N1-Alkyl pyrimidinediones as non-nucleoside inhibitors of HIV-1 reverse transcriptase

Article abstract of DOI:10.1016/j.bmcl.2010.01.085

A series of N1-alkyl pyrimidinediones were designed, synthesized and evaluated as HIV-1 non-nucleoside reverse transcriptase inhibitors (NNRTIs). Our efforts identified compound 10b, which represents the lead compound in this series with pharmacokinetics

Full text of DOI:10.1016/j.bmcl.2010.01.085

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1589–1592
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
N1-Alkyl pyrimidinediones as non-nucleoside inhibitors of HIV-1
reverse transcriptase
b
a
b
b
a
Michael L. Mitchell a, , Jong Chan Son , Hongyan Guo , Yun-A Im , Eun Jung Cho , Jianhong Wang ,
Jaclyn Hayes a, Michael Wang a, Amber Paul a, Eric B. Lansdon a, James M. Chen a, Doris Graupe a,
Gerry Rhodes a, Gong-Xin He a, Romas Geleziunas a, Lianhong Xu a, Choung U. Kim a
*
a Gilead Sciences, Inc., 333 Lakeside Drive, Foster City, CA 94404, USA
b Korea Research Institute of Chemical Technology, PO Box 107, Yuseong, Daejeon 305-600, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of N1-alkyl pyrimidinediones were designed, synthesized and evaluated as HIV-1 non-nucleoside
reverse transcriptase inhibitors (NNRTIs). Our efforts identified compound 10b, which represents the lead
compound in this series with pharmacokinetics and antiviral potency that may support once-daily
dosing.
Received 4 December 2009
Revised 11 January 2010
Accepted 13 January 2010
Available online 21 January 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Human immunodeficiency virus
Reverse transcriptase
Non-nucleoside
Pyrimidinedione
Nonnucleoside reverse transcriptase inhibitors (NNRTIs) of
human immunodeficiency virus type 1 (HIV-1) are a key compo-
nent of highly active antiretroviral therapy (HAART) for the long-
term management of HIV infection. Efavirenz is the most widely
prescribed drug in this class. Under sub-optimal conditions, treat-
ment with efavirenz can lead to the development of resistant
viruses that carry the reverse transcriptase (RT) mutation K103N.
Efavirenz also causes CNS adverse events and is teratogenic.1 These
liabilities underscore the need for NNRTIs that are safer and better
tolerated. Our goal was to identify an NNRTI that possessed an
improved safety profile with the convenience of once-daily dosing.
In the preceding Letter,2 we discussed SAR and properties of
N1-heteoaryl/aryl pyrimidinediones. We herein report the synthe-
sis and evaluation of N1-alkyl pyrimidinedione as HIV-1 NNRTIs.
To improve the intrinsic stability of this class of compounds, we
first ascertained the stability of the pyrimidinedione core. We dis-
covered that removal of the N1 substituent, which yielded 8b, pro-
duced a molecule that was highly resistant to oxidation and
glucuronidation in human microsomes. It also displayed favorable
pharmacokinetics in dog but lacked antiviral activity (Table 1). To
restore antiviral activity, we began by adding small alkyl groups at
N1. We discovered that the addition of N1-ethyl (compound 10b)
produced a compound that had good HIV-1 antiviral activity
(EC50 = 6.1 nM). This discovery prompted us to further explore
SAR of N1-alkyl pyrimindindiones. The series of N-alkylated ana-
logs that were prepared during these studies are illustrated in
Scheme 1.
Three pyrimidinedione cores were prepared in a similar manner
(Scheme 1). 5-Alkyl-barbituric acids3 1a–c were converted to
trichloropyrimidines 2a–c by treatment with phosphorus oxychlo-
ride. Reaction of 2a–c with sodium methoxide furnished 4-chloro-
2,6-dimethoxypyrimidines 3a–d.4 Coupling partner
6
was
prepared from benzonitrile 4 by bromination and followed by
displacement of the resulting bromide using potassium cynanide.
The anion of benzonitrile 6, generated with sodium hydride, was
reacted with chloropyrimidines 3a–d, and subsequent oxidization
of the resulting nitrile with oxygen produced dimethoxypyrimi-
dines 7a–c.5 Deprotection using acetyl bromide gave the core
pyrimidinediones 8a–c that are C5-substituted with ethyl, isopro-
pyl or tert-butyl groups. The core pyrimidinediones were then
regioselectively alkylated using various alkyl halides and
potassium carbonate in DMF.
We also examined different linkers between the pyrimidinedi-
one and phenyl moieties to determine whether this linker plays
any role in potency and/or stability. The carbonyl linker was
replaced with moieties of varying polarity as well as with an ether
linker to investigate linker flexibility. The synthesis of these ana-
logs is illustrated in Scheme 2.
Dimethoxypyrimidine 7b was treated with Deoxylfluor,6 depro-
tected with acetyl bromide and alkylated to provide difluorometh-
ylene analog 12. The racemic monofluoro analog 16 was prepared
* Corresponding author. Tel.: +1 650 522 5185; fax: +1 650 522 5899.
E-mail address: mmitchell@gilead.com (M.L. Mitchell).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.01.085

1590
M. L. Mitchell et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1589–1592
Table 1
Activity, protein binding and pharmacokinetic parameters of N-alkyl pyrmidinediones
Compd
L
R1
R2
EC50 (nM)a fold change
Y181C
%Protein bindingb
t1/2 (h)
Dog PKc
WT
K103N
Vss (L/kg)
%F
Efavirenz
Etravirine
8b
10a
10b
10c
10d
9
11
—
—
—
—
—
—
H
1.1
1.6
>500
187
6.1
3.6
3.0
11.7
10
2.6
2.6
>1
>2.7
9.6
>58
12
>43
>50
4.3
8.1
2.6
35
0.9
>1
99.9
99.8
ND
ND
8.87
6.4
ND
6.5
ND
1.6
0.45
ND
2.25
1.40
2.23
1.53
ND
ND
13d
ND
ND
69e
ND
28f
ND
ND
ND
ND
ND
CO
CO
CO
CO
CO
CO
CO
CF2
CHF
O
iPr
iPr
iPr
iPr
iPr
Et
tBu
iPr
iPr
iPr
Me
Et
Pr
Bu
Et
Et
Et
Et
Et
>2.7
6.4
>36
10
>43
2.7
9.0
>24
>14
85.9
83.4
99.1
99.9
85.1
99.3
98.7
98.0
96.0
1.67
6.1
1.42
ND
ND
ND
ND
12
13
16
4.2
21
35
ND
ND
ND
ND = not determined.
a
Values are means of at least two experiments in MT-2 cells.
Value determined from HSA column.
Phamacokinetic values determined from IV (0.5 mg/kg) or PO solution dosing in beagle dogs in triplicate.
%F determined with 4 mg/kg oral dose.
%F determined with 0.5 mg/kg oral dose.
%F determined with 2 mg/kg oral dose.
b
c
d
e
f
Scheme 1. Reagents and conditions: (a) diethylaniline or 2,6-lutidine, POCl3, 90%; (b) NaOMe, MeOH, 65%; (c) NaH, DMF; NaH, O2, 85%; (d) AcBr, 70 °C, 87%; (e) alkyl halide,
K2CO3, DMF.
from 10b by reduction with sodium borohydride and fluorination
with DAST. Ether linked analog 16 was prepared by coupling of
phenol 14 and chloropyrimidine 2b, followed by benzyl protec-
tion/deprotection to aid purification and finally N1 alkylation.
The data for the prepared analogs are illustrated in Table 1.
Addition of alkyl chains of increasing length to 8b, from methyl
to n-butyl improved potency against WT HIV-1. The n-butyl analog
displayed the best antiviral activity in the series (compare 10a 10b,
10c and 10d). Not surprisingly, protein binding also increased with
alkyl length.
Curiously, the resistance profile of the n-propyl analog 10c is
inferior to that of the ethyl analog 10b or n-butyl analog 10d. X-
ray crystallographic studies were performed to attempt to under-
stand these differences.7 Co-crystal structures with wild type RT
showed that the pyrimidinedione cores of 10b, 10c, and 10d bind
in similar orientations but subtle differences exist in the direction
of the alkyl group termini (Fig. 1).
HIV-1 RT with closely-related pyrimidinedione NNRTIs.8 The
authors observed that the larger 5-position isopropyl group in
MKC-442 and TNK-651 force Y181 into a conformation enabling
strong interactions with the 6-phenyl ring of the inhibitor whereas
the smaller 5-position methyl group in HEPT cannot; this interac-
tion leads to greater potency for MCK-442 and TNK-651 relative to
HEPT.8,9 Although 10b contains the 5-position isopropyl group,
Y181 adopts a similar conformation in the co-crystal structures
of HEPT and 10b. Thus, conformational switching of Y181 may also
depend on the size of the N1 substituent as the co-crystal struc-
tures for 10c and 10d show Y181 adopts the ‘up’ conformation
(as with MKC-442 and TNK-651) and the ‘down’ conformation with
10b that contains the smaller N1-ethyl group.
The movement of the Y181 residue also has been observed in
co-crystal structures with a different structural class of NNRTIs
bound to RT.10 Similar to an analysis of benzophenone NNRTI ana-
logs, the ‘down’ position of Y181 may contribute to a favorable
resistance profile. Compound 10b relies less on the Y181 side chain
for binding in the pocket than 10c or 10d. With the side chain ‘up’,
the n-propyl has a poor resistance profile. The n-butyl regains the
activity profile of 10b in part due to the longer alkyl group
Another difference observed between the co-crystal structures
is that Y181 is flipped ‘down’ when 10b is bound but ‘up’ with
10c or 10d. This conformational switching of Y181 has been ob-
served by Stuart and co-workers in the co-crystal structures of

M. L. Mitchell et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1589–1592
1591
It was noteworthy that, the ethyl (10b), n-butyl (10d) and
cyclopropylmethyl (10f) analogs displayed comparable potencies
and PK properties but 10b was less protein bound. We therefore
fixed ethyl as the N1 alkyl substituent and examined C5
analogs.
The 5-alkyl substituent was explored by changing the isopropyl
group of 10b to ethyl (9) or tert-butyl (11). Compound 10b had a
superior potency profile compared to 9 or 11, suggesting the ethyl
group may not occupy the NNRTI pocket optimally whereas the
tert-butyl group may restrict ligand mobility, compromising po-
tency against the Y181C mutant. Additionally, in vivo glucuronida-
tion of compound 9 was observed for the first time in this series
and resulted in suboptimal pharmacokinetics. Thus, we fixed the
N1 and C5 positions and investigated other linkers.
Among the linkers investigated, the hydrophobic CF2 group
found in 12 yielded the most potent compound in terms of anti-
viral activity. Unfortunately, compound 12 was found to be unsta-
ble when tested in microsomes (data not shown). The racemic
CHF linker (13) showed improved in vitro stability relative to
12 but lost antiviral potency. Analog 16 with an ether linker
showed a small loss in potency for wild type and Y181C viruses
but lost significant activity against K103N virus. Thus, of the link-
ers examined, the carbonyl linker offers the best balance of po-
tency and stability.
Compound 10b possesses the best balance of antiviral activity
against wild type and mutant viruses, low protein binding and
good pharmacokinetic properties. However, 10b is highly crystal-
line (Table 2) and has low solubility. Thus, prodrugs of compound
10b were prepared according to Scheme 3 to attempt to reduce
crystallinity and/or improve solubility.
Scheme 2. Reagents and conditions: (a) deoxyfluor, cat. EtOH, 110 °C, 73%; (b)
AcBr, 60 °C, 85%; (c) EtI, K2CO3, DMF; (d) NaBH4, EtOH, 44%; (e) DAST, CH2Cl2, 68%;
(f) NaH, DMF, 88%; (g) BnOH, NaH, DMF, 40%; (h) H2, 10% Pd/C, EtOAc/EtOH, 29%; (i)
EtI, K2CO3, DMF, 28%.
Alkylation of 10b with bromomethyl acetate12/potassium car-
bonate or chloromethyl isopropylcarbonate13/sodium hydride gave
prodrugs 17 and 18, respectively. Reaction of 10b with chloro-
methyl di-tert-butylphosphate14 and potassium carbonate gave
di-tert-butyl phosphate 19 that was deprotected with HCl in diox-
ane to give phosphate prodrug 20.
The melting point, solubility and PK parameters of the prodrugs
17, 18 and 20 are shown in Table 2. Methoxy acetate (16) and POC
(17) prodrugs showed lower melting points than 10b and higher
oral bioavailability when administered as a suspension. Methoxy
phosphate 20 displayed a ꢀ1000-fold increase in solubility and
Figure 1. Overlay of X-ray co-crystal structures of wild type reverse transcriptase
and compounds 10b (green), 10c (yellow), and 10d (cyan).
anchoring 10c within the NNRTI binding pocket. The combination
of terminal methyl directionality differences with an apparent
change in the NNRTI binding pocket could perhaps lead to re-
stricted translational flexibility and diminished potency against
mutant viruses.11
Scheme 3. Reagents and conditions: (a) bromomethyl acetate, K2CO3, DMF, 64%;
(b) POC-Cl, NaH, DMF, 92%; (c) chloromethyl di-tert-butylphosphate, K2CO3, DMF,
50 °C, 61%; (d) TFA, CH2Cl2, 64%.
Table 2
Beagle dog pharmacokinetic parameters of parent 10b and prodrugs 17, 18 and 20 dosed from suspension versus tablet formulation (5 mg/kg)
Compds
MPa (°C)
Solb
(l
g/mL)
Suspension
AUCinf (nM h)
Tablet
F (%)
AUCinf (nM h)
F (%)
10b
17
18
230.3
167.0
154.7
ND
11
2.2
<1
13,131
14,704
54,022
18,765
35.8
48.9
108.2
75.2
13,485
8781
4543
ND
45.1
35.3
24.0
ND
20
12,000
a
Value determined using a differential scanning calorimeter (DSC).
Aqueous solubility measured in 50 mM phosphate buffer.
b

1592
M. L. Mitchell et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1589–1592
Acknowledgements
The authors wish to thank Manoj Desai, Gene Eisenberg, Nilima
Kutty, Candra Lancaster, and Stephanie Leavitt for their
contributions.
References and notes
1. (a) Arendt, G.; de Nocker, D.; von Giessen, H. J.; Nolting, T. Expert Opin. Drug Saf.
2007, 6, 147; (b) Fumaz, C. R.; Muñoz-Moreno, J. A.; Moltó, J.; Negredo, E.;
Ferrer, M. J.; Sirera, G.; Pérez-Alvarez, N.; Gómez, G.; Burger, D.; Clotet, B. J.
Acquir. Immune Defic. Syndr. 2005, 38, 560; (c) SUSTIVAÒ (efavirenz) capsules
and tablets: Package insert, Bristol-Myers Squibb Co, Princeton, NJ, USA, 2008.
packageinserts.Bms/com/pi/pi_sustiva.pdf.
2. Mitchell, M. L.; Son, J. C.; Lee, I. Y.; Lee, C.-K.; Kim, H. S.; Wang, J.; Hayes, J.;
Wang, M.; Paul, A.; Lansdon, E. B.; Chen, J. M.; Eisenberg, G.; Geleziunas, R.; Xu,
L.; Kim, C. U. Bioorg. Med. Chem. Lett. 2010, 20, 1585.
4. Ji, L.; Chen, F.-E.; Feng, X.-Q.; De Clercq, E.; Balzarini, J.; Pannecouque, C. Chem.
Pharm. Bull. 2006, 54, 1248.
Figure 2. Comparison of exposure of compound 10b from 5 mg/kg dose of parent
(10b) or prodrugs 17 and 18 when dosing as tablet.
5. Danel, K.; Pedersen, E. B.; Nielsen, C. Synthesis 1997, 9, 1021.
6. Lal, G. S.; Pez, G. P.; Pesaresi, R. J.; Prozonic, F. M.; Cheng, H. J. Org. Chem. 1999,
64, 7048.
7. Coordinates have been deposited into the PDB. For compounds 10b, 10c and
10d, accession codes are 3LAL, 3LAM, and 3LAN, respectively.
8. Hopkins, A. L.; Ren, J.; Esnouf, R. M.; Willcox, B. E.; Jones, Y.; Ross, C.; Miyasaka,
T.; Walker, R. T.; Tanaka, H.; Stammers, D. K.; Stuart, D. I. J. Med. Chem. 1996, 39,
1589. and references cited therein.
good oral bioavailability in dogs but resulted in emesis at doses
higher than 5 mg/kg.
In contrast, when micronized material (D50 = 3l) was prepared
9. Ren, J.; Stammers, D. K. Virus Res. 2008, 134, 157.
10. Ren, J.; Chamberlain, P. P.; Stamp, A.; Short, S. A.; Weaver, K. L.; Romines, K. R.;
Hazen, R.; Freeman, A.; Ferris, R. G.; Andrews, C. W.; Boone, L.; Chan, J. H.;
Stammers, D. K. J. Med. Chem. 2008, 51, 5000.
11. Das, K.; Clark, A. D., Jr.; Lewi, P. J.; Heeres, J.; De Jonge, M. R.; Koymans, L. M.;
Vinkers, H. M.; Daeyaert, F.; Ludovici, D. W.; Kukla, M. J. J. Med. Chem. 2004, 47,
2550.
12. Yamaoka; Y; Roberts, R. D.; Stella, V. J. J. Pharm. Sci. 1983, 72, 400.
13. Boehme, H.; Budde, J. Synthesis 1971, 11, 588.
14. Krise, J. P.; Zygmunt, J.; Georg, G. I.; Stella, V. J. J. Med. Chem. 1999, 42,
3094.
and dosed in tablet form in dogs, the parent 10b gave superior oral
bioavailability compared to 17 and 18 (Fig. 2).
Compound 10b (GS-9441) shows comparable protein adjusted
antiviral potency (EC50 = 16 nM) against wild type HIV as etravirine
(EC50 = 11 nM) but may have the advantage of once-daily adminis-
tration.15 While the resistance profile may be non-optimal, based
on biological activity against wild type virus, favorable protein
binding and good pharmacokinetic properties that suggest once-
daily dosing, further profiling of compound 10b for clinical evalu-
ation is on-going.
15. In the presence of 50% human serum, compound 10b exhibits a 2.7-fold shift in
antiviral potency for wild type and mutant viruses.