Structure-activity and crystallographic analysis of a new class of non- amide-based thrombin inhibitor

Article abstract of DOI:10.1016/S0960-894X(99)00617-4

The structure-activity relationships of a novel series of non-amide- based thrombin inhibitors are described. Exploration of the P2 and the aryl binding region for this series has identified optimal groups for achieving nanomolar potency. The binding modes of these optimal groups have been confirmed by X-ray structural analysis.

Full text of DOI:10.1016/S0960-894X(99)00617-4

Bioorganic & Medicinal Chemistry Letters 10 (2000) 79±82
Structure±Activity and Crystallographic Analysis of a New Class
of Non-amide-Based Thrombin Inhibitor
Tianbao Lu, Richard M. Soll, Carl R. Illig, Roger Bone, Larry Murphy, John Spurlino,
F. Raymond Salemme and Bruce E. Tomczuk*
3-Dimensional Pharmaceutical, Inc., Eagleview Corporate Center, Exton, PA 19341, USA
Received 3 August 1999; accepted 28 October 1999
AbstractÐThe structure±activity relationships of a novel series of non-amide-based thrombin inhibitors are described. Exploration
of the P2 and the aryl binding region for this series has identi®ed optimal groups for achieving nanomolar potency. The binding
modes of these optimal groups have been con®rmed by X-ray structural analysis. # 1999 Elsevier Science Ltd. All rights reserved.
The serine protease, thrombin, occupies a pivotal role in
the coagulation cascade. Thrombin is involved in (1) the
cleavage of ®brinogen to release ®brin, (2) the cross-
linking of ®brin to form clots, (3) the stimulation of
platelet aggregation, another major component of clots,
and (4) the autoampli®cation of the coagulation cascade
to produce additional active thrombin.1±3 There are cur-
rently several active-site thrombin inhibitors (1±4) which
are being pursued, with argatroban (Novastan, 1) being
the furthest in phase III development. Closely related
derivatives of argatroban, such as the benzamidine UK
156406 (2)4 and the amidrazone derivative LB-30057 (3)5
are aminoacid-based templates. The peptidomimetic
pyrazinone L-375378 (4) has also been reported.6 We
were intrigued prior to these literature reports of a non-
peptide-based series exempli®ed by compound 5.7 This
compound possessed impressive potency despite its
structural simplicity, denoted by the absence of an elec-
trophilic functionality, an aminoacid framework, and
chirality. In a previous report,8 we reported on 12, an
amidinopiperidine derivative related to compound 5.
Herein, we expand on the structure±activity relation-
ships and crystallographic analysis of substitutions in
the central phenyl template and in the arylsulfonate
group of 11.
sodium bicarbonate with arylsulfonyl chlorides. The
piperazine portion of the targeted compounds is derived
from isonipecotic acid (7), which is N-protected (di-t-
butyl dicarbonate (1 equiv), NaHCO3 (2 equiv), diox-
ane:H2O (1:1), 24 h (91%)) and reduced to the alcohol
ꢀ
.
8 (BH3 THF (1 equiv), THF, 0 to 25 C (84%)). Alcohol
8 was mesylated (MsCl (1 equiv), NEt3, CH2Cl2 (93%))
and coupled to the sodium salt of 6 (generated with
NaH (1.1 equiv), DMF) at 50 ꢀC for 3 h (69%) to pro-
duce piperidines 9. Alternatively, Mitsunobu coupling
(PPh3 (1.5 equiv), DEAD (1.5 equiv), THF, (60±85%))
was used between phenol 6 and alcohol 8 to give 9. For
analogue 30, containing a hydroxymethyl group on the
central phenyl template, the piperidine intermediate (9)
of 27 was reduced (LiBH4 (1.5 equiv), THF (61%)).
Deprotection (4N HCl in dioxane, 2 h (79%)) and
treatment with aminoiminomethanesulfonic acid (2
equiv, NEt3, DMF, 12 h (49%)) provided amidino-
piperidines 10. Nitroaryl derivatives 14, 20 and 23
were reduced to anilino derivatives 15, 21 and 24 by
catalytic hydrogenation using Pd/C as a catalyst at
room temperature.
The synthesis of these amidinopiperidines is straight-
forward, as exempli®ed in Scheme 1.9 The symmetrical
resorcinols are monosulfonated (6) in good yields by
using a biphasic mixture of diethyl ether and saturated
Keywords: anticoagulants; enzyme inhibitors; substituent e€ects;
X-ray crystal structures.
*Corresponding author. Tel.: +1-610-458-6058; fax: +1-610-458-
8249; e-mail: tomczuk@3dp.com
0960-894X/00/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(99)00617-4

80
T. Lu et al. / Bioorg. Med. Chem. Lett. 10 (2000) 79±82
Table 2. E€ects of substitutions on the central template
Compound
X
Ki thrombin (nM)
12
26
27
28
29
30
31
CH3
H
4.6‹1
190‹11
3700‹200
44‹2
39‹2
490‹52
30‹1
CO2CH3
OCH3
CH2CH3
CH2OH
Cl
Scheme 1. Synthesis of amidinopiperidines.
Compounds 11±31 were evaluated for the inhibition of
thrombin using standard chromogenic assays (Tables 1
and 2).10 There are several structure±activity relation-
ships in Table 1 that should be noted. First, the para-
substitutions (23, 24) are the least tolerated compared to
their corresponding meta- and ortho-substituents. Sec-
ond, the meta-substituents (17, 18, 20, 21) are less
potent than the corresponding ortho-substituents (12±
15), with the exception of the 3-amino derivative (21)
which was slightly more potent than 15. Third, the
ortho-substituents (12±16), in general, are the most
potent compounds, with the exception of the 2-amino
(15) derivative, which was the least potent compound in
this group. Substitution of the 2 position with a small
sampling of groups with a wide range of steric, hydro-
philic, and electronic properties did not appear to be a
major determinant of potency. As observed in X-ray
structures (Fig. 1), the phenyl ring is found in a per-
pendicular stacking arrangement with Trp215 as found
in this study and others.11 The ortho-substituent is
invariably directed away from the enzyme surface and is
solvent exposed which would predict the non-dis-
criminatory characteristic of this SAR.
in this region as exempli®ed by the 740-fold loss of
potency for the methoxycarbonyl group (27). On the
other hand, sterically small groups such as methyl (12),
ethyl (29), methoxy (28) and chloro (31) appear to be
tolerated (5±44 nM), although the methyl group appears
optimal. It is noteworthy to observe that minimal steric
occupancy at this position, such as hydrogen (26) resul-
ted in a 50-fold loss of potency versus 12. The e€ect of
the increased volume of the substitution under the six-
ties loop in the thrombin-inhibitor complex is clearly
seen in the X-ray crystal structures of 12, 26 and 29.13
There is almost no structural rearrangement of either
the protein or inhibitor position seen with this substitu-
tion, although the cavity of the P2 pocket is ®lled by the
additional volume (Fig. 1). Homologation of the methyl
to an ethyl group on the central phenyl template (29)
resulted in a decrease in potency (eightfold). Although
additional volume is ®lled, the energy required to
maintain the dihedral angle along with increased van
der Waals overlap more than o€sets the gain from the
reduced surface exposure (Fig. 2). The close contact of
the terminal CH3 of the ethyl group (3.3 A) from the
imidazole ring of His57 accounts for the majority of the
unfavorable energy. Additionally, the preferred angle
for the dihedral angle of the ethyl substitution is 90ꢀ
as compared to the observed dihedral angle of 55ꢀ,
which may also contribute to the unfavorable energy.
The central phenyl template binds to the S2 region of
thrombin (Table 2).12 There are severe steric constraints
Table 1. E€ects of arylsulfonate substitution
Compound
R1
R2
R3
Ki thrombin (nM)
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
H
Cl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
NO2
NH2
H
32‹2
4.6‹1
14‹1
16‹2
52‹2
28‹1
39‹2
445‹39
36‹2
115‹11
32‹2
22‹1
327‹14
399‹26
35‹2
CF3
NO2
NH2
CO2CH3
H
H
Cl
H
H
H
H
Cl
CF3
CH3
NO2
NH2
Cl
H
H
H
H
Figure 1. X-ray structure of methyl (12, blue) and hydrogen (26,
orange) on central phenyl template in thrombin.
-CHˆCH±CHˆCH-

T. Lu et al. / Bioorg. Med. Chem. Lett. 10 (2000) 79±82
81
are solvent exposed whereas substituents in the 3 and 4
position may clash with the protein speci®city pocket.
The 10-fold variance in potency within ortho substitu-
tion may re¯ect a combination of proximity to the six-
ties loop as well as optimal edge to face interaction with
tryptophan 215.
Although there have been several reports on the use of
similar aryl sca€olds in thrombin inhibitor design, lim-
ited SAR has been reported. Optimal for activity is a
rigid guanidine backbone,8 methyl group substitution
on the orcinol, and ortho-substitution on the aryl sul-
fonate. The present data, together with our earlier work,
completes the SAR on this series of potent non-peptidic,
guanidino-containing compounds as it relates to in vitro
thrombin inhibition.
Figure 2. X-ray structures of methyl (12, green) and ethyl (29, purple)
on central phenyl template in thrombin.
Acknowledgement
Modeling studies of 27 in which the methoxycarbonyl
group was constrained to a co-planar arrangement with
the phenyl ring (presumed energy minimum) indicated a
severe steric interaction with the S2 pocket of thrombin.
Even when twisted 90ꢀ out of the phenyl plane, one of
the oxygens of the methoxycarbonyl group is still less
than 2.5 A from the imidazole ring of His57, which is
held in place by a hydrogen bonding network. The
hydrogen-donating hydroxymethylene group (30) resul-
ted in a 100-fold loss of potency versus 12 despite the
potential for a water-mediated hydrogen bond with the
enzyme in this region.
We wish to thank Stephen Eisennagel for MS and
HPLC determinations as well as Mike Kolpak of US
Bioscience for NMR spectra.
References and Notes
1. Colman, R. W.; Marder, V. J.; Salzman, E. W.; Hirsch, J. In
Hemostasis and Thrombosis: Basic Principles and Clinical
Practice; 3rd ed.; Coleman, R. W.; Hirsch, J.; Marder, V. J.;
Salzman, E. W.; Eds.; J. B. Lippincott: Philadelphia, 1994, p 3.
2. Colman, R.W.; Cook, J.; Niewiarowski, S. In Hemostasis
and Thrombosis: Basic Principles and Clinical Practice; 3rd ed.;
Coleman, R. W.; Hirsch, J.; Marder, V. J.; Salzman, E. W.;
Eds.; J. B. Lippincott: Philadelphia, 1994, p. 508.
3. Mann, K. G. In Hemostasis and Thrombosis: Basic Princi-
ples and Clinical Practice; 3rd ed.; Coleman, R. W.; Hirsch, J.;
Marder, V. J.; Salzman, E. W.; Eds.; J. B. Lippincott: Phila-
delphia, 1994, p 184.
4. Allen, M.; Abel, S. M.; Barber, C. G.; Cussans, N. J.;
Danilewicz, J. C.; Ellis, D.; Hawkeswood, E.; Herron, M.;
Holland, S.; Fox, D. N. A.; James, K.; Kobylecki, R. J.;
Overington, J. P.; Pandit, J.; Parmar, H.; Powling, M. J.;
Rance, D. J.; Taylor, W.; Shepperson, N. B. Abstracts of
Papers, 215th National Meeting American Chemical Society,
Dallas, TX, 1998; American Chemical Society: Washington,
DC, 1998; MEDI 200.
5. Lee, K.; Hwang, S. Y.; Hong, S.; Hong, C. Y.; Lee, C.-S.;
Shin, Y.; Kim, S.; Yun, M.; Yoo, Y. J.; Kang, M.; Oh, Y. S.
Bioorg. Med. Chem. 1998, 6, 869.
6. Sanderson, P. E. J.; Lyle, T. A.; Cutrona, K. J.; Dyer, D. L.;
Dorsey, B. D.; McDonough, C. M.; Naylor-Olsen, A. M.;
Chen, I.-W.; Chen, Z.; Cook, J. J.; Cooper, C. M.; Gardell, S.
J.; Hare, T. R.; Krueger, J. A.; Lewis, S. D.; Lin, J. H.; Lucas,
B. J., Jr.; Lyle, E. A.; Lynch, J. J.; Stranieri, M. T.; Vastag, K.;
Yan, Y.; Shafer, J. A.; Vacca, J. P. J. Med. Chem. 1998, 41,
4466.
7. von der Saal, W.; Heck, R.; Leinert, H.; Poll, T.; Stegmeier,
K.; Michel, H. WO 94/20467, 1994.
8. Lu, T.; Tomczuk, B.; Illig, C. R.; Bone, R.; Murphy, L.;
Spurlino, J.; Salemme, F. R.; Soll, R. M. Bioorg. Med. Chem.
Lett. 1998, 8, 1595.
In general, compounds 11±31 showed excellent screen-
ing selectivity (0% inhibition at equal to or greater than
1 mM) against a panel of other serine proteases, includ-
ing Factor Xa, plasmin, urokinase, trypsin, chymo-
trypsin, and elastase. The most fully characterized
analogue, 12, showed greater than 3-orders of magni-
tude selectivity for thrombin over these other serine
proteases, as previously reported.8 This selectivity
against other serine proteases can be readily rationalized
by docking 12 into the active sites of the published X-ray
structures of Factor Xa, trypsin, and plasmin.14 In
Factor Xa, the replacement of Leu99 by Tyr99 within
the S2 region causes a bad contact with the methyl
group on the phenyl template. The structure of trypsin,
which lacks the capping residues of the sixties loop and
the residues for the aryl binding region, would predict
the low anity. Plasmin is even less de®ned in the S2
and aryl binding regions.
The present paper de®nes allowable groups in the
present series projecting into the S2 and S3 thrombin
speci®city pockets. With respect to the S3 pocket, sub-
stitution of the 2 position with a small sampling of
groups spanning hydrophobicity/hydrophilicity and
electron donation/electron withdrawing properties
a€ects potency the least. These results contrast the
dependencies on 3-substitution and the profound e€ects
of 4-substitution reported in earlier work.8 Crystal-
lographic analysis of 12 bound to thrombin (Fig. 1)
shows that substituents ortho to the sulfonate linkage
9. Lu, T.; Illig, C. R.; Tomczuk, B.; Soll, R. M.; Subasinghe,
N. L.; Bone, R. F. WO 97/11693, 1997.
10. Inhibition constants were determined at 37 ꢀC in a 96-well
format using a Molecular Devices plate reader. Varying

82
T. Lu et al. / Bioorg. Med. Chem. Lett. 10 (2000) 79±82
concentrations of inhibitors in 10 mL dimethylsulfoxide were
added to wells with 280 mL of assay bu€er (pH 7.5) which
contained 50 mM HEPES, 0.2 M NaCl, 1% dimethylsulf-
oxide, 0.05% B-octyl-glucoside and substrate and incubated
for 30 min at 37 ꢀC. Reactions were initiated by the addition of
10 mL of enzyme in assay bu€er without dimethylsulfoxide
and substrate and the change in absorbance at 405 nm was
monitored for 30 min. Apparent dissociation constants (Ki
app) were obtained as the inverse slope from plots of the ratio
of initial velocity in the presence of inhibitor to initial velocity
in the absence of inhibitor as a function concentration. The
inhibition constant (Ki) was calculated using the equation Ki ˆ
Ki app/(1+S/Km) where S is the substrate concentration Km
values for each enzyme±substrate pair were determined from
double reciprocal plots using the same ®nal bu€er bu€er con-
ditions as Ki determinations. Substrate, substrate concentra-
tion, and Km value for human a-thrombin were succinyl-Ala-
Ala-Pro-Arg-p-nitroanilide, 100 and 320 mM, respectively.
11. Engh, R. A.; Branstetter, H.; Sucher, G.; Eichinger, A.;
Baumann, U.; Bode, W.; Huber, R.; Poll, T.; Rudolph, R.;
von der Saal, W. Structure 1996, 4, 1353.
12. Bone, R.; Lu, T.; Illig, C. R.; Soll, R.; Spurlino, J. J. Med.
Chem. 1998, 41, 2068.
13. Crystallographic data for 11: R=20.0% at 1.90 A resolu-
tion; 26: R=17.25% at 2.3 A resolution; 29: R=20.5% at
1.90 A resolution. 29: R=20.5% at 1.90 A resolution.
14. The X-ray structures of Factor Xa (1FAX), trypsin (1AOJ),
and plasmin (1BML) are available in the Protein Database.