Discovery of allosteric modulators of factor XIa by targeting hydrophobic domains adjacent to its heparin-binding site

Article abstract of DOI:10.1021/jm301757v

To discover promising sulfated allosteric modulators (SAMs) of glycosaminoglycan-binding proteins (GBPs), such as human factor XIa (FXIa), we screened a library of 26 synthetic, sulfated quinazolin-4(3H)-ones (QAOs) resulting in the identification of six molecules that reduced the V _max of substrate hydrolysis without influencing the K_M. Mutagenesis of residues of the heparin-binding site (HBS) of FXIa introduced a nearly 5-fold loss in inhibition potency supporting recognition of an allosteric site. Fluorescence studies showed a sigmoidal binding profile indicating highly cooperative binding. Competition with a positively charged, heparin-binding polymer did not fully nullify inhibition suggesting importance of hydrophobic forces to binding. This discovery suggests the operation of a dual-element recognition process, which relies on an initial Coulombic attraction of anionic SAMs to the cationic HBS of FXIa that forms a locked complex through tight interaction with an adjacent hydrophobic patch. The dual-element strategy may be widely applicable for discovering SAMs of other GBPs.

Full text of DOI:10.1021/jm301757v

Article
pubs.acs.org/jmc
Discovery of Allosteric Modulators of Factor XIa by Targeting
Hydrophobic Domains Adjacent to Its Heparin-Binding Site
Rajesh Karuturi,† Rami A. Al-Horani,† Shrenik C. Mehta,† David Gailani,‡ and Umesh R. Desai*,†
†Department of Medicinal Chemistry and Institute for Structural Biology and Drug Discovery, Virginia Commonwealth University,
Richmond, Virginia 23298, United States
‡Departments of Pathology, Immunology and Microbiology, Vanderbilt University Medical Center, Nashville, Tennessee 37203,
United States
S
* Supporting Information
ABSTRACT: To discover promising sulfated allosteric modulators (SAMs)
of glycosaminoglycan-binding proteins (GBPs), such as human factor XIa
(FXIa), we screened a library of 26 synthetic, sulfated quinazolin-4(3H)-ones
(QAOs) resulting in the identification of six molecules that reduced the Vmax
of substrate hydrolysis without influencing the KM. Mutagenesis of residues of
the heparin-binding site (HBS) of FXIa introduced a nearly 5-fold loss in
inhibition potency supporting recognition of an allosteric site. Fluorescence
studies showed a sigmoidal binding profile indicating highly cooperative
binding. Competition with a positively charged, heparin-binding polymer did
not fully nullify inhibition suggesting importance of hydrophobic forces to
binding. This discovery suggests the operation of a dual-element recognition process, which relies on an initial Coulombic
attraction of anionic SAMs to the cationic HBS of FXIa that forms a locked complex through tight interaction with an adjacent
hydrophobic patch. The dual-element strategy may be widely applicable for discovering SAMs of other GBPs.
Examples of these include oligosaccharides containing
sulfate10−13 and phosphate14 groups, aptamers,15 sulfated-
linked cyclitols,16 and dendritic polyglycerol sulfates.17
Although none has reached the clinic as yet, a pentaphosphate
oligosaccharide called PI-88 is being evaluated in several clinical
trials.
Another concept being advanced to capitalize on the
possibility of numerous druggable GAG−protein interactions
is sulfate-decorated nonsaccharide small molecules. These
include sulfated flavans18,19 and flavones,20−22 sulfated
xanthones,23 and sulfated tetrahydroisoquinolines.24,25 Sulfated
flavans were the first small, nonsugar molecules designed
utilizing a de novo structure-based approach to modulate
antithrombin function.18,19 In contrast, a pharmacophore-based
approach was used to design sulfated isoquinolines,24,25 while
the rest were investigated as analogues of the parent sulfated
flavans. Although promising, the affinity of most of the
designed, sulfated, nonsaccharide, small molecules was found
to be high (>250 μM).18−25
INTRODUCTION
■
Glycosaminoglycan (GAG)-binding proteins (GBPs) play
critical roles in a number of physiological and pathological
responses such as coagulation, immune regulation, angio-
genesis, morphogenesis, viral infection, and cancer.1−4 The
modulation of these responses by GAGs such as heparin,
heparan sulfate (HS), and chondroitin sulfate offers major
opportunities for discovering a large number of therapeutic
agents considering that GAGs present a wide range of
structures for protein recognition. Yet, only one interaction,
the heparin−antithrombin interaction,5 has yielded a clinically
useful agent.
A major problem in discovering GAG-based molecules, for
example, heparin oligomers, as drugs is the rather poor
specificity of their interaction with proteins.6−8 Although the
presence of certain rare residues in heparin/heparan sulfate
(H/HS) chains, such as 3-O-sulfated glucosamine or 2-O-
sulfated glucuronic acid, is believed to indicate specificity, the
majority of HS−protein interactions are likely to be non-
selective because the nature of forces that govern these
interactions is primarily electrostatic, which is nondirectional
and operational over long distances.8,9 This implies that H/HS
recognize practically any collection of electropositive residues,
that is, a group of arginines and lysines, which severely limits
the discovery of highly selective modulators.
Despite these advances, the field suffers severely from the
lack of a generalizable strategy for the rational design of
modulators of GAG−protein interactions. In fact, no real
“design” has been utilized in developing the polymeric
polyanions,9−16 while the computational design of sulfated,
small molecule antithrombin activators18,24 is highly case-
specific. A rational, more broadly applicable strategy would
A number of researchers have attempted the design of H/HS
mimetics that rely on a limited number of anionic groups (e.g.,
sulfate, carboxylate, phosphate) on a smaller saccharide scaffold
to alleviate the difficulties of nonspecificity of H/HS polymers.
Received: November 28, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
performed using silica gel (200−400 mesh, 60 Å) from Sigma−
Aldrich. Chemical reactions sensitive to air or moisture were carried
out under nitrogen atmosphere in oven-dried glassware. Reagent
solutions, unless otherwise noted, were handled under a nitrogen
atmosphere using syringe techniques. Flash chromatography was
performed using the Teledyne ISCO (Lincoln, NE) Combiflash RF
system and disposable normal silica cartridges of 30−50 μm particle
size, 230−400 mesh size, and 60 Å pore size. The flow rate of the
mobile phase was in the range 18−35 mL/min, and mobile phase
gradients of ethyl acetate/hexanes and CH2Cl2/CH3OH were used to
elute compounds.
greatly help in developing selective sulfated modulators of
GBPs.
We recently studied the mechanism of inhibition of human
α-thrombin by sulfated benzofurans,26 small functional
mimetics of GAGs, and found that these molecules recognize
a single arginine residue that is located adjacent to a
hydrophobic domain.27 This gave rise to the possibility of
designing molecules that specifically target the hydrophobic
(hP) domain surrounding the heparin-binding site (HBS) of
proteins. In fact, coagulation enzymes possess hP domains of
varying sizes around their strongly electropositive HBSs. For
example, thrombin contains two clusters of hP residues near its
HBS,27,28 while the corresponding site in factor XIa is
significantly different. Thus, we reasoned that it should be
possible to discover molecules, herein called as sulfated
allosteric modulators (SAMs), by exploiting differential
recognition of hP patches around the HBS. We reasoned that
the strategy would involve (1) initial attraction of an anionic
sulfate group present on a SAM to one or more arginines/
lysines present in the HBS of a heparin-binding protein (HBP)
followed by (2) recognition of an adjacent hP patch on the
HBP to form a complex (Figure 1). Enzymes devoid of either
Proteins and Chromogenic Substrates. Human plasma
proteases including thrombin, factor Xa, factor IXa, factor XIa, and
active-site labeled FXIa, that is, FXIa-DEGR, were obtained from
Haematologic Technologies (Essex Junction, VT). Bovine α-
chymotrypsin and bovine trypsin were obtained from Sigma−Aldrich
(St. Louis, MO). Stock solutions of factor XIa, thrombin, trypsin, and
chymotrypsin were prepared in 50 mM Tris−HCl buffer, pH 7.4,
containing 150 mM NaCl, 0.1% PEG8000, and 0.02% Tween80. Stock
solutions of factor Xa were prepared in 20 mM Tris−HCl buffer, pH
7.4, containing 100 mM NaCl, 2.5 mM CaCl2, 0.1% PEG8000, and
0.02% Tween80. Chromogenic substrates, Spectrozyme TH (H-D-
cyclohexylalanyl-Ala-Arg-p-nitroanilide), Spectrozyme factor Xa (Me-
thoxycarbonyl-D-cyclohexylglycyl-Gly-Arg-p-nitroanilide), and Spec-
trozyme CTY were obtained from American Diagnostica (Greenwich,
CT). Factor XIa chromogenic substrate (S-2366, L-PyroGlu-Pro-Arg-
p-nitroanilide·HCl) and trypsin substrate (S-2222, Benzoyl-Ile-Glu(γ-
OH and -OCH3)-Gly-Arg-p-nitroanilide·HCl) were obtained from
Diapharma (West Chester, OH). Bovine unfractionated heparin
(UFH) was purchased from Sigma−Aldrich (St. Louis, MO). Pooled
normal human plasma for coagulation assays was purchased from
Valley Biomedical (Winchester, VA). Activated partial thromboplastin
time reagent containing ellagic acid, thromboplastin-D, and 25 mM
CaCl2 was obtained from Fisher Diagnostics (Middletown, VA).
1
Chemical Characterization of Compounds. H and 13C NMR
were recorded on a Bruker-400 MHz spectrometer in either CDCl3,
CD3OD, acetone-d6, DMSO-d6, or D2O. Signals, in parts per million
(ppm), are either relative to the internal standard or to the residual
peak of the solvent. The NMR data are reported as chemical shift
(ppm), multiplicity of signal (s = singlet, d = doublet, t = triplet, q =
quartet, dd = doublet of doublet, m = multiplet), coupling constants
(Hz), and integration. Electrospray ionization mass spectrometry (ESI-
MS) of compounds were recorded using a Waters Acquity TQD MS
spectrometer in positive or negative ion modes. Samples were
dissolved in methanol and infused at a rate of 20 μL/min. For high-
resolution MS (HRMS) measurements, a Perkin−Elmer AxION 2
TOF MS was used in negative ion mode. Ionization conditions on
both instruments were optimized for each compound to maximize the
ionization of the parent ion. Generally, the extractor voltage was set to
3 V, the Rf lens voltage was 0.1 V, the source block temperature was
set to 150 °C, and the desolvation temperature was about 250 °C. The
purity of each final compound was greater than 95% as determined by
ultraperformance liquid chromatography−mass spectrometry (UPLC-
MS).
General Procedure for Synthesis of Sulfated Quinazolin-
4(3H)-ones. Sulfation of phenolic precursors was achieved using
microwave-assisted chemical sulfation as described earlier.29,30 Briefly,
to a stirred solution of polyphenol in anhydrous CH3CN (1−5 mL) at
room temperature, Et3N (10 equiv per −OH group) and Me3N/SO3
complex (6 equiv per −OH) were added. The reaction vessel was
sealed and microwaved (CEM Discover, Cary, NC) for 30 min at 90
°C. The reaction mixture was cooled and transferred to a round-
bottom flask, and the volume was reduced as much as possible under
low pressure conditions at 25 °C. The reaction mixture was then
directly loaded on to a flash chromatography column and purified
using a dichloromethane and methanol solvent system (5−20%) to
obtain the sulfated quinazolin-4(3H)-ones (QAOs). The samples were
concentrated and reloaded onto a SP Sephadex C-25 column for
sodium exchange. Appropriate fractions were pooled, concentrated in
vacuo, and lyophilized to obtain a white powder. Spectral character-
Figure 1. Strategy for the design of a SAM of a GBP exploiting the
difference in hydrophobicity (hP, shown as light colored patch) on the
periphery of a HBS (shown as blue ellipse with positive charges). A
SAM binds an enzyme (shown by red dashed line), e.g., E1, only if it
contains both hP and HBS. E2 and E3 do not recognize the SAM
because of an absence of either hP or HBS. This generates selectivity
of recognition.
the HBS or the hP domain would not bind the SAM and hence
escape inhibition. Only enzymes possessing the HBS and an
appropriate hP domain will be targeted by SAMs. The potency
of inhibition will be dependent on the complementarity of
SAM’s hydrophobic scaffold with the hP domain on the
enzyme. In essence, this strategy revolves around electronic
steering of a small molecule to the HBS of the protein due to
nondirectionality of an initial, weak ionic bond followed by
filtering and tight locking of an optimal hydrophobic SAM
scaffold. This work describes that success of this strategy in the
discovery of SAMs of human factor XIa.
EXPERIMENTAL PROCEDURES
■
Chemicals, Reagents, and Analytical Chemistry. Anhydrous
CH2Cl2, THF, CH3CN, DMF, DMA, and acetone were purchased
from Sigma−Aldrich (Milwaukee, WI) or Fisher (Pittsburgh, PA) and
used as such. Other solvents used were of reagent gradient and used as
purchased. Analytical thin layer chromatography (TLC) was
performed using UNIPLATE silica gel GHLF 250 μm precoated
plates (ANALTECH, Newark, DE). Column chromatography was
B
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
istics of all the sulfated compounds 3aS−3gS, 7S−20S, and 25S−29S
55.49, 50.14, 26.23, 25.01, 20.54. MS (ESI) calculated for
C37H29N7Na2O12S2 [(M − Na)]−, m/z 850.78, found [(M − Na)]−,
m/z 850.39.
are listed below. See the Supporting Information for details.
1
(3aS). H NMR (D2O, 400 MHz): 7.89−7.87 (m, 1H), 7.69−7.63
(m, 3H), 7.56−7.52 (m, 1H), 7.48−7.45 (m, 2H), 7.36 (t, J = 7.8 Hz,
1H). 13C NMR (D2O, 100 MHz): 151.41, 135.56, 133.50, 130.56,
127.52, 125.74, 125.40, 124.97, 120.78, 119.42. MS (ESI) calculated
for C14H9N2NaO5S, [(M − H)]−, m/z 339.01, found [(M − Na)]−,
m/z 317.17.
(10S). 1H NMR (D2O, 400 MHz): 7.86 (s, 1 H), 7.38−7.07 (m, 10
H), 6.86−6.70 (m, 3 H), 5.06 (s, 2 H), 4.35 (s, 2 H), 3.62 (d, J = 6.4
Hz, 6 H), 1.73 (s, 2 H), 1.36 (s, 2 H). 13C NMR (D2O, 100 MHz):
165.04, 164.63, 157.86, 157.63, 150.37, 150.31, 149.74, 149.40, 142.95,
142.28, 142.16, 134.29, 134.18, 133.67, 133.45, 126.16, 125.71, 125.56,
124.38, 122.40, 121.59, 120.91, 113.39, 113.31, 111.94, 55.67, 55.63,
50.09, 42.42, 26.23, 25.01. MS (ESI) calculated for C37H31N7Na2O12S2
[(M − Na)]−, m/z, 852.79, found [(M − Na)]−, m/z 852.40.
(11S). 1H NMR (D2O, 400 MHz): 8.11 (s, 1 H), 7.68 (s, 1 H), 7.59
(s, 1 H), 7.47−7.33 (m, 4 H), 7.27−7.19 (m, 3 H), 7.12−7.08 (m, 2
H), 7.03−6.99 (m, 3 H), 6.83−6.81 (m, 2 H), 5.25 (s, 2 H), 4.55 (t, J
= 6.4 Hz, 2 H), 4.03 (t, J = 6.4 Hz, 2 H), 2.27−2.24 (m, 2 H). 13C
NMR (D2O, 100 MHz): 167.17, 166.94, 160.26, 160.16, 158.53,
158.51, 152.81, 143.73, 140.30, 140.27, 134.74, 134.65, 130.27, 130.18,
128.70, 127.60, 127.48, 125.69, 124.31, 124.22, 120.64, 120.45, 118.57,
116.07, 115.99, 115.92, 115.87, 64.81, 60.94, 48.13, 30.01. MS (ESI)
calculated for C34H25N7Na2O10S2 [(M − Na)]−, m/z 778.71, found
[(M − Na)]−, m/z 778.11.
1
(3bS). H NMR (DMSO-d6, 400 MHz): 8.23 (d, J = 8.7 Hz, 1 H),
8.11 (m, 2 H), 7.66 (m, 2 H), 7.35 (m, 1 H), 7.26 (d, J = 8.8 Hz, 1 H),
6.87 (d, J = 8.7 Hz, 1 H). 13C NMR (DMSO-d6, 100 MHz): 170.5,
158.15, 139.1, 129.49, 128.64, 126.66, 125.73, 119.38, 115.09. MS
(ESI) calculated for C14H9N2NaO5S, [(M − H)]−, m/z 339.01, found
[(M − Na)]−, m/z 317.17.
(3cS). 1H NMR (D2O, 400 MHz): 8.13−8.11 (m, 1 H), 7.86−7.82
(m, 1 H), 7.72−7.68 (m, 3 H), 7.54 (t, J = 8.0 Hz, 1 H), 7.48 (t, J =
2.1 Hz, 1 H). 13C NMR (D2O, 100 MHz): 152.17, 135.76, 134.72,
127.81, 125.96, 118.56, 118.31. MS (ESI) calculated for
C14H8N2Na2O9S2 [(M − Na)]−, m/z 435.33, found [(M − Na)]−,
m/z 435.10.
(3dS). 1H NMR (D2O, 400 MHz): 7.96 (d, J = 8.0 Hz, 1H), 7.71 (t,
J = 7.7 Hz, 1 H), 7.53 (d, J = 8.2 Hz, 1 H), 7.47−7.35 (m, 4 H), 3.90
(s, 3 H). 13C NMR (100 MHz, D2O): 151.45, 142.88, 135.63, 130.40,
127.48, 125.86, 122.82, 120.64, 119.49, 112.57, 56.27. MS (ESI)
calculated for C15H11N2NaO6S [(M − Na)]−, m/z 347.31, found [(M
− Na)]−, m/z 347.16.
1
(12S). H NMR (D2O, 400 MHz): 7.96 (s, 1 H), 7.84(s, 1 H),
7.57(d, J = 8.5 Hz, 2 H), 7.44 (d, J = 7.2 Hz, 1 H), 7.17−6.9 (m, 9 H),
6.8−6.5 (m, 3 H), 5.24 (s, 2 H), 4.31 (s, 2 H), 3.6 (s, 2 H), 1.69 (s, 2
H), 1.28 (s, 2 H). 13C NMR (D2O, 100 MHz): 165.27, 158.43, 157.87,
153.18, 151.46, 149.68, 149.43, 142.81, 138.07, 133.99, 129.51, 126.61,
125.61, 125.45, 125.17, 123.55, 120.86, 113.19, 65.89, 50.15, 42.43,
26.12, 24.91. ESI-TOF (HRMS) calculated for C35H27N7Na2O10S2
[(M − Na)]−, m/z 792.1164, found [(M − Na)]−, m/z 792.1194 and
[(M − 2Na)]2−, m/z 384.5643.
(3eS). 1H NMR (D2O, 400 MHz): 7.70−7.68 (m, 1 H), 7.60−7.54
(m, 2 H), 7.45−7.42 (m, 1 H), 7.34 (d, J = 8.1 Hz, 1 H), 7.24−7.20
(m, 1 H), 6.95 (d, J = 8.8 Hz, 1 H), 3.78 (s, 3H). 13C NMR (D2O, 100
MHz): 154.27, 139.79, 135.42, 127.07, 126.69, 125.60, 123.91, 121.70,
118.89, 113.17, 56.08. MS (ESI) calculated for C15H11N2NaO6S [(M
− Na)]−, m/z 347.31, found [(M − Na)]−, m/z 347.16.
(3fS). 1H NMR (D2O, 400 MHz): 7.90 (d, J = 0.9 Hz, 1 H), 7.65 (t,
J = 1.3 Hz, 1 H), 7.63 (d, J = 1.3 Hz, 1 H), 7.36 (t, J = 7.2 Hz, 1 H),
7.04 (s, 2 H), 3.86 (s, 6 H). 13C NMR (D2O, 100 MHz): 154.24,
135.52, 132.18, 129.95, 127.47, 125.74, 119.42, 105.18, 56.44. MS
(ESI) calculated for C16H13N2NaO7S [(M − Na)]−, m/z 377.34,
found [(M − Na)]−, m/z 377.17.
(13S). 1H NMR (D2O, 400 MHz): 7.79−7.75 (m, 3 H), 7.61(d, J =
7.8 Hz, 2 H), 7.17−7.02 (m, 9 H), 6.81−6.69 (m, 3 H), 5.13 (s, 2H),
4.25 (s, 2 H), 3.61 (s, 2 H), 1.65 (s, 2 H), 1.27 (s, 2 H). 13C NMR
(D2O, 100 MHz): 165.41, 164.97, 158.34, 158.12, 153.52, 153.46,
149.64, 149.1, 142.82, 133.85, 133.62, 133.52, 129.57, 126.55, 126.40,
125.25, 124.78, 122.47, 120.93, 120.86, 113.38, 113.25, 65.97, 52.1,
50.07, 26.12, 24.91. ESI-TOF (HRMS) calculated for
C35H27N7Na2O10S2 [(M − Na)]−, m/z 792.1164, found [(M −
Na)]−, m/z 792.1172 and [(M − 2Na)]2−, m/z 384.5628.
1
(3gS). H NMR (D2O, 400 MHz): 7.94 (s, 1 H), 7.79−7.72 (m, 2
H), 7.52−7.38 (m, 4 H). 13C NMR (D2O, 100 MHz): 154.18, 149.45,
135.61, 131.95, 127.73, 126.00, 125.85, 119.45, 116.27, 55.61, 42.50,
38.73. MS (ESI) calculated for C14H8N2Na2O9S2 [(M − Na)]−, m/z
435.33, found [(M − Na)]−, m/z 435.16.
(14S). H NMR (D2O, 400 MHz): 7.84 (s, 1 H), 7.78 (d, J = 8.6
1
Hz, 2 H), 7.60(s, 1 H), 7.40 (d, J = 7.6 Hz, 2 H), 7.22−6.99 (m, 8 H),
6.78 (d, J = 7.8 Hz, 1 H), 6.68−6.66 (m, 2 H), 5.15 (s, 2 H), 4.27 (s, 2
H), 3.65 (s, 2 H), 1.68 (s, 2 H), 1.30 (s, 2 H). 13C NMR (D2O, 100
MHz): 165.40, 158.44, 153.90, 151.74, 149.35, 143.14, 138.08, 134.1,
133.73, 129.86, 126.85, 125.65, 125.06, 123.73, 122.72, 121.19, 121.01,
113.66, 66.33, 59.61, 50.33, 26.36, 25.24. MS (ESI) calculated for
C35H27N7Na2O10S2 [(M − Na)]−, m/z 792.76, found [(M − Na)]−,
m/z 792.14.
1
(7S). H NMR (D2O, 400 MHz): 8.39−8.37 (m, 1 H), 8.33 (s, 1
H), 8.26 (t, J = 1.9 Hz, 1 H), 8.20 (t, J = 1.9 Hz, 1 H), 8.16−8.15 (m, 1
H), 8.02 (d, J = 8.1 Hz, 2H), 7.89−7.82 (m, 4 H), 7.55−7.50 (m, 3
H), 7.41−7.30 (m, 2 H), 7.24−7.21 (m, 1 H), 5.76 (s, 2 H), 4.61 (t, J
= 6.4 Hz, 2 H), 4.64 (t, J = 6.9 Hz, 2 H), 2.25 (s, 3 H), 2.07−2.03 (m,
2 H), 1.84 (m, 2 H). 13C NMR (D2O, 100 MHz): 169.29, 166.11,
165.81, 158.70, 157.85, 153.85, 151.10, 150.82, 142.04, 138.84, 138.36,
134.40, 134.11, 129.69, 128.81, 127.58, 127.38, 126.98, 125.46, 124.80,
124.22, 123.27, 123.20, 123.16, 122.95, 121.29, 120.44, 114.51, 114.50.
MS (ESI) calculated for C37H30N7NaO8S [(M − Na)]−, m/z 732.73,
found [(M − Na)]−, m/z 732.38.
1
(15S). H NMR (D2O, 400 MHz): 7.85 (d, J = 8.7 Hz, 2 H), 7.78
(s, 1 H), 7.65 (d, J = 8.6 Hz, 2 H) 7.23−7.05 (m, 9 H), 6.97−6.94 (m,
1 H), 6.38−6.72 (m, 2 H), 5.19 (s, 2 H), 4.27 (s, 2 H), 3.57 (s, 2 H),
1.68 (s, 2 H), 1.21 (s, 2 H), 0.95 (s, 2 H). 13C NMR (D2O, 100 MHz):
165.06, 158.48, 158.27, 153.49, 153.36, 149.79, 149.42, 142.84, 133.85,
129.63, 129.53, 126.55, 125.59, 122.45, 121.0, 120.79, 113.45, 112.38,
66.33, 58.85, 50.23, 28.92, 27.05, 22.31. ESI-TOF (HRMS) calculated
for C36H29N7Na2O10S2 [(M − Na)]−, m/z 806.1321, found [(M −
Na)]−, m/z 806.1327 and [(M − 2Na)]2−, m/z 391.5690.
1
(8S). H NMR (D2O, 400 MHz): 7.53−6.53 (m, 16 H), 5.11 (s, 2
H), 4.24 (s, 2 H), 3.62 (s, 2 H), 1.65 (s, 2 H), 1.26 (s, 2 H). 13C NMR
(D2O, 100 MHz): 154.62, 151.52, 151.37, 151.02, 149.67, 149.38,
145.07, 142.68, 142.42, 138.21, 138.08, 136.39, 134.55, 133.70, 133.54,
130.28, 129.57, 127.21, 126.54, 126.31, 125.64, 125.31, 125.07, 124.33,
123.18, 122.26, 122.08, 121.17, 120.83, 119.94, 118.94, 113.42, 113.32,
65.92, 59.25, 55.53, 50.04, 39.56, 26.28, 25.00. MS (ESI) calculated for
C35H27N7Na2O10S2 [(M − Na)]−, m/z, 792.74, found [(M − Na)]−,
m/z 792.36.
(16S). 1H NMR (D2O, 400 MHz): 7.89 (d, J = 8.6 Hz, 2 H), 7.73−
7.69 (m, 3 H), 7.23−7.10 (m, 9 H), 6.96−6.94 (m, 1 H), 6.77−6.69
(m, 2 H), 5.15 (s, 2 H), 4.16 (s, 2 H), 3.57 (s, 2 H), 1.50 (s, 2 H), 1.02
(s, 2 H), 0.80 (s, 4 H). 13C NMR (D2O, 100 MHz): 158.57, 158.40,
153.54, 149.92, 149.70, 142.78, 133.99, 129.67, 129.52, 126.26, 125.74,
122.37, 121.0, 120.81, 113.54, 66.37, 55.49, 42.41, 29.33, 25.55, 24.75,
12.32. ESI-TOF (HRMS) calculated for C37H31N7Na2O10S2 [(M −
Na)]−, m/z 820.1477, found [(M − Na)]−, m/z 820.1467 and [(M −
2Na)]2−, m/z 398.5775.
(9S). 1H NMR (D2O, 400 MHz): 7.97 (s, 1 H), 7.87 (s, 1 H), 7.58−
7.53 (m, 2 H), 7.24−6.99 (m, 9 H), 6.74−6.59 (m, 3 H), 5.25 (s, 2 H),
4.33 (t, J = 5.9 Hz, 2 H), 3.70−3.65 (m, 2 H), 2.23 (s, 3 H), 1.72 (s, 2
H), 1.32 (s, 2 H). 13C NMR (100 MHz, D2O): 172.29, 165.30, 165.09,
157.94, 157.09, 151.76, 151.51, 150.50, 149.65, 149.31, 142.83, 139.07,
138.07, 133.82, 129.67, 126.72, 125.74, 125,64, 125.30, 123.59, 122.46,
122.17, 120.87, 118.43, 118.24, 116.86, 113.57, 113.47, 65.96, 59.15,
1
(17S). H NMR (D2O, 400 MHz): 8.17 (s, 1 H), 7.87 (s, 1 H),
7.34−7.13(m, 10 H), 6.96−6.87 (m, 2 H), 6.68−6.57 (m, 3 H), 5.38
(s, 2 H), 4.24 (s, 2 H), 3.74 (s, 2 H), 3.64 (s, 3 H), 1.81 (s, 2 H), 1.43
(s, 2 H). 13C NMR (D2O, 100 MHz): 165.40, 158.05, 157.44, 153.43,
C
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
151.25, 148.62, 142.76, 139.62, 134.24, 133.92, 129.63, 126.66, 126.39,
125.95, 125.32, 123.47, 122.83, 122.24, 122.06, 120.73, 113.27, 112.38,
66.33, 55.85, 50.27, 26.13, 24.98. MS (ESI) calculated for
C36H29N7Na2O11S2 [(M − Na)]−, m/z 822.79, found [(M − Na)]−,
m/z 822.16.
(29S). 1H NMR (D2O, 400 MHz): 7.88 (d, J = 8.6 Hz, 2 H), 7.74−
7.67 (m, 3 H), 7.26−7.07 (m, 7 H), 6.94−6.90 (m, 2 H), 6.64−6.62
(m, 3 H), 5.22 (s, 2 H), 4.21 (s, 2 H), 3.50 (s, 2 H), 1.66 (s, 2 H), 1.16
(s, 2 H), 0.91 (s, 2 H). 13C NMR (D2O, 100 MHz): 165.28, 158.55,
158.11, 153.59, 153.42, 149.59, 134.11, 133.69, 129.70, 129.58, 126.41,
122.00, 121.07, 120.95, 113.28, 113.16, 65.73, 48.20, 42.41, 28.94,
22.10, 12.34. ESI-TOF (HRMS) calculated for C36H29N7Na2O10S2
[(M − Na)]−, m/z 806.1321, found [(M − Na)]−, m/z 806.1316 and
[(M − 2Na)]2−, m/z 391.5695.
Direct Inhibition of Factor XIa by Sulfated QAOs. A
chromogenic substrate hydrolysis assay using a microplate reader
(FlexStation III, Molecular Devices) was used to measure direct
inhibition of FXIa, as described earlier.31 Generally, each well of the
96-well microplate had 85 μL of pH 7.4 buffer (50 mM Tris−HCl, 150
mM NaCl, 0.1% PEG8000, and 0.02% Tween80) to which 5 μL of a
potential FXIa inhibitor (or solvent reference) was added followed by
5 μL of FXIa to give 0.765 nM in the well. After 10 min incubation at
37 °C, 5 μL of FXIa substrate S2366 (345 μM in the well) was rapidly
added, and the residual FXIa activity was measured from the initial rate
of increase in absorbance at 405 nm. Stocks of FXIa inhibitors were 20
mM, which were then serially diluted to give 12 different aliquots in
the plate wells. Relative residual FXIa activity at each concentration of
the inhibitor was calculated from the ratio of FXIa activity in the
presence and absence of the inhibitor. Logistic eq 1 was used to fit the
dose dependence of the residual proteinases activity to obtain the
potency (IC50) and efficacy (ΔY) of inhibition. In this equation, Y is
the ratio of residual factor XIa activity in the presence of inhibitor to
that in its absence (fractional residual activity), YM and Y0 are the
maximum and minimum possible values of the fractional residual
proteinase activity, IC50 is the concentration of the inhibitor that
results in 50% inhibition of enzyme activity, and HS is the Hill slope.
Nonlinear curve fitting resulted in YM, Y0, IC50 and HS values:
(18S). 1H NMR (D2O, 400 MHz): 8.52−8.50 (m, 2 H), 8.33−8.31
(m, 3 H), 8.02−7.97 (m, 2 H), 7.83−7.77 (m, 4 H), 7.48−7.44 (m, 5
H), 5.77 (s, 2 H), 4.62−4.57 (m, 4 H), 2.20−2.14 (m, 2 H), 1.96−1.91
(m, 2 H). 13C NMR(D2O, 100 MHz): 167.93, 167.39, 160.51, 159.86,
156.45, 154.71, 152.91, 152.73, 144.29, 141.07, 135.32, 135.23, 135.12,
10.83, 128.62, 128.35, 128.02, 126.18, 124.58, 124.53, 122.06, 118.89,
118.09, 116.27, 115.90, 67.63, 61.01, 28.13, 26.93. MS (ESI) calculated
for C35H26N7Na3O14S3 [(M − Na)]−, m/z 910.78, found m/z 910.33.
(19S). 1H NMR (CD3OD, 400 MHz): 8.62 (s, 1 H), 8.45 (d, J = 2.1
Hz, 2 H), 8.34 (t, J = 1.5 Hz, 1 H), 8.14−8.04 (m, 3 H), 7.92−7.82
(m, 4 H), 7.57 (m, 2 H), 7.38 (t, J = 2.1 Hz, 1 H), 7.25 (t, J = 2.1 Hz, 1
H), 5.84 (s, 2 H), 4.66−4.58 (m, 4 H), 2.33 (s, 3 H), 2.22−2.17 (m, 2
H), 1.95−1.91 (m, 2 H). 13C NMR (CD3OD, 100 MHz): 171.02,
168.11, 167.48, 159.77, 159.72, 154.78, 152.97, 152.70, 141.29, 140.82,
135.22, 135.11, 128.72, 128.57, 128.38, 128.30, 124.68, 124.54, 119.83,
119.11, 118.59, 118.28, 118.02, 116.42, 116.37, 67.66, 28.11, 26.86,
21.0. MS (ESI) calculated for C37H28N7Na3O16S3 [(M − 2SO3
−
Na)]−, m/z 764.80, found [(M − 2SO3 − Na)]−, m/z 764.37.
(20S). 1H NMR (D2O, 400 MHz): 8.61 (s, 1 H), 7.35−7.09 (m, 11
H), 6.80−6.70 (m, 3 H), 5.87 (s, 2 H), 4.72 (s, 2 H), 4.61 (s, 2 H),
2.25−2.21 (m, 2 H), 1.99−1.93 (m, 2 H). 13C NMR (D2O, 100
MHz): 168.12, 167.53, 160.10, 159.82, 154.84, 154.64, 153.03, 152.85,
143.93, 141.06, 140.84, 135.28, 135.10, 128.77, 128.63, 128.36, 128.32,
127.18, 124.60, 118.96, 118.54, 118.19, 118.00, 116.42, 116.40, 67.71,
60.67, 28.24, 26.93, 24.99. MS (ESI) calculated for C35H25N7Na4O18S4
[(M − 3SO3 − Na+)]−, m/z 719.83, found [(M − 3SO3 − Na+)]−, m/
z 722.30.
(25S). 1H NMR (D2O, 400 MHz): 8.12 (s, 1 H), 7.82−7.16 (m, 10
H), 6.48 (d, J = 2.1 Hz, 1 H), 6.36 (d, J = 2.1 Hz, 1 H), 5.47 (s, 2H),
4.56 (t, J = 6.1 Hz, 2 H), 4.12 (t, J = 6.4 Hz, 2 H), 3.59 (s, 3 H), 2.0−
1.9 (m, 2 H), 1.68−1.61 (m, 2 H). 13C NMR (D2O, 100 MHz):
173.48, 159.61, 156.42, 156.11, 154.9, 151.38, 149.79, 148.65, 143.65,
143.33, 138.44, 134.38, 133.46, 129.87, 127.24, 127.14, 125.87, 125.79,
125.02, 123.73, 123.11, 122.55, 121.14, 116.09, 113.66, 110.12, 101.10,
100.42, 66.47, 62.52, 56.04, 50.46, 26.19, 25.03. MS (ESI) calculated
for C37H27N5Na4O21S4 [(M − 3Na)]−, m/z 1028.89, found [(M −
3Na)]−, m/z 1028.82.
YM − Y0
Y = Y0 +
50)×HS
1 + 10(log[I] −log IC
0
(1)
Inhibition of Other Proteases of the Coagulation and
Digestive Systems. The inhibition potential of 500 μM SAM
(16S, 15S, 13S) against coagulation enzymes including thrombin and
factor Xa and digestive enzymes including trypsin and chymotrypsin
was evaluated using chromogenic substrate hydrolysis assays reported
in the literature.26,28,31 These assays were performed using substrates
appropriate for the enzyme being studied under conditions closest to
the physiological condition (37 °C and pH 7.4), except for thrombin,
which was performed at 25 °C and pH 7.4. The concentrations of
enzymes and substrates in microplate wells, respectively, were as
follows: 6 nM and 50 μM for thrombin; 1.09 nM and 125 μM for
factor Xa; 2.5 ng/mL and 80 μM for bovine trypsin; and 500 ng/mL
and 240 μM for bovine chymotrypsin. The ratio of the proteolytic
activity of an enzyme in the presence of the sulfated QAO to that in its
absence was used to calculate percent inhibition (%).
Michaelis−Menten Kinetics of Substrate Hydrolysis in
Presence of 14S. The initial rate of S2366 hydrolysis by FXIa was
obtained from the linear increase in absorbance at 405 nm
corresponding to less than 10% consumption of S2366. The initial
rate was measured as a function of various concentrations of the
substrate (0.01−1.6 mM) in the presence of fixed concentration of
14S in 50 mM Tris−HCl buffer, pH 7.4, containing 150 mM NaCl,
0.1% PEG8000, and 0.02% Tween80 at 37 °C. The data was fitted
using the standard Michaelis−Menten eq 2 to determine the KM and
(26S). 1H NMR (D2O, 400 MHz): 8.15 (s, 1 H), 7.57−7.05 (m, 10
H), 6.51 (d, J = 2.1 Hz, 1 H), 6.36 (d, J = 2.1 Hz, 1 H), 5.48 (s, 2 H),
4.56 (t, J = 6.2 Hz, 2 H), 4.12 (t, J = 6.4 Hz, 2 H), 3.49 (s, 3 H), 3.42
(s, 3 H), 2.04−1.98 (m, 2 H), 1.89−1.71 (m, 2 H). 13C NMR (D2O,
100 MHz): 173.48, 159.61, 156.42, 156.11, 154.9, 151.38, 149.79,
148.65, 143.65, 143.33, 138.44, 134.38, 133.46, 129.87, 127.24, 127.14,
125.87, 125.79, 125.02, 123.73, 123.11, 122.55, 121.14, 116.09, 113.66,
110.12, 101.10, 100.42, 66.47, 62.52, 56.04, 50.46, 26.19, 25.03. MS
(ESI) calculated for C38H29N5Na4O22S4 [(M − 3Na)]−, m/z 1058.88,
found [(M − 3Na)]−, m/z 1058.90 [(M + Na)]−.
(27S). 1H NMR (D2O, 400 MHz): 8.12 (s, 1 H), 7.75−7.14 (m, 10
H), 6.52−6.51 (m, 1 H), 6.30−6.28 (m, 1 H), 5.46 (s, 2 H), 4.56−4.53
(m, 2 H), 4.09−4.06 (m, 2 H), 3.5 (s, 2 H), 2.39 (s, 3 H), 2.01−2.0
(m, 2 H), 1.60−1.59 (m, 2 H). 13C NMR (D2O, 100 MHz): 173.31,
172.72, 159.53, 156.41, 156.18, 154.88, 151.81, 150.74, 148.60, 143.57,
143.38, 139.38, 134.47, 133.38, 127.42, 127.29, 126.04, 124.99, 123.22,
122.52, 119.00, 118.79, 117.35, 116.09, 113.79, 109.98, 100.89, 100.38,
66.49, 62.49, 55.98, 50.44, 26.24, 25.01, 20.64. MS (ESI) calculated
forC39H29N5Na4O23S4 [(M − 3Na)]−, m/z 1086.93, found [(M −
3Na)]−, m/z 1086.90.
Vmax
:
V
max[S]
V =
i
KM + [S]
(2)
1
(28S). H NMR (D2O, 400 MHz): 7.76 (d, J = 8.12 Hz, 4 H),
7.28−7.18 (m, 6 H), 7.05−6.91 (m, 6 H), 6.80−6.77 (m, 2 H), 4.92
(s, 4 H), 4.87 (s, 4 H). 13C NMR (D2O, 100 MHz): 164.76, 158.22,
153.35, 149.42, 142.91, 133.90, 133.82, 129.71, 126.67, 125.32, 125.22,
120.98, 112.88, 59.43, 49.90. ESI-TOF (HRMS) calculated for
C36H26N10Na2O10S2 [(M − Na)]−, m/z 845.1178, found [(M −
Na)]−, m/z 845.1142 and [(M − 2Na)]2−, m/z 411.0621.
Thermodynamics of Sulfated QAOs Binding to Human
Factor XIa. Fluorescence experiments were performed using a QM4
spectrofluorometer (Photon Technology International, Birmingham,
NJ) in 50 mM Tris−HCl buffer of pH 7.4 containing 150 mM NaCl
and 0.1% PEG8000 at 37 °C. Fluorescence emission spectra of active-
site labeled FXIa, that is, FXIa-DEGR (250 nM), in the absence and
D
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 1. Synthesis of Monomeric Sulfated QAOs
a
Conditions: (a) NaHSO3, DMA, reflux/overnight, 65−80%; (b) SO3/Me3N, TEA, CH3CN, microwave/30 min, 85−90%.
presence of saturating concentrations of sulfated QAOs (13S−16S)
were recorded using an excitation wavelength of 345 nm. The emission
scan range was set to 350−600 nm in increments of 1 nm. The
excitation and emission slit widths were set at 1 mm and 1.5 mm,
respectively. The equilibrium dissociation constant (KD) of the
sulfated QAOs−FXIa complex was measured using the change in
the fluorescence of the active site dansyl group due to binding.
Titrations were performed by adding aliquots of a solution of sulfated
QAOs (13S−16S) in the above buffer to a fixed concentration of
FXIa-DEGR (250 nM) and monitoring the change in the fluorescence
of FXIa-DEGR at 547 nm (λex = 345 nm). The slit widths on the
excitation and emission side were 1 and 1.5 mm, respectively. The
change in fluorescence at 547 nm was fitted using the standard Hill
equation for ligand binding (eq 3) for cooperative binding to obtain
the apparent dissociation constant (KD,app) of binding. In this equation,
ΔF represents the change in fluorescence following addition of the
sulfated QAO from the initial fluorescence (F0), while ΔFmax
represents the maximal change in fluorescence. The Hill coefficient
“n” is a measure of the cooperativity of binding. Each measurement
was performed at least twice:
175, and 179 in the chymotrypsinogen numbering system), were
changed to alanine individually and as a group using a Quick Change
kit (Stratagene, La Jolla, CA). cDNAs in expression vector pJVCMV
were used to transfect HEK-293 cells as described. Stably expressing
clones were expanded in 175 cm2 flasks, and serum-free media
(Cellgro Complete, Mediatech, Herndon, VA) was collected every 48
h, supplemented with benzamidine (5 mM) and stored at −80 °C
pending purification. Recombinant FXI was purified from conditioned
media on an anti-fXI IgG 1G5.12 affinity column. After loading, the
column was washed with 25 mM Tris−HCl, pH 7.4, 100 mM NaCl,
and eluted with 2 M NaSCN in the same buffer. Protein containing
fractions were pooled, concentrated, and dialyzed, and protein
concentrations were determined by dye-binding assay (Bio-Rad).
Recombinant wild-type FXI or FXI-CD (∼200−300 μg/mL) was
activated with 5 μg/mL FXIIa at 37 °C. Completion of activation was
confirmed by sodium dodecyl sulfate polyacrylamide gel electro-
phoresis (SDS-PAGE). Activated preparations were passed over a
1G5.12 column to separate the protease from FXIIa. In the case of
proteins prepared in FXI-Ser-362,482, the catalytic domains bind to
the column, while the heavy chain passes through.
ΔF
0
[sulfated QAO]n
= ΔFmax
×
RESULTS AND DISCUSSION
F
(KD,app)n + [sulfated QAO]n
■
(3)
Synthesis of the Library of QAOs. To put the dual
element hypothesis of Figure 1 to test, we synthesized a library
of 26 QAOs including 7 monomers and 19 dimers containing
1−4 sulfate groups. The QAO core structure is a well-known
hP scaffold with three-dimensional similarity to the flavonoid
scaffold studied earlier as a HBS ligand.18−22 The QAO core
scaffold was synthesized using a condensation reaction between
anthranilamide and suitably substituted benzaldehyde34 to
obtain QAO monomers 3a−3g containing one or two phenolic
groups (Scheme 1). This is a fairly well-established thermal
cyclodehydration reaction. Yet, the reaction yields decrease as
the number of phenolic groups increase restricting the library
diversity to two phenolic groups on the scaffold. Monomers
3a−3g were sulfated under microwave conditions29,30 to yield
the corresponding sulfated derivatives 3aS−3gS in 85−90%
yields. Sulfation changes the physicochemical characteristics of
the QAO scaffold by introducing significant water solubility.
Water-soluble sulfated molecules have been previously purified
using size-exclusion chromatography.29,30 However, the con-
siderable hydrophobicity of the QAO scaffold allowed for the
use of a traditional dichloromethane−methanol solvent system
using flash chromatography. This greatly eased handling of
these highly polar compounds. The sulfated QAOs were
Prothrombin Time (PT) and Activated Partial Thromboplas-
tinTime (APTT). Clotting time was measured in a standard one-stage
recalcification assay with a BBL Fibrosystem fibrometer (Becton−
Dickinson, Sparles, MD). For PT assays, thromboplastin-D was
reconstituted according to the manufacturer’s directions and warmed
to 37 °C. A 10 μL amount of sulfated quinazolinones (13S, 15S, and
16S), to give the desired concentration, was brought up to 100 μL
with citrated human plasma, incubated for 30 s at 37 °C followed by
addition of 200 μL of prewarmed thromboplastin-D. For the APTT
assay, 10 μL of the same molecules was mixed with 90 μL of citrated
human plasma and 100 μL of prewarmed APTT reagent (0.2% ellagic
acid). After incubation for 4 min at 37 °C, clotting was initiated by
adding 100 μL of prewarmed 25 mM CaCl2, and the time to clot was
noted. The data were fit to a quadratic trend line, which was used to
determine the concentration of the inhibitor necessary to double the
clotting time. Clotting time in the absence of an anticoagulant was
determined in a similar fashion using 10 μL of deionized water and/or
the appropriate organic vehicle and was found to be 17.6 s for PT and
39.9 s for APTT.
Mutagenesis and Expression of Recombinant Factor XIa.
Mutations in the FXIa 170 helix were introduced into a modified
human FXI cDNA (FXI-Ser-362,482), which contains serine
substitutions for Cys362 and Cys482 (FXI numbering), as described
earlier in several papers.32,33 A disulfide bond between these residues
connects the heavy chains and catalytic domains after cleavage at the
activation site, and eliminating the bond allows the catalytic domain of
FXIa (FXIa-CD) to separate from the heavy chain. The basic residues
of the FXIa 170-helix, Lys529, Arg530, Arg532, Lys535 and Lys539 in
the FXI numbering system (corresponding to residues 170, 171, 173,
1
characterized using H and 13C NMR and ESI-MS techniques.
To develop a more diverse library, dimerization of the
monomeric scaffold was desirable. A rather simple tool for
generating a dimeric equivalent is the copper-catalyzed azide−
E
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of Sulfated QAOs 7S−20S
a
Conditions: (a) CuSO4·5H2O (1 mol %), sodium ascorbate (5 mol %), DMF/H2O (1:1), room temperature (rt)/overnight, 80−95%; (b) SO3/
Me3N, TEA, CH3CN, microwave/30 min, 85−90%.
a
Scheme 3. Synthesis of Sulfated QAOs 25S−27S
a
Conditions: (a) K2CO3, propargyl bromide, DMF, rt/2 h, 85−90%; (b) K2CO3, MeI, DMF, rt/4 h, 85−90%; (c) 3N HCl, acetone, reflux/
overnight, 55−60%; (d) CuSO4·5H2O (1 mol %), sodium ascorbate (5 mol %), DMF/H2O (1:1), rt/overnight, 80−95%; (e) SO3/Me3N, Et3N,
CH3CN, microwave/30 min, 85−90%.
alkyne cycloaddition (CuAAC) reaction.35 Thus, QAO alkynes
4a−5d and azides 6a−6h (Scheme 2) were synthesized from
corresponding monomers 3a−3d using acetylation of phenolic
group followed by standard nucleophilic displacement strategy
and and deacetylation in excellent yields (see the Supporting
Information). Varying the linker length (n = 3−6) in the azide
6 scaffold afforded an opportunity of expanding the diversity of
the library. CuAAC in the presence of aqueous CuSO4 (1 mol
%) and sodium ascorbate (5 mol %) gave 1,2,3-triazoles 7−20
(Scheme 2). The high reaction yields (80−95%) observed for
the QAO series coupled with an essentially single product
suggested 1,4-substitution of the triazole, as established in the
literature.35 Intermediates 7−20 were sulfated using the
microwave conditions developed for monomers to obtain
sulfated QAOs 7S−20S (Scheme 2). The reaction resulted in
sulfation of each available −OH group to give a single
F
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 4. Synthesis of Sulfated QAO 28S
a
Conditions: (a) CuSO4·5H2O (2 mol %), sodium ascorbate (10 mol %), DMF/H2O (1:1), rt/overnight, 80%; (b) SO3/Me3N, Et3N, CH3CN,
microwave/30 min, 90%.
a
Scheme 5. Synthesis of Sulfated QAO 29S
a
Conditions: (a) CpRuCl(PPh3)2, dioxane, 60 °C/overnight, 80%; (b) SO3/Me3N, Et3N, CH3CN, microwave/30 min, 90%.
persulfated product, which simplified purification resulting in
80−90% yields.
In combination, the library of 26 sulfated QAOs was
synthesized through the use of simple synthetic tools in
generally high yields. Structural diversity in this family of
sulfated hydrophobic small molecules arises from the number
of sulfate groups on the core scaffold (1−4), the position of
sulfate groups (either 2, 3, 4, or 5 on the aromatic ring), the
type of core scaffolds (QAO or flavonoid), the type of linker (1
or 2 triazoles), the length of linker (8−11 intervening atoms),
and the geometry of the linker (1,4- or 1,5-triazole). This is a
first small library of potential GAG mimetics that is based on a
fully synthetic, heterocyclic scaffold. The success of this
synthetic venture indicates the feasibility of the development
of other novel GAG mimetics based on a sulfated hydrophobic
scaffold.
Compounds 7S−20S contain an identical QAO unit at either
end and will be referred to herein as homo-“click” dimers,
although these are not truly symmetric. To explore further
structural dependence, inhibitors 25S−27S were synthesized
(Scheme 3). These contain a flavonoid core on one end and a
QAO core on the other (hetero-“click” dimers). The major
structural difference between the two types of click dimers is
the higher sulfation level of the flavonoid scaffold. Utilizing an
intramolecular H-bond-dependent protection−deprotection
strategy, flavonoid 21b was transformed into propargyl
derivative 24, which gave 25−27 upon CuAAC with azides
6a, 6d, and 6h (Scheme 3). Sulfation of these molecules
resulted in flavonoid−QAO hetero-click dimers 25S−27S.
To study the dependence of inhibition on the geometry of
the linker, a double-click sulfated QAO dimer 28S was
synthesized using a slight modification of the strategy
developed for homo-click dimers (Scheme 4). Inhibitor 28S
contains two 1,4-triazole units, instead of one 1,4-triazole
moiety in 7S−21S and 25S−27S. Finally, geometric isomerism
in the linker was also studied. Whereas all click dimers
contained a 1,4-substituted triazole moiety, 29S contained a
1,5-substituted triazole moiety. Inhibitor 29S was synthesized
from phenolic precursor 29, which was synthesized in high
yields using a ruthenium-catalyzed cycloaddition36 of azide 6f
and alkyne 5B (Scheme 5). Comparison of 1H NMR spectra of
regioisomers 15 and 29 shows that the methylenic protons
attached to the triazole moiety in each case was significantly
different. Whereas 1,4-substituted triazole 15 showed the CH2
at 5.81 δ, 29 displays the corresponding signal at 5.96 δ
confirming the difference in geometries.
Inhibition of Human Factor XIa by Sulfated QAOs. The
library of sulfated QAOs was screened for inhibition of human
FXIa and other coagulation enzymes using chromogenic
substrate hydrolysis assay, as described earlier.22,28,31 The
sigmoidal decrease in the initial rate of protease activity (on a
semilog plot) as a function of ligand concentration was fitted
using the logistic dose−response equation to calculate the IC50.
Of the 26 sulfated QAOs studied, 16 exhibited inhibition of
factor XIa at pH 7.4 and 37 °C (Figure 2A) suggesting a good
“hit” ratio arising from the novel SAM discovery strategy. Most
importantly, the unsulfated precursors of the 26 sulfated
molecules did not inhibit factor XIa at all highlighting the
−
importance of the anionic group (OSO3 ) as a recognition
element.
The range of inhibitory potency was found to be reasonable
(50 to >1000 μM), while the efficacy for nearly all inhibitors
was very high (>85%). Inhibitor 26S displayed an efficacy of
approximately 50%, which supports the possibility of structure-
G
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
This is an unusual observation. Typically, conformational
flexibility reduces the potency of inhibition. The result suggests
that an extended linker probably serves to place the two QAO
scaffolds better in two hydrophobic regions of FXIa.
Comparison of 28S and 16S also supports this conclusion.
Whereas 28S contains two triazole rings, 15S contains only one
triazole moiety on an otherwise identical base scaffold. Both
inhibitors contain a minimum of 10 linker atoms, yet the
additional triazole of 28S is likely to impart significant rigidity
to the linker in comparison to that for 15S. Inhibitor 28S is
approximately 6-fold less potent than 15S (Tables 1 and 2).
Another factor that appears to play a role is the geometry of
the triazole ring. Whereas 1,4-triazole containing 15S displayed
a potency of 59 μM, the potency dropped to 94 μM with 29S,
which contained a 1,5-triazole moiety. Likewise, nonsulfate
substitutions, such as acetyl (e.g., 9S and 19S) or methoxy (e.g.,
17S and 26S), also introduce variations in inhibitor potency
suggesting a small but significant contribution of these positions
in binding.
To assess whether the sulfated QAOs inhibited serine
proteases related to human FXIa, we screened the inhibitors
against thrombin, factor Xa, trypsin, and chymotrypsin.
Screening was performed using appropriate chromogenic
substrates, as described earlier.28,31 Essentially, no inhibition
was observed at concentrations as high as 500 μM (Figure 2B).
This suggested high selectivity of sulfated QAOs for targeting
human FXIa. This is not completely unexpected as the design
strategy (Figure 1) should inherently engineer selectivity. Both
the HBS and hP domain(s) are expected to be structurally
different even on related enzymes; thus, engineering enhanced
selectivity. In comparison, enzymes that possess only on one
type of binding site (either HBS or hP) should not be targeted
by sulfated QAOs.
Overall, the work led to identification of six inhibitors (8S,
13S−16S, and 28S) that displayed IC50 values less than 100
μM (Tables 1 and 2). All six molecules bear one sulfate group
at either end of the molecule in a rather symmetric location.
Molecules that are either more or less sulfated than these six are
much less potent. The inhibition potency was significantly
dependent on the substitution pattern of the sulfated QAO
scaffold.
Comparison of Inhibition Potency of Sulfated QAOs
With GAGs and GAG Mimetics. The best SAM discovered in
this study, 16S, displayed an IC50 of 52 μM (Table 1). This
potency represents an excellent lead for the discovery of
advanced molecules. Yet, this potency compares favorably to
several sulfated molecules designed so far in the literature or
available naturally. For example, rationally designed small
mimetics of heparin pentasaccharide displayed antithrombin
affinities in the range 10−1000 μM.18,19,22,24,25 Similarly, a HS
octasaccharide that recognizes herpes simplex virus glycopro-
tein D binds with an affinity of 19 μM.37 Another GAG
oligomer, that is, a dermatan sulfate (DS) hexasaccharide, binds
to its target, heparin cofactor II, with an affinity of 20 μM.38,39
Both these GAG oligomers, the HS octasaccharide and DS
hexasaccharide, are thought to possess high selectivity for their
targets and yet display relatively moderate affinities. We predict
that rational optimization of 16S through computational and/
or structure-based approaches will yield significant increase in
potency. Such approaches have in fact resulted in better GAG
mimetics, as exemplified by the 1000-fold increase in potency
shown by sulfated benzofurans dimers from their monomeric
precursors.26,40 Our recent report on sulfated pentagalloyl
Figure 2. (A) Representative profiles of direct inhibition of human
factor XIa (FXIa) by sulfated QAOs. The inhibition of FXIa was
measured spectrophotometrically through a S2366 hydrolysis assay at
pH 7.4 and 37 °C. Solid lines represent sigmoidal fits to the data to
obtain IC50, YM, and Y0 using eq 1, as described in the Experimental
Procedures. (B) Proteolytic activity of human thrombin (Thr), factor
Xa (FXa), trypsin (Tryp), and chymotrypsin (ChTryp) by 500 μM
sulfated QAO (16S, 15S, and 13S) using the chromogenic substrate
assay. The assays were performed using substrates appropriate for the
enzyme being studied under conditions closest to the physiological
condition. The ratio of the proteolytic activity of an enzyme in the
presence of the sulfated QAO to that in its absence was used to
determine percent activity (%).
dependent “allosteric modulation”. Molecules with only one
sulfate group, that is, all monomeric sulfated QAOs and dimeric
7S, were found to be inactive. Molecules containing two sulfate
groups (8S−17S, 28S, and 29S) were most active (52−320
μM), while those with three or more sulfate groups (18S−20S
and 25S−27S) were progressively less potent (Tables 1 and 2).
Among the inhibitors that contain two sulfate groups, those
with −OSO3− at the meta position on both rings are less potent
than those with meta/para sulfate substitution, which in turn
are less potent than para/para sulfate substitution. Comparative
examples of the three categories include 9S, 14S, and 13S,
respectively, with affinities of 153, 91, and 82 μM (Table 1).
That a sulfate group at the para position is favored is also borne
out by comparison of 25S, 26S, and 27S (Table 2). The
observation that para substitution is more favored than either
meta or ortho substitution is not unusual considering steric
influence. Although the literature on sulfated, yet hydrophobic,
small molecules is not extensive, our earlier work on sulfated
benzofurans also suggests that the sterically most accessible 5-
sulfate is the most preferred.26,27
Within the para/para disubstituted series, increasing the
length of the linker from four (13S) to five (15S) to six
methylenes (16S) increases the potency gradually (∼80 → ∼50
μM). The trend is also shown by the meta/meta disubstituted
inhibitors 11S (three atom linker) and 8S (four atoms), which
display IC50 values of 320 and 102 μM, respectively (Table 1).
H
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Table 1. Inhibition of Human Factor XIa by Sulfated QAOs 7S−20S
inhibitor
R1
R2
−H
−H
−H
−OSO3
−H
R3
−H
−H
−OH
−H
−H
−H
−H
−H
−H
−H
−H
−OSO3
−OSO3
−OSO3
R1′
R2′
R3′
−H
−H
−OSO3
−H
−H
−H
−H
−H
−H
−H
−H
−H
−OSO3
−OSO3
n
IC50 (μM)
ΔY %
b
−
−
7S
−OAc
−OSO3
−OSO3
−OSO3
−OSO3
−OAc
−OCH3
−OSO3
−H
−H
−OSO3
−H
−H
−OCH3
−H
−H
−H
−OSO3
−H
−OSO3
−OSO3
−H
−OSO3
−OSO3
−OSO3
−H
4
4
4
4
3
4
4
4
5
6
4
4
4
4
>1000
NA
c
−
−
8S
102
153
139
320
159
82
2
1
1
2
2
100
96
6
5
−
9S
1
1
4
3
2
−
−
10S
11S
12S
13S
14S
15S
16S
17S
18S
19S
20S
−OCH3
95
−
−
−OSO3
−OSO3
−H
−H
−H
91
−
−
−
−H
88
−
−
−
−
−OSO3
−OSO3
−OSO3
−OSO3
−H
−H
−H
−H
1
2
1
1
91
−
91
103
89
−
−
−
59
2
1
5
2
2
−H
52
97
−
−
−
−
−OSO3
−OSO3
−OSO3
−OSO3
182
213
273
1
2
3
88
−
−
−
−
−OSO3
−OAc
−OSO3
93
−
−
−H
−H
91
−
>1000
NA
a
The IC50, HS, and ΔY values were obtained following nonlinear regression analysis of direct inhibition of factor Xa. Inhibition was monitored by
b
c
spectrophotometric measurement of residual proteases activity (see the Experimental Procedures). Not applicable. Errors represent 1 SE.
a
Table 2. Inhibition of Human Factor XIa by Sulfated QAOs 25S−29S
inhibitor
R1
R2
−H
−OSO3
−H
−
R3
−H
−H
−OSO3
−
IC50 (μM)
ΔY %
b
−
25S
26S
27S
28S
29S
−OSO3
>1000
NA
50 13
c
−
−OCH3
−OAc
−
216 27
−
407 10
90
98
94
6
3
1
305
3
−
−
−
90 0.2
a
The IC50, HS, and ΔY values were obtained following nonlinear regression analysis of direct inhibition of factor Xa. Inhibition was monitored by
b
c
spectrophotometric measurement of residual proteases activity (see the Experimental Procedures). Not applicable. Errors represent 1 SE.
glucoside inhibition of FXIa through its HBS also bodes well
for design of second generation SAMs from the leads identified
in this work.41 A critical component to such an improvement
would be the cocrystal structure of a SAM with human FXIa.42
Inhibition Potency of Sulfated QAOs in Human
Plasma. To assess whether the chromogenic substrate-based
inhibition of human factor XIa by sulfated QAOs translates into
activity against macromolecular substrates, we studied antico-
agulant activity in human plasma. Two assays, the prothrombin
and activated partial thromboplastin time assays (PT and
APTT, respectively), are typically utilized to identify an
inhibitor’s ability to retard the extrinsic and intrinsic
I
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. Human plasma anticoagulation by sulfated QAOs 13S (A), 15S (B), and 16S (C). The time to clot was measured in either the APTT assay
(solid symbols) or the PT assay (open symbols) in the presence of varying concentrations of the three sulfated QAOs. Solid lines are trend lines,
which were used to calculate the concentration of the anticoagulant that is expected to double the clotting time. The mean of two experiments are
reported.
coagulation signal. A dose-dependent prolongation of APTT is
observed in the presence of sulfated QAOs 13S, 15S, and 16S
(Figure 3). A 2-fold increase in APTT was observed in the
range 0.95−1.06 mM for the three molecules (Table 3), which
Table 3. Effect of Sulfated QAOs on Human Plasma Clotting
a
Times
inhibitor
2 × APTT (μM)
2 × PT (μM)
13S
15S
16S
979
1119
1684
>2000
1062
950
a
Prolongation of clotting time as a function of concentration of
Figure 4. Michaelis−Menten kinetics of S2366 hydrolysis by human
factor XIa in the presence of sulfated QAO 14S. The initial rate of
hydrolysis at various substrate concentrations was measured
spectrophotometrically in pH 7.4 buffer at 37 °C. Solid lines represent
nonlinear regressional fits to the data by the standard Michaelis−
sulfated quinazolinones in either the activated partial thromboplastin
time assay (APTT) or the prothrombin time assay (PT). Clotting
assays were performed in duplicate (SE ≤ 10%) as described in the
Experimental Procedures.
Menten equation to yield KM and Vmax
.
is about 12−18-fold less active in comparison to the potency in
buffer. This is typical of many anticoagulants and arises
primarily from binding to human serum albumin.43,44 In the PT
assay, only 13S appears to exhibit reasonable dose-dependent
prolongation of clotting time with a 2-fold increase calculated
to occur at about 1.12 mM. Inhibitors 15S and 16S are much
less active in the PT assay. Although the coagulation cascade is
a highly intertwined network of proteolytic reactions, it is
generally accepted that the primary effect of FXIa is in the
intrinsic pathway. This implies that inhibition of FXIa is likely
to more dramatically affect the APTT in comparison to the PT.
Thus, the results observed for 15S and 16S further support the
selective inhibition of FXIa. The effect of 13S on both the
APTT and PT is interesting and intriguing and requires further
experimental study.
Mechanism of Inhibition of Sulfated QAOs. Inherent in
the design strategy is the expectation that SAMs should
function as allosteric modifiers of proteolytic activity. To assess
this, the kinetics of chromogenic substrate S2366 hydrolysis by
factor XIa in the presence of 14S was studied. The plot of initial
rate as a function of Spectrozyme FXIa concentration displayed
a characteristic hyperbolic profile (Figure 4), which was fitted
using the standard Michaelis−Menten equation to derive the
KM and Vmax of factor XIa activity. The KM for Spectrozyme
FXIa was found to be 0.31 0.03 mM, which did not change
much as the concentration of 14S increased to 135 μM (0.30
0.07 mM). In constrast, the Vmax decreased from 43.3 1.5 to
5.1 0.5 mAU/min as the concentration of 14S increased from
0 to 135 μM. Thus, while the affinity of small chromogenic
substrate remains unaffected by 14S binding, the proteolytic
activity decreases. This is characteristic of a noncompetitive
mechanism of factor XIa inhibition, and most H/HS mimetics
reported in the literature to date, such as sulfated low molecular
weight lignins45 and sulfated benzofurans,26,40 have exhibited
such a mechanism.
Sulfated QAOs Engage Residues of the HBS Present
on the Catalytic Domain of Factor XIa. Heparin binds to
FXIa in two sites, in the A3 domain (Lys252, Lys253, and
Lys255) and in the catalytic domain (Lys529, Arg530, Arg532,
Lys535, and Lys539). To identify whether sulfated QAOs
engage the A3 domain or the catalytic domain, we studied
inhibition of human FXIa containing only the catalytic domain.
This domain and all site-directed mutants were expressed and
purified, as described earlier.32,33 The catalytic domain alone
(FXIa-CD) was inhibited by a sulfated QAO, 16S, as potently
as the wild-type enzyme indicating that the A3 domain is not
necessary for activity (not shown). Replacement of Arg530
alone by Ala in FXIa-CD resulted in an increase of 2-fold in the
IC50 (Figure 5). Further replacement of Lys529, Arg530, and
Arg532 to Ala each reduced the potency by approximately 5-
fold (Figure 5). At the same time, Lys535Ala and Lys539Ala
exhibited no change in the IC50 of 16S from that of the wild-
type FXIa-CD (not shown). This implies that one or more
J
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 5. Dose−response profiles for 16S inhibition of recombinant
wild-type FXIa catalytic domain (FXIa-CD, solid circles), Arg530Ala
FXIa-CD single mutant (open circles), and Lys529Ala/Arg530Ala/
Arg532Ala FXIa-CD triple mutant (shaded circles) in 50 mM Tris−
HCl buffer pH 7.4 containing 150 mM NaCl, 0.1% PEG8000, and
0.02% Tween80 at 37 °C. Solid lines represent sigmoidal dose−
response fits of eq 1.
Figure 6. Effect of Polybrene on the proteolytic activity of full-length
wild-type factor XIa (shaded bars), FXIa-CD (open bars), and FXIa-
CD containing Lys529Ala/Arg530Ala/Arg532Ala triple mutations
(solid bars). The proteolytic activities of each enzyme in the presence
of 30 μM Polybrene and 3 μM 16S were measured using S2366
hydrolysis. See the text for details.
residues of the HBS present on the catalytic domain of FXIa
is(are) involved in binding to 16S. The loss of affinity for 16S
with the triple mutation is moderate. In comparison,
replacement of a single arginine on thrombin (Arg173)
introduced a defect of approximately 22-fold for a sulfated
benzofuran,27 while for sulfated low molecular weight lignins,
the loss in potency was in the range 2−8-fold.28 This implies
that the loss in potency for sulfated QAOs is more similar to
that of sulfated low molecular weight lignins than that for
sulfated benzofurans. Sulfated lignins contain several sulfate
groups, which bind to more than one arginine/lysine residues
on thrombin.28 In a similar manner, the two sulfate groups of
QAOs likely bind to two arginine/lysine loci on factor XIa. The
site-directed mutagenesis data shows that one of these
electropositive loci is the HBS. The other locus remains to
be identified. In combination with the noncompetitive
mechanism of inhibition, the results suggest that sulfated
QAOs recognize an exosite on FXIa away from the active site
or allosterically inhibit the enzyme.
Factor XIa Inhibition by Sulfated QAOs Is Not
Neutralized by a Competing Electropositive Polymer.
A critical element of the proposed design principle is that hP
domain(s) on the target enzyme are involved in selective
inhibition. To assess this aspect, we studied the effect of
Polybrene on the proteolytic activity of factor XIa. Polybrene
has been regularly used to neutralize heparin since the early
1990s.46,47 Basically, the multiple positive charges of Polybrene
neutralize the numerous sulfate and carboxylate groups of
heparin resulting in an antidote effect. If sulfated QAOs bind to
factor XIa only through electrostatic sulfate−arginine/lysine
forces and not through significant hydrophobic forces, then
Polybrene should neutralize their inhibitory effects. Figure 6
shows the 16S inhibition of the catalytic activity of wild-type
FXIa, FXIa-CD, and FXIa-CD containing triple mutation
(Lys529Ala/Arg530Ala/Arg532Ala) in the presence of Poly-
brene. The proteolytic activity of both the full length, wild-type
FXIa, FXIa-CD, and FXIa-CD containing triple mutations is
reduced to 4.2 1.0, 6.3 0.7, and 4.9 1.2%, respectively, in
the presence of 16S. This is as expected due to saturation of
FXIa by 16S. Addition of 30 μM Polybrene, which rapidly
neutralizes heparin’s anticoagulant effect, changes the inhibition
level to 9.2 0.2% (wild-type FXIa), 13.1 0.4% (FXIa-CD),
have been nearly fully reversed by the high concentration of
Polybrene. Thus, 16S appears to utilize primarily hydrophobic
forces in binding to FXIa. Also evident from these results is the
role of the HBS residues. The loss in inhibition by the addition
of Polybrene is nearly 2-fold for full length, wild-type FXIa and
FXIa-CD but is 1.45-fold for the FXIa-CD with triple
mutations. This implies that, when a small group of positively
charged residues are eliminated, the contribution of electro-
static forces decreases, as expected. Both of these aspects are
key elements of the proposed design principle and are expected
to contribute to the selectivity observed in targeting FXIa
(above). Yet, these results should be tempered with the
recognition that Polybrene is not selective for sulfated QAO.
Unfortunately, no agent is currently available that can
selectively compete with such sulfated and yet hydrophobic
molecules. The best avenue for establishing the proposed dual-
element design strategy is likely to be X-ray structure
determination of a FXIa−SAM cocomplex.
Sulfated QAOs Bind to Factor XIa in a Classic, Direct
Allosteric Manner. The studies performed so far indicate that
sulfated QAO inhibit human FXIa by utilizing an allosteric site.
However, the HBS on the catalytic domain of FXIa, although
allosteric, is within 22 Å of the active site. The molecular size of
sulfated QAOs is also fairly large raising a minor concern
whether the inhibition is truly through an allosteric, cooperative
process. To test this possibility, we studied the interaction of
sulfated QAOs with FXIa irreversibly blocked at the active site
using a dansylated EGR peptide, that is, FXIa-DEGR. Figure 7A
shows the fluorescence emission spectra of FXIa alone and in
complex with a SAM (13S−16S). The fluorescence of the
active site dansyl group increases in the presence of each
sulfated QAO. The maximal increase in fluorescence is
dependent on the type of sulfated QAO. In addition, the
emission maximum of FXIa-DEGR shifts from approximately
550 nm in the absence of sulfated QAOs to ∼505 nm in their
presence suggesting a major blueshift. Both the increase in
fluorescence intensity and the blueshift in λem suggest strong
perturbation of the electrostatic environment around the active
site of FXIa following interaction with sulfated QAOs.
To further investigate the interaction, the fluorescence of
FXIa-DEGR was monitored as a function of sulfated QAO
concentration. Figure 7B shows the profiles of the titrations for
the most potent sulated QAOs 13S−16S. The profiles reveal a
characteristic sigmoidal dependence on the concentration of
sulfated QAO. This is a striking observation and strongly
and 7.1
0.1% (FXIa-CD with triple mutations). The
reduction in inhibition shows that the antidote does affect
FXIa−16S interaction, but the influence is mostly marginal. A
primarily electrostatic interaction between FXIa and 16S should
K
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
SIGNIFICANCE
■
Human FXIa is a key coagulation protease and is being
increasingly regarded as the most optimal enzyme to target for
developing new antithrombotics that are devoid of bleeding
consequences.32,33,48,49 Studies show that reducing the levels of
active FXI significantly reduces clotting tendency. Most
importantly, this reduction in activity leaves the normal
hemostasis unaffected. Additionally, ischemic stroke, a delete-
rious effect, is significantly reduced in FXI-deficient patients. In
contrast, venous thrombosis arises from higher natural levels of
FXI. Thus, the sulfated QAOs discovered in this work will be
key leads in designing more potent inhibitors of an enzyme
thought to be safer to target.
To date, only a handful of small molecule inhibitors of FXIa
have been reported. These molecules are orthosteric, active site
directed inhibitors.50−56 Among these, the reversible inhibitors
contain a strongly basic group that recognizes the active site
Asp189 of FXIa. In contrast, the irreversible inhibitors
covalently target the active site Ser195 through the use of
boronic acid or ketothiazole groups. Sulfated QAOs are
significant because these function through an allosteric site.
Recently, we reported the first allosteric inhibitor of human
FXIa, sulfated pentagalloyl glucoside, that directly binds to one
of the enzyme’s HBSs.41 The two allosteric inhibitors, sulfated
QAO and sulfated pentagalloyl glucoside, are structurally
completely different and appear to utilize different allosteric
sites. This implies that multiple opportunities are available for
exploiting allosteric inhibition of FXIa. In comparison to
orthosteric inhibition, allosteric modulation of an enzyme’s
activity offers the promise of controlling the efficacy of
inhibition (ΔY). For example, inhibitor 26S displays only
50% efficacy, while 13S−16S display efficacy greater than 85%
(Tables 1 and 2). This implies that it may be possible to design
new sulfated QAOs with variable levels of efficacies that likely
offer a finer control over the anticoagulation state of plasma.
The discovery of sulfated QAOs, in combination with work
on sulfated benzofurans,26,27 supports the dual-element design
concept present in Figure 1. This is a simple approach,
amenable to computational modeling, and likely to be very
useful. A major promise of this approach is high target
selectivity. Because GAGs primarily function through exosites,
the proposed SAM design strategy is expected to be applicable
to many GBPs such as serpins, growth factors, chemokines, and
viral envelope glycoproteins.
Figure 7. (A) Changes in the fluorescence emission spectrum of
dansylated factor XIa (FXIa-DEGR) induced by the binding of
sulfated QAOs. Spectra were recorded in 50 mM Tris−HCl buffer of
pH 7.4 containing 150 mM NaCl and 0.1% PEG8000 at 37 °C. (B)
Fractional change in fluorescence of FXIa-DEGR at 547 nm (λex = 345
nm) as a function of the concentration of sulfated QAOs (13S−16S).
Solid line represents nonlinear regressional fit to the data using the
standard Hill eq 3 to obtain the ΔFmax, Hill coefficient “n”, and KD of
binding.
suggestive of a cooperative binding process. The profile can be
fitted well by the standard, three-parameter Hill equation,
which gives the maximal fluorescence change (ΔFmax), the Hill
coefficient (n), and the apparent dissociation constant (KD) of
binding (Table 4). By using this equation, the four sulfated
Table 4. Thermodynamics of Sulfated QAOs Binding to
a
Human FXIa
inhibitor
ΔFmax (%)
n
KD (μM)
b
13S
14S
15S
16S
66
57
72
90
2
2
2
3
7.9 1.9
9.0 1.1
6.8 2.7
6.4 2.0
84 12
91
37
38
2
8
5
ASSOCIATED CONTENT
■
a
S
* Supporting Information
Titrations were performed by adding aliquots of a solution of sulfated
QAOs (13S−16S) to 250 nM FXIa−DEGR in 50 mM Tris−HCl
buffer of pH 7.4 containing 150 mM NaCl and 0.1% PEG8000 at 37
°C and monitoring the change in the fluorescence of FXIa−DEGR at
547 nm (λEX = 345 nm). The ΔFmax, the Hill coefficient “n”, and KD
General procedures for the synthesis of polyphenolic
quinazolin-4(3H)-ones 3a−3g, 7−20, and 25−29 and other
intermediate compounds. 1H and 13C NMR spectra and HRMS
spectra for select compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
b
were obtained by nonlinear regressional fit of the data by eq 3. Error
=
1 SE.
AUTHOR INFORMATION
QAOs were found to bind with an affinity of 37−91 μM, which
compare favorably with the IC50 measured above. The Hill
coefficients were calculated to be in the range 6.4−9.0, which
support a strongly cooperative, allosteric interaction. Overall,
the thermodynamic fluorescence study demonstrates that
sulfated QAOs bind to FXIa through a classic, allosteric
interaction process.
■
Corresponding Author
*Phone: (804) 828-7328; fax: (804) 827-3664; e-mail:
urdesai@vcu.edu.
Notes
The authors declare no competing financial interest.
L
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(15) Nimjee, S. M.; Oney, S.; Volovyk, Z.; Bompiani, K. M.; Long, S.
B.; Hoffman, M.; Sullenger, B. A. Synergistic effect of aptamers that
inhibit exosites 1 and 2 on thrombin. RNA 2009, 15, 2105−2111.
(16) Freeman, C.; Liu, L.; Banwell, M. G.; Brown, K. J.; Bezos, A.;
Ferro, V.; Parish, C. R. Use of sulfated linked cyclitols as heparan
sulfate mimetics to probe the heparin/heparan sulfate binding
specificity of proteins. J. Biol. Chem. 2005, 280, 8842−8849.
ACKNOWLEDGMENTS
■
This work was supported by grants HL090586 and HL107152
from the National Institutes of Health to U.R.D.
ABBREVIATIONS USED
■
APTT, activated partial thromboplastin time; CuAAC, copper-
catalyzed azide−alkyne cycloaddition reaction; DS, dermatan
sulfate; FXI, factor XI; FXIa, factor XIa; FXIa−CD, catalytic
domain of FXIa; FXIIIa, factor XIIIa; GAG, glycosaminoglycan;
GBP, GAG-binding protein; HBS, heparin-binding site; hP,
hydrophobic; HS, heparan sulfate; QAO, quinazolin-4(3H)-
ones; PEG, polyethylene glycol; PT, prothrombin time; SAM,
sulfated allosteric modulator
(17) Turk, H.; Haag, R.; Alban, S. Dendritic polyglycerol sulfates as
̈
new heparin analogues and potent inhibitors of the complement
system. Bioconjugate Chem. 2004, 15, 162−167.
(18) Gunnarsson, G. T.; Desai, U. R. Interaction of designed sulfated
flavanoids with antithrombin: Lessons on the design of organic
activators. J. Med. Chem. 2002, 45, 4460−4470.
(19) Gunnarsson, G. T.; Desai, U. R. Designing small, nonsugar
activators of antithrombin using hydropathic interaction analyses. J.
Med. Chem. 2002, 45, 1233−1243.
REFERENCES
(20) Correia-da-Silva, M.; Sousa, E.; Duarte, B.; Marques, F.; Cunha-
Ribeiro, L. M.; Pinto, M. M. Dual anticoagulant/antiplatelet
persulfated small molecules. Eur. J. Med. Chem. 2011, 46, 2347−2358.
(21) Correia-da-Silva, M.; Sousa, E.; Duarte, B.; Marques, F.;
Carvalho, F.; Cunha-Ribeiro, L. M.; Pinto, M. M. Flavonoids with an
oligopolysulfated moiety: A new class of anticoagulant agents. J. Med.
Chem. 2011, 54, 95−106.
(22) Gunnarsson, G. T.; Riaz, M.; Adams, J.; Desai, U. R. Synthesis of
per-sulfated flavonoids using 2,2,2-trichloro ethyl protecting group and
their factor Xa inhibition potential. Bioorg. Med. Chem. 2005, 13,
1783−1789.
(23) Correia-da-Silva, M.; Sousa, E.; Duarte, B.; Marques, F.;
Carvalho, F.; Cunha-Ribeiro, L. M.; Pinto, M. M. Polysulfated
xanthones: Multipathway development of a new generation of dual
anticoagulant/antiplatelet agents. J. Med. Chem. 2011, 54, 5373−5384.
(24) Al-Horani, R. A.; Liang, A.; Desai, U. R. Designing
nonsaccharide, allosteric activators of antithrombin for accelerated
inhibition of factor Xa. J. Med. Chem. 2011, 54, 6125−6138.
(25) Raghuraman, A.; Liang, A.; Krishnasamy, C.; Lauck, T.;
Gunnarsson, G. T.; Desai, U. R. On designing non-saccharide,
allosteric activators of antithrombin. Eur. J. Med. Chem. 2009, 44,
2626−2631.
(26) Sidhu, P. S.; Liang, A.; Mehta, A. Y.; Abdel Aziz, M. H.; Zhou,
Q.; Desai, U. R. Rational design of potent, small, synthetic allosteric
inhibitors of thrombin. J. Med. Chem. 2011, 54, 5522−5531.
(27) Abdel Aziz, M. H.; Sidhu, P. S.; Liang, A.; Kim, J. Y.; Mosier, P.
D.; Zhou, Q.; Farrell, D. H.; Desai, U. R. Designing allosteric
regulators of thrombin. Monosulfated benzofuran dimers selectively
interact with arg173 of exosite II to induce inhibition. J. Med. Chem.
2012, 55, 6888−6897.
■
(1) Sarrazin, S.; Lamanna, W. C.; Esko, J. D. Heparan sulfate
proteoglycans. Cold Spring Harbor Perspect. Biol. 2011, 3, a004952.
(2) Maeda, N.; Ishii, M.; Nishimura, K.; Kamimura, K. Functions of
chondroitin sulfate and heparan sulfate in the developing brain.
Neurochem. Res. 2011, 36, 1228−1240.
(3) Olson, S. T.; Richard, B.; Izaguirre, G.; Schedin-Weiss, S.;
Gettins, P. G. Molecular mechanisms of antithrombin-heparin
regulation of blood clotting proteinases. A paradigm for understanding
proteinase regulation by serpin family protein proteinase inhibitors.
Biochimie 2010, 92, 1587−1596.
(4) Gandhi, N. S.; Mancera, R. L. The structure of glycosaminogly-
cans and their interactions with proteins. Chem. Biol. Drug Des. 2008,
72, 455−482.
(5) Desai, U. R. New antithrombin-based anticoagulants. Med. Res.
Rev. 2004, 24, 151−181.
(6) Kreuger, J.; Spillmann, D.; Lindahl, U. Interactions between
heparan sulfate and proteins: the concept of specificity. J. Cell Biol.
2006, 174, 323−327.
(7) Lindahl, U. Heparan sulfate-protein interactions − A concept for
drug design. Thromb. Haemostasis 2007, 98, 109−115.
(8) Mosier, P. D.; Krishnasamy, C.; Kellogg, G. E.; Desai, U. R. On
the specificity of heparin/heparan sulfate binding to proteins. Anion-
binding sites on antithrombin and thrombin are fundamentally
different. PLoS One 2012, 7, e48632.
(9) Coulomb forces have a r−1 (r = distance between two ions)
relationship, whereas van der Waals forces have a r−3 to r−6
dependence.
(10) Petitou, M.; Nancy-Portebois, V.; Dubreucq, G.; Motte, V.;
Meuleman, D.; de Kort, M.; van Boeckel, C. A.; Vogel, G. M.; Wisse, J.
A. From heparin to EP217609: The long way to a new
pentasaccharide-based neutralisable anticoagulant with an unprece-
dented pharmacological profile. Thromb. Haemostasis 2009, 102, 804−
810.
(11) Saxena, K.; Schieborr, U.; Anderka, O.; Duchardt-Ferner, E.;
Elshorst, B.; Gande, S. L.; Janzon, J.; Kudlinzki, D.; Sreeramulu, S.;
Dreyer, M. K.; Wendt, K. U.; Herbert, C.; Duchaussoy, P.; Bianciotto,
M.; Driguez, P. A.; Lassalle, G.; Savi, P.; Mohammadi, M.; Bono, F.;
Schwalbe, H. Influence of heparin mimetics on assembly of the
FGF·FGFR4 signaling complex. J. Biol. Chem. 2010, 285, 26628−
26640.
(28) Abdel Aziz, M. H.; Mosier, P. D.; Desai, U. R. Identification of
the site of binding of sulfated, low molecular weight lignins on
thrombin. Biochem. Biophys. Res. Commun. 2011, 413, 348−352.
(29) Al-Horani, R. A.; Desai, U. R. Chemical sulfation of small
molecules-advances and challenges. Tetrahedron 2010, 66, 2907−2918.
(30) Raghuraman, A.; Riaz, M.; Hindle, M.; Desai, U. R. Rapid and
efficient microwave-assisted synthesis of highly sulfated organic
scaffolds. Tetrahedron Lett. 2007, 48, 6754−6758.
(31) Henry, B. L.; Thakkar, J. N.; Liang, A.; Desai, U. R. Sulfated, low
molecular weight lignins inhibit a select group of heparin-binding
serine proteases. Biochem. Biophys. Res. Commun. 2012, 417, 382−386.
(32) Smith, S. B.; Gailani, D. Update on the physiology and
pathology of factor IX activation by factor XIa. Expert Rev. Hematol.
2008, 1, 87−98.
́
(12) Sarrazin, S.; Bonnaffe, D.; Lubineau, A.; Lortat-Jacob, H.
Heparan sulfate mimicry: A synthetic glycoconjugate that recognizes
the heparin binding domain of interferon-gamma inhibits the cytokine
activity. J. Biol. Chem. 2005, 280, 37558−37564.
(33) Emsley, J.; McEwan, P. A.; Gailani, D. Structure and function of
factor XI. Blood 2010, 115, 2569−2577.
(13) Lubineau, A.; Lortat-Jacob, H.; Gavard, O.; Sarrazin, S.;
́
Bonnaffe, D. Synthesis of tailor-made glycoconjugate mimetics of
(34) Xia, Y.; Yang, Z. Y.; Hour, M. J.; Kuo, S. C.; Xia, P.; Bastow, K.
F.; Nakanishi, Y.; Nampoothiri, P.; Hackl, T.; Hamel, E.; Lee, K. H.
Antitumor agents. Part 204: Synthesis and biological evaluation of 2-
aryl quinazolinones. Bioorg. Med. Chem. Lett. 2001, 11, 1193−1196.
(35) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. A
stepwise Huisgen cycloaddition process: Copper(I)-catalyzed regiose-
heparan sulfate that bind IFN-γ in the nanomolar range. Chem.Eur.
J. 2004, 10, 4265−4282.
(14) Ferro, V.; Dredge, K.; Liu, L.; Hammond, E.; Bytheway, I.; Li,
C.; Johnstone, K.; Karoli, T.; Davis, K.; Copeman, E.; Gautam, A. PI-
88 and novel heparan sulfate mimetics inhibit angiogenesis. Semin.
Thromb. Hemostasis 2007, 33, 557−568.
M
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
lective “ligation” of azides and terminal alkynes. Angew. Chem., Int. Ed.
2002, 41, 2596−2599.
(36) Zhang, Li.; Chen, X.; Xue, P.; Sun, H. H. Y.; Williams, I. D.;
Sharpless, K. B.; Fokin, V. V.; Jia, G. Ruthenium-catalyzed cyclo-
addition of alkynes and organic azides. J. Am. Chem. Soc. 2005, 127,
15998−15999.
(37) Copeland, R.; Balasubramaniam, A.; Tiwari, V.; Zhang, F.;
Bridges, A.; Linhardt, R. J.; Shukla, D.; Liu, J. Using a 3-O-sulfated
heparin octasaccharide to inhibit the entry of herpes simplex virus type
1. Biochemistry 2008, 47, 5774−5783.
(38) Tollefsen, D. M. The interaction of glycosaminoglycans with
heparin cofactor II: Structure and activity of a high-affinity dermatan
sulfate hexasaccharide. Adv. Exp. Med. Biol. 1992, 313, 167−176.
(39) Raghuraman, A.; Mosier, P. D.; Desai, U. R. Understanding
dermatan sulfate-heparin cofactor II interaction through virtual library
screening. ACS Med. Chem. Lett. 2010, 1, 281−285.
(54) Schumacher, W. A.; Seiler, S. E.; Steinbacher, T. E.; Stewart, A.
B.; Bostwick, J. S.; Hartl, K. S.; Liu, E. C.; Ogletree, M. L.
Antithrombotic and hemostatic effects of a small molecule factor Xia
inhibitor in rats. Eur. J. Pharmacol. 2007, 570, 167−174.
(55) Buchanan, M. S.; Carroll, A. R.; Wessling, D.; Jobling, M.;
Avery, V. M.; Davis, R. A.; Feng, Y.; Xue, Y.; Oster, L.; Fex, T.;
Deinum, J.; Hooper, J. N.; Quinn, R. J.; Clavatadine, A. A natural
product with selective recognition and irreversible inhibition of factor
XIa. J. Med. Chem. 2008, 51, 3583−3587.
(56) Hanessian, S.; Larsson, A.; Fex, T.; Knecht, W.; Blomberg, N.
Design and synthesis of macrocyclic indoles targeting blood
coagulation cascade Factor XIa. Bioorg. Med. Chem. Lett. 2010, 20,
6925−6928.
(40) Verghese, J.; Liang, A.; Sidhu, P. S.; Hindle, M.; Zhou, Q.; Desai,
U. R. First steps in the direction of synthetic, allosteric, direct
inhibitors of thrombin and factor Xa. Bioorg. Med. Chem. Lett. 2009,
19, 4126−4129.
(41) Al-Horani, R. A.; Ponnusamy, P.; Mehta, A. Y.; Gailani, D.;
Desai, U. R. Sulfated pentagalloylglucoside is a potent, allosteric, and
selective inhibitor of factor XIa. J. Med. Chem. 2013, 56, 867−878.
(42) Jin, L.; Pandey, P.; Babine, R. E.; Gorga, J. C.; Seidl, K. J.;
Gelfand, E.; Weaver, D. T.; Abdel-Meguid, S. S.; Strickler, J. E. Crystal
structures of the FXIa catalytic domain in complex with ecotin mutants
reveal substrate-like interactions. J. Biol. Chem. 2005, 280, 4704−4712.
(43) Yacobi, A.; Udall, J. A.; Levy, G. Serum-protein binding as a
determinant of warfarin body clearance and anticoagulant effect. Clin.
Pharmacol. Ther. 1976, 19, 552−558.
(44) Kragh-Hansen, U. Structure and ligand binding properties of
human serum albumin. Dan. Med. Bull. 1990, 37, 57−84.
(45) Henry, B. L.; Monien, B. H.; Bock, P. E.; Desai, U. R. A novel
allosteric pathway of thrombin inhibition. Exosite II mediated potent
inhibition of thrombin by chemo-enzymatic, sulfated dehydropolymers
of 4-hydroxycinnamic acids. J. Biol. Chem. 2007, 282, 31891−31899.
(46) Rotteveel, R. C.; Roozendaal, K. J.; Weijers, R. N.; Eijsman, L.
Influence of heparin, protamine and Polybrene on the time integral of
thrombin generation (endogenous thrombin potential). Haemostasis
1996, 26, 1−10.
(47) Harsfalvi, J.; Pfliegler, G.; Udvardy, M.; Boda, Z.; Tornai, I.; Rak,
K. The use of Polybrene for heparin neutralization in protein C activity
assay. Blood Coagulation Fibrinolysis 1990, 1, 357−361.
(48) Goto, S. Factor XIa as a possible new target of antithrombotic
therapy. J. Thromb. Haemostasis 2006, 4, 1494−1495.
(49) Schumacher, W. A.; Luettgen, J. M.; Quan, M. L.; Seiffert, D. A.
Inhibition of Factor Xia as a new approach to anticoagulation.
Arterioscler., Thromb., Vasc. Biol. 2010, 30, 388−392.
(50) Wong, P. C.; Crain, E. J.; Watson, C. A.; Schumacher, W. A. A
small-molecule factor XIa inhibitor produces antithrombotic efficacy
with minimal bleeding time prolongation in rabbits. J. Thromb.
Thrombolysis 2011, 32, 129−137.
(51) Lin, J.; Deng, H.; Jin, L.; Pandey, P.; Quinn, J.; Cantin, S.;
Rynkiewicz, M. J.; Gorga, J. C.; Bibbins, F.; Celatka, C. A.; Nagafuji, P.;
Bannister, T. D.; Meyers, H. V.; Babine, R. E.; Hayward, N. J.; Weaver,
D.; Benjamin, H.; Stassen, F.; Abdel-Meguid, S. S.; Strickler, J. E.
Design, synthesis, and biological evaluation of peptidomimetic
inhibitors of factor XIa as novel anticoagulants. J. Med. Chem. 2006,
49, 7781−7791.
(52) Deng, H.; Bannister, T. D.; Jin, L.; Babine, R. E.; Quinn, J.;
Nagafuji, P.; Celatka, C. A.; Lin, J.; Lazarova, T. I.; Rynkiewicz, M. J.;
Bibbins, F.; Pandey, P.; Gorga, J. C.; Meyers, H. V.; Abdel-Meguid, S.
S.; Strickler, J. E. Synthesis, SAR exploration, and X-ray crystal
structures of factor XIa inhibitors containing an α-ketothiazole
arginine. Bioorg. Med. Chem. Lett. 2006, 16, 3049−3054.
(53) Lazarova, T. I.; Jin, L.; Rynkiewicz, M.; Gorga, J. C.; Bibbins, F.;
Meyers, H. V.; Babine, R.; Strickler, J. Synthesis and in vitro biological
evaluation of aryl boronic acids as potential inhibitors of factor XIa.
Bioorg. Med. Chem. Lett. 2006, 16, 5022−5027.
N
dx.doi.org/10.1021/jm301757v | J. Med. Chem. XXXX, XXX, XXX−XXX