Design, synthesis, and biological evaluation of LNA nucleosides as adenosine A3 receptor ligands

Article abstract of DOI:10.1016/j.bmc.2007.05.056

We have prepared a series of adenosine analogs based on the bicyclo[2.2.1]heptane scaffold of locked nucleic acid (LNA) and tested them for both agonist and antagonist activity at the adenosine A₃ receptor. The design of these derivatives was b

Full text of DOI:10.1016/j.bmc.2007.05.056

Bioorganic & Medicinal Chemistry 15 (2007) 5440–5447
Design, synthesis, and biological evaluation of LNA
nucleosides as adenosine A3 receptor ligands
Jacob Ravn,* Katrine Qvortrup, Christoph Rosenbohm and Troels Koch
Santaris Pharma A/S, Bøge Alle 3, DK-2970 Hørsholm, Denmark
Received 6 November 2006; revised 21 May 2007; accepted 23 May 2007
Available online 26 May 2007
Abstract—We have prepared a series of adenosine analogs based on the bicyclo[2.2.1]heptane scaffold of locked nucleic acid (LNA)
and tested them for both agonist and antagonist activity at the adenosine A3 receptor. The design of these derivatives was based on
the known A3 agonist IB-MECA and related compounds. Modifications thus include the 50-uronamides and N6-(3-iodobenzyl)
derivatives. In this way we have prepared analogs of known A3 agonists with the sugar ring restricted in an N-conformation.
For comparison we have also prepared 20-O-methyl derivatives of IB-MECA. The LNA nucleosides showed no agonist activity
but some of them are potent antagonists. The 20-O-methyl derivative of IB-MECA is an agonist with similar potency as the parent
compound.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
The adenosine receptors (AR) are members of the family
of G-protein coupled receptors. Four distinct subtypes
of ARs are known, the A1, A2A, A2B, and A3 subtypes,
respectively. Adenosine receptors are widely distributed
throughout the body and therefore play a role in a wide
range of physiological functions and responses.1
In recent years the A3AR has been the focus of increased
interest because of its involvement in several clinical
indications. Ligands at the A3AR, in particular agonists,
are therefore of significant therapeutic value.2,3 For
example, in cancer the reference A3AR agonist IB-
MECA, 14 (Fig. 1), has been shown to inhibit tumor
growth in different in vivo models5,6 and is currently in
phase II clinical trial for colorectal cancer.7 Further-
more, A3AR agonists are believed to have beneficial
effects on different kinds of ischemia, lung injury, and
arthritis. A3AR antagonists on the other hand have been
suggested as anti-glaucoma, anti-inflammatory agents
and for the treatment of renal failure and asthma.2,3
Traditionally, agonists for the human A3AR have been
adenosine derivatives.8 Adenosine itself is an intermedi-
Figure 1.
Keywords: LNA; Amino-LNA; Adenosine A3 receptor; IB-MECA.
*
Corresponding author. Tel.: +45 4517 9817; fax: +45 4517 9898;
e-mail: jra@santaris.com
ately potent agonist with an EC50 value of 290 nM.9
Through manipulation at different sites of adenosine it
is possible to obtain highly potent and selective ligands
Present address: Department of Chemistry, University of Copenha-
gen, DK-2100 Copenhagen, Denmark
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.05.056

J. Ravn et al. / Bioorg. Med. Chem. 15 (2007) 5440–5447
5441
R1
R1
been explored using ligands with conformationally re-
stricted ribose rings. In this way the high-affinity agonist
5 was discovered by replacement of the ribose ring with
a bicyclo[3.1.0]hexane scaffold in which the carbocyclic
pseudosugar ring is restricted in a North-conformation
indicating a conformational preference of the receptor.14
A direct comparison of an N- and S-conformation could
be made after the preparation of 6 and 7 with binding
constants (Ki) of 404 and 63,000 nM, respectively.15
From this study it is apparent that the N-conformation
of the sugar ring is preferred at the A3AR. Interestingly,
the spirolactam 8 was shown to be an antagonist with
reasonable affinity.16 This complete loss of agonist activ-
ity could be attributed to the conformational restriction
of either the sugar ring or the amide group. The hydro-
gen bonding ability of the latter has been shown to be
important for activation of the A3AR.17
NH
NH
N
N
N
N
N
N
R2
R2
N
N
O
O
O
NH
OH
OH
10, R1 = H, R2 = CH2OH,
14, R1 = H, R2 = CH2OH,
11, R1 = H, R2 = CONHMe,
15, R1 = H, R2 = CONHMe,
12, R1 = 3-iodobenzyl, R2 = CH2OH
13, R1 = 3-iodobenzyl, R2 = CONHMe
16, R1 = 3-iodobenzyl, R2 = CH2OH
17, R1 = 3-iodobenzyl, R2 = CONHMe
R1
NH
N
N
R2
N
N
O
OH
O
18, R1 = H, R2 = CH2OH,
19, R1 = H, R2 = CONHMe,
In the present work, we wanted to further study the ef-
fect of conformationally restricted adenosine analogs by
the introduction of locked nucleic acid (LNA18) adeno-
sine analogs (Fig. 2). LNA is based on a 2,5-dioxabicy-
clo[2.2.1]heptane scaffold constraining the sugar ring in
an almost complete North-conformation.19 In addition,
we have prepared a series of 20-amino-LNA adenosine
analogs, containing the 2-oxa-5-azabicyclo[2.2.1]hep-
tane scaffold, thereby combining the conformational
characteristics of the LNA scaffold and a hydrogen
bond donor at the 20-heteroatom. Although LNA is
a bicyclic structure, we will use the conventional
nucleoside numbering in order to compare with other
nucleosides. Because the 20-oxygen in LNA is alkylated,
we have also prepared a few of the monocyclic 20-O-
methyl adenosine derivatives. These compounds will
contribute to an understanding of the role of the
20-OH group as it, with a minimal of structural change,
removes the 20-OH hydrogen which is believed to be
crucial in binding to the receptor. For all scaffolds we
have prepared the four possible combinations of
N-methyl-50-uronamide and N6-(3-iodobenzyl) modifica-
tions (Fig. 2) for a direct comparison with the reference
compound IB-MECA (1).
20, R1 = 3-iodobenzyl, R2 = CH2OH
21, R1 = 3-iodobenzyl, R2 = CONHMe
Figure 2.
at the A3AR, although the boundary between full ago-
nists, partial agonists, and antagonists is highly depen-
dent on the exact combination of modifications.8
Generally, N6-alkylated 50-uronamide derivatives are
amongst the most potent ligands reported. Further sub-
stitution at the 2-position is tolerated, for example, Cl-
IB-MECA 2,10 and can lead to increased selectivity
but can also convert an agonist into an antagonist. It
is generally believed that one or both of the 20- and 30-
hydroxy groups are required for A3AR activation
though no conclusive study has been made. Both the
20-deoxy-20-fluoro and the 30-deoxy-30-fluoro derivatives
311,12 and 412,13 of Cl-IB-MECA showed a dramatic de-
crease in binding affinity toward the A3AR suggesting a
role for the hydroxy groups as hydrogen bond donors,
not hydrogen bond acceptors.
The influence of the ribose ring conformation on
binding affinity and activation of the A3AR has also
I
NH2
N
NH2
N
NH2
N
HN
N
N
N
N
N
N
N
N
N
O
O
O
HN
HN
HN
HO
N
i
N
N
N
O
O
O
O
i
ii
iii
O
O
O
O
OBn
OBn
OH
11
OH
22
23
13
ii
I
NH2
N
HN
N
N
N
N
HO
HO
N
N
N
O
iii
O
O
O
OH
OH
10
12
Scheme 1. Reagents and conditions: (i) (a) PhI(OAc)2, TEMPO, CH3CN/H2O, 40 ꢁC; (b) SOCl2, EtOH, rt; (c) MeNH2, MeOH, rt; (ii) HCO2NH4,
Pd(OH)2/C, MeOH, reflux; (iii) (a) 3-iodobenzyl bromide, DMF, 70 ꢁC; (b) NH4OH, MeOH, 50 ꢁC.

5442
J. Ravn et al. / Bioorg. Med. Chem. 15 (2007) 5440–5447
2. Results and discussion
used to prepare the corresponding 20-O-methyl adenosine
analogs 19–21 (Scheme 3) from commercially available
18.
2.1. Chemistry
For the preparation of LNA adenosine analogs the 30-O-
benzyl derivative 2220 was used as starting material
(Scheme 1). Treating this compound with diacetoxyiodo-
benzene and a catalytic amount of 2,2,6,6-tetramethylpi-
peridine-1-oxyl (TEMPO) gave the corresponding 50-
uronic acid which was converted to the N-methyl-50-uro-
namide23 using SOCl2/EtOH followed by treatment with
excess methylamine. Debenzylation via palladium cata-
lyzed transfer hydrogenation then gave the fully deprotec-
ted N-methyl-50-uronamide 11. Alkylation of N-1 of this
compound with 3-iodobenzyl bromide followed by ring-
opening/ring-closure under alkaline conditions (Dimroth
rearrangement21 where N-1 becomes the exocyclic N6)
gave the N6-(3-iodobenzyl)-N-methyl-50-uronamide 13.
Similar N6-alkylation of LNA adenosine 10 (obtained
from 2220) finally gave the N6-(3-iodobenzyl)-50-hydroxy
nucleoside 12. We have recently reported the synthesis
of 20-amino-LNA purine nucleosides, including the
adenosine 14.22 The starting compound 2422 was used in
this study to prepare a series of 20-amino-LNA adenosine
analogs (Scheme 2). Boc-protection ofthe 20-aminogroup
was followed by removal of the 50-O-mesylate via the
benzoyl ester to give the key intermediate 25. Conversion
of this intermediate to the 50-uronamide 15 was
accomplished using the same method as for 11. Because
the Boc group is lost during treatment with SOCl2,
N6-alkylation with 3-iodobenzyl bromide was applied to
nucleoside 25to give 26 followed by oxidation to the
50-uronamide. Again, methanesulfonic acid was used for
the final debenzylation giving 17. For the synthesis of
the 50-hydroxy analog 16 on the other hand, the sequence
was simply reversed, starting with catalytic transfer
hydrogenation followed by alkylation. Finally, the same
synthetic strategy used in the preparation of 11–13 was
2.2. Adenosine A3 receptor activity
The prepared compounds were all tested for activity at
the human adenosine A3 receptor. AequoScreenTM
(Euroscreen, Belgium) cell lines expressing the A3
human recombinant receptor were used as a functional
assay to estimate both agonist and antagonist activity
in terms of EC50 and IC50 values, respectively. The
results for all tested compounds are shown in Table 1.
In this assay, the reference agonist IB-MECA (1) was
measured to have an EC50 of 2.0 nM, which is in good
accordance with literature data.23
NH2
N
NH2
N
N
N
N
N
O
HO
HN
N
N
O
O
NH
NH
ii, v
OH
14
OH
15
ref 23
NHBz
NH2
N
N
N
N
N
MsO
HO
N
N
N
i
O
O
NH
NBoc
OBn
OBn
24
25
iii
v, i ii, vi
I
I
HN
HN
N
NH2
N
NH2
N
N
N
N
N
N
N
N
N
N
N
HO
HO
O
N
HO
HN
O
O
N
N
i
O
O
NBoc
NH
OBn
OH
OH
O
OH
19
O
ii
26
16
18
ii, iv
ii
I
HN
N
N
N
I
I
N
O
HN
HN
HN
N
N
N
O
N
N
O
HO
HN
N
N
N
NH
O
O
OH
17
OH
20
O
OH O
Scheme 3. Reagents and conditions: (i) (a) Boc2O, Et3N, DCM, rt; (b)
NaOBz, DMSO, 100 ꢁC; (c) NH3, MeOH, rt; (ii) (a) PhI(OAc)2,
TEMPO, CH3CN/H2O, 40 ꢁC; (b) SOCl2, EtOH, rt; (c) MeNH2,
MeOH, rt; (iii) (a) 3-iodobenzyl bromide, DMF, 70 ꢁC; (b) NH4OH,
MeOH, 50 ꢁC; (iv) MsOH, DCM, rt; (v) HCO2NH4, Pd(OH)2/C,
MeOH, reflux; (vi) TFA, rt.
21
Scheme 2. Reagents and conditions: (i) (a) PhI(OAc)2, TEMPO,
CH3CN/H2O, 40 ꢁC; (b) SOCl2, EtOH, rt; (c) MeNH2, MeOH, rt; (ii)
(a) 3-iodobenzyl bromide, DMF, 70 ꢁC; (b) NH4OH, MeOH,
50 ꢁC.

J. Ravn et al. / Bioorg. Med. Chem. 15 (2007) 5440–5447
5443
The purpose of this study was to investigate the effect of
introducing a 20-O,40-C-methylene linkage into a known
A3 receptor agonist. As can be seen from the data for the
LNA nucleosides 10–13 this modification completely
eliminates the agonist activity but leaving a moderate
antagonist activity, for example, 150 nM for 12. Surpris-
ingly, the LNA analog of IB-MECA, 13, has neither
agonist, nor antagonist activity. The introduction of
the 20-O,40-C-methylene linkage has two distinct effects:
(i) locking the sugar furanose ring in an N-type confor-
mation and (ii) alkylating the 20-oxygen, thus removing
a potential hydrogen bond donor. In order to further
investigate this relationship, we prepared a small series
of 20-amino-LNA adenosines 14–17. These compounds
effectively preserve the conformational restriction but
reintroduce a hydrogen bond donor at the 20-position.
This exchange of the 20-oxygen with an NH does not
restore any agonist activity. The introduction of the
20-amino group does, however, increase the antagonist
activity, especially for the 3-iodobenzyl analogs, result-
ing in very potent antagonists 16 and 17. These results
indicate that the lack of agonist activity in the LNA
nucleoside compounds may not be due to removal of
the 20-hydroxy hydrogen, but is rather caused by restric-
tion of the furanose ring into an unfavorable conforma-
tion. In order to further elaborate on this, we prepared a
series of 20-O-methyl adenosine derivatives 18–21, thus
removing the 20-OH hydrogen but leaving the confor-
mational flexibility of the furanose ring essentially
intact. From the receptor data for these compounds it
is clear that they all bind efficiently to the receptor. It
also seems that the N-methyl-50-uronamide is required
for agonist activity. Notably, in the direct comparison
of 21 and 1, it is apparent that the 20-O-methyl group
does not affect the agonist activity. This clearly indicates
that the binding of 1 to the receptor does not involve the
20-OH group as a hydrogen bond donor. This is in obvi-
ous contrast to the work of Jeong and co-workers.11
They concluded that the diminished activity of the cor-
responding 20-deoxy-20-fluoro compound 3 was caused
by the removal of the 20-OH group as a hydrogen bond
donor. As an alternative, we would like to propose that
the fluorine atom in 3 shifts the conformation of the
furanose ring, causing a drop in activity. These results
clearly suggest that the lack of agonist activity of the
LNA and 20-amino-LNA adenosine derivatives is
caused by the restriction of conformation. This is sur-
prising considering that the carbocyclic nucleoside 5,
in which the pseudosugar ring is restricted into a confor-
mation close to that of the LNA nucleosides, is an extre-
mely potent agonist. The furanose ring of LNA
nucleosides has been shown to be locked in an 30-endo
conformation19 (Fig. 3) which is also the preferred con-
formation of 20-deoxy-20-fluoro nucleosides.24 The pseu-
dofuranose ring of
5
is locked in an 20-exo
conformation.14 The antagonist data for especially the
20-amino-LNAs 16 and 17 show that the compounds
still bind to the receptor, although with lower affinity.
Apparently, the 30-endo conformation (LNA and 20-
deoxy-20-fluoro) does not activate the receptor, while
the 20-exo (6) conformation does. It must be concluded
that the activation of the receptor is highly dependent
on the furanose conformation and that only small
changes in this conformation have a detrimental effect
on the activity. The increased antagonist activity of 20-
amino-LNA adenosines over LNA adenosines is also
interesting. Considering the lack of involvement of the
20-OH hydrogen in the binding of IB-MECA, it is
unlikely that the reintroduction of a hydrogen (–O– to
–NH–) is responsible for this higher potency. Instead,
a potential positive charge on the 20-nitrogen might
influence the binding mode.
In order to investigate the issue of selectivity we also
measured the agonist and antagonist effect of the five
most potent compounds 16–18, 20, and 21 against the
closely related adenosine A1 receptor (Table 1). None
of the tested compounds showed any activity up to a
concentration of 3 lmol and consequently must be
Figure 3.
Table 1. In vitro functional agonist (EC50, EMAX) and antagonist (IC50, IMAX) data for adenosine analogs at the human adenosine A3 and A1
receptors
R1
R2
A3EC50 (nM)
EMAX (%)
IMAX (%)
A1 EC50 (nM)
A1 IC50 (nM)
a
A3 IC50 (nM)
Compound
1
10
11
12
13
14
15
16
17
18
19
20
21
3-Iodobenzyl
H
CONHMe
CH2OH
2.0
100b
ND
>105
>105
150
>105
>105
170
8.2
ND
ND
ND
ND
>105
>105
>105
>105
>105
>105
>105
>105
>105
453
H
CONHMe
CH2OH
ND
ND
ND
ND
3-Iodobenzyl
3-Iodobenzyl
H
100
CONHMe
CH2OH
ND
ND
ND
ND
H
CONHMe
CH2OH
100
100
100
100
ND
ND
3-Iodobenzyl
3-Iodobenzyl
H
>3000
>3000
>3000
ND
>3000
>3000
>3000
ND
CONHMe
CH2OH
12
15
H
CONHMe
CH2OH
CONHMe
100
100
ND
1.7
3-Iodobenzyl
3-Iodobenzyl
>105
1.9
100
>3000
>3000
>3000
>3000
ND
ND, not determined.
a IC50 was not determined for compounds showing agonist activity.
b Reference level.

5444
J. Ravn et al. / Bioorg. Med. Chem. 15 (2007) 5440–5447
considered as highly selective A3 ligands, at least over
the A1 receptor.
Ph), 6.07 (1H, s, H-10), 4.89 (1H, s, H-20), 4.71 (1H, s,
H-30), 4.68 (1H, d, J = 11.9 Hz, PhCH2), 4.63 (1H, d,
J = 11.9 Hz, PhCH2), 4.39 (1H, d, J = 8.2 Hz, Ha-100),
3.98 (1H, d, J = 8.2 Hz, Hb-100), 2.70 (3H, d,
J = 4.7 Hz, CH3). ESI-MS m/z 397.2 (MH+).
In conclusion, we have prepared a series of LNA and 20-
amino-LNA analogs of IB-MECA. Human adenosine
A3 receptor activation data showed that these deriva-
tives are not agonists, but fairly potent antagonists.
Comparisons with the corresponding 20-O-methyl aden-
osine analogs indicated that this lack of agonist activity
is caused by conformational restriction of the furanose
ring, not by alkylation of the 20-oxygen. Highly potent
antagonists were obtained when the 20-oxygen in the
LNA adenosines was replaced with an NH-group.
3.1.2.
(1R,3R,4R,7S)-(7-Hydroxy-3-(adenin-9-yl)-2,5-
dioxabicyclo[2:2:1]hept-1-yl)-N-methylcarboxamide (11).
Compound 23 (2.5 g, 6.3 mmol) was dissolved in meth-
anol (50 mL) and added HCO2NH4 (4.0 g, 63 mmol)
and Pd(OH)2 20% on carbon (0.5 g, 0.58 mmol). The
mixture was stirred at reflux for 16 h and then filtered
through a pad of Celite. Evaporation of the solvent un-
der reduced pressure gave 11 as a brown solid (1.7 g,
90%). The crude product was used without further puri-
fication in the next step. A 140 mg sample was purified
by RP-HPLC to give 75 mg of pure material. 1H
NMR (DMSO-d6, 300 MHz) d 8.36 (1H, s, H-2), 8.30
(1H, q, J = 4.7 Hz, NHMe), 8.18 (1H, s, H-8), 7.27
(2H, s, NH2), 6.01 (1H, br s, OH), 6.00 (1H, s, H-10),
4.60 (1H, s, H-20), 4.57 (1H, s, H-30), 4.35 (1H, d,
J = 8.0 Hz, Ha-100), 3.91 (1H, d, J = 8.0 Hz, Hb-100),
2.68 (3H, d, J = 4.7 Hz, CH3). CI HR-MS: theoretical
mass (C12H14N6O4): 307.11547 [M+H]+, measured
mass: 307.115598.
3. Experimental
3.1. General procedures
˚
Anhydrous solvents were dried over 4 A molecular
sieves (4–8 mesh). Reagents were used as supplied. 20-
O-methyl adenosine was purchased from ChemGenes
(Wilmington, MA), 2,6-dichloropurine was purchased
from ChangChem Ltd (Nantong, China), and all other
reagents were purchased through Sigma–Aldrich Den-
mark. Dry column vacuum chromatography (DCVC)
was performed according to the published procedure.25
Reverse phase preparative HPLC (RP-HPLC) was per-
formed on a Biotage C18HS 12+S column. A gradient
of 20–100% acetonitrile in water was used as eluent.
1H NMR spectra were recorded on a Bruker AC-300
(300 MHz) spectrometer. Electrospray ionization mass
spectra were recorded on an Agilent 1100 series LC–
MS. High resolution mass spectra (Fast Atom Bom-
bardment, FAB or Chemical Ionization, CI) were re-
corded by Kent Mass Spectrometry (Kent, UK).
Human Adenosine A3 receptor assay was run at Euro-
screen, Belgium.
3.1.3. (1S,3R,4R,7S)-7-Hydroxy-1-hydroxymethyl-3-(N6-
(3-iodobenzyl)-adenin-9-yl)-2,5-dioxabicyclo[2:2:1]heptane
(12). To a solution of 1020 (1.0 g, 3.6 mmol) in anhy-
drous DMF (50 mL) was added 3-iodobenzyl bromide
(2.13 g, 7.2 mmol) and the mixture was stirred at 70 ꢁC
for 16 h. Additional 3-iodobenzyl bromide (1.1 g,
3.6 mmol) was added and stirring continued for 16 h.
The solvent was removed under reduced pressure and
the residual oil was triturated with a mixture of ether
and acetone (4:1). The resulting brown gum was redis-
solved in methanol (15 mL) and NH4OH (32%, aq)
(10 mL), and stirred for 16 h. The solvents were re-
moved under reduced pressure and the residue was puri-
fied by DCVC (d = 6 cm, 0–20% MeOH in EtOAc,
100 mL fractions, 0.5% increase pr fraction) to give 12
3.1.1. (1R,3R,4R,7S)-(7-Benzyloxy-3-(adenin-9-yl)-2,5-
dioxabicyclo[2:2:1]hept-1-yl)-N-methylcarboxamide (23).
To a solution of 2220 (5.0 g, 13.5 mmol) in acetonitrile
(25 mL) and H2O (25 mL) were added (diacetoxyiodo)-
benzene (9.6 g, 29.7 mmol) and TEMPO (0.64 g,
4.1 mmol), and the mixture was stirred at 40 ꢁC for
2 h. The solvents were removed under reduced pressure
and the residue was triturated with diethyl ether. The
white precipitate was filtered off, washed with ether,
and dried under vacuum (4.9 g, 94%). The intermediate
(4.4 g, 11.5 mmol) was dissolved in 99.9% ethanol
(60 mL) and cooled to 0 ꢁC. SOCl2 was added dropwise
over 10 min and stirring was continued at room temper-
ature for 3 h. The mixture was concentrated in vacuo
and the residual solvent was coevaporated with metha-
nol. The residual oil was redissolved in a 2 M solution
of methylamine in methanol (40 mL, 80 mmol) and the
mixture was stirred at room temperature for 16 h. The
solvent and reagent were removed under reduced pres-
sure and the residue was coevaporated twice with meth-
anol to give the target product 23 as a brown solid foam
(5.1 g, 99%). 1H NMR (DMSO-d6, 300 MHz) d 8.35
(1H, q, J = 4.7 Hz, NHMe), 8.31 (1H, s, H-2), 8.15
(1H, s, H-8), 7.37 (2H, s, NH2), 7.32–7.29 (5H, m,
1
as a white solid (0.45 g, 25%). H NMR (DMSO-d6,
300 MHz) d 8.48 (1H, br s, CH2NH), 8.27 (1H, s, H-
2), 8.22 (1H, s, H-8), 7.72 (1H, br s, 3-I-Ph), 7.59 (1H,
d, J = 7.9 Hz, 3-I-Ph), 7.37 (1H, d, J = 7.7 Hz, 3-I-Ph),
7.10 (1H, t, J = 7.8 Hz, 3-I-Ph), 5.92 (1H, s, H-10),
5.70 (1H, br s, OH), 5.06 (1H, br s, OH), 4.67 (2H, br
s, 3-I-PhCH2), 4.43 (1H, s, H-20 or H-30), 4.26 (1H, s,
H-20 or H-30), 3.92 (1H, d, J = 8.2 Hz, Ha-100), 3.79–
3.76 (3H, m, H-50, Hb-100). FAB HR-MS: theoretical
mass (C18H18IN5O4): 496.04817 [M+H]+, measured
mass: 496.047325.
3.1.4. (1R,3R,4R,7S)-(7-Hydroxy-3-(N6-(3-iodobenzyl)-
adenin-9-yl)-2,5-dioxabicyclo[2:2:1]hept-1-yl)-N-methyl-
carboxamide (13). Same procedure as for the synthesis of
12 using crude 11 (90 mg, 0.29 mmol) as starting mate-
rial. The crude product was purified by DCVC
(d = 3 cm, 0–8% MeOH in DCM, 25 mL fractions,
0.5% increase pr fraction) to give 13 as a white solid
(40 mg, 26%). 1H NMR (CDCl3, 300 MHz) d 8.40
(1H, s, CH2NH), 7.84 (1H, s, H-2 or H-8), 7.72 (1H,
s, H-2 or H-8), 7.62 (1H, d, J = 7.9 Hz, 3-I-Ph), 7.39

J. Ravn et al. / Bioorg. Med. Chem. 15 (2007) 5440–5447
5445
(1H, d, J = 7.3 Hz, 3-I-Ph), 7.26 (1H, s, 3-I-Ph), 7.06
(1H, t, J = 7.8 Hz, 3-I-Ph), 6.89 (1H, br s, NHMe),
6.27 (1H, br s, OH), 6.10 (1H, s, H-10), 4.83 (2H, br s,
3-I-PhCH2), 4.80 (1H, s, H-20 or H-30), 4.75 (1H, s, H-
20 or H-30), 4.54 (1H, d, J = 8.4 Hz, Ha-100), 4.10 (1H,
d, J = 8.4 Hz, Hb-100), 2.90 (3H, d, J = 5.0 Hz, CH3).
FAB HR-MS: theoretical mass (C19H19IN6O4):
523.05907 [M+H]+, measured mass: 523.05852.
0.96 mmol), and the mixture was stirred at 70 ꢁC for
16 h. Additional 3-iodobenzyl bromide (142 mg,
0.48 mmol) was added and stirring continued for 16 h.
The solvent was removed under reduced pressure and
the residual oil was triturated with a mixture of ether
and acetone (4:1). The resulting brown gum was redis-
solved in methanol (2 mL), and NH4OH (32%, aq)
(2 mL) and stirred for 16 h. The solvents were removed
under reduced pressure and the residue was redissolved
in TFA (4 mL). The mixture was stirred at room tem-
perature for 1 h and then evaporated to a yellow oil.
The crude product was purified by RP-HPLC to give
16 (16 mg, 10%). 1H NMR (DMSO-d6, 300 MHz) d
8.43 (1H, br s, CH2NH), 8.26 (1H, s, H-2), 8.21 (1H,
s, H-8), 7.73 (1H, br s, 3-I-Ph), 7.57 (1H, d,
J = 7.7 Hz, 3-I-Ph), 7.35 (1H, d, J = 7.6, Hz, 3-I-Ph),
7.11 (1H, t, J = 7.7 Hz, 3-I-Ph), 5.83 (1H, s, H-10),
5.35 (1H, br s, OH), 4.97 (1H, br s, OH), 4.67 (2H, br
s, 3-I-PhCH2), 4.10 (1H, s, H-30), 3.74 (2H, br s, H-50),
3.54 (1H, s, H-20), 2.99 (1H, d, J = 10.0 Hz, Ha-100),
2.74 (1H, d, J = 10.0 Hz, Hb-100). FAB HR-MS: theoret-
ical mass (C18H19IN6O3): 495.06416 [M+H]+, measured
mass: 495.063836.
3.1.5. (1R,3R,4R,7S)-7-Benzyloxy-5-tert-butoxycarbonyl-
1-hydroxymethyl-3-(adenine-9-yl)-2-oxa-5-aza-bicyclo
[2:2:1]heptane (25). To a solution of 2422 (500 mg,
0.91 mmol) in DCM (20 mL) and triethylamine (3 mL)
was added di-tert-butyl dicarbonate (418 lL,
1.82 mmol) and the solution was stirred at room temper-
ature for 2 h. The solution was washed with saturated aq
NaHCO3 (20 mL), dried over Na2SO4, filtered, and
evaporated to
a yellow oil (ESI-MS m/z 651.2
[M+H]+). The oil was redissolved in DMSO (10 mL)
and added NaOBz (525 mg, 3.6 mmol) and the mixture
was stirred at 90 ꢁC for 2 h. H2O (20 mL) was added
and the mixture was extracted with DCM (3· 50 mL).
The combined organic phase was washed with brine
(2· 50 mL), dried over Na2SO4, filtered, and evaporated
to a yellow solid foam (ESI-MS m/z 677.1 [M+H]+). The
solid was dissolved in saturated NH3 in methanol
(30 mL) and stirred at room temperature for 16 h. The
solvent was removed in vacuo and the residue was puri-
fied by DCVC (d = 2 cm, 0–5% MeOH in DCM, 25 mL
fractions, 0.5% increase pr fraction) to give 32 as a white
3.1.8. (1R,3R,4R,7S)-7-Benzyloxy-5-tert-butoxycarbonyl-
1-hydroxymethyl-3-(N6-(3-iodobenzyl)-adenine-9-yl)-2-
oxa-5-aza-bicyclo[2:2:1]heptane (26). Compound 25
(300 mg, 0.64 mmol) was converted to 26 using the same
procedures as for the synthesis of 12. The crude product
was purified by DCVC (d = 2 cm, 0–8% MeOH in
DCM, 25 mL fractions, 0.5% increase pr fraction) to
give 26 as a white solid (130 mg, 30%). 1H NMR
(CDCl3, 300 MHz, 2 rotamers 2:3) d 8.36, 8.35 (1H, s,
CH2NH), 7.84, 7.80 (1H, s, H-8), 7.74, 7.73 (1H, s, H-
2), 7.61 (1H, d, J = 7.9 Hz, 3-I-Ph), 7.32–7.27 (7H, m,
Ph, 3-I-Ph), 7.06 (1H, t, J = 7.8 Hz, 3-I-Ph), 6.03, 5.95
(1H, s, H-10), 4.92, 4.69 (1H, s, H-20), 4.80 (2H, br s,
3-I-PhCH2) 4.62–4.55 (3H, m, H-30, CH2Ph,), 4.03,
4.01 (1H, d, J = 12.5 Hz, CH2OH), 4.03, 3.90 (1H, d,
J = 12.5 Hz, CH2OH), 3.64, 3.60 (1H, d, J = 10.1 Hz,
Ha-100), 3.41, 3.35 (1H, d, J = 10.1 Hz, Hb-100), 1.49,
1.46 (9H, s, C(CH3)3); ESI-MS m/z 685.3 [M+H]+.
1
solid (0.22 g, 50%). H NMR (DMSO-d6, 300 MHz, 2
rotamers 1:1) d 8.22, 8.22 (1H, s, H-2), 8.17, 8.14 (1H,
s, H-8), 7.34 (2H, br s, NH2), 7.29–7.24 (5H, m, Ph),
5.98, 5.95 (1H, s, H-10), 5.22 (1H, m, OH), 4.97, 4.73
(1H, s, H-20), 4.63-4.49 (2H, m, CH2Ph), 4.31, 4.30
(1H, s, H-30), 3.81–3.73 (2H, m, CH2OH), 3.45 (1H, d,
J = 10.1 Hz, Ha-100), 3.26 (1H, d, J = 10.1 Hz, Hb-100),
1.43 (9H, s, C(CH3)3); ESI-MS m/z 469.1 [M+H]+.
3.1.6. (1R,3R,4R,7S)-(7-Hydroxy-3-(adenin-9-yl)-2-oxa-
5-azabicyclo[2:2:1]hept-1-yl)-N-methylcarboxamide (15).
Compound 25 (300 mg, 0.64 mmol) was converted to 15
using the same procedures as for the synthesis of 23 and
11. The crude product was purified by RP-HPLC to give
39 mg (21%) of pure material. 1H NMR (DMSO-d6,
300 MHz) d 8.34 (1H, s, H-2), 8.17 (1H, s, H-8), 8.16
(1H, br s, NHMe), 7.32 (2H, s, NH2), 5.88 (1H, s, H-
10), 4.32 (1H, s, H-30), 3.69 (1H, s, H-20), 3.50 (1H, d,
J = 10.3 Hz, Ha-100), 2.85 (1H, d, J = 10.3 Hz, Hb-100),
2.68 (3H, d, J = 4.6 Hz, CH3). CI HR-MS: theoretical
mass (C12H15N7O3): 306.13146 [M+H]+, measured
mass: 306.131584.
3.1.9. (1R,3R,4R,7S)-(7-Hydroxy-3-(N6-(3-iodobenzyl)-
adenin-9-yl)-2-oxa-5-azabicyclo[2:2:1]hept-1-yl)-N-meth-
ylcarboxamide (17). To a solution of 26 (110 mg
0.16 mmol) in acetonitrile (2 mL) and H2O (2 mL) were
added (diacetoxyiodo)benzene (320 mg, 1.0 mmol), and
TEMPO (15 mg, 0.1 mmol) and the mixture was stirred
at 40 ꢁC for 2 h. The solvents were removed under
reduced pressure and the residue was triturated with
diethyl ether. The white precipitate was filtered off,
washed with ether, and dried under vacuum. The inter-
mediate was dissolved in 99.9% ethanol (5 mL) and
cooled to 0 ꢁC. SOCl2 (0.26 mL, 3.5 mmol) was added
and stirring was continued at room temperature for
3 h. The mixture was concentrated in vacuo and the
residual solvent was coevaporated with methanol. The
residual oil was redissolved in a 2 M solution of methyl-
amine in methanol (5 mL, 10 mmol) and the mixture
was stirred at room temperature for 16 h. The solvent
and reagent were removed under reduced pressure to
give the intermediate as a brown foam. The crude
3.1.7. (1S,3R,4R,7S)-7-Hydroxy-1-hydroxymethyl-3-(N6-
(3-iodobenzyl)-adenin-9-yl)-2-oxa-5-azabicyclo[2:2:1]hep-
tane (16). Compound 25 (150 mg, 0.32 mmol) was
dissolved in methanol (4 mL) and added HCO2NH4
(101 mg, 1.60 mmol) and Pd(OH)2 20% on carbon
(0.05 g, 0.05 mmol). The mixture was stirred at reflux
for 5 h and then filtered through a pad of Celite. The sol-
vent was removed under reduced pressure to give a
white solid. The solid was redissolved in anhydrous
DMF (5 mL), added 3-iodobenzyl bromide (285 mg,

5446
J. Ravn et al. / Bioorg. Med. Chem. 15 (2007) 5440–5447
intermediate was redissolved in anhydrous DCM (4 mL)
and added methanesulfonic acid (2 mL mmol), and the
mixture was stirred at room temperature for 3 h. The
reaction mixture was diluted with anhydrous DCM
(10 mL), cooled to 0 ꢁC, and neutralized by the careful
addition of triethylamine (4.5 mL). The mixture was
washed with saturated aq NaHCO3 (10 mL), and brine
(10 mL) and then dried over Na2SO4. The solvent was
removed under reduced pressure to give a yellow foam.
The crude final product was purified by RP-HPLC to give
17(14 mg).1 HNMR(DMSO-d6, 300 MHz)d8.48(1H, br
s,CH2NH),8.38(1H,s,H-2),8.24(1H,s,H-8),8.10(1H,br
s,NHMe),7.73(1H,s,3-I-Ph),7.60(1H,d,J = 7.9 Hz,3-I-
Ph), 7.38 (1H, d, J = 7.6 Hz, 3-I-Ph), 7.11 (1H, t,
J = 7.7 Hz, 3-I-Ph), 5.92 (1H, s, H-10), 5.67 (1H, br s,
OH), 4.68 (2H, br s, 3-I-PhCH2), 4.36 (1H, s, H-30), 3.77
(1H, s, H-20), 3.52 (1H, d, J = 10.4 Hz, Ha-100), 2.89 (1H,
d, J = 10.4 Hz, Hb-100), 2.68 (3H, d, J = 4.6 Hz, CH3).
FAB HR-MS: theoretical mass (C19H20IN7O3):
522.07506 [M+H]+, measured mass: 522.073557.
4.81 (2H, br s, 3-I-PhCH2), 4.73 (1H, dd, J = 7.5 Hz
and 4.6 Hz, H-20), 4.60 (1H, d, J = 4.6 Hz, H-30), 4.33
(1H, s, H-40), 3.93 (1H, d, J = 12.9 Hz, Ha-50), 3.74
(1H, t, J = 12.1 Hz, Hb-50), 3.33 (3H, s, OCH3). FAB
HR-MS: theoretical mass (C18H20IN5O4): 498.06382
[M+H]+, measured mass: 498.064358.
3.1.12. N6-(3-Iodobenzyl)-20-O-methyladenosine-N-methyl-
50-uronamide (21). From 19 (200 mg, 0.65 mmol) using
the same procedure as for 13. The crude product was
purified by DCVC (d = 2 cm, 0–6% MeOH in DCM,
25 mL fractions, 0.5% increase pr fraction) to give 21
as a white solid (164 mg, 48%). 1H NMR (CDCl3,
300 MHz) d 8.34 (1H, s, H-2), 7.82 (1H, s, H-8 or 3-I-
Ph), 7.73 (1H, s, H-8 or 3-I-Ph), 7.57 (1H, d,
J = 8.1 Hz, 3-I-Ph), 7.32 (1H, d, J = 7.3 Hz, 3-I-Ph),
7.04 (1H, t, J = 7.8 Hz, 3-I-Ph), 5.99 (1H, d,
J = 8.0 Hz, H-10), 4.79 (2H, br s, 3-I-PhCH2), 4.62
(1H, s, H-40), 4.56 (1H, d, J = 4.4 Hz, H-30), 4.47 (1H,
dd, J = 7.4 and 4.4 Hz, H-20), 3.31 (3H, s, OCH3), 2.91
(3H, s, NHCH3). FAB HR-MS: theoretical mass
(C19H21IN6O4): 525.07472 [M+H]+, measured mass:
525.074340.
3.1.10. 20-O-Methyladenosine-N-methyl-50-uronamide (19).
To a solution of 18 (3.0 g, 10.7 mmol) in acetonitrile
(25 mL) and H2O (25 mL) were added (diacetoxyiodo)-
benzene (10.3 g, 32.0 mmol) and TEMPO (0.50 g,
3.2 mmol) and the mixture was stirred at 40 ꢁC for 4 h.
The white precipitate was filtered off, washed with aceto-
nitrile, and dried under vacuum (1.50 g). The reaction
mixture was evaporated and the residue was triturated
with diethyl ether. The white precipitate was filtered
off, washed with ether, and dried under vacuum
(1.45 g). The two portions were combined to give a total
yield of the uronic acid of 2.95 g, 94%. The intermediate
(1.37 g, 4.6 mmol) was dissolved in 99.9% ethanol
(30 mL) and cooled to 0 ꢁC. SOCl2 (3.4 mL, 46 mmol)
was added dropwise over 10 min and stirring was con-
tinued at room temperature for 3 h after which a white
precipitate was formed. The mixture was concentrated
in vacuo and the residual solvent was coevaporated with
methanol to give a white solid material. The material
was redissolved in a 2 M solution of methylamine in
methanol (40 mL, 80 mmol) and the mixture was stirred
at room temperature for 16 h. The white precipitate was
filtered off, washed with diethyl ether, and dried under
vacuum to give the target product 19 as a white solid
3.2. Human adenosine A3 receptor assay
AequoScreenTM (Euroscreen, Belgium) cell lines
expressing the A3 human recombinant receptor and
the promiscuous G protein Ga16 were used throughout
the study. AequoScreenTM cells were cultured following
recommended conditions for at least one week prior
to the test. The day before the test, cells were har-
vested with PBS-EDTA, washed, and re-suspended
in BSA-DMEM–F12 (Dulbecco’s modified Eagle’s
medium—Ham’s F12 with 0.1% BSA). Suspended
cells were then incubated at room temperature with
Coelenterazine H overnight. For agonist and antago-
nist testing, 50 ll of the cell suspension was injected
onto 50 ll of the test compound or control in 96-well
plates, and the resulting emission of light measured
for the determination of cell activation. After the first
reading, and following an incubation time of 15–
30 min, 100 ll of the reference agonist (IB-MECA)
at a concentration equal to the EC80 of the day of
the experiment was injected onto the cell suspension
containing the test compounds. The resulting emission
of light was measured for the determination of
antagonistic effects. Light emission for agonist and
antagonist tests was recorded using a Hamamatsu
FDSS-6000 reader. For agonist data, percentages of
activation were calculated on the basis of the activa-
tion (luminescence data) induced by the reference ago-
1
(1.34 g, 94%) H NMR (DMSO-d6, 300 MHz) d 8.85
(1H, m, NHCH3), 8.47 (1H, s, H-2), 8.24 (1H, s, H-8),
7.43 (2H, br s, NH2), 6.11 (1H, d, J = 7.1 Hz, OH),
5.79 (1H, d, J = 3.9 Hz, H-10), 4.45-4.36 (3H, m, H-20,
H-30 and H-40), 3.24 (3H, s, OCH3), 2.72 (3H, d,
J = 4.3 Hz, NHCH3). CI HR-MS: theoretical mass
(C12H16N6O4): 309.13112 [M+H]+, measured mass:
309.131270.
nist at
a saturating concentration (EC100). For
antagonist data, percentages of inhibition were calcu-
lated on the basis of the activation (luminescence
data) induced by the reference agonist at a concentra-
tion equal to the EC80. The test compounds were
tested as duplicate determinations at eight concentra-
tions of 100, 10, 1, 0.1 lM and 10, 1, 0.1, and
0.01 nM for agonist activity and 50, 5, 0.5, 0.05 lM
and 5, 0.5, 0.05, and 0.005 nM for antagonist activ-
ity. Dose–response data, EC50/IC50, from test com-
pound were analyzed with XLfit (IDBS) software.
3.1.11. N6-(3-Iodobenzyl)-20-O-methyladenosine (20).
From 18 (200 mg, 0.71 mmol) using the same procedure
as for 12. The crude product was purified by RP-HPLC
to give 20 (60 mg, 17%). 1H NMR (CDCl3, 300 MHz) d
8.36 (1H, s, H-2), 7.72 (1H, s, H-8 or 3-I-Ph), 7.71 (1H,
s, H-8 or 3-I-Ph), 7.61 (1H, d, J = 7.9 Hz, 3-I-Ph), 7.28
(1H, d, J = 7.7 Hz, 3-I-Ph), 7.05 (1H, t, J = 7.8 Hz, 3-
I-Ph), 6.82 (1H, br d, J = 6.8 Hz, 30-OH), 6.69 (1H, br
t, J = 6.7 Hz, 50-OH), 5.81 (1H, d, J = 7.5 Hz, H-10),

J. Ravn et al. / Bioorg. Med. Chem. 15 (2007) 5440–5447
5447
10. Kim, H. O.; Ji, X. D.; Siddiqi, S. M.; Olah, M. E.; Stiles,
G. L.; Jacobson, K. A. J. Med. Chem. 1994, 37,
3614–3621.
11. Kim, H. O.; Park, J. G.; Moon, H. R.; Gunaga, P.; Lim,
M. H.; Chun, M. W.; Jacobson, K. A.; Kim, H. D.; Jeong,
L. S. Nucleosides Nucleotides Nucleic Acids 2003, 22,
927–930.
12. Jeong, L. S.; Lee, H. W.; Jacobson, K. A.; Lee, S. K.;
Chun, M. W. Nucleic Acids Symp. Ser. 2005, 49,
31–32.
13. Lim, M. H.; Kim, H. O.; Moon, H. R.; Lee, S. J.; Chun,
M. W.; Gao, Z. G.; Melman, N.; Jacobson, K. A.; Kim, J.
H.; Jeong, L. S. Bioorg. Med. Chem. Lett. 2003, 13,
817–820.
14. Tchilibon, S.; Joshi, B. V.; Kim, S.-K.; Duong, H. T.;
Gao, Z.-G.; Jacobson, K. A. J. Med. Chem. 2005, 48,
1745–1758.
15. Jacobson, K. A.; Ji, X.; Li, A. H.; Melman, N.; Siddiqui,
M. A.; Shin, K. J.; Marquez, V. E.; Ravi, R. G. J. Med.
Chem. 2000, 43, 2196–2203.
16. Gao, Z. G.; Kim, S. K.; Biadatti, T.; Chen, W.; Lee, K.;
Barak, D.; Kim, S. G.; Johnson, C. R.; Jacobson, K. A.
J. Med. Chem. 2002, 45, 4471–4484.
17. Gao, Z. G.; Joshi, B. V.; Klutz, A. M.; Kim, S. K.; Lee, H.
W.; Kim, H. O.; Jeong, L. S.; Jacobson, K. A. Bioorg.
Med. Chem. Lett. 2006, 16, 596–601.
3.3. Human adenosine A1 receptor assay
GPCR-expressing clones were transfected to coexpress
human adenosine A3 receptor (GenBank Accession No.
L22607), mitochondrial apoaequorin, and Ga16. Cells
were collected from plates with PBS containing 5 mM
EDTA, pelleted, and re-suspended at 107 cells/ml in
DMEM-F12 medium. Cells were then incubated with
5 lM Coelenterazine H (Molecular Probes) for 4 h at
room temperature, washed in DMEM-F12 culture med-
ium, and re-suspended at
a
concentration of
2 · 106 cells/ml. Cells were then mixed with test molecules
and light emission by the aequorin is recorded with a
Hamamatsu FDSS6000 luminometer for 30 s. The test
compounds were tested as duplicate determinations at
eight concentrations of 3000, 1000, 300, 100, 30, 10, 3,
and 1 nM. Data were analyzed with the PRISM software
(GraphPad Prism Software, San Diego, CA) using non-
linear regression applied to a sigmoidal dose–response
model. Results are expressed as percent of the agonist re-
sponse. Controls include assays using cells not expressing
GPCR (mock transfected), in order to exclude possible
non-specific effects of the candidate compound.
18. ‘LNA’ is defined as an oligonucleotide containing one or
more 20-O,40-C-methylene-b-D-ribofuranosyl nucleotide
monomers. The term ‘20-amino-LNA’ is used herein for
LNA analogs, where the 20-oxygen has been substituted
by a nitrogen atom. The term ‘LNA nucleoside’ (e.g.,
adenosine) refers to a 20-O,40-C-methylene-b-D-ribofur-
anosyl nucleoside.
19. Obika, S.; Nanbu, D.; Hari, Y.; Morio, J. A. K.; In, Y.;
Ishida, T.; Imanishi, T. Tetrahedron Lett. 1997, 38,
8735–8738.
References and notes
1. Palmer, T. M.; Stiles, G. L. Neuropharmacology 1995, 34,
683–694.
2. Jacobson, K. A.; Gao, Z. G. Nat. Rev. Drug Discov. 2006,
5, 247–264.
3. Fishman, P.; Bar-Yehuda, S. Curr. Top. Med. Chem. 2003,
3, 463–469.
4. Gallo-Rodriguez, C.; Ji, X. D.; Melman, N.; Siegman, B.
D.; Sanders, L. H.; Orlina, J.; Fischer, B.; Pu, Q.; Olah,
M. E.; van Galen, P. J. J. Med. Chem. 1994, 37, 636–646.
5. Fishman, P.; Bar-Yehuda, S.; Ardon, E.; Rath-Wolfson,
L.; Barrer, F.; Ochaion, A.; Madi, L. Anticancer Res.
2003, 23, 2077–2083.
6. Merimsky, O.; Bar-Yehuda, S.; Madi, L.; Fishman, P.
Drug Dev. Res. 2003, 58, 386–389.
7. Stemmer, S.; Shani, A.; Klein, B.; Silverman, M.; Lorber,
I.; Farbstein, M.; Shmueli, E.; Figer, A. J. Clin. Oncol.
(Meeting Abstracts) 2006, 24, 13082.
20. Koshkin, A. A.; Fensholdt, J.; Pfundheller, H. M.;
Lomholt, C. J. Org. Chem. 2001, 66, 8504–8512.
21. Fuji, T.; Terahara, N. Chem. Pharm. Bull. 1986, 34,
1094–1107.
22. Ravn, J.; Rosenbohm, C.; Christensen, S. M.; Koch,
T. Nucleosides Nucleotides Nucleic Acids 2006, 25,
843–847.
23. Jacobson, K. A.; Park, K. S.; Jiang, J. L.; Kim, Y. C.;
Olah, M. E.; Stiles, G. L.; Ji, X. D. Neuropharmacology
1997, 36, 1157–1165.
24. Kawasaki, A. M.; Casper, M. D.; Freier, S. M.; Lesnik, E.
A.; Zounes, M. C.; Cummins, L. L.; Gonzalez, C.; Cook,
P. D. J. Med. Chem. 1993, 36, 831–841.
8. Joshi, B. V.; Jacobson, K. A. Curr. Top. Med. Chem. 2005,
5, 1275–1295.
9. Fredholm, B. B.; Irenius, E.; Kull, B.; Schulte, G.
Biochem. Pharmacol. 2001, 61, 443–448.
25. Pedersen, D. S.; Rosenbohm, C. Synthesis 2001, 16,
2431–2434.