Discovery of ritonavir, a potent inhibitor of HIV protease with high oral bioavailability and clinical efficacy

Article abstract of DOI:10.1021/jm970636+

The structure-activity studies leading to the potent and clinically efficacious HIV protease inhibitor ritonavir are described. Beginning with the moderately potent and orally bioavailable inhibitor A-80987, systematic investigation of peripheral (P3 and P2') heterocyclic groups designed to decrease the rate of hepatic metabolism provided analogues with improved pharmacokinetic properties after oral dosing in rats. Replacement of pyridyl groups with thiazoles provided increased chemical stability toward oxidation while maintaining sufficient aqueous solubility for oral absorption. Optimization of hydrophobic interactions with the HIV protease active site produced ritonavir, with excellent in vitro potency (EC₅₀ = 0.02 μM) and high and sustained plasma concentrations after oral administration in four species. Details of the discovery and preclinical development of ritonavir are described.

Full text of DOI:10.1021/jm970636+

602
J . Med. Chem. 1998, 41, 602-617
Discover y of Riton a vir , a P oten t In h ibitor of HIV P r otea se w ith High Or a l
Bioa va ila bility a n d Clin ica l Effica cy
Dale J . Kempf,* Hing L. Sham, Kennan C. Marsh, Charles A. Flentge, David Betebenner, Brian E. Green,
Edith McDonald, Sudthida Vasavanonda, Ayda Saldivar, Norman E. Wideburg, Warren M. Kati, Lisa Ruiz,
Chen Zhao, LynnMarie Fino, J ean Patterson, Akhteruzzaman Molla, J acob J . Plattner, and Daniel W. Norbeck
Pharmaceutical Products Division, Abbott Laboratories, D-47D, AP52, 200 Abbott Park Road, Abbott Park, Illinois 60064
Received September 22, 1997
The structure-activity studies leading to the potent and clinically efficacious HIV protease
inhibitor ritonavir are described. Beginning with the moderately potent and orally bioavailable
inhibitor A-80987, systematic investigation of peripheral (P3 and P2′) heterocyclic groups
designed to decrease the rate of hepatic metabolism provided analogues with improved
pharmacokinetic properties after oral dosing in rats. Replacement of pyridyl groups with
thiazoles provided increased chemical stability toward oxidation while maintaining sufficient
aqueous solubility for oral absorption. Optimization of hydrophobic interactions with the HIV
protease active site produced ritonavir, with excellent in vitro potency (EC50 ) 0.02 µM) and
high and sustained plasma concentrations after oral administration in four species. Details
of the discovery and preclinical development of ritonavir are described.
In tr od u ction
The human immunodeficiency virus encodes an as-
partic proteinase (HIV protease) which is responsible
for the posttranslational proteolytic processing of the
gag and gag-pol polyprotein gene products into mature,
functional proteins.1-3 Accompanying these processing
F igu r e 1. Design of symmetry-based inhibitors of HIV
events, the nascent viral particles undergo a morpho-
logical transformation from an immature, noninfectious
form into mature, infectious virions.4 The critical role
of functional HIV protease in the HIV replication cycle
was initially demonstrated by site-directed mutagenesis
wherein mutation of the catalytic aspartate residues to
either asparagine5 or alanine6 generated viral particles
which remained in the immature form. Importantly,
these particles were incapable of establishing a new
round of infection in susceptible T-lymphocytes.5 Simi-
lar behavior was subsequently observed through the
blockade of HIV protease by small-molecule inhibitors,
thus validating HIV protease as a target for drug design
for acquired immune deficiency syndrome.7-10
protease.
the polyprotein substrates of HIV protease. As a result,
the identification of agents with high potency and
specificity for HIV protease with the proper physico-
chemical and pharmacodynamic properties for use as
oral agents has proven challenging due to the low
intestinal absorption and rapid hepatic clearance ob-
served with most peptidomimetics.18 Recently, sub-
stantial progress in the identification of agents with
high oral bioavailability has been made.19-22
We recently reported the design of novel inhibitors
of HIV protease based on the C2-symmetric structure
of HIV protease. Beginning with the tetrahedral inter-
mediate for cleavage of an asymmetric dipeptide sub-
strate, duplication of the N-terminus by a rotation about
a C2 axis bisecting the carbon-nitrogen single bond led
to the conceptualization of the symmetric or pseudo-
symmetric core diamines 1 and 2 (Figure 1).23,24
Structure-activity studies on derivatives of 1 and 2 led
initially to the identification of A-77003,24,25 which,
though displaying low oral bioavailability, possessed
adequate anti-HIV activity and aqueous solubility for
intravenous examination in HIV patients.26 Further
studies of the influence of structural features on the oral
bioavailability of this series produced the truncated
inhibitor A-80987,27 which demonstrated good oral
bioavailability in both animal models and in humans.19
Although A-80987 represented a substantial improve-
ment in oral absorption over A-77003, the short circu-
lating half-life of both inhibitors mirrored that of most
peptidomimetics. Examination of the metabolic fate of
A-77003 and A-80987 in vitro and in vivo revealed that
The classification of HIV protease and other retroviral
proteases as members of the aspartic proteinase class
was proposed on the basis of sequence alignment11 and
confirmed by X-ray crystallographic studies of both
engineered12 and synthetic13 protein. The retroviral
proteases (including HIV protease) are unique within
this class in that functional protease exists as a C2-
symmetric homodimer with a single active site. Each
monomeric unit contributes one of the conserved cata-
lytic triads (Asp-Thr-Gly) common to the aspartic pro-
teinase class. The wide availability of three-dimen-
sional structures of complexes of HIV protease with
bound inhibitors has prompted the use of this informa-
tion for structure-based design of novel inhibitors.14-17
Nonetheless, the majority of HIV protease inhibitor
classes are peptidomimetics, reflecting the structure of
* To whom correspondence should be addressed.
S0022-2623(97)00636-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/22/1998

Discovery of Ritonavir
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4 603
each inhibitor was primarily metabolized and cleared
through N-oxidation of one or both of the pyridyl end
groups.19 Recently we communicated the identification
of ritonavir (ABT-538), a symmetry-based HIV protease
inhibitor with a substantially reduced rate of metabo-
lism and high oral bioavailability in animals and
humans.19 Phase I/II clinical trials with ABT-538
produced a rapid28,29 and profound reduction of circulat-
ing viral RNA concomitant with a large increase in CD4
cell levels.30,31 Phase III trials also provided a signifi-
cant increase in time to AIDS-defining event or death32
and provided the basis for rapid licensing of ritonavir
for treatment of HIV infection. Here we describe the
details of the preclinical structure-activity studies
beginning with A-80987 and leading to the identification
of ritonavir.
Resu lts
The HIV protease inhibitors were evaluated in two
standard biological assays as previously described.25 The
first assay measured the inhibition of purified, recom-
binant HIV-1 protease using a fluorogenic substrate.36
IC50 values for inhibitors with potency >1 nM are
provided in the tables. The majority of the compounds
inhibited HIV protease at subnanomolar concentrations
and thus acted as active site titrants. Accurate Ki
values were not determined. Rather, the percent inhi-
bition at a given inhibitor concentration (0.5 or 1.0 nM)
is reported. This value provided a qualitative compari-
son between inhibitors, but varied in different experi-
ments due to slightly different concentrations of HIV
protease. The second assay measured the ability of the
inhibitors to block the spread of HIV-13B in the im-
mortalized human T-cell line MT4 by measuring the
cytopathic effect of HIV in those cells by uptake of a
tetrazolium dye.37 EC50 values represent the mean of
two triplicate determinations using 10 and 32 TCID50
per well, respectively. Potency differences of <2-fold are
not considered outside of experimental error. In this
assay, the activity of saquinavir38 and indinavir20
ranged from 0.01 to 0.05 µM. In general, the dose-
response curves for each analogue were similar, and
EC90 values were 3-4-fold higher than EC50 values.
Compound cytotoxicity (CCIC50) was determined in
uninfected MT4 cells.
Ch em istr y
We33 and others34 have previously described the
synthesis of the (S,S,S)-diamino alcohol 2. For ana-
logues of A-80987, differentiation of the two amino
groups of 2 was required. This was accomplished either
via statistical monoacylation as previously reported33
or via regioselective acylation of the azoxaborylidine 3
(T. Sowin et al., unpublished results). Subsequently,
an improved synthesis of 2 yielded large quantities of
the monoprotected diamine 4, which was conveniently
differentiated.35 To attach heterocyclic carbamates to
bind in the S2′ region of the active site, intermediates
2-4 were acylated with the p-nitrophenyl carbonates
of the corresponding heterocyclic carbinols (RCH2OH)
to produce monoamines 5 and/or 6 following either
chromatographic separation or deprotection. For the P2
and P3 groups, a variety of heterocyclic N-substituted
valine esters were required. Various heterocyclic alco-
hols or amines were prepared and acylated with L-valine
methyl ester activated as either the corresponding
isocyanate or p-nitrophenyl carbamate. Hydrolysis of
the esters followed by carbodiimide-mediated coupling
to 5 or 6 produced the desired HIV protease inhibitors
with the different peripheral groups (P3/P2 and P2′′,
respectively) either proximal or distal to the central
hydroxyl group of core unit 2. General procedures are
provided in the Experimental Section. The synthetic
methods leading to the inhibitors are tabulated in the
Supporting Information, along with yields and physical
characterization of the final products.
The aqueous solubility and pharmacokinetic profiles
of many of the inhibitors were also evaluated. Esti-
mated solubility in pH 7.4 phosphate and pH 4 acetate
buffers were measured as previously described.25 Plasma
concentrations of inhibitor following either a 5 mg/kg
intravenous (iv, n ) 2) or 10 mg/kg oral (n ) 3) dose in
the rat were quantitated by HPLC analysis of parent
compound as described.25 Plasma half-life (T1/2) was
calculated following the iv dose. The time (Tmax) and
value (Cmax) of highest plasma concentration and the
area under the time and plasma concentration curve
(AUC) were measured following the oral dose. Percent
oral bioavailability (F) was calculated by comparison of
the dose-normalized AUC values for iv and oral doses.
Solubility and pharmacokinetic parameters are pre-
sented in the tables, along with ratio of maximum

604 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4
Kempf et al.
concentration to the in vitro anti-HIV activity (Cmax
EC50).
/
5-methyl- and 6-methylpyridines was similar to the
6-methyl-2-pyridines 9-12. However, the pharmaco-
kinetic profiles of the 3-pyridines were consistently
poorer than the corresponding 2-pyridines. This trend
was observed previously with unsubstituted P3 pyridyl
groups25,33 and may reflect more rapid oxidation of the
less hindered 3-pyridyl nitrogen.
Str u ctu r e-Activity Str a tegies. In examining the
structure-activity relationships of analogues of A-80987,
we identified the dual goal of improving anti-HIV
activity and slowing the rate of metabolism. Underlying
this premise was the supposition that significant lower-
ing of metabolism would not only increase oral bioavail-
ability but also affect plasma half-life. Longer half-life
values would lead to higher trough levels upon repeated
dosing, thereby retaining constant levels well in excess
of the EC50. We speculated that maintenance of high
plasma levels would not only produce a more profound
antiviral effect in vivo but should also delay the time
before the emergence of resistant mutants due to more
complete suppression of viral replication.39 Modification
of the pyridyl groups of A-80987 to reduce the rate of
oxidative metabolism was examined within the frame-
work of previous studies in this series that identified
structural features promoting higher oral bioavailabil-
ity, including sufficient aqueous solubility, limited
hydrogen-bonding functionality, and limited size.33 We
reasoned that at least one basic nitrogen was required
for sufficient solubility in a low pH dosing regimen.
Modifications of the pyridyl groups of A-80987 included
the addition of substituents designed to slow oxidative
metabolism either sterically or electronically. Addition-
ally, we pursued replacing both pyridyl groups with
other heterocycles of sufficient basicity to contribute to
aqueous solubility. Initial studies suggested that meth-
yl substitution on the P3 2-pyridyl group of A-80987
improved activity against HIV protease,33 and we
anticipated further improvements as we examined other
substituents (vide infra).
We also examined substituents that modified the
electronic nature of the P3 pyridyl group of A-80987,
including the 5-methoxy-2-pyridyl analogues 34-37 and
the 5- and 6-methoxy-3-pyridyl analogues 38-45. Com-
pounds 34-45 were uniformly more active against HIV
than A-80987. In most cases, the additional methoxy
group provided sufficient aqueous solubility. None of
the analogues matched the pharmacokinetic profile of
A-80987, however. Compound 36 showed the best
properties, with a Cmax of 2.17 µM, a Cmax/EC50 ratio of
22, and oral bioavailability of 17%. As with the above
alkyl analogues, the plasma concentrations of 34-45
declined rapidly, again presumably due to rapid me-
tabolism. In contrast to 34-45, the aminopyridyl
analogues 46-53 were less active than A-80987.
5-Amino-3-pyridyl inhibitors 50-51 were the most
potent and showed excellent aqueous solubility at pH
4.0. However, oral administration to rats produced only
low plasma levels.
Oth er P 3 Heter ocycles. Concurrent with the above
studies, we explored the replacement of the P3 pyridyl
group of A-80987 with other heterocycles. The results
of those studies are provided in Tables 1 and 2. Pyrimi-
dine analogues 54-65 (Table 1) showed equivalent
activity to the corresponding pyridines (cf. 56-59 vs
7-10). Furthermore, several inhibitors in this series
showed significant oral bioavailability. Most promising
was compound 63, which was 5-fold more potent than
Su bstitu ted P 3 P yr id yl An a logu es. The activity
of 6-alkyl-2-pyridyl analogues of A-80987 (compound 5)
is shown in Table 1. In each case, both regioisomers,
resulting from attachment of the substituted valine
either proximal or distal to the hydroxyl group of 2, were
prepared. Carbamate, urea, and N-methylurea linkers
between the P3 heterocycle and the P2 valine were
examined. As observed previously,24,27 carbamate ana-
logues showed slightly but consistently higher activity,
both in the HIV protease assay and the antiviral assay,
than the corresponding ureas or N-methylureas. All of
the 6-alkylpyridines were more potent than the corre-
sponding unsubstituted pyridines, e.g., A-80987.33 Eth-
yl substitution appeared optimal for this series (cf. 13
and 14) with the larger isopropyl and tert-butyl sub-
stituents offering no advantage. Indeed, substitution
with either isopropyl or tert-butyl led in many cases to
inhibitors with increased cellular toxicity and lower
aqueous solubility. Of the subset of inhibitors of this
type that were examined, compounds with larger alkyl
groups showed lower Cmax and generally lower oral
bioavailability than A-80987, although Cmax/EC50 ratios
exceeded that of 80987 in some cases (e.g., 13) because
of increased potency. None of the analogues displayed
a pharmacokinetic profile equal to that of A-80987.
Notably, substitution with 6-alkyl groups did not sig-
nificantly affect the circulating half-life, suggesting that
pyridine oxidation was not being sufficiently retarded.
P3 methyl-substituted 3-pyridyl analogues are repre-
sented by compounds 25-33. The activity of both
A-80987 and gave plasma levels in excess of 2 µM (Cmax
/
EC50 > 50). Further improvements over A-80987 were
observed upon incorporation of five-membered hetero-
cycles, specifically thiazoles and oxazoles, at the P3
position (Table 2). To a first approximation, the struc-
ture-activity relationships of the 2-alkyl-4-thiazolyl
analogues 66-79 paralleled that of the pyridyl inhibi-
tors 5-24. In detail, however, the SAR diverged in
two important ways. First, whereas in the pyridyl
series alkyl groups larger than ethyl offered no advan-
tage, the following trend was observed in the 4-thiazolyl
series: i-Pr ≈ t-Bu > Et > Me > H. Second, with larger
substituents on the thiazolyl groups, the activity of the
N-methylurea analogues (Q ) NCH3) became strongly
dependent upon the regioisomeric position of the core
hydroxyl group. Thus, compounds 77 and 79, in which
the hydroxyl group was situated distal to the P2 valine
residue (B ) OH), were 10-fold more potent against HIV
than the corresponding regioisomers 76 and 78, which
contained the hydroxyl group proximal to the P2 valine
(A ) OH). This regioisomeric divergence was not
observed in carbamate analogues (cf. 74-75). The
culmination of these two trends resulted in the identi-
fication of compound 77 (A-83962),19 which was ca.
8-fold more potent than A-80987 and 20-fold more
potent than the corresponding pyridyl N-methylurea
A-80613 (compound 6).33 Although the oral bioavail-
ability of 77 in rats was moderate (12%), the Cmax/EC50
ratio exceeded that of A-80987 by almost 2-fold. More

Discovery of Ritonavir
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4 605
Ta ble 1. Activity of P3 Pyridine Analogues
solubility (µg/mL)
% inhibn EC50 CCIC50
pH
4.0
pH
7.4
T1/2 Cmax Tmax
(h) (µM) (h) (µg‚h/mL) (%) EC50
AUC
F
Cmax/
no.
R1
R2
X
Y
Z
A
B
(nM)
61 (0.5) 0.22
OH 55 (0.5) 0.58
80 (0.5) 0.10
(µM)
(µM)
5
6
7
8
9
H
H
CH3
CH3
CH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CH3
CH3
CH3
CH3
CH3
OCH3
OCH3
OCH3
OCH3
H
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
CH
CH NCH3
CH
CH
CH NCH3 OH
CH NCH3
CH NH
CH NH
CH
CH
CH NCH3 OH
CH NCH3
CH
CH
CH NCH3 OH H
CH NCH3
CH
CH
CH NCH3 OH H
CH NCH3
N
N
N
N
N
N
N
N
N
CH
CH
O
OH
H
OH
H
H
>100
>100
>100
122
8.6 1.90 4.11 0.25
178 1.22 2.18 0.15
3.5 0.74 1.15 0.33
11
101
137
1.98
1.22
0.61
0.90
0.95
1.66
26
22
7.5 12
16
16
32
19
3.8
O
O
H
OH 65 (0.5) 0.051 >100
1.95 1.74 0.10
0.22 2.42 0.25
0.68 3.39 0.29
34
7
13
H
43 (0.5) 0.33
OH 56 (0.5) 0.26
69 (0.5) 0.30
OH 76 (0.5) 0.34
76 (0.5) 0.019
81
59
75
10 CH3
11 CH3
12 CH3
13 Et
14 Et
15 Et
H
OH
H
OH
H
H
85
91
137
14
1.7
1.2
>100
O
O
H
82
4.5 0.74 1.26 0.25
0.5 0.39 0.59 0.20
18
0.81
0.47
0.73
16
8.0 26
23
66
OH 78 (0.5) 0.023 >100
H
63 (0.5) 0.067
OH 69 (0.5) 0.25
57
59
781
0.52 1.53 0.15
23
16 Et
H
OH
H
17 i-Pr
18 i-Pr
19 i-Pr
20 i-Pr
21 t-Bu
22 t-Bu
23 t-Bu
24 t-Bu
25 CH3
26 CH3
27 CH3
28 CH3
29
30
31
32
33
34
35
36
37
38 OCH3
39 OCH3
40 OCH3
41 OCH3
42
43
44
45
46
47
48
49
O
O
H
81 (0.5) 0.049 >100
46
71
2.2 1.31 0.31 0.25
2.3 0.57 0.20 0.25
0.31
0.14
8.3
5.6
6
7
OH 83 (0.5) 0.027
62 (0.5) 0.18
OH 71 (0.5) 0.18
76 (0.5) 0.064
OH 73 (0.5) 0.055
65 (0.5) 0.60
OH 52 (0.5) 0.31
20
19
19
19
21
19
19
H
OH
H
N
N
N
N
O
O
H
H
OH
H
CH
CH
CH
CH
CH
CH
CH
CH
CH
N
N
N
N
CH
CH
CH
CH
O
O
H
69 (0.5) 0.092 >100
98
29
85
6.2 0.28 0.48 0.30
1.0 0.38 0.33 0.15
3.4
0.21
0.07
5.9
1.6
5
7
OH 80 (0.5) 0.051 >100
H
OH 63 (0.5) 0.12
H
OH 56 (1.0) 0.057 >100
H
OH 67 (1.0) 0.13
H
H
NCH3 OH
NCH3
O
O
47 (0.5) 0.48
>100
>100
H
OH
H
532
48
0.53 0.69 0.10
0.31
0.24
0.27
0.29
0.06
9.7
6.3
6.4 14
6.7
2.1
6
5
H
H
H
H
H
H
H
H
H
59 (1.0) 0.087 >100
3.3 0.57 0.45 0.10
10
43
50
0.95 0.78 0.15
0.44 1.22 0.10
0.35 0.31 0.20
NCH3 OH
60 (1.0) 0.41
>100
>100
>100
65
736
870
3
2
NCH3
NH
O
H
OH
OH
H
65 (1.0) 0.63
72 (0.5) 0.039
48
39
77
4.9 0.46 1.20 0.20
8.5 0.25 0.29 0.10
3.3 0.60 2.17 0.15
0.38
0.05
0.82
0.29
0.07
0.03
0.05
0.05
0.16
0.03
0.31
0.32
7.7 31
O
OH 74 (0.5) 0.049
60 (0.5) 0.10
OH 60 (0.5) 0.076
75
61
87
1.3
17
6
22
CH NCH3 OH
CH NCH3
N
N
N
N
N
N
N
N
CH
CH
H
H
OH
H
30
66
15
0.57 1.05 0.10
1.19 0.30 0.10
0.80 0.20 0.10
5.7 14
O
O
H
79 (0.5) 0.061 >100
329
27
19
612
3.2
46
127
200
1.5
0.9
0.8
0.4
5
4
2
1
H
H
H
OH 79 (0.5) 0.046 >100
H
OH 64 (0.5) 0.16
H
NCH3 OH
NCH3
O
O
NCH3 OH
NCH3
O
O
53 (1.0) 0.27
>100
>100
1.8 0.17 0.44 0.10
H
OH
H
58
0.28 0.18 0.15
0.6 0.43 0.56 0.15
5.3 0.33 0.19 0.10
H
H
H
H
H
H
H
H
OCH3 CH
OCH3 CH
OCH3 CH
OCH3 CH
NH2
NH2
NH2
NH2
H
96 (0.5) 0.054 >100
3.3 10
0.7 10
7.9 29
OH 83 (0.5) 0.019
84 (0.5) 0.063
OH 84 (0.5) 0.068
57 (0.5) 0.85
OH 59 (0.5) 0.69
56 (1.0) 0.77
OH 48 (0.5) 0.34
50 (0.5) 0.35
OH 63 (0.5) 0.37
52 (0.5) 7.60
OH 53 (0.5) 9.65
60 (0.5) 0.20
OH 58 (0.5) 0.11
55
58
H
18
69
0.32 1.83 0.10
0.22 1.75 0.10
H
OH
H
61
10
26
N
N
N
N
CH
CH
CH
CH
N
N
N
N
N
N
N
N
N
H
>100
>100
>100
>100
>100
>100
89
>100
>100
>100
CH NCH3 OH H
CH NCH3
H
OH
H
OH
H
OH
H
50 NH2
51 NH2
52
53
54
55
56 CH3
57 CH3
58 CH3
59 CH3
60
61
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
O
O
O
O
O
O
O
O
H
862
739
22
33
0.72 0.20 0.10
0.50 0.14 0.10
0.02
0.03
0.5
0.3
1.0
0
H
H
H
H
H
NH2
NH2
H
H
H
H
H
H
CH3
CH3
CH3
CH3
CH3
CH3
H
H
169
99
38
24
0.23 1.03 0.20
0.20 0.50 0.20
0.21
0.10
4.5
2.4
5
4
OH
H
H
69 (0.5) 0.097 >100
45
5.9
OH 73 (0.5) 0.071
41 (0.5) 1.44
OH 64 (0.5) 0.26
64 (0.5) 0.22
OH 67 (0.5) 0.10
59
>100
>100
>100
96
55
13
1.15 1.33 0.10
0.64
10
19
NCH3 OH H
NCH3
O
O
O
O
H
OH
H
OH
H
H
H
H
77
38
14
11
0.18 0.35 0.20
0.21 0.86 0.30
0.30 0.35 0.20
0.43 2.06 0.10
0.30 1.52 0.15
0.30 1.13 0.25
0.07
0.19
0.12
0.48
0.55
0.33
2.8
6.0
5.7
10
17
2
9
5
53
2
62 CH3
63 CH3
64 CH3
65 CH3
H
66 (0.5) 0.071 >100
N
N
N
OH 70 (0.5) 0.039 >100
H
OH 67 (0.5) 0.12
562
719
769
112
93
227
NCH3 OH
NCH3
50 (0.5) 0.66
>100
>100
H
6.4 10
importantly, the half-life of 77 (1.03 h) was longer than
that observed for any previous N-methylurea-based
inhibitor. Of the analogous 4-oxazolyl inhibitors 80-
83, compound 83 showed potency and oral bioavailabil-
ity similar to those of 77 with the exception that 83 was
absorbed and eliminated more rapidly than 77.
Encouraged by the above results, we examined 2-sub-
stituents of varying polarity. Methoxymethyl substitu-

606 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4
Ta ble 2. Activity of P3 Five-membered Heterocyclic Analogues
Kempf et al.
solubility (µg/mL)
%
inhibn EC50 CCIC50
(nM) (µM) (µM)
pH
4.0
pH
7.4
T1/2 Cmax Tmax
AUC
F
Cmax/
no.
R
X
Y
Z
Q
A
B
(h) (µM) (h) (µg.h/mL) (%) EC50
66
67
68 CH3
69 CH3
70 Et
71 Et
72 Et
73 Et
74 i-Pr
75 i-Pr
76 i-Pr
77 i-Pr
78 t-Bu
79 t-Bu
80 i-Pr
81 i-Pr
82 i-Pr
H
H
N
S
S
S
S
S
S
S
S
S
S
S
S
S
S
O
O
O
O
S
S
S
S
S
S
S
S
S
CH
CH
N
N
N
N
N
N
N
N
N
N
O
O
O
O
N
N
CH
NCH3 OH
NCH3
NCH3 OH
H
41 (0.5) 0.74
OH 82 (0.5) 0.49
55 (0.5) 0.21
OH 72 (0.5) 0.18
75 (0.5) 0.065
OH 76 (0.5) 0.042
58 (0.5) 0.29
OH 73 (0.5) 0.26
83
>100
>100
>100
59
240
204
115
205
53
1.40 3.13 0.37
0.67 0.64 0.10
1.49 2.01 0.25
0.72 2.10 0.15
1.66
0.29
1.08
1.06
0.42
0.33
0.87
1.05
0.22
0.05
0.95
0.53
0.51
0.23
0.02
0.20
0.09
1.10
0.41
0.17
0.63
35
5.2
15
26
9.5 15
7.4 31
20
17
3.0 43
1.6 21
24
12
9.3
5.3 13
n/a 10
5.0 24
3.3
18
4
1
10
12
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
S
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
S
H
62
25
36
H
NCH3
H
O
O
OH
H
H
1.6
0.09 0.49 0.99 0.10
65
61
60
21
188
243
13
1.6
12
28
0.8
2.4
9.4
7.9
4.6
4.0
1.8
0.7
0.75 1.30 0.15
0.74 1.46 0.20
1.20 1.49 0.25
0.68 0.47 0.25
0.24 0.21 0.10
0.75 1.65 0.25
1.03 0.98 0.33
0.36 0.83 0.33
0.29 0.38 0.33
NCH3 OH
H
5
6
NCH3
H
O
O
OH
H
H
70 (0.5) 0.011 >100
OH 72 (0.5) 0.010
63 (0.5) 0.29
OH 78 (0.5) 0.029
60 (0.5) 0.37
OH 71 (0.5) 0.029
19
32
47
19
17
6.4
NCH3 OH
NCH3
NCH3 OH H
H
80
66
49
52
11
3.8
727
252
6
34
2
H
NCH3
H
O
O
OH
H
H
71 (0.5) 0.018 >100
ns
0.17 0.10
OH 86 (0.5) 0.020
60 (0.5) 0.033
OH 79 (0.5) 0.061
61 (0.5) 0.16
OH 87 (0.5) 0.068
52
62
63
44
100
0.33 0.46 0.20
0.29 0.29 0.10
0.22 2.13 0.25
0.91 0.32 0.10
0.28 0.38 0.10
0.27 1.79 0.23
NCH3 OH
H
264
39
9
35
2
6
19
83 i-Pr
84 i-Pr
85 i-Pr
NCH3
CH2
CH2
O
H
OH
H
OH
H
H
9.0
3.4
10
34
175
3.9
10
86 MeOCH2
87 MeOCH2
88 MeOCH2
89 MeOCH2
90 (Me)2N
91 4-morph
92 4-morph
93
94
95 Et
96 Et
97 i-Pr
98 i-Pr
99
100
101 MeOCH2
102 MeOCH2
103
104
105 t-Bu
106 t-Bu
107 t-Bu
108 t-Bu
109 CH3O
110 CH3O
H
66 (0.5) 0.096 >100
O
OH 69 (0.5) 0.076 >100
H
OH 69 (0.5) 0.076
H
H
NCH3 OH
54 (0.5) 0.33
61
65
53
29
61
77
8.8
0.56 2.29 0.50
0.39 1.35 0.20
1.24
0.49
16
8.0 18
7
NCH3
O
O
H
OH
OH
H
66 (0.5) 0.089
73 (0.5) 0.033
14
63
71
1.5
12
7.5
2.0
0.35 0.09 0.15
0.32 0.00 n/a
0.54 3.33 0.18
0.66 2.53 0.15
0.03
0.00
2.05
0.57
0.49
0.01
0.03
0.02
1.1
0
35
3
0
9
O
OH 70 (0.5) 0.033
56 (0.5) 0.37
OH 56 (0.5) 0.13
68 (0.5) 0.29
H
H
O
O
OH
H
H
>100
>100
66
S
6.8
8.5 19
CH
CH
CH
CH
CH
CH
CH
CH
O
O
OH
H
H
1.2
1.1
7.9
<0.01 1.23 0.34 0.28
9.2
0.5
0.8
0.7
1
2
1.0
2
S
S
S
O
O
S
S
S
OH 71 (0.5) 0.091 >100
63 (0.5) 0.10 >100
OH 71 (0.5) 0.076 >100
<0.01 0.19 0.16 0.10
O
O
OH
H
H
1.0
0.62 0.10 0.20
0.46 0.15 0.10
H
H
O
O
OH
H
H
59 (0.5) 0.54
OH 62 (0.5) 0.22
64 (0.5) 0.18
>100
>100
>100
O
O
OH
H
H
0.37 0.38 0.15
0.42
0.02
0.36
1.53
11
0.6
9.2
2
1
4
OH 72 (0.5) 0.085 >100
H
OH 59 (0.5) 0.081
H
2.4
<0.01 0.27 0.12 0.10
H
H
C(CH3) CH2O OH
C(CH3) CH2O
67 (0.5) 0.11
>100
0.45 0.41 0.45
S
H
61
nsa 0.96 n/a
n/a 12
CH
CH
CH
CH
CH
CH
N
N
N
N
O
O
O
O
OH
H
67 (0.5) 0.068 >100
OH 74 (0.5) 0.073
49 (0.5) 1.01
OH 80 (0.5) 0.048
H 70 (0.5) 0.22
19
56
55
61
NCH3 OH
H
NCH3
H
O
O
OH
H
OH 76 (0.5) 0.084 >100
a
ns: no iv sample; compound was too insoluble for iv dosing and was only dosed orally.
tion provided inhibitors (compounds 86-89) of similar
potency to the corresponding ethyl-substituted ana-
logues 70-73 and moderate oral bioavailability. 2-
Aminothiazoles showed only moderate anti-HIV activity
(data not shown), analogous to the aminopyridyl ana-
logues (vide supra). 2-Dialkylamino-substituted thi-
azoles, in contrast, demonstrated excellent potency (cf.
90-92); however, the oral bioavailability of 91 and 92
was very low. Isomeric thiazoles, oxazoles, and isox-
azoles were also examined as P3 pyridyl replacements.
2-Thiazolyl inhibitors 93-94 showed similar activity to
the corresponding unsubstituted 4-thiazolyl inhibitors
(66-67) and promising oral bioavailability. In contrast,
both the potency and bioavailability of the 5-thiazolyl
and 5-oxazolyl inhibitors 95-104 were inferior to that
of the 4-thiazolyl analogues. Isoxazolyl-based inhibitors
105-110 showed good anti-HIV activity in vitro but
were not evaluated for pharmacokinetic profile due to
anticipated lack of aqueous solubility.
Heter ocyclic P 2′ P yr id yl Rep la cem en ts. Given
the propensity of A-80987 to undergo metabolic oxida-
tion at both the P3 and P2′ pyridyl groups,19 we
concurrently investigated the replacement of the P2′
pyridyl group by other heterocyclic groups (Table 3).
Earlier studies had demonstrated that a hydrogen-bond-
accepting group two atoms from the linking methylene
was critical for tight binding to the HIV protease active
site. Furthermore, substituents on the P2′ heterocycle
were generally not tolerated (vide infra). 5-Pyrimidinyl
substitution provided inhibitors (compounds 111-113)
of equal potency to the pyridyl counterparts, but with
lower bioavailability. Furanyl analogues 114-117

Discovery of Ritonavir
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4 607
Ta ble 3. Activity of P2′ Heterocyclic HIV Protease Inhibitors
solubility (µg/mL)
% inhibn EC50 CCIC50
pH
4.0
pH
7.4
T1/2 Cmax Tmax
(h) (µM) (h) (µg‚h/mL) (%) EC50
AUC
F
Cmax/
no.
R
Q
X-Y
Z
A
B
(nM)
(µM)
(µM)
111
112
113
114
115
116 CH3
117 CH3
118
119
H
O
CH-N
N
N
N
O
O
O
O
N
N
OH
OH
H
OH
OH
OH
H
H
H
52 (0.5) 0.25
48 (0.5) 0.68
>100
>100
>100
>100
58
>100
>100
>100
>100
39
437
444
5.9
166
15
251
271
3.3
44
0.9
0.6
2.2
4.2
0.50 1.18 0.20
0.35 2.91 0.20
0.87 0.91 0.30
0.25 0.28
1.07 1.02 0.40
0.22 0.00 n/a
0.28
0.87
0.22
0.23
0.50
0.00
0.02
1.49
4.24
6.7
13
4.7
4.3
2.0
1.1
1.0
0.0
3.3
5.7
3.7
H
H
H
H
NCH3 CH-N
NCH3 CH-N
O
OH 60 (0.5) 0.45
5.1
n/a
9.3
0.0
n/a
n/a
47
CH
H
H
H
78 (0.5) 0.23
64 (0.5) 1.05
72 (0.5) 0.090
nsa
NCH3 CH
O
O
O
O
CH
CH
S
OH 75 (0.5) 0.047
54 (0.5) 0.52
OH 59 (0.5) 0.55
ns
ns
0.16 0.14
2.98 0.14
H
H
OH
H
H
3.7
7.3
S
1.23 2.04 0.37
a
ns: no iv sample; compound was too insoluble for iv dosing and was only dosed orally.
showed excellent potency but, with the exception of
compound 115, very poor bioavailability, possibly due
to low aqueous solubility. In contrast, 5-thiazolyl
analogues 118 and 119 showed excellent pharmaco-
kinetic properties following oral dosing, despite limited
solubility (compound 118 was too insoluble for iv dos-
ing). Key to the large AUC following oral dosing of 119
(A-81525)19 was a prolonged absorption profile and
significantly longer half-life. Consequently, although
the Cmax/EC50 ratio for 119 was modest, plasma levels
were detectable >6 h following a 10 mg/kg dose in rats.
Id en tifica tion of Riton a vir . From the above stud-
ies, we ascertained that the pharmacokinetic profile of
analogues of A-80987 could be improved by replacement
of either pyridyl moiety with more highly electron-
deficient heterocycles, in particular thiazolyl or oxazolyl
groups. Furthermore, alkyl substitution on the P3
heterocycle led to an improvement in anti-HIV activity
in vitro. We therefore anticipated that combining the
elements of A-83962 (77) and A-81525 (119) would
provide inhibitors of unique potency and pharmaco-
kinetic properties. However, due to the higher pKa of
thiazole, the aqueous solubility of both 77 and 119 was
significantly lower than that of A-80987 at pH 4.0,
potentially posing a barrier to efficient absorption.33,40
Compounds 120 and 121 displayed excellent anti-HIV
potency. However, neither inhibitor was soluble enough
in an acceptable formulation for intravenous dosing, and
oral administration provided very low plasma levels
(Table 4). Previous studies had shown that an N-
methylurea linkage between the P3 and P2 groups
provided higher solubility than a carbamate linkage.25,33
Accordingly, the N-methylurea inhibitors 122 and 123
could be dosed in solution. Compound 123 (ABT-538,
ritonavir) retained the potency of A-83962 and gave
plasma levels after oral administration >100-fold in
excess of the EC50. Furthermore, the levels of ritonavir
in plasma declined much more slowly than previous
inhibitors, and levels in rats exceeded the in vitro EC50
for >8 h following a single 10 mg/kg dose (Figure 2). In
contrast, levels of A-80987 exceeded the EC50 for only 2
h. The pharmacokinetic behavior of ritonavir was
further investigated in mice, beagle dogs, and cynomol-
gus monkeys. The plasma concentration curves are
shown in Figure 2. Pharmacokinetic parameters have
been published elsewhere.19 In each species, levels were
sustained in excess of the EC50 for >8 to >12 h. The
extended pharmacokinetic profile of ritonavir correlated
with a suppression of the rate of metabolism in vitro
using dog liver microsomes.19 Compounds 77 and 119
were metabolized 5 times more slowly than A-80987.
The rate of metabolism of ritonavir, in which both
pyridyl groups have been replaced with the more
electron-deficient thiazole, was ca. 20-fold lower than
that of A-80987.19
An a logu es of Riton a vir . Following the identifica-
tion of ritonavir (123), we initiated a detailed investiga-
tion of the structure-activity relationships of analogous
inhibitors. Results of those studies are provided in
Table 4. Not unexpectantly, substitution of thiazole
with oxazole at either the P3 or P2′ position provided
inhibitors (124-127) of similar profile to ritonavir, both
in potency and pharmacokinetic behavior. Each of the
oxazole analogues 125-127 showed more rapid absorp-
tion than ritonavir, potentially due to higher aqueous
solubility. However, plasma concentrations of the ox-
azolyl inhibitors were maintained for a shorter period
of time. A variety of 2-substituents on the P3 hetero-
cycle were examined (compounds 128-139). As antici-
pated, activity was systematically enhanced with in-
creasing substitution up to isopropyl (compounds 128-
132). Increasing the size even further (compound 133)
was detrimental, but both cyclopropyl- and cyclobutyl-
substituted inhibitors (134-137) were comparable in
activity to the isopropyl analogues 122 and 123. Like
ritonavir, most of the alkyl substituted analogues (130-
132, 134-135, 137) showed excellent pharmacokinetics
in the rat model, with oral bioavailabilities of 57-100%,
Cmax values in excess of 3 µM, AUC values of g9 µg‚h/
mL, and, in several cases, Cmax/EC50 values of >100. In
contrast, the morpholinyl-substituted inhibitors 138-
139, while retaining high potency, showed very low oral
bioavailability, a difference also observed in the P2′
pyridyl analogues 91-92 (vide supra). We also exam-
ined the P3 5-thiazolyl compounds 140-141. Although
the potency of 140-141 declined with respect to the
4-thiazolyl counterparts 122-123, the pharmacokinetic
profile of 141 was notable in that high Cmax and AUC
values were obtained even though the compound had
limited solubility and was thus dosed as a suspension.
We next examined changes at the N-methylurea
linker concurrent with modifications of the P2 amino

608 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4
Ta ble 4. Analogues of Ritonavir
Kempf et al.
solubility (µg/mL)
%
AUC
T1/2 Cmax Tmax (µg‚h/
(h) (µM) (h) mL) (%) EC50
inhibn EC50 CCIC50
pH
4.0
pH
7.4
F
Cmax/
no.
R1
X
Y
Q
AA
Val
Z
U
R2
A
B
(nM)
75 (0.5) 0.010
OH 79 (0.5) 0.005
67 (0.5) 0.19
OH 79 (0.5) 0.025
H 55 (0.5) 0.98
OH 80 (0.5) 0.043
OH 86 (0.5) 0.013
OH 70 (0.5) <0.032 >100
OH 75 (0.5) 0.79 62
OH 82 (0.5) 0.077 >100
(µM)
(µM)
120 i-Pr
121 i-Pr
122 i-Pr
123 i-Pr
124 i-Pr
125 i-Pr
126 i-Pr
127 i-Pr
128 CH3
129 CH3
130 Et
S
CH
CH
O
O
S
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
H
OH
H
OH
H
OH
H
H
H
H
H
OH
H
H
H
OH
H
OH
H
OH
H
OH
H
OH
H
H
OH
OH
H
OH
H
H
OH
H
H
H
H
OH
H
OH
H
OH
H
H
H
H
H
OH
OH
H
H
100
51
56
57
57
61
61
0.28
4.6
7.1
0.01 nsa 0.00 n/a
1.9
3.7
5.3
0.00
0.04 0.10 0.01 n/a
2.0 2.29 1.75 1.70 16
1.2 2.62 2.0 7.75 78
0
0
10
12
S
S
S
O
O
S
O
S
O
S
S
O
S
S
S
S
S
S
S
CH
CH
S
S
S
S
S
S
S
S
O
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
O
O
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Ala
Ala
Ala
Ala
Ala
Ala
Ala
Ala
Ala
Ala
Ala
Ala
â-Ala
â-Ala
â-Ala
â-Ala
Gly
Gly
Gly
Val
Val
Val
Val
S
S
S
S
S
O
O
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
CH
CH
CH
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
ns
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
H
6.9
105
34
22
57
21
9.3
32
0.56 1.59 0.58 2.65 23
127
0.26 2.05 0.15 1.85 35 >64
H
57 (0.5) 0.28
63
59
100
19
nd
5.12 1.00 9.37 167
18
117
458
131 Et
132 Et
OH 75 (0.5) 0.053
OH 75 (0.5) 0.018
OH 68 (0.5) 0.14
0.8 6.17 1.20 14.17 70
2.0 8.01 0.33 8.91 57
133 3-pent
134 c-Pr
135 c-Pr
136 c-Bu
137 c-Bu
138 4-morph
139 4-morph
140 i-Pr
141 i-Pr
142 i-Pr
143 i-Pr
144 i-Pr
145 i-Pr
146 i-Pr
147 i-Pr
148 i-Pr
149 i-Pr
150 i-Pr
151 i-Pr
152 i-Pr
153 i-Pr
154 i-Pr
155 i-Pr
156 i-Pr
157 i-Pr
158 i-Pr
159 i-Pr
160 i-Pr
161 i-Pr
162 i-Pr
163 i-Pr
164 i-Pr
165 i-Pr
166 i-Pr
167 i-Pr
168 i-Pr
169 i-Pr
170 i-Pr
171 i-Pr
172 i-Pr
173 i-Pr
174 i-Pr
175 i-Pr
176 i-Pr
177 i-Pr
178 i-Pr
179 i-Pr
180 i-Pr
181 i-Pr
H
53 (0.5) 0.15
OH 69 (0.5) 0.047
50 (0.5) 0.23
OH 71 (0.5) 0.039
45
48
20
26
0.53 3.98 2.30 10.70 175
1.1 6.46 2.17 21.86 58
27
138
H
0.40 2.48 1.20 6.29 60
64
0
7
CH
CH
S
O
O
O
O
H
72 (0.5) 0.080 >100
0.9
5.3
1.1
2.9
ns
ns
0.00 n/a
0.17 0.10 0.04 n/a
0.00 n/a
OH 71 (0.5) 0.025
1.9 1.1
OH 64 (0.5) 0.14
44 (0.5) 1.40
60
36
22
14
24
19
20
63
56
61
52
100
H
S
0.50 5.80 0.92 14.27 151
ns 3.33 3.80 9.85 n/a
42
14
CH NEt
CH NEt
CH NcPr
CH NcPr
CH
CH
H
4.6
4.3
2.8
4.3
2.8
2.8
OH 72 (0.5) 0.25
OH 44 (0.5) 0.28
H
H
1.9
2.72
O
O
48 (0.5) 1.69
OH 74 (0.5) 0.10
56 (0.5) 1.35
0.53 1.09 0.40 1.23 19
11
CH NCH3
CH NCH3
CH NCH3
CH NEt
CH NEt
CH NPr
CH NiBu
CH NnBu
CH NcPr
CH NcPr
CH NCH3
CH NCH3
CH NcPr
CH NcPr
CH NEt
CH NPr
CH NiBu
CH CH2O
CH CH2O
CH
H
26
14
OH 70 (0.5) 0.039
OH 66 (0.5) 0.059
0.90 1.68 0.50 1.60 31
4.0 3.60 0.25 2.86 23
44
61
H
65 (0.5) 0.089 >100
OH 73 (0.5) 0.014
OH 50 (0.5) 0.14
OH 54 (0.5) 0.18
OH 50 (0.5) 0.29
52
19
17
ns
0.00 n/a
0.00
0
0
17
H
9.4 14.0
45
41
57
55
59
48
13
16
56
58
8.0
5.3
21
33
OH 57 (0.5) 0.55
H
OH
H
OH
0.61 5.19 1.80 11.99 76
9
1.3
1.1
2.2
2.0
0.96 0.63 0.10 0.19 7.1
0.34 0.65 0.28 0.40 9.4
0
0
73
14
13.5
17.9
7.9
8.4
OH 66 (1.0) 2.1
41
OH
OH
1.2
1.6
1.0
1.2
35
15
OH 69 (0.5) 0.10
>100
25
>100
>100
25
18
18
59
H
H
73 (0.5) 0.017
50 (0.5) 0.82
ns
0.00 n/a
0.00
0
0
CH
OH 61 (0.5) 0.21
OH 69 (0.5) 0.18
OH 50 (0.5) 0.55
OH 37 (0.5) 1.1
OH 36 (0.5) 0.81
CH CH2NMe Val
CH CH2NEt Val
CH CH2NPr Val
CH CH2NMe Ala
CH CH2NEt Ala
CH CH2NPr Ala
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
CH NCH3
H
H
H
H
H
H
OH
OH
H
2.3
4.1
2.0
2.9
24
18
22
18
Val
Val
Val
Val
Val
Val
Ala
CH CH3 OH
CH CH3
CH i-Pr OH
63 (0.5) 0.075
5.7
3.7
2.9
2.8
0.46 2.08 0.70 2.37 28
0.44 0.65 1.20 0.91 23
28
6
H
OH 68 (0.5) 0.11
1.5 1.5
H
19
CH i-Pr
H
H
H
H
OH 68 (0.5) 0.20
OH 71 (0.5) 0.009
OH 82 (0.5) 0.012
19
28
57
O
O
O
H
H
H
6.0
36
4.6
20
0.59 0.37 0.20 0.27 7.7
0.40 0.49 0.10 0.18 3.8
44
42
OH 64 (0.5) 0.042 >100
a
ns: no iv sample; compound was too insoluble for iv dosing and was only dosed orally.
acid of ritonavir. The rationale for this study was 2-fold.
First, previous studies had identified N-demethylation
of the urea linkers of A-77003 as a major metabolic
pathway in addition to pyridine N-oxidation.41 Second,
valine to isoleucine mutants at position 32 in HIV
protease were selected by serial passage of HIV in the

Discovery of Ritonavir
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4 609
based HIV protease inhibitor with good potency and
moderate oral bioavailability33 to the identification of
ABT-538 (ritonavir), which displays excellent potency
and pharmacokinetic properties. Key to the discovery
of ritonavir was the modification of chemical functional-
ity within A-80987 to groups more stable to oxidative
metabolism by the cytochrome P450 enzymes present
in the intestine and liver. In a multifaceted approach,
both steric and electronic modifications to the oxida-
tively susceptible pyridyl groups of A-80987 were in-
vestigated. Ultimately, the replacement of the pyridyl
groups with less electron rich thiazolyl and oxazolyl
groups led to substantial reductions in the metabolic
degradation and excretion of the inhibitors.19 As a
result, high plasma levels were achieved in several
animal species and humans19 and sustained for hours
following a single dose. The improved pharmacokinetic
properties of ritonavir appear to be related not only to
increased chemical stability but also to metabolic inhi-
bition. Ritonavir has recently been shown to be not only
metabolized by but also a potent inhibitor of the 3A4
isozyme of cytochrome P450 (CYP 3A4).44 Inhibition of
CYP by ritonavir is associated with the unhindered
nitrogen atom on the unsubstituted P2′ 5-thiazolyl
group, which binds directly to the heme in the CYP
active site.45 However, the sole presence of the 5-thi-
azolyl group is insufficient, since A-81525 (compound
119) was 10-fold less effective at inhibiting CYP than
ritonavir.45 The difference in CYP inhibition between
A-81525 and ritonavir correlates to the more rapid rate
of metabolism of A-81525 in liver microsomes, presum-
ably at the pyridyl nitrogen of A-81525. Although not
measured, it is likely that rapid CYP-mediated metabo-
lism significantly compromises the oral bioavailability
of virtually all of the analogues of A-80987 described in
Tables 1 and 2. Furthermore, the attachment of sub-
stituents that may be more susceptible to oxidation to
the P3-thiazolyl group of ritonavir provided analogues
(e.g., compounds 138 and 139, containing morpholinyl
groups) that displayed very low plasma levels. It is
likely therefore that a combination of direct heme
binding and chemical stability contributes to the favor-
able pharmacokinetic profile of ritonavir via potent
F igu r e 2. Mean plasma concentrations of ritonavir (ABT-538)
after a single oral dose. Open circles, rat (10 mg/kg); closed
circles, dog (10 mg/kg); open squares, mouse (25 mg/kg); closed
squares, monkey (10 mg/kg).
presence of A-77003.42 It was our hope that reducing
the size of the P2 valine, which is common to both
A-77003 and ritonavir and interacts directly with valine-
32 via a hydrophobic contact,43 might produce analogues
that would not select this mutation. Inhibitors utilizing
various P2 groups and urea linkers are shown in Table
4 (compounds 142-164). Although substantial improve-
ments over ritonavir were not realized, several observa-
tions are of note. The activity of N-ethyl- and, in
particular, N-cyclopropylurea-valine analogues 142-
145 declined significantly compared to that of the
corresponding N-methylureas (122-123). In contrast,
with alanine as a P2 residue, the antiviral potency of
N-ethylurea 152 improved upon the potency of the
N-methylurea 149 and was indistinguishable from that
of ritonavir. Larger ureas were not tolerated, even with
a P2 alanine (compounds 153-157). Interestingly, the
P2 alanine carbamates 146-147 were much less potent
than either the N-methyl- or N-ethylureas and demon-
strated a preference for one regioisomer (147) over the
other, in contrast to the P2 valine carbamates 120-121.
Several of the alanine-based inhibitors, particularly 157,
showed excellent pharmacokinetic properties; however,
the potent N-ethylurea-alanine inhibitor 152 was too
insoluble to be dosed as a solution and gave no plasma
levels following oral administration.
More drastic changes led uniformly to a significant
fall in potency, including incorporation of â-alanine and
glycine at the P2 position and extension or contraction
of the P3-P2 linker functionality (compounds 158-
174). An exception to this trend was compound 166,
the extended analogue of 121. Although 166 showed
excellent antiviral potency, like 121 it was highly
insoluble and gave no plasma levels after oral dosing.
Finally, we briefly examined analogues of ritonavir with
changes at the P2′ 5-thiazolyl group. Alkyl substitution
on the thiazole ring resulted in decreased activity
(compounds 175-178). In contrast, incorporation of an
unsubstituted P2′ 5-isoxazolyl group gave inhibitors
(179-181) of very high potency but of relatively low oral
bioavailability.
inhibition of CYP in the liver and/or intestine.
A
consequence of this inhibition is that significant drug-
drug interactions may be observed between ritonavir
and other CYP3A4 substrates. Recently we have shown
that this interaction can be useful for the elevation of
the plasma concentrations of other investigational HIV
protease inhibitors (e.g., saquinavir,38 indinavir,20 nelfi-
navir,21 and VX-47822), which are also primarily me-
tabolized by CYP3A.45 This profound pharmacokinetic
enhancement by co-dosing with ritonavir may have
implications for the control of HIV infection with dual
protease inhibitor therapy and has served as the basis
for ongoing clinical trials utilizing dual protease inhibi-
tor therapy (ritonavir plus saquinavir, ritonavir plus
nelfinavir, and ritonavir plus indinavir).
Although replacement of the pyridyl groups of A-80987
with thiazolyl groups markedly slowed metabolism, the
aqueous solubility of the resulting inhibitors, in par-
ticular at pH 4, suffered from the lower pKa of thiazole
vs pyridine. Consequently, the intestinal absorption of
analogues such as compounds 120 and 121 were ap-
Discu ssion
These results describe the systematic structure-
activity studies which led from A-80987, a symmetry-

610 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4
Kempf et al.
parently solubility-restricted. Modification of the car-
bamate linking group to the more soluble N-methylurea
linker25 was sufficient to increase oral bioavailability
from 0% (120) to 78% (123, ritonavir). If, however,
ritonavir was dosed as a suspension rather than a
solution (using organic cosolvents), no plasma levels
were detected.19 These results suggest that a solubility
of ca. 1-2 µg/mL must be exceeded in compounds of this
type to permit intestinal absorption. Only two com-
pounds (118 and 141) that were dosed as suspensions
in rats gave significant plasma levels. In general,
increased aqueous solubility of HIV protease inhibitors
has been associated with higher oral bioavailabil-
ity.33,46,47 However, the superior oral bioavailability of
ritonavir suggests that if a solubility threshold that
allows formulation as a solution is exceeded, metabolism
may play a more important role than aqueous solubility
in determining the overall bioavailability and the dura-
tion of plasma levels in vivo.
markedly pronounced in ritonavir compared to previous
analogues. The importance assigned to these measures
was based on the assumption that high levels of any
inhibitor must be maintained in the plasma of patients
to exert a strong and durable antiviral effect in vivo.
Recently, this assumption has been supported by the
quantification of high viral turnover in patients28,29 and
in observing that the rate of appearance of mutations
in HIV protease in patients who received ritonavir
monotherapy inversely correlated with the trough plasma
concentrations in those patients.39 The maintenance of
high plasma levels is even more important with inhibi-
tors that are highly bound to human serum proteins.49,50
In separate studies (A. Molla et al., unpublished re-
sults), we discovered that high (g99%) binding of
ritonavir to human serum proteins attenuates the
antiviral activity of ritonavir by ca. 20-fold. Addition
of human serum to in vitro HIV cultures may give a
more accurate projection of the in vivo potency of HIV
protease inhibitors. Comparison of this value to the
sustained plasma concentrations of a given inhibitor in
humans may provide a useful predictor for antiviral
efficacy in vivo.
This study details the preclinical studies that led to
the identification of the HIV protease inhibitor ritonavir.
Clinical trials with ritonavir have produced dramatic
declines in plasma HIV RNA and substantial recovery
in CD4 cell levels in HIV-infected individuals.30,31 Ac-
companying these changes in surrogate markers, ad-
ministration of ritonavir is associated with increased
life-expectancy and time to AIDS-defining clinical
events32 and has recently been licensed by the Food and
Drug Administration for the therapy of HIV infection.
Ongoing clinical studies using ritonavir in combination
with reverse transcriptase inhibitors display evidence
of completely blocking the replication of HIV in vivo.51
Furthermore, dual therapy with ritonavir and other
protease inhibitors is under active investigation as an
effective, convenient regimen.52 The important role of
ritonavir and other HIV protease inhibitors in the
management of HIV infection will thus continue to be
demonstrated.
The antiviral activity of ritonavir in MT4 cells was
ca. 10-fold greater than that of A-80987. This difference
was due primarily to increased potency against HIV
protease (Ki values of 15 and 164 pM, respectively) and
appears to be a result of the additional hydrophobic
interaction of the P3 isopropylthiazolyl group of ritonavir
with the side chain of valine-82 (V82) of HIV protease.
The preliminary X-ray crystal structure of ritonavir
bound to HIV protease19 confirms this interaction and
rationalizes the structure-activity relationships ob-
served upon changing the size of the alkyl group at the
2-position of the P3 thiazolyl group in compounds 66-
79. Thus, both methyl groups of the isopropylthiazolyl
group of ritonavir (and presumably 77) interact with
V82. In the tert-butyl analogue (compound 79), the
additional methyl group would project away from V82
and would not be expected to provide an increase in
activity over compound 77. The close interaction of the
P3 isopropyl group with V82 also appears to affect the
response of HIV to selective pressure during treatment
with ritonavir. HIV strains mutated at valine-82 (to
alanine, phenylalanine, and threonine) have been shown
to be the primary species present during the initial
rebound of plasma HIV RNA during suboptimal mono-
therapy with ritonavir.39
Exp er im en ta l Section
Within this series of C2 symmetry-based inhibitors,
the structure of ritonavir appears to be optimized for
both anti-HIV activity and pharmacokinetics. Most
perturbations in structure led to compounds of inferior
properties. Analogues with better or comparable po-
tency (e.g., compounds 121, 139, 152, 166, 179, and 180)
generally suffered from diminished oral bioavailability.
Only the very close bisoxazolyl analogue 127 was nearly
equivalent to ritonavir. Interestingly, the P2 D-valinyl
diastereomer of ritonavir was recently discovered to be
virtually indistinguishable from ritonavir in activity,
pharmacokinetic properties and binding/inhibition of
cytochrome P450.48 Both diastereomers were shown to
adopt a similar binding orientation within the active site
of HIV protease.
Syn th esis of Mon oa m in es 5 a n d 6. Synthesis of 5a and
6a have been reported.33 Monoamines 5b-g and 6b-e,g were
prepared according to the following general procedure de-
scribed for 5d and 6d . Monoamines 6f and 6h were prepared
according to the general procedure described for 6h .
(2S,3S,5S)-2-Am in o-5-[N-[[(5-t h ia zolyl)m et h oxy]ca r -
bon yl]a m in o]-1,6-d ip h en yl-3-h yd r oxyh exa n e (5d ) a n d
(2S ,3S ,5S )-5-Am in o-2-[N -[[(5-t h ia zolyl)m e t h oxy]ca r -
bon yl]a m in o]-1,6-d ip h en yl-3-h yd r oxyh exa n e (6d ). 5-(Hy-
droxymethyl)thiazole (182) was prepared by a modification of
the procedure of Mashraqui and Keehn53 as follows. A solution
of 7.5 g (0.123 mol) of thioformamide and 18.54 g (0.123 mol)
of ethyl 2-chloro-2-formylacetate in 250 mL of dry acetone was
heated at reflux for 2 h. The solvent was removed in vacuo,
and the residue was purified by silica gel chromatography
using CHCl3 as an eluent to provide 11.6 g (60%) of ethyl
thiazole-5-carboxylate (Rf ) 0.25) as a light yellow oil. 1H
NMR (CDCl3) δ 1.39 (t, J ) 7 Hz, 3 H), 4.38 (q, J ) 7 Hz, 2 H),
8.50 (s, 1 H), 8.95 (s, 1 H). A mixture of 76 mmol of LiAlH4 in
250 mL of THF was cooled in an ice bath and treated with a
solution of 11.82 g (75.7 mmol) of the above ester in 100 mL
of THF dropwise over 1.5 h. After addition, the reaction
mixture was stirred for 1 h and treated cautiously with 2.9
In this study, we used a ratio of the Cmax obtained
after a 10 mg/kg oral dose in rats to the in vitro EC50
for anti-HIV activity as one measure by which inhibitors
were evaluated. An additional feature not reflected in
that ratio is the duration of plasma levels, which was

Discovery of Ritonavir
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4 611
mL of water, 2.9 mL of 15% NaOH, and 8.7 mL of water. The
solid salts were filtered, and the filtrate was set aside. The
crude salts were heated at reflux in 100 mL of EtOAc for 30
min. The resulting mixture was filtered, and the two filtrates
were combined, dried over Na2SO4, and concentrated in vacuo.
The product was purified by silica gel chromatography eluting
sequentially with 0%, 2%, 4% CH3OH in CHCl3, to provide a
75% yield of 182 (Rf ) 0.3, 4% CH3OH in CHCl3) which
solidified upon standing: NMR (CDCl3) δ 4.92 (s, 2 H), 7.78
(s, 1 H), 8.77 (s, 1 H); CIMS m/z 116 (M + H)+.
carboxaldehyde (183) as a colorless liquid. 1H NMR (CDCl3)
δ 7.89 (s, 1H), 8.12 (s, 1H), 9.87 (s, 1H); CIMS m/z 98 (M +
H)+. A solution of 627 mg (6.46 mmol) of 183 in CH3OH (10
mL) at 0 °C was treated under argon with 247 mg (6.46 mol)
of NaBH4. After 5 min, the reaction was quenched with
acetone and the solvent was removed in vacuo. Silica gel
chromatography using a gradient of CH3OH-CH2Cl2 (5%,
10%) afforded 408 mg (64%) of 5-(hydroxymethyl)oxazole as a
colorless oil: 1H NMR (CDCl3) δ 2.03 (t, J ) 6.0 Hz, 1H), 4.70
(d, J ) 6.0 Hz, 2H), 7.04 (s, 1H), 7.87 (s, 1H); CIMS m/z 117
(M + NH4)+, 100 (M + H)+. Further reaction according to the
A solution of 3.11 g (27 mmol) of 182 and excess 4-methyl-
morpholine in 100 mL of CH2Cl2 was cooled to 0 °C and treated
with 8.2 g (41 mmol) of p-nitrophenyl chloroformate. After
being stirred for 1 h, the reaction mixture was diluted with
CHCl3, washed successively with 1 N HCl, saturated aqueous
NaHCO3, and saturated brine, dried over Na2SO4, and con-
centrated in vacuo. The residue was purified by silica gel
chromatography (1-2% CH3OH/CHCl3) to provide 5.9 g (78%)
of (5-thiazolyl)methyl (4-nitrophenyl)carbonate (Rf 0.5, 4%
CH3OH/CHCl3) as a yellow solid: NMR (CDCl3) δ 5.53 (s, 2
H), 7.39 (dt, J ) 9, 3 Hz, 2 H), 8.01 (s, 1 H), 8.29 (dt, J ) 9, 3
Hz, 2 H), 8.90 (s, 1 H); CIMS m/z 281 (M + H)+. A solution of
500 mg (1.76 mmol) of diamine 233 and 480 mg (1.71 mmol) of
the above (4-nitrophenyl)carbonate in 20 mL of THF was
stirred at ambient temperature for 4 h. After removal of the
solvent in vacuo, the residue was purified by silica gel
chromatography using first 2% and then 5% CH3OH in CHCl3
to provide a mixture of 6d and 5d . Silica gel chromatography
of the mixture using a gradient of 0, 1, 2% CH3OH in 98:2
i-PrNH2-CHCl3 provided 110 mg (16%) of 6d (Rf 0.48, 96:2:2
CHCl3-CH3OH-i-PrNH2) and 185 mg (28%) of 5d (Rf 0.44,
96:2:2 CHCl3-CH3OH-i-PrNH2). 5d : NMR (CDCl3) δ 1.55
(dt, J ) 14, 8 Hz, 1 H), 1.74 (m, 1 H), 2.44 (dd, J ) 15, 1 Hz,
1 H), 2.75-3.0 (m, 4 H), 3.44 (m, 1 H), 4.00 (br t, 1 H), 5.28
(m, 3 H), 7.1-7.4 (m, 10 H), 7.86 (s, 1 H), 8.80 (s, 1 H); CIMS
m/z 426 (M + H)+. 6d : NMR (CDCl3) δ 1.3-1.6 (m, 2 H), 2.40
(dd, J ) 14, 8 Hz, 1 H), 2.78 (dd, J ) 5 Hz, 1 H), 2.88 (d, J )
7 Hz, 2 H), 3.01 (m, 1 H), 3.72 (br q, 1 H), 3.81 (br d, J ) 10
Hz, 1 H), 5.28 (s, 2 H), 5.34 (br d, J ) 9 Hz, 1 H), 7.07 (br d,
J ) 7 Hz, 2 H), 7.15-7.35 (m, 8 H), 7.87 (s, 1 H), 8.80 (s, 1 H);
CIMS m/z 426 (M + H)+.
1
procedure outlined above produced 5e and 6e. 5e: H NMR
(DMSO-d6) δ 1.16-1.30 (m, 1H), 1.36-1.47 (m, 1H), 2.56-2.66
(m, 2H), 2.75-2.85 (m, 1H), 2.89-3.01 (m, 1H), 3.53-3.71 (m,
3H), 4.97 (d, J ) 2.4 Hz, 2H), 7.01 (d, J ) 9 Hz, 1H), 7.11-
7.32 (m, 14H), 8.36 (s, 1H); CIMS m/z 410 (M + H)+. 6e: 1H
NMR (DMSO-d6) δ 1.20 (m, 1 H), 1.42 (m, 1 H), 2.60 (m, 2 H),
2.80 (m, 1 H), 2.90 (m, 1 H), 4.90 (AB, J ) 12, 10 Hz, 2 H),
4.70 (s, 1 H), 7.20 (m, 14 H), 8.36 (s, 1 H); CIMS m/z 410 (M
+ H)+.
(2S,3S,5S)-2-Am in o-5-[N -[[(2-isop r op yl-5-t h ia zolyl)-
m et h oxy]ca r b on yl]a m in o]-1,6-d ip h en yl-3-h yd r oxyh ex-
a n e (5g) a n d (2S,3S,5S)-5-Am in o-2-[N-[[(2-isop r op yl-5-
t h ia zoly l)m e t h ox y]c a r b o n y l]a m in o ]-1,6-d ip h e n y l-3-
h yd r oxyh exa n e (6g). Prepared from isobutyramide as
described for 5d /6d . 5g: 7%; Rf 0.25 (1% i-PrNH2/CHCl3); 1H
NMR (DMSO-d6) δ 1.27 (d, J ) 7 Hz, 6 H), 1.5-1.7 (m, 2 H),
2.41 (dd, J ) 14, 9 Hz, 1 H), 2.56 (dd, J ) 14, 9 Hz, 1 H),
2.6-2.8 (m, 4 H), 3.8 (br m, 2 H), 4.45 (br, 1 H), 5.06 (s, 2 H),
7.1-7.3 (m, 11 H), 7.59 (s, 1 H); CIMS m/z 468 (M + H)+. 6g:
11%; Rf 0.37 (1% i-PrNH2/CHCl3); 1H NMR (DMSO-d6) δ 1.28
(d, J ) 7 Hz, 6 H), 1.44 (m, 1 H), 1.63 (m, 1 H), 2.46 (dd, J )
12, 5 Hz, 1 H), 2.55-2.65 (m, 3 H), 2.80 (dd, J ) 12, 5 Hz, 1
H), 2.95 (br m, 1 H), 3.55-3.7 (m, 3 H), 5.06 (d, J ) 13 Hz, 1
H), 5.11 (d, J ) 13 Hz, 1 H), 6.98 (d, J ) 9 Hz, 1 H), 7.1-7.3
(m, 10 H), 7.60 (s, 1 H); CIMS m/z 468 (M + H)+.
(2S,3S,5S)-5-Am in o-2-[N-[[(5-isoxa zolyl)m eth oxy]ca r -
bon yl]am in o]-1,6-diph en yl-3-h ydr oxyh exan e (6h ). A mix-
ture of 1.54 g (5.41 mmol) of diamine 233 and 0.673 g (5.41
mmol) of phenylboric acid in anhydrous toluene (130 mL) was
heated at reflux under argon for 2 h with removal of H2O by
a Dean-Stark trap. The resulting yellow solution was allowed
to cool, and the solvent was removed in vacuo to give an oil
which solidified upon standing. The residue was taken up in
90 mL of THF, cooled to -40 °C, and treated dropwise under
Ar atmosphere over a period of 1 h with 1.11 g (3.78 mmol) of
(5-isoxazolyl)methyl (4-nitrophenyl)carbonate (prepared from
5-(hydroxymethyl)isoxazole in 40 mL of THF). The solution
was allowed to warm to -20 °C over the next 0.5 h and then
was stirred at 0 °C for 2.5 h and at room temperature for 1 h.
After removal of the solvent in vacuo, the residue was taken
up in EtOAc (200 mL), washed sequentially with 5% aqueous
K2CO3 (4 × 25 mL) and saturated brine (25 mL), dried over
Na2SO4, and concentrated in vacuo. Silica gel chromatography
of the residue using a gradient of CH3OH in CHCl3 (2%, 4%,
6%) afforded a mixture of the desired product and its regio-
isomer. Purification of the mixture on two consecutive 250 g
SiO2 columns (deactivated with 1% i-PrNH2-CH2Cl2) with a
gradient of i-PrNH2-CH2Cl2 (0.5%, 1%) afforded 6h as a sticky
solid (0.730 g, 1.78 mmol, 33%): 1H NMR (DMSO-d6) δ 1.17-
1.57 (m, 5H), 2.56-2.69 (m, 2H), 2.75-2.86 (m, 1H), 2.89-
3.00 (m, 2H), 3.53-3.71 (m, 3H), 5.06 (s, 2H), 6.32 (d, J ) 2.4
Hz, 1H), 7.11-7.30 (m, 10H), 8.54 (d, J ) 2.4 Hz, 1H); CIMS
m/z 410 (M + H)+.
(2S,3S,5S)-2-Am in o-5-[N-[[(5-pyr im idin yl)m eth oxy]car -
bon yl]a m in o]-1,6-d ip h en yl-3-h yd r oxyh exa n e (5b) a n d
(2S,3S,5S)-5-Am in o-2-[N-[[(5-p yr im id in yl)m et h oxy]ca r -
bon yl]a m in o]-1,6-d ip h en yl-3-h yd r oxyh exa n e (6b). 5b:
28%; 1H NMR (CDCl3) δ 1.57 (m, 1 H), 1.75 (m, 1 H), 2.44 (dd,
1 H), 2.78-2.89 (m, 3 H), 2.96 (m, 1 H), 3.44 (m, 1 H), 4.02
(m, 1 H), 5.10 (q, AB, 2 H), 5.40 (br, 1 H), 7.07-7.33 (m, 10
1
H), 8.73, (s, 2 H), 9.18 (s, 1 H). 6b: 13%; H NMR (CDCl3) δ
1.43 (m, 1 H), 1.53 (m, 1 H), 2.43 (dd, 1 H), 2.78 (dd, 1 H),
2.88 (d, 2 H), 3.04 (m, 1 H), 3.71 (dd, 1 H), 3.83 (d, 1 H), 5.09
(s, 2 H), 5.37 (br d, 1 H), 7.07-7.32 (m, 10 H), 8.74, (s, 2 H),
9.19 (s, 1 H).
(2S ,3S ,5S )-2-Am in o-5-[N -[[(3-fu r a n yl)m e t h oxy]ca r -
bon yl]a m in o]-1,6-d ip h en yl-3-h yd r oxyh exa n e (5c) a n d
(2S,3S,5S)-5-Am in o-2-[N-[[(3-fu r a n yl)m eth oxy]ca r bon yl]-
a m in o]-1,6-d ip h en yl-3-h yd r oxyh exa n e (6c). 5c: 36%; 1H
NMR (CDCl3) δ 1.57 (m, 1 H), 1.75 (d, 1 H), 2.45 (dd, 1 H),
2.80-2.87 (m, 3 H), 2.96 (m, 1 H), 3.46 (m, 1 H), 4.02 (m, 1
H), 4.95 (q, AB, 2 H), 5.14 (br s, 1 H), 7.13-7.43 (m, 13 H). 6c:
17%; 1H NMR (CDCl3) δ 1.42 (m, 1 H), 1.53 (d, 1 H), 2.40 (dd,
1 H), 2.78 (dd, 1 H), 2.88 (d, 2 H), 3.02 (m, 1 H), 3.71 (dd, 1 H),
4.96 (s, 2 H), 5.28 (br d, 1 H), 7.07-7.46 (m, 13 H).
(2S,3S,5S)-2-Am in o-5-[N-[[(5-oxa zolyl)m et h oxy]ca r -
bon yl]a m in o]-1,6-d ip h en yl-3-h yd r oxyh exa n e (5e) a n d
(2S,3S,5S)-5-Am in o-2-[N-[[(5-oxazolyl)m eth oxy]car bon yl]-
a m in o]-1,6-d ip h en yl-3-h yd r oxyh exa n e (6e). A flask was
charged with 1.02 g (5.96 mmol) of 5-(diethoxymethyl)oxazole54
and cooled to 0 °C. A solution of trifluoroacetic acid/CH2Cl2
(1:1 (v/v), 6.7 mL) and H2O (0.39 mL) was added, and the
solution was stirred at 0 °C for 10 min. The solvent was
removed in vacuo, and the residue was azeotroped with
toluene. Silica gel chromatography using a gradient of EtOAc-
hexane (20%, 30%, 40%) afforded 0.344 g (59%) of 5-oxazole-
Rep r esen ta tive P r oced u r es for P r ep a r a tion of P 3
Heter ocyclic En d P ieces. Heterocyclic end pieces were
prepared according to literature methods or according to the
following procedures.
P r ocedu r e A. N-Meth yl-N-[(6-eth yl-2-pyr idin yl)m eth yl]-
a m in e (184). A solution of 6 g of 4-hydroxy-1-hexene (pre-
pared from propionaldehyde and allylmagnesium bromide) and
18 g of ethyl chloro(hydroxyimino)oxalate in 150 mL of ether
was treated over 2.5 h with 21 mL of triethylamine. After
addition, stirring was continued for 3 h, and the solvent was

612 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4
Kempf et al.
removed in vacuo. Silica gel chromatography using ethyl
acetate in CH2Cl2 provided 5.5 g of ethyl 5-(2-hydroxybutyl)-
2-isoxazoline-3-carboxylate (184a ): 1H NMR (CDCl3) δ 0.96
(t, J ) 7 Hz, 3 H), 1.37 (t, J ) 7 Hz, 3 H), 1.55 (m, 2 H), 1.77
(ddd, J ) 14, 6, 3 Hz, 1 H), 1.90 (m, 1 H), 2.93 (td, J ) 17, 8
Hz, 1 H), 3.33 (dd, J ) 17, 11 Hz, 1 H), 3.78 (m, 1 H), 4.35 (q,
J ) 7 Hz, 2 H), 5.02 (m, 1 H). Swern oxidation (as described
below for the preparation of 184e) of 184a followed by silica
gel chromatography using 5% ethyl acetate in CH2Cl2 provided
5.05 g (92%) of ethyl 5-(2-oxobutyl)-2-isoxazoline-3-carboxylate
(184b): 1H NMR (CDCl3) δ 1.07 (t, J ) 7 Hz, 3 H), 1.36 (t, J
) 7 Hz, 3 H), 2.49 (q, J ) 7 Hz, 2 H), 2.71 (dd, J ) 17, 7 Hz,
1 H), 2.87 (dd, J ) 18, 8 Hz, 1 H), 3.03 (dd, J ) 17, 6 Hz, 1 H),
3.43 (dd, J ) 18, 11 Hz, 1 H), 4.35 (q, J ) 7 Hz, 2 H), 5.16 (m,
1 H). A solution of 1.05 g (4.9 mmol) of 184b in 150 mL of
ethanol was treated with 4 mL of Raney nickel and 1 mL of
48% HBF4 and stirred under 1 atm of H2. After 6 h, the
mixture was filtered, concentrated in vacuo, and purified by
silica gel chromatography using 10% ethyl acetate in CH2Cl2
to provide 430 mg (48%) of ethyl 6-ethylpyridine-2-carboxylate
(184c): 1H NMR (CDCl3) δ 1.34 (t, J ) 8 Hz, 3 H), 1.43 (t, J
) 7 Hz, 3 H), 2.95 (q, J ) 8 Hz, 2 H), 4.48 (q, J ) 7 Hz, 2 H),
7.36 (d, J ) 7 Hz, 1 H), 7.74 (t, J ) 7 Hz, 1 H), 7.94 (d, J ) 7
Hz, 1 H). A solution of 1.2 g (6.7 mmol) of 184c in 30 mL of
THF was cooled to 0 °C and treated with 6.7 mL (6.7 mmol) of
1 M LiAlH4 in THF. After 1 h, the mixture was sequentially
quenched with 0.5 mL of water, 15% NaOH, and 1 mL of
water, filtered, and concentrated in vacuo. Silica gel chroma-
tography using methanol in CH2Cl2 provided 870 mg of 6-ethyl-
2-(hydroxymethyl)pyridine (184d ): 1H NMR (CDCl3) δ 1.33
(t, J ) 8 Hz, 3 H), 2.87 (q, J ) 8 Hz, 2 H), 4.25 (br s, 1 H), 4.74
(s, 2 H), 7.07 (d, J ) 7 Hz, 1 H), 7.11 (d, J ) 7 Hz, 1 H), 7.64
(t, J ) 7 Hz, 1 H). A solution of 0.67 mL (9.4 mmol) of dry
DMSO in 10 mL of CH2Cl2 was cooled to -78 °C, treated
dropwise with 0.39 mL (4.5 mmol) of oxalyl chloride, stirred
under N2 for 15 min, treated with a solution of 530 mg (3.9
mmol) of 184d in 10 mL of CH2Cl2, and stirred for 20 min.
After dropwise addition of 2.6 mL (18.9 mmol) of triethylamine,
the solution was stirred for 20 min, quenched with water, and
extracted with three portions of CH2Cl2. The combined organic
layers were dried over Na2SO4 and concentrated, and the
residue was purified on silica gel using ethyl acetate in CH2Cl2
to provide 465 mg (89%) of 6-ethyl-2-pyridinecarboxaldehyde
(184e): 1H NMR (CDCl3) δ 1.36 (t, J ) 7 Hz, 3 H), 2.94 (q, J
) 7 Hz, 2 H), 7.40 (m, 1 H), 7.79 (m, 2 H), 10.07 (s, 1 H); CIMS
m/z 136 (M + H)+. A solution of 460 mg (3.4 mmol) of 184e in
32 mL of methanol in a high-pressure flask was treated with
6 mL of methylamine. After 16 h of stirring, the solution was
treated with 115 mg of 10% Pd/C and stirred under 4 atm of
H2 for 4 h. After filtration, the filtrate was concentrated in
vacuo and purified by silica gel chromatography using 94:4:2
CH2Cl2-CH3OH-isopropylamine mixtures to provide 379 mg
(74%) of 184: 1H NMR (CDCl3) δ 1.30 (t, J ) 7 Hz, 3 H), 2.54
(s, 3 H), 2.80 (q, J ) 7 Hz, 2 H), 3.00 (br, 1 H), 3.90 (s, 2 H),
7.06 (d, J ) 8 Hz, 1 H), 7.13 (d, J ) 8 Hz, 1 H), 7.57 (t, J ) 8
Hz, 1 H); CIMS m/z 151 (M + H)+.
di-tert-butyl dicarbonate (450 mg) was added. The reaction
mixture was stirred for an additional hour at ambient tem-
perature and stored at 0 °C for 16 h. After concentration of
the solvent, silica gel chromatography using a gradient (5%,
10%) of EtOAc in CH2Cl2 provided 1.85 g (79%) of methyl
6-[(tert-butyloxycarbonyl)amino]nicotinate, which was reduced
in a manner analogous to the preparation of 185 to provide
1.064 g of 186.
P r oced u r e D. N-Meth yl-N-[(2-isop r op yl-4-th ia zolyl)-
m eth yl]a m in e (187). A suspension of 100 g (1.15 mol) of
isobutyramide in 4 L of diethyl ether was stirred vigorously
and treated in portions with 51 g (0.115 mol) of P4S10
. The
resulting mixture was stirred at ambient temperature for 2
h, filtered, and concentrated in vacuo to provide 94.2 g (80%)
of crude 2-methylpropane thioamide (187a ), which was used
without further purification: 1H NMR (DMSO-d6) δ 1.08 (d, J
) 7 Hz, 6 H), 2.78 (heptet, J ) 7 Hz, 1 H), 9.06 (br, 1 H), 9.30
(br, 1 H); CIMS m/z 104 (M + H)+. A mixture of 94 g (0.91
mol) of 187a , 116 g (0.91 mol) of 1,3-dichloroacetone, and 110
g (0.91 mol) of MgSO4 in 1.6 L of acetone was heated at reflux
for 3.5 h. The resulting mixture was allowed to cool and
filtered, and the solvent was removed in vacuo to provide crude
4-(chloromethyl)-2-isopropylthiazole hydrochloride (187b) as
a yellow oil: 1H NMR (DMSO-d6) δ 1.32 (d, J ) 7 Hz, 6 H),
3.27 (heptet, J ) 7 Hz, 1 H), 4.78 (s, 2 H), 7.61 (s, 1 H); CIMS
m/z 176 (M + H)+. A solution of 40 g of 187b in 100 mL of
water was added dropwise with stirring to 400 mL of 40%
aqueous methylamine. The resulting solution was stirred for
1 h and then concentrated in vacuo. The residue was taken
up in CHCl3, dried over Na2SO4, and concentrated in vacuo.
Purification of the residue by silica gel chromatography using
10% CH3OH in CHCl3 provided 21.35 g (55%) of 187: 1H NMR
(DMSO-d6) δ 1.34 (d, J ) 7 Hz, 6 H), 2.56 (s, 3 H), 3.30 (heptet,
J ) 7 Hz, 1 H), 4.16 (s, 2 H), 7.63 (s, 1 H); CIMS m/z 171 (M
+ H)+.
P r oced u r e E. 4-(Hyd r oxym eth yl)-2-isop r op ylth ia zole
(188). A solution of 2.35 g (23 mmol) of 187a and 2.89 mL
(23 mmol) of ethyl bromopyruvate in 75 mL of acetone was
treated with excess MgSO4 and heated at reflux for 2.5 h. The
resulting mixture was allowed to cool, filtered, and concen-
trated in vacuo to an oil, which was taken up in CHCl3, washed
sequentially with aqueous NaHCO3 and brine, dried over
Na2SO4, and concentrated. The residue was purified by
chromatography on silica gel using CHCl3 as an eluent to
provide 3.96 g (86%) of ethyl 2-isopropyl-4-thiazolecarboxylate
(188a , Rf 0.21, CHCl3) as an oil: 1H NMR (CDCl3) δ 1.41 (t, J
) 8 Hz, 3 H), 1.42 (d, J ) 7 Hz, 6 H), 3.43 (heptet, J ) 7 Hz,
1 H), 4.41 (q, J ) 8 Hz, 2 H), 8.05 (s, 1 H); CIMS m/z 200 (M
+ H)+. A solution of 10 mL (10 mmol) of LiAlH4 in PhCH3
was diluted in a dry flask under N2 atmosphere with 75 mL
of THF. The resulting mixture was cooled to 0 °C and treated
dropwise with a solution of 3.96 g (20 mmol) of 188a in 10 mL
of THF. After addition, the solution was stirred at 0 °C for 3
h, diluted with ether, and treated with a small amount of
aqueous Rochelle’s salt. After stirring, the slurry was filtered
and washed with EtOAc, and the combined filtrates were
concentrated in vacuo. The residue was purified by silica gel
chromatography using 2% CH3OH in CHCl3 to provide 2.18 g
(69%) of 188 (Rf 0.58, 4% CH3OH in CHCl3): 1H NMR (CDCl3)
δ 1.39 (d, J ) 7 Hz, 6 H), 2.94 (br, 1 H), 3.31 (heptet, J ) 7
Hz, 1 H), 4.74 (s, 2 H), 7.04 (s, 1 H); CIMS m/z 158 (M + H)+.
P r ocedu r e B. 6-Eth yl-2-(h ydr oxym eth yl)pyr idin e (185).
Ethyl 6-ethylpyridine-2-carboxylate was prepared according
to the procedure of Kanemasa et al.55 A solution of the above
ester (1.2 g, 6.7 mmol) in 30 mL of THF was cooled to 0 °C
and treated with 6.7 mL of 1 M LiAlH4 in THF. The resulting
solution was stirred for 1 h, quenched by sequential addition
of 0.5 mL of H2O, 0.5 mL of 15% NaOH, and 1 mL of H2O,
filtered, and concentrated in vacuo. Flash chromatography
using 5-10% MeOH-CHCl3 provided 870 mg (95%) of 185:
1H NMR (CDCl3) δ 1.33 (t, J ) 7 Hz, 3 H), 2.87 (q, J ) 7 Hz,
2 H), 4.22 (br, 1 H), 4.74 (s, 2 H), 7.06 (d, J ) 8 Hz, 1 H), 7.11
(d, J ) 8 Hz, 1 H), 7.63 (t, J ) 8 Hz, 1 H).
P r oced u r e C. 6-[(ter t-Bu tyloxyca r bon yl)a m in o]-3-(h y-
d r oxym eth yl)p yr id in e (186). A solution of 2.013 g (9.32
mmol) of methyl 6-aminonicotinate, 2.235 g (10.25 mmol) of
di-tert-butyl dicarbonate, and 122 mg (1.0 mmol) of 4-(di-
methylamino)pyridine (DMAP) in acetonitrile (80 mL) was
stirred at ambient temperature for 4 h, and then additional
P r oced u r e F . 4-(Hyd r oxym eth yl)-2-isop r op yloxa zole
(189). A mixture of 10 g (63 mmol) of L-serine methyl ester
hydrochloride in 280 mL of dichloromethane was cooled to 0
°C and treated with 22 mL (153 mmol) of triethylamine. The
resulting mixture was treated dropwise with 8.1 mL (76 mmol)
of isobutyryl chloride, stirred for 20 min at 0 °C and at ambient
temperature for 40 min, partitioned between dichloromethane
and aqueous HCl, washed with aqueous NaHCO3 and satu-
rated brine, dried over Na2SO4, and concentrated in vacuo.
Silica gel chromatography using 2% methanol in CH2Cl2
provided 6.33 g (53%) of N-isobutyryl-L-serine methyl ester
(189a ): 1H NMR (CDCl3) δ 1.20 (d, J ) 7 Hz, 6 H), 2.47 (heptet,
J ) 7 Hz, 1 H), 2.57 (t, J ) 6 Hz, 1 H), 3.80 (s, 3 H), 3.96 (m,

Discovery of Ritonavir
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4 613
2 H), 4.68 (m, 1 H), 6.42 (br d, 1 H); CIMS m/z 190 (M + H)+.
In a dry, three-neck flask, 90 mL of anhydrous CH2Cl2 was
cooled under Ar atmosphere to -78 °C and treated with 6.6
mL (39 mmol) of trifluorosulfonic anhydride. The resulting
solution was treated dropwise with a solution of 10.5 g (52
mmol) of diphenyl sulfoxide56 in 90 mL of CH2Cl2, stirred for
30 min, and then treated sequentially with 55 g (260 mmol)
of K3PO4 and a solution of 4.9 g (26 mmol) of 189a in 50 mL
of CH2Cl2. After being stirred for 30 min, the mixture was
warmed to 0 °C for 20 min, quenched with water, extracted
with CH2Cl2, dried over Na2SO4, and concentrated in vacuo.
Silica gel chromatography using ethyl acetate in CH2Cl2
provided 3.5 g (79%) of methyl 2-isopropyl-2-oxazoline-4-
carboxylate (189b): 1H NMR (CDCl3) δ 1.22 (d, J ) 7 Hz, 6
H), 2.64 (heptet, J ) 7 Hz, 1 H), 3.79 (s, 3 H), 4.39 (dd, J )
10, 8 Hz, 1 H), 4.47 (t, J ) 8 Hz, 1 H), 4.72 (dd, J ) 10, 8 Hz,
1 H); CIMS m/z 172 (M + H)+. A solution of 2.53 g (14.8 mmol)
of 189b in 100 mL of anhydrous toluene was treated with 7.4
g of NiO2 and heated under Ar atmosphere to 60 °C for 30
min. After being allowed to cool, the mixture was filtered,
concentrated in vacuo, and purified by silica gel chromatog-
raphy using ethyl acetate in CH2Cl2 to provide 940 mg (38%)
of methyl 2-isopropyl-1,3-oxazole-4-carboxylate (189c): 1H
NMR (CDCl3) δ 1.38 (d, J ) 7 Hz, 6 H), 3.15 (heptet, J ) 7
Hz, 1 H), 3.91 (s, 3 H), 8.16 (s, 1 H); CIMS m/z 170 (M + H)+.
A solution of 1.20 g (7.1 mmol) of 189c in 100 mL of THF was
cooled to -25 °C and treated with 7.1 mL (7.1 mmol) of lithium
aluminum hydride in THF. After 10 min, the solution was
sequentially quenched with water (0.5 mL), 15% NaOH (0.5
mL), and water (1.07 mL) and filtered. The residue was
washed with ethyl acetate, and the combined filtrates were
concentrated in vacuo and purified by silica gel chromatogra-
phy using methanol in CH2Cl2 to give 0.74 g (74%) of 189: 1H
NMR (CDCl3) δ 1.85 (d, J ) 7 Hz, 6 H), 2.47 (t, J ) 6 Hz, 1
H), 3.08 (heptet, J ) 7 Hz, 1 H), 4.58 (d, J ) 6 Hz, 2 H), 7.49
(s, 1 H); CIMS m/z 142 (M + H)+.
P r oced u r e G. 2-Isop r op yl-4-[(N-m eth yla m in o)m eth yl]-
oxa zole (190). A mixture of isobutyramide (9.88 g, 112
mmol), 1,3-dichloroacetone (10.0 g, 75 mmol), NaHCO3 (9.43
g, 112 mmol), and MgSO4 (18.0 g, 150 mmol) in 130 mL of
acetone was heated at reflux under argon for 63 h. After being
allowed to cool to ambient temperature, the resulting mixture
was filtered and concentrated in vacuo to a dark brown
semisolid. The residue was purified by silica gel chromatog-
raphy using a gradient of EtOAc-CH2Cl2 (5%, 10%, 20%, 40%)
to obtain 6.06 g (46%) of 4-(chloromethyl)-4-hydroxy-2-isopro-
pyloxazoline (190a ) as an orange liquid: 1H NMR (CDCl3) δ
1.20-1.28 (m, 6H), 2.56-2.72 (m, 1H), 3.70 (s, 2H), 4.18 (d, J
) 9.6 Hz, 1H), 4.38 (d, J ) 9.6 Hz, 1H); CIMS m/z 178, 180 (M
+ H)+. A solution of 4.88 g (27.5 mmol) of 190a in 1,2-
dichloroethane (20 mL) was added to a solution of 2.40 mL
(32.9 mol) of SOCl2 in 1,2-dichloroethane (80 mL) at 0 °C under
argon. The resulting solution was heated to 70 °C for 15 min,
allowed to cool to ambient temperature, and concentrated in
vacuo to give crude 4-(chloromethyl)-2-isopropyloxazole (190b,
4.20 g, 0.0263 mol, 96%) as a brown semisolid: 1H NMR
(CDCl3) δ 1.36 (d, J ) 7.5 Hz, 6H), 3.03-3.18 (m, 1H), 4.50 (s,
2H), 7.56 (s, 1H); CIMS m/z 160, 162 (M + H)+. Compound
190b was treated according to procedure A above to provide
190 in 71% yield: 1H NMR (CDCl3) δ 1.33 (d, J ) 6.9 Hz, 6H),
2.46 (s, 3H), 2.99-3.14 (m, 1H), 3.64 (s, 2H), 7.42 (s, 1H); CIMS
m/z 155 (M + H)+, 172 (M + NH4)+.
of THF was cooled to 0 °C under N2 and treated dropwise with
3.24 mL (8.84 mmol) of triethyl phosphonoacetate. After 10
min, a solution of the above aldehyde in 25 mL of THF was
added, and the solution was allowed to stir for 25 min while
warming to ambient temperature. After quenching with 100
mL of saturated NH4Cl, the mixture was extracted with three
portions of EtOAc, dried over Na2SO4, and concentrated in
vacuo. Flash chromatography using 5-10% EtOAc in hexane
provided 1.61 g (81%) of 191: Rf 0.58 (20% EtOAc-hexane);
1H NMR (CDCl3) δ 1.33 (t, J ) 7 Hz, 3 H), 1.42 (d, J ) 7 Hz,
6 H), 3.32 (heptet, J ) 7 Hz, 1 H), 4.26 (q, J ) 7 Hz, 2 H), 6.75
(d, J ) 15 Hz, 1 H), 7.29 (s, 1 H), 7.58 (d, J ) 15 Hz, 1 H).
P r oced u r e I. 2-(4-Mor p h olin yl)-4-(h yd r oxym eth yl)-
th ia zole (192). A solution of 3.35 g (18.8 mmol) of thiocar-
bonyldiimidazole in 100 mL of THF was treated with 0.82 mL
(9.4 mmol) of morpholine. After being stirred at ambient
temperature for 3.5 h, an additional 0.82 mL portion of
morpholine was added, and stirring was continued. After 6
h, the solution was treated with excess concentrated aqueous
ammonia and stirred overnight. The resulting solution was
concentrated in vacuo, taken up in CHCl3, separated from the
aqueous phase, dried over Na2SO4, and concentrated. Puri-
fication of the residue by silica gel chromatography using
EtOAc provided 1.85 g (76%) of 4-[(amino)thiocarbonyl]mor-
pholine (Rf 0.17,10% CH3OH in CHCl3) as a white solid: 1H
NMR (CDCl3) δ 3.76 (m, 4 H), 3.83 (m, 4 H), 5.75 (br, 2 H);
CIMS m/z 147 (M + H)+. The above thiourea was converted
as described above for 188 to give 192 (Rf 0.16, 4% CH3OH in
CHCl3) in 34% overall yield: 1H NMR (CDCl3) δ 2.44 (br, 1
H), 3.46 (t, J ) 5 Hz, 4 H), 3.81 (t, J ) 5 Hz, 1 H), 4.55 (br s,
2 H), 6.45 (s, 1 H); CIMS m/z 200 (M + H)+.
P r oced u r e J . 2-(Hyd r oxym eth yl)-4-isop r op ylth ia zole
(193). A mixture of 2.11 g (12.8 mmol) of 1-bromo-3-methyl-
butan-2-one, 1.0 g (12.8 mmol) of ethyl thiooxamate, and 1.70
g (14 mmol) of MgSO4 in 50 mL of acetone was heated at reflux
for 3 h. After being allowed to cool, the mixture was filtered,
concentrated in vacuo, and purified by silica gel chromatog-
raphy using CHCl3 to provide 0.29 g (11%) of ethyl 4-isopropyl-
2-thiazolecarboxylate (Rf 0.9, 4% CH3OH in CHCl3): 1H NMR
(DMSO-d6) δ 1.27 (d, J ) 7 Hz, 6 H), 1.32 (t, J ) 7 Hz, 3 H),
3.12 (heptet, J ) 7 Hz, 1 H), 4.37 (q, J ) 7 Hz, 2 H), 7.73 (s,
1 H); CIMS m/z 200 (M + H)+. The above ester was reduced
as described for compound 188 to give a 96% yield of 193 (Rf
0.3, 5% CH3OH in CHCl3).
P r oced u r e K. 5-(Hyd r oxym eth yl)-2-isop r op ylth ia zole
(194). A mixture of 5.0 g (48 mmol) of thioisobutyramide, 7.3
g (48 mmol) of ethyl R-chloroformylacetate, and 5.8 g (48 mmol)
of MgSO4 in 200 mL of acetone was heated to reflux for 3 h.
After being allowed to cool, the mixture was filtered, concen-
trated in vacuo, and purified by flash chromatography using
9:1 hexane-EtOAc to give 8.0 g (83%) of ethyl 2-isopropyl-5-
thiazolecarboxylate. A solution of 4.0 g (20 mmol) of this ester
in 20 mL of anhydrous THF was added dropwise to 20 mL (20
mmol) of a cooled (0 °C) 1 M solution of LiAlH4 in THF. After
addition, the solution was stirred for 1 h, allowed to warm,
and quenched with solid Rochelle’s salt followed by ethyl
acetate until bubbling ceased. The mixture was filtered and
concentrated in vacuo. Flash chromatography using 3%
MeOH in CHCl3 provided 0.8 g (25%) of 194: 1H NMR (DMSO-
d6) δ 1.30 (d, J ) 7 Hz, 6 H), 3.22 (heptet, J ) 7 Hz, 1 H), 4.61
(dd, J ) 6, 1 Hz, 2 H), 5.45 (t, J ) 6 Hz, 1 H), 7.48 (t, J ) 1
Hz, 1 H); CIMS m/z 158 (M + H)+.
P r oced u r e L. 5-ter t-Bu t yl-3-(h yd r oxym et h yl)isox-
a zole (195). A solution of 5.0 g (25 mmol) of ethyl 5-tert-
butylisoxazole-3-carboxylate (Maybridge) in 150 mL of THF
was cooled to -30 °C and treated with 25 mL (25 mmol) of
lithium aluminum hydride in THF. After 30 min, the solution
was sequentially quenched with 10 mL of acetone, partitioned
between ethyl acetate and saturated Rochelle’s salt, dried over
Na2SO4, and concentrated in vacuo. Silica gel chromatography
using ethyl acetate in CH2Cl2 provided 3.65 g (93%) of 195:
1H NMR (CDCl3) δ 1.34 (s, 9 H), 2.05 (t, J ) 6 Hz, 1 H), 4.73
(d, J ) 6 Hz, 2 H), 6.00 (s, 1 H); CIMS m/z 156 (M + H)+.
P r oced u r e H . E t h yl 3-(2-Isop r op yl-4-t h ia zolyl)p r o-
p en oa te (191). A solution of 3.10 g (15.6 mmol) of 188a in
50 mL of dichloromethane was cooled to -78 °C and treated
dropwise over 1.5 h with 15.6 mL of diisobutylaluminum
hydride (1.5 M in toluene). After addition, the solution was
stirred for 0.5 h and quenched with 5 mL of methanol followed
by 15 mL of Rochelle’s salt. After being allowed to warm, the
mixture was filtered, and the filter cake was washed with
chloroform. The combined filtrate was washed with aqueous
Rochelle’s salt, dried over Na2SO4, and concentrated in vacuo
to give 1.37 g (56%) of crude 2-isopropyl-4-thiazolecarbox-
aldehyde. A slurry of 10 mmol of prewashed NaH in 25 mL

614 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4
Kempf et al.
P r oced u r e M. N-Meth yl-N-[(5-ter t-bu tyl-3-isoxa zolyl)-
m eth yl]a m in e (196). A solution of 1.0 g (6.4 mmol) of 195
and 1.25 mL (9.0 mmol) of triethylamine in 60 mL of CH2Cl2
was cooled to 0 °C and treated with 0.61 mL (7.7 mmol) of
methanesulfonyl chloride. After 10 min, the solvent was
removed in vacuo. The residue was purified by silica gel
chromatography to provide 1.49 g (99%) of 5-tert-butyl-3-
[(methanesulfonyloxy)methyl]isoxazole (196a): 1H NMR (CDCl3)
δ 1.36 (s, 9 H), 3.08 (s, 3 H), 5.26 (s, 2 H), 6.10 (s, 1 H); CIMS
m/z 251 (M + NH4)+. A solution of 754 mg (3.2 mmol) of 196a
in 7.5 mL of anhydrous methylamine was stirred in a high-
pressure flask at ambient temperature for 4.5 days. After
removal of the solvent in vacuo, the residue was purified by
silica gel chromatography using CH2Cl2-CH3OH-isopropyl-
amine to provide 490 mg (90%) of 196: 1H NMR (CDCl3) δ
1.34 (s, 9 H), 2.48 (s, 3 H), 3.78 (s, 2 H), 5.94 (s, 1 H); CIMS
m/z 169 (M + H)+.
P r oced u r e N. 3-Meth oxy-5-(h yd r oxym eth yl)isoxa zole
(197). To a solution of 2 g (13.9 mmol) of methyl 3-hydroxy-
5-isoxazolecarboxylate in 40 mL of benzene was added 4 g of
AgCO3 and 2.5 mL of CH3I. The reaction mixture was stirred
at ambient temperature in the dark for 16 h. The resulting
mixture was filtered through a pad of Celite, and the filtrate
was concentrated in vacuo. Silica gel chromatography using
1:1 ethyl acetate-hexane provided 1.61 g (74%) of methyl
3-methoxy-5-isoxazolecarboxylate (197a ): 1H NMR (CDCl3) δ
3.95 (s, 3 H), 4.03 (s, 3 H), 6.53 (s, 1 H). To a solution of 1.6
g (10.2 mmol) of 197a in 20 mL of THF at -60 °C was added
1.6 mL (1.6 mmol) of LiAlH4 in THF. After 15 min, the
reaction was carefully quenched with water, and extractive
workup followed by silica gel chromatography using 5% MeOH
in CHCl3 provided 0.85 g (65%) of 197: 1H NMR (CDCl3) δ
1.95 (t, J ) 6 Hz, 1 H), 3.96 (s, 2 H), 4.66 (d, J ) 6 Hz, 2 H),
5.88 (s, 1 H); CIMS m/z 130 (M + H)+.
P r oced u r e O. N-Cyclop r op yl-N-[(2-isop r op yl-4-t h i-
a zolyl)m eth yl]a m in e (198). A solution of 1.8 g (10.2 mmol)
of 187b in 10 mL of CHCl3 was added dropwise with stirring
to 10 mL of cyclopropylamine. The resulting solution was
stirred at ambient temperature for 16 h, concentrated in vacuo,
and purified by silica gel chromatography using 5% CH3OH
in CHCl3 to provide 0.39 g (19%) of 198: 1H NMR (DMSO-d6)
δ 0.24 (m, 2 H), 0.35 (m, 2H),1.30 (d, J ) 7 Hz, 6 H), 2.10 (tt,
J ) 12, 3 Hz, 1 H), 3.23 (heptet, J ) 7 Hz, 1 H), 3.77 (s, 2 H),
7.21 (s, 1 H); CIMS m/z 197 (M + H)+.
P r oced u r e P . 5-(H yd r oxyet h yl)-2-isop r op ylt h ia zole
(199). Ester 200 was reduced with LiAlH4 in the manner
described for 194 to give 199 in 47% yield: 1H NMR (CDCl3)
δ 1.40 (d, J ) 7 Hz, 6 H), 2.96 (t, J ) 6 Hz, 2 H), 3.30 (heptet,
J ) 7 Hz, 1 H), 3.93 (t, J ) 6 Hz, 2 H), 6.83 (s, 1 H); CIMS m/z
172 (M + H)+.
P r oced u r e Q. Eth yl (2-Isop r op yl-4-th ia zolyl)a ceta te
(200). A mixture of 1.5 g (14.5 mmol) of thioisobutyramide,
2.4 g (14.5 mmol) of ethyl R-chloroformyl acetate and 1.75 g
(14.5 mmol) of MgSO4 in 50 mL of acetone was heated to reflux
for 16 h. After being allowed to cool, the mixture was filtered,
concentrated in vacuo, and purified by flash chromatography
using 1% MeOH in CHCl3 to give 1.57 g (51%) of 200: 1H NMR
(DMSO-d6) δ 1.19 (t, J ) 7 Hz, 3 H), 1.30 (d, J ) 7 Hz, 6 H),
3.24 (heptet, J ) 7 Hz, 1 H), 3.75 (s, 2 H), 4.08 (q, J ) 7 Hz,
2 H), 7.30 (s, 1 H); CIMS m/z 214 (M + H)+.
P r oced u r e R . N-E t h yl-N-[(2-isop r op yl-4-t h ia zolyl)-
eth yl]a m in e (201). A solution of 0.20 g (1.2 mmol) of 199 in
5 mL of THF was cooled to 0 °C and treated with 0.09 mL (1.2
mmol) of methanesulfonyl chloride and 0.16 mL (1.2 mmol) of
triethylamine. After being stirred for 30 min, the solution was
treated with 10 mL of 40% aqueous ethylamine, heated to
reflux for 4 h, and then stirred at ambient temperature for 16
h. The resulting mixture was diluted with EtOAc, washed
sequentially with aqueous NaHCO3 and H2O, dried over
Na2SO4, and concentrated in vacuo to provide crude 201 which
was of sufficient purity for further reaction: 1H NMR (CDCl3)
δ 1.11 (t, J ) 7 Hz, 3 H), 1.38 (d, J ) 7 Hz, 6 H), 2.70 (q, J )
7 Hz, 2 H), 2.96 (m, 4 H), 3.30 (heptet, J ) 7 Hz, 1 H), 6.80 (s,
1 H); CIMS m/z 199 (M + H)+.
Representative procedures for the synthesis of N-substituted
amino esters. Preparation of the esters which are precursors
leading to compounds 5-12 has been reported.33 Other esters
were prepared by the following representative procedures.
Spectral characterization for the esters is provided as Sup-
porting Information.
P r oced u r e S.
N-[[N-Met h yl-N-[(2-isop r op yl-4-t h i-
a zolyl)m eth yl]a m in o]ca r bon yl]-L-va lin e Meth yl Ester
(202). A solution of 66 g (0.328 mol) of 4-nitrophenyl chloro-
formate in 1.2 L of CH2Cl2 was cooled to 0 °C and treated with
L-valine methyl ester hydrochloride. The resulting mixture
was treated slowly, with stirring, with 69 mL (0.63 mol) of
4-methylmorpholine. The resulting solution was allowed to
slowly warm to ambient temperature and was stirred over-
night. After washing with three portions of 10% aqueous
NaHCO3, the solution was dried over Na2SO4 and concentrated
in vacuo. The residue was purified by silica gel chromatog-
raphy by eluting with CHCl3 to provide N-[[(4-nitrophenyl)-
oxy]carbonyl]-L-valine methyl ester: 1H NMR (DMSO-d6) δ
0.94 (d, J ) 7 Hz, 3 H), 0.95 (d, J ) 7 Hz, 3 H), 2.12 (octet, J
) 7 Hz, 1 H), 3.69 (s, 3 H), 4.01 (dd, J ) 8, 6 Hz, 1 H), 7.41
(dt, J ) 9, 3 Hz, 2 H), 8.27 (dt, J ) 9, 3 Hz, 2 H), 8.53 (d, J )
8 Hz, 1 H); CIMS m/z 314 (M + NH4)+. A solution of 15.7 g
(92 mmol) of 187 in 200 mL of THF was combined with a
solution of 20.5 g (69 mmol) of the above p-nitrophenyl
carbamate. The resulting solution was treated with 1.6 g of
DMAP and 12.9 mL (92 mmol) of triethylamine, heated at
reflux for 2 h, allowed to cool, and concentrated in vacuo. The
residue was taken up in CH2Cl2, washed extensively with 5%
aqueous K2CO3, dried over Na2SO4, and concentrated in vacuo.
Purification by silica gel chromatography using CHCl3 as an
eluent provided 16.3 g (54%) of 202: Rf 0.26 (5% CH3OH/
CHCl3); 1H NMR (DMSO-d6) δ 0.88 (d, J ) 7 Hz, 3 H), 0.92 (d,
J ) 7 Hz, 3 H), 1.32 (d, J ) 7 Hz, 6 H), 2.05 (octet, J ) 7 Hz,
1 H), 2.86 (s, 3 H), 3.25 (heptet, J ) 7 Hz, 1 H), 3.61 (s, 3 H),
3.96 (dd, J ) 8, 7 Hz, 1 H), 4.44 (AA′, 2 H), 6.58 (d, J ) 8 Hz,
1 H), 7.24 (s, 1 H); CIMS m/z 328 (M + H)+.
P r oced u r e T. N-[[(2-Isop r op yl-4-th ia zolyl)m eth oxy]-
ca r bon yl]-L-a la n in e Meth yl Ester (203). A solution of 1.12
g (5.56 mmol) of 4-nitrophenyl chloroformate in 20 mL of
CH2Cl2 was cooled to 0 °C and treated sequentially with 0.8 g
(5.1 mmol) of 188 and 0.6 mL (5.6 mmol) of 4-methylmorpho-
line. The resulting solution was stirred at 0 °C for 1 h, diluted
with CH2Cl2, washed with three portions of aqueous NaHCO3,
dried over Na2SO4, and concentrated in vacuo to give crude
(4-(hydroxymethyl)-2-isopropylthiazolyl)methyl 4-nitrophenyl
carbonate. A portion (0.53 g, 1.65 mmol) of the above carbon-
ate was taken up in 20 mL of CHCl3, treated with 0.23 g (1.67
mmol) of L-alanine methyl ester hydrochloride and 0.36 mL
(3.3 mmol) of 4-methylmorpholine, and heated at reflux for
16 h. After being allowed to cool, the solvent was removed in
vacuo, and the residue was purified by silica gel chromatog-
raphy using 2% CH3OH in CHCl3 to provide 0.45 g (94%) of
1
203: Rf 0.43 (5% CH3OH in CH2Cl2); H NMR (DMSO-d6) δ
1.26 (d, J ) 8 Hz, 3 H), 1.32 (d, J ) 7 Hz, 6 H), 3.27 (heptet,
J ) 7 Hz, 1 H), 3.63 (s, 3 H), 4.10 (pentet, J ) 8 Hz, 1 H), 5.02
(s, 2 H), 7.47 (s, 1 H), 7.81 (d, J ) 8 Hz, 1 H); CIMS m/z 287
(M + H)+.
P r oced u r e U. N-[[(2-Isop r op yl-5-th ia zolyl)m eth oxy-
ca r bon yl]-L-va lin e Meth yl Ester (204). A solution of 0.40
g (2.75 mmol) of 194, 2.9 mmol of R-N-carbonylvaline methyl
ester24 and 0.28 mmol of DMAP in 15 mL of CH2Cl2 was heated
at reflux for 5 h. The resulting solution was washed succes-
sively with 10% citric acid, aqueous NaHCO3, and brine, dried
over Na2SO4, and concentrated in vacuo. Silica gel chroma-
tography of the residue using 5% EtOAc in CHCl3 provided
0.25 g (29%) of 204 (Rf 0.8, 5% CH3OH in CHCl3): 1H NMR δ
0.89 (d, J ) 7 Hz, 6 H), 0.95 (d, J ) 7 Hz, 3 H), 0.97 (d, J ) 7
Hz, 3 H), 2.14 (m, 1 H), 3.33 (heptet, J ) 7 Hz, 1 H), 3.74 (s,
3 H), 4.30 (dd, J ) 9, 5 Hz, 1 H), 5.23 (s, 2 H), 5.25 (br d, 1 H),
7.63 (s, 1 H); CIMS m/z 315 (M + H)+.
P r oced u r e V. N-[[N-Eth yl-N-[(2-isop r op yl-4-th ia zolyl)-
m eth yl]a m in o]ca r bon yl]glycin e Eth yl Ester (205).
A
solution of 1.47 g (8 mmol) of N-ethyl-N-[(2-isopropyl-4-

Discovery of Ritonavir
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4 615
thiazolyl)methyl]amine (prepared according to procedure D)
and 0.9 mL (8 mmol) of ethyl isocyanatoacetate in 30 mL of
THF was heated at reflux for 16 h. After being allowed to
cool, the solution was concentrated in vacuo, taken up in
CH2Cl2, washed with two portions of 10% aqueous NaHCO3
and one portion of saturated brine, dried over Na2SO4, and
concentrated. Silica gel chromatography using 0.5% methanol
in CHCl3 provided 0.95 g (38%) of 205: Rf 0.70 (10% CH3OH/
CHCl3).
with LiOH and coupled to the above amine using EDC/HOBT
as described in procedure W to provide 84 in 46% yield.
Biologica l Eva lu a tion . Procedures for measuring the
inhibition of recombinant HIV protease, the inhibition of HIV-
1IIIB in MT4 cells, the aqueous solubility in pH 4.0 and 7.4
buffers, and the pharmacokinetic profile of the inhibitors in
rats after oral and iv dosing have been described previ-
ously.19,24,25,33 For pharmacokinetic analysis, compounds were
formulated in 5% dextrose containing 20% ethanol, 30%
propylene glycol, and 2 equiv of methanesulfonic acid.
Rep r esen ta tive P r oced u r es for Hyd r olysis a n d Cou -
p lin g of Su b st it u t ed Am in o Acid s. P r oced u r e W.
(2S,3S,5S)-5-[N-[N-[[N-Meth yl-N-[(2-isopr opyl-4-th iazolyl)-
m e t h yl]a m in o]ca r b on yl]va lin yl]a m in o]-2-[N -[[(5-t h i-
a zolyl)m e t h oxy]ca r b on yl]a m in o]-1,6-d ip h e n yl-3-h y-
d r oxyh exa n e (ABT-538, Com p ou n d 123). A solution of
1.42 g (4.3 mmol) of 202 in 17 mL of dioxane was treated with
17.3 mL of 0.50 M aqueous LiOH. The resulting solution was
stirred at ambient temperature for 30 min, treated with 8.7
mL of 1 M HCl, and concentrated in vacuo. The residue was
taken up in CH2Cl2, washed with water, dried over Na2SO4,
and concentrated in vacuo to provide 1.1 g (81%) of crude
N-[[N-methyl-N-[(2-isopropyl-4-thiazolyl)methyl]amino]car-
bonyl]-L-valine: CIMS m/z 314 (M + H)+. A solution of 70
mg (0.223 mmol) of the above acid, 79 mg (0.186 mmol) of 6d ,
30 mg (0.223 mmol) of 1-hydroxybenzotriazole hydrate (HOBT),
and 51 mg (0.266 mmol) of N-ethyl-N′-[(dimethylamino)propyl]-
carbodiimide (EDC) in 2 mL of THF was stirred at ambient
temperature for 16 h. The resulting solution was concentrated
in vacuo, and the residue was purified by silica gel chroma-
tography using 97:3 CH2Cl2-CH3OH to provide 100 mg (74%)
of 123 (Rf 0.4, 95:5 CH2Cl2-CH3OH) as a solid: mp 61-63
°C; CIMS m/z 721 (M + H)+.
P r oced u r e X. (2S,3S,5S)-5-[N-[N-[[(6-Am in o-3-p yr id i-
n yl)m eth oxy]car bon yl]valin yl]am in o]-2-[N-[[(3-pyr idin yl)-
m et h oxy]ca r b on yl]a m in o]-1,6-d ip h en yl-3-h yd r oxyh ex-
a n e (53). N-[[[5-[(tert-Butyloxycarbonyl)amino]-3-pyridinyl]-
methoxy]carbonyl]-L-valine (115 mg, 0.31 mmol) was coupled
to 6d (105 mg, 0.25 mmol) according to procedure W above to
give, following flash chromatography using 5-7% MeOH/
CHCl3, 175 mg (91%) of the product amide. A portion of this
amide (148 mg, 0.19 mmol) was taken up in 4 mL of 1:1
trifluoroacetic acid/CH2Cl2 and stirred at ambient temperature
for 1 h. The solvent was removed in vacuo, and the residue
was partitioned between EtOAc and saturated aqueous NaH-
CO3, dried over Na2SO4, and concentrated. Flash chromatog-
raphy using MeOH/CHCl3 provided 85 mg (66%) of 53. CIMS
m/z 669 (M + H)+.
Ack n ow led gm en t. Details of the X-ray crystal
structure of ritonavir bound to HIV protease were
provided by Chang Park, Xiang-Peng Kong, and Kent
Stewart. The assistance of the Structural Chemistry
Division, Abbott Laboratories, in providing NMR and
mass spectra is gratefully acknowledged.
Su p p or tin g In for m a tion Ava ila ble: Tabular compilation
of the physical and analytical data for protease inhibitors and
spectral characterization of intermediate substituted amino-
acyl esters (20 pages). Ordering information is given on any
current masthead page.
Refer en ces
(1) Debouck, C.; Gorniak, J . G.; Strickler, J . E.; Meek, T. D.; Metcalf,
B. W.; Rosenberg, M. Human Immunodeficiency Virus Protease
Expressed in Escherichia coli Exhibits Autoprocessing and
Specific Maturation of the gag Precursor. Proc. Natl. Acad. Sci.
U.S.A. 1987, 84, 8903-8906.
(2) Graves, M. C.; Lim, J . J .; Heimer, E. P.; Kramer, R. A. An 11-
kDa Form of Human Immunodeficiency Virus Protease Ex-
pressed in Escherichia coli is Sufficient for Enzymic Activity.
Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 2449-2453.
(3) Darke, P. L.; Nutt, R. F.; Brady, S. F.; Garsky, V. M.; Ciccarone,
T. M.; Leu, C.-T.; Lumma, P. K.; Freidinger, R. M.; Veber, D.
F.; Sigal, I. S. HIV-1 Protease Specificity of Peptide Cleavage is
Sufficient for Processing of gag and pol Polyproteins. Biochem.
Biophys. Res. Commun. 1988, 156, 297.
(4) Gottlinger, H. G.; Sodroski, J . G.; Haseltine, W. A. Role of Capsid
Precursor Processing and Myristoylation in Morphogenesis and
Infectivity of Human Immunodeficiency Virus Type 1. Proc. Natl.
Acad. Sci. U.S.A. 1989, 86, 5781-5785.
(5) Kohl, N. E.; Emini, E. A.; Schleif, W. A.; Davis, L. J .; Heimbach,
J . C.; Dixon, R. A. F.; Scolnick, E. M.; Sigal, I. S. Active Human
Immunodeficiency Virus Protease is Required for Viral Infectiv-
ity. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 4686-4690.
(6) Le Grice, S. F.; Mills, J .; Mous, J . Active Site Mutagenesis of
the AIDS Virus Protease and its Alleviation by trans Comple-
mentation. EMBO J . 1988, 7, 2547-2553.
(7) Kempf, D.; Sham, H. HIV Protease Inhibitors. Curr. Pharm. Des.
1996, 2, 225-246.
(8) West, M. L.; Fairlie, D. P. Targeting HIV-1 Proteasesa Test of
Drug Design Methodologies. Trends Pharm. Sci. 1995, 16, 67-
75.
(9) Thaisrivongs, S. HIV Protease Inhibitors. Annu. Rep. Med.
Chem. 1994, 29, 133-144.
(10) Darke, P. L.; Huff, J . R. HIV Protease as an Inhibitor Target
for the Treatment of AIDS. Adv. Pharmacol. 1994, 25, 399-454.
(11) Pearl, L. H.; Taylor, W. R. Sequence Specificity of Retroviral
Proteases. Nature 1987, 328, 482.
(12) Navia, M. A.; Fitzgerald, P. M. D.; McKeever, B. M.; Leu, C.-T.;
Heimbach, J . C.; Herber, W. K.; Sigal, I. S.; Darke, P. L.;
Springer, J . P. Three-Dimensional Structure of Aspartyl Pro-
tease from Human Immunodeficiency Virus HIV-1. Nature 1989,
337, 615-620.
P r oced u r e Y. (2S,3S,5S)-5-[N-[N-[[N-Met h yl-N-[(2-
m e t h yl-4-t h ia zolyl)m e t h yl]a m in o]ca r b on yl]va lin yl]-
a m in o]-2-[N-[[(3-p yr id in yl)m et h oxy]ca r b on yl]a m in o]-
1,6-d ip h en yl-3-h yd r oxyh exa n e (69). A solution of N-[[N-
methyl-N-[(2-methyl-4-thiazolyl)methyl]amino]carbonyl]-
valine (66 mg, 0.23 mmol) and 39 mg (0.28 mmol) of p-
nitrophenol in 2 mL of THF was treated with 58 mg (0.65
mmol) of dicyclohexylcarbodiimide and stirred at ambient
temperature. After 16 h, the mixture was filtered, treated with
65 mg (0.16 mmol) of 6a , and stirred under N2 at ambient
temperature. After 16 h, the solution was concentrated in
vacuo and purified by silica gel chromatography using 3%
CH3OH in CHCl3 to provide 104 mg (95%) of 69.
P r oced u r e Z. (2S,3S,5S)-2-[N-[N-[3-(2-Isop r op yl-4-
t h ia zolyl)p r op a n oyl]va lin yl]a m in o]-5-[N-[[(3-p yr id yl)-
m et h oxy]ca r b on yl]a m in o]-1,6-d ip h en yl-3-h yd r oxyh ex-
a n e (84). Boc-L-valine was coupled to 5a as described above,
and the Boc group was removed with HCl/dioxane followed
by neutralization to the free amine. A solution of 225 mg (1
mmol) of 191 in 10 mL of freshly distilled CH3OH and 1 mL
of dry THF was treated with 50 mg of Mg metal. After being
stirred for 20 min, the mixture was poured over cold 2 N HCl,
neutralized to pH 8, extracted with EtOAc, dried over Na2SO4,
and concentrated in vacuo. Flash chromatography using 10%
EtOAc in hexane provided 105 mg (50%) of ethyl 3-(2-
isopropyl-4-thiazolyl)propanoate. This ester was hydrolyzed
(13) Wlodawer, A.; Miller, M.; J askolski, M.; Sathyanarayana, B. K.;
Baldwin, E.; Weber, I. T.; Selk, L. M.; Clawson, L.; Schneider,
J .; Kent, S. B. H. Conserved Folding in Retroviral Proteases:
Crystal Structure of a Synthetic HIV-1 Protease. Science 1989,
245, 616-621.
(14) Holloway, M. K.; Wai, J . M. Structure-Based Design of Human
Immunodeficiency Virus-1 Protease InhibitorssCorrelating Cal-
culated Energy With Activity. ACS Symp. Ser. 1995, 589, 36-
50.
(15) Erickson, J .; Kempf, D. J . Structure-Based Design of Symmetric
Inhibitors of HIV-1 Protease. Arch. Virol. 1994, 9, 19-29.
(16) Clare, M. HIV Protease: Structure-Based Design. Perspect. Drug
Discovery Des. 1993, 1, 49-68.
(17) Wlodawer, A.; Erickson, J . W. Structure-Based Inhibitors of
HIV-1 Protease. Annu. Rev. Biochem. 1993, 62, 543-585.

616 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4
Kempf et al.
(18) Plattner, J . J .; Norbeck, D. W. Obstacles to Drug Development
from Peptide Leads. In Drug Discovery Technologies; Clark, R.,
Moos, W. H., Eds.; Ellis Horwood Ltd.: Chichester, 1990; pp 92-
126.
(19) Kempf, D. J .; Marsh, K. C.; Denissen, J . F.; McDonald, E.;
Vasavanonda, S.; Flentge, C. A.; Green, B. E.; Fino, L.; Park, C.
H.; Kong, X.-P.; Wideburg, N. E.; Saldivar, A.; Ruiz, L.; Kati,
W. M.; Sham, H. L.; Robins, T.; Stewart, K. D.; Hsu, A.; Plattner,
J . J .; Leonard, J . M.; Norbeck, D. W. ABT-538 is a Potent
Inhibitor of Human Immunodeficiency Virus Protease and has
High Oral Bioavailability in Humans. Proc. Natl. Acad. Sci.
U.S.A. 1995, 92, 2484-2488.
(20) Vacca, J . P.; Dorsey, B. D.; Schleif, W. A.; Levin, R. B.; McDaniel,
S. L.; Darke, P. L.; Zugay, J .; Quintero, J . C.; Blahy, O. M.; Roth,
E.; Sardana, V. V.; Schlabach, A. J .; Graham, P. I.; Condra, J .
H.; Gotlib, L.; Holloway, M. K.; Lin, J .; Chen, I.-W.; Vastag, K.;
Ostovic, D.; Anderson, P. S.; Emini, E. A.; Huff, J . R.
L-735,524: An Orally Bioavailable Human Immunodeficiency
Virus Type 1 Protease Inhibitor. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 4096-4100.
(21) Kalish, V.; Kaldor, S.; Shetty, B.; Tatlock, J .; Davies, J .;
Hammond, M.; Dressman, B.; Fritz, J .; Appelt, K.; Reich, S.;
Musick, L.; Wu, B.-W.; Su, K. Iterative Protein Structure-Based
Drug Design and Synthesis of HIV Protease Inhibitors. Eur. J .
Med. Chem. 1995, 30, S 201-S 214.
(22) Kim, E. E.; Baker, C. T.; Dwyer, M. D.; Murcko, M. A.; Rao, B.
G.; Tung, R. D.; Navia, M. A. Crystal Structure of HIV-1 Protease
in Complex With VX-478, a Potent and Orally Bioavailable
Inhibitor of the Enzyme. J . Am. Chem. Soc. 1995, 117, 1181-
1182.
(23) Kempf, D. J .; Norbeck, D. W.; Codacovi, L.; Wang, X. C.;
Kohlbrenner, W. E.; Wideburg, N. E.; Paul, D. A.; Knigge, M.
F.; Vasavanonda, S.; Craig-Kennard, A.; Saldivar, A.; Rosen-
brook, W., J r.; Clement, J . J .; Plattner, J . J .; Erickson, J .
Structure-Based, C2 Symmetric Inhibitors of HIV Protease. J .
Med. Chem. 1990, 33, 2687-2689.
(24) Kempf, D. J .; Codacovi, L.; Wang, X. C.; Kohlbrenner, W. E.;
Wideburg, N. E.; Saldivar, A.; Vasavanonda, S.; Marsh, K. C.;
Bryant, P.; Sham, H. L.; Green, B. E.; Betebenner, D. A.;
Erickson, J .; Norbeck, D. W. Symmetry-Based Inhibitors of HIV
Protease. Structure-Activity Studies of Acylated 2,4-Diamino-
1,5-diphenyl-3-hydroxypentane and 2,5-Diamino-1,6-diphenyl-
hexane-3,4-diol. J . Med. Chem. 1993, 36, 320-330.
(25) Kempf, D. J .; Marsh, K. C.; Paul, D. A.; Knigge, M. F.; Norbeck,
D. W.; Kohlbrenner, W. E.; Codacovi, L.; Vasavanonda, S.;
Bryant, P.; Wang, X. C.; Wideburg, N. E.; Clement, J . J .;
Plattner, J . J .; Erickson, J . Antiviral and Pharmacokinetic
Properties of C2-Symmetric Inhibitors of the Human Immuno-
deficiency Virus Type 1 Protease. Antimicrob. Agents Chemother.
1991, 35, 2209-2214.
(26) Reedijk, M.; Boucher, C. A. B.; Van Bommel, T.; Ho, D. D.; Tzeng,
T. B.; Sereni, D.; Veyssier, P.; J urriaans, S.; Granneman, R.;
Hsu, A.; Leonard, J . M.; Danner, S. A. Safety, Pharmacokinetics,
and Antiviral Activity of A77003, a C2 Symmetry-Based Human
Immunodeficiency Virus Protease Inhibitor. Antimicrob. Agents
Chemother. 1995, 39, 1559-1564.
(27) Kempf, D. J . Design of Symmetry-Based, Peptidomimetic Inhibi-
tors of HIV Protease. Methods Enzymol. 1994, 241, 334-354.
(28) Ho, D. D.; Neumann, A. U.; Perelson, A. S.; Chen, W.; Leonard,
J . M.; Markowitz, M. Rapid Turnover of Plasma Virions and CD4
Lymphocytes In HIV-1 Infection. Nature 1995, 373, 123-126.
(29) Wei, X.; Ghosh, S. K.; Taylor, M. E.; J ohnson, V. A.; Emini, E.
A.; Deutsch, P.; Lifson, J . D.; Bonhoeffer, S.; Nowak, M. A.;
Hahn, B. H.; Saag, M. S.; Shaw, G. M. Viral Dynamics in Human
Immunodeficiency Virus Type 1 Infection. Nature 1995, 373,
117-122.
(30) Danner, S. A.; Carr, A.; Leonard, J . M.; Lehman, L. M.; Gudiol,
F.; Gonzales, J .; Raventos, A.; Rubio, R.; Bouza, E.; Pintado, V.;
Aguado, A. G.; De Lomas, J . G.; Delgado, R.; Borleffs, J . C. C.;
Hsu, A.; Valdes, J . M.; Boucher, C. A. B.; Cooper, D. A.; Gimeno,
C.; Clotet, B.; Tor, J .; Ferrer, E.; Martinez, P. L.; Moreno, S.;
Zancada, G.; Alcami, J .; Noriega, A. R.; Pulido, F.; Glassman,
H. N. A Short-Term Study of the Safety, Pharmacokinetics, and
Efficacy of Ritonavir, an Inhibitor of HIV-1 Protease. N. Engl.
J . Med. 1995, 333, 1528-1533.
(31) Markowitz, M.; Saag, M.; Powderly, W. G.; Hurley, A. M.; Hsu,
A.; Valdes, J . M.; Henry, D.; Sattler, F.; La Marca, A.; Leonard,
J . M.; Ho, D. D. A Preliminary Study of Ritonavir, an Inhibitor
of HIV-1 Protease, to Treat HIV-1 Infection. N. Engl. J . Med.
1995, 333, 1534-1539.
(33) Kempf, D. J .; Marsh, K. C.; Fino, L. C.; Bryant, P.; Craig-
Kennard, A.; Sham, H. L.; Zhao, C.; Vasavanonda, S.; Kohlbren-
ner, W. E.; Wideburg, N. E.; Saldivar, A.; Green, B. E.; Herrin,
T.; Norbeck, D. W. Design of Orally Bioavailable, Symmetry-
Based Inhibitors of HIV Protease. Bioorg. Med. Chem. 1994, 2,
847-858.
(34) Ghosh, A. K.; McKee, S. P.; Thompson, W. J .; Darke, P. L.; Zugay,
J . C. Potent HIV-1 Protease Inhibitors: Stereoselective Synthe-
sis of a Dipeptide Mimic. J . Org. Chem. 1993, 58, 1025-1029.
(35) Stuk, T. L.; Haight, A. R.; Scarpetti, D.; Allen, M. S.; Menzia, J .
A.; Robbins, T. A.; Parekh, S. I.; Langridge, D. C.; Tien, J .-H.
J .; Pariza, R. J .; Kerdesky, F. A. J . An Efficient Stereocontrolled
Strategy for the Synthesis of Hydroxyethylene Dipeptide
Isosteres. J . Org. Chem. 1994, 59, 4040-4041.
(36) Matayoshi, E. D.; Wang, G. T.; Krafft, G. A.; Erickson, J . Novel
Fluorogenic Substrates for Assaying Retroviral Proteases by
Resonance Energy Transfer. Science 1990, 247, 954-958.
(37) Pauwels, R.; Balzarini, J .; Baba, M.; Snoeck, R.; Schols, D.;
Herdewijn, P.; Desmyter, J .; De Clercq, E. Rapid and Automated
Tetrazolium-Based Colorimetric Assay for the Detection of Anti-
HIV Compounds. J . Virol. Methods 1988, 20, 309-321.
(38) Roberts, N. A.; Martin, J . A.; Kinchington, D.; Broadhurst, A.
V.; Craig, J . C.; Duncan, I. B.; Galpin, S. A.; Handa, B. K.; Kay,
J .; Kro¨hn, A.; Lambert, R. W.; Merrett, J . H.; Mills, J . S.; Parkes,
K. E. B.; Redshaw, S.; Ritchie, A. J .; Taylor, D. L.; Thomas, G.
J .; Machin, P. J . Rational Design of Peptide-Based HIV Pro-
teinase Inhibitors. Science 1990, 248, 358-361.
(39) Molla, A.; Korneyeva, M.; Gao, Q.; Vasavanonda, S.; Schipper,
P. J .; Mo, H.-M.; Markowitz, M.; Chernyavskiy, T.; Niu, P.;
Lyons, N.; Hsu, A.; Granneman, G. R.; Ho, D. D.; Boucher, C.
A. B.; Leonard, J . M.; Norbeck, D. W.; Kempf, D. J . Ordered
Accumulation of Mutations in HIV Protease Confers Resistance
to Ritonavir. Nature Med. 1996, 2, 760-766.
(40) Kempf, D. J . Progress in the Discovery of Orally Bioavailable
Inhibitors of HIV Protease. Perspect. Drug Discovery Des. 1995,
2, 427-436.
(41) Denissen, J .; Marsh, K.; Grabowski, B.; J ohnson, M. Metabolism
and Disposition of the HIV Protease Inhibitor A-77003. Pharm.
Res. 1993, 10, S-376.
(42) Kaplan, A.; Michael, S.; Wehbie, R.; Knigge, M.; Paul, D.; Everitt,
L.; Kempf, D.; Norbeck, D.; Erickson, J .; Swanstrom, R. Selection
of Multiple Human Immunodeficiency Virus Type 1 Variants
which Encode Viral Proteases with Decreased Sensitivity to an
Inhibitor of the Viral Protease. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 5597-5601.
(43) Hosur, M. V.; Bhat, T. N.; Kempf, D. J .; Baldwin, E. T.; Liu, B.;
Gulnik, S.; Wideburg, N. E.; Norbeck, D. W.; Appelt, K.;
Erickson, J . W. Influence of Stereochemistry on Activity and
Binding Modes for C2 Symmetry-Based Inhibitors of HIV-1
Protease. J . Am. Chem. Soc. 1994, 116, 847-855.
(44) Kumar, G. N.; Grabowski, B.; Lee, R.; Denissen, J . F. Hepatic
Drug-Metabolizing Activities in Rats after 14 Days of Oral
Administration of the Human Immunodeficiency Virus-Type 1
Protease Inhibitor Ritonavir (ABT-538). Drug Metab. Dispos.
1996, 24, 615-617.
(45) Kempf, D. J .; Marsh, K. C.; Kumar, G.; Rodrigues, A. D.;
Denissen, J . F.; McDonald, E.; Kukulka, M. J .; Hsu, A.; Gran-
neman, G. R.; Baroldi, P. A.; Sun, E.; Pizzuti, D.; Plattner, J . J .;
Norbeck, D. W.; Leonard, J . M. Pharmacokinetic Enhancement
of Inhibitors of the Human Immunodeficiency Virus Protease
by Coadministration with Ritonavir. Antimicrob. Agents Chemo-
ther. 1997, 41, 654-660.
(46) Hungate, R. W.; Chen, J . L.; Starbuck, K. E.; Vacca, J . P.;
McDaniel, S. L.; Levin, R. B.; Dorsey, B. D.; Guare, J . P.;
Holloway, M. K.; Whitter, W.; Darke, P. L.; Zugay, J . A.; Schleif,
W. A.; Emini, E. A.; Quintero, J . C.; Lin, J . H.; Chen, I.-W.;
Anderson, P. S.; Huff, J . R. Synthesis, Antiviral Activity and
Bioavailability Studies of γ-Lactam Derived HIV Protease
Inhibitors. Bioorg. Med. Chem. 1994, 2, 859-879.
(47) Dorsey, B. D.; Levin, R. B.; McDaniel, S. L.; Vacca, J . P.; Guare,
J . P.; Darke, P. L.; Zugay, J . A.; Emini, E. A.; Schleif, W. A.;
Quintero, J . C.; Lin, J . H.; Chen, I.-W.; Holloway, M. K.;
Fitzgerald, P. M. D.; Axel, M. G.; Ostovic, D.; Anderson, P. S.;
Huff, J . R. L-735,524: The Design of a Potent and Orally
Bioavailable HIV Protease Inhibitor. J . Med. Chem. 1994, 37,
3443-3451.
(48) Kempf, D. J .; Molla, A.; Marsh, K. C.; Park, C.; Rodrigues, A.
D.; Korneyeva, M.; Vasavanonda, S.; McDonald, E.; Flentge, C.
A.; Muchmore, S.; Wideburg, N. E.; Saldivar, A.; Cooper, A.; Kati,
W. M.; Stewart, K. D.; Norbeck, D. W. Lack of Stereospecificity
in the Binding of the P2 Amino Acid of Ritonavir to HIV
Protease. Bioorg. Med. Chem. Lett 1997, 7, 699-704.
(49) Kageyama, S.; Anderson, B. D.; Hoesterey, B. L.; Hayashi, H.;
Kiso, Y.; Flora, K. P.; Mitsuya, H. Protein Binding of Human
Immunodeficiency Virus Protease Inhibitor KNI-272 and Alter-
ation of its In Vitro Antiretroviral Activity in the Presence of
High Concentrations of Proteins. Antimicrob. Agents Chemother.
1994, 38, 1107-1111.
(32) Heath-Chiozzi, M.; Leonard, J .; Sun, E.; Lehman, L.; Henry, D.;
Kravcik, S.; Mills, R.; Potthoff, A. The Advanced HIV Ritonavir
Study Group; Cameron, D. W. Ritonavir Clinical Benefit Cor-
relates with HIV RNA and CD4 Cell Levels in Advanced HIV
Illness. XI International Conference on AIDS, Vancouver, Canada,
J une 1996, Abstract We.B.3127.

Discovery of Ritonavir
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 4 617
(50) Bilello, J . A.; Bilello, P. A.; Prichard, M.; Robins, T.; Drusano,
G. L. Reduction of the In Vitro Activity of A77003, an Inhibitor
of Human Immunodeficiency Virus Protease, by Human Serum
a1 Acid Glycoprotein. J . Infect. Dis. 1995, 171, 546-551.
(51) Markowitz, M.; Cao, Y.; Vesanen, M.; Talal, A.; Nixon, D.;
Hurley, A.; O’Donovan, R.; Racz, P.; Tenner-Racz, K.; Ho, D. D.
Recent HIV Infection Treated with AZT, 3TC, and a Potent
Protease Inhibitor. 4th Conference on Retroviruses and Op-
portunistic Infections, Washington, D.C., J an 22-26, 1997,
Abstract LB8.
(52) Cohen, C.; Sun, E.; Cameron, W.; McMahon, D.; Farthing, C.;
Poretz, D.; Markowitz, M.; Follansbee, S.; Mellors, J .; Ho, D.;
Hsu, A.; Maki, R.; Salgo, M.; Leonard, J . Ritonavir-Saquinavir
Combination Treatment in HIV-Infected Patients. 36th Inter-
science Conference on Antimicrobial Agents and Chemotherapy,
New Orleans, LA, Sept 15-18, 1996, Abstract LB7b.
(53) Mashraqui, S. H.; Keehn, P. M. Cyclophanes. 14. Synthesis,
Structural Assignment, and Conformational Properties of
[2.2](2,5)Oxazolo and Thazolophanes. J . Am. Chem. Soc. 1982,
104, 4461-4465.
(54) Scho¨llkopf, U.; Schro¨de, R. 2-Unsubstituted Oxazoles from
a-Metalated Isocyanides and Acylating Agents. Angew. Chem.,
Int. Ed. Engl. 1971, 10, 333.
(55) Kanemasa, S.; Asai, Y.; Tanaka, J . Synthesis of Substituted
Pyridines by a Cycloaddition Route Using Nitrile Oxides and
Homoallyl Alcohols. Bull. Chem. Soc. J pn. 1991, 64, 375-380.
(56) Yokokawa, F.; Hamada, Y.; Shioiri, T. Synthesis of 2,4-Disub-
stituted Oxazoles. Synlett 1992, 153.
J M970636+