Structure-Activity Relationship of Purine Ribonucleosides for Inhibition of Hepatitis C Virus RNA-Dependent RNA Polymerase

Article abstract of DOI:10.1021/jm030424e

As part of a continued effort to identify inhibitors of hepatitis C viral (HCV) replication, we report here the synthesis and evaluation of a series of nucleoside analogues and their corresponding triphosphates. Nucleosides were evaluated for their ability to inhibit HCV RNA replication in a cell-based, subgenomic replicon system, while nucleoside triphosphates were evaluated for their ability to inhibit in vitro RNA synthesis mediated by the HCV RNA-dependent RNA polymerase, NS5B. 2′-C-Methyladenosine and 2′-C-methylguanosine were identified as potent inhibitors of HCV RNA replication, and the corresponding triphosphates were found to be potent inhibitors of HCV NS5B-mediated RNA synthesis. The data generated in the cell-based assay demonstrated a fairly stringent structure- activity relationship around the active nucleosides. Increase in steric bulk beyond methyl on C2, change in the stereo- or regiochemistry of the methyl substituent, or change of identity of the heterobase beyond that of the endogenous adenine or guanine was found to lead to loss of inhibitory activity. The results highlight the importance of the ribo configuration 2′- and 3′-hydroxy pharmacophores for inhibition of HCV RNA replication in the cell-based assay and demonstrate that inclusion of the 2′ -C-methylribonucleoside pharmacophore leads to increased resistance to adenosine deaminase and purine nucleoside phosphorylase mediated metabolism.

Full text of DOI:10.1021/jm030424e

J . Med. Chem. 2004, 47, 2283-2295
2283
Str u ctu r e-Activity Rela tion sh ip of P u r in e Ribon u cleosid es for In h ibition of
Hep a titis C Vir u s RNA-Dep en d en t RNA P olym er a se
Anne B. Eldrup,*,† Charles R. Allerson,† C. Frank Bennett,† Sanjib Bera,† Balkrishen Bhat,† Neelima Bhat,†
Michele R. Bosserman,‡ J ennifer Brooks,† Christine Burlein,‡ Steven S. Carroll,‡ P. Dan Cook,† Krista L. Getty,‡
Malcolm MacCoss,§ Daniel R. McMasters,§ David B. Olsen,‡ Thazha P. Prakash,† Marija Prhavc,† Quanlai Song,†
J oanne E. Tomassini,‡ and J ie Xia†
Department of Medicinal Chemistry, Isis Pharmaceuticals, 2280 Faraday Avenue, Carlsbad, California 92008, Department of
Biological Chemistry, Merck Research Laboratories, West Point, Pennsylvania 19486, and Department of Medicinal Chemistry,
Merck Research Laboratories, Rahway, New J ersey 07065
Received September 2, 2003
As part of a continued effort to identify inhibitors of hepatitis C viral (HCV) replication, we
report here the synthesis and evaluation of a series of nucleoside analogues and their
corresponding triphosphates. Nucleosides were evaluated for their ability to inhibit HCV RNA
replication in a cell-based, subgenomic replicon system, while nucleoside triphosphates were
evaluated for their ability to inhibit in vitro RNA synthesis mediated by the HCV RNA-
dependent RNA polymerase, NS5B. 2′-C-Methyladenosine and 2′-C-methylguanosine were
identified as potent inhibitors of HCV RNA replication, and the corresponding triphosphates
were found to be potent inhibitors of HCV NS5B-mediated RNA synthesis. The data generated
in the cell-based assay demonstrated a fairly stringent structure-activity relationship around
the active nucleosides. Increase in steric bulk beyond methyl on C2, change in the stereo- or
regiochemistry of the methyl substituent, or change of identity of the heterobase beyond that
of the endogenous adenine or guanine was found to lead to loss of inhibitory activity. The results
highlight the importance of the ribo configuration 2′- and 3′-hydroxy pharmacophores for
inhibition of HCV RNA replication in the cell-based assay and demonstrate that inclusion of
the 2′-C-methylribonucleoside pharmacophore leads to increased resistance to adenosine
deaminase and purine nucleoside phosphorylase mediated metabolism.
In tr od u ction
therapies. The RNA-dependent RNA polymerase, en-
coded at the 3′-terminal portion of the HCV genome, is
essential to viral replication and thus represents a valid
target for therapeutic intervention by design of specific
inhibitors.
The hepatitis C virus (HCV) is the pathogen associ-
ated with the majority of sporadic and transfusion-
related non-A and non-B hepatitis infections. While
often asymptomatic, HCV infection can progress to
chronic hepatitis, leading to liver cirrhosis and some-
times hepatocellular carcinoma. Current estimates sug-
gest that 170 million people worldwide suffer from HCV
infection. At present, treatment options comprise im-
munotherapy using recombinant interferon-R in com-
bination with ribavirin. The clinical benefit of this
treatment is limited, and a vaccine has not yet been
developed. Consequently, HCV infection represents a
significant health problem in need of more effective
therapies.
The hepatitis C virus genome consists of a single open
reading frame encoding a >3000 amino acid protein that
is processed into multiple viral proteins including the
NS2/3 autoprotease, the NS3 serine protease and NT-
Pase/helicase, and NS5B, the RNA-dependent RNA
polymerase (RdRp).1-3 Significant effort has been di-
rected toward the identification of specific inhibitors of
the HCV protease NS3,4-9 revealing several, very potent
inhibitors. The clinical evaluation of these inhibitors is
ongoing and hence has yet to result in marketed
Nucleoside inhibitors (NIs) of virally encoded poly-
merases have been validated for other viral targets, such
as HIV. While a number of non-nucleoside inhibitors
(NNIs) of NS5B-mediated HCV RNA replication have
been reported,10-13 only a few NIs have been identi-
fied.14-16 Thus far, most NS5B NIs described have
involved some alteration of the 2′- or 3′-position. Virtu-
ally all NIs of virally encoded polymerases described
thus far elicit their effect through “chain termination”.
In order for a nucleoside triphosphate to effect chain
termination, the nucleoside triphosphate would need to
initially serve as a substrate for HCV NS5B, leading to
incorporation of the nucleoside monophosphate and
subsequent translocation within the active site in
preparation for extension of the growing RNA chain.
Second, the nucleoside triphosphate would need to
encompass one or more structural elements that would
effect “chain termination”, that is, significantly impede
the further elongation of the RNA chain. Structural
elements that influence the 3′-position sterically or
electronically would be expected to affect elongation
efficiency because the 3′-hydroxyl group participates
directly in phosphodiester formation as part of the
elongation process. The requirement for HCV NS5B to
be able to distinguish ribonucleoside triphosphates from
* To whom correspondence should be addressed. Phone: 760 603
3852. Fax: 760 603 4654. E-mail: aeldrup@isisph.com.
† Isis Pharmaceuticals.
‡ Merck Research Laboratories, PA.
§ Merck Research Laboratories, NJ .
10.1021/jm030424e CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/26/2004

2284 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9
Ta ble 1. Inhibitory Potency of 2′- and 3′-Modified Nucleoside
Eldrup et al.
Ta ble 2. Inhibitory Potency of 2′-Modified Nucleoside
Triphosphates on NS5B-Catalyzed RNA Synthesis (NS5B IC50)
and Inhibitory Potency of 2′-Modified Nucleosides on HCV RNA
Triphosphates on NS5B-Catalyzed RNA Synthesis (NS5B IC50
and Inhibitory Potency of 2′- and 3′-Modified Nucleosides on
HCV RNA Replication in a Cell-Based, Subgenomic Replicon
)
Replication in a Cell-Based, Subgenomic Replicon Assay
Assay (Replicon EC50
)
(Replicon EC50)
NS5B
replicona
NS5B
replicona
a
b
c
d
B
IC50, µM
EC50, µM
a
b
B
IC50, µM
EC50, µM
1
2
3
4
5
6
H
H
H
H
H
H
OH
OH
OH
OH
OCH3
OCH3
H
H
H
H
H
H
H
H
F
F
OH
OH
adenine
guanine
adenine
guanine
adenine
guanine
22
0.6
28
1.8
47
1.6
46
50
5
6
7
8
9
10
11
12
H
H
H
H
OH
OH
CH3
CH3
OCH3
OCH3
F
F
H
H
OH
OH
adenine
guanine
adenine
guanine
adenine
guanine
adenine
guanine
47
1.6
>50
substrate
50
20
1.9
0.13
>50
>50
>50
>50
>50
>50
0.3
1.2b
>50
>50
>50
a
b
Replicon data represent the value at 24 h. Although toxicity
3.5
was not observed at up to 50 µM in replicon cells at 24 h,
significant cytotoxicity was observed following longer-term culture
in several cell lines.
a
Replicon data represent the value at 24 h.
Sch em e 1. Synthetic Route to 2′-C-methylguanosine,
12a
2′-deoxynucleoside triphosphates would make it likely
that the polymerase be highly refined in its ability to
recognize the 2′-position of nucleoside triphosphates.
As part of a continued effort to identify competitive
inhibitors of HCV NS5B, we report here the synthesis
and evaluation of a series of nucleoside analogues and
their corresponding nucleoside triphosphates. On the
basis of the considerations outlined above, our initial
SAR was focused on the evaluation of modifications at
the 2′ and 3′ positions of the nucleoside (triphosphate).
Since some polymerases are known to promote high
error rates as a way to provide diversity to the viral
genome,17 we considered it possible that HCV NS5B
would be somewhat amenable to changes in the purine
heterocycle. Hence, in addition to the SAR at nucleoside
2′ and 3′ positions, a limited SAR was performed to
evaluate the effect of modulation of hydrogen-bonding
capacity. The ability of the nucleosides to inhibit HCV
RNA replication was evaluated in a cell-based, subge-
nomic HCV replicon assay as previously described.14
Nucleoside triphosphates were evaluated for their abil-
ity to inhibit HCV NS5B-mediated RNA elongation as
previously described.14 For the adenosine derivative that
proved active in the cell-based assay, the ability to resist
adenosine deaminase (ADA) mediated conversion to the
corresponding inosine derivative was assessed. Simi-
larly, the two most active purine derivatives were
evaluated for their resistance to purine nucleoside
phosphorylase (PNP) mediated phosphorolysis.
a
Reagents and conditions: (i) 2-amino-6-chloropurine, DBU and
trimethylsilyl triflate in actonitrile, 4 h at 65 °C; (ii) methanolic
ammonia, 7 h at room temperature; (iii) 2-mercaptoethanol and
sodium methoxide in methanol, 7 h at reflux.
similar strategy was applied (Scheme 1). Briefly, gly-
cosylation of 2-C-methyl-1,2,3,5-tetra-O-benzoyl-D-ri-
bose21 (13) under Vorbru¨ggen conditions yielded the
benzoyl-protected 2-amino-6-chloropurine derivative (14)
under stereocontrol induced by transient 1,2-acyloxo-
nium ion formation. Deprotection of this derivative gave
the ribonucleoside (15), which was readily converted to
2′-C-methylguanosine (12) by substitution of the 6-chlo-
ro substituent with hydroxide.
Ch em istr y
Addition of organometallics to 2-keto ribofuranose
intermediates is known to proceed stereoselectively with
addition of the alkyl group to the â face, leading to
formation of the ribo derivatives, as illustrated by the
synthesis of 12 (Scheme 1).21 Consequently, to explore
the effect of inversion of stereochemistry at C2′, we
utilized “the nucleoside route” proceeding via the previ-
ously reported and unstable 2-keto derivative.22 Thus,
adenosine was 3′,5′-O-protected with the bifunctional
reagent 1,3-dichloro-1,1,3,3,-teraisopropyldisiloxane
(TIPDS dichloride) to give 1623 (Scheme 2), which was
subsequently oxidized under J ones conditions to give
the 2-keto intermediate (17). This derivative was re-
acted with methylmagnesium bromide to afford the
For the initial SAR, a series of nucleoside triphos-
phates modified at the 2′ or 3′ positions (Tables 1 and
2) were synthesized from the corresponding nucleosides
and purified according to the general procedure for
triphosphate synthesis, purification, and QC described
in the Experimental Section. The following nucleosides
were synthesized according to available literature: 3′-
deoxy-3′-fluoroadenosine (3),18 3′-deoxy-3′-fluorogua-
nosine (4),18 2′-deoxy-2′-fluoroadenosine (7),19 and 2′-
deoxy-2′-fluoroguanosine (8)19 (Table 1). The nucleoside
2′-C-methyladenosine (11, Table 2) was synthesized
using a route similar to that described in ref 20. For
the corresponding guanosine derivative (12, Table 2), a

Purine Ribonucleosides
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9 2285
Sch em e 2. Synthetic Route to 2′-C-
methylarabinoadenosine, 19a
Sch em e 4. Synthetic Route to 2′-C-methyl-2′-O-
methyladenosine, 29a
a
Reagents and conditions: (i) CrO3, pyridine, acetic anhydride
in dichloromethane, 30 min at 0 °C; (ii) methylmagnesium bromide
in diethyl ether/tetrahydrofuran, 4 h at -30 to -50 °C; (iii)
hydrogen fluoride/triethylamine in dichloromethane/acetonitrile,
3 h at room temperature.
a
Reagents and conditions: (i) KOH, 18-crown-6 and MeI in
tetrahydrofuran, 3 h at room temperature; (ii) Pd/C, hydrogen,
dichloromethane/methanol/acetic acid, 24 h at room temperature;
(iii) acetic anhydride-pyridine (1:1), DMAP, 16 h at room temper-
ature; (iv) acetic anhydride, concentrated H2SO4, 1 h at 0 °C; (v)
6-chloropurine, trimethylsilyl triflate, DBU, acetonitrile, 5 h at
65 °C; (vi) methanolic ammonia, 16 h at room temperature.
Sch em e 3. Synthetic Route to 3′-C-methyladenosine,
23a
The effect of methylation of the 2′-hydroxyl group was
investigated by the synthesis of 2′-C-methyl-2′-O-meth-
yladenosine (29) according to the route depicted in
Scheme 4. Starting with the previously reported 3,5-
bis-O-(2,4-dichlorophenylmethyl)-2-C-methyl-1-O-meth-
yl-R-D-ribofuranose (25),26,27 methylation was performed
under phase-transfer conditions (KOH, 18-crown-6 and
methyl iodide) to give 26. The subsequent change of
protection groups from 2,6-dichlorobenzyl to acetyl
proceeded via catalytic hydrogenation and acetylation
under standard conditions to give 27. Insertion of the
acetate leaving group at C1 was performed under acidic
conditions to yield the key intermediate 28. Coupling
under Vorbru¨ggen conditions was performed to yield a
mixture of R/â diastereoisomers. Separation of the
diastereoisomeric mixture was accomplished after depro-
tection and substitution of the 6-chloro functionality,
using methanolic ammonia at elevated temperature, to
give the desired â anomer of 2′-C-methyl-2′-O-methyl-
adenosine (29) as well as its R anomer. In an effort to
further explore the structural effects at the furanose C2′,
2′-C-ethyladenosine was also synthesized (36, Scheme
5). An initial attempt to employ the strategy used in
the synthesis of the purine 2′-C-methylribonucleosides
11 and 12 failed, possibly because of failure to form the
prerequisite EtTiCl3 species. As an alternative to using
this relatively nonbasic, chemospecific titanium reagent,
we opted to instead use the ribose precursor with the
less-sensitive dichlorobenzyl protecting group, which
allowed the use of the more reactive EtMgBr. Synthesis
of 31 was subsequently achieved via the stereospecific
addition of the ethyl substituent to the â face of the
previously described 3,5-bis-O-(2,4-dichlorophenyl-
methyl)-1-O-methyl-R-D-erythro-pentofuranose-2-ulose
(30)26,27 using EtMgBr in diethyl ether. Benzoylation of
a
Reagents and conditions: (i) 6-chlorpurine, DBU and trimeth-
ylsilyl triflate in acetonitrile, 3 h at 60 °C; (ii) liquid ammonia in
dioxane, 14 h at 80 °C; (iii) ammonium formate and Pd/C in
methanol, 12 h at reflux.
TIPDS-protected 2′-C-methylarabino derivative (18)
under stereocontrol arising from steric hindrance of the
â face of the sugar by the purine heterocycle. Subse-
quent deprotection with hydrogen fluoride gave the
desired 2′-C-methylarabinoadenosine (19). To assess the
effect of change in regiochemistry for the methyl sub-
stituent, 3′-C-methyladenosine (23) and 3′-C-methylxy-
loadenosine (24) were synthesized. 3′-C-Methyladenos-
ine (23) was synthesized from 2024 according to the
procedure outlined in Scheme 3. Stereospecific coupling
of 20 with 6-chloropurine under Vorbru¨ggen conditions
yielded the fully protected derivative 21. Subsequent
ammonolysis with liquid ammonia at elevated temper-
ature gave 22 and was followed by deprotection under
hydrogen-transfer conditions to give the desired 3′-C-
methyladenosine (23). 3′-C-Methylxyloadenosine (24)
was synthesized by a method similar to that described
in ref 25.

2286 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9
Sch em e 5. Synthetic Route to 2′-C-ethyladenosine, 36a
Eldrup et al.
Sch em e 6. Synthetic Route to Heterobase Modified
2′-C-methylribonucleosides, 38-41a
a
Reagents and conditions: (i) Dess-Martin periodinane, dichlo-
romethane, 48 h at room temperature; (ii) EtMgBr, diethyl ether,
4 h at -78 to -15 °C; (iii) benzoyl chloride, DMAP, triethylamine,
dichloromethane, 48 h at 50 °C; (iv) Pd/C, hydrogen, methanol/
acetic acid (5:2); (v) benzoyl chloride, DMAP, triethylamine,
dichloromethane, 3 h at room temperature; (vi) acetic acid, acetic
anhydride, H2SO4, overnight at room temperature; (vii) 6-chloro-
purine, trimethylsilyl triflate, DBU, acetonitrile, 1.5 h at 50 °C;
(viii) ammonia/dioxane (5:1), overnight at 50 °C.
a
Reagents and conditions: (i) 1 N aqueous sodium hydroxide,
2 h at reflux; (ii) Pd/C and triethylamine in methanol under
hydrogen, overnight at room temperature; (iii) ammonium hy-
droxide, overnight at 80 °C.
Resu lts
the sterically hindered 2-hydroxyl group was performed
under forcing conditions, followed by catalytic hydro-
genation and subsequent esterification of the 5-hydroxyl
functionality to give 33. Hydrolysis of the ketal was
succeeded by acetylation to establish the acetate leaving
group at C1 (necessary for stereospecific coupling of the
purine base) and simultaneous esterification of the
2-hydroxyl group to give the derivative 34. This deriva-
tive was coupled under Vorbru¨ggen conditions to gener-
ate the protected derivative 35. Ammonolysis and
deprotection were achieved in a single step using liquid
ammonia at elevated temperature to give the desired
2′-C-ethyladenosine (36).
A series of nucleoside triphosphates, modified in the
furanosyl 2′- or 3′-positions, were screened against the
purified HCV NS5B in order to assess their ability to
inhibit HCV NS5B-mediated RNA synthesis (Table 1).
The corresponding nucleosides were screened in a cell-
based, subgenomic replicon assay for the ability to
inhibit HCV RNA replication; in this assay, activity
would be dependent on both cellular uptake and the
cell’s ability to metabolize the nucleoside to its tri-
phosphate (Table 1). While none of the screened nucleo-
sides demonstrated significant activity in the cell-based
assay (EC50 values ranging from 46 to >50 µM) [apart
from 3′-deoxy-3′-fluoroadenosine (EC50 ) 1.2 µM) which
was quite cytotoxic in long-term cell culture], several
nucleoside triphosphates in the guanine series proved
to be low- or submicromolar inhibitors of HCV NS5B-
mediated RNA synthesis. As might be expected, a
number of nucleoside triphosphates lacking the 3′-
hydroxyl moiety were potent inhibitors of HCV NS5B-
mediated RNA synthesis as illustrated by the deriva-
tives 3′-deoxyguanosine (Table 1, 2, IC50 ) 0.6 µM) and
3′-deoxy-3′-fluoroguanosine (Table 1, 4, IC50 ) 1.8 µM).
The observed inhibition of RNA synthesis most likely
occurs via incorporation of the nucleoside monophos-
phates and subsequent inability to support further
extension of the growing RNA chain (chain termination).
The purine ribonucleoside derivatives 38-41 (Scheme
6) were prepared in order to evaluate the effect of
changes in the hydrogen bond acceptor and donor sites
of the purine heterobase. 2′-C-Methylinosine (38) was
readily obtained from hydrolytic substitution on the
6-chloropurine derivative 37,20 while the purine deriva-
tive 39 was available through catalytic hydrogenation
of the same 6-chloropurine precursor. The remaining
members of this series were prepared from the 2-amino-
6-chloropurine-2′-C-methyl ribonucleoside 15 either
through ammonolysis to yield the 2,6-diaminopurine
derivative (40) or by catalytic hydrogenation to yield the
2-aminopurine derivative (41).

Purine Ribonucleosides
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9 2287
Ta ble 3. Inhibitory Potency of Modified Nucleoside
Somewhat surprisingly, alkylation of the 2′-hydroxyl
Triphosphates, Modified by Inclusion of Alkyl Substituents To
Give Alkyl-Derivatized Ribo, Ara, and Xylo Derivatives, on
NS5B-Catalyzed RNA Synthesis (NS5B IC50) and Inhibitory
Potency of Nucleosides, Modified by Inclusion of Alkyl
Substituents To Give Alkyl-Derivatized Ribo, Ara, and Xylo
Derivatives, on HCV RNA Replication in a Cell-Based,
group in guanosine triphosphate to the corresponding
2′-O-methylguanosine triphosphate (Table 1, 6, IC50
)
1.6 µM) also resulted in a potent inhibitor of RNA
synthesis. In contrast, 2′-O-methyladenosine triphos-
phate (Table 1, 5, IC50 ) 47 µM) failed to inhibit HCV
NS5B-mediated RNA synthesis.
Subgenomic Replicon Assay (Replicon EC50
)
Mod ifica tion s of th e Ribose 2-P osition . On the
basis of the finding that modification of the 2′-position
of ribonucleoside triphosphates could produce potent
chain terminators, we elected to further explore the
structural space around this position. While 2′-O-
methylguanosine triphosphate (Table 2, 6) was found
to be a potent inhibitor of HCV NS5B-mediated RNA
synthesis, the corresponding 2′-deoxy-2′-fluoroguanosine
triphosphate derivative (Table 2, 8, substrate) supported
RNA elongation following its incorporation into the RNA
chain. As before, the corresponding adenosine derivative
2′-deoxy-2′-fluoroadenosine triphosphate (Table 2, 7,
IC50 > 50 µM) derivative was found to be inactive. Since
both the 2′-O-methyl and 2′-deoxy-2′-fluoro modifica-
tions eliminate the 2′-hydroxy hydrogen bond donor, it
might be speculated that the ability of 6 to effect chain
termination is caused by steric interference between the
2′-O-methyl substituent and either the polymerase or
the incoming nucleoside triphosphate. The 2′-fluoro
modification might not cause sufficient steric hindrance
to trigger inhibition of RNA synthesis by such a mech-
anism. Alternatively, the disparity in activity might be
the result of differences in conformational preferences
of the two modified furanoses.28 To probe the effect of
substituents on the â face of the nucleoside, we tested
nucleosides and nucleoside triphosphates in which the
2′-hydroxyl group was inverted from the endogenous
ribo configuration to the arabino orientation (Table 2,
9 and 10). With both adenosine and guanosine deriva-
tives, activity was absent in the cell-based assay and,
for the triphosphates, was significantly reduced against
HCV NS5B. Only the arabinoguanosine triphosphate
displayed a weak inhibitory effect on HCV NS5B-
mediated RNA synthesis (Table 2, 10, IC50 ) 20 µM).
Thus, it appeared that while increased steric bulk
around the 2′-position could create nonobligate chain
terminators, exclusion of the hydrogen bond donor/
acceptor in the 2′-ribo position generally led to signifi-
cantly reduced, or in some cases complete abrogation
of inhibitory activity. This, in combination with the
observation that potency of the 3′-deoxy modified nu-
cleosides in the enzymatic assay did not translate into
inhibition in the cell-based assay, led us to examine the
2′-C-methyl derivatives 11 and 12 (Table 2). In both of
these derivatives, the ribo configuration of the 2′- and
3′-hydroxyl groups were maintained as recognition
elements, while the â face of the nucleoside was steri-
cally compromised by addition of a 2′-C-methyl sub-
stituent. This class of compounds demonstrated potent
activity in the cell-based assay, as well as against HCV
NS5B. For 2′-C-methyladenosine (Table 2, 11), inhibi-
tion of the enzyme by the triphosphate was moderate
(IC50 ) 1.9 µM), while the activity observed in the cell-
based assay was comparably better (EC50 ) 0.3 µM).
The opposite relationship was observed for the corre-
sponding 2′-C-methylguanosine derivative (IC50 ) 0.13
µM, EC50 ) 3.5 µM) (Table 2, 12).
NS5B
IC50, µM
replicona
EC50, µM
a
b
c
d
11
19
23
24
29
36
CH3
OH
H
H
CH3
Et
OH
CH3
OH
OH
OCH3
OH
H
H
OH
OH
OH
CH3
OH
OH
1.9
>50
>50
>50
>50
>50
0.3
>50
>50
>50
>50
>50
CH3
OH
H
H
a
Replicon data represent value at 24 h.
Meth yl Su bstitu tion s a t th e Ribose C2 a n d C3.
Structurally related analogues of the active adenosine
based inhibitor 11 (Table 3) were synthesized in order
to assess the effect of changes in stereo- and regiochem-
istry of the furanose moiety. The importance of the
stereochemistry at C2′ was demonstrated through the
synthesis and testing of the arabino derivative, which
was found to be inactive in the cell-based assay (EC50
> 50 µM) (19, Table 3). Likewise, the corresponding
triphosphate was found to be inactive against HCV
NS5B (IC50 > 50 µM), indicating that the lack of cell-
based activity is at least in part due to the enzyme’s
inability to accept this modification. The importance of
regiochemistry in the context of methyl substitution was
examined by synthesis of 3′-C-methyladenosine (23,
Table 3) and 3′-C-methylxyloadenosine (24, Table 3).
Both compounds tested inactive in the cell-based assay
(EC50 > 50 µM). Similarly, the corresponding triphos-
phates were synthesized and found to be incapable of
inhibiting HCV NS5B-mediated RNA synthesis (IC50
>
50 µM).
Effect of Meth yla tion of th e 2′-Hyd r oxyl Gr ou p
a n d In cr ea sed Ster ic Bu lk a t C2′. On the basis of
the inactivity of the arabino derivative 19, the 2′-
hydroxyl group of 11 appeared to be an important
structural element for activity. This was further inves-
tigated by synthesizing and evaluating the inhibitory
activity of 2′-C-methyl-2′-O-methyladenosine (29, Table
3). The nucleoside failed to inhibit HCV RNA replication
in the cell-based replicon assay (EC50 > 50 µM), as did
its R-anomer (EC50 > 50 µM) (data not shown). Fur-
thermore, the corresponding triphosphate displayed no
ability to inhibit HCV NS5B-mediated RNA synthesis
(IC50 > 50 µM). To more thoroughly examine the effect
of increased steric bulk around C2′, 2′-C-ethyladenosine
(36) and the corresponding triphosphate were synthe-
sized. The change in steric bulk was found to be
deleterious to both cell-based (EC50 > 50 µM) and
enzymatic activities (IC50 > 50 µM).
Mod ifica tion s of th e P u r in e Heter oba se. To probe
the tolerance for changes in hydrogen-bonding capacity
of the purine heterobase, 2′-C-methyl modified ribo-
nucleosides were synthesized in which hydrogen bond

2288 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9
Ta ble 4. Inhibitory Potency of Heterobase Modified Nucleoside
Eldrup et al.
bond donor in the purine 2-position to give the 2,6-
diaminopurine derivative 40 resulted in moderate activ-
ity in the cell-based assay (EC50 ) 7.0 µM). The
corresponding triphosphate proved to be a low-micro-
molar inhibitor of RNA synthesis (IC50 ) 2.6 µM).
Triphosphates on NS5B-Catalyzed RNA Synthesis (NS5B IC50
and Inhibitory Potency of Heterobase Modified Nucleosides on
HCV RNA Replication in a Cell-Based, Subgenomic Replicon
)
Assay (Replicon EC50
)
Ca lcu la t ion of Con for m a t ion a l P r efer en ce for
11, 19, 23, a n d 24. The relative energies of Northern-
and Southern-type conformers were calculated for 2′-
C-methyladenosine (11), 2′-C-methylarabinoadenosine
(19), 3′-C-methyladenosine (23), and 3′-C-methylxyload-
enosine (24) as described in the Experimental Section
(Table 5, Figure 1). Whereas 19 and 23 were calculated
to prefer a Southern, C2′-endo conformation by 2-3
kcal/mol, 11 and 24 displayed a similarly strong prefer-
ence for a Northern conformation according to the
computation (see Table 5). For comparison, the North-
ern and Southern conformers of adenosine were calcu-
lated by this method to lie much closer in energy, with
a Southern C2′-endo, C3′-exo twist conformer lying ca.
0.5 kcal/mol below the Northern C3′-endo conformer, in
general agreement with NMR population analyses (data
not shown).28,29 Thus, the conformational bias appears
to be driven by the preference of the methyl group for
an equatorial orientation.
NS5B
IC50, µM
replicona
EC50, µM
X
Y
11
12
38
39
40
41
NH2
OH
OH
H
NH2
H
H
NH2
H
H
NH2
NH2
1.9
0.13
4.0
nd
2.6
46
0.3
3.5
>50
>50
7.0
>50
a
Replicon data represent the value at 24 h.
Ta ble 5. Comparison of the Calculated Lowest-Energy
Northern- and Southern-Type Conformers for the Structures in
Figure 1a
Northern
Southern
compd E, kcal mol-1 P, deg τ, deg E,kcal mol-1 P, deg τ, deg
11
19
23
24
0
15.6 38.9
14.9 37.2
31.6 38.0
26.4 36.5
2.52
0
0
155.2 35.7
165.0 36.1
180.4 38.8
184.1 37.3
En zym a tic Con ver sion s Med ia ted by P u r in e
Nu cleosid e P h osp h or yla se a n d Ad en osin e Dea m i-
n a se. Pharmacokinetic data obtained in rat have dem-
onstrated the oral bioavailability of 2′-C-methylgua-
nosine (12) to be 82%.16 In contrast, the analogous
adenosine derivative (11) was found to be undetectable
in plasma after per oral dosing, with very rapid clear-
ance (>200 mL min-1 kg-1) observed after intravenous
dosing.14 The limited bioavailability of 2′-C-methylad-
enosine suggests rapid metabolism of this derivative.
To further investigate this possibility, we looked at the
ability of 2′-C-methyladenosine (11) and 2′-C-meth-
ylguanosine (12) to function as substrates for relevant
metabolic enzymes. Susceptibility of 11 and 12 to
cleavage by purine nucleoside phosphorylase (PNP) to
give the nucleoside 1′-phosphate and the purine het-
erocycle was examined at 1.8 and 18 U/mL at 27 h. For
2.25
2.91
0
1.99
a
E is the energy of the conformer relative to the MMFFs (ꢀ )
50) global minimum. P is the pseudorotational angle in degrees,
and τ is the puckering amplitude in degrees.
donors and/or acceptors were either omitted or added
(38-41, Table 4). Deletion of the 6-amino group to give
the purine ribonucleoside derivative (39) led to complete
loss of activity in the cell-based assay (EC50 > 50 µM).
Similarly, the inosine (38) and the 2-aminopurine (41)
derivatives were also unable to inhibit HCV RNA
replication in the cell-based assay (EC50 > 50 µM), while
the corresponding triphosphates were either low-micro-
molar or weak inhibitors of HCV NS5B (IC50 ) 4.0 and
46 µM, respectively). Addition of an amino hydrogen
F igu r e 1. Methyl substitution of furanose alters the conformational preference, thereby changing the positioning of the 2′- and
3′-hydroxyl groups. Of the methyl substitutions examined, only 11 retains the ribo configuration 2′- and 3′-hydroxy structure
elements as well as the preference for C3′-endo conformation of the furanose.

Purine Ribonucleosides
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9 2289
submicromolar activity against the isolated HCV NS5B
polymerase. Nucleoside triphosphates based on guanine
were generally more active than the corresponding
adenine-based derivatives against the NS5B poly-
merase. The underlying reasons for the functional
difference between adenosine and guanosine triphos-
phate derivatives remain elusive at this time. Deletion
of the 3′-hydroxyl group of guanosine to give 3′-deoxy
or 3′-deoxy-3′-fluoro derivatives resulted in activity at
the enzyme level; however, this activity generally did
not translate to activity in the cell-based assay. The
inability of the 3′-deoxy modified nucleosides to elicit
activity in the cell-based assay is an observation that,
in our experience, extends to various classes of nucleo-
sides, both purine- and pyrimidine-derived, in which the
3′-hydroxyl group is excluded (unpublished results). In
contrast to the activity generally observed with 3′-
deoxyguanosine triphosphates, modification of the 2′-
position of nucleoside triphosphates did not generally
lead to potent inhibitors of HCV NS5B-mediated syn-
thesis. Of the derivatives examined herein, only 2′-O-
methylguanosine triphosphate (6), 2′-C-methyladeno-
sine triphosphate (11), and 2′-C-methylguanosine triphos-
phate (12) displayed significant inhibitory effect on HCV
NS5B-mediated RNA synthesis. For 2′-O-methylgua-
nosine, the observed activity of the nucleoside triphos-
phate did not translate to a matching activity of the
nucleoside in the cellular assay. To gain a better
understanding of the underlying reason for the observed
inactivity of the 2′-O-methyl derivative 6 in the cell-
based assay, we evaluated cellular uptake and intrac-
ellular metabolism for the nucleoside 6. Uptake and
metabolic conversion to the corresponding triphosphate
(in hepatoma cells) was found to be inefficient (data not
shown). Thus, inefficient cellular uptake and metabo-
lism of 2′-O-methylguanosine (6) might provide an
explanation for the lack of activity of this compound in
the cellular assay. The ability of HCV NS5B to discrimi-
nate against nucleoside triphosphates modified in the
2′-position would not be surprising in light of the need
to avoid the incorporation of 2′-deoxynucleoside mono-
phosphates into the elongated RNA product. However,
on the basis of the results presented herein, it appears
that HCV NS5B is able to recognize 3′-deoxy modified
nucleoside triphosphates and that inactivity of the 3′-
deoxynucleosides in the cell-based assay may be the
result of upstream selectivity provided by nucleoside
transporters and/or nucleoside kinases. Further evalu-
ation of the potent 3′-deoxynucleoside triphosphate
NS5B inhibitors, as well as (monophosphate)prodrug
approaches to enhance cellular uptake and metabolic
profile of the corresponding nucleosides, will be de-
scribed elsewhere.
F igu r e 2. (a) Phosphorolysis of the glycosidic bond of nucleo-
sides catalyzed by purine nucleoside phosphorylase (PNP).
Nucleosides (200 µM) were incubated in the presence of
enzyme (1.8 or 18 units/mL) for 27 h followed by analysis for
product formation using HPLC as described in the experimen-
tal procedures. Under the experimental conditions, adenosine,
2′-C-methyladenosine (11), and guanosine were substrates for
PNP-mediated phosphorolysis, while no conversion of 2′-C-
methylguanosine (12) to product was detected. (b) Adenosine
deaminase-catalyzed conversion of nucleosides. Adenosine and
2′-C-methyladenosine (11) (25 µM) were incubated in the
presence of ADA, followed by analysis for product formation
using HPLC as described in experimental procedures. Under
reaction conditions that include 1.25 units/mL ADA, 2′-C-
methyladenosine was converted to 2′-C-methylinosine.
12, no cleavage of the purine was observed (Figure 2a),
whereas unmodified guanosine was efficiently cleaved
(82% and 87%, respectively, at 1.8 and 18 U/mL PNP).
In comparison, 11 was efficiently converted (37%) at the
higher enzyme concentration, while unmodified adenos-
ine was almost quantitatively converted (Figure 2a).
Similarly, we evaluated the ability of the adenosine
derivative 11 to function as a substrate for adenosine
deaminase (ADA) mediated conversion to 2′-C-methyli-
nosine derivative (Figure 2b). Under conditions in which
unmodified adenosine was 25% converted to inosine, 2′-
C-methyladenosine (11) underwent no appreciable con-
version. However, at higher enzyme concentrations both
adenosine and 11 could be quantitatively converted to
their corresponding inosine derivatives.
While 2′-C-methyladenosine triphosphate was found
to be a moderately active inhibitor of HCV NS5B-
mediated RNA synthesis (IC50 ) 1.9 µM), the corre-
sponding nucleoside was found to be comparably more
active in the cell-based assay (EC50 ) 0.3 µM). The
reverse relationship between the two activities was
observed for the guanosine derivative (IC50 ) 0.13 µM;
EC50 ) 3.5 µM) (11 and 12, Table 2). The observed
opposing trends are likely to reflect differences in
efficiency of nucleoside uptake and/or intracellular
metabolism to the corresponding triphosphate for the
Discu ssion
Of the various modifications to the nucleoside 2′- and
3′-positions examined herein, only the introduction of
the 2′-C-methyl substituent led to potent inhibition of
HCV RNA replication in the cell-based, subgenomic
replicon assay. However, several additional 2′- and 3′-
modified nucleoside triphosphates demonstrated low or

2290 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9
Eldrup et al.
two derivatives. Indeed, uptake and metabolism of the
guanosine derivative 12 to the triphosphate in hepato-
ma cells was found to be lacking,16 whereas the corre-
sponding adenosine derivative 11 was taken up and
processed efficiently under similar conditions.14 Hence,
for these two derivatives, potency in the cellular assay
appears to correlate with efficiency of cellular uptake
and metabolism to the triphosphate. The fact that cell-
based potency tracks with cellular uptake and metabolic
conversion to the triphosphate suggests that there may
be some room for improving the potency of 2′-C-
methylguanosine (12), perhaps through application of
(monophosphate)prodrug approaches. Inefficient uptake
of the guanine derivatives would be in line with general
observations made during the advancement of this
work, indicating a general lower level of uptake for
guanosine nucleoside analogues compared to that of
other ribonucleoside derivatives (unpublished results).
inactivity of 2′-O-methyladenosine (5) in the enzymatic
assay. The importance of the 2′-C-methyl pharmaco-
phore is accentuated by the complete loss of activity
upon addition of a single methylene unit in the closely
related ethyl derivative (36, Table 3). This derivative
was inactive both as a nucleoside in the cellular assay
and as a triphosphate in the HCV NS5B assay. Since
both the 2′-C-methyladenosine (12) and the 2′-C-ethyl-
adenosine (36) derivatives are calculated to exhibit a
strong preference for C3′-endo (results not shown),
conformational differences cannot explain the observed
inactivity of the 2′-C-ethyl derivative. Instead, it appears
that the enzyme is extremely sensitive to the steric
environment surrounding the C2′ position and that the
ethyl substituent per se is not tolerated by HCV NS5B.
A model derived from the NS5B crystal structure and
the æ6 RdRp initiation complex16 is consistent with a
tight steric environment around C2′ of the nucleoside
triphosphate prior to elongation. Effective inhibition of
HCV NS5B may require striking a delicate balance
between too little (2′-H) and too much (2′-C-ethyl) steric
bulk, and the size of the methyl substituent may be
nearly optimal in this respect.
A change in stereochemistry at C2′, as illustrated by
2′-C-methylarabinoadenosine (19), was found to result
in inactivity in both the cellular and enzymatic assays
(19, Table 3). The change of stereochemistry at C2′
results in transfer of the 2′-hydroxyl element to the â
face of the nucleoside, along with a change in the
calculated conformational preference to favor C2′-endo.
This conformational bias is in contrast to that of the
potent inhibitor 2′-C-methyladenosine (11), which was
calculated to prefer a C3′-endo conformation (Table 5,
Figure 1). The inability of the triphosphate of 19 to
inhibit HCV NS5B-mediated RNA synthesis is also in
line with the modest activity of arabinoadenosine (9)
triphosphate, in which the 2′-hydroxyl group is also
positioned on the â face of the nucleoside. The regio-
isomeric 3′-C-methyladenosine (23, Table 3) and 3′-C-
methylxyloadenosine (24, Table 3) both tested inactive
in the cellular assay as well as in the HCV NS5B assay
as triphosphates. Because 3′-deoxynucleosides are also
inactive in the cellular assay (1-4, Table 1), the
inactivity of derivative 24 in the same cellular assay
might be interpreted as confirmation of the requirement
for a hydroxyl group with proper stereochemical orien-
tation at C3 for effective inhibition. However, the
guanosine derived 3′-deoxynucleoside triphosphates (2
and 4, Table 1) are all active inhibitors of HCV RNA
synthesis despite the lack of the 3′-hydroxyl functional-
ity. In contrast, the nucleoside triphosphate of 24 was
unable to inhibit this enzyme-mediated process. More-
over, 3′-C-methyladenosine (23), which maintains both
hydroxyl groups present in the endogenous adenosine
substrate for the polymerase (ATP), was likewise unable
to act as an inhibitor (or a substrate) of HCV NS5B-
mediated RNA synthesis. The calculated conformational
preference of 23 (C2′-endo/C3′-exo) also differs from that
of 11 (C3′-endo), which may explain, at least in part,
the difference in ability to interact with HCV NS5B
(Table 5, Figure 1).
Some modification of the hydrogen-bonding capacity
of the purine heterocycle was found to be tolerated by
HCV NS5B without resulting in a complete loss of
activity. Simple changes such as the deletion of the
2-amino group of 12 to give the inosine derivative 38
(Table 3), as well as the insertion of an additional
hydrogen bond donor in 11 to give the 2,6-diaminopu-
rine derivative 40 (Table 4), were found to create
moderate nucleoside triphosphate inhibitors of HCV
NS5B-mediated RNA synthesis. Yet, for only one of
these derivatives, 40, did the ability of the nucleoside
triphosphate to inhibit RNA synthesis equate with an
ability of the nucleoside to inhibit HCV RNA replication
in the cell-based assay. However, the observed activity
of 40 in the cell-based assay is likely a result of efficient
conversion to the corresponding guanosine derivative
by adenosine deaminase (data not shown). Thus, the
results suggest that HCV NS5B itself has a limited
ability to discriminate against changes in the heterobase
hydrogen-bonding pattern but that high levels of selec-
tion against such modifications can occur upstream from
HCV NS5B, in the nucleoside metabolic pathway,
possibly by nucleoside receptors and/or nucleoside ki-
nases.
For the two derivatives with demonstrated ability to
inhibit HCV RNA replication, 11 and 12, our data
suggest a likely correlation between enzymatic stability
and oral bioavailability in rodents.14,16 2′-C-methylgua-
nosine (12) was found to be inert to cleavage of the
heterobase by purine nucleoside phosphorylase (PNP),
while 2′-C-methyladenosine was found to be somewhat
susceptible to cleavage by PNP (Figure 2a). Moreover,
conversion of 2′-C-methyladenosine (11) to the corre-
sponding inosine derivative by adenosine deaminase
was quite efficient (Figure 2b), indicating that both PNP
and ADA mediated conversions may contribute to
limiting the oral bioavailability of the adenosine deriva-
tive.
Methylation of the active 2′-C-methyladenosine (12)
to generate 2′-C-methyl-2′-O-methyl derivative 29 (Table
3) led to a loss in activity in the cell-based assay. The
corresponding triphosphate was also unable to inhibit
RNA synthesis by HCV NS5B, suggesting that in the
context of adenosine derivatives methylation of the 2′-
hydroxyl interferes with recognition by HCV NS5B. This
interpretation is further supported by the observed
In conclusion, 2′-C-methyladenosine and 2′-C-meth-
ylguanosine have been identified as potent inhibitors
of HCV RNA replication in a subgenomic replicon assay

Purine Ribonucleosides
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9 2291
Gen er a l P r oced u r e for th e Syn th esis of Nu cleosid e
Tr ip h osp h a tes. To the appropriate, unprotected nucleoside
(0.05 mmol) was added 4 Å molecular sieves (about 20-30)
(Mallinckrodt 4494) and trimethyl phosphate (stored over
sieves) (0.5 mL). The mixture was stirred overnight in a sealed
container. It was then cooled to 0 °C, and phosphorus oxy-
chloride (0.0070 mL, 0.075 mmol) was added via syringe. The
mixture was stirred for 3 h at 0 °C, and then tributylamine
(0.060 mL, 0.25 mmol), tributylammonium pyrophosphate
(0.25 mmol, 90.8 mg), and acetonitrile (stored over sieves) (0.25
mL) were added. The mixture was stirred for an additional
30 min at 0 °C, the sealed vial was then opened, and the
reaction was quenched by addition of TEAB (1 M) (0.5 mL)
and water (5 mL). The reaction mixture was purified on anion
exchange HPLC and desalted by reverse-phase HPLC as
described below. Mass and purity were confirmed by LC-MS
as described vide supra. Typical yields following this procedure
ranged from 5 to 15 mg of triphosphate.
Gen er a l P r oced u r e for P u r ifica tion a n d QC of 5′-
Tr ip h osp h a te Der iva tives. The triphosphate derivatives
were purified by anion exchange (AX) chromatography using
a 30 mm × 100 mm Mono Q column (Pharmacia) with a buffer
system of 50 mM Tris, pH 8. Elution gradients were typically
from 40 mM NaCl to 0.8 M NaCl in two column volumes at
6.5 mL/min. Appropriate fractions from anion exchange chro-
matography were collected and desalted by reverse-phase (RP)
chromatography using a Luna C18 250 mm × 21 mm column
(Phenomenex) with a flow rate of 10 mL/min. Elution gradients
were generally from 1% to 95% methanol in 14 min at a
constant concentration of 50 mM triethylammonium acetate
(TEAA), pH 7. Mass spectra of the purified triphosphates were
determined using on-line HPLC mass spectrometry on a
Hewlett-Packard (Palo Alto, CA) MSD 1100. A Phenomenex
Luna (C18(2)), 150 mm × 2 mm, plus 30 mm × 2 mm guard
column, 3 mm particle size, was used for RP HPLC. A 0-50%
linear gradient (15 min) of acetonitrile in 20 mM TEAA, pH
7, was performed in series with mass spectral detection in the
negative ionization mode. Nitrogen gas and a pneumatic
nebulizer were used to generate the electrospray. The mass
range of 150-900 was sampled. Molecular masses were
determined using the HP Chemstation analysis package. The
purity of the purified triphosphates was determined by
analytical RP and AX HPLC. RP HPLC with a Phenomonex
Luna or J upiter column (250 mm × 4.6 mm), 5 mm particle
size, was typically run with a 2-70% acetonitrile gradient in
15 min in 100 mM TEAA, pH 7. AX HPLC was performed on
a 5 mm × 16 mm Mono Q column (Pharmacia). Triphosphates
were eluted with a gradient of 0-0.4 M NaCl at constant
concentration of 50 mM Tris, pH 8. The purity of the triph-
osphates was generally >90%.
Assa y for In h ibition of HCV RNA Rep lica tion in Cells.
Inhibition of HCV RNA replication was assayed in a subge-
nomic bicistronic replicon assay in HB-1 cells by an in situ
ribonuclease protection assay (RPA), as previously described
at 24 h.14 The replicon EC50 values are the average of at least
two separate potency determination (Tables 1-4). The cyto-
toxicity of the nucleosides was determined by MTS at 24 h, as
described in ref 14, and the CC50 values were >50 µM (data
not shown).
Assa y for In h ibition of NS5B-Med ia ted RNA Syn th esis
in Vitr o. Inhibition of the enzymatic activity of HCV NS5B∆21
by the corresponding nucleoside triphosphates was determined
as previously described.14 The IC50 values for all active
compounds are the averages of at least three separate deter-
minations.
Assa y for P NP -Med ia ted P h osp h or olysis. Assay reac-
tions (100 mM HEPES, pH 7.0, 50 mM sodium phosphate, pH
7.0, 200 µM nucleoside) were initiated with addition of 1.8 or
18 U/mL PNP (Sigma N-3514, isolated from human blood).
Reaction mixtures were incubated at 4 °C for 27 h. RP-HPLC
analysis utilized a C18 column (Vydac C218TP, 150 × 4.6 mm,
5 µM) and guard column (Perkin-Elmer 0711-0092, 15 mm ×
3.2 mm, 7 µM) with 10 mM potassium phosphate, 2 mM TBA
(buffer A) and 50% methanol, 10 mM potassium phosphate, 2
harbored in hepatoma cells. While the SAR of the 2′-
and 3′-ribonucleoside triphosphate positions appears to
be quite stringent, some flexibility exists for modifica-
tion of the purine heterobase moiety. Our findings
indicate that selectivity against heterobase modifica-
tions in the cellular assay rely in part on specificity
provided by uptake mechanisms and/or kinase activity.
Thus, the use of heterobases that have previously
demonstrated efficient uptake and conversion to the
triphosphate in a ribonucleoside context might prove
advantageous in the future design of more potent
inhibitors of HCV RNA replication. While modifications
of the heterobase in ribonucleosides might be expected
to lead to mutagenic events, the 2′-C-methyl pharma-
cophore appears to prevent recognition by human poly-
merases,14 which may permit the use of nucleobases
otherwise considered unsuitable. Obviously, the success
of such an approach would rely on the design of
completely metabolically stable 2′-C-methylribonucleo-
side analogues; any ability to function as substrate for
ribonucleotide reductase and/or purine nucleoside phos-
phorylase would open the door for transglycosylation
and subsequent incorporation of modified purine nu-
cleosides into the host genome. Additionally, the enzy-
matic studies presented herein suggest that the oral
bioavailability of 2′-C-methyladenosine (11) may be
limited by its conversion to the corresponding inosine
derivative assisted by adenosine deaminase (ADA) as
well as by PNP-mediated lability of the adenine het-
erobase. Prodrug approaches may provide one way to
minimize conversion by ADA to a level that would
permit oral dosing of this compound. More intriguing,
however, is the existence of heterobase modifications
that are known to be resistant to ADA activity.30 The
use of such heterobase modifications may pave the road
for 2′-C-methyladenosine derivatives displaying greater
potency, improved metabolic profile, and significantly
enhanced oral bioavailability relative to 2′-C-methylad-
enosine.
Exp er im en ta l Section
Gen er a l Meth od s. The following chemicals were used as
received: 3′-deoxyadenosine (1) (Sigma D3394), 3′-deoxygua-
nosine (2) (Sigma D7285), 2′-O-methyladenosine (5) (Sigma
M9886), 2′-O-methylguanosine (6) (Yamasa Biochemicals
8189), arabinofuranosyladenine (9) (Sigma A5762), arabino-
furanosylguanine (10) (ICN 198924), 2-amino-6-chloro-
purine (Aldrich, 34,230-0), 6-chloropurine (Aldrich, 51,161-7),
3,5-bis-O-(2,4-dichlorophenylmethyl)-1-O-methyl-R-D-ribofura-
nose (Lipomed NUC-55), and tributylammonium pyrophos-
phate (Sigma P 8533). TLC was performed on silica 60 (Merck
5554 aluminum sheet), and column chromatograpy was per-
formed on silica 60 (230-400 mesh ASTM) (Merck 9385). The
4 Å molecular sieves (Mallinckrodt 4494) were activated prior
to use (120 °C in vacuo). 1H and 13C NMR spectra were
obtained at 200 MHz (Varian Mercury VX) in 5 mm tubes
unless otherwise indicated. Chemical shifts are positive in the
low-field direction. FAB mass spectra were recorded on a J EOL
Hx110/110 mass spectrometer. All nucleosides screened in the
call-based assay were purified by preparative RP HPLC using
a reverse-phase column (Phenomenex, Luna 10 µm, 21.2 mm
× 250 mm, A ) water, B ) acetonitrile, 2-95% B in 60 min,
flow 3 mL min-1). Nucleosides were analyzed for purity by
analytical RP HPLC (Phenominex C18 aqua 5 µm, 150 mm ×
4.6 mm column, A ) 100 mM triethylammonium acetate, pH
7, B ) 10% 100 mM triethylammonium acetate in acetonitrile,
100% B in 25 min, flow 1 mL min-1). The purity of the
nucleosides screened was generally greater than 98%.

2292 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9
Eldrup et al.
mM TBA (buffer B), 1 mL/min flow rate. A 50 µL reaction
volume injection and elution over the gradient yielded quan-
tifiable nucleoside peaks (1-5% B over 15 min; 5-75% B over
2.5 min; hold 75% for 1 min); monitoring was at 260 nm.
adenosine 1623 (20.0 g, 39.2 mmol) was added, and the mixture
was stirred at 0 °C for another 30 min. The mixture was
diluted with ethyl acetate (1000 mL) and filtered through a
pad of silica gel (200 g). The filtrate was washed with saturated
aqueous sodium bicarbonate (2 × 200 mL), dried over mag-
nesium sulfate, and evaporated in vacuo to a solid (20.5 g). To
a solution of this keto derivative in dry THF (100 mL) was
added methylmagnesium bromide (150 mmol in 250 mL of
diethyl ether) at -78 °C. The mixture was stirred at -50 to
-30 °C for 4 h, and acetone (20 mL) was then added to quench
the reaction. The mixture was washed with saturated aqueous
ammonium chloride (200 mL), dried over sodium sulfate, and
evaporated in vacuo to a solid (15.0 g). This solid (1.00 g, 1.90
mmol) was dissolved in dichloromethane/acetonitrile (20 mL,
1:1), and triethylamine (1 mL) and Et3N‚3HF (1 mL) were
added. The mixture was stirred at room temperature for 3 h
and then evaporated in vacuo. The residue was purified on
silica gel using dichloromethane/acetone/methanol (20:5:2) as
the eluent. Fractions containing the product were pooled and
evaporated in vacuo to give the desired product (0.15 g, 21%
Assa y for ADA-Med ia ted Con ver sion . Assay reactions
included 5 mM Tris-HCl, pH 7.4, 5 mM potassium phosphate,
0.00625 or 1.25 U/mL adenosine deaminase (ADA, Sigma
A-5168, type IX from calf spleen). Reactions were initiated with
addition of 25 µM nucleoside. Reaction mixtures were incu-
bated at room temperature for 30 min (0.006 25 U/mL ADA)
or 24 h (1.25 U/mL ADA) before the enzyme was heat-
inactivated (75 °C for 15 min). RP-HPLC analysis utilized a
SUPELCOSIL LC-18-S column (Supelco 58931, 150 mm × 4.6
mm, 5 µM) with 97.5%/2.5% 50 mM potassium phosphate pH
4.4/methanol (buffer A) and 80%/20% 50 mM potassium
phosphate pH 4.4/methanol (buffer B), 1 mL/min flow rate. A
25 µL reaction volume injection and elution over the gradient
yielded quantifiable nucleoside peaks (0% B hold for 3 min,
0-100% B over 9 min, 100% hold for 3 min, 100-60% over
4.5 min); monitoring was at 254 and 280 nm. Injections of
adenosine and inosine yielded linear standard curves allowing
for the quantitation of substrate utilization and product
formation.
1
for three steps). H NMR (DMSO-d6): δ 8.19 (s, 1H), 8.12 (s,
1H), 7.21 (s br, 2H), 6.15 (d, 1H, J ) 4.8 Hz), 5.93 (s, 1H), 5.63
(s, 1H), 5.36 (br, 1H), 3.90 (m, 1H), 3.81 (m, 1H), 3.65 (m, 2H),
1.13 (s, 3H). HRMS (FAB): calcd for
282.1202, found 282.1208.
C
11H15N5O4 + H+
Ca lcu la tion s. Nucleoside conformational preferences were
calculated by molecular mechanics using the MMFFs force
field and a dielectric constant, ꢀ, of 50. For each nucleoside,
1000 conformers were generated using the J G distance geom-
etry program31 and minimized to low gradient using Batch-
Min.32
9-(2-O-Acetyl-3,5-bis-O-ben zyl-3-C-m eth yl-â-D-r ibofu r a -
n osyl)-6-ch lor op u r in e (21). To a precooled (0 °C) mixture
of 2024 (2.86 g, 6.67 mmol), 6-chloropurine (1.24 g, 8.00 mmol),
and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) (3.00 mL, 20
mmol) in anhydrous acetonitrile (27 mL) was added trimeth-
ylsilyl triflate (4.82 mL, 26.7 mmol). The mixture was heated
at 60 °C for 3 h and subsequently quenched by dropwise
addition of saturated aqueous sodium bicarbonate. The mix-
ture was diluted with ethyl acetate (400 mL), washed with
saturated aqueous sodium bicarbonate (2 × 300 mL), dried
over sodium sulfate, and evaporated in vacuo. The crude
product was purified on silica using hexane/ethyl acetate (3:
1) to give the desired compound (1.60 g, 46%) as a slightly tan
foam. 1H NMR (CDCl3): δ 8.71 (s, 1H), 8.55 (s, 1H), 7.40-
7.26 (m, 10H), 6.51 (d, 1H), 5.99 (d, 1H), 4.73-4.56 (m, 4H),
4.43 (m, 1H), 3.81 (dd, 1H,), 3.59 (dd, 1H), 2.04 (s, 3H), 1.51
(s, 3H). MS (ESI): calcd for C27H27N4O5 + Na+ 545.1568, found
545.0.
2-Am in o-6-ch lor o-9-(2-C-m eth yl-2,3,5-tr i-O-ben zoyl-â-
D-r ibofu r a n osyl)p u r in e (14). To a precooled (0 °C) solution
of 13 (1.00 g, 1.72 mmol), 2-amino-6-chloropurine (0.32 g, 1.87
mmol), and 1,8-diazabicycl[5.4.0]undec-7-ene (DBU) (0.77 mL,
5.10 mmol) in anhydrous acetonitrile (20 mL) was added
trimethylsilyl triflate (1.25 mL, 6.88 mmol) dropwise. The
reaction mixture was heated at 65 °C for 4 h, allowed to come
to room temperature, poured into saturated aqueous sodium
bicarbonate (150 mL), and extracted with dichloromethane (3
× 100 mL). The combined organic phase was dried over sodium
sulfate and evaporated in vacuo. The residue was purified over
silica gel using hexane/ethyl acetate (7:3) as the eluent to give
the desired compound (0.68 g, 63%) as a colorless foam. 1H
NMR (CDCl3): δ 8.15-7.30 (m, 16H), 6.65 (s, 1H), 6.40 (d, J
) 6.6 Hz, 1H), 5.35 (s, 2H), 5.08 (m, 1H), 4.84-4.73 (m, 2H),
1.60 (s, 3H). MS (ESI): calcd for C32H26ClN5O7 + H+ 628.1599,
found 628.0.
9-(3,5-Bis-O-b en zyl-3-C-m et h yl-â-D-r ib ofu r a n osyl)a d -
en in e (22). A solution of 21 (680 mg, 1.30 mmol) in a 1:1
mixture of anhydrous 1,4-dioxane and liquid ammonia (30 mL)
was sealed in a stainless steel reaction vessel and heated at
80 °C for 14 h. After cooling, the vessel was opened and the
solvents were allowed to evaporate. The resulting residue was
dissolved in dichloromethane (200 mL), washed with brine (2
× 200 mL), dried over magnesium sulfate, and evaporated in
vacuo. The crude material was purified on silica gel using
dichloromethane/methanol (19:1) to give the desired compound
2-Am in o-6-ch lor o-9-(2-C-m eth yl-â-D-r ibofu r a n osyl)p u -
r in e (15). A solution of 14 (0.65 g, 1.03 mmol) in saturated
methanolic ammonia (15 mL) was stirred in a sealed container
for 7 h. The solvent was removed, and the residue was purified
on silica gel using methanol/dichloromethane (1:9) as the
eluent. Fractions containing the product were pooled and
evaporated in vacuo to give the desired compound (0.30 g, 92%)
1
1
(0.58 g, 97%) as a colorless foam. H NMR (CDCl3): δ 8.22 (s,
as a colorless powder. H NMR (methanol-d4): δ 8.56 (s, 1H),
1H), 8.03 (s, 1H), 7.40-7.26 (m, 10H), 6.21 (s br, 2H), 6.14 (d,
1H), 4.65-4.53 (m, 5H), 4.44 (m, 1H), 3.78 (dd, 1H), 3.59 (dd,
1H), 1.56 (s, 3H). MS (ESI): calcd for C25H27N5O4 + H+
462.2141, found 462.1.
6.01 (s, 1H), 4.20 (d, J ) 8.8 Hz, 1H), 4.07-3.98 (m, 2H), 3.85
(dd, J ) 12.6 and 3.0 Hz, 1H), 0.97 (s, 3H). MS (ESI): calcd
for C11H14ClN5O7 + H+ 316.0813, found 315.9.
2′-C-Meth ylgu a n osin e (12). To a solution of 15 (0.20 g,
0.63 mmol) and 2-mercaptoethanol (0.22 mL, 3.15 mmol) in
methanol (20 mL) was added sodium methoxide (0.18 g, 3.15
mmol), and the mixture was heated under reflux for 7 h. The
solution was cooled to room temperature and neutralized with
acetic acid. The solvent was evaporated in vacuo and the
residue was purified on silica gel using dichloromethane/
methanol (17:3) as the eluent to give the desired compound
(0.11 g, 59%) as a colorless solid. 1H NMR (DMSO-d6): δ 10.94
(bs, 1H), 8.04 (s, 1H), 6.72 (s, 2H), 5.73 (s, 1H), 4.00-3.62 (m,
4H), 0.81 (s, 3H). HRMS (FAB): calcd for C11H15N5O5 + H+
298.1151, found 298.1149.
6-Am in o-9-(2-C-m et h yl-â-D-a r a b in ofu r a n osyl)p u r in e
(19). A mixture of chromium(VI) oxide (10.0 g, 100 mmol),
pyridine (16.2 mL, 200 mmol), and acetic anhydride (9.50 mL,
100 mmol) in dichloromethane (250 mL) was stirred at 0 °C
for 30 min. 3′,5′-O-(1,1,3,3-Tetraisopropyldisiloxane-1,3-diyl)-
9-(3-C-Meth yl-â-D-r ibofu r a n osyl)a d en in e (23). A mix-
ture of 22 (0.29 g, 0.62 mmol) and ammonium formate (312
mg, 4.94 mmol) in methanol (10 mL) was sparged with argon
for 10 min and then treated with 10% Pd/C (50 mg). The
mixture was heated at reflux for 12 h, cooled to room
temperature, and filtered through a pad of Celite. The crude
material was purified on silica gel using dichloromethane/
methanol (1:9), evaporated in vacuo to a pale glass, and
lyophilized from water (2 × 10 mL) to give the desired
compound (61 mg, 35%) as a colorless solid. 1H NMR (DMSO-
d6): δ 8.34 (s, 1H), 8.12 (s, 1H), 7.40 (s br, 2H), 5.83 (d, 1H),
4.45 (d, 1H), 3.89 (m, 1H), 3.70-3.48 (m, 2H), 1.31 (s, 3H).
HRMS (FAB): calcd for C11H15N5O4 + H+ 282.1202, found
282.1197.
1-O-Meth yl-2-C-m eth yl-2-O-m eth yl-3,5-bis-O-(2,4-dich lo-
r op h en ylm eth yl)-R-D-r ibofu r a n ose (26). To a solution of

Purine Ribonucleosides
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9 2293
2526,27 (2.00 g, 4.03 mmol) and 18-crown-6 (200 mg) in dry THF
(10 mL) was added powdered KOH (1.13 g, 20.2 mmol). The
resulting mixture was stirred at room temperature for 60 min,
and methyl iodide (0.86 g, 6.05 mmol) was added. The reaction
mixture was stirred for another 3 h, the solids were filtered
off, and the filtrate was concentrated in vacuo. The crude
product was purified on silica gel using hexane/ethyl acetate
(2:1) as the eluent to give the desired compound (2.00 g, 97%)
of Dess-Martin periodinane (10.1 g, 23.0 mmol) in anhydrous
dichloromethane (50 mL) was added a solution of 30 (3.70 g,
7.60 mmol) in dichloromethane (50 mL) dropwise. The reaction
mixture was then stirred at room temperature for 48 h. The
mixture was diluted with diethyl ether (200 mL) and poured
into a chilled solution of Na2S2O3‚5H2O (18 g) in saturated
aqueous sodium bicarbonate (150 mL). The organic phase was
washed with saturated aqueous sodium bicarbonate (100 mL),
then water (150 mL), and then brine (150 mL). The organic
phase was dried over sodium sulfate and evaporated in vacuo
to give the keto derivative, which was used for the next
reaction without purification. To diethyl ether (120 mL) at -78
°C was slowly added EtMgBr (3.0 M, 12.6 mL) followed by the
keto derivative (obtained vide infra) in diethyl ether (75 mL).
The reaction mixture was stirred at -78 °C for 30 min and
then allowed to come to -15 °C and stirred for 4 h. The
reaction mixture was poured into saturated aqueous am-
monium chloride. The organic phase was washed with brine,
dried over sodium sulfate and evaporated in vacuo. The residue
was purified over silica gel using hexane/ethyl acetate (9:1)
1
as an oil. H NMR (CDCl3): δ 7.50-7.16 (m, 6 H), 4.82 (d, 1
H, J ) 13.8 Hz), 4.66-4.51 (m, 4 H), 4.26 (q, 1 H), 3.65 (dd, 1
H), 3.54 (d, 2 H), 3.46 (s, 3 H), 3.40 (s, 3 H), 1.40 (s, 3 H). MS
(ESI) calcd for C22H24Cl4O5 + H+ 509.0456, found 509.5.
1-O-Meth yl-2-C-m eth yl-2-O-m eth yl-3,5-d i-O-a cetyl-R-D-
r ibofu r a n ose (27). A solution of compound 26 (2.00 g, 3.92
mmol) and Pd/C (10%) in a mixture of dichloromethane/
methanol/acetic acid (15 mL, 1:1:1) was stirred under hydrogen
at 1 atm overnight. The solids were filtered off, and the filtrate
was concentrated in vacuo. The crude compound was purified
on silica using dichloromethane/methanol (10:1) to give the
desired compound (0.65 g, 75%). This compound (0.16 g) was
treated with acetic anhydride/pyridine (4 mL, 1:1) in the
presence of 4-(dimethylamino)pyridine (2 mg) at room tem-
perature overnight. The reaction mixture was quenched with
methanol and concentrated in vacuo, and the crude material
was purified on silica gel using hexane/ethyl acetate (1:1) to
1
as the eluent to give the desired compound (1.30 g, 34%). H
NMR (CDCl3): δ 7.46-7.17 (m, 6H), 4.82-4.55 (m, 5H), 4.15
(q, J ) 4.4, 8.8 Hz, 1H), 3.64 (d, J ) 4.4 Hz, 2H), 3.45 (s, 3H),
3.26 (s, 1H), 1.62 (m, 2H), 0.99 (t, J ) 7.2 Hz, 3H). MS (FAB):
calcd for C22H24Cl4O5 + Li+ 517.17, found 517.10
give the desired compound as
a
colorless oil. 1H NMR
1-O-Meth yl-2-C-eth yl-2-O-ben zoyl-R-D-r ibofu r a n osid e
(32). To a solution of 4-(dimethylamino)pyridine (DMAP) (0.34
g, 2.80 mmol) and triethylamine (4.40 mL) in dichloromethane
(10 mL) was added benzoyl chloride (0.64 mL, 5.50 mmol)
followed by 31 (0.65 g, 1.20 mmol) in dichloromethane (10 mL).
The reaction mixture was heated at 50 °C for 48 h. The
solution was cooled to room temperature and poured into a
solution of diethyl ether (150 mL) and saturated aqueous
sodium bicarbonate (150 mL). The organic phase was washed
with aqueous sodium bicarbonate (2 × 100 mL) and then with
brine (100 mL), dried over sodium sulfate, and evaporated in
vacuo. The crude product was purified on silica gel using ethyl
acetate/hexane (1:9) as the eluent to give the desired compound
(0.70 g, 95%). A suspension of this benzoate derivative and
Pd/C (10%) (0.70 g) in methanol (5 mL) and acetic acid (2 mL)
was stirred under hydrogen at 1 atm overnight. The solution
was filtered, evaporated in vacuo, and coevaporated from
toluene. The residue was purified over silica gel using ethyl
acetate/hexane (3:7) as the eluent to give the desired compound
(0.11 g, 32%) as a colorless oil. 1H NMR (CDCl3): δ 8.07-8.02
(m, 2H), 7.61-7.44 (m, 3H), 4.92 (d, J ) 4.8 Hz, 1H), 4.67 (s,
1H), 4.19 (q, J ) 4.6, 8.8 Hz, 1H), 3.88 (m, 2H), 3.49 (s, 3H),
3.11 (s, 1H), 2.31 (t, 1H), 1.74 (m, 2H), 1.03 (t, J ) 7.2 Hz,
3H). MS (FAB): calcd for C15H21O6 + Li+ 303.14, found 303.20.
(CDCl3): δ 4.80 (d, J ) 4.8 Hz, 1H), 4.64 (s, 1H), 4.50-4.40
(m, 1H), 4.30-4.20 (m, 2H), 3.42 (s, 3H), 3.35 (s, 3H), 2.20 (s,
3H), 2.05 (s, 3H), 1.40 (s, 3H). MS (ESI) calcd for C12H20O7 +
H+ 299.1107, found 299.1.
1-O-Acetyl-2-C-m eth yl-2-O-m eth yl-3,5-d i-O-a cetyl-D-r i-
bofu r a n ose (28). To a precooled (0 °C) solution of compound
27 (0.23 g, 0.74 mmol) in anhydrous acetic anhydride (3.12
mL) was added concentrated sulfuric acid (0.071 mL) in
anhydrous acetic anhydride (3.12 mL) dropwise. The reaction
mixture was stirred at 0 °C for 1 h. The mixture was
neutralized with cold saturated aqueous sodium bicarbonate
solution and extracted with ethyl acetate (3 × 50 mL). The
organic phase was washed with saturated aqueous sodium
bicarbonate (2 × 100 mL) and then with brine (100 mL), dried
over magnesium sulfate, and evaporated in vacuo. The crude
product was purified on silica gel using hexane/ethyl acetate
(4:1) as the eluent to give the desired compound (0.22 g, 98%)
as an anomeric mixture. 1H NMR (CDCl3): δ 6.99 (s, 1H), 6.20
(s, 1H), 5.54-5.47 (m, 1H), 5.36 (d, J ) 4.0 Hz, 1H), 5.19 (d, J
) 6.7 Hz, 1H), 4.54 (dd, 1H), 4.40-4.07 (m, 4H), 3.38 (s, 3H),
3.35 (s, 3H), 2.17 (s, 3H), 2.15 (s, 3H), 2.13 (s, 3H), 2.08 (s,
3H), 2.06 (s, 3H), 2.04 (s, 3H), 1.30 (s, 3H), 1.27 (s, 3H). MS
(ESI) calcd for C13H20O8 + H+ 327.1056, found 327.2.
2′-C-Meth yl-2′-O-m eth yla d en osin e (29). To a precooled
(0 °C) mixture of 28 (0.25 g, 0.76 mmol), 6-chloropurine (0.11
g, 0.70 mmol), and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU)
(0.29 mL) in anhydrous acetonitrile (6.0 mL) was added
trimethylsilyl triflate (0.42 mL, 2.18 mmol). The reaction
mixture was then stirred for 4 h at 65 °C, evaporated in vacuo,
and purified on silica using hexane/ethyl acetate (1:1) as the
eluent to give the desired intermediate (0.17 g, 56%) as an
anomeric mixture. This compound was treated with methan-
olic ammonia (4 mL) at room temperature overnight. The
reaction mixture was evaporated in vacuo and purified on silica
gel using dichloromethane/methanol (5:1) as the eluent to give
the desired compound 29 (7.5 mg), the R anomer (3.5 mg), and
a mixture of compound 29 and the R anomer (21 mg) in 32%
total yield as white solids. The overall R/â anomeric ratio was
2:1.
1-O-Met h yl-2-C-et h yl-2,5-d i-O-b en zoyl-R-D-r ib ofu r a -
n ose (33). To a solution of 4-(dimethylamino)pyridine (DMAP)
(0.10 g, 0.82 mmol) and triethylamine (0.80 mL) in dichlo-
romethane (5 mL) was added benzoyl chloride (0.26 mL, 2.20
mmol) followed by 32 in dichloromethane (5 mL). The reaction
mixture was stirred at room temperature for 3 h. This solution
was poured into a mixture of diethyl ether (100 mL) and
saturated aqueous sodium bicarbonate (100 mL). The organic
phase was washed with aqueous sodium bicarbonate (100 mL)
and brine (100 mL), dried over sodium sulfate, and evaporated
in vacuo. The residue was purified on silica gel using hexane/
ethyl acetate (4:1) as the eluent to give the desired compound
1
(0.16 g, 94%). H NMR (CDCl3): δ 8.09-8.01 (m, 4H), 7.60-
7.37 (m, 6H), 5.11 (d, J ) 4.6 Hz, 1H), 4.75-4.45 (m, 5H), 3.52
(s, 3H), 1.78 (m, 2H), 1.05 (t, J ) 7.4 Hz, 3H). MS (FAB): calcd
for C22H25O7 + Li+ 407.17, found 407.20
1
R-An om er . H NMR (methanol-d4): δ 8.26 (s, 1H), 8.18 (s,
1H), 6.18 (s, 1H), 4.30-4.18 (m, 1H), 4.06 (d, 1H), 3.86 (dd,
1H), 3.65 (dd, 1H), 3.15 (s, 3H), 1.56 (s, 3H). MS (ESI): calcd
for C24H34N10O8 + Na+ (2M + Na+) 613.2459, found 612.9.
1-O-Acet yl-2-C-et h yl-2,5-d i-O-b en zoyl-3-O-a cet yl-D-r i-
bofu r a n ose (34). To a solution of the dibenzoyl derivative 33
in acetic acid (5 mL) were added acetic anhydride (1.0 mL)
and sulfuric acid (0.2 mL) at 0 °C. The reaction mixture was
stirred at room temperature overnight. The solution was
poured into ice and extracted with dichloromethane (3 × 40
mL). The combined organic phase was washed with saturated
aqueous sodium bicarbonate (2 × 50 mL), dried over sodium
sulfate, and evaporated in vacuo. The residue was purified over
1
â-An om er . H NMR (methanol-d4): δ 8.59 (s, 1H), 8.20 (s,
1H), 6.27 (s, 1H), 4.21 (d, 1 H), 4.07-3.99 (m, 2 H), 3.89-3.82
(m, 1H), 3.54 (s, 3H), 0.95 (s, 3H). HRMS (FAB): calcd for
C
12H17N5O4 + H+ 296.1359, found 296.1359.
1-O-Met h yl-2-C-et h yl-3,5-b is-(2,4-d ich lor op h en ylm e-
th yl)-R-D-r ibofu r a n ose (31). To a precooled (0 °C) suspension

2294 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9
Eldrup et al.
silica gel using hexane/ethyl acetate (3:1) as the eluent to give
the desired compound (0.085 g, 51%) as an anomeric mixture.
1H NMR (CDCl3): δ 8.09-7.34 (m, 10 H), 6.61 (s, 1H), 6.51 (s,
1H), 5.80 (d, J ) 7.8 Hz, 1H), 5.51 (d, J ) 3.2 Hz, 1H), 4.73-
4.38 (m, 6H), 2.41-2.02 (m, 4H), 2.19 (s, 3H), 2.16 (s, 3H), 1.99
(s, 1H), 1.92 (s, 3H). LRMS (FAB): calcd for C25H27O9 + Li+
477.17, found 477.20
6.58 (s br, 1H), 5.25 (d, 1H), 5.16 (t, 1H), 5.16 (s, 1H), 3.60-
4.06 (overlapping m, 4H), 0.81 (s, 3H). HRMS (FAB): calcd
for C11H15N5O4 + H+ 282.1202, found 282.1199.
Ap p en d ix
Abbr evia tion s. ADA, adenosine deaminase; DBU,
1,8-diazabicyclo[5.4.0]undec-7-ene; DMAP, 4-(dimethyl-
amino)pyridine; HCV, hepatitis C virus; NS5B, non-
structural protein 5B; RdRp, RNA-dependent RNA
polymerase; NNI, non-nucleoside inhibitor; NI, nucleo-
side inhibitor; HIV, human immunodeficiency virus;
NTP, nucleoside triphosphate, AX, anion exchange; RP,
reverse phase; HPLC, high-performance liquid chroma-
tography; RPA, ribonuclease protection assay; PNP,
purine nucleoside phosphorylase; TIPDS dichloride, 1,3-
dichloro-1,1,3,3,-tetraisopropyldisiloxane.
6-Am in o-9-(2-C-eth yl-â-D-r ibofu r a n osyl)p u r in e (36). To
precooled (0 °C) solution of 34 (0.085 g, 0.18 mmol),
a
6-chloropurine (0.042 g, 0.27 mmol), and 1,8-diazabicyclo[5.4.0]-
undec-7-ene (DBU) (0.08 mL, 0.55 mmol) in acetonitrile (8 mL)
was added trimethylsilyl triflate (0.13 mL, 0.76 mmol). The
reaction mixture was stirred at 50 °C for 1.5 h. The solution
was allowed to come to room temperature, poured into
saturated aqueous sodium bicarbonate (100 mL), and extracted
with ethyl acetate (100 mL). The organic phase was washed
with aqueous sodium bicarbonate (2 × 50 mL) and brine (100
mL), dried over sodium sulfate, and evaporated in vacuo. The
residue was purified on silica gel using hexane/ethyl acetate
(7:3) to give the desired compound as a colorless powder. A
solution of this compound (35) in dioxane (2 mL) and liquid
ammonia (8 mL) was heated in a Parr bomb at 60 °C overnight.
The solution was evaporated to dryness and the residue was
purified on silica gel using dichloromethane/methanol (9:1) as
the eluent to give the desired compound (13.0 mg, 16%). 1H
NMR (methanol-d4): δ 8.48 (s, 1H), 8.19 (s, 1H), 6.12 (s, 1H),
4.38 (d, J ) 8.8 Hz, 1H), 4.08-3.96 (m, 2H), 3.85 (dd, J ) 3.2,
12.6 Hz, 1H), 1.44-1.06 (m, 2H), 0.78 (t, J ) 7.4 Hz, 3H).
HRMS (FAB): calcd for C12H17N5O4 296.1358, found 296.1355.
2′-C-Meth ylin osin e (38). A solution of 37 (0.02 g, 0.067
mmol) in 1 N aqueous sodium hydroxide (2 mL) was heated
at reflux for 2 h. The mixture was cooled, neutralized with
HCl (1 N, aqueous), and evaporated in vacuo. The residue was
purified on silica gel using methanol/dichloromethane (1:9
through 1:4) as the eluent. Fractions containing the desired
compound were pooled and evaporated in vacuo to give the
desired product (8.0 mg, 42.3%) as a colorless powder after
freeze-drying. 1H NMR (DMSO-d6): δ 12.15 (s br, 1H), 8.40
(s, 1H), 8.01 (s, 1H), 5.85 (s, 1H), 5.15-5.21 (overlapping t, d,
and s, 3H), 3.65-4.13 (overlapping m, 4H), 0.71 (s, 3H).
(2-C-Meth yl-â-D-r ibofu r a n osyl)p u r in e (39). A mixture of
37 (0.020 g, 0.067 mmol), triethylamine (0.1 mL), and Pd (10%
on carbon) (0.01 g) in methanol (5.0 mL) was stirred under
hydrogen overnight. The mixture was filtered through Celite,
evaporated in vacuo, and purified on silica gel using methanol/
dichloromethane (1:9) as the eluent. Fractions containing the
desired compound were pooled and evaporated in vacuo to give
the desired product (12 mg, 67%) as a colorless powder after
freeze-drying. 1H NMR (acetonitrile-d3): δ 9.10 (s, 1H), 8.93
(s, 1H), 8.70 (s, 1H), 6.19 (s, 1H), 4.01 (m, 2H), 3.66-4.02
(overlapping m, 1H), 3.83 (dd, J ) 13.0 and 3.4 Hz, 1H), 0.88
(s, 3H). HRMS (FAB): calcd for C10H15N4O4 + H+ 268.1126,
found 268.1129.
2,6-Dia m in o-9-(2-C-m et h yl-â-D-r ib ofu r a n osyl)p u r in e
(40). To 15 (0.10 g, 0.32 mmol) was added ammonium
hydroxide (5 mL), and the resulting mixture was heated to 80
°C overnight in a sealed container. The mixture was evapo-
rated in vacuo and purified on silica using methanol/dichlo-
romethane (1:9 through 1:4) as the eluent. Fractions contain-
ing the desired product were pooled and evaporated in vacuo
to give the desired product (0.054 g, 57.5%) as a colorless
powder after freeze-drying. 1H NMR (methanol-d4): δ 8.13 (s,
1H), 5.91 (s, 1H), 4.19 (d, J ) 8.8 Hz, 1H), 4.01 (m, 2H), 3.83
(dd, J ) 13.0 and 3.4 Hz, 1H), 0.94 (s, 3H). HRMS (FAB): calcd
for C11H17N6O4 + H+ 297.1311, found 297.1305.
Refer en ces
(1) Hwang, S. B.; Park, K.-J .; Kim, Y.-S.; Sung, Y. C.; Lai, M. M. C.
Hepatitis
C virus NS5B protein is a membrane-associated
phosphoprotein with a predominantly perinuclear localization.
Virology 1997, 227, 439-446.
(2) Behrens, S.-E.; Tomei, L.; De Francesco, R. Identification and
properties of the RNA-dependent RNA polymerase of hepatitis
C virus. EMBO J . 1996, 15, 12-22.
(3) Takamizawa, A.; Mori, C.; Fuke, I.; Manabe, S.; Murakami, S.;
Fujita, J .; Onishi, E.; Andoh, T.; Yoshida, I.; Okayama, H.
Structure and organization of the hepatitis C virus genome
isolated from human carriers. J . Virol. 1991, 65, 1105-1113.
(4) Colarusso, S.; Gerlach, B.; Koch, U.; Muraglia, E.; Conte, I.;
Stansfield, I.; Matassa, V. G.; Narjes, F. Evolution, synthesis
and SAR of tripeptide R-ketoacid inhibitors of the hepatitis C
virus NS3/NS4A serine protease. Bioorg. Med. Chem. Lett. 2002,
12, 705-708.
(5) Narjes, F.; Koehler, K. F.; Koch, U.; Gerlach, B.; Colarusso, S.;
Steinkuhler, C.; Brunetti, M.; Altamura, S.; De Francesco, R.;
Matassa, V. G. A designed P1 cysteine mimetic for covalent and
non-covalent inhibitors of HCV NS3 protease. Bioorg. Med.
Chem. Lett. 2002, 12, 701-704.
(6) Steinkuhler, C.; Koch, U.; Narjes, F.; Matassa, V. G. Hepatitis
C virus serine protease inhibitors: current progress and future
challenges. Curr. Med. Chem. 2001, 8, 919-932.
(7) Llinas-Brunet, M.; Bailey, M.; Fazal, G.; Goulet, S.; Halmos, T.;
Laplante, S.; Maurice, R.; Poirier, M.; Poupart, M.-A.; Thibeault,
D.; Wernic, D.; Lamarre, D. Peptide-based inhibitors of the
hepatitis C virus serine protease. Bioorg. Med. Chem. Lett. 1998,
8, 1713-1718.
(8) Llinas-Brunet, M.; Bailey, M.; Deziel, R.; Fazal, G.; Gorys, V.;
Goulet, S.; Halmos, T.; Maurice, R.; Poirier, M.; Poupart, M.-A.;
Rancourt, J .; Thibeault, D.; Wernic, D.; Lamarre, D. Studies on
the C-terminal of hexapeptide inhibitors of the hepatitis C virus
serine protease. Bioorg. Med. Chem. Lett. 1998, 8, 2719-2724.
(9) Lamarre, D.; Anderson, P. C.; Balley, M.; Beaulieu, P.; Bolger,
G.; Bonneau, P.; Bos, M.; Cameron, D. R.; Cartier, M,; Cordin-
gley, M. G.; Faucher A.-M.; Goudreau, N.; Kawai, S. H.; Kukolj,
G.; Lagace´, l.; Laplante, S. R.; Narjes, H.; Poupart, M.-A.;
Rancourt, G.; Sentjens, R. E.; St. George, R.; Simoneau, B.;
Stelnmann, G.; Thibeault, D.; Tsantrzos, Y. S.; Weldon, S. M.;
Yong, C.-L.; Llinas-Brunet, M. Nature 2003, 426, 186-189.
(10) Chan, L.; Reddy, T. J .; Proulx, M.; Das, S. K.; Pereira, O.; Wang,
W.; Siddiqui, A.; Yannopoulos, C. G.; Poisson, C.; Turcotte, N.;
Drouin, A.; Alaoui-Ismaili, M. H.; Bethell, R.; Hamel, M.;
L’Heureux, L.; Bilimoria, D.; Nguyen-Ba, N. Identification of
N,N-Disubstituted Phenylalanines as a Novel Class of Inhibitors
of Hepatitis C NS5B Polymerase. J . Med. Chem. 2003, 46, 1283-
1285.
(11) Wang, M.; Ng, K. K. S.; Cherney, M. M.; Chan, L.; Yannopoulos,
C. G.; Bedard, J .; Morin, N.; Nguyen-Ba, N.; Alaoui-Ismaili, M.
H.; Bethell, R. C.; J ames, M. N. G. Non-nucleoside Analogue
Inhibitors Bind to an Allosteric Site on HCV NS5B Polymerase.
J . Biol. Chem. 2003, 278, 9489-9495.
(12) Dhanak, D.; Duffy, K. J .; J ohnston, V. K.; Lin-Goerke, J .; Darcy,
M.; Shaw, A. N.; Gu, B.; Silverman, C.; Gates, A. T.; Nonnema-
cher, M. R.; Earnshaw, D. L.; Casper, D. J .; Kaura, A.; Baker,
A.; Greenwood, C.; Gutshall, L. L.; Maley, D.; DelVecchio, A.;
Macarron, R.; Hofmann, G. A.; Alnoah, Z.; Cheng, H.-Y.; Chan,
G.; Khandekar, S.; Keenan, R. M.; Sarisky, R. T. Identification
and Biological Characterization of Heterocyclic Inhibitors of the
Hepatitis C Virus RNA-Dependent RNA Polymerase. J . Biol.
Chem. 2002, 277, 38322-38327.
2-Am in o-9-(2-C-m eth yl-â-D-r ibofu r a n osyl)p u r in e (41).
A mixture of 15 (0.02 g, 0.064 mmol), triethylamine (0.1 mL),
and Pd/C (0.01 g) in methanol (5.0 mL) was stirred under
hydrogen overnight. The mixture was filtered through Celite,
evaporated in vacuo, and purified on silica gel using methanol/
dichloromethane (1:9) as the eluent. Fractions containing the
desired compound were pooled and evaporated in vacuo to give
the desired product (0.014 g, 78%) as a colorless powder after
1
freeze-drying. H NMR (DMSO-d6): 8.60 (s, 1H), 8.40 (s, 1H),

Purine Ribonucleosides
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 9 2295
(13) Gu, B.; J ohnston, V. K.; Gutshall, L. L.; Nguyen, T. T.; Gontarek,
R. R.; Darcy, M. G.; Tedesco, R.; Dhanak, D.; Duffy, K. J .; Kao,
C. C.; Sarisky, R. T. Arresting Initiation of Hepatitis C Virus
RNA Synthesis Using Heterocyclic Derivatives. J . Biol. Chem.
2003, 278, 16602-16607.
(14) Carroll, S. S.; Tomassini, J . E.; Bosserman, M.; Getty, K.;
Stahlhut, M. W.; Eldrup, A. B.; Bhat, B.; Hall, D.; Simcoe, A.
L.; LaFemina, R.; Rutkowski, C. A.; Wolanski, B.; Yang, Z.;
Migliaccio, G.; De Francesco, R.; Kuo, L. C.; MacCoss, M.; Olsen,
D. B. Inhibition of Hepatitis C Virus RNA Replication by 2′-
Modified Nucleoside Analogs. J . Biol. Chem. 2003, 278, 11979-
11984.
(15) Shim, J .; Larson, G.; Lai, V.; Naim, S.; Wu, J . Z. Canonical 3′-
deoxyribonucleotides as chainterminator for HCV NS5B RNA-
dependent RNA polymerase. Antiviral Res. 2003, 243-251.
(16) Migliaccio, G.; Tomassini, J . E.; Carroll, S. S.; Tomei, L.;
Altamura, S.; Bhat, B.; Bartholomew, L.; Bosserman, M.; Cec-
cacci, A.; Colwell, L. F.; Cortese, R.; De Francesco, R.; Eldrup,
A. B.; Getty, K.; Hou, W. S.; LaFemina, R.; Ludmere, S. W.;
MacCoss, M.; MacMasters, D. R.; Stahlhut, M. W.; Olsen, D. B.;
Hazuda, D. J ..; Flores, O. Identification and validation of
ribonucleoside analogues as hepatitis C antiviral agents. J . Biol.
Chem. 2003, 278, 49164-49170.
(17) Crotty, S.; Cameron, C. E.; Andino, R. RNA virus error catas-
trophe: direct molecular test by using ribavirin. Proc. Natl.
Acad. Sci. U.S.A. 2001, 98, 6895-6900.
(18) Pankiewicz, K. W.; Krzeminski, J .; Watanabe, K. A. Nucleosides.
164. Studies directed toward the synthesis of 2′-deoxy-2′-
substituted arabino nucleosides. 10. Synthesis of 2′-â-fluoro- and
3′-R-fluoro-substituted guanine nucleosides. Effect of sugar
conformational shifts on nucleophilic displacement of the 2′-
hydroxy and 3′-hydroxy group with DAST. J . Org. Chem. 1992,
57, 7315-7321.
(19) Kawasaki, A. M.; Casper, M. D.; Freier, S. M.; Lesnik, E. A.;
Zounes, M. C.; Cummins, L. L.; Gonzalez, C.; Cook, P. D.
Uniformly modified 2′-deoxy-2′-fluoro phosphorothioate oligo-
nucleotides as nuclease-resistant antisense compounds with high
affinity and specificity for RNA targets. J . Med. Chem. 1993,
36, 831-841.
(20) Franchetti, P.; Cappellacci, L.; Marchetti, S.; Trincavelli, L.;
Martini, C.; Mazzoni, M. R.; Lucacchini, A.; Grifantini, M. 2′-
C-Methyl Analogs of Selective Adenosine Receptor Agonists:
Synthesis and Binding Studies. J . Med. Chem. 1998, 41, 1708-
1715.
(23) Markiewicz, W. T.; Wiewiorowski, M. A new type of silyl
protecting groups in nucleoside chemistry. Nucleic Acids Res.,
Spec. Publ. 1978, 4, 185-188.
(24) Schmit, C.; Bevierre, M.-O.; De Mesmaeker, A.; Altmann, K.-H.
The effects of 2′- and 3′-alkyl substituents on oligonucleotide
hybridization and stability. Bioorg. Med. Chem. Lett. 1994, 4,
1969-1974.
(25) Koole, L. H.; Buck, H. M.; Vial, J . M.; Chattopadhyaya, J .
Synthesis and conformation of 9-(3′-C-methyl-â-D-xylo-furano-
syl)adenine and 3′-C-methyladenosine, two sugar-methylated
nucleoside analogs. Acta Chem. Scand. 1989, 43, 665-669.
(26) Carroll, S. S.; Lafemina, R. L.; Hall, D. L.; Himmelberger, A.
L.; Kuo, L. C.; Maccoss, M.; Olsen, D. B.; Rutkowski, C. A.;
Tomassini, J . E.; An, H.; Bhat, B.; Bhat, N.; Cook, P. D.; Eldrup,
A. B.; Guinosso, C. J .; Prhavc, M.; Prakash, T. P. Preparation
of nucleoside derivatives as inhibitors of RNA-dependent RNA
viral polymerase. PCT Int. Appl. WO 02/57425, 2002; 235 pp.
(27) Carroll, S. S.; Maccoss, M.; Olsen, D. B.; Bhat, B.; Bhat, N.; Cook,
P. D.; Eldrup, A. B.; Prakash, T. P.; Prhavc, M.; Song, Q.
Preparation of nucleoside derivatives as inhibitors of RNA-
dependent RNA viral polymerase. PCT Int. Appl. WO 02/57287,
2002; 85 pp.
(28) Plavec, J .; Tong, W.; Chattopadhyaya, J . How do the gauche and
anomeric effects drive the pseudorotational equilibrium of the
pentofuranose moiety of nucleosides? J . Am. Chem. Soc. 1993,
115, 9734-9746.
(29) Ford, H., J r.; Dai, F.; Mu, L.; Siddiqui, M. A.; Nicklaus, M. C.;
Anderson, L.; Marquez, V. E.; Barchi, J . J . Adenosine Deaminase
Prefers a Distinct Sugar Ring Conformation for Binding and
Catalysis: Kinetic and Structural Studies. Biochemistry 2000,
39, 2581-2592.
(30) Seela, F.; Muth, H. P.; Bindig, U. Synthesis of 6-substituted
7-carbapurine 2′,3′-dideoxynucleosides. Solid-liquid phase-
transfer glycosylation of 4-chloropyrrolo[2,3-d]pyrimidine and
deoxygenation of its 2′-deoxyribofuranoside. Synthesis 1988,
670-674.
(31) Kearsley, S. K. Unpublished in-house distance geometry pro-
gram. In Distance Geometry and Molecular Conformation, 2nd
ed.; Crippen, C. M., Havel, T. F. Eds.; Research Studio Press,
Wiley: New York.
(32) Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W.
C. MacroModelsan integrated software system for modeling
organic and bioorganic molecules using molecular mechanics.
J . Comput. Chem. 1990, 11, 440-467.
(21) Harry-O’kuru, R. E.; Smith, J . M.; Wolfe, M. S. A Short, Flexible
Route toward 2′-C-Branched Ribonucleosides. J . Org. Chem.
1997, 62, 1754-1759.
(22) Samano, V.; Robins, M. J . Nucleic acid related compounds. 60.
Mild periodinane oxidation of protected nucleosides to give 2′-
and 3′-ketonucleosides. The first isolation of a purine 2′-deoxy-
3′-ketonucleoside derivative. J . Org. Chem. 1990, 55, 5186-5188.
J M030424E