Phosphoramidate ProTides of 2′-C-methylguanosine as highly potent inhibitors of hepatitis C virus. Study of their in vitro and in vivo properties

Article abstract of DOI:10.1021/jm1003792

Hepatitis C virus infection constitutes a serious health problem in need of more effective therapies. Nucleoside analogues with improved exposure, efficacy, and selectivity are recognized as likely key components of future HCV therapy. 2?-C-Methylguanosine triphosphate has been known as a potent inhibitor of HCV RNA polymerase for some time, but the parent nucleoside is only moderately active due to poor intracellular phosphorylation. We herein report the application of phosphoramidate ProTide technology to bypass the rate-limiting initial phosphorylation of this nucleoside. Over 30 novel ProTides are reported, with variations in the aryl, ester, and amino acid regions. l-Alanine compounds are recognized as potent and selective inhibitors of HCV in replicon assay but lack rodent plasma stability despite considerable ester variation. Amino acid variation retaining the lead benzyl ester moiety gives an increase in rodent stability but at the cost of potency. Finally l-valine esters with ester variation lead to potent, stable compounds. Pharmacokinetic studies on these agents in the mouse reveal liver exposure to the bioactive triphosphate species following single oral dosing. Systemic exposure of the ProTide and parent nucleoside are low, indicating possible low toxicity in vivo, while liver concentrations of the active species may be predictive of efficacy in the clinic. This represents one of the most thorough cross-species studies of ProTides to date.

Full text of DOI:10.1021/jm1003792

J. Med. Chem. 2010, 53, 4949–4957 4949
DOI: 10.1021/jm1003792
Phosphoramidate ProTides of 20-C-Methylguanosine as Highly Potent Inhibitors of
Hepatitis C Virus. Study of Their in Vitro and in Vivo Properties
Christopher McGuigan,*,† Arnaud Gilles,† Karolina Madela,† Mohamed Aljarah,† Sabrina Holl,† Sarah Jones,†
John Vernachio,‡ Jeff Hutchins,‡ Brenda Ames,‡ K. Dawn Bryant,‡ Elena Gorovits,‡ Babita Ganguly,‡ Damound Hunley,‡
Andrea Hall,‡ Alexander Kolykhalov,‡ Yule Liu,‡ Jerry Muhammad,‡ Nicholas Raja,‡ Robin Walters,‡ Jin Wang,‡
Stanley Chamberlain,‡ and Geoffrey Henson‡
†Welsh School of Pharmacy, Cardiff University, King Edward VII Avenue, Cardiff CF10 3NB, U.K., and
‡Inhibitex Inc., 9005 Westside Parkway, Alpharetta, Georgia 30004
Received March 26, 2010
Hepatitis C virus infection constitutes a serious health problem in need of more effective therapies.
Nucleoside analogues with improved exposure, efficacy, and selectivity are recognized as likely key
components of future HCV therapy. 20-C-Methylguanosine triphosphate has been known as a potent
inhibitor of HCV RNA polymerase for some time, but the parent nucleoside is only moderately active
due to poor intracellular phosphorylation. We herein report the application of phosphoramidate
ProTide technology to bypass the rate-limiting initial phosphorylation of this nucleoside. Over 30 novel
ProTides are reported, with variations in the aryl, ester, and amino acid regions. L-Alanine compounds
are recognized as potent and selective inhibitors of HCV in replicon assay but lack rodent plasma
stability despite considerable ester variation. Amino acid variation retaining the lead benzyl ester moiety
gives an increase in rodent stability but at the cost of potency. Finally L-valine esters with ester variation
lead to potent, stable compounds. Pharmacokinetic studies on these agents in the mouse reveal liver
exposure to the bioactive triphosphate species following single oral dosing. Systemic exposure of the
ProTide and parent nucleoside are low, indicating possible low toxicity in vivo, while liver concentra-
tions of the active species may be predictive of efficacy in the clinic. This represents one of the most
thorough cross-species studies of ProTides to date.
Introduction
Several families of nucleoside analogues have emerged with
apparent selectivity for HCV, these include 40-modified agents
such as 40-azidocytidine (R1479, 1a, Figure 1) developed as
an oral pro-drug by Roche4 and 20-C-methyl adenosine (1b,
Figure 1) and related deaza compounds developed by Merck.5
Interestingly, the Merck team observed that the guanine
analogue 1c was more potent (>10-fold) as an inhibitor of the
HCV RNA polymerase, as its bioactive 50-triphosphate form,
than was the triphosphate of the adenosine lead 1b. However,
poor cell entry by 1c coupled with apparently poor phospho-
rylation lead to it being >10-fold less active than 1b in HCV
replicon (1b EC50 0.26 μM, 1c 3.5 μM).5
Unfortunately, despite the reasonable in vitro potency and
selectivity of 1b, the compound could not be progressed due to
it being a good substrate for metabolic enzymes (adenosine
deaminase and purine nucleoside phosphorylase), leading to
low oral bioavailability.5
To overcome these PK liabilities of 1b, the Merck team
progressed with several unnatural base modifications, such as
the 7-deaza family, leading to their late preclinical candidate
MK-0608. As recently highlighted, a lack of data on this family
for over three years now may place a query over their progres-
sion.3 We were concerned that such base modifications could
carry with them the possibility of toxicity, and we sought to
retainanatural base andtoseektoovercomethe limitationsof
the guanosine compound 1c in particular. Because the first
phosphorylation of 1c was considered to be rate-limiting, we
An estimated 200 million people, or ca. 3.5% of the world’s
population, are chronically infected with the hepatitis C virus
(HCVa) and at risk of developing life threatening liver disease
such as cirrhosis or liver carcinoma. HCV infection is the
major cause of liver transplantation in industrialized coun-
tries. The current therapy for HCV consists of pegylated
interferon and ribavirin, neither of which are specific inhibi-
tors of HCV, and a regimen which has side effects and limited
efficacy in at least half of the patient population.1
The HCV genome offers several clear targets for specific
anti-HCV therapy; these include an RNA polymerase and a
serine protease, both of which have attracted considerable
academic and industrial attention.2
The RNA dependent HCV polymerase is considered to be
an attractive target for therapy on account of a high degree of
conservation across the six HCV genotypes, and agents tar-
geted at the active site, such as nucleoside analogues, may
be particularly advantageous with respect to the barrier to
resistance.3
*To whom correspondence should be addressed. Phone/Fax: þ44 29
20874537. E-mail: mcguigan@cardiff.ac.uk.
a Abbreviations: HCV, hepatitis C virus; TLC, thin layer chromato-
graphy; HPLC, high performance, liquid chromatography; ClogP,
calculated logarithm of octanol-water partition coefficient; PK, phar-
macokinetics; SI, selectivity index; DNAUC, dose normalized area
under the curve; NTP, nucleoside triphosphate; THF, tetrahydrofuran.
r
2010 American Chemical Society
Published on Web 06/08/2010
pubs.acs.org/jmc

4950 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
McGuigan et al.
wondered if a pro-drug of the 50-monophosphate of 1c may be
active, and thus we applied our phosphoramidate ProTide
approach6 to this nucleoside. We first introduced this ProTide
motif in 19967 and have exemplified it considerably,6 includ-
ing very recently to 1a8 and briefly to 1b-c.9 Our approach
has also been adopted by Pharmasset for their 20-modified
nucleoside HCV family10 and adapted to acyclic nucleoside
phosphonates by Gilead.11 The Pharmasset nucleoside is
also 20-C-methyl based, but is a uridine compound, also with
a 20-R-fluorine. This company are using a ProTide based on
isopropylalanine phenyl phosphate,12 rather parallel to our
early work on anti-HIV ProTides13 Idenix have also pursued a
ProTide approach on 1C, although the precise structure of
their lead IDX-184 has yet to issue.14
Following an early indication that the ProTides may suc-
cessfully impact on 1c,9 we herein report an extensive study
thereof. We report over 30 novel analogues with extensive
variation in the aryl, ester, and amino acid regions of the Pro-
Tide, and we study these extensively across species for their
stability. We note that the benzyl alanine motif continues7-9
to exhibit good potency in replicon, but there is poor stability
and PK in this series, which can only be solved by a novel
combined variation in both the amino acid and ester regions
as we will describe. The basic structure of a phosphoramidate
motif is shown in Figure 2.
Chemistry
The target ProTides of 1c were prepared using phosphoro-
chloridate chemistry as we have extensively reported.6-9 To
aid in both the 50-regioselectivity of phosphorylation and the
general organic solubility of the nucleoside, we investigated
various protecting groups for the 20,30-diol unit. One key issue
is the ability to remove this protecting group after construc-
tion of the 50-ProTide unit, and the relative chemical fragility
of this group dictates the nature of the diol protecting group
that might be acceptable. Indeed, the somewhat surprising
acid-stable nature of ProTides15 means that acid sensitive diol
protection may be acceptable, and this led us to the use of iso-
propylidene protection.
Figure 1. Some anti-HCV nucleosides.
Thus 1c was allowed to react with acetone and perchloric
acid at ambient temperatureovernight togive1d(Scheme1)in
93% yield. This proved to be significantly more soluble than
1c and readily reacted with the appropriate phosphorochlo-
ridate reagents6-9,15 in organic solvents such as THF (Scheme 1).
Various bases could be used to aid the reaction; the Grignard
Figure 2. A general ProTide structure.
Scheme 1a
a For 2a-n, 3a-g, 4a-j, see Tables 1, 3, 5.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4951
Table 1. HCV Replicon Activity of L-Alanines Ester Derivatives
Table 2. Plasma Stability of L-Alanine Estersa
% remaining in plasma at 2-4 °C
ProTide aryl
ester
isomersa
EC50/μMb CC50/μM
1c
3.5
>100
>50
ProTide
ester
time
human cyno canine rat mouse
2a
2b
2c
2d
2e
2f
1-Nap Me
1-Nap nBu
38:62
0.21
0.10
0.091
0.17
0.12
0.07
0.045
0.057
0.062
0.17
0.09
0.17
0.095
0.08
2a
2b
Me
nBu
30 min
30 min
4 h
ND
100
97
ND
98
ND
95
ND
0.4
0
1.7
0
43:57
37:63
40:60
38:62
48:52
34:66
31:69
45:55
43:57
44:56
>100
ND
1-Nap nPnt
1-Nap iPr
81
94
90
98
0
>100
>100
>100
>100
>100
>100
>50
2c
nPnt
30 min
4 h
100
96
0
0
1-Nap 2-Bu
1-Nap cPnt
1-Nap cHx
1-Nap tBuCH2
1-Nap Bn
94
98
0
0
2d
2e
2g
iPr
30 min
30 min
30 min
4 h
ND
ND
100
100
ND
58
ND
ND
100
100
ND
54
ND
ND
100
100
ND
67
ND
ND
2.5
0.5
ND
1.3
ND
0.7
14
10
2.9
2g
2h
2i
2-Bu
cHx
0
2j
2-Nap Bn
1-Nap BnCH2
2h
2i
2jb
2m
tBuCH2 30 min
0.3
1.1
0.7
2.2
2k
2l
>100
>100
>100
>100
Bn
Bn
30 min
30 min
1-Nap BnCH2CH2 40:60
1-Nap R/S PhEt
1-Nap S PhEt
ND
86
ND
100
ND
97
2m
2n
33, 14, 35, 18c
42:58
(R/S)-1- 30 min
PhEt
a Isomer ratio from HPLC; polar:nonpolar. b Replicon data for geno-
type 1b in HUH7 cells with 48 h exposure. c Four stereoisomers.
4 h
S-PhEt 30 min
4 h
86
88
93
97
92
0.3
0.8
0
0
2n
100
99
100
100
3.1
0
reagent tBuMgCl was among the most effective, and prefer-
ably used in ca. 2 molar equivalence, along with the phos-
phorochloridate. Reaction overnight at ambient temperature
followed by flash column chromatography on silica gave the
isopropylidene protected intermediates (Scheme 1) in 30-
88% yield. These were deprotected using 60% acetic acid at
95 °C overnight. The relative stability of the ProTide motif
under these conditions was taken as further support of their
surprising acid stability and may be taken as further evidence
of their possible compatibility with oral dosing. The final
compounds 2a-n, 3a-g, and 4a-j (Tables 1, 3, and 5) were
each isolated after column chromatography, sometimes fol-
lowed by preparative TLC and/or HPLC. Standard methods
are given below and spectroscopic and analytical data on 2a
only; full detail on other analogues is given as Supporting
Information. In each case, the final products were isolated as
mixtures of diastereoisomers at the phosphate, as evidenced
by multiple (in most cases 2) peaks in the P-31 NMR and also
in the HPLC. The approximate ratio of peaks (fast eluting on
reverse phase, i.e., polar to nonpolar) as determined by HPLC
is given in Tables 1, 3, and 5, where peaks were sufficiently
clearly resolved. The approximate ratios from P-31 NMR
roughly mirrored those determined by HPLC and allowed a
correlation between HPLC mobility/polarity and NMR shift,
such that the most predominant isomer in most cases was the
downfield NMR peak, corresponding to the more nonpolar
stereoisomer. The HPLC retention time data confirmed that
the ProTides were in every case considerably more lipophilic
than parent1c. Calculatedlipophilicity (ClogP) values16 for 1c
and 2a are -2.6 and -0.52, respectively, indicating a roughly
100-fold lipophilicityenhancement for 2a. BecauseMerck had
noted apparently poor cell uptake by 1c,5 we considered the
enhancement in this parameter for the ProTides potentially
useful in vivo. However, the figure for the methylalanine ester
2a was still somewhat lower than that considered optimal
for passive cell uptake, and thus we varied the ester moiety to
enhance lipophilicity further and also to probe the putative
ester cleavage step considered to be key to the in vivo activation
of these agents.6,15 In the L-alanine series 2a-n, the lengthening,
branching, and aromatization of the ester moiety lead to agents
such as the benzyl ester 2i with a ClogP of 1.2 and branched
analogues of this, such as 2m and 2n, with perhaps near ideal
values of 1.5. Notably, in the case of 2m, the introduction of a
further variable chiral center in the ester led to a mixture of four
diastereomers, as evidenced by P-31 NMR and HPLC.
a ND: not determined. b 2-Naphthyl.
Besides ester variation, we prepared one analogue, 2j, with
2-naphthyl as the aryl moiety in place of 1-naphthyl. We also
varied the amino acid unit from L-alanine (2a-n), to valine
(3a), leucine (3b), methionine (3c-d), phenylalanine (3e),
isoleucine(3f), and the unnaturalachiralaminoacid dimethyl-
glycine (3g) (Table 3). In most cases, these were prepared as
the benzyl ester, except for 3d, being the iPr ester of Met.
As will be discussed below, L-valine emerged as a promising
amino acid from these earlier studies. This is notable, as it had
previously been observed by us to be among the least effective
amino acids when applied to several families of ProTides.6,17
However, in this case, promising biological data necessitated
the preparation of a family of L-valine esters (4a-j, Table 5).
These included some unusual esters not previously widely
reported for ProTides: the substituted benzyl family (4f-i)
and the neopentyl compound (4j).
Activity in Replicon Assay and Plasma Stability
As is common in anti-HCV drug development, we used
subgenomic HCV replicon as a primary biological readout.
Data are reported for the initial L-alanine family 2a-n in
Table 1, along with comparator data for 1c in this assay. Each
data point represents the mean of at least three independent
assays. As noted in Table 1, all of the L-alanine ProTides were
activesub-μM inthe replicon assay, while 1cwas onlyactiveat
3.5 μM. Thus, the initial family of esters was 4-50-fold more
potent than 1c. This was taken as an early indication that the
ProTides successfully improved the intracellular delivery of the
eventual 50-triphosphate pharmacophore of 1c, presumably
by a combination of enhanced passive diffusion of the lipo-
philic ProTides into cells and the delivery of the monophos-
phate therein by a nucleoside kinase-independent process.
There was no clear correlation between ProTide lipophilicity
and biological potency in this family, with several diverse
esters showing activity in the 40-60 nM range.
All the ProTides displayed minimal toxicity in the Huh7
cells expressing the HCV genotype 1b bicistronic subgenomic
replicon. When replicon 1b cells were incubated with the
different ProTides for 72 h and cell viability was measured
using the CellTiter-Glo assay (Promega, Madison, WI), CC50
values greater than 100 μM were routinely observed. This was
true even for the most potent compounds like 2g, thus leading

4952 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
McGuigan et al.
Table 4. Plasma Stabilities of Amino Acid Derivatives
Table 3. HCV Replicon Activity of Different Amino Acid Derivativesa
isomersb EC50/μMb CC50/μM
% remaining in plasma at 2-4 °C
ProTide
aryl
ester
AA
3a
3b
3c
3d
3e
3f
1-Nap Bn
1-Nap Bn
1-Nap Me L-Met
L-Val
L-Leu
32:68
NAc
0.76
0.12
0.28
0.34
1.2
>100
>50
ProTide ester AA
time human cyno canine rat mouse
3a
Bn L-Val
30 min
4 h
89
66
2
31:69
36:64
34:66
NAc
>100
>100
>100
>50
1-Nap iPr
1-Nap Bn
1-Nap Bn
1-Nap Bn
L-Met
L-Phe
L-Ile
3b
3c
3d
Bn L-Leu 30 min
Me L-Met 30 min
iPr L-Met 30 min
4 h
34
76
32
1
0.9
1.5
77
79
96
93
100
100
2.3
0.5
3g
Me2Gly
31:69
ND
a NA: not available. ND: not determined. b See Table 1. c Peaks
merged on HPLC.
3e
3f
Bn L-Phe
Bn L-Ile
30 min
30 min
4 h
100
95
0
3g
Bn Me2Gly 30 min 100
4 h 41
92
83
100
100
0
0
to high selectivity indexes in these cells. Several compounds
were also tested against MT4, CEM, HepG2, Huh7, HEL,
293, IEC-6, and CaCo2 cells. All cell lines tested were in log
phase of growth throughout the cell cytotoxicity experiments
(data not shown). A similar lack of toxicity was observed
in these diverse cell lines. For example, compound 2n had
CC50 values >100 μM for all the above cell lines except MT4
(CC50=87 μM) and CEM (CC50=82 μM). This gives 2n an
SI of ca. g1000.
0
Table 5. HCV Replicon Activity of L-Valine Ester Derivativesa
ProTide aryl ester AA
isomersb EC50/μMb CC50/μM
1-Nap Bn Val
2-Nap Bn Val
3a
4a
4b
4c
4d
32:68
56:44
0.76
1.7
>100
>100
>100
>100
ND
1-Nap Bn
1-Nap cHx
D-Val 68:32
65:35
NAc
>3
1.0
>3
Val
Thus, to progress hit to lead selection, we sought other
assays that may be informative of in vivo PK. Prior to
investigating the PK of the various ProTides, we examined
the plasma stability of the L-alanine esters in mouse, rat, dog,
monkey, and human plasma (Table 2). The plasma stability
studies were carried out at 2-4 °C because they were intended
to confirm sufficient stability for calibration curves, not
necessarily for understanding stability at the physiologically
more relevant 37 °C. Perhaps not surprisingly, the various
esters of the L-alanine ProTides all were very unstable in
rodent plasma even at these lower temperatures. The data
are reported as % remaining after either 30 min or 4 h. All the
L-alanine esters synthesized were essentially completely
cleaved in 30 min. The most stable esters were the branched
isopropyl and 2-butyl esters, which still were greater than 80%
consumed in 30 min even at this lower temperature. Because
of this profound plasma instability, it became clear we would
not be able to generate calibration curves in rodent plasma
and as a result we would not be able to measure ProTide
plasma concentrations in either mouse or rat. On the other
hand, the L-alanine esters for the most part had sufficient
plasma stability in other species to generate calibration curves
and measure exposure. The benzyl ester 2i was the notable
exception, with nearly 40% degradation in 30 min in human,
monkey, and dog plasma. However, substitution at the
benzylic position, as in the case of the 2-phenylethyl derivative
2m and 2n, led to substantially improved plasma stability in
these three species with up to 100% remaining even after 4 h.
Because modification of the ester group did not readily
provide rodent plasma stability, and because being able to do
initial pharmacokinetic studies in rodents was considered
desirable, the amino acid core of the ProTides was varied to
investigate both potency and plasma stability. Thus, as noted
in Table 3, we prepared several ProTides, mainly as their
benzyl esters, of six different amino acids. Benzyl esters were
selected because we believed they provided the best chance
of showing potency in the replicon assay. We particularly
selected bulky amino acids, partly to enhance lipophilicity and
partly to seek to induce some steric hindrance to degradation
of the ProTide. As noted in Table 3, this lead to a significant
(2-10-fold) loss of activity in the best cases and a 20-fold
decrease in the worst case. Disubstitution of the R carbon
of the amino acid lead to a 20-fold decrease in activity, while
1-Nap (R/S)-1- Val
PhEt
4e
1-Nap BnCH2- Val
CH2
1.7
>100
4f
4g
4h
4i
1-Nap o-ClBn Val
1-Nap m-ClBn Val
1-Nap p-ClBn Val
1-Nap o-MeBn Val
1-Nap tBuCH2 Val
33:67
0.43
0.68
0.62
1.0
>100
ND
ND
31, 25, 44
35:65
36:64
NAc
100
>100
4j
1.5
a NA: not available. ND: not determined. b See Table 1. c See Table 3.
amino acids such as the L-leucine derivative 3b were the
most potent EC50 = 120 nM. Branching at the β carbon of
the amino acid as in the L-valine and L-isoleucine derivatives
3a and 3f, resulted in a 10-fold decrease in activity, however,
these ProTides retained sub-μM activity and thus were still
more active than 1c. Compound 3b was ca. 30 times more
active than 1c.
The plasma stability of these alternate amino acid ProTides
was examined in the hope of finding improved rodent plasma
stability. The results of testing in mouse plasma for com-
pounds 3a-3g are reported in Table 4. Disubstitution of the
R carbon of the amino acid, 3g, did not noticeably improve
mouse plasma stability relative to the L-alanine derivative 2i.
Longer chain amino acids also failed to provide plasma
stability. However, branching at the β carbon did provide
significantly improved plasma stability. Thus, the Val and Ile
derivatives 3a and 3f provided 66% and 95% stability in
mouse plasma even after 4 h. This stability was measured at
2-4 °C and clearly would be worse at 37 °C. However, this
level of stability would at least allow generation of calibration
curves and subsequent measurement of plasma ProTide levels
in rodents. These branched ProTides maintained their stabi-
lity in the plasma from other species, but their activity in the
replicon assay was somewhat less than that of the L-alanine
analogues.
Thus, we subsequently prepared a small family of L-valine
analogues with ester variations to examine whether we could
improve antiviral activity and maintain rodent plasma stabi-
lity. The derivatives made are listed in Table 5 and are
compared to compound 3a. Table 5 includes a 2-naphthyl
derivative 4a and the D-valine derivative 4b, both of which are

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4953
Table 6. Plasma Stability of L-Valine Esters
Table 7. Multiple Species Liver and Intestinal S9 Data for Selected
Compounds with 60 min Incubations (% Remaining)
% remaining in plasma at 2-4 °C
time human cyno canine rat mouse
ProTide
S9% remaining
human
dog
monkey
rat
ND
ProTide
ester
Bna
2a
liver (
intestine (
liver (
intestine (
liver (
intestine (
liver (
intestine (
liver (
intestine (
liver (
intestine (
liver (
intestine (
liver (
intestine (
liver (
intestine (
liver (
intestine (
liver (
intestine (
20/40
60/63
27/41
65/85
16/50
42/48
7/59
67/85
50/54
81/91
66/68
53/80
82/100
49/87
42/56
16/53
8/29
ND
ND
4a
30 min
4 h
93
100
70
100
95
95
99
49
22
82
56
99
76
51
8
ND
2d
2e
2h
2i
0/86
48/87
ND
25/28
77/76
ND
4b
Bnb
30 min
4 h
30 min
68
80
69
61
82
97
4c
4d
4f
4g
4h
4i
cHx
95
95
100
69
79
94
67
77
ND
ND
(S)-1-PhEt 30 min
97
77
100
100
90
16/88
20/68
0/32a
0/30a
ND
5/31
57/86
0/0a
30/46a
ND
ND
o-ClBn
m-ClBn
p-ClBn
o-MeBn
tBuCH2
30 min
30 min
30 min
30 min
30 min
28/49
27/41a
0/49a
14/29
24/33
0/78
100
95
89
100
91
100
79
2j
33/28
23/27
56/75
56/72
14/48
12/33
27/40
57/86
4/15a
0/0a
4j
ND
a 2-Naphthol derivative. b D-Val derivative.
3a
3b
3c
3e
3f
0/77a
25/77a
ND
8/62
65/81
ND
24/70
0/44a
8/28a
3/14
less active in replicon assay. The other ester derivatives, 4c-4j,
are similar in activity to the lead compound 3a, with simple
aliphatic esters perhaps falling off slightly in potency and
benzyl esters with an electron withdrawing group ortho, meta,
orpara, all being somewhatmore potent. However, branching
at the benzylic position, as in the (R/S)-1-phenylethyl deriva-
tive 4d, resulted in a substantial decrease in activity, which was
not observed in the L-alanine series.
Plasma stabilities were examined across this series and are
listed in Table 6. A number of interesting observations can be
made from the rodent plasma stability data. While the 2-naphthol
derivative 4a has similar stability to the 1-naphthol derivative,
the D-valine benzyl ester 4b was observed to be less stable than
the L-valine analogue in rodent plasma. The other L-valine
esters maintained rodent, and other species, plasma stabilities
with the possible exceptions of the ortho chloro and ortho
methyl substituted benzyl esters 4f and 4i, which showed
minor decreases in stability relative to the para-substituted
and unsubstituted benzyl esters.
ND
ND
ND
ND
42/59
0/20a
0/5a
ND
ND
ND
ND
ND
ND
37/78
71/70
29/83
52/74
64/84
87/96
0/89a
29/95
a Incubated for 30 min. (refers to with/without cofactor.
demonstrated greater stability in intestinal S9 preps than in
liver, as was desired (Table 7). In this general sense, these
compounds could be considered as liver targeting. The lone
exception was the benzyl ester of L-alanine, compound 2i,
which was highly cleaved in the intestine. Stability in intestinal
cells was considered important because we wanted the Pro-
Tide to pass through the gut and intestine and be absorbed
and cleaved by the liver. The intestinal stabilities, as measured
by % remaining after 60 min incubations, ranged from 0% to
65%. The most stable of this set of compounds was 2d, the
isopropyl ester of L-alanine.
Discussion of in Vitro S9 Stability Studies
As noted above, the compounds in Table 7 were all
incubated with and without cofactor. Looking across the data
in the table, there is a trend toward greater cofactor depen-
dence in the liver than in the intestine. Because P450 isozymes
require a cofactor, and because P450 levels are high in liver,
this is taken as a further subtle indication of liver targeting.
The benzyl ester derivatives such as compound 3a showed a
large degree of cofactor dependent liver cleavage, as did the
neopentyl ester derivative compound 2h.
Human, dog, cynomolgus monkey, and rat liver and intes-
tinal S9 stability data, as measured by disappearance of parent
compound, were obtained for selected compounds (Table 7).
Mouse data were not collected because of difficulty in obtain-
ing mouse liver S9 preparations. All samples were run with
and without cofactor (NADPH) and expressed as % remain-
ing compound after 60 min incubations. Because NADPH is a
cofactor for P450 isozymes, metabolism in the absence of
cofactor was considered non-P450 dependent metabolism.
We were interested in identifying which species compared
closest to human S9 data and which compounds gave the most
desirable profile in human intestinal and liver S9 experiments.
Compounds were compared across species mainly by look-
ing at the ratios of liver vs intestinal metabolism as opposed to
absolute numbers in each category. When considered this
way, it became clear that rat and monkey were reasonable
comparators to the human data, but that the dog in vitro data
was the outlier. As an example of this, compound 2d (Table 7)
shows more liver cleavage than intestinal cleavage in human,
monkey, and rat, but in the dog it is reversed.
Mouse DMPK Results and Discussion
A seriesof ester and amino acid derivatives were selected for
in vivo studies based on a combination of factors including
potency, plasma stability, and SAR. Primarily L-valine deri-
vatives were investigated because they had the most rodent
plasma stability. The in vivo work focused on oral dosing not
only because it is a desired route of administration in gene-
ral but also because HCV is a liver disease and absorbed
ProTide would pass directly into the liver via the portal vein.
For HCV, the general problem of first pass metabolism may
work in our favor to produce high levels of the active agent.
The overall goal would be to find a compound with low
systemic nucleoside exposure but good therapeutic liver tri-
phosphate levels.
Human S9 data were collected for the majority of project
compounds but only a portion of the data are reported here.
We were looking for compounds with good stability in human
intestinal S9 preparations, but with rapid cleavage in liver, in
the presence of cofactor. In addition to looking at the ratio
between liver and intestinal cleavage, we also paid attention
to the overall degree of cleavage. The majority of compounds
For all the ProTides, systemic levels of the parent ProTides
were not detectable in the mouse (data not shown). Even for
the L-valine derivatives, which demonstrated much improved

4954 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
McGuigan et al.
Table 8. Mouse Systemic Nucleoside Levels from 50 mg/kg PO Dose of ProTide
ProTide
C
max ng/mL
Clast ng/mL (h)
T1/2
h
AUC0-t ng h/mL
3
AUC0-inf ng h/mL
3
DNAUC
1c
2h
2n
3a
4c
4f
5203
262
1380
504
185
518
388
339
668
652
22 (24 h)
103 (6 h)
222 (6 h)
130 (6 h)
1.5 (24 h)
202 (6 h)
388 (6 h)
43 (24 h)
688 (6 h)
244 (6 h)
5.4
5.2
1.8
2.3
3.0
3.5
20390
1008
3722
1611
1664
1600
1650
3990
2590
2269
20567
1920
4269
2012
1670
2447
411
38
85
40
33
49
4g
4h
4i
>33
88
6.9
>6
3.0
4420
52
>45
4j
rodent plasma stability, ProTides were not observed in sys-
temic circulation.
Table 9. Comparison of 20-C-Methyl Guanosine Triphosphate Liver
Levels from 50 mg/kg PO Dose of ProTide
Table 8 reports the plasma levels of parent nucleoside (1c)
that result from a po dose of ProTide. Mice were orally dosed
at 50 mg/kg in an experimental formulation of 5% DMA,
20% Solutol HS 15, 20% PEG 400, 55% 50 mM sodium
acetate, pH 4.0. Plasma nucleoside levels were measured at
0.25, 0.5, 1, 2, 4, 6, and 24 h and are compared to the oral PK
of the parent nucleoside itself. The parent nucleoside (1c)
demonstrates only modest bioavailability from a 50 mg/kg
oral dose. However, the percent oral bioavailability, calcu-
lated from the po and iv (data not shown) DNAUCs values, is
quite high (F=85%). The dose normalized AUC (DNAUC)
Cmax Clast
ProTide ng/g (h) ng/mL (h)
AUC 0-6 AUC 0-24
h ng/g
h ng/g
3
36868
DNAUC
3
1c
2h
3470 (2)
370 (6)
390 (24)
370 (6)
1158 (6)
134 (16)
470 (6)
74 (24)
290 (6)
16510
1584
6980
2596
1940
815
737
32
2n, no. 1 1950 (4)
140
108
97
2n no. 2
670 (4)
512 (4)
275 (4)
290 (6)
216 (6)
225 (4)
91 (6)
5390
3860
3a
4c
4f
4g
4h
4i
77
18
880
166 (6)
920
18
for the nucleoside 1c is 411 ng h/mL/mg/kg, which is higher
3
4j
than the DNAUCs for any of the different ProTides. It should
be noted that the oral doses of the different compounds were
not molar equivalent doses, so the reported values are not
exact comparisons. However, it is clear from the Table 8 data
that all the ProTides result in lower systemic exposure of
parent nucleoside. The ProTides all have nucleoside DNAUCs
that are all 5-10-fold lower than parent nucleoside. This is
consistent with the substantially lower nucleoside Cmax values
that were also observed. It is considered possible that these
lower systemic exposures to the parent nucleoside may lead to
lower in vivo toxicity.
levels are achieved. Table 9 reports the 20-C-methyl guanosine
triphosphate levels in mouse liver. Liver levels were deter-
mined at0.25, 0.5, 1, 2,4, and 6h. For three compounds, 1c, 4c,
and 4h, the 24 h liver levels were also measured. The NTP
DNAUC values were calculated based on actual values not
projections to 24 h. This results in significantly lower DNAUC
values for several of the compounds with only 6 h data points.
Once again, these doses are nonequivalent on a molar basis,
and as a result the triphosphate levels of 1c are approximately
2-fold higher than if dosed at an equivalent level.
Comparing the ProTides to each other, the highest nucleoside
Cmax and DNAUC values belong to the L-alanine derivative 2n,
while the lowest come from the L-valine derivative 4c. The
selection of the ester group also makes a significant diffe-
rence in Cmax and DNAUC. For the two L-alanine derivatives
2h and 2n, the (S)-1-phenylethyl ester 2n resulted in a 5-fold
increase in Cmax and 2-fold increase in DNAUC over the
neopentyl ester 2h. A similar comparison in the L-valine series
is not possible because the (S)-1-phenyl ethyl ester in the L-valine
series was inactive and not tested. Comparison of the neopentyl
esters in the L-alanine, and L-valine series shows a higher
nucleoside Cmax and DNAUC for the L-valine (4j) than the
L-alanine (2h). The benzyl ester derivatives 3a, 4f, 4g, 4h, and 4i,
were the most potent compounds in the L-valine series. These
compounds along with the primary alcohol ester 4j, produced
the highest Cmax values for parent nucleoside. Interestingly, the
secondary alcohol derivative 4c has a much lower Cmax and over
all nucleoside exposure than the other L-valine esters.
As mentioned above, we have observed the nucleoside 1c to
have excellent oral bioavailabilty in the mouse (F = ∼85%).
It is rapidly absorbed, and its metabolism in mouse liver is
modest, leading to similar exposure from an oral dose as from
an iv dose. On the other hand, oral bioavailability in the cyno-
molgus monkey is poor (F ∼ 10%, data not shown). Clearly,
there are major differences in how the mouse and nonhuman
primates absorb and metabolize 20-C-methylguanosineand its
ProTides, and the mouse may not be an appropriate model for
studying 20-C-methylguanosine ProTide pharmacokinetics.
However, liver triphosphate levels can be measured in the mouse
at multiple time points so very valuable data can be obtained.
The liver triphosphate levels from the parent nucleoside in
the mouse were substantially higher than those from the Pro-
Tides, mimicking the systemic nucleoside levels. Even accoun-
ting for dosing differences, and extrapolating out the 6 h
DNAUC values to 24 h, all ProTides provide 2-5-fold less
liver triphosphate, as measured by DNAUC, than does the
nucleoside itself.
Although the very potent neopentyl L-alanine derivative 2h
showed low systemic levels of nucleoside, as was desired, because
of their inherent rodent plasma instability, the L-alanine
derivatives were not further studied in the mouse. Several of
the L-alanines were later studied in cynomolgous monkeys,
and their data will be reported elsewhere.
However, in considering the ratio of the triphosphate
DNAUC to the nucleoside DNAUC, several ProTides have
similar, to perhaps better, ratios than the nucleoside 1c. The
ProTides2n, 4a, and2chave triphosphate DNAUCtonucleo-
side DNAUC ratios of about2, similarto1c. FortheL-alanine
derivative 2h, or the L-valine derivative 3a, the DNAUC values
are calculated only using 6 h data. If the DNAUC values are
Although maintaining low systemic nucleoside levels is
desirable, it only makes sense if sufficient liver triphosphate

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4955
extrapolated out to 24 h, it is quite possible that the triphos-
phate to nucleoside ratio for these two compounds would be
better than that obtained for the nucleoside.
The best overall triphosphate exposure appears to come
from the (S)-1-phenyl ethyl ester of L-alanine phosphorami-
date (2n). This compound was examined in an additional
experiment out to 16 h. A certain amount of interexperiment
variation was observed due to the extreme difficulty of mea-
suring triphosphate levels in the liver. The DNAUC from the
and by MEDAC Ltd., Surrey. 20-C-Methyl guanosine 50-triphos-
phate was purchased from Carbosynth, Berkshire, UK.
Compound purity was assured by a combination of high field
multinuclear NMR (H,C,P) and HPLC. Purity by the latter was
always >95% with no detectable parent nucleoside, for all final
products.
Replicon Potency. Huh7 cells expressing the HCV genotype
1b bicistronic subgenomic relicon (Apath, LLC, Brooklyn, NY)
were seeded into white 96-well plates (Nunc/VWR) at a density
of 2 ꢀ 104 cells/well in medium without G-418. A Stacker
multidrop liquid dispenser (MTX Lab Systems, Vienna, VA)
was employed to ensure uniform and fast cell seeding into
multiple plates. Then 18-24 h after cell plating, inhibitors were
added and cells were incubated for additional 48 h. Compounds
were tested in triplicates and quadruplicates at 3ꢀ or 4ꢀ serial
dilutions over a range of 0.0001-10 μM concentrations. HCV
replication was monitored by Renilla luciferase reporter activity
assay using Renilla luciferase reporter (Promega, Madison, WI)
and a Veritas luminometer (Turner Biosystems, Sunnyvale,
CA). Then 50% inhibitory concentration (EC50) values were
calculated as the concentration of compound that results in 50%
decrease in the reporter expression as compared to untreated
cells. The values were determined by nonlinear regression (four-
parameter sigmoidal curve fitting) analysis. For any one IC50
determination, the replicon assay was run in triplicate and stan-
dard deviation was calculated from the three repeats and this
standard deviation is usually less than 20% of the IC50. A second
standard deviation is determined when multiple IC50s are
obtained from different batches of replicon cells. When more
than three independent IC50 values have been obtained, this
second standard deviation is calculated. Because of the diffe-
rences in the different batches of replicon cells, this standard
deviation is often ca. 100% of the IC50.
Cytotoxicity Assays. The same cells used in the replicon assay
were seeded into 96-well plates at a density of 2 ꢀ 104 cells per
well. Twenty-four h after cell plating, 11 serial 2ꢀ compound
dilutions, starting with 100 μM, were applied to the testing
plates (three repeats per compound dilution). Each testing plate
was run with a “no-compound” control. Incubation with com-
pounds was continued at 37 °C in a CO2 incubator for 72 h. The
cells were in log phase of growth throughout the cell cytotoxicity
experiments. To determine cell viability, the CellTiter-Glo assay
(Promega, Madison, WI) was performed according to the manu-
facturer’s protocol. The compound concentration resulting in
50% luminescent signal was reported as the CC50 concentration.
Plasma Stability. Stability experiments were performed in
duplicate using human, cynomolgus monkey, canine, rat, and
mouse plasma (Bioreclamation, Inc., Long island, NY). Pro-
Tides were added to a final concentration of 1 μg/mL in 1 mL of
plasma preincubated at 2-4 °C. The reaction mixture was
maintained at 2-4 °C, and 50 μL samples were taken at 30 or
240 min of incubation and transferred to a 96-well V-bottom
plate. Then 200 μL of ice cold acetonitrile was added to each
sample. The precipitated samples were centrifuged at 2500 rpm,
4 °C, for 20 min in a Sorvall RT6000S centrifuge (Thermo
Scientific, Waltham MA). Then 50 μL of supernatant from each
sample was transferred to a 96-deep well plate followed by the
addition of 50 μL of H2O to each sample. Samples were covered,
mixed well by vortexing, and maintained at 2-8 °C before and
during analysis. Fifteen μL of each test sample was analyzed for
compound concentrations by LC-MS/MS. Liquid chromatog-
raphy was performed with an Agilent 1100 series HPLC system
equipped with a Synergi 4 μm Polar-RP, 30 mm ꢀ 2.0 mm
column (Phenomenex, Torrance, CA). Linear gradient (100%
mobile phase A (H2O þ 0.1% HCOOH) to 100% B (acetonitrile
þ 0.1% HCOOH) over 3 min, flow rate 1.0 mL/min) was used
for the analyte elution. The HPLC system was coupled to an
API 4000 triple quadrupole mass spectrometer (Applied Bio-
systems, Framingham, MA). Mass spectrometry analysis was
performed in positive ion mode (MHþ 659.2, transition 166.1).
second experiment was 140 ng h/mL/mg/kg. Nucleoside sys-
3
temic exposure was also measured in the second experiment,
and the DNAUC for nucleoside was 71 ng h/mL/mg/kg.
3
In summary of the mouse DMPK work, the ProTides
produce significantly lower liver triphosphate levels than the
parent nucleoside, however, several compounds could be
identified with improved ratios of liver triphosphate to sys-
temic nucleoside levels. The mouse is considered a nonoptimal
model for studying ProTides of 20-C-methyl guanosine be-
cause of the observed rapid breakdown of the ProTides in
the mouse and because of the good absorption and meta-
bolic stability of the resulting 20-C-methyl guanosine. How-
ever, very useful data were obtained comparing different
ProTides to each other and parent nucleoside. Further work
on ProTides and other pro-drugs of 20-C-methyl guanosine is
underway in the cynomolgus monkey, and will be reported
elsewhere.
Conclusions
We have reported herein an extensive ProTide study of
20-C-methylguanosine, with variations in the aryl, ester, and
amino acid region of the ProTide. In almost every case, the
ProTides are more active in the HCV replicon assay, than the
parent nucleoside, often by 10-30-fold. We have extensively
studied the compounds for stability in plasma from multiple
species and also in liver and intestinal preparations from seve-
ral species. A combination of ester and amino acid changes
gives rodent plasma-stable compounds. Several analogues
were further evaluated for oral PK in mouse and reveal rapid
uptake and metabolism to triphosphate in the liver. The data
are somewhat suggestive of liver targeting by these ProTides.
This, and their inherent potency and selectivity against HCV
in vitro, suggests that their continued preclinical evaluation is
warranted.
Experimental Section
General. Anhydrous solvents were purchased from Aldrich
and used without further purification. All reactions were carried
out under an argon atmosphere. Reactions were monitored with
analytical TLC on silica gel 60-F254 precoated aluminum plates
and visualized under UV (254 nm) and/or with 31P NMR
spectra. Column chromatography was performed on silica gel
(35-70 μM). Proton (1H), carbon (13C), and phosphorus (31P)
NMR spectra were recorded on a Bruker Avance 500 spectro-
meter at 25 °C. Spectra were autocalibrated to the deuterated
solvent peak and all 13C NMR, 31P NMR were proton-decoupled.
Analytical and semipreparative HPLC were conducted by Varian
Prostar (LC Workstation-Varian prostar 335 LC detector) using
Varian Polaris C18-A (10 μM) as an analytic column and Varian
Polaris C18-A (10 μM) as a semipreparative column; elution was
performed using a mobile phase consisting of water/acetonitrile
in gradient (system 1, 100/0 to 0/100 v/v in 30 min) or water/methanol
(system 2, 100/0 to 0/100 v/v in 30 min). High resolution mass
spectra was performed as a service by Cardiff University, using
electrospray (ES). CHN microanalysis were performed as a
service by the School of Pharmacy at the University of London

4956 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
McGuigan et al.
Data was analyzed using Analyst v1.4.2 software (Applied
Biosystems, Framingham, MA), Microsoft Office Excel 2007,
and GraphPad Prism 5 (GraphPad Software, La Jolla, CA
S9 Clearance Assay. All liver and intestinal S9 tissue fractions
were purchased from Xenotech (Lenexa, KS) and stored at -80 °C
until time of use. NRS (NADPH-Regenerating System) was
purchased from BD Gentest (BD Biosciences, San Jose, CA)
and was used as cofactor. Reagents including NRS components
and liver or intestinal S9 tissue fractions were thawed and
immediately placed on ice. S9 fractions, 50 mM potassium
phosphate buffer with or without NRS were placed in wells of
a 96-well deep-well plate for final concentrations of 1 mg/mL S9,
1.3 mM NADP, 3.3 mM glucose-6-phosphate, 0.4 U/mL glu-
cose-6-phosphate dehydrogenase, and 3.3 mM magnesium
chloride. The plate was preincubated in an orbital shaker at 37 °C
for 10 min. The reaction was then initiated by addition of com-
pound at a final concentration of 10 μM. The suspensions were
thoroughly mixed by pipetting and an initial sample was with-
drawn and added to equal volume of cold acetonitrile (for a
calculation of starting concentration). The plate was placed
back in the 37 °C shaker, and at 60 min a sample was withdrawn
and added to equal volume cold acetonitrile. Following quench-
ing by acetonitrile, samples were kept on ice and precipitate
was pelleted in all samples by centrifugation at 3000 rpm for
15 min. Supernatant was removed and transferred to vials for
HPLC analysis.
Pharmacokinetic Studies. Compounds were formulated in
5% dimethylacetamide, 20% Solutol HS15, 20% polyethylene
glycol 400, 55% 50 mM sodium acetate, (pH 4) (D/S/P) and
administered by oral gavage to female ICR mice (Taconic
Farms, Germantown, NY). Blood and tissue collections were
performed at various time points as terminal procedures. Blood
was collected into EDTA containing tubes (1.6 mg EDTA/mL
blood, Sarstedt, Inc., Newton, NC), and the plasma was sepa-
rated by centrifugation within 30 min of collection. Liver sam-
ples were immediately frozen upon collection in liquid nitrogen.
Plasma and liver samples were stored frozen at e -80 °C prior
to analysis.
Measurement of Plasma Nucleoside Concentrations. Plasma
samples were prepared for analysis as follows. First, 400 μL of
50 mM ammonium acetate was added to 100 μL of each plasma
sample. Calibration curves were prepared by serial dilution of a
stock solution of 20-C-methyl guanosine (1c) into control plas-
ma. Solid phase extraction of the samples was performed with
H2O-Philic speeddisk columns (J.T. Baker) which were pre-
viously solvated with 1 mL of methanol followed by equilibra-
tion with 1 mL of 50 mM ammonium acetate. The columns were
rinsed with 1 mL of 50 mM ammonium acetate:methanol (95:5,
v/v), and samples were eluted with 1 mL of methanol:ammo-
nium hydroxide (95:5, v/v). The samples were dried under
nitrogen and reconstituted in 80 μL of H2O. A volume of 10 μL
of each test sample was analyzed for 20-C-methyl guanosine (1c)
concentrations by liquid chromatography coupled to tandem
mass spectrometry (LC-MS/MS). Liquid chromatography was
performed with an Agilent 1100 series HPLC system equipped
with a Betasil 2.1 mm ꢀ 100 mm 5 μm column (Thermo Scientific).
Mass spectrometry analysis was in positive ion mode as des-
cribed above.
Measurement of Liver Triphosphate Concentrations. Liver
samples were snap-frozen in liquid nitrogen and pulverized with
a steel mortar and pestle on dry ice with liquid nitrogen. Frozen,
pulverized tissue samples were weighed and 5 equiv of 70% metha-
nol containing 50 mM thenoyltrifluoroacetone was added.
Samples were then homogenized twice in a bead mill (FastPrep
Homogenizer) using Silicone-Carbide Sharp Particles (BioSpec
Products, Inc., Bartlesville, OK) at 4 °C at an agitating speed of
6 for 30 s. The homogenates were centrifuged at 15000 rpm for
30 min at 4 °C. The supernatants were collected and dried in
Masterblock polypropylene deep-well plates (SPEware Corp.,
Baldwin Park, CA) under a stream of nitrogen. The dried extracts
were reconstituted with 100 μL of 10 mM N,N-dimethylhexyl-
amine, 3 mM ammonium formate in H2O, vortexed, and cen-
trifuged at 3500 rpm for 30 min at 4 °C. Calibration curves were
constructed by spiking varying concentrations of 20-C-methyl
guanosine 50-triphosphate (Carbosynth, Berkshire, UK) into
control liver samples prior to homogenization. Fifteen μL of each
test sample was analyzed for 20-C-methyl guanosine 50-triphos-
phate concentrations by liquid chromatography coupled to
tandem mass spectrometry (LC-MS/MS). Reverse phase liquid
chromatography was performed with an Agilent 1100 series
HPLC system equipped with an. XTerra MS C18, 3.5 μM,
2.1 mm ꢀ 50 mm Column (Waters, Milford, MA). Mass spec-
trometry analysis was performed in negative ion mode and data
were analyzed as above. The levels (peak areas) of endogenous
adenosine 50-triphosphate and guanosine 50-triphosphate were
also monitored as internal quality controls for the liver samples.
Pharmacokinetic Analysis. Plasma and liver concentration
data was analyzed for standard pharmacokinetic parameters
using Win-Nonlin v5.2 software (Pharsight, St. Louis, MO).
Standard Procedure A: ProTide Synthesis. To the 20,30-O,O-
isopropylidene-20-C-β-methylguanosine (1 equiv) in anhydrous
THF (6 mL/g of nucleoside) was added dropwise a solution of
tBuMgCl in THF (1M, 2 equiv). After stirring for 20 min, a
solution of the phosphorochloridate (2 equiv) in anhydrous
THF (6 mL/g of phosphorochloridate) was slowly added. The
resulting solution was stirred overnight at room temperature,
then the solvent was removed under reduced pressure. The resul-
ting foamy residue was purified by silica gel column chromato-
graphy (eluent: CHCl3/MeOH 92:8, v/v) to afford the pure
protected ProTide.
Standard Procedure B: ProTide Deprotection. A solution of
the protected ProTide in a 60% acetic acid in water solution
(∼10 mL/200 mg of protected protide) was stirred at 95 °C
overnight. The solvent was then removed under reduced pres-
sure to dryness (can also be coevaporated with hexane or toluene
to remove traces of AcOH), and the resulting residue was puri-
fied by silica gel column chromatography (eluent: CHCl3/MeOH
9:1, v/v) to give after lyophilization the pure ProTide.
Example: Synthesis of 20-C-Methylguanosine-50-O-[naphthyl-
(methyloxy-L-alaninyl)]phosphate (2a). Step 1: Synthesis of the
protected ProTide
Prepared using standard procedure A. Starting from 181 mg of
protected nucleoside. Yield: 37%.
1
31P NMR (202 MHz, CD3OD) δ 4.49, 4.40. H NMR (500
MHz, CD3OD) δ 8.17 (2d, J=8.0, 1H, H8-napht), 7.85 (m, 2H,
H5-napht and H8), 7.70 (m, 1H, H4-napht), 7.58-7.48 (m, 3H,
H7, H6, H2-napht), 7.42 (m, 1H, H3-napht), 6.08 and 6.07 (s, 1H,
0
0
0
0
H1 ), 4.62-4.54 (m, 2H, H3 and H5 ), 4.48-4.36 (m, 2H, H4 and
00
H5 ), 4.09 (m, 1H, HR), 3.61 (s, 3H, CH3 ester), 1.59 and 1.57 (2s,
3H, CH3 isopropylidene), 1.36 (m, 6H, CH3 isopropylidene and
CH3 Ala), 1.03 and 1.00 (s, 3H, 20-CH3).
Step 2: Deprotection of the ProTide
Prepared using standard procedure B. Starting from 70.5 mg.
Yield: 45% (30 mg).
1
31P NMR (202 MHz, CD3OD) δ 4.36, 4.27. H NMR (500
MHz, CD3OD) δ 8.18 (m, 1H, H8-napht), 7.93-7.83 (m, 2H,
H5-napht and H8), 7.69 (m, 1H, H4-napht), 7.56-7.49 (m, 3H,
0
H7, H6, H2-napht), 7.40 (m, 1H, H3-napht), 5.94 (s, 1H, H1 ),
0
00
0
0
4.57 (m, 2H, H5 and H5 ), 4.29-4.23 (m, 2H, H4 and H3 ),
4.10-3.98 (m, 1H, HR), 3.56 and 3.55 (2s, 3H, CH3 ester), 1.30
(m, 3H, CH3 Ala), 1.01 and 0.99 (2s, 3H, 20-CH3). HPLC, system
1, RT 13.00, 13.43 min.
Acknowledgment. We thank Helen Murphy for secretarial
assistance.
Supporting Information Available: Preparative methods,
spectroscopic and analytical data on target compounds. This
material is available free of charge via the Internet at http://
pubs.acs.org.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4957
References
(9) McGuigan, C.; Perrone, P.; Madela, K.; Neyts, J. The phosphor-
amidate ProTide approach greatly enhances the activity of beta-20-
C-methylpurines against Hepatitis C Virus. Bioorg. Med. Chem.
Lett. 2009, 19, 4316–4320.
(10) Furman, P. A.; Wang, P. Y.; Niu, C. R.; Bao, D. H. Symonds, W.;
Nagarathnam, D.; Steuer, H. M.; Rachakonda, S.; Ross, B. S.;
Otto, M. J.; Sofia, M. PSI-7851: A novel liver-targeting nucleotide
prodrug for the treatment of Hepatitis C. Hepatology 2009, 48
(Suppl), presented at 59th Annual Meeting of the American Association
for the Study of Liver Diseases, Oct 31-Nov 4, 2008, San Francisco,
CA, poster 1901.
(11) Birkus, G.; Kutty, N.; He, G.-X.; Mulato, A.; Lee, W.; McDermott,
M; Cihlar, T. Activation of GS-7340 and other Tenofovir phos-
phonoamidate prodrugs by human proteases. Antiviral Res. 2007,
74, A57.
(12) Notably, the agent has recently been resolved into its separate
diastereoisomers. See: Chun, B. K.; Rachakonda, S.; Du, J. F.;
Khan, N.; Shi, J. X.; Stee, W.; Cleary, D.; Ross, B. S.; Sofia, M. J.
An efficient and diastereoselective synthesis of PSI-6130: a clini-
cally efficacious inhibitor of HCV NS5BpPolymerase. J. Org.
Chem. 2009, 74, 6819–6824.
(1) Hoofnagle, J. H.; Seef, L. B. Peginterferon and ribavirin for chronic
hepatitis C. N. Engl. J. Med. 2006, 355, 2444–2451.
(2) Soriano, V.; Madejon, A.; Vispo, E.; Labarga, P.; Garcia-Sama-
niego, J.; Martin-Carbonero, L.; Sheldon, J.; Bottecchia, M.;
Tuma, P.; Barreiro, P. Emerging drugs for hepatitis C. Expert
Opin. Emerging Drugs 2008, 13, 1–19.
(3) Brown, N. A. Progress towards improving antiviral therapy for
hepatitis C with hepatitis C virus polymerase inhibitors. Part 1:
Nucleoside analogues. Expert Opin. Invest. Drugs 2009, 18, 709–
725.
(4) Klumpp, K.; Leveque, V.; Le Pogam, S.; Ma, H.; Jiang, W.-R.;
Kang, H.; Granycome, C.; Singer, M.; Laxton, C.; Hang, J. Q.;
Sarma, K.; Smith, D. B.; Heindl, D.; Hobbs, C. J.; Merrett, J. H.;
Symons, J.; Cammack, N.; Martin, J. A.; Devos, R.; Najera, I. The
novel nucleoside analog R1479 (40-azidocytidine) is a potent
inhibitor of NS5B-dependent RNA synthesis and hepatitis C virus
replication in cell culture. J. Biol. Chem. 2006, 281, 3793–3799.
(5) Eldrup, A. B.; Prhavc, Ml.; Brooks, J.; Bhat, B.; Prakash, T. P.;
Song, Q.; Bera, S.; Bhat, N.; Dande, P.; Cook, P. D.; 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.; Langen,
D. V.; Wolanski, B.; Olsen, D. B. Structure-Activity Relationship
of heterobase modified 20-C-methyl ribonucleosides as inhibitors of
hepatitis C virus RNA replicaton. J. Med. Chem. 2004, 47, 5284–
5297.
(13) McGuigan, C.; Sutton, P. W.; Cahard, D.; Turner, K.; O’Leary,
G.; Wang, Y.; Gumbleton, M.; De Clercq, E.; Balzarini, J. Synthe-
sis, anti-human immunodeficiency virus activity and esterase labi-
lity of some novel carboxylic ester-modified phosphoramidate
derivatives of stavudine (d4T). Antivir. Chem. Chemother. 1998,
9, 473–479.
ꢀ
(14) Zhou, X. J.; Pietropaolo, K.; Sullivan-Bolyai, J.; Kuca, B.; Liu, W.;
(6) Cahard, D.; McGuigan, C.; Balzarini, J. Aryloxy phosphorami-
date Triesters as Pro-Tides. Mini-Rev. Med. Chem. 2004, 4, 371–
382.
(7) McGuigan, C.; Cahard, D.; Sheeka, H. M.; De Clercq, E.; Balzarini,
J. Aryl phosphoramidate derivatives of d4T have improved anti-
HIV efficacy in tissue culture and may act by the generation of
a novel intracellular metabolite. J. Med. Chem. 1996, 39, 1748–
1753.
Xu, L.; Belanger, B.; Khan S.; Mayers. D. IDX184, A liver-targeted
nucleotide HCV polymerase inhibitor: Results of a first-in-man
safety and pharmacokinetic study. Presented at EASL, 25th April
2009, poster no. 966.
(15) Derudas, M.; Carta, D.; Brancale, A.; Vanpouille, C.; Lisco, A.;
Margolis, L.; Balzarini, J.; McGuigan, C. The application of
phosphoramidate ProTide technology to Acyclovir confers anti-
HIV inhibition. J. Med. Chem. 2009, 52, 5520–5530.
(8) McGuigan, C.; Kelleher, M. R.; Perrone, P.; Mulready, S.; Luoni,
G.; Daverio, F.; Klumpp, K.; Smith, D. B. The application of
phosphoramidate ProTide technology to the potent anti-HCV
aompound 40-azidocytidine. Bioorg. Med. Chem. Lett. 2009, 19,
4250–4254.
(16) ClogP values were calculated using ChemOffice Ultra 11.0.
(17) McGuigan, C.; Tsang, H. W.; Cahard, D.; Turner, K.; Velazquez,
S.; Salgado, A.; Bidois, L.; Naesens, L.; De Clercq, E.; Balzarini, J.
Phosphoramidate derivatives of d4T as inhibitiors of HIV: the
effect of amino acid variation. Antiviral Res. 1997, 35, 195–204.