The design and evaluation of heparin-binding foldamers

Article abstract of DOI:10.1002/anie.200501279

Controlling blood loss: The synthesis of a series of arylamide oligomers 1, which interact with heparin is described. The heparin-small molecule interaction inhibits the heparin's ability to activate antithrombin. These arylamide oligomers are the first e

Full text of DOI:10.1002/anie.200501279

Angewandte
Chemie
Protein Interactions
DOI: 10.1002/anie.200501279
The Design and Evaluation of Heparin-Binding
Foldamers**
Sungwook Choi, Dylan J. Clements,
Vojislava Pophristic, Ivaylo Ivanov,
Satyavani Vemparala, Joel S. Bennett, Michael L. Klein,
Jeffrey D. Winkler, and William F. DeGrado*
Foldamers, nonbiological oligomers with well-defined secon-
dary or tertiary structures,[1,2] provide novel templates for the
design of biologically active molecules that compete for a
[*] S. Choi, Dr. I. Ivanov, Dr. S. Vemparala, Prof. Dr. M. L. Klein,
Prof. Dr. J. D. Winkler, Prof. Dr. W. F. DeGrado
Department ofChemistry, University ofPennsylvania
Philadelphia, PA 19104-6323 (USA)
Fax: (+1)215-573-7229
E-mail: wdegrado@mail.med.upenn.edu
S. Choi, D. J. Clements, Prof. Dr. W. F. DeGrado
Department ofBiochemistry and Biophysics
University ofPennsylvania
Philadelphia, PA 19104-6059 (USA)
Prof. Dr. V. Pophristic
Department ofChemistry and Biochemistry
University ofthe Sciences in Philadelphia
Philadelphia, PA 19104 (USA)
Prof. Dr. J. S. Bennett
Department ofMedicine, Hematology and Oncology Division
University ofPennsylvania
Philadelphia, PA 19104 (USA)
[**] We thank the National Institutes ofHealth (grant no. 65803),
PolyMedix, Inc., and Burroughs Wellcome Fund (fellowship to I.I)
for financial support
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2005, 44, 6685 –6689
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6685

Communications
variety of protein–protein[3] and protein–membrane interac-
Furthermore, by comparing various heparin-binding pro-
teins, Cardin and Weintraub classified two consensus sequen-
ces responsible for binding, XBBXBX and XBBBXXBX,
where B is a basic residue and X is any other amino acid. It
was suggested that these sequence motifs fold along one face
of an a helix or b sheet.[17] Most heparin-binding sites have a
bipolar structure, in which the basic amino acids are facing
one side; neutral or lipophilic amino acids generally face the
opposite side, regardless of the secondary structure.[18]
A few medium-sized peptides such as protamine ana-
logues,[10,19] antithrombin-derived peptides,[20] and polyargi-
nine peptides,[21] have been reported as potent heparin
antidotes. Small-molecule inhibitors of heparin may alleviate
complications associated with peptide-based heparin anti-
dotes, such as proteolytic stability, distribution, and difficulty
of scale-up. We have previously reported the synthesis of
amphiphilic arylamide oligomers that mimic the biological
properties of antimicrobial peptides and proteins.[4] Herein,
we expand this strategy towards the design of small molecules
that strongly interact with highly sulfated heparin.
tions.[4] Their semirigid structures and their adjustable lengths
mean that they provide excellent starting points for the
elaboration of protein mimics that might be difficult to design
based on small-molecule scaffolds. Here we describe the
design of aryl amide oligomers that compete for heparin–
protein interactions. Heparin, a linear and highly sulfated
polysaccharide, is a crucial component in a variety of
biological processes mediated by specific heparin–protein
interactions including blood coagulation, viral infection, and
cell growth. Heparin has become a commonly used clinical
anticoagulant to prevent and treat thrombotic diseases.[5,6]
However, bleeding complications including hemorrhage and
heparin-induced thrombocytopenia (HIT), which is
a
common immune-mediated disorder, are major adverse
effects associated with heparin therapy.[7] Therefore, frequent
coagulation monitoring may be necessary to minimize the risk
of life-threatening hemorrhages resulting from an overdose,
while at the same time maximizing the anticoagulant efficacy.
Low-molecular-weight (LMW) heparins, like unfractio-
nated (UF) heparin, consist of repeating units of l-iduronic
acid and d-glucosamine (Scheme 1a). LMW heparins have
Our design is based on a 1,3-substituted arylamide
oligomer with additional hydrogen-bonding substituents
included to increase solubility and conformational rigidity
and to provide appropriately spaced cationic groups to
interact with heparin. The thioether and ether groups of the
generic oligomer 1 provide internal hydrogen-bond acceptors,
while simultaneously providing points of attachment for
additional functional groups.
Scheme 1. a) The major repeating unit and b) the antithrombin III
binding pentasaccharide ofheparin.
resolved some of the problems associated with UF-heparin
use, in that they have a more predictable dose response, an
improved bioavailability, and a longer half-life.[8] Protamine is
used extensively as a clinical heparin antidote to neutralize
the anticoagulation function of heparin following cardiovas-
cular surgery[9] but it is not prescribed for use with LMW-
heparin therapy. Protamine treatment can also cause several
side effects mediated by nonimmunologic and immunoglo-
bulin-mediated pathways.[10] Thus, much safer and more
effective agents for neutralizing the anticoagulant function
of heparin are currently of great interest.
The interaction between heparin and antithrombin, a
plasma serine proteinase inhibitor that is the major inhibitor
of the coagulation cascade, facilitates a conformational
change in antithrombin. This conformational change accel-
erates inhibition of coagulation factors such as thrombin and
factor Xa.[11,12] Through X-ray crystallographic analysis of the
antithrombin–pentasaccharide complex[13,14] and several stud-
ies of site-directed mutagenesis of antithrombin,[15,16] it has
been shown that the negatively charged pentasaccharide
binds to the basic amino acids (lysine and arginine) of
antithrombin.
Initially, a molecular dynamics study on binding of one of
the arylamide analogues (shown in Figure 1, compound 8c) to
heparin in aqueous solution was undertaken. Analysis of the
obtained trajectory suggests that these compounds might
interact strongly with the pentasaccharide sequence of
heparin (PDB code: 1E03). Owing to the extensive charge
complementarity between the inhibitor and the heparin
analogue, the binding is dominated by electrostatic interac-
tions. At early stages in the trajectory, the two guanidium
cations of the terminal side chains (R4, Table 1) anchor
compound 8c to the pentasaccharide in a nearly parallel
À
À
mode, by closely interacting with CH2 OSO3 groups of the
two terminal pentasaccharide sugars (Figure 1a). The other
positively (NH3 ) and negatively (COOÀ, OSO3À) charged
+
groups of compound 8c and the pentasaccharide, respectively,
are then aligned. The limited flexibility of the pentasaccharide
(Figure 1a) allows these groups to assume optimal positions,
thereby maximizing the attractive interactions (Figure 1b)
and hydrogen bonding without incurring an overly large loss
in conformational entropy, which might occur with more
flexible small peptides. All the side chains of compound 8c
are involved in the binding, with the terminal ones playing an
6686 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6685 –6689

Angewandte
Chemie
Table 1: Biological activity ofsmall molecules 8a–d.
Compound R4
R5
IC50 [mm] KB [mm][a] HC50 [mm][b]
8a
8b
H
H
NH2
256
6.7
3.2
>1540
>1363
77.9
8c
8d
NH2
NH2
22.5
28.1
1.8
2.0
>1087
927
[a] The dissociation constant (KB) was measured by Schild plot analysis by using
the anti-Factor Xa assay. [b] The HC50 value (measurement ofhemolytic activity)
was obtained by measuring 50% lysis ofhuman erythrocytes.
Figure 1. a) The initial (top) and final (after 7 ns, bottom) conforma-
tions ofthe aryl amide oligomer–pentasaccharide heparin system. In
the initial stage ofthe simulation, the aryl amide backbone undergoes
an internal twist, which facilitates the anchoring of the terminal aryl
amide groups to pentasaccharide. Color code: S: yellow; C: cyan; N:
blue; O: red. Water molecules and hydrogen atoms are not shown for
clarity. b) The final aryl amide oligomer–pentasaccharide conformation,
with electrostatic potential surfaces[29] shown in red (aryl amide
oligomer) and green (pentasaccharide).
Scheme 2. a) Boc-NH(CH2)2OH, DEAD, PPh3, THF; b) 2n LiOH,
MeOH, THF; c) TsCl, DIEA, CH2Cl2; d) Boc-NH(CH2)2SH, DIEA,
CH2Cl2; e) SnCl2·2H2O, NaOAc·3H2O, EtOH, 788C; f) FmocCl, pyri-
dine, THF, 08C; g) 2, (COCl)2, pyridine, DMF (cat.), THF, then 6, Et3N,
DMAP, CH2Cl2; h) 20% PIP in DMF; i) for 8a: 50% TFA in CH2Cl2; f or
8b and 8a: N,N’-bis(tert-butoxycarbonyl)-1H-pyrazole-1-carboxamidine,
DIEA, MeCN/H2O, then 50% TFA in CH2Cl2; f or8c and 8d: RCO2H,
POCl3, pyridine, À158C, then 50% TFA in CH2Cl2. Boc=tert-butoxycar-
bonyl, DEAD=diethylazodocarboxylate, DIEA=N,N-diisopropylethyl-
amine, DMAP=4-dimethylaminopyridine, DMF=N,N-dimethylform-
amide, Fmoc=9-fluorenylmethoxycarbonyl, PIP=piperidine, TFA=tri-
fluoroacetic acid, THF=tetrahydrofuran, Ts=toluene-4-sulfonyl.
additional role of initial anchoring and aligning. However,
some of the pentasaccharide charged groups are pointing
away from the complex, toward the solution.
Based on these observations, we synthesized a series of
aryl amide oligomers, as described in Scheme 2.
We examined the interaction between compounds 8a–d
and LMW heparin (Lovenox, MW = 4500 Daltons) and UF
heparin, respectively, both of which contain an antithrombin-
binding domain (pentasaccharide, Scheme 1b). The ability of
the compounds to compete with antithrombin for binding to
Lovenox at physiological salt concentration (0.15m NaCl) was
assessed by a standard anti-factor Xa chromogenic assay
(Figure 2).
Angew. Chem. Int. Ed. 2005, 44, 6685 –6689
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6687

Communications
a dose-dependent manner. The Schild plot of inhibition for
compounds 8a–d establishes that they are competitive
inhibitors for heparin–antithrombin formation with dissocia-
tion constants (KB) of 6.7, 3.2, 1.8, and 2.0 mm, respectively.
These data also show that 8c was the most potent inhibitor
among compounds 8a–d.
To further address the specificity and affinity of these
foldamers in a stringent biological medium, we examined
their abilities to inhibit clotting in human plasma by measur-
ing the neutralization of UF heparin by 8a–d and protamine
in an activated partial thromboplastin time (aPTT) clotting
assay. The study was conducted with UF heparin because
conventional aPTT clotting assays are a relatively insensitive
method for measuring the activity of LMW heparin. As
shown in Figure 2b, 8a–d and protamine efficiently reversed
heparin-induced aPTT clotting. Also, similar patterns of
neutralization of heparin by 8a–d were observed, which were
consistent with those from the factor Xa assay. The most
active compounds 8c and 8d were only two- to threefold less
potent than protamine (average MW = 5100 Daltons), despite
the fact that these are small molecules.
Finally, as an initial test of toxicity, we measured
hemolytic activities of the compounds against human eryth-
rocytes. None of these small molecules lysed human red cells
at concentrations as high as 1000 mm.
Figure 2. Dose-dependent neutralization ofa) LMW heparin and b) UF
heparin by compounds 8a–d. a) Factor Xa assay. b) aPTT clotting assay
in the presence ofUF heparin (2.42 mgmLÀ1).
In summary, we have described the synthesis of a series of
arylamide oligomers, which interact with LMW heparin and
UF heparin. This interaction inhibits the ability of LMW
heparin to activate antithrombin and, in this way, affects the
biological function of LMW heparin. These arylamide
oligomers are the first example of low-molecular-weight
antagonists that inhibit the anticoagulant function of heparin.
Compound 8a served as an initial lead compound to
inhibit the function of Lovenox. Although it exhibited a low
in vitro inhibition (IC50 = 256 mm, Table 1) against Lovenox,
we were inspired to investigate the possible role of the basic
side chain of the compound in its interaction with polysulfated
heparin. The conversion of the side chains of 8a to give the
polyguanidinylated derivative 8b resulted in a threefold
increase in potency. This result is consistent with observations
reported by Linhardt and co-workers. They suggested that
arginine binds about 2.5 times more tightly than lysine to the
sulfate groups of heparin due to distinct hydrogen bonding.[22]
Therefore, we reasoned that the increased activity of com-
pound 8b arose from differences in hydrogen bonding
between ammonium or guanidinium cations and the sulfate
groups of LMW heparin.
Experimental Section
The synthesis of all compounds is described in the Supporting
Information. The factor Xa assay in the presence of Lovenox and
inhibitor was measured in a chromogenic assay by using reagents
purchased from DiaPharma. In brief, the reagents were reconstituted
in a 0.02m tris(hydroxymethyl)aminomethane (Tris) buffer at pH 8.4.
Antithrombin was prepared with 0.336m NaCl for a physiological salt
concentration (0.15m) in the final reaction. Antithrombin
(0.08 IUmLÀ1
; ;
50 mL) was mixed with Lovenox (11.2 mgmLÀ1
To obtain more potent heparin inhibitors, further chem-
ical modification was necessary. The introduction of extra
positively charged substituents at the terminal amines gave
compounds 8c and 8d, which exhibited a greater than tenfold
increase in potency compared to that of 8a. Additionally, 8d
was found to have a similar in vitro efficacy to 8c, in that both
compounds provide significant improvement and have similar
inhibition constants (IC50 = 28.1 and 22.5 mm, respectively,
Table 1). The introduction of additional positive charges
enhanced the potency of the compounds, as expected from the
molecular dynamics study.
To estimate the binding affinities of compounds 8a–d,
Schild plot analysis, which measures agonist dose–response
curve shifts by a competitive antagonist, was used to evaluate
their capacity to interact with LMW heparin. Table 1 lists the
dissociation constants (KB) of these four compounds. These
compounds prevent the binding of heparin to antithrombin in
1 mL). After the incubation for 5 min, different concentrations of
inhibitor (1 mL) were added to each row of wells; the wells were then
gently shaken and incubated for 20 min. Factor Xa (2.8 nkatmLÀ1
;
50 mL) was then added and the reaction mixture was shaken and
incubated for an additional 10 min. Substrate (S-2765; 10 mL) was
added and the 96-well plate was then shaken and read every 30 s for
7 min at 405 nm on a ThermoLabsystems Multiskan Spectrum. In the
assay for Schild plot analysis, we followed the same protocol as for the
anti-factor Xa assay except the concentrations of Lovenox and
inhibitors were varied.
aPTT activity was resolved by using a BBL Fibrometer (Becton
Dickinson, San Jose, CA) and a commercial kit containing Citrex I
Control Plasma and Cephalinex-Activated PTT Reagent (Bio/Data
Corporation, Horsham, PA). In brief, heparin (5 mL) and antagonist
(5 mL) were preincubated (378C on a Fibrometerꢀs heating block) for
10 s prior to addition of control plasma (0.1 mL) into the test cuvette.
This mixture was allowed to incubate for 2 min at 378C. Meanwhile,
25 mm CaCl2 (0.2 mL) was set to preincubate in a separate cuvette.
Cephalinex (0.1 mL) was pipetted into the test cuvette containing the
6688 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6685 –6689

Angewandte
Chemie
plasma and incubated for exactly 5 min. The preincubated CaCl2
(0.1 mL) was added with simultaneously starting of the Fibrometerꢀs
timer. Once a clot formed, the timer stopped and the clotting time was
recorded. Biochemicals not provided with the kit were heparin
sodium salt from bovine intestinal mucosa (Sigma, H-0777,
150 UnitsmgÀ1) and protamine sulfate from salmon (Sigma, p4020).
The details of the interaction between heparin and compound 8c
were investigated by using molecular dynamics simulations. A
pentasaccharide heparin sequence (charge À11) was placed in a box
(55.69 ” 53.09 ” 60.04Š) with compound 8c (+8 charges, Table 1) and
three Na+ ions to balance the charge. The system was solvated by
TIP3P water,[23] thereby resulting in a size of ꢀ 17500 atoms.
Parameters for compound 8c were derived from the newly developed
force field for aryl amide oligomers,[24] to ensure proper description of
the backbone conformational dynamics, in combination with semi-
empirical (MOPAC) charges.[25] Heparin parameters were developed
partly by analogy with the existing Amber parameters,[26] augmented
[22] J. R. Fromm, R. E. Hileman, E. E. O. Caldwell, J. M. Weiler,
R. J. Linhardt, Arch. Biochem. Biophys. 1995, 323, 279.
[23] W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey,
M. L. Klein, J. Chem. Phys. 1983, 79, 926.
[24] V. Pophristic, S. Vemparala, I. Ivanov, Z. Liu, M. L. Klein, W. F.
DeGrado, J. Phys. Chem. B, unpublished results.
[25] A. Jakalian, B. L. Bush, D. B. Jack, C. I. Bayly, J. Comput. Chem.
2000, 21, 132.
[26] D. A. Pearlman, D. A. Case, J. W. Caldwell, W. S. Ross, I. T. E.
Cheatham, S. DeBolt, D. Ferguson, G. Seibel, P. Kollman,
Comput. Phys. Commun. 1995, 91, 1.
[27] C. J. M. Huige, C. Altona, J. Comput. Chem. 1995, 16, 56.
[28] L. Kale, R. Skeel, M. Bhandarkar, R. Brunner, A. Gursoy, N.
Krawetz, J. Phillips, A. Shinozaki, K. S. Varadarajan, J. Comput.
Phys. 1999, 151, 283.
[29] N. A. Baker, D. Sept, S. Joseph, M. J. Holst, J. A. McCammon,
Proc. Natl. Acad. Sci. USA 2001, 98, 10037.
by parameters for the SO4 groups,[27] partly from semiempirical
À
calculations.[25] The system was equilibrated for 1 ns and the analysis
was performed based on a subsequent 6 ns production run. The
simulations were carried out by using the NAMD2 software[28] with
full electrostatics by use of the Particle Mesh Ewald method and a
time step of 1.5 fs.
Received: April 12, 2005
Published online: August 11, 2005
Keywords: antithrombin · heparin · inhibitors · oligomers ·
.
protein interactions
[1] D. J. Hill, M. J. Mio, R. B. Prince, T. S. Hughes, J. S. Moore,
Chem. Rev. 2001, 101, 3893.
[2] S. H. Gellman, Acc. Chem. Res. 1998, 31, 173.
[3] J. A. Kritzer, M. E. Hodsdon, A. Schepartz, J. Am. Chem. Soc.
2005, 127, 4118.
[4] D. Liu, S. Choi, B. Chen, R. J. Doerksen, D. J. Clements, J. D.
Winkler, M. L. Klein, W. F. DeGrado, Angew. Chem. 2004, 116,
1178; Angew. Chem. Int. Ed. 2004, 43, 1158.
[5] I. Capila, R. J. Linhardt, Angew. Chem. 2002, 114, 426; Angew.
Chem. Int. Ed. 2002, 41, 390.
[6] R. Lever, C. P. Page, Nat. Rev. Drug Discovery 2002, 1, 140.
[7] T. E. Warkentin, M. N. Levine, J. Hirsh, P. Horsewood, R. S.
Roberts, M. Gent, J. G. Kelton, N. Engl. J. Med. 1995, 332, 1330.
[8] J. I. Weitz, N. Engl. J. Med. 1997, 337, 688.
[9] G. R. Jones, R. Hashim, D. M. Power, Biochim. Biophys. Acta
1986, 883, 69.
[10] L.-C. Chang, J. F. Liang, H.-F. Lee, L. M. Lee, V. C. Yang, AAPS
Pharmsci. 2001, 3(2), article 18.
[11] R. D. Rosenberg, P. S. Damus, J. Biol. Chem. 1973, 248, 6490.
[12] S. T. Olson, Y.-J. Chuang, Trends Cardiovasc. Med. 2002, 12, 331.
[13] L. Jin, J. P. Abrahams, R. Skinner, M. Petitou, R. N. Pike, R. W.
Carrell, Proc. Natl. Acad. Sci. USA 1997, 94, 14683.
[14] D. J. D. Johnson, J. A. Huntington, Biochemistry 2003, 42, 8712.
[15] S. Schedin-Weiss, V. Arocas, S. C. Bock, S. T. Olson, I. Bjork,
Biochemistry 2002, 41, 12369.
[16] S. Schedin-Weiss, U. R. Desai, S. C. Bock, S. T. Olson, I. Bjork,
Biochemistry 2004, 43, 675.
[17] A. D. Cardin, H. J. Weintraub, Arteriosclerosis 1989, 9, 21.
[18] H. Margalit, N. Fischer, S. A. Ben-Sasson, J. Biol. Chem. 1993,
268, 19228.
[19] T. W. Wakefield, P. C. Andrews, S. K. Wrobleski, J. Surg. Res.
1996, 63, 280.
[20] S. Onoue, Y. Nemoto, S. Harada, T. Yajima, K. Kashimoto, Life
Sci. 2003, 73, 2793.
[21] J. R. Fromm, R. E. Hileman, E. E. O. Caldwell, J. M. Weiler,
R. J. Linhardt, Arch. Biochem. Biophys. 1997, 343, 92.
Angew. Chem. Int. Ed. 2005, 44, 6685 –6689
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6689