Design, Synthesis, and Pharmacological Characterization of a Neutral, Non-Prodrug Thrombin Inhibitor with Good Oral Pharmacokinetics

This is an open access article published under a Creative Commons Non-Commercial No

Derivative Works (CC-BY-NC-ND) Attribution License, which permits copying and

redistribution of the article, and creation of adaptations, all for non-commercial purposes.

pubs.acs.org/jmc

Drug Annotation

Design, Synthesis, and Pharmacological Characterization of a

Neutral, Non-Prodrug Thrombin Inhibitor with Good Oral

Pharmacokinetics

Alexander Hillisch,* Kersten M. Gericke, Swen Allerheiligen, Susanne Roehrig, Martina Schaefer,

Adrian Tersteegen, Simone Schulz, Philip Lienau, Mark Gnoth, Vera Puetter, Roman C. Hillig,

and Stefan Heitmeier

Cite This: https://dx.doi.org/10.1021/acs.jmedchem.0c01035

Read Online

sı

*

Metrics & More

Article Recommendations

Supporting Information

ACCESS

ABSTRACT: Despite extensive research on small molecule thrombin

inhibitors for oral application in the past decades, only a single double

prodrug with very modest oral bioavailability has reached human therapy

as a marketed drug. We have undertaken major eﬀorts to identify neutral,

non-prodrug inhibitors. Using a holistic analysis of all available internal

data, we were able to build computational models and apply these for the

selection of a lead series with the highest possibility of achieving oral

bioavailability. In our design, we relied on protein structure knowledge to

address potency and identiﬁed a small window of favorable

physicochemical properties to balance absorption and metabolic stability.

Protein structure information on the pregnane X receptor helped in

overcoming a persistent cytochrome P450 3A4 induction problem. The

selected compound series was optimized to a highly potent, neutral, non-

prodrug thrombin inhibitor by designing, synthesizing, and testing

derivatives. The resulting optimized compound, BAY1217224, has reached ﬁrst clinical trials, which have conﬁrmed the desired

pharmacokinetic properties.

In this paper, we analyze the molecular features that lead to a

signiﬁcantly lower oral druggability of thrombin in comparison

to FXa and describe the drug discovery eﬀorts at Bayer AG to

identify neutral, orally available thrombin inhibitors.

INTRODUCTION

■

Thrombin (or blood coagulation factor IIa) can be regarded as

the key target in the blood coagulation cascade. The chase for

oral thrombin inhibitors thus began in the late 1970s, leading

to ximelagatran,¹and have culminated in only one marketed

product, dabigatran etexilate²(Figure 1), a benzamidine

double prodrug with only very modest oral bioavailability. In

the late 1990s alternative targets, such as FXa, were explored

with the discovery that the highly basic, cell permeation-

preventing benzamidine group can be replaced by a neutral

chloro or methoxyaryl group.³Both aspects led to the fact that

thrombin drug discovery was overtaken by FXa research,

producing compounds with clearly superior pharmacokinetics.

Those compounds (rivaroxaban,⁴apixaban,⁵edoxaban,⁶

betrixaban⁷) ﬁnally entered the market and are highly

successful remedies. Results from FXa research were trans-

ferred back to thrombin with some success. Merck and

AstraZeneca have reported on neutral thrombin inhibi-

tors.⁸⁻¹⁰The neutral prodrug AZD8165 entered clinical trials

but showed short half-life in healthy male volunteers. To the

best of our knowledge, currently there is no neutral thrombin

inhibitor in clinical development.

RESULTS AND DISCUSSION

■

Druggability of Thrombin. The Bayer AG search for

novel thrombin inhibitors started with a high-throughput

screening campaign of our in-house compound pool. The

approach was complemented with protein structure-based

design. These eﬀorts resulted in four distinct chemical series.

During lead optimization phases, several thousand compounds

were synthesized and characterized.¹¹⁻¹⁶However, we had to

face up to the signiﬁcantly lower druggability of thrombin in

comparison to FXa.

Received: June 17, 2020

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

A

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

Figure 1. Selected small molecule, oral thrombin inhibitors.

Figure 2. SiteMap analysis of selected serine proteases from the blood coagulation cascade.

The druggability of thrombin has been analyzed in detail and

classiﬁed as being “diﬃcult”.¹⁷Upon analysis of the X-ray

structures of selected serine proteases from the blood

coagulation cascade using the SiteMap¹⁸approach, it becomes

evident that FXa has the binding site with the highest oral

druggability (Dscore 1.08, volume 280 Å³), followed by

thrombin (Dscore 1.05, volume 412 Å³) and FXIa (Dscore

1.05, volume 312 Å³, Table S1). Both thrombin and FXa stand

out from other proteases of the coagulation cascade due to

their higher hydrophobicity (Figure 2, orange volume). The

main diﬀerence between thrombin and FXa is the signiﬁcantly

larger size of the thrombin pocket which is attributed to the

deep S2 pocket. The S1 pockets are nearly identical for the

three top-ranked proteases, whereas FVIIa and FIXa both

diﬀer in one amino acid from the other three proteases: Ala190

is substituted by a serine, forming a hydrogen bond with a

deeply buried water molecule and preventing its displacement

by P1 chloroaryl-containing inhibitors. The substrate binding

sites of the latter two proteases can be regarded as orally

undruggable (Table S1, Supporting Information).

compound of this pool stood out, capturing our attention

(Figure 3).

Figure 3. Structure of compound 1 as weak thrombin inhibitor with

interesting physicochemical properties.

Compound 1 has convincing metabolic clearance (in vivo iv

blood clearance, rat, 1.4 L h⁻¹kg⁻¹) and cell permeability (in

vitro Caco-2 P_app(A−B) cell permeation, 62 nm/s). Unfortu-

nately, this compound is only a relatively weak thrombin

inhibitor (IC₅₀= 220 nM). With a clogP of 3.2 and a PSA of

76 Å², 1 served as a prototype with respect to physicochemical

properties. From the analysis of all our patented examples, we

speculated that those with a signiﬁcantly reduced lipophilicity

(clogP = 2.5−3.5) and a PSA not larger than approximately

105 Å²would result in compounds with a higher chance of oral

bioavailability.

The task was then to ﬁnd the right structural inhibitor series

that would allow subnanomolar thrombin inhibition within the

frame of physicochemical constraints.

To do so, we analyzed the substrate binding pocket of

thrombin with WaterMap,^22,23a molecular dynamics approach

that characterizes solvent structure and energetics within a

ligand binding pocket of a target protein. By use of

inhomogeneous solvation theory,²⁴the system interaction

energy and excess entropy for each water molecule in the

Lead Finding and Lead Series Selection. In 2009, we

initiated a novel attempt aimed at the diﬃcult, but presumably

resolvable, task of identifying non-prodrug, orally bioavailable

inhibitors of thrombin.

First, we wanted to identify a window of physicochemical

properties that would lead to favorable in vitro and

subsequently in vivo pharmacokinetics. The average clogP,¹⁹

topological polar surface area (tPSA),²⁰and corrected

21

molecular weight (MW_corr

)

for 521 exempliﬁed compounds

in our patents¹¹⁻¹⁶is 5.1, 84 Å², and 488 Da, respectively. This

nicely reﬂects the properties of the large, hydrophobic

thrombin substrate binding pocket. The properties of one

B

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

substrate binding pocket are calculated. In principle, the less

energetically (enthalpically and/or entropically) favorable

expelled waters are, the more favorable is their contribution

to the binding free energy. In Figure 4, hydration sites with the

site, site 31, was identiﬁed on the border of the S4 pocket (ΔG

= 3.5 kcal/mol), along with a site within the S3 pocket (site 34,

ΔG = 3.4 kcal/mol).

As a result of these calculations, we hypothesized that the

ideal thrombin inhibitor would occupy only the S1, S2, and S4

pockets; to keep lipophilicity, PSA, and molecular weight low,

it would not extend with larger substituents into other

hydrophobic subpockets. Following that theory, we cocrystal-

lized representative compounds from all four chemical series

with thrombin and analyzed their binding features in light of

the “three-pocket hypothesis”.

The physicochemical properties and some SAR features of

the four series (Figure 5) can be explained as follows (for

detailed physicochemical values, see Table S3, Supporting

Information):

Indazoles. Branching from a central indazole core, this

series occupies the S1 pocket with a chloroaryl residue and the

S4 pocket with an aromatic or cycloaliphatic residue. Due to

the geometry of the indazole, the S2 pocket can only be

addressed by a “one-heavy-atom” substituent at position 4. But

with that substituent, the S2 pocket is not completely ﬁlled.

Substitution with chlorine (3) diminishes inhibition slightly

from an IC₅₀of 173 nM (for 4-unsubstituted 2) to 225 nM,

whereas a “two-heavy-atom” methoxy substitution (4) leads to

a loss of activity by a factor of 3 (IC₅₀= 464 nM) (Figure 5) as

the aromatic CH at position 3 of the indazole pushes the

methoxy carbon toward the 5-position. This conformation is

energetically favored by 2.8 kcal/mol (OPLS2005 force

ﬁeld)²⁵but does not ﬁt into the S2 pocket (Figure 6). The

overall properties are that this class is relatively lipophilic (av.

clogP = 4.6) with a small PSA of 68 Å²(Table S3, Supporting

Information).

Figure 4. WaterMap hydration sites in the substrate binding pocket of

thrombin. Sites are color- and size-coded with regard to the modeled

hydration free energy (high ΔG large, reddish; low ΔG small,

greenish).

color-coded modeled highest free energy (ΔG) are displayed

(for detailed values, see Table S2, Supporting Information).

The analysis suggested that the top three hydration sites (site

3, ΔG = 6.9 kcal/mol; site 36, ΔG = 6.6 kcal/mol; site 1, ΔG =

5.6 kcal/mol) with the highest free energy gain upon complex

formation are the ones located in the S1, S2, and S4 pockets of

thrombin. Hydration site 8 (ΔG = 5.5 kcal/mol) was not

further considered since it lies somewhat outside the space

ligands typically occupy. Two less important hydration sites (5

and 11) lie on the edges of a hypothetical triangle that

connects the S1, S2, and S4 pockets. An interesting hydration

Oxazolidinones. The main pharmacophoric residues are a

chlorothiophene as P1 and a ﬁve- or six-membered heterocycle

or heteroaryl, as P4 residue (e.g., 5). The central phenyl

residue allows addressing of the S2 pocket. However, complete

ﬁlling of this pocket requires a long, ﬂexible, at least “three-

Figure 5. X-ray structures of potential lead compounds in complex with thrombin. All four lead series were discovered by internal HTS at Bayer.

The red x highlights unoccupied volume in the S2 pocket. Compound surface colors represent suitability of lead structure to archive highly potent

orally bioavailable thrombin inhibitors (red, low; yellow, medium; green, high) and show the degree of ﬁlling of the S2 pocket.

C

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

improve selectivity toward FXa inhibition. In contrast to the

other three series, oxazolidinones bind less deeply in the

substrate pocket. They occupy volume around the more

exposed ester binding pocket (EBP) and the S3 pocket and do

not directly contact His57 and Ser195. The average

physicochemical properties, namely, low clogP (3.2), medium

PSA, and a lipophilic ligand eﬃciency²⁶(LLE) of 4.0, of the

oxazolidinones are attractive, being superior to the other three

series.

Imidazoles. The imidazole core is substituted with two

aromatic moieties occupying the S1 pocket and the EBP, a

subpocket on the rim of the S1 pocket directly above the

disulﬁde bond formed by Cys191 and Cys220. An aliphatic

branched residue ﬁlls the S2 pocket; the S4 pocket is occupied

by diﬀerent bulky substituents (6). Filling of the EBP is

essential for activity. The resulting high mass (av 521 Da) and

lipophilicity (av clogP = 5.6) contribute to the low average

ligand eﬃciency²⁷(LE = 0.26) and LLE (1.7) which are the

worst within these four series.

Figure 6. Comparison of indazole vs benzoxazole ﬁt in the S2 pocket

(gray).The blue arrow shows the diﬀerence in internal energy

between two conceivable conformations each.

heavy-atom” group. A rigid methoxy substituent (5) does not

completely ﬁll the S2 pocket (Figure 5). P2 residues usually

a

Table 1. Initial SAR of Rigid P4 Residues

a

n.d.: not determined.

D

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

Benzoxazoles. Branching from a central benzoxazole core,

this series occupies the S1 pocket with a chloroaryl residue and

the S4 pocket with diﬀerent substituents. In contrast to the

indazoles, the S2 pocket can be ﬁlled perfectly with a “two-

heavy-atom” residue like methoxy. This substituent seems to

be the local optimum (Figure 5). The oxygen at position 1 of

benzoxazole pushes the methoxy carbon toward the 6-position.

This conformation is energetically favored by 1.5 kcal/mol and

ﬁts exactly into the S2 pocket (Figure 6). Derivatives lacking

substitution at position 7 (7) or with a “one-heavy-atom”

substituent (8 and 9) are signiﬁcantly less potent. Interestingly,

an ethyl residue prefers conformations in which the terminal

carbon is positioned above or below the aromatic plane; 11

thus loses activity by a factor of nearly 10 in comparison to the

methoxy derivative 10. N-Propyl substitution (12) leads to a

loss of inhibitory activity by a factor of 80 relative to 10. The

average properties of this series (high lipophilicity, clogP = 5.1;

low LLE, 1.7) are unfavorable.

Despite the poor physicochemical properties of the

benzoxazole compound class, it was selected as the most

promising of the four classes. Our decision was based on the

perfect ﬁt of the methoxy-substituted core to the S2 pocket, in

conjunction with a mass eﬃcient addressing of the S1 pocket

by a chlorobenzyl residue and a binding mode that buries the

compound deep inside the substrate pocket. We planned to

occupy the S4 pocket with rigid substituents, which should

result in high-aﬃnity compounds, and to decorate future

compounds with more polar binding substituents, to reduce

the overall lipophilicity and stay within the window of

physicochemical properties, presumably leading to good oral

pharmacokinetics. Other series with apparently more attractive

physicochemical properties of, at that time, existing com-

pounds (e.g., oxazolidinones) were deprioritized in favor of

benzoxazoles which showed a more attractive binding mode

and facilitated planning of virtual compounds in terms of their

physicochemical properties.

Initial Lead Optimization. Lead optimization started with

the branched P4 benzoxazole 10 (Table 1). This compound is

characterized by ﬂexible side chains as P3/P4 residues and a

high lipophilicity. To gain potency and introduce more

polarity, we substituted the ﬂexible P4 residue by heterocycles.

An unsubstituted pyrrolidine resulted in an only moderately

active 13. Docking studies suggested substitution at the 2-

and/or 3-position of the pyrrolidine to replace water molecules

buried in the S3 and S4 pockets (Figure 4). Dimethyl

substitution at either position (14 and 15) signiﬁcantly

increased potency and lipophilicity. To introduce polarity,

further reduce ﬂexibility, and make use of more appropriate

exit vectors for substituents that enable displacement of

energetically unfavorable water molecules (Figure 7), we

changed to morpholine. The unsubstituted morpholine

derivative 16 displayed only weak activity. Since we intended

to ﬁll the S4 pocket completely, 3,3-dimethyl substitution was

introduced (17) which increased inhibition to an IC₅₀of 27

nM. This design hypothesis was veriﬁed with an X-ray

cocrystal structure of 17 with thrombin (see Table S4,

Supporting Information). To install polarity, the S4 pocket

was analyzed with regard to exposed hydrogen-bond acceptor

functionalities. The closest carbonyl backbone group is

Glu97A. Considering that the backbone NH of Glu97A is

not hydrogen-bonded intramolecularly within thrombin, we

expected some ﬂexibility of the amide group and hoped to be

able to form a hydrogen bond with a hydroxyl group at

Figure 7. Modeled complex of 16 in thrombin (based on the X-ray

structure of 10 in complex with thrombin). S1, S2, and S4 pockets,

the oxyanion hole (OAH), and the ester binding pocket (EBP) are

displayed.

position 4 of an aliphatic six-membered ring. The resulting 4-

hydroxypiperidine derivative 18 was only modestly active.

However, clogP was signiﬁcantly reduced to 2.3 and rat in vitro

hepatocyte clearance was improved (1.3 L h⁻¹kg⁻¹). Hoping

to increase thrombin inhibition, the piperidine was substituted

at position 3 with a geminal dimethyl group (19) or a methyl

group (20a), displacing a buried water molecule in the S4

pocket. 20a, having a low molecular weight (MW = 414 Da)

and balanced polarity (PSA = 88 Å², clogP = 2.8), was a

signiﬁcant step forward toward orally active thrombin

inhibitors, and it was broadly characterized.

Characterization of 20a. 20a showed strong thrombin

inhibition (biochemical IC₅₀= 10 nM; in human plasma, IC₅₀

= 60 nM) and a doubling of clotting time (prothrombin time,

PT) at 6 μM (EC₂₀₀). It had excellent permeation properties

[P_app(A−B) = 126 nm/s, Caco-2 assay] and a moderate in vitro

hepatocyte clearance (2.3 L h⁻¹kg⁻¹). In an in vivo rat

pharmacokinetic study, blood clearance, t_1/2, and oral

bioavailability were found to be moderate (1.7 L h⁻¹kg⁻¹,

1.3 h, and 46%, respectively; see Table 5). 20a did not show

any CYP inhibition, up to the highest concentration tested (20

μM), of the major enzymes, but a strong CYP3A4 induction

eﬀect via the pregnane X receptor (PXR) pathway was

discovered. The onset of CYP3A4 induction in human

hepatocytes was above 370 ng/mL (Table 5), and trans-

activation of human PXR started at a minimum eﬃcacious

concentration (MEC) of 0.69 μM (∼300 ng/mL) and was

regarded as major optimization parameter.

To overcome PXR binding and thus CYP3A4 induction, we

carried out cocrystallization experiments with the PXR ligand

binding domain and several compounds, among them 17 and

20a; however, cocrystallization was successful only for 17.

Although the weak density indicated occupancy of only 0.70,

the chlorophenyl residue and the dimethylmorpholine group

could be positioned unambiguously (Figure 8). The

compound binds to the PXR with the chlorophenyl moiety

pinched between Trp299 and Phe288 and forming a stacking

interaction with Trp299. The chlorine is in close contact with

Val211 and Leu209. The benzoxazole nitrogen serves as a

hydrogen-bond acceptor interacting with the Gln285 side

E

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

Figure 8. X-ray structure of 17 bound to the PXR ligand binding domain (left, PDB code 6TFI) and induced ﬁt docking pose of 20a in the PXR

(right).

chain. The dimethylmorpholine occupies a hydrophobic

pocket formed by Leu206, Leu240, and Ile414. Nevertheless,

this X-ray structure did not explain the strong PXR binding of

20a. Assuming a similar binding mode on the PXR for the

dimethylmorpholine and 4-hydroxypiperidine P4 residues, the

latter would position a polar hydroxyl group right between

three hydrophobic amino acid side chains. Since this was

regarded unlikely, an induced ﬁt docking with Glide SP²⁸on

the PXR:17 complex was performed. His407, Arg410, and

Gln285 were treated ﬂexibly. As result, the docking suggested a

diﬀerent orientation of the piperidine in comparison to the

morpholine moiety (Figure 8). While the piperidinyl methyl

group occupies a subpocket formed by Leu410 and Ile414, the

4-hydroxyl group forms a strong hydrogen bond with His407,

explaining the CYP induction properties with a somewhat

diﬀerent binding mode than that observed for 17. This binding

hypothesis is further supported by the analysis of 19 in the

PXR assay showing a signiﬁcant reduction in PXR trans-

activation (MEC = 6.3 μM, 9-fold relative to 20a; Table 1).

The additional methyl group is thought to shield the hydroxyl

group and thus would prevent hydrogen-bond formation with

His407. The speciﬁc binding mode of 20a with the PXR was

not observed for 19 when the same induced ﬁt docking

protocol was applied.

Exploring the SAR of P1 and the Bicyclic Core. To

evaluate the planned modiﬁcations of P1, we kept P4 constant

with substituent 3-methyl-4-hydroxypiperidine. Compounds in

Table 2 are racemic mixtures of two isomers with either cis or

trans conﬁguration of the piperidine substituents. In compar-

ison to the single most active isomer 20a, the racemate 20rac

loses a factor of ∼2.5 in thrombin inhibition, while in vitro

pharmacokinetic parameters were very similar.

To introduce some polarity in P1, m-chlorobenzyl was

replaced by the four possible chloropyridine analogues keeping

the chlorine position constant. As expected, 4-chloro-2-pyridyl

(21) was the most potent derivative but, nevertheless, loses a

factor of ∼9 in thrombin inhibition compared to 20rac. Of the

four P1 chloropyridine isomers, 21 is the only isomer leading

to solvent exposure of the pyridine nitrogen when bound to

thrombin. The other three isomers were signiﬁcantly weaker

with respect to thrombin inhibition. Thrombin lacks any

hydrogen-bond donor functionalities in the S1 pocket. The

pyridine nitrogen in 22 and 23 is close to polar groups

(Asp189 and Gly219 backbone carbonyl, respectively) leading

to electrostatic repulsion and consequently inactivity. In 24,

the pyridine nitrogen is placed in a lipophilic environment

resulting in a moderate loss of activity compared to 20rac. The

position para to the chlorine residue in m-chlorobenzyl was

further explored with ﬂuorine and hydroxyl substituents. The

P1 hydroxyl proton of 26 was expected to form a hydrogen

bond with the backbone carbonyl of Gly219, similar to the

bioisosteric 5-chloroindole of FXa inhibitors.²⁹The additional

hydroxyl and ﬂuorine group led to a moderate decrease in

thrombin aﬃnity. To reduce potential metabolic attack at the

benzylic P1 position, methyl, triﬂuoromethyl, cyclopropyl, and

cyclobutyl were introduced. The R-conﬁgured P1 residue 27b

with a methyl substituent only lost a factor of 2 with regard to

thrombin inhibition (vs 20rac). To vary hydrogen-bond

acceptor strength at both polar atoms of the core and to

reduce the hydrogen-bond strength to Gln285 in the PXR

(Figure 8), derivatives such as the inverted benzoxazole 31 and

triazolopyridine 32 were synthesized. Interestingly and

unexpectedly, the inverted benzoxazole was nearly inactive

while the triazolopyridine lost a factor of ∼24 in thrombin

inhibition. Quantum mechanical calculations at the OPBE/cc-

pvtz level retrospectively could explain the signiﬁcantly

reduced hydrogen bond acceptor strengths of the inverted

benzoxazole oxygen. In addition, the benzoxazole nitrogen is a

polar strong H-bond acceptor but not involved in hydrogen

bonding with thrombin which in addition most likely

contributes to the signiﬁcantly reduced binding aﬃnity.

Lead Optimization. The initial plans for further lead

optimization could then be summarized as follows.

(i) Attempt to increase thrombin inhibition with conforma-

tionally restrained P4 residues displacing or replacing two

water molecules in the S4 pocket and optionally also in the S3

pocket.

(ii) Attempt to decrease PXR binding to reduce CYP3A4

induction by modulating hydrogen-bond acceptor strength of

the benzoxazole nitrogen to diminish contact with Gln285

and/or by introducing polarity in P1 to reduce binding in the

hydrophobic cleft between Trp299 and Phe288.

(iii) Attempt to maintain the balance between absorption

and metabolic clearance. Since 20a already displayed a good

balance of main pharmacokinetic optimization parameters,

derivatives with physicochemical properties similar to 20a and

within the originally deﬁned window (clogP = 2.5−3.5, PSA <

105 Å²) would be synthesized. Analysis of several hundred

derivatives indicated that P4 residues containing one hydroxyl

group, one hydrogen-bond acceptor, and a few (2−4)

additional carbon atoms would lead to the desired properties.

In addition, we intended to block the P1 benzyl position with

small substituents to decrease metabolic attack.

F

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

a

Table 2. P1 and Core Modiﬁcations to Address PXR Binding

a

n.d.: not determined.

In summary, the P1 and core SAR study identiﬁed several

modiﬁcations that oﬀered a path forward to solve the PXR-

mediated CYP3A4 induction issue and increase metabolic

stability. Selected derivatives, designed with insight from X-ray

structural analysis, corroborated our PXR binding mode and

reduced PXR transactivation. Of special interest in that regard

was the introduction of the 4-chloro-2-pyridyl residue (21,

PXR MEC = 3.6 μM) as polar spot in a hydrophobic cleft of

G

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

a

Table 3. Exploring P4 with Piperidine and Morpholine for Potent Substituents in Deﬁned Physicochemical Space

a

b

n.d.: not determined. Single isomer of unknown conﬁguration.

H

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

the PXR (Phe288 and Trp299) and the triazolopyridine core

(32, PXR MEC = 20 μM) to diminish the hydrogen-bond

strength with Gln285 in the PXR. Introduction of a methyl

substituent at the benzylic P1 position increased metabolic

stability somewhat. However, all compound modiﬁcations

outlined in Table 2 led to a decrease in thrombin inhibition

relative to 20rac. Thus, potency had to be regained with

altered P4 residues to be able to combine the promising P1

modiﬁcations and ﬁnally increase thrombin inhibition.

from hydroxymethyl (41) to hydroxyethyl (42a). Since both

derivatives still transactivated the PXR, we designed derivatives

that positioned the hydroxyl functionality rigidly in a way that

would enhance binding to thrombin (hydrogen bond to

Thr172, Glu217) but not to the PXR (no hydrogen bond with

His407). Especially the resulting 3-hydroxycyclobutyl deriva-

tive (43) was a potent thrombin inhibitor (IC₅₀= 2 nM). PXR

transactivation, however, was not diminished by using the rigid

substituents. To further explore these results, 42a was

cocrystallized with thrombin, leading to a surprising result

with respect to stereochemistry of the most active stereo-

isomer. The electron density clearly identiﬁed the P4 S,S-

isomer 42a as the slightly more potent (IC₅₀= 4 nM) relative

to the corresponding R,R-isomer 42b (IC₅₀= 6 nM). Isomer

42a positions the hydroxyethyl residue in an equatorial

conformation with its terminal hydroxyl group hydrogen-

bonded to Tyr60A (Figure 9). An X-ray cocrystal structure of

Optimization of the P4 Residue. The ﬁrst approach

toward more potent P4 residues involved substitution at

position 2 of the piperidine with hydrophobic substituents of

growing steric bulk to displace water molecules in the S3

pocket (Figure 4, site 34). Thrombin inhibition decreased with

methyl substitution (33, Table 3) relative to 20a. Ethyl and

cyclopropyl (34 and 35) substituents gave the desired increase

as these were presumably bulky enough to displace water in

the S3 pocket. But since these derivatives were too lipophilic

(clogP larger than 3.5), clearance was signiﬁcantly increased

and 35 still showed strong PXR transactivation. Addition of a

second hydroxyl function to P3 resulted in the highly potent

36 (IC₅₀= 270 pM) with very low clearance but unfortunately

strongly reduced Caco-2 permeation. Omitting the 4-hydroxyl

group (37a, 37b) led to potent, highly permeable but too

lipophilic compounds with strong PXR transactivation and

high metabolic clearance in rat hepatocytes. In summary,

adding or omitting additional hydroxyl groups in the 4-

hydroxypiperidine subseries resulted in compounds that were

either too lipophilic or too hydrophilic. To reduce the

granularity of polar chemical functionalities in P4, we elected

to integrate an ether oxygen into the ring and limit substituents

to one hydroxyl function. Thus, piperidine was exchanged by

morpholine. In addition, the benzylic position was “blocked”

with a methyl substituent (R-stereoisomer) to increase

metabolic stability, as outlined above.

Figure 9. X-ray cocrystal structure of 42a with thrombin (PDB code

6ZUX).

Substitution at position 2 of the morpholine with methyl and

2-hydroxyethyl residues gave compound 38 showing promising

in vivo PK properties such as 38% oral bioavailability in rats

(Table 5). This particular P4 residue was designed to ﬁll the S4

pocket (Figure 4, site 1) with the methyl group and

simultaneously replace a water molecule contacting the

Tyr60A hydroxyl and the backbone Glu97A carbonyl oxygen

(Figure 4, site 31) with the terminal hydroxyl function.

Thrombin inhibition was still regarded as insuﬃcient. Further

substitution with a methyl or ethyl group at position 3 of the

morpholine led to extremely potent inhibitors (39 and 40, 0.5

and 0.4 nM thrombin IC₅₀). 40 was cocrystallized with

thrombin, with X-ray analysis conﬁrming the stereochemistry

of the most potent stereoisomer and the anticipated binding

mode. Since the aﬃnity gain from methyl to ethyl was small,

we speculate that the rigidiﬁcation with the additional

substituent conformationally rearranges the hydroxyethyl/

methyl pharmacophore in an energetically favorable way for

binding. Water displacement (site 34, Figure 4) plays a minor

role in this case. Unfortunately, the lipophilicity increased

signiﬁcantly, leading to high metabolic clearance. To reduce

lipophilicity, the P4 morpholine was substituted with a methyl

group at position 5 and a hydroxyalkyl residue at position 3.

This functional group was expected to form a hydrogen bond

with the carboxylate side chain of Thr172 or Glu217 in the S3

pocket. 40 still showed strong PXR transactivation (MEC =

0.26 μM) and weaker CYP3A4 induction in the hepatocyte

assay (NOEL = 1111 ng/mL). Thrombin inhibition increased

42b revealed that the R,R-isomer binds with the methyl group

in the S4 pocket. Instead of hydrogen bonding to residues in

the S3 pocket, the terminal hydroxyl group forms an

intramolecular hydrogen bond with the compound’s amide

oxygen and occupies the S3 pocket (results not shown; see

Table S4, Supporting Information).

Consolidating ADMET and Potency Enhancing

Groups. The most potent P4 substituents were subsequently

combined with somewhat more polar cores and P1 residues to

address ADMET issues like clearance and CYP induction. As

shown in Table 4, 4-chloro-2-pyridyl P1 derivatives (44, 45,

46) showed a reduction in PXR transactivation relative to 42a.

In the cases of 44 and 45 this also led to the desired complete

loss of CYP3A4 induction as measured in the hepatocyte assay

(NOEL > 10 000 ng/mL, Table 5). Thus, single modiﬁcations

(e.g., of P4 or P1) that slightly decrease PXR aﬃnity contribute

synergistically. Such combinations can even lead to a

signiﬁcantly reduced PXR transactivation. 45, with convincing

in vitro pharmacokinetic properties, clearly demonstrated the

success of the combination strategy. However, thrombin

inhibition was still deemed to be insuﬃcient. The other two

P1 4-chloro-2-pyridyl derivatives (44, 46) were potent but less

attractive with respect to clearance in hepatocytes.

A similar picture emerged for combinations of the

triazolopyridine core with the chlorobenzyl P1 residue (47,

48, 49). Thrombin inhibition was even weaker relative to the

benzoxazole core derivatives, and PXR transactivation was

I

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

a

Table 4. Combination of Potent P4 Residues with Core and P1 Modiﬁcations

a

b

Single isomer of unknown conﬁguration, n.d.: not determined.

reduced but metabolic stability was not suﬃcient. 47 showed

no detectable CYP3A4 induction (Table 5).

hydrophobic and hydrogen-bond donor functionalities. The

only hydrogen-bond donors in the direct vicinity of 42a were

the two backbone NH groups forming the oxyanion hole

(Gly193 and Ser195). Interestingly 28, as a mixture of eight

isomers, gave a hint that positioning a triﬂuoromethyl group in

this region led to a decrease in thrombin aﬃnity. Molecular

modeling studies helped to explain the observed eﬀect by

showing that only one ﬂuorine atom is needed as acceptor,

whereas the other two ﬂuorines result in electrostatic

repulsion. Therefore, monoﬂuoromethyl derivatives were

Nevertheless, with these compounds we were close to an

acceptable overall pharmacokinetic proﬁle (Table 5). However,

thrombin inhibition was considered to be about 1 order of

magnitude too weak. The challenge was to devise modiﬁca-

tions that would increase thrombin binding without changing

the physicochemical properties signiﬁcantly. A thorough

analysis of the cocrystal structure of 42a with thrombin was

performed. The binding site surface was scanned for

J

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

Table 5. In Vivo Pharmacokinetic Properties (iv Bolus and po Rats) and CYP3A4 Induction of Selected Compounds

d

MRT_iv

[h]

AUC_norm

CL_b

Blood/plasma

ratio

V_ss

t_1/2

CYP3A4 induction, NOEL

[ng/mL]

b c

,

Compd

[kg·h/L]

[L h⁻¹kg⁻¹

]

[L/kg]

[h]

F

[%]

1.9

a

Dabigatran

(etexilate, po)

1.6

3.0

0.59

n.d.

0.54

1.2

n.d.

20a

1.1

1.8

1.2

1.3

0.46

0.51

0.36

1.1

0.48

0.33

0.32

0.47

0.23

0.39

0.16

0.33

0.34

1.7

2.9

3.9

2.4

4.0

2.4

4.6

3.2

3.1

1.24

1.07

0.88

1.08

1.35

0.94

0.97

2.4

5.6

3.7

2.7

2.0

1.3

2.2

3.3

3.5

1.3

2.7

1.0

1.1

0.44

0.5

0.32

1.1

1.4

46

38

370

n.d.

1111

n.d.

>10000

n.d.

38

40

41

44

45

47

50

51

n.d.

38

1.2

>10000

a

The iv experiments were performed with dabigatran (zwitterionic compound, no prodrug), and po experiments were performed with dabigatran

b

c

d

etexilate. Oral bioavailability. n.d.: not determined. NOEL: no eﬀect level.

thought to be slightly more polar (compared to methyl

substituents) in a position that should enhance thrombin

binding and further block the benzylic position to metabolic

attack. Several derivatives were prepared, and while thrombin

inhibition was indeed increased by a factor of about 5 (relative

to the ﬂuorine lacking derivatives), the eﬀects on clearance and

CYP induction were not conclusive or counterproductive

throughout this subseries. The binding mode of 51 was

conﬁrmed by X-ray cocrystal structural analysis and is

essentially identical to 42a, with the ﬂuorine substituent in

close vicinity to the oxyanion backbone NH groups of Gly193

(2.3 Å) and Ser195 (2.2 Å). 51 was characterized in depth.

Characterization of the Clinical Candidate 51. 51

showed convincing overall properties with subnanomolar

thrombin inhibition (IC₅₀= 0.71 nM), excellent Caco-2

permeability, low plasma protein binding across species (e.g.,

binding of 78% in rat and human) and a moderate stability in

hepatocytes and in vivo clearance. Although PXR trans-

activation was still detectable (MEC = 1.9 μM, Table 4), no

CYP3A4 induction in human hepatocytes (NOEL > 10000

ng/mL, Table 5) was found. The reasons for this discrepancy

are unknown. In general, a good correlation between PXR

transactivation and CYP induction in hepatocytes was

observed for the underlying compound class (compare Tables

4 and 5). 51 does not aﬀect related serine proteases (such as

trypsin, plasmin, tPA, FVIIa, FXa, FXIa, and plasma kallikrein)

at concentrations up to 20 μM, demonstrating >10 000-fold

selectivity for thrombin relative to these other homologous

serine proteases. In rat, rabbit, and dog plasma, IC₅₀values of

10 nM, 20 nM, and 6.5 nM, respectively, were measured.

The prolongation of the prothrombin time (PT) and the

activated partial thromboplastin time (aPTT) were measured

to determine the in vitro anticoagulant activity of 51 (deﬁned

as the inhibitor concentration required to double the time to

ﬁbrin formation). In human plasma, the clotting times were

doubled at a concentration of 1.1 μM (PT) and of 0.70 μM

(aPTT).

Figure 10. In vivo antithrombotic eﬀects of 51 (n = 6−10, SEM):

(A) thrombus weight reduction in an arteriovenous shunt model in

rat (n = 6 SEM); (B) simultaneously conducted tail bleeding time

model in the same animals; (C) thrombus weight reduction in a

venous thrombosis model (“Wessler model”) in rat; *p < 0.05, ***p <

0.0001.

demonstrated dose-dependent thrombus weight reduction

after po administration (Figure 10C).

Due to its overall favorable and balanced pharmacokinetic

and pharmacodynamic properties, 51 (BAY1217224, Table 5)

was selected for a clinical phase 1 study. Human PK was

predicted using allometric scaling based on three species

leading to a peak/trough ratio of 3.4 after BID dosing. Since

there is no need for a prodrug approach, lower intraindividual

diﬀerences are expected, and the low renal excretion rate

makes it an excellent candidate for use in patients with

impaired renal function, a diﬃcult-to-treat population to date.

Compound 51 was orally administered to 22 healthy

volunteers in three increasing doses. Peak plasma concen-

The in vivo antithrombotic eﬀects of 51 were determined in

in vivo models for arterial and venous thrombosis after iv and

po application. In the arteriovenous shunt model in

anesthetized rats, 51 reduced the thrombus weight in a dose-

dependent manner after iv administration (Figure 10A). When

the bleeding time was determined at the tail, mild prolongation

(<2-fold) was observed (Figure 10B). In another thrombosis

model with venous stasis in the vena cava in rats, 51

trations were reached early after oral administration (t_max

=

0.75 h) and a low-to-medium variability of AUC and C_maxwas

observed. The terminal half-life ranged from 9.5 to 13 h.

K

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

a

Scheme 1. General Synthetic Pathway for the Benzoxazole Derivatives

a

Reagents and conditions: (a) R′ = H, Ag₂CO₃, benzoic acid, acetonitrile, 60 °C; (b) R′ = Cl, Hu

̈

nig’s base, DMF, rt; (c) 1 N NaOH, 1,4-dioxane,

rt; (d) HATU, Hu

̈

nig’s base, DMF, rt.

̀

Polak−Ribiere conjugate gradient, convergence threshold 0.5) in

CONCLUSION

■

̈

MacroModel (Schrodinger LLC).

Induced Fit Docking³¹in the PXR. PXR active compounds were

docked using Glide in the ﬁrst step of the induced ﬁt protocol. The X-

ray structure of the PXR in complex with 17 was used for the docking

experiments. To generate a diverse ensemble of ligand poses, reduced

ligand and protein van der Waals radii (scaling factor 0.5) were used

during the docking step. For each pose, a Prime structure prediction

was then used to accommodate the ligand by reorienting nearby side

chains (within a radius of 5 Å). Especially His407, Arg410, and

Gln285 were treated ﬂexibly in the docking protocol. These residues

and the ligand were then energy-minimized. Each PXR active

compound was re-docked into its corresponding low-energy protein

conformations (energy window of 30 kcal/mol to the lowest energy

conformation), and the resulting complexes were ranked according to

GlideScore.

Chemistry. We describe the synthesis of 51 here in detail.

Synthesis of the other derivatives (10, 13−19, 20a/20b, 21−26, 27a/

27b, 28−32, 33a/33b, 34−36, 37a/37b, 38−41, 42a/42b, 43−50,

52) is described in Supporting Information or in the ﬁled patent

applications.^15,32−34The key step in the synthesis of the benzoxazoles

V (Scheme 1), the introduction of the benzylic amine moiety, could

be successfully carried out starting from either the unsubstituted (R′ =

H) or the prefunctionalized (R′ = Cl) benzoxazole I. In the ﬁrst case,

the silver-mediated direct amination methodology developed by

Chang and co-workers³⁵led to the desired compounds II, usually in

moderate yield. The second possible access was the generally higher

yielding S_NAr reaction via the chlorinated benzoxazole I.³⁶Afterward,

saponiﬁcation of esters II under standard conditions resulted in high

yields of the corresponding acids III which were then subjected to

amide-coupling conditions using diﬀerent cyclic secondary amines IV

and HATU as coupling agent to obtain the ﬁnal benzoxazoles V in

good yields.

The synthesis of clinical candidate 51 (BAY1217224, Scheme 2)

started from commercially available vanillic acid (53) which was

converted in ﬁve steps and high overall yield into key building block

benzoxazole 58. The required benzylic amine 65 was produced from

commercially available 3-chloroacetophenone (59) which was α-

brominated and then subjected to a bromine−ﬂuorine exchange to

yield 61. Reductive amination and Boc protection led to racemic 63

which was then separated into enantiomers via preparative chiral

HPLC chromatography to give the desired S-enantiomer 64. Finally,

benzylic amine hydrochloride 65 was obtained via Boc deprotection

under acidic conditions. The chiral morpholine building block 74 was

synthesized starting from ^L-aspartic acid (66). After conversion into

bis-methyl ester 67 and benzyl protection of the amine, reduction to

diol 69 was carried out using lithium aluminum hydride. Treatment of

diol 69 with 2-chloropropionyl chloride set the stage for cyclization

under basic conditions to provide oxomorpholine 71. Subsequent

carbonyl reduction, separation of the diastereoisomers via HPLC, and

benzyl deprotection gave the required building block 73. Finally,

reaction of benzoxazole 58 with benzylic amine 65, followed by ester

saponiﬁcation and coupling with chiral secondary amine 73, resulted

in 51.

The design of BAY1217224, an orally bioavailable, selective,

neutral, non-prodrug thrombin inhibitor, has been described.

To the best of our knowledge, this is the ﬁrst molecule

identiﬁed with such properties. Key to the discovery of

BAY1217224 was the careful selection of the benzoxazole lead

series using X-ray crystallography, computed water energetics

in the substrate pocket, and evaluation of a number of diﬀerent

potential leads in a variety of assays. In the optimization of the

benzoxazole lead series, we focused the properties of the

prepared compounds to a small window of favorable

physicochemical properties to balance absorption and

metabolic stability. A persistent cytochrome P450 3A4

induction issue was encountered during the optimization.

The problem was addressed using PXR cocrystallization,

careful SAR analysis, and induced ﬁt docking studies.

Structural elements leading to decreased PXR aﬃnity were

identiﬁed and could be combined with features reinstalling

thrombin inhibition. The resulting IC₅₀= 0.7 nM thrombin

inhibitor BAY1217224 is highly selective and has excellent

Caco-2 permeability, moderate clearance, and no CYP

induction potential in hepatocytes. The compound has reached

clinical trials which have conﬁrmed oral bioavailability in

healthy male volunteers.

EXPERIMENTAL SECTION

■

Druggability Assessment. For the modeling of oral druggability,

the SiteMap¹⁸program was applied. Fifteen site points were required

to detect a ligand binding pocket. The standard grid for site points

was used, together with the more restrictive hydrophobicity deﬁnition.

The Dscore was used to compare oral druggability of diﬀerent serine

̈

proteases. X-ray structures were prepared using Schrodinger’s Maestro

protein preparation routine.³⁰

Physicochemical Property Estimation. Lipophilicity of com-

pounds was estimated using clogP.¹⁹Polar surface area was

characterized with the tPSA²⁰descriptor. The corrected molecular

weight (MW_corr)

²¹was used to quantify molecule size. MW_corrapplies

appropriate weight corrections for molecules with halogens because

these atoms have an unproportionally low volume compared to their

atomic weight. Thus, it is a surrogate parameter for the molecular

volume and related to diﬀusion rates.

Docking in Thrombin. Glide SP version 3.5²⁸(Schrodinger

̈

LLC) was used to dock compounds into prepared thrombin X-ray

structures. The X-ray structure of thrombin in complex with 10 was

used for the majority of the docking studies. The ligand was placed in

the center of a 22 Å box to calculate the interaction grid. The van der

Waals scaling factor was set to 0.8 and the partial charge cutoﬀ to

0.15. The ligands were docked ﬂexibly and nonplanar amide bonds

were penalized. 10 000 poses per docking run were sampled. The

complex with the most probable binding pose after manual inspection

of top ranked poses was energy-minimized using the OPLS2005 force

ﬁeld²⁵(dielectric constant 1.0, constant dielectric, solvent water,

General Procedures. All commercial reagents and catalysts were

used as provided by the commercial supplier without puriﬁcation.

L

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

a

Scheme 2. Synthesis of 51 (BAY1217224)

a

Reagents and conditions: (a) SOCl₂, MeOH, reﬂux, 100%; (b) fuming HNO₃, HOAc, rt, 80%; (c) H₂(1 bar), 10% Pd/C (5 wt %), THF/EtOH

(2:1), rt, 96%; (d) pyridine, reﬂux, 96%; (e) SOCl₂, DMF (cat.), 70 °C, 82%; (f) Br₂, DCM, 82%; (g) KF, ZnF₂, TBAF·3 H₂O, acetonitrile, 80 °C,

58% or Selectﬂuor, MeOH, 110 °C, microwave, 86%; (h) (i) NH₃, Ti(Oi-Pr)₄, EtOH, rt, (ii) NaBH₄, (iii) 4 N HCl in 1,4-dioxane, 1,4-dioxane, rt;

(i) Boc₂O, triethylamine, DCM, rt; (j) chiral chromatography, 23% over ﬁve steps; (k) 4 N HCl in 1,4-dioxane, 1,4-dioxane, rt, 100%; (l) SOCl₂,

MeOH, reﬂux; (m) benzaldehyde, NaBH(OAc)₃, triethylamine, DCM, 50% over two steps; (n) LiAlH₄, THF, reﬂux, 78%; (o) triethylamine, i-

PrOH, 0 °C, 90%; (p) KOt-Bu, i-PrOH, 0 °C to rt, 90%; (q) borane−dimethyl sulﬁde, THF, rt; (r) H₂(1 bar), 10% Pd/C (5 wt %), 20%

Pd(OH)₂/C (5 wt %), EtOH, 99%; (s) Hunig’s base, DMF, rt, 92%; (t) 1 N NaOH, 1,4-dioxane, rt, 88%; (u) HATU, Hunig’s base, DMF, rt, 70%.

̈ ̈

Solvents for synthesis, extraction, and chromatography were reagent

grade and used as received. Moisture-sensitive reactions were carried

out under an atmosphere of argon, and anhydrous solvents were used

as provided by the commercial supplier. ¹H NMR spectra were

recorded on Bruker Avance spectrometers operating at 400 MHz.

Chemical shifts (δ) are reported in ppm relative to TMS as an internal

standard. The descriptions of the coupling patterns of ¹H NMR

signals are based on the optical appearance of the signals and do not

necessarily reﬂect the physically correct interpretation. In general, the

chemical shift information refers to the center of the signal. LC−MS

analysis was performed using the respective methods 1a−5a, as noted.

Unless otherwise indicated, all compounds have ≥95% purity.

Method 1a. Instrument: Waters ACQUITY SQD UPLC system;

column, Waters Acquity UPLC HSS T3 1.8 μm, 50 mm × 1 mm;

mobile phase A, 1 L of water + 0.25 mL of 99% strength formic acid;

mobile phase B, 1 L of acetonitrile + 0.25 mL of 99% strength formic

M

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

acid; gradient, 0.0 min 90% A → 1.2 min 5% A → 2.0 min 5% A;

oven, 50 °C; ﬂow rate, 0.40 mL/min; UV detection, 208−400 nm.

Method 2a. Instrument, Micromass Quattro Premier with Waters

UPLC Acquity; column, Thermo Hypersil GOLD 1.9 μm, 50 mm × 1

mm; mobile phase A, 1 L of water + 0.5 mL of 50% strength formic

acid; mobile phase B, 1 L of acetonitrile + 0.5 mL of 50% strength

formic acid; gradient, 0.0 min 97% A → 0.5 min 97% A → 3.2 min 5%

A → 4.0 min 5% A; oven, 50 °C; ﬂow rate, 0.3 mL/min; UV

detection, 210 nm.

Method 3a. MS instrument, Waters (Micromass) Quattro Micro;

HPLC instrument, Agilent 1100 series; column, YMC-Triart C18 3

μm, 50 mm × 3 mm; mobile phase A, 1 L of water + 0.01 mol of

ammonium carbonate; mobile phase B, 1 L of acetonitrile; gradient,

0.0 min 100% A → 2.75 min 5% A → 4.5 min 5% A; oven, 40 °C;

ﬂow rate, 1.25 mL/min; UV detection, 210 nm.

Method 4a. MS instrument, Waters (Micromass) QM; HPLC

instrument, Agilent 1100 series; column, Agient ZORBAX Extend-

C18 3.5 μm, 3.0 mm × 50 mm; mobile phase A, 1 L of water + 0.01

mol of ammonium carbonate, mobile phase B, 1 L of acetonitrile;

gradient, 0.0 min 98% A → 0.2 min 98% A → 3.0 min 5% A → 4.5

min 5% A; oven, 40 °C; ﬂow rate, 1.75 mL/min; UV detection, 210

nm.

Method 5a. Instrument, Waters ACQUITY SQD UPLC system;

column, Waters Acquity UPLC HSS T3 1.8 μm, 50 mm × 1 mm;

mobile phase A, 1 L of water + 0.25 mL of 99% strength formic acid,

mobile phase B, 1 L of acetonitrile + 0.25 mL of 99% strength formic

acid; gradient, 0.0 min 95% A → 6.0 min 5% A → 7.5 min 5% A;

oven, 50 °C; ﬂow rate, 0.35 mL/min; UV detection, 210−400 nm.

GC−MS analysis was performed using the respective methods 1b and

2b, as noted.

Method 1b. Instrument, Thermo DFS, Trace GC Ultra; column,

Restek RTX-35, 15 m × 200 μm × 0.33 μm; constant helium ﬂow

rate, 1.20 mL/min; oven, 60 °C; inlet, 220 °C; gradient, 60 °C, 30

°C/min → 300 °C (maintained for 3.33 min).

Method 2b. Instrument, Micromass GCT, GC6890; column,

Restek RTX-35, 15 m × 200 μm × 0.33 μm; constant helium ﬂow

rate, 0.88 mL/min; oven, 70 °C; inlet, 250 °C; gradient, 70 °C, 30

°C/min → 310 °C (maintained for 3 min).

Method 1c. MS analysis was performed using method 1c.

Instrument, Thermo Fisher-Scientiﬁc DSQ; chemical ionization;

reactant gas, ammonia; source temperature, 200 °C; ionization

energy 70 eV. Preparative enantiomer/diastereomer separation on a

chiral phase was performed using the respective methods 1d and 2d,

as noted.

Method 1d. Phase, Daicel Chiralpak AD-H SFC, 10 μm 250 mm ×

20 mm, mobile phase, carbon dioxide/ethanol 70:30; ﬂow rate, 100

mL/min, makeup ﬂow rate, 30 mL/min, backpressure, 80 bar;

temperature, 40 °C; UV detection, 220 nm.

Method 2d. Phase, Daicel Chiralpak AY-H, 5 μm 250 mm × 20

mm, mobile phase, isohexane/ethanol 90:10; ﬂow rate, 40 mL/min;

temperature, 40 °C; UV detection, 220 nm. Analytical enantiomer/

diastereomer separation on a chiral phase was performed using the

respective methods 1e and 2e, as noted.

Method 1e. Phase, Daicel Chiralpak AS-H, 5 μm 250 mm × 4.6

mm, mobile phase, isohexane/isopropanol 50:50; ﬂow rate, 1 mL/

min; temperature, 30 °C; UV detection, 220 nm.

Method 2e. Phase, Daicel Chiralpak AD-H SFC, 5 μm 250 mm ×

4.6 mm, mobile phase, carbon dioxide/ethanol 70:30; ﬂow rate, 3

mL/min; temperature, 30 °C; UV detection, 220 nm. Assessment of

optical rotation [α] was performed using an Anton Paar polarimeter

MCP200 with parameters (solvent, wavelength, temperature) as

indicated. Preparative normal phase chromatography (column

chromatography/ﬂash chromatography/MPLC) was performed

using Biotage Isolera chromatography systems with Biotage silica

cartridges or silica gel 60 (230−400 mesh) in combination with glass

columns/frits.

Methyl 4-Hydroxy-3-methoxybenzoate (54). 4-Hydroxy-3-

methoxybenzoic acid 53 (60.0 g, 357 mmol) in methanol (500

mL) was treated at 0 °C dropwise with thionyl chloride (91.1 mL,

1.25 mol, 3.5 equiv). After complete addition, the solution was stirred

under reﬂux for 2 h. The reaction mixture was concentrated under

reduced pressure to aﬀord 54 which was used without further

puriﬁcation within the next step. Yield: 65.0 g (100%). ¹H NMR (400

MHz, DMSO-d₆): δ = 9.97 (br s, 1H), 7.31−7.55 (m, 2H), 6.86 (d, J

= 8.1 Hz, 1H), 3.81 (s, 3H), 3.79 (s, 3H). LC−MS (method 2a): t_R

(min) = 0.78. MS (ESI+): m/z = 183 [M + H]⁺.

Methyl 4-Hydroxy-3-methoxy-5-nitrobenzoate (55). Methyl

4-hydroxy-3-methoxybenzoate 54 (320.0 g, 1.70 mol) in conc acetic

acid (1.00 L) was treated at 20 °C (slight ice cooling) dropwise with

65% nitric acid (177 mL, 2.56 mol, 1.5 equiv) [CAUTION: reaction

exhibits a very slow exotherm which can lead to a severe run-away

reaction due to accumulation. Therefore, the reaction temperature

should always be monitored and kept at or below 20 °C]. After

complete addition, the mixture was stirred for another 15 min until

completion. The formed suspension was ﬁltered under reduced

pressure and the solid was washed with water to aﬀord after drying

under reduced pressure 55 which was used without further

puriﬁcation within the next step. Yield: 290 g (73%). ¹H NMR

(400 MHz, DMSO-d₆): δ = 11.43 (br s, 1H), 8.02 (d, J = 2.0 Hz,

1H), 7.66 (d, J = 2.0 Hz, 1H), 3.95 (s, 3H), 3.86 (s, 3H). LC−MS

(method 1a): t_R(min) = 0.83. MS (ESI+): m/z = 226 [M + H]⁺.

Methyl 3-Amino-4-hydroxy-5-methoxybenzoate (56). Meth-

yl 4-hydroxy-3-methoxy-5-nitrobenzoate 55 (50.0 g, 220 mmol) in

methanol (1.00 L) and THF (500 mL) was treated with 10% Pd/C

(4.73 g) and then stirred for 48 h under 1 atm of hydrogen. The

reaction mixture was ﬁltered over Celite under vacuum and the ﬁltrate

was concentrated under reduced pressure to aﬀord 56 which was used

without further puriﬁcation within the next step. Yield: 42.5 g (96%).

¹H NMR (400 MHz, DMSO-d₆): δ = 9.97 (br s, 1H), 7.31−7.55 (m,

2H), 6.86 (d, J = 8.1 Hz, 1H), 3.81 (s, 3H), 3.79 (s, 3H). LC−MS

(method 1a): t_R(min) = 0.46. MS (ESI+): m/z = 197 [M + H]⁺.

Methyl 7-Methoxy-2-thioxo-2,3-dihydro-1,3-benzoxazole-

5-carboxylate (57). Methyl 3-amino-4-hydroxy-5-methoxybenzoate

(20.0 g, 101 mmol) and potassium O-ethyl dithiocarbonate (17.9 g,

112 mmol, 1.1 equiv) were dissolved in pyridine (400 mL), and the

solution was stirred under reﬂux for 3 h (analogously to³⁶). The

reaction mixture was then cooled and poured onto a mixture of ice

(600 g) and concentrated aqueous hydrogen chloride solution (60

mL). The solid formed was ﬁltered oﬀ under reduced pressure and

washed with water (5 × 200 mL). The solid was dried initially at 50

°C/40 mbar and then under high vacuum to aﬀord 57. Yield: 23.3 g

(96%). ¹H NMR (400 MHz, DMSO-d₆): δ = 14.1 (br s, 1H), 7.45 (d,

J = 1.0 Hz, 1H), 7.32 (d, J = 1.2 Hz, 1H), 4.00 (s, 3H), 3.88 (s, 3H).

LC−MS (method 1a): t_R(min) = 0.79. MS (ESI+): m/z = 240 [M +

H]⁺.

Methyl 2-Chloro-7-methoxy-1,3-benzoxazole-5-carboxy-

late (58). 7-Methoxy-2-thioxo-2,3-dihydro-1,3-benzoxazole-5-carbox-

ylate (150 g, 627 mmol) was suspended in thionyl chloride (450 mL),

catalytic amounts of N,N-dimethylformamide (1.0 mL) were added,

and the mixture was then stirred for 3 h (analogously to ref 36). More

N,N-dimethylformamide (1.0 mL) was added, and the mixture was

stirred at 70 °C until the evolution of gas had ceased (about 4 h). The

reaction solution was concentrated under reduced pressure, and the

residue was coevaporated with dichloromethane (3 × 200 mL) to

completely remove the thionyl chloride. The solid was dried under

high vacuum and then puriﬁed by column chromatography on silica

gel (dichloromethane). Alternatively, the crude product can also be

1

used further directly. Yield: 125.6 g (82%). H NMR (400 MHz,

chloroform-d): δ = 7.99 (d, J = 1.3 Hz, 1H), 7.62 (d, J = 1.2 Hz, 1H),

4.07 (s, 3H), 3.96 (s, 3H). LC−MS (method 1a): t_R(min) = 1.00.

MS (ESI+): m/z = 242 [M + H]⁺.

2-Bromo-1-(3-chlorophenyl)ethanone (60). 1-(3-

Chlorophenyl)ethanone (126 mL, 970 mmol) in dichloromethane

(1.10 L) was treated at 20−25 °C dropwise with a solution of

bromine (49.8 mL, 970 mmol, 1.0 equiv) in dichloromethane (300

mL). After stirring for 30 min, ice-cold water (600 mL) was added to

the reaction mixture. After extraction and phase separation, the

organic layer was dried over sodium sulfate, ﬁltered, and concentrated

under reduced pressure. The residue was dissolved in tert-butyl

methyl ether (150 mL) and then treated with n-heptane (1.00 L). The

N

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

solution was cooled under stirring to −10 °C, the precipitated solid

was ﬁltered under reduced pressure and the residue washed with cold

n-heptane (100 mL). The solid was dried initially under high vacuum

to aﬀord 60. Yield: 166 g (73%). ¹H NMR (400 MHz, DMSO-d₆): δ

= 8.03 (t, J = 1.8 Hz, 1H), 7.93−7.98 (m, 1H), 7.76 (ddd, J = 0.9, 2.1,

8.0 Hz, 1H), 7.54−7.64 (m, 1H), 4.99 (s, 2H). LC−MS (method 2a):

t_R(min) = 2.29. MS (ESI+): m/z = 233 [M + H]⁺.

borohydride (110 g, 2.91 mol) was added in four portions, and the

mixture was stirred at room temperature for 36 h. Further sodium

borohydride (29.4 g, 776 mmol) was added, and the mixture was

stirred at room temperature for 1 h. The reaction solution was poured

into 2 M aqueous ammonia solution (4.85 L), and the precipitated

salts were then ﬁltered oﬀ over a frit under reduced pressure. tert-

Butyl methyl ether (14 L) and water (50 L) were added to the ﬁltrate,

the mixture was extracted, and 5% aqueous sodium chloride solution

was then added to facilitate phase separation. After phase separation,

the aqueous phase was re-extracted with tert-butyl methyl ether (5 L)

and the combined organic phases were dried over sodium sulfate,

ﬁltered, and concentrated under reduced pressure. The crude product

was dissolved in diethyl ether/tetrahydrofuran (10:1; 1.1 L), and a 4

N solution of hydrogen chloride in 1,4-dioxane (385 mL) was then

added with stirring and ice cooling. The precipitated white solid was

ﬁltered oﬀ under reduced pressure, washed with a little diethyl ether,

and dried under high vacuum. Yield: 261 g (77%). LC−MS (method

4a): t_R(min) = 1.93. MS (ESI+): m/z = 174 [M + H − HCl]⁺.

rac-tert-Butyl [1-(3-Chlorophenyl)-2-ﬂuoroethyl]carbamate

(63). Method 1. rac-1-(3-Chlorophenyl)-2-ﬂuoroethanamine 62

(7.47 g, 43.3 mmol) was suspended in dichloromethane (150 mL).

First triethylamine (9.14 g, 12.6 mL, 90.4 mmol) and then di-tert-

butyl dicarbonate (10.3 g, 47.3 mmol) were added, and the mixture

was stirred at room temperature overnight. The reaction solution was

then washed with 0.5 N aqueous hydrogen chloride solution (100

mL), saturated aqueous sodium bicarbonate solution (100 mL), and

water (100 mL). The organic phase was dried over sodium sulfate,

ﬁltered, and concentrated under reduced pressure. The crude product

was puriﬁed by preparative RP-HPLC (water/acetonitrile). Yield:

1-(3-Chlorophenyl)-2-ﬂuoroethanone (61). Method 1. Tetra-

n-butylammonium ﬂuoride trihydrate (50.7 g, 161 mmol), zinc

ﬂuoride (22.1 g, 214 mmol), and potassium ﬂuoride (6.22 g, 107

mmol) were initially charged in acetonitrile (850 mL) and stirred at

80 °C for 1 h (analogously to ref 37). 2-Bromo-1-(3-chlorophenyl)-

ethanone 60 (50.0 g, 214 mmol) in acetonitrile (210 mL) was then

added dropwise at this temperature over a period of 3 h, and the

mixture was subsequently stirred at 80 °C for a further 3 h. The

reaction solution was cooled to room temperature, and the

precipitated salts were ﬁltered oﬀ over a glass frit. The ﬁltrate was

concentrated under reduced pressure, water was added to the residue,

and the mixture was extracted repeatedly with tert-butyl methyl ether.

Further precipitated salts were removed by ﬁltration. The organic

phases were dried over magnesium sulfate, ﬁltered, and concentrated

under reduced pressure. The crude product was then puriﬁed by ﬂash

chromatography on silica gel (petroleum ether/ethyl acetate 10:1) to

1

aﬀord 61. Yield: 27.0 g (58%, purity 80%). H NMR (400 MHz,

DMSO-d₆): δ = 7.91 (t, J = 1.7 Hz, 1H), 7.85 (d, J = 7.8 Hz, 1H),

7.73−7.80 (m, 1H), 7.55−7.63 (m, 1H), 5.84 (d, J = 45.9 Hz, 2H).

GC−MS (method 1B): t_R(min) = 3.73. MS (EI+): m/z = 172 [M]⁺.

Method 2. 1-(Chloromethyl)-4-ﬂuoro-1,4-diazoniabicyclo[2.2.2]-

octane bistetraﬂuoroborate (Selectﬂuor) (45.8 g, 129 mmol) was

added to 1-(3-chlorophenyl)ethanone (10.0 g, 64.7 mmol) in

methanol (80.0 mL) and then, in 10 portions, stirred in the

microwave (Biotage Synthesizer) at 110 °C for 2.5 h (analogously to

ref 38). Water (5 mL) was then added to each portion, and the

portions were stirred in the microwave at 110 °C for 1 h. The reaction

mixtures were then combined, the methanol was removed under

reduced pressure, and the residue was diluted with water and then

extracted with ethyl acetate. The organic phase was dried over sodium

sulfate, ﬁltered, and concentrated under reduced pressure. The crude

product was used for the next step without further puriﬁcation. Yield:

12.5 g (86%, purity 77%). GC−MS (method 2b): t_R(min) = 3.56.

MS (EI+): m/z = 172 [M]⁺.

rac-1-(3-Chlorophenyl)-2-ﬂuoroethanamine Hydrochloride

(62). Method 1. Under an atmosphere of argon, titanium

tetraisopropoxide (6.52 g, 6.87 mL, 23.0 mmol) was added to 1-(3-

chlorophenyl)-2-ﬂuoroethanone 61 (2.00 g, 11.5 mmol) in 2 M

ethanolic ammonia solution (28.7 mL, 57.4 mmol), and the mixture

was stirred at room temperature for 16 h. Afterward, sodium

borohydride (654 mg, 17.3 mmol) was added and the mixture was

stirred at room temperature for 5 h. Further sodium borohydride (350

mg, 9.25 mmol) was added, and the mixture was stirred at room

temperature overnight. The reaction solution was poured into 25%

strength aqueous ammonia solution (100 mL) and then ﬁltered

through Celite. tert-Butyl methyl ether (200 mL) was added to the

ﬁltrate, and the mixture was extracted. After phase separation, the

aqueous phase was extracted with tert-butyl methyl ether (100 mL).

The combined organic phases were dried over sodium sulfate, ﬁltered,

and concentrated under reduced pressure. The crude product was

dissolved in diethyl ether/tetrahydrofuran (5:1; 60 mL), and a 4 N

solution of hydrogen chloride in 1,4-dioxane (10.0 mL) was then

added. The solid formed was ﬁltered oﬀ under reduced pressure,

washed with a little diethyl ether, and dried under high vacuum. Yield:

1.54 g (63%). ¹H NMR (400 MHz, DMSO-d₆): δ = 8.98 (br s, 3H),

7.70 (s, 1H), 7.60−7.47 (m, 3H), 4.88−4.67 (m, 3H). LC−MS

(method 4a): t_R(min) = 1.94. MS (ESI+): m/z = 174 [M + H −

HCl]⁺.

1

7.23 g (61%). H NMR (400 MHz, DMSO-d₆): δ = 7.72 (br d, J =

8.6 Hz, 1H), 7.25−7.51 (m, 4H), 4.76−5.00 (m, 1H), 4.45−4.61 (m,

1H), 4.30−4.44 (m, 1H), 1.38 (s, 9H). LC−MS (method 1a): t_R

(min) = 1.10. MS (ESI+): m/z = 218 [M + H − C₄H₉]⁺.

Method 2. Under an atmosphere of argon, rac-1-(3-chlorophenyl)-

2-ﬂuoroethanamine 62 (41.5 g, 198 mmol) was suspended in

dichloromethane (200 mL), and subsequently ﬁrst triethylamine

(80.0 g, 110 mL, 790 mmol) and then dichloromethane (200 mL)

were added. Afterward, di-tert-butyl dicarbonate (31.0 g, 142 mmol)

in dichloromethane (100 mL) was added, and the mixture was stirred

at room temperature overnight. Further di-tert-butyl dicarbonate

(9.91 g, 45.4 mmol) was added, and the mixture was stirred until

conversion was almost complete (monitored by TLC). The reaction

solution was washed with 1 N aqueous hydrogen chloride solution (2

× 100 mL) and saturated aqueous sodium bicarbonate solution (100

mL). The organic phase was dried over sodium sulfate, ﬁltered, and

concentrated under reduced pressure. Yield: 51.0 g (94%). LC−MS

(method 5a): t_R(min) = 3.25. MS (ESI+): m/z = 218 [M + H −

C₄H₉]⁺.

tert-Butyl [(1S)-1-(3-Chlorophenyl)-2-ﬂuoroethyl]carbamate

(64). Enantiomer separation on a chiral phase of 7.23 g of the

compound (rac)-63 according to method 2d gave 3.05 g of

enantiomerically pure (R)-isomer and 3.05 g of enantiomerically

1

pure (S)-isomer (64). (R)-isomer: H NMR (400 MHz, DMSO-d₆):

δ = 7.72 (br d, J = 8.7 Hz, 1H), 7.45 (s, 1H), 7.26−7.41 (m, 2H),

7.26−7.30 (m, 1H), 4.81−4.97 (m, 1H), 4.45−4.57 (m, 1H), 4.32−

4.44 (m, 1H), 1.19−1.46 (m, 9H). HPLC (method 1e): t_R(min) =

5.01, 99.0% ee. LC−MS (method 5a): t_R(min) = 3.26. MS (ESI+):

m/z = 218 [M + H − C₄H₉]⁺. (S)-isomer (64): ¹H NMR (400 MHz,

DMSO-d₆): δ = 7.72 (br d, J = 8.56 Hz, 1H), 7.45 (s, 1H), 7.30−7.40

(m, 3H), 4.80−4.98 (m, 1H), 4.45−4.58 (m, 1H), 4.32−4.44 (m,

1H), 1.20−1.44 (m, 9H). HPLC (method 1e): t_R(min) = 7.46,

99.0% ee. LC−MS (method 7A): t_R(min) = 3.26. MS (ESI+): m/z =

218 [M + H − C₄H₉]⁺.

(1S)-1-(3-Chlorophenyl)-2-ﬂuoroethanamine Hydrochloride

(65). (S)-tert-Butyl [1-(3-chlorophenyl)-2-ﬂuoroethyl]carbamate

(64) (17.3 g, 63.2 mmol) was initially charged in 1,4-dioxane (50

mL), and 4 N hydrogen chloride solution in 1,4-dioxane (79 mL, 316

mmol) was then added at room temperature. A solid formed after a

short period of time. 1,4-Dioxane (250 mL) and then 4 N hydrogen

chloride solution in 1,4-dioxane (31.6 mL, 126 mmol) were added,

Method 2. Titanium tetraisopropoxide (1.10 kg, 1.16 L, 3.88 mol)

was added dropwise to 1-(3-chlorophenyl)-2-ﬂuoroethanone 61 (335

g, 1.94 mol) in 2 M ethanolic ammonia solution (4.85 L, 9.71 mol)

(the temperature was kept at 20 °C by ice cooling), and the mixture

was stirred at room temperature overnight. At 10 °C, sodium

O

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

and the mixture was stirred at room temperature overnight. The

suspension formed was concentrated completely under reduced

pressure, the residue was triturated with tert-butyl methyl ether (200

mL), ﬁltered, and the ﬁlter residue was washed with tert-butyl methyl

ether (2 × 50 mL). The solid formed was dried under high vacuum.

Yield: 13.2 g (99%). ¹H NMR (400 MHz, DMSO-d₆): δ = 8.91 (br s,

3H), 7.68 (s, 1H), 7.39−7.59 (m, 3H), 4.64−4.93 (m, 3H). Optical

rotation: = 27.06° (c = 0.51, methanol). LC−MS (method 4a): t_R

(min) = 1.94. MS (ESI+): m/z = 174 [M + H]⁺.

Dimethyl ^L-Aspartate Hydrochloride (67). L-Aspartic acid

(1.00 kg, 7.51 mol) in MeOH (6.00 L) was treated dropwise with

thionyl chloride (1.20 L, 16.5 mol) at 0 °C (addition time: 1 h).

Afterward, the reaction mixture was stirred for 7 h under reﬂux. The

reaction mixture was then concentrated under reduced pressure, and

the crude product was used for the next step without further

puriﬁcation. Yield: 1.50 kg (100%). ¹H NMR (400 MHz, DMSO-d₆):

δ = 8.40−8.82 (m, 3H), 4.17−4.43 (m, 1H), 3.74 (s, 3H), 3.66 (s,

3H), 2.89−3.10 (m, 2H). GC−MS (method 1b): t_R(min) = 3.25. MS

(method 1c): m/z = 162 [M + H]⁺.

about 1:1). LC−MS (method 1a): t_R(min) = 0.65 (diastereomer 1),

t_R(min) = 0.67 (diastereomer 2). MS (ESI+): m/z = 286 [M + H]⁺.

(5S)-4-Benzyl-5-(2-hydroxyethyl)-2-methylmorpholin-3-one

(71). N-Benzyl-2-chloro-N-[(2S)-1,4-dihydroxybutan-2-yl]-

propanamide 71 (71.8 g, 206 mmol, purity 82%) was initially

charged in isopropanol (1.30 L), and the mixture was cooled to 0 °C.

Afterward, potassium tert-butoxide (92.4 g, 824 mmol) was added in

one portion, and the mixture was stirred at 0 °C for 30 min. The

reaction solution was allowed to warm to room temperature, and the

isopropanol was removed under reduced pressure. The residue was

taken up in ethyl acetate (500 mL). Water (600 mL) was added, the

mixture was extracted, and after phase separation, the aqueous phase

was extracted with ethyl acetate (2 × 300 mL). The organic phases

were dried over sodium sulfate, ﬁltered, and concentrated under

reduced pressure. The crude product was used for the next step

without further puriﬁcation. Yield: 58.6 g (quant, purity 90%,

diastereomer ratio about 3:2). LC−MS (method 2a): t_R(min) =

1.51 (diastereomer 1), t_R(min) = 1.53 (diastereomer 2). MS (ESI+):

m/z = 250 [M + H]⁺.

2-[(3S,6S)-4-Benzyl-6-methylmorpholin-3-yl]ethanol (72).

(5S)-4-Benzyl-5-(2-hydroxyethyl)-2-methylmorpholin-3-one 72

(30.0 g, 108 mmol) was initially charged in tetrahydrofuran (1.10

L), 2 M borane−dimethyl sulﬁde complex solution in tetrahydrofuran

(217 mL, 433 mmol) was added under argon, and the mixture was

stirred under reﬂux for 2 h. The mixture was subsequently cooled to 0

°C, methanol (200 mL) was added carefully, and the mixture was

stirred under reﬂux for 30 min. The mixture was subsequently

concentrated completely under reduced pressure, and the residue was

taken up in acetonitrile and subjected to puriﬁcation and diastereomer

separation by preparative RP-HPLC (acetonitrile/water, isocratic).

Here, the target compound eluted as second component, (3S,6S)-

diastereoisomer 2. Yield: 12.1 g (47%); (3S,6R)-diastereoisomer 1.

Yield: 6.23 g (24%).

Dimethyl N-Benzyl-^L-aspartate (68). Dimethyl L-aspartate

hydrochloride 67 (375 g, 1.90 mol) in dichloromethane (4.80 L)

was treated at room temperature with benzaldehyde (183 mL, 1.80

mol, 0.95 equiv) and triethylamine (264 mL, 1.90 mol, 1.00 equiv)

and stirred for 2 h. Afterward, sodium triacetoxyborohydride (482 g,

2.27 mol, 1.20 equiv) was added at 0 °C and the reaction mixture was

stirred at room temperature overnight. The reaction mixture was

washed with 2 N aqueous hydrogen chloride solution (3 × 1.0 L), and

then the aqueous phase was adjusted to pH = 8−9 with 50% aqueous

sodium hydroxide solution. The aqueous phase was extracted with

ethyl acetate (2 × 1.0 L), and the organic phase was dried over

sodium sulfate, ﬁltered, and concentrated under reduced pressure.

The crude product was used for the next step without further

1

puriﬁcation. Yield: 288 g (61%). H NMR (400 MHz, DMSO-d₆): δ

1

(3S,6S)-Diastereoisomer 2. H NMR (400 MHz, DMSO-d₆): = δ

= 7.14−7.36 (m, 5H), 3.73−3.82 (m, 1H), 3.63 (s, 3H), 3.47−3.60

(m, 4H), 3.27−3.38 (m, 1H), 2.66−2.75 (m, 1H), 2.56−2.64 (m,

1H). LC−MS (method 1a): t_R(min) = 0.45. MS (ESI+): m/z = 251

[M + H]⁺.

7.13−7.39 (m, 5H), 4.42 (t, J = 5.1 Hz, 1H), 3.36−3.72 (m, 7H),

2.56−2.64 (m, 1H), 2.27−2.36 (m, 1H), 2.14−2.26 (m, 1H), 1.60−

1.84 (m, 2H), 1.00 (d, J = 6.11 Hz, 3H). LC−MS (method 3a): t_R

(min) = 2.33 min. MS (ESI+): m/z = 236 [M + H]⁺.

(2S)-2-(Benzylamino)butane-1,4-diol (69). Dimethyl N-benzyl-

L-aspartate 68 (69.0 g, 275 mmol) in tetrahydrofuran (1.00 L) was

treated at 0 °C with 2.4 N lithium aluminum hydride in

tetrahydrofuran (330 mL, 792 mmol, 2.88 equiv). After complete

addition, the reaction mixture was stirred for 2 h under reﬂux.

Afterward, the reaction mixture was quenched dropwise at 0 °C with

water (30 mL), then with 15% aqueous sodium hydroxide solution

(30 mL), and ﬁnally with water (90 mL). After 15 min, sodium sulfate

(100 g) was added and the reaction mixture was stirred for another 15

min. The reaction mixture was ﬁltered under reduced pressure, and

the residue was washed with tetrahydrofuran (250 mL) and hot

isopropanol (500 mL). The combined ﬁltrates were concentrated

under reduced pressure, and the crude product was used for the next

1

(3S,6R)-Diastereoisomer 1. H NMR (400 MHz, DMSO-d₆): δ =

7.07−7.50 (m, 5H), 4.49 (t, J = 5.0 Hz, 1H), 4.10 (d, J = 13.69 Hz,

1H), 3.76 (dd, J = 3.4, 11.3 Hz, 1H), 3.37−3.57 (m, 3H), 3.21−3.33

(m, 1H), 2.95 (d, J = 13.5 Hz, 1H), 2.27 (dt, J = 3.55, 7.03 Hz, 1H),

1.80 (dtd, J = 2.6, 7.2, 14.3 Hz, 1H), 1.68 (dd, J = 10.4, 11.4 Hz, 1H),

1.48 (qd, J = 7.0, 13.9 Hz, 1H), 0.94 (d, J = 6.4 Hz, 3H). LC−MS

(method 3a): t_R(min) = 2.23. MS (ESI+): m/z = 236 [M + H]⁺.

2-[(3S,6S)-6-Methylmorpholin-3-yl]ethanol (73). 2-[(3S,6S)-

4-Benzyl-6-methylmorpholin-3-yl]ethanol 72 (58.0 g, 246 mmol) was

initially charged in ethanol (1.50 L), 10% palladium on carbon (2.90

g) and 20% palladium hydroxide on carbon (2.90 g) were added

under argon atmosphere, and the mixture was then stirred under an

atmosphere of hydrogen at standard pressure overnight. The reaction

solution was ﬁltered through Celite, and the ﬁlter residue was washed

with hot ethanol (100 mL). The ﬁltrate was concentrated under

reduced pressure and the product was dried under high vacuum.

1

step without further puriﬁcation. H NMR (400 MHz, DMSO-d₆): δ

= 7.10−7.37 (m, 5H), 4.50 (br s, 1H), 3.72 (d, J = 1.5 Hz, 2H), 3.50

(dt, J = 2.57, 6.4 Hz, 2H), 3.37−3.44 (m, 1H), 3.34 (s, 3H), 2.57−

2.65 (m, 1H), 1.43−1.60 (m, 2H). LC−MS (method 4a): t_R(min) =

1.82. MS (ESI+): m/z = 195 [M + H]⁺.

1

Yield: 35.5 g (99%). H NMR (400 MHz, chloroform-d): δ = 3.80−

3.90 (m, 2H), 3.67−3.76 (m, 1H), 3.59−3.66 (m, 1H), 3.42−3.56

(m, 1H), 3.12 (br s, 2H), 2.89−3.00 (m, 1H), 2.75−2.87 (m, 1H),

2.61−2.73 (m, 1H), 2.23−2.41 (m, 1H), 1.45 (dd, J = 3.5, 14.6 Hz,

1H), 1.15 (d, J = 6.3 Hz, 3H). Optical rotation: 89.7° (c = 0.565,

chloroform). LC−MS (method 4a): t_R(min) = 0.54. MS (ESI+): m/z

= 146 [M + H]⁺. LC−MS (method 1a): MS (ESI+): m/z = 146 [M +

H]⁺.

N-Benzyl-2-chloro-N-[(2S)-1,4-dihydroxybutan-2-yl]-

propanamide (70). (2S)-2-(Benzyl-amino)butane-1,4-diol 69 (45.1

g, 199 mmol, purity 86%) was initially charged in isopropanol (1.00

L), the mixture was cooled to 0 °C, and triethylamine (40.2 g, 55.4

mL, 397 mmol) was added. Afterward, (rac)-2-chloropropionyl

chloride (37.8 g, 29.6 mL, 298 mmol) was added dropwise. After

30 min of stirring, further (rac)-2-chloropropionyl chloride (18.9 g,

14.8 mL, 149 mmol) was added dropwise, and the reaction solution

was allowed to warm to room temperature and then concentrated

under reduced pressure. The residue was taken up in ethyl acetate

(1.00 L) and washed with water. The organic phase was dried over

sodium sulfate, ﬁltered, and concentrated under reduced pressure.

The crude product was used for the next step without further

puriﬁcation. Yield: 71.8 g (quant, purity 82%, diastereomer ratio

Methyl 2-{[(1S)-1-(3-Chlorophenyl)-2-ﬂuoroethyl]amino}-7-

methoxy-1,3-benzoxazole-5-carboxylate (74). Under argon

atmosphere, methyl 2-chloro-7-methoxy-1,3-benzoxazole-5-carboxy-

late 58 (125 g, 517 mmol) was initially charged in N,N-

dimethylformamide (850 mL), and (1S)-1-(3-chlorophenyl)-2-

ﬂuoroethanamine hydrochloride 65 (114 g, 543 mmol) and N,N-

diisopropylethylamine (360 mL, 2.07 mol) were added at room

temperature. The reaction solution was stirred at 70 °C (oil bath

P

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

temperature) for 4 h, (1S)-1-(3-chlorophenyl)-2-ﬂuoroethanamine

hydrochloride 65 (8.69 g, 41.4 mmol) was then added, and the

mixture was stirred at room temperature overnight (about 14 h). A

further (1S)-1-(3-chlorophenyl)-2-ﬂuoroethanamine hydrochloride

65 (1.09 g, 5.17 mmol) was added, and the mixture was stirred at

70 °C (oil bath temperature) for 2 h and then concentrated under

reduced pressure. The residue was taken up in tert-butyl methyl ether

(2.0 L), and the organic phase was washed with water (3 × 1.0 L).

Under reduced pressure, the organic phase was concentrated to about

one-third of its original volume, and the precipitated solid was then

ﬁltered oﬀ under reduced pressure. The solid precipitated in the

aqueous phase was likewise ﬁltered oﬀ under reduced pressure and

then washed with tert-butyl methyl ether (2 × 100 mL). The

combined solids were dried under high vacuum. Yield: 181 g (92%).

¹H NMR (400 MHz, DMSO-d₆) δ = 9.10 (d, J = 8.7 Hz, 1H), 7.59 (s,

were diluted with reservoir solution to a ﬁnal concentration of 5 mM

inhibitor. Crystals of thrombin were soaked overnight in this solution

at 10 °C.

Thrombin Data Collection, Processing, and Reﬁnement. For

data collection, a crystal was plunged into a solution containing 15%

glycerol and was mounted at 100 K on a Rigaku 007 diﬀractometer

equipped with a MAR image plate detector. Only 51 (BAY1217224)

was mounted at 100 K on a Bruker Proteum 235 system equipped

with a CCD detector. All data were processed using MOSFLM³⁹and

SCALA⁴⁰from the CCP4 suite.⁴¹Compound 51 (BAY1217224) was

processed using SAINT and SADABS from the Proteum suite (Bruker

AXS). Structure solution was performed using PHASER⁴²with a

known in-house structure. Reﬁnement was performed using

REFMAC5.⁴³The ligands were generated using PRODRG⁴⁴and

docked into the electron density within COOT.⁴⁵Data collection and

reﬁnement statistics for all data sets are summarized in Table S5. The

absolute structures were determined without any doubt in all complex

X-ray structures.

1H), 7.36−7.52 (m, 4H), 7.31 (d, J = 1.1 Hz, 1H), 5.16−5.37 (m,

1H), 4.50−4.82 (m, 2H), 3.96 (s, 3H), 3.84 (s, 3H). Optical rotation:

77.20° (c = 0.465, methanol). LC−MS (method 4a): t_R(min) = 2.57.

MS (ESI+): m/z = 379 [M + H]⁺.

PXR Expression, Puriﬁcation, and Cocrystallization. An

expression construct identical to the one reported by Wang et al.,⁴⁶

coding for a peptide of the nuclear receptor coactivator NCOA1

(residues 678−700) fused via a GGSGG linker to the C-terminus of

the ligand binding domain of the PXR (residues 129−434), was

expressed in E. coli, puriﬁed via aﬃnity chromatography (HisTrap, 5

mL), ion-exchange chromatography (Mono S HR 10/10), and gel

ﬁltration [Superdex 75 26/60, 20 mM Tris buﬀer, pH 7.8, 250 mM

NaCl, 2.5 mM EDTA, 5% (v/v) glycerol, 5 mM DTT], concentrated

to 12.1 mg/mL, and shock-frozen in liquid nitrogen. For

cocrystallization with compound 17, the PXR was thawed,

supplemented with 25 mM compound (from a 500 mM stock

solution in DMSO), and incubated at 277 K for 24 h. Crystals were

formed using the vapor diﬀusion method in hanging drops consisting

of 1 μL of protein−ligand solution and 1 μL of reservoir solution [100

mM imidazole, pH 8.0, 20% (v/v) MPD (2-methyl-2,4-pentanediol)].

PXR Data Collection, Processing, and Reﬁnement. For data

collection, crystals were brieﬂy immersed in cryo buﬀer [100 mM

imidazole, pH 8.0, 30% (v/v) MPD] and shock-frozen in liquid

nitrogen. A diﬀraction data set to 1.85 Å resolution was collected on

beamline BL14.1 operated by the Helmholtz-Zentrum Berlin (HZB)

at the BESSY II electron storage ring (Berlin-Adlershof, Germany)⁴⁷

and processed using XDS⁴⁸via the graphical user interface

XDSAPP.⁴⁹The structure was solved by rigid body reﬁnement with

REFMAC5⁴³with PDB entry 3HVL as starting model. The initial

model was rebuilt in COOT⁴⁵and reﬁned using REFMAC5.

Diﬀerence electron density indicated binding of the ligand only in

one of the two PXR molecules in this crystal form (chain B). The

ligand parameters were calculated using PRODRG,⁴⁴and the ligand

was ﬁtted to the electron density using COOT. The ﬁnal ligand

occupancy was adjusted to 0.70. Density in the ligand binding pocket

of chain A was modeled as MPD (artifact from the crystallization

reservoir solution) which may also be present as a minor component

in chain B (here not modeled). Data collection and reﬁnement

statistics are summarized in Table S5.

PXR Transactivation Assay. To evaluate PXR nuclear receptor

transactivation, a screening luciferase reporter gene assay in human

HepG2 cells was applied. HepG2 cells were seeded in a 384-well plate

and cultivated at 37 °C in humidiﬁed air with 5% CO₂. 48 h prior to

readout, cells were transiently co-transfected with a vector for human

PXR and a luciferase reporter gene under the control of the human

CYP3A4 promotor with FuGENE HD transfection reagent (Promega,

Madison, WI, USA) according to the manufacturer’s instructions. 24

h after seeding, the cells were treated with test compound (2 nM to

50 μM, 0.5% DMSO) prepared by a 1:3 serial dilution. Rifampicin

was incubated in the same manner, as a positive control. In addition,

for normalization of the luminescence signal, cells were incubated

with rifampicin at a concentration of 50 or 16.7 μM corresponding to

100% activation, as well as with DMSO for background luminescence

corresponding to 0% activation (n = 32 wells each). After treatment,

cells were lysed and incubated with the luciferase substrate ONE-Glo

reagent (Promega, Madison, WI, USA) according to the manufac-

2-{[{[(1S)-1-(3-Chlorophenyl)-2-ﬂuoroethyl]amino}-7-me-

thoxy-1,3-benzoxazole-5-carboxylic Acid (75). Methyl 2-{[(1S)-

1-(3-chlorophenyl)-2-ﬂuoroethyl]amino}-7-methoxy-1,3-benzoxa-

zole-5-carboxylate 74 (181 g, 478 mmol) was initially charged in 1,4-

dioxane (2.4 L), and a cold (about 8 °C) solution of 45% strength

aqueous sodium hydroxide solution (287 mL, 4.78 mol) in water

(2.35 L) was then added. The mixture was stirred at room

temperature for 4 h and then diluted with water (2.0 L). The

reaction solution was extracted with tert-butyl methyl ether (2 × 1.0

L), ice was added to the aqueous phase, and the aqueous phase was

acidiﬁed with concentrated hydrogen chloride solution (about 450

mL). The mixture was stirred at 10 °C for 15 min, the precipitated

solid was ﬁltered oﬀ under reduced pressure, and the residue was

washed with water (2 × 1.0 L), left at room temperature overnight,

and then dried at 60 °C for 4 h and then at 50 °C under reduced

pressure until the mass remained constant. Yield: 172 g (98%, purity

89%). ¹H NMR (400 MHz, DMSO-d₆): δ = 12.9 (br s, 1H), 9.06 (d,

J = 8.8 Hz, 1H), 7.59 (s, 1H), 7.35−7.51 (m, 4H), 7.31 (d, J = 1.3 Hz,

1H), 5.18−5.33 (m, 1H), 4.51−4.82 (m, 2H), 3.95 (s, 3H). LC−MS

(method 1a): t_R(min) = 0.92. MS (ESI+): m/z = 365 [M + H]⁺.

(2-{[(1S)-1-(3-Chlorophenyl)-2-ﬂuoroethyl]amino}-7-me-

thoxy-1,3-benzoxazol-5-yl)[(2S,5S)-5-(2-hydroxyethyl)-2-

methylmorpholin-4-yl]methanone (51). 2-{[-{[(1S)-1-(3-Chlor-

ophenyl)-2-ﬂuoroethyl]amino}-7-methoxy-1,3-benzoxazole-5-carbox-

ylic acid 75 (500 mg, 1.23 mmol, purity 89%) and 2-[(3S,6S)-6-

methylmorpholin-3-yl]ethanol 73 (232 mg, 1.60 mmol) were initially

charged in N,N-dimethylformamide (20.0 mL), and N,N-diisopropy-

lethylamine (477 mg, 643 μL, 3.69 mmol) was added. Afterward,

HATU (654 mg, 1.72 mmol) was added at room temperature, and

the mixture was stirred for 14 h and then puriﬁed without any further

workup by preparative RP-HPLC (acetonitrile/water). Traces of the

minor isomer were removed by SFC on a chiral phase according to

1

method 1d from the product obtained. Yield: 423 mg (70%). H

NMR (400 MHz, DMSO-d₆): δ = 8.98 (br s, 1H), 7.58 (s, 1H),

7.32−7.51 (m, 3H), 6.85 (br d, J = 7.21 Hz, 1H), 6.70 (br s, 1H),

5.16−5.32 (m, 1H), 4.13−4.82 (m, 4H), 3.92 (s, 3H), 3.20−3.83 (m,

10H), 2.59−3.08 (m, 1H), 1.71−2.04 (m, 2H), 0.87−1.23 (m, 3H).

Optical rotation: 63.79° (c = 0.625, chloroform). HPLC (method 2e):

t_R(min) = 14.0, 99.9% de. LC−MS (method 1a): t_R(min) = 0.93.

MS (ESI+): m/z = 492 [M + H]⁺.

Thrombin Crystallization and Inhibitor Complex Forma-

tion. α-Thrombin (Haemochrom Diagnostica, Germany) was

dissolved in 20 mM phosphate buﬀer, pH 7.5, 350 mM NaCl, and

2 mM benzamidine to a ﬁnal concentration of 10 mg/mL. A 4-fold

excess of hirudin (Bachem) was added, and the mixture was incubated

for 2 h on ice. Crystallization was performed using vapor diﬀusion in

sitting drops, using equal volumes of protein solution and reservoir

solution (0.02 M phosphate buﬀer, pH 7.5, 27% PEG 8000, 100 mM

NaCl). Thrombin seeds were added to the ﬁnal drop. Crystals with a

size of ∼0.2−0.4 mm grew overnight at 10 °C. All thrombin inhibitors

were dissolved to give 50 mM DMSO stock solutions. DMSO stocks

Q

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

turer’s instructions, and the luminescence signal was detected using a

plate reader. The assay was validated with a standard set of inducers

(e.g., bosentan, carbamazepine) and noninducers (e.g., penicillin V,

ﬂuconazole). From the dose−response curves, a concentration-

dependent increase in the luciferase activity above 4-fold standard

deviation of the DMSO background was classiﬁed as PXR

transactivation and is represented as the minimum eﬃcacious

concentration (MEC).

Caco-2 Permeability Assay. Caco-2 cells (DSMZ, German

Resource Centre for Biological Material, Braunschweig, Germany)

were seeded at a density of 4 × 10⁴cells/well on 24-well

polycarbonate insert plates, 0.4 μm pore size (Corning Life Sciences,

Lowell, MA), and maintained for 15 d in Dulbecco’s modiﬁed Eagle

medium supplemented with 10% fetal BSA, 20 mM glucose, 10 mg/

500 mL streptomycin, and 10000 IU/500 mL penicillin in a

humidiﬁed incubator at 8% CO₂. Before the assay was run, the

culture medium was replaced by transport buﬀer consisting of Hanks’

balanced salt solution (pH 7.4) supplemented with 10 mM HEPES

and 20 mM glucose. Stock solutions of test compounds were prepared

in DMSO and diluted 100-fold with transport buﬀer. These solutions

were applied to either the apical or the basolateral compartment, and

1% DMSO was added to the buﬀer in the trans-compartment. After

an incubation of 2 h at 37 °C, samples were taken from both

compartments and analyzed by LCMS/MS. Transepithelial electrical

resistance (TEER) was measured in Hanks’ balanced salt solution

before and after the transport studies using an STX 100 TEER

electrode (World Precision Instruments, Berlin, Germany). To ensure

the integrity of the cell monolayer, TEER values were determined

before and after an experiment. The acceptance criterion for

acceptable batches of cell monolayers was a TEER value of 600 Ω.

Human Hepatocyte Assay for CYP3A4 Induction. Human

hepatocytes were seeded at a density of ∼40 000 cells/96 wells in a

collagen sandwich and cultured for 1 d before compound treatment.

Cells were treated with a 1:3 serial dilution of eight concentrations for

2 consecutive days with media change every day. After compound

treatment for 48 h, cells were lysed and mRNA was prepared by a

state-of-the-art magnetic beads technique. In brief, approximately 24 h

after the last treatment, hepatocytes were harvested for mRNA

isolation. Thus, cell culture medium was removed from each well, cells

were washed with 150 μL of supplement-free cell culture medium

prior to cell lysis, cells were lysed with 100 μL of lysis buﬀer

containing proteinase K (50 μg/μL, ﬁnal concentration in well: 30

ng/μL), and ﬁnal cell lysates were stored at −80 °C. mRNA was

isolated using the Dynabeads mRNA Direct Kit (Life Technologies,

Germany). 150 μL of each cell lysate was mixed with 100 μg magnetic

beads and incubated for a few minutes; then, the supernatant was

removed and the beads were washed twice with washing buﬀer A and

twice with washing buﬀer B with 200 and 100 μL per well,

respectively. Single-stranded cDNA was prepared from mRNA with

the high capacity RNA-to-cDNA kit (Life Technologies). 20 μL RT

Master Mix [comprised of 10× RT buﬀer, 25× deoxyNTPs, 10×

Random Hexamers, RNase inhibitor (20 U/μL), MultiScribe reverse

transcriptase (50 U/μL), and RNase-free water] was added to each

well and transcribed for 60 min at 37 °C using the Gene Amp PCR

P

_appwas calculated as (V_r/P₀)·(1/S)·(P₂/t), where V_ris the volume of

medium in the receiver chamber, P₀is the measured peak height or

peak area of the test drug in the donor chamber at t₀, S is the surface

area of the monolayer, P₂is the measured peak height or peak area of

the test drug in the acceptor chamber after an incubation of 2 h, and t

is the incubation time (2 h). The eﬄux ratio was calculated as

P

_app(B−A)/P_app(A−B), where P_app(B−A) and P_app(A−B) represent

the apparent permeability of test compound from the basolateral to

apical and apical to basolateral side of the cellular monolayer,

respectively. Cell data are expressed as arithmetic means and SD. The

SD for the eﬄux ratio was calculated using linear error propagation.

In Vivo Pharmacokinetics in Rats. Animal studies were

conducted in accordance with the German Animal Protection Act

(Deutsches Tierschutzgesetz). For the determination of pharmacoki-

netic parameters, male Wistar rats strain Hannover (Harlan

Winkelmann, Borchen, Germany) were catheterized under isoﬂurane

(Delta Select, Dreieich, Germany) anesthesia. One day after

implantation of the catheter, compounds were administered either

intravenously by direct injection into the lateral tail vein at doses of

0.3 to 0.5 mg/kg, 2 mL/kg in plasma containing 2% DMSO, or orally

at a dose of 1 mg/kg, 5 mL/kg (10% ethanol/40% PEG/50%

demineralized water, v/v/v) by gavage using a calibrated glass syringe.

Blood samples were collected at 0.033, 0.083, 0.25, 0.5, 1, 2, 3, 5, 7,

and 24 h postdosing for iv and 0.083, 0.25, 0.5, 1, 2, 3, 5, 7, and 24 h

postdosing for po, via the catheter. After centrifugation, plasma was

precipitated by the addition of acetonitrile.

̈

System 9700 thermocycling program (Biometra, Gottingen, Ger-

many). The prepared cDNA samples were stored at −80 °C prior to

analysis by quantitative RT-PCR, which was carried out on a

QuantStudio7 Flex PCR system (Applied Biosystems) according to

the manufacturer’s protocol. A primer mix was prepared for each gene

expression assay. A typical primer mix contained TaqMan Fast

Advanced Master Mix (1×), gene expression assay (1×, 900 nM

forward and reverse primers), and RNase-free water and was added to

the cDNA. The relative quantity of the target cDNA compared to that

of the control cDNA (actin, tubulin) was determined by the ΔΔCt

method. Relative quantiﬁcation measures the change in mRNA

expression in a test sample relative to that in a control sample (e.g.,

vehicle treated). CYP induction was calculated based on the ΔΔCt

method and expressed as fold induction over the vehicle-treated

control.

In Vitro Clearance Determinations with Rat Hepatocytes.

Incubations with rat hepatocytes (freshly prepared from male Whistar

rats) were performed at 37 °C, pH 7.4, in a total volume of 1.5 mL

using a modiﬁed Janus robotic system (PerkinElmer). The incubation

mixtures contained 1 × 10⁶cells/mL (corrected, according to viability

of the cells, determined via microscopy after staining with trypan

blue), 1 μM substrate, and Williams’ medium E (Sigma, product no.

W1878). The ﬁnal acetonitrile concentration was <1%. Aliquots of

125 μL were withdrawn from the incubation mixture after 2, 10, 20,

30, 50, 70, and 90 min and dispensed in a 96-well plate containing

acetonitrile (250 μL) to stop the reaction. After centrifugation,

supernatants were analyzed by LC−MS/MS (AB Sciex Triple Quad

5500). The calculation of in vitro clearance values from half-life data

using rat hepatocytes, reﬂecting substrate depletion, was performed

using the following equation: CL′_intrinsic[mL/(min·kg)] = (0.693/in

vitro t_1/2[min])·(liver weight [g liver/kg body mass])·(cell no. [1.1 ×

10⁸]/liver weight [g])/(cell no. [1 × 10⁶]/incubation volume [mL]).

The CL_bloodwas estimated using the nonrestricted well-stirred model:

CL_bloodwell-stirred [L/(h·kg)] = (Q_H[L/(h·kg)]·CL′_intrinsic[L/(h·

kg)])/(Q_H[L/(h·kg)] + CL′_intrinsic[L/(h·kg)]). For calculations, the

following values were used: rat speciﬁc liver weight of 32 g/kg body

mass, hepatic blood ﬂow of 4.2 L/(h·kg); the cell number in the liver

was estimated to be 1.1 × 10⁸cells/g liver.

The supernatants were subjected to HPLC performed on an

Agilent 1200 liquid chromatography system (Agilent Technologies,

Waldbronn, Germany). The mobile phase consisted of 10 mM

ammonium acetate (pH 3.0) and acetonitrile, and a linear gradient

from 20% to 90% acetonitrile (v/v) within 2 min was applied.

Tandem mass spectrometry was performed on an API 3000, 4000, or

5500 triple quadrupole mass spectrometer (Applied Biosystems,

Darmstadt, Germany) connected to the HPLC system through a

TurboIonSpray interface. Pharmacokinetic parameters were calculated

from plasma concentrations with Kinex, a validated in-house (Bayer

AG, Wuppertal, Germany) calculation program.

In all sets of experiments, blood samples were collected in

heparinized syringes following exsanguination by a cut through the

carotid artery under deep isoﬂurane anesthesia (approximately 2.5%

v/v). Blood cells were separated from plasma by centrifugation, and

the plasma was stored for further analyses.

Biochemical Assays for Determination of Thrombin

Inhibition and Selectivity against Other Serine Proteases.

Inhibitory potency and/or selectivity of test compounds were

determined. The assays are based on the ﬂuorescent detection of

aminomethylcoumarine (AMC), released from the ﬂuorogenic

peptidic protease substrates upon protease catalyzed cleavage.

R

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

The active proteases or zymogenes, typically puriﬁed from human

plasma, and corresponding substrates are commercially available. All

enzymes and substrates are diluted in assay buﬀer (50 mM Tris/HCl,

pH 7.4, 100 mM NaCl, 5 mM CaCl₂, 0.1% BSA).

°C, and the bleeding from the cut was observed for 15 min. The time

until the bleeding ceased for at least 30 s was determined.

Rat Venous Stasis Model. Thrombus formation was induced in

anesthetized rats (n = 6−10 per dose group) as described previously,

with minor modiﬁcations. The abdominal vena cava was exposed, and

two loose sutures (8−10 mm apart) were placed below the left renal

venous branch. Compound 51 dissolved in PEG/water/glycerol (996

g/100 g/60 g), or vehicle, was given by oral gavage (po)

administration 90 min before thrombus induction. Thromboplastin

(0.5 mg/kg) was injected into a femoral vein, and after 15 s, the

proximal and distal sutures were tied; 15 min later, the ligated

segment was removed, and the thrombus was withdrawn and weighed.

Serine protease assays comprise of the following enzymes and

substrates. The ﬁnal assay concentrations are given: thrombin

(Kordia; 0.02 nM), Boc-Asp(OBzl)-Pro-Arg-AMC (Bachem I-1560;

5 μM); factor Xa (Kordia; 1.3 nM), Boc-Ile-Glu-Gly-Arg-AMC

(Bachem I-1100; 5 μM); factor XIa (Kordia; 0.15 nM), Boc-

Glu(OBzl)-Ala-Arg-AMC (Bachem I-1575; 5 μM); plasmin (Kordia;

0.1 μg/mL, 1.2 nM), MeOSuc-Ala-Phe-Lys-AMC (Bachem I-1275;

50 μM); tissue plasminogen activator (tPA, Loxo; 2 nM), CH₃SO₂-D-

Phe-Gly-Arg-AMC (Pentapharm 091-06; 5 μM); Kallikrein (Kordia;

0.2 nM), H-Pro-Phe-Arg-AMC (Bachem I-1295; 5 μM); trypsin

(Sigma; 2.5 ng/mL, 0.1 nM), substrate Boc-Ile-Glu-Gly-Arg-AMC

(Bachem I-1100; 5 μM)

Coupled protease reactions are applied for assaying factor VIIa:

factor VIIa (Kordia; 0.5 pM), tissue factor (1:1000 RecombiPlasT-

in2G, Bedford, USA), factor X (Kordia; 0.07 U/mL, 12 nM), factor

Xa substrate Boc-Ile-Glu-Gly-Arg-AMC (Bachem I-1100; 50 μM);

factor X is premixed with factor Xa substrate.

For determination of test compound potency, the enzyme and

corresponding substrate dilutions are used to perform protease assays.

To 384 well microtiter plates (white, Greiner), containing 1 μL/well

serial dilutions of test or reference compounds, 20 μL of assay buﬀer,

20 μL of enzyme dilution, and 20 μL of substrate (mix) are added.

Control reactions do not contain test compound (DMSO only). After

incubation for typically 30 min (linear reaction kinetics) at room

temperature, ﬂuorescence (ex 360 nm, em 465 nm) is measured in a

microtiter plate ﬂuorescence reader (e.g., Tecan Saﬁre II). IC₅₀values

are determined by plotting log test compound concentration against

the percentage protease activity.

Determination of the Anticoagulatory Activity. The anti-

coagulatory activity of test substances was determined in vitro in

human plasma, rabbit plasma, and rat plasma. Blood was drawn oﬀ in

a mixing ratio of sodium citrate/blood of 1:9 using 0.11 M sodium

citrate solution as receiver. Immediately thereafter, the blood was

mixed thoroughly and centrifuged at ∼4000g for 15 min to separate

the plasma.

ASSOCIATED CONTENT

* Supporting Information

The Supporting Information is available free of charge at

https://pubs.acs.org/doi/10.1021/acs.jmedchem.0c01035.

■

sı

(A) Synthesis of compounds 10, 13−19, 20a/20b, 21−

26, 27a/27b, 28−32, 33a/33b, 34−36, 37a/37b, 38−

1

41, 42a/42b, 43−50, and 52; (B) H NMR data of

intermediates 53−58, 60−62, 64−65, 67−69, 72−75,

and BAY1217224 (51); (C) Tables S1−S4 (PDF)

Molecular formula strings and some data (XLSX)

Accession Codes

The coordinates and structure factors have been deposited to

the Protein Data Bank (PDB) with the following accession

codes for thrombin cocomplex structures: 6ZUG (compound

10), 6ZUH (compound 17), 6ZUN (compound 20a), 6ZUU

(compound 31), 6ZUW (compound 40), 6ZV8 (compound

51), 6ZUX (compound 42a), 6ZV7 (compound 42b), and

6TFI (PXR cocrystal structure with cmpd 17). Authors will

release the atomic coordinates upon article publication.

AUTHOR INFORMATION

Corresponding Author

■

The prothrombin time (PT, synonyms: thromboplastin time, quick

test) was determined in the presence of various concentrations of test

substance using a commercial test kit [Neoplastin (Boehringer,

Mannheim, Germany) or Hemoliance RecombiPlasTin (Instrumen-

tation Laboratory, Lexington, MA)]. Test compounds were incubated

with the plasma at 37 °C for 3 min. Coagulation was then started by

the addition of thromboplastin, and the time when coagulation

occurred was determined. The concentration of test substance which

Alexander Hillisch − Research and Development, Bayer AG,

Pharmaceuticals, 42103 Wuppertal, Germany; orcid.org/

0000-0002-4228-2347; Email: alexander.hillisch@bayer.com

Authors

Kersten M. Gericke − Research and Development, Bayer AG,

Pharmaceuticals, 42103 Wuppertal, Germany

Swen Allerheiligen − Research and Development, Bayer AG,

Pharmaceuticals, 42103 Wuppertal, Germany

Susanne Roehrig − Research and Development, Bayer AG,

Pharmaceuticals, 42103 Wuppertal, Germany

Martina Schaefer − Research and Development, Bayer AG,

Pharmaceuticals, 13342 Berlin, Germany

Adrian Tersteegen − Research and Development, Bayer AG,

Pharmaceuticals, 42103 Wuppertal, Germany

Simone Schulz − Research and Development, Bayer AG,

Pharmaceuticals, 42103 Wuppertal, Germany

Philip Lienau − Research and Development, Bayer AG,

Pharmaceuticals, 13342 Berlin, Germany

Mark Gnoth − Research and Development, Bayer AG,

Pharmaceuticals, 42103 Wuppertal, Germany

Vera Puetter − Research and Development, Bayer AG,

Pharmaceuticals, 13342 Berlin, Germany

Roman C. Hillig − Research and Development, Bayer AG,

Pharmaceuticals, 13342 Berlin, Germany

led to a doubling of the PT is reported as the EC₂₀₀

.

Arteriovenous Shunt and Hemorrhage Model (Combimo-

del Rat). Animal studies were conducted in accordance with the

German Animal Protection Act (Deutsches Tierschutzgesetz). Fasting

male rats (strain: HsdCpd:WU) with a weight of 300−350 g were

anesthetized using Inactin (150−180 mg/kg). Thrombus formation

was initiated in an arteriovenous shunt in accordance with the method

described by Christopher N. Berry;⁵⁰thus, the left jugular vein and

the right carotid artery were exposed and connected by an

extracorporeal shunt, using a polyethylene tube (PE 60) with a

length of 10 cm. In the middle, this polyethylene tube was attached to

a further polyethylene tube (PE 160) with a length of 3 cm which

contained a roughened nylon thread arranged to form a loop, to

provide a thrombogenic surface. Compound 51 was administered at

three dose levels as iv bolus with subsequent continuous iv infusion to

exclude the impact of pharmacokinetic diﬀerences on the in vivo

assessment. 20 min after the iv bolus administration of 51, the

extracorporeal circulation was opened for 15 min. Subsequently, the

shunt was removed and the thrombus on the nylon thread was

weighed.

Stefan Heitmeier − Research and Development, Bayer AG,

Pharmaceuticals, 42103 Wuppertal, Germany

To determine the bleeding time, immediately after opening of the

shunt circulation, the tip of the tail of the rat was cut by 3 mm using a

razor blade. The tail was placed into physiological saline kept at 37

Complete contact information is available at:

S

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

(9) Chobanian, H. R.; Pio, B.; Guo, Y.; Shen, H.; Huffman, M. A.;

Madeira, M.; Salituro, G.; Terebetski, J. L.; Ormes, J.; Jochnowitz, N.;

Hoos, L.; Zhou, Y.; Lewis, D.; Hawes, B.; Mitnaul, L.; O’Neill, K.;

Ellsworth, K.; Wang, L.; Biftu, T.; Duffy, J. L. Improved stability of

proline-derived direct thrombin inhibitors through hydroxyl to

heterocycle replacement. ACS Med. Chem. Lett. 2015, 6, 553−557.

(10) Abrahamsson, K.; Andersson, P.; Bergman, J.; Bredberg, U.;

https://pubs.acs.org/10.1021/acs.jmedchem.0c01035

Notes

The authors declare the following competing ﬁnancial

interest(s): All authors are employed at Bayer AG or a

subsidiary of Bayer AG.

̊

Branalt, J.; Egnell, A.-C.; Eriksson, U.; Gustafsson, D.; Hoffman, K.-J.;

ACKNOWLEDGMENTS

Nielsen, S.; Nilsson, I.; Pehrsson, S.; Polla, M. O.; Skjaeret, T.;

■

̈

̊

Strimfors, M.; Wern, C.; Olwegard-Halvarsson, M.; Ortengren, Y.

Discovery of AZD8165 − a clinical candidate from a novel series of

neutral thrombin inhibitors. MedChemComm 2016, 7, 272−281.

(11) Allerheiligen, S.; Bauser, M.; Schirok, H.; Haerter, M.; Siegel,

S.; Rester, U.; Gerdes, C.; Heitmeier, S.; von Degenfeld, G.;

Buchmueller, A.; Dittrich-Wengenroth, E.; Saatmann, U.;

Strassburger, J.; Gnoth, J. M.; Lang, D. Oxazolidinones.

WO2008155069, 2008.

We thank Petra Rothhaupt and Ivonne Herms for excellent

technical support with PXR puriﬁcation and crystallization and

the staﬀ of beamline BL14.1 at the Helmholtz-Zentrum Berlin

(HZB) for excellent support during data collection. We thank

Kay Greenﬁeld for critical review of the manuscript.

ABBREVIATIONS USED

■

(12) Bauser, M.; Buchmueller, A.; Bruens, A.; von Degenfeld, G.;

Dittrich-Wengenroth, E.; Gerdes, C.; Gnoth, M.; Gnoth, J.;

Gottschling, D.; Heitmeier, S.; Hendrix, M.; Koebberling, J.; Lang,

D.; Rester, U.; Saatmann, U.; Schneider, D.; Siegel, S.; Strassburger,

J.; Wilmen, A.; Tersteegen, A. Acylaminoimidazoles. WO2007082666,

2007.

(13) Bauser, M.; Buchmueller, A.; von Degenfeld, G.; Dittrich-

Wengenroth, E.; Gerdes, C.; Gnoth, M.; Gnoth, J.; Gottschling, D.;

Heitmeier, S.; Hendrix, M.; Koebberling, J.; Lang, D.; Rester, U.;

Saatmann, U.; Tersteegen, A.; Bruens, A. Acyl Aminoimidazoles and

Acyl Aminothiazoles. WO2008043533 2007.

(14) Bauser, M.; Buchmueller, A.; von Degenfeld, G.; Dittrich-

Wengenroth, E.; Gerdes, C.; Gnoth, M.; Gnoth, J.; Gottschling, D.;

Heitmeier, S.; Hendrix, M.; Koebberling, J.; Lang, D.; Rester, U.;

Saatmann, U.; Tersteegen, A.; Bruens, A. Acylaminopyrazoles.

WO2008046527, 2008.

(15) Schneider, D.; Buchmueller, A.; Dittrich-Wengenroth, E.;

Gerdes, C.; Gnoth, M.; Gnoth, J.; Heitmeier, S.; Hendrix, M.; Rester,

U.; Saatmann, U.; Siegel, S. Benzoxazoles. WO2007140982, 2007.

(16) Siegel, S.; Schneider, D.; Buchmueller, A.; Dittrich-

Wengenroth, E.; Gerdes, C.; Gnoth, M.; Gnoth, J.; Heitmeier, S.;

Hendrix, M.; Rester, U.; Saatmann, U. 2-(Heteroaryl) Alkyl Indazole

6-phenyl and Thienyl Methyl Amide. WO2008006479, 2008.

(17) Cheng, A. C.; Coleman, R. G.; Smyth, K. T.; Cao, Q.; Soulard,

P.; Caffrey, D. R.; Salzberg, A. C.; Huang, E. S. Structure-based

maximal affinity model predicts small-molecule druggability. Nat.

Biotechnol. 2007, 25, 71−75.

(18) Halgren, T. A. Identifying and characterizing binding sites and

assessing druggability. J. Chem. Inf. Model. 2009, 49, 377−389.

(19) Leo, A. Calculating logPoct from structures. Chem. Rev. 1993,

93, 1281.

(20) Ertl, P.; Rohde, B.; Selzer, P. Fast calculation of molecular polar

surface area as a sum of fragment-based contributions and its

application to the prediction of drug transport properties. J. Med.

Chem. 2000, 43, 3714−3717.

(21) Lobell, M.; Hendrix, M.; Hinzen, B.; Keldenich, J.; Meier, H.;

Schmeck, C.; Schohe-Loop, R.; Wunberg, T.; Hillisch, A. In silico

ADMET traffic lights as a tool for the prioritization of HTS hits.

ChemMedChem 2006, 1, 1229−1236.

(22) Abel, R.; Young, T.; Farid, R.; Berne, B. J.; Friesner, R. A. Role

of the active-site solvent in the thermodynamics of factor Xa ligand

binding. J. Am. Chem. Soc. 2008, 130, 2817−2831.

FVIIa, factor VIIa; FIXa, factor IXa; FXa, factor Xa; FXIa,

factor XIa; Caco-2, colon carcinoma cell line; EBP, ester

binding pocket; MEC, minimum eﬃcacious concentration;

OAH, oxyanion hole; S1, S2, S3, S4 pockets, substrate binding

pockets of proteinases; P1, P2, P3, P4, substrate residues of

proteinases; LLE, lipophilic ligand eﬃciency; CYP3A4,

cytochrome P450 3A4; HATU, O-(7-azabenzotriazol-1-yl)-

N,N,N′,N′-tetramethyluronium hexaﬂuorophosphate

REFERENCES

■

(1) Eriksson, U. G.; Bredberg, U.; Hoffmann, K. J.; Thuresson, A.;

Gabrielsson, M.; Ericsson, H.; Ahnoff, M.; Gislen, K.; Fager, G.;

Gustafsson, D. Absorption, distribution, metabolism, and excretion of

ximelagatran, an oral direct thrombin inhibitor, in rats, dogs, and

humans. Drug Metab. Dispos. 2003, 31, 294−305.

(2) Hauel, N. H.; Nar, H.; Priepke, H.; Ries, U.; Stassen, J. M.;

Wienen, W. Structure-based design of novel potent nonpeptide

thrombin inhibitors. J. Med. Chem. 2002, 45, 1757−1766.

(3) Straub, A.; Roehrig, S.; Hillisch, A. Oral, direct thrombin and

factor Xa inhibitors: the replacement for warfarin, leeches, and pig

intestines? Angew. Chem., Int. Ed. 2011, 50, 4574−4590.

(4) Roehrig, S.; Straub, A.; Pohlmann, J.; Lampe, T.; Pernerstorfer,

J.; Schlemmer, K. H.; Reinemer, P.; Perzborn, E. Discovery of the

novel antithrombotic agent 5-chloro-N-({(5S)-2-oxo-3- [4-(3-oxo-

morpholin-4-yl)phenyl]-1,3-oxazolidin-5-yl}methyl)thiophene- 2-car-

boxamide (BAY 59-7939): an oral, direct factor Xa inhibitor. J. Med.

Chem. 2005, 48, 5900−5908.

(5) Wong, P. C.; Pinto, D. J.; Zhang, D. Preclinical discovery of

apixaban, a direct and orally bioavailable factor Xa inhibitor. J.

Thromb. Thrombolysis 2011, 31, 478−492.

(6) Chan, L.; Pisano, M. Edoxaban (Savaysa): A factor Xa inhibitor.

P T 2015, 40, 651−695.

(7) Zhang, P.; Huang, W.; Wang, L.; Bao, L.; Jia, Z. J.; Bauer, S. M.;

Goldman, E. A.; Probst, G. D.; Song, Y.; Su, T.; Fan, J.; Wu, Y.; Li,

W.; Woolfrey, J.; Sinha, U.; Wong, P. W.; Edwards, S. T.; Arfsten, A.

E.; Clizbe, L. A.; Kanter, J.; Pandey, A.; Park, G.; Hutchaleelaha, A.;

Lambing, J. L.; Hollenbach, S. J.; Scarborough, R. M.; Zhu, B. Y.

Discovery of betrixaban (PRT054021), N-(5-chloropyridin-2-yl)-2-

(4-(N,N-dimethylcarbamimidoyl)benzamido)-5-methoxybenz amide,

a highly potent, selective, and orally efficacious factor Xa inhibitor.

Bioorg. Med. Chem. Lett. 2009, 19, 2179−2185.

(8) Burgey, C. S.; Robinson, K. A.; Lyle, T. A.; Nantermet, P. G.;

Selnick, H. G.; Isaacs, R. C.; Lewis, S. D.; Lucas, B. J.; Krueger, J. A.;

Singh, R.; Miller-Stein, C.; White, R. B.; Wong, B.; Lyle, E. A.;

Stranieri, M. T.; Cook, J. J.; McMasters, D. R.; Pellicore, J. M.; Pal, S.;

Wallace, A. A.; Clayton, F. C.; Bohn, D.; Welsh, D. C.; Lynch, J. J., Jr.;

Yan, Y.; Chen, Z.; Kuo, L.; Gardell, S. J.; Shafer, J. A.; Vacca, J. P.

Pharmacokinetic optimization of 3-amino-6-chloropyrazinone acet-

amide thrombin inhibitors. Implementation of P3 pyridine N-oxides

to deliver an orally bioavailable series containing P1 N-benzylamides.

Bioorg. Med. Chem. Lett. 2003, 13, 1353−1357.

(23) Young, T.; Abel, R.; Kim, B.; Berne, B. J.; Friesner, R. A. Motifs

for molecular recognition exploiting hydrophobic enclosure in

protein-ligand binding. Proc. Natl. Acad. Sci. U. S. A. 2007, 104,

808−813.

(24) Lazaridis, T. Inhomogeneous fluid approach to solvation

thermodynamics. I. Theory. J. Phys. Chem. B 1998, 102, 3531−3541.

(25) Jorgensen, W. L.; Maxwell, D. S.; Tirado-Rives, J. Development

and testing of the OPLS all-atom force field on conformational

energetics and properties of organic liquids. J. Am. Chem. Soc. 1996,

118, 11225−11236.

T

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Drug Annotation

(26) Ryckmans, T.; Edwards, M. P.; Horne, V. A.; Correia, A. M.;

Owen, D. R.; Thompson, L. R.; Tran, I.; Tutt, M. F.; Young, T. Rapid

assessment of a novel series of selective CB(2) agonists using parallel

synthesis protocols: A Lipophilic Efficiency (LipE) analysis. Bioorg.

Med. Chem. Lett. 2009, 19, 4406−4409.

(44) Schuttelkopf, A. W.; van Aalten, D. M. PRODRG: a tool for

̈

high-throughput crystallography of protein-ligand complexes. Acta

Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60, 1355−1363.

(45) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features

and development of Coot. Acta Crystallogr., Sect. D: Biol. Crystallogr.

2010, 66, 486−501.

(27) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: a

useful metric for lead selection. Drug Discovery Today 2004, 9, 430−

431.

(46) Wang, W.; Prosise, W. W.; Chen, J.; Taremi, S. S.; Le, H. V.;

Madison, V.; Cui, X.; Thomas, A.; Cheng, K. C.; Lesburg, C. A.

Construction and characterization of a fully active PXR/SRC-1

tethered protein with increased stability. Protein Eng., Des. Sel. 2008,

21, 425−433.

(28) Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren, T. A.;

Klicic, J. J.; Mainz, D. T.; Repasky, M. P.; Knoll, E. H.; Shelley, M.;

Perry, J. K.; Shaw, D. E.; Francis, P.; Shenkin, P. S. Glide: a new

approach for rapid, accurate docking and scoring. 1. Method and

assessment of docking accuracy. J. Med. Chem. 2004, 47, 1739−1749.

(29) Nagata, T.; Yoshino, T.; Haginoya, N.; Yoshikawa, K.; Isobe,

Y.; Furugohri, T.; Kanno, H. Cycloalkanediamine derivatives as novel

blood coagulation factor Xa inhibitors. Bioorg. Med. Chem. Lett. 2007,

17, 4683−4688.

(47) Mueller, U.; Darowski, N.; Fuchs, M. R.; Forster, R.; Hellmig,

M.; Paithankar, K. S.; Puhringer, S.; Steffien, M.; Zocher, G.; Weiss,

M. S. Facilities for macromolecular crystallography at the Helmholtz-

Zentrum Berlin. J. Synchrotron Radiat. 2012, 19, 442−449.

(48) Kabsch, W. Xds. Acta Crystallogr., Sect. D: Biol. Crystallogr.

2010, 66, 125−132.

(49) Sparta, K. M.; Krug, M.; Heinemann, U.; Muller, U.; Weiss, M.

̈

(30) Sastry, G. M.; Adzhigirey, M.; Day, T.; Annabhimoju, R.;

Sherman, W. Protein and ligand preparation: parameters, protocols,

and influence on virtual screening enrichments. J. Comput.-Aided Mol.

Des. 2013, 27, 221−234.

S. XDSAPP2.0. J. Appl. Crystallogr. 2016, 49, 1085−1092.

(50) Berry, C. N.; Girard, D.; Lochot, S.; Lecoffre, C.

Antithrombotic actions of argatroban in rat models of venous,

’mixed ’ and arterial thrombosis, and its effects on the tail transection

bleeding time. Br. J. Pharmacol. 1994, 113, 1209−1214.

(31) Sherman, W.; Day, T.; Jacobson, M. P.; Friesner, R. A.; Farid,

R. Novel procedure for modeling ligand/receptor induced fit effects. J.

Med. Chem. 2006, 49, 534−553.

(32) Allerheiligen, S.; Buchmuller, A.; Engel, K.; Gerdes, C.; Gericke,

̈

K. M.; Gerisch, M.; Heitmeier, S.; Hillisch, A.; Kinzel, T.; Lienau, P.;

̈

Riedl, B.; Rohrig, S.; Schmidt, M. V.; Strassburger, J.; Tersteegen, A.

Substituted Benzoxazoles. WO2014195231 2013.

̈

(33) Allerheiligen, S.; Buchmuller, A.; Engel, K.; Gerdes, C.; Gericke,

K. M.; Gerisch, M.; Heitmeier, S.; Hillisch, A.; Kinzel, T.; Lienau, P.;

̈

Riedl, B.; Rohrig, S.; Schmidt, M. V.; Strassburger, J.; Tersteegen, A.

Triazolopyridines as Thrombin Inhibitors for the Treatment of

Thromboembolic Diseases. WO2014195244, 2013.

(34) Allerheiligen, S.; Buchmuller, A.; Engel, K.; Gerdes, C.; Gericke,

̈

K. M.; Gerisch, M.; Heitmeier, S.; Hillisch, A.; Kinzel, T.; Lienau, P.;

̈

Riedl, B.; Rohrig, S.; Schmidt, M. V.; Strassburger, J. Substituted

Benzoxazoles. WO2014195230, 2013.

(35) Cho, S. H.; Kim, J. Y.; Lee, S. Y.; Chang, S. Silver-mediated

direct amination of benzoxazoles: tuning the amino group source

from formamides to parent amines. Angew. Chem., Int. Ed. 2009, 48,

9127−9130.

(36) Lok, R.; Leone, R. E.; Williams, A. J. Facile rearrangements of

alkynylamino heterocycles with noble metal cations. J. Org. Chem.

1996, 61, 3289−3297.

(37) Chen, Z.; Zhu, W.; Zheng, Z.; Zou, X. One-pot α -nucleophilic

fluorination of acetophenones in a deep eutectic solvent. J. Fluorine

Chem. 2010, 131, 340−344.

(38) Thvedt, T. H. K.; Fuglseth, E.; Sundby, E.; Hoff, B. H.

Microwave assisted fluorination: an improved method for side chain

fluorination of substituted 1-arylethanones. Tetrahedron 2009, 65,

9550−9556.

(39) Battye, T. G.; Kontogiannis, L.; Johnson, O.; Powell, H. R.;

Leslie, A. G. iMOSFLM: a new graphical interface for diffraction-

image processing with MOSFLM. Acta Crystallogr., Sect. D: Biol.

Crystallogr. 2011, 67, 271−281.

(40) Evans, P. R. Scaling and assessment of data quality. Acta

Crystallogr., Sect. D: Biol. Crystallogr. 2006, D62, 72−82.

(41) Winn, M. D.; Ballard, C. C.; Cowtan, K. D.; Dodson, E. J.;

Emsley, P.; Evans, P. R.; Keegan, R. M.; Krissinel, E. B.; Leslie, A. G.;

McCoy, A.; McNicholas, S. J.; Murshudov, G. N.; Pannu, N. S.;

Potterton, E. A.; Powell, H. R.; Read, R. J.; Vagin, A.; Wilson, K. S.

Overview of the CCP4 suite and current developments. Acta

Crystallogr., Sect. D: Biol. Crystallogr. 2011, 67, 235−242.

(42) Read, R. J. Pushing the boundaries of molecular replacement

with maximum likelihood. Acta Crystallogr., Sect. D: Biol. Crystallogr.

2001, 57, 1373−1382.

(43) Winn, M. D.; Murshudov, G. N.; Papiz, M. Z. Macromolecular

TLS refinement in REFMAC at moderate resolutions. Methods

Enzymol. 2003, 374, 300−321.

U

https://dx.doi.org/10.1021/acs.jmedchem.0c01035

J. Med. Chem. XXXX, XXX, XXX−XXX

Article Doi

Design, Synthesis, and Pharmacological Characterization of a Neutral, Non-Prodrug Thrombin Inhibitor with Good Oral Pharmacokinetics

DOI: 10.1021/acs.jmedchem.0c01035

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.jmedchem.0c01035

Full text of DOI:10.1021/acs.jmedchem.0c01035

Products guided by the article

R&D Labs maybe for 1272528-51-4

Relevant to this article

Hot Product