Novel inhibitors of HIV protease: Design, synthesis and biological evaluation of picomolar inhibitors containing cyclic P1/P2 scaffolds

Article abstract of DOI:10.1016/S0960-894X(00)00163-3

A novel series of HIV protease inhibitors containing cyclic P1/P2 scaffolds has been synthesized and evaluated for biological activity. The trans 3,5-dibenzyl-2-oxo pyrrolidinone ring system resulted in a 50 pM enzyme inhibitor against HIV protease in vitro when combined with an indanolamine derived P'-backbone. This compound also shows comparable activity to currently marketed drugs in the MT-4 cell-based antiviral assay. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00163-3

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1159±1162
Novel Inhibitors of HIV Protease:
Design, Synthesis and Biological Evaluation of Picomolar
Inhibitors Containing Cyclic P1/P2 Sca€olds
Andrew Spaltenstein, a,* Merrick R. Almond, a,y William J. Bock, a,{
Darryl G. Cleary, a,y Eric S. Fur®ne, a Richard J. Hazen, a Wieslaw M. Kazmierski, a
Francesco G. Salituro, b Roger D. Tung b and Lois L. Wright a
aGlaxo Wellcome Inc, 5 Moore Drive, RTP, NC 27709, USA
bVertex Pharmaceuticals, 3030 Waverly Street, Cambridge, MA 02139, USA
Received 17 December 1999; accepted 8 March 2000
AbstractÐA novel series of HIV protease inhibitors containing cyclic P1/P2 sca€olds has been synthesized and evaluated for bio-
logical activity. The trans 3,5-dibenzyl-2-oxo pyrrolidinone ring system resulted in a 50 pM enzyme inhibitor against HIV protease
in vitro when combined with an indanolamine derived P0-backbone. This compound also shows comparable activity to currently
marketed drugs in the MT-4 cell-based antiviral assay. # 2000 Elsevier Science Ltd. All rights reserved.
Highly active antiretroviral therapy (HAART) is the
currently recommended approach for the treatment of
HIV infection.1 HAART usually consists of a three-
drug cocktail containing at least two components that
inhibit di€erent viral targets. While some drug combi-
nations targeting exclusively the viral reverse tran-
scriptase have become available recently,2 protease
inhibitors (PIs) still constitute the cornerstone of most
treatment regimens. Such combination therapy has been
strikingly successful over the past 5 years, drastically
reducing AIDS related complications and deaths.1,3
However, some aspects of the current generation PIs
remain problematic. The generally large daily doses
required for complete viral suppression, can lead to
undesired side e€ects, limited patient adherence to ther-
apy, and high cost.4 Moreover, treatment failure due to
resistance to the currently available PIs is increasing,
which limits future treatment options because of cross
resistance among the other available PI-therapies.5 For
these reasons we have been engaged in a next generation
PI program aimed at identifying new, structurally
diverse agents with signi®cantly increased potency and
unique resistance pro®les.
Recently, we disclosed the design of a novel cyclic P1/P2
sca€old6 that was incorporated into the hydro-
xyethylene isostere backbone of amprenavir (Fig. 1).
Our initial success prompted us to investigate the use of
these and similar cyclic sub-units in combination with
other successful HIV PI backbones. In this paper we
explore the initial structure±activity relationship (SAR)
of a series of compounds that combine these novel cyclic
units with the P10±P30 indanolamine-based arrangement
®rst utilized by researchers at Merck.7
Our previous results in the sulfonamide series suggested
that the ®ve-membered 2-pyrrolidinone (A) sca€old
may be superior to conformationally somewhat more
¯exible six-membered systems such as pyrimidinones or
morpholinones. For this reason, we initially focused on
the smaller ring lactam as well as a set of new, isosteric
sca€olds such as 1l6-isothiazolidine-1,1-dione (B), 2-
imidazolidinone (C), and 1l6,2,5-thiadiazolidine-1,1-
dione (D) (Fig. 2).
Chemistry
The 2-pyrrolidinone A and 2-imidazolidinone C sca€olds
were synthesized according to previously published pro-
cedures.6 The synthesis of the 1l6-isothiazolidine-1,1-
dione B is outlined in Scheme 1. Cbz-phenylalanine was
converted to the protected amino aldehyde via LAH
*Corresponding author. Tel.: +1-919-483-2438; fax: +1-919-483-
6053; e-mail: as11513@glaxowellcome.com
yCurrent address: Triangle Pharmaceuticals, Durham, NC, USA
{Current address: Aventis, Tucson, AZ, USA
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00163-3

1160
A. Spaltenstein et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1159±1162
Figure 1.
lation, deprotection path used previously for the 2-
pyrrolidinone ring system.6
The 1l6,2,5-thiadiazolidine-1,1-dione sca€old D was
prepared as detailed in Scheme 2, via borane reduction of
l-phenylalanine benzylamide to the diamine 5, followed
by treatment with sulfamide in re¯uxing pyridine in the
presence of excess triethylamine/DMAP,13 a€ording 6 in
80% yield.
Figure 2.
Coupling of the P1/P2 cyclic sca€olds to the known
epoxide 714 was initially accomplished in DMF at 80ꢀC
using sodium hydride (Scheme 3). Unfortunately, these
harsh conditions took a heavy toll on the stereochemical
integrity of the P2-substituent of sca€olds A and B,
leading to equilibration (approximately 4:1 and 3:1 mix-
tures of trans to cis isomers for A and B, respectively).
While this lack of stereochemical control was not accep-
table for extended SAR studies, this route did, after
careful chromatographic separation, provide an enriched
sample of the cis isomer for comparison purposes. After
considerable experimentation, a general solution to this
epimerization problem was found with the use of the
highly hindered P4 phosphazene base.15 Thus combining
all ingredients (phosphazene base added last) at low
reduction of the corresponding Weinreb amide8 (75%,
two steps). Ole®nation, using the modi®ed Peterson
conditions published by Thompson et al.,9 a€orded the
desired allyl amine (70%), which was converted to
thioacetate 1 (79%) by radical induced sulfur addition
to the terminal ole®n position (thiolacetic acid/AIBN in
CCl4/hn=254 nm).10 Oxidation/chlorination with
chlorine gas in acetic acid/aq hydrochloric acid11 pro-
duced sulfonyl chloride 2 (95%).12 The Cbz protecting
group was removed with 48% HBr/AcOH and the
resulting material was cyclized to the desired 1l6-iso-
thiazolidine-1,1-dione 3 by treating with triethyl amine
in THF (58%, two steps). Introduction of the P2 ben-
zylsubstituent to a€ord 4 followed the protection, alky-
Scheme 1. (a) (i) EDC/MeNHOMe/TEA; (ii) LAH/THF; (b) (i)
TMSCH2MgBr; (ii) BF3ÂEt2O; (c) AcSH/AIBN/hn; (d) Cl2/AcOH/
HClaq; (e) HBr/AcOH; (f) TEA/THF; (g) Boc2O/THF; (h) LDA/BnBr
/THF; (i) TFA/CH2Cl2.
Scheme 2. (a) EDC/BnNH2/TEA; (b) TFA; (c) BH3ÂMe2S/THF; (d)
SO2(NH2)2/pyridine/TEA/DMAP, r¯.
Scheme 3.

A. Spaltenstein et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1159±1162
Table 1. Inhibition of HIV protease and antiviral activity of 8a±c Table 2. Biological activity of various P1/P2 sca€olds
1161
X
Ki (nM)
IC50 (mM)
Entry
R1
R2
Conf. OH
K (nM)a
IC50 (mM)b
8a
H
H
Bn
H
H
H
H
Bn
[S]
[R]
[S]
[S]
330
4800
0.05
>0.5c
>50
>50
0.3
n/ad
0.05
0.3
epi-8a
8b
8c
Amprenavir
Ritonavir
Nel®navir
Indinavir
0.04
0.02
0.01
0.07
0.04
0.10
0.21
0.05
0.04
0.02
16
n/a
0.3
1.5
Saquinavir
aKi, enzyme inhibition constant.20
bIC50, antiviral inhibition in MT-4 cell culture.21
cRepresents lower limit estimate.19
dn/a, not available.
1.2
0.8
temperature and carefully warming to 0 ꢀC over 2 h
a€orded the pure trans alkylation product 8b in 65%
yield with no trace of the undesired cis isomer 8c.20
Results and Discussion
In order to establish the preferred stereochemistry of the
central hydroxy functionality for these new backbones,
we ®rst synthesized the truncated cyclic sca€old (without
P2 substituent) and coupled it to both epoxide isomers17
7 to a€ord 8a and its backbone OH-epimer. Table 1
illustrates that there is at least a 10-fold preference of the
S over the R con®guration in terms of enzyme inhibition.
This observation is consistent with the data reported for
the related Indinavir backbone.18 Encouraged by the
considerable potency of these P2-truncated inhibitors,
we predicted substantial room for exploration both in
terms of a suitable P2 substituent as well as regarding the
exact nature of the cyclic sca€old. As anticipated, intro-
duction of a P2 benzyl substituent led to a dramatic
increase in enzyme potency, producing a compound with
50 pM activity against HIV protease. This compound
also showed 300 nM antiviral activity in a MT-4 cell-
based assay. Table 1 furthermore con®rms predictions
derived from molecular modeling that the preferred
con®guration of the P2 substituent relative to the 5-
benzyl substituent on the 2-pyrrolidinone ring system is
trans (8b vs 8c).
this interaction. Based on molecular modeling, the
decreased potency of the 2-imidazolidinone (C) and
1l6,2,5-thiadiazolidine-1,1-dione (D) ring systems was
less surprising. In both cases, the P2 substituent is
attached to a partially sp2 hybridized nitrogen atom
which projects it into the corresponding active site
pocket at a similar angle to that of the less active cis
isomer in the 2-pyrrolidinone series. The 10-fold loss in
potency for these isosteres is therefore consistent with
the results obtained for 8c.
In summary, we have discovered a new and highly
potent cyclic HIV protease inhibitor sca€old. Lead
compound 8b, is roughly equipotent with the currently
marketed HIV protease inhibitors and modeling studies
indicate the potential for further optimization in the P1,
P2 and P10 positions. We have initiated work in all of
these areas and will report our progress in future
accounts.
Attempts to optimize and simplify the cyclic P1/P2 scaf-
folds with isosteric replacements in the 2- and 3-positions
proved more challenging (Table 2). The replacement of
the carbonyl functionality by a sulfonyl group led to a
300-fold increase in Ki value. While the carbonyl oxygen
has been shown6 to participate in a crucial hydrogen
bond to the `¯ap water' in the active site, we had hoped
that the pair of sp2 hybridized oxygens in the sulfonyl
replacement might be able to maintain or even improve
References and Notes
1. CDC Morbidity and Mortality Weekly Reports 1999, 48, 459.
2. For example see: Opravil, M.; Hirschel, B.; Lazzarin, A.;
Furrer, H. J.; Chave, J. P.; Bern, E.; 39th Interscience Con-

1162
A. Spaltenstein et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1159±1162
ferences on Antimicrobial Agents and Chemotherapy (ICAAC),
San Francisco, 1999, Abstract #283.
Scholey, K. J. Med. Chem. 1994, 37, 3023. We found that the
addition of excess triethylamine/DMAP is critical for this
transformation. Pyridine alone a€orded only trace amounts of
the desired product after extended reaction times..
14. Askin, D.; Eng, K. K.; Rossen, K.; Purick, R. M.; Wells, K.
M.; Volante, R. P.; Reider, P. J. Tetrahedron Lett. 1994, 35, 673.
15. (a) Schwesinger, R.; Schlemper, H. Angew. Chem. 1987, 99,
1212. (b) Pietzonka, T.; Seebach, D. Chem. Ber. 1991, 124,
1837. P4 phosphazene base is commercially available from
Fluka.
16. As evidenced by TLC and NMR analysis of the crude
reaction mixtures.
17. Askin, D.; Eng, K. K.; Rossen, K.; Purick, R. M.; Wells,
K. M.; Volante, R. P.; Reider, P. J. Tetrahedron Lett. 1994, 35,
673.
3. CDC Morbidity and Mortality Weekly Reports 1997, 46, 861.
4. (a) Kakuda, T. N.; Strubble, K. A.; Piscitelli, S. C. J.
Health-Syst. Pharm. 1998, 55, 233. (b) Hoetelmans, (c) R. M.
W.; Meehirst, P. L.; Mulder, J. W.; Burger, D. M.; Koks, C.
H. W.; Beijnen, J. H. Pharmacy World Sci. 1997, 19, 159.
5. (a) Molla, A.; Richard G. G.; Sun, E.; Kempf, D. J. Anti-
viral Res. 1998, 39, 1. (b) Fur®ne, E. S. In Handbook of
Experimental Pharmacology; von der Helm, K. Korant, B.
(Eds.); in press.
6. Salituro, F. G.; Baker, C. T.; Court, J. J.; Deininger, D. D.;
Kim, E. E.; Li, B.; Novak, P. M.; Rao, B. G.; Pazhanisamy,
S.; Porter, M. D.; Schairer, W. C.; Tung, R. D. Bioorg. Med.
Chem. Lett. 1998, 8, 3637.
7. Lin, J. H.; Ostovic, D.; Vacca, J. P. Pharmaceut. Biotechnol.
1998, 11233.
18. Vacca, J. P.; Guare, J. P.; deSolma, S. J.; Sanders, W. M.;
Giuliani, E. A.; Young, S. D.; Darke, P. L.; Zugay, J.; Sigal, I.
S.; Schleif, W. A.; Quintero, J. C.; Emini, E. A.; Andreson, P.
S.; Hu€, J. R. J. Med. Chem. 1991, 34, 1225.
19. Determination of an accurate Ki for this compound was
not possible because all of the enzyme inhibition by this
sample could be accounted for by the presence of a small
amount (<10%) of the corresponding trans isomer 8b.
Therefore, we assumed that the Ki for the cis isomer must be
at least 10±100-fold greater than that for the trans isomer.
20. Toth, M. V.; Marshall, G. R. Int. J. Pep. Protein Res.
1990, 36, 544.
8. Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815.
9. Thompson, W. J.; Fitzgerald, P. M.; Holloway, M. K.;
Emini, E. A.; Darke, P. L.; Schwering, J. E.; Homnick, C. F.;
Nunberg, J.; Springer, J. P.; Hu€, J. R. J. Med. Chem. 1992,
35, 1685.
10. (a) Cunneen, J. I. J. Chem. Soc. 1947, 134. (b) Pale-Gros-
demange, C.; Simon, E. S.; Prime, K. L.; Whitesides, G. M. J.
Am. Chem. Soc. 1991, 113, 12.
11. Stewart, J. M.; Cordts, H. P. J. Am. Chem. Soc. 1952, 74, 5880.
12. Temperature control (0 ꢀC) was found to be important in
order to avoid chlorination of the phenyl ring.
13. Castro, J. L.; Baker, R.; Guiblin, A. R.; Hobbs, S. C.;
Jenkins, M. R.; Russell, M. G. N.; Beer, M. S.; Stanton, J. A.;
21. 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.