4,1-Benzoxazepinone analogues of efavirenz (Sustiva) as HIV-1 reverse transcriptase inhibitors

Article abstract of DOI:10.1016/S0960-894X(01)00239-6

A series of 4,1-benzoxazepinone analogues of efavirenz (Sustiva) as potent NNRTIs has been discovered. The cis-3-alkylbenzoxazepinones are more potent then the trans isomers and can be synthesized preferentially by a novel stereoselective cyclization. The best compounds are potent orally bioavailable inhibitors of both wild-type HIV-1 and its clinically relevant K103N mutant virus, but are highly protein-bound in human plasma.

Full text of DOI:10.1016/S0960-894X(01)00239-6

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1389–1392
4,1-Benzoxazepinone Analogues of Efavirenz (SustivaTM) as
HIV-1 Reverse Transcriptase Inhibitors
Anthony J. Cocuzza,* Dennis R. Chidester, Beverly C. Cordova, Ronald M. Klabe,
Susan Jeffrey, Sharon Diamond, Carolyn A. Weigelt, Soo S. Ko, Lee T. Bacheler,
Susan K. Erickson-Viitanen and James D. Rodgers
DuPont Pharmaceuticals Company, Experimental Station, E336/141, PO Box 80336, Wilmington, DE 19880-0336, USA
Received 22 January 2001; accepted 28 March 2001
Abstract—A series of 4,1-benzoxazepinone analogues of efavirenz (SustivaTM) as potent NNRTIs has been discovered. The cis-3-
alkylbenzoxazepinones are more potent then the trans isomers and can be synthesized preferentially by a novel stereoselective
cyclization. The best compounds are potent orally bioavailable inhibitors of both wild-type HIV-1 and its clinically relevant K103N
mutant virus, but are highly protein-bound in human plasma. # 2001 DuPont Pharmaceuticals Company. Published by Elsevier
Science Ltd. All rights reserved.
Effective therapy for the treatment of HIV-1 infection
and AIDS requires a combination of antiviral drugs.
Currently, the standard of care for antiretroviral naıve
patients is efavirenz or an HIV protease inhibitor and
two nucleoside reverse transcriptase inhibitors. Efavir-
enz,1 a non-nucleoside reverse transcriptase inhibitor
(NNRTI), has demonstrated clinical efficacy in both
antiretroviral naıve and experienced patients and may
also provide an option of a protease-sparing regimen
when used with two nucleoside reverse transcriptase
(RT) inhibitors.2 Although HIV infection is generally
well controlled by efavirenz-containing regimens, a
number of patients develop resistance to the drug
through a mutation in the viral reverse transcriptase. In
these patients, the K103N mutation is present in over
90% of the RT sequences examined.3 In an effort to
develop a second-generation drug with improved activ-
ity against K103N and other resistant mutants, an SAR
investigation was launched to determine the effect of
varying the aromatic substitution pattern of efavirenz as
well as modifying and replacing the acetylenic side
chain.4À6 In addition, other ring systems have been exam-
ined. Investigation of the analogous quinazolinones resul-
ted in four compounds (DPC 961, DPC 963, DPC 082,
and DPC 083) which are currently under clinical
investigation.7À9 This report describes the synthesis and
preclinical evaluation of a new class of efavirenz-related
NNRTIs, the 4,1-benzoxazepinones (1 and 2).
We synthesized 4,1-benzoxazepinone NNRTIs by a
route analogous to the preparation of efavirenz and its
analogues.4,5,10 As has been previously described, treat-
ment of the readily available trityl protected 2-amino-
trifluoroacetophenones 3 with lithium acetylides affor-
ded acetylenic alcohols 4 (Scheme 1). Detritylation pro-
vided acetylenic amino alcohols 5, while lithium
aluminum hydride reduction of the acetylene followed
by deprotection afforded trans olefinic aminoalcohols 6.
These amino alcohols, which had previously been con-
verted to efavirenz and its benzoxazine analogues by
cyclization with phosgene, were converted to 4,1-benz-
oxazepinones 1 and 2 by N-acylation with 2-bromo-
acylbromides followed by sodium hydride mediated
intramolecular O-alkylation. Use of bromoacetyl bro-
mide provided high yields of 3-unsubstituted 4,1-benz-
oxazepinones while homologous 2-bromoacyl bromides
yielded difficult to separate mixtures of two diastere-
omeric 3-alkyl-4,1-benzoxazepinones.
*Corresponding author. Fax: +1-302-695-6670; e-mail: anthony.j.
cocuzza@dupontpharma.com
0960-894X/01/$ - see front matter # 2001 DuPont Pharmaceuticals Company. Published by Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00239-6

1390
A. J. Cocuzza et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1389–1392
The difficulty in separating the 3-alkyl-4,1-benzox-
azepinone diastereomers led us to investigate a stereo-
selective synthesis. By carefully monitoring the sodium
hydride-mediated cyclization reaction depicted in
Scheme 2 by TLC and NMR we found that one dia-
stereomeric 2-bromoamide 7 cyclized much faster than
the other. If the reaction was not allowed to proceed to
completion the more potent benzoxazepinone 1f, in
which the 3-methyl is cis to the 5-trifluoromethyl group,
formed in preference to the trans isomer 1g. Since this
strategy could at best afford a 50% yield of 1f, we
sought reaction conditions which would promote the
conversion of both diastereomeric bromoamides 7 into
the desired cis-benzoxazepinone 1f. In particular, reac-
tion conditions which caused the two diastereomeric
bromoamides to interconvert at a rate faster than the
cyclization step could funnel both bromoamide inter-
mediates into the desired cis isomer. Such a fast inter-
conversion of haloamides could possibly be effected by
tandem SN2 reactions promoted by the addition of
excess bromide or iodide. 2-Bromopropionamide 7 was
subjected to a number of such conditions, one of which
proved particularly effective in promoting a diastereo-
selective transformation. Treatment of 7 with cesium
carbonate (1.5 equiv) and lithium iodide (2 equiv) in
DMF at room temperature afforded a 93:7 ratio of the
cis-benzoxazepinone 1f to the trans-benzoxazepinone
1g.11 The structure of 1f was confirmed by X-ray crys-
tallography, while the relative stereochemistry of later
analogues could be determined by the 19F NMR reso-
nance of the trifluoromethyl group (À79 d for the cis-
benzoxazepinones, À74.5 d for the trans in CDCl3). The
procedures described above were used to synthesize a
number of 3-alkylbenzoxazepinones. The sodium
hydride cyclization procedure was initially used to pre-
pare mixtures of the two diastereomeric 3-alkyl-ben-
zoxazepinones, and once the trans isomers were found
less potent, the cesium carbonate/lithium iodide cycli-
zation conditions were used to prepare stereoselectively
the cis-4,1-benzoxazepinones (Table 1).
The ability of 4,1-benzoxazepinones to inhibit HIV-1
reverse transcriptase in an in vitro enzyme assay
(enzyme IC50) and to inhibit the wild-type RF strain of
HIV-1 (wild-type IC90) in a whole cell assay is repre-
sented in Table 1. In addition, most compounds were
examined for their ability to inhibit a virus containing
the clinically relevant K103N mutation (K103N IC90).
Aromatic substitutions which had proven optimal in
other series were investigated and compounds with
potent activity were found in 7-chloro, 7-fluoro, and
6,7-difluoro substituted compounds. Compounds in
which the acetylene or olefin side chain was terminally
substituted with ethyl, isopropyl, or cyclopropyl groups
(1b, 1c, and 1e) were potent antivirals, however, acet-
ylenes substituted with larger phenyl (1a) or 3-furanyl
(1d) groups resulted in some loss of activity. 3-Unsub-
stituted benzoxazepinones 1b, 1c, 1e, and 2e demon-
strated potency against wild-type virus, but their ability
to inhibit replication of the K103N mutant was poor.
When a 3-methyl or 3-ethyl group was introduced cis to
the trifluoromethyl group, benzoxazepinones with
potent antiviral activity against both wild-type and the
K103N mutant viruses were obtained. The antiviral
potency of 1f, 2f, and 2h is comparable to quinazoli-
nones DPC 961 and DPC 083. The introduction of lar-
ger 3-alkyl groups (1i, 1j, and 1k) resulted in some loss
of wild-type activity. When compared to their cis-ste-
reoisomers, benzoxazepinones in which the 3-alkyl is
trans to the trifluoromethyl group suffered loss in wild-
type or K103N potency. For example, 1g exhibited far
less potency against the mutant virus than its isomer
1f.15
Scheme 1. Reagents and conditions: (a) n-BuLi, alkylacetylene, THF,
0 ꢀC, 15 min; (b) LAH, THF, rt, 16 h; (c) concd HCl, MeOH, rt, 30
min; (d) 2-bromoacyl bromide, pyridine, ether, 1 h; (e) NaH, DMF, rt,
3–24 h.
Two benzoxazepinone NNRTIs, acetylene 1f and its
olefin analogue 2f were selected for in vivo evaluation.
Rhesus monkeys were given either a single oral 10 mg/kg
Scheme 2.

A. J. Cocuzza et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1389–1392
Table 1. Structure and biological activity of 4,1-benzoxazepinones
1391
Compound
X
R1
R2
Enzyme
IC50 (nM)12
Wild-type
IC90 (nM)13
K103N
IC90 (nM)13
Efavirenz14
38
32
23
1.9
2.0
2.1
24.3
9.1
9.7
47.5
9.1
3.9
2.2
7.0
8.4
17.2
20.7
16.8
2.5
49
10
27
DPC 96114
DPC 08314
1a
1b
1c
1d
1e
2e
1f14
1g14
1h
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-Cl
7-F
7-F
7-F
7-F
7-F
7-F
6,7-diF
6,7-diF
Phenyl
Ethyl
Isopropyl
3-Furanyl
H
H
H
H
H
3300
964
423
2510
432
108
82
249
391
799
597
235
46
185
1940
1140
95
171
129
151
193
173
4100
1763
784
ND
1395
362
29
1130
139
ND
ND
ND
35
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
Cyclopropyl
H
cis Me
trans Me
cis Et
cis nPr
cis iPr
cis CH2CF3
cis Me
cis Et
trans Me
trans Et
cis Me
cis Et
cis Me
cis Et
cis Me
cis Me
1i
1j
1k
2f14
2h
7.8
13
1l
1m
1n
1p
2n
2p
1q
189.4
87.4
5.2
5.3
7.6
2.5
5.8
5.2
ND
ND
333
117
118
70
75
81
2q
References and Notes
Table 2. Pharmacokinetic parameters of 1f and 2f in rhesus monkeys
after a 10 mg/kg po dose
Compd
Cmax
(mM)
Tmax
(h)
AUCT
(mM h)
T1=2
C24 h
(mM)
N
1. Young, S. D.; Britcher, S. F.; Tran, L. O.; Payne,
L. S.; Lumma, W. C.; Lyle, T. A.; Huff, J. R.; Ander-
son, P. S.; Olsen, D. B.; Carroll, S. S.; Pettibone, D. J.;
O’Brien, J. A.; Ball, R. G.; Balani, S. K.; Lin, J. H.;
Chen, I-W.; Scheif, W. A.; Sardana, V. V.; Long, W. J.;
Brynes, V. W.; Emini, E. A. Antimicrob. Agents Che-
mother. 1995, 39, 2602.
1f
2f
1.67
2.36
6
11
30
50.8
N.D.
16
1.14
1.4
1
3
2. Panel on Clinical Practices for the Treatment of HIV
Infection. Guidelines for the Use of Antiretroviral Agents in
HIV-Infected Adults and Adolescents; February 5, 2001; http://
www.hivatis.org
3. Bacheler, L.; Weislow, O.; Snyder, S.; Hanna, G.;
D’Aquila, R. The SUSTIVA Resistance Study Team.
Abstract, Presented at the 12th World AIDS Conference,
Geneva, Switzerland, June 29–July 3, 1998.
4. Patel, M.; Ko, S. S.; McHugh, R. J.; Markwalder, J. A.;
Srivastava, A. S.; Cordova, B. C.; Klabe, R. M.; Erickson-
Viitanen, S.; Trainor, G. L.; Seitz, S. P. Bioorg. Med. Chem.
Lett. 1999, 9, 2805.
dose of 1f or 2f and the pharmacokinetic results are
summarized in Table 2. The total plasma concentration
of these drugs at 24 h was sufficient to be considered for
development. However, equilibrium dialysis protein
binding experiments showed that these compounds were
highly protein bound in human serum (>99.7% bound)
so that the free fraction of drug at 24 h would be less
than the IC90 for the K103N mutation.16 Therefore,
these compounds would not be considered viable can-
didates for development.
5. Patel, M.; McHugh, R. J.; Cordova, B. C.; Klabe, R. M.;
Erickson-Viitanen, S.; Trainor, G. L.; Ko, S. S. Bioorg. Med.
Chem. Lett. 1999, 9, 3221.
6. Cocuzza, A. J. Bioorg. Med. Chem. Lett. in press.
7. Corbett, J. W.; Ko, S. S.; Rodgers, J. D.; Gearhart, L. A.;
Magnus, N. A.; Bacheler, L. T.; Diamond, S.; Jeffrey, S.;
Klabe, R. M.; Cordova, B. C.; Garber, S.; Logue, K.; Trainor,
G. L.; Anderson, P. S.; Erickson-Vittanen, S. K. J. Med.
Chem. 2000, 43, 2019.
We have discovered a new class of non-nucleoside reverse
transcriptase inhibitor: the 4,1-benzoxazepinones. The
cis-3-alkyl-4,1-benzoxazepinones are inhibitors of both
wild-type HIV-1 and the clinically significant mutant
K103N with potency comparable to efavirenz and the
quinazolinone clinical candidates. Two compounds
selected for further study (1f and 2f) exhibited favorable
pharmacokinetic profiles in rhesus monkeys but were
highly protein bound in human plasma.
8. Corbett, J. W.; Ko, S. S.; Rodgers, J. D.; Jeffrey, S.;
Bacheler, L. T.; Klabe, R. M.; Diamond, S.; Lai, C.-M.;

1392
A. J. Cocuzza et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1389–1392
Rabel, S. R.; Saye, J. A.; Adams, S. P.; Trainor, G. L.;
Anderson, P. S.; Erickson-Vittanen, S. K. Antimicrob. Agents
Chemother. 1999, 43, 2893.
from ethyl acetate/hexane afforded 860 mg (47%) of 1f as a
single diastereomer. Additional material could be obtained by
chromatography of the mother liquor.
9. Corbett, J. W.; Gearhart, L. A.; Ko, S. S.; Rodgers, J. D.;
Cordova, B. C.; Klabe, R. M.; Erickson-Vittanen, S. K.
Bioorg. Med. Chem. Lett. 2000, 10, 193.
10. Pierce, M. E.; Parsons, R. L.; Radesca, L. A.; Lo, Y. S.;
Silverman, S.; Moore, J. R.; Islam, Q.; Choudhury, A.; For-
tunak, J. M. d.; Nguyen, D.; Luo, C.; Morgan, S. J.; Davis,
W. P.; Confalone, P. N.; Chen, C.; Tillyer, R. D.; Frey, L.;
Tan, L.; Xu, F.; Zhao, D.; Thompson, A. S.; Corley, E. G.;
Grabowski, E. J. J.; Reamer, R.; Reider, P. J. J. Org. Chem.
1998, 63, 8536.
11. Experimental procedure for compound 1f: To a 0 ꢀC
solution of 4f (X=4-Cl, R1=cycPr) (1.54 g, 5.3 mmol) and
pyridine (0.527 mL, 6.5 mmol) in anhydrous ether (85 mL)
was added 2-bromopropionyl bromide (0.640 mL, 6.1 mmol).
After stirring for 30 min, the reaction mixture was washed
with water, sat. sodium bicarbonate, and brine, dried over
magnesium sulfate and evaporated. The intermediate bro-
moamide 7 was dissolved in dry DMF (85 mL), lithium iodide
(1.48 g, 11.1 mmol) and cesium carbonate (2.7 g, 8.2 mmol)
were added, and the reaction mixture was stirred at room
temperature for 48 h. The reaction mixture was diluted with
water and extracted with ether (3Â). The combined extracts
were washed with brine, dried and evaporated to a crude pro-
duct (1.5 g) of a 93:7 mixture of 1f and 1g. Crystallization
12. Compounds were assayed for enzyme inhibitory activity
(IC50) according to the protocol descibed in ref 7.
13. Compounds were assayed for whole cell based antiviral
activity (IC90) according to the protocol described in: Bache-
ler, L. T.; Paul, M.; Jadhav, P. K.; Otto, M.; Miller, J. Anti-
viral. Chem. Chemother. 1994, 5, 111.
14. The data presented for efavirenz, DPC 961, DPC 083, 1f,
1g, and 2f reflect values obtained for a single enantiomer. All
other compounds were tested as racemic mixtures. The biolo-
gical evaluation of each enantiomer of efavirenz, quinazoli-
nones, and 4,1-benzoxazepinones determined that only the S
enantiomer is active.
15. The lower antiviral potency of the 3-unsubstituted and
trans-3-alkylbenzoxazepinone compared to the cis-3-alkylben-
zoxazepinones is not easily explained by comparing the mini-
mum energy conformations of efavirenz, 1f, and 1g as
determined by X-ray crystallography and molecular modeling
studies. While key structural features of efavirenz overlap
better with the more potent 1f than 1g, these differences are
small.
16. Although protein-binding shift assays were routinely run
on our compounds, these studies failed to predict the high
values determined by the more direct equilibrium dialysis
experiments.