Inhibitors of hepatitis C virus NS3·4A protease. Part 3: P 2 proline variants

Article abstract of DOI:10.1016/j.bmcl.2004.01.078

We recently described the identification of an optimized α-ketoamide warhead for our series of HCV NS3·4A inhibitors. We report herein a series of HCV protease inhibitors incorporating 3-alkyl-substituted prolines in P₂. These compounds show ex

Full text of DOI:10.1016/j.bmcl.2004.01.078

Bioorganic & Medicinal Chemistry Letters 14 (2004) 1939–1942
.
Inhibitors of hepatitis C virus NS3 4A protease. Part 3:
P2 proline variants
Robert B. Perni,* Luc J. Farmer,* Kevin M. Cottrell, John J. Court,
Lawrence F. Courtney, David D. Deininger, Cynthia A. Gates, Scott L. Harbeson,
Joseph L. Kim, Chao Lin, Kai Lin, Yu-Ping Luong, John P. Maxwell, Mark A. Murcko,
Janos Pitlik, B. Govinda Rao, Wayne C. Schairer, Roger D. Tung, John H. Van Drie,
Keith Wilson and John A. Thomson
Vertex Pharmaceuticals Inc., 130 Waverly Street, Cambridge, MA 02139, USA
Received 26 November 2003; accepted 26 January 2004
.
Abstract—We recently described the identification of an optimized a-ketoamide warhead for our series of HCV NS3 4Ainhibitors.
We report herein a series of HCV protease inhibitors incorporating 3-alkyl-substituted prolines in P2. These compounds show
exceptional enzymatic and cellular potency given their relatively small size. The marked enhancement of activity of these 3-sub-
stituted proline derivatives relative to previously reported 4-hydroxyproline derivatives constitutes additional evidence for the
.
importance of the S2 binding pocket as the defining pharmacophore for inhibition of the NS3 4Aenzyme.
# 2004 Elsevier Ltd. All rights reserved.
.
The hepatitis C NS3 4Aprotease is an important and
important factor to address. Recently reported series of
compounds displayed molecular weights ranging from
700–1000,3 clearly above the optimal range for most
orally delivered drugs. The enzyme S2 pocket has been
demonstrated to be important for inhibitor recognition,
and bulky substituents have been identified as optimal
at that position, contributing to the overall high mole-
cular weight of the inhibitors.3 We herein report on a
attractive target for developing anti-viral therapies for
HCV.1 We recently reported on the design of a series of
.
potent NS3 4Aprotease inhibitors based on a tetrapep-
tide scaffold.2 Our design efforts have continued to
focus on optimizing potency while improving the phy-
sicochemical properties of these molecules to enable
convenient oral dosing of the inhibitor. Size is an
Keywords: Hepatitis C; Protease inhibitors.
* Corresponding author. Tel.: +1-617-444-6237; fax: +1-617-444-6766; e-mail: robert_perni@vrtx.com
0960-894X/$ - see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.01.078

1940
R. B. Perni et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1939–1942
series of potent tetrapeptide inhibitors utilizing a 3-alkyl
proline P2 moiety that retains significant binding
potency despite their relatively small molecular size.
3-position methyl group to be subtly tilted away from
the active site due to the position of the histidine. The
initial interpretation of density near catalytic Asp 81
and His 139 as a structural water appears to be incorrect
in view of this result. The similar enzymatic activity of 2
relative to 1 (1.4 mM versus 0.92 mM respectively,
cLogP=4.4) provided the impetus to explore this series
of 3-substituted proline analogues in detail.
Analogues derived from 4-hydroxyproline such as 1 and
32 provide good binding to the S2 pocket, but result in
inhibitors possessing high cLogP values (e.g., 5.4 and
6.4 for 1 and 3, respectively) and additional metaboli-
cally vulnerable functionality. Inhibitors such as 3,
which possess an obligatory charge have thus far dis-
played reduced cellular activity.2b
The syntheses of 3-functionalized prolines 9a–g are
described in Schemes 1 and 2. Cuprate addition,6 of
These observations led us to inspect other possibilities
for P2 proline substitution that would not possess the
liabilities afforded by 4-hydroxyproline derivatives.
During our analysis of a crystallographic co-complex of
.
1 with the NS3 4Aprotein we observed poorly defined
electron density in the region where we expected to find
the catalytic histidine. We initially interpreted the elec-
tron density map of this area close to the aspartate (Asp
81) and serine (Ser 139), as representing a structural
water molecule. Additional refinement of the structure is
shown in Figure 1, with a putative bound water molecule
(green sphere). Molecular modeling studies based on the
X-ray structures of 4-hydroxyproline-substituted inhi-
bitors indicated that a vector from the 3-position of the
P2 proline ring might allow access to the active site via a
relatively small substituent (also shown in Fig. 1). We
hypothesized that the putative crystallographically-defined
water molecule could then be displaced with a 1–4 car-
bon group extending from the a-face of the proline 3-
position, potentially boosting the binding affinity by ꢀ2
kcal/mol from the entropic gain from displacing that
bound water4 thereby compensating for affinity lost
upon removal of the 4-substituent. To test this hypoth-
esis, we prepared 2 from commercially available 3-
methyl proline as shown in Scheme 4.5 The affinity of 2
was determined to be 1.4 mM, consistent with expecta-
tions. The yellow structure in Figure 2 depicts the
modeled orientation of the bound inhibitor if the histi-
dine were displaced from its normal orientation. The
actual X-ray structure of the co-complex showed that
the ligand bound similarly to the modeled conformation
with the catalytic histidine in the normal conformation
(green structure in Fig. 2). In this structure we found the
.
Figure 2. X-ray structure of 2 (green) bound to the tNS3 4Aprotein
relative to the predicted model structure of 2 from Figure 1 (yellow).
.
Scheme 1. (a) 2-vinyl, 2-propenyl or 2-butenyl-MgBr, CuBr DMS, ether,
ꢁ20 ꢂC then TMSCl, ꢁ78 ꢂC (65%, 73%, 84%); (b) 10% Pd–C, H2, 1
atm, EtOH (90%, 92%, 89%); (c) R00ZnBr, THF, ꢁ30 ꢂC, BF3OEt2
then TMSCl (64%, 40%, 37%); (d) PhCH(Li)SPh, BuLi, TMEDA,
ꢁ78 ꢂC (45%); (e) Ra-Ni, acetone/water (1:1), reflux, 12 h (83%).
Scheme 2. (a) HCl gas, EtOAc, ꢁ20 ꢂC (80–90%); (b) LAH, THF,
reflux (85–90%); (c) CbzCl, K2CO3, THF/H2O (1:1) (60–85%); (d)
Jones, acetone, (70–80%); (e) isobutylene, H2SO4 cat., DCM (67–
85%); (f) 10% Pd–C, H2, 1 atm, EtOAc (90–95%).
.
Figure 1. Predicted structure of 2 (white) bound to the tNS3 4A
protein relative to X-ray structure of 1 (purple).

R. B. Perni et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1939–1942
1941
Scheme 3. (a) PtO2, EtOH/AcOH/H2O (7:2:1), H2, 50 psi; (b) CbzCl, Na2CO3, acetone/H2O (1:1) (90%, two steps); (c) isobutylene, H2SO4 cat.,
CH2Cl2; (d) 10% Pd–C, H2, 1 atm, EtOH (84%, two steps).
Scheme 4. (a) See ref 5.
vinyl, 2-propenyl and 2-butenylmagnesium bromide in
the presence of copper bromide dimethylsulfide complex
and trimethylsilyl chloride on pyroglutaminol7 4 pro-
vided, after hydrogenation, compounds 5a, b and c. The
1,4-addition of organozinc to 4 afforded intermediates
6a, b, and c.8 Finally, the 3-benzyl pyroglutaminol 7
was prepared via the addition of 2-lithio benzylphenyl
sulfide9 to compound 4 followed by Raney-nickel
reduction.
The 3-cyclohexylproline derivative, 12, was prepared
according to Scheme 3 in four steps from commercially
available 3-phenylproline. Compounds 13–15 were pre-
pared according to literature procedures.11,12
The final tetrapeptides 16–28 (Tables 1 and 2) were then
elaborated using the same methodology as previously
reported5 (Scheme 4).
We produced our initial set of inhibitors (16–20) on a
0
scaffold containing a phenylalanine P1 terminus in
The 3-substituted pyroglutaminols 5a–c, 6a–c and 7
were converted to proline derivatives 9a–g in six steps
(Scheme 2). Tandem removal of the silyl group and t-
butyl carbamate in the presence of anhydrous hydro-
chloric acid in ethyl acetate generated pyroglutaminol
intermediates, which, upon LAH reduction in refluxing
THF, provided hydroxymethyl pyrrolidine 8a–g. Pro-
tection of the 3-alkyl hydroxymethyl pyrrolidines as the
benzylic carbamate10,11 with benzyloxycarbonyl chlo-
ride in the presence of K2CO3 followed by Jones oxida-
tion gave the corresponding Cbz-protected prolines that
were esterified with isobutylene in dichloromethane with
a catalytic amount of sulfuric acid. The resulting carba-
mate esters were hydrogenated to give the pivotal pro-
line ester scaffolds 9a–g.
order to maximize binding potency at the possible
expense of cellular activity. With the exception of 18
these closely related inhibitors possess essentially iden-
.
tical activity versus the NS3 4Adimer ( Table 1). It is
unclear why the n-propyl derivative, 18, is 5- to 10-fold
less potent than the other compounds. This result,
obviously, does not appear related to steric demand given
Table 2. Enzymatic and cellular results from optimized 3-substituted
proline derivatives
Table 1. 3-Substituted proline structure–activity summary
R2
Ki (mM)a
IC50 (mM)b
CC50 (mM)c
21
22
23
24
25
26
27
28
29
CH2CH3
CH(CH3)2
cyclohexyl
CH(CH3)CH2CH3
CH(CH2CH3)2
cyclopentyl
C(CH3)3
0.15
0.09
0.10
0.10
0.33
0.25
0.46
0.31
1.2
0.45
0.91
0.76
—
>3
2.9
>3
1.1
—
30
59
28
—
74
43
74
22
—
R2
Ki (mM)a
CH2C6H5
H
16
17
18
19
20
CH2CH3
CH2CH=CH2
CH2CH2CH3
CH(CH3)CH2CH3
CO2CH3
0.22
0.21
1.2
0.12
0.18
a HPLC based enzymatic assay performed as described in ref 2b.
b Cellular HCV replicon13 assay performed as described in ref 2b.
c Cytotoxicity of the compounds (CC50) was determined by an MTS-
based cell viability assay.
a HPLC based enzymatic assay performed as described in ref 2b.

1942
R. B. Perni et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1939–1942
that other compounds containing small substituents on
the 3-position of the P2 proline were nearly as potent at
many similar compounds with large 4-hydroxyproline
groups (e.g., 19 versus 3). Nevertheless, these initial
results were extremely encouraging.
Chemistry; King F. D., Oxford, A. W., Eds. Elsevier
Science B.V., 2002, p 215.
2. (a) Perni, R. B.; Britt, S. D.; Court, J. J.; Courtney, L. F.;
Deininger, D. D.; Farmer, L. J.; Gates, C. A.; Harbeson,
S. L.; Kim, J. L.; Landro, J. A.; Levin, R. B.; Luong,
Y.-P.; O’Malley, E. T.; Pitlik, J.; Rao, B. G.; Schairer,
W. C.; Thomson, J. A.; Tung, R. D.; Van Drie, J. H.;
Wei, Y. Bioorg. Med. Chem. Lett. 2003, 13, 4059. (b) Perni,
R. B.; Pitlik, J.; Britt, S. D.; Court, J. C.; Courtney, L. F.;
Deininger, D. D.; Farmer, L. J.; Gates, C. A.; Harbeson,
S. L.; Lin, C.; Lin, K.; Moon, Y.-C.; Luong, Y.-P.;
O’Malley, E. T.; Rao, B. G.; Thomson, J. A.; Tung, R. D.;
Van Drie, J. H.; Wei, Y. Bioorg. Med. Chem. Lett. 2004,
14, 1441.
Having developed smaller P2 containing inhibitors that
possess significant enzymatic potency we embarked on a
broader survey of potential non-polar 3-position analo-
gues on an optimized scaffold that bears no charge and
would presumably be better able to penetrate cells.
Those results are presented in Table 2.
3. (a) For recent examples see: Johansson, A.; Poliakov, A.;
Akerblom, E.; Wiklund, K.; Lindeberg, G.; Winiwarter,
S.; Danielson, U. H.; Samuelsson, B.; Hallberg, A.
Bioorg. Med. Chem. 2003, 11, 2551. (b) Beevers, R.; Carr,
M. G.; Jones, P. S.; Jordon, S.; Kay, P. B.; Lazell, R. C.;
Raynham, T. M. Bioorg. Med. Chem. Lett. 2002, 12, 641.
4. Dunitz, J. D. Science 1994, 670, 264.
5. Perni, R.; Court, J.; O’Malley, E.; Rao, B. G., PCT WO
01/74768 2001.
6. Herdeis, C.; Huberman, H. P. Tetrahedron: Asymmetry
1992, 3, 1213.
7. Corey, E. J.; Jautelat, M.; Oppolzer, W. Tetrahedron Lett.
1967, 8, 2325.
8. Hanson, M. W.; Rieke, R. D. J. Am. Chem. Soc. 1995,
117, 10775.
9. Hashimoto, M.; Hashimoto, K.; Shirahama, H. Tetra-
hedron 1996, 52, 1931.
This series of proline variants, 21–29, was based on a
02b
14
scaffold comprised of optimized P1 and P3, P4 sub-
stituents. The resulting inhibitors demonstrated class-
wide binding potency relative to carboxy-terminal inhi-
bitors, 16–20. Again, there is little spread in the data
with inhibitory constants ranging from 0.09 mM to 0.46
mM with no visible steric trends in this, albeit limited,
series. As hoped significant cellular activity13 is evident
in this series. Several compounds exhibit IC50 values of
<1 mM in an HCV replicon assay. While 21 incorpo-
rates the smallest proline substituent of the series shown
in Table 2, an ethyl group, it is among the most potent
compounds, along with 22 and 23, in the replicon cel-
lular assay (IC50=0.45 mM, cLogP=5.5) indicating that
size alone is not the major determinant of activity. The
unsubstituted proline P2 derivative, 29 Ki=1.2 mM),
provides a baseline comparator that clearly demon-
strates the advantage of incorporating a substituent at
the 3-position relative to no substitution. The combin-
ation of a small proline moiety and optimized P3–P4
substituents results in a molecule with good cellular
activity probably due to an improved balance between
enzymatic binding affinity and overall physicochemical
properties.
10. Thottathil, J. K.; Moniot, J. L.; Mueller, R. H.; Wong,
M. K. Y.; Kissick, T. P. J. Org. Chem. 1986, 51, 3140.
11. Holladay, M. W.; Lin, W. L.; May, C. S.; Garvey, D. S.;
Witte, D. G.; Miller, T. R.; Wolfram, C. A. W.; Nadzan,
A . M.J. Med. Chem. 1991, 34, 457.
12. Parr, I. B.; Boehlein, S. K.; Dribben, A. B.; Schuster,
S. M.; Richards, N. G. J. J. Med. Chem. 1996, 39, 2367.
13. Lohmann, V.; Korner, F.; Koch, J.-O.; Herian, U.;
Theilmann, L.; Bartenschlager, R. Science 1999, 285, 110.
14. Increasing lipophilicity of the groups at the P3 and P4
positions using hydrocarbon-substituted amino acids
optimized the hydrophobic interaction with the S3 and S4
pockets. SAR results indicate that, in general, cyclohexyl
glycine is optimal at P4 and t-butyl glycine is optimal
at P3. These results are consistent with those pre-
viously reported in: Ingallinella, P.; Altamura, S.; Bianchi,
E.; Taliani, M.; Ingenito, R.; Cortese, R.; De Francesco,
R.; Steinkuhler, C.; Pessi, A. Biochemistry 1998, 37,
8906.
15. (a) The results presented in this manuscript form the basis
for the design of the fused-bicyclic proline derivatives
described in: Yip, Y.; Victor, F.; Lamar, J.; Johnson,
R.; Wang, Q. M.; Barket, D.; Glass, J.; Jin, L.; Liu, L.;
Venable, D.; Wakulchik, M.; Xie, C.; Heinz, B.; Villar-
real, E.; Colacino, J.; Yumibe, N.; Tebbe, M.; Munroe, J.;
Chen, S.-H. Bioorg. Med. Chem. Lett. 2004, 14, 251.
(b) Victor, F.; Lamar, J.; Snyder, N.; Yip, Y.; Guo, D.;
Yumibe, N.; Johnson, R.; Wang, Q. M.; Glass, J. I.;
Chen, S.-H. Bioorg. Med. Chem. Lett. 2004, 14, 257.
Although the original hypothesis of targeting a sub-
stituent from the P2 proline into the NS3 catalytic site
was not possible as initially formulated, exploiting the
subtleties of the enzyme inhibitor interactions at this
site led to the discovery of a series potent inhibitors with
compact P2 substituents and improved physicochemical
parameters relative to the corresponding 4-hydroxy-
proline-derived analogues. Further investigation of
this series is in progress and will be reported in due
course.15
References and notes
1. Narjes, F.; Koch, U.; Steinkuhler, C. Exp. Opin. Invest.
Drugs 2003, 12, 153. For a comprehensive recent review
see: Perni, R. B.; Kwong, A. D. In Progress in Medicinal