Cycloalkylpyranones and cycloalkyldihydropyrones as HIV protease inhibitors: Exploring the impact of ring size on structure-activity relationships

Article abstract of DOI:10.1021/jm960296c

Previously, 3-substituted cycloalkylpyranones, such as 2d, have proven to be effective inhibitors of HIV protease. In an initial series of 3-(1- phenylpropyl) derivatives with various cycloalkyl ring sizes, the cyclooctyl analog was the most potent. We be

Full text of DOI:10.1021/jm960296c

J . Med. Chem. 1996, 39, 4125-4130
4125
Cycloa lk ylp yr a n on es a n d Cycloa lk yld ih yd r op yr on es a s HIV P r otea se
In h ibitor s: Exp lor in g th e Im p a ct of Rin g Size on Str u ctu r e-Activity
Rela tion sh ip s
Karen R. Romines,*,† J eanette K. Morris,† W. J effrey Howe,‡ Paul K. Tomich,§ Miao-Miao Horng,§
Kong-Teck Chong, Roger R. Hinshaw, David J . Anderson,| J oseph W. Strohbach,† Steve R. Turner,† and
Steve A. Mizsak†
Structural, Analytical and Medicinal Chemistry Research, Computer-Aided Drug Discovery, Chemical and Biological
Screening, Cancer and Infectious Diseases Research, and Medicinal Chemistry Research, Pharmacia and Upjohn, Inc.,
Kalamazoo, Michigan 49001
Received April 19, 1996X
Previously, 3-substituted cycloalkylpyranones, such as 2d , have proven to be effective inhibitors
of HIV protease. In an initial series of 3-(1-phenylpropyl) derivatives with various cycloalkyl
ring sizes, the cyclooctyl analog was the most potent. We became interested in exploring the
influence of other structural changes, such as substitution on the phenyl ring and saturation
of the 5,6-double bond, on the cycloalkyl ring size structure-activity relationship (SAR).
Saturation of the 5,6-double bond in the pyrone ring significantly impacts the SAR, altering
the optimal ring size from eight to six. Substitution of a sulfonamide at the meta position of
the phenyl ring dramatically increases the potency of these inhibitors, but it does not change
the optimal ring size in either the cycloalkylpyranone or the cycloalkyldihydropyrone series.
This work has led to the identification of compounds with superb binding affinity for the HIV
protease (Ki values in the 10-50 pM range). In addition, the cycloalkyldihydropyrones showed
excellent antiviral activity in cell culture, with ED50 values as low as 1 µM.
One strategy for controlling human immunodeficiency
virus (HIV) infection is inhibition of the HIV protease,
which plays a critical role in viral maturation.1 The
search for an inexpensive inhibitor with good oral
bioavailability at Pharmacia and Upjohn, Inc. began
with a screening program, and this resulted in the
discovery of a class of low molecular weight, non-peptidic
lead structures, the 4-hydroxycoumarins (e.g. 1).2,3 One
of the templates which we have since developed from
this lead structure is the cyclooctylpyranone 2d . We
chose to replace the rigid benzene ring of the coumarin
with conformationally-flexible cycloalkyl rings of various
sizes, and when we prepared the first series of cyclo-
alkylpyranone derivatives, this strategy resulted in a
dramatic increase in enzyme binding affinity. In both
this first series of analogs and a similar subsequent
series, it was clear that the optimal size of the cycloalkyl
ring was eight. X-ray crystal structures of 2c and 2d
complexed with HIV protease were obtained, and these
showed that the cycloalkyl ring folded into the S1 pocket
of the protease. These crystal structures also illustrated
that the cyclooctyl ring was better able to fill the S1
region than the cycloheptyl ring.4 As we developed this
new template, one of the issues of interest to us was
the impact of structural changes in the cycloalkylpyra-
nones, such as substitution on the phenyl ring and
saturation of the 5,6-double bond in the pyrone ring,
on the optimal size of the cycloalkyl ring.
of an aryl sulfonamide substituent at this position has,
in many cases, resulted in significant antiviral activity
in addition to potent enzyme inhibitory activity.6 We
were interested in exploring the impact of this substitu-
tion on the optimal cycloalkyl ring size. A sulfonamide
substituent which had been associated with good en-
zyme binding affinity in this and related templates6b,7
was selected, and the compounds of interest were
prepared in three steps from the known cycloalkylpyra-
nones 3a -e8 (Scheme 1).5,6 Substitution of the sulfon-
amide at the meta position dramatically improved the
enzyme binding affinity of these analogs relative to the
unsubstituted series.9,10 The optimal cycloalkyl ring
sizes, however, were unchanged, and the seven- and
eight-membered derivatives (5c,d ) were the most po-
tent. Addition of a P3′ substituent does not appear to
influence the structure-activity relationship (SAR) of
the P1 substituent in this template.
Recently, substitution at the meta position on the
phenyl ring of 2d has been shown to allow substituents
to reach the S3′ pocket, and this is associated with
enhanced enzyme binding affinity.5,6 In particular, use
We were also interested in the impact of a modifica-
tion much closer to the P1 cycloalkyl ring: saturation
of the 5,6-double bond in the pyrone ring. The cyclo-
alkyldihydropyrones were formed by catalytic hydroge-
nation of the 5,6-double bond in the cycloalkylpyranones
3a -d (Scheme 2). We found that hydrogenation of this
tetrasubstituted double bond required fairly rigorous
† Structural, Analytical and Medicinal Chemistry Research.
‡ Computer-Aided Drug Discovery.
§ Chemical and Biological Screening.
Cancer and Infectious Diseases Research.
| Medicinal Chemistry Research.
X Abstract published in Advance ACS Abstracts, September 1, 1996.
S0022-2623(96)00296-8 CCC: $12.00 © 1996 American Chemical Society

4126 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 20
Notes
Sch em e 1a
Sch em e 3a
a
(a) 3-NO2C6H4CHO, AlCl3; (b) Et3Al, CuBr-Me2S; (c) Pd/C,
H2, EtOH; (d) 4-cyano-2-pyridinesulfonyl chloride; pyridine, CH2Cl2.
a
(a) 3-EtCH(OH)C6H4NHCO2CH2Ph, p-TsOH; (b) Pd/C, cyclo-
hexene; (c) 4-cyano-2-pyridinesulfonyl chloride, pyridine, CH2Cl2.
predicted the 6-membered ring compound 7d would
have the best binding affinity, and this also was in
accord with the experimental results. In both series of
compounds, visual inspection of the modeled lowest
energy bound conformations indicated that the in-
creased activity of the 8-membered unsaturated and the
6-membered saturated analogs was explained by the
ability of the cycloalkyl ring to bend into the S1 pocket
and achieve maximal hydrophobic contact without also
inducing strain in the ring, as illustrated in Figures 1
and 2. The most potent compound in the unsaturated
series (2d ) had a better binding affinity than the most
potent compound of the saturated series (7b), and again,
this is most likely due to the greater degree of hydro-
phobic interaction between the cyclooctyl ring of 2d and
the S1 pocket of the enzyme relative to the cyclohexyl
ring of 7b, as illustrated in Figure 3.
Sch em e 2a
a
(a) PtO2, H2, HOAc; (b) PhCHO, AlCl3; (c) Et3Al, CuBr-Me2S.
Since the binding affinity of the first series of cyclo-
alkyldihydropyrones was relatively weak, we were
interested in preparing the m-sulfonamide-substituted
derivatives of these compounds. Synthesis of these
analogs was similar to the unsubstituted analogs above
(Scheme 3).11 In this case, the saturated intermediates
6a -d were condensed with 3-nitrobenzaldehyde. After
the C-3R ethyl substituent was in place, the nitro group
was reduced and the resulting aniline was coupled with
4-cyano-2-pyridinesulfonyl chloride to give 9a -d . As
expected, substitution of the sulfonamide at the meta
position greatly enhanced the HIV protease inhibitory
activity9 of these compounds relative to the unsubsti-
tuted analogs, presumably via interaction of the aryl-
sulfonamide with the S3′ region of the HIV protease.
In addition, the optimal ring size of these compounds
was six, just as it was in the unsubstituted series.
conditions, but we were able to selectively form the
desired intermediates 6a -d . 1H NMR experiments
indicate the cis isomer was formed exclusively. These
intermediates were then substituted at the C-3 position
via condensation with benzaldehyde, followed by con-
jugate addition of the C-3R ethyl substituent.11 The
enzyme binding affinity of analogs with ring sizes of five
to eight is shown in Scheme 2.12 In this case, the best
inhibitor was the cyclohexyl derivative 7b, not the
cyclooctyl derivative 7d . Clearly, saturation of the 5,6-
double bond of the central pyrone ring has a significant
impact on the ring size SAR.
To further explore the difference in ring size SAR for
the saturated and unsaturated series, we turned to
molecular modeling techniques. A Monte Carlo search
for low-energy conformations on each ring was carried
out with the ligand bound and unbound. The energy
difference between the two forms was used as an
approximate binding energy.13 For the cycloalkylpyra-
nones, binding energy calculations predicted the 8-mem-
bered ring analog 2d would have the best binding
affinity, and this was the case experimentally. For the
cycloalkyldihydropyrones, binding energy calculations
It is interesting to note that there appears to be some
flexibility in the identity of the P1′ substituent of the
cycloalkyldihydropyrones. We prepared the tert-butyl
derivatives of both the unsubstituted (10) and the
sulfonamide-substituted (11) cyclooctyldihydropyrone in
a manner analogous to that described above.15 In both
cases, the enzyme binding affinity of these compounds
was only somewhat less than that of the corresponding

Notes
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 20 4127
F igu r e 1. Overlay of the lowest energy bound conformations found using Monte Carlo searching of models of cycloalkylpyranones
in the binding site of HIV protease. The 4-hydroxypyrone ring, which forms hydrogen-bonding interactions with the catalytic
aspartic acids (CdO) and the isoluecine residues (OH) in the active site of the protease, is in the center of each molecule, with
oxygen atoms highlighted in red. The cycloalkyl rings are on the left side. Molecule coloring is according to the size of the
cycloalkyl ring: 5 (2a ) is purple, 6 (2b) is green, 7 (2c) is yellow, 8 (2d ) is blue, and 10 (2e) is white. The dot surface represents
the van der Waals surface of the HIV protease binding site as determined in an X-ray crystal structure of the protease with 2d .
The labels S2 through S2′ denote the location of binding pockets typically occupied by side chains of peptidic inhibitors. According
to these models, the cyclooctyl ring (blue) achieves the best hydrophobic interaction with the S1 region.
ethyl analogs 7d and 9d , respectively. The identity of
the sulfonamide substituent, however, appears to be
more important to the enzyme binding affinity of the
cycloalkyldihydropyrones. Compound 12, which has a
4-cyanophenyl substituent, is far less potent than the
corresponding 4-cyano-2-pyridyl analog 9b.16
little antiviral activity. The exception was the 8-mem-
bered ring analog 5d (ED50 ) 2.6 µM). In contrast, the
cycloalkyldihydropyrones generally demonstrated very
good antiviral activity.18 The most potent compound
was the 6-membered ring analog 9b (ED50 ) 0.95 µM;
ED90 ) 2.5 µM). The observed antiviral activity is not
due to cytotoxicity, as the CCTD50 values for all of the
compounds in Table 1 were above 30 µM.
In summary, saturation of the 5,6-double bond in the
pyrone ring of the cycloalkylpyranone template signifi-
cantly impacts the cycloalkyl ring SAR, altering the
optimal ring size from eight to six. Substitution of a
sulfonamide at the meta position of the phenyl ring,
however, does not change the optimal ring size in either
the cycloalkylpyranone or the cycloalkyldihydropyrone
series. Both of the sulfonamide-substituted series
demonstrated superb binding affinity for the HIV pro-
tease, with Ki values for the best compounds in the 10-
50 pM range. In addition, the cycloalkyldihydropyrones
showed excellent antiviral activity in cell culture.
Exp er im en ta l Section
Melting points are uncorrected. 1H NMR spectra were
measured on a Bruker AM 300 (300 MHz) instrument using
tetramethylsilane as an internal standard. All other physical
data were measured by the Analytical Chemistry group of
Pharmacia and Upjohn, Inc. Flash chromatography was
performed on 230-400 mesh silica gel 60.
All of the compounds prepared above were screened
for antiviral activity in HIV-1 infected H-9 cells.17 As
expected, the unsubstituted analogs 2a -e, 7a -d , and
10 showed little or no antiviral activity at concentrations
up to 30 µM. The results for the sulfonamide-substi-
tuted compounds, listed in Table 1, were somewhat
surprising. Despite excellent enzyme binding affinity,
the cycloalkylpyranones 5a -e on the whole showed
4-Cya n o-2-p yr id in esu lfon yl Ch lor id e.19 5-Cyano-2-mer-
captopyridine20 (2.5 g, 18.4 mmol) was suspended in 1 N HCl
(35 mL) and cooled in an ice bath. Chlorine was bubbled
through the suspension for 1 h. The reaction mixture was then

4128 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 20
Notes
F igu r e 2. Overlay of the lowest energy bound conformations found using Monte Carlo searching of models of cycloalkyldihy-
dropyrones in the binding site of HIV protease. Labeling is analogous to that used in Figure 1, with the molecules colored according
to ring size: 5 (7a ) is purple, 6 (7b) is green, 7 (7c) is yellow, and 8 (7d ) is blue. Relative to Figure 1, the view is rotated slightly
around the horizontal axis to allow a better view of the cycloalkyl rings. According to these models, the cyclohexyl ring (green)
achieves the best hydrophobic interaction with the S1 region.
F igu r e 3. Modeled binding orientation of the 8-membered ring analog 2d (blue), predicted to have the tightest binding in the
cycloalkylpyrone series, and the 6-membered analog 7b (green), predicted to have the tightest binding in the cycloalkyldihydro-
pyrone series. The apparent greater degree of projection of the blue ring into the S1 pocket relative to the green ring may account
for the higher potency of 2b measured in the enzyme inhibitory activity assay.
filtered, and the filter cake was washed with ice water (100
mL). After drying in vacuo, a pale yellow solid (2.27 g) was
obtained: mp 51-54 °C. This intermediate displayed limited
stability and was therefore used immediately in the next
reaction. Alternatively, it could be stored at -78 °C under
nitrogen for up to several months with only marginal loss in

Notes
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 20 4129
Ta ble 1. Biological Activity of m-Sulfonamide-Substituted
Analogs
2-pyridinesulfonyl chloride (0.11 g, 0.55 mmol), and pyridine
(0.089 mL, 1.10 mmol) in CH2Cl2 (3 mL) was stirred at room
temperature for 1.5 h. Column chromatography of the crude
reaction mixture on 60 g of silica gel (elution with 1-2.5%
MeOH/CHCl3) gave 0.100 g of a white foam. Crystallization
from CH2Cl2/hexane gave 0.051 g (21%) of 5a as white
crystals: mp 128-131 °C; 1H NMR (CDCl3) δ 8.98-8.92 (m, 1
H), 8.17-7.99 (m, 2 H), 7.22-7.17 (m, 3 H), 7.08-7.02 (m, 1
H), 6.89 (br s, 1 H), 5.75 (s, 1 H), 4.11 (t, J ) 7.9 Hz, 1 H),
2.85-2.75 (m, 2 H), 2.67-2.59 (m, 2 H), 2.16-2.01 (m, 4 H),
0.89 (t, J ) 7.3 Hz, 3 H) ppm; MS (EI) m/z 451 (M+). Anal.
(C23H21N3O5S) C, H, N.
compd
Ki (nM)a
ED50 (µM)b
5a
5b
5c
5d
5e
9a
9b
9c
9d
11
12
0.075
0.065
0.018
0.007
0.25
0.147
0.050
0.19
8.6
>3c
>3c
2.6
>3c
2.9
0.95
2.0
2.9
0.175
0.20
15
2.5
Rep r esen ta tive P r oced u r e for th e P r ep a r a tion of 6a -
d. Tetr ah ydr ocyclopen ta[b]pyr an -2,4(3H,4aH)-dion e (6a).
A solution of PtO2 (2.3 g) and 3a (9.13 g, 60.0 mmol) in HOAc
(150 mL) was hydrogenated at 50 psi for 1.5 h. The reaction
mixture was then filtered through Celite and concentrated to
a brown solid. Column chromatography on 150 g of silica gel
(elution with 1-10% MeOH/CHCl3) gave 2.91 g of a yellow
solid. Crystallization from CH2Cl2/hexane gave 2.32 g (25%)
>3c
a
b
HIV protease inhibitory activity. Antiviral activity in HIV-
1-infected H-9 cells. c <10% inhibition at 3 µM.
chemical integrity. However, decomposition with evolution of
gas was noted occasionally upon warming to room tempera-
ture.
1
of 6a as white crystals: mp 127-132 °C; H NMR (CDCl3) δ
3-[(Ben zyloxyca r bon yl)a m in o]-r-eth ylben zyl Alcoh ol.
A mixture of 3-nitropropiophenone (7.17 g, 40.0 mmol) and
10% Pd/C (0.5 g) in MeOH (150 mL) was placed on a Parr
hydrogenation apparatus (initial pressure 50 psi) for 3 h. The
reaction mixture was then filtered through Celite and con-
centrated in vacuo to give 6.09 g of crude 1-(3-aminophenyl)-
1-propanol as a brown oil, which was used without further
purification. This intermediate was immediately dissolved in
150 mL of CH2Cl2, and (i-Pr)2NEt (8.4 mL, 48.0 mmol) was
added. Benzyl chloroformate (6.3 mL, 44.0 mmol) was then
added slowly, and the resulting solution was stirred at room
temperature for 18 h. Cold HCl (0.5 N) was added, and the
layers were separated. The aqueous layer was extracted with
CH2Cl2 (100 mL), and the combined organic layers were
concentrated to give 13.8 g of a brown oil. Column chroma-
tography on 200 g of silica gel (elution with 40-50% Et2O/
hexane) gave an oil which was crystallized from EtOAc/hexane
to yield 10.0 g (88%) of the title compound as a white
crystalline solid: mp 88-91 °C; 1H NMR (CDCl3) δ 7.42-7.33
(m, 6 H), 7.29-7.24 (m, 2 H), 7.06-7.03 (m, 1 H), 6.71 (br s, 1
H), 5.20 (s, 2 H), 4.57 (t, J ) 6.5 Hz, 1 H), 1.91 (br s, 1 H),
1.84-1.68 (m, 2 H), 0.91 (t, J ) 7.4 Hz, 3 H) ppm.
5.10-5.06 (m, 1 H), 3.57 (d, J ) 17.3 Hz, 1 H), 3.38 (d, J )
17.3 Hz, 1 H), 2.91-2.84 (m, 1 H), 2.21-2.08 (m, 2 H), 1.99-
1.78 (m, 4 H) ppm; 13C NMR (CD3OD) δ 178.0, 171.9, 83.7,
42.8, 34.2, 31.1, 24.2 ppm; MS (EI) m/z 154 (M+). Anal.
(C8H10O3) C, H.
Rep r esen ta tive P r oced u r e for th e P r ep a r a tion of 7a -
d , 8a -d , a n d 10. 5,6,7,7a -Tetr a h yd r o-4-h yd r oxy-3-(1-
p h en ylp r op yl)-4a H-cyclop en ta [b]p yr a n -2-on e (7a ). A so-
lution of AlCl3 (0.346 g, 2.59 mmol) in THF (4 mL) was added
to a solution of 6a (0.200 g, 1.30 mmol) and benzaldehyde (0.13
mL, 1.31 mmol) in THF (8 mL). The resulting mixture was
stirred at room temperature for 2.5 h, Na2CO3‚10H2O (0.50 g)
was added, and the mixture was stirred an additional 20 min.
The mixture was then dried over MgSO4, filtered through
Celite, and concentrated to yield 0.263 g of yellow oil. This
material was immediately dissolved in a solution of CuBr-
Me2S (0.80 g, 0.389 mmol) in THF (4 mL), and Et3Al (1.4 mL
of 1.0 M in hexane) was added. The reaction mixture was then
stirred at room temperature for 18 h, quenched with water,
and diluted with Et2O. The organic layer was separated,
filtered through Celite, and concentrated to give 0.145 g of
clear oil. Column chromatography (elution with 10-30%
EtOAc/hexane) yielded 0.043 g of white solid. Crystallization
from CH2Cl2/hexane gave 0.026 g (7%) of 7a as white crys-
R ep r esen t a t ive P r oced u r e for t h e P r ep a r a t ion of
2e a n d 4a -e. [3-[1-(6,7-Dih yd r o-4-h yd r oxy-2-oxo-2(5H)-
cyclop en ta [b]p yr a n -3-yl)p r op yl]p h en yl]ca r ba m ic Acid ,
P h en ylm eth yl Ester (4a ). A mixture of 3a (0.806 g, 5.3
mmol), 3-[(benzyloxycarbonyl)amino]-R-ethylbenzyl alcohol (1.51
g, 5.3 mmol), and p-toluenesulfonic acid (0.32 g, 1.6 mmol) in
toluene (50 mL) was heated to reflux for 5 h in a flask equipped
with a Dean-Stark trap. The reaction mixture was then
diluted with EtOAc, washed with water and saturated NaH-
CO3, and concentrated to 2.17 g of an oil. Column chroma-
tography on 75 g of silica gel (elution with 20-50% EtOAc/
hexane and 5% MeOH/CHCl3) gave 0.317 g (14%) of 4a as a
1
tals: mp 94-98 °C; H NMR (CDCl3) δ 7.38-7.19 (m, 5 H),
4.12-4.09 (m, 0.4 H), 3.67-3.65 (m, 0.6 H), 3.28-3.20 (m, 2
H), 2.61-2.50 (m, 2 H), 2.14-1.87 (m, 6 H), 0.72-0.67 (m, 3
H) ppm; MS (EI) m/z 272 (M+); HRMS (EI) calcd for C17H20O3
272.1412, found 272.1410.
Rep r esen ta tive P r oced u r e for th e P r ep a r a tion of 9a -
d , 11, a n d 12. 4-Cya n o-N-[3-[1-(5,6,7,7a -tetr a h yd r o-4-
h ydr oxy-2-oxo-4aH-cyclopen ta[b]pyr an -3-yl)pr opyl]ph en -
yl]-2-p yr id in esu lfon a m id e (9a ). A solution of 8a (0.300 g,
0.945 mmol) and 10% Pd/C (0.15 g) in EtOH (50 mL) was
hydrogenated at an initial pressure of 50 psi for 2 h. The
reaction mixture was then filtered through Celite and con-
centrated in vacuo to give 0.269 g (99%) of an aniline
1
beige solid: mp 84-88 °C; H NMR (CDCl3) δ 7.37-7.23 (m,
8 H), 7.13 (d, J ) 7.4 Hz, 1 H), 6.70 (s, 1 H), 5.18 (s, 2 H), 4.24
(t, J ) 7.8 Hz, 1 H), 2.77-2.72 (m, 2 H), 2.60-2.55 (m, 2 H),
2.16-1.98 (m, 4 H), 0.98 (t, J ) 7.3 Hz, 3 H) ppm; 13C NMR
(CDCl3) δ 166.6, 163.6, 163.4, 153.3, 143.7, 138.4, 135.9, 129.8,
128.6, 128.3, 128.2, 122.8, 118.0, 117.3, 111.3, 104.7, 67.0, 41.8,
31.2, 25.5, 24.0, 19.8, 12.5 ppm; MS (EI) m/z 419 (M+). Anal.
(C25H25NO5) C, H, N.
1
intermediate as a gray foam: mp 75-79 °C; H NMR (CD3-
OD) δ 6.94 (t, J ) 7.7 Hz, 1 H), 6.83-6.72 (m, 2 H), 6.52-6.50
(m, 1 H), 4.89-4.71 (m, 1 H), 3.95-3.89 (m, 1 H), 2.67-2.56
(m, 1 H), 2.29-2.07 (m, 2 H), 2.07-1.86 (m, 4 H), 1.82-1.59
(m, 2 H), 0.87 (dt, J ) 1.5, 7.3 Hz, 3 H) ppm; MS (EI) m/z 287
(M+). The aniline (0.254 g, 0.884 mmol) was promptly dis-
solved in CH2Cl2 (5 mL), and 4-cyano-2-pyridinesulfonyl
chloride (0.179 g, 0.884 mmol) and pyridine (0.143 mL, 1.77
mmol) were added to the solution. The resulting mixture was
stirred at room temperature for 18 h and then purified by
column chromatography on 60 g of silica gel (elution with
0-5% MeOH/CHCl3) to give 0.233 g of a white foam. Crystal-
lization from CH2Cl2 gave 0.167 g (42%) of 9a as white
crystals: mp 124-127 °C; 1H NMR (CDCl3) δ 9.00-8.94 (m, 1
H), 8.12-7.95 (m, 2 H), 7.51 (s, 0.5 H), 7.39 (s, 0.5 H), 7.24-
7.01 (m, 4 H), 5.11-5.08 (m, 1 H), 3.84-3.81 (m, 1 H), 3.24-
3.21 (m, 1 H), 2.92-2.87 (m, 1 H), 2.15-2.10 (m, 1 H), 2.00-
Rep r esen ta tive P r oced u r e for th e P r ep a r a tion of 5a -
e. 4-Cya n o-N-[3-[1-(6,7-d ih yd r o-4-h yd r oxy-2-oxo-2(5H)-
cyclop en t a [b]p yr a n -3-yl)p r op yl]p h en yl]-2-p yr id in esu l-
fon a m id e (5a ). A mixture of 4a (0.23 g, 0.55 mmol) and 10%
Pd/C (0.15 g) in cyclohexene (4 mL) and EtOH (4 mL) was
warmed to reflux for 1.5 h. The mixture was then filtered
through Celite, washed with EtOH, and concentrated to give
0.17 g (100%) of an aniline intermediate as a yellow oil: 1H
NMR (CD3OD) δ 7.23-7.16 (m, 1 H), 6.96 (t, J ) 7.7 Hz, 1 H),
6.84-6.78 (m, 1 H), 6.56-6.52 (m, 1 H), 4.10-4.05 (m, 1 H),
2.76-2.62 (m, 4 H), 2.36-2.21 (m, 2 H), 2.12-2.02 (m, 2 H),
0.87 (t, J ) 7.3 Hz, 3 H) ppm; MS (EI) m/z 285 (M+). A mixture
of the above aniline intermediate (0.17 g, 0.55 mmol), 4-cyano-

4130 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 20
Notes
1.61 (m, 6 H), 0.69-0.63 (m, 2 H) ppm; MS (EI) m/z 453 (M+).
tions of Ki values much less than half the enzyme concentration,
the usual lower limit for IC50 determination. Although we
cannot calculate the IC50 values using this method, since these
inhibitors are competitive (as shown by early kinetics work2 and
by crystal structures of inhibitors complexed with the active site
of HIV protease4-6), by definition the IC50 values are identical
to the Ki values.
Anal. (C23H23N3O5S) C, H, N.
Su p p or tin g In for m a tion Ava ila ble: Yields and spectral
data for compounds 2e, 4b-e, 5b-e, 6b-d , 7b-d , 8a -d , 9b-
d , and 10-12 (9 pages). Ordering information is given on any
current masthead page.
(10) The dramatic increase in binding affinity of the meta sulfon-
amide substituted derivatives is presumably due to the ability
of the sulfonamide aryl substituent, in this case the 4-cyano-2-
pyridyl group, to efficiently interact with the S3′ pocket of the
protease. In addition, X-ray crystal structures of similar com-
pounds have shown that the aryl group occupying the S3′ region
is able to form π-stacking interactions with the Arg 8 residue of
the enzyme. See ref 6.
(11) These synthetic procedures were adapted from routes developed
for related templates. Romero, D. L.; Tommasi, R. A.; J ana-
kiraman, M. N.; Strohbach, J . W.; Turner, S. R.; Biles, C.; Morge,
R. A.; J ohnson, P. D.; Aristoff, P. A.; Tomich, P. K.; Lynn, J . C.;
Horng, M.-M.; Chong, K.-T.; Hinshaw, R. R.; Howe, W. J .; Finzel,
B. C.; Watenpaugh, K. D.; Thaisrivongs, S. J . Med. Chem.,
submitted.
Refer en ces
(1) (a) Thaisrivongs, S. HIV Protease Inhibitors. Annu. Rep. Med.
Chem. 1994, 17, 133-144. (b) Redshaw, S. Inhibitors of HIV
Proteinase. Exp. Opin. Invest. Drugs 1994, 3, 273-286.
(2) Thaisrivongs, S.; Tomich, P. K.; Watenpaugh, K. D.; Chong, K.-
T.; Howe, W. J .; Yang, C.-P.; Strohbach, J . W.; Turner, S. T.;
McGrath, J . P.; Bohanon, M. J .; Lynn, J . C.; Mulichak, A. M.;
Spinelli, P. A.; Hinshaw, R. R.; Pagano, P. J .; Moon, J . B.;
Ruwart, M. J .; Wilkinson, K. F.; Rush, B. D.; Zipp, G. L.; Dalga,
R. J .; Schwende, F. J .; Howard, G. M.; Padbury, G. E.; Toth, L.
N.; Zhao, Z.; Koeplinger, K. A.; Kakuk, T. J .; Cole, S. L.; Zaya,
R. M.; Piper, R. C.; J effrey, P. Structure-Based Design of HIV
Protease Inhibitors: 4-Hydroxycoumarins and 4-Hydroxy-2-
pyrones as Non-peptidic Inhibitors. J . Med. Chem. 1994, 37,
3200-3204.
(12) In vitro enzyme kinetics for these compounds were performed
as described in ref 2.
(13) Energy-based conformational searching was carried out with the
Monte Carlo multiple minimum (MCMM) facility of Batchmin
v. 4.5,14 using the Amber* united atom forcefield, the PRCG
minimizer, and the surface area solvation approximation, as
implemented in BatchMin. A model of each compound was
constructed in the HIV-2 protease binding site, starting from
the X-ray crystal structure of the complex of 2c. Each model
was then subjected to 1000 steps of torsional variation of the
cycloalkyl ring bonds, followed by energy minimization of the
ligand within the field of an 8 Å nonmoving shell of protein
atoms, to arrive at an energy-ordered list of “bound” conforma-
tions for each ligand. The ligand model was then moved away
from the protein and the procedure was repeated. The nonmov-
ing shell of protein atoms contributed a solvation component to
the overall “unbound” energies, even though it did not contribute
an interaction energy with the ligand. The difference between
the lowest bound and unbound energies became the approximate
binding energy for that compound. The entire procedure was
repeated, using the X-ray structure of the complex of 2d as the
source of protein atom positions. The binding energies calcu-
lated in the two runs were then averaged.
(14) Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W.
C. MacroModel -- An Integrated Software System for Modeling
Organic and Bioorganic Molecules using Molecular Mechanics.
J . Comput. Chem. 1990, 11, 440-467.
(15) The intermediates for 10 and 11 were prepared using a tert-
butyl organozinc reagent instead of Et3Al. For the synthesis of
this reagent, see: J ubert, C.; Knochel, P. Preparation of Poly-
functional Nitro Olefins and Nitroalkanes Using the Copper-
Zinc Reagents RCu(CN)ZnI. J . Org. Chem. 1992, 57, 5431-
5438.
(3) For a review of the 4-hydroxycoumarins and related 4-hydroxy-
pyrone templates, see: Romines, K. R.; Chrusciel, R. A. 4-Hy-
droxypyrones and Related Templates as Nonpeptidic HIV Pro-
tease Inhibitors. Current Med. Chem. 1995, 2, 825-838 and
references cited therein.
(4) Romines, K. R.; Watenpaugh, K. D.; Tomich, P. K.; Howe, W.
J .; Morris, J . K.; Lovasz, K. D.; Mulichak, A. M.; Finzel, B. C.;
Lynn, J . C.; Horng, M. M.; Schwende, F. J .; Ruwart, M. J .; Zipp,
G. L.; Chong, K.-T.; Dolak, L. A.; Toth, L. N.; Howard, G. M.;
Rush, B. D.; Wilkinson, K. F.; Possert, P. L.; Dalga, R. J .;
Hinshaw, R. R. Use of Medium-Sized Cycloalkyl Rings to
Enhance Secondary Binding: Discovery of a New Class of HIV
Protease Inhibitors. J . Med. Chem. 1995, 38, 1884-1891.
(5) Romines, K. R.; Watenpaugh, K. D.; Howe, W. J .; Tomich, P.
K.; Lovasz, K. D.; Morris, J . K.; J anakiraman, M. N.; Lynn, J .
C.; Horng, M.-M.; Chong, K.-T.; Hinshaw, R. R.; Dolak, L. A.
Structure-Based Design of Nonpeptidic HIV Protease Inhibitors
from a Cyclooctylpyranone Lead Structure. J . Med. Chem. 1995,
38, 4463-4473.
(6) (a) Skulnick, H. I.; J ohnson, P. D.; Howe, W. J .; Tomich, P. K.;
Chong, K.-T.; Watenpaugh, K. D.; J anakiraman, M. N.; Dolak,
L. A.; McGrath, J . P.; Lynn, J . C.; Horng, M.-M.; Hinshaw, R.
R.; Zipp, G. L.; Ruwart, M. J .; Schwende, F. J .; Zhong, W.-Z.;
Padbury, G. E.; Dalga, R. J .; Shiou, L.; Possert, P. L.; Rush, B.
D.; Wilkinson, K. F.; Howard, G. M.; Toth, L. N.; Williams, M.
G.; Kakuk, T. J .; Cole, S. L.; Zaya, R. M.; Lovasz, K. D.; Morris,
J . K.; Romines, K. R.; Thaisrivongs, S.; Aristoff, P. A. J . Med.
Chem. 1995, 38, 4968-4971. (b) Skulnick, H. I.; J ohnson, P.
D.; Aristoff, P. A.; Morris, J . K.; Lovasz, K. D.; Howe, W. J .;
Watenpaugh, K. D.; J anakiraman, M. N.; Anderson, D. J .;
Reischer, R. J .; Schwartz, T. M.; Banitt, L. S.; Tomich, P. K.;
Lynn, J . C.; Horng, M.-M.; Chong, K.-T.; Hinshaw, R. R.; Dolak,
L. A.; Seest, E. P.; Schwende, F. J .; Rush, B. D.; Howard, G. M.;
Toth, L. N.; Wilkinson, K. R.; Kakuk, T. J .; J ohnson, C. W.; Cole,
S. L.; Zaya, R. M.; Zipp, G. L.; Possert, P. L.; Dalga, R. J .;
Romines, K. R. Structure-Based Design of Nonpeptidic HIV
Protease Inhibitors: The Sulfonamide-Substituted Cyclooc-
tylpyranones. J . Med. Chem., submitted.
(16) We have observed this effect in other templates.6b,7 It suggests
that there is an additional component in the binding interaction
between the S3′ region of the enzyme and the pyridyl sulfon-
amide substituent relative to the phenyl sulfonamide substitu-
ents, but the specific nature of such interactions is unknown at
this time.
(17) Antiviral activity was measured as described in Chong, K.-T.;
Pagano, P. J .; Hinshaw, R. R. Bisheteroarylpiperazine Reverse
Transcriptase Inhibitor in Combination with 3′-Azido-3′-Deoxy-
thymidine or 2′,3′-Dideoxycytidine Synergistically Inhibits Hu-
(7) J . W. Strohbach, S. R. Turner, and S. Thaisrivongs. Unpublished
Results.
(8) Effenberger, F.; Ziegler, T.; Scho¨nwa¨lder, K.-H.; Kesmarszky,
T.; Bauer, B. Die Acylierung von (Trimethylsilyl)enolethern mit
Malonyldichlorid -- Darstellung von 4-Hydroxy-2H-pyran-2-onen.
Chem. Ber. 1986, 119, 3394-3404.
(9) In vitro enzyme kinetics were performed with a fused recombi-
nant enzyme instead of the dimeric native protease to obviate
enzyme dissociation. To obtain these lower Ki values, assays
were incubated for 72 h at a lower enzyme concentration (0.2-
0.5 nM). Otherwise the assay conditions were as described in
ref 2. We note that this is not an initial rate assay, but an end-
point assay. In the absence of inhibitor, more than 99.5% of
the substrate is converted to product; the percent conversion in
the presence of inhibitor depends on inhibitor concentration. In
our assay, the end points are measured for up to 96 different
inhibitor concentrations, and the Ki value is then calculated
using an equation in which the enzyme rate equation is solved
as a function of substrate concentration. When many enzyme
concentrations are used, this technique allows accurate calcula-
man Immunodeficiency Virus Type
1 Replication In Vitro.
Antimicrob. Agents Chemother. 1994, 38, 288-293.
(18) It has been our experience that the translation of enzyme binding
affinity, as indicated by the Ki values, into antiviral activity in
cell culture cannot be directly correlated to any single factor,
such as protein binding, cell membrane permeability, or solubil-
ity, just to name a few. This issue continues to be an area of
ongoing research.
(19) Prepared using the procedure of Roblin and Clapp: Roblin, R.
O.; Clapp, J . W. J . Am. Chem. Soc. 1950, 72, 4890-4892.
(20) Forrest, H. S.; Walker, J . Chemotherapeutic Agents of the
Sulphone Type. Part V. 2:5-Disubstituted Derivatives of Pyr-
idine. The Preparation of Heterocyclic Sulfonamides. J . Chem.
Soc. 1948, 1939-1945.
J M960296C