figure 2. (a) C20-endo conformation of adenosine — syn and anti
rotomers are readily interconverted; (b) C30-endo conformation of xylo
adenosine — anti rotomer favoured due to repulsion between purine
and C30-OH.
Figure 3. C30-endo (a) and C20-endo conformations (b) of 30F-xylo-
adenosine sighting along the C30–C40 bond. The favourable gauche
interaction between the C30–F and C40–O40 bonds (a) forces the
nucleoside into the C30-endo conformation.
occur at both aromatic termini but rotation of the
purine into the required syn conformation cannot take
place. The presence of water in the enzyme active site
would then permit the hydrolysis of the proposed cova-
lent xyloADPR–enzyme complex resulting in the release
of xyloADPR from the enzyme.
The concentration of cyclic 30F-ADP ribose required to
elicit half-maximal Ca2+ release (EC50) was calculated
as 1.5Æ0.5 mM (Fig. 4). By comparison, the EC50 values
for cyclic ADP ribose and cyclic 20-deoxy-ADP ribose
have previously been reported as 32 and 58 nM,
respectively.33 The high value obtained for the EC50 of 5
compares with that determined for the agonist cyclic 30-
deoxy-ADP ribose 2 (EC50=5 mM) and with that
reported for the antagonist cyclic 30-OMe-ADP ribose 3
(IC50=5 mM).33 Both 3 and 5 possess a substituent at
the 30-position that functions only as a hydrogen bond
acceptor. While the importance of the 30-OH group in
the binding of cADPR to the receptor has been pre-
viously established by the biological evaluation of 2 and
3, the low binding affinity exhibited by the 30-modified
cADPR congeners 2, 3, and 5 can now be ascribed, at
least in part, to their inability to act as hydrogen bond
donors. Furthermore, the antagonist activity observed
for 3 can be tentatively attributed to the larger steric
volume of the methyl group. Unfortunately, for the
other analogues 6 and 7 the influence upon receptor
recognition and Ca2+ release by a reversal in the ste-
reochemistry of the 30-group (with any potential change
in nucleoside/macrocycle conformation) could not be
assessed since cyclic xyloADP ribose and cyclic 30F-
xyloADP ribose were considered too unstable for bio-
logical assay.
In nucleosides that incorporate one fluorine substituent
on the sugar moiety, the gauche effect arising from the
interaction between the various electronegative sub-
stituents is the predominant effect that drives the N/S
equilibrium strongly towards one of the puckered
forms.52 In 30F-xylo-adenosine, the strong electro-
negativity of the C30 fluorine atom located above the
plane of the sugar ring pushes the equilibrium towards a
conformation in which the C–F bond adopts a pre-
ferential gauche orientation with the C40-ring oxygen
bond. This gauche orientation only occurs in the C30-
endo (N-type) conformation and this form is therefore
preferred. The conformational analogy with the xylo-
adenosine nucleoside unit allows similar arguments to
be applied in rationalising the difficulty in cyclising 30F-
xyloNAD+ 24. In 30F-adenosine, however, the fluorine
atom is located below the plane of the sugar ring. The
ring must adopt a C20-endo conformation in order to
attain a favourable gauche orientation between the C–F
bond and the C40-ring oxygen bond, thus driving the
overall equilibrium towards the S-type conformation52
(Fig. 3).
4.Conclusion
It was therefore anticipated that anti/syn interconver-
sion in the prime target analogue 30F-NAD+ 29 would
occur more readily than in either of the two xylo ana-
logues 14 and 24 permitting enzymatic cyclisation to
occur. Indeed, upon incubation of 30F-NAD+ 29 with
the Aplysia enzyme, cyclic 30F-ADP ribose 5 was suc-
cessfully formed in yields comparable with the cyclisa-
tion of the natural substrate NAD+ and the product
was stable.
Starting from simple sugar precursors and adenine, we
have prepared analogues of adenosine monophosphate
that were coupled to diphenyl phosphate-activated
nicotinamide mononucleotide. XyloNAD+ 14, 30F-
xyloNAD+ 24 failed to cyclise in a productive fashion
when incubated with A. californica ADP-ribosyl cyclase
while 30F-NAD+ 29 yielded cyclic 30F-ADP ribose 5.
The inability of 14 and 24 to cyclise was attributed to
restricted purine rotation around the glycosidic bond by
0
Cyclic 30F-ADP ribose 5 was evaluated for Ca2+ release
activity in SUH in accordance with a previously pub-
lished procedure.33 Upon incubation of 5 and SUH in
the presence of Fluo-3, a dye that exhibits fluores-
cence upon sequestering calcium, the release of Ca2+
was observed in the SUH. Cyclic 30F-ADP ribose was
therefore shown to be an agonist of the cADPR
receptor.
stericrepulsion between the 3 -substituent on the b-face
and the purine moiety. This repulsion was exacerbated
by the C30-endo conformation of the xylo-nucleosides
bringing the C30-group into closer proximity with the
purine. Calcium-release assay revealed 5 to be a rela-
tively weak agonist. In conjunction with biological data
previously obtained for cyclic 30-deoxy-ADP ribose 2
and cyclic 30-OMe-ADP ribose 3, this result suggests

480
C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
Figure 4. Concentration–response curves for cyclic 30F-ADP ribose for calcium release in sea urchin egg homogenate.
that the ability of a 30-substituent of cADPR or an
analogue to donate a hydrogen bond is crucial for good
cADPR receptor recognition.
enzymes were purchased from Aldrich/Sigma. Reactions
that required anhydrous conditions were conducted
under dry nitrogen. Dichloromethane and pyridine were
distilled over CaH2 and stored over molecular sieves and
KOH, respectively. Dioxane was distilled over BaO and
stored over activated molecular sieves, as was DMF.
THF was freshly distilled over Na and benzophenone.
NMR analyses were carried out on a 400 MHz Varian
spectrometer. The d values are stated in ppm and J
values are given in Hertz (Hz). For 1H NMR, the values
are referenced on the TMS signal in CDCl3 (0 ppm), on
the HOD signal in D2O (4.65 ppm), on the
CHD2COCD3 in acetone-d6 (2.09 ppm) and on
CHD2SOCD3 in DMSO-d6 (2.40 ppm). For 13C NMR,
the values are referenced on the carbon signal in CDCl3
(77.00 ppm) and on added CH3CN in D2O (1.30 ppm).
The signal of CFCl3 is used to reference 19F NMR (0.00
ppm). For 31P NMR, the values are referenced on the
H3PO4 signal in D2O (0.00 ppm). Mass spectra were
recorded at the University of Bath. Ultraviolet (UV)
absorbance was measured with a Perkin-Elmer Lambda
3 UV/VIS spectrophotometer. Melting points were
determined using a Reichert-Jung Thermo Galen Kofler
Block and are uncorrected.
5.Experimental
5.1. General
Thin layer chromatography (TLC) was carried out
using percolated plates (Merck TLC aluminium sheet
silica 60 F254Art no. 5554). TLC plates were visualised
by ultraviolet light (254 nm) and immersion in ani-
saldehyde stain (by volume: 93% ethanol, 3.5% sulfuric
acid, 1% acetic acid and 2.5% anisaldehyde) followed
by heating. Flash chromatography (SGC) utilised
Merck Sorbsil C60 silica gel and Merck GPR solvents
used without distillation. Ion exchange chromatography
was performed using an unjacketed Glass Econo-col-
umn FPLC column (20Â1.0 cm internal diameter) and
BioRad AG-MP1 strong anion-exhange resin (func-
tional group RCH2Me3N+; matrix styrene divi-
nylbenzene) with a concave gradient (0 min: 100%
MQ!20 min: 100% MQ!40 min: 96% MQ!60 min:
84% MQ!75 min: 68% MQ!105 min: 0% MQ!120
min: 0% MQ) of 150 mM TFA. HPLC was performed
on a Shimadzu LC-6A chromatograph with UV detec-
tor operating at 254 nm or on a Varian ProStar system
with UV detector operating at 254 or 314 nm. HPLC
chromatography used a combination of a Partisil 10 mm
SAX guard column (10Â0.46 cm) and either a Supelco
(20Â0.46 cm) 10 mm SAX HPLC column with an iso-
cratic elution using phosphate buffer (50 mM), pH 3
containing 5% MeOH at a flow rate of 2 mL/min or a
Technicol (10Â0.46 cm) 10 mm SAX HPLC column
with an isocratic elution using at 1 mL/min of the same
buffer.
5.2. Chemistry
5.2.1. 5-O-Benzyl-1,2,3-tri-O-benzoyl-D-xylofuranose (9).
5-O-Benzyl-1,2-O-isopropylidene-a-d-xylofuranose35
(1.10 g, 3.92 mmol) was added to a pre-cooled (0 ꢀC)
mixture of trifluoroacetic acid (36 mL) and water (4
mL). The solution was stirred at 0 ꢀC for 45 min after
which TLC (4:1 CHCl3/Me2CO) indicated the forma-
tion of two major products (Rf=baseline, Rf=0.09), a
trace of SM (Rf=0.57) and a second minor product
(Rf=0.38). The reaction mixture was poured into tolu-
ene (100 mL) and concentrated. The residue was azeo-
troped with toluene a further five times to yield a
colourless viscous oil (1.15 g).
All chemicals including solvents, reagents, buffers and

C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
481
The oil was dissolved in dry pyridine (25 mL), cooled to
0 ꢀC under an atmosphere of dry N2 and benzoyl chlor-
ide (3.1 equiv, 12.20 mmol, 1.41 mL) added dropwise.
After 18 h at rt, TLC (3:1 hexane/EtOAc) indicated the
formation of several products. Benzoyl chloride (2.0
equiv, 7.84 mmol, 0.90 mL) was added and the solution
stirred for a further 11 h. TLC indicated mainly one
product (Rf=0.38) but a trace of SM was still present.
The addition of benzoyl chloride (0.5 equiv, 1.96 mmol,
0.23 mL) with stirring for a further 13 h resulted in the
complete conversion to one spot (Rf=0.38). The reac-
tion mixture was concentrated and co-evaporated from
toluene (3Â). The residue was dissolved in DCM and
washed with saturated aqueous CuSO4 and water. The
J3À4=6.4, H-3), 6.91 (1H, d, J1À2=4.7, H-1), 7.30–7.65
(9H, m, Ar), 7.90–8.15 (6H, m, Ar); 13C NMR
(100.5 MHz; CDCl3) dC 60.86 (C-5), 75.46 (C-3), 76.72
(C-2), 79.06 (C-4), 93.94 (C-1), 128.62, 128.68, 128.82,
128.86, 128.92, 129.02, 129.43, 129.48, 130.01, 130.07,
130.17, 130.19, 130.22, 133.73, 133.78, 133.85, 134.03,
134.07 (18ÂAr), 165.24, 165.53, 166.37 (3ÂC¼O); MS
m/z (relative intensity) FAB+ 341 (100, M+H+–
C6H5COOH), 462 (1, M+H+); HRMS (FAB+) calcd
C26H22O8 (M+H+) 462.1315, found 462.1285.
5.2.3. N6-Benzoyl-9-(20,30-di-O-benzoyl-ꢀ-D-xylofurano-
syl)-adenine (11). A suspension of N6-benzoyladenine
(1.15 equiv, 2.49 mmol, 597 mg) in HMDS (15 mL) and
pyridine (0.4 mL) was refluxed under an atmosphere of
dry N2. After 68 h, the homogeneous yellow solution
was concentrated and azeotroped with dry toluene (2Â)
to yield a gummy residue. A solution of dry 1,2,3-tri-O-
benzoyl-d-xylofuranose 10 (1.005 g, 2.17 mmol) in DCE
(20 mL) was added to the bis-silylated benzoyladenine
with the careful exclusion of moisture. TMS-OTf (2.2
equiv, 4.67 mmol, 0.93 mL) was added under dry N2
and the solution heated under reflux. After 1 h, TLC
(1:1 CHCl3/Me2CO) indicated the formation of a major
product (Rf=0.55) and a minor product (Rf=0.13). 2-D
TLC was helpful due to streaking. The solution was
heated under reflux for a further 3 h but no change in
composition occurred. The solution was allowed to cool
and diluted with DCM. Washing with water (2Â), dry-
ing (MgSO4) and concentration yielded an orange foam
(1.72 g). Purification by SGC [CHCl3 followed by a lin-
ear gradient of 7:1!1:1 (v/v) CHCl3/Me2CO] yielded
the protected nucleoside as a yellow foam (820 mg,
65%). 1H NMR (399.8 MHz; CDCl3) dH 3.90, 4.05 (2H,
organicphase was dried (MgSO ) and concentrated to
4
yield an orange oil (4.31 g). Purification by SGC [linear
gradient; 15:1!4:1 (v/v) hexane/EtOAc+0.5% (v/v)
MeOH] yielded the title compound as a 2:1 mixture of
a- and b-anomers (1.65 g, 76%). a: 1H NMR
(399.8 MHz; CDCl3) dH 3.69, 3.78 (2H, ABX,
2JAB=9.8, JAX=4.9, JBX=4.4, H-5, H-50), 4.51 (2H,
s, CH2–Ar), 4.90–4.94 (1H, m, H-4), 5.97 (1H, dd,
J2À1=4.9, J2À3=5.4, H-2), 6.13 (1H, dd, J3À4=6.3,
J3À2=5.9, H-3), 6.95 (1H, d, J1À2 4.9, H-1), 7.10–7.60
(14H, m, Ar), 7.95–8.10 (6H, m, Ar). 13C NMR
(100.5 MHz; CDCl3) dC 68.57 (C-5), 74.42 (CH2–Ar),
3
3
1
75.72 (C-3), 77.00 (C-2), 78.06 (C4), 94.54 (C-1). b: H
NMR (399.8 MHz; CDCl3) dH 3.85–3.92 (2H, m, H-5,
H-50), 4.44, 4.52 (2H, AB, JAB=12.2, CH2–Ar), 4.94–
4.98 (1H, m, H-4), 5.79 (1H, s, H-2), 5.91 (1H, d,
J3À4=4.9, H-3), 6.67 (1H, s, H-1), 7.10–7.60 (14H, m,
Ar), 7.95–8.10 (6H, m, Ar). 13C NMR (100.5 MHz;
CDCl3) dC 68.83 (C-5), 74.26 (CH2–Ar), 75.03 (C-3),
80.69 (C-2), 82.54 (C-4), 100.06 (C-1). a+b 128.96,
129.03, 129.07, 129.12, 129.27, 129.38, 129.52, 129.69,
129.87, 130.04, 130.09, 130.46, 130.51, 130.57, 130.64,
134.11, 134.19, 134.32, 134.48, 138.23, 138.28 (24xAr),
165.42, 165.61, 165.73, 165.99, 166.08, 166.45 (6ÂC¼O).
ABX, JAB=12.9, JAX=3.5, JBX=2.9, H-50A, H-50B),
2
3
3
4.79 (1H, ‘dt’ (ddd), J4À5ꢁJ4À5 =3.2, J4À3=6.4, H-40),
0
6.00 (1H, ‘t’ (dd), J3À4–J3À2=6.0, H-30), 6.27 (1H, d,
J1À2=5.3, H-10), 6.56 (1H, ‘t’ (dd), J2À1–J2À3=5.6, H-
20), 7.35–7.55 (9H, m, Ar), 7.83 (1H, br d, NH), 7.98
(2H, d, J 7.4, CHCC¼O), 8.02 (2H, d, J=9.4,
CHCC¼O), 8.12 (2H, d, J=8.8, CHCC¼O), 8.26 (1H,
d, J=2.3, H-2), 8.68 (1H, s, H-8); 13C NMR
(100.5 MHz; CDCl3) dC 60.91 (C-50), 75.07 (C-30), 77.52
(C-20), 81.04 (C-40), 87.69 (C-10), 127.59 (C-2), 127.86,
128.48, 128.58, 128.73, 128.87, 130.14, 130.17, 130.69,
131.86, 132.28, 133.21, 133.98, 135.76, 136.07, 141.33
(Ade), 150.47 (Ade), 151.38 (Ade, C-8), 160.08 (Ade),
163.17, 165.08, 166.19, 170.37; MS m/z (relative inten-
sity) FAB+ 341(100, M+H+–C6H5COOH), 580(59,
M+H+); HRMS (FAB+) calcd C26H22O8 (M+H+)
580.1832, found 580.1851. UV lMAX=279 nm (EtOH).
5.2.2. 1,2,3-tri-O-Benzoyl-D-xylofuranose (10). 5-O-Ben-
zyl-1,2,3-tri-O-benzoyl-d-xylofuranose 9 (1.652 g, 2.99
mmol) was dissolved in a mixture of EtOH (30 mL) and
cyclohexene (15 mL). Palladium hydroxide on carbon
[0.4 g, ꢁ1/4 (w/w)] was added and the suspension
refluxed for 1 h. TLC (3:1 hexane/EtOAc) indicated the
formation of one product (Rf=0.12) and the complete
consumption of starting material. The Pd(OH)2/C was
filtered off through Celite, washed thoroughly with
CHCl3, and the combined organic fraction con-
centrated. Purification by SGC [linear gradient; 9:1!3:1
(v/v) hexane/EtOAc+0.5% (v/v) MeOH] yielded the
title compound as an inseparable anomeric mixture (b/
a) (1:2 ratio) (1.271 g, 92%). b: H NMR (399.8 MHz;
5.2.4. xylo-Adenosine-50-phosphate (13). N6-Benzoyl-9-
(20,30-di-O-benzoyl-b-d-xylofuranosyl) adenine 11 (446
mg, 0.77 mmol) was dissolved in triethylphosphate (3
mL) and the yellow solution heated gently under N2
until some vapour was seen to condense on the neck of
the flask. The solution was cooled to 0 ꢀC and phos-
phorous oxychloride (1.1 equiv, 0.85 mmol, 0.079 mL)
added dropwise with stirring. After 1 h, TLC (1:1
CHCl3/Me2CO) indicated the complete consumption of
starting material and the formation of a very polar UV
active spot on the TLC baseline. MQ water (90 mL) was
1
CDCl3) dH 2.57 (1H, br s, OH), 3.80–4.03 (2H, m, H-5,
H-50), 4.84–4.88 (1H, m, H-4), 5.81 (2H, br s, H-2, H-3),
6.63 (1H, s, H-1), 7.30–7.65 (9H, m, Ar), 7.90–8.15 (6H,
m, Ar); 13C NMR (100.5 MHz; CDCl3) dC 61.48 (C-5),
75.05 (C-3), 80.60 (C-2), 83.98 (C-4), 99.61 (C-1),
164.99, 165.09, 165.79 (3ÂC¼O). a: 1H NMR
(399.8 MHz; CDCl3) dH 2.57 (1H, br s, OH), 3.80–3.88
(2H, m, H-5, H-50), 4.80 (1H, ‘dt’ (ddd),
0
J4À5ꢁJ4À5 =4.4, J4À3=6.7, H-4), 5.91 (1H, ‘t’ (dd),
J2À1=4.7, J2À3=5.6, H-2), 6.07 (1H, ‘t’ (dd), J3À2=5.6,

482
C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
added to the reaction mixture followed by ether (90
mL). The resultant white precipitate, identified by TLC
as the polar product was filtered off, washed with water
and dried under vacuum. MS for 12 m/z (relative inten-
sity) FAB+ 240(100), 556(61, M+–C6H5CO–), 659 (62,
M+); HRMS (FAB+) calcd C31H26N5O10P1 (M+)
659.1417, found 659.1424.
free acid) in dry MeOH (5 mL). The resulting solution
was stirred at rt for 30 min, concentrated and then co-
evaporated from dry DMF (3Â). The off-white solid was
added to a solution of the activated 20,30-di-O-acetyl-
NMN in a mixture of dry DMF (1.0 mL) and pyridine
(2.0 mL). After 36 h at rt, the reaction mixture was con-
centrated and dissolved in a pre-cooled (0 ꢀC) mixture of
MQ water and concentrated ammonia [20:1 (v/v), 2 mL
total volume]. The flask was maintained at 0 ꢀC for 6 h
and concentrated. The residue was dissolved in MQ
water, washed with CHCl3 (2Â) and concentrated
(T<20 ꢀC). The solid was re-dissolved in MQ water and
the pH of the solution brought carefully to pH8 with
1 M NaOH. Purification was achieved by ion-exchange
chromatography on BioRad AG-MP1 resin using a
concave gradient of 150 mM TFA (0 min: 100%
MQ!20 min: 100% MQ!40 min: 96% MQ!60 min:
84% MQ!75 min: 68% MQ!105 min: 0% MQ!120
min: 0% MQ). The title compound eluted at 17–25%
TFA. Fractions containing >95% of the product (as
determined by HPLC) were pooled. Concentration and
co-evaporation (3Â) from MQ water, followed by lyo-
philisation yielded the product (free acid) as a white,
The crude protected nucleotide was dissolved in a solu-
tion of methanol (50 mL) through which ammonia gas
had been bubbled for 10 min. The yellow solution was
cautiously heated at ꢁ50 ꢀC in a sealed tube for 36 h.
Monitoring by HPLC (10 mm Partisil SAX anion-
exchange column (15 cm), isocratic elution 0.05 M
KH2PO4/5% MeOH/pH 3) revealed the relatively rapid
formation of a peak with RT=3.3 min (N6-benzoyl-
xyloAMP) and the gradual formation of a peak with
RT=2.6 min. The reaction mixture was concentrated,
leaving a yellow solid. The solid was redissolved in MQ
water (80 mL) and brought to pH 8 with 1 M NaOH.
Purification was effected by ion-exchange chromato-
graphy on BioRad AG-MP1 resin using a concave gra-
dient of 150 mM TFA (0 min: 100% MQ!20 min:
100% MQ!40 min: 96% MQ!60 min: 84% MQ!75
min: 68% MQ!105 min: 0% MQ!120 min: 0% MQ).
The product eluted at 28–45% TFA. Fractions con-
taining >95% of the product (as determined by HPLC)
were pooled. Concentration and co-evaporation (3Â)
from MQ water yielded the product as a colourless,
1
fluffy solid in 12% yield. H NMR (399.8 MHz; D2O)
dH 4.20–4.75 (10H, m, xylo- and ribofuranoside pro-
tons), 6.13 (1H, s, H-10), 6.16 (1H, d, J=5.0, H-100), 8.28
(1H, ‘t’ (dd), J=6.4, HNIC-5), 8.41 (1H, s, H2), 8.57
(1H, br s, H8), 8.92 (1H, d, J=7.9, HNIC-4), 9.28 (1H,
d, J=5.3, HNIC-6), 9.41 (1H, s, HNIC-2); 31P NMR
(161.9 MHz; D2O) dP À10.57!À10.02 (m, P-O-P). MS
m/z (relative intensity) FAB+ 665(30).
1
vitreous solid. lMAX=258 nm. H NMR (399.8 MHz;
D2O) dH 3.95–4.04 (2H, m, H-50A, H-50B), 4.30 (1H, 0s,
0
H-30), 4.42 (1H, d, J4 À5 =4.7, H-4 ), 4.52 (1H, s, H-2 ),
5.90 (1H, s, H-10), 8.00 (1H, s, H-8), 8.30 (1H, s, H-2);
31P NMR (161.9 MHz; D2O) dP 5.31 (s, decoupled);
0
0
5.2.6. 5-O-Benzoyl-1,2-O-isopropylidene-ꢁ-D-xylofura-
nose (16).43 A solution of 1,2-O-isopropylidene-a-d-
xylofuranose 15 (25.0 g, 0.131 mol) in dry pyridine (125
mL) was cooled to À20 ꢀC under dry N2. A solution of
benzoyl chloride (1.0 equiv, 0.131 mol, 18.48 g, 15.3
mL) in dry pyridine (70 mL) was added dropwise with
rapid stirring over 2 h at À20 ꢀC. TLC (9:1 CHCl3/
Me2CO) indicated a very small quantity of starting
material (Rf=0.07), one main product (Rf=0.52) and a
small side-product (Rf=0.89). The orange suspension
was diluted with DCM and washed with aqueous
ammonium chloride and water. The organic phase was
separated, dried (MgSO4) and concentrated. The residue
was co-evaporated from toluene (3Â) to remove traces of
pyridine. The clear orange oil (50.1 g>100%) was pur-
ified by SGC (CHCl3 followed by a linear gradient of
20:1!15:1 (v/v) CHCl3/Me2CO) to yield an off-white
solid (27.1 g, 70.2%). Mp [pentane/ether, 1:1 (v/v)] 83–
0
5.24 (t, JPÀCÀH5 =6.3, coupled). MS m/z (relative inten-
sity) FAB+ 348 (40); HRMS (FAB+) calcd
C10H15N5O7P (M+) 348.0709, found 348.0680.
5.2.5. xylo-Adenosine 50-diphospho-500-ribofuranoside-10-
nicotinamide (xyloNAD+) (14).Activation of NMN. A
solution of NMN (50 mg, 0.15 mmol, free acid) in MQ
water (0.3 mL) was added dropwise to a pre-cooled
(0 ꢀC), rapidly stirred, mixture of acetic anhydride (6
mL) in pyridine (10 mL). After 48 h at 0 ꢀC, the reaction
was concentrated (T<20 ꢀC) and the residue stirred
with a mixture of MQ water and pyridine [20:1 (v/v), 1
mL total volume] for 20 min at rt. The solvent was
removed in vacuo, and the residue evaporated from dry
DMF (3Â). Diphenyl phosphorochloridate (2.0 equiv,
0.30 mmol, 0.062 mL) and tri-n-butylamine (2.0 equiv,
0.30 mmol, 0.072 mL) were added to a solution of the
freshly prepared 2,3-di-O-acetyl-NMN in a mixture of
DMF (1.0 mL) and dioxane (2.0 mL) under an atmos-
phere of dry N2. The solution was stirred at rt for 1 h
after which the solvent was removed in vacuo. Excess
diphenyl phosphorochloridate was removed by repeated
trituration (3Â) with cold, dry ether before finally co-
evaporating with dry DMF (3Â).
85 ꢀC. H NMR data consistent with literature data.43
1
MS m/z (relative intensity) FAB+ 237 (92, M+H+–
Me2CO), 295 (100, M+H+); HRMS (FAB+) calcd
C15H19O6 (M+H+) 295.1182, found 295.1186.
5.2.7. 5-O-Benzoyl-1,2-O-isopropylidene-ꢁ-D-ribofura-
nose (17). Chromium trioxide (4.0 equiv, 0.190 mol,
19.03 g) was added to a stirred solution of pyridine (8.0
equiv, 0.386 mol, 30.8 mL) in DCM (400 mL) at rt. The
red solution was stirred rapidly at rt for 15 min and then
cooled to 0 ꢀC. 5-O-Benzoyl-1,2-O-isopropylidene-a-d-
xylofuranose 16 (14.0 g, 0.048 mol, 1.0 equiv) was
added, followed immediately by acetic anhydride (4.0
Coupling. Tri-n-octylamine (2.0 equiv, 0.334 mmol,
0.118 g, 0.146 mL) was added dropwise to a suspension
of xylo-adenosine-50-phosphate (58 mg, 0.167 mmol,

C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
483
equiv, 0.190 mol, 18.0 mL). The suspension was stirred
at 0 ꢀC for 10 min. A small aliquot of the reaction mix-
ture was removed, co-evaporated from toluene (3Â) and
re-dissolved in DCM. TLC (9:1 CHCl3/Me2CO) indi-
cated that all the starting material (Rf=0.52) had been
consumed and that only a less polar streaking spot was
present. The reaction mixture was poured into EtOAc (1
L) and the precipitate filtered off through a Buchner
funnel containing a layer of silica (2 cm) over Celite (1
cm). The precipitate was washed thoroughly with
EtOAc and the combined organic fraction concentrated
(T<25 ꢀC) to yield a brown mobile oil (20.5 g>100%).
J2À1=3.8, H-1), 7.43 (2H, ‘t’ (dd), JAr2ÀAr1=7.9,
JAr2ÀAr3=7.2, Ar-2), 7.56 (1H, ‘t’ (dd), JAr3ÀAr2 7.2, Ar-
3), 8.05 (2H, d, JAr1ÀAr2 7.6, Ar-1); 13C NMR
(100.5 MHz; CDCl3) dC 26.07 (isopropylidene CH3),
26.56 (isopropylidene CH3), 61.20 (d, JC-5-F=10.7, C-5),
77.53 (d, JC-4-F=18.3, C-4), 82.28 (d, JC-2-F=32.0, C-2),
93.81 (d, JC-3-F=183.9, C-3), 104.82 (C-1), 112.08 (iso-
propylidene C(CH3)2), 128.07 (Ar), 129.30 (Ar), 129.45
(Ar), 132.89 (Ar quat.), 165.67 (C¼O); 19F NMR
(376.2 MHz; CDCl3) dF À208.89 (ddd, JF-3=50.1, JF-
4=30.8, JF-2=11.0); MS m/z (relative intensity) FAB+
239 (77, M+H+ –CH3COCH3), 297 (100, M+H+);
HRMS (FAB+) calcd C15H17O5F1 (M+H+) 296.1060,
found 296.1051.
The crude 5-O-benzoyl-1,2-O-isopropylidene-a-d-3-
ulose was dissolved in THF (120 mL), cooled to 0 ꢀC
and water (96 mL) added. Sodium borohydride (1.1
equiv, 0.052 mol, 1.98 g) was added in five small por-
tions over 15 min. TLC (9:1 CHCl3/Me2CO) indicated
the formation of a new material (Rf=0.61) and the
complete consumption of the 3-ulose intermediate. The
reaction was diluted with CHCl3 and washed with
water. The aqueous phase was extracted repeatedly (3Â)
with CHCl3 and the combined organic fractions dried
(MgSO4) and concentrated. Purification by SGC
[CHCl3 followed by 9:1 (v/v) CHCl3/Me2CO] yielded
the product as a white solid (11.37 g, 81%). Mp (ether/
pentane 1:1 v/v)=81–82 ꢀC. 1H NMR (399.8 MHz;
CDCl3) dH 1.37 (3H, s, isopropylidene CH3), 1.59 (3H,
s, isopropylidene CH3), 2.84 (1H, d, JOH-3=9.1, OH),
3.90–3.96 (1H, m, H-3), 4.11 (1H, ddd, J4À5=2.8,
5.2.9. 5-O-Benzoyl-3-deoxy-3-fluoro-D-xylofuranose (19).
5-O-Benzoyl-3-deoxy-3-fluoro-1,2-O-isopropylidene-d-
xylofuranose 18 (3.24 g, 19.0 mmol) was dissolved in a
mixture of acetic acid (80 mL) and water (20 mL) and
heated to 80 ꢀC. After 72 h, TLC (9:1 CHCl3/Me2CO)
indicated only a small quantity of SM (Rf=0.84) and
one main product (Rf=0.16). The reaction was con-
centrated and co-evaporated from toluene (3Â). Pur-
ification by SGC [linear gradient; 15:1!7:1 (v/v)
CHCl3/Me2CO] yielded the product as a colourless oil
1
(2.10 g, 75%). a: H NMR (399.8 MHz; CD3COCD3) d
2.97–2.03 (1H, br s, OH), 4.18–4.30 (1H, m, H-2),
H
4.41–4.60 (3H, m, H-5, H-50, H-4), 5.02 (1H, dm, J3-
F=52.6, H-3), 5.70 (1H, d, J1À2=6.4, H-1), 7.47 (2H, ‘t’
(dd), JAr2ÀAr1=7.3, JAr2ÀAr3=7.6, Ar-2), 7.59 (1H, ‘t’
(dd), JAr3ÀAr2=7.2, Ar-3), 7.99 (2H, d, JAr1ÀAr2=7.4,
Ar-1); a 19F NMR (376.2 MHz; CDCl3) dF À201.95
(ddd, JF-2=12.4, JF-4=28.9, JF-3=51.0); a 13C NMR
(100.5 MHz; CD3COCD3) dC 64.60 (d, JC-5-F=13.7, C-
5), 76.37 (d, JC-2-F=25.9, C-2), 79.5 (d, JC-4-F=19.1, C-
4), 97.52 (d, JC-3-F=185.4, C-3), 97.61 (d, JC-1-F=8.4,
0
J4À5 =5.6, J4À3=11.7, H-4), 4.45, 4.70 (2H, ABX,
2JAB=12.3, JAX=4.1, JBX 2.3, H-5, H-50) 4.61, (1H,
‘t’ (dd), J2À1=4.1, J2À3=4.7, H-2), 5.84 (1H, d,
J1À2=3.9, H-1), 7.42 (2H, ‘t’ (dd), JAr2ÀAr1=7.6,
JAr2ÀAr3=7.6, Ar-2), 7.55 (1H, ‘t’ (dd), JAr3ÀAr2=7.6,
Ar-3), 8.05 (2H, d, JAr1ÀAr2=7.6, Ar-1); 13C NMR
(100.5 MHz; CDCl3) dC 26.40 (2Â isopropylidene CH3),
63.22 (C-5), 71.87 (C-3), 78.12 (C-2+C-4), 103.74 (C-1),
112.49 (isopropylidene C(CH3)2), 128.04 (Ar), 129.38
(Ar), 129.43 (Ar), 132.84 (Ar quat.), 166.07 (C¼O); MS
m/z (relative intensity) FAB+ 237 (100, M+H+-
CH3COCH3), 295(40, M+H+); HRMS (FAB+) calcd
C15H18O6 (M+H+) 294.1104, found 294.1097.
3
3
1
C-1); b: H NMR (399.8 MHz; CD3COCD3) dH 2.97–
2.03 (1H, br s, OH), 4.18–4.30 (1H, m, H-4), 4.37–4.62
(3H, m, H-2, H-5, H-50), 5.10 (1H, d, J3-F=53.0, H-3),
5.46 (1H, s, H-1), 7.47 (2H, ‘t’ (dd), JAr2-Ar1=7.3, JAr2-
Ar3=7.6, Ar-2), 7.59 (1H, ‘t’ (dd), JAr3-Ar2=7.2, Ar-3),
7.99 (2H, d, JAr1ÀAr2=7.4, Ar-1);
(376.2 MHz; CDCl3) dF À204.22 (ddd, JF-2=15.2, JF-
4=26.2, JF-3=51.0);
13C NMR (100.5 MHz;
b
19F NMR
b
5.2.8.
5-O-Benzoyl-3-deoxy-3-fluoro-1,2-O-isopropyli-
CD3COCD3) dC 63.50 (d, JC-5-F=12.2, C-5), 76.50 (d,
JC-2-F=19.0, C-2), 80.82 (d, JC-4-F=26.2, C-4), 98.53 (d,
JC-3-F=182.3, C-3), 104.77 (C-1); a+b 129.62, 129.64,
130.48, 131.17, 131.25, 134.20, 134.24, 166.64, 166.7.;
MS m/z (relative intensity) FAB+ 239 (100, M+H+
–H2O), 257 (100, M+H+); HRMS (FAB+) calcd
C12H14O5F1 (M+H+) 257.0825, found 257.0804.
dene-D-xylofuranose (18). Diethylaminosulfur tri-
fluoride (1.1 equiv, 0.262 mol, 4.22 g, 3.46 mL) was
added slowly to a mixture of 5-O-benzoyl-1,2-O-iso-
propylidene-a-d-ribofuranose (7.00 g, 23.8 mmol) and
pyridine (3.6 equiv, 85.8 mmol, 6.79 g, 6.94 mL) in dry
DCM (120 mL) at À10 ꢀC under an atmosphere of dry
N2. After 4 days at rt, TLC (9:1 CHCl3/Me2CO) indi-
cated only a small quantity of SM and formation of one
main product (Rf=0.86). The orange solution was very
cautiously washed with water, dried (MgSO4) and con-
centrated to yield an orange oil (8.1 g, >100%). Puri-
fication by SGC [linear gradient; 15:1!11:1 (v/v)
hexane/EtOAc+0.5% (v/v) MeOH] yielded the title
compound as a pale yellow oil (5.71 g, 81%).1H NMR
(399.8 MHz; CDCl3) dH 1.33 (3H, isopropylidene CH3),
1.50 (3H, isopropylidene CH3), 4.50–4.63 (3H, m, H-5,
H-50, H-4), 4.73 (1H, dd, J2-F=11.1, J2À1=3.8, H-2),
5.05 (1H, dd, J3-F=50.4, J3À4 1.8, H-3), 6.03 (1H, d,
5.2.10. 1,2,5-Tri-O-benzoyl-3-deoxy-3-fluoro-D-xylofura-
nose (20). Benzoyl chloride (2.0 equiv, 0.164 mol, 2.30
g, 1.90 mL) was added to a pre-cooled (0 ꢀC) solution of
5-O-benzoyl-3-deoxy-3-fluoro-d-xylofuranose 19 (2.10
g, 8.20 mmol) and DMAP (75 mg, cat.) in dry pyridine
(100 mL) in four portions over a period of 8 h under an
atmosphere of dry N2. After a further 16 h at rt, TLC
(3:1 hexane/EtOAc) indicated the formation of several
products. DMAP (100 mg) was added to the reaction
mixture followed by benzoyl chloride (1.0 equiv, 8.20
mmol, 0.85 mL) dropwise and the reaction stirred for a

484
C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
further 24 h at rt. TLC indicated mainly two products
(Rf=0.38 and 0.43). The reaction mixture was con-
centrated and co-evaporated from toluene (2Â). The
resultant orange solid was re-dissolved in CHCl3 and
washed with saturated aqueous CuSO4 and water. Dry-
ing (MgSO4) and concentration yielded a dark orange
oil (6.81 g, >100%). Purification by SGC [linear gra-
dient; 15:1!3:1 (v/v) hexane/EtOAc+0.5% (v/v)
MeOH] yielded a 1:1 mixture of a- and b-anomers as a
colourless oil (1.61 g, 42%).
tube for 36 h after which TLC (4:1 CHCl3/Me2CO)
indicated the formation of one main product (Rf=0.30).
The reaction mixture was concentrated and the residue
dissolved in MQ water. The solution was washed with
ether (1Â) and the nucleoside-containing aqueous phase
concentrated to yield the title compound as a yellow
powder (160 mg, 59%). lMAX=259 nm (EtOH). 1H
NMR (399.8 MHz; CD3SO2CD3) dH 3.20–3.50 (1H, br
2
3
s, OH), 3.71, 3.77 (2H, ABX, JAB=11.4, JAX=6.1,
JBX=5.5, H-50A, H-50B), 4.32 (1H, dsex, J4 -F=28.4,
3
0
J4 À5A =6.1, J4 À5B =6.0, J4 À3 3.5, H-40), 4.76 (1H, ‘dt’
0
0
0
0
0
0
1
0
0
0
0
0
0
a: H NMR (399.8 MHz; CDCl3) dH 4.57–4.96 (3H, m,
(ddd), J2 -F=15.8, J2 À3 =4.4, J2 À1 =2.1, H-2 ), 5.11
H-5, H-50, H-4), 5.60 (1H, ddd, J3-F=53.0, J3À4=4.7,
J3À2=3.2, H-3), 5.82 (1H, ddd, J2-F=22.2, J2À3=5.0,
J2À1=3.2, H-2), 6.91 (1H, d, J1À2=5.0, H-1), 7.30–8.13
(15H, m, Ar); a 19F NMR (376.2 MHz; CDCl3) dF
À202.92 (‘dt’ (ddd), JF-4=21.4, JF-2=22.9, JF-3=53.4);
a: 13C NMR (100.5 MHz; CDCl3) dC 61.5 (d, JC-5-
F=11.6, C-5), 77.30 (2C, 2d, JC-2-F=JC-4-F 28.0, C-2, C-
4), 93.95 (d, JC-3-F=189.4, C-3), 94.26 (C-1); b 1H
NMR (399.8 MHz; CDCl3) dH 4.57–4.96 (3H, m, H-5,
H-50, H-4), 5.38 (1H, dd, J3-F 50.4, J3À4 4.1, H-3), 5.80
(1H, d, J2-F 11.0, H-2), 6.67 (1H, s, H-1), 7.30–8.13
(15H, m, Ar); b: 19F NMR (376.2 MHz; CDCl3) dF
(1H, ddd, J3 -F=51.8, J3 À4 =3.2, J3 À2 =1.8, H-3 ), 5.95
0
0
0
0
0
0
0
0
0
(1H, d, J1 À2 =2.3, H-1 ), 7.36 (1H, br s, NH), 8.10 (1H,
s, H-8), 8.17 (1H, s, H-2); 19F NMR (376.2 MHz;
0
0
CD3SO2CD3) dF À201.26 (ddd, JF-3 =51.0, JF-4 =27.5,
JF-2 =15.1); 13C NMR (100.5 MHz; CD3SO2CD3) dC
0
58.17 (d, JC-5 -F10.3, C-50), 77.47 (d, JC-2 -F=25.7, C-20),
0
0
81.80 (d, JC-4 -F=19.3, C-40), 88.57 (d, JC-1 -F=1.9, C-
0
0
10), 95.50 (d, JC-3 -F=183.0, C-30), 118.63 (C-5), 138.26
0
(C-8), 149.02 (C-4), 152.62 (C-2), 155.88 (C-6); MS m/z
(relative intensity) FAB+ 270 (100, M+H+); HRMS
(FAB+) calcd C10H12N5O3F1 (M+) 269.0924, found
269.0930.
À203.58 (ddd, JF-2=10.7, JF-4=25.9, JF-3=48.8); b: 13
C
NMR (100.5 MHz; CDCl3) dC 62.37 (d, JC-5-F=12.9, C-
5), 79.26 (d, JC-2-F=30.3, C-2), 81.08 (d, JC-4-F=20.0,
C-4), 92.37 (d, JC-3-F=191.3, C-3), 99.08 (C-1); a+b:
128.21, 128.23, 128.29, 128.36, 128.38, 128.50, 129.37,
129.60, 129.67, 129.79, 129.87, 130.00, 133.09, 133.12,
133.34, 133.48, 133.54, 133.83, 164.42, 164.50, 164.59,
164.81, 165.95, 167.27. MS m/z (relative intensity) Elec-
trospray 343 (70, M–PhCOO), 487 (100, M+Na+), 951
(90, 2 M+ Na+).
5.2.13. 30-Deoxy-30-fluoro-xylo-adenosine monophosphate
(23). 30-Deoxy-30-fluoro-xylo-adenosine 22 (120 mg,
0.446 mmol) was dissolved in triethylphosphate (4 mL)
by gentle warming with a heat-gun until some vapour
was seen to condense on the neck of the round-bottom
flask. The stirred solution was cooled to 0 ꢀC and phos-
phorous oxychloride (1.2 equiv, 0.535 mmol, 82 mg,
0.050 mL) added dropwise under dry N2. The solution
was allowed to warm slowly to rt and stirred overnight.
HPLC (SAX) indicated the formation of a new peak
(RT=3.1 min) and approximately 50% starting material
(RT=2.8 min). The yellow solution was cooled to 0 ꢀC
and phosphorous oxychloride (0.6 equiv, 0.268 mmol,
41 mg, 0.025 mL) added dropwise. After 18 h at rt,
HPLC revealed the composition of the reaction mixture
to still consist of ꢁ20% of the nucleoside SM. Further
addition of phosphorous oxychloride (0.2 equiv, 0.089
mmol, 14 mg, 8.3 mL) and stirring for 12 h at rt failed to
drive the reaction to completion. MQ water (150 mL)
was cooled to 0 ꢀC with rapid stirring and the reaction
mixture added dropwise. After 30 min, the aqueous
phase was washed with EtOAc(6 Â) to remove triethyl-
phosphate. The solution was brought to pH=11 with
1.0 M NaOH and purified by chromatography using
AG-MP1 strong ion-exchange resin. The title com-
pound eluted at 52–70% of 150 mM TFA and was iso-
5.2.11. N6-Benzoyl-9-(20-O-benzoyl-30-deoxy-30-fluoro-
ꢀ-D-xylofuranosyl)-adenine (21). Compound 21 was
obtained as described above for compound 11 starting
from 452 mg (0.973 mmol) of 20, yielding a pale yellow
oil (125 mg, 22%). Rf=0.44 (95:5 CHCl3/MeOH).
l
MAX=280 nm (EtOH). 1H NMR (399.8 MHz; CDCl3)
dH 4.73–4.86 (3H, m, H-40, H-50A, H-50B), 5.42 (1H, dd,
0
0
0
0
0
J3 -F=50.1, J3 À4 =2.0, H-3 ), 5.920 (1H, d, J2 -F=13.3,
H-20), 6.52 (1H, d, J1 À2 =1.5, H-1 ), 7.25–7.50 (7H, m,
Ar), 7.54–7.65 (2H, m, Ar), 8.03–8.06 (6H, m, Ar), 8.30
(1H, s, H-8), 8.76 (1H, s, H-2), 9.31 (1H, br s, NH); 19F
0
0
0
NMR (376.2 MHz; CDCl3) dF À201.80 (ddd, JF-2
=12.4, JF-4 =30.3, JF-3 =49.6); 13C NMR (100.5 MHz;
0
0
CDCl3) dC 61.43 (d, JC-5 -F=10.3, C-50), 80.21 (d, JC-4 -
0
0
F=31.1, C-40), 80.39 (d, JC-2 -F=19.3, C-20), 87.39 (C-
0
10), 93.44 (d, JC-3 -F=197.5, C-30), 128.83, 128.23,
0
128.66, 128.69, 128.87, 129.46, 129.96, 130.20, 131.64,
133.63, 134.34, 136.09, 139.80 (Ade, C-8), 139.88 (Ade),
151.20 (Ade), 151.64 (Ade, C-2), 160.04 (Ade), 163.06,
164.78, 166.26 (3ÂC¼O); MS m/z (relative intensity)
FAB+ 581 (100, M+);
lated as a vitreous, colourless solid (104mg, 67%)
1
l
MAX=259 nm (water, pH=7). H NMR (399.8 MHz;
D2O) dH 4.18–4.31 (2H, m, H-50A, H-50B), 4.68 (1H, ‘dtt’
0
0
0
0
0
0
00
(dddd), J4 -F=29.0, J4 -C5 -O-P=6.7, J4 -5A ꢁJ4 À5B =4.1,
H-40), 4.79 (1H, d, J2 -F=13.1, H-20), 5.22 (1H, d, J3 -F
50.7, H-30), 6.14 (1H, s, H-10), 8.36 (1H, s, H-8), 8.37
(1H, s, H-2); 19F NMR (376.2 MHz; D2O) dF À202.36
0
0
5.2.12. 30-Deoxy-30-fluoro-xylo-adenosine (22). N6-Ben-
zoyl-9-(20 -O-benzoyl-30 -deoxy-30 -fluoro-b-d-xylofur-
anosyl)adenine (21) (582 mg, 1.00 mmol) was dissolved
in a solution of methanol (50 mL) through which
ammonia gas had been bubbled for 10 min. The yellow
solution was cautiously heated at ꢁ50 ꢀC in a sealed
(ddd, JF-3 =51.0, JF-4 =28.9, JF-2 =13.8); 31P NMR
(161.9 MHz; D2O) dP 1.55 (s, decoupled); MS m/z
(relative intensity) FAB+ 329(100, MÀ2H+–H2O), 350
(13, M+H+); HRMS (FAB+) calcd C10H13N5O6F1P1
(M+) 349.0587, found 349.0586.
0
0
0

C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
485
5.2.14. 30F-xylo-Adenosine 50-diphospho-500-ribofurano-
side-100-nicotinamide (30F-xyloNAD+) (24). Compound
24 was obtained as described above for compound 14
starting from 50.0 mg (0.143 mmol, free acid) of 23 to
give a white, fluffy solid (22.8 mmol, 16%, quantified by
5.2.16. N6-Benzoyl-9-(20-O-benzoyl-30-deoxy-30-fluoro-
ꢀ-D-ribofuranosyl) adenine (26). Compound 26 was
obtained as described above for compound 11 starting
from 265 mg (6.59 mmol) of 25, yielding a yellow foam
(194 mg, 51%). Rf=0.74 (9:1 CHCl3/MeOH);
1
1
UV); lMAX=258 nm. H NMR (399.8 MHz; D2O) dH
lMAX=280 nm (EtOH); H NMR (399.8 MHz; CDCl3)
4.18–4.50 (8H, m, F-xylo- and ribofuranoside protons),
0
4.63–4.74 (1H, m), 5.26 (1H, ddd, J3 -F=51.0,
dH 4.55–4.88 (3H, m, H-4, H-5), 6.00 (1H, ddd, J3-F
=53.6, J3À4=4.7, J3À2=2.3, H-3), 6.37 (1H, ddd, J2-F
=17.6, J2À1=5.7, J2À3=5.0, H-2), 6.46 (1H, d, J1À2=6.4,
H-1), 7.35–7.62 (6H, m, Ar), 8.04–8.11 (4H, m, Ar); 19F
NMR (376.2 MHz; CDCl3) dF À199.91 (ddd, JF-3=53.7,
JF-4=23.4, JF-2=17.9); 13C NMR (100.5 MHz; CDCl3)
dC 63.33 (d, JC-5-F=8.4, C-5), 73.64 (d, JC-2-F=14.8, C-
2), 81.49 (d, JC-4-F=24.5, C-4), 86.23 (C-1), 89.55 (d, JC-
3-F=190.7, C-3), 123.89 (Ar), 128.03 (Ar), 128.41 (Ar),
128.65 (Ar), 128.78 (Ar), 128.87 (Ar), 129.06 (Ar), 129.32
(Ar), 129.83 (Ar), 129.93 (Ar), 130.20 (Ar), 133.03 (Ar),
133.62 (Ar), 133.79 (Ar), 134.15 (Ar), 142.01 (Ade),
149.89 (Ade), 151.94 (Ade), 153.07 (Ade), 164.58 (C¼O),
165.40 (C¼O), 166.14 (C¼O); MS m/z (relative intensity)
FAB+ 343 (53, M+H+–(Bz-N6-Ade)), 582 (100,
M+H+), 583 (33, M+2H+);
0
0
0
0
0
0
0
J3 À4 =3.2, J3 –2 =1.8, H-3 ), 6.00 (1H, d, J1 À2 =2.3, H-
10), 6.04 (1H, d, J=5.5, H-100), 8.13 (1H, s, H-8), 8.18
(1H, dd, J=6.1, J=7.9, HNIC-5), 8.20 (1H, s, H-2), 8.80
(1H, d, J=8.2, HNIC-6), 9.15 (1H, d, J=6.1, HNIC-4),
9.29 (1H, d, J=5.3, HNIC-2), 19F NMR (376.2 MHz;
0
0
0
D2O) dF À202.03 (ddd, JF-3 =51.0, JF-4 =26.6, JF-2
=13.2); 31P NMR (161.9 MHz; D2O) dP À10.3 (d, JP-O-
P=20.6), À10.6 (d, JP-O-P=21.3); MS m/z (relative
intensity) FAB+ 666 (35, M+H+).
5.2.15. 1-O-Acetyl-2,5-di-O-benzoyl-3-deoxy-3-fluoro-D-
ribofuranose (25). Compound 25 was synthesised in 12
steps according to the procedure of Mikhailopulo et
al.44 b: TLC (3:1 hexane/EtOAc) Rf=0.35, colourless
1
oil (210 mg, quant); H NMR (399.8 MHz; CDCl3) dH
1.97 (3H, s, COCH3), 4.45–4.71 (2H, m, H-5), 4.72 (1H,
0
5.2.17. 30-Deoxy-30-fluoro-adenosine (27). Compound 27
was obtained as described above for compound 22
starting from 220 mg (0.378 mmol) of 26, to yield an
orange powder (132 mg, >100%). Rf=0.37 (4:1
‘dq’ (dddd), J4-F=19.0, J4À5=4.4, J4À5 =4.4, J4À3=4.1,
H-4), 5.38 (1H, ddd, J3-F=55.0, J3À4=5.5, J3À2=1.5, H-
3), 5.60 (1H, ddd, J2-F=8.2, J2À3=5.0, J2À1=2.3, H-2),
6.41 (1H, ‘t’ (dd), J1-F=2.1, J1À2=1.8, H-1), 7.35–7.62
(6H, m, Ar), 8.04–8.11 (4H, m, Ar); 19F NMR
(376.2 MHz; CDCl3) dF À209.25 (ddd, JF-3=51.0, JF-4
=17.9, JF-2=8.3); 13C NMR (100.5MHz; CDCl3) dC 21.1
(COCH3), 63.17 (d, JC-5-F=5.1, C-5), 75.00 (d, JC-2-F
=14.2, C-2), 82.31 (d, JC-4-F=25.1, C-4), 89.24 (d, JC-3-
F=194.6, C-3), 98.13 (C-1), 128.36 (Ar), 128.41 (Ar),
128.50 (Ar), 129.49 (Ar), 129.57 (Ar), 129.75 (Ar),
129.82 (Ar), 133.27 (Ar), 133.34 (Ar), 133.57 (Ar),
133.61 (Ar), 164.97 (C¼O), 165.08 (C¼O), 165.67
(C¼O), 165.71 (C¼O), 168.91 (C¼O), 169.66 (C¼O);
MS m/z (relative intensity) FAB+ 343 (100, M+H+-
AcOH), 402 (1, M+); HRMS (FAB+) calcd
C21H19F1O7 (M+) 402.1115, found 402.1131.
CHCl3/MeOH);
(399.8 MHz; CD3COCD3) dH 3.61–3.68 (2H, m, H-50),
l
MAX=259nm (EtOH); 1H NMR
0
0
0
0
0
4.29 (1H, ‘dt’ (ddd), J4 -F=27.5, J4 À3 =3.5, J4 À5 =3.5,
H-40), 4.95 (1H, ddd, J2 -F=36.4, J2 À1 =7.8, J2 À3 =4.1,
0
0
0
0
0
H-20), 5.10 (1H, dd, J3 -F=54.8, J3 À4 =4.4, H-3 ), 5.95
0
0
0
0
0
0
0
(1H, d, J1 -2 =8.2, H-1 ), 7.44 (1H, br s, NH), 8.15 (1H,
s, H-8), 8.39 (1H, s, H-2); 19F NMR (376.2 MHz;
CD3COCD3) dF À197.54 (ddd, JF-3=53.7, JF-4=26.2,
JF-2=26.2); 13C NMR (100.5 MHz; CD3COCD3) dC
61.02 (d, JC-5-F=10.9, C-5), 71.87 (d, JC-2-F=16.1, C-2),
83.98 (d, JC-4-F=21.2, C-4), 86.79 (C-1), 93.15 (d, JC-3-F
=181.7, C-3), 119.28 (Ade), 140.00 (Ade), 149.01 (Ade),
152.32 (Ade), 156.11 (Ade); MS m/z (relative intensity)
Electrospray 270 (100, M+H+).
a: TLC (3:1 hexane/EtOAc) Rf=0.39, white foam (715
mg, 91%); 1H NMR (399.8 MHz; CDCl3) dH 1.97
(3H, s, COCH3), 4.45–4.70 (2H, m, H-5), 4.85 (1H,
5.2.18. 30-Deoxy-30-fluoroadenosine 50-monophosphate
(28). Compound 28 was obtained as described above
for compound 23 starting from 90 mg (0.334 mmol) of
27, to yield an vitreous, colourless solid (97mg, 83%).
0
‘ddt’ (dddd), J4-F=26.1, J4À5=3.8, J4À5 =3.8,
J4À3=1.5, H-4), 5.36 (1H, ‘dt’ (ddd), J2ÀF=21.7,
J2À3=4.7, J2À1=5.0, H-2), 5.47 (1H, ‘dt’ (ddd), J3-
F=57.0, J3À2=4.7, J3À4=1.5, H-3), 6.65 (1H, d,
J1À2=4.4, H-1), 7.35–7.62 (6H, m, Ar), 8.04–8.11 (4H,
m, Ar); 19F NMR (376.2 MHz; CDCl3) dF À195.64
(ddd, JF-3=56.5, JF-4=26.0, JF-2=22.0); 13C NMR
(100.5 MHz; CDCl3) dC 20.79 (COCH3), 61.30 (d, JC-
5-F=9.7, C-5), 71.61 (d, JC-2-F=15.5, C-2), 80.72 (d,
JC-4-F=25.1, C-4), 88.30 (d, JC-3-F=191.3, C-3), 93.89
(C-1), 128.36 (Ar), 128.41 (Ar), 128.50 (Ar), 129.49
(Ar), 129.57 (Ar), 129.75 (Ar), 129.82 (Ar), 133.27
(Ar), 133.34 (Ar), 133.57 (Ar), 133.61 (Ar), 164.97
(C¼O), 165.08 (C¼O), 165.67 (C¼O), 165.71 (C¼O),
168.91 (C¼O), 169.66 (C¼O); MS m/z (relative inten-
sity) FAB+ 343 (100, M+H+–AcOH), 402 (1, M+);
HRMS (FAB+) calcd C21H19F1O7 (M+) 402.1115,
found 402.1131.
l
MAX=258nm (H2O, pH=6). 1H NMR (399.8MHz;
D2O) dH 4.07–4.21 (2H, m, H-50), 4.62–4.72 (1H, m, J4 -F
0
=26.9, H-40), 4.93 (1H, ddd, J2 -F=24.6, J2 À1 =4.7,
0
0
0
0
0
0
0
0
0
J2 À3 =4.1, H-2 ), 5.32 (1H, dd, J3 ÀF=53.3, J3 À4 =4.1,
H-30), 6.21 (1H, d, J1 À2 =7.9, H-1 ), 8.39 (1H, s, H-8),
8.59 (1H, s, H-2); 19F NMR (376.2 MHz; D2O) dF
À198.97 (ddd, JF-3=50.9, JF-4=24.8, JF-2=24.8);
31P NMR (161.9 MHz; D2O) dP 1.34 (s, decoupled);
MS m/z (relative intensity) FAB+ 349 (0.5,M+),
369 (5, MÀ3H++Na+); HRMS (FAB+) calcd
C10H13N5O6F1P1 (M+) 349.0588, found 349.0582.
0
0
0
5.2.19. 30F-Adenosine 50-diphospho-500-ribofuranoside-100-
nicotinamide (30F-NAD+) (29). Compound 29 was
obtained as described above for compound 14 starting
from 61.0 mg (0.175 mmol, free acid) of 23 to give a
white, fluffy solid (0.055mmol, 31%, quantified by UV);

486
C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
1
l
MAX=259 nm (H2O, pH=7); H NMR (399.8MHz;
9. Okamoto, H. Mol. Cell. Biochem. 1999, 193, 115.
10. Lee, H. C.; Munshi, C.; Graeff, R. Mol. Cell. Biochem.
1999, 193, 89.
11. De Flora, A.; Guida, L.; Franco, L.; Zocchi, E. Inter. J.
Biochem. Cell Biol. 1997, 29, 1149.
12. Graeff, R.; Munshi, C.; Aarhus, R.; Johns, M.; Lee, H. C.
J. Biol. Chem. 2001, 276, 12169.
13. Matsumura, N.; Tanuma, S. Biochem. Biophys. Res.
Commun. 1998, 253, 246.
D2O) dH 4.20–4.75 (9H, m, ribofuranoside protons),
5.05–5.20 (1H, m, H-30), 5.88 (1H, m, H-10), 5.90 (1H,
m, H-100), 7.55 (1H, s, H8), 8.00 (1H, ‘t’ (dd), J=6.4,
HNIC-5), 8.27 (1H, br s, H2), 8.63 (1H, d, J=7.9, HNIC
-
4), 8.91 (1H, d, J=5.3, HNIC-6), 9.12 (1H, s, HNIC-2);
19F NMR (376.2 MHz; D2O) dF À198.97 (ddd, JF-3
=50.9, JF-4=24.8, JF-2=24.8); 31P NMR (161.9 MHz;
D2O) dP –10.57! À10.02 (m, P-O-P).
14. Cakir-Kiefer, C.; Muller-Steffner, H.; Schuber, F. Bio-
chem. J. 2000, 349, 203.
5.2.20. Cyclic 30F-adenosine 50-diphosphate ribose (30).
The conditions developed for the preparation of the
cADPR analogues were also employed for the synthesis of
cyclic 30-deoxy-30-fluoro-ADP ribose. Incubation of the
NAD+ analogue 29 (5 mmol, quantified by UV) in 25 mM
HEPES buffer (10 mL) at pH 6.8 in the presence of Aply-
sia cyclase (20 mL of a 25 mM HEPES solution of enzyme,
ꢁ2 units) was carried out at rt and the formation of
product was followed by analytical anion exchange
HPLC. HPLC revealed the formation of a new peak
(RT=4.4 min) along with the nicotinamide peak (RT=2.2
min). Purification was achieved by ion-exchange chroma-
tography on BioRad AG-MP1 resin using a concave
gradient of 150 mM TFA (0 min: 100% MQ!20 min:
100% MQ!40 min: 96% MQ!60 min: 84% MQ!75
min: 68% MQ!105 min: 0% MQ!120 min: 0% MQ).
The title compound eluted at 15–17% TFA. Fractions
containing >95% of the product (as determined by
HPLC) were pooled. Concentration and co-evaporation
(3Â) from MQ water, followed by lyophilisation yielded
the product (free acid) as a white, fluffy solid (2.3mmol,
15. Tohgo, A.; Munakata, H.; Takasawa, S.; Nata, K.;
Akiyama, T.; Hayashi, N.; Okamoto, H. J. Biol. Chem.
1997, 272, 3879.
16. Lee, H. C.; Aarhus, R. Cell. Regul. 1991, 2, 203.
17. Zhang, F. J.; Gu, Q. M.; Sih, C. J. Bioorg. Med. Chem.
1999, 7, 653.
18. Lee, H. C.; Aahrus, R.; Levitt, D. Nat. Struct. Biol. 1994,
1, 143.
19. Prasad, G. S.; Levitt, D. G.; Lee, H. C.; Stout, C. D.
Proteins 1996, 24, 138.
20. Prasad, G. S.; McRee, R. E.; Stura, E. A.; Levitt, D. G.;
Lee, H. C.; Stout, C. D. Nat. Struct. Biol. 1996, 3, 957.
21. Munshi, C.; Aarhus, R.; Graeff, R.; Lee, H. C. J. Biol.
Chem. 2000, 275, 21566.
22. Sauve, A. A.; Deng, H. T.; Angeletti, R. H.; Schramm,
V. L. J. Am. Chem. Soc. 2000, 122, 7855.
23. Sauve, A. A.; Munshi, C.; Aarhus, R.; Lee, H. C.;
Schramm, V. L. Biochemistry 1998, 37, 13239.
24. Migaud, M. E.; Pederick, R. L.; Bailey, V. C.; Potter,
B. V. L. Biochemistry 1999, 38, 9105.
25. Wall, K. A.; Klis, M.; Kormet, J.; Coyle, D.; Ame, J. C.;
Jacobson, M. K.; Slama, J. T. Biochem. J. 1998, 335, 631.
26. Ashamu, G. A.; Galione, A.; Potter, B. V. L. J. Chem.
Soc. Chem. Commun. 1995, 18, 1929.
1
46%, quantified by UV); H NMR (399.8 MHz; D2O)
27. Sethi, J. K.; Empson, R. M.; Bailey, V. C.; Potter, B. V. L.;
Galione, A. J. Biol. Chem. 1997, 272, 16358.
28. Bailey, V. C.; Sauve, A. A.; Galione, A.; Potter, B. V. L.
J. Chem. Soc. Chem. Commun. 1997, 7, 695.
29. Bailey, V. C.; Sethi, J. K.; Fortt, S. M.; Potter, B. V. L.;
Galione, A. Chem. Biol. 1997, 4, 51.
dH 3.30–4.75 (9H, m, fluororibo- and ribofuranoside pro-
tons), 5.36 (1H, m, H-30), 5.74 (1H, m, H-20 or H-40), 6.13
(1H, d, J 7.9, H-10), 6.16 (1H, d, J 3.5, H-100), 8.41 (1H, s,
H8), 9.02 (1H, s, H2); 19F NMR (376.2MHz; D2O) dF
À200.99 (ddd, JF-3=53.7, JF-4=26.1, JF-2=27.6); 31P
NMR (161.9 MHz; D2O) dP À9.00! À8.50 (m, P-O-P).
30. Wong, L.; Aarhus, R.; Lee, H. C. Biochim. Biophys. Acta
1999, 1472, 555.
31. Vu, C. Q.; Lu, P. J.; Chen, C. S.; Sih, C. J. Biol. Chem.
1996, 271, 4747.
Acknowledgements
32. Bailey, V. C.; Fortt, S. M.; Summerhill, R. J.; Potter,
B. V. L.; Galione, A. FEBS Lett. 1996, 379, 227.
33. Ashamu, G. A.; Sethi, J. K.; Galione, A.; Potter, B. V. L.
Biochemistry 1997, 36, 9505.
We acknowledge Project Grant support from the Well-
come Trust (No 055709)
34. Ashamu, G. A.; Galione, A.; Potter, B. V. L. J. Chem.
Soc. Chem. Commun. 1995, 13, 1359.
35. Marwood, R. D.; Shuto, S.; Jenkins, D. J.; Potter, B. V. L.
J. Chem. Soc. Chem. Commun. 2000, 219.
References and notes
1. Guse, A. H. J. Mol. Med. 2000, 78, 26.
2. Guse, A. H. Cell. Signal. 1999, 11, 309.
3. Prakash, Y. S.; Kannan, M. S.; Walseth, T. F.; Sieck,
G. C. Am. J. Physiol. 1998, 274, C1653.
36. Hanessian, S.; Liak, T. J. Synthesis 1981, 396.
37. Vorbruggen, H.; Krolikiewicz, K.; Bennua, B. Chem. Ber.
1981, 114, 1234.
38. Yoshikawa, M.; Kato, T.; Takenishi, T. Tetrahedron Lett.
1967, 50, 5068.
39. Hughes, N. A.; Kenner, G. W.; Todd, A. R. J. Chem. Soc.
1957, 735, 3733.
40. Fawcett, C. P.; Agre, N. O. J. Biol. Chem. 1961, 2237,
1709.
41. Lee, W. J. J. Heterocycl. Chem. 1971, 8, 179.
42. Colman, R. F. Methods Enzymol. 1997, 280, 186.
43. Gosselin, G.; Puech, F.; Genudellac, C.; Imbach, J. L.
Carbohydr. Res. 1993, 249, 1.
44. Mikhailopulo, I. A.; Poopeiko, N. E.; Pricota, T. I.;
Sivets, G. G.; Kvasyuk, E. I.; Balzarini, J.; DeClercq, E.
J. Med. Chem. 1991, 34, 2195.
4. Higashida, H.; Hashii, M.; Yokoyama, S.; Hoshi, N.;
Asai, K.; Kato, T. J. Neurochem. 2001, 76, 321.
5. Cancela, J. M.; Gerasimenko, O. V.; Gerasimenko, J. V.;
Tepikin, A. V.; Petersen, O. H. EMBO J. 2000, 19, 2549.
6. da Silva, C. P.; Guse, A. H. Biochim. Biophys. Acta 2000,
1498, 122.
7. Bootman, M. D.; Collins, T. J.; Peppiatt, C. M.; Pro-
thero, L. S.; MacKenzie, L.; De Smet, P.; Travers, M.;
Tovey, S. C.; Seo, J. T.; Berridge, M. J.; Ciccolini, F.;
Lipp, P. Semin. Cell Devel. Biol. 2001, 12, 3.
8. Franco, L.; Guida, L.; Bruzzone, S.; Zocchi, E.; Usai, C.;
De Flora, A. FASEB J. 1998, 14, 1507.

C. J. W. Mort et al. / Bioorg. Med. Chem. 12 (2004) 475–487
487
45. Ashamu, G. A. PhD Thesis, University of Bath, 1997.
46. Ludemann, H. D. Biophys. Struct. Mech. 1975, 1, 121.
47. Akhrem, A. A. Org. Mag. Res. 1979, 12, 247.
48. Ekiel, I. Acta Biochim. Pol. 1979, 26, 435.
49. Pearlman, D. A.; Kim, S. H. J. Biomol. Struct. Dyn. 1985,
3, 99.
50. Wada, T.; Inageda, K.; Aritomo, K.; Nishina, H.; Taka-
hashi, K.; Katada, T.; Sekine, M. Nucleosides Nucleotides
1995, 14, 1301.
51. Smith, P. E.; Tanner, J. J. J. Mol. Recog. 2000, 13, 27.
52. Thibaudeau, C.; Plavec, J.; Chattopadhyaya, J. J. Org.
Chem. 1998, 63, 4967.