Definitive Metabolite Identification Coupled with Automated Ligand Identification System (ALIS) Technology: A Novel Approach to Uncover Structure-Activity Relationships and Guide Drug Design in a Factor IXa Inhibitor Program

Article abstract of DOI:10.1021/acs.jmedchem.5b01293

A potent and selective Factor IXa (FIXa) inhibitor was subjected to a series of liver microsomal incubations, which generated a number of metabolites. Using automated ligand identification system-affinity selection (ALIS-AS) methodology, metabolites in the incubation mixture were prioritized by their binding affinities to the FIXa protein. Microgram quantities of the metabolites of interest were then isolated through microisolation analytical capabilities, and structurally characterized using MicroCryoProbe heteronuclear 2D NMR techniques. The isolated metabolites recovered from the NMR experiments were then submitted directly to an in vitro FIXa enzymatic assay. The order of the metabolites' binding affinity to the Factor IXa protein from the ALIS assay was completely consistent with the enzymatic assay results. This work showcases an innovative and efficient approach to uncover structure-activity relationships (SARs) and guide drug design via microisolation-structural characterization and ALIS capabilities.

Full text of DOI:10.1021/acs.jmedchem.5b01293

Subscriber access provided by ORTA DOGU TEKNIK UNIVERSITESI KUTUPHANESI
Article
Definitive Metabolite Identification Coupled with Automated
Ligand Identification System (ALIS) Technology: A Novel
Approach to Uncover Structure–Activity Relationships
and Guide Drug Design in a Factor IXa Inhibitor Program
Ting Zhang, Yong Liu, Xianshu Yang, Gary E. Martin, Huifang Yao, Jackie Shang, Randal M.
Bugianesi, Kenneth P. Ellsworth, Lisa M. Sonatore, Peter Nizner, Edward C. Sherer, Susan E.
Hill, Ian W. Knemeyer, Wayne M. Geissler, Peter J. Dandliker, Roy Helmy, and Harold B. Wood
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 12 Feb 2016
Downloaded from http://pubs.acs.org on February 13, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Definitive Metabolite Identification Coupled with
Automated Ligand Identification System (ALIS)
Technology: A Novel Approach to Uncover
Structure–Activity Relationships and Guide Drug
Design in a Factor IXa Inhibitor Program
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
,
,
Ting Zhang
*
‡, Yong Liu †, Xianshu Yang#, Gary E. Martin†, Huifang Yao†, Jackie Shangǁ,
*
Randal M. Bugianesi§, Kenneth P. Ellsworth§, Lisa M. Sonatore§, Peter Nizner§, Edward C.
Sherer╚, Susan E. Hill┴, Ian W. Knemeyer┴, Wayne M. Geissler§, Peter J. Dandliker#, Roy
Helmy†, and Harold B. Wood‡
‡Department of Medicinal Chemistry, Merck Research Laboratories, Merck & Co., Inc., PO Box
2000, Rahway, NJ 07065, USA
†Department of Process and Analytical Chemistry, Merck Research Laboratories , Merck &
Co., Inc., PO Box 2000, Rahway, NJ 07065, USA
#Department of Pharmacology, Merck Research Laboratories, Merck & Co., Inc., 33 Avenue
Louis Pasteur, Boston, Massachusetts 02115, USA
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 49
1
2
ǁ Department of Pharmacokinetics, Pharmacodynamics and Drug Metabolism, Merck Research
Laboratories, Merck & Co., Inc., 2000 Galloping Hill Road, Kenilworth, NJ 07033, USA
3
4
5
6
7
8
9
§Department of Pharmacology, Merck Research Laboratories, Merck & Co., Inc., 2000
Galloping Hill Road, Kenilworth, NJ 07033, USA
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
╚ Department of Chemistry Modeling and Informatics, Merck Research Laboratories , Merck &
Co., Inc., PO Box 2000, Rahway, NJ 07065, USA
┴Department of Pharmacokinetics, Pharmacodynamics and Drug Metabolism, Merck Research
Laboratories, Merck & Co., Inc., Merck Research Laboratories, 33 Avenue Louis Pasteur,
Boston, MA 02115 USA
KEYWORDS: Definitive Metabolite Identification; ALIS; Factor IXa (FIXa); SAR;
MicroCryoProbe NMR
ABSTRACT. A potent and selective Factor IXa (FIXa) inhibitor was subjected to a series of
liver microsomal incubations, which generated a number of metabolites. Using automated ligand
identification systemꢀaffinity selection (ALISꢀAS) methodology, metabolites in the incubation
mixture were prioritized by their binding affinities to the FIXa protein. Microgram quantities of
the metabolites of interest were then isolated through microisolation analytical capabilities, and
structurally characterized using MicroCryoProbe heteronuclear 2D NMR techniques. The
isolated metabolites recovered from the NMR experiments were then submitted directly to an in
vitro FIXa enzymatic assay. The order of the metabolites’ binding affinity to the Factor IXa
ACS Paragon Plus Environment

Page 3 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
protein from the ALIS assay was completely consistent with the enzymatic assay results. This
work showcases an innovative and efficient approach to uncover structure–activity relationships
(SARs) and guide drug design via microisolationꢀstructural characterization and ALIS
capabilities.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
• Introduction
Characterization of drug metabolites provides discovery scientists with information on the
sites of metabolism of a structural lead series.1ꢀ5 Such knowledge is especially relevant when the
pharmacokinetic (PK) profile of the molecule is suboptimal. Medicinal chemists could also
benefit from such information in future analog designs intended to improve PK properties by
blocking those sites of metabolism.6 In addition, metabolites sometimes are more potent and can
have even better drug profiles than the parent molecule itself. Hence, a comprehensive
understanding of the location(s) in a molecule where metabolism occurs could lead to new drug
design and expanded intellectual property space.7ꢀ8
In drug discovery research, metabolites are typically generated in vitro from hepatocytes
and/or microsomes, or they may be isolated in vivo from body fluids such as urine and/or blood
samples from human and preclinical species. Traditionally, metabolite characterization has relied
heavily on liquid chromatography–mass spectrometry (LCꢀMS) techniques for sensitivity
reasons. With the advancement of high resolution (HR)ꢀMS/MS capabilities, investigators can
define the metabolically modified region(s) of the molecule and propose likely metabolite
structures based on the molecular weight change and MS/MS fragmentation pattern.9ꢀ16
Definitive structural assignment of a particular metabolite requires a sample of the metabolite
that is amenable to NMR analysis. Generally, a suitable metabolite sample can be prepared from
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 49
1
2
3
4
5
6
7
8
9
large scale incubation, followed by isolation and NMR characterization. After identifying the
structure of such metabolite, if the metabolite is of interest, medicinal chemists may then need to
design a synthetic sequence to produce larger quantities of the material for additional in vitro and
in vivo studies.6,17 Conversely, if a given metabolite is not of interest, then the cycle begins again
for another metabolite.6,17 The overall process is not targeted and consequently requires a
significant and continuing resource commitment. Because of these limitations and the low
abundance of the minor metabolites in an incubation mixture, typically only a selected number of
metabolites, usually the major ones, are isolated, fully characterized, and synthesized for further
studies. However, there remains the distinct possibility that one or more of the minor metabolites
could represent missed and new structure–activity relationships (SARs) for better drug design.
The relatively recent development of 1.7 mm MicroCryoProbe NMR capabilities has
significantly reduced the quantity of a metabolite that must be isolated for NMR
characterization.18ꢀ20 This enhanced analytical capability will open new avenues for the
characterization and investigation of minor metabolites that may embody potentially new SAR.
Recently, the automated ligand identification system (ALIS), an affinity selection−mass
spectrometry based (ASꢀMS) platform has been used to identify hits from mixtureꢀbased
combinatorial libraries.21ꢀ26 The ALIS methodology coupled with the microisolation and
structure elucidation capabilities can be applied to investigate metabolism incubation mixtures,
rank order the binding affinity of metabolites, and identify metabolites that warrant structural
characterization and synthetic elaboration. Ultimately, these novel analytical capabilities provide
tools that can help guide medicinal chemists to synthesize only the desired compound(s) and
improve the discovery time cycle for novel therapeutic leads.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 5 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Advancements in the area of in vitro pharmacologic testing now allow biologic potency
measurements to be undertaken with low microgram (ꢀg) quantities of samples. Typical
metabolite microisolation techniques afford quantities of metabolites that, when prepared for
NMR characterization using 1.7 mm MicroCryoProbe techniques, allow direct utilization of the
NMR sample in the biological assay following structural characterization. The metabolite in the
NMR sample can be quantitated by either NMR or HPLCꢀUV/MS techniques.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We report herein an example showcasing the power of combining the aforementioned
technologies in an antiꢀthrombotic drug discovery program,27,28 which enabled the quick
identification of the metabolites of interest and uncovered new SAR for improving potency.
After microsomal incubation of parent compound, ALISꢀAS technology was used to guide and
prioritize definitive metabolite isolation and identification, which ultimately led to the discovery
of an active minor metabolite with much higher binding affinity than parent compound.
Following isolation on microgram scale and NMR structural characterization, the intrinsic
activity of the µg quantity of the metabolite was confirmed in an enzymatic activity assay.
• Chemistry
Compound 7 was synthesized via the synthetic route shown in Scheme 1. Starting from a
heterocyclic halide 1, Pdꢀcatalyzed αꢀarylation with compound 2 led to nitrile 3.29 Compound 3
was then reduced using literature methods to give primary amine 4.30 In parallel, carboxylic acid
5 was converted to the acid chloride 6, which was then coupled with amine 4 to give the final
compound as a racemate. Finally, chiral resolution afforded FIXa inhibitor 7. Titrant compound
8 was synthesized using similar methods.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 49
1
2
Scheme 1a
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aReagents and conditions: (i) Pd(OAc)2, 2,8,9ꢀtriisobutylꢀ2,5,8,9ꢀtetraazaꢀ1ꢀphosphabicyclo[3.3.3]undecane, NaHMDS in THF,
dioxane, 100 C, 3 h; (ii) NaBH4, NiCl2, EtOH, THF, 25 C, overnight; (iii) oxalyl chloride, cat. DMF, DCM, 0ꢀ25 C, 2 h; (iv)
EtOH, THF, 25 oC, 1 h; (v) Chiral separation on IC column.
o
o
o
Results and Discussion
Previously, our group reported a discovery research effort focused on the development of a
novel class of antiꢀthrombotic Factor IXa (FIXa) inhibitors.27,28 A highꢀthroughput screening
campaign was conducted and a benzimidazole hit was identified. The discovery chemistry team
developed the benzimidazole hit into a highly potent and selective FIXa inhibitor, 7, which
exhibited a robust response in a pharmacodynamic (PD) assay and demonstrated the
characteristics of a true FIXa inhibitor.28 As shown in Table 1, 7 has good human Factor IXa
(hFIXa) potency and is highly selective against offꢀtarget human Factor Xa (hFXa). However,
when subjected to rat pharmacokinetics (PK) studies, high in vivo clearance of 7 was observed
(Table 1). Moreover, high in vitro intrinsic clearance (CLint) values were also observed in
human and rat liver microsomal and hepatocyte incubations.31
Table 1. Potency and rat PK data of 7
ACS Paragon Plus Environment

Page 7 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
H
N
O
Cl
N
N
H
N
6
Cl
N
N
7
F
8
7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
hFIXa Ki (nM)a
hFXa Ki (nM)a
Ratioa
308
F% a
AUCN (po) a,d
ꢀMꢁh/mpk
0.0970
T1/2
h
CLa
mL/min/kg
36.5
9.11
2798
11.1
1.24
Nb = 3
SDc = 1.5
N = 3
N = 2
SD = 1.2
N = 2
N = 2
N = 2
SD = 737
SD = 0.010
SD = 0.30
SD = 2.1
a) mean values; b) N: number of times data acquired; c) SD: standard deviation; d) AUCN: AUC normalized by dose.
XICꢀHLM
7
M538d
M538c
M538b
M538a
XICꢀRLM
7
M538d
M538c
M538b
M538a
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 49
1
2
3
4
5
6
Figure 1. Extracted Ion Chromatograms (XIC) showing metabolites arising from human liver
microsomal (HLM) and rat liver microsomal (RLM) incubations of 7.
7
8
9
To better understand the high clearance of 7, the compound was exposed to human and rat
microsomal incubations and the results are shown in Figure 1. Extensive oxidation of 7 was
observed when incubated with both human and rat microsomes. Results from LCMSꢀMS
analysis partially narrowed down the site of oxidation for each metabolite (Figure 2). MS/MS
spectra and proposed fragmenation patterns are included in Supporting Information.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Proposed relative structures for major oxidative metabolites with chromatographic
retention times from the chromatograms shown in Figure 1.
While the original goal of the metabolite identification study was to improve PK, we were
intrigued by the extensive oxidative metabolism observed in multiple regions of the molecule.
Each metabolite represents a potential direction for lead optimization. To utilize the results of
ACS Paragon Plus Environment

Page 9 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the microsomal incubation effectively, we would need to know the exact structure of each
metabolite. However, pursuing the isolation and characterization of all of the metabolites is
extremely resource intensive and timeꢀconsuming. Therefore, methodology that can help
prioritize the metabolites for isolation and structural characterization is highly desirable. This
need prompted us to explore ALISꢀAS experiments.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Affinityꢀbased competition experiments (ACE) are based on equilibrium affinity selection,
whereby a fixed concentration of a mixture of compounds and serially increasing concentrations
of a known competitor ligand (titrant) are incubated together with the target protein. The amount
of each compound bound to the target protein is then analyzed by ALISꢀAS. The MS signals of
the ligands and the competitor from such experiments reflect the equilibrium concentrations of
each proteinꢀligand complex, thereby providing information about the equilibrium dissociation
constant (Kd) of each component in a complex mixture. As the titrant concentration is increased,
MS signals for FIXaꢀbound metabolites are correspondingly reduced in a sigmoidal response. A
curveꢀfitting algorithm is used to calculate the concentration of titrant at which metabolite MS
signal is reduced by 50% compared to a control sample without titrant (ACE50). The ability of
the titrant to competitively displace the targetꢀbound ligands reveals their binding affinity to the
same site of protein and allows relative binding affinity ranking of the ligands.22ꢀ26 In an ACE50
curve, a higher value indicates a ligand with higher affinity for the protein reflected by a higher
titrant concentration required to displace the compound of interest from the binding site.
ALIS has demonstrated its unique capability in screening labelꢀfree mixtureꢀbased libraries
and finding hits that were subsequently developed into lead compounds in various classes of
protein targets.32ꢀ44 The strength of ALIS in measuring binding affinity directly from a mixture of
compounds can be brought into the workflow of metabolite identification. Instead of isolating
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 49
1
2
3
4
5
6
7
8
9
and characterizing every metabolite and profiling them individually, one can use the FIXa
protein and a crude metabolite mixture from rat liver microsomes to rank order binding affinities
via ALIS screening. The binding affinity of the metabolite provides an indication of whether a
metabolite is potentially active. Consequently, the ALIS screening methodology was utilized to
prioritize metabolites of 7 with highest affinity for FIXa for isolation and subsequent structural
characterization.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
FIXa is a structurally wellꢀdefined enzyme target.45 Its enzymatic activities have been
characterized in several different assays.46 In ALIS screening, a truncated human FIXa protein
(415 aa, E.coli derived) was used.
Compound 8, bearing fluoro (F) substitution on the benzimidazole moiety, is a close analog of
7. The hFIXa Ki of 8 is 10.6 nM (N = 3, SD = 5.6), which is in the same range as that measured
for 7 (Ki = 9.11 nM). Therefore, 8 was selected as the titrant in the ALIS competition assay.
Based on their structures, 7, 8, and the metabolites of 7 were assumed to bind to the same site of
the FIXa protein. The binding affinity of 8 can be utilized to assess the binding affinity of the
metabolites of 7 for the FIXa protein in ALIS competition experiments, allowing the metabolites
to be rankꢀordered.
The characterization of the binding mode and binding affinity of metabolites of 7 with FIXa
protein was analyzed using ACE described briefly above. In these experiments, a constant
concentration of the RLM metabolite mixture derived from 7 and serially increasing
ACS Paragon Plus Environment

Page 11 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
concentrations of the titrant, 8, were incubated with the FIXa protein. Protein−ligand complexes
were separated from the unbound compounds by SEC. The proteinꢀbound complex of 7, its
metabolites, and the titrant, 8, were then dissociated from the protein using reverseꢀphase
chromatography and finally quantitated by mass spectrometry. The ACE result of 7 and its
metabolites in the presence of the titrant was consistent with the assumption that they share a
common binding site of the FIXa protein.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As shown in Figure 3, 7 and its metabolites were displaced from the binding site of FIXa
protein by the titrant. M538a has a higher ACE50 value than 7, indicating it has greater binding
affinity in the RLM metabolite mixture derived from 7, and therefore requires higher
concentration of the titrant, 8, for displacement from the binding site. Conversely, M538b had a
lower ACE50 value than 7, indicating that M538b has a weaker binding affinity. Finally, M538c
and M538d have even weaker binding affinities than M538b. Therefore, the ALIS binding
affinity ranking from strongest to weakest is: M538a>7>M538b> M538c, M538d. We expected
this to be consistent with their respective inhibitory activity in the FIXa enzymatic assay. The
rankꢀordered binding affinity provides the needed guidance for the prioritization of definitive
metabolite isolation and identification work.
Due to its high binding affinity to FIXa protein, M538a was considered to be a high priority
metabolite for isolation and structural characterization. Due to its low abundance, which could
pose isolation and characterization challenges using conventional techniques, M538a was
subjected to the aforementioned microisolation methodology and MicroCryoProbe NMR
characterization. In addition, M538b was selected for definitive structural elucidation due to its
close binding affinity compared to compound 7. Finally, as a negative control to validate whether
ALIS binding affinity is aligned with FIXa enzymatic assay potency, the structural elucidation of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 49
1
2
3
4
5
6
7
8
9
one of the weaker binding metabolites, M538c, was also undertaken. It should be noted that
M538b and M538c were also selected for PK reasons. In a previous human hepatocyte
incubation experiment (not discussed in this report), M538b was shown to be the major
metabolite; whereas in the present human liver microsomal (HLM) incubation, M538c was
shown to be the major metabolite (Figure 1). These two metabolites represent 7’s major
oxidative metabolism outcome in human in vitro systems.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. ALIS Screen using a crude RLM incubation metabolite mixture. ALIS binding
affinity ranking was: M538a > 7 > M538b > M538c, M538d. At each concentration,
two replicates were performed and data averaged.
As was noted above, incubation of 7 with liver microsomes resulted in the formation of
metabolites M538a, M538b, and M538c; M538a was a minor metabolite relative to M538b and
M538c. In order to obtain sufficient quantities of M538a to allow NMR characterization,
ACS Paragon Plus Environment

Page 13 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
incubation protocols were optimized. It was found that incubations of 7 at 100 ꢀM concentration
with dog liver microsomes gave the best conversion of 7 to M538a (Figure 4). The same protocol
also produced acceptable amounts of M538b and M538c. Even though dog microsomal
incubation provided the highest yield of M538a, the absolute conversion was only 0.14% and
only about 25 µg of M538a was formed in the large scale incubation mixture. Isolation of µg
quantities of metabolites from a complex in vitro/in vivo matrix poses a significant analytical
challenge and a systematic approach for microisolation must be implemented. First, due to the
limited amounts of metabolites present in the matrix, an analytical scale prepꢀHPLC system
equipped with an analytical HPLC column will need to be employed. Second, analytical control
for purity evaluation and concentration of different aliquots should be in place in every step of
the isolation protocol such as workꢀup, fraction pooling, and solvent removal to ensure best
recovery possible. Next, the buffer used in the isolation (which in most cases is acidic) can cause
decomposition of metabolites during the solvent removal step. A probe stability study must be
conducted to determine the proper drying procedure to ensure the integrity of metabolites and
minimize degradation. This microisolation approach described herein was applied in our study.
The final step in the isolation sequence involved dissolving the isolated metabolite in a suitable
deuterated solvent for the NMR work and evaporating the sample to dryness a final time.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 49
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Dog liver microsomal incubation of 7.
By applying the microisolation protocol, microgram quantities of M538a, M538b, and M538c
were successfully isolated and the site of oxidation was partially defined by LCꢀMS. The
structures were then unequivocally established by heteronuclear 2D NMR methods (Table 2).
Table 2. Isolation and the characterized structures of metabolites of 7.
H
23
Cl
8
N
22
18
24
16
O
21
25
20
14
12
N
17
15
26
9
7
6
NH
19
13
10
N
31
30
27
28
11
2
32
N
1
5
Cl
3
N
29
4
F
Metabolite name
Amount (µg)
Purity by LC-UV
Structure*
ACS Paragon Plus Environment

Page 15 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
M538a
12
100%
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
M538b
M538c
30
96%
95%
114
*Tentative stereochemistry based on the known stereochemistry of 7.
Structural characterization of low microgram quantities of drug metabolites by NMR has long
been a challenging undertaking due to the inherent insensitivity of NMR spectroscopy as
compared, for example, to mass spectrometry.5,47 There has been a continual advancement in
NMR probe design over the past several decades.18,48,49 The biggest advance insofar as using
NMR to characterize drug metabolites quite likely was the development of a 1.7 mm
MicroCryoProbe in 2008ꢀ2009, the performance of which has been rigorously benchmarked for
small samples.19 When 1.7 mm cryoprobe technology is employed in tandem with a pure shift
HSQC (PSꢀHSQC)50ꢀ52 NMR spectrum, samples of a few micrograms of an isolated drug
metabolite can be successfully characterized in very reasonable time.20 The sensitivity of this
approach can be further enhanced by intentionally folding the PSꢀHSQC spectrum in the F1
dimension to either acquire the data in a shorter time or with a commensurate increase in s/n for
the same expenditure of spectrometer time.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 49
1
2
3
4
5
6
7
8
9
Structural characterization is greatly facilitated by the concerted utilization of accurate mass
MS/MS and the acquisition of pure shift HSQC data. The MS/MS data can be utilized to
localize the site of oxidative metabolism as denoted by the bracketed segments of the structures
shown in Figure 2. While oneꢀdimensional proton spectra may be severely complicated by
overlapping, extraneous peaks from the chromatographic isolation, e.g. column bleed,
contaminating species eluting from the column from previous chromatographic isolations, etc.,
separating resonances in the second frequency domain of a pure shift HSQC spectrum sidesteps
many of the inherent problems of metabolite characterization when only proton homonuclear
methods are employed. Fortuitously, 13C shifts are quite empirical in nature, which allows the
13C shift perturbation of protonated carbons flanking the site of oxidation to be calculated. Using
this approach, the site of hydroxylation can be unequivocally pinpointed. When necessary, pure
shift HSQC data can be augmented by longꢀrange heteronuclear shift correlation data, e.g. LRꢀ
HSQMBC53 and IDRꢀ(Inverted Direct Response)ꢀHSQCꢀTOCSY54 data when there is sufficient
sample available.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The sample of M538a was the smallest of the three metabolites with ~12 µg available for
analysis by NMR.20 From a PSꢀHSQC “fingerprint”, it was obvious that one of the protonated
carbons of the fluorophenyl moiety was absent from the spectrum of the metabolite.
Both fluoro and hydroxyl substituents shift the carbon ortho to the substituent on a phenyl ring
upfield. Hence, as shown for the potential 30ꢀhydroxyl metabolite, the C29 resonance between
the fluoro and hydroxyl groups would be expected to resonate considerably upfield near 100
ppm. Since none of the protonated aromatic carbons in the PSꢀHSQC spectrum of M538a
resonated upfield of 115.9 ppm, 30ꢀhyroxylation was ruled out. Hydroxylation at the 31ꢀposition
(see Supporting Information) would require all three protonated carbons of the fluorophenyl
ACS Paragon Plus Environment

Page 17 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
moiety to resonate in the range of 114ꢀ117 ppm, which also allowed this possibility to be ruled
out. Hence, it was necessary to differentiate between 27ꢀ and 29ꢀhydroxylation. The structures
were differentiated on the basis of an IDRꢀHSQCꢀTOCSY spectrum54 acquired using a 20 ms
mixing time. 27ꢀHydroxylation would lead to three contiguous protonated carbons whereas 29ꢀ
hydroxylation would only require two adjacent protonated carbons. The latter case was
confirmed with the IDRꢀHSQCꢀTOCSY data, establishing the structure of M538a as the 29ꢀ
hydroxy, which was also consistent with the chemical shifts of the protonated aromatic carbon
resonances in the PSꢀHSQC spectrum.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Similarly, the more abundant metabolites, M538b and M538c, were characterized via PSꢀ
HSQC and LRꢀHSQMBC54 (detailed discussions in Supporting Information).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 49
1
2
3
4
5
6
7
8
9
Lastly, the three isolated and structurally characterized ꢀgꢀquantity metabolites were submitted
to the FIXa enzymatic assay at a modified lowꢀconcentration titration range in order to
accurately assess the intrinsic potencies of each metabolite. As shown in Figure 5, the FIXa
enzymatic potency rankꢀordering of M538a, 7, M538b, and M538c aligns remarkably well with
ALIS FIXa binding affinity ranking shown in Figure 3.55 This observation further validated the
premise that using the resourceꢀsparing ALISꢀAS alone on a metabolite mixture can directly
rankꢀorder metabolite potency against the target protein. The ALISꢀAS assay results can then be
used to prioritize further work on the isolation, structural characterization, and synthesis of
potentially active metabolites, as well as additional SAR study followꢀup on the metabolite(s) of
interest. This approach circumvents tedious isolation and structural characterization of every
metabolite and instead directs the discovery chemistry effort more effectively. In this particular
case, for example, the more potent metabolite, M538a, became a new opportunity for lead
optimization.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 19 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5: FIXa enzymatic assay titration of 7 and its metabolites. *IC50 values represent single
determinations performed with two replicates per concentration.
Conclusions
In conclusion, we have showcased a novel and efficient approach of coupling definitive
metabolite isolation/identification and ALIS technology as a means of identifying new SARs in
an antiꢀthrombotic drug discovery research program. The ALISꢀAS technology enabled
preliminary affinity ranking using a microsomal incubation mixture without the need of isolating
each metabolite. The affinity ranking results were then used to guide and prioritize definitive
metabolite isolation and subsequent structural characterization. The isolation of pure metabolites
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 49
1
2
3
4
in microgram quantities was made possible by microisolation capabilities; and the structure
5
6
7
elucidation of ꢀg scale metabolites was enabled by MicroCyroProbe NMR. Lastly, the intrinsic
8
9
potencies of the ꢀg quantities of metabolites were confirmed in a FIXa enzymatic assay and
proven to be consistent with their ALIS binding affinities towards the FIXa protein. These results
demonstrate that ALISꢀAS binding affinity ranking can be used to predict the intrinsic potencies
of metabolites in an incubation mixture without isolation and characterization as prerequisites.
With the intrinsic potency ranking from an ACE experiment as an initial guide, a more focused
effort in definitive metabolite identification can then ensue, leading to quick identification of
region(s) on the parent molecule as a lead optimization focus. This approach avoids the
traditional tedious isolation, characterization, resynthesis and in vitro potency assessment for
each metabolite in order to identify metabolite(s) of interest. Moreover, as demonstrated herein,
ALIS can also reveal highly potent metabolites, some of which could be missed by traditional
approaches if such metabolites are present at very low abundance. It is in those highly potent but
otherwise overlooked metabolites where potentially new structure–activity relationships that can
lead to better drug design may be uncovered. Ultimately, we have demonstrated for the first time
an endꢀtoꢀend work flow using microgram quantities of materials, from microsomal
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
incubation/isolation, and NMR structure elucidation, to in vitro enzymatic assay. This practice
eliminates the lengthy and expensive synthetic steps in making pure metabolites for biological
assay, which has considerable potential to speed up the process of drug discovery.
Experimental Section
Chemistry:
ACS Paragon Plus Environment

Page 21 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
All reactions were carried out employing standard chemical techniques in ambient atmosphere
unless otherwise noted. Solvents used for extraction, washing, and chromatography were HPLC
grade. Analytical HPLC was performed on a Waters Acquity Ultra Performance LC with UV
detection at 230 and 254 nm along with ELSD and ESIꢀMS, with all final compounds showing
>95% purity and a parent mass ion consistent with the desired structure. (HPLC conditions:
Column: Supelco Ascentis Express C18, 2.7 ꢀM, 3.0 × 100 mm. Sovents: A = water/0.05%
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
TFA; B = Acetonitrile/0.05% TFA. Gradient: 10ꢀ99% B in 4 min. Flow: 1.0 mL/min, 2 ꢀL
injection. UV: 200ꢀ400 nm. MS Conditions: Scan Range: 170ꢀ900 amu. ESI positive. Run time 5
min.) All routine NMR spectra were recorded on a 500 MHz Varian INOVA 500 instrument
unless noted otherwise. 1H chemical shifts are reported as δ values in ppm relative to the residual
solvent peak (CD3OD = 3.30, acetoneꢀd6 = 2.04). Automated flash column chromatography was
performed on a Teledyne ISCO Combiflash Rf system.
2,6-Dichloro-4-(3-methyl-1H-1,2,4-triazol-1-yl)benzoic acid (5)
Step A: Methyl 2,4,6ꢀtrichlorobenzoate (10)
Into a 5000 mL 3ꢀnecked roundꢀbottom flask was placed 2,4,6ꢀtrichlorobenzoic acid 9 (200 g,
887 mmol), N,Nꢀdimethylformamide (2000 mL), potassium carbonate (368 g, 2.66 mol), and
o
CH3I (379 g, 2.67 mol). The resulting solution was stirred for 12 h at 25 C. It was then diluted
with 3000 mL of water and extracted with 3 × 1500 mL of ethyl acetate. The organic layers were
combined, washed with 1 × 2000 mL of brine, dried over anhydrous sodium sulfate and
concentrated under vacuum. This resulted in 192 g (86%) of methyl 2,4,6ꢀtrichlorobenzoate 10
as a light yellow oil.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 49
1
2
3
4
Step B: Methyl 2,6ꢀdichloroꢀ4ꢀ((diphenylmethylene)amino)benzoate (11)
5
6
7
8
9
Into a 5000ꢀmL 3ꢀnecked roundꢀbottom flask purged and maintained with an inert atmosphere
of nitrogen was placed methyl 2,4,6ꢀtrichlorobenzoate 10 (192 g, 802 mmol), toluene (2000 mL),
Cs2CO3 (789 g), Xantphos (46.7 g, 80.7 mmol), diphenylmethanimine (173 g, 956 mmol) and
Pd(OAc)2 (9.1 g, 40.5 mmol). The resulting solution was heated to reflux for 12 h in an oil bath.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
o
The reaction mixture was cooled to 25 C with a water/ice bath. The solid was filtered out and
washed with toluene. The combined filtrate was concentrated under vacuum. The residue was
applied onto a silica gel column and eluted with ethyl acetate/petroleum ether (1:20). This
resulted in 208.1 g (64%) of methyl 2,6ꢀdichloroꢀ4ꢀ[(diphenylmethylidene)amino]benzoate 11 as
a colorless oil.
Step C: Methyl 4ꢀaminoꢀ2,6ꢀdichlorobenzoate (12)
Into a 3000ꢀmL 3ꢀnecked roundꢀbottom flask was placed methyl 2,6ꢀdichloroꢀ4ꢀ
[(diphenylmethylidene)amino]benzoate 11 (208 g, 542 mmol), tetrahydrofuran (1250 mL), and
o
2N hydrogen chloride (504 mL). The resulting solution was stirred for 3 h at 25 C. The pH
value of the solution was adjusted to 9 with aq. sodium bicarbonate and extracted with 3 × 1000
mL of ethyl acetate. The organic layers were combined, dried over anhydrous sodium sulfate,
and concentrated under vacuum. The residue was applied onto a silica gel column and eluted
with ethyl acetate/petroleum ether (1:20). This resulted in 87 g (69%) of methyl 4ꢀaminoꢀ2,6ꢀ
dichlorobenzoate 12 as a white solid.
Step D: Methyl 2,6ꢀdichloroꢀ4ꢀhydrazinylbenzoate hydrochloride (13)
ACS Paragon Plus Environment

Page 23 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Into a 2000ꢀmL 3ꢀnecked roundꢀbottom flask was placed methyl 4ꢀaminoꢀ2,6ꢀ
dichlorobenzoate 12 (87 g, 395 mmol), and HCl (37%, 870 mL). This was followed by the
addition of a solution of NaNO2 (27.3 g, 396 mmol) in water (30 mL) dropwise with stirring at 0
oC over 15 min. To this was added a solution of SnCl2 (210 g, 1.11 mol) in HCl (37%, 200 mL)
dropwise with stirring at 0 oC over 30 min. The resulting solution was stirred for 2 h at 0 oC in a
water/ice bath. The solid was filtered out and washed with 100 mL of DCM. The combined
filtrate was concentrated under vacuum. This resulted in 85.5 g (76%) of methyl 2,6ꢀdichloroꢀ4ꢀ
hydrazinylbenzoate hydrochloride 13 as a white solid.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Step E: Methyl 2,6ꢀdichloroꢀ4ꢀ(3ꢀmethylꢀ1Hꢀ1,2,4ꢀtriazolꢀ1ꢀyl)benzoate (14)
Into a 2000ꢀmL 3ꢀnecked roundꢀbottom flask purged and maintained with an inert atmosphere
of nitrogen was placed ethyl ethanecarboximidate hydrochloride (39.0 g, 315 mmol) and
dichloromethane (1000 mL). This was followed by the addition of triethylamine (38.2 g, 378
mmol) dropwise with stirring at ꢀ40 oC over 30 minutes. The resulting solution was stirred for 4 h
o
at 0 C in a water/ice bath. To this was added methyl 2,6ꢀdichloroꢀ4ꢀhydrazinylbenzoate
hydrochloride 13 (85.5 g, 315 mmol), pyridine (99.69 g, 1.26 mol), and
(diethoxymethoxy)ethane (934 g, 6.30 mol). The resulting solution was concentrated under
vacuum. The residual solution was stirred for 16 h while the temperature was maintained at 80 oC
o
in an oil bath. The reaction mixture was cooled to 25 C with a water/ice bath and then poured
into 1000 mL of EtOAc/water (1:1). The aqueous phase was extracted with 3 x 1000 mL of ethyl
acetate. The organic layers were combined, dried over anhydrous sodium sulfate and
concentrated under vacuum. The crude product (80 g) was purified with FlashꢀPrepꢀHPLC with
the following conditions: column, C18ꢀSG; mobile phase, AcCN/aq. NH4HCO3 (0.1%) = 40/60
increasing to AcCN/aq. NH4HCO3 (0.1%) = 55/45 within 25 min; detector, UV 254 nm. This
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 49
1
2
3
4
resulted in 35.7 g (38%) of methyl 2,6ꢀdichloroꢀ4ꢀ(3ꢀmethylꢀ1Hꢀ1,2,4ꢀtriazolꢀ1ꢀyl)benzoate 14
5
6
as a white solid.
7
8
9
Step F: 2,6ꢀDichloroꢀ4ꢀ(3ꢀmethylꢀ1Hꢀ1,2,4ꢀtriazolꢀ1ꢀyl)benzoic acid (5)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Into a 500ꢀmL 3ꢀnecked roundꢀbottom flask was placed methyl 2,6ꢀdichloroꢀ4ꢀ(3ꢀmethylꢀ1Hꢀ
1,2,4ꢀtriazolꢀ1ꢀyl)benzoate 14 (14.7 g, 51.4 mmol), MeOH/H2O/THF (1:1:4, 240 mL) and
LiOHꢁH2O (10.8 g, 257 mmol). The resulting solution was stirred for 12 h at 60 oC in an oil bath.
o
The reaction mixture was cooled to 25 C, then concentrated under vacuum. The residual
solution was diluted with 100 mL of water, then washed with 200 mL of ethyl acetate. The
aqueous layer was adjusted to pH 3 with HCl (2 M). The solid was collected by filtration and
dried. This resulted in 9.8 g (67%) of 2,6ꢀdichloroꢀ4ꢀ(3ꢀmethylꢀ1Hꢀ1,2,4ꢀtriazolꢀ1ꢀyl)benzoic
1
acid 5 as a white solid. LCꢀMS: m/z = 272 [M+H]+. H NMR (Bruker 400 MHz, DMSOꢀd6,
ppm): δ 14.29 (s, 1H), 9.31 (s, 1H), 8.06 (s, 2H), 2.38 (s, 3H).
(R)-2,6-dichloro-N-(2-(3-fluorophenyl)-2-(2-methyl-1H-benzo[d]imidazol-5-yl)ethyl)-4-(3-
methyl-1H-1,2,4-triazol-1-yl)benzamide (7).
Step A: (±)ꢀ2ꢀ(3ꢀFluorophenyl)ꢀ2ꢀ(2ꢀmethylꢀ1Hꢀbenzo[d]imidazolꢀ5ꢀyl)acetonitrile (3).
The reactor was charged with palladium(II) acetate (0.160 g, 0.711 mmol), 2,8,9ꢀtriisobutylꢀ
2,5,8,9ꢀtetraazaꢀ1ꢀphosphabicyclo[3.3.3]undecane (0.507 mL, 1.42 mmol), and 5ꢀbromoꢀ2ꢀ
methylꢀ1Hꢀbenzo[d]imidazole 1 (1.5 g, 7.11 mmol). The vessel was evacuated and refilled with
N2 for three cycles and degassed dioxane (14.2 mL) was added, followed by NaHMDS (28.4
mL, 28.4 mmol). The reaction was stirred at 25 °C for 20 min before 2ꢀ(3ꢀ
fluorophenyl)acetonitrile 2 (73.3 mg, 0.543 mmol) was added. The reaction was then capped and
heated at 100 °C for 3 h. More NaHMDS (6 mL) was added and the reaction was stirred for
ACS Paragon Plus Environment

Page 25 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
another 1 h. Additional NaHMDS (4 mL) was added and the reaction was stirred for another 1 h
before it was cooled to room temperature, and quenched with water. The aqueous layer was
extracted with EtOAc (3 × 300 mL), washed with brine (300 mL), dried over MgSO4, filtered,
and the filtrate was concentrated in vacuo. The crude material was purified by chromatography
on silica gel (eluent: 0ꢀ100% A in B. A: 10% MeOH in DCM; B: DCM.) to give 1.62 g (±)ꢀ2ꢀ(3ꢀ
Fluorophenyl)ꢀ2ꢀ(2ꢀmethylꢀ1Hꢀbenzo[d]imidazolꢀ5ꢀyl)acetonitrile 3 as a tan gue. LCꢀMS: m/z =
266.0 [M+H]+.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Step B: 2,6ꢀDichloroꢀ4ꢀ(3ꢀmethylꢀ1Hꢀ1,2,4ꢀtriazolꢀ1ꢀyl)benzoyl chloride (6).
At 0 °C, in a preꢀweighed flask containing carboxylic acid 5 (2.28 g, 8.38 mmol) in 3 mL
DCM, was added oxalyl chloride (1.50 mL, 16.8 mmol), followed by one drop of DMF. After
0.5 h, the reaction was warmed to 25 °C and stirred for 2 h. The reaction mixture was
concentrated and the excess oxalyl chloride was azeotropically removed with toluene. LCꢀMS:
m/z = 289.9 [M+H]+.
Step C: (±)ꢀ2ꢀ(3ꢀFluorophenyl)ꢀ2ꢀ(2ꢀmethylꢀ1Hꢀbenzo[d]imidazolꢀ5ꢀyl)ethanamine (4).
Under N2, to the step A product 3 (1.62 g, 6.11 mmol) was added anhydrous nickel(II) chloride
(0.791 g, 6.11 mmol) in anhydrous ethanol (15.3 mL) and THF (15.3 mL) was added sodium
borohydride (1.39 g, 36.6 mmol). The reaction was stirred at 25°C overnight. LCꢀMS: m/z =
270.1 [M+H]+.
Step
D:
(R)ꢀ2,6ꢀDichloroꢀNꢀ(2ꢀ(3ꢀfluorophenyl)ꢀ2ꢀ(2ꢀmethylꢀ1Hꢀbenzo[d]imidazolꢀ5ꢀ
yl)ethyl)ꢀ4ꢀ(3ꢀmethylꢀ1Hꢀ1,2,4ꢀtriazolꢀ1ꢀyl)benzamide (7)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 49
1
2
3
4
5
6
7
8
9
The reaction mixture from step C was added to the step B product 6 (2.66 g, 9.16 mmol). The
vial containing the step C reaction was rinsed with a total of 5 mL THF that was also added to
the step B product 6. The mixture was stirred at 25 °C for 1 h. The reaction was then quenched
with a saturated NaHCO3 aqueous solution (200 mL), diluted with 300 mL EtOAc, agitated, and
filtered. The filtrate was collected, and the phases separated. The aqueous phase was extracted
with EtOAc (3 × 300 mL). The combined aqueous layers were washed with brine (300 mL),
dried over MgSO4, filtered, and concentrated in vacuo. The crude material was purified by
chromatography on silica gel (eluent: 0ꢀ100% A in B. A: 10% MeOH in DCM; B: DCM.) to
yield 0.88 g racemic product. LCꢀMS: m/z = 523.0 [M+H]+. ¹H NMR (500 MHz, CD₃OD): δ
8.97 (s, 1 H); 7.79 (s, 2 H); 7.46 (s, 1 H); 7.39 (t, J = 8.3 Hz, 1 H); 7.33ꢀ7.25 (m, 1 H); 7.22ꢀ7.09
(m, 3 H); 6.91 (td, J = 8.5, 2.4 Hz, 1 H); 4.55 (t, J = 8.1 Hz, 1 H); 4.16ꢀ4.00 (m, 2 H); 2.53 (s, 3
H); 2.39 (s, 3 H).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chiral separation: IC column (2 × 15 cm), 35% ethanol (0.1% DEA)/CO2, 50 mL/min, 100
bar, 220 nm, injection volume: 1 mL; 9 mg/mL in methanol. Fastꢀeluting enantiomer A, 378 mg;
slowꢀeluting enantiomer B, 379 mg. The title compound 7 is enantiomer B. (The assignment of
stereochemistry was based on known XꢀRay coꢀcrystallization data between FIXa enzyme and
analogs of 7)27,28
(S)-2,6-Dichloro-N-(2-(7-fluoro-2-methyl-1H-benzo[d]imidazol-5-yl)-2-(3-
fluorophenyl)ethyl)-4-(3-methyl-1H-1,2,4-triazol-1-yl)benzamide (8)
Step A: 5ꢀBromoꢀ7ꢀfluoroꢀ2ꢀmethylꢀ1Hꢀbenzo[d]imidazole (16)
A solution of 5ꢀbromoꢀ3ꢀfluorobenzeneꢀ1,2ꢀdiamine 15 (1 g, 4.88 mmol), acetic acid (2.93 g,
48.8 mmol) in PPA (polyphosphoric acid) (4.88 mL) was heated to 135ꢀ140 oC for 4 h under N2.
ACS Paragon Plus Environment

Page 27 of 49
Journal of Medicinal Chemistry
1
2
The reaction was then cooled to 25 oC and transferred into 25 mL of ice cold water. Charcoal
(about 1 g) was added and the mixture was filtered over a Celite bed, which was rinsed with
water. The filtrate was basified with aq. ammonia solution until pH = 10~12. The basified
reaction mixture was extracted with ethyl acetate (3 × 100 mL), and the combined ethyl acetate
layers were dried over MgSO4 and concentrated to a yellow solid residue. To the residue was
added hexanes (5 mL) and the mixture was filtered to reveal 0.93 g 5ꢀbromoꢀ7ꢀfluoroꢀ2ꢀmethylꢀ
1Hꢀbenzo[d]imidazole 16 as a white solid. LCꢀMS: m/z = 228.9 [M+H]+.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Step B: 2ꢀ(7ꢀFluoroꢀ2ꢀmethylꢀ1Hꢀbenzo[d]imidazolꢀ5ꢀyl)ꢀ2ꢀ(3ꢀfluorophenyl)acetonitrile (17)
The reactor was charged with palladium(II) acetate (0.037 g, 0.166 mmol), 2,8,9ꢀtriisobutylꢀ
2,5,8,9ꢀtetraazaꢀ1ꢀphosphabicyclo[3.3.3]undecane (114 mg, 0.332 mmol), and 5ꢀbromoꢀ7ꢀfluoroꢀ
2ꢀmethylꢀ1Hꢀbenzo[d]imidazole 16 (380 mg, 1.66 mmol). The vessel was evacuated and refilled
with N2 for 3 cycles and degassed dioxane (3.32 mL) was added, followed by NaHMDS (6.64
mL, 6.64 mmol). The reaction was stirred at 25 °C for 20 min before 2ꢀ(3ꢀ
fluorophenyl)acetonitrile (269 mg, 1.99 mmol) was added. The reaction was then capped and
heated at 100 °C for 6 h. The reaction was cooled to room temperature, and quenched with water.
The aqueous layer was extracted with EtOAc (3 × 100 mL), washed with brine (100 mL), dried
over MgSO4, filtered, and the filtrate was concentrated in vacuo. The crude material was passed
through a silica short column using eluent 0ꢀ100% A in B (A: 10% MeOH in DCM; B: DCM.) to
yield a crude product, which was used directly in the next step. LCꢀMS: m/z = 284.1 [M+H]+.
Step C: 2ꢀ(7ꢀFluoroꢀ2ꢀmethylꢀ1Hꢀbenzo[d]imidazolꢀ5ꢀyl)ꢀ2ꢀ(3ꢀfluorophenyl)ethanꢀ1ꢀamine (18)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 49
1
2
3
4
5
6
7
8
9
Under N2, to product 17 (180 mg, 0.635 mmol), anhydrous nickel(II) chloride (82 mg, 0.635
mmol) in anhydrous ethanol (1.6 mL) and THF (1.6 mL) was added sodium borohydride (144
mg, 3.81 mmol). The reaction was stirred at 25 °C overnight to produce 18.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Step
D:
(S)ꢀ2,6ꢀDichloroꢀNꢀ(2ꢀ(7ꢀfluoroꢀ2ꢀmethylꢀ1Hꢀbenzo[d]imidazolꢀ5ꢀyl)ꢀ2ꢀ(3ꢀ
fluorophenyl)ethyl)ꢀ4ꢀ(3ꢀmethylꢀ1Hꢀ1,2,4ꢀtriazolꢀ1ꢀyl)benzamide (8)
The reaction mixture containing 18 from above was then transferred to a vial containing
compound 6 (119 mg, 0.318 mmol). The vial containing the reaction mixture from above was
rinsed with a total of 2 mL THF and then added to the vial containing compound 6. The mixture
was stirred at 25 °C for 1 h. The reaction was then quenched with saturated NaHCO3 aqueous
solution, diluted with 10 mL EtOAc, agitated, and filtered. The filtrate was collected, and phases
were separated. The aqueous phase was extracted with EtOAc (3 × 10 mL). The combined
aqueous layers were washed with brine, dried over MgSO4, filtered and concentrated in vacuo.
The crude material was purified by chromatography on silica gel (eluent: 0ꢀ100% A in B. A:
10% MeOH in DCM; B: DCM.) to yield 90 mg racemic product containing impurity. This crude
product was subjected to chiral resolution and purity upgrade: IC column (2 × 15 cm), 45% 2:1
ACN:isopropanol(0.1% DEA)/CO2, 60 mL/min, 100 bar, 220 nm, inj vol.: 0.5 mL; 8 mg/mL 1:2
DCM:methanol. Faster eluent, enantiomer A, 25 mg; Slower eluent, enantiomer B, 22 mg. The
title compound 8 is enantiomer B. LCꢀMS: m/z = 541.1 [M+H]+. ¹H NMR (500 MHz, acetoneꢀ
d6): δ 11.8 (brs, 1 H), 9.02 (s, 1 H); 8.07 (s, 1 H); 7.85 (s, 2 H); 7.37 (t, J = 5.5 Hz, 2 H); 7.31ꢀ
7.24 (m, 2 H); 6.99 (m, 2 H); 4.62 (t, J = 7.9 Hz, 1 H); 4.25ꢀ4.12 (m, 2 H); 2.55 (s, 3 H); 2.38 (s,
3 H).
FIXa/FXa Enzymatic Assay Protocols:
ACS Paragon Plus Environment

Page 29 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
•
Materials:
Buffer components PEGꢀ8000, HEPES, and CaCl2•H2O were obtained from Hampton
Research. Sodium chloride and nꢀAcetylꢀKPRꢀAFC (AFC 7ꢀamidoꢀ4ꢀtrifluoroꢀ
=
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
methylcoumarin) were obtained from Sigma. CH3SO2ꢀDꢀcyclohexylGlyꢀGlyꢀArgꢀAFCꢁAcOH),
was obtained from CPC Scientific Custom Peptide. Human FXa was purchased from Sekisui
Diagnostics. Recombinant human Factor IXa enzyme was purchased from Haemtech
Technologies and further purified via a ResS column.
•
Methods:
Test compounds were titrated with DMSO as the diluent in a 10ꢀpoint dose response in a
separate step followed by a 60ꢀfold dilution into the reaction. Assay buffer contained 50 mM
HEPES, pH 7.4, 5 mM CaCl2•2H2O, 150 mM NaCl, 0.1% PEGꢀ8000. Assay pipetting steps to
384 well microplates were performed with a Hamilton Microlab Star robotic platform.
Enzyme activity was measured by recording the increase in fluorescence signal after cleavage
of a labeled tripeptide and subsequent release of the free fluorophore AFC as follows.
Human Factor IXa protein was incubated with test compound for 30 minutes followed by the
addition of peptide substrate CH3SO2ꢀDꢀCHGꢀGlyꢀArgꢀAFCꢁAcOH, resulting in a 60ꢀfold
dilution of compound and final concentration of 2 nM FIXa and 250 ꢀM substrate. After
incubation at ambient temperature for 1 h, the fluorescence was measured on an Envision
Multilabel reader with 405 nm excitation and 510 nm emission filters.
FXa activity was determined using the same methodology with final concentrations of 15 nM
FXa and 100 ꢀM nꢀAcetylꢀKPRꢀAFC.
Inhibitory activity (IC50) of test compounds was determined from a dose titration using a four
parameter logistic fit, and subsequently converted to Ki using the ChengꢀPrusoff equation (Ki =
ACS Paragon Plus Environment