Synthetic chemistry-led creation of a difluorinated biaryl ether non-nucleoside reverse transcriptase inhibitor

Article abstract of DOI:10.1039/b714091f

In the process of developing a series of novel, fluorinated biaryl ether NNRTIs, we fortuitously discovered derivative 20, which possesses excellent potency against both wild-type and clinically relevant mutations of the reverse transcriptase enzyme. The

Full text of DOI:10.1039/b714091f

View Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
Synthetic chemistry-led creation of a difluorinated biaryl ether non-nucleoside
reverse transcriptase inhibitor†
Lyn H. Jones,* Amy Randall, Oscar Barba‡ and Matthew D. Selby
Received 13th September 2007, Accepted 24th September 2007
First published as an Advance Article on the web 2nd October 2007
DOI: 10.1039/b714091f
In the process of developing a series of novel, fluorinated
biaryl ether NNRTIs, we fortuitously discovered derivative
20, which possesses excellent potency against both wild-type
and clinically relevant mutations of the reverse transcriptase
enzyme.
Introduction
Reverse transcriptase (RT) is an essential enzyme in the infectious
life cycle of HIV and inhibition of this enzyme has shown great
utility in the treatment of HIV. Unfortunately, established drugs
called non-nucleoside reverse transcriptase inhibitors (NNRTIs),
which bind in an allosteric pocket of the enzyme, are susceptible
to viral resistance caused by single-point mutations in RT. As a
result, there is still an unmet medical need for NNRTIs that do
not suffer from rapid drug resistance.1 Strategies have therefore
focussed on developing NNRTIs that retain significant activity
against the most clinically relevant mutations (particularly K103N
and Y181C), which reduce their susceptibility to the mutant virus.
To this end we have worked on a series of biaryl ethers2
that possess intrinsically high affinity for the RT enzyme and
we desired a number of fluorinated derivatives to further build
our structure–activity relationships (SARs) within this class,
particularly with regards to their mutant profile. Fluorine imparts
many desirable characteristics into drugs as it can modulate both
the pharmacokinetic and pharmacodynamic properties of the
compound.3 Incorporating fluorine into molecules selectively and
safely can represent a significant synthetic challenge and therefore
methods for the generation of fluorinated ‘fragments’ are highly
desirable. Our chemistry has focussed on the incorporation of
fluorine into the B ring of the biaryl ether, and our originally
planned targets are illustrated in Fig. 1. We discovered that during
our synthetic exploits we were able to generate a pair of isomeric
difluorinated NNRTIs (3 and 20), which possessed surprisingly
different activities.
Fig. 1 Biaryl ether NNRTIs.
Scheme 1 Reagents and conditions: (i) 3-chloro-5-fluoro-benzonitrile,
Cs2CO3, DMF, 80 ◦C, 28% (6), 100% (7); (ii) BBr3, DCM, rt, 95% (8),
78% (9); (iii) 10, K2CO3, NaI, DMF, 40 ◦C, 21% (1), 50% (2).
fluoro-benzonitrile provided the biaryl ether 6. Treatment with
boron tribromide revealed the phenol 8 which was subsequently
alkylated with the known chloride 102 to provide 1. The fluorinated
derivative 2 was prepared in the same manner using commercially
available 5-fluoro-2-methoxy phenol 5 as starting material. Inter-
estingly, although both 1 and 2 possess very good potency against
the wild-type enzyme (Table 1), neither is capable of inhibiting the
clinically relevant mutations K103N or Y181C to levels that are
within ten-fold of the wild-type value.
Results and discussion
Our initial target was the non-fluorinated derivative 1, which
would allow us to assess the effect, if any, of fluorinating ring
B. 1 was prepared in three simple steps from 2-methoxy phenol
4 (Scheme 1). Arylation with commercially available 3-chloro-5-
Table 1 Potency of 1 and 2 for wild-type, K103N and Y181C RTa
Discovery Chemistry, Sandwich Laboratories, Pfizer Global Research and
Development, Ramsgate Road, Sandwich, Kent, UK CT13 9NJ. E-mail: lyn.
jones@pfizer.com.; Fax: +44 (0)1304 651821; Tel: +44 (0)1304 644256
† Electronic supplementary information (ESI) available: Experimental
details and COSY spectrum for compound 19. See DOI: 10.1039/b714091f
Compound
Wt IC50/nM
K103N IC50/nM
Y181C IC50/nM
1
2
30
51
602
1510
303
533
a IC50 values are geomeans of 3 measurements.
‡ Current address: OSI Pharmaceuticals, Watlington Rd, Oxford, UK OX4
6LT.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 3431–3433 | 3431
©

View Online
Table 2 Potency of 3 and 20 for wild-type, K103N and Y181C RTa
We further pursued difluorinated derivative 3 to build our SAR
in this series. Using similar chemistry we planned to prepare 3
(Fig. 1) from the known phenol 12 which itself had been prepared
previously in several steps from 2,4-difluoroanisole (Scheme 2).4
This procedure relied on protecting the most acidic 3-position
of 2,4-difluoroanisole with a trimethylsilyl group. We initially
hypothesized that it may be possible to prepare intermediate
12 in a simpler manner via direct halogen–lithium exchange of
commercially available 1-bromo-3,5-difluoro-2-methoxy benzene
(17), quenching with an appropriate boronic ester followed by
oxidation with hydrogen peroxide to provide the desired phenol 12.
Previous work by Schlosser and Heiss had suggested that this route
would be feasible since treatment of 1-bromo-3,5-difluorobenzene
with n-butyl lithium at −75 ◦C in THF followed by CO2 gives
3,5-difluorobenzoic acid in 91% yield.5
Compound
Wt IC50/nM
K103N IC50/nM
Y181C IC50/nM
3
20
5450
4.3
67 300
12
45 100
19
a IC50 values are geomeans of 3 measurements.
exchange chemistry with commercially available bromide 17
(Scheme 4). Interestingly, treatment of 17 with n-butyl lithium
in THF, followed by a quench with triisopropyl borate and acid
hydrolysis gave boronic acid 18 exclusively. Subsequent oxidation
with hydrogen peroxide furnished the phenol 19 as the only
regioisomer (proved by COSY NMR—see ESI†). Although we
had not originally set out to prepare this isomer, we decided to
push intermediate 19 through to the regioisomeric NNRTI using
similar chemistry to that described above (Scheme 4). Fortuitously,
derivative 20 possessed excellent potency against both wild-type
RT and the K103N and Y181C mutants (Table 2).7
Scheme 2 Reagents and conditions: (i) nBuLi, THF, −78 ◦C then
(CH3)3SiCl, 94%; (ii) nBuLi, THF, −78 ◦C then DMF, 46% (7 : 1 ratio
of ortho and meta-substituted aldehydes); (iii) TBAF, CF3CH2OH, THF,
43%; (iv) mCPBA, DCM then KOH aq. 64%.4
However, we were also aware of the work by Bridges et al., which
showed that treatment of the isomeric bromide 13 (Scheme 3) with
n-butyl lithium at −78 ◦C followed by CO2 can give a mixture of
three products depending on the choice of solvent.6 In diethyl
ether, 13 gave 14 almost exclusively in 83% yield, but in THF 14,
15 and 16 were formed in a ratio of 1 : 2 : 2 respectively and in
75% total yield. In this case an autometalation mechanism was
proposed, whereby the initially formed organo-lithium can act as
a base towards unlithiated substrate, extracting the most acidic
proton (the 3-proton from the starting bromide) to give 14 and
15 upon quenching. This process consumes half an equivalent of
n-butyl lithium, the remaining half is then available to deprotonate
the reduced starting material, resulting in the formation of 16 upon
quenching with CO2.
Scheme 4 Reagents and conditions: (i) nBuLi, THF, −78 ◦C then
(iPrO)3B then HCl aq., 43%; (ii) H2O2 aq., AcOH, THF, rt, 59%; (iii) a.
3-chloro-5-fluoro-benzonitrile, Cs2CO3, DMF, reflux, 34%, b. BBr3, DCM,
rt, 50%, c. 10, K2CO3, NaI, DMF, 40 ◦C, 84%.
We still desired a method of preparing our original difluorinated
isomer 3, particularly one which avoided the route outlined in
Scheme 2 to intermediate 12. Following the precedence set by
Bridges, we then repeated the halogen–lithium exchange chemistry
on bromide 17 in diethyl ether (Scheme 5). Pleasingly, treatment
of the resulting boronic acid with hydrogen peroxide provided
the desired phenol 12 exclusively. Arylation, demethylation and
alkylation follows the same chemistry described previously to
Scheme 3 Reagents and conditions: (i) nBuLi, THF or Et2O, −78 ◦C then
Scheme 5 Reagents and conditions: (i) a. nBuLi, Et2O, −78 ◦C then
(iPrO)3B then HCl aq., b. H2O2 aq., AcOH, THF, rt, 62% over two steps;
(ii) a. 3-chloro-5-fluoro-benzonitrile, Cs2CO3, DMF, reflux, 75%, b. BBr3,
DCM, rt, 70%, c. K2CO3, NaI, DMF, 40 ◦C, 40%.
CO2.6
Despite being a little unsure how the key step in our synthe-
sis would proceed, we decided to pursue the halogen–lithium
3432 | Org. Biomol. Chem., 2007, 5, 3431–3433
This journal is
The Royal Society of Chemistry 2007
©

View Online
provide the difluorinated derivative 3. Interestingly, this derivative
was over 1000-fold less potent than its isomer 20 (Table 2).
also thank Torren Peakman for performing the COSY NMR
on 19.
Conclusions
References
We have developed short synthetic routes to successfully furnish
fluorinated biaryl ether NNRTIs.8 Halogen–lithium exchange
chemistry led to an alternative arrangement of substituents on ring
B of the biaryl ether, resulting in the preparation of the difluori-
nated derivative 20, which possesses excellent potency against the
wild-type RT enzyme and retains impressive activity against the
clinically relevant mutations K103N and Y181C. This fortuitous
discovery illustrates the essential role synthetic chemistry has to
play in the drug discovery/design process. Further development
of 20, and a detailed structure–function analysis of this series, is
ongoing work within our group and will be reported in due course.
1 E. De Clercq, Chem. Biodiversity, 2004, 1, 44.
2 L. H. Jones, D. S. Middleton, C. E. Mowbray, S. D. Newman and
D. H. Williams, WO2006067587; GSK have reported a related series of
benzophenone NNRTIs; J. H. Chan, G. A. Freeman, J. H. Tidwell, K. R.
Romines, L. T. Schaller, J. R. Cowan, S. S. Gonzales, G. S. Lowell, C. W.
Andrews III, D. J. Reynolds, M. St Clair, R. J. Hazen, R. G. Ferris, K. L.
Creech, G. B. Roberts, S. A. Short, K. Weaver, G. W. Koszalka and L. R.
Boone, J. Med. Chem., 2004, 47, 1175.
3 K. L. Kirk, J. Fluorine Chem., 2006, 127, 1013.
4 G. T. Chen, M. King, F. Gusovsky, C. R. Creveling, J. W. Daly, B.-H.
Chen, J.-y. Nie and K. L. Kirk, J. Med. Chem., 1993, 36, 3947.
5 M. Schlosser and C. Heiss, Eur. J. Org. Chem., 2003, 4618.
6 A. J. Bridges, W. C. Patt and T. M. Stickney, J. Org. Chem., 1990, 55,
773.
7 Roche and Merck have recently reported similarly substituted biaryl
ether NNRTIs: B. T. Silva and Z. K. Sweeney, US20070088015 (Roche);
J. P. Dunn, D. R. Hirschfeld, T. Silva, Z. K. Sweeney and H. Vora,
US2005239881 (Roche); S. A. Saggar, J. T. Sisko, T. J. Tucker, R. M.
Tynebor, D.-S. Su and N. J. Anthony, US20070021442 (Merck).
8 All compounds (1, 2, 3 and 20) are non-cytotoxic, i.e. CC30 > 30 lM.
Acknowledgements
We thank Romu Corbau, Ian Burr, Alex Martin and Amy
Thomas for determining the potencies of these compounds. We
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 3431–3433 | 3433
©