Bicyclic nucleoside inhibitors of Varicella-Zoster virus: The effect of branching in the p-alkylphenyl side chain

Article abstract of DOI:10.1016/j.bmcl.2005.05.071

Further to the discovery of bicyclic furanopyrimidine nucleoside analogues (BCNAs) as potent anti-VZV agents, a branched series of this family of compounds was synthesised. The aim was to study the impact of the geometry and steric hindrance in the side c

Full text of DOI:10.1016/j.bmcl.2005.05.071

Bioorganic & Medicinal Chemistry Letters 15 (2005) 3791–3796
Bicyclic nucleoside inhibitors of Varicella–Zoster virus:
The effect of branching in the p-alkylphenyl side chain
Giovanna Luoni,a Christopher McGuigan,a,* Graciela Andrei,b Robert Snoeck,b
Erik De Clercqb and Jan Balzarinib
aWelsh School of Pharmacy, Cardiff University, King Edward VII Avenue, Cardiff CF10 3XF, UK
bRega Institute for Medical Research, Katholieke Universiteit Leuven, B-3000 Leuven, Belgium
Received 8 March 2005; revised 6 May 2005; accepted 11 May 2005
Available online 29 June 2005
Abstract—Further to the discovery of bicyclic furanopyrimidine nucleoside analogues (BCNAs) as potent anti-VZV agents, a
branched series of this family of compounds was synthesised. The aim was to study the impact of the geometry and steric hindrance
in the side chain as well as lipophilic role of this moiety on biological activity. The results showed a detrimental effect of branching
on antiviral activity, with a different magnitude depending on the position of branching in the side chain. This study again showed
the importance of this moiety for biological activity, as well as the limited efficacy of the ClogP value as a tool for predicting the
potency of BCNAs, while suggesting an alternative rationale behind the design of future series.
Ó 2005 Elsevier Ltd. All rights reserved.
We have already reported the discovery of an entirely new
class of potent antiviral deoxynucleoside analogues bear-
ing an unusual bicyclic base.1 These bicyclic furanopyrim-
idines have revealed to be potent and selective inhibitors
of Varicella–Zoster virus replication and several efforts
have been made in our group to investigate the structural
requirements needed for optimal antiviral activity.
Modifications of the side chain, the base and the sugar
moiety have been considered.2 In particular, previous
SAR studies have revealed as an absolute requirement
for antiviral activity—a long lipophilic alkyl or alkyl-
phenyl side chain; in fact, modifications involving a sig-
nificant increase of the polarity in the side chain
decreased the activity of the corresponding analogues.
Figure 1. Bicyclic furanopyrimidines (BCNAs) belonging to the most
active series and their corresponding ClogP values.
As a result, the potency of these nucleosides seems to
be strictly related to their ClogP values, with an optimal
value falling between 2.5 and 3.5.
under development for use in shingles by FermaVir
Pharmaceuticals.
Among all the derivatives synthesised, the most active
series reported so far is the one bearing a phenyl group
in the side chain (Fig. 1). In particular the p-pentylphe-
nyl derivative (4) displays an EC50 below 1 nM and a
selectivity index value of ca. 1,000,000.3 It is currently
To understand the role of the alkyl group in biological
activity, we here report the synthesis and biological eval-
uation of a new class of BCNA derivatives characterised
by a branched side chain (Fig. 2). This represents the
first example of branched p-alkylphenyl BCNAs.
These modifications would be of interest for investigat-
ing the effect of steric hindrance and conformational
restriction and also of the, previously predictive, ClogP
Keywords: VZV; BCNA; Nucleoside analogues; ClogP; Shingles.
*
Corresponding author. Tel./fax: +44 2920 874537; e-mail:
mcguigan@cardiff.ac.uk
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.05.071

3792
G. Luoni et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3791–3796
Figure 2. Target derivatives and their corresponding ClogP value.
Figure 3. Synthesis of EDU (17). Reagents and conditions: (i) HC„CSiMe3, DIPEA, Pd(Ph3P)4, CuI, DMF, rt, overnight, 81%; (ii) NH4OH,
MeOH, rt, overnight, 75%.
value. Compound 14 was particularly indicative for our
purposes, having the same ClogP value as the most ac-
tive p-pentylphenyl BCNA (4).
uridine (17) as the synthon for subsequent coupling with
appropriate aryl iodides (Fig. 3).
Thus, the desired aryl iodides were synthesised via an
aromatic iodination of the appropriate alkylbenzene in
the presence of 2 equiv of I2 and 2 equiv of AgNO2,4
which led to compounds in good yield as a mixture of
ortho and para regioisomers, the ratio of which was
The general synthetic pathway for the synthesis of bicy-
clic furanopyrimidines (BCNAs) involves a Pd-catalysed
coupling between 5-iodo-20-deoxyuridine (IDU, 15) and
the appropriate aryl acetylenes.1 Unfortunately, the aryl
acetylenes required for the synthesis of these new ana-
logues were not commercially available.
1
determined by H NMR (Fig. 4).
Due to the fact that it was not possible to separate the
regioisomers by flash chromatography, the aryl iodides
were used as mixtures for the following synthetic steps.
Therefore, we considered the building of the alkynyl
moiety on the nucleoside, leading to 5-ethynyl-20-deoxy-
Figure 4. Synthesis of the required aryl iodides (18–22), corresponding yields and para/ortho ratios.

G. Luoni et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3791–3796
3793
The alkylphenyl starting materials necessary for the syn-
thesis of compounds 20 and 22 were not commercially
available and they had to be synthesised.
The product 28 was successfully isolated and NMR
analysis excluded the presence of any side product,
which may have resulted from the attack of any residual
tert-BuLi on styrene.
As reported in the literature by Wei and Taylor,5 styrene
faces attack by a range of organolithium reagents and
subsequent treatment of the intermediate anion with
MeOH gives access to several alkylbenzene analogues.
This approach was chosen for the synthesis of the de-
sired (3,3-dimethyl)butylbenzene (25) (Fig. 5).
Converse to the iodination with molecular iodine, a dif-
ferent synthetic pathway was preferred for the last out-
standing p-iodocyclohexylbenzene (30), and this was
achieved in 57% yield, starting from cyclohexylaniline
via formation of the corresponding diazonium salt
(Fig. 7).7
The same synthetic pathway was chosen for the prepara-
tion of (4-methyl)pentylbenzene (28). As shown in
Figure 6 the required iso-propyl lithium 29 was prepared
by reacting iso-propyl iodide with excess tert-BuLi in
diethyl ether at ꢀ78 °C. Such an exchange is indicated
to be irreversible in the presence of 2 equiv of tert-BuLi.6
After generation of the desired primary alkyl lithium
derivative, the residual tert-BuLi was eventually elimi-
nated by proton abstraction from diethyl ether, by sim-
ply allowing the reaction to stand at room temperature
for ca. 1 h. The reaction mixture then reacted with sty-
rene to afford 28, following the previously described
procedure.
Once all the necessary p-alkylaryliodides were success-
fully synthesised, they were subsequently coupled with
EDU (17) in the presence of Pd(Ph3P)4, CuI and trieth-
ylamine to achieve the corresponding 5-arylethynyl-20-
deoxyuridines (31–38). A copper-catalysed cyclisation
then led to the formation of bicyclic furanopyrimidine
ring derivatives (7–14) (Fig. 8).8
While 5-arylethynyl-20-deoxyuridines 31 and 38 were
isolated and characterised, the two synthetic steps were
carried out in a one-pot procedure for the synthesis of
other cyclised compounds in the series, and yields dis-
played in Figure 8 refer to EDU as the starting material.
The key spectroscopic evidence for the cyclisation is the
disappearance of the NH peak in the 1H NMR (ca.
11.6 ppm) and the presence of a H-5 peak in the region
7.10–7.30 ppm, depending on the aryl substituent.
Compared to the yields achieved for 7 and 11 (respec-
tively, 63% and 52% as overall yields from EDU), we
observed a decrease in the efficiency for the one-pot pro-
cedure. Besides the inevitable loss of material due to a
more difficult purification, these lower yields could be
due to the formation of 5,6-disubstituted analogues as
side products.
Figure 5. Synthesis of (3,3-dimethyl)butylbenzene (25). Reagents and
conditions: (i) tert-BuLi, Et2O, ꢀ78 °C, 30 min; (ii) MeOH, ꢀ78 °C,
30 min, 86%.
As shown in Figure 9, during the synthesis of 13, the
corresponding 5,6-disubstituted analogue 39 was, in
fact, isolated and characterised. Analogous fluorescent
compounds at similar Rfs were detected by TLC in all
the other reaction mixtures carried out in one pot.
These disubstituted derivatives are believed to be the
result of a side reaction occurring during the one-pot
cyclisation. Indeed, the suggested mechanism for their
formation is predicted to involve the reactivity of the
5-arylethynyl nucleoside intermediates, the triple bond
of which would react with the Pd-complex formed with
the remaining excess of aryl iodide. Given this, the
isolation of the 5-arylethynyl intermediates would pre-
vent the formation of these side products, as supported
by the higher yields obtained for 7 and 11.
Figure 6. Synthesis of 4-methylpentylbenzene (28). Reagents and
conditions: (i) tert-BuLi, Et2O, ꢀ78 °C, 10 min, rt, 1 h; (ii) styrene (23),
Et2O, ꢀ25 °C, 30 min; iii) MeOH, ꢀ25 °C, rt, 30 min, 57%.
Figure 7. Synthesis of p-iodocyclohexylbenzene (30).

3794
G. Luoni et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3791–3796
Figure 8. Synthesis of branched derivatives (7–14). Reagents and conditions: (i) ArI, Et3N, Pd(Ph3P)4, CuI, DMF, rt, overnight; (ii) MeOH, CuI,
reflux, 4 h.
Figure 9. Synthesis of the target structure 13. The 5,6-disubstituted analogue 39 was also isolated.
Furthermore, the fact that the synthetic aryl iodides 18–
20 and 22 were used as a mixture of ortho and para
regioisomers for subsequent coupling with EDU could
also have contributed to the low yields in the coupling
reaction. Although we failed to isolate any ortho bicyclic
furanopyrimidine analogue, the possibility that the 5-(o-
alkylaryl)ethynyl derivatives had been formed cannot be
ruled out. As a matter of fact, a parallel ongoing work
has found that the cyclisation step of some ortho-substi-
tuted 5-(alkylaryl)ethynyl analogues appears to require
more severe conditions to achieve good yields, while
no significant difficulties had been encountered at the
stage of coupling. On account of this, the amount of
the cyclised ortho isomer may not have been consider-
able. Yet, taking into account the relatively poor yields
of the whole process for 8–10 and 13, it seems likely that
these conditions could have affected, to some extent, the
yields of the coupling reactions. All of the spectroscopic

G. Luoni et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3791–3796
3795
Table 1. Antiviral activities for compounds 7–14; samples were dissolved in neat DMSO and diluted with biological media to generate ca. 50–200 lM
stock solutions containing 1% DMSO
ClogPb
EC50 (lM)
MCCd (lM)
CC50 (lM)e
a
Compound
R
VZV OKA
VZV YS
TKꢀVZV 07/1c
TKꢀVZV YS/Rc
7
CH(CH3)2
1.78
2.31
2.84
3.37
2.18
2.71
3.37
2.97
6.9
0.048
0.0017
>80
7.9
0.051
0.0011
NDf
>20
NDf
0.31
3.1
>20
>200
>50
>80
>50
>80
P80
>20
>50
>2
P50
>200
P50
400
45
>200
>200
>200
>200
140
8
CH2CH(CH3)2
C2H4CH(CH3)2
C3H6CH(CH3)2
C(CH3)3
9
>2
10
11
12
13
14
NDf
>50
NDf
NDf
>20
>50
>80
P50
400
CH2C(CH3)3
C2H4C(CH3)3
C6H11
0.33
1.4
P80
>20
>200
>200
ACV
1
—
—
1.5
0.09
3.1
0.07
40
>50
>20
>5
53
>20
>20
>5
>200
200
>400
123
188
C2H5
C3H7
C4H9
C5H11
C6H13
C7H15
Ph
1.38
1.91
2.44
2.97
3.50
4.0
2
0.01
0.008
0.0005
0.0001
0.0001
0.003
0.032
P50
P20
P20
20
3
0.0022
0.0003
0.0005
0.0054
0.031
>200
>200
18
4
>20
>5
>5
>5
5
6
40
>5
>5
>5
>5
5
>200
18
>200
2.24
a EC50, effective concentration (lM), required to reduce virus plaque formation by 50%.
b Values calculated using ClogP version 1.0.0. Biobyte, P.O. Box 517, Claremont, CA 91711, USA.
c TKꢀ, thymidine kinase deficient.
d MCC, minimal cytotoxic concentration (lM), required to alter microscopically detectable cell morphology.
e CC50, 50% cytotoxic concentration, required to inhibit Hel cell growth by 50%.
f ND, not determined.
data confirm the isolated products 7–14 to be pure para
isomers.
the iso-propyl derivative 7 and n-propylphenyl BCNA
2 (ca. 700-fold). This suggests a detrimental effect of
branching at the position a to the phenyl group, sup-
ported further by the inactivity of 11.
The synthesised compounds 7–14 were evaluated for their
ability to inhibit the replication of VZV in cell culture.
Table 1 contains data relating to two strains of thymidine
kinase-competent (TK+) VZV and also two strains of
thymidine kinase-deficient (TKꢀ) VZV, with data also
given for aciclovir (ACV) as a reference compound.
When we compare the results for 10 and 13 to that of n-
hexylphenyl BCNA (5), we observe a striking difference
in activity between compounds that have comparable
ClogP values. On this basis, the result for the cyclohexyl
derivative 14 is particularly interesting. Indeed, although
14 and the extremely potent n-pentylphenyl (4) deriva-
tives have the same ClogP value, they show a difference
in activity of 2800-fold. It is also of interest to consider
the result for the biphenyl derivative 40, which displays
an antiviral activity that remains between that of 4 and
14, yet we find that the ClogP values of the compounds
are only slightly different.
Data corresponding to the p-alkylphenyl series (1–6) are
also included in the table, as well as the results related to
compound 40, which has a biphenyl group as the side
chain. Cytotoxicity data are given for each compound
in two assays.
In general, all the compounds show an activity lower
than those of the potent n-alkylphenyl BCNA homo-
logues, indicating a negative effect of the branching.
This is particularly evident in the tert series, where the
only active derivative is 13, while 11 and 12 did not show
any activity at the highest subtoxic concentrations tested
(50 and 80 lM, respectively). For the iso and tert series,
there seems to be a tendency for an increase of activity
with the increase of the chain length. Indeed, in the iso
series, the potency increases by ca. 140-fold from 7 to
8 and by ca. 30-fold from 8 to 9. Similarly, in the tert
series, while 11 and 12 did not show any significant
activity, 13 is endowed with an antiviral potency higher
than that of aciclovir. In this context, the loss of activity
of the iso-hexyl derivative (10) is particularly surprising
and it indicates that the branching at the C-5 position on
the alkylphenyl side chain is not tolerated. The fact that
the branching could have a different influence depending
on its position in the alkylphenyl side chain can be sup-
ported further by a large difference in activity between
Although these losses of activity could be explained by
difference in the geometry and steric hindrance of the side
chain between the derivatives, the reliability of compari-
son between the ClogP values of bicyclic nucleoside ana-
logues with different structures could also be questioned.
To assess the possibility to use the ClogP value for
predicting the optimal antiviral activity of BCNAs for
different series, it was decided to carryout an experimen-
tal measurement of their logP values, choosing the
n-pentylphenyl 4 as the reference.
To carry out the assay, equal amounts of water and n-
octanol (99+%) were mixed overnight to attain a mutual
saturation. After separation, an appropriate amount of
4 was dissolved in n-octanol at a concentration lower
than 1 mM. However, when the water was added
and the two phases were mixed by shaking for 6 h,

3796
G. Luoni et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3791–3796
2. McGuigan, C.; Brancale, A.; Barucki, H.; Srinivasan, S.;
Jones, G.; Pathirana, R.; Carangio, A.; Blewett, S.; Luoni,
G.; Bidet, O.; Jukes, A.; Jarvis, C.; Andrei, G.; Snoeck, R.;
De Clercq, E.; Balzarini, J. Antiviral Chem. Chemother. 2001,
12, 27.
3. McGuigan, C.; Barucki, H.; Blewett, S.; Carangio, A.;
Erichsen, J. T.; Andrei, G.; Snoeck, R.; De Clercq, E.;
Balzarini, J. J. Med. Chem. 2000, 43, 4993.
4. Sy, W. W.; Lodge, B. A. Tetrahedron Lett. 1989, 30, 3769.
5. Wei, X.; Taylor, R. J. K. J. Chem. Soc., Chem. Commun.
1996, 2, 187.
6. Bailey, W. F.; Punzalan, E. R. J. Org. Chem. 1990, 55,
5404.
UV spectroscopy did not show any significant difference
in absorbance from the octanol solution before and after
the extraction with water, indicating that no detectable
amount of compound was transferred through the
aqueous phase.
An aliquot of the aqueous layer was also analysed, but
no UV spectrum could be detected.
In view of these results, 1 mg of 4 was first suspended in
100 ml water, and left stirring for 6 h, and occasionally
sonicated. The suspension was filtered giving a clear
solution. However, when an aliquot of this aqueous
solution was analysed, no significant absorbance could
be detected and the UV spectrum could not be
registered. Therefore, due to the pronounced lipophilic-
ity of the compound, the experimental measurement of
the logP value of 4 could not be carried out using this
method and the available software for the calculation
of logP remains the most convenient tool to estimate
the logP value of such derivatives.
7. Kaszynski, P.; Sienkowska, M.; Benin, V. Tetrahedron
2000, 56, 165.
8. General experimental procedure for the one-pot coupling and
cyclisation given for compound (8). To a stirred solution of
17 (450 mg, 1.77 mmol) in anhydrous dimethylformamide
(10 ml), at room temperature and under a nitrogen atmo-
sphere, triethylamine (0.27 ml, 1.95 mmol), 18 (554.5 mg,
2.13 mmol),
tetrakis(triphenylphosphine)palladium(0)
(102.6 mg, 0.089 mmol) and copper(I) iodide (33.8 mg,
0.18 mmol) were added. The reaction mixture was stirred
for 15 h, after which time TLC (EtOAc) showed complete
conversion of the starting material. Et3N (5 ml) and CuI
(101.4 mg) were added and the reaction mixture was then
heated at 80 °C. The reaction was monitored by TLC
(EtOAc/MeOH 95:5) and after 4 h the solvent was removed
in high vacuo. The resulting residue was dissolved in
dichloromethane/methanol (1:1), and then excess Amberlite
In conclusion, the branching of the alkylaryl side chain
of BCNAs leads to a decrease in their biological activity
against VZV, with a different impact depending on the
position of the branching in the chain. In view of these
results, it is clear that ClogP is not the only factor
influencing the biological activity, the geometry and
steric hindrance in the alkylaryl side chain are also
important factors to be considered in predicting the
antiviral activity of BCNAs.
ꢀ
IRA-400 (HCO3 form) was added and the mixture was
stirred at room temperature for 30 min. The reaction
mixture was then filtered and washed with methanol and
the combined filtrate was evaporated to dryness. The crude
residue was purified by flash chromatography (EtOAc), and
the product was loaded absorbed on silica, after being
dissolved in hot methanol. The appropriate fractions were
combined and the solvent was removed in vacuo. A second
purification was necessary (CHCl3/MeOH 95:5) and the
recovered product was subsequently crystallised with
MeOH to yield the product as a white solid (136.5 mg,
Acknowledgments
The authors are grateful to Mrs. Anita Camps and Miss
Lies Vandenheurck for excellent technical assistance.
The research was supported by grants from the Fonds
voor Geneeskundig Wetenschappelijk Onderzoek Vla-
anderen (G. 0267.04) and the Belgian Geconcerteerde
Onderzoeksacties. We also thank Helen Murphy for
excellent secretarial assistance.
1
20%). H NMR (DMSO-d6; 300 MHz) d 8.85 (1H, s, H-4),
7.74 (2H, H-A)-7.29 (2H, H-B) (AB system, J = 8.1 Hz),
7.22 (1H, s, H-5), 6.19 (1H, wt, J = 6.0 Hz, H-10), 5.32 (1H,
d, J = 4.2 Hz, 30-OH), 5.21 (1H, t, J = 5.0 Hz, 50-OH),
4.28–4.25 (1H, m, H-30), 3.94–3.93 (1H, m, H-40), 3.75–3.61
(2H, m, H-50), 2.48–2.38 (3H, m, H-20, a-CH2), 2.15–2.07,
(1H, m, H-20), 1.90–1.82, (1H, m, CH), 0.87 (6H, d,
J = 6.6 Hz, CH3). 13C NMR (DMSO-d6; 75 MHz) d 22.5
(CH3), 29.9 (CH), 41.6, 44.7, 61.0 (C-20, C-50, a-CH2), 69.9
(C-30), 87.9, 88.5 (C-10, C-40), 99.1 (C-5), 107.2 (C-4a),
124.8, 130.0 (C-Ph), 126.3 (C-ipso), 138.2 (C-4), 143.2
(C-para), 154.1, 154.2 (C-6, C-2), 171.4 (C-7a).
References and notes
´
1. McGuigan, C.; Yarnold, C. J.; Jones, G.; Velazquez, S.;
Barucki, H.; Brancale, A.; Andrei, G.; Snoeck, R.; De
Clercq, E.; Balzarini, J. J. Med. Chem. 1999, 42, 4479.