Design, synthesis and biological evaluation of selective boron-containing thrombin inhibitors

Article abstract of DOI:10.1016/S0968-0896(99)00069-3

Based on the structural comparison of the S-1 pocket in different trypsin-like serine proteases, a series of Boc-D-trimethylsilylalanine-proline-boro-X pinanediol derivatives, with boro-X being different amino boronic acids, have been synthesised as inhib

Full text of DOI:10.1016/S0968-0896(99)00069-3

Bioorganic & Medicinal Chemistry 7 (1999) 1295±1307
Design, Synthesis and Biological Evaluation of Selective
Boron-containing Thrombin Inhibitors
Anette Wienand,* Claus Ehrhardt, Rainer Metternich and Carlo Tapparelliy
Novartis Pharma AG, Research, CH-4002 Basel, Switzerland
Received 20 July 1998; accepted 30 November 1998
AbstractÐBased on the structural comparison of the S-1 pocket in di€erent trypsin-like serine proteases, a series of Boc-d-tri-
methylsilylalanine-proline-boro-X pinanediol derivatives, with boro-X being di€erent amino boronic acids, have been synthesised
as inhibitors of thrombin. The in¯uence of hydrogen donor/acceptor properties of di€erent residues in the P-1 side chain of these
inhibitors on the selectivity pro®le has been investigated. This study con®rmed the structure-based working hypothesis: The
hydrophobic/hydrophilic character of amino acid residues 190 and 213 in the neighbourhood of Asp 189 in the S-1 pocket of
thrombin (Ala/Val), trypsin (Ser/Val) and plasmin (Ser/Thr) de®ne the speci®city for the interaction with di€erent P-1 residues of
the inhibitors. Many of the synthesised compounds demonstrate potent antithrombin activity with Boc-d-trimethylsilylalanine-
proline-boro-methoxypropylglycine pinanediol (9) being the most selective thrombin inhibitor of this series. # 1999 Elsevier
Science Ltd. All rights reserved.
Introduction
inhibitors. They act as `reaction intermediate analogues'
forming a tetrahedral complex with the active site Ser
195. Positively charged residues like Arg or Lys in P-1
and Pro in P-2 increase the anity of these compounds
for thrombin substantially.2,3 However, at the same
time, the selectivity of these peptide analogues over
other trypsin-like serine proteases is generally low.
Thrombin is a trypsin-like serine protease that plays a
central role in both haemostasis and thrombosis. In the
coagulation cascade, thrombin is the ®nal key enzyme.
It is responsible for the cleavage of soluble ®brinogen,
releasing ®brinopeptides A and B and generating ®brin,
which then can polymerise to form a haemostatic plug.
Because of its importance in thrombosis, thrombin
represents a prominent pharmacological target.
It has been shown that peptide boronic acid inhibitors
with non-basic P-1 side chains can improve the selectiv-
ity for thrombin over trypsin and plasmin.4±8 The ®rst
compound of this type was Z-d-Phe-Pro-methoxy-
propylboroglycine.4,5,9±11 The selectivity of this com-
pound was rationalised on the basis of the 3-D
structures of the di€erent proteins, showing that the
hydrophobic methoxypropyl side chain can be better
accomodated in the S-1 pocket of thrombin than in
those of trypsin and plasmin.11
Anticoagulants like heparins or coumarins act indir-
ectly, heparins by activation of endogenous plasma
proteins that inhibit thrombin and other proteases of
the coagulation cascade, coumarins by inhibiting the
hepatic synthesis of vitamin K-dependent proteins
including thrombin. These indirect mechanisms account
to a large part for the limitations of these agents as
therapeutics, because careful monitoring is necessary
and often side e€ects can be observed. Therefore, direct-
acting thrombin inhibitors that control blood clotting
and prevent both arterial and venous thrombosis are
attractive drug candidates.1
In this paper we further elaborate this hypothesis. Sev-
eral compounds of the type Boc-d-TMSal-Pro-boro-X
pinanediol (Boc-d-trimethylsilylalanine-proline-boro-X
pinanediol, with boro-X being di€erent amino boronic
acids) were designed and synthesised which, in general,
show a high selectivity over trypsin and plasmin while
maintaining thrombin anities in the nanomolar range.
Peptide boronic acids based on a ®brinogen-like
sequence are known as very potent direct-acting thrombin
Key words: Anticoagulants; thrombin inhibitors; peptide boronates;
molecular modelling.
Inhibitor Design
* Corresponding author. Tel.: +41 61 324 2096; fax: +41 61 324 3036;
ye-mail: anette.wienand@pharma.novartis.com
Deceased.
The S-1 pockets (Fig. 1(a)) of the serine proteases
thrombin, plasmin and trypsin are structurally similar:
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00069-3

1296
A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
Figure 1. (a) X-ray crystal structure of thrombin from the thrombin/PPACK complex (PDB entry 1ppb17,33). The inhibitor has been removed, water
305 located in a pocket formed by amino acids 190, 213, 228 is shown as red sphere, the pockets S-1, S-2 and S-3 are indicated. (b) Models of the
complexes thrombin/18 and (c) thrombin/6. For the structure calculations water 305 has been removed and the boronic acid has been replaced by a
COCF3-group.

A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
1297
they all contain an aspartic acid (Asp 189) at the bottom
of the S-1 pocket which can form a salt bridge with basic
side chains of substrates or inhibitors. However, there are
two amino acid positions in close proximity to Asp 189,
namely 190 and 213, which are di€erent for the three
enzymes: Ala/Val in thrombin, Ser/Val in trypsin and
Ser/Thr in plasmin, i.e. this region is most hydrophobic
in thrombin and most polar in plasmin. X-ray structures
of thrombin and trypsin show that the small pocket
formed by amino acids 190, 213 and the conserved Tyr
228 is occupied by a water molecule (water 305 in Fig.
1(a)). For optimal stabilisation, this water molecule has to
accept a hydrogen bond from the P-1 side chain of the
inhibitor. Although there is not yet an X-ray structure of
plasmin available, one can conclude from protein homo-
logy building that these features are also valid for plasmin.
Further modi®cations were generated by the application
of the MCSS (Multicopy Simultaneous Search)15,16
methodology to the X-ray structure of thrombin. In this
computational method several thousands of test parti-
cles are seeded at random positions within the active site
of a protein and then minimised simultaneously. As a
result of these calculations a map of optimal binding
sites for the respective test particles can be obtained.
For our purposes benzenes and N-methylacetamides
were seeded into the S-1 pocket and then minimised
using the experimental thrombin structure from the
thrombin/PPACK-complex.17 In order to generate new
structural motives for the P-1 side chain, we tried to
connect the resulting particles manually to the Gly-Ca-
atom of the tri¯uoromethylketone template structure
Boc-d-TMSal-Pro-Gly-C(ˆO)CF3 using the graphics
program INSIGHT II.16 This template was generated
by ®rst modelling Boc-d-TMSal-Pro-Arg-C(ˆO)CF3,
minimising within DISCOVER,16 and then deleting
the Arg side chain. A tri¯uoromethylketone was used
instead of the boronic acid since there was no boron
atom available in the DISCOVER force ®eld. Within
the approximation of the force ®eld this should not
cause signi®cant deviations.
Based on these considerations, we formulated the fol-
lowing working hypothesis: Hydrophobic side chains in
P-1 or side chains containing pure hydrogen bond
acceptors near the water molecule in the small pocket
formed by amino acids 190, 213 and 228 should
decrease binding to the three serine proteases. Due to
the nature of the amino acids in positions 190 and 213
the e€ect should be most pronounced for plasmin and
least important for thrombin. If such a side chain would
still contain a hydrogen bond donor to Asp 189, it
should be possible to produce selective inhibitors which
do not lose too much of their thrombin anity.
A benzene and a modi®ed N-methylacetamide could be
successfully attached to the template in the right dis-
tance resulting in compounds 6 and 18. The DIS-
COVER-minimised structures are shown in Figure 1(b)
and (c). In 18 one carbon has been replaced by a NH2
group to provide a hydrogen bond to Asp 189. Both
compounds are expected to disturb the water molecule
in the pocket close to Asp 189, while still donating a
hydrogen bond to Asp 189. In the case of compound 6,
this `hydrogen bond' has to be provided by an aromatic
C H group.
To exploit and test the above hypothesis, we synthesised
several peptide boronates of type Boc-d-TMSal-Pro-
boro-X pinanediol with constant positions P-3 and P-2
and variations in P-1.
d-Trimethylsilyl alanine (d-TMSal) in P-3 was used as
bioisosteric replacement of d-phenylalanine to reduce
the metabolic instability of our molecules.12 In addition
to the standard compounds 16 (guanidinopropyl in P-
1),13 15 (aminobutyl in P-1)14 and 9 (methoxypropyl in
P-1),10 we also synthesised the derivatives 8, 13 and 17
(see schemes for the di€erent side chains). The inhibi-
tion constants of these compounds for the enzymes
thrombin, plasmin and trypsin were determined by a
chromogenic assay and are summarised in Table 1.
Chemistry
The thrombin inhibitors described in this paper were
synthesised using a convergent synthesis. All of them
have the dipeptide Boc-d-TMSal-Pro-OH in common
which can be combined with the di€erent boroamino
acids in a single synthetic operation (see Scheme 1).
Table 1. Dissociation constants (Ki) of the inhibitors were assessed in
a chromogenic assay with human thrombin, human plasmin and
bovine pancreatic trypsin. The selectivity for the inhibition for
thrombin over plasmin or over trypsin is compared on the basis of
the ratios Ki,PLAS/Ki,THR or Ki,TRY/Ki,THR (in parentheses). Mean
values (‹SD, n=3)
Synthesis of the ®rst building block, Boc-D-TMSal-Pro-
ONSu 3
(S)-N-(t-butoxycarbonyl)-b-(trimethylsilyl)alanine
1
(Boc-d-TMSal-OH)3 was obtained using the Schoell-
kopf procedure as described in the literature.11 After
activation with p-nitrophenol it was coupled with pro-
line to yield 2, which was again activated with N-
hydroxysuccinimide (see Scheme 2) to give Boc-d-
TMSal-Pro-ONSu 33 which can be stored at 4 ꢀC for
several weeks.
Ki [nM]
Compd Thrombin
Plasmin
Trypsin
5485 (4648)
>0.1 mM (>141)
>0.1 mM (>800)
>0.1 mM (>12000)
802 (61.6)
6
7
8
9
13
15
16
17
18
1.18
708
123
8.4
13
0.039
>0.1 mM (>85000)
>0.1 mM (>141)
>0.1 mM (>800)
>0.1 mM (>12000)
>0.1 mM (>7692)
3.51 (90)
0.902 (22)
94.1 (1255)
>0.1 mM (>21000)
Synthesis of the second building block 5 and coupling
with 3
0.39 (10)
0.123 (3)
1.62 (21.6)
0.041
0.075
4.75
All the described amino boronic acids have been syn-
thesised using a procedure developed by Matteson:18±20
1632 (344)

1298
A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
Scheme 1.
Scheme 2.
A suitable substituted ole®n was treated with catechol
borane and the crude product was reacted with (+)-
pinanediol to give the borone pinanediol esters 4. The
following insertion of `CHCl' generated a single dia-
stereomer due to the optically pure pinanediol, resulting
in the chloro substituted homologues 5. However, the
reaction did usually not go to completion. The unreacted
starting material could be removed by chromatography
after the subsequent one-pot-reaction: substitution of Cl
by N(SiMe3)2, acidic hydrolysis and coupling with the
dipeptide building block Boc-d-TMSal-Pro-ONSu 3
(see Scheme 3). Using this methodology, the thrombin
inhibitors 6±9 and the precursors 10±12 were obtained.
ladium on carbon. Further treatment of 14 and 15 with
cyanamide yielded the arginine and homoarginine
derivatives 16 or 17 (see Scheme 5). The urea analogue
of 16 was obtained after the reaction of 14 with potas-
sium cyanate (see Scheme 6).
Biology
Estimates of the inhibition constants (Ki) of the inhibi-
tors 6±9, 13 and 15±18 with thrombin, plasmin and
trypsin were determined by a chromogenic assay. All
compounds tested were potent competitive inhibitors of
thrombin with Ki values in the nanomolar or picomolar
range (see Table 1).
Further elaboration of compounds 10±12
Thrombin inhibitor 13 with a hydroxy group in P-1 was
synthesised from 10 by removal of the t-butyldimethyl-
silyl protecting group with Bu4NF (see Scheme 4). The
boroornithine (14) and borolysine (15) derivatives were
synthesised by substitution of the bromine in 11 and 12
with azide followed by hydrogenation catalysed by pal-
In the series 8, 13, 15 the compound with a Lys side
chain in P-1 (15) has, as expected, the highest anity for
thrombin, but at the same time the lowest selectivity.
Replacement of NH2 by OH (13) reduces the thrombin
anity but increases the selectivity against plasmin and
trypsin. The same e€ect can be seen if the amino-function
Scheme 3. RˆC6H5 (6; 33%), p-Cl-C6H5 (7; 26%), ꢁCH2†2OME (8; 58%), CH2OMe (9; 57%), (CH2)2OSiButMe2 (10; 35%), CH2Br (11; 69%),
(CH2)2Br (12; 68%)

A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
1299
Scheme 4.
Scheme 5.
Scheme 6.
is replaced by a methoxy group (8). The comparison of
8 and 9, where the di€erence is one CH2-group in the
side chain, indicates that the chain length has also a
strong in¯uence on thrombin anity.
chain length (17) or replacing one guanidino NH by O
(18) reduces the anity for thrombin but on the other
side increases the selectivity over trypsin and plasmin.
Compound 6 with a homophenylalanine side chain in P-
1 is very active as inhibitor of thrombin and also very
selective for thrombin over plasmin and trypsin. With a
chloro substituent in p-position of the aromatic ring (7)
The Arg-based derivatives 16, 17 and 18 show a similar
pattern. Arg in P-1 (16) is the most active compound for
thrombin but has only a weak selectivity. Increasing the

1300
A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
the Ki value decreased about 600 times. Therefore, the
substitution pattern of an aromatic ring in P-1 would be
very interesting to investigate, but since we found that
both 6 and 7 are also very potent inhibitors of chymo-
trypsin, we did not invest any further work in this class
of compounds.
As a consequence, the thrombin anity is comparable
to that of compound 16, and the increase in selectivity is
smaller than in the case of compound 18.
The Lys side chain of compound 15 donates three
hydrogen bonds: To the carbonyl oxygens of Ala 190
and Gly 219 and via a water molecule to Asp 189 on the
bottom of the S-1 pocket [cf. ref 22]. This water mole-
cule is in contact with another one which corresponds to
water 305 in Figure 1(a). The hydroxyl group in the
sidechain of compound 13 can only donate one hydrogen
bond so that it cannot take part in all the favourable
hydrophilic interactions shown by the Lys side chain.
As a consequence, the Ki for thrombin of 13 (13 nM) is
higher than that of compound 15 (0.039 nM). Again, the
di€erence between 13 and 15 is more pronounced for
trypsin and plasmin leading to higher selectivity.
Discussion
The thrombin anity and selectivity of the compounds
listed in Table 1 corresponds to the expectations we
derived from analysis of our models.
Recently, an X-ray structure of Moc-d-Dpa-Pro-
boroMpg (morpholinocarbonyl-d-diphenylalanine-pro-
line-boromethoxypropylglycine) with thrombin was
solved.21 This compound contains a propylmethoxy
group in P-1 like compound 9. The structure showsÐin
accordance with our hypothesisÐthat the P-1 side chain
points into the S-1 channel, in the direction of the
pocket formed by amino acids 190/213/228. The fact
that this pocket is more hydrophobic in thrombin com-
pared to trypsin and plasmin allows for a rationalisation
of the selectivity of compound 9, compared to, for
example, 15. However, the X-ray analysis revealed an
additional feature which we did not take into account.
A water molecule forms a hydrogen bond bridge from
the oxygen of the propylmethoxy group to Gly 216 and
219. Compound 8 with a methoxybutyl P-1 side chainÐ
one methylene group longerÐcannot form this hydrogen
bond and therefore binds to thrombin with lower anity.
A phenethyl side chain in P-1 (compound 6) resulted in
a highly active and selective thrombin inhibitor. So far
there is no X-ray crystallographic structure with such a
P-1 residue available. However, an X-ray analysis for a
compound with a penta¯uoroethylketon as transition
state containing a Trp side chain in P-1 was carried
out.23 The indole ring points into the S-1 pocket and the
phenyl ring of the indole moiety interacts with the
phenyl ring of Tyr 228 in a similar fashion as shown in
our model of compound 6 for the phenethyl group (Fig.
1(c)). Another X-ray structure of thrombin complexed
with a non-transition state inhibitor containing a 4-
aminopyridine moiety in P-1 was solved.24 Despite the
repulsion of the pyridine nitrogen and the oxygens of
Asp 189 the aromatic ring points again into the S-1
pocket in a similar way as described above.
Crystallographic analyses of Ac-d-Phe-Pro-boroArg-
OH and related compounds have been published by
Weber et al.22 They give a detailed analysis of the bind-
ing mode for Arg and Lys side chains: One guanidino
nitrogen of the Arg-side chain forms a hydrogen bond
to a water molecule in the small pocket formed by
amino acids 190, 213 and Tyr 228, corresponding to
water 305 in Figure 1(a). According to our modelling
studies one hydrogen of this water molecule should be
directed to the carbonyl of Phe 227 and the second one
towards the center of the aromatic ring of Tyr 228. Such
a water molecule is optimally stabilised by a hydrogen
bond donor and disturbed by a hydrogen bond accep-
tor. This argument explains very clearly the di€erence
between the serine protease binding properties of com-
pound 16 with an Arg side chain in P-1 and its oxygen
analogue 18. Compound 18 (Fig. 1(b)) destabilises the
water molecule in the 190/213/228 pocket in all three
enzymes. Again, this e€ect is much more pronounced
for trypsin and especially plasmin compared to throm-
bin. In agreement with these ®ndings, the anity for
thrombin is reduced by a factor of about 100, while the
anities for trypsin decrease 13,000-fold and for plas-
min more than 100,000-fold.
A chloro substituent in para position on the phenylring
of compound 6 would come quite close to the oxygens
of Asp 189 and would cause a electrostatic repulsion. In
agreement with these predictions compound 7 is about
600-fold less active for thrombin compared to com-
pound 6.
Replacement of the amidino group in the boroArg
compound DuP 714 by substituted 5- and 6-membered
heterocycles also yields selective nanomolar thrombin
inhibitors.8 The most active of this series where those
with amino groups attached to the ring systems even
though X-ray results suggest that the amino group does
not interact with Asp 189.8
Conclusion
The interaction of P-1 residues with Asp 189 in trypsin
like serine proteases is rather complicated. Depending
on the nature of the P-1 residue several amino acids in
the neighbourhood of Asp 189 and up to two water
molecules are involved in addition. Although we did not
include all these aspects into our modelling studies the
original hypothesis based on a controlled perturbation
of a water molecule located in the 190/213/228-pocket
led as expected to selective but still potent thrombin
inhibitors.
In compound 17 with a P-1 side chain one methylene
group longer than in 16, this elongation is responsible
for a slight increase of the hydrophobic character of this
P-1 residue compared to compound 16. However, the
disturbing e€ect is much weaker than in the case of 18.

A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
1301

1302
A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
Experimental
4H), 3.50±3.85 (m, 2H), 4.42±4.60 (m, 1H), 4.77±4.90
(m, 1H), 5.10±5.22 (m, 1H).
All compounds were characterised by 300 MHz proton
NMR using a Bruker 360 FT NMR spectrometer. Che-
mical shifts are expressed as ppm down®eld from tetra-
methylsilane; J values are expressed in Hz. The 1H
NMR data of the compounds 6±18 are summarised in
Table 2, the signals for the di€erent side chains can be
found after the experimental details for the compounds.
Fast atom bombardment mass spectrometry (Xe, 8 keV)
on a VG 70-SE mass spectrometer was used for the
characterisation of the reported compounds. Melting
points were determined on a BUCHI 535 instrument.
For the chromatographic puri®cation, the ¯ash chro-
matography technique was applied using 230±400 mesh
silica gel.
Boc-D-TMSal-Pro-NH-CH[(CH2)2C6H5]B-OPin 6
(+)-Pinanediol-2-phenyl-ethane-1-boronate. Styrene
(3.47g, 30.0 mmol) and catecholborane (3.6 g, 30.0mmol)
are stirred for 20 h at 100 ꢀC, yielding a yellow-brown
oil. This oil is added to 5.0 g (30.0 mmol) of (+)-pina-
nediol in THF and stirred overnight. After ¯ash chro-
matography (hexane:EtOAc, 8:2) 6.2 g (73%) of (+)-
pinanediol-2-phenyl-ethane-1-boronate is obtained: EI
MS 284 (M+); 1H NMR (CDCl3) d 0.83 (s, 3H), 1.00 (d,
J=12 Hz, 1H), 1.19 (t, J=9 Hz, 2H), 1.28 (s, 3H), 1.37
(s, 3H), 1.79±1.86 (m, 1H), 1.86±1.92 (m, 1H), 2.03 (t,
J=6 Hz, 1H), 2.12±2.20 (m, 1H), 2.28±2.38 (m, 1H), 2.78
(t, J=9 Hz, 2H), 4.23±4.28 (m, 1H), 7.12±7.29 (m, 5H).
Boc-D-TMSal-Pro-ONSu 3
Boc-D-TMSal-Pro-NH-CH[(CH2)2C6H5]B-OPin 6. Over
a 20 min period, 15 mL (24.0 mmol) of n-buthyllithium
(1.6 M in hexane) are added to a precooled ( 100 ꢀC)
solution of 2.1 mL of CH2Cl2 and 37 mL of THF. After
stirring for 30 min at 100 ꢀC, 6.2 g (21.8 mmol) of (+)-
pinanediol-2-phenyl-ethane-1-boronate in 15 mL of
THF is added over a 20 min period. The reaction mix-
ture is again stirred for 1 h at 100 ꢀC, after which
1.51 g (10.9 mmol) of ZnCl2 in 12 mL of THF is added.
The reaction mixture is allowed to warm up to room
temperature overnight. After concentration in vacuo the
residue is dissolved in Et2O:H2O. The organic layer is
dried over Na2SO4 and concentrated in vacuo. 6.9 g
(95%) of (+)-pinanediol-(S)-1-chloro-3-phenyl-propane-
1-boronate is obtained as yellow-orange oil.
Boc-d-TMSal-OH 1 (29.7 g, 113.7 mmol) and p-nitro-
phenol (19.0 g, 136.3 mmol) are dissolved in EtOAc.
After cooling to 0 ꢀC, 23.4 g (113.6 mmol) of DCC is
added and the mixture is stirred for 1 h at 0 ꢀC and then
for 15 h at room temperature. A precipitate is formed,
which is ®ltered o€ and washed with EtOAc and the
®ltrate is concentrated in vacuo. The resulting oil is
puri®ed by ¯ash chromatography (hexane:EtOAc, 9:1).
Boc-d-TMSal-ONp is obtained as white crystals.
Boc-d-TMSal-ONp (51.6 g, 113.7 mmol) is dissolved in
THF and an aqueous solution of equimolar amounts of
proline and Et3N is added. After 20 h at room tem-
perature, the THF is removed in vacuo and the aqueous
residue is diluted with water and then extracted several
times with EtOAc. The pH of the aqueous layer is
adjusted to 3 by adding 10% citric acid and extracted
several times with EtOAc. The combined organic layers
are washed with brine, dried over Na2SO4 and con-
centrated in vacuo. The colourless oil is crystallised
from Et2O:hexane to give 31.0 g (76%) of Boc-d-
TMSal-Pro-OH 2 as a white crystalline compound, mp
176 ꢀC.
5.1 mL (8.09 mmol) of n-buthyllithium (1.6 M in hex-
ane) is added to a precooled ( 78 ꢀC) solution of 1.7 mL
(8.09 mmol) of hexamethyldisilazane in 15 mL of THF.
The reaction mixture is stirred 1 h at room temperature,
then cooled down to 78 ꢀC again. 2.68 g (8.09 mmol) of
(+)-pinanediol-(S)-1-chloro-3-phenyl-propane-1-boronate
dissolved in 4 mL of THF is added. After stirring for 1 h
at 78 ꢀC, the solution is warmed up overnight to room
temperature. After recooling to 78 ꢀC, 3 mol equiva-
lents of HCl in dioxane are added. The mixture is stirred
for 1 h at 78 ꢀC, then for 2 h at room temperature.
After cooling down to 20 ꢀC, a solution of 3.7 g
(8.09 mmol) of dipeptide 3 in 15 mL of CH2Cl2 is added,
followed by the addition of 2.3 mL (18.2 mmol) of tri-
ethylamine. The mixture is stirred for 1 h at 20 ꢀC and
for 2 h at room temperature, then ®ltered. After con-
centration in vacuo, the residue is dissolved in
Et2O:H2O. The organic layer is dried over Na2SO4 and
concentrated in vacuo. After ¯ash chromatography
(hexane:EtOAc, 6:4), 1.76 g (33%) of Boc-d-TMSal-
Pro-NH-CH[(CH2)2C6H5]B-OPin 6 is obtained: [a]d20
88.6ꢀ (c 0.5 in MeOH); FAB MS 654 (MH+); 1H
Compound 2 (31.0 g, 86.5 mmol) is dissolved in 350 mL
of EtOAc. After cooling to 0 ꢀC, 12 g (103.5 mmol) of
N-hydroxysuccinimide and 18 g (86.5 mmol) of DCC
are added. The mixture is stirred for 3 h at 0 ꢀC and then
for an additional 15 h at room temperature. The mixture
is recooled to 0 ꢀC, the dicyclohexylurea is ®ltered o€
and washed several times with EtOAc. The ®ltrate is
washed with aqueous 0.1 M Na2CO3 and then with
aqueous 2% KHSO4. After drying over Na2SO4 and
concentration in vacuo, 38.8 g (98%) of Boc-d-TMSal-
Pro-ONSu 3 is obtained as a white foam: FAB MS 456
(MH+); 1H NMR (CDCl3) d 0.07 (s, 9H), 0.96 (d,
J=7.2 Hz, 2H), 1.41 (s, 9H), 2.00±2.45 (m, 4H), 2.82 (s,

A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
1303
NMR (DMSO, 120 ꢀC) d 2.54±2.79 (m, 2H), 7.09±7.17
(m, 2H), 7.19±7.26 (m, 3H).
(12.9 mmol) of (+)-pinanediol-(S)-1-chloro-5-methoxy-
pentane-1-boronate in 7 mL of THF is added and the
mixture is stirred for 1 h at 78 ꢀC, then for 15 h at
room temperature. After this period, the reaction mix-
ture is cooled to 78 ꢀC again. 8.4 mL of HCl in dioxan
(4.6 N solution, 38.6 mmol) is added and the solution
stirred for 1 h at 78 ꢀC and then for 2 h at room tem-
Boc-D-TMSal-Pro-NH-CH[(CH2)2(p-Cl-C6H4)]B-OPin 7
is obtained analogously: [a]2d0 83.2ꢀ (c 0.5ꢀin MeOH);
FAB MS 688 (M+); 1H NMR (DMSO, 150 C) d (2.77±
2.86 (m, 2H), 7.16 (d, J=7.8 Hz), 7.23 (d, J=7.8 Hz).
perature. The mixture is cooled to
20 ꢀC, 5.8 g
Boc-D-TMSal-Pro-NH-CH[(CH2)4OMe]B-OPin 8
(12.9 mmol) of 3 in 20 mL of CH2Cl2 is added, followed
by 3.58 mL (25.7 mmol) of triethylamine to start the
coupling reaction. After stirring at 15ꢀC for 1 h, the
mixture is stirred for 2 h at room temperature. The mix-
ture is then ®ltered over Hy¯o and concentrated in vacuo.
The residue is diluted with Et2O:H2O and extracted
several times with Et2O. After drying over Na2SO4 and
concentration in vacuo, the product is puri®ed by ¯ash
chromatography (EtOAc) and 2.2 g (59%) of Boc-d-
TMSal-Pro-NH-CH[(CH2)4OMe]B-OPin 8 is obtained as
white foam: [a]2d0 80.6ꢀ (c 0.5 in MeOH); FAB MS 636
(+)-Pinanediol-4-methoxy-butane-1-boronate. To 2.1 g
(68.1 mmol) of NaH (80% suspension in mineral oil)
suspended in 30 mL of diglyme 5.04 mL (68 mmol) of 3-
butene-1-ol is added. After 3 h, the evolution of H2 is
stopped and the suspension is cooled to 0 ꢀC. 3.8 mL
(60 mmol) of methyliodide is added and the solution is
stirred at ambient temperature for 2 h. Destillation (bp
70 ꢀC) yields 3.8 g (73%) of 1-methoxy-3-butene.
(M+); H NMR (DMSO, 120 ꢀC) d 1.20±1.61 (m, 6H);
1
3.24 (s, 3H), 3.30±3.35 (m, 2H).
Analogously, Boc-d-TMSal-Pro-NH-CH[(CH2)3OMe]
B-OPin 9 is obtained: [a]2d0 48.8ꢀ (c 0.25 in CH2Cl2);
FAB MS 622 (M+); H NMR (DMSO, 120 ꢀC) d 1.15±
1
2.35 (m, 4H), 3.20 (s, 3H), 3.25±3.35 (m, 2H).
1-Methoxy-3-butene (6.4 g, 73.3 mmol) is reacted with
catecholborane (8.9 g, 74.3 mmol) at 100 ꢀC over 16 h.
The crude product is added to 12.9 g (74.3 mmol) of
(+)-pinanediol dissolved in 85 mL of THF. After two
days at room temperature, the THF is removed in
vacuo and the residue is puri®ed by ¯ash chromato-
graphy (hexane:EtOAc, 85:15) to give 10.9 g (55%) of
(+)-pinanediol-4-methoxy-butane-1-boronate as a col-
Boc-D-TMSal-Pro-NH-CH[(CH2)4OH]B-OPin 13
1
ourless oil: EI MS 266 (M+); H NMR (DMSO) d 0.73
(t, J=7.6 Hz, 2H), 0.82 (s, 3H), 1.00 (d, J=10.1 Hz,
1H), 1.25 (s, 3H), 1.31 (s, 3H), 1.33±1.43 (m, 2H), 1.45±
1.54 (m, 2H), 1.68 (ddd, J=13.7 Hz, 3.4 Hz, 2.1 Hz,
1H), 1.83 (m, 1H), 1.95 (t, J=5.7 Hz, 1H), 2.14±2.22 (m,
1H), 2.25±2.34 (m, 1H), 3.20 (s, 3H), 3.28 (t, J=6.3 Hz,
2H), 4.28 (dd, J=8.0 Hz, 2.1 Hz, 1H).
(+)-Pinanediol-4-t-butyldimethylsiloxy-butane-1-boronate.
3-Butene-1-ol (10.5 mL, 120 mmol), t-butyldimethyl-
silylchloride (22.3 g, 144 mmol), and imidazole (20.4 g,
300 mmol) are dissolved in 60 mL of DMF and stirred
overnight at 35 ꢀC. The two layers are separated, the
product layer is washed with 2 N tartaric acid, water
and brine. After drying over Na2SO4 and concentration
in vacuo, 21 g (94%) of 1-t-butyldimethylsiloxy-3-
butene is obtained.
Boc-D-TMSal-Pro-NH-CH[(CH2)4OMe]B-OPin
8.
CH2Cl2 (1.4 mL) in 25 mL of THF is cooled to
100 ꢀC and 10.3 mL (16.5 mmol) of n-butyllithium (1.6
M in hexane) is added over 20 min. After 30 min at
100 ꢀC, a cold ( 78 ꢀC) solution of 4.0 g (15 mmol) of
(+)-pinanediol-4-methoxy-butane-1-boronate in 9 mL
of THF is added. After an additional 1 h at 100 ꢀC,
1.04 g (7.55 mmol) of anhydrous ZnCl2 in 9 mL of THF
is added. After 15 min at 100 ꢀC, the reaction mixture
is warmed to room temperature. The solvent is removed
in vacuo, the residue diluted with hexane:water and
extracted several times with hexane. After drying over
Na2SO4 and removal of the solvent in vacuo, 4.7 g
(99%) of (+)-pinanediol-(S)-1-chloro-5-methoxy-pen-
tane-1-boronate is obtained as a yellow oil.
1-t-Butyldimethylsiloxy-3-butene (7.0 g, 37.5 mmol) is
stirred for 24 h with 4.75 g (37.5 mmol) of catecholbor-
ane at 125 ꢀC, yielding a yellow-brown oil. This oil is
added to 6.5 g (37.5 mmol) of (+)-pinanediol in 55 mL
THF and stirred overnight. After ¯ash chromatography
(hexane:EtOAc, 98:2) 9.2 g (67%) of (+)-pinanediol-4-
t-butyldimethyl-siloxy-butane-1-boronate is obtained:
1
EI MS 366 (M+); H NMR (CDCl3) d 0.05 (s, 6H),
0.80±0.90 (m, 2H), 0.85 (s, 3H), 0.90 (s, 9H), 1.1 (m, 1H),
1.25 (s, 3H), 1.38 (s, 3H), 1.40±1.60 (m, 5H), 1.76±2.42
(m, 4H), 3.55±3.65 (m, 2H), 4.20±4.28 (m, 1H).
A solution of 2.7 mL (12.9 mmol) of LiN(SiMe3)2 (1 M
in THF) in 20 mL of THF is cooled to 78 ꢀC and
8.07mL (12.9 mmol) of n-buthyllithium (1.6 M in hexane)
is added. The solution is warmed to room temperature
Boc-D-TMSal-Pro-NH-CH[(CH2)4OSi(t-Bu)Me2]B-OPin
10. Over a 20 min period 17.5 mL (27.6 mmol) of n-
buthyllithium (1.6 M in hexane) is added to a precooled
and then cooled down to
78 ꢀC again. 4.05 g

1304
A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
( 100 ꢀC) solution of 3 mL of CH2Cl2 and 50 mL of
THF. After stirring 15 min at 100 ꢀC, 9.2 g (25.1 mmol)
of (+)-pinanediol-4-t-butyldimethyl-siloxy-butane-1-
boronate in 25 mL of THF is added over a 20 min pe-
riod. The reaction mixture is again stirred for 15 min at
100 ꢀC, after which 1.8 g (12.6 mmol) of ZnCl2 (solid)
is added. The reaction mixture is allowed to warm up to
room temperature overnight. After concentration in
vacuo the residue is dissolved in Et2O:H2O. The organic
layer is dried over Na2SO4 and concentrated in vacuo.
9.5 g (91%) of (+)-pinanediol-(S)-1-chloro-5-t-butyl-
dimethyl-siloxy-butane-1-boronate is obtained as yel-
low-orange oil.
(+)-Pinanediol-3-bromo-propane-1-boronate. 3-Bromo-
1-propene (5.1 mL, 60 mmol) is reacted with catechol-
borane (6.5 mL, 60 mmol) at 100 ꢀC over 16 h. The
crude product is added to (+)-pinanediol (10.4 g,
60 mmol) dissolved in 60 mL of THF. After 1 day at
room temperature, the THF is removed in vacuo
and the residue is puri®ed by ¯ash chromatography
(hexane:EtOAc, 95:5) to give 12.7 g (70%) of (+)-pina-
nediol-3-bromo-propane-1-boronate as a colourless oil:
1
MS (CI) 300 (M+); H NMR (CDCl3) d 0.85 (s, 3H),
0.97 (t, J=7.2 Hz, 2H), 1.09 (d, J=11 Hz, 1H), 1.30 (s,
3H), 1.38 (s, 3H), 1.81±1.87 (m, 1H), 1.89±1.95 (m, 1H),
1.96±2.07 (m, 3H), 2.18±2.27 (m, 1H), 2.29±2.38 (m,
1H), 3.44 (t, J=7.2 Hz, 2H), 4.24±4.28 (m, 1H).
n-Buthyllithium (5.6 mL, 9,04 mmol, 1.6 M in hexane) is
added to a precooled ( 78 ꢀC) solution of 1.91 mL
(9.04 mmol) of hexamethyldisilazane in 13 mL of THF.
The reaction mixture is stirred 1 h at room temperature,
then cooled down to 78 ꢀC again. 5.0 g (9.04 mmol) of
(+)-pinanediol-(S)-1-chloro-5-t-butyldimethylsiloxy-
butane-1-boronate in 5 mL of THF is added. After stir-
ring for 1 h at 78 ꢀC, the solution is warmed up over-
night to room temperature. After cooling to 78 ꢀC
again, 3 mol equivalents of HCl in dioxane are added.
The mixture is stirred for 1 h at 78 ꢀC, then for 2 h at
room temperature. After cooling down to 20 ꢀC, a
solution of 4.1 g (9.04 mmol) of dipeptide 3 in 15 mL of
CH2Cl2 is added, followed by the addition of 2.55 mL
(18.1 mmol) of triethylamine. The mixture is stirred for
1 h at 20 ꢀC and for 2 h at room temperature, then ®l-
tered. After concentration in vacuo, the residue is dis-
solved in Et2O:H2O. The organic layer is dried over
Na2SO4 and concentrated in vacuo. After ¯ash chro-
matography (hexane:EtOAc, 6:4), 2.3 g (35%) of Boc-d-
TMSal-Pro-NH-CH[(CH2)4OSi(t-Bu)Me2]B-OPin 10 is
obtained, which was not entirely pure but used without
further puri®cation for the next step: FAB MS 736
(M+); 1H NMR (DMSO, 100 ꢀC) d 0.02 (s, 6H), 0.88 (s,
9H), 1.20±2.35 (m, 6H), 3.53±3.60 (m, 2H).
Boc-D-TMSal-Pro-NH-CH[(CH2)4Br]B-OPin 11. CH2Cl2
(4.1 mL) in 60 mL of THF is cooled to 100 ꢀC and
29.2 mL (46.4 mmol) of n-butyllithium (1.6 M in hex-
ane) is added over 20 min. After 30 min at 100 ꢀC, a
cold ( 78 ꢀC) solution of 12.7 g (42.2 mmol) of (+)-
pinanediol-3-bromo-propane-1-boronate in 25 mL of
THF is added. After an additional 1 h at 100 ꢀC, 2.92 g
(21.1 mmol) of anhydrous ZnCl2 in 25 mL of THF is
added. After 15 min at 100 ꢀC, the reaction mixture
is warmed to room temperature and stirred overnight.
The solvent is removed in vacuo, the residue diluted with
Et2O:H2O and extracted several times with Et2O. After
drying over Na2SO4 and removal of the solvent in vacuo,
12.3 g (83%) of (+)-pinanediol-(S)-1-chloro-4-bromo-
pentane-1-boronate is obtained as a yellow oil.
A solution of 5.5 mL (26.4 mmol) of LiN(SiMe3)2 (1 M
in THF) in 40 mL of THF is cooled to 78 ꢀC and
16.5 mL (26.4 mmol) of n-buthyllithium (1.6 M in hex-
ane) is added. The solution is warmed to room tem-
perature and then cooled down to 78 ꢀC again. 9.23 g
(26.4 mmol) of (+)-pinanediol-(S)-1-chloro-4-bromo-
pentane-1-boronate in 20 mL of THF is added and the
mixture is stirred for 1 h at 78 ꢀC, then for 19 h at
room temperature. After this period, the reaction mix-
ture is cooled to 78 ꢀC again. 17.7 mL (4.6 N solution,
81.4 mmol) of HCl in dioxan is added and the solution
stirred for 1 h at 78 ꢀC and then for 2 h at room
temperature. Compound 3 (9.52 g, 26.4 mmol) dissolved
in 15 mL of THF is cooled to 20 ꢀC and 3.4 mL
(26.4 mmol) of pivalic acid and 4 mL (26.4 mmol) of
triethylamine are added. After 2 h at 20 ꢀC the pre-
Boc-D-TMSal-Pro-NH-CH[(CH2)4OH]B-OPin
13.
Compound 10 (600 mg, 0.816 mmol) is dissolved in
6 mL of THF and tetrabutylammonium ¯uoride (514mg,
1.62 mmol) is added. After stirring overnight at ambient
temperature, the mixture is concentrated in vacuo and
dissolved in Et2O:H2O. The organic layer is dried over
Na2SO4 and concentrated in vacuo. The side product is
removed with EtOAc on silica gel, then the product is
eluted with EtOAc:EtOH (9:1), yielding 320 mg (63%)
of Boc-d-TMSal-Pro-NH-CH[(CH2)4 OH]B-OPin 13 as
a white foam: [a]d20 72.4ꢀ (c 0.5 inꢀMeOH); FAB MS
622 (MH+); 1H NMR (DMSO, 120 C) d 1.19±1.60 (m,
6H), 2.85±2.93 (m, 2H), 3.70±3.75 (m, 1H).
cipitated NEt3 HCl is ®ltered o€. At 20 ꢀC the thus
.
prepared mixed anhydride and 7.7 mL (52.8 mmol) of
triethylamine are added to the reaction mixture above.
After stirring at 20 ꢀC for 1 h, the mixture is stirred for
3 h at room temperature, then diluted with Et2O:H2O
and extracted several times with Et2O. After drying over
Na2SO4 and concentration in vacuo, the product is
puri®ed by ¯ash chromatography (hexane:EtOAc, 1:1)
and 12.2 g (69%) of Boc-d-TMSal-Pro-NH-CH[(CH2)3
Br]B-OPin 11 is obtained as white foam: FAB MS 670
Boc-D-TMSal-Pro-BoroOrn-OPin 14
1
(MH+); H NMR (CDCl3); d 1.20±2.45 (m, 4H), 3.32±
3.42 (m, 2H).
Analogously, Boc-d-TMSal-Pro-NH-CH[(CH2)4Br]B-
1
OPin 12 is prepared: FAB MS 684 (MH+); H NMR
(CDCl3); d 1.25±1.58 (m, 6H), 2.92±3.11 (m, 2H).

A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
1305
Boc-D-TMSal-Pro-BoroOrn-OPin 14. Compound 11
(1.8 g, 2.68 mmol) is dissolved in 7 mL of DMSO and
358 mg (5.37 mmol) of sodium azide is added. The
mixture is stirred for 16 h at room temperature. EtOAc:
ice water is added, and the solution extracted several
times with EtOAc. After drying over Na2SO4 and con-
centration in vacuo, the resulting oil is crystallised to
give 1.4 g (82%) of Boc-d-TMSal-Pro-NH-CH[(CH2)3-
N3]B-OPin as a white crystalline compound: mp 60±
62 ꢀC.
Compound 14 (242.6 mg, 0.4 mmol) and 1 N HCl
(0.4 mL) in H2O is warmed to 50 ꢀC. After 5 min,
33.2 mg (0.4 mmol) of potassium cyanate is added in
small portions. After 3.5 h at 50 ꢀC, ice-water is added,
the product extracted with EtOAc, dried over Na2SO4
and concentrated in vacuo. After ¯ash chromatography
(EtOAc:EtOH, 8:2), 110 mg (53%) of Boc-d-TMSal-
Pro-NH-CH[(CH2)3NHC(O)NH2]B-OPin 18 is obtained:
[a]d20 70.8ꢀ (c 0.5 in MeOH); FAB MS 650 (MH+); 1H
NMR (DMSO, 150 ꢀC) d 1.20±1.60 (m, 4H), 2.94±3.02
(m, 2H), 4.78±4.90 (m, 2H), 5.41±5.50 (m, 2H).
The azide (1.4 g, 2.21 mmol) discussed above is dissolved
in 50 mL EtOAc and hydrogenated in the presence of
0.25g of 10% Pd/C. After 16h, the catalyst is removed
and the solution is concentrated in vacuo. The resulting oil
is crystallised to give 1.07 g (80%) of Boc-d-TMSal-Pro-
BoroOrn-OPin 14 as white crystals: mp 200±202 ꢀC; [a]d20
11.6ꢀ (c 0.5 in MeOH); FAB MS 607 (MH+); 1H NMR
(DMSO, 120 ꢀC) d 1.20±2.01 (m, 4H), 2.70±2.80 (m, 2H).
Enzyme inhibition kinetics
Human a-thrombin was puri®ed from human plasma
according to published methods,25 which involved the
puri®cation of prothrombin,26 the generation of throm-
bin from prothrombin by using the venom from Oxy-
uranus scutellatus,27 and the puri®cation of thrombin on
S-Sepharose (Pharmacia, Uppsala, Sweden).28 The iso-
lated thrombin was free of proteolytically degraded
forms, as judged by SDS-polyacrylamide electropho-
resis, and was found to be >95% active as assessed by
active site titration with 4-methylumbelliferyl-p-guani-
dinobenzoate (Sigma).29 Bovine pancreatic trypsin
(Sigma) and human plasmin (KABI, Moelndal, Swe-
den) were titrated by the same method. The substrates
Pefachrome TH (2-AcOH-H-d-CHG-Ala-Arg-pNA),
Pefachrome PL (2-AcOH-H-d-Ala-CHT-Lys-pNA), and
Pefachrome TRY (Bz-Gly-Arg-pNA) were from Pentafarm.
Analogously, Boc-d-TMSal-Pro-NH-CH[(CH2)4NH2]
B-OPin 15 was prepared: mp 128±129 ꢀC; [a]d20 59.6ꢀ (c
1.0 in MeOH); FAB MS 621 (MH+); 1H NMR
(CDCl3) d 1.20±2.40 (m, 6H), 2.65±2.60 (m, 2H).
Boc-D-TMSal-Pro-BoroArg-OPin 16
Inhibitors were dissolved in cremophor:ethanol (1:1) or
Me2SO and diluted with distilled water to yield a 1 mM
stock solution. Further dilutions were made into the
assay bu€er (100 mM sodium phosphate bu€er, pH 7.4,
containing 100 mM NaCl and 0.1% bovine serum albu-
min). Kinetic assays were performed at 25 ꢀC using a
microwell plate; each well contained 50 mL of substrate,
100 mL of inhibitor, and 100 mL of enzyme in bu€er.
Final concentration of substrate and enzyme were as
follows: 160 pM a-thrombin and 100 mM Pefachrome
TH (Km=6.9 mM), 800 pM human plasmin and 200 mM
Pefachrome PL (Km=66.4 mM), and 260 pM bovine
pancreatic trypsin and 500 mM Pefachrome TRY
(Km=167.7 mM). Km values were determined as pre-
viously described.30 Assays were initiated by adding
enzyme to solutions containing inhibitor and substrate.
The release of p-nitroaniline by hydrolysis of the sub-
strate was followed for 30 min by measuring the increase
in optical density at 405 nm with a Thermomax micro-
well kinetic reader (Molecular Devices, Menlo Park
CA). When the inhibited steady-state rate was achieved
rapidly, the inhibition constant (Ki) was determined by
®tting the data by weighted linear regression to the
Dixon equation.31 For slow, tight-binding inhibitors,
the mechanism of inhibition could be described by
Scheme 7, where E, S, P and I are the enzyme, substrate,
product (p-nitroaniline) and inhibitor, respectively, and
kon and ko€ are the association and dissociation rate
constants for the inhibition. Progress-curve data for the
formation of p-nitroaniline in the presence of di€erent
concentrations of inhibitor were ®tted by non-linear
regression to the equation for the mechanism presented
Compound 14 (250 mg, 0.41 mmol) and benzenesulfonic
acid (65.2 mg, 0.42 mmol) were dissolved in 2 mL of
ethanol and stirred for 10 min at room temperature.
Then 86.6 mg (2.06 mmol) of cyanamide is added and
the reaction mixture is stirred for seven days. Every day
the solution is concentrated and fresh solvent is added.
After seven days the solvent is removed in vacuo, the
residue is dissolved in methanol and ®ltrated over
Sephadex LA 20. 260 mg (78%) of Boc-d-TMSal-Pro-
BoroArg-OPin 18 was obtained as a white foam: [a]d20
45.3ꢀ (c 1.0 in CH2Cl2); FAB MS 649 (MH+); H
1
NMR (DMSO, 150 ꢀC) d 1.20±2.10 (m, 4H), 2.73±2.80
(m, 2H); 6.13±6.24 (m, 3H), 7.02±7.13 (m, 1H).
Analogously, Boc-d-TMSal-Pro-BoroHArg-OPin 17 is
prepared: [a]2d0 40.8 (c 0.5 in CH2Cl2); FAB MS 663
(MH+); 1H NMR (DMSO, 120 ꢀC) d 1.20±2.32 (m, 6H),
2.70±2.80 (m, 2H); 6.68±6.78 (m, 3H), 7.09±7.15 (m, 1H).
Boc-D-TMSal-Pro-NH-CH[(CH2)3NHC(O)NH2]B-OPin
18

1306
A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
C.E. wants to thank Professor M. Karplus and his
group at Harvard University for the introduction to the
MCSS methodology.
References
1. For reviews see: (a) Stone, S. R. Trends Cardiovasc. Med.
1995, 5, 134. (b) Kimball, S. D. Curr. Pharm. Des. 1995, 1,
441. (c) Tapparelli, C.; Metternich, R.; Ehrhardt, C.; Cook, N.
S. Trends Pharmacol. Sci. 1993, 14, 366. (d) Das, J.; Kimball,
S. D. Bioorg. Med. Chem. 1995, 3, 999.
Scheme 7.
2. Kettner, C.; Mersinger, L.; Knabb, R. J. Biol. Chem. 1990,
265, 18289.
3. Metternich, R. European Patent Application 0471651,
1992.
4. Elgedy. S.; Deadman, J.; Patel, G.; Green, D.; Chino, N.;
Goodmin, C. A.; Scully, M. F.; Kakkar, V. V.; Claeson, G.
Tetrahedron Lett. 1992, 33, 4209.
5. Deadman, J. J.; Elgedy, S.; Godwin, C. A.; Green, D.;
Baban, J. A.; Patel, G.; Skordalakes, E.; Chino, N.; Claeson,
G.; Kakkar, V. V.; Scully, M. F. J. Med. Chem. 1995, 38,
1511.
in Scheme 7. These analyses yielded estimates for the
apparent values of kon, ko€, and Ki which were corrected
for the presence of substrate as described by Morrison
and Walsh32 to give the true values. The kinetic con-
stants were determined in triplicate and expressed as
mean‹SD.
Molecular modelling
6. Lee, S.-H.; Alexander, R. S.; Smallwood, A.; Trievel, R.;
Mersinger, L.; Weber, P. C.; Kettner, C. Biochemistry 1997,
36, 13180.
7. Deadman, J.; Claeson, G.; Scully, M. F. J. Enzyme Inhibi-
tion 1995, 9, 29.
8. Dominguez, C.; Carini, D. J.; Weber, P. C.; Knabb, R. M.;
Alexander, R. S.; Kettner, C. A.; Wexler, R. R. Bioorg. Med.
Chem. Lett. 1997, 7, 79.
9. Claeson, G.; Philipp, M. H. W.; Metternich, R. US Patent
Application, 07-406 663, 1989.
For building, manipulation and display of molecular
structures the graphic program INSIGHT II, version
2.516 was used.
All structural calculations were based on the protein
structure taken from the thrombin/PPACK-complex
solved by Bode et al. (PDB entry 1ppb,17,33). The water
molecule occupying the hydrophobic pocket close to
Asp 189 (residue 305) was removed.
10. Claeson, G.; Philipp, M.; Agner, E.; Scully, M. F.; Met-
ternich, R.; Kakkar, V. V. Biochem. J. 1993, 290, 309.
11. Tapparelli, C.; Metternich, R.; Ehrhardt, C. J. Biol. Chem.
1993, 268, 4734.
12. Weidmann, B. Chimia 1992, 46, 312.
13. Cook, N. S.; Tapparelli, C.; Metternich, R. Can. J. Phy-
siol. Pharmacol. 1994, 72, S151.
Structures of thrombin/ligand complexes were mini-
mised by DISCOVER 2.7.16 The inhibitors and protein
residues within 6.0 A around the inhibitor were kept
¯exible during the minimisation.
In DISCOVER the CVFF all hydrogen force ®eld
applying a distance dependent dielectric constant with a
scaling factor of 2.0 was used. The boronic acid part
of the inhibitors was replaced by a tri¯uoromethyl-
methanolate forming a covalent bond between the oxy-
gen of Ser 195 and the tetrahedral methanolate carbon.
14. Tapparelli, C.; Metternich, R.; Gfeller, P.; Gafner, B.;
Powling, M. Thromb. Haemostasis 1995, 73, 641.
15. Miranker, A.; Karplus, M. Protein: Struct. Funct. Genetics
1991, 11, 29.
16. Molecular Simulations Inc., San Diego, Ca 92121-3752.
17. Bode, W.; Mayr, I.; Baumann, U.; Huber, R.; Stone, S.
R.; Hofsteenge, J. EMBO J. 1989, 8, 3467.
18. Matteson, D. S. Tetrahedron 1989, 45, 1859, and refer-
ences therein.
19. Matteson, D. S.; Jesthi, P. K.; Sadhu, K. M. Organome-
tallics 1984, 3, 1284.
20. Matteson, D. S.; Majumdar, D. Organometallics 1983, 2,
1529.
21. Skordalakes, E.; Tyrell, R.; Elgedy, S.; Goodwin, C. A.;
Green, D.; Dodson, G.; Scully, M. F.; Freyssinet, J.-M. H.;
Kakkar, V. V.; Deadman, J. J. J. Am. Chem. Soc. 1997, 119,
9935.
22. Weber, P. C.; Lee, S.-L.; Lewandowski, F. A.; Schadt, M.
C.; Chang, C.-H.; Kettner, C. A. Biochemistry 1995, 34, 3750.
23. Malikayil, J. A.; Burkhart, J. P.; Schreuder, H. A.; Broersma,
Jr., R. J.; Tardif, C.; Kutcher, III, L. W.; Mehdi, S.; Schatzman,
H. L.; Neises, B.; Peet, N. P. Biochemistry 1997, 36, 1034.
24. Von der Saal, W.; Kucznierz, R.; Leinert, H.; Engh, R. A.
Bioorg. Med. Chem. Lett. 1997, 7, 1283; see also: Feng, D. M.;
Gardell, S. J.; Lewis, S. D.; Bock, M. G.; Chen, Z.; Freidinger,
R. M.; Naylor-Olsen, A. M.; Ramjit, H. G.; Woltmann, R.;
Baskin, E. P.; Lynch, J. J.; Lucas, R.; Shafer, J. A.; Dancheck,
K. B.; Chen, I.-W.; Mao, S.-S.; Krueger, J. A.; Hare, T. R.;
Mulichak, A. M.; Vacca, J. P. J. Med. Chem. 1997, 40, 3726.
MCSS calculations15,16 using the above mentioned
thrombin structure were done for N-methylacetamide
(ACAM) and benzene (BENZ) as test particles within
the S-1-pocket only. The particles were seeded in a
sphere with a radius of 10 A centered at the position
of the (removed) oxygen of water 305. The parameter 19
polar hydrogen force ®eld was used.
For ACAM one calculation with 600 particles and 12
others with 1000 particles each were carried out resulting
in 102 di€erent minima. For benzene, 5 calculations
with 400 and another 5 with 600 particles were done
yielding 15 unique minima. Every single calculation
consisted of 10 cycles of simultaneous minimisations.
Acknowledgements
We would like to thank T. Than, H. Kaiser, W. Ludi, P.
Gfeller and B. Gafner for their skilful technical assistance.

A. Wienand et al. / Bioorg. Med. Chem. 7 (1999) 1295±1307
1307
25. Stone, S. R.; Hofsteenge, J. Biochemistry 1986, 25,
4622.
26. Miletich, J. P.; Broze, G. J.; Majerus, P. W. Anal. Bio-
chem. 1980, 105, 304.
27. Owen, W. G., Jackson, C. M. Thromb. Res. 1973, 3, 705.
28. Lundblad, R. L. Biochemistry 1971, 10, 2501.
29. Jameson, G. W.; Roberts, D. V.; Adams, R. W.; Kyle, W.
S. A.; Elmore, D. T. Biochem. J. 1973, 131, 107.
30. Hofsteenge, J.; Taguchi, H.; Stone, S. S. Biochem. J. 1986,
237, 243.
31. Dixon, M. Biochem. J. 1953, 55, 170.
32. Morrison, J. F.; Walsh, C. T. Adv. Enzymol. Areas Mol.
Biol. 1988, 61, 201.
33. Bernstein, F. C.; Koetzle, T. F.; Williams, G. J. B.; Meyer,
E. F.; Brice, M. D.; Rodgers, J. R.; Kennard, O.; Shima-
nouchi, T.; Tasumi, M. J. Mol. Biol. 1977, 112, 535.