Optimization of pyrrolidinone based HIV protease inhibitors

Article abstract of DOI:10.1016/j.bmcl.2004.10.029

The synthesis and enhanced antiviral activity of a series of P1′-optimized 3,5-dibenzyl pyrrolidinone based HIV protease inhibitors is described. Optimization of P1-substituted pyrrolidinone based HIV protease inhibitors has yielded analogs with very pote

Full text of DOI:10.1016/j.bmcl.2004.10.029

Bioorganic & Medicinal Chemistry Letters 15 (2005) 81–84
Optimization of pyrrolidinone based HIV protease inhibitors
Ronald G. Sherrill,* C. Webster Andrews, William J. Bock,
Ronda G. Davis-Ward, Eric S. Furfine, Richard J. Hazen, Randy D. Rutkowske,
Andrew Spaltenstein and Lois L. Wright
GlaxoSmithKline, 5 Moore Drive, Research Triangle Park, NC 27709, USA
Received 1 September 2004; revised 7 October 2004; accepted 9 October 2004
Available online 11 November 2004
Abstract—Optimization of P1-substituted pyrrolidinone based HIV protease inhibitors has yielded analogs with very potent anti-
viral activity.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
P1
P1'
The HIV pandemic continues unabated. However, in the
western world, HAART (Highly Active Anti-Retroviral
Therapy) has made dramatic impacts on both the mort-
ality and morbidity associated with HIV infection. HIV
protease inhibitors were pivotal in early triple combina-
tion therapy trials and continue to be a critical compo-
nent of HIV therapy.1
OH
OH
H
N
N
O
O
1a
Figure 1.
Our efforts in this field of research led to the generation
of a novel series of inhibitors where a 3,5-disubstituted
pyrrolidinone occupies the P1/P2 pocket of the active
site of the viral protease.2 The 2,5-dibenzyl substituted
pyrrolidinone 1a (Fig. 1) demonstrated exceptional en-
zyme inhibition (IC50 = 0.050nM). Unfortunately, the
antiviral activity (ED50 = 0.72lM) of this analog in cell
based assays failed to match prototypical inhibitors such
as the marketed drug, amprenavir (ED50 = 0.15lM).
rationalized that the para-position of the P1-benzyl
group would be near to the exterior of the active site
and therefore might tolerate additional functionaliza-
tion (Fig. 2). Similarly, Merck researchers identified
the para-position of the P10-benzyl3 as having similar at-
tributes in an identical hydroxyethylene backbone.
Researchers at Merck have previously reported that po-
lar, hydrophilic, phenolic substituents tethered to the P1
or P10 phenyl groups of indinavir progenitors could en-
hance antiviral potency.3 Similarly, we sought to optim-
ize substitutions at these positions within our inhibitor
in order to demonstrate increased antiviral potency.
Using molecular modeling derived from the crystal
structure4 of a similar analog bound to the HIV protease
protein where the P1 position was unsubstituted, we
2. Chemistry
The general synthetic strategy for construction of these
analogs (see Fig. 3) involved a diastereoselective synthe-
sis of the P1/P2 pyrrolidinone scaffolds 7a and 7c fol-
lowed by addition to epoxides 8a and 8b. Preparation
of the N1-unsubstituted 3,5-dibenzyl pyrrolidinone
intermediates was accomplished through Wittig olefina-
tion of either N-Boc-L-phenylalaninal (3a) or N-Boc-O-
benzyl-L-tyrosinal (3b) with known phosphorane 2,5
available through alkylation of methyl triphenylphosph-
orylideneacetate with benzyl bromide. As expected, the
Keyword: HIV protease inhibitors.
*
Corresponding author. Tel.: +1 919 483 6281; fax: +1 919 483
6053; e-mail: ron.g.sherrill@gsk.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.10.029

82
R. G. Sherrill et al. / Bioorg. Med. Chem. Lett. 15 (2005) 81–84
X
PPh3
3a (X=H)
+
3b (X=OBn)
CO2Me
H
NHBOC
2
O
a
b
X
4a (X=H)
4b (X=OBn)
BOCNH
CO2Me
X
X
Figure 2.
BOCNH
CO2Me
BOCNH
CO2Me
5a (X=H)
5:1
2:1
6a (X=H)
5b (X=OH)
6b (X=OH)
olefination yielded exclusively the E-isomer of enoates
4a and 4b in good yields. Diastereoselectivity, induced
through 1,3-allylic strain,6 was achieved in the subse-
quent catalytic reduction of the olefin bond to give pre-
dominantly the desired anti-isomer. Reduction of
tyrosine derived enoate 4b provided a 2:1 mixture of anti
and syn isomers, 5b and 6b, in essentially quantitative
yield with concomitant loss of the benzyl protecting
group. On the other hand, phenylalanine derived enoate
4a provided a 5:1 mixture of anti and syn isomers 5a and
6a, although we could not readily rationalize the im-
proved diastereoselectivity observed. Neither mixture
of diastereomers could be resolved by chromatography
at this juncture.
c
X
NH
O
7a (X=H)
Y
7b (X=OH)
d
7c (X=OMom)
e
O
O
N
Acidic deprotection of the Boc group on the diastereo-
meric mix of 5 and 6 followed by basification to induce
cyclization provided pyrrolidinones 7a and 7b. The min-
or cis epimer of 2,5-dibenzyl pyrrolidinones 7a was chro-
matographically removed using TLC mesh silica gel.
The relative stereochemistry of 7a was assigned based
on 2D-NMR analysis and later confirmed through sin-
gle-crystal X-ray diffraction. We had previously estab-
lished the preferred stereochemistry as the trans
isomer.2 The phenolic pyrrolidinone 7b was inseparable
from the contaminating diastereomer. Therefore, the 2:1
mixture of trans and cis isomers was carried through the
remainder of the sequence following installation of the
methoxymethyl ether protecting group on the phenol
as 7c.
O
X
Y
8a Y=OBn
8b Y=H
OH
OH
H
N
N
O
O
1b
1d
(X=H, Y=OBn)
(X=H, Y=OH)
f
g or h or i or j
g or h
1f,h,j,l−p (X=H, Y=OR)
or
1e
(X=OH, Y=H)
1c,g,i,k (X=OR, Y=H)
Figure 3. (a) CHCl3, rt; for 4a: 77%, for 4b 71%; (b) 5% Pd/C, H2
(1atm), MeOH; for 5a/6a and 5b/6b 100%; (c) i. TFA/CH2Cl2 (1:1); ii.
K2CO3 (4equiv), MeOH; for 7a 69%, for 7b 100%; (d) NaH, DMF,
methoxymethyl chloride, 74%; (e) i. P4-phosphazene, ꢀ20°C, THF; ii.
21–26equiv concd HCl, IPA, rt; for 1b: 85 and 93%, respectively; for
1e: 76% and 60%, respectively; (f) Pd(OH)2/C, EtOH, 92%, for 1d; (g)
Cs2CO3, dioxane, halide Z, 80°C, 28–98%; for 1c Z = benzyl bromide,
for 1f and 1g Z = N-chloroethyl-morpholine, for 1j and 1k Z = N-
(bromoacetyl)morpholine,9 for 1m Z = 2-bromoethyl 2-methoxyethyl
ether, for 1n Z = 4-morpholinecarboxylic acid, 2-bromoethyl ester;10
(h) for 1h and 1i; i. methyl bromoacetate Cs2CO3, dioxane, 80°C, 63–
64%; ii. LiBH4, THF 36–67%; (i) for 1l, 2-pyridinemethanol, DEAD,
PPh3, THF, 16%; (j) i. methyl bromoacetate Cs2CO3, dioxane, 80° C,
64%; for 1o: ii. THF/H2O/TEA 3/1.5/0.75 at reflux; 39%; iii. EDC,
HOBT, DMF, N-methylpiperazine, 24%; or for 1p: ii. 2M NH3 in
MeOH, 64%.
Epoxide 8a was prepared identically to known epoxide
8b, used in the manufacturing of indinavir.7 However,
3-[4-(benzyloxy)phenyl]propionic acid8 was substituted
for phenylpropionic acid as starting material.
Addition of pyrrolidinones 7a and 7c to epoxides 8a and
8b, respectively, using P4-phosphazene base at ꢀ20°C
followed by acid deprotection of the acetonide with
HCl generated the complete hydroxyethylene backbone
of analogs 1b and 1e. The deprotection protocol con-
comitantly removed the methoxymethyl ether protecting
group at P1 yielding phenol analog 1e. The benzyl pro-
tecting group at the P10 position was removed using
catalytic hydrogenation to provide phenol 1d.

R. G. Sherrill et al. / Bioorg. Med. Chem. Lett. 15 (2005) 81–84
83
Tethered analogs 1f,g,j,k,m,n were prepared through
alkylation of the phenol and the appropriate halide with
Cs2CO3 in dioxane at 80°C. Analog 1l was prepared via
Mitsonobu coupling of 1d with 2-pyridylmethanol. Eth-
oxy analogs 1h and 1i were prepared via alkylation of
phenols 1d and 1e, respectively, with methyl bromoace-
tate followed by LiBH4 reduction of the intermediate
ester. Similarly, amide analogs 1o and 1p were prepared
by initial alkylation of phenol 1d with methyl bromoace-
tate. Direct aminolysis of the intermediate ester yielded
acetamide 1p whereas basic hydrolysis of the ester and
carbodiimide coupling of the acid with N-methylpiper-
azine yielded 1o.
an identical ethanolamine backbone:3 the phenols, 1d
and 1e; the morpholinoethoxy ethers, 1f and 1g; the eth-
oxy ethers 1h and 1i; and the morpholineacetamides, 1j
and 1k.
Each of the P1 substituted phenols demonstrated de-
creased enzyme potency. Antiviral potency relative to
the unsubstituted phenyl analog 1a, was either similar
or decreased excepting morpholineacetamide 1k and
phenol 1e, which both demonstrated increased antiviral
activity of 0.20lM versus 0.72lM from the parent. Of
particular note, the morpholinoethoxy analog 1g dis-
played a greater than 20-fold loss in enzyme potency with
a concomitant loss of antiviral activity to 0.96lM. The
measured loss of enzyme and antiviral potency at P1
could not be rationalized by the presence of the contam-
inating one third inactive diastereomer in these analogs.
Apparently, the greater conformational constraint inher-
ent in the pyrrolidinone system did not allow optimal
placement of even small, flexible substituents at this posi-
tion. For these reasons, we elected to focus the remainder
of our efforts on P10 substituted analogs (Table 1).
3. Results and discussion
Initially, we sought to focus our efforts toward either the
P1 or P10 position through synthesis of a representative
set of comparative analogs. For this purpose, we chose
four substituents which had previously imparted in-
creased antiviral efficacy in HIV protease inhibitors with
Table 1.
X
Y
OH
OH
H
N
N
O
O
1a-p
Analoga
Xb
Y
Ki (nM)c
IC50 (lM)d
amprenavir
0.04
0.05
0.11
0.13
0.06
0.13
0.15 ( 0.01)
0.72 ( 0.11)
0.92 ( 0.09)
0.55 ( 0.05)
0.32 ( 0.04)
0.20 ( 0.05)
1a
1b
1c
1d
1e
H
H
H
OBn
H
OBn
H
OH
H
OH
1f
H
0.14
1.30
0.083 ( 0.02)
0.96 ( 0.14)
—OCH2CH2-N
O
1g
H
—OCH2CH2-N
O
1h
1i
H
–OCH2CH2OH
–OCH2CH2OH
H
0.02
0.30
0.094 ( 0.01)
0.48 ( 0.03)
1j
H
0.02
0.10
0.07
0.063 ( 0.02)
0.20 ( 0.02)
0.38 ( 0.04)
—OCH2CON
O
1k
1l
H
—OCH2CON
O
—OCH2
N
H
1m
1n
H
H
–O(CH2CH2O)2CH3
0.02
0.05
0.093 ( 0.02)
0.24 ( 0.02)
—OCH2CH2OC(O)N
O
1o
1p
H
H
0.05
0.07
0.13 ( 0.03)
0.36 ( 0.03)
—OCH2CON
NMe
–OCH2CONH2
a All compounds were >95% pure by 1H NMR and HPLC.
b Analogs where X 5 H are 2:1 mixture of trans:cis pyrrolidinone.
c Ki, enzyme inhibition constant,11 %CV 0.15–1.15.
d IC50, antiviral inhibition against HIV-1 in MT-4 cell culture;12 NP6.

84
R. G. Sherrill et al. / Bioorg. Med. Chem. Lett. 15 (2005) 81–84
The P10 substituted phenols, in contrast, demonstrated
similar or increased enzyme inhibition relative to the
unsubstituted analog. More importantly, each of the
P10 analogs with the exception of the benzyl analog 1b
demonstrated increased antiviral activity in our cell
based assay. As observed for the P1 substituted analog,
the morpholinoethoxy analog 1f also demonstrated the
most dramatic loss in enzyme potency, 0.14 versus
0.05nM. However, an enhanced antiviral potency was
recorded for this compound, 0.08lM versus 0.72lM
for the parent. Particularly noteworthy, analogs 1f, 1h,
1j and 1m all demonstrated somewhat improved antivi-
ral efficacy relative to the marketed protease inhibitor,
amprenavir.
2. Spaltenstein, A.; Almond, M. R.; Bock, W. J.; Cleary, D.
G.; Furfine, E. S.; Hazen, R. J.; Kazmierski, W. M.;
Salituro, F. G.; Tung, R. D.; Wright, L. L. Bioorg. Med.
Chem. Lett. 2000, 10, 1159–1162.
3. Thompson, W. J.; Fitzgerald, P. M. D.; Holloway, M. K.;
Emini, E. A.; Darke, P. L.; McKeever, B. M.; Schleif, W.
A.; Quintero, J. C.; Zugay, J.; Tucker, T. J.; Schwering, J.
E.; Homnick, C. F.; Nunberg, J.; Springer, J. P.; Huff, J.
R. J. Med. Chem. 1992, 35, 1685–1701.
4. Kazmierski, W. M.; Furfine, E.; Gray-Nunez, Y.; Spal-
tenstein, A.; Wright, L. Bioorg. Med. Chem. Lett. 2004, 14,
5689–5692.
5. Bestmann, H. J.; Schulz, H. Tetrahedron Lett. 1960, 4, 5–6.
6. Kauppinen, P. M.; Koskinen, A. M. P. Tetrahedron Lett.
1997, 38, 3103–3106.
7. Askin, D.; Eng, K. K.; Rossen, K.; Purick, R. M.; Wells,
K. M.; Volante, R. P.; Reider, P. Tetrahedron Lett. 1994,
35, 673–676.
8. Xue, C.; He, X.; Corbett, R. L.; Roderick, J.; Wasserman,
Z. R.; Liu, R.; Jaffee, B. D.; Covington, M. B.; Qian, M.;
Trzaskos, J. M.; Newton, R. C.; Magolda, R. L.; Wexler,
R. R.; Decicco, C. P. J. Med. Chem. 2001, 44, 3351–3354.
9. Vloon, W. J.; Kruk, C.; Pandit, U. K.; Hofs, H. P.; McVie,
J. G. J. Med. Chem. 1987, 30, 20–24.
In summary, we have disclosed a series of optimized 2,5-
dibenzyl pyrrolidine based HIV protease inhibitors
which demonstrate improved antiviral potency through
modifications at P10. These analogs demonstrated com-
parable antiviral efficacy to currently marketed agents,
such as amprenavir.
10. Jilani, J. A. Eur. Patent Appl. EP 1231209 A1, 2002.
11. Toth, M. V.; Marshall, G. R. Int. J. Pept. Protein Res.
1990, 36, 544–550.
References and notes
12. Daluge, S. M.; Purifoy, D. J. M.; Savina, P. M.; St. Clair,
M. H.; Parry, N. R.; Dev, I. K.; Novak, P.; Ayers, K. M.;
Reardon, J. E. Antimicrob. Agents Chemother. 1994, 38,
1590–1603.
1. Palella, F. J., Jr.; Chmiel, J. S.; Moorman, A. C.;
Holmberg, S. D. AIDS (London England) 2002, 16,
1617–1626.