New potent and selective A1 adenosine receptor agonists

Article abstract of DOI:10.1016/j.bmc.2004.07.004

Thiirane analogs of ENAdo have been synthesised and found to be extremely potent and selective A₁ adenosine receptor agonists. Thiirane analogs of ENAdo have been synthesised and found to be extremely potent and selective A₁ adenosin

Full text of DOI:10.1016/j.bmc.2004.07.004

Bioorganic & Medicinal Chemistry 12(2004) 4877–4884
New potent and selective A1 adenosine receptor agonists
Sally A. Hutchinson,a, Stephen P. Baker,c Joel Lindend and Peter J. Scammellsa,b,
*
aSchool of Biological and Chemical Sciences, Deakin University, Geelong, VIC 3217, Australia
bDepartment of Medicinal Chemistry, Victorian College of Pharmacy, Monash University, 381 Royal Parade, Parkville,
VIC 3052, Australia
cDepartment of Pharmacology and Therapeutics, University of Florida, College of Medicine, Box 100267, Gainesville, FL 32610, USA
dDepartments of Medicine (Cardiology) and Molecular Physiology and Biological Physics, University of Virginia, Charlottesville,
VA22908, USA
Received 1 June 2004; revised 2 July 2004; accepted 2 July 2004
Available online 4 August 2004
Abstract—Thiirane analogs of ENAdo have been synthesised and found to be extremely potent and selective A1 adenosine receptor
agonists.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
attenuation of necrotic injury that is observed when
the myocardium has been exposed to brief periods of is-
chaemia before a prolonged ischaemic period.4 Adeno-
sine and A1AR agonists, administered prior to an
ischaemic insult, can reduce infarct size, arrhythmia
and improve post-ischaemic cardiac function. Precise
mechanisms linking A1AR stimulation to cardioprotec-
tion are not clear. A1ARs are expressed in cardiac myo-
cytes and their stimulation is thought to produce
beneficial cellular metabolic effects, attenuating ATP
depletion and reducing cellular acidosis, thereby pro-
longing cell viability during ischaemia.4 These discover-
ies have stimulated interest in the development of agents
capable of simulating ischaemic preconditioning and
affording myocardial protection. A1AR agonists may
also act as cardioprotective agents due to their ability
to inhibit lipolysis and reduce plasma fatty acid and tri-
glyceride levels.4
Adenosine is an effective agent for the treatment of par-
oxysmal supraventricular tachycardias (PSVT).1,2 It acts
via extracellular receptors (A1, A2A, A2B, A3) to mediate
a range of physiological effects. Cardiovascular effects
include a slowing of heart rate (negative chronotropic
effect), prolonged of AV nodal conduction (negative
dromotropic effect), depressed atrial contractility (nega-
tive inotropic effect), anti-b-adrenergic effects and in-
creased coronary artery blood flow. The A1 adenosine
receptors (A1ARs) mediate adenosineÕs negative dromo-
tropic effects, which are responsible for termination of
supraventricular tachycardias. The use of adenosine
for the therapy of such arrhythmias has some draw-
backs, which result from its short half life (e.g., recur-
rence of arrhythmias) and its lack of receptor subtype
selectivity (e.g., vasodilatory effects). As a result, it has
been postulated that longer acting adenosine agonists,
which are selective for the A1AR may be
advantageous.2,3
In the search for potent and selective A1AR agonists we
reported the design and synthesis of ENAdo (1).5,6
Other adenosine analogs with oxygenated cycloalkyl
N6-substituents [e.g., CVT-510 (2) and GR79236 (3)]
have also been found to confer high selectivity for the
A1AR.7–9 The more potent S-endo isomer of ENAdo
(1) was found to degrade on silica to form polar break-
down products.6 This process was believed to involve
epoxide opening and cyclisation on N1. A similar proc-
ess was observed for 1,3-dipropyl-8-[2-(5,6-epoxy)nor-
bornyl]xanthine (ENX). In this case, the epoxide was
involved in an intramolecular cyclisation with N9 under
Adenosine is also involved in an endogenous cardiopro-
tective phenomenon known as ischaemic precondition-
ing (IPC). Ischaemic preconditioning involves the
Keywords: Adenosine; Agonist; A1 adenosine receptor (A1AR).
*
Corresponding author. Tel.: +61-3-9903-9542; fax: +61-3-9903-
9582; e-mail: peter.scammells@vcp.monash.edu.au
Present address: Division of Textile and Fibre Technology, CSIRO,
Belmont, VIC 3216, Australia.
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.07.004

4878
S. A. Hutchinson et al. / Bioorg. Med. Chem. 12 (2004) 4877–4884
O
NH
N
NH
N
NH
N
O
HO
N
N
N
N
N
N
N
N
N
OH
OH
OH
O
O
O
HO OH
HO OH
HO OH
3 GR79236
1 ENAdo
2 CVT-510
acidic conditions.10 In contrast, compounds of this type
proved to be surprisingly resistant to nucleophilic ring
opening, presumably due to the norbornyl hindering
the approach of the nucleophile.11 In this study we tar-
geted thiirane analogs of ENAdo (1), which were hoped
to be more resistant to acid-catalysed ring opening and
cyclisation. Thiiranes are generally less reactive towards
electrophilic reagents than the corresponding epoxides
as a result of the lower polarity of the C–S bond (relative
to C–O) and lower electron density of the sulfur.12,13
Additionally, epithionorbornanes have been reported
to be stable to column chromatography and no decom-
position or isomerisation was observed at 130ꢁC
(GPLC).14
norborn-5-en-2-yl)adenosine15 using the approach out-
lined in Scheme 1. The thiirane moiety was introduced
using methodology for the conversion of alkenes to
episulfides developed by Bombala and Ley.16 In this
approach, the alkene is treated with phthalimide-N-
sulfenyl chloride (10) and the resultant addition product
is reduced with lithium aluminium hydride to afford the
corresponding episulfide. Phthalimide-N-sulfenyl chlo-
ride (10) was prepared by reacting potassium phthali-
mide (8) with sulfur monochloride (S2Cl2) at room
temperature to produce N,N-dithiobis(phthalimide)
(9). Bubbling chlorine gas through the solution of this
species (9) in chloroform gave the desired sulfenyl chlo-
ride (10). In our hands, this reaction required more for-
cing conditions, a longer reaction time and proceeded in
lower yield than was previously reported.16 Once in
hand, phthalimide-N-sulfenyl chloride (10) was reacted
with the alkene moiety of N6-(exo-5,6-epithionorborn-
2-yl)adenosine (4) to afford the expected addition prod-
uct (5) in 74% yield. Reaction of this species with
LiAlH4 effected the desired cyclisation and afforded
2. Results and discussion
The initial target compound, N6-(exo-5,6-epithionor-
born-2-yl)adenosine (7), was prepared from N6-(exo-
O
N
S
S
O
2R/2S
2R/2S
Cl
NH
NH
N
NH
N
N
N
N
N
N
(i)
(ii)
N
N
N
N
N
TBSO
TBSO
RO
O
O
O
TBSO OTBS
TBSO OTBS
RO OR
6 R = TBS
7 R = H
4
5
(iii)
O
O
O
O
S2Cl2
Cl2
NK
N S S N
N SCl
O
O
O
O
8
9
10
Scheme 1. Reagents: (i) Phth-SCl, NEt3, MeOH; (ii) LiAlH4, THF; (iii) NBu4F, Et2O.

S. A. Hutchinson et al. / Bioorg. Med. Chem. 12 (2004) 4877–4884
4879
HS
(i)-(ii)
S
see
text
2S
HN
OBn
HN
(iii)
OBn
HN
OBn
11
12
13
O
O
O
2S
NH
Cl
S
N
N
N
N
S
N
N
X
N
X
RO
N
NH2
14
RO
O
O
NiPr2Et
RO OR
RO OR
15a X=R=H
15b X=F, R = TBS
16a X=R=H
16b X=F, R=TBS
16c R = H
(iv)
Scheme 2. Reagents: (i) Phth-SCl, NEt3, MeOH; (ii) LiAlH4, THF; (iii) KOH, MeOH/H2O; (iv) NH4F, MeOH.
the thiirane (6) in reasonable yield (54%). Deprotection
of the silyl groups with tetrabutylammonium fluoride
(TBAF) in tetrahydrofuran afforded the thiirane ana-
logue of ENAdo, N6-(exo-5,6-epithionorborn-2-yl)
adenosine (7) in good yield (84%).
useful as the by-products of hydrogenation are carbon
dioxide and toluene, which are readily separated. There
have been a variety of H-donors, which have been re-
ported in the deprotection of CBZ groups including
H2,18 cyclohexene,19 formic acid,20 cis-decalin,21 1,4-cy-
clohexadiene22 and ammonia.23 Catalytic hydrogenoly-
sis with Pd/C and H2 gas failed to remove the CBZ
moiety. This result was not unexpected as the deprotec-
tion of sulfur containing amino acids can fail due to cat-
alyst poisoning.21 It had been reported that cleavage of
CBZ groups from peptides that possess an amino acid
containing sulfur such as methionine or cysteine, could
be effected by catalytic transfer hydrogenation with pal-
ladium black (Pd0) as catalyst and with formic acid as the
H-donor.24 Unfortunately when the thiirane (12) was re-
acted with Pd0 no CBZ cleavage was observed, and the
thiirane (12) was found to have undergone rearrange-
ment. The mechanism of rearrangement presumably in-
volved acid-catalysed ring opening of the thiirane and
subsequent rearrangement to 13. Likewise, attempted
deprotection of the CBZ group with trimethylsilyl iodide
in acetonitrile resulted in rearrangement to 13.
In order to facilitate the synthesis of more complex pu-
rine and ribose modified adenosine with optically pure
N6-(5,6-epithionorborn2-y-l) groups, a more convergent
synthesis was developed. In the approach the 5,6-epith-
ionorborn-2-yl and purine riboside units were assembled
separately and coupled late in the synthesis. More spe-
cifically,
2S-endo-amino-5,6-exo-epithionorbornane
(14) was prepared and coupled with the appropriately
substituted 6-chloropurine riboside prior to deprotec-
tion of the ribose hydroxyl groups (Scheme 2). 2S-
endo-Amino-5,6-exo-epithionorbornane
(14),
was
synthesised from 2S-endo-aminonorborn-5-ene using
a sequence involving amine protection, episulfide forma-
tion and deprotection. Benzyloxycarbonyl (CBZ) was
employed as the protecting group for the amine. This
group was chosen as it was expected to be stable to
the subsequent reaction conditions (electrophilic addi-
tion of the sulfenyl chloride and LiAlH4 reduction)
and able to be cleaved without disrupting the thiirane
(e.g., by hydrogenolysis). The CBZ protecting group
was incorporated in good yield (80%) by reaction of
the aminonorbornene with benzylchloroformate and
2M NaOH in CH2Cl2.17 Reaction of protected 2S-
endo-aminonorborn-5-ene (11) with phthalimide-N-sul-
fenyl chloride (10) afforded the expected adduct, 2S-
endo-benzyloxycarbonylamino-5-endo-chloro-6-exo-(N-
phthali- midothio)norbornane in 57% yield. Reduction
with LiAlH4 produced the protected 2-amino-5,6-epith-
ionorbornane (12) in good yield (79%).
It had been reported that cleavage of CBZ groups could
be achieved by reaction with 40% aqueous potassium
hydroxide (KOH) in MeOH at reflux.18 The optimum
ratio of KOH solution to methanol was found to be 1:1
for the best solubility of the starting material (12). A rea-
sonable yield (ꢀ72%) of a mixture of the deprotected
amine, 2S-endo-amino-5,6-exo-epithionorbornane (14)
and the by-product of the reaction, benzyl alcohol was
obtained. The ratio of the desired product (14) to benzyl
alcohol in the extract was 1:1.5, the recovery of less of the
amine (14) than the benzyl alcohol presumably due to the
relatively high solubility of the amine (14) in the water
layer. This amine was used without further purification.
Removal of the CBZ group proved more problematic.
Catalytic hydrogenolysis is the most common method
for cleavage of CBZ groups. This method is particularly
Coupling of the crude amine (14) with 6-chloropurine ri-
boside (15a) was completed using HunigÕs base as acid
¨

4880
S. A. Hutchinson et al. / Bioorg. Med. Chem. 12 (2004) 4877–4884
scavenger and t-BuOH as the solvent.15 Purification by
column chromatography afforded N6-(2S-endo-5,6-
epithionorborn-2-yl)adenosine (16a).
Table 3. Affinity of 16c (nM) at human adenosine receptors
Compound
A1
A2A
A2B
A3
16c
0.15 0.09
812 163
>100,000
175 10
The introduction of a 2-fluoro group on the purine ring
had been observed to increase selectivity and efficacy for
the A1AR in previous studies.15 The 2-fluoro purine
(15b) was synthesised as previously described, and cou-
pled to the crude amine (14). Deprotection of the TBS
groups was achieved with NH4F and gave the desired
compound for evaluation, 2-fluoro-N6-(2S-endo-5,6-
epithionorborn-2-yl)adenosine (16c).
compounds was less potent than CGS21680 for stimu-
lating cAMP accumulation and each produced less than
the maximal response. This indicates that these com-
pounds are partial agonists at the A2AAR.
These data show that compounds 7, 16a and 16c have
high affinity and potency for the A1AR while acting as
full agonists at this receptor. As expected, the endo-iso-
mer (16a) has a higher affinity for the A1AR than the
exo-isomer (7). Although 2-fluoro substitution of the
endo-isomer did not change the affinity for the A1AR,
its potency for inhibiting cAMP accumulation increased
slightly. In contrast, these compounds show relatively
low affinity and potency at the A2AAR and are partial
agonists at this receptor. Compound 16c has the highest
selectivity for the A1AR as compared to the A2AAR
based upon affinity (1482-fold) or potency for the cellu-
lar response (1167-fold). Overall, 16c has the favourable
characteristics of high affinity, potency and selectivity
for the A1AR suggesting that additional derivatives
may optimise selective agonist properties for this recep-
tor subtype.
Compounds 7, 16a and 16c were tested for affinity and
agonist activity at the A1AR in DDT cells (Table 1).
Compared to N6-cyclopentyladenosine (CPA), com-
pound 7 has a slightly lower affinity (1.5-fold) whereas
16a and 16b have six and ninefold higher affinities,
respectively. These derivatives were more potent (3–12-
fold) than CPA for inhibiting (ꢁ)isoproterenol-stimu-
lated cAMP accumulation and all produced the same
maximal response as CPA. At the A2AAR in PC-12
cells, compounds 7, 16a and 16c have low micromolar
affinities that were 13.6–14.8-fold lower than for the
standard CGS21680 (Table 2). Similarly, each of the
Table 1. Ki, IC50 and intrinsic activity (IA) (nM) for adenosine
derivatives at the A1AR from DDT1 MF-2cells
Based upon the data for compound 16c using DDT and
PC-12cells, this compound was further tested for affin-
ity at the human recombinant A1, A2A, A2B and A3ARÕs
(Table 3). This compound had the highest affinity (sub-
nanomolar) at the A1AR and high selectivity for this
receptor subtype. Its affinity for the human A1AR was
18-fold higher than for the A1AR in DDT cells (ham-
ster) highlighting the species difference in this receptor.
Compound
Ki
IC50
2.4 1.2 (16)
IA
CPA
7
18 3 (7)
27 3 (6)
3.6 1 (5)
2.7 1 (5)
1.00
0.89 0.02 (9)
1.2 0.7 (4)
0.43 0.19 (9)
1.00 0.03 (9)
1.04 0.06 (4)
1.03 0.03 (9)
16a
16c
The affinity (Ki values) was calculated from the concentration of the
compounds that inhibited [3H]CPX binding to the A1AR by 50% in
DDT cell membranes. The potency (IC50 values) is the concentration
of compounds that inhibited (ꢁ)isoproterenol (1lM) stimulated
cAMP accumulation by 50% in DDT cells. The intrinsic activity is the
maximal inhibition of (ꢁ)isoproterenol-stimulated cAMP accumula-
tion as compared to the maximum inhibition by CPA, which was set as
1.00. Numbers in parentheses are the N. The maximal CPA inhibition
of (ꢁ)isoproterenol-stimulated cAMP accumulation was 85 2%.
3. Conclusions
In conclusion, an efficient convergent approach of the
synthesis of N6-(5,6-epithionorborn-2-yl)adenosines
has been developed. The N6-(5,6-epithionorborn-2-yl)
substituent has been found to confer high A1AR po-
tency and selectivity. N6-(2S-endo-5,6-Epithionorborn-
2-yl)-2-fluoroadenosine (16c) was found to be one of
the most potent and selective A1AR agonists reported
to date.
Table 2. Ki, EC50 and intrinsic activity (IA) (nM) for adenosine
derivatives at the A2AAR from PC-12cells
4. Experimental
Compound
Ki
EC50
IA
1H and 13C NMR spectra were obtained with a Varian
300MHz Unity Plus spectrometer at 288.976 and
75.437MHz, respectively; or with a JEOL 400MHz
Eclipse Plus FT NMR spectrometer at 399.784 and
100.534MHz, respectively. High resolution ES MS data
was collected using a Bruker BioApex 47e FTMS with
an Analytica Electrospray Source. Melting points were
determined on a Reichert Hot-Stage microscope and
are uncorrected. Electrospray mass spectra were ob-
tained on a Micromass Platform instrument with Mass-
lynx software. Thin layer chromatography was carried
out using 0.2mm plates using Merck silica gel 60 F254.
CGS21680
7
284 56 (6)
25 4 (15)
541 99 (6)
636 76 (6)
502 502 (8)
1.00
0.77 0.05 (6)
*
3866 142(4)
4207 857 (5)
4003 1331 (3)
16a
16c
0.76 0.08 (6)**
0.76 0.06 (8) *
The Ki values were calculated from the concentration of the com-
pounds that inhibited [3H]ZM241385 binding to the A2AAR by 50% in
PC-12cell membranes. The EC
values are the concentration of
50
compounds that elicited 50% of the maximal cAMP accumulated in
PC-12cells. The intrinsic activity is the maximal stimulation of cAMP
accumulation as compared to the maximum produced by CGS 21680,
which was set at 1.00. Numbers in parentheses are the N. The maximal
CGS 21680-stimulated cAMP accumulation was 11.5 1.8nmol/mg
protein/10min. *p < 0.02and **p < 0.007 compared to CGS 21680.

S. A. Hutchinson et al. / Bioorg. Med. Chem. 12 (2004) 4877–4884
4881
Column chromatography was achieved using Merck sil-
ica gel 60, particle size 0.063–0.200mm, 70–230mesh.
H-100/H-400), 2.53 (s, 1H, H-100/H-400), 2.82 (dd, 1H, H-500,
H-600), 3.70 (d, 1H, H-5a0/H-5b0), 3.88 (d, 1H, H-5a0/H-
5b0), 4.04–4.19 (m, 1H, H-200), 4.26 (s, 1H, H-40), 4.43
4.1. N6-exo-(endo-5-Chloro-exo-6-(N-phthalimido-
thio)norborn-2-yl)-20,30,50-tris-O-TBS-adenosine (5)
(d, 1H, H-30), 4.92(br s, 1H, H-2 ), 5.84 (d, 1H, H-10),
0
6.40 (br s, 1H, NH), 7.91 (br s, 1H, H-2/H-8), 8.20 (br s,
1H, H-2/H-8). 13C NMR (CDCl3): d 25.6, 33.7, 35.1,
37.7, 38.1, 38.5, 54.9, 63.4, 72.6, 75.5, 88.2, 91.2, 121.3,
141.6, 149.1, 153.2, 155.1. HR-ES MS calcd for
C17H22N5O4S+ (M+1) 392.1393, found 392.1388.
Phthalimide-N-sulfenyl chloride (10, 77mg, 0.36mmol)
was dissolved in CHCl3 (1mL) and added dropwise to
4 (92mg, 0.13mmol) in CDCl3 (7mL). The reaction
was stirred for 100h then evaporated in vacuo and chro-
matographed (petroleum ether/EtOAc, 4:1) to afford 5
(88mg, 74%) as an oil. 1H NMR (CDCl3): d ꢁ0.30,
ꢁ0.27, ꢁ0.05, ꢁ0.03, 0.14, 0.15 (6·s, 6·3H, SiCH3),
0.76, 0.78, 0.96 (3·s, 3·3H, But), 0.95–0.99 (m, 1H,
H-7a00), 1.11–1.13 (m, 1H, H-3x00/H-3n00), 1.55–1.67 (m,
4.4. NN0-Dithiobis(phthalimide) (9)16
To a stirred suspension of potassium phthalimide (8,
5.11g, 28mmol) in CH2Cl2 (25mL) at 0ꢁC was added
a solution of S2Cl2 (1.1mL, 14mmol) in CH2Cl2
(5mL). The reaction was allowed to warm to rt then
heated briefly to reflux. After 3h the reaction mixture
was filtered, the filtrate was evaporated in vacuo and
washed with petroleum ether (2·10mL) and water
(2·10mL) then dried in vacuo. Recrystallisation from
CHCl3 afforded the title compound 9 product (4.78g,
49%) as translucent needle like crystals (mp 232–
00
1H, H-3x00/H-3n00), 1.81–1.92(m, 1H, H-7s ), 2.43 (br
s, 1H, H-400/H-100), 3.43–3.47 (m, 1H, H-600), 3.83 (dd,
1H, H-5a0/H-5b0), 4.04 (dd, 1H, H-5a0/H-5b0), 4.15 (t,
1H, H-40), 4.23–4.27 (m, 2H, H-500), 4.39 (t, 1H, H-30),
4.83 (t, 1H, H-20), 4.86 (br s, 1H, H-200), 5.99 (d, 1H,
H-10), 7.77–8.02(m, 6H, ArH, H-2, H-8). 13C NMR
(CDCl3): d ꢁ5.5, ꢁ5.2, ꢁ5.1, ꢁ4.7, ꢁ4.2, ꢁ4.1, 17.8,
18.0, 18.5, 26.2, 26.4, 26.6, 30.9, 32.2, 33.7, 45.2, 60.0,
64.0, 64.0, 73.9, 77.3, 87.5, 89.3, 121.2, 124.8, 128.6,
133.3, 134.3, 136.1, 140.8, 155.2, 170.8.
1
234ꢁC) (lit.16 229–230ꢁC). H NMR (CDCl3): d 8.01
(s, ArH). 13C NMR (CD3COCD3): d 124.9, 136.2, 206.2.
4.5. Phthalimide-N-sulfenyl chloride (10)16
4.2. N6-exo-(exo-5,6-Epithionorborn-2-yl)-20,30,50-tris-O-
TBS-adenosine (6)
Compound 9 (3.391g, 9.5mmol) was suspended in
CHCl3 (40mL). The reaction was heated to 55ꢁC and
Cl2 gas was bubbled through for a total of 13.5h over
two days. Nitrogen was bubbled through the solution
and the reaction was evaporated in vacuo to afford the
product 10 (2.75g, 100%) as a yellow solid (mp 115–
118ꢁC) (lit.16 mp 115–117ꢁC). 1H NMR (CDCl3): d
7.75–7.93 (m, 2H, ArH), 8.02–8.06 (m, 2H, ArH). 13C
NMR (CDCl3): d 124.9, 135.7, 165.5.
Compound 5 (230mg, 0.25mmol) dissolved in THF
(4mL) was cooled to ꢁ78ꢁC and LiAlH4 in Et2O
(1M, 0.3mL) was added dropwise. The reaction was al-
lowed to stir at this temperature for 15min then gradu-
ally warmed to rt. The reaction was quenched with
water (15mL) then extracted with Et2O (3·15mL).
The organic layer was dried (MgSO4) and evaporated
in vacuo. Purification by column chromatography
(petroleum ether/EtOAc, 4:1) to afford 6 (99mg, 54%)
4.6. 2S-endo-Benzyloxycarbonylaminonorborn-5-ene (11)
1
as an oil. H NMR (CDCl3): d ꢁ0.24, ꢁ0.21, ꢁ0.05,
ꢁ0.04, 0.09, 0.13 (6·s, 6·3H, SiCH3), 0.80, 0.92, 0.95
(3·s, 3·3H, But), 0.91–0.96 (m, 1H, H-7a00), 1.25–1.36
2S-endo-Aminonorborn-5-ene hydrochloride (0.998g,
6.85mmol) in CH2Cl2 (30mL) and 2M NaOH
(10.3mL) were cooled to 0ꢁC. Benzyl chloroformate
(1.0mL, 7.0mmol) in CH2Cl2 (10mL) was added over
10min. The reaction was stirred for 3h at room temper-
ature then the organic layer was separated, washed suc-
cessively with dilute HCl (20mL) and water (30mL),
dried (MgSO4) and evaporated in vacuo to afford a vis-
cous oil. Column chromatography (petroleum ether/Et-
OAc, 10:1) afforded the title compound 11 (1.483g, 89%)
00
(m, 1H, H-3x00/H-3n00), 1.69–1.72(m, 1H, H-3x /H-
3n00), 2.02–2.07 (m, 1H, H-7s00), 2.56 (br s, 1H, H-400/H-
100), 2.62–2.70 (m, 2H, H-500/H-600), 2.93 (br s, 1H,
H-400/H-100), 3.77 (dd, 1H, H-5a0/H-5b0), 4.04–4.14 (m,
2H, H-200, H-5a0/H-5b0), 4.28–4.33 (m, 1H, H-40), 4.64–
4.70 (m, 1H, H-30), 4.64–4.70 (t, 1H, H-20), 5.75 (br s,
1H, NH), 6.00–6.03 (m, 1H, H-10), 8.12(s, 1H, H-2/H-
8), 8.37 (s, 1H, H-2/H-8). 13C NMR (CDCl3): d ꢁ5.4,
ꢁ5.1, ꢁ5.0, ꢁ4.8, ꢁ4.7, ꢁ4.4, 17.8, 18.0, 18.5, 22.7,
25.6, 25.8, 26.0, 29.7, 32.2, 32.6, 44.0, 47.9, 52.6, 62.5,
72.0, 75.9, 85.5, 88.2, 120.0, 132.9, 139.0, 152.8, 153.6.
1
as a white solid (mp 41–42ꢁC). H NMR (CDCl3): d
0.66–0.72(m, 1H, H-3x), 1.32(dd, 1H, H-7s), 1.46
(dd, 1H, H-7a), 2.16–2.24 (m, 1H, H-3n), 2.82 (br s,
1H, H-4), 3.04 (br s, 1H, H-1), 4.29–4.37 (m, 1H, H-
2), 4.50 (br s, 1H, NH), 5.07 (s, 2H, CH2), 5.98–6.01
(m, 1H, H-5), 6.32–6.35 (m, 1H, H-6), 7.26–7.34 (m,
5H, ArH). 13C NMR (CDCl3): d 34.8, 42.0, 45.7, 48.0,
50.4, 65.8, 127.4, 127.5, 127.9, 131.0, 136.2, 139.1, 155.5.
4.3. N6-(exo-5,6-Epithionorborn-2-yl)adenosine (7)
The thiirane 6 (95mg, 0.13mmol) was stirred in for 24h in
a 1M solution of TBAF in THF (0.6mL). The reaction
mixture was evaporated in vacuo and purification by re-
verse phase column chromatography (MeOH/H2O/
NH3, 60:40:0.1) then on silica gel (CHCl3/MeOH, 9:1) af-
forded the title compound 7 (43mg, 84%) as a transparent
4.7. 2S-endo-Benzyloxycarbonylamino-5-endo-chloro-6-
exo-(N-phthalimidothio)norbornane
1
oil. H NMR (CDCl3): d 0.88 (d, 1H, H-7a00), 1.56–1.62
Phthalimide-N-sulfenyl chloride (10, 1.296g, 5.33mmol)
was dissolved in CH2Cl2 (8mL) and added dropwise to a
(m, 2H, H-3n00, H-3x00), 2.01 (d, 1H, H-7s00), 2.52 (s, 1H,

4882
S. A. Hutchinson et al. / Bioorg. Med. Chem. 12 (2004) 4877–4884
1
ratio of 1.5:1, as a pale yellow oil. H NMR (CDCl3):
solution of 11 (1.195g, 4.91mmol) in CH2Cl2 (7mL).
The reaction was stirred for 21h then filtered and chro-
matographed, (petroleum ether/EtOAc, 10:1) to afford
the desired product (1.38g, 57%) as a white foam (mp
d 0.74–0.80 (m, 2H, H-3n/H-7a), 1.54–1.58 (m, 1H, H-
7s), 2.06–2.14 (m, 1H, H-3x), 2.39–2.42 (m, 2H, H-1,
H-4), 2.89 (d, 1H, H-5/H-6), 3.10 (d, 1H, H-5/H-6),
3.44–3.49 (m, 1H, H-2). Benzyl alcohol 1.99 (br s, 1H,
OH), 4.67 (s, 2H, CH2), 7.27–7.37 (m, 5H, ArH). 13C
NMR (CDCl3): d 28.0, 33.3, 37.6, 38.1, 38.7, 43.2,
53.9. Benzyl alcohol 65.1, 126.9, 127.5, 128.5, 141.0.
1
59–62ꢁC). H NMR (CDCl3): d 1.46–1.53 (m, 1H, H-
3n), 1.64 (d, 1H, H-7s), 1.94–2.03 (m, 2H, H-3x/H-7a),
2.48 (br s, 1H, H-4), 2.58 (br s, 1H, H-1), 3.56–3.59
(m, 1H, H-6), 4.07–4.12 (m, 1H, H-2), 4.29–4.33 (m,
1H, H-5), 4.92–5.07 (m, 3H, CH2/NH), 7.17–7.28 (m,
5H, ArH CBZ), 7.80–7.86 (m, 4H, ArH phthalimide).
13C NMR (CDCl3): d 28.7, 35.6, 44.6, 44.9, 51.8, 54.0,
66.1, 66.8, 123.4, 123.9, 128.1, 128.5, 131.7, 134.1,
134.6, 136.1, 155.9, 168.2.
4.11. N6-(2S-endo-5,6-Epithionorborn-2-yl)adenosine
(16a)
A mixture of 6-chloropurine riboside (15a, 47mg,
0.16mmol), 2S-endo-amino-5,6-exo-epithionorbornane
(49mg, ꢀ0.18mmol) and N(i-Pr)2Et (0.3mL, 1.7mmol)
was heated to reflux for 19h. The reaction was evapo-
rated in vacuo and purified by column chromatography
(CHCl3/MeOH, 95:5) afforded the title compound 16a
(17mg, 27%) as a transparent solid (mp 204ꢁC dec).
1H NMR (CDCl3): d 0.92(d, 1H, H-7a 00), 1.11–1.28
(m, 1H, H-3n00), 1.64 (d, 1H, H-7s00), 2.29–2.38 (m, 1H,
H-3x00), 2.51 (d, 1H, H-400), 2.91 (d, 1H, H-100), 2.95 (d,
1H, H-500), 3.03 (d, 1H, H-600), 3.65 (d, 1H, H-5a0/H-
4.8. 2S-endo-Benzyloxycarbonylamino-5,6-exo-epithio-
norbornane (12)
2S-endo-Benzyloxycarbonylamino-5-endo-chloro-6-exo-
(N-phth-alimidothio)norbornane (0.671g, 1.47mmol)
was dissolved in THF (15mL) and added dropwise to
a stirred suspension of LiAlH4 (56mg, 1.48mmol) in
THF (6mL) at ꢁ78ꢁC. The reaction was maintained
at that temperature for 10min then allowed to warm
to ꢁ25ꢁC when it was quenched with water (8mL).
The mixture was extracted with Et2O dried (Na2SO4)
and evaporated in vacuo to afford a pink semi-solid.
Purification with column chromatography (petroleum
ether/EtOAc, 4:1) packed on SiO2 gave 12 (318mg,
79%) as a white solid (mp 66–69ꢁC). 1H NMR (CDCl3):
d 0.79–0.90 (m, 2H, H-3n/H-7a), 1.58 (d, 1H, H-7s),
2.17–2.27 (m, 1H, H-3x), 2.43–2.45(m, 1H, H-4),
2.72–2.75(m, 1H, H-1), 2.84 (d, 1H, H-/5H-6), 2.97
(d, 1H, H-5/H-6), 4.12–4.20 (m, 1H, H-2), 4.92–5.02
(br s, 1H, NH), 5.10 (s, 2H, CH2), 7.29–7.35(m, 5H,
ArH). 13C NMR (CDCl3): d 27.3, 32.8, 35.9, 37.1,
37.9, 41.1, 53.1, 66.7, 128.07, 128.12, 128.5, 136.2, 155.9.
0
5b0), 3.92(d, 1H, H-5a /H-5b0), 4.27 (s, 1H, H-40), 4.35
(d, 1H, H-30), 4.58–4.66 (m, 1H, H-200), 4.87 (t, 1H, H-
20), 5.77 (d, 1H, H-10), 7.84 (br s, 1H, H-2/H-8), 8.22
(br s, 1H, H-2/H-8). HR-ES MS calcd for
C17H22N5O4S+ (M+1) 392.1393, found 392.1382.
4.12. 20,30,50-Tris-O-TBS-2-fluoro-N6-(2S-endo-5,6-
epithionorborn-2-yl)adenosine (16b)
A mixture of 15b (135mg, 0.21mmol), the amine (14,
23mg, ꢀ0.16mmol) and N(i-Pr)2Et (0.2mL, 1.2mmol)
in t-BuOH (7mL) was heated to reflux for 17h. The
reaction mixture was evaporated in vacuo and purified
by column chromatography (hexane/EtOAc, 30:1) af-
forded the title compound 16b (43mg, 36%) as a trans-
parent oil. 1H NMR (CDCl3): d ꢁ0.10, 0.00, 0.10,
0.11, 0.12, 0.13, (6·s, 6·3H, SiMe3), 0.85, 0.93, 0.95
(3·s, 3·3H, But), 0.86–0.91 (m, 1H, H-3n00), 1.12(d,
1H, H-7s00/H-7a00), 1.66 (d, 1H, H-7s00/H-7a00), 2.27–2.38
(m, 1H, H-3x00), 2.51 (br s, 1H, H-100/H-400), 2.89 (d,
1H, H-500/H-600), 2.96 (br s, 1H, H-100/H-400), 3.05 (d,
1H, H-500/H-600), 3.79 (dd, 1H, H-5a0/H-5b0), 4.05 (dd,
1H, H-5a0/H-5b0), 4.12–4.14 (m, 1H, H-40), 4.31 (t,
1H, H-30), 4.64 (t, 1H, H-20), 4.69–4.77 (br m, 1H, H-
200), 5.91 (d, 1H, H-10), 8.17 (br s, 1H, H-8). 13C NMR
(CDCl3): d ꢁ5.5, ꢁ5.4, ꢁ5.0, ꢁ4.8, ꢁ4.7, ꢁ4.4, 17.9,
18.0, 18.5, 25.7, 25.8, 26.0, 33.1, 35.8, 37.3, 38.0, 41.1,
53.1, 62.1, 71.3, 75.6, 85.1, 88.8, 118.1, 138.7, 149.7
(J=18.9Hz), 155.8 (J=20.2Hz), 159.3 (J=211Hz).
4.9. 2S-endo-Benzyloxycarbonylamino-7-thionorbor-5-
ene (13)
The thiirane 12 (35mg, 0.13mmol) was dissolved in
4.4% formic acid in MeOH (10mL) and passed through
a 1.5cm plug of Pd0 at ꢀ1mL/min twice. The resultant
material was evaporated in vacuo and purification by
column chromatography (petroleum ether/EtOAc, 4:1)
afforded 13 as a transparent oil (28mg, 80%). 1H
NMR (CDCl3): d 0.66–0.72(m, 1H, H-3n), 1.45–1.48
(m, 1H, H-7a), 2.11–2.25 (m, 1H, H-3x), 2.83 (br s,
1H, H-1/H-4), 3.04 (br s, 1H, H-1/H-4), 4.28–4.43 (m,
1H, H-2), 5.07 (br s, 2H, CH2), 5.99–6.02(m, 1H, H-
5/H-6), 6.33–6.36 (m, 1H, H-5/H-6), 7.30–7.34 (m, 5H,
ArH). 13C NMR (CDCl3): d 35.7,42.5, 46.2, 48.6, 50.9,
66.5, 128.1, 128.1, 128.5, 131.5, 136.3, 140.0, 155.8.
4.10. 2S-endo-Amino-5,6-exo-epithionorbornane (14)
4.13. N6-(2S-endo-5,6-Epithionorborn-2-yl)-2-fluoroaden-
osine (16c)
A mixture of 2S-endo-benzyloxycarbonylamino-5,6-exo-
epithionorbornane 12 (89mg, 0.32mmol) in 40% KOH
(4mL) and MeOH (4mL) was heated at reflux for 6h.
The MeOH was removed in vacuo and the mixture
was extracted with Et2O (3·25mL), dried (Na2SO4)
and evaporated in vacuo to afford a mixture of benzyl
alcohol and the desired product 14 (49mg, ꢀ72%) in a
Compound 16b (39mg, 0.05mmol) and NH4F (115mg,
3.1mmol) were heated to ꢀ60ꢁC in MeOH (4mL) for
23h. The reaction mixture was evaporated in vacuo
and purified by column chromatography (CHCl3/
MeOH, 95:5) to yield 16c (16mg, 75%) as a white solid
(mp 249–251ꢁC dec). 1H NMR (CD3OD): d 0.97 (d, 1H,

S. A. Hutchinson et al. / Bioorg. Med. Chem. 12 (2004) 4877–4884
4883
H-7a00), 0.86–0.91 (m, 1H, H-3n00), 1.66 (d, 1H, H-7s00),
2.27–2.38 (m, 1H, H-3x00), 2.51 (br s, 1H, H-100/H-400),
2.89 (d, 1H, H-500/H-600), 2.96 (br s, 1H, H-100/H-400),
3.05 (d, 1H, H-500/H-600), 3.79 (dd, 1H, H-5a0/H-5b0),
4.05 (dd, 1H, H-5a0/H-5b0), 4.12–4.14 (m, 1H, H-40),
4.31 (t, 1H, H-30), 4.64 (t, 1H, H-20), 4.69–4.77 (br m,
1H, H-200), 5.91 (d, 1H, H-10), 8.17 (br s, 1H, H-8).
13C NMR (CDCl3): d 28.1, 33.7, 36.0, 37.9, 39.4, 42.4,
54.5, 63.2, 72.2, 75.5, 87.7, 90.8, 119.4, 141.4, 145.8
(J=23.3Hz), 157.7 (J=20.4Hz), 160.5 (J=212Hz). ES
MS calcd 410.1298 (M+H), found 410.1294.
ands; [125I]-ABA (A1AR), [3H]-ZM241385 (A2AAR),
125I]-ABOPX (A2BAR), [125I]-ABA (A3AR), as de-
[
scribed previously.26 The Ki values determined for hu-
man A1 and A3 receptors represent binding to the G
protein coupled high affinity conformational state that
is detected by the use of an agonist radioligand.27 Ki val-
ues were calculated from the IC50 as described
previously.28
Acknowledgements
The authors wish to thank Ms. Heidi Figler for perform-
ing the receptor binding assays and the Centre for Chiral
and Molecular Technologies at Deakin University for
financial support. An Australian Postgraduate Award
for S.A.H. is also gratefully acknowledged.
5. Drug solutions
Stock solutions of the compounds were prepared in
DSMO at a concentration of 10mM and stored frozen.
On the day of use, the stock solution was diluted with
HankÕs balanced salt solution for cAMP assays or
50mM Tris–HCl buffer at pH7.4 for receptor assays.
Control incubations contained the same final concentra-
tion of DMSO.
References and notes
1. Lerman, B. B.; Belardinelli, L. Circulation 1991, 83, 1499–
1509.
2. Wilbur, S. L.; Marchlinski, F. E. Am. J. Cardiol. 1997, 79,
30–37.
6. Cell culture and biochemical assays
3. Clark, K. L.; Merkel, L. Emer. Drugs 2000, 5, 89–108.
4. Sorbera, L. A.; Castaner, J.; Martin, L.; Bayes, M. Drugs
Future 2002, 27, 846–849.
5. Scammells, P. J.; Baker, S. P.; Belardinelli, L.; Olsson, R.
A.; Russell, R. A.; Wright, D. M. Tetrahedron 1996, 52,
4735–4744.
6. Scammells, P. J.; Baker, S. P.; Belardinelli, L.; Olsson, R.
A.; Russell, R. A.; Knevitt, S. A. Bioorg. Med. Chem. Lett.
1996, 6, 811–814.
7. Hutchinson, S. A.; Baker, S. P.; Scammells, P. J. Bioorg.
Med. Chem. Lett. 1999, 9, 933–936.
8. Snowdy, S.; Liang, H. X.; Blackburn, B.; Lum, R.;
Nelson, M.; Wang, L.; Pfister, J.; Sharma, B. P.; Wolff, A.;
Belardinelli, L. Br. J. Pharmacol. 1999, 126, 137–146.
9. Gurden, M. F.; Coates, J.; Ellis, F.; Evans, B.; Foster, M.;
Hornby, E.; Kennedy, I.; Martin, D. P.; Strong, P.;
Vardey, C. J.; Wheeldon, A. Br. J. Pharmacol. 1993, 109,
693–698.
Hamster smooth muscle DDT1 MF-2(DDT) and rat
pheochromcytoma PC-12cells were grown as monolay-
ers and used one day preconfluent. Cell membranes for
receptor assays and intact cell suspensions for cAMP as-
says were prepared as reported.15 The Ki values for the
compounds at the A1AR in DDT cell membranes and
the A2AAR in PC-12membranes were determined from
displacement of specific [3H]-8-cyclopentyl-1,3-di-
propylxanthine and [3H]ZM241385, respectively. The
concentration of compounds that inhibited (ꢁ)isopro-
terenol-stimulated cAMP by 50% (IC50) in DDT cells
and stimulated half-maximal cAMP accumulation
(EC50) in PC-12cells was determined as previously
described.15
7. Data analysis
10. Pfister, J. R.; Berlardinelli, L.; Lee, G.; Lum, R. T.;
Milner, P.; Stanley, W. C.; Linden, J.; Baker, S. P.;
Schreiner, G. J. Med. Chem. 1997, 40, 1773–1778.
11. Hutchinson, S. A.; Scammells, P. J., unpublished results.
12. Sandler, M. Chem. Rev. 1966, 66, 297–339.
13. Soller, U. In Small Ring Heterocycles; Hasser, A., Ed.;
Wiley: New York, 1983; Vol. 42, p 333.
14. Fujisawa, T.; Kobori, T. Chem. Lett. 1972, 1065–1068.
15. Hutchinson, S. A.; Baker, S. P.; Scammells, P. J. Bioorg.
Med. Chem. 2002, 10, 1115–1122.
16. Bombala, M. U.; Ley, S. V. J. Chem. Soc., Perkin Trans. 1
1979, 3013–3016.
17. Krow, G. R.; Cannon, K. C.; Carey, J. T.; Ma, H.;
Raghavachari, R.; Szczepanski, S. W. J. Org. Chem. 1988,
53, 2665–2668.
The IC50 and EC50 values for cAMP accumulation were
determined by nonlinear regression analysis of the con-
centration–response using GraphPad software (San
Diego, CA, USA). The concentration of compounds
that inhibited specific radioligand binding by 50%
(IC50) was determined by nonlinear regression analysis
using the GraphPad software. The Ki values were calcu-
lated from the IC50 using the Cheng and Prusoff conver-
sion.25 The intrinsic activity (IA) of the compounds is
expressed as a fraction of the maximal response for each
compound as compared to the maximum for the stand-
ards comprising CPA and CGS 21680 for the A1 and
A2AAR responses, respectively. Statistical analysis of
the data was performed using the StudentÕs t-test and
differences were considered significant if p < 0.05.
18. Sajiki, H. Tetrahedron Lett. 1995, 36, 3465–3468.
19. Jackson, A. E.; Johnstone, A. W. Synthesis 1976, 685–
687.
20. Sivanandaiah, K. M.; Gurusiddappa, S. J. Chem. Res. (S)
1979, 108–109.
21. Okada, Y.; Ohta, N. Chem. Pharm. Bull. 1982, 30, 581–
585.
8. Radioligand binding to recombinant human receptors
Competitive binding assay were performed using rec-
ombinant human receptors with the following radiolig-
22. Felix, A. M.; Heimer, E. P.; Lambros, T. J.; Tzougraki,
C.; Meienhofer, J. J. Org. Chem. 1978, 43, 4194–4196.

4884
S. A. Hutchinson et al. / Bioorg. Med. Chem. 12 (2004) 4877–4884
23. Meienhofer, J.; Kuromizu, K. Tetrahedron Lett. 1974,
3259–3262.
26. Linden, J.; Thai, T.; Figler, H.; Jin, X.; Robeva, A. S. Mol.
Pharmacol. 1999, 56, 705–713.
24. ElAmin, B.; Anantharamaiah, M.; Royer, G. P.; Means,
G. E. J. Org. Chem. 1979, 44, 3442–3444.
25. Cheng, Y.; Prusoff, S. Biochem. Pharmacol. 1973, 22,
3099.
27. Murphree, L. J.; Marshall, M. A.; Rieger, J. M.;
MacDonald, T. L.; Linden, J. Mol. Pharmacol. 2002, 61,
455–462.
28. Linden, J. J. Cyclic Nucl. Res. 1982, 8, 163–172.