Design, synthesis and evaluation of graftable thrombin inhibitors for the preparation of blood-compatible polymer materials

Article abstract of DOI:10.1039/b510239a

Piperazinyl-amide derivatives of N-α-(3-trifluoromethyl- benzenesulfonyl)-L-arginine (1) were synthesized as graftable thrombin inhibitors. The possible disturbance of biological activity due to a variable spacer-arm fixed on the N-4 piperazinyl position was evaluated in vitro, against human a-thrombin, and in blood coagulation assay. Molecular modelling (in silico analysis) and X-ray diffraction studies of thrombin-inhibitor complexes were also performed. The fixation of bioactive molecules on poly(butylene terephthalate) (PBT) and poly(ethylene terephthalate) (PET) membranes was performed by wet chemistry treatment and evaluated by XPS analysis. Surface grafting of inhibitor Id improved the membrane hemocompatibility by reducing blood clot formation on the modified surface. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b510239a

View Article Online / Journal Homepage / Table of Contents for this issue
A R T I C L E
Design, synthesis and evaluation of graftable thrombin inhibitors for
the preparation of blood-compatible polymer materials
Claudio Salvagnini,a Catherine Michaux,b Julie Remiche,c Johan Wouters,b Paulette Charlierc
and Jacqueline Marchand-Brynaert*a
a Unite´ de Chimie Organique et Me´dicinale, Universite´ catholique de Louvain, Baˆtiment
Lavoisier, place Louis Pasteur 1, B-1348, Louvain-la-Neuve, Belgium.
E-mail: marchand@chim.ucl.ac.be; Fax: +32-10-474168; Tel: +32-10-472740
b Laboratoire de Chimie Biologique Structurale, Faculte´s Universitaires Notre Dame de la Paix,
rue de Bruxelles, B-5000, Namur, Belgium
c Laboratoire de Cristallographie des Prote´ines, Universite´ de Lie`ge, Baˆtiment B5, Alle´e de la
chimie, B-4000, Sart-Tilman, Lie`ge, Belgium
Received 20th July 2005, Accepted 3rd October 2005
First published as an Advance Article on the web 19th October 2005
Piperazinyl-amide derivatives of N-a-(3-trifluoromethyl-benzenesulfonyl)-L-arginine (1) were synthesized as graftable
thrombin inhibitors. The possible disturbance of biological activity due to a variable spacer-arm fixed on the N-4
piperazinyl position was evaluated in vitro, against human a-thrombin, and in blood coagulation assay. Molecular
modelling (in silico analysis) and X-ray diffraction studies of thrombin-inhibitor complexes were also performed. The
fixation of bioactive molecules on poly(butylene terephthalate) (PBT) and poly(ethylene terephthalate) (PET)
membranes was performed by wet chemistry treatment and evaluated by XPS analysis. Surface grafting of inhibitor
1d improved the membrane hemocompatibility by reducing blood clot formation on the modified surface.
1. Introduction
2. Results
2.1 Design of the thrombin inhibitors
The design of biomaterials endowed with high blood compat-
ibility (i.e. preventing clot formation) represents a substantial
challenge.1 The usual strategies to improve hemocompatibility
of material devices consist of selective surface treatments like the
grafting of zwitterionic components as biomembrane mimics,2
coating with heparin3 or thrombomodulin,4 two natural an-
ticoagulants, and the immobilization of extracellular matrix
proteins or active small peptides for promoting the growth of
endothelial cells.5 Another approach, less exploited till now,
makes use of synthetic drugs that are active in preventing
blood clotting. For instance, a thrombin inhibitor derived from
Argatroban (MD-805)6 was graft-polymerized on the surface of
polyurethane membranes.7,8 A p-amino-benzamidine derivative
was covalently immobilized on maleic anhydride copolymer thin
films by nucleophilic substitution via a spacer.9
In our laboratory, we are interested in the biocompatibi-
lization of polymer materials by their surface derivatization
with biologically active molecules of synthetic origin, designed
on the basis of medicinal chemistry data. Polyesters, such as
poly(ethylene terephthalate) (PET) and poly(butylene tereph-
thalate) (PBT) were chosen as model substrates. For instance,
the surface grafting of peptidomimetics of the Arg-Gly-Asp
sequence on PET membranes resulted in materials showing
interesting cell adhesion properties.10,11 Now, we consider the
possibility to improve the blood compatibility of polyesters by
the covalent grafting of thrombin inhibitors. For that purpose,
we have synthesized and evaluated a novel family of thrombin
inhibitors equipped with a spacer arm for polymer surface
anchorage. This study reports on the design, preparation, surface
grafting and in vitro evaluation of the inhibitory activity and
anticoagulant effect of this family of compounds based on the
L-arginine template. Taking into account possible disturbance
due to the spacer arm, in silico evaluation of the interaction
with the protein active site has been performed, as well as the
X-ray diffraction study of one representative thrombin-inhibitor
complex.
Clot formation results from a complex sequence of biochemical
events including the blood coagulation cascade in which throm-
bin plays a central role. Thrombin is a serine protease involved
in the conversion of soluble fibrinogen into insoluble fibrin and
in the activation of platelet aggregation.12 This key protein has
become the principal target in the discovery of novel antithrom-
botic agents.13,14 Historically, two research lines have been
developed in parallel: (i) the steric inhibitors (non-covalently
bound) based on two simple lead compounds, namely N-tosyl-
L-arginine methyl ester (TAME, Fig. 1),15 and benzamidine;16
(ii) the electrophilic inhibitors (covalently bound) derived from
the tripeptide motif (D)-Phe-Pro-Arg,17 with the peptide arginal
Efegatran18 and the chloromethylene ketone PPACK,19 as leads.
The only two synthetic thrombin inhibitors, Argatroban20 and
Ximelagatran,21 available at the moment in the market for the
treatment of coagulation complications were developed from the
first class of inhibitors.
In our context of biomaterials, the surface bound inhibitor is
not allowed to be processed by the enzyme. We thus selected
one representative of the first class as a model compound.
We considered a simplified structure of Argatroban 1 (Fig. 1)
featuring the following characteristics: (i) the guanidyl function
of the L-arginine skeleton, for specific ionic interaction with Asp-
189 in the S1 enzyme pocket—since oral bioavailability14 is not a
relevant problem in the field of hemocompatible materials, this
strongly basic function could be maintained as a recognizing
function; (ii) the lipophilic sulfonamide moiety, for interaction
with the S2–S3 binding sites—this substituent was also used
to introduce a trifluoromethyl group as a useful spectroscopic
tag22 for XPS (X-ray photoelectron spectroscopy) analysis of the
biomaterial surface and quantification of the amount of grafted
inhibitor; (iii) the piperidinyl amide moiety (X = CH2), stable
towards processing by the active serine of thrombin and blood
esterases—it has been proven that the acid function placed on
the piperidine ring of Argatroban is not absolutely necessary for
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0
4 2 0 9
©

View Article Online
Fig. 1 Thrombin inhibitors.
activity.14 Moreover, by removing the carboxylic acid and methyl
group in the piperidine ring of Artagroban, we also avoided the
control of two chiral centers.
We further transformed the piperidine ring into a piperazine
(X = NH) with the view to introduce a versatile anchorage
point for different spacer arms, without creating a new chiral
center. Thus, we chose to fix the spacer arm on the amide
substituent; this was also the strategy selected by Ito and
Imanishi23 who transformed the acid function of Argatroban
into a polymerizable acrylamide residue.
The effect of the piperazine motif on the activity, and the
effect of the spacer arm on the positioning in the enzyme cavity,
had to be controlled before using molecules 1 in our biomaterial
application. Accordingly, we prepared the reference compound
1a, the parent compound 1b, and two derivatives 1c, and 1d,
bearing respectively a short and a long spacer. The terminal
function of the spacer is a primary amine for coupling to
polymers displaying activated hydroxyl or carboxyl functions,
following previously established protocols.22,24
2.2 Chemistry
The target molecules 1 were constructed from protected argi-
nine 2 (N-a-Boc-N-x-nitro-L-arginine) in five steps outlined in
Scheme 1. After activation of the carboxyl function, piperidine
(X = CH2), N-(benzyloxycarbonyl)piperazine (X = NCO2Bn)
and piperazines 8 and 13 (X = N–spacer–NH–CO2PNB) were
coupled to give the corresponding amides 3a–d (Table 1). After
deprotection of the tert-butoxycarbonyl group (Boc), reaction
of the free amine 4 with m-(trifluoromethyl)benzenesulfonyl
chloride led to compounds 5a–d (Table 1). The final compounds
1a–d (Table 1) were isolated as acetate salts after cleavage of the
protective groups by catalytic hydrogenation.
Piperazines 8 and 13 equipped with the protected amino-
propyl and aminododecyl spacers were independently pre-
pared according to Scheme 2, by using classical chemistry.
The piperazine carrying the C3 spacer was prepared from 3-
bromopropylamine by first protecting the amino group, then
substitution of the bromine with Boc-piperazine, and finally
Boc deprotection. For the preparation of the piperazine carrying
the C12 spacer, the sequence starting from 12-aminododecanoic
Scheme 1 General synthesis of compounds 1a–d: (i) iBuO2CCl, Et3N,
THF, −20 ◦C, 15 min; (ii) azacyclohexane, THF, −20 ◦C to 20 ◦C, 1 h;
(iii) TFA–CH2Cl2 1 : 1, 20 ◦C, 2 h; (iv) 3-trifluoromethyl-benzenesulfonyl
chloride, Et3N, CH2Cl2, 0 ◦C to 20 ◦C, 4 h; (v) H2, Pd/C, EtOH–AcOH
(3 : 1), 50 ◦C, 12 h.
acid involved the protection of the amino group, the reduction
of the carboxylic acid to the corresponding alcohol, and its
activation for the substitution with Boc-piperazine.
Intermediates and final products of Schemes 1 and 2 were
fully characterized by IR, NMR and Mass spectroscopy (see
Experimental).
2.3 Biological activity
The inhibition of human thrombin (Ki) by compounds 1a–
d was measured spectrophotometrically using as a competi-
tor the chromogenic substrate H-D-phenylalanyl-L-pipecolyl-L-
arginine-p-nitroanilide dihydrochloride (S-2238). The thrombin
activity was evaluated from the hydrolysis rate of S-2238
measured at 405 nm (production of p-nitroaniline). All the
compounds were active at the micromolar level, similarly to
TAME (Table 2). The presence of a basic function on the amide
substituent, and the use of this function to fix an amino-alkyl
chain, did not significantly modify the activity.
Table 1 Yields of inhibitors and intermediates
Xa
3, (%)
4, (%)
5, (%)
1, (%)
CH2
N–CO2Bn
N–H
N–(CH2)3–NH–CO2–PNB
N–(CH2)3–NH2
N–(CH2)12–NH–CO2–PNB
N–(CH2)12–NH2
a, 70
b, 63
—
c, 60
—
d, 65
—
a, quant.
b, quant.
—
c, quant.
—
d, quant.
—
a, 70
b, 70
—
c, 75
—
d, 75
—
a, 95
—
b, 90
—
c, 90
—
d, 90
The effect of compounds 1a–d on clot formation on polymer
materials surface was also evaluated with human whole blood.
Two representative substrates were considered, a cell culture
support made of PET and a blood filtration membrane made of
a Bn = CH2Ph; PNB = CH2–Ph–p-NO2.
4 2 1 0
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0

View Article Online
Scheme 2 Synthesis of the piperazine linked spacer arms: (i) NaOH 1 N, p-nitro benzylchloroformate, THF, 0–5 ◦C; (ii) Boc-piperazine, CH2Cl2,
DIPEA, 60 ◦C; (iii) HCl 1 N, HOAc; (iv) BH3·THF, MeOH, HCl; (v) MeSO2Cl, pyridine, CH2Cl2, 0–5 ◦C; (vi) NaI, Boc-piperazine, DIPEA, CH3CN,
85 ◦C; (vii) TFA–CH2Cl2 1 : 1.
Table 2 Inhibition of human thrombin
2.4 Docking study
Our purpose was to localize, by molecular modelling, the pos-
sible positions of the spacer arm of compounds 1 in the thrombin
cavity. Indeed, after grafting on a solid support via the spacer
terminus, the inhibitor should still be able to enter and fit in the
enzyme active site.
Thrombin is formed by an A chain of 36 amino acids and
a B chain of 259 amino acids connected by a disulfur bridge.
This enzyme contains the catalytic triad (Asp-102, His-57, Ser-
195) characteristic of the chymotrypsin family. Moreover, three
important binding pockets have been identified: the specific
pocket S1 with Asp-189, the hydrophobic proximal pocket S2
(also called P-pocket) and a larger hydrophobic distal pocket S3
(also called D-pocket).25,26
In our docking study, the coordinates of the human thrombin
crystal structure complexed with Argatroban (1DWC in PDB)
were used.27 In order to test the two docking algorithms
(Autodock28 and Gold29), Argatroban was firstly docked in
thrombin. The resulting complexes are in good agreement with
the crystal structure (data not shown).
Compound
Ki/lMa
1a
4–8
1b
17–24
8–18
25–39
5–6
1c
1d
TAME
a Reported values are the result of two independent measurements.
PBT. The weight of clot was measured on native polymers and on
polymers coated with inhibitors, under standardized conditions,
using heparin (Hep) as a control (Table 3).
Compounds 1a–d were found to be active anticoagulants in
our experimental conditions (about 100 nmol cm−2 of inhibitor
on apparent support surface) reducing the weight of clot on both
substrates by about 25%, while TAME was similarly active on
PBT only.
The most active compound of the series, 1c, was then docked
using the two methods. To account for amino acids flexibility
of the active site, the solutions common to the two programs
were then refined with the Discover3 module.30 Only the three
water molecules close to the catalytic triad in the crystal complex
were considered during the run. The geometry of the ligand was
optimized in vacuo using quantum mechanics and particularly
the HF/6-31+G(d,p). method and the 6-31+G** basis set
(Gaussian98 program).31 The guanidyl and terminal amine
groups of 1c were considered as protonated at physiological
pH. The theoretical pKa of the piperazine moiety was calculated
with the PALLAS program (CompuDrug Chemistry), and gave
a value of 7.6 for the alkylamine. That means that about 50% of
the ligand is not protonated. Here, we have considered this form
but the conclusions are the same for the protonated state of the
piperazine moiety.
Table 3 Inhibitor coating effect on blood clot formation
Sample
Clot weight/mga
PET
27.5 1.5
11.4 0.5
20.1 1.4
19.6 4.2
21.6 3.0
27.0 2.0
29.9 1.0
8.0 3.2
20.5 1.0
22.0 3.2
23.8 1.5
23.0 3.1
PET Hep
PET 1a
PET 1c
PET 1d
PET TAME
PBT
PBT Hep
PBT 1a
PBT 1c
PBT 1d
PBT TAME
a Results are the mean of 3 independent experiments
deviation.
standard
Two binding modes, being the two best ranked solutions of
both programs, were identified (Fig. 2). In the first mode (A,
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0
4 2 1 1

View Article Online
Fig. 2 Two potential binding modes of 1c into the binding site of human thrombin, as deduced by docking and energy minimization studies. Water
molecules and hydrogen atoms were omitted for clarity.
“internal”), the spacer points into the binding site of the enzyme,
while in the second one (B, “external”), the spacer points toward
the surface of the enzyme.
to the different pockets of the active site of a-thrombin was
very similar to that of inhibitor MD-805 (Argatroban) as was
described in structure 1DWC27 and was close to the “internal
mode” described above (see superimposition in Fig. 4). Such
experimental observation validates our theoretical approach.
The orientation of compound 1c in the active site of a-
thrombin is described in Fig. 3B and the compared distances in
inhibitor binding are listed in Table 4. The guanidyl group (N22,
N21 and N19) was oriented to the S-pocket but was unable to
form H-bonds with the OD1 and OD2 groups of the Asp-189
side chain, as it is in the a-thrombin–Argatroban complex or
in the “internal mode”. The sulfonamide moiety lay in the D-
pocket with the O24 atom of the sulfone group H-bonded to
the amide backbone of Gly-219. The carbonyl O15 and amide
N12 were both H-bonded to the peptidic backbone of Gly-
216, as described in the a-thrombin–Argatroban complex. In
the “internal mode”, Gly-216 was slightly shifted. Its carbonyl
In the “internal” mode, compound 1c approximately lay in
the same orientation as Argatroban. The guanidyl function was
close to Asp-189 in the S-pocket and interacted by H-bonding
with Gly-219 and a water molecule. The lipophilic sulfonamide
moiety lay in the D-pocket and the sulfone group interacted with
Gly-219. The piperazine group and its spacer arm were buried in
the P-pocket where H-bond was observed between Ser-195 and
the amine group. The amide carbonyl moiety was also involved
in H-bonding with Gly-216.
In the “external” mode, the lipophilic sulfonamide and
piperazine moieties were located in the P- and D-pockets
respectively. The guanidyl function still lay in the S-pocket. The
sulfone group was involved in H-bonding with Ser-195 and His-
57, and the amide carbonyl with Gly-219. The trifluoromethyl
group interacted with Lys-60f and a water molecule; the
protonated amine interacted with Glu-97a and was close to
the enzyme surface. Moreover, p–p stacking interaction was
observed between Trp-60d and the phenyl group.
˚
backbone is moved about 1 A from its position observed in the a-
thrombin–Argatroban complex and further from the inhibitor
amide N12 atom, resulting in the loss of the hydrogen bond.
The loss of one hydrogen bond induced a different positioning
of the piperazine cycle, bound in the P-pocket, in comparison
with the two X-ray structures where it makes a hydrophobic
interaction with the catalytic His-57. The spacer with its amine
terminal group (N9) also had a different orientation in the X-ray
structure compared to the model. While in the “internal mode”
structure we observed H-bonding between the OG hydroxyl
function of the active serine and the OE1 and OE2 groups of the
Glu-192 side chain, the X-ray structure revealed an unexpected
interaction with the NZ terminal amine group of the Lys-60f
side chain. Since this spacer does not exist in Argatroban, this
kind of interaction was unpredictable by theoretical methods.
The two complexes were found to be stable and have very
close binding energies, −79.2 and −79.6 kcal mol−1 respectively.
Thus, the docking studies clearly showed that both orientations
are possible from the energetic point of view.
2.5 X-Ray diffraction
Several crystal structures27,32–36 of thrombin complexed with
covalent and non -covalent inhibitors have been disclosed, since
the seminal work of Bode et al.19 A single crystal of the novel
compound 1c and human a-thrombin was prepared by using
the hanging drop vapour diffusion technique, in the presence
˚
of hirudin peptide fragment; this crystal diffracted to 1.65 A
2.6 Inhibitors grafting on surface
resolution.
In the X-ray complex structure, a well-defined electron density
for the enzyme at the catalytic site and for the inhibitor was found
(Fig. 3A). Not surprisingly, the binding mode of compound 1c
Transformation of hydroxyl group into sulfonic ester is a well-
knownprocessinorganicsynthesisforalcoholactivationtowards
nucleophilic substitution. This strategy was already successfully
4 2 1 2
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0

View Article Online
Fig. 3 (2Fo–Fc) electron density map of compound 1c, at the active site of the human a-thrombin complex, at 1.0r contour level (A). Interactions
of compound 1c, at the active site of the human a-thrombin complex (B). The inhibitor is in the same orientation as in Fig. 2A. The CPK colour
code has been used for both protein and inhibitor atoms, with compound 1c atoms drawn in larger spheres. Shortest contacts with the protein within
˚
a distance of less than 4.1 A are shown with dotted lines.
Fig. 4 Superimposed active sites of crystal structure of the a-thrombin complex with compound 1c (in CPK color code), the “internal” binding
mode model (Fig. 2A) (in magenta) and the crystal structure of the a-thrombin complex with compound MD-805 (Argatroban) from structure
1DWC (in cyan). Compound 1c and MD-805 atoms are drawn in larger sphere.
used in our laboratory for polymer functionalization.22 Chain-
end hydroxyl groups of PET and PBT membranes were activated
with p-toluenesulfonyl chloride (TsCl) and pyridine. The incu-
bation of activated membranes with inhibitor solutions should
provide sulfonyl ester displacement and covalent fixation of the
active molecules on the surfaces (Scheme 3). The efficiency of
the binding was verified by XPS analysis. Blank samples were
prepared by processing samples without the activating agent
p-toluenesulfonyl chloride. The F detection on blank samples
could be attributed to unspecific adsorption. The corrected
values of F/C (difference between activated and blank samples)
allowed the quantification of grafted inhibitors (Table 5). The
fixation of compound 1c on both surfaces was not successful: on
PET the difference in F detection on blank and activated samples
(act) was not significant, while on PBT no F was detected. On
the other hand, compound 1d resulted in a good level of fixation,
with a higher ratio on PET than on PBT. From the F/C corrected
values we could determine the concentration of grafted signals
˚
in the interface domain analyzed by XPS (50–100 A depth or
about 10 atomic layers). These values (% derivatization) were
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0
4 2 1 3

View Article Online
Table 4 Compared distances in inhibitor binding
Table 6 Inhibitor grafting effect on blood clot formation
Clot weight/mga
˚
Distance/A
Inhibitor/protein atom
identification
PET 1d blank
PET 1d act
PET 1c blank
PET 1c act
PBT 1d blank
PBT 1d act
PBT 1c blank
PBT 1c act
29.7 0.4
27.5 1.5
31.3 3.4
31.1 5.7
28.5 4.1
21.7 4.5
29.3 1.4
29.7 0.6
1DWC
1W7G
Internal mode
Xa N22
Asp-189 OD1
Asp-189 OD2
Ala-190 O
3.05
2.93
3.63
4.06
4.50
5.07
5.11
4.36
3.76
5.48
3.55
3.76
3.01
2.85
2.99
—
3.58
4.08
4.12
3.31
4.93
5.37
6.11
5.19
3.44
6.43
3.26
3.21
3.02
2.73
—
3.33
2.93
2.97
3.80
2.95
4.28
2.73
3.97
4.62
5.01
3.11
3.02
3.29
5.05
—
Trp-215 O
Gly-219 O
Xa N21
Asp-189 OD1
Asp-189 OD2
Ala-190 O
a Results are the mean of 3 independent experiments
deviation.
standard
Trp-215 O
Gly-219 O
Ala-190 O
Xa N19
Xa O24
Xa O15
Xa N12
Xa O34
Xa N9
The PBT membrane is tortuous and highly porous; the lack of
crystallographic and density data in this case did not allow a
similar calculation of surface molecular concentration.
The different morphology of the two polymers may explain
the systematically lower fixation ratios observed in the web-like
PBT membrane in comparison with the smooth PET membrane.
Grafting phenomena occur in the solvent accessible surface of
the membrane (open surface) that, in the case of the particular
morphology of the considered PBT membrane, may not coincide
with the XPS analyzed portion of the sample. Indeed, grafted
signals, mostly present in the polymer matrix for instance, could
be underestimated by the spectroscopic analysis.
The evaluation of materials hemocompatibility was per-
formed by measuring the weight of clot formed on the supports
after static incubation with blood in standardized conditions.
The blood clot formation on samples grafted with compound
1d was found to be systematically lower than on blank samples
(Table 6). The decrease in blood clot formation was more
important on PBT (around 24%) than on PET (7%). No effect
on blood clot formation was observed on samples treated with
compound 1c. This finding was not surprising as low fixation of
1c was detected on PET and no fixation at all was observed on
PBT.
Gly-219 N
Gly-216 N
Gly-216 O
Ser-195 OG
Ser-195 OG
His-57 O
5.21
3.50
2.81
6.59
4.80
2.96
4.71
4.48
2.99
2.63
—
—
—
—
Lys-60f NZ
Glu-192 OE1
Glu-192 OE1
a The inhibitor name is abbreviated by X and represents compound MD-
805 (Argatroban) in structure 1DWC and compound 1c (this study) in
structure 1W7G and “internal mode”, respectively. Atom numbers refer
to Fig. 4 atom labels.
transformed into pmol cm−2 as previously described for PET.37
Indeed this membrane is dense, regular and semi-crystalline.
3. Conclusion
Although significantly less active than Argatroban, the designed
inhibitors 1 were found to be sufficiently active for our bio-
material application. The lack of condensed piperidine ring
on the N-terminal sulfonyl benzene group, and carboxyl and
methyl groups on the C-terminal piperidine ring, in fact caused
a weaker inhibition potency. However the activities were found
to be similar to amide derivatives of N,a-substituted-L-arginine
precursors of Argatroban15 and benzamidine derivatives devel-
oped for material hemocompatibilization.9
Our novel thrombin inhibitors were designed to be grafted
on the surface of biomaterials via a spacer arm, either short
(compound 1c) or long (compound 1d). From docking and X-
ray studies, two potential binding modes of 1c in thrombin
were identified. The docking studies clearly showed that both
orientations are possible, from an energeticpoint ofview. Indeed,
Scheme 3 Inhibitors grafting on polyesters surface: (i) TsCl, pyridine,
acetone, 20–60 ◦C, 1 h; (ii) NH2R (inhibitors), PBS (pH 7.2)–CH3CN 1 :
1, 2 h, 20 ◦C. PET: n = 2; PBT: n = 4.
Table 5 Atomic surface composition of modified polymers analyzed by XPS
C 1s
O 1s
F 1s
N 1s
Si 2p
F/C
Corr F/C
Der% a
pmol cm−2b
PET 1c blank
PET 1c act
PET 1d blank
PET 1d act
PBT 1c blank
PBT 1c act
PBT 1d blank
PBT 1d act
72.1
73.4
69.0
69.5
73.5
74.5
72.1
71.2
27.0
25.6
27.7
25.5
26.2
25.2
24.2
25.2
0.2
0.3
0.7
1.3
0.0
0.0
0.2
0.6
0.2
0.5
1.5
1.9
0.2
0.2
0.8
1.0
0.5
0.3
1.1
1.8
0.1
0.1
2.8
2.1
0.003
0.004
0.010
0.019
0
0.001
0.009
0
0.4
3.3
0
11
94
0
0
0.003
0.008
0.005
1.9
—
a Percentage of covalently derivatized monomer units. b Concentration of surface fixed inhibitors.
4 2 1 4
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0

View Article Online
the two complexes are stable and have very similar binding
energies. The second mode, called “external mode”, where the
molecular spacer is close to the enzyme surface, showed that
the grafted inhibitor would still be able to bind the active site
of thrombin. Docking studies of analogues of 1c bearing longer
spacers ((CH2)4 and (CH2)5) were also performed and confirmed
the possibility of observing both the A (internal) and B (external)
modes of binding (data not shown). Interestingly, for those
longer spacers, even in the so-called “internal mode” observed
experimentally, the spacer was also able to join the enzyme
surface by displacing Trp-60d (data not shown). The same
situation should occur with compound 1d bearing a (CH2)12
spacer arm. Moreover, from biological data, we observed that
the substitution of a CH2 group with a NH function in position
4 of the piperidine ring, as well as the introduction in the same
position of a spacer arm, does not change significantly the
activity of the inhibitor (Table 1).
All the experimental and theoretical results confirmed our
initial hypothesis that an inhibitor carrying a suitable spacer
arm in position 4 of the piperazinyl-amide moiety should be
able to enter and fit the active site of thrombin providing the
specific biological activity that we required for the development
of blood-compatible materials. The grafting of our synthetic
thrombin inhibitors on a polyester surface was performed and
validated by XPS thanks to the incorporation of a fluorine tag
into their structures: fixation of compound 1d with a long spacer
was more efficient than the grafting of 1c equipped with a short
arm. Moreover, due to the morphology of the polymers chosen
as models for this study, the ratio of PET surface derivatization
was higher than that of PBT.
(0.43 g, 3.13 mmol, 1 equiv) was then added dropwise at −20
◦
to −25 C. After 30 min azacyclohexane (3.13 mmol, 1 equiv)
in dry THF (2 mL) was added dropwise at −20 ◦C and stirred
for an additional 20 min. The solution was allowed to warm
up to room temperature and stirred for 1 h. The solvent was
then evaporated and the solid residue partitioned between
EtOAc and brine. The organic layer was dried over MgSO4 and
concentration under vacuum gave crude 3 that was purified by
column chromatography on SiO2 (EtOAc) with a yield of about
65%.
5-(3-Nitroguanidino)-2(S)-tert-butoxycarbonylamino-1-(piper-
idin-1-yl)pentan-1-one (3a). White amorphous solid; Rf = 0.5
(SiO2, EtOAc, UV); IR mmax (CH2Cl2 liquid film on NaCl) 3292,
2932, 2857, 1700, 1624, 1522, 1445, 1366, 1254, 1165 cm−1; dH
(300 MHz, CDCl3, Me4Si) 1.42 (9H, s), 1.56 (2H, m), 1.65 (6H,
m), 1.74 (2H, m), 3.27 (1H, m), 3.36 (2H, m), 3.52 (1H, m),
3.59 (2H, m), 4.58 (1H, m), 5.90 (1H, d, J = 7.0 Hz), 7.87 (2H,
s), 8.95 (1H, s); dC (50 MHz, CDCl3, Me4Si) 24.2, 24.3, 25.3,
26.2, 28.3, 31.3, 40.1, 43.2, 46.4, 48.5, 80.3, 156.6, 159.3, 169.5;
MS (ESI) m/z 387.0 (M+, 100%); HRMS (ESI) m/z: 425.1912
(calcd for C16H30N6O5K = 425.1915).
5-(3-Nitroguanidino)-2(S)-tert-butoxycarbonylamino-1-[4-(ben-
zyloxycarbonyl)piperazin-1-yl]pentan-1-one
(3b). White
amorphous solid; Rf = 0.1 (SiO2, EtOAc, UV); IR mmax (KBr)
3305, 2979, 2932, 1700, 1635, 1430, 1251, 1231, 1165 cm−1; dH
(300 MHz, CDCl3, Me4Si) 1.42 (9H, s), 1.64 (2H, m), 1.72 (2H,
m), 3.20 (1H, m), 3.30–3.60 (8H, m), 3.65 (1H, m), 4.57 (1H,
m), 5.13 (2H, s), 5.80 (1H, m), 7.34 (5H, m), 7.74 (2H, m), 8.87
(1H, m); dC (50 MHz, CDCl3, Me4Si) 24.5, 28.3, 30.7, 40.5,
41.9, 43.6, 43.7, 45.2, 49.2, 67.5, 80.2, 127.8, 128.1, 128.5, 136.3,
154.9, 156.1, 159.5, 170.5; MS (APCI) m/z 522.0 (M+, 100%),
465.9 (50%), 422.1 (7%); Anal. calcd. for C23H35N7O7·0.5H2O:
C, 52.18; H, 6.63; N, 17.98%. Found: C, 52.32; H, 6.85; N,
18.04%.
The grafting of compound 1d on both PET and PBT
membranes was found to reduce significantly the blood clot
formation. The inhibitor bound on the surface is still active
and its surface concentration of about 90 pmol cm−2 seems
to be sufficient to inhibit the coagulation cascade. Modifi-
cation of the grafting protocol and other surface activation
techniques are under investigation to increase the surface
concentration of bioactive signals and hopefully improve the
materials performance. Nevertheless, the proposed technique
appeared to be a very promising method for the development
of hemocompatible polymer materials, by mild, non-aggressive
wet-chemistry treatments.
{(S)-1-[4-(3-(4-Nitrobenzyloxycarbonyl)aminopropyl)pipera-
zine-1-carbonyl]-4-nitroguanidino-butyl}carbamic acid tert-
butyl ester (3c). Beige amorphous solid; Rf = 0.1 (SiO2,
isopropanol–CH2Cl2 10 : 90, UV); IR mmax (CH2Cl2 liquid film
on NaCl) 3305, 2926, 2857, 1704, 1629, 1521, 1445, 1347,
1254 cm−1; dH (500 MHz, CDCl3, Me4Si) 1.43 (9H, s), 1.63 (2H,
m), 1.71 (4H, m), 2.42 (2H, m), 2.46 (2H, m), 2.50 (2H, m),
3.25 (1H, m), 3.30 (2H, m), 3.43 (2H, m), 3.61 (3H, m), 4.56
(1H, m), 5.18 (2H, s), 5.58 (1H, t, J = 5.7 Hz), 5.81 (1H, d, J =
8.3 Hz), 7.50 (2H, d, J = 8.4 Hz), 7.59 (2H, m), 8.21 (2H, d,
J = 8.4 Hz), 8.68 (1H, m); dC (75 MHz, CDCl3, Me4Si) 24.2,
26.0, 28.2, 31.2, 40.1, 42.0, 45.3, 48.6, 52.4, 52.9, 56.1, 64.9, 80.4,
123.6, 128.0, 144.1, 147.4, 155.8, 156.5, 159.3, 169.8; MS (APCI)
m/z 624.2 (M+, 100%); Anal. calcd. for C26H41N9O9·H2O: C,
48.66; H, 6.76; N, 19.65%. Found: C, 48.62; H, 6.16; N, 19.05%.
4. Experimental
4.1 Chemistry
Reagents and solvents were purchased from Acros Chim-
ica, Aldrich or Fluka. Tetrahydrofuran was dried over
sodium/benzophenone, then distilled. Column chromatography
was carried out with silica gel 60 (70–230 mesh ASTM) supplied
by Merck. The IR spectra were recorded with a Perkin-Elmer
1710 instrument, only the most significant adsorption bands
being reported. The mass spectra were obtained with a Finnigan
MAT TSQ-70 instrument. The high resolution mass spectra
(HRMS) were performed at the University of Mons, Belgium
(Professor R. Flammang). The microanalyses were performed
at the Christopher Ingold Laboratories of University College,
London (Dr A. Stones). The melting points were determined
with an Electrothermal microscope and are uncorrected. The 1H
and 13C NMR spectra were recorded on Varian Gemini 300 (at
300 MHz for proton and 75 MHz for carbon) or Bruker AM-
500 spectrometers (at 500 MHz for proton and 125 MHz for
carbon); the chemical shifts were reported in ppm (d) downfield
from tetramethylsilane (TMS).
{12-[4-((S)-2-tert-Butoxycarbonylamino-5-nitroguanidino-
pentanoyl)piperazin-1-yl]-dodecyl}carbamic acid 4-nitrobenzyl
ester (3d). Yellow amorphous solid; Rf = 0.5 (SiO2, EtOAc,
UV); IR mmax (CH2Cl2 liquid film on NaCl) 3354, 2928, 2855,
1708, 1650, 1523, 1437, 1347, 1254 cm−1; dH (300 MHz, CDCl3,
Me4Si) 1.26 (12H, m), 1.43 (9H, s), 1.49 (4H, m), 1.64 (2H, m),
1.75 (2H, m), 2.36 (2H, t, J = 7.2 Hz), 2.45 (4H, m), 3.19 (2H,
q, J = 6.6 Hz), 3.26 (1H, m), 3.46 (2H, m), 3.64 (3H, m), 4.56
(1H, m), 4.85 (1H, m), 5.19 (2H, s), 5.85 (1H, d, J = 8.2 Hz),
7.51 (2H, d, J = 8.7 Hz), 7.80 (2H, m), 8.21 (2H, d, J = 8.7
Hz), 8.84 (1H, m); dC (50 MHz, CDCl3, Me4Si) 26.8, 27.5, 28.0,
29.3, 29.6, 29.9, 41.2, 43.6, 52.9, 58.7, 65.0, 71.5, 123.6, 128.0,
144.3, 147.6, 155.5, 155.7, 159.8, 168.0; MS (APCI) m/z 750.3
(M+, 100%); Anal. calcd. for C35H59N9O9·0.5H2O: C, 55.39; H,
7.97; N, 16.61%. Found: C, 55.70; H, 8.00; N, 16.14%.
4.1.1 General procedure for the preparation of 5-(3-nitrogua-
nidino)-2(S)-tert-butoxycarbonylamino-1-(azacyclohexan-1-yl)-
pentan-1-one (3a–c). N-a-Boc-N-a-nitro-L-arginine (1.00 g,
3.13 mmol, 1 equiv) was dissolved in dry THF (10 mL) and
Et3N (0.32 g, 3.13 mmol, 1 equiv). Isobutyl chloroformate
4.1.2 General procedure for the preparation of 2(S)-amino-5-
nitroguanidino-1-(azacyclohexan-1-yl)pentan-1-one (4). Prod-
uct 3 (1.25 mmol) was dissolved in a 1 : 1 mixture of TFA–CH2Cl2
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0
4 2 1 5

View Article Online
(10 mL) and stirred at room temperature for 3 h. The solution
was concentrated at reduced pressure and the obtained orange
gel was suspended and triturated in Et2O until precipitation
of the trifluoroacetate salt as a crystalline solid occurred. The
precipitate was washed with Et2O and used directly in the
next reaction (quantitative yield). For spectroscopic analysis the
product was converted to the free amine by dissolution in NaOH
1 N, and extraction of the aqueous solution with EtOAc (about
90% product recovery).
7.5 Hz), 8.06 (1H, d), 8.07 (1H, s), 8.79 (1H, m); dC (50 MHz,
CDCl3, Me4Si) 23.9, 24.6, 25.2, 26.2, 30.1, 40.5, 43.5, 46.4, 52.6,
123.2 (JCF = 272.6 Hz), 124.2 (JCF = 3.8 Hz), 129.4 (JCF
=
3.4 Hz), 130.0, 130.8, 131.5 (JCF = 33.3 Hz), 140.9, 159.5,
168.3; MS (APCI) m/z 495.2 (M+, 100%); Anal. calcd. for
C18H25N6O5SF3·0.5H2O: C, 42.94; H, 5.40; N, 16.69; S, 6.37%.
Found: C, 42.78; H, 4.97; N, 16.23, S, 6.77%.
3-Trifluoromethyl-benzenesulfonic acid {4-nitroguanidino-
1(S)-[4-(benzyloxycarbonyl)-piperazin-1-yl-carbonyl]butyl}amide
(5b). White amorphous solid; Rf = 0.2 (SiO2, EtOAc, UV);
IR mmax (KBr) 3234, 2927, 2863, 1699, 1636, 1429, 1327, 1231,
1166, 1132 cm−1; dH (300 MHz, CDCl3, Me4Si) 1.60 (2H, m),
1.78 (2H, m), 3.0–3.6 (10H, m), 4.18 (1H, m), 5.08 (2H, s), 6.77
(1H, br s), 7.30 (5H, m), 7.40 (2H, m), 7.60 (1H, t, J = 7.6 Hz),
7.78 (1H, d, J = 7.6 Hz), 8.02 (1H, d, J = 7.6 Hz), 8.04 (1H,
s), 8.60 (1H, br s); dC (50 MHz, CDCl3, Me4Si) 24.2, 29.9, 40.4,
41.8, 43.0, 43.5, 44.9, 52.6, 67.5, 123.1 (JCF = 273 Hz), 124.0,
2(S)-Amino-5-nitroguanidino-1-(piperidin-1-yl)pentan-1-one
(4a). White amorphous solid; IR mmax (KBr) 3292, 2937, 2857,
1617, 1439, 1257 cm−1; dH (300 MHz, D2O) 1.47 (2H, m), 1.64
(6H, m), 1.91 (2H, m), 3.29–3.42 (4H, m), 3.51 (1H, m), 3.67
(1H, m), 4.57 (1H, t, J = 5.7 Hz); dC (50 MHz, D2O) 23.9,
25.2, 25.6, 26.3, 27.5, 30.6, 40.4, 44.4, 47.2, 50.7, 159.1, 167.1;
MS (ESI) m/z 287.1 (M+, 100%); HRMS (ESI) m/z: 309.1643
(calcd for C11H22N6O3Na = 309.1651).
4-(2-(S)-Amino-5-nitroguanidino-pentanoyl)piperazine-1-car-
boxylic acid benzyl ester (4b). White solid; IR mmax (KBr) 3290,
3055, 2938, 1700, 1638, 1430, 1265, 1232 cm−1; dH (300 MHz,
CDCl3, Me4Si) 1.47 (2H, m), 1.70 (4H, m), 3.30 (2H, m),
3.35–3.65 (8H, m), 3.82 (1H, m), 5.15 (2H, s), 7.36 (5H, m),
7.80–8.40 (3H, br); dC (50 MHz, acetone, Me4Si) 25.1, 29.1,
41.4, 43.6, 44.7, 44.9, 46.6, 51.9, 68.2, 129.1, 129.2, 129.7, 138.5,
156.2, 161.2, 170.5; MS (APCI) m/z 422.4 (M+, 100%); HRMS
(ESI) m/z: 422.2172 (calcd for C18H28N7O5 = 422.2152).
{3-[4-(2-(S)-Amino-5-nitroguanidino-pentanoyl)piperazin-1-
yl]propyl}carbamic acid 4-nitrobenzyl ester (4c). Brown
amorphous solid; IR mmax (KBr) 3600–3200, 1701, 1638, 1551,
1329 cm−1; dH (300 MHz, D2O) 1.66 (2H, m), 1.90 (4H, m), 3.00
(1H, m), 3.20 (5H, m), 3.27 (3H, m), 3.59 (4H, m), 4.05 (1H,
m), 4.53 (1H, m), 5.17 (2H, s), 7.53 (2H, d, J = 8.1 Hz), 8.19
(2H, d, J = 8.1 Hz); dC (75 MHz, CDCl3, Me4Si) 25.1, 26.1,
29.7, 40.3, 40.7, 42.3, 45.2, 50.7, 52.7, 53.3, 56.3, 65.0, 123.6,
128.0, 144.2, 147.4, 155.8, 159.6, 173.2; MS (APCI) m/z 524.2
(M+, 100%); Anal. calcd. for C21H33N9O7: C, 48.18; H, 6.35; N,
24.08%. Found: C, 48.14; H, 6.43; N, 24.15%.
127.7, 127.8, 128.1, 128.4, 129.2, 129.9, 130.6, 131.4 (JCF
=
33 Hz), 136.0, 140.7, 154.8, 158.9, 168.8; MS (APCI) m/z 630.1
(M+, 100%); Anal. calcd. for C25H30N7O7SF3: C, 47.69; H, 4.80;
N, 15.57; S, 5.09%. Found: C, 47.52; H, 4.85; N, 15.13; S, 5.13%.
(3-{4-[(S)-5-Nitroguanidino-2-(3-trifluoromethyl-benzenesul-
fonylamino)pentanoyl]piperazin-1-yl}propyl)carbamic acid 4-
nitrobenzyl ester (5c). Beige amorphous solid; Rf = 0.5 (SiO2,
CH2Cl2–isopropanol 9 : 1, UV); IR mmax (CH2Cl2 liquid film on
NaCl) 3308, 2946, 1716, 1635, 1608, 1522, 1436, 1348, 1327,
1263, 1166 cm−1; dH (300 MHz, CDCl3, Me4Si) 1.66 (2H, m),
1.80 (2H, m), 1.97–2.06 (2H, m), 2.30–2.38, (2H, m), 2.34 (4H,
m), 3.26 (5H, m), 3.34 (3H, m), 4.15 (1H, m), 5.18 (2H, s), 5.66
(1H, t), 6.62 (1H, br s), 7.51 (2H, d, J = 8.7 Hz), 7.66 (1H, t,
J = 8.1 Hz), 7.81 (1H, d, J = 8.1 Hz), 8.04 (1H, d, J = 8.1 Hz),
8.06 (1H, s), 8.21 (2H, d, J = 8.7 Hz); dC (75 MHz, CDCl3, Me4Si)
25.3, 26.0, 29.8, 40.0, 42.0, 45.1, 52.2, 52.7, 52.8, 55.9, 64.8,
123.0 (JCF = 273 Hz), 123.5, 124.0 (JCF = 4 Hz), 126.4 (JCF
=
4 Hz), 127.9, 129.8, 130.7, 131.1 (JCF = 32 Hz), 144.4, 146.1,
147.2, 155.8, 159.4, 168.2; MS (ESI) m/z 732.2 (M+, 100%);
Anal. Calcd. for C28H36N9O9SF3: C, 45.96; H, 4.96; N, 17.23%.
Found: C, 46.06; H, 5.09.
{12-[4-((S)-2-Amino-5-nitroguanidino-pentanoyl)piperazin-1-
yl]dodecyl}carbamic acid 4-nitrobenzyl ester (4d). Yellow
amorphous solid; IR mmax (CH2Cl2 liquid film on NaCl) 3300,
2928, 2855, 1708, 1690, 1521, 1347, 1256 cm−1; dH (300 MHz,
D2O) 1.18 (16H, m), 1.20 (2H, m), 1.43 (2H, m), 1.69 (4H, m),
1.91 (2H, m), 3.11 (6H, m), 3.29 (2H, m), 3.64 (4H, m), 4.55
(1H, m), 5.17 (2H, s), 7.54 (2H, d, J = 8.1 Hz), 8.22 (2H, d, J =
8.1 Hz); dC (50 MHz, CDCl3, Me4Si) 24.2, 24.3, 24.4, 27.3, 27.5,
40.3, 40.8, 41.6, 43.3, 51.2, 51.6, 51.8, 51.9, 57.3, 57.5, 59.3,
65.1, 73.1, 124.4, 129.0, 145.2, 146.7, 156.8, 162.0, 168.7; MS
(APCI) m/z 650.5 (M+, 100%); HRMS (ESI): 650.4001 (calcd
for C30H52N9O7 = 650.3990).
4.1.3 General procedure for the preparation of 3-trifluoro-
methyl-benzenesulfonic acid [4-guanidino-1(S)-(azacyclohexan-
1-ylcarbonyl)butyl]amide (5). Product 4 (0.54 mmol, 1 equiv)
was dissolved CH2Cl2 (5 mL) and Et3N (0.1688 g, 1.67 mmol,
3.1 equiv). Then 3-trifluoromethyl-benzenesulfonyl chloride (0.
1527 g, 0.64 mmol, 1.2 equiv) in CH2Cl2 (1 mL) was added
dropwise with stirring at 0–5 ◦C. The solution was stirred for
additional 2 h at room temperature, then the mixture was washed
with brine, dried over MgSO4, and evaporated under reduced
pressure. The residue was purified by column chromatography
on SiO2 (EtOAc) with a yield of about 75%.
(12-{4-[(S)-5-Nitroguanidino-2-(3-trifluoromethyl-benzenesul-
fonylamino)pentanoyl]piperazin-1-yl}dodecyl)carbamic
acid
4-nitrobenzyl ester (5d). Orange amorphous solid; Rf = 0.4
(SiO2, CH2Cl2–isopropanol 9 : 1, UV); IR mmax (CH2Cl2 liquid
film on NaCl) 3314, 2928, 2854, 1707, 1632, 1523, 1431, 1347,
1327, 1263, 1166, 1134 cm−1; dH (300 MHz, CDCl3, Me4Si) 1.24
(16H, m), 1.39 (2H, m), 1.50 (2H, m), 1.62 (2H, m), 1.85 (2H,
m), 1.98 (2H, m), 2.22 (2H, m), 2.35 (2H, m), 3.19 (2H, q, J =
6.6 Hz), 3.25 (2H, m), 3.36 (4H, m), 4.10 (1H, m), 4.91 (1H, m),
5.18 (2H, s), 6.60 (1H, m), 7.50 (2H, d, J = 8.7 Hz), 7.68 (1H, t,
J = 7.5 Hz), 7.81 (1H, d, J = 7.5 Hz), 8.04 (1H, d, J = 8.7 Hz),
8.07 (1H, s), 8.20 (2H, d, J = 8.4 Hz); dC (50 MHz, CDCl3,
Me4Si) 24.7, 25.6, 26.8, 27.0, 27.6, 29.5, 29.8, 30.1, 40.4, 41.4,
42.5, 45.5, 52.5, 52.6, 53.1, 58.4, 65.2, 123.2 (JCF = 273 Hz),
123.8, 124.4 (JCF = 4 Hz), 128.2, 129.7 (JCF = 4 Hz), 130.2,
130.9, 131.6 (JCF = 34 Hz), 140.5, 144.4, 146.6, 147.6, 155.9,
159.3, 168.0; MS (APCI) m/z 858.5 (M+, 100%); HRMS (ESI):
880.3596 (calcd for C37H54N9O9SF3Na: 880.3615).
4.1.4 General procedure for the preparation of 3-(trifluoro-
methyl)benzenesulfonic acid [4-guanidino-1(S)-(azacyclohexan-
1-ylcarbonyl)butyl]amide (1). Product 5 (1.01 mmol, 1 equiv)
was dissolved in a 3 : 1 mixture of EtOH–AcOH (10 mL).
The flask was purged with nitrogen and the catalyst Pd/C 10%
(0.091 g, 0.086 mmol, 0.085 equiv) was added. The reaction
mixture was then stirred at 50 ◦C for 12 h under hydrogen
atmosphere. The catalyst was filtered off and the solution was
concentrated under reduced pressure. The solid residue was
crystallized from a solution of H2O–MeOH 1 : 1 and washed
3-Trifluoromethyl-benzenesulfonic acid [4-nitroguanidino-1(S)-
(piperidin-1-ylcarbonyl)butyl]amide (5a). White amorphous
solid; Rf = 0.2 (SiO2, EtOAc, UV); IR mmax (KBr) 3228, 2942,
2857, 1628, 1430, 1326, 1256, 1164, 1129 cm−1; dH (300 MHz,
CDCl3, Me4Si) 1.23 (2H, m), 1.51 (4H, m), 1.65 (2H, m), 1.86
(2H, m), 3.16 (3H, m), 3.37 (3H, m), 4.11 (1H, m), 6.53 (1H,
m), 7.61 (2H, m), 7.68 (1H, t, J = 7.5 Hz), 7.83 (1H, d, J =
4 2 1 6
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0

View Article Online
with Et2O. The solvent was evaporated and the product was
isolated as acetate salt (about 90% yield).
Rf = 0.5 (SiO2, cyclohexane–EtOAc 5 : 4, UV); IR mmax (KBr)
3417, 3339, 2938, 2848, 1705, 1607, 1521, 1347, 1245, 859 cm−1;
dH (300 MHz, CDCl3, Me4Si) 2.1 (2H, quint, J = 6.4 Hz),
3.37 (2H, t, J = 6.4 Hz), 3.46 (2H, t, J = 6.4 Hz), 5.07 (1H, s),
5.20 (2H, s), 7.51 (2H, d, J = 8.5 Hz), 8.22 (2H, d, J = 8.5 Hz);
dC (50 MHz, CDCl3, Me4Si) 30.5, 32.3, 39.5, 65.1, 123.6, 128.0,
143.8, 147.6, 155.7; MS (APCI) m/z 318.9 (M 81Br+, 80%), 316.9
(M 79Br+, 80%); Anal. calcd. for C11H13N2O4Br: C, 41.66; H, 4.13;
N, 8.83%. Found: C, 41.80; H, 4.13; N, 8.81%.
3-Trifluoromethyl-benzenesulfonic acid [4-guanidino-1(S)-
(piperidin-1-yl-carbonyl)butyl]amide (1a). White amorphous
solid; IR mmax (KBr) 3333, 3200, 2944, 2853, 1639, 1325, 1163,
1122 cm−1; dH (300 MHz, CD3OD) 1.23 (2H, m), 1.43 (4H, m),
1.61 (4H, m), 3.01–3.21 (6H, m), 4.18 (1H, m), 7.67 (1H, t, J =
8.1 Hz), 7.83 (1H, d, J = 7.8 Hz), 7.99 (1H, d), 8.00 (1H, s); dC
(50 MHz, CD3OD) 25.1, 25.9, 26.4, 27.5, 31.2, 41.9, 44.3, 47.5,
4-[3-(4-Nitrobenzyloxycarbonylamino)propyl]piperazine-1-
carboxylic acid tert-butyl ester (7). Boc-piperazine (1.372 g,
7.22 mmol, 1 equiv) and compound 6 (2.517 g, 7.94 mmol,
1.1 equiv.) were dissolved in dry CH2Cl2 (40 mL) and DIPEA
(4 mL, 20.4 mmol, 3 equiv). The mixture was heated at reflux
for 2 h. The solvent was then evaporated under reduced pressure
and the gel obtained was extracted with EtOAc. The organic
layer was then washed with brine, dried over MgSO4 and
evaporated under reduced pressure. The residue was purified by
column chromatography to obtain pure compound 7 (2.72 g),
as a yellow gel (84% yield). Rf = 0.3 (SiO2, EtOAc–acetone 2 :
1, UV); IR mmax (CH2Cl2 liquid film on NaCl) 2947, 1719, 1685,
1523, 1349, 1241 cm−1; dH (300 MHz, CDCl3, Me4Si) 1.47 (9H,
s), 1.72 (2H, quint, J = 6.5 Hz), 2.40 (4H, t, J = 4.8 Hz), 2.47
(2H, m), 3.32 (2H, m), 3.44 (4H, t, J = 4.8 Hz), 5.20 (2H, s),
6.09 (1H, t), 7.52 (2H, d, J = 8.8 Hz), 8.22 (2H, d, J = 8.8 Hz);
dC (50 MHz, CDCl3, Me4Si) 25.7, 28.4, 40.6, 43.6, 52.9, 56.8,
64.8, 79.7, 123.5, 127.9, 144.2, 147.3, 154.5, 155.7; MS (APCI)
m/z 423.2 (M+, 100%); Anal. calcd. for C20H30N4O6·H2O: C,
54.53; H, 7.32; N, 12.72%. Found: C, 54.57; H, 7.36; N, 12.65%.
53.7, 124.7 (JCF = 276.6 Hz), 125.3 (JCF = 3.8 Hz), 130.5 (JCF
=
3.6 Hz), 131.3, 132.1, 132.4 (JCF = 33.1 Hz), 143.1, 158.8, 170.1;
MS (APCI) m/z 450.3 (M+, 100%); HRMS (ESI) m/z: 450.1762
(calcd for C18H27N5O3SF3: 450.1787).
3-Trifluoromethyl-benzenesulfonic acid [4-guanidino-1(S)-
(piperazin-1-ylcarbonyl)butyl]amide (1b). White amorphous
solid; IR mmax (KBr) 3500–3000, 1657, 1549, 1423, 1322,
1163 cm−1; dH (300 MHz, CD3OD) 1.37 (2H, m), 1.44 (2H,
m), 2.66 (1H, m), 2.79 (2H, m), 2.92 (3H, m), 3.26 (2H, m), 3.48
(2H, m), 4.10 (1H, m), 7.56 (1H, t, J = 7.5 Hz), 7.69 (1H, d, J =
7.5 Hz), 7.88 (1H, d), 7.89 (1H, s); dC (50 MHz, CD3OD)
25.1, 30.9, 40.9, 41.8, 44.5, 44.9, 45.1, 53.6, 125.2, 130.5,
131.6, 132.0, 132.8, 143.5, 158.8, 171.0 (CF3 not visi-
ble); MS (APCI) m/z 451.3 (M+, 100%); Anal. calcd. for
C17H25N6O3SF3·CH3COOH·4H2O: C, 39.18; H, 6.40; N, 14.4%.
Found: C, 38.85; H, 6.12; N, 13.97%.
N-{(S)-1-[4-(3-Aminopropyl)piperazine-1-carbonyl]-4-guani-
dino-butyl}-3-trifluoromethyl-benzenesulfonamide (1c). Beige
amorphous solid; IR mmax (CH2Cl2 liquid film on NaCl) 3395,
1636, 1558, 1412, 1328, 1165 cm−1; dH (500 MHz, CD3OD) 1.61
(6H, m), 2.16 (2H, m), 2.22 (1H, m), 2.29 (3H, m), 2.63 (2H, t, J
= 7.4 Hz), 3.13 (2H, m), 3.27–3.61 (4H, m), 4.00 (1H, m), 7.60
(1H, t, J = 8.1 Hz), 7.69 (1H, d, J = 8.1 Hz), 8.00 (1H, d, J =
8.1 Hz), 8.04 (1H, s); dC (125 MHz, CDCl3) 26.9, 30.5, 32.9,
41.0, 42.2, 42.7, 46.2, 53.8, 54.3, 57.1, 58.4, 124.5 (JCF = 270 Hz),
125.1 (JCF = 4 Hz), 130.0 (JCF = 4 Hz), 131.4, 132.0, 132.1 (JCF
= 37 Hz), 149.2, 158.7, 170.4; MS (ESI) m/z 508.2 (M+, 100%);
Anal. calcd. for C20H32N7O3SF3·2CH3COOH·1.5H2O: C, 44.03;
H, 6.62; N, 14.98%. Found: C, 44.37; H, 6.32; N, 14.68%.
(3-Piperazin-1-yl-propyl)carbamic acid 4-nitrobenzyl ester
(8). A solution of HCl 1 N in AcOH (56 mL, 8 equiv)
was added to product 7 (2.95 g, 7 mmol, 1 equiv) and the
reaction mixture was stirred at room temperature for 2 h. The
solution was concentrated by evaporation at reduced pressure
to obtain a brown gel. This residue was several times triturated
and suspended in Et2O until a beige solid suspension was
observed. The solvent was then evaporated to obtain product
8 as hydrochloride salt (quantitative yield). To isolate the free
amine, the product was dissolved in water and the solution was
brought to pH 10 with NaOH 1 N. The aqueous solution was
then extracted 5 times with EtOAc, the organic phases were
combined, dried over MgSO4 and evaporated to give compound
8 (2.28 g) as a free amine (92% product recovery). IR mmax (KBr)
3316, 3214, 2944, 2801, 1718, 1607, 1521, 1347, 1260 cm−1; dH
(200 MHz, CDCl3, Me4Si) 1.63 (2H, quint, J = 6.4 Hz), 1.94
(1H, s), 2.37 (6H, m), 2.83 (4H, t, J = 4.9 Hz), 3.23 (2H, q, J =
6.0 Hz), 5.12 (2H, s), 6.36 (1H, t, J = 6.0 Hz), 7.44 (2H, d,
J = 8.8 Hz), 8.14 (2H, d, J = 8.8 Hz); dC (75 MHz, CDCl3,
Me4Si) 25.9, 41.5, 46.5, 54.7, 58.2, 65.3, 124.1, 128.4, 144.8,
147.8, 156.2; MS (APCI) m/z 323.3 (M+, 100%); Anal. calcd.
for C15H22N4O4·1.2HCl: C, 49.21; H, 6.34; N, 15.31%. Found:
C, 49.50; H, 6.19; N, 14.80%.
N-{(S)-1-[4-(12-Aminododecyl)piperazine-1-carbonyl]-4-guani-
dino-butyl}-3-trifluoromethyl-benzenesulfonamide
(1d).
Orange amorphous solid; IR mmax (KBr) 3500–2800, 1647, 1560,
1407, 1328, 1280, 1166, 1134 cm−1; dH (300 MHz, CD3OD) 1.24
(16H, m), 1.41 (3H, m), 1.58 (6H, m), 1.92 (1H, m), 2.04 (1H,
m), 2.22 (2H, t, J = 7.5 Hz), 2.37 (4H, m), 2.82 (2H, t, J = 7.5
Hz), 3.14 (3H, m), 3.34 (2H, m), 4.23 (1H, m), 7.68 (1H, t, J =
7.5 Hz), 7.81 (1H, d, J = 7.5 Hz), 8.02 (1H, d, J = 7.5 Hz), 8.05
(1H, s); dC (50 MHz, CD3OD) 26.0, 27.6, 28.6, 28.8, 30.4, 30.7,
30.8, 31.2, 40.8, 41.8, 42.8, 46.3, 53.7, 54.3, 59.5, 59.7, 125.1 (JCF
= 260 Hz), 125.2 (JCF = 4 Hz), 130.5 (JCF = 4 Hz), 131.6, 131.8
(JCF = 32 Hz), 132.3, 143.4, 158.9, 170.5; MS (ESI) m/z 634.4
(M+, 100%); Anal. calcd. for C29H50N7O3SF3·3CH3COOH: C,
51.65; H, 7.68; N, 11.21%. Found: C, 52.17; H, 7.76; N, 12.05%.
4.1.6 Preparation of dodecyl spacer arm moiety.
4.1.5 Preparation of propyl spacer arm moiety.
12-(4-Nitrobenzyloxycarbonylamino)dodecanoic acid (9).
12-Aminododecanoic acid (250 mg, 1.16 mmol, 1 equiv) was
dissolved in THF (3 mL), water (3 mL) and NaOH 1 N (1.14 mL,
1 equiv). 4-Nitrobenzyl chloroformate (246 mg, 1.16 mmol,
1 equiv) in THF (1 mL) and NaOH 1 N (1.3 mL, 1.1 equiv)
were added simultaneously to the reaction mixture at 0 ◦C. The
solution was stirred at 0 ◦C for 10 min and then for 2 h at
room temperature, checking pH was above 10. HCl 1 N was
added to the reaction mixture to reach pH 1, and the solution
was then extracted twice with EtOAc. The organic layers
were combined, washed with HCl 1 N and water, dried over
MgSO4, and evaporated under reduced pressure to give a yellow
solid that was purified by column chromatography on SiO2 to
obtain pure product 9 (0.345 g) as a white solid (75% yield).
Mp = 107.9–108.9 ◦C; Rf = 0.3 (SiO2, EtOAc–cyclohexane
(3-Bromopropyl)-carbamic acid 4-nitrobenzyl ester (6). 3-
Bromopropyl amine (2 g, 9 mmol, 1 equiv) was dissolved in water
(20 mL), THF (20 mL) and NaOH 1 N (9 mL, 9 mmol, 1 equiv).
4-Nitrobenzyl chloroformate (1.97 g, 9 mmol, 1 equiv) dissolved
in cold◦THF (5 mL) was added dropwise to the reaction mixture
at 0–5 C, simultaneously with NaOH 1 N (10 mL, 10 mmol,
1.1 equiv). The solution was then allowed to warm up at room
temperature and stirred for 2 h, checking the pH to be above
10. The solvent was then removed by evaporation at reduced
pressure, the residue was dissolved in EtOAc and washed twice
with brine. The aqueous phases were extracted once with fresh
EtOAc. The organic phases were combined and dried over
MgSO4 and evaporated at reduced pressure to give pure com-
pound 6 (2.85 g) as white solid (98% yield). Mp = 71.9–72.3 ◦C;
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0
4 2 1 7

View Article Online
3 : 5, UV); IR mmax (CH2Cl2 liquid film on NaCl) 3367, 2920, 2851,
1725, 1689, 1613, 1527, 1351, 1256, 1246 cm−1; dH (300 MHz,
CDCl3, Me4Si) 1.27 (14H, m), 1.52 (2H, m), 1.64 (2H, m), 2.36
(2H, t, J = 7.1 Hz), 3.21 (2H, m), 4.84 (1H, m), 5.20 (2H, s),
7.52 (2H, d, J = 8.6 Hz), 8.22 (2H, d, J = 8.6 Hz); dC (75 MHz,
CDCl3, Me4Si) 24.7, 26.7, 29.0, 29.1, 29.2, 29.3, 29.4, 29.9, 33.8,
41.3, 65.1, 123.7, 128.0, 144.1, 150.2, 155.7, 178.2; MS (APCI)
m/z 395.0 (M+, 100%); Anal. calcd. for C20H30N2O6: C, 60.90;
H, 7.65; N, 7.10%. Found: C, 60.55; H, 7.45; N, 6.76%.
(12-Hydroxydodecyl)carbamic acid 4-nitrobenzyl ester (10).
Compound 9 (0.220 g, 0.56 mmol, 1 equiv) was dissolved in dry
THF (10 mL) and the solution was cooled down to 0 ◦C. BH3 1 M
in THF (1.11 mL, 1.11 mmol, 2 equiv) was added dropwise to
the solution that was then stirred for additional 2 h at 0 ◦C. The
reaction mixture was quenched by dropwise addition of AcOH–
MeOH 1 : 9. The mixture was then concentrated and the residue
dissolved in EtOAc. The organic phase was washed with water,
dried over MgSO4, and concentrated under reduced pressure.
The residue was further purified by column chromatography on
SiO2 to give pure 10 (0.193 g) as a white solid (91% yield). Mp =
103.7–104.7 ◦C; Rf = 0.6 (SiO2, EtOAc–cyclohexane 4 : 5,
UV); IR mmax (CH2Cl2 liquid film on NaCl) 3400, 2987, 2865,
1688 cm−1; dH (300 MHz, CDCl3, Me4Si) 1.27 (16H, m), 1.55
(4H, m), 3.21 (2H, q, J = 6.6 Hz), 3.65 (2H, t, J = 6.6 Hz), 4.83
(1H, m), 5.20 (2H, s), 7.52 (2H, d, J = 8.7 Hz), 8.22 (2H, d, J =
8.7 Hz); dC (75 MHz, CDCl3, Me4Si) 25.8, 26.7, 29.2, 29.3,
29.4, 29.5, 29.6, 29.9, 32.8, 41.3, 63.1, 65.0, 123.7, 128.0, 144.1,
150.2, 155.7; MS (APCI) m/z 381 (M+, 100%); Anal. calcd. for
C20H32N2O5: C, 63.14; H, 8.48; N, 7.36%. Found: C, 63.76; H,
8.78; N, 7.02%.
Methanesulfonic acid 12-(4-nitrobenzyloxycarbonylamino)-
dodecyl ester (11). Compound 10 (10.24 g, 26.91 mmol, 1 equiv)
was dissolved in CH2Cl2 (200 mL). Pyridine (4.22 g, 53.83 mmol,
2 equiv) and methanesulfonyl chloride (4.63 g, 40.37 mmol,
1.5 equiv) were added dropwise at 0 ◦C. The solution was stirred
at 0 ◦C for 3 h, and for an additional 12 h at room temperature.
The mixture was then washed with water, and HCl 1 N, dried
over MgSO4 and concentrated at reduced pressure. The residue
was purified by column chromatography on SiO2 to give pure
compound 11 (9.33 g) as a white solid (76% yield). Mp = 80.9–
82.9 ◦C; Rf = 0.3 (SiO2, CH2Cl2–EtOAc 99 : 1, UV); IR mmax
(CH2Cl2 liquid film on NaCl) 3370, 2921, 2851, 1690, 1614,
1531, 1355, 1264, 1167 cm−1; dH (300 MHz, CDCl3, Me4Si) 1.27
(14H, m), 1.37 (2H, m), 1.52 (2H, m), 1.75 (2H, tt, J = 7.0 Hz),
3.01 (3H, s), 3.21 (2H, q, J = 6.9 Hz), 4.23 (2H, t, J = 6.8 Hz),
4.83 (1H, br s), 5.20 (2H, s), 7.51 (2H, d, J = 8.6 Hz), 8.22 (2H,
d, J = 8.6 Hz); dC (50 MHz, CDCl3, Me4Si) 25.4, 26.6, 28.9,
29.1, 29.2, 29.3, 29.4, 29.5, 29.6, 29.9, 37.3, 41.2, 64.9, 70.2,
123.7, 128.0, 144.2, 147.6, 155.8; MS (APCI) m/z 459.1 (M+,
100%); Anal. calcd. for C21H34N2O7S: C, 55.00; H, 7.47; N, 6.11;
S, 6.99%. Found: C, 54.85; H, 7.40; N, 6.19; S, 6.46%.
4-[12-(4-Nitrobenzyloxycarbonylamino)dodecyl]piperazine-1-
carboxylic acid tert-butyl ester (12). Compound 11 (0.100 g,
0.22 mmol, 1 equiv), NaI (0.033 g, 0.22 mmol, 1 equiv), and
Boc-piperazine (0.041 g, 0.22 mmol, 1 equiv) were dissolved
in CH3CN (5 mL) and water (0.2 mL). The reaction mixture
was stirred at 80 ◦C for 30 min, then ethyldiisopropylamine
(0.028 g, 0.22 mmol, 1 equiv) was added. After 12 h at 80 ◦C the
solvent was removed by evaporation and the residue dissolved
in CH2Cl2. The organic phase was washed with water, dried
over MgSO4, and concentrated at reduced pressure. The residue
was purified by column chromatography on SiO2 to give pure
compound 12 (0.108 g) as a yellow solid (90% yield). Mp =
62.6–63.6 ◦C; Rf = 0.5 (SiO2, CH2Cl2/isopropanol 96 : 4, UV);
IR mmax (CH2Cl2 liquid film on NaCl) 3320, 2930, 2855, 1696,
1601, 1523, 1413, 1347, 1252, 1163 cm−1; dH (300 MHz, CDCl3,
Me4Si) 1.24 (16H, m), 1.44 (13H, m), 2.30 (2H, m), 2.34 (4H,
m), 3.17 (2H, q, J = 6.6 Hz), 3.42 (4H, m), 5.02 (1H, m), 5.17
(2H, s), 7.49 (2H, d, J = 8.1 Hz), 8.19 (2H, d, J = 8.7 Hz);
dC (50 MHz, CDCl3, Me4Si) 26.9, 27.7, 28.6, 29.4, 29.7, 29.9,
30.1, 41.5, 43.8, 53.2, 58.9, 65.2, 79.7, 123.8, 128.2, 144.4, 147.8,
154.9, 156.1; MS (APCI) m/z 549.9 (M+, 100%); Anal. calcd.
for C29H48N4O6·0.3H2O: C, 62.84; H, 8.95; N, 10.11%; Found:
C, 62.76; H, 8.85; N, 10.07%.
(12-Piperazin-1-yl-dodecyl)carbamic acid 4-nitrobenzyl ester
(13). Compound 12 (0.497 g, 0.91 mmol, 1 equiv) was dissolved
in a 1 : 1 mixture of TFA–CH2Cl2 (7 mL) and the solution
was stirred at room temperature for 2 h. The solution was
concentrated under reduced pressure to give a brown gel that was
triturated and suspended in Et2O until precipitation of a brown
solid occurred. The solvent was evaporated and compound 13
(0.510 g) was recovered as the trifluoroacetate salt (quantitative
yield). Mp = 79.5–80.5 ◦C; IR mmax (CH2Cl2 liquid film on NaCl)
3354, 2920, 2850, 1690, 1612, 1528, 1350 cm−1; dH (300 MHz,
CDCl3, Me4Si) 1.25 (16H, m), 1.49 (4H, m), 2.33 (2H, t, J = 7.7
Hz), 2.47 (4H, m), 2.96 (4H, m), 3.19 (2H, q, J = 6.7 Hz), 3.99
(1H, m), 4.93 (1H, m), 5.18 (2H, s), 7.50 (2H, d, J = 8.4 Hz),
8.21 (2H, d, J = 8.7 Hz); dC (50 MHz, CDCl3, Me4Si) 25.3, 26.6,
26.9, 29.1, 29.3, 29.4, 29.6, 29.9, 30.7, 41.2, 42.6, 49.3, 57.9, 65.0,
123.6, 128.0, 144.3, 147.6, 155.9; MS (APCI) m/z 449.4 (M+,
100%); HRMS (ESI) m/z 449.3119 (calcd. for C24H41N4O4 =
449.3128).
4.2 Polymer surface modification
The polymer materials considered were poly(ethylene terephtha-
late) (PET) track-etched micro-porous membrane (Whatman
S.A., thickness 12 lm, density 1.39 g cm−3, pore diameter
0.49 lm, pore density 1.45 × 106 pore cm−2) and poly(butylene
terephthalate) (PBT) melt blown filtration membrane (Johns
Manville, thickness 133 lm, mean flow pore 5 lm, basis weight
82 gsm). The surface chemical composition of the derivatized
polymers was determined by XPS with a SSX 100/206 photo-
electron spectrometer from Surface Science Instruments (USA)
equipped with a monochromatized microfocus Al X-ray source
(powered at 20 mA and 10 kV). The pressure in the analysis
chamber was around 10−6 Pa. The angle between the surface
normal and the axis of the analyser lens was 55◦. The analysed
area was approximately 1.4 mm2 and the pass energy was set
at 150 eV. In these conditions, the resolution determined by
the full width at half maximum (FWHM) of the Au 4f7/2 peak
was around 1.6 eV. A flood gun set at 10 eV and a Ni grid
placed 3 mm above the sample surface were used for charge
stabilization. The binding energies were calculated with respect
to the C–(C,H) component of the C 1s peak fixed at 284.8 eV.
Data treatment was performed with the CasaXPS program
(Casa Software Ltd, UK). Molar fractions were calculated using
peak areas normalised on the basis of acquisition parameters
and sensitivity factors provided by the manufacturer.
4.2.1 Activation and coupling. Native polyester mem-
branes (14) were cut into disks of 13 mm of diameter. The
polymer samples (10 disks) were stirred at 20 ◦C (PBT) or
60 ◦C (PET) for 1 h into a mixture of acetone (50 mL), pyridine
(1.06 mL) and p-toluenesulfonyl chloride (2.5 g). The samples
were washed successively with acetone (50 mL, 5 min) and
water (50 mL, 5 min). The resulted activated samples (15) were
individually immersed in 1 mL of a 10−3 M solution of inhibitor
in a phosphate buffer (PBS, pH 7.2)–CH3CN mixture (1 : 1)
and incubated for 2 h at 20 ◦C with shaking. The samples were
washed with PBS–CH3CN (1 mL, 2 × 10 min), water (1 mL,
2 × 5 min), 5 × 10−3 M HCl (1 mL, 2 × 5 min) and water (1 mL,
2 × 10 min). The resulting inhibitor grafted membranes (16)
were finally dried over filter paper and analysed by XPS within
24 h. The blank samples were prepared in the same way but
without addition of p-toluenesulfonyl chloride in the activation
step.
4.2.2 XPS analysis. The surface molar fractions were
measured by XPS analysis. The concentration of F tagged
4 2 1 8
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0

View Article Online
molecules on the surface was calculated from the F/C ratio.
Considering for instance the polymer repeat unit of PET –
CO–C6H4–CO2–CH2CH2–O– (C10H8O4) and the derivatized
(by grafting of compound 1d) chain unit –CO–C6H4–CO2–
CH2CH2–1d (C39H57F3N7O6S) we calculated the percentage of
derivatized units as follows. For a mixture of 96.7% (C10H8O4)
and 3.3% (C39H57F3N7O6S), F/C × 100 = 9.9 × 100/1095.7 =
0.904 (experimental value 0.900). Similarly, the corrected F/C
value of 0.005 for compound 1d on PBT gave a derivatization
ratio of 1.9%. We previously calculated on the basis of PET
crystallographic data and simple geometrical considerations37
an average of 1.72 × 1015 PET monomer units per cm2 of
surface covering 10 atomic layers (the depth analysed by
XPS) or about 2850 pmol cm−2. Thus, 3.3% of derivatized
surface monomer units corresponds to about 90 pmol cm−2 of
fixed compound 1d. In the same way we calculated a surface
concentration of compound 1c of about 10 pmol cm−2 for a
corrected F/C value of 0.001 (0.4% derivatization). The lack of
crystallographic data on the PBT membrane did not allow an
analogue calculation of signal surface concentration.
on conformational and non-bonded contact information from
the CSD database. Parameters: Popsiz = 100; maxops = 100000;
niche size = 2.
Autodock uses a hybrid method called Lamarckian Genetic
Algorithm (genetic algorithm coupled withalocal search) to pre-
dict the interaction of flexible ligands with rigid macromolecular
targets. The scoring function includes van der Waals, coulombic
electrostatic, directional hydrogen bonding, entropy of ligand
binding and desolvation contributions. Parameters: Runs = 200;
Population size = 50; Number of generations = 27000.
Discover3 uses the molecular mechanics to optimize the
conformation of the ligand–protein complex and evaluate the
interaction energy (the enthalpic contribution of the binding)
which is the sum of coulombic and van der Waals terms (DEcb
and DEvdw). The backbone is moved following force constants;
side chains and water molecules move freely.
Forcefield: CVFF; Dielectric constant: 1*r; Criteria conver-
gence: 10 kcal mol−1 for the Steepest Descent algorithm, 0.01
for the Conjugate Gradient one.
The thrombin amino acid residues were numbered by the
chymotrypsin(ogen) numbering system as suggested by Bode
et al.19
4.3 Biological assays
4.3.1 Enzyme kinetics. Substrate H-D-phenylalanyl-L-
pipecolyl-L-arginine-p-nitroanilide dihydrochloride (S2238) was
obtained from Chromogenix. Human thrombin was obtained
from Hypen BioMed. Enzyme kinetics studies were performed
on a Cary210 spectrophotometer.
4.5 Crystallization, data collection and structure refinement
Human a-thrombin was obtained from Kordia Life Sciences
(Leiden, The Netherlands) and its anion-binding exosite in-
hibitor, the hirudin peptide fragment (54–65), was obtained
from Bachem (Bubendorf, Switzerland). Crystals of human a-
thrombin were grown in the presence of hirudin and compound
1c using the hanging drop vapour diffusion technique at
293 K. Prior to the crystallization, the protein was solubilized to
approximately 10 mg ml−1 in 0.36 M NaCl, 96 mM pH 6.5 citrate
buffer, containing hirudin at a concentration of 3 mM. Drops
were prepared by mixing 1 ll of protein solution with 1 ll
of compound 1c at a final concentration of 2.9 mM to 1 ll
of reservoir solution. The reservoir solution contained 20%
PEGMME 5000, 0.36 M NaCl in 0.1 M HEPES buffer
pH 7.5. Compound 1c was additionally further diffused into
the crystal for 72 h from a 44 mM solution containing 22%
PEGMME 5000 and 0.36 M NaCl to a final concentration of
16 mM.
The inhibition activities were measured by adapting the
protocol developed by Lottemberg et al.:38 100 lL of the
solution of substrate S2238 (0.1 mM in water) and 100 lL
of water solution of the tested compound (50–200 lM final
concentration) were diluted in 2 mL of Tris-Hepes Buffer (Tris
0.01 M, Hepes 0.01 M, NaCl 0.5 M, PEG 6000 0.1% m/v,
pH 7.8). Finally 10 lL of water solution of thrombin (10
NIH mL−1, 75.2 nM) were added to start the reaction. The
appearance of the substrate hydrolysis product (p-nitroaniline)
was measured at 405 nm as a function of time. Plots of V/Vi
versus inhibitor concentration (ratio of hydrolysis in the absence
and in the presence of inhibitors) gave the inhibition constants
indicated in Table 2.
4.3.2 Clot formation. For the evaluation of material hemo-
compatibility we used a simple test based on the weight
measurement of the blood clot formed on given substrates
(polymer squares of 4 × 4 cm fixed on a cylindrical glass support
of 2.6 cm internal diameter) after blood contact and incubation.
Human blood was withdrawn from a healthy volunteer, collected
in citrated tubes and used within a day. For coating experiments
200 lL of a 10−3M solution of the tested compound were placed
on the polymer surface that was then dried overnight. After
weighing the samples, 200 lL of human blood were placed on
the surface, 20 lL of CaCl2 were added to start coagulation
One crystal was flash frozen in liquid nitrogen after rapid
soaking in a cryoprotectant solution containing about 50%
glycerol in the crystallization buffer. Data were collected at
ESRF (European Synchrotron Radiation Facility, Grenoble,
˚
France) on beamline BM30a at a wavelength value of 0.9797 A.
All data were recorded on a MarCCD detector, and intensities
were indexed and integrated using MOSFLM version 6.01.39
This single crystal, belonging to space group C2 with unit cell
˚
˚
˚
dimensions a = 70.51 A, b = 71.34 A, c = 72.56 A and b =
◦
˚
100.59 , diffracted to 1.65 A resolution. The scaling of the
intensity data was accomplished with SCALA of the CCP4
program suite,40 and all corresponding statistics are given in
Table 7.
◦
and the samples were incubated statically for 1.5 h at 37 C in
a saturated water atmosphere and 5% CO2. The samples were
then washed with water (3 × 2 mL), the clot was fixed with
formaldehyde (2 mL of a 5% solution in water) and then washed
with water (2 mL). The samples were dried in oven (37 ◦C) until
constant weight. The sample weight differences gave the weight
of the formed clot. Native polymer coated with heparin (Leo,
5000 I.E./U.I./mL, 200 lL) was used as anti-clotting reference
sample.
The phasing procedure for solving the structure was the
molecular replacement method using the AMoRe package.41
The structure of the complex human a-thrombin/hirugen (PDB
code 1HTG)42 was used as the search model for the rotation and
translation searches. Model building with TURBO-FRODO43
and refinement with CNS 1.144 gave final overall crystallographic
R factors of 24.3% (working) and 27.2% (free), with values in
the outer shell of 27.5% and 31.7%, respectively, for 2333 protein
atoms, 34 inhibitor atoms and 187 water molecules. The statistics
of refinement are summarized in Table 7. The inhibitor density
4.4 Docking study
Molecular modelling studies were performed on a Silicon
Graphics Octane2 workstation.
Gold implements a genetic algorithm allowing the protein–
ligand docking with full ligand and partial protein flexibility.
Indeed, conformation of some amino acids (Ser, Thr and Lys)
are optimized during the run. The energy function is partly based
was very clear as shown in Fig. 3A: the refined temperature
2
˚
factors for the inhibitor range from 15 (O15 atom) to 30 A .
Co-ordinates and structure factors have been deposited in
the Protein Data Bank (accession code 1W7G). Figures were
prepared with MOLSCRIPT/RASTER3D.45
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0
4 2 1 9

View Article Online
Table 7 X-Ray data collection parameters and refinement statistics
9 M.-F. Gouzy, C. Sperling, K. Salchert, T. Pompe, U. Streller, P.
Uhlmann, C. Rauwolf, F. Simon, F. Bohme, B. Voit and C. Werner,
Biomaterials, 2004, 25, 3493.
Data Ccollection
10 J. Marchand-Brynaert, E. Detrait, O. Noiset, T. Boxus, Y.-J. Schnei-
der and C. Remacle, Biomaterials, 1999, 20, 1773.
11 S. Biltresse, M. Attolini, G. Dive, A. Cordi, G. C. Tucker and J.
Marchand-Brynaert, Bioorg. Med. Chem., 2004, 12, 5379.
12 J. I. Weitz and M. Crowther, Thromb. Res., 2002, 106, 275.
13 F. Thorstensson, I. Kvarnstroem, D. Musil, I. Nilsson and B.
Samuelsson, J. Med. Chem., 2003, 46, 1165.
14 J.-M. Altenburger, G. Y. Lassalle, M. Matrougui, D. Galtier, J.-C.
Jetha, Z. Bocskei, C. N. Berry, C. Lunven, J. Lorrain, J.-P. Herault,
P. Schaeffer, S. E. O’Connor and J.-M. Herbert, Bioorg. Med. Chem.,
2004, 12, 1713.
15 S. Okamoto, K. Kinjo, A. Hijikata, R. Kikumoto, Y. Tamao and K.
Ohkubo, T. S., J. Med. Chem., 1980, 23, 827.
16 C. R. W. Guimaraes and R. Bicca de Alencastro, J. Med. Chem.,
2002, 45, 4995.
17 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.
18 S. Bajusz, E. Szell, D. Bagdy, E. Barabas, G. Horvath, M. Dioszegi,
Z. Fittler, G. Szabo, A. Juhasz and E. Tomori, J. Med. Chem., 1990,
33, 1729.
19 W. Bode, I. Mayr, U. Baumann, R. Huber, S. R. Stone and J.
Hofsteenge, EMBO J., 1989, 8, 3467.
20 K. McKeage and G. L. Plosker, Drugs, 2001, 61, 515.
21 C. W. Francis, Best Practice & Research Clinical Haematology, 2004,
17, 139.
22 S. Biltresse, D. Descamps, C. Henneuse-Boxus and J. Marchand-
Brynaert, J. Polym. Sci., Part A: Polym. Chem., 2002, 40,
770.
23 Y. Ito and Y. Imanishi, J. Biomater. Appl., 1992, 6, 293.
24 S. Devouge, C. Salvagnini and J. Marchand-Brynaert, Bioorg. Med.
Chem. Lett., 2005, 15, 3252.
25 R. E. Babine and S. L. Bender, Chem. Rev., 1997, 97, 1359.
26 D. K. Jones-Hertzog and W. L. Jorgensen, J. Med. Chem., 1997, 40,
1539.
27 D. W. Banner and P. Hadvary, J. Biol. Chem., 1991, 266, 20085.
28 G. M. Morris, D. S. Goodsell, R. S. Halliday, R. Huey, W. E.
Hart, R. K. Belew and A. J. Olson, J. Comput. Chem., 1998, 19,
1639.
˚
Wavelength/A
Space group
0.979657 (BM30a at ESRF)
C2
Unit-cell parameters
˚
a/A
70.510
71.343
72.555
100.59
˚
b/A
˚
c/A
b/◦
˚
Resolution range/A
Overall reflections
Highest resolution shell
No of unique reflections
Data completeness (%)
Redundancy
24.84–1.65
1.69–1.65
40937
96.5 (93)
4.8 (2.7)
9.2 (23.7)
15.7 (4.3)
Rsym (%)
I/r(I)
Refinement
˚
Resolution range/A
Completeness (working + test)
(%)
19.78–1.65 (1.75–1.65)
95.1 (90.5)
No. of unique reflections (I >
40496 (6084)
2r(I))
R (working test) (%)
Rfree (test set) (%)
Protein atoms (non-H)
Heterogen atoms
24.3 (27.5)
27.2 (31.7)
2333
34
187
16.2
16.0
Solvent atoms
2
˚
B values from Wilson plot/A
2
˚
Mean B value (overall)/A
Estimated coordinate errors (low
˚
resolution cutoff of 5.0 A)
From Luzzati plot/A
From sigma A/A
Deviations from ideal geometry
˚
0.23
0.09
˚
˚
Bond lengths/A
0.006
1.4
23.8
0.91
1W7G
Bond angles/◦
Dihedral angles/◦
Improper angles/◦
PDB entry code
29 G. Jones, P. Willett, R. C. Glen, A. R. Leach and R. Taylor, Gold,
Astex Technology, Cambridge, UK, 2001.
30 Accelrys, Discover3, Accelrys Inc., San Diego, CA, USA, 1998.
31 M. J. Frisch, G. W. Trucks and H. B. Schlegel, Gaussian 98, Gaussian
Inc., Pittsburgh, PA, USA, 2001.
Acknowledgements
32 Z. Chen, Y. Li, A. M. Mulichak, S. D. Lewis and J. A. Shafer, Arch.
Biochem. Biophys., 1995, 322, 198.
This work was supported by FNRS (Fonds National de la
Recherche Scientifique, IISN 4.4505.00, Belgium), the Commu-
naute´ Franc¸aise de Belgique (Action de Recherche Concerte´e
no. 99/04-240) and the University of Lie`ge (Fonds spe´ciaux,
cre´dit classique 2001, R.CFRA.0460). The authors thank the
“Unite´ de chimie des interfaces, UCL” for access to the XPS
facilities. C. S. thanks Professors J. Fastrez and P. Rouxhet
for helpful discussion; Dr R. Touillaux for NMR spectra
interpretation; G. Nollet and S. Gharbi for providing key
synthetic intermediates. J. R. thanks Dr E. Sauvage for help in
the crystallographic study. J. M.-B. is senior research associate
of the FNRS.
33 A. Dahlgren, P.-O. Johansson, I. Kvarnstrom, D. Musil, I. Nilsson
and B. Samuelsson, Bioorg. Med. Chem., 2002, 10, 1829.
34 D. Noeteberg, J. Brnalt, I. Kvarnstroem, M. Linschoten, D. Musil,
J.-E. Nystroem, G. Zuccarello and B. Samuelsson, J. Med. Chem.,
2000, 43, 1705.
35 J. W. Nilsson, I. Kvarnstroem, D. Musil, I. Nilsson and B. Samules-
son, J. Med. Chem., 2003, 46, 3985.
36 S. Hanessian, E. Balaux, D. Musil, L.-L. Olsson and I. Nilsson,
Bioorg. Med. Chem. Lett., 2000, 10, 243.
37 J. Marchand-Brynaert, Surface Functionalization of Polymer Mem-
branes, ed. T. Smith Sorensen, Surface Chemistry and Electrochem-
istry of Membranes, Marcel Dekker Inc., New York-Basel, 1999,
pp. 91–124.
38 R. Lottemberg, J. A. Hall, J. W. Fenton, II and C. M. Jackson,
Thromb. Res., 1982, 28, 313.
39 A. G. R. W. Leslie, Crystallogr. Comput., 1991, 5, 50.
40 Collaborative Computational Project, Acta Crystallogr., Sect. D,
1994, 50, 760.
41 J. Navaza, Acta Crystallogr., Sect. A, 1994, 50, 157.
42 E. Skrzypczak-Jankun, V. E. Carperos, K. G. Ravichandran, A.
Tulinsky, M. Westbrook and J. M. Maraganore, J. Mol. Biol., 1991,
221, 1379.
References
1 C. P. Sharma, J. Biomater. Appl., 2001, 15, 359.
2 Z. M. Yang, L. Wang, J. Yuan, J. Shen and S. C. Lin, J. Biomater.
Sci., Polym. Ed., 2003, 14, 707.
3 W.-C. Lin, T.-Y. Liu and M.-C. Yang, Biomaterials, 2004, 25, 1947.
4 C. Sperling, K. Salchert, U. Streller and C. Werner, Biomaterials,
2004, 25, 5101.
5 G. W. Fussell and S. L. Cooper, Biomaterials, 2004, 25, 2971.
6 L. R. Bush, Cardiovasc. Drug Rev., 1991, 9, 247.
7 Y. Ito, L. S. Liu, R. Matsuo and Y. Imanishi, J. Biomed. Mater. Res.,
Part A, 1992, 26, 1065.
8 H. Noguchi, H. Iwata and Y. Ikada, J. Biomed. Mater. Res., Part A,
1998, 39, 621.
43 A. Roussel and C. Cambillau, Silicon Graphics Geometry Partner
Directory, Mountain View, CA, USA, 1989.
44 A. T. Bru¨nger, P. D. Adams, G. M. Clove, W. L. Delano, P. Gros,
R. W. Grosse-Kunsteve, J.-S. Jiang, J. Kuszewski, M. Nilges and
N. S. Pannu, Acta Crystallogr., Sect. D, 1998, 54, 905.
45 P. Kraulis, J. Appl. Crystallogr., 1991, 946.
4 2 2 0
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 2 0 9 – 4 2 2 0