Atropisomer Control in Macrocyclic Factor VIIa Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.6b00244

Incorporation of a methyl group onto a macrocyclic FVIIa inhibitor improves potency 10-fold but is accompanied by atropisomerism due to restricted bond rotation in the macrocyclic structure, as demonstrated by NMR studies. We designed a conformational constraint favoring the desired atropisomer in which this methyl group interacts with the S2 pocket of FVIIa. A macrocyclic inhibitor incorporating this constraint was prepared and demonstrated by NMR to reside predominantly in the desired conformation. This modification improved potency 180-fold relative to the unsubstituted, racemic macrocycle and improved selectivity. An X-ray crystal structure of a closely related analogue in the FVIIa active site was obtained and matches the NMR and modeled conformations, confirming that this conformational constraint does indeed direct the methyl group into the S2 pocket as designed. The resulting rationally designed, conformationally stable template enables further optimization of these macrocyclic inhibitors.

Full text of DOI:10.1021/acs.jmedchem.6b00244

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Atropisomer Control in Macrocyclic Factor VIIa Inhibitors
Peter W Glunz, Luciano Mueller, Daniel L. Cheney, Vladimir Ladziata, Yan Zou,
Nicholas R. Wurtz, Anzhi Wei, Pancras C Wong, Ruth R. Wexler, and E. Scott Priestley
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00244 • Publication Date (Web): 25 Mar 2016
Downloaded from http://pubs.acs.org on March 29, 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 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Atropisomer Control in Macrocyclic Factor VIIa
Inhibitors
Peter W. Glunz,* Luciano Mueller, Daniel L. Cheney, Vladimir Ladziata, Yan Zou,† Nicholas R. Wurtz,
Anzhi Wei, Pancras C. Wong, Ruth R. Wexler and E. Scott Priestley
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
BristolꢀMyers Squibb Research & Development, 311 PenningtonꢀRocky Hill Road, Pennington, NJ
08534.
RECEIVED DATE (to be automatically inserted after your manuscript is accepted if required
according to the journal that you are submitting your paper to)
Running title: Atropisomerism in FVIIa macrocycles.
†Deceased. Date of death: 03/21/2014.
ABSTRACT. Incorporation of a methyl group onto a macrocyclic FVIIa inhibitor improves potency 10ꢀ
fold, but is accompanied by atropisomerism due to restricted bond rotation in the macrocyclic structure,
as demonstrated by NMR studies. We designed a conformational constraint favoring the desired
atropisomer in which this methyl group interacts with the S2 pocket of FVIIa. A macrocyclic inhibitor
incorporating this constraint was prepared and demonstrated by NMR to reside predominantly in the
desired conformation. This modification improved potency 180ꢀfold relative to the unsubstituted,
racemic macrocycle and improved selectivity. An Xꢀray crystal structure of a closely related analog in
the FVIIa active site was obtained and matches the NMR and modeled conformations, confirming that
this conformational constraint does indeed direct the methyl group into the S2 pocket as designed. The
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 44
1
2
3
4
resulting rationally designed, conformationally stable template enables further optimization of these
macrocyclic inhibitors.
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
Introduction
Atropisomers are stereoisomers with axial chirality arising from restricted rotation around a single
bond. Ōki has proposed a minimum 1000 second halfꢀlife criterion for atropisomers, sufficient
conformational stability for chemical separation of the individual conformers.1 Often atropisomerism is
observed in biaryl systems in which steric hindrance, usually at the ortho position, prevents free rotation
of the two aryl rings.2 Constraining a molecule into a cyclic structure can also give rise to
atropisomerism as in some natural products. Complestatin incorporates a phenylꢀindole linkage into a
17ꢀmembered macrocycle, which restricts rotation leading to atropisomers.3 Similarly, Vancomycin
contains a pair of atropisomeric biaryl ethers constrained into the CD and DE macrocyclic ring systems,
as well as an atropisomeric biaryl bond connecting the AB ring.4
In drug discovery, the presence of atropisomers can complicate synthesis, analytical characterization,
and evaluation of biological and ADME properties, and introduces development challenges.5 All of
these issues can be further exacerbated by equilibration of the atropisomers. Strategies for addressing
atropisomers in drug discovery include introducing symmetry to remove the chiral axis, reducing steric
hindrance to allow for rapid interchange of rotameric states, or stabilization of the atropisomer by either
increasing steric hindrance to prevent equilibration or introducing an additional stereochemical feature
to favor the desired atropisomeric state.6
Macrocyclic molecules have gained attention in drug design for their ability to present diverse
functionality and stereochemical complexity in a conformationally defined structure.7 Recently, we have
reported utilizing macrocyclization to restrict a FVIIa inhibitor into its bioactive conformation identified
ACS Paragon Plus Environment

Page 3 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
from the enzymeꢀbound Xꢀray crystal structure of an uncyclized inhibitor.8 The bound Xꢀray crystal
structure and enhanced potency of the macrocycle relative to the uncyclized inhibitor provided evidence
that the desired conformation was attained.
8
9
In this macrocyclic chemotype, we sought to exploit an interaction with the S2 pocket of FVIIa, an
important interaction employed in many FVIIa inhibitors.9 Incorporation of a methyl group at the 3ꢀ
position of the core phenyl ring of macrocycle 2 and a closely related carbamateꢀlinked macrocycle 5
improved potency 5ꢀ and 10ꢀfold, respectively, relative to unsubstituted macrocycles 1 and 4. However,
NMR spectra of these macrocycles showed apparent conformational heterogeneity, due to
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
atropisomerism arising from restricted rotation of the unsymmetrically substituted phenyl ring
constrained in a macrocycle. In addition to the complexity observed in the NMR spectra of these and
related macrocyclic compounds, the atropisomers were oftentimes separable by chromatography, but
equilibrated prior to analysis. Thus, this conformational heterogeneity confounded isolation, structural
characterization, and purity assessment of macrocycles bearing this P2 substituent. These effects of
atropisomerism introduced a roadblock in optimizing this series of macrocycles. In addition, the
observed ~1:1 ratio of conformers suggested that the methyl group was oriented to engage the S2 pocket
50% of the time and thus is not fully contributing to binding affinity and selectivity for FVIIa. In this
paper, we characterize the conformational heterogeneity of these macrocycle systems, including the
interchange of their conformers by NMR. We then report a structureꢀbased design approach to overcome
the atropisomer issue preventing the advancement of this series through conformational control of the
macrocyclic phenylglycine core to a single atropisomer, while exploiting fully the S2 interaction.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 44
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 1. Macrocyclic FVIIa inhibitors.
Results and Discussion
NMR characterization of macrocycle 3. The solutionꢀphase structure of the unsubstituted, amideꢀ
linked macrocycle 3 in DMSO was determined by NMR (Figure 2) and was found to be similar to the
crystal structure of the same molecule bound in FVIIa.8 Rapid interconversion (on the NMR time scale)
of the hydrocarbon chain in the macrocycle between the chair and saddle conformers was observed.
Inspection of the electronꢀdensity map of the crystal structure of macrocycle 3 suggested that both the
saddle and chair conformers may be populated as well in the FVIIaꢀbound state.
The assignment of resolved NMR signals for the P2 phenyl protons made it possible to monitor the
rotation of this ring, and thus the exchange of rotameric states. Via a series of NOEꢀexchange
experiments at temperatures ranging from 22 to 49 °C, the lifetimes of the rotameric states were
measured and ranged from 58 seconds (t½ = 40 s) to 9 seconds (t½ = 6 s) at 22 and 49 °C, respectively
(see Supplemental Materials for further details). This corresponds to an activation barrier of ~12.4
kcal/mol.10 Despite the lack of substitution, the slow interchange of the phenyl ring, on the order of one
minute at 22 °C, shows that rotation of the phenyl ring is considerably restricted by incorporation into
the macrocyclic structure even in the absence of an additional ring substituent. These results for the
unsubstituted macrocycle 3 are consistent with apparent atropisomers observed for methylꢀsubstituted
macrocycle 2.
ACS Paragon Plus Environment

Page 5 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Figure 2. A. Representative NMR solution structures of macrocycle 3, saddle conformation (green) and
chair conformation (orange). B. Superposition of the same NMR conformers with electron density map
(initial FoꢀFc at 3 RMSD) of the crystal structure of macrocycle 3 bound to factor VIIa suggests that
averaging of the saddle and chair conformers takes place in the enzymeꢀbound form as well.11
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
NMR Characterization of macrocycle 5. The conformational properties of methylꢀsubstituted,
carbamateꢀlinked macrocycle 5 were studied by NMR. The observed structural heterogeneity gave rise
to most protons producing two resonances of comparable intensity. For example, two peaks of almost
equal intensities are evident for the P2 methyl group (2.26 and 2.40 ppm), corresponding to the up and
down orientations of the P2 methyl relative to the plane of the macrocycle as depicted in Figure 3. A
modified NOEꢀdifference experiment was employed to prove that the observed peak splittings were due
to a pair of slowly interconverting conformers. Ringꢀflip rates could be extracted from NOEꢀdifference
spectra at a series of temperatures. Arrhenius fit of exchange rates produced an activation energy of the
phenyl ring flip of 12.4 kcal/mol for compound 5 (for details see Supplemental Materials). This
activation energy for ring flip is remarkably similar to the previously determined activation energy in
compound 3. The primary difference in the Arrhenius fits of NOESY data in compounds 3 and 5 were
the exchange factors A, which were 13.5 in compound 5 and 17.2 in compound 3. This difference in
exchange rate accounted for the much slower phenyl ring flips in compound 5 at room temperature
which produced an estimated life time of a rotameric state in the macrocyclic ring of compound 5 of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 44
1
2
3
4
5
6
7
8
approximately 2600 seconds (t½ = ~ 1800 s). Thus, the modified NOE difference data demonstrate that
the rotation of the phenyl group in 5 is restricted in the macrocycle ring and at room temperature this
substituted phenyl group does not readily equilibrate, leading to the observed atropisomerism. The
atropisomeric stability of macrocycle 5, and related structures, is moderate; conformations neither
rapidly convert nor retain longꢀterm stereochemical integrity. This characteristic, referred to as Class 2
atropisomerism, is a potential liability in drug development due to interconversion of stereoisomers on
the timeframe of isolation, analytical characterization, assaying and storage.5 Thus, we considered this
conformational heterogeneity a hindrance to advancing this chemotype.
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
O
O
Me
Me
NH2
O
NH
NH2
O
NH
N
N
H
H
N
N
N
N
H
H
O
O
5-down
5-up
Figure 3. Depictions of the two atropisomers of the P2 phenyl in macrocycle 5: Cꢀ3 methyl down (left)
and Cꢀ3 methyl up (right) relative to the plane of the macrocycle. In the NOEꢀdifference experiments the
methyl group in one conformer is irradiated and the exchange peak is measured.
Computational studies to design conformational constraint to enforce desired atropisomer. We
set out to design a means to constrain the macrocycle inhibitor to a single, desired atropisomer, and
hypothesized that given its close proximity to the P2 phenyl, substitution on the carbamate linker might
direct the 3ꢀposition methyl to the desired down orientation. Computationally, we examined the effect of
methyl substitution and its stereochemistry at both the benzylic and homobenzylic positions. To simplify
the quantum chemical calculations, we reduced the P1 aminoisoquinoline to phenyl and restricted our
attention to rotational states of the P2 phenyl (6-10, Figure 4), where we estimated the relative energy
ACS Paragon Plus Environment

Page 7 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
costs of positioning the 3ꢀposition methyl down or up. The two rotamers of 6 (the truncated version of
macrocycle 5 lacking substitution on the carbamate linker) showed similar calculated energies,
consistent with the equal rotameric populations observed in the solution NMR spectrum of macrocycle
5. Substitution at the benzylic position had a larger effect on the calculated energy difference between
the rotameric states, with the stereochemistry of the methyl substituent determining the rotameric
preference. The 3ꢀposition methyl group pointing down, which represents the desired atropisomer, is
favored by 1.9 kcal/mol in the Rꢀstereoisomer (7). In contrast, the Sꢀstereoisomer (8) favored the methyl
group pointing up, the undesired atropisomer, by 3.7 kcal/mol. Substitution at the homobenzylic position
had a smaller calculated effect on the rotameric preference of the P2 phenyl. The Rꢀstereoisomer (9)
weakly favored (0.8 kcal/mol) the desired atropisomer, whereas the Sꢀstereoisomer (10) weakly favored
(0.5 kcal/mol) the undesired atropisomer. Overall, these calculations suggested that an Rꢀmethyl
substituent at the benzylic position of the carbamate linker would stabilize the desired P2 atropisomer.
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 44
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. Energies for rotameric states of the P2 phenyl (Cꢀ3 methyl is oriented either down or up
relative to the macrocycle ring) in truncated model systems 6ꢀ10, panels BꢀF. (R)ꢀBenzylic methyl
(panel C) favors the desired, methylꢀdown conformation by 1.9 kcal/mol. The aminoisoquinoline group
was replaced with a phenyl moiety to simplify calculations. Molecular graphics were generated with the
program Pymol.11
Chemistry. The synthesis of macrocycle 5 with methyl substitution on the P2 phenyl is shown in
Scheme 1. Carbamate 12 was prepared by coupling of phenethanol 11 and 3ꢀisocyanatobenzonitrile.
Conversion of the resultant bromo intermediate 12 to the boronic acid 13 was accomplished by Miyaura
ACS Paragon Plus Environment

Page 9 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
borylation with bis(neopentyl glycolato)diboron, followed by hydrolysis. Petasis threeꢀcomponent
coupling12 between boronic acid 13, tertꢀbutyl Nꢀ(6ꢀaminoisoquinolinꢀ1ꢀyl)ꢀNꢀ[(tertꢀ
butoxy)carbonyl]carbamate,13 and glyoxylic acid monohydrate assembled the phenylglycine core (14).
Reduction of the nitrile functionality afforded amino intermediate 15, which served as the
macrocyclization precursor. Macrocyclization was accomplished via syringeꢀpump addition of 15 into
PyBop, DMAP and TEA, which after TFA deprotection afforded 5.
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
Scheme 1. Synthesis of macrocycle 5a
aReagents and conditions: (a) NaH, THF; 3ꢀisocyanatobenzonitrile, 62 %; (b) bis(neopentyl
glycolato)diboron, Pd(dppf)Cl2, DMSO, 64 %; (c) tertꢀbutyl Nꢀ(6ꢀaminoisoquinolinꢀ1ꢀyl)ꢀNꢀ[(tertꢀ
butoxy)carbonyl]carbamate, glyoxylic acid monohydrate, CH3CN, DMF, 34 %; (d) H2, 10% PdꢀC,
MeOH, 45 %; (e) Syringeꢀpump addition into PyBop, TEA, DMAP; TFA, 42 %.
Macrocycles 26 and 27, containing the (R)ꢀbenzylic methyl, were synthesized as shown in Scheme 2.
Allylation of iodobenzene 16, followed by Ru(III)ꢀcatalyzed oxidation afforded phenylacetic acid 18.
Treatment of this acid with oxalyl chloride afforded the acid chloride, which was coupled with (R)ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 44
1
2
3
4
5
6
7
8
benzyloxazolidinꢀ2ꢀone chiral auxiliary to yield imide 19. Alkylation of 19 with methyl iodide afforded
intermediate 20. Cleavage of the chiral auxiliary, followed by reduction of the resultant acid
functionality (21), provided alcohol 22. Subjecting this intermediate to Miyaura borylation conditions,
followed by hydrolysis, afforded boronic acid 23. A oneꢀpot Petasis threeꢀcomponent coupling/amide
coupling afforded the macrocyclization precursor 24 in a single step. In this procedure, boronic acid 23,
tertꢀbutyl Nꢀ(6ꢀaminoisoquinolinꢀ1ꢀyl)ꢀNꢀ[(tertꢀbutoxy)carbonyl]carbamate, and glyoxylic acid
monohydrate are heated to generate the phenylglycine derivative. To this reaction mixture, is added 3ꢀ
(aminomethyl)aniline and BOP to accomplish amide formation, yielding aminoalcohol 24. Reaction of
24 with phosgene, followed by syringeꢀpump addition of the activated species into a solution of TEA,
closed the macrocycle to give 25. Separation of diastereomers via chiral chromatography and
deprotection afforded diastereomeric macrocycles 26 and 27.
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
Scheme 2. Synthesis of macrocycles 26a,b and 27a,ba
ACS Paragon Plus Environment

Page 11 of 44
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
aReagents and conditions: (a) iꢀPrMgCl, LiCl, CuCN, THF; allyl bromide, 100 %; (b) RuCl3, NaIO4,
CCl4, H2O, 85 %; (c) (COCl)2, CH2Cl2; (d) (R)ꢀbenzyloxazolidinꢀ2ꢀone, BuLi, THF, 78 %, 2 steps; (e)
NaHMDS, THF; MeI, 65 %; (f) LiOH, H2O2, THF, 98 %; (g) BH3, THF, 97 %; (h) 5,5,5',5'ꢀtetramethylꢀ
2,2'ꢀbi(1,3,2ꢀdioxaborinane), Pd(dppf)Cl2, KOAc, DMSO, 71 %; (i) tertꢀbutyl Nꢀ(6ꢀaminoisoquinolinꢀ1ꢀ
yl)ꢀNꢀ[(tertꢀbutoxy)carbonyl]carbamate, glyoxylic acid monohydrate, CH3CN/DMF; 3ꢀ
(aminomethyl)aniline or 3ꢀ((methylamino)methyl)aniline, BOP, TEA (24a, 55 %; 24b, 51 %); (j)
phosgene, CH3CN, DCM, DMPU; syringe pump addition into TEA, DCM, (25a, 41 %; 25b, 49 %); (k)
separation of diastereomers; TFA, (26a, 41 % and 27a, 37 %; 26b, 46 % and 27b, 37 %).
TF/FVIIa activity of substituted analogs. The substituted macrocycles were tested for inhibition of
TF/FVIIa and a panel of serine proteases, and the data are summarized in Table 1. Unsubstituted
macrocycle 4 has a FVIIa Ki of 510 nM, with similar inhibitory activity against FXa and modest
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 44
1
2
3
4
5
6
7
8
selectivity versus the other serine proteases. Introduction of the 3ꢀposition methyl substituent (5)
improves potency 10ꢀfold. Introduction of a benzylic methyl group with Rꢀstereochemistry (28)
improves potency 4ꢀfold relative to 4; both compounds are 1:1 mixtures of stereoisomers at the
phenylglycine αꢀcenter. Rꢀstereochemistry at the benzylic position is 10ꢀfold more potent than the Sꢀ
stereochemistry (28 vs. 29). The observed preference for the Rꢀmethyl substituent is fortunate as it
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
suggested that we might not only control the atropisomeric preference for the subsequent substitution of
the P2ꢀphenyl group as predicted in our calculations, but might also derive an additional binding
interaction from this group. Combining these two substituents, the 3ꢀposition and benzylic methyl
groups, affords the potent inhibitor 27a with TF/FVIIa Ki = 1.4 nM. The phenylglycine Cꢀα
diastereoisomer 26a is very weakly active (Ki = 4510 nM), as is consistent with previously prepared
macrocycles 1ꢀ3,8 in which activity resides in the Rꢀstereoisomer. Nꢀmethylation of the amide (27b)
provides a compound with similar activity relative to the secondary amide 27a. Macrocycle 27b is
significantly more selective against FXa, FXIa, thrombin and trypsin than the unsubstituted macrocycle
4 and singly methylꢀsubstituted macrocycles 5 and 28. Much of this >1000ꢀfold selectivity is due to
improvement in activity against FVIIa. These macrocyclic inhibitors lacking substitution on the benzyl
amide portion, which points to the S1’ pocket, are consistently potent against tissue kallikrein.
Selectivity against this serine protease family is important, as it is implicated in cancerꢀrelated
processes14 and many other physiological functions.15 The unsubstituted macrocycle 4 favors tissue
kallikrein by 15 fold relative to FVIIa. The optimized macrocycle 27b improves to 28ꢀfold selective
against tissue kallikrein, owing to its increased potency against FVIIa. Substitution on the benyzlamide
moiety to interact with the S1’ pocket provides a means to improve tissue kallikrein selectivity, which
will be presented separately.16
Macrocycle 27a is 16ꢀfold more potent than methylꢀsubstituted macrocycle 5 and 180ꢀfold more
potent than the unsubstituted macrocycle 4, after accounting for stereochemistry. Overall, this observed
ACS Paragon Plus Environment

Page 13 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
180ꢀfold increase in potency is larger than the combined 40ꢀfold increase in potency predicted from the
individual contributions of each of the two methyl substituents: 3ꢀposition methyl (10ꢀfold, 5 vs. 4) and
Rꢀbenzylic methyl (4ꢀfold, 28 vs. 4). This result suggests there may be cooperativity of these two
interactions.
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
Table 1. SAR for substituted macrocycles.a
Ki (nM)
FXIa
Cmpd
Phg
W
X
Y
Z
TF/FVIIa
FXa
thrombin
trypsin
tissue
kall
Stereochem
R/S
R/S
R/S
R/S
R
H
H
H
H
H
H
510
46
610
5700
>13000
>13000
4700
3500
4500
35
6
4
Me
H
H
2600
10800
5
Me
H
H
H
140
1450
4510
1150
1.4
4500
9100
>6200
>6200
>10000
>10000
>10000
>10000
25
4
28
H
Me
H
H
6800
>11000
>14200
>14200
>14200
>14200
7400
29
Me
Me
Me
Me
Me
Me
Me
Me
H
>13300
>13300
>13300
>13300
>11500
>11500
>11500
>11500
31
154
25
36
26a
26b
27a
27b
R
H
Me
H
S
H
S
H
Me
1.3
a Ki values were averaged from two experiments. TF/FVIIa assays were performed with recombinant human enzyme, while
other enzyme assays were performed with purified human enzymes. All assays were run at 25 °C except tissue kallikrein 1,
which was run at 37 °C. Detailed descriptions of the enzyme assay protocols have been reported.17 Tissue kall = tissue
kallikrein 1.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 44
1
2
3
4
5
6
7
8
NMR structure of macrocycle 27a. The NMR solution structure of 27a in DMSO was determined
by dynamic simulated annealing in XPLOR starting from a randomized 3Dꢀstructure of 27a, using 28
distance restraints (Figure 5). The structure shows predominantly a single conformation for the
macrocyclic ring system, suggesting that the benzylic methyl group in fact restricts the unsymmetrically
substituted phenylglycine phenyl to a single atropisomer. 1HꢀNMR of 27a in 90% aqueous/10% DMSO
shows a single macrocycle conformation and similar chemical shift and Jꢀcoupling constant data when
compared to 100% DMSO NMR data (see Supporting Information) suggesting preservation of the
solution structure in an aqueous environment. This conformational control observed for 27a is in
contrast to the unsubstitutedꢀlinker analog 5, which shows a 1:1 ratio of conformations (vide supra).
Importantly, compound 27a in solution adopts the desired conformation in which the 3ꢀmethyl group of
the phenylglycine core is pointing down relative to the plane of the macrocycle ring. In the binding site
of FVIIa, this atropisomer orients the methyl group into the S2 pocket. As expected, the orientation of
the isoquinoline subunit is not well defined in the NMR solution structure. The conformation of this
moiety is to a large extent determined by interactions with the S1 pocket of FVIIa.
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. NMR structural studies of 27a indicate the presence of a single dominant macrocycle
conformation, placing the P2 methyl group in the desired down orientation.11
ACS Paragon Plus Environment

Page 15 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
X-ray structure of related macrocycle 30. The Xꢀray crystal structure of a closely related analog of
27a containing the same 3ꢀposition and Rꢀbenzylic methyl groups (macrocycle 30), was obtained in
FVIIa (Figure 6A). Figure 6B, shows the superposition of the bound Xꢀray structure of 30 with the
NMR solution structure of 27a and the lowestꢀenergy conformation from the computational studies (7-
down, Figure 4). All three structures show very close alignment of the macrocycle core. The P2 rotamer
predicted as energetically stable in our computational studies corresponded to both the dominant
solution structure and the enzymeꢀbound structure observed in the crystal. A subsequent and broader
force fieldꢀbased conformational analysis was performed and estimated the bound conformation to also
be the most probable in aqueous solution (74%; see Experimental Section and Supporting Information).
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
As shown in Figure 6A, 30 adopts the desired conformation in which the 3ꢀposition methyl group is
pointing down relative to the plane of the macrocycle ring. In this orientation, the methyl group interacts
with the lipophilic S2 pocket. The close similarity between the bound structure of 30 in FVIIa and the
solution structure of 27a suggests that the latter compound is also making this interaction with the S2
pocket. The increased potency of 27a relative to the unsubstituted macrocycle 5 provides further support
that this interaction is indeed formed.
NH2
O
NH
OH
N
Me
N
N
F
H
O
30
Figure 6. A. Crystallographic structure of 30 (structure shown in inset) bound in factor VIIa (2.06 Å
resolution). The initial FoꢀFc electron density map is depicted within 1.5 Å of the ligand at a contour of
3.0 RMSD.18 B. Superposition of the bound conformation of 30 in factor VIIa (magenta) with the NMR
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 44
1
2
3
4
solution structures of 27a (green) and lowest energy conformation of a truncated structure of 27a (7-
down, see Figure 4C) in which the isoquinoline has been replaced with a phenyl (orange).11
5
6
7
8
9
Conclusion. We have shown that incorporation of a methyl group in the 3ꢀposition of the
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
phenylglycine core of a macrocyclic FVIIa inhibitor improves potency 10ꢀfold, but is accompanied by
atropisomerism due to restricted rotation of the phenyl group in the macrocyclic structure. Based upon
computational studies, we proposed that incorporating an Rꢀbenzylic methyl group on the linker portion
of the macrocycle would lock the adjacent phenyl ring of the macrocycle into the desired atropisomer in
which the 3ꢀposition methyl points down, interacting with the S2 pocket of FVIIa. The macrocyclic
inhibitor 27a incorporating the Rꢀbenzylic methyl was prepared and was demonstrated by NMR to exist
in the desired atropisomer. The Rꢀbenzylic methyl group in combination with the P2ꢀmethyl group (27a)
has a significant effect on potency—180ꢀfold relative to the unsubstituted macrocycle 4. An Xꢀray
crystal structure of a closely related analog of 27a in FVIIa was obtained and matches the NMR
determined conformation in solution as well as the modeled confirmation of 27a. In addition, this FVIIaꢀ
bound Xꢀray structure shows that this conformational constraint does indeed direct the 3ꢀMe group into
the S2 pocket as designed. Importantly, through the addition of this conformational constraint we have
overcome the atropisomer issue that was hindering the advancement of this macrocyclic chemotype and
significantly increased potency and selectivity in the process. Furthermore, this constraint has provided a
rigid macrocyclic scaffold with wellꢀdefined vectors for substitution to make key interactions with the
FVIIa active site, enabling further optimization, which will be disclosed separately.
Experimental Section
ACS Paragon Plus Environment

Page 17 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
General Methods for Chemistry. All reactions were carried out using commercial grade reagents
and solvents. Solution ratios express a volume relationship, unless stated otherwise. NMR chemical
shifts (δ) are reported in parts per million relative to internal TMS. Flash chromatography was carried
8
9
out on ISCO CombiFlash systems using preꢀpacked silica cartridges and eluted with gradients of the
specified solvents. Preparative reverse phase high pressure liquid chromatography (HPLC) was carried
out on C18 HPLC columns using methanol/water gradients containing 0.1% trifluoroacetic acid. Purity
of all final compounds was determined to be ≥ 95% by analytical HPLC using the following conditions:
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
Method A: Waters Sunfire C18 column (3.5 ꢀM, 3.0 x 150 mm); Method B: Waters Xbridge Phenyl
column (3.5 ꢀM, 3.0 x 150 mm). Eluted at 1 mL/min with a 12 min gradient from 90% solvent A to
100% solvent B (solvent A: 5% acetonitrile, 95% water, 0.05% TFA; solvent B: 95% acetonitrile, 5%
water, 0.05% TFA), monitoring UV absorbance at 220 nm.
4-Bromo-2-methylphenethyl 3-cyanophenylcarbamate (12)
NaH (150 mg, 60% dispersion in oil) was added portion wise to a solution of 2ꢀ(4ꢀbromoꢀ2ꢀ
methylphenyl)ethanol (570 mg, 2.65 mmol) in THF (26.5 mL) and the mixture was stirred for 30 min.
The solution was cooled to ꢀ78°C and 3ꢀisocyanatobenzonitrile (382 mg, 2.65 mmol) was added in one
portion. The cooling bath was removed and the reaction was stirred for 2 h. Water (100 mL) was added
to the reaction mixture and it was extracted with EtOAc (2X). The combined organics phase was washed
with brine, dried (Na2SO4) and concentrated. The crude product was purified by flash chromatography
1
(0 to 100 % EtOAc/hexanes gradient) to yield 12 (590 mg, 62 % yield) as a colorless solid. H NMR
(400 MHz, DMSOꢀd6) δ ppm 9.97 (s, 1 H) 7.85 (s, 1 H) 7.70 (d, J=8.1 Hz, 1 H) 7.41 ꢀ 7.55 (m, 2 H)
7.39 (d, J=1.8 Hz, 1 H) 7.32 (dd, J=8.2, 1.9 Hz, 1 H) 7.17 (d, J=8.3 Hz, 1 H) 4.28 (t, J=7.0 Hz, 2 H)
2.92 (t, J=7.0 Hz, 2 H) 2.30 (s, 3 H).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 44
1
2
3
4
5
6
4-(2-(3-Cyanophenylcarbamoyloxy)ethyl)-3-methylphenylboronic acid (13)
7
8
9
To a flask containing 12 (730 mg, 2.0 mmol), 5,5',5'ꢀtetramethylꢀ[2,2']bi[[1,3,2]dioxaborinanyl] (496
mg, 2.2 mmol), potassium acetate (868 mg, 8.84 mmol) and Pd(dppf)Cl2·DCM (82 mg, 0.10 mmol),
was added DMSO (10 mL). The reaction mixture was degassed with three cycles of vacuum followed
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
by argon backfill. The reaction mixture was heated for 2 h at 80 °C. The mixture was diluted with water
and extracted with diethyl ether (3X). The combined organic phase was dried (MgSO4) and
concentrated. The residue was dissolved in a mixture of diethyl ether (10 mL), DCM (1 mL), and EtOAc
(1 mL). Diethanolamine (210 mg, 2.0 mmol) in isopropanol (5 mL) was added, and the reaction mixture
was stirred overnight at rt. The reaction mixture was concentrated, and the residue was purified by
preparative HPLC to give 13 (420 mg, 64 % yield) as a brown oil. MS (ESI) m/z 347.1 (M+Na)+; 1H
NMR (400 MHz, CD3OD) δ ppm 7.77 (s, 1 H) 7.56 (d, J=7.8 Hz, 1 H) 7.36 ꢀ 7.52 (m, 2 H) 7.33 (t,
J=8.0 Hz, 1 H) 7.24 (d, J=7.6 Hz, 1 H) 7.11 (d, J=7.3 Hz, 1 H) 4.26 (t, J=7.1 Hz, 2 H) 2.94 (t, J=7.1 Hz,
2 H) 2.29 (s, 3 H).
2-(5-(Bis(tert-butoxycarbonyl)amino)naphthalen-2-ylamino)-2-(4-(2-(3-
cyanophenylcarbamoyloxy)ethyl)-3-methylphenyl)acetic acid (14)
A solution of 13 (420 mg, 1.30 mmol), tertꢀbutyl Nꢀ(6ꢀaminoisoquinolinꢀ1ꢀyl)ꢀNꢀ[(tertꢀ
butoxy)carbonyl]carbamate (466 mg, 1.30 mmol) and glyoxylic acid monohydrate (143 mg, 1.30 mmol)
in acetonitrile (4 mL) and DMF (0.4 mL), were heated in a microwave reactor at 100 °C for 10 min. The
mixture was concentrated and the residue was purified by flash chromatography (0 to 10% MeOH in
CH2Cl2 gradient) to afford 14 (308 mg, 34 % yield) as a tan solid. MS (ESI) m/z 696.15 (M+H)+.
ACS Paragon Plus Environment

Page 19 of 44
Journal of Medicinal Chemistry
1
2
3
2-(4-(2-(3-(Aminomethyl)phenylcarbamoyloxy)ethyl)-3-methylphenyl)-2-(5-(bis(tert-
4
5
butoxycarbonyl)amino)naphthalen-2-ylamino)acetic acid (15)
6
7
8
9
To a solution of 14 (308 mg, 0.44 mmol) in MeOH (20 mL), was added Raney Ni. The mixture was
degassed (evacuated and flushed with Ar), then was stirred under an atmosphere of H2 (60 psi) for 15 h.
The mixture was filtered and concentrated, then purified via preparative HPLC to give 15 (140 mg, 45
% yield) as a yellow solid. MS (ESI) m/z 700.15 (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
2-(1-Amino-isoquinolin-6-ylamino)-20-methyl-13-oxa-4,11-diazatricyclo[14.2.2.16,10]henicosa-
1(19),6,8,10(21),16(20),17-hexaene-3,12-dione trifluoroacetic acid salt (5)
A solution of 15 (140 mg, 0.20 mmol) in DMF (5 mL) was added dropwise over 1 h via a syringe pump
into a solution of PyBOP (208 mg, 0.40 mmol), DMAP (172 mg, 1.0 mmol) and DIEA (0.17 mL, 1.0
mmol) in DCM (50 mL) at reflux. The mixture was stirred at rt for 15 h, then was concentrated. The
residue was dissolved in DCM and was treated with TFA (2 mL) and H2O (5 drops). The mixture was
stirred at rt for 2 h, then was concentrated. The residue was purified by preparative HPLC to afford 5 (40
mg, 42 % yield) as a white solid. MS (ESI) m/z 482.20 (M+H)+; HRMS (ESI) m/z 482.2191 (0.82 ppm);
1H NMR (500MHz, CD3OD) δ 8.07 (dd, J=9.1, 6.1 Hz, 1H), 7.50 (s, 0.5H), 7.42 (dd, J=7.8, 1.8 Hz,
0.5H), 7.34 ꢀ 7.25 (m, 2H), 7.21 (ddd, J=16.4, 9.1, 2.2 Hz, 1H), 7.17 ꢀ 7.11 (m, 2H), 6.94 ꢀ 6.82 (m, 2H),
6.75 ꢀ 6.64 (m, 2H), 6.18 (d, J=14.0 Hz, 1H), 5.14 (d, J=8.3 Hz, 1H), 4.96 ꢀ 4.88 (m, 0.5H), 4.77 ꢀ 4.69
(m, 0.5H), 4.55 (br. s., 1H), 4.20 ꢀ 3.96 (m, 2H), 3.21 ꢀ 3.10 (m, 1H), 2.88 ꢀ 2.75 (m, 1H), 2.51 (s, 1.6H),
2.36 (s, 1.4H), ~1:1 atropisomers; Purity: 99.6% (method A), 98.7% (method B).
1-Allyl-4-bromo-2-methylbenzene (17)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 44
1
2
3
To a solution of 4ꢀbromoꢀ1ꢀiodoꢀ2ꢀmethylbenzene (5 g, 16.8 mmol) in THF (100 mL) at ꢀ20 °C, was
4
5
6
7
8
9
added isopropylmagnesium chloride (2M, THF) (15 mL, 30.0 mmol). The mixture was stirred at ꢀ20 °C
for 45 min. A solution of lithium chloride (1.71 g, 40.4 mmol) and copper(I) cyanide (1.81 g, 20.21
mmol) in THF (40 mL) was added. The pale green solution was stirred for 10 min at ꢀ10 °C, then allyl
bromide (4.37 mL, 50.5 mmol) was added. The mixture was stirred at ꢀ10 °C for 10 min. The reaction
was quenched with sat. NH4Cl. and diluted with EtOAc. The organic phase was washed with H2O, 1N
HCl and brine. The organic phase was dried (Na2SO4), filtered through a 1" pad of SiO2 and
concentrated to afford 17 (3.55 g, 100 % yield) as a pale yellow oil. 1H NMR (400 MHz, CDCl3) δ ppm
7.24 ꢀ 7.31 (m, 2 H) 7.00 (d, J=7.9 Hz, 1 H) 5.86 ꢀ 5.96 (m, 1 H) 5.07 (dd, J=10.1, 1.3 Hz, 1 H) 4.97 (dd,
J=17.1, 1.3 Hz, 1 H) 3.31 (d, J=6.2 Hz, 2 H) 2.26 (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
2-(4-Bromo-2-methylphenyl)acetic acid (18)
To a solution of 17 (3.55 g, 16.8 mmol) in CCl4 (100 mL), water (150 mL) and acetonitrile (100 mL) at
rt, was added Ruthenium (III) chloride hydrate (0.523 g, 2.52 mmol) and sodium periodate (18.0 g, 84
mmol). The suspension was stirred vigorously for 2.5 h, then was diluted with H2O and DCM. The
mixture was filtered through celite. The phases were separated. The aqueous phase was extracted with
DCM (2X). The combined organic extract was washed with H2O and brine, dried (Na2SO4) and
concentrated. The resultant black solid was partitioned between Et2O and 0.1 N NaOH. The organic
phase was extracted with H2O, then the combined aqueous extract was acidified with 12 N HCl. The
aqueous phase was extracted with EtOAc (3X). The combined organic extract was washed with H2O and
brine, dried (Na2SO4), filtered through a 1" pad of SiO2 and concentrated to afford 18 (3.26 g, 85 %
yield) as an offꢀwhite solid. MS (ESI) m/z 229.2 (M+H)+; 1H NMR (400 MHz, CDCl3) δ ppm 7.35 (d,
J=1.8 Hz, 1 H) 7.30 (dd, J=8.1, 2.0 Hz, 1 H) 7.06 (d, J=7.9 Hz, 1 H) 3.62 (s, 2 H) 2.30 (s, 3 H).
ACS Paragon Plus Environment

Page 21 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
(R)-4-Benzyl-3-(2-(4-bromo-2-methylphenyl)acetyl)oxazolidin-2-one (19)
6
7
8
9
To a solution of 18 (2.77 g, 12.1 mmol) in DCM (40 mL) at rt, were added oxalyl chloride (1.16 mL,
13.3 mmol) and DMF (0.01 mL). The mixture was stirred at rt for 2 h, then concentrated to an orange oil
(acid chloride). To a solution of (R)ꢀ4ꢀbenzyloxazolidinꢀ2ꢀone (2.36 g, 13.3 mmol) in THF (40 mL) at ꢀ
78 °C, was added butyllithium (1.6 M, hexane) (8.31 mL, 13.3 mmol). The mixture was stirred for 5
min, then a solution of the acid chloride (12.1 mmol) in THF (10 mL). The mixture was stirred at ꢀ78 °C
for 1 h, then was quenched with sat. NH4Cl. The mixture was diluted with EtOAc and washed with H2O
and brine, dried (Na2SO4) and concentrated. The crude product was purified by flash chromatography
(gradient from 0 to 80% ethyl acetate/hexanes). The product was suspended in hexanes (20 mL), then
the product was collected and dried in vacuo to afford 19 (3.65 g, 78 % yield) as a colorless, crystalline
solid. MS (ESI) m/z 387.9 (M+H)+; 1H NMR (400 MHz, CD3OD) δ ppm 7.37 (d, J=1.7 Hz, 1 H) 7.27 ꢀ
7.35 (m, 4 H) 7.16 ꢀ 7.20 (m, 2 H) 7.05 (d, J=8.2 Hz, 1 H) 4.65 ꢀ 4.72 (m, 1 H) 4.16 ꢀ 4.33 (m, 4 H) 3.31
(dd, J=13.5, 3.0 Hz, 1 H) 2.78 (dd, J=13.2, 9.9 Hz, 1 H) 2.28 (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
(R)-4-Benzyl-3-((R)-2-(4-bromo-2-methylphenyl)propanoyl)oxazolidin-2-one (20)
To a solution of 19 (2.00 g, 5.15 mmol) in THF (20 mL) at ꢀ78 °C, was added NaHMDS (1M, THF)
(5.67 mL, 5.67 mmol). The pale orange solution was stirred at ꢀ78 °C for 65 min, then methyl iodide
(1.61 mL, 25.8 mmol) was added. The mixture was stirred at ꢀ78 °C for 1.5 h, then was allowed to
slowly warm to rt over 4h. The reaction was quenched with sat. NH4Cl. The mixture was diluted with
EtOAc, then washed with H2O and sat. Na2SO3, dried (Na2SO4) and concentrated. The crude product
was purified by flash chromatography (gradient from 0 to 40 % ethyl acetate/hexanes) to afford 20 (1.35
g, 65 % yield) as a viscous colorless oil. MS (ESI) m/z 401.8 (M+H)+; 1H NMR (400 MHz, CDCl3) δ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 44
1
2
3
4
5
6
7
ppm 7.30 ꢀ 7.37 (m, 3 H) 7.22 ꢀ 7.29 (m, 4 H) 7.01 (d, J=8.4 Hz, 1 H) 5.09 (q, J=6.9 Hz, 1 H) 4.62 ꢀ 4.67
(m, 1 H) 4.06 ꢀ 4.15 (m, 2 H) 3.37 (dd, J=13.4, 3.3 Hz, 1 H) 2.79 (dd, J=13.4, 9.9 Hz, 1 H) 2.39 (s, 3 H)
1.44 (d, J=7.0 Hz, 3 H).
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
(R)-2-(4-Bromo-2-methylphenyl)propanoic acid (21)
To a solution of 20 (1.66 g, 4.13 mmol) in THF (20 mL) and water (6.7 mL) at 0 °C, was added a
solution of lithium peroxide (prepared by adding hydrogen peroxide (1.81 mL, 20.6 mmol) to LiOH
(0.148 g, 6.19 mmol) in water (7 mL)), dropwise. The mixture was stirred at 0 °C, for 1 h. The reaction
was quenched with sat. Na2SO3 (~10 mL), then the volatiles were removed in vacuo. The mixture was
diluted with H2O, then the aqueous solution was extracted with DCM (2X). The aqueous was acidified
with conc. HCl, then was extracted with EtOAc (2X). The combined organic extracts were washed with
brine, dried (Na2SO4) and concentrated to afford 21 (983 mg, 98 % yield) as a colorless oil. MS (ESI)
m/z 243.1 (M+H)+; 1H NMR (400 MHz, CDCl3) δ ppm 7.30 ꢀ 7.34 (m, 2 H) 7.16 (d, J=9.2 Hz, 1 H) 3.93
(q, J=7.0 Hz, 1 H) 2.35 (s, 3 H) 1.48 (d, J=7.5 Hz, 3 H).
(R)-2-(4-Bromo-2-methylphenyl)propan-1-ol (22)
To a solution of 21 (978 mg, 4.02 mmol) in THF (20 mL) at 0 °C, was added BoraneꢀTHF (1 M, THF)
(12.1 mL, 12.1 mmol), dropwise. The mixture was removed from the ice bath and stirred at rt for 65 h.
The reaction mixture was poured into H2O (50 mL), then 1N HCl (5 mL) and EtOAc were added. The
mixture was extracted with EtOAc (2X). The combined organic extract was washed with 0.1 N HCl,
H2O and brine, dried (NaSO4), filtered through a 1" pad of SiO2 and concentrated to afford 22 (898 mg,
97 % yield) as a colorless oil. MS (ESI) m/z 211.1 (M ꢀ H2O + H)+; 1H NMR (400 MHz, CDCl3) δ ppm
ACS Paragon Plus Environment

Page 23 of 44
Journal of Medicinal Chemistry
1
2
3
7.29 ꢀ 7.34 (m, 2 H) 7.06 ꢀ 7.09 (m, 1 H) 3.65 ꢀ 3.75 (m, 2 H) 3.16 ꢀ 3.25 (m, 1 H) 2.33 (s, 3 H) 1.32 (t,
4
5
J=5.9 Hz, 1 H) 1.22 (d, J=6.6 Hz, 3 H).
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
(R)-(4-(1-Hydroxypropan-2-yl)-3-methylphenyl)boronic acid (23)
To a sealed tube was added S-7 (500 mg, 2.18 mmol), 5,5,5',5'ꢀtetramethylꢀ2,2'ꢀbi(1,3,2ꢀdioxaborinane)
(592 mg, 2.62 mmol), potassium acetate (535 mg, 5.46 mmol), Pd(dppf)Cl2 (359 mg, 0.436 mmol) in
DMSO (2 mL). The tube was filled with Ar and stirred at 85 °C for 3 h. The mixture was quenched with
water, and extracted with EtOAc 3x20 ml. The combined organic layer was filtered through silica gel
and concentrated. The product was purified via preparative HPLC to afford 23 (300 mg, 71 % yield) as a
colorless solid. MS (ESI) m/z 177.1 (M–OH)+; 1H NMR (400 MHz, CD3OD) δ 7.35 ꢀ 7.61 (m, 2 H),
7.16 (d, J=7.91 Hz, 1 H), 4.80 (s, 2 H), 3.66 (dd, J=10.8, 5.9 Hz, 1 H), 3.52 (dd, J=10.6, 7.9 Hz, 1 H),
3.15 ꢀ 3.21 (m, 1 H), 2.34 (s, 3 H), 1.24 (d, J=7.03 Hz, 3 H).
tert-Butyl N-{6-[({[(3-aminophenyl)methyl]carbamoyl}({4-[(2R)-1-hydroxypropan-2-yl]-3-
methylphenyl})methyl)amino]isoquinolin-1-yl}-N-[(tert-butoxy)carbonyl]carbamate (24a)
To a mixture of (R)ꢀ4ꢀ(1ꢀhydroxypropanꢀ2ꢀyl)ꢀ3ꢀmethylphenylboronic acid (23) (64.8 mg, 0.334 mmol),
tertꢀbutyl Nꢀ(6ꢀaminoisoquinolinꢀ1ꢀyl)ꢀNꢀ[(tertꢀbutoxy)carbonyl]carbamate (120 mg, 0.334 mmol), and
glyoxylic acid monohydrate (30.7 mg, 0.334 mmol), were added DMF (0.67 mL) and acetonitrile (2
mL). The heterogeneous mixture was stirred at 80 °C for 2 h to give an orange solution, which was
cooled to rt. To this mixture were added 3ꢀ(aminomethyl)aniline (44.9 mg, 0.367 mmol), DMF (2 mL),
BOP (162 mg, 0.367 mmol) and TEA (0.233 mL, 1.67 mmol). The mixture was stirred at rt for 1 h. The
mixture was partitioned between EtOAc and water. The aqueous phase was extracted with EtOAc. The
combined organic phase was washed with H2O and brine, dried (Na2SO4) and concentrated. The crude
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 44
1
2
3
product was purified by flash chromatography (1 to 10% methanol/methylene chloride gradient) to
afford 24a (122 mg, 55 % yield) as an offꢀwhite solid. MS (ESI) m/z 670.5 (M+H)+; 1H NMR (500MHz,
CD3OD) δ 8.06 (d, J=5.8 Hz, 1H), 7.64 (d, J=9.1 Hz, 1H), 7.44 (d, J=5.8 Hz, 1H), 7.38 ꢀ 7.34 (m, 2H),
7.29 ꢀ 7.23 (m, 2H), 6.90 (t, J=7.7 Hz, 1H), 6.72 (d, J=1.9 Hz, 1H), 6.57 ꢀ 6.52 (m, 1H), 6.52 (s, 1H),
6.47 (d, J=7.7 Hz, 1H), 5.04 (s, 1H), 4.29 (s, 2H), 3.66 (dd, J=10.9, 6.5 Hz, 1H), 3.56 (dd, J=10.7, 7.4
Hz, 1H), 3.23 ꢀ 3.16 (m, 1H), 2.37 (s, 3H), 1.28 (s, 18H), 1.27 ꢀ 1.21 (m, 3H).
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
tert-Butyl N-{6-[({[(3-aminophenyl)methyl](methyl)carbamoyl}({4-[(2R)-1-hydroxypropan-2-yl]-3-
methylphenyl})methyl)amino]isoquinolin-1-yl}-N-[(tert-butoxy)carbonyl]carbamate (24b)
Substituting 3ꢀ((methylamino)methyl)aniline for 3ꢀ(aminomethyl)aniline afforded 24b (117 mg, 51 %
yield) as an offꢀwhite solid. MS (ESI) m/z 684.5 (M+H)+; 1H NMR (500MHz, CD3OD) δ 8.03 (dd,
J=18.0, 5.9 Hz, 1H), 7.65 ꢀ 7.53 (m, 1H), 7.49 ꢀ 7.35 (m, 2H), 7.35 ꢀ 7.13 (m, 3H), 7.13 ꢀ 6.94 (m, 1H),
6.80 ꢀ 6.44 (m, 4H), 5.64 ꢀ 5.48 (m, 1H), 4.65 ꢀ 4.45 (m, 2H), 3.69 ꢀ 3.64 (m, 1H), 3.58 ꢀ 3.53 (m, 1H),
3.22 ꢀ 3.15 (m, 1H), 3.05 (s, 3H), 2.39 ꢀ 2.34 (m, 3H), 1.30 ꢀ 1.25 (m, 18H), 1.24 ꢀ 1.20 (m, 3H).
(2S,15R)-2-[(1-Aminoisoquinolin-6-yl)amino]-15,17-dimethyl-13-oxa-4,11-
diazatricyclo[14.2.2.16,10]henicosa-1(18),6,8,10(21),16,19-hexaene-3,12-dione, TFA salt (26a)
(2R,15R)-2-[(1-Aminoisoquinolin-6-yl)amino]-15,17-dimethyl-13-oxa-4,11-
diazatricyclo[14.2.2.16,10]henicosa-1(18),6,8,10(21),16,19-hexaene-3,12-dione, TFA salt (27a)
To a solution of 24a (120 mg, 0.179 mmol) in acetonitrile (1.5 mL) and DCM (3 mL) at 0 °C, was added
phosgene (15% in toluene) (0.118 mL, 0.179 mmol), dropwise. The resultant suspension was stirred at 0
°C for 20 min. The suspension was bubbled with Ar for 30 min, then was treated with DMPU (0.1 mL)
ACS Paragon Plus Environment

Page 25 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
to give a clear solution. The solution was added dropwise via a syringe pump into a solution of TEA
(0.125 mL, 0.896 mmol) in DCM (100 mL) over 3 h. The reaction mixture was stirred at rt for 2 h. The
reaction mixture was washed with water and brine, dried (Na2SO4) and concentrated. The crude product
was purified by flash chromatography (0 to 100% ethyl acetate/hexanes gradient) to afford 25a (51 mg,
41 % yield) as an orange solid. MS (ESI) m/z 696.5 (M+H)+.
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
The mixture of diastereomers was separated by chiral prep HPLC (Regis Technologies, WhelkꢀO1
10/100 FEC 25 cm X 21.1 mm, 80% IPA in heptane, 20 mL/min, RT: 6.04 and 9.06 min). Then, each
diastereomer was dissolved in TFA (2 mL), stirred at rt for 30 min, concentrated, and the residue was
purified by preparative HPLC.
Macrocycle 26a: Peak 1. (18.2 mg, 17 % yield from 24a, 41 % yield from 25a) white solid. MS (ESI)
m/z 496.2 (M+H)+; HRMS (ESI) m/z 496.2342 (ꢀ0.32 ppm); 1H NMR (500MHz, CD3OD) δ 8.63 (br. s.,
1H), 8.07 (d, J=9.1 Hz, 1H), 7.45 (d, J=1.7 Hz, 1H), 7.34 ꢀ 7.32 (m, 1H), 7.31 (d, J=1.7 Hz, 1H), 7.29 ꢀ
7.24 (m, 1H), 7.21 (dd, J=9.2, 2.3 Hz, 1H), 7.14 (t, J=7.7 Hz, 1H), 6.90 (dd, J=7.7, 0.6 Hz, 1H), 6.87 (d,
J=7.2 Hz, 1H), 6.72 (d, J=1.9 Hz, 1H), 6.69 (dd, J=8.0, 1.1 Hz, 1H), 6.18 (s, 1H), 5.13 (s, 1H), 4.81 ꢀ
4.74 (m, 1H), 4.63 (dd, J=11.0, 3.3 Hz, 1H), 4.19 (dd, J=10.7, 5.8 Hz, 1H), 4.05 (dd, J=16.1, 4.3 Hz,
1H), 3.50 ꢀ 3.42 (m, 1H), 2.48 (s, 3H), 1.38 (d, J=6.9 Hz, 3H); Purity: 100 % (method A), 100 %
(method B).
Macrocycle 27a: Peak 2. (16.3 mg, 15 % yield from 24a, 37 % yield from 25a) white solid. MS (ESI)
m/z 496.2 (M+H)+; HRMS (ESI) m/z 496.2347 (0.79 ppm); 1H NMR (500MHz, CD3OD) δ 8.67 (br. s.,
1H), 8.06 (d, J=9.4 Hz, 1H), 7.52 (dd, J=8.0, 1.7 Hz, 1H), 7.43 (d, J=8.0 Hz, 1H), 7.31 (d, J=7.2 Hz,
1H), 7.25 ꢀ 7.17 (m, 2H), 7.14 (t, J=7.7 Hz, 1H), 6.90 (dd, J=7.7, 0.6 Hz, 1H), 6.86 (d, J=6.9 Hz, 1H),
6.72 (d, J=1.9 Hz, 1H), 6.67 (dd, J=8.0, 1.1 Hz, 1H), 6.11 (s, 1H), 5.16 (s, 1H), 4.77 (dd, J=16.4, 7.6 Hz,
1H), 4.65 ꢀ 4.54 (m, 1H), 4.21 ꢀ 4.13 (m, 1H), 4.07 (dd, J=16.5, 4.4 Hz, 1H), 3.55 ꢀ 3.41 (m, 1H), 2.34
(s, 3H), 1.32 (d, J=6.9 Hz, 3H); Purity: 99.8 % (method A), 96.6 % (method B).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 44
1
2
3
4
5
6
(2S,15R)-2-[(1-Aminoisoquinolin-6-yl)amino]-4,15,17-trimethyl-13-oxa-4,11-
7
8
diazatricyclo[14.2.2.16,10]henicosa-1(18),6,8,10(21),16,19-hexaene-3,12-dione, TFA salt (26b)
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
(2R,15R)-2-[(1-Aminoisoquinolin-6-yl)amino]-4,15,17-trimethyl-13-oxa-4,11-
diazatricyclo[14.2.2.16,10]henicosa-1(18),6,8,10(21),16,19-hexaene-3,12-dione, TFA salt (27b)
According to the procedure for the cyclization of 24a, 24b (115 mg, 0.168 mmol) was cyclized to afford
25b (59 mg, 49 % yield) as an orange solid. MS (ESI) m/z 710.5 (M+H)+. The mixture of diastereomers
was separated by chiral prep HPLC (Regis Technologies, WhelkꢀO1 10/100 FEC 25 cm X 21.1 mm,
50% (1:1 MeOH/EtOH) in heptane, 20 mL/min, RT: 6.13 and 9.92 min). Then, each diastereomer was
dissolved in TFA (2 mL), stirred at rt for 30 min, concentrated, and the residue was purified by
preparative HPLC.
Macrocycle 26b: Peak 1. (23.7 mg, 23 % yield from 24b, 46% from 25b) white solid. MS (ESI) m/z
510.2 (M+H)+; HRMS (ESI) m/z 510.2489 (ꢀ2.2 ppm); 1H NMR (500 MHz, CD3OD) δ 8.05 (d, J=9.35
Hz, 1H), 7.58 (d, J=1.38 Hz, 1H), 7.31 (t, J=7.84 Hz, 2H), 7.14ꢀ7.25 (m, 3H), 6.88ꢀ6.95 (m, 2H), 6.84
(d, J=2.20 Hz, 1H), 6.70 (dd, J=1.10, 7.70 Hz, 1H), 6.04 (s, 1H), 5.72 (s, 1H), 5.43 (d, J=16.23 Hz, 1H),
4.87ꢀ4.94 (m, 1H), 3.84ꢀ3.97 (m, 2H), 3.40ꢀ3.48 (m, 1H), 3.25 (s, 3H), 2.49 (s, 3H), 1.43 (d, J=7.15 Hz,
3H); Purity: 99.6 % (method A), 96.6 % (method B).
Macrocycle 27b: Peak 2. (19.0 mg, 18 % yield from 24b, 37% from 25b) white solid. MS (ESI) m/z
510.2 (M+H)+; HRMS (ESI) m/z 510.2492 (ꢀ1.4 ppm); 1H NMR (500 MHz, CD3OD) δ 8.05 (d, J=9.35
Hz, 1H), 7.63 (dd, J=1.79, 7.84 Hz, 1H), 7.43 (d, J=7.98 Hz, 1H), 7.31 (d, J=7.15 Hz, 1H), 7.23 (d,
J=1.38 Hz, 1H), 7.21 (dd, J=2.48, 9.35 Hz, 1H), 7.18 (t, J=7.84 Hz, 1H), 6.91 (d, J=6.88 Hz, 2H), 6.84
(d, J=2.20 Hz, 1H), 6.69 (dd, J=0.96, 7.84 Hz, 1H), 5.93 (s, 1H), 5.74 (s, 1H), 5.45 (d, J=16.51 Hz, 1H),
ACS Paragon Plus Environment

Page 27 of 44
Journal of Medicinal Chemistry
1
2
3
4
4.61ꢀ4.69 (m, 1H), 3.96ꢀ4.03 (m, 1H), 3.91 (d, J=16.51 Hz, 1H), 3.46ꢀ3.55 (m, 1H), 3.29 (s, 3H), 2.34
(s, 3H), 1.31 (d, J=6.88 Hz, 3H); Purity: 100 % (method A), 100 % (method B).
5
6
7
8
Crystallization, Data Collection, and Structure Refinement.
9
Procedures followed those described in Priestley et al.8 Briefly, recombinant fullꢀlength FVIIa was
purchased from Novo Nordisk and the GLA domain (residues 1−44) was removed by cathepsin G
digestion. Single crystals suitable for Xꢀray analysis were obtained at 4 °C by hanging drop vapor
diffusion with the reservoir containing 20% PEG 6000, 20 mM CaCl2, 0.1 M MES (pH 6.0). Data for a
crystal of the FVIIa/compound 30 complex were collected at 17ꢀID located at the Advanced Photon
Source (APS) using an ADSC Quantumꢀ210 CCD detector. Image data were processed with the
program HKLꢀ2000.19 Refinement was carried out with CNX.20 The structure was reꢀrefined prior to
deposition using BUSTER (GlobalPhasing, Ltd., Cambridge, UK), GRADE (GlobalPhasing, Ltd.) for
the ligand restraint files, RHOFIT (GlobalPhasing, Ltd.) for ligand placement, and COOT21 for electron
density map fitting. Coordinates and structure amplitudes for compound 30 have been deposited in the
PDB (accession code of 5I46). Compound 3 was previously deposited8 with accession code of 4ZXY.
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
Computational Chemistry.
Quantum chemical calculations on model systems (Figure 4). The P2 rotamers were manually built
based on previous crystallographic structures of factor VIIa complexed with the phenyl glycine
chemotype and molecular modeling.8 The aminoisoquinoline was replaced by phenyl to make the
calculations more tractable. Geometry optimizations were carried out using B3LYP/ 6ꢀ31G** as
implemented with defaults in the software program Gaussian.22 Final single point energies were
computed at the RIꢀMP2/ccꢀpVTZ level of theory using QꢀChem v4.223 in the gas phase.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 44
1
2
3
4
5
6
7
8
Conformational search and Boltzmann distribution. A force field conformational search and
Boltzmann distribution analysis was conducted using a customized script controlling the software
program MacroModel, and suggested the bioactive conformation of 7 was highly likely in solution.24
For the purpose of completeness two sampling methods were utilized in parallel, the Macrocycle
Sampling25 and mixed Low Mode – Monte Carlo Multiple Minimum (using the LMCS, MCMM
keywords) algorithms with the OPLS 2.1 forcefield and SGB Born continuum solvent model.26
Conformations having ≤ 0.6 RMSD Å with respect to atomic coordinates of any other conformation
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
after best fit superposition were deemed redundant and removed. Only conformations within 10 kcal of
the observed global minimum were retained. Conformational free energies were estimated based on the
rigid – rotor approximation and frequency calculations as implemented in MacroModel, together with
the forcefield and solvation energies according to the equation:
Eq 2. ε = Gibbs free energy ≈ Eff + Esolv + Hvib + Htrans + Hrot – (Svib + Strans + Srot )*298 K
Where
Eff = OPLS 2005 forcefield energy
Esolv = Solvation energy from the SGB continuum model
Hvib, trans, rot are the vibrational, rotational, and translational enthalpies
Svib, trans, rot are the vibrational, rotational, and translational entropies
The Boltzmann probability of the jth conformation of N conformations was estimated according to the
equation:
ε j kT
e−
Eq 3.
P(Conf j ) =
N
ε
i / kT
e−
∑
i
ACS Paragon Plus Environment

Page 29 of 44
Journal of Medicinal Chemistry
1
2
3
Where
4
5
6
ε = the conformational free energy estimated from Eq. 2,
7
8
9
T = 298 °C,
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
k = the Boltzmann constant
N = the total number of conformations under consideration. Conformations having an excess of 10 kcal
are neglected.
In equation 3, the numerator is referred to as the Boltzmann weight for the jth conformation and the
denominator is referred to as the molecular partition function, or the sum of all such weights for all
conformations.
NMR Experimental
NMR sample preparation. NMR samples were prepared by dissolution of lyophilized powder of
macrocyclic FVIIa inhibitors in DMSOꢀd6 (CILꢀ34, D 99.96%) in the concentration range of 3 – 7 mM
and placed in a 5 mm NMR tube.
NMR data acquisition of unsubstituted macrocycle 3. A sample of 3 was prepared at 3.0 mM in
DMSOꢀd6. The following NMR spectra were recorded on a Varian Inova 600 MHz spectrometer which
was equipped with a Cold probe: a) experiments at 25 °C; HCꢀHSQC, HCꢀHMBC, PECOSY, NOESY
(τm = 150 ms & 300 ms), TOCSY (τm = 70 ms), b) chemical exchange experiments; NOESY (τm = 1000
ms) recorded at 22, 26.5, 31.0, 35.5, 41.0, 44.5, 49.0 °C. The spectra were manually assigned in Felix
using the ACD/NMR27ꢀpredicted chemical shifts as starting points and entered into a Felix database.
ACS Paragon Plus Environment