Synthesis and evaluation against hepatitis C virus of 7-deaza analogues of 2′-C-methyl-6-O-methyl guanosine nucleoside and l-Alanine ester phosphoramidates

Article abstract of DOI:10.1016/j.bmcl.2012.12.004

7-Deazapurines are known to possess broad antiviral activity, however the 2′-C-methylguanosine analogue displays poor cell permeation and limited phosphorylation, thus is not an efficient inhibitor of hepatitis C virus (HCV) replication. We previously rep

Full text of DOI:10.1016/j.bmcl.2012.12.004

Bioorganic & Medicinal Chemistry Letters 23 (2013) 2260–2264
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and evaluation against hepatitis C virus of 7-deaza
analogues of 20-C-methyl-6-O-methyl guanosine nucleoside
and L-Alanine ester phosphoramidates
a
b
Claire Bourdin a, Christopher McGuigan a, , Andrea Brancale , Stanley Chamberlain ,
John Vernachio b, Jeff Hutchins b, Elena Gorovits b, Alexander Kolykhalov b, Jerry Muhammad b,
Joseph Patti b, Geoffrey Henson b, Blair Bleiman b, K. Dawn Bryant b, Babita Ganguly b, Damound Hunley b,
Aleksandr Obikhod b, C. Robin Walters b, Jin Wang b, Changalvala V. S. Ramamurty b,
Srinivas K. Battina b, C. Srivinas Rao c
⇑
a School of Pharmaceutical Sciences, Cardiff University, King Edward VII Avenue, Cardiff CF10 3NB, UK
b Inhibitex Inc., 9005 Westside Parkway, GA 30009, USA
c CiVentiChem, 1001 Sheldon Drive, NC 27513, USA
a r t i c l e i n f o
a b s t r a c t
7-Deazapurines are known to possess broad antiviral activity, however the 20-C-methylguanosine ana-
logue displays poor cell permeation and limited phosphorylation, thus is not an efficient inhibitor of hep-
atitis C virus (HCV) replication. We previously reported the 6-O-methyl entity as a prodrug moiety to
increase liphophilicity of guanine nucleosides and the ProTide approach applied to 20-C-methyl-6-O-
methylguanosine has lead to potent HCV inhibitors now in clinical trials. In this Letter, we report the syn-
thesis and biological evaluation of 20-C-methyl-6-O-methyl-7-deaza guanosine and ProTide derivatives.
In contrast to prior studies, removal of the N-7 of the nucleobase entirely negates anti-HCV activity com-
pared to the 20-C-methyl-6-O-methylguanosine analogues. To understand better this significant loss of
activity, enzymatic assays and molecular modeling were carried out and suggested 20-C-methyl-6-O-
methyl-7-deaza guanosine and related ProTides do not act as efficient prodrugs of the free nucleotide,
in marked contrast to the case of the parent guanine analogue.
Article history:
Received 26 October 2012
Revised 3 December 2012
Accepted 4 December 2012
Available online 20 December 2012
Keywords:
HCV
NS5b polymerase
20-C-Methyl-6-O-methyl-7-deaza guanosine
ProTide
Phosphoramidate
Ó 2012 Elsevier Ltd. All rights reserved.
Identified in 1989,1 Hepatitis C is an infectious disease affecting
approximately 180 million people worldwide.2 The standard treat-
ment of care currently administered to patients consists of the gen-
eral antiviral agent ribavirin combined with pegylated interferon
and one of the recently discovered non-nucleoside inhibitors tela-
previr or boceprevir. This treatment is not only long but has limited
efficacy3 and causes numerous side effects,4 hence the need of
more efficient therapy. Several modified purine nucleosides, such
as b-20-C-methyl purines and nucleobase modified purines, have
exhibited significant anti-HCV activity.5 After phosphorylation to
their triphosphates b-20-C-methyl modified nucleosides are direct
inhibitors of HCV replication,6 and they are known as Direct Acting
Antivirals (DAAs). This includes several 7-deaza purines, particu-
larly b-20-C-methyl-7-deaza adenosine (1) (Fig. 1) which is
tive (EC50 >100 lM, CC50 >100 lM). The loss of inhibitory activity is
likely to be due to its poor cell uptake and inefficient intracellular
metabolism to its 50-triphosphate form.6,7 This membrane perme-
ation issue was previously addressed and modifications at the
6-position of nucleobase lead us to develop b-20-C-methyl-6-O-
methyl guanosine (2) (Fig. 1), being however fivefold less potent
in HCV replicon than its b-20-C-methyl guanosine analogue.8 Nev-
ertheless the ProTide technology applied to this nucleoside not
only boosted the lipophilicity but also enabled by-pass of the lim-
iting first phosphorylation step, delivering efficiently inside the
host cell the corresponding 50-monophosphate form.9 The
L-Ala-
nine neopentyl phosphoramidate of b-20-C-methyl-6-O-methyl
guanosine (3) (Fig. 1) ,8 also known as INX-08189 or BMS-
986094, is one of the examples of a ProTide developed for HCV
treatment.
reported to be a potent inhibitor of HCV replication (EC50
=
0.25
l
M, CC50 >100
l
M),5,6 whereas its guanosine analogue is inac-
It has previously been reported that both 20-C-methyl guanosine
7-deaza and 7-aza guanosine triphosphates exhibited sub-micro-
molar inhibitory potencies against the polymerase (respectively
IC50 = 0.12
not discriminate between 7-aza and 7-deaza derivatives. We thus
lM and IC50 = 0.13 l
M).6 Hence, the polymerase does
⇑
Corresponding author. Tel./fax: +44 (0)29 208 74537.
E-mail address: mcguigan@cardiff.ac.uk (C. McGuigan).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.12.004

C. Bourdin et al. / Bioorg. Med. Chem. Lett. 23 (2013) 2260–2264
2261
Figure 1. Structures of potent anti-HCV b-20-C-methyl nucleosides (1–2) and ProTide (3).
decided to synthesise 7-deaza analogues of 20-C-methyl-6-O-
methyl guanosine and related ProTides in order to increase inhibi-
tory potency against HCV.
nucleoside 8, three 50-ProTides (9–11) bearing the
L-Alanine as
amino acid moiety were synthesized following our standard proce-
dure12 using tert-butyl magnesium chloride as a base13 (Fig. 2).
Purification by column chromatography and preparative TLC were
required to obtain pure compounds. NMR and analytical HPLC con-
firmed the structure with purity of at least 95% for all phosphoram-
idates. They were each obtained in 12–13% yield as a mixture of
two diastereoisomers, which were only separated in the case of
b-20-C-Methyl-6-O-methyl-7-deaza guanosine nucleoside (8)
was synthesized in two stages: first the formation of the nucleo-
base (4) (Fig. 2), followed by coupling with the commercial 3,5-
bis-O-(2,4-dichlorophenylmethyl)-20-C-methyl-1-O-methyl-
ribofuranose (5) (Fig. 2).
a-D-
7-Deaza-6-chloroguanine (4) was formed in moderate yield
over two steps as described in the literature.10,11 Displacement of
the 6-chlorine of 4 by sodium methoxide was undertaken under
reflux for 5 days but lead mainly to degradation and generation
of 6-methoxy-7-deaza guanine in very poor yield (3%). It was then
decided to perform this step directly at the nucleoside level. Con-
densation with the ribofuranose moiety (5) was performed, as de-
scribed in the literature,6 by in situ conversion to the 1-bromo
analogue followed by subsequent reaction with 7-deaza-6-chloro-
guanine potassium salt (formed in situ with potassium hydroxide
powder) to afford only the b-anomer of 6 (Fig. 2). Treatment of 6
with boron trichloride in dichloromethane resulted in the depro-
tection of dichlorophenylmethyl group to yield 20-C-methyl-6-
chloro-7-deazaguanosine (7). The latter was then converted to
the 6-O-methyl analogue by substitution of the chlorine with
sodium methoxide in methanol under reflux,8 affording 20-C-
methyl-6-O-methoxy-7-deaza guanosine (8) after column
chromatography, in 50–60% yield. Without prior protection of the
the L-alanine-O-cyclohexyl ester ProTide (9a and 9b).
The three phosphoramidates (9–11) and their parent nucleoside
(8) were evaluated in a HCV replicon assay (EC50) and for cytotox-
icity in Huh-7 cells (CC50) (Table 1). Table 1 compares the biologi-
cal results of synthesized compounds (9–11) and their 20-C-
methyl-6-O-methyl guanosine analogues (2, 3, 12–13).
In vitro results (Table 1) suggest that 20-C-methyl-6-O-methyl-
7-deaza guanosine (8) is not an inhibitor of HCV replication
whereas its corresponding ProTides (9–11) boost the antiviral
activity from 3- to 10-fold. So, to some extent the ProTide approach
applied to 20-C-methyl-6-O-methyl-7-deaza guanosine is effective
and the results confirm the delivery of the 50-triphosphate inside
the cells, whereas the N-7 modification of the parent nucleoside
(8) does not show the improvement of potency desired and either
lacks of cell permeation or is a poor substrate for nucleoside ki-
nases responsible for phosphorylation. The separated diastereoiso-
mers 9a and 9b exhibit similar potency in replicon assays showing
little if any influence of phosphorus stereochemistry on bioactiva-
Figure 2. Synthesis of 20-C-methyl-6-O-methyl-7-deaza guanosine (8) and ProTides (9–11). Reagents and conditions: (a) anhyd CH2Cl2, HBr (33% in acetic acid, 6.7 equiv),
0 °C to rt, 2 h; anhyd ACN, KOH (3.0 equiv), TDA-1 (0.2 equiv), rt, 1 h; (b) anhyd CH2Cl2, BCl3 (10 equiv), ꢀ78 °C, 2 h then ꢀ20 °C, 2 h 30 min; (c) anhyd MeOH, NaOMe
(3.0 equiv), reflux, overnight; (d) anhyd THF, tBuMgCl (1.2 equiv), naphthyl L-Alanine ester phosphochloridate (1.2 equiv), rt, overnight.

2262
C. Bourdin et al. / Bioorg. Med. Chem. Lett. 23 (2013) 2260–2264
Table 1
Biological activity of 20-C-methyl-6-O-methyl-7-deaza guanosine (8) and 20-C-methyl-6-O-methyl guanosine (2) and their corresponding phosphoramidates
R
20-C-Methyl-6-O-methyl-7-deaza guanosine
20-C-Methyl-6-O-methyl guanosine
Compd
EC50
(lM)
CC50
(l
M)
Compd
EC50
(
lM)
CC50 (lM)
8
>100
13.3
8.9
12.9
27.3
>100
18
13
27
27
2
3
0.03
—
0.01
0.02
>100
6
—
6
11
cHex
cHex
CH2tBu
Bn
9a
9b
10
11
12
—
3
13
Figure 3. (a) UV spectra of conversion of 20-C-methyl-6-O-methyl-7-deaza guanosine (8) into 20-C-methyl-7-deaza guanosine (14) by adenosine deaminase assay; (b)
docking of b-20-C-methyl-6-O-methyl-7-deaza guanosine 50-monophosphate (grey sticks) in ADAL-1 (green line: natural substrate adenosine 50-monophosphate).
tion and potency. However, comparing the 7-deaza to the parent
guanine series, it is remarkable that the removal of the N-7 of
the nucleobase leads to a loss of potency over 1000-fold in the case
of the neopentyl and benzyl derivatives (3 vs 10 and 13 vs 11).
Their ClogP values being similar (3.14 to 4.00), cell uptake is prob-
ably not the main issue in this case.
phosphates,16 predicted that the later is
a relatively good
substrate (Fig. 3b). Thus, the b-20-C-methyl-7-deaza guanosine 50-
monophosphate should be delivered intracellularly.
To further investigate if 20-C-methyl-6-O-methyl-7-deaza gua-
nosine phosphoramidates deliver the corresponding 50-monophos-
phate intracellularly and thus may lead to the bioactive 20-C-
methyl-7-deaza guanosine 50-triphosphate, enzymatic experi-
ments were carried out to mimic the different stages of the puta-
tive mechanism of ProTide activation (Fig. 4). After incubation
with carboxypeptidase Y (CPY) in Trizma buffer, enzymatic cleav-
age of the amino acid ester of phosphoramidate 10 was monitored
by 31P NMR. Spectra were recorded regularly and selected spectra
are shown in Figure 5. The two peaks at dP ꢁ4 ppm correspond to
the two diastereoisomers of phosphoramidate 10. The two phos-
phorus signals decrease over time confirming conversion of 10 to
the amino acyl phosphate 15 lacking both the naphthyl and neo-
pentyl ester moieties represented by the peak at 7.10 ppm
(Fig. 5). While 50% of the diastereoisomer corresponding to dP
ꢁ4.04 ppm is converted after 5 h, more than 9 h is necessary for
50% of the disatereoisomer which peak appears at dP ꢁ3.55 ppm,
to be converted to its amino acyl phosphate derivative 15. This dif-
ference in metabolism rate is probably a result of different fitting in
the enzymatic catalytic site. These observations correspond to a
metabolic conversion rate of 10 over 20-times slower than its ana-
logue 3.8 This may partly explain the considerable loss of activity
exhibited in replicon assays. It is to some extent rather surprising
that the remote 7-deaza modification has such an impact on the
deesterification step on the ProTide. Compound 15 is a key metab-
olite for further enzymatic conversion to the corresponding 50-
monophosphate 16 by phosphoramidase type enzymes.
However, phosphoramidates 9–11 are somewhat cytotoxic to-
wards Huh-7 cells resulting in a poor selectivity index in compar-
ison to their of 20-C-methyl-6-O-methyl guanosine analogues.
Because the 20-C-methyl modification of 7-deaza adenosine re-
sulted in significant loss of cellular toxicity7 and neither 20-C-
methyl guanosine nor its 7-deaza analogue exhibited cytotoxicity
in Huh-7 cells,6 and because 7-deaza guanosine lacks antiviral
properties in cell cultures but is reported to be orally active
in vivo against numerous RNA viruses without high toxicity,14
the cytotoxicity may result from the 6-O-methyl modification. A
similar trend has already been reported by our group8 and this
modification would ideally be metabolised to its guanine deriva-
tive affording the delivery of the potent 20-C-methyl-7-deaza gua-
nosine 50-triphosphate (IC50 = 0.12 M).6 However, our key
l
observation is that the removal of the 7-aza moiety greatly reduces
anti-HCV efficacy in this series of nucleosides and activity is not
greatly restored by the ProTide.
To further understand the impact of the methine group replac-
ing the N-7 of the nucleobase in the biological activation pathway,
some enzymatic assays and molecular docking were carried out
with the 7-deaza analogue (10) of 3.
Preliminary UV experiments were carried out incubating 8 with
adenosine deaminase (ADA). As expected, the conversion of b-20-C-
methyl-6-O-methyl-7-deaza guanosine into b-20-C-methyl-7-dea-
za guanosine does not occur at a nucleoside level (Fig. 3a). How-
ever molecular docking of b-20-C-methyl-6-O-methyl-7-deaza
guanosine 50-monophosphate in an homology model of adenosine
deaminase-like protein1 (ADAL-1),15 recently reported as the en-
zyme involved in metabolism of 6-O/N-substituted purine mono-
To explain this difference of reactivity towards carboxypeti-
dase-type enzymes, some molecular docking was performed using
cathepsin A (PNB 1YSC) whose structure and substrate specificity
is similar to CPY.17 Preliminary docking of phosphoramidate 10
predicted that the Sp isomer fits best in the cathepsin pocket, thus

C. Bourdin et al. / Bioorg. Med. Chem. Lett. 23 (2013) 2260–2264
2263
Figure 4. Putative mechanism of activation of 20-C-methyl-6-O-methyl-7-deaza guanosine phosphoramidates.
which lies towards the front of the pocket and is quite far from the
essential catalytic residues Gly52, Gly53 and Ser146, whereas the
carbonyl group of 3 faces inside the catalytic pocket resulting in
faster ester cleavage. This would suggest that a hydrogen bond
with the N-7 in the nucleobase pocket is probably essential to lock
the nucleobase in a configuration that enables a good fit of the
phosphoramidate carbonyl into the catalytic site of carboxyester-
ase-type enzymes. Thus, modeling data support the slow release
of amino acid monophosphate 15 seen during CPY assays.
The second step of activation was investigated in order to con-
firm the release of the 50-monophosphate 16. ProTide 10 was incu-
bated with Huh-7 cell lysates. Mass spectra and 31P NMR spectra
(data not shown) were recorded and suggested the formation of
the desired 50-monophosphate 16 (M+Na+ = 413). Nevertheless
conversion of 10 is very slow and even after 48 h, 10 remains the
main specie present in the medium (M+H+ = 658, M+Na+ = 680,
M+K+ = 696). This result is in correlation with the predictive dock-
ing in Human Hint-1 (PNB 1KPF) (Fig. 6b), where the phosphorus
lies far from the optimal position near the catalytic residues based
on adenosine monophosphate, the natural substrate.
Figure 5. Carboxypetidase Y assay on phosphoramidate 10.
To conclude, we report the synthesis of 7-deaza derivatives of b-
20-C-methyl-6-O-methyl guanosine. Despite the efficiency of the
ProTide technology, replacement of N-7 by a methine group leads
to a significant loss of potency against HCV and loss of antiviral
selectivity which cannot be salvaged by the ProTide motif. A much
Sp isomers of both 10 and analogue 3 were docked into the cathep-
sin active site (Fig. 6a). The removal of N-7 provokes an undesirable
effect on the position of the amino acid ester carbonyl group of 10,
Figure 6. (a) Docking of SP isomers of 3 (blue) and 10 (yellow) in the catalytic site of cathepsin A (red: oxygen, pink: phosphorus). (b) Docking of aminoacyl intermediate of 10
(grey stick) and adenosine 50-monophosphate (AMP, red line) in the catalytic site of human Hint-1.

2264
C. Bourdin et al. / Bioorg. Med. Chem. Lett. 23 (2013) 2260–2264
3. Walker, M. P.; Appleby, T. C.; Zhong, W.; Lau, J. Y. N.; Hong, Z. Antiviral Chem.
Chemother. 2003, 14, 1.
4. Hoofnagle, J. H.; Seef, L. B. N. Engl. J. Med. 2006, 355, 2444.
5. Vittori, S.; Dal Ben, D.; Lambertucci, C.; Marucci, G.; Volpini, R.; Cristalli, G. Curr.
Med. Chem. 2006, 13, 3529.
6. Eldrup, A. B.; Prhavc, M.; Brooks, J.; Bhat, B.; Prakash, T. P.; Song, Q.; Bera, S.;
Bhat, N.; Dande, P.; Dan Cook, P.; Bennett, C. F.; Carroll, S. S.; Ball, R. G.;
Bosserman, M.; Burlein, C.; Colwell, L. F.; Fay, J. F.; Flores, O. A.; Getty, K.;
LaFemina, R. L.; Leone, J.; MacCoss, M.; McMasters, D. R.; Tomassini, J. E.;
VonLangen, D.; Wolanski, B.; Olsen, D. B. J. Med. Chem. 2004, 47, 5284.
7. Olsen, D. B.; Eldrup, A. B.; Bartholomew, L.; Bhat, B.; Bosserman, M. R.; Ceccacci,
A.; Colwell, L. F.; Fay, J. F.; Flores, O. A.; Getty, K. L.; Grobler, J. A.; LaFemina, R.
L.; Markel, E. J.; Migliaccio, G.; Prhavc, M.; Stahlhut, M. W.; Tomassini, J. E.;
MacCoss, M.; Hazuda, D. J.; Carroll, S. S. Antimicrob. Agents Chemother. 2004, 48,
3944.
reduced rate of processing of the ProTide to the key amino acyl
intermediate and the slow release of the 50-monophosphate specie
may be the reasons for this outcome, since the conversion to b-20-
C-methyl-7-deaza guanosine 50-monophosphate by ADAL-1 should
be efficient. The remarkable loss of potency at the ProTide level on
formation of the 7-deaza species is a surprising outcome of this
study and it shows the need to tune the ProTide motif, and under-
stand its interplay with the nucleoside, on each occasion. More-
over, the lack of activity of these 7-deaza guanosine derivatives
is not expected to be due to differences of affinity at the polymer-
ase level compared to guanosine derivatives.6
8. McGuigan, C.; Madela, K.; Aljarah, M.; Gilles, A.; Brancale, A.; Zonta, N.;
Chamberlain, S.; Vernachio, J.; Hutchins, J.; Hall, A.; Ames, B.; Gorovits, E.;
Ganguly, B.; Kolykhalov, A.; Wang, J.; Muhammad, J.; Patti, J. M.; Henson, G.
Bioorg. Med. Chem. Lett. 2010, 20, 4850.
9. Hersh, M. R.; Kuhn, J. G.; Philips, J. L. Cancer Chemother. Pharmacol. 2004, 4, 371.
10. Flechter, T. M.; Cathers, B. E.; Ravikumar, K. S.; Mamiya, B. M.; Kerwin, S. M.
Bioorg. Chem. 2001, 29, 36.
11. Tangeda, S. J.; Jyostna, S.; Garlapati, A. Eur. J. Med. Chem. 2010, 45, 1453.
12. Uchiyama, M.; Aso, Y.; Noyori, R.; Hayakawa, Y. J. Org. Chem. 1993, 58, 373.
13. McGuigan, C.; Perrone, P.; Madela, K.; Neyts, J. Bioorg. Med. Chem. Lett. 2009, 19,
4316.
Acknowledgments
This work was funded by Inhibitex and BMS. We would like to
thank Dr. Brancale for his help with the modeling and Helen Mur-
phy for secretarial assistance.
Supplementary data
14. Smee, D. F.; Alaghamandan, H. A.; Gilbert, J.; Burger, R. A.; Jin, A.; Sharma, B. S.;
Ramasamy, K.; Revankar, G. R.; Cottam, H. B.; Jolley, W. B. Antimicrob. Agents
Chemother. 1991, 35, 152.
15. Meneghesso, S.; Vanderlinden, E.; Brancale, A.; Balzarini, J.; Naesens, L.;
McGuigan, C. Chem. Med. Chem. 2013, in press.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.
12.004.
16. Murakami, E.; Bao, H.; Mosley, R. T.; Du, J.; Sofia, M. J.; Furman, P. A. J. Med.
Chem. 2011, 54, 5902.
References and notes
17. (a) Elslisger, M.; Pshezhetsky, A. V.; Vinogradova, M. V.; Svedas, V. K.; Potier, M.
Biochemistry 1996, 35, 14899–148909; (b) Hiraiwa, M. Cell Mol. Life Sci. 1999,
56, 894.
1. Choo, Q.-L.; Richman, K. H.; Han, J. H., et al Proc. Natl. Acad. Sci. U.S.A. 1991, 88,
2451.
2. Cohen, J. Science 1999, 285, 26.