Protein-Induced Change in Ligand Protonation during Trypsin and Thrombin Binding: Hint on Differences in Selectivity Determinants of Both Proteins?

Article abstract of DOI:10.1021/acs.jmedchem.9b02061

Trypsin and thrombin, structurally similar serine proteases, recognize different substrates; thrombin cleaves after Arg, whereas trypsin cleaves after Lys/Arg. Both recognize basic substrate headgroups via Asp189 at the bottom of the S1 pocket. By crystal

Full text of DOI:10.1021/acs.jmedchem.9b02061

Subscriber access provided by UNIV OF CAMBRIDGE
Article
Protein-induced Change in Ligand Protonation during Trypsin and Thrombin
Binding: Hint on Differences in Selectivity Determinants of both Proteins?
Khang Ngo, Chelsey Collins-Kautz, Stefan Gerstenecker, Björn Wagner, Andreas Heine, and Gerhard Klebe
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b02061 • Publication Date (Web): 03 Feb 2020
Downloaded from pubs.acs.org on February 10, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Protein-induced Change in Ligand Protonation during Trypsin and
Thrombin Binding: Hint on Differences in Selectivity Determinants
of both Proteins?
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
†
†
‡
Khang Ngo , Chelsey Collins-Kautz , Stefan Gerstenecker , Björn Wagner , Andreas
†
*,†
Heine , Gerhard Klebe
†Institute of Pharmaceutical Chemistry, Philipps-University Marburg, Marbacher Weg 6,
35032 Marburg, Germany
‡Pharma Research Non-Clinical Safety, F. Hoffmann-La Roche AG, 4070 Basel
(Switzerland)
ABSTRACT
Trypsin and thrombin, structurally similar serine proteases, recognize different
substrates; thrombin cleaves after Arg whereas trypsin after Lys/Arg. Both recognize
basic substrate headgroups via Asp189 at the bottom of the S1-pocket. By
crystallography and ITC, we studied a series of D-Phe/-D-DiPhe-Pro-(amino)pyridines.
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Identical ligand pairs show the same binding poses. Surprisingly, one ligand binds to
trypsin in protonated, to thrombin in unprotonated state at P1 along with differences in the
residual solvation pattern. While trypsin binding is mediated by an ordered water
molecule, in thrombin water is scattered over three hydration sites. Although having highly
similar S1-pockets, our results suggest different electrostatic properties of Asp189
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
possibly contributing to the selectivity determinant. Thrombin binds a specific Na -ion next
to Asp189 which is absent in trypsin. The electrostatic properties across the S1-pocket
are further attenuated by charged Glu192 at the rim of S1 in thrombin which is replaced
by uncharged Gln192 in trypsin.
Keywords: Trypsin & Thrombin, Selectivity, Protonation States, Isothermal Titration
Calorimetry, X-ray Crystallography, Protein-Ligand Interactions
ACS Paragon Plus Environment
2

Page 3 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
The success of a drug molecule in therapy strongly depends on its selectivity profile
toward a given protein target. Experience shows that in some cases high selectivity is the
prerequisite for an effective therapy without unwanted side effects, whereas in other
cases a mixed profile toward several closely related isoforms of a target family is
desired.1–3 Nevertheless, as long as highly selective tool compounds are not available,
this intriguing question about the best-suited profile can hardly be answered. Therefore,
the issue of sufficient drug selectivity is increasingly moving into the focus of
contemporary drug research.4–6
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
First of all, appropriate steric fit, optimal shape complementarity and high-affinity binding
are usually defined as prime strategies in the development of drug molecules with
sufficient selectivity. Such structural and energetic aspects of mutual protein-ligand
recognition have to be accomplished to guarantee potent binding. They are established
by non-covalent interactions, e.g., by forming hydrogen bonds, ionic interactions,
7
hydrophobic contacts or van-der-Waals interactions. The resulting affinity of the formed
complex is then characterized by the dissociation constant K which is related to thermodynamic
d
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
parameters such as Gibb’s free energy G°, binding enthalpy H° and entropy -TS°.8
Furthermore, before a complex assembles, both binding partners are solvated with a solvation shell
of water molecules. Some of these are in loose contact with the protein or ligand, others establish
strong hydrogen bonds in one or the other states. In all cases, the inventory of available degrees of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
9
freedom of the binding partners involving water molecules will change. Furthermore, water
molecules can actively participate in binding and mediate contacts between protein and ligand.
Changes in the solvation pattern during the entire association/dissociation process will also play
an essential role in protein-ligand recognition and have an impact on affinity and particularly on
binding kinetics.8,10,11 Thus, the sole consideration of the formed end-state complexes is usually
not sufficient to explain affinity.
However, are these criteria to improve affinity toward a given isoform of a target family
over another family member sufficient to achieve sufficient selectivity? Definitely, affinity
enhancing features are a necessary criterion but are they also sufficient for target
selectivity discrimination? Likely, more factors have to be regarded to finally achieve the
desired selectivity profile. In order to equip drug molecules by rational concepts with an
appropriate selectivity profile, we must first elucidate which features the proteins have at
disposal to evolve their selectivity determinants.
In the present investigation, we selected the well-studied and highly similar serine
proteases trypsin and thrombin for a case study. The first is an important enzyme in
ACS Paragon Plus Environment
4

Page 5 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
digestion, the second displays the key enzyme in blood clotting.12 Particularly their S1
specificity pockets, known to be most important for selectivity discrimination, are, on first
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
13
sight, highly similar in architecture. In literature, the only obvious residue exchange of
Ser190 by Ala in the pocket is made responsible for discrimination.14–16 On first glance,
as the Ser/Ala exchange sacrifices the option to form an additional hydrogen bond either
directly to a bound substrate or ligand molecule, or mediated via active-site water
molecules. This might modulate the residual solvation structure in the S1 pocket. In the
present study, we investigated the difference in binding of a congeneric series of D-Phe-
Pro-(amino)pyridine inhibitors with slightly deviating basic character of their C-terminal P1
residue toward thrombin and trypsin by X-ray crystallography and by isothermal titration
calorimetry (ITC). Both proteases recognize substrates with basic P1 residues via binding
to Asp189 at the bottom of their S1 pocket. Surprisingly, thrombin is highly selective to
recognize natural P1-Arg substrates whereas trypsin possesses a much broader
substrate specificity and cleaves unselectively after P1-Arg or P1-Lys residues.14,17,18 One
obvious difference between these two basic amino acids is a significant difference in their
pK values. Nevertheless, such values can be strongly modulated depending on the local
a
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
environment in their binding pockets. E.g. in a hydrophobic pocket, either acidic or basic
ligand head groups will less likely adopt a charged state whereas a basic ligand group
becomes easily positively charged by protonation, when bound to a negatively charged
residue in a more polar binding site. The opposite is the case if the ligand bears an acidic
group and the protein provides basic residues such as Arg or Lys in the binding pocket.7
Consequently, the protein environment is designed in a way to induce significant pKa
shifts of titratable groups of the bound ligand. However, is there any difference caused by
the Asp189 residue in trypsin and thrombin that might take a differing effect on the
observed pKa shift of titratable groups of a bound substrate or inhibitor molecule? Since
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ITC can determine overall changes in the protonation inventory of a binding reaction, we studied
our compound series with deviating basic character in parallel against trypsin and thrombin.
Remarkably, a difference is found indicating that one ligand of the series binds to trypsin in
protonated state at its basic anchor group, whereas the same ligand is accommodated in the S1
pocket of thrombin in neutral state. This experimental observation suggests that the two Asp
residues in both proteases show different electrostatic properties and supposedly contribute to
features of the two proteases responsible to achieve selectivity discrimination in the otherwise
sterically nearly identical S1 pockets.
ACS Paragon Plus Environment
6

Page 7 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
RESULTS AND DISCUSSION
For our comparative study, we designed the following congeneric series of substrate
analogue inhibitors 1-4 with structurally different P1 (amino)pyridine residues possessing
a slightly deviating basic character (Table 1). In case of 4, we attached a D-DiPhe
substituent at P3 position to enhance potency which made characterization by ITC easier
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(s. below). The listed binding constants and pK values of the basic head groups were
a
determined by an enzyme kinetic assay and potentiometric pK measurements (s.
a
Experimental Section).
Table 1. Inhibition Towards Thrombin and Trypsin by Compounds 1-4.
pKab
a
a
K / nM
Thrombin)
K / µM
i
i
Inhibitor
basic head
group
(
(Trypsin)
Ph
N
1
2
NH2
H N
28 ± 1
2
3
9 ± 4
O
O
N
H
O
N
6.4
NH2
Ph
N
3
79 ± 46
62 ± 10
6.7
NH2
H N
2
N
H
O
N
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Ph
N
1
40 ± 15
47 ± 7
6.3
5.0
H N
2
3
4
N
O
N
H
O
NH2
Ph
Ph
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
1
213 ± 39
497 ± 36
H2N
O
N
H
O
N
aK
values were determined using an enzyme kinetic fluorescence assay and are given as arithmetic mean
i
b
from three measurements ( SD). pK values were obtained by potentiometric titration.
a
Synthesis. The general synthesis route for D-Phe/D-DiPhe-Pro based inhibitors was
already described previously by us and other groups.19–22 For the present study, we
focused on the same D-Phe/D-DiPhe-Pro scaffold by introducing different aminopyridines
ACS Paragon Plus Environment
8

Page 9 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
for inhibitors 1-3 and unsubstituted pyridine for compound 4 as P1 building blocks. In the
following, we will describe the synthesis of precursor 11 which was used for the amide
coupling reaction with Boc-D-Phe-Pro-OH leading to the target inhibitor 1. Inhibitors 2-4
were synthesized from commercially available precursors 12-14. Boc-D-Phe-Pro-OH was
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
23
afforded according to a protocol of Steinmetzer et al. , the analogously Boc-D-DiPhe-
19
Pro-OH was obtained by adapting a similar procedure as outlined by Baum et al. The
P1 building block 11 is a known fragment that was synthesized according to a patented
procedure published by Hoffmann-La Roche Inc. in 2003. There, the synthesis was
performed using gaseous reagents which were reacted under high pressure (up to
24
20 bar). To avoid such drastic conditions, a synthesis of 11 was developed under milder
conditions also suited on smaller scale (see Scheme 1).
Scheme 1. Synthesis of precursor 11.
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
OH
O
O
OH
N3
NHBoc
Br
a
b
c, d
e
HO
N
OH
Br
N
Br
Br
N
Br
Br
N
Br
Br
N
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
NHBoc
NH2
NH2
f
g
2 HCl
H N
N
H N
N
NH2
2
2
10
11
Reagents and conditions: a) POBr , 130-150 °C, 2 h then MeOH, 0 °C-rt, 15 h, 74%; b)
3
NaBH , MeOH, 65 °C, 3 h, 91%; c) PBr , 1,4-dioxane, 40 °C, 3 h; d) NaN , DMF, rt, 24 h,
4
3
3
6
6% (over 2 steps); e) PMe , THF, 0 °C-rt, 3 h then Boc O, NEt , THF, 0 °C-rt, 3 h, 50%
3
2
3
;
f) 25% (w/v) NH3(aq), Cu O, DMEDA, K CO , ethylene glycol, 60 °C, 16 h, 63%; g) 4 M
2
2
3
HCl in 1,4-dioxane, rt, 3 h, 95%.
Precursor 11 was obtained from citrazinic acid 5 which was converted into 6 via
bromination and esterification.25 Methyl ester 6 was reduced to the corresponding
26
alcohol 7, followed by bromination and azidation to 8. Compound 8 was used for a one-
pot Staudinger reaction and Boc-protection to afford the Boc-amine 9.27,28 In the next
step, the two amino groups could be introduced via an Ullmann reaction to provide
29
intermediate 10 according to a procedure described by Elmkaddem et al. Cleavage of
the protecting group was performed under acidic conditions to yield the P1 building
block 11 as double HCl salt for the subsequent synthesis of inhibitor 1. Due to the minor
ACS Paragon Plus Environment
10

Page 11 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
nucleophilicity of the aromatic amino groups in compound 11, the primary aliphatic amino
function reacted exclusively by forming the wanted amide.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In Scheme 2 the amide coupling is described for each precursor 11-14 (12 and 13 were
commercially available from ChemPUR, 14 from Sigma Aldrich) to the respective Boc-D-
Phe/D-DiPhe-Pro derivative 15-18. In the final step, intermediates 15-18 were
deprotected using TFA to the desired inhibitors 1-4 as double TFA salts.
Scheme 2. Synthesis of inhibitors 1-4.
Ph
R4
O
Ph
R4
O
R1
R3
Boc
N
R1
N
R1
a
b
H N
2
Y
X
N
H
H2N
Y
X
Y
X
N
H
N
O
3
O
3
R
H
R
R2
R2
R2
11 X = N
12 X = N
15 X = N
16 X = N
1 X = N
2 X = N
Y = C
Y = C
Y = C
Y = C
Y = C
Y = C
1
1
1
1
1
1
R = NH
R = NHBoc
R = NH
R = NHBoc
R = NH
R = NHBoc
2
2
2
2
2
2
2
2
2
2
2
2
R = NH
R = H
R = NH
R = H
R = NH
R = H
4
4
4
4
R = H
R = H
R = H
R = H
1
3 X = C
14 X = N
17 X = C
18 X = N
3 X = C
4
X = N
Y = C
Y = N
Y = C
Y = N
Y = C
Y = N
1
1
1
1
1
1
R = H
R = H
R = H
R = H
R = H
R = H
2
2
2
2
2
2
R = H
R = H
R = H
R = H
R = H
R = H
3
3
3
4
R = NHBoc
R = NHBoc
R = NHBoc
R = Ph
4
4
R = H
R = H
Reagents and conditions: a) Boc-D-Phe-Pro-OH (for 15-17), Boc-D-DiPhe-Pro-OH (for
18), HOBt, EDC · HCl, DMF, DIPEA, 0 °C-rt, 18-24 h, 65-78%; b) TFA, rt, 2 h, 78-92%.
3
R exists only for 3, 13 and 17.
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
X-ray Crystallography. The four D-Phe/D-DiPhe-Pro-based thrombin and trypsin
inhibitors were analyzed by X-ray crystallography. Eight crystal structures were obtained
with a resolution of 0.95-1.57 Å.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The diaminopyridine structure of inhibitor 1 (Figure 1A) uses its polar head group to
interact with Asp189 at the bottom of the S1 pocket of either thrombin or trypsin. The
ortho amino groups make contacts to Gly219 and the water molecule W2 whereas the
pyridine-N undergoes a water mediated interaction with Asp189 via W1. In case of
thrombin, additional interactions are experienced by the “left” exocyclic amino group with
the carboxylate group of Glu192. Furthermore, this residue adopts a back-folded
conformation to participate via the interstitial water molecule W3 in the binding of the P3
group and via W4 in the formation of a contact to the P2 group In trypsin, this contact is
.
not found and residue 192 is here replaced by a corresponding Gln which is disordered
in this structure (Figure 1E). Due to the exchange of Ala190 to Ser in the trypsin complex,
the Ser190 sidechain interacts directly with water molecule W2 and the “right” amino
group. This amino group performs additionally an interaction to the backbone carbonyl
group of Ser190 and to the carboxylate group of Asp189. Structurally, the back-folding of
ACS Paragon Plus Environment
12

Page 13 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Glu192 in thrombin to undergo a water-mediated contact to 1 seems to be replaced in
trypsin by the unique Ser190Oγ H-bond to the ligand, which is further supported by a
contact to the more firmly fixed water molecule W2 in the latter protease.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The P1 residue of inhibitor 2 is bound in two conformations (Figure 1B, 76% for I and
24% for II) in the S1 pocket of thrombin. Thus, it formally mimics the binding mode of 1.
In orientation I, the amino group is orientated toward Tyr228 by forming a hydrogen bond
with the water molecule W2, in the other orientation, its amino group points in the direction
of Gly219 while the pyridine-N interacts with Asp189 via water molecule W1 in both
conformations. A related binding mode can be assigned to the trypsin structure
(Figure 1F), however, here ligand 2 adopts only one conformation by addressing W2 and
Ser190Oγ on the “right” side via its amino group, and it is involved in an additional direct
H-bond to the carboxylate of Asp189. A very similar interaction pattern is found for 1 and
2
toward Tyr228 as observed in thrombin forming water-mediated interactions of the
amino group or the pyridine-type N to the carboxylic function of Asp189 and the water
molecule W2 hosted on top of the phenyl moiety of Tyr228. Also in the thrombin-inhibitor
complex with 2, the residue Glu192 is in a back-folded conformation to interact via water
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
molecule W3 to the P3 group and via W4 to the P2 group. In trypsin such contacts are
missing and the corresponding Gln192 does not adopt a folded geometry to allow
contacting the P1 ligand head group. Possibly, the charged Glu192 in thrombin is
responsible for the observed dual orientation of the head group, even so the conformer
hydrogen-bonding this residue is the lower populated one. Analogously to the trypsin
complex of 1, Ser190 is able to form the same interactions to W2 and the amino group of
the ligand, which is furthermore in both proteases in contact to Asp189 and the backbone
carbonyl group of then residue in position 190.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Inhibitor 3 is a regio-isomeric aminopyridine derivative of compound 2, in which the P1
moiety is para-substituted by the amino group while the pyridine-N is in meta position with
respect to the D-Phe-Pro scaffold. The thrombin complex (Figure 1C) indicates a direct
hydrogen bond to Asp189 complemented by further interactions to the backbone carbonyl
oxygens of Ala190 and Gly219, while the pyridine-N interacts with W2. Glu192
establishes the same interaction geometry as found for the complexes with inhibitor 1 and
2
by water-mediated contacts to the P2 and P3 moiety via W3 and W4. For the P1 group,
a similar binding pose can also be observed in the trypsin complex (Figure 1G) in which
ACS Paragon Plus Environment
14

Page 15 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
either direct or water-mediated (W1) interactions are formed to both oxygens of the
carboxylate of Asp189. Furthermore, in the trypsin complex with inhibitor 3, the Ser190
sidechain interacts directly with the pyridine-N. Due to this direct contact to Ser190, W2
is shifted to the rear of the pocket but still interacts with the pyridine-N and with the
hydroxy group of Ser190. The sidechain of Gln192 is disordered and not visible in the
trypsin complex with ligand 3. Surprisingly, there is a second conformation observed for
the aromatic ring of the terminal phenylalanine attached at the P3 position, which,
however, has no influence on the binding mode of the P1 group.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The refined interactions of inhibitor 4 are comparable in both cases and do not suggest
any structural differences in the S1 pocket among the complexes with the two enzymes.
Figure 1D shows the binding mode of inhibitor 4 in thrombin. This ligand is the analog of
a compound studied in a previous contribution by us that possesses a D-diphenylalanine
as P3 group, but with respect to P1 and P2 it describes the same binding mode as the D-
21
monophenylalanine derivative. As reported earlier, the unsubstituted pyridine ring of the
P1 head group forms an indirect interaction with Asp189 mediated by two water
molecules (W1 and W5). They are found at sites, which are not fully occupied (each with
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
0% occupancy) and show a close mutual distance of 2.0 Å along with reasonable B-
2
factors (23-24 Å ). An additional water molecule (W6) is found at a site, which was set to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
50% occupancy with a distance of 2.1 Å to W1 and an acceptable B-factor (23 Å ). A
similar disordered water pattern between the pyridine-type nitrogen and Asp189 has been
described in our previously determined structure of the D-Phe-Pro analog with thrombin.21
Remarkably, we observed for 4 a very similar binding mode as for the aminopyridine
inhibitors 1 and 2, however in all analogous trypsin complexes the pyridine-N interacts
with Asp189 via a single and fully occupied water molecule found in one distinct spatial
position. Residue Ser190 indicates two conformations which are possibly only resolved
in this complex due to the achieved high resolution of 0.95 Å. Conformation I (occupancy
of 79%) orientates to water molecule W2 while the other (II, occupancy of 21%) is in
contact with the sidechain of Tyr228. We can therefore summarize that the major
difference between the thrombin and trypsin complexes with 4 are the following: In
thrombin, the contacts between 4 and Asp189 are formed via a network of hydrogen-
bonds, however refinement suggests that water is here distributed over several hydration
ACS Paragon Plus Environment
16

Page 17 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
sites whereas in trypsin the carboxylate of Asp189 reinforces an ordered contact
mediated by a single water molecule.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Binding modes of inhibitors 6HSX (1, A), 6T3Q (2, B), 6T4A (3, C) and 6TDT (4, D) in
the S1 pocket of thrombin (left) compared to the structures of trypsin (right) by inhibitors 6T0M (1,
E), 6T0P (2, F), 6T5W (3, G) and 6SY3 (4, H). The inhibitors are represented in light brown
whereas the protein residues are displayed in green. Heteroatoms are color-coded by atom type.
Water molecules are colored in red and favorable hydrogen bonds are shown as dashed lines.
ACS Paragon Plus Environment
18

Page 19 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The distances in the complex of thrombin between the water molecules (W1, W5, W6) at the
bottom of the S1 site are given in Å and are colored in red. In the trypsin complexes with 1 (E)
and 3 (G), Gln192 is not visible due to the weakly defined mF -DF electron density of this residue
o
c
after refinement. Likely it is scattered over multiple orientations.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ITC Data Collection. For all inhibitors of the series, we can assume that the polar head
groups become at least partly protonated at the pyridine-N upon binding to the S1 pocket.
This is in agreement with our previous studies, which suggested significant pK shifts of
a
30
the basic anchor groups once bound next to the charged protein residues. Accordingly,
we studied the entire series by ITC titrations in different buffers to record putatively
overlaid changes in protonation states. The values obtained for the different buffers are
listed in Table 2.
Table 2. Thermodynamic data of ∆H° from direct ITC titrations in thrombin and trypsin.
Thrombin
Trypsin
∆H° /
∆H° /
∆H° /
kJ mol-1
Tricine
∆H° /
kJ mol-1
Tris
∆H° /
kJ mol-1
HEPES
∆H° /
kJ mol-1
Tricine
∆H° /
kJ mol-1
Tris
kJ mol-1
Phosphate
kJ mol-1
HEPES
Inhibitor
-
39.1 ±
-30.5 ±
0.3
-17.7 ±
0.5
1
-107.0 ± 1.0 -87.5 ± 0.7 -80.6 ± 0.8 -64.7 ± 1.1
1.7
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
-
36.9 ±
0.9
-27.2 ±
1.8
-13.0 ±
1.0
2
-83.7 ± 0.5 -64.9 ± 0.1 -54.5 ± 0.5 -42.3 ± 0.2
-36.8 ±
0.5
-29.5 ±
0.2
-16.4 ±
1.0
3
4
-97.6 ± 0.3 -76.7 ± 0.4 -32.0 ± 0.7 -22.9 ± 0.9
-30.8 ± 1.1 -26.8 ± 1.3 -24.8 ± 0.5 -24.3 ± 0.9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
-
-
-
-35.7 ±
-18.5 ±
0.3
4a
-23.2 ± 1.2
-
-
-27.1 ± 0.6
-1.0 ± 1.0
1.9
a
Standard deviations were determined from triplicate measurements. ∆H° is given as the
observed enthalpy by a control experiment using 19 as a displacement ligand (Table S2) which
confirm the buffer independence of inhibitor 4 in thrombin. For 4 both direct titration and
a
displacement titration (4 ) are listed. The displacement titration was performed in phosphate and
Tris buffer since their ionization enthalpies have the largest difference of the chosen buffers.31,32
Due to the low binding affinity of 4 in trypsin, the values of 4 were obtained by displacement
titration.
As a pronounced buffer dependence has been observed, we can definitely conclude
that changes in protonation state are given for the individual ligands (s. Figure 2 and
discussion below). To relate the observed quantities to well-defined reference values in
aqueous solution, we determined the pK properties of our four ligands experimentally in
a
aqueous solution (Table 3). As suggested by the crystal structures and the order/disorder
phenomena observed next to Asp189, the unsubstituted derivative 4 indicates different
ACS Paragon Plus Environment
20

Page 21 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
binding properties in thrombin and trypsin. We therefore investigated particularly this
ligand in more detail by ITC experiments (s. below).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 3. Protonation Effects of Inhibitors 1-4 in Thrombin and Trypsin.
+
pKa1b pKa2b
Inhibitor
Δn
H
+
Δn
H
(Thrombin)
a
(
Trypsin)
1
2
3
4
0.94
0.95
0.87
0.15
0.81
6.4
6.7
6.3
5.0
7.5
7.5
7.5
6.6
0.90
0.78
1.30
aThe protonation state of all four inhibitors were determined by direct titrations except of 4 in
b
trypsin. The latter measurement was performed by a displacement titration. The pK values were
a
determined by potentiometric titration and describe the acidity of the P1 head group (pK 1) and
a
the N-terminal amino group (pK 2) in protonated state.
a
Protonation effects recorded by ITC. Table 3 lists the overall changes of the protonation
inventory of the respective inhibitors which describes the exchange of protons (Δn +
)
H
during the binding reaction between ligand, protein and buffer. The measured heat signals
were plotted against the heat-of-ionization of the respective buffer and a linear correlation
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
is indicated. The slope of the linear fit (Figure 2) gives the molar amount of protons
transferred during the entire binding process. It can be concluded that the aminopyridine
derivatives 1, 2 and 3 receive between 0.78 to 0.94 moles of protons upon
accommodation to the S1 pocket of both proteins.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. The diagrams (A: thrombin, B: trypsin) display the experimentally determined
enthalpies ΔH°obs of the inhibitors 1-4 under different buffer conditions. They are plotted
32
-1
31
-
against the heat-of-ionization ΔH° (Phosphate : 3.60 kJ mol , HEPES : 21.01 kJ mol
ion
1
31
-1
32
-1
, Tricine : 31.97 kJ mol , Tris : 47.53 kJ mol ) of the respective buffers. The slopes of
the linear fit give the molar amounts of protons that are transferred between buffer, protein
and ligand upon binding (Table 3). The intercept of these plots is reported as ‘buffer
corrected’ value (Figure 3, Table 4).
We studied the protonation inventory of this type of inhibitors already in detail in
previous contributions.19,21,33–38 The investigated D-Phe-Pro derivatives contained a P1
ACS Paragon Plus Environment
22

Page 23 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
benzamidino group, which is permanently charged both in the buffer and in the protein-
bound state. Additionally, the N-terminal group is partially charged in the buffer and
reveals full protonation in the complex. Furthermore, in all complexes His57 releases
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0
.50-0.60 moles of protons upon ligand accommodation.33 Thus, overall, the proton
release from His57 and the entrapment by the ligand’s terminal P3 amino group is virtually
compensated, leaving no net protonation effect for this type of ligands.34,35 We assume,
the same inventory applies for the latter groups in 1-4.
However, as a difference to the permanently charged P1-benzamidino derivatives, we
studied here the weakly basic (amino)pyridine derivatives 1-4 and their P1 moieties
possess pK values of 5.0-6.7 in aqueous solution. According to the Henderson-
a
Hasselbalch equation, all four inhibitors 1-4 will be largely deprotonated under the applied
buffer conditions at pH 7.8 prior to protein binding (maximally ~ 5-10% may be protonated
in the buffer). Based on the results listed in Table 3, we assume the ligands experience
a pK shift upon binding, reinforced by the spatial neighborhood to the negatively charged
a
Asp189 residue. This leads to full protonation of the pyridine-N in 1-3 for both, trypsin and
thrombin. Obviously, ligand 4 behaves differently and shows protonation only in trypsin
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
Table 3, Figure 2). Schiebel et al. could prove full protonation of aniline in trypsin, despite
30
of its rather low pK of 4.0, by a neutron diffraction study. Obviously, Asp189 has such
a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a strong influence on the protonation state of the bound ligand and probably polarizes its
neighborhood so strongly that a pK shift of several orders of magnitude is induced,
a
triggering full protonation of the ligand, even in case of a weakly basic aniline or pyridine
30
residue. For this reason, we conclude that the inhibitors 1-3 entrap 0.78-0.95 moles of
protons upon binding in both serine proteases besides the mutual compensation effect of
the terminal amino function (pK = 7.5 for 1-3,) which also becomes fully protonated upon
a
binding along with the release of 0.5-0.6 moles of protons from His57.
Surprisingly, binding of inhibitor 4 reveals a deviating protonation inventory for both
enzymes. Even though both complexes show almost identical binding poses in the S1
pocket, the unsubstituted pyridine derivative 4 binds, as expected, in protonated state at
its pyridine-N to trypsin. Since the pyridine-N in 4 (pK = 5.0) is less basic than the head
a
groups of 1-3 (pK = 6.3-6.7), a slightly larger molar amount of protons must be
a
entrapped. As mentioned above, we decided to study the ligand with a second phenyl
group at P3 to increase the binding affinity toward the enzymes. By this, also the pKa
ACS Paragon Plus Environment
24

Page 25 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
value of the N-terminal amino group is slightly shifted compared to 1-3. This may also
give rise to a slightly altered protonation inventory. Finally, due to lower affinity, the
protonation effect of 4 during binding to trypsin could not be determined by direct ITC
titration. Instead, we had to apply displacement titrations for 4 in three different buffers to
trace the changes in protonation state. Possibly, the displacement titrations increase the
estimated inaccuracy of this experiment due to error propagation. Nevertheless, the
recorded value of 1.30 moles of protons clearly indicates full protonation of the pyridine-
N head group of 4 in trypsin.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Unexpectedly, 4 shows a different protonation effect upon binding to thrombin, as no
entrapment of a proton was registered at the pyridine-N head group. In a former study by
21
Biela et al. , we could investigate the analogous D-Phe-Pro ligand and it showed no
obvious structural influence on the binding pose in the S1 pocket, thus we believe that
the binding mode and protonation inventory of the D-Phe-Pro analog and 4 can be directly
compared to the aminopyridine derivatives 1-3. Overall, only 0.15 moles of protons are
transferred upon binding of 4 to thrombin (note that a minor difference can result from the
slightly shifted pK value observed for the D-DiPhe derivative), an amount which, as it falls
a
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
close to the experimental accuracy, speaks for no protonation of 4 at the pyridine-N head
group in the complex with thrombin. This observation agrees with the data of Biela et al.,
who also did not observe any protonation of the P1 pyridine head group at their D-Phe-
Pro analog. However, they had to study the protonation inventory by displacement
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
21
titration. Similarly to a protocol by Biela et al. , we titrated the strong binding ligand 19 to
displace 4 (Table S2) as a control experiment and also by this displacement protocol no
+
change in the protonation state could be observed (Δn = -0.09 mol).
H
a
b
Table 4. Thermodynamic data of ∆H°, ∆G° and -T∆S° from direct ITC titrations in
thrombin and trypsin.
Thrombin
Trypsin
ACS Paragon Plus Environment
26

Page 27 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
∆
H° /
∆G°/
kJ mol
-T∆S° /
Kd /
nM
∆H° /
∆G° /
-T∆S° /
kJ mol
Kd /
µM
-1
-1
-1
-1
-1
-1
Inhibitor kJ mol
kJ mol
69.6
51.6
61.7
1.2
kJ mol
-56.2
-56.1
kJ mol
-23.8
-25.3
-24.9
-17.7c
1
2
3
4
-109.4
-86.0
-98.2
-30.6
-39.8
-34.4
-36.5
-29.4
100
916
32.4
68
38
30.8
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
400
-53.6
-61.9c
a
28.7
43
3681
44.3c
757c
All results are presented as buffer-corrected values. ∆G° is given as the mean over the
results by titrations in different buffers (Phosphate, HEPES, Tricine and TRIS; in case of
b
trypsin without phosphate). -T∆S° has been calculated as the difference between ∆G°
c
and the buffer corrected ∆H° values. These values were obtained by displacement
titrations.
Figure 3. Thermodynamic signatures of the inhibitors 1-4 in complex with thrombin (A)
and trypsin (B). The results are represented as buffer-corrected values listed in Table 4.
They were determined by direct titration in thrombin. Notably, due to the relatively weak
affinity of the studied ligands toward trypsin (K values between 40-500 µM) all ITC
i
measurements with trypsin have been performed as displacement titrations.
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 75
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Comparison of the thermodynamic signatures. In the following, the buffer corrected
values, which result from a linear extrapolation to hypothetical buffer conditions with zero
heat-of-ionization (Figure 3 and Table 4) will be discussed. Based on the ITC data, we
computed K values and they match within 3x standard deviation with the K values from
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
d
i
our enzyme kinetic assay (Table 1 and Table 4). Apart from the fact that very different
signals are evaluated to obtain the inhibitory potency, the meaning of K as ITC
d
dissociation constant and K as inhibitory constant from an enzyme assay are not
i
identical.
The three aminopyridine inhibitors 1-3 show similar signatures in their thermodynamic
profiles toward thrombin. The Gibb’s free energy ΔG° of all ligands falls into a range of -
-1
34 to -40 kJ mol resulting from an enthalpically more favored binding profile. This can
be explained by the charge-assisted interactions to Asp189, sometimes mediated by
interstitial water molecules. Further interactions are formed by the amino groups of 1 with
Gly219 and Glu192. Furthermore, contacts with the water molecules W1 and W2 are
found. In summary, these interactions give rise to a very strong enthalpic signature (ΔH°
ACS Paragon Plus Environment
28

Page 29 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
-
1
=
8
-109.4 kJ mol ) of 1, closely followed by the other aminopyridine derivatives (ΔH° = -
-1
-1
6.0 kJ mol for 2 and ΔH° = -98.2 kJ mol for 3), which exhibit only a single amino group
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
for interactions. The ligands are the most enthalpic binders found so far with this
scaffold.20,21,36–38 The stronger enthalpic signature compared to the analogous
benzamidine derivative results from the enthalpically less-expensive desolvation of the
P1 head groups in 1-3 which are uncharged in solution prior to protein binding and only
become charged in the binding pocket. This is definitely a desirable scenario, as
39
enthalpically binding ligands have been favored as development candidates and the
mere change of the protonation state at the protein binding site helps to design molecules
with improved bioavailability as charged species exhibit unfavorable PK properties. The
large enthalpic signature is compensated by an unfavorable entropic contribution.
-1
Ligands 1 and 3 are separated by a -TΔΔS°3→1 difference of -7.9 kJ mol , whereas 2
differs even more in entropy (-TΔΔS°2→1 = -18.0 kJ mol and -TΔΔS°2→3 = -10.1 kJ mol-1).
Its slightly increased entropic advantage over 1 and 3 results from the fact that 2 is able
to bind with two alternative conformations (Figure 1B), which is, with respect to the
inventory of degrees of freedom, entropically beneficial.
-1
ACS Paragon Plus Environment
29