Design, synthesis and biological evaluation of thrombin inhibitors based on a pyridine scaffold

Article abstract of DOI:10.1039/b705344d

A series of 2,4-disubstituted pyridine derivatives has been designed, synthesised and evaluated as thrombin inhibitors. A Grignard exchange reaction was used to introduce various benzoyl substituents in position 4 of the pyridine ring, where they serve as

Full text of DOI:10.1039/b705344d

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Design, synthesis and biological evaluation of thrombin inhibitors based on a
pyridine scaffold†
David Blomberg,a Tomas Fex,b Yafeng Xue,b Kay Brickmannb and Jan Kihlberg*a,b
Received 10th April 2007, Accepted 18th June 2007
First published as an Advance Article on the web 10th July 2007
DOI: 10.1039/b705344d
A series of 2,4-disubstituted pyridine derivatives has been designed, synthesised and evaluated as
thrombin inhibitors. A Grignard exchange reaction was used to introduce various benzoyl substituents
in position 4 of the pyridine ring, where they serve as P3 residues in binding to thrombin. In position 2
of the pyridine ring, a para-amidinobenzylamine moiety was incorporated as P1 residue by an SNAr
reaction using ammonia as nucleophile followed by a reductive amination. A crystal structure obtained
for one of the compounds in the active site of thrombin revealed that the basic amidine group of the
inhibitor was anchored to Asp 189 at the bottom of the S1 pocket. A comparison with melagatran,
bound in the active site of thrombin, revealed a good shape match but lack of hydrogen bonding
possibilities in the S2–S3 region for the thrombin inhibitors reported in this study.
as anticoagulants have been developed.5–8 However, until now
Introduction
the most advanced oral DTI is melagatran, which reached the
Blood coagulation is regulated through an intricate network of
market as its dual prodrug ximelagatran, but unfortunately was
later withdrawn.9
biological factors where thrombin (factor IIa) acts as a key enzyme
by converting soluble fibrinogen into an insoluble matrix of fibrin.
Thrombin belongs to the serine protease family where a serine
residue in the active site catalyses the cleavage of amide bonds in
protein and peptide substrates. It was cocrystallised with the ir-
reversible inhibitor PPACK (D-Phe-Pro-Arg chloromethylketone)
and the 3-dimensional structure was determined in 1989.1,2 The
substrate binding pockets are usually designated S1, S2 and S3 and
bind to the corresponding P1, P2 and P3 amino acid residues in
the substrate. The enzyme is highly specific for cleaving after basic
P1 residues, especially arginine, the positively charged guanidine
group interacting with Asp 189 at the bottom of the S1 pocket.
The S2 pocket is lined by the side chains of Tyr 60A and Trp
60D, located in the insertion loop characteristic for thrombin, and
prefers hydrophobic amino acids e.g. proline in the P2 position of
the substrate. The S3 pocket, containing lipophilic and aromatic
moieties as the main interaction points, is known to interact
favourably with aromatic P3 groups.3–5
Today the major drugs used as anticoagulants are warfarin
(Waran) and heparin, which inhibit the blood coagulation cascade
by two different pathways. Since both therapies are associated with
drawbacks, development of new types of anticoagulants is highly
desirable.4 One very attractive opportunity is to develop oral direct
thrombin inhibitors (oral DTIs) using structure based design start-
ing from the crystal structure of the thrombin–PPACK complex.3
By combination of structure based design with co-crystallisation
of the inhibitors with thrombin, several compounds with potential
As part of our studies focused on developing new chemistry
around substituted pyridines,10,11 we recognised the possibility to
apply our methodology to the synthesis of potential thrombin
inhibitors. Applying the P1-P2-P3 nomenclature, it was envi-
sioned that the pyridine ring could serve as a P2 scaffold.
A p-amidinobenzylamine residue, known from many thrombin
inhibitors,12–17 was considered suitable as a P1 substituent also in
our case. Various benzoyl groups, with small substituents were
chosen to fill the S3 pocket. Benzoyl groups containing either a
m-methyl or an o-methoxy substituent provided a good fit in the
structure based design and were also predicted to be compatible
with the planned synthetic route. Glide docking of one of the
target inhibitors, compound 20a, displays the interactions between
the amidine and Asp 189 as well as a potential hydrogen bond
between the benzoyl carbonyl group and the phenolic hydroxyl
group in Tyr 60A (Fig. 1).18 Inhibitors 20a–c were prepared via
a 14 step synthetic sequence, and the three designed compounds
were evaluated as thrombin inhibitors in an enzymatic assay.
aOrganic Chemistry, Department of Chemistry, Umea˚ University, SE-901
87 Umea˚, Sweden. E-mail: jan.kihlberg@chem.umu.se
bAstraZeneca R&D Mo¨lndal, SE-431 83 Mo¨lndal, Sweden
† Electronic supplementary information (ESI) available: Copies of 1H and
13C NMR spectra and lists of spectral data for all isolated compounds.
HRMS for all isolated compounds. Reversed phase HPLC chromatograms
for compounds 20a–c. Data collection and refinement statistics for the X-
ray structure are also available. See DOI: 10.1039/b705344d
Fig. 1 Compound 20a flexibly docked into the active site of thrombin
(PDB code 1K22) using standard precision Glide with a rigid receptor.
The molecular surface is coloured by atom type, where red, blue, yellow
and white denote oxygen, nitrogen, sulfur and carbon, respectively.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 2599–2605 | 2599
©

View Article Online
periodinane.22 The electron withdrawing effect of the carbonyl
group at position 4 of the pyridine ring in 3a–c was expected to
enhance the reactivity in the following SNAr reaction at position 2
using benzylamine IV as nucleophile.
Results and discussion
Synthetic route to the thrombin inhibitors
The designed thrombin inhibitors I contain a central pyridine
scaffold appropriately substituted at positions 2 and 4 (Fig. 2). A
retrosynthetic analysis10,19 revealed 2-fluoro-4-iodopyridine20 (II)
as a suitable scaffold with the possibility to introduce substituents
both at positions 2 and 4. Introduction of a substituent at position
4 of the pyridine ring requires an iodo–magnesium Grignard
exchange reaction, followed by quenching with an appropriately
substituted benzaldehyde III as electrophile.21 Introduction of
a substituent at position 2 of the pyridine ring should be
accomplished via a nucleophilic aromatic substitution reaction
(SNAr) using benzylamine IV as nucleophile. The nucleophilic
aromatic substitution was planned as the second transformation
due to anticipated selectivity issues between fluorine and iodine,
if the SNAr reaction was performed as the first step. The desired
benzamidine moiety in thrombin inhibitors I was intended to be
obtained from the cyano group in IV.
In a first attempt to introduce the benzylamine moiety, an excess
of 4-cyanobenzylamine was reacted with ketone 3a in pyridine at
100 ◦C for 30 minutes using microwave irradiation (Scheme 2).
A colour change of the reaction mixture was observed but only
starting material 3a was detected by LCMS analysis. Increasing the
temperature to 130 ◦C did not alter the outcome of the reaction.
To investigate if the cyano functionality in 4-cyanobenzylamine
was the reason for the failure of this nucleophilic substitution, 4-
bromobenzylamine was instead used as nucleophile. The bromo
functionality was then intended to be transformed to the desired
cyano moiety at a later stage via a Rosemund von Braun reaction.23
Hence, excess 4-bromobenzylamine and ketone 3a were heated to
130 ◦C in pyridine using microwave irradiation for 30 minutes.
Now the desired substitution product 5 was formed, but only
in an unsatisfactory yield (<10%, LCMS analysis) accompanied
by nearly equal amounts of the undesired decomposition product
aminopyridine 6a. The LCMS analysis also revealed that the major
part of starting material 3a was left unchanged.
Fig. 2 A retrosynthetic analysis suggesting that thrombin inhibitors I
could be generated from 2-fluoro-4-iodopyridine (II), substituted ben-
zaldehydes III and 4-cyanobenzylamine (IV).
Scheme 2 (i) 4-Cyanobenzylamine, pyridine, 100–130 ◦C, microwave irra-
diation; (ii) 4-bromobenzylamine, pyridine, 130 ◦C, microwave irradiation.
As a first step in the synthetic route, 2-fluoro-4-iodopyridine was
dissolved in THF and treated with isopropyl magnesium chloride
to achieve an iodo–magnesium exchange, which was followed by
addition of a substituted benzaldehyde (3 examples) to react with
the formed Grignard reagent (Scheme 1). This gave secondary
alcohols 2a–c (86–87%), which were subsequently oxidised to
the corresponding ketones 3a–c (86–88%) using Dess–Martin
Due to the unexpected problems with the nucleophilic substitu-
tion reactions using benzylamines outlined above, other synthetic
alternatives were investigated. Use of ammonia as nucleophile in
reactions with ketones 3a–c was expected to afford aminopyridines
6a–c (Scheme 3). The strategy was then to couple 6a–c reductively
with a benzaldehyde, preferably 4-cyanobenzaldehyde, as the next
step was planned to be the transformation of a cyano group to the
desired amidines. In order to evaluate this route, fluoropyridine 3a
was treated with 25% ammonia in water and heated in a sealed steel
cylinder to ∼150 ◦C for 10 h. This gave aminopyridine 6a together
with approximately equimolar amounts of the corresponding
imine. The ketone–imine mixture was dissolved in 10% citric
acid (aq.) to allow hydrolysis of the imine which afforded ketone
6a in excellent yield (94%). Several conditions24–26 for reductive
amination of 6a with 4-cyanobenzaldehyde were investigated
and use of sodium triacetoxyborohydride and triethylamine in
dichloromethane27 turned out to be the most promising. However,
after Boc-protection,25 aminopyridine 8a was only obtained in a
disappointingly low yield (12%). We reasoned that the low yield
Scheme 1 (i) i-PrMgCl, substituted benzaldehyde, THF (86–87%); (ii)
Dess–Martin periodinane, CH2Cl2 (86–88%).
2600 | Org. Biomol. Chem., 2007, 5, 2599–2605
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Scheme 3 (i) 25% NH3 in water ∼150 ◦C (94–99%); (ii) 4-cyanobenzalde-
hyde, triethylamine, Na(OAc)3BH, toluene, CH2Cl2; (iii) Boc2O, DMAP,
CH2Cl2 (64–78%, from 6a–c).
in the reductive amination was most likely due to difficulties in
formation of the imine, rather than problems in the reduction of
the imine to the amine. Therefore a strategy was attempted where
the imine was preformed and then reduced.28 Aminopyridine 6a,
4-cyanobenzaldehyde and triethylamine were refluxed for 10 h
in a Dean–Stark apparatus with toluene as solvent, followed by
reduction with sodium triacetoxyborohydride. The Dean–Stark
procedure and the reduction was repeated twice, then the generated
secondary amine 7a was protected with a Boc group which gave
8a in a good yield (77%, from 6a). Next, fluoropyridines 3b–
c were treated with aqueous ammonia as described for 3a to
generate aminopyridines 6b–c (96–99%) after acidic hydrolysis.
Ketones 6b–c were then subjected to reductive amination followed
by Boc-protection, as described above, to afford compounds 8b–
c (64–78%, from 6b–c). With the three structural components
corresponding to P1, P2 and P3 motifs of thrombin inhibitors
in place, only transformation of the cyano group to the desired
amidine remained to be accomplished.
Unfortunately, attempts to convert the cyano functionality
in 8a to the desired amidine with lithium hexamethyldisilazide
(LiHMDS) in THF failed. This was also the case for the Pinnerer
reaction using hydrochloric acid in ethanol followed by addition of
ammonia in methanol.29 However, reacting 8a with hydroxylamine
to afford the corresponding amidine oxime followed by O-
acylation using acetic anhydride in acetic acid afforded compound
9 (Scheme 4). By subjecting 9 to hydrogenation using a catalytic
amount of palladium on activated carbon in methanol the labile
nitrogen–oxygen bond was cleaved to generate the amidine, but
simultaneously the ketone was reduced to the corresponding
alcohol. The highly polar amidine was then protected with a
Boc-group using Boc2O to give amidine 18a in modest and
unpredictable yields (0–22%, from 8a).30 Two reasons were found
for the variable outcome in the formation of amidine 18a. The
oxime corresponding to 9 was formed when ketone 8a was treated
with hydroxylamine, which then resulted in the undesired amine
10 after hydrogenation and protection. In addition, reduction
of the keto functionality in 9, not only to the corresponding
alcohol, but also further to the deoxygenated compound 11 was
observed. The amount of deoxygenated product was strongly
dependent on the reaction time and choice of solvent. Reduction
in acetic acid gave methylene derivative 11 as the major product,
whereas use of methanol in combination with short reaction times
favoured formation of alcohol 18a.31 The lack of reliability in the
transformation of 9 into 18a required further investigations in
Scheme 4 (i) a: H2NOH·HCl, DIPEA, ethanol, reflux; b: acetic anhy-
dride, acetic acid; (ii) a: H2 (1 atm), Pd/C, methanol; b: Boc2O, CH2Cl2.
order to establish a robust synthetic route. As the encountered
problems, at least in part, seemed to originate from the keto
functionality present in compounds 8a–c, the ketones were first
reduced to the corresponding alcohols 12a–c using sodium boro-
hydride (Scheme 5). Subsequently, the alcohols were protected as
silylethers using tert-butyldimethylsilyl trifluoromethanesulfonate
and collidine in dichloromethane to afford 13a–c (82–84%, from
8a–c). The cyano groups of 13a–c were then converted to
hydroxy amidines 14a–c and acylated to afford 15a–c as described
above. Cleavage of the labile nitrogen–oxygen bond was achieved
by hydrogenation for 35 minutes at atmospheric pressure with
catalytic amounts of palladium on activated carbon in methanol
to afford amidines 16a–c. To our satisfaction the silyl ethers proved
considerably more stable during hydrogenation than ketone 9, and
the undesired deoxygenation side product was detected only in
trace amounts (LCMS). Protection of the amidines with a Boc
group afforded 17a–c (46–66%, from 13a–c) and the silyl ethers
were then cleaved using tetrabutylammonium fluoride (TBAF) to
afford alcohols 18a–c. Oxidation to ketones 19a–c was thereafter
achieved by treatment with Dess–Martin periodinane. Finally,
the Boc group was removed from the thrombin inhibitors in
a mixture of trifluoroacetic acid and dichloromethane followed
by purification with reversed phase HPLC to afford the target
compounds 20a–c (33–53%, from 17a–c). The synthetic route to
thrombin inhibitors 20a–c was thus accomplished over 14 steps
with an overall yield ranging from 10–14%.
Biological evaluation of the thrombin inhibitors
Compounds 20a–c were evaluated as thrombin inhibitors in an
enzymatic assay32 (Table 1). All three compounds showed only
modest inhibition. The results indicate that binding to thrombin
is influenced by the substitution pattern of the benzoyl group,
i. e. the P3 residue. The ortho-methoxy group in 20b decreases
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 2599–2605 | 2601
©

View Article Online
Scheme 5 (i) NaBH4, methanol; (ii) TBDMSOTf, collidine, CH2Cl2 (82–84%, from 8); (iii) H2NOH·HCl, DIPEA, ethanol; (iv) acetic anhydride, acetic
acid; (v) H2 (1 atm), Pd/C, methanol; (vi) Boc2O, CH2Cl2 (46–66%, from 13); (vii) TBAF, THF; (viii) Dess–Martin periodinane, CH2Cl2; (ix) TFA,
CH2Cl2 (33–53%, from 17).
Table 1 Evaluation of compounds 20a–c in an enzymatic assay
way than was borne out in practice (cf. Fig. 1 and 3a). The
hydrogen bond between the benzoyl carbonyl group and the
phenolic hydroxyl group in Tyr 60A, also suggested from dockings,
was not formed in the cocrystal. It is possible that elongation of the
benzylamine moiety to a phenylethylamine could give a compound
with a better fit to the enzyme, e.g. filling out more of the S2 and
S3 pockets and also reaching some of the important hydrogen
bonds.
Compound
IC50/lM
Experimental
20a
20b
20c
R = H
15
>44.4a
11
R = o-OMe
(2-Fluoro-pyridin-4-yl)-phenyl-methanol (2a),
(2-fluoro-pyridin-4-yl)-(2-methoxy-phenyl)-methanol (2b) and
(2-fluoro-pyridin-4-yl)-m-tolyl-methanol (2c)
R = m-Me
a The highest concentration applied in the assay was 44.4 lM.
General procedure. 2-Fluoro-4-iodopyridine (1 equiv.,
2.9 mmol) was dissolved in THF (3 mL) and treated with
isopropyl magnesium chloride (1.3 equiv., 2 M in THF, 3.8 mmol)
and stirred for 1 h without cooling the exothermic reaction.
The benzaldehyde (if solid, dissolved in THF, 1 mL) (1.3 equiv.,
3.8 mmol) was added and the reaction was stirred at room
temperature for 12 h. The reaction was quenched with NH4Cl
(aq., sat.) and diluted with NaHCO3 (aq., sat.). The aqueous
phase was extracted with EtOAc and the combined organic layers
were dried and concentrated under reduced pressure. The residue
was purified by flash chromatography EtOAc–heptane 1 : 4 → 1 :
2 to give alcohol 2a (86%), 2b (converted to 3b as crude product),
or 2c (87%).
binding affinity while the meta-methyl group in 20c results in an
equally active inhibitor as unsubstituted 20a. Inspection of the X-
ray structure (Fig. 3) suggests that an ortho-methoxy group should
be well accommodated but apparently it has a negative influence
on binding of the benzoyl group in the S3 pocket.
Crystal structure of 20c in the active site of thrombin
Even though the affinity was moderate, a crystal structure of 20c
in the active site of thrombin could be obtained (Fig. 3a). As
expected, the amidine group was found to anchor 20c in the S1
pocket of thrombin by binding to Asp 189. A comparison between
the crystal structures of 20c and the potent thrombin inhibitor
melagatran (PDB code 1k22) reveals quite a good shape match
between the two inhibitors (Fig. 3b). However, melagatran forms
three hydrogen bonds to the backbone of the enzyme in the S2–S3
region, in addition to its interaction with Asp189 in the S1 pocket
(Fig. 3c). Such hydrogen bonds are absent for 20c (Fig. 3a) and
this difference most likely explains the relatively low potency of
20c. In general it can be concluded that the hydrogen bonds in the
S2–S3 region of thrombin are of great importance for potency and
should be considered as a high priority in the design of thrombin
inhibitors.
(2-Fluoro-pyridin-4-yl)-phenyl-methanone (3a),
(2-fluoro-pyridin-4-yl)-(2-methoxy-phenyl)-methanone (3b) and
(2-fluoro-pyridin-4-yl)-m-tolyl-methanone (3c)
General procedure. Alcohol 2a, 2b or 2c (1 equiv., 2.5 mmol)
was dissolved in CH2Cl2 (40 mL) and treated with Dess–Martin
periodinane (1.2 equiv., 15 wt% solution in CH2Cl2) at room
temperature. After 0.5 h, Na2S2O5 (21.7 mmol) dissolved in
NaHCO3 (aq., sat.) was added and the mixture was vigorously
stirred until two clear phases were obtained. The aqueous layer
was extracted with CH2Cl2 and the combined organic layers was
washed with NaHCO3 (aq., sat.), dried and concentrated under
A comparison of the crystal structure with results from Glide
dockings shows that docked 20a fill the S2 pocket in a better
2602 | Org. Biomol. Chem., 2007, 5, 2599–2605
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
(2-Amino-pyridin-4-yl)-phenyl-methanone (6a),
(2-amino-pyridin-4-yl)-(2-methoxy-phenyl)-methanone (6b) and
(2-amino-pyridin-4-yl)-m-tolyl-methanone (6c)
General procedure. Ketone 3a, 3b or 3c (2 mmol) was added
to a sealed teflon coated steel cylinder followed by addition of
25% ammonia solution in water (25 mL) and heated on a sand
bath (∼150 ◦C) for 15 h. After cooling to room temperature, the
aqueous phase was extracted with EtOAc. The combined organic
layers were concentrated under reduced pressure and the residue
was treated with citric acid (10%, 50 mL) for 3 h. The aqueous layer
was made basic (pH ∼ 8) with NaHCO3 (s) and extracted with
EtOAc. The combined organic layers were dried and concentrated
under reduced pressure to give amine 6a (94%), 6b (96%) or 6c
(99%).
(4-Benzoyl-pyridin-2-yl)-(4-cyano-benzyl)-carbamic acid tert-butyl
ester (8a), (4-cyano-benzyl)-[4-(2-methoxy-benzoyl)-pyridin-2-yl]-
carbamic acid tert-butyl ester (8b) and (4-cyano-benzyl)-[4-(3-
methyl-benzoyl)-pyridin-2-yl]-carbamic acid tert-butyl ester (8c)
General procedure. Amine 6a, 6b or 6c (1.0 equiv., 2 mmol) was
mixed with 4-cyano benzaldehyde (1.3 equiv.) and triethylamine
(1.5 equiv.) in toluene (120 mL). The mixture was heated to
reflux in Dean–Stark equipment for 10 h, followed by removal
of the remaining solvent under reduced pressure. The residue
was dissolved in CH2Cl2 (10 mL) and treated with sodium
triacetoxyborohydride (1.3 equiv.) and the reaction was stirred for
3 h at room temperature. The reaction was concentrated under
reduced pressure and the residue was subjected to the above
described procedure two additional times. Then the reaction was
quenched with NaHCO3 (aq., sat.) and extracted with CH2Cl2.
The combined organic phases were dried and concentrated under
reduced pressure. The residue was purified by flash chromatog-
raphy EtOAc–heptane 1 : 3 → 1 : 1 to give aminopyridine 7a,
7b or 7c. The aminopyridine was dissolved in CH2Cl2 (40 mL)
and treated with Boc2O (5 equiv.), DMAP (cat., 0.1 equiv.).
After 24 h, the reaction was washed with NaHCO3 (aq., sat.)
and extracted with CH2Cl2. The combined organic phases were
dried and concentrated under reduced pressure. The residue was
purified by flash chromatography EtOAc–heptane 1 : 3 to give
Boc-protected aminopyridine 8a (77% from 6a), 8b (64% from 6b)
or 8c (78% from 6c).
{4-[(tert-Butyl-dimethyl-silanyloxy)-phenyl-methyl]-pyridin-2-yl}-
(4-cyano-benzyl)-carbamic acid tert-butyl ester (13a),
{4-[(tert-butyl-dimethyl-silanyloxy)-(2-methoxy-phenyl)-methyl]-
pyridin-2-yl}-(4-cyano-benzyl)-carbamic acid tert-butyl ester (13b)
and {4-[(tert-butyl-dimethyl-silanyloxy)-m-tolyl-methyl]-pyridin-
2-yl}-(4-cyano-benzyl)-carbamic acid tert-butyl ester (13c)
Fig. 3 (a) The crystal structure of thrombin inhibitor 20c in the active
site of thrombin. (b) An overlay of melagatran (green) and 20c (blue) when
bound by thrombin. (c) Melagatran bound in the active site of thrombin
with the hydrogen bonding network in the S2 and S3 pockets displayed (red
dashed lines). In the S2–S3 region hydrogen bonds are formed between two
of the nitrogen atoms and one of the carbonyl oxygen atoms in melagatran
and the backbone of the enzyme.
General procedure. Ketone 8a, 8b or 8c (0.17 mmol, 1.0 equiv.)
was dissolved in MeOH (5 mL) and treated with NaBH4
(2.0 equiv.) at room temperature. The reaction was stirred for
10 minutes and then the solvent was removed under reduced
pressure. The residue was dissolved in CH2Cl2 and washed with
brine. The organic layer was dried and concentrated under reduced
pressure to afford alcohols 12a, 12b or 12c. The alcohol was dis-
solved in CH2Cl2 (5 mL) and treated with 2,4,6-trimethylpyridine
(1.3 equiv.) and tert-butyldimethylsilyl trifluoromethanesulfonate
reduced pressure. The residue was purified by flash chromatogra-
phy EtOAc–heptane 1 : 5 to give ketone 3a (88%), 3b (80%, from
1) or 3c (86%).
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 2599–2605 | 2603
©

View Article Online
(1.3 equiv.) for 40 minutes at room temperature. The reaction was
quenched with NaHCO3 (aq., sat.) and the aqueous layer was
extracted with CH2Cl2. The combined organic layers were dried
and concentrated under reduced pressure. The residue was purified
by flash chromatography EtOAc–heptane 1 : 4 to give 13a (82%),
13b (84%), or 13c (82%).
(aq., sat.). The aqueous layer was extracted with CH2Cl2 and
the combined organic layers were dried and concentrated under
reduced pressure. The residue was purified by flash chromatog-
raphy EtOAc–heptane 1 : 4 → 1 : 2 to afford ketone 19a, 19b
or 19c. The ketone was dissolved in CH2Cl2 (1 mL) and treated
with trifluoroacetic acid (3 mL) for 35 minutes. The solvent was
removed under reduced pressure and the residue was purified by
preparative reversed phase HPLC to give amidines 20a (33%), 20b
(53%) or 20c (52%).
[4-(tert-Butoxycarbonylamino-imino-methyl)-benzyl]-{4-[(tert-
butyl-dimethyl-silanyloxy)-phenyl-methyl]-pyridin-2-yl}-carbamic
acid tert-butyl ester (17a), [4-(tert-butoxycarbonylamino-imino-
methyl)-benzyl]-{4-[(tert-butyl-dimethyl-silanyloxy)-(2-methoxy-
phenyl)-methyl]-pyridin-2-yl}-carbamic acid tert-butyl ester (17b)
and [4-(tert-butoxycarbonylamino-imino-methyl)-benzyl]-{4-[(tert-
butyl-dimethyl-silanyloxy)-m-tolyl-methyl]-pyridin-2-yl}-carbamic
acid tert-butyl ester (17c)
Thrombin inhibitor measurements32. The thrombin inhibitor
potency of compounds 20a–c was measured with a chromogenic
substrate method in a robotic microplate processor, using 96-
well, half volume microtiter plates. The linear absorbance increase
values were determined by measurements at 405 nm (37 ◦C) during
40 minutes with melagatran as control substance. The IC50 values
were calculated by fitting the data to a three parameter equation
by Microsoft XLfit.
General procedure. Cyano compound 13a, 13b or 13c
(0.47 mmol, 1.0 equiv.) was dissolved in EtOH (20 mL) and
treated with H2NOH·HCl (10 equiv.) and diisopropylethylamine
(20 equiv.) under reflux for 1.5 h. The solvent was removed under
reduced pressure and coevaporated with CHCl3 two times to
give hydroxy amidine 14a, 14b or 14c. The hydroxy amidine was
dissolved in acetic acid (20 mL) followed by addition of Ac2O
(30 equiv.) at room temperature. After 30 minutes, the solvent was
removed under reduced pressure with toluene as azeotrope and
the residue was taken up in EtOAc and washed with NaHCO3
(aq., sat.). The organic layer was dried and concentrated under
reduced pressure from toluene two times to afford 15a, 15b or 15c.
The acetylated hydroxy amidine was dissolved in MeOH (50 mL)
and Pd/C (100 wt%, compared to 13a, 13b or 13c) was added and
the reaction was put under H2 atmosphere at normal pressure.
The reduction was vigorously stirred for 30 minutes, followed by
removal of Pd/C by filtration through a pad of celite, which was
rinsed with MeOH–AcOH 10 : 1. The solvent was removed with
toluene as azeotrope and the residue was dissolved in EtOAc and
washed with Na2CO3 (aq., sat.). The combined organic layers
were dried and concentrated under reduced pressure to afford
16a, 16b or 16c. The amidine was dissolved in CH2Cl2 (10 mL)
and treated with Boc2O (1.1 equiv.) for 5 minutes followed by
addition of NaHCO3 (aq., sat.). The aqueous phase was extracted
with EtOAc and the combined organic layers were dried and
concentrated under reduced pressure. The residue was purified
by flash chromatography EtOAc–heptane 1 : 4 to give 17a (66%),
17b (46%) or 17c (57%).
Crystallisation and X-ray structure determination. Crystalli-
sation and soaking of compound: human a-thrombin (factor
IIa) was purchased from “Enzyme Research Laboratories”. Pre-
formed thrombin–hirudin complex (stored at 4 degrees) was used
for crystallisation. Apo-crystals were grown using a hanging-
drop method with micro-seeding. The drop was made by mixing
1.5 lL of the thrombin–hirudin complex and 1.5 lL of reservoir
solution containing: 0.05 M sodium phosphate buffer (pH 7.3),
28% PEG8000. The soaking was carried out by adding powder
of the compound directly to the drop and leaving it at room
temperature for 7 weeks.
Data collection and structure determination: the diffraction
data were collected from a frozen crystal on a MarCCD detector
mounted on a microfocus rotating anode generator FR-E Super-
Bright from Rigaku. The crystal-to-detector distance was set to
be 140 mm and 104 images were collected with oscillation of 1
degree. The data were processed with MOSFLM33 and programs
in the CCP4 suite.34 Refinement and model re-building was carried
out using programs REFMAC535 and COOT.36 The structure has
been deposited with PDB code 2PKS.
Acknowledgements
This work was funded by grants from the Swedish Research
Council, the Go¨ran Gustafsson Foundation for Research in
Natural Sciences and Medicine, and AstraZeneca R&D Mo¨lndal.
We also thank Isabella Feierberg at AstraZeneca R&D Mo¨lndal
for help with the docking.
4-[(4-Benzoyl-pyridin-2-ylamino)-methyl]-benzamidine acetate salt
(20a), 4-{[4-(2-methoxy-benzoyl)-pyridin-2-ylamino]-methyl}-
benzamidine acetate salt (20b) and 4-{[4-(3-methyl-benzoyl)-
pyridin-2-ylamino]-methyl}-benzamidine acetate salt (20c)
References
1 W. Bode, I. Mayr, U. Baumann, R. Huber, S. R. Stone and J. Hofsteenge,
EMBO J., 1989, 8, 3467.
General procedure. Compound 17a, 17b, or 17c (0.2 mmol,
1.0 equiv.) was dissolved in THF (4 mL) at room temperature
and treated with tetrabutylammonium fluoride hydrate (2.0 equiv.)
for 1 h. The solvent was removed under reduced pressure and
the residue was filtered through a path of silica gel (EtOAc–
heptane, 2 : 1 as eluent) to afford alcohols 18a, 18b or 18c. The
afforded alcohol was dissolved in CH2Cl2 (4 mL) and treated with
Dess–Martin periodinane (1.3 equiv.) at room temperature for 5
minutes followed by addition of Na2S2O5 (4 equiv.) in NaHCO3
2 W. Bode, D. Turk and A. Karshikov, Protein Sci., 1992, 1, 426.
3 R. Pfau, Curr. Opin. Drug Discovery Dev., 2003, 6, 437.
4 S. Srivastava, L. N. Goswami and D. K. Dikshit, Med. Res. Rev., 2005,
25, 66.
5 T. Steinmetzer and J. Stu¨rzebecher, Curr. Med. Chem., 2004, 11, 2297.
6 C. A. Coburn, Expert Opin. Ther. Pat., 2001, 11, 721.
7 J. C. Danilewicz, S. M. Abel, A. D. Brown, P. V. Fish, E. Hawkeswood,
S. J. Holland, K. James, A. B. McElroy, J. Overington, M. J. Powling
and D. J. Rance, J. Med. Chem., 2002, 45, 2432.
8 T. Steinmetzer, J. Hauptmann and J. Stu¨rzebecher, Expert Opin. Invest.
Drugs, 2001, 10, 845.
2604 | Org. Biomol. Chem., 2007, 5, 2599–2605
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
9 D. Gustafsson, R. Bylund, T. Antonsson, I. Nilsson, J. E. Nystro¨m,
U. Eriksson, U. Bredberg and A. C. Teger-Nilsson, Nat. Rev. Drug
Discovery, 2004, 3, 649.
10 D. Blomberg, K. Brickmann and J. Kihlberg, Tetrahedron, 2006, 62,
10937.
11 S. Saitton, J. Kihlberg and K. Luthman, Tetrahedron, 2004, 60, 6113.
12 N. H. Hauel, H. Nar, H. Priepke, U. Ries, J. M. Stassen and W. Wienen,
J. Med. Chem., 2002, 45, 1757.
13 J. Z. Ho, T. S. Gibson and J. E. Semple, Bioorg. Med. Chem. Lett., 2002,
12, 743.
14 S. Levesque, Y. St-Denis, B. Bachand, P. Preveille, L. Leblond, P. D.
Winocour, J. J. Edmunds, J. R. Rubin and M. A. Siddiqui, Bioorg. Med.
Chem. Lett., 2001, 11, 3161.
15 A. Linusson, J. Gottfries, T. Olsson, E. Ornskov, S. Folestad, B. Norden
and S. Wold, J. Med. Chem., 2001, 44, 3424.
16 D. Noteberg, J. Bra˚nalt, I. Kvarnstro¨m, M. Linschoten, D. Musil, J. E.
Nystro¨m, G. Zuccarello and B. Samuelsson, J. Med. Chem., 2000, 43,
1705.
B. B. Chen, H. G. Stenmark, H. W. Wu, D. P. Spangler, M. Clare, B. N.
Desai, I. K. Khanna, M. N. Nguyen, T. Duffin, V. W. Engleman, M. B.
Finn, S. K. Freeman, M. L. Hanneke, J. L. Keene, J. A. Klover, G. A.
Nickols, M. A. Nickols, C. N. Steininger, M. Westlin, W. Westlin, Y. X.
Yu, Y. P. Wang, C. R. Dalton and S. A. Norring, Bioorg. Med. Chem.
Lett., 2006, 16, 839.
25 W. J. Pitts, J. Wityak, J. M. Smallheer, A. E. Tobin, J. W. Jetter, J. S.
Buynitsky, P. P. Harlow, K. A. Solomon, M. H. Corjay, S. A. Mousa,
R. R. Wexler and P. K. Jadhav, J. Med. Chem., 2000, 43, 27.
26 J. A. Wendt, H. W. Wu, H. G. Stenmark, M. L. Boys, V. L. Downs, T. D.
Penning, B. B. Chen, Y. P. Wang, T. Duffin, M. B. Finn, J. L. Keene,
V. W. Engleman, S. K. Freeman, M. L. Hanneke, K. E. Shannon, M. A.
Nickols, C. N. Steininger, M. Westlin, J. A. Klover, W. Westlin, G. A.
Nickols and M. A. Russell, Bioorg. Med. Chem. Lett., 2006, 16, 845.
27 A. F. AbdelMagid, K. G. Carson, B. D. Harris, C. A. Maryanoff and
R. D. Shah, J. Org. Chem., 1996, 61, 3849.
28 J. S. Yang, Y. D. Lin, Y. H. Chang and S. S. Wang, J. Org. Chem., 2005,
70, 6066.
17 M. R. Wiley, N. Y. Chirgadze, D. K. Clawson, T. J. Craft, D. S.
GiffordMoore, N. D. Jones, J. L. Olkowski, L. C. Weir and G. F. Smith,
Bioorg. Med. Chem. Lett., 1996, 6, 2387.
29 J. J. Vanden Eynde, A. Mayence, M. T. Johnson, T. L. Huang, M. S.
Collins, P. D. Walzer, M. T. Cushion and I. O. Donkor, Med. Chem.
Res., 2005, 14, 143.
18 R. A. Friesner, J. L. Banks, R. B. Murphy, T. A. Halgren, J. J. Klicic,
D. T. Mainz, M. P. Repasky, E. H. Knoll, M. Shelley, J. K. Perry, D. E.
Shaw, P. Francis and P. S. Shenkin, J. Med. Chem., 2004, 47, 1739.
19 S. Saitton, A. L. Del Tredici, N. Mohell, R. C. Vollinga, D. Bostro¨m, J.
Kihlberg and K. Luthman, J. Med. Chem., 2004, 47, 6595.
20 P. Rocca, C. Cochennec, F. Marsais, L. Thomas-dit-Dumont, M.
Mallet, A. Godard and G. Queguiner, J. Org. Chem., 1993, 58, 7832.
21 P. Knochel, W. Dohle, N. Gommermann, F. F. Kneisel, F. Kopp, T.
Korn, I. Sapountzis and V. A. Vu, Angew. Chem., Int. Ed., 2003, 42,
4302.
22 D. B. Dess and J. C. Martin, J. Org. Chem., 1983, 48, 4155.
23 C. F. Koelsch and A. G. Whitney, J. Org. Chem., 1941, 6, 795.
24 M. L. Boys, L. A. Schretzman, N. S. Chandrakumar, M. B. Tollefson,
S. B. Mohler, V. L. Downs, T. D. Penning, M. A. Russell, J. A. Wendt,
30 M. A. Ismail, R. Brun, J. D. Easterbrook, F. A. Tanious, W. D. Wilson
and D. W. Boykin, J. Med. Chem., 2003, 46, 4761.
31 For experimental procedure see compound 17a. 1H NMR and 13C
NMR for compounds 11 and 18a are presented in the ESI.† HRMS
for compound 11 (calcd for C30H37N4O4 (M + H) 517.2815, found
517.2827) and compound 18a (calcd for C30H37N4O5 (M + H) 533.2764,
found 533.2764).
32 J. W. Nilsson, I. Kvarnstro¨m, D. Musil, I. Nilsson and B. Samulesson,
J. Med. Chem., 2003, 46, 3985.
33 A. G. W. Leslie, Acta Crystallogr., 1999, D55, 1696.
34 S. Bailey, Acta Crystallogr., 1994, D50, 760.
35 G. N. Murshudov, A. A. Vagin, A. Lebedev, K. S. Wilson and E. J.
Dodson, Acta Crystallogr., 1999, D55, 247.
36 P. Emsley and K. Cowtan, Acta Crystallogr., 2004, D60, 2126.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 2599–2605 | 2605
©