Fragment-Linking Approach Using 19F NMR Spectroscopy to Obtain Highly Potent and Selective Inhibitors of β-Secretase

Article abstract of DOI:10.1021/acs.jmedchem.5b01917

Fragment-based drug discovery (FBDD) has become a widely used tool in small-molecule drug discovery efforts. One of the most commonly used biophysical methods in detecting weak binding of fragments is nuclear magnetic resonance (NMR) spectroscopy. In part

Full text of DOI:10.1021/acs.jmedchem.5b01917

Subscriber access provided by University of Otago Library
Article
A Fragment Linking Approach Using 19F NMR Spectroscopy
to Obtain Highly Potent and Selective Inhibitors of #-secretase.
John Brad Jordan, Douglas A. Whittington, Michael D. Bartberger,
E. Allen Sickmier, Kui Chen, Yuan Cheng, and Ted Judd
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01917 • Publication Date (Web): 15 Mar 2016
Downloaded from http://pubs.acs.org on March 16, 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 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
9
A Fragment-Linking Approach Using F NMR
Spectroscopy to Obtain Highly Potent and Selective
Inhibitors of β-Secretase
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
John B. Jordan*, Douglas A. Whittington, Michael D. Bartberger, E. Allen Sickmier, Kui Chen,
Yuan Cheng and Ted Judd
Therapeutic Discovery, Amgen, Inc. One Amgen Center Drive, Thousand Oaks, CA
91320
RECEIVED DATE
*
To whom correspondence should be addressed. Tel: 805-313-5647. Fax: 805-480-1337. E-
mail:brad.jordan@amgen.com
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
Fragment-based drug discovery (FBDD) has become a widely used tool in small molecule drug
discovery efforts. One of the most commonly used biophysical methods in detecting weak
binding of fragments is nuclear magnetic resonance (NMR) spectroscopy. In particular, FBDD
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
9
1
performed with F NMR-based methods has been shown to provide several advantages over H
NMR using traditional magnetization-transfer and/or two-dimensional methods. Here, we
1
9
demonstrate the utility and power of F-based fragment screening by detailing the identification
1
9
of a second-site fragment through F NMR screening that binds to a specific pocket of the
aspartic acid protease, β-secretase (BACE-1). The identification of this second-site fragment
allowed the undertaking of a fragment-linking approach, which ultimately yielded a molecule
exhibiting a more than 360-fold increase in potency while maintaining reasonable ligand
efficiency and gaining much improved selectivity over cathepsin-D (CatD). X-ray
crystallographic studies of the molecules demonstrated that the linked fragments exhibited
binding modes consistent with those predicted from the targeted screening approach, through-
space NMR data, and molecular modeling.
ACS Paragon Plus Environment

Page 3 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
Fragment-based drug discovery (FBDD) has rapidly gained momentum in the discovery and
development of small molecule therapeutics. Over the past ten years, FBDD has become a
common workstream in the path to find novel chemical matter, often for challenging targets, and
a number of molecules in various phases of development have been derived from an initial
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
1
fragment screening hit . With the advent of FBDD came the necessity for highly sensitive
biophysical methods aimed at detecting the weak binding of small molecules normally observed
in fragment-based approaches. As such, various solution-phase NMR techniques and adaptations
to surface plasmon resonance (SPR) methods (as well as other methods) have been utilized, both
of which now comprise the majority of the biophysical toolbox used in high-throughput fragment
screening.
The use of NMR in screening for fragment binding has historically utilized two main
approaches: protein-detected and ligand-detected experiments. In protein-detected experiments,
pioneered by Fesik and coworkers, the target of interest is typically labeled with the NMR-active
1
5
N nucleus, and two-dimensional heteronuclear single-quantum coherence (HSQC) experiments
2
are used to monitor the protein signals for changes upon ligand binding . Conversely, typical
ligand-detected approaches use unlabeled protein at low concentrations and rely on
magnetization transfer pathways from either the protein itself (saturation transfer difference, or
STD) or from bulk water (Water-Ligand Observe via Gradient SpectroscopY, or Water-LOGSY)
3
, 4
to observe binding of the small molecule to the target of interest
. More recently, we and
1
9
others have discussed the use of F NMR spectroscopy in FBDD as a powerful tool in detecting
weak binding of small molecules, and even as a potentially general strategy for fragment
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
screening by NMR . The utility of using the fluorine nucleus as a highly sensitive detection
4
, 5, 6, 7
19
method of weak binding is not a new development
. However, here the use of F-based
fragment screening is extended to so called “second-site” approaches, in which a known ligand is
used to block one region of a target’s pocket while another region is probed for small molecule
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
9
binding using F NMR. In this example, this approach is applied to β-secretase, a potentially
high-value target implicated in the onset and progression of Alzheimer’s disease (AD).
BACE-1 is an aspartic protease that functions in the cleavage process of amyloid precursor
protein (APP), an amyloidogenic pathway that ultimately yields shorter fragments of the Aβ
peptide, including the pathogenic species Aβ40 and Aβ42, which are believed to play a primary
8, 9, 10, 11
role in the etiology of AD
. The identification of potent ligands for BACE-1 has been met
1
2, 13, 14
with some success by fragment-based approaches
. A number of papers have detailed
results of fragment screening against BACE-1 by various methods, and many of those fragments
have been optimized through structure-based design to yield potent, and even in vivo efficacious,
1
2, 13, 14, 15, 16
inhibitors of BACE-1
. In addition, recent publications have detailed extensive
structure activity relationship (SAR) efforts to obtain suitable selectivity of BACE-1 inhibitors
1
7
against cathepsin D (CatD), a related aspartic protease residing in lysosomes . It has been
observed that the acidic environment of lysosomes results in the accumulation of traditionally
17
charged BACE-1 inhibitors , resulting in much higher local concentrations of the molecules and
thereby presenting problems with off-target inhibition of other lysosomal enzymes, such as
CatD. Off-target inhibition of CatD is of great potential concern, having been proposed to be
18
involved in ocular as well as neurodegenerative side-effects . For this reason, a very high
degree of selectivity against CatD and other lysosomal enzymes is necessary, and thus, CatD
ACS Paragon Plus Environment

Page 5 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
17, 19
inhibition is a commonly used selectivity filter for small molecule inhibitors of BACE-1
.
There are some distinct structural differences in the binding pockets between BACE-1 and CatD,
the most prominent being the nature and composition of the S3 pocket in BACE-1. Utilization of
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
17
the S3 pocket by appropriate chemical matter has been shown to enhance CatD selectivity . We
postulated that further extension of a lead molecule past the S3 pocket and down into the
S3subpocket would afford additional selectivity. Thus, one of our primary aims in this work was to
use FBDD to find a second-site ligand to the S3subpocket of BACE-1 that could subsequently be
linked to the core compound that would afford both an increase in potency and, importantly, a
higher degree of CatD selectivity.
2
One of the founding principles of FBDD, as originally described using SAR by NMR , is that
leads can be rapidly derived from weakly binding fragments due to the fact that the binding
2
0, 21, 22
energies of multiples fragments can be additive.
Thus, linking a second-site ligand with
an affinity of ~ 1mM with a first-site ligand could theoretically yield a potency gain of three
22
orders of magnitude. This, of course, is neglecting any effects on affinity due to entropic
factors or from the linker design itself. The fragment-linking approach has been well
documented over the last decade and has proven to be a viable approach to small molecule drug
2
0, 21, 22, 23
discovery
. However, these approaches are not trivial, and many times several iterations
of linkers are required to provide optimal conformations of the ligand. It is often more common
to find that the observed fragment linking additivity is less than the theoretical additivity than it
2
, 23
is to find that they meet or exceed the theoretical gains in affinity
.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In attempts to apply fragment-based methods to the design of potent and selective small
molecule inhibitors of BACE-1, a molecule derived from a core fragment and resulting from
early-stage SAR was used to serve as blocking compound for a large portion of the BACE-1
pocket, thereby allowing a “second-site” fragment screen to be conducted. Using this method of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
9
fragment linking, F-NMR based fragment screening has been applied to identify chemical
matter that binds specifically to the S3 and S3subpocket of BACE-1. Upon identification of
fragments binding the “blocked” conformation of BACE-1, inter-ligand NOE NMR experiments
and molecular modeling led to the chemical linkage of the two compounds (core + fragment) and
yielded compounds with both increased on-target potency (~350x) as well as improved
selectivity (~2000x) against CatD.
Results and Discussion
1
9
The use of F NMR in screening campaigns, particularly in fragment-based approaches, has
4
, 6, 7
19
been previously detailed
. In FBDD, the use of F NMR has many benefits; the major one
being the utilization of the fluorine nucleus for a very sensitive detection tool for weak binding.
1
9
Other attributes of using F as a detection tool have been reviewed extensively by Dalvit and
4
, 7
19
others . In the realm of sensitivity, the F nucleus is almost as sensitive as that of a proton
1
(
with a gyromagnetic ratio about 0.83 that of H), and the lack of a protonated background in
1
9
F-NMR enables the use of solvents, buffers, or detergents that would normally interfere with a
1
typical H-based NMR screening campaign; thereby increasing the effective sensitivity. This
high sensitivity allows the detection of small quantities of compounds (∼20 ꢀM) in very short
19
experiment times (3–4 min). Also, The F nucleus is characterized by a large chemical shift
dispersion (∼200 ppm) and very narrow line widths (ꢁw1/2 ∼ 1–2 Hz) that allow screening to be
ACS Paragon Plus Environment

Page 7 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
1
conducted on large pools of compounds without the concern of signal overlap . Standard H-
based NMR fragment screening experiments such as STD or LOGSY are performed in pools of
2
4 19
five compounds or less . F-NMR, however, allows the screen to be conducted on pools of
compounds containing up to 20 compounds-all possessing unique chemical shifts. This
effectively negates the need for any specific pooling strategy when designing libraries.
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
In approaching the second-site fragment screen for BACE-1, a suitable blocking molecule was
needed to facilitate the identification of a true S3 pocket binder. Previously, Cheng et al.
presented data demonstrating the elaboration of a 2-aminoquinoline fragment (K ~ 900 µM) to
D
1
3
a molecule with an affinity of 11 pM . In a semi-parallel effort to the traditional structure-based
1
9
design efforts in this study, a second-site fragment screen was attempted using our in-house F
fragment library in an attempt to find additional chemical matter that would allow the linking of
a core molecule to a fragment that was targeted to occupy the S3 binding site of BACE-1.
Compound 58 from Cheng et al. (compound 1 in Table 1) was used in the fragment screens as a
blocking compound to occupy the majority of the binding site of BACE-1 (K ~ 16 nM) while
D
1
9
leaving the S3 and S3subpocket regions accessible (Figure 1A). Next, the F-NMR fragment screen
was performed with saturating concentrations (~ 5 µM) of compound 1 in the experimental
19
buffer.The NMR signals of the F nucleus are extremely sensitive to changes in the environment
5
surrounding the fluorine . When the free and protein-bound fractions of the fluorine containing
compound experience significantly different environments, broadening of the NMR signals
occurs with a concomitant decrease in signal intensity. In all, the screen yielded seven
fragments that exhibited binding to the target in the presence of the blocking compound. As
depicted in Figure 2, the fluorine NMR signals of the ligands that bind to BACE-1 can
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
experience significant linebroadening, resulting in an obvious decrease in signal intensity upon
binding to the larger protein. Following the screen, the binder with highest affinity (as
determined by the effect on the NMR lineshape and by SPR) was used in a competition binding
assay to determine if the fragment indeed bound to the desired pocket. To do this, a more potent
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
17
inhibitor (Compound 2, K ~ 600 pM, Table 1) from the aminooxazoline xanthene series that
D
occupied all of the binding pockets in question (Figure 1B) was used as a competitor with the
blocking compound 1 and the newly identified fragment (compound 3). As evidenced in Figure
2, the fragment alone failed to bind the protein, but clearly did so in the presence of the blocking
19
compound. Upon the addition of the competitor compound (compound 2), the F NMR signal
from the fragment became sharp again, suggesting that the competitor molecule replaced a
substantial fraction of the fragment in the binding site, resulting in a higher concentration of free
compound (fragment) in solution and a sharpening of the NMR signal. In addition, the fragment
failed to exhibit binding to CatD both in the presence and absence of the blocking compound
1
9
(
Figure 2). These data were highly suggestive that the F fragment was indeed binding to the S3
and/or S3subpocket. It should be noted that the displacement of the fragment was not complete in
the competition assay; this was likely due to the poor solubility of the competitor molecule.
However, in all cases tested, addition of the more potent inhibitor that fully occupied the S3 site
1
9
resulted in significant sharpening of the F signals, indicating the increased presence of unbound
ligand.
In order to determine the affinity of each fragment to BACE-1, surface plasmon resonance (SPR)
experiments were used in the presence of saturating concentrations of the blocking compound
(compound 1, 5 µM). Three molecules exhibited binding to the blocked form of BACE-1 with
ACS Paragon Plus Environment

Page 9 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
affinities in the 100-300 µM range, with four more compounds exhibiting much weaker binding
(Supplementary Figure 1). In particular, compound 3 stood out in its affinity to BACE-1 in the
presence of the blocking compound. Furthermore, this compound exhibited undetectable binding
to BACE-1 in the absence of the blocking compound (Figure 3, in agreement with the NMR data
(Figure 2), and suggesting that the presence of the blocking compound helped to define the
binding pocket for the fragment.
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
Taken together, these results were highly suggestive that compound 3 was indeed binding
specifically to the S3 and S3subpocket of BACE-1. Based on the theory of fragment additivity, two
fragments can be linked together to ultimately form a more potent compound. The extent of
potency increases, however, is dictated by factors such as linker design, rigid body entropy, and
even synergistic binding. A simplistic view to approximate the benefits of linking two fragment
2
0, 21, 23
is the simple multiplicative effect of the two affinities (K *K )
.However, it affinities
A
B
commonly fail to reach the optimal potency due to imperfect linker design or other entropic
effects. Due to the high relative affinity of compound 3 compared to the other fragments (Figure
S1), this molecule was selected as the lead fragment, and further experiments were conducted to
determine if a fragment-linking strategy was viable.
With the knowledge that compound 3 likely bound specifically to the S3 site, a strategy was
designed that would allow the linkage of compound 3 to a core molecule enabling direct access
to the S3 and S3subpocket regions. From historical structural data, it was known that replacement of
the pyridine ring of compound 1 with a toluyl ring at the 6-position of the core aminoquinoline
allowed the ortho-methyl group to flip, and face directly toward the S3 pocket (compound 5, and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
structure in Fig 1C). This particular trajectory provided a direct line into the S3 pocket with
which the linked molecule could possibly gain additional potency and selectivity against CatD.
25
In order to test this hypothesis inter-ligand NOE (ILNOE) NMR experiments were used in an
attempt to identify the proximity and orientation of the bound fragment relative to another
compound of known binding mode (Figure 4). For these experiments, a weaker 2-
aminoquinoline tool molecule with a toluyl substitution at the 6 position of 2-aminoquinoline
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
compound 4) was used as a probe (K ~ 20 µM) to allow for strong NOEs between the methyl
D
19
group and any protons of the F fragment. As shown in Figure 4, NOEs were clearly observed
between the ortho- methyl group of the tool compound (4) and the 1, 2, and 3 protons of the
chloropyridine moiety of the fragment (3). These data, along with structural data from compound
5 suggested that the ortho-methyl group of the tool compound and chloropyridine moiety of the
19
fragment were within 5-6 Å from each other and that the fluorophenyl group of the F fragment
likely extended down into the hydrophobic S3subpocket of BACE-1. These results served to guide
modeling efforts in attempts to design adequate linkers of these two compounds.
Using the spatial and orientation data obtained from the ILNOE experiments, molecular
modeling studies were used to propose potential linking approaches to attach compound 3 onto
the already potent core 2-aminoquinoline (compound 5). With the observed orientation of the
toluyl substituent (providing a nearly direct trajectory into the S3 region) coupled with the likely
binding orientation of a benzamide S3 moiety (i.e. engagement of the Gly230 carbonyl by the
amide NH and occupancy of the S3subpocket by the para-substituted arene) various rigidified
linking agents were computationally probed via docking of the lowest energy adduct conformers
into the BACE-1 active site. In all, four hybrid molecules, possessing linkers which were
ACS Paragon Plus Environment

Page 11 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
predicted to dispose the resultant pendent benzamide substituent in a manner described above,
were proposed for synthesis (Figure 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
Upon completion of the syntheses of compounds 6-9, the compounds were tested for binding to
BACE-1 by SPR. All of the compounds exhibited a reasonable improvement in affinity (Figure
6, Table 2). However, conjugation of the fragment to the core molecule did result in a modest
decrease in ligand efficiency (LE) (Table 2) with compound 9 staying close to the generally
targeted LE of 0.3. To test for potency and selectivity both in vitro and in vivo, the compounds
were tested in BACE-1 enzyme and cellular assays along with a CatD enzyme assay. (Table 2
and Figure S2). All compounds exhibited reasonable improvements in IC values compared to
50
the parent compound both in vitro and in vivo. However, compound 9 exhibited a roughly 350-
fold gain in its affinity for BACE and only a 16-fold cell shift (compared with a cell shift of 380
1
3
for the parent molecule).
Most importantly, however, this particular compound exhibited
~
2000-fold increase in selectivity against CatD (in both enzyme and cell assays) with respect to
the parent compound (see data in Table 2 and Figure S2).
To thoroughly investigate the binding mode of the linked molecules, the compounds were
1
9
submitted for X-ray crystallography in order to determine if the fluorophenyl group of the F
fragment was indeed binding in the S3 site. High quality X-ray structures were determined for
two of the compounds thereby demonstrating unequivocally that the binding mode of the final
compounds was consistent with the experimental results. In the crystal structure of BACE-1 with
compound 6 (Figure 7a), the aminoquinoline portion of the inhibitor engages the catalytic
aspartic acid residues as expected and the flexible aliphatic chain extending off from the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
aminoquinoline bends around and occupies the S1’ and S2’ pockets and forms a hydrogen bond
with Gly34, as observed in the crystal structure of compound 5. As predicted, the alkynyl
linkage is oriented such that the cyclopropyl moiety binds in the S1 pocket and the fluorophenyl
ring occupies the S3 pocket and part of the S3subpocket. In this position, compound 6 forms a
hydrogen bond between its amide NH and the backbone carbonyl oxygen atom of Gly230 and
between its amide carbonyl oxygen atom and a water molecule that in turn interacts with the
backbone of Phe108. In the crystal structure of BACE-1 with compound 9 (Figure 7B), many
similar features are observed including interactions with the protein made by the aminoquinoline
portion of the inhibitor derived from the original blocking compound. In place of the alkyne
linker of compound 6, however, compound 9 contains two methylene groups that allow a nearly
90° torsion angle to be made, enabling the pyridine ring to sit on the surface of the protein in the
S1 pocket. The edge of the pyridine ring also makes van der Waals interactions with the face of
the aminoquinoline ring. The amide NH of the inhibitor forms a hydrogen bond to Gly230 and
while the amide moiety largely fills the S3 pocket, the fluorophenyl ring pushes deeply into the
S3subpocket. These extensive interactions between compound 9 and BACE-1 likely account for its
excellent in vitro and in vivo potency, and the deeper penetration of the fluorophenyl ring of
compound 9 into the S3subpocket, relative to compound 6, likely results in its higher selectivity
over CatD. The crystallographic results also demonstrate why certain compounds, such as
compound 3, have stronger binding interactions in the presence of the blocking compound than
in its absence.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 13 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Conclusion
1
9
We have demonstrated here the utility of F-NMR in a fragment-based approach to designing
potent and selective small molecule inhibitors of BACE-1 using a fragment linking approach.
The final compound presented here was derived from a 2-aminoquinoline fragment originally
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
1
4
proposed by Astex and structure-based drug design was used to optimize this compound into a
1
3
19
highly potent molecule . Here, we demonstrate that F-NMR based fragment approaches are a
rapid means to identify additional chemical matter that can be used in fragment-linking
approaches to allow a higher degree of molecular diversity.
Materials and Methods
NMR Spectroscopy. All NMR experiments were performed on a Bruker Avance III NMR
1
spectrometers (Bruker Biospin, Billerica, MA) operating at a H frequency of 500.13 MHz or
1
9
8
00.12 MHz. F experiments were conducted using a SEF cryogenic probe equipped for
1
9
direct F detection while proton experiments were acquired at 800 MHz on a TCI cryoprobe.
19
1
One dimensional F spectra were acquired for each sample at 283 K using H decoupling with a
spectra width of 71428 Hz, an acquisition time of 917 ms, and 128 scans with a relaxation delay
of 1 s. This yielded experiment times on the order of 6 min each, and allowing 2 min for initial
temperature equilibration. This experimental set up allowed all reference and screen spectra to be
1
9
acquired in less than 24 h. The F fragment screen for BACE-1 was conducted in 50 mM
Sodium Acetate, pH 5.0, 100 mM NaCl, and 5% D2O (for field frequency lock). Protein and
compound concentrations for the fragment screen were 10 µM and 20 µM, respectively.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
25
Two-dimensional Inter-ligand NOE (ILNOE) experiments were conducted in a similar buffer
1
1
(
except using deuterated components) using a standard two-dimensional H- H NOESY pulse
sequence on an 800 MHz Bruker Avance III NMR spectrometer using a 500 ms mixing time and
spectral widths of 8169 Hz (in F1 and F2) and 2048 and 512 data points in the direct and indirect
dimensions, respectively.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
All data were processed using Topspin 3.2 (Bruker Biospin, Billerica, MA) and were then
compared visually to the reference spectra using the spectral overlay feature. Hits were identified
by signal intensity and/or chemical shift changes. Since each compound in a pool had a unique
chemical shift, hit identification was straightforward, and hit compounds could be identified by
simply matching the chemical shift of the hit compound to that found in the compound database.
Surface Plasmon Resonance Spectroscopy
Dissociation constant (K ) measurements were performed on Biacore S51 and SensiQ Pioneer
D
FE SPR instruments (GE Healthcare and SensiQ Technologies, respectively). BACE-1 protein
and inhibitors for SPR measurements were generated in-house; all other reagents were purchased
from GE Healthcare, SensiQ or Sigma-Aldrich. Glycosylated BACE-1 was reacted with sodium
periodate to oxidize cis-diol groups on sugar chains to aldehydes. The oxidized BACE-1 was
immobilized at high density (10000–12000 RU) onto either a CM5 or COOH5 sensor chip (GE
Healthcare and SensiQ Technologies, respectively) using aldehyde coupling chemistry and
resulted in surface activities close to 100% based on reference inhibitor binding. The
immobilization running buffer consisted of 10 mM HEPES pH 7.4 with 150 mM NaCl and
immobilization steps consisted of a 3–4 min EDC/NHS activation step [200 mM 1-ethyl-3-(3-
ACS Paragon Plus Environment

Page 15 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
dimethylaminopropyl) carbodiimide hydrochloride, 50 mM N-hydroxysuccinimide], 7 min 5
mM carbohydrizide in water, 7 min 1 M ethanolamine hydrochloride pH 8.5, 10 min 20 ꢀg/mL
oxidized BACE-1 in 10 mM sodium acetate pH 4.0, 7 min 100 mM sodium cyanoborohydride in
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
1
00 mM sodium acetate pH 4.0, 3 × 30 s 50 mM glycine pH 9.5 and 3 × 30 s 1 M NaCl in 100
mM NaHCO pH 9.5. The reference spot for all SPR experiments consisted of the blank surface
3
of the sensor chip treated in an identical manner as the other flow cell spot excluding the addition
of protein.
For K measurements, the buffer was replaced with 50 mM sodium acetate pH 5.0, 150 mM
D
NaCl, 0.005 (v/v) Tween-20, and 2% (v/v) DMSO. Compound stocks prepared in DMSO
(typically 20 mM) were serially diluted in running buffer and injected over the immobilized
BACE-1. Association and dissociation time were typically set to 45 s and 90 s, respectively.
However, in the case of the more potent compounds, dissociation times were extended in some
cased to 20 minutes. All SPR experiments were performed at 25 °C with a flow rate of 30
ꢀ
L/min and data collection rate of 10 Hz.
The data were processed and analyzed using either Scrubber-2 analysis software (BioLogic
Software, Campbell, Australia) or QDat (SensiQ Technologies, Oklahoma City, OK). The
sample response observed on the reference spot was subtracted from the sample response with
immobilized BACE-1 to correct for systematic noise and baseline drift. Data was solvent
corrected and the response from blank injections was used to double-reference the binding data.
The data were molecular weight normalized and K values established using either simple steady
D
analysis or single-site kinetic fit using a mass transport term (km).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Molecular Modeling.
Low-energy conformers for a variety of potential adducts possessing various rigidifying
elements (alkynyl, cyclopropyl, etc. linkages) were generated with the MMFF94 force field as
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
26
implemented in the MOE program suite while constraining the aminoquinoline warhead and
S1’-S2’ atoms to their positions found in the crystal structure of 1 (Figure 1). The resultant
conformers were then subjected to MM-GBSA minimization in the BACE-1 active site (rigid
27
protein model) using the AMBER94 / GAFF force field as implemented in AMBER 9.0
Adducts possessing the best predicted binding affinities and which best presented the pendent
benzamide NH in an orientation favorable for hydrogen bonding to the Gly230 carbonyl oxygen
were proposed for synthesis.
X-ray Crystallography. The catalytic domain of human BACE-1 was expressed as inclusion
2
8
bodies, purified, and apo crystals were grown as described previously . The co-crystal
structures of BACE-1 with compounds 1,2, and 5 bound were obtained by soaking apo crystals
in modified mother liquor solution (25 % PEG 5000 MME, 0.1 M sodium citrate, pH 6.6, 0.2 M
ammonium iodide, 3% (v/v) dimethylsulfoxide) containing 0.5 mM compound for 6 hours at
room temperature. The crystals were then transferred briefly to the same solution supplemented
with 20% (v/v) glycerol prior to flash-cooling for data collection. In order to obtain co-crystal
structures of BACE-1 with compounds 6 and 9, concentrated protein at ~6 mg/mL was mixed
with 0.5 mM inhibitor and incubated on ice for 90 min. The protein-inhibitor complexes were
crystallized by hanging drop vapor diffusion at room temperature by mixing 1 µL of protein with
1 µL of precipitant solution (1.4-1.5 M ammonium sulfate, 0.2 M lithium chloride and either 0.1
M bis-tris (pH 5.5) or 0.1 M MES (pH 6.0)); wells contained 0.5 mL of precipitant solution.
ACS Paragon Plus Environment

Page 17 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Single crystals appeared within 1 - 2 weeks. Prior to data collection, crystals were transferred
briefly into cryo solutions consisting of mother liquor supplemented with 20% (v/v) glycerol and
flash-cooled in liquid nitrogen. Diffraction data for co-crystals with compounds 9 and 12 were
collected at SER-CAT beam line 22-BM at the Advanced Photon Source using a MAR225 CCD
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
detector and λ = 1.0000 Å. Apo BACE-1 crystals soaked with compounds 1, 2, and 5 had data
collected at the Advanced Light Source, beamline 5.0.2, using an ADSC Q315 CCD detector and
2
9
λ = 1.0000 Å. Images were processed with HKL2000 and the structures were refined using
3
0
31
REFMAC . Model building was performed with COOT . Data collection and refinement
statistics appear in Table S1 of the Supporting Information.
BACE-1 Enzymatic Assay
BACE-1and CatD enzymatic activity was determined by the enhancement of fluorescence
intensity upon enzymatic cleavage of the fluorescence resonance energy transfer substrate. The
BACE-1 recognition and cleavage sequence of the substrate is derived from the reported
32
literature , and the fluorophore and quencher dyes were attached to side chain of Lys residues at
3
3
the termini of the substrate peptide. The human recombinant BACE-1 assay was performed in
50 mM acetate, pH 4.5, 8% DMSO, 100 ꢀM Genepol, and 0.002% Brij-35. In dose–response
IC assays, 10 point 1:3 serial dilutions of compound in DMSO were pre-incubated with the
5
0
enzyme for 60 min at room temperature. Subsequently, the substrate was added to initiate the
reaction. After 60 min at room temperature, the reaction was stopped by addition of 0.1 M Tris
base to raise the pH above the enzyme active range, and the increase of fluorescence intensity
was measured on Safire II microplate reader (Tecan, Männedorf, Switzerland).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Chemistry. Unless otherwise noted, all materials were obtained from commercial suppliers and
were used without further purification. All final compounds were purified to ≥95% purity as
determined by high-performance liquid chromatography (HPLC).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 19 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Synthesis of N-(1-(3-(2-bromophenyl) prop-2-yn-1-yl)cyclopropyl)-4-fluorobenzamide used
in the synthesis of Compound 6.
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 20 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Step 1: Synthesis of 3-((tert-butyldimethylsilyl) oxy) propanenitrile: To a solution of 3-
hydroxypropanenitrile (10 g, 141 mmol, 1 equiv) in tetrahydrofuran (200 mL) at 0 °C, was added
imidazole (23.94 g, 352 mmol, 2.5 equiv) followed by tert-butyldimethylsilyl chloride (TBDMS-
Cl) (25.4 g, 169 mmol, 1.2 equiv). After the addition was completed, the reaction mixture was
stirred at room temperature for 16 h. After completion of the reaction (monitored by TLC), the
mixture was quenched with water (20 mL) and extracted with ethyl acetate (2 x 100 mL). The
combined organic layers were washed with brine (50 mL), dried over Na SO and concentrated
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
4
in vacuo to give the crude material as light-yellow oil. The crude material was purified by flash
column chromatography (silica gel, 230-400 mesh), eluting with a gradient of 0 to 10% ethyl
acetate in hexane to provide 3-((tert-butyldimethylsilyl) oxy) propanenitrile (20 g, 108 mmol, 77
1
% yield) as a colorless liquid. H NMR (400 MHz, Chloroform-d) δ 3.86 (t, J = 6.4 Hz, 2H),
2.56 (t, J = 6.3 Hz, 2H), 0.93 (s, 9H), 0.11 (s, 6H).
Step 2: Synthesis of 1-(2-((tert-butyldimethylsilyl)oxy)ethyl)cyclopropan-1-amine: To a
solution of 3-((tert-butyldimethylsilyl) oxy) propanenitrile (20 g, 108 mmol, 1.0 equiv) in diethyl
ether (200 mL), was added titanium (IV) isopropoxide (53.8 mL, 183 mmol, 1.7 equiv) at 0 °C.
Ethylmagnesium bromide (90 mL, 270 mmol, 2.5 equiv, 3.0 M in diethyl ether) was added
slowly. After addition, the reaction mixture was stirred at room temperature for 1 h and again
cooled to 0 °C. To this mixture, boron trifluoride etherate was added (54.7 mL, 216 mmol, 2.0
equiv, 50 % solution) dropwise. After addition, the reaction mixture was stirred at room
temperature for 1 h. After completion of the reaction (monitored by TLC), the mixture was
quenched with 10% sodium hydroxide solution (150 mL) and extracted with ethyl acetate (2 x
1
50 mL). The combined organic layer was washed with brine (50 mL), dried over Na SO and
2 4
ACS Paragon Plus Environment

Page 21 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
concentrated in vacuo to give the crude material as brown liquid. The crude material was purified
by flash chromatography column (silica gel, 230-400 mesh), eluting with a gradient of 0 to 10%
Ethyl acetate in hexane to provide 1-(2-((tert-butyldimethylsilyl)oxy)ethyl)cyclopropan-1-amine
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
1
(
10 g, 46.4 mmol, 43 % yield) as a brown liquid. H NMR (300 MHz, Chloroform-d) δ 3.89 (d, J
= 5.9 Hz, 2H), 2.92 (s, 2H), 1.69 (d, J = 5.9 Hz, 2H), 0.91 (s, 9H), 0.77 – 0.74 (m, 2H), 0.54 –
.51 (m, 2H), 0.12 (s, 6H).
0
Step 3: Synthesis of N-(1-(2-((tert-butyldimethylsilyl) oxy) ethyl) cyclopropyl)-4-
fluorobenzamide: To a solution of 1-(2-((tert-butyldimethylsilyl)oxy)ethyl)cyclopropan-1-
amine (10 g, 46.4 mmol, 1.0 equiv) and triethylamine (9.71 mL, 69.6 mmol, 1.5 equiv) in
dichloromethane (100 mL) cooled to 0 °C, was added 4-fluorobenzoyl chloride(8.83 g, 55.7
mmol, 1.2 equiv). After addition, the reaction mixture was stirred at room temperature for 3 h.
After completion of the reaction (monitored by TLC), the mixture was quenched with water (100
mL) and extracted with dichloromethane (2 x 50 mL). The combined organic layer was washed
with brine (50 mL), dried over Na SO and concentrated in vacuo to give the crude material
2
4
(
light-yellow oil). The crude material was purified by flash column chromatography (silica gel,
2
30-400 mesh), eluting with a gradient of 0 to 10% ethyl acetate in hexane, to provide N-(1-(2-
((tert-butyldimethylsilyl) oxy) ethyl) cyclopropyl)-4-fluorobenzamide (6 g, 17.78 mmol, 38.3 %
1
yield) as light yellow solid. MS (ESI positive ion) m/z: 338.2 (M+1). H NMR (300 MHz,
Chloroform-d) δ 7.76 – 7.72 (m, 2H), 7.12 – 7.07 (m, 2H), 6.54 (s, 1H), 3.82 (t, J = 5.9 Hz, 2H),
1.90 (t, J = 6.1 Hz, 2H), 0.96 – 0.94 (m, 2H), 0.87 (s, 9H), 0.84 – 0.80 (m, 2H), 0.04 (s, 6H).
Step 4: Synthesis of 4-fluoro-N-(1-(2-hydroxyethyl)cyclopropyl)benzamide: To a solution of
N-(1-(2-((tert-butyldimethylsilyl) oxy) ethyl) cyclopropyl)-4-fluorobenzamide (6 g, 17.78 mmol,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
.0 equiv ) in tetrahydrofuran (60 mL) was added tetra-N-butylammonium fluoride (35.6 mL,
5.56 mmol, 2.0 equiv, 1.0 M in tetrahydrofuran) at room temperature and stirred at room
3
temperature for 3 h. After completion of the reaction (monitored by TLC), the reaction mixture
was quenched with water (60 mL) and extracted with ethylacetate (2 x 60 mL). The organic
layer was washed with brine (30 mL) and dried over Na SO . The solution was filtered and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
4
concentrated in vacuo to give the crude material (light-yellow solid), which was purified by flash
column chromatography (silica gel, 230-400 mesh), eluting with a gradient of 0 to 30% Ethyl
acetate in hexane to provide 4-fluoro-N-(1-(2-hydroxyethyl)cyclopropyl)benzamide (2.8 g, 12.54
1
mmol, 70.6 % yield) as an off white solid. MS (ESI positive ion) m/z: 224.2 (M+1). H NMR
(400 MHz, Chloroform-d) δ 7.79 – 7.75 (m, 2H), 7.15 – 7.10 (m, 2H), 6.54 (s, 1H), 3.76 – 3.73
(m, 2H), 1.79 – 1.76 (m, 2H), 1.00 – 0.90 (m, 4H).
Step 5: Synthesis of 4-fluoro-N-(1-(2-oxoethyl)cyclopropyl)benzamide: To a solution of 4-
fluoro-N-(1-(2-hydroxyethyl)cyclopropyl)benzamide (1.0 g, 4.48 mmol, 1.0 equiv) in
dichloromethane (30 mL), was added Dess-Martin periodinane (2.85 g, 6.72 mmol, 1.5 equiv) at
0 °C and then stirred at room temperature for 3 h. After completion of the reaction (monitored by
TLC), the reaction was quenched with 10 % sodium bicarbonate (50 mL) and extracted with
dichloromethnae (2 x 50 mL). The organic layer was washed with brine (10 mL) and dried over
Na SO . The solution was filtered and concentrated in vacuo to give the crude material (light-
2
4
yellow gum) which was purified by flash column chromatography (silica gel, 230-400 mesh),
eluting with a gradient of 0 to 20% ethyl acetate in hexane to provide 4-fluoro-N-(1-(2-
oxoethyl)cyclopropyl)benzamide (0.5 g, 2.26 mmol, 50.5 % yield), as an off white solid.
ACS Paragon Plus Environment

Page 23 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
MS (ESI positive ion) m/z: 222.2 (M+1). H NMR (400 MHz, Chloroform-d) δ 9.85 (s, 1H),
7
.76 – 7.71 (m, 2H), 7.13 – 7.08 (m, 2H), 6.77 (s, 1H), 2.92 (s, 2H), 1.04 – 1.01 (m, 2H), 0.87 –
.85 (m, 2H).
0
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 6: Synthesis of 4-fluoro-N-(1-(prop-2-yn-1-yl) cyclopropyl) benzamide: To a solution of
-fluoro-N-(1-(2-oxoethyl)cyclopropyl)benzamide (500 mg, 2.26 mmol, 1.0 equiv) and
4
potassium carbonate (1.2 g, 9.04 mmol, 4.0 equiv) in methanol (10 mL), was added dimethyl (1-
diazo-2-oxo-propyl) phosphonate (521 mg, 2.71 mmol, 1.2 equiv) at 0 °C. After addition, the
reaction mixture was stirred at room temperature for 4 h. After completion of the reaction
(monitored by TLC), the mixture was quenched with water (15 mL) and extracted with ethyl
acetate (2 x 20 mL). The organic layer was washed with brine (5 mL) and dried over Na SO .
2
4
The solution was filtered and concentrated in vacuo to give the crude material (light-yellow gum)
which was purified by flash column chromatography (silica gel, 230-400 mesh), eluting with a
gradient of 0 to 10% Ethyl acetate in hexane to provide 4-fluoro-N-(1-(prop-2-yn-1-yl)
cyclopropyl) benzamide (300 mg, 1.38 mmol, 61.1 % yield), as a light yellow solid.
1
MS (ESI positive ion) m/z: 218.2 (M+1). H NMR (300 MHz, Chloroform-d) δ 7.78 – 7.74 (m,
2H), 7.14 – 7.08 (m, 2H), 6.55 (s, 1H), 2.70 (s, 2H), 2.00 (t, J = 2.6 Hz, 1H), 0.99 – 0.94 (m,
4H).
Step 7: Synthesis of N-(1-(3-(2-bromophenyl) prop-2-yn-1-yl)cyclopropyl)-4-
fluorobenzamide: To a solution of 4-fluoro-N-(1-(prop-2-yn-1-yl) cyclopropyl) benzamide (300
mg, 1.381 mmol, 1.0 equiv) in tetrahydrofuran (10 mL), was added 1-iodo-2-bromo-benzene
(391 mg, 1.381 mmol, 1.0 equiv), triethylamine (210 mg, 2.071 mmol, 1.5 equiv), triphenyl
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
phosphine (18.11 mg, 0.069 mmol, 0.05 equiv) and copper(I) iodide (26.3 mg, 0.138 mmol, 0.1
equiv) at room temperature. The reaction mixture was degassed with nitrogen,
bis(triphenylphosphine)palladium(II)chloride (48.5 mg, 0.069 mmol, 0.05 equiv) was added, and
the reaction mixture stirred at room temperature for 4 h. After completion (monitored by TLC),
the reaction was quenched with water (15 mL), and extracted with ethyl acetate (2 x 10 mL).
The organic layer was washed with brine (3 mL), and the organic extract was dried over Na SO .
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
4
The solution was filtered and concentrated in vacuo to give the crude material. The crude product
was purified by flash column chromatography (silica gel, 230-400 mesh), eluting with a gradient
of 0 to 10% Ethyl acetate in hexane to provide N-(1-(3-(2-bromophenyl) prop-2-yn-1-
yl)cyclopropyl)-4-fluorobenzamide (250 mg, 0.672 mmol, 48.6 % yield) as an off white solid.
1
MS (ESI positive ion) m/z: 372.0 (M+). H NMR (300 MHz, Chloroform-d) δ 7.81 – 7.75 (m,
2H), 7.56 (d, J = 8.0 Hz, 1H), 7.43 – 7.40 (m, 1H), 7.24 – 7.07 (m, 4H), 6.70 (s, 1H), 2.95 (s,
2H), 1.06 – 0.96 (m, 4H).
Synthesis of Compound 6.
ACS Paragon Plus Environment

Page 25 of 69
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
Step 1: Synthesis of 6-bromo-2-((4-methoxybenzyl)amino)quinoline-3-carbaldehyde. To a
solution of 6-bromo-2-chloroquinoline-3-carbaldehyde (10 g, 37.0 mmol, 1.0 equiv) in absolute
ethanol (200 mL in 1 L sealed tube) was added 4-methoxy benzyl amine (10.56 mL, 81 mmol,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
.2 equiv) at room temperature and the mixture was heated to 90 °C for 24 h. The reaction was
not completed as indicated by thin layer chromatography (TLC). This reaction mixture was
cooled to room temperature and added 4-methoxy benzyl amine (4.8 mL, 37 mmol, 1.0 equiv),
then heated to 95 °C for additional 24 h. Once the reaction was completed (monitored by TLC),
the reaction mixture was cooled to room temperature and aqueous 6N HCl was added (50 mL)
and then heated to 90 °C for 24 h. After completion of the reaction (monitored by LC-MS), the
reaction mixture was cooled to room temperature and ethanol was removed under vacuum. The
resulting solid was washed with water (3 x 30 mL), methanol (2 x 5 mL) and dried to afford 6-
bromo-2-((4-methoxybenzyl) amino) quinoline-3-carbaldehyde (2, 10 g, 72.9%) as a yellow
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
solid. MS (ESI positive ion) m/z: 371.0, 373.0 (M+, M+2). H NMR (400 MHz, Chloroform-d) δ
10.05 (s, 1H), 8.49 (br s, 1H), 8.24 (s, 1H), 7.88 – 7.74 (m, 3H), 7.38 (d, J = 8.4 Hz, 2H), 6.90 (d,
J = 8.4 Hz, 2H), 4.88 (s, 2H), 3.85 (s, 3H).
Step 2: Synthesis of methyl 3-(6-bromo-2-((4-methoxybenzyl) amino) quinolin-3-yl)
acrylate: To a solution of 6-bromo-2-((4-methoxybenzyl)amino)quinoline-3-carbaldehyde (5 g,
13.47 mmol, 1.0 equiv) in anhydrous dicholormethane (100 mL) at 0 °C was added methyl 2-
(triphenylphosphoranylidene)acetate (4.50 g, 13.47 mmol, 1.0 equiv) under a nitrogen
atmosphere and the reaction was stirred at room temperature for 3 h. After completion of the
reaction (monitored by TLC), the reaction mixture was concentrated under vacuum. The crude
material was purified by flash column chromatography (silica gel, 230-400 mesh) using 15 to
2
0% of ethyl acetate in petroleum ether as an eluent to afford methyl 3-(6-bromo-2-((4-
methoxybenzyl)amino)quinolin-3-yl)acrylate (3.76 g, 65.3%) as a yellow solid. MS (ESI positive
1
ion) m/z: 427.0, 429.0 (M+, M+2). H NMR (300 MHz, Chloroform-d) δ 7.85 (s, 1H), 7.75 –
ACS Paragon Plus Environment

Page 27 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7
.61 (m, 4H), 7.35 (d, J = 8.76 Hz, 2H), 6.93 – 6.87 (m, 2H), 6.47 (d, J = 15.7 Hz, 1H), 5.01 (br
s, 1H), 4.74 (d, J = 5.1 Hz, 2H), 3.82 (s, 6H).
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 3: Synthesis of (E)-3-(6-bromo-2-((4-methoxybenzyl)amino)quinolin-3-yl)acrylic acid:
To a solution of compound methyl 3-(6-bromo-2-((4-methoxybenzyl) amino) quinolin-3-yl)
acrylate (3.6 g, 8.43 mmol, 1.0 equiv) in tetrahydrofuran (40 mL) and water (40 mL) was added
lithium hydroxide mono hydrate (0.404 g, 16.85 mmol, 2.0 equiv) at room temperature. After
addition, the reaction mixture was stirred at room temperature for 4 h. After completion of the
reaction (monitored by TLC), the reaction was quenched with water (40 mL) and extracted with
ethyl acetate (2 x 60 mL). The combined organic extract was dried over Na SO and
2
4
concentrated in vacuo to provide (E)-3-(6-bromo-2-((4-methoxybenzyl)amino)quinolin-3-
yl)acrylic acid (3.4 g, 8.23 mmol, 98 % yield) as a light yellow solid. MS (ESI positive ion) m/z:
1
4
13 (M+). H NMR (300 MHz, DMSO-d ) δ 8.59 (s, 1H), 8.06 (s, 1H), 7.90 – 7.85 (m, 3H), 7.40
6
(d, J = 8.6 Hz, 2H), 6.91 (d, J = 8.6 Hz, 2H), 6.64 (d, J = 15.3 Hz, 1H), 4.85 (s, 2H), 3.72 (s, 3H).
Step 4: Synthesis of (E)-3-(6-bromo-2-((4-methoxybenzyl)amino)quinolin-3-yl)-N-(3,3-
dimethylbutyl)acrylamide: To a solution of (E)-3-(6-bromo-2-((4-
methoxybenzyl)amino)quinolin-3-yl)acrylic acid (3.4 g, 8.23 mmol, 1.0 equiv) in
dichloromethane (100 mL), was added triethylamine (5.73 mL, 41.1 mmol, 5.0 equiv) and 2,4,6-
tripropyl-1,3,5,2,4,6-trioxatriphosphinane 2,4,6-trioxide (7.85 g, 12.34 mmol, 1.5 equiv) in 50%
Ethyl acetate/dichloromethane at room temperature. The mixture was then cooled to 0 °C and
3,3-dimethylbutan-1-amine (0.999 g, 9.87 mmol, 1.2 equiv) was added. After addition, the
reaction mixture was warmed up to room temperature and stirred for 16 h. After completion
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
monitored by TLC) of the reaction, the reaction was quenched with water (60 mL) and extracted
with ethyl acetate (2 x 60 mL). The organic layer was washed with brine solution (30 mL), and
the organic extract was dried over Na SO . The solution was filtered and concentrated in vacuo
2
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
to give the crude material as a light yellow solid. The crude material was purified by flash
column chromatography (silica gel, 230-400 mesh), eluting with a gradient of 0 to 30% Ethyl
acetate in hexane, to provide (E)-3-(6-bromo-2-((4-methoxybenzyl)amino)quinolin-3-yl)-N-(3,3-
dimethylbutyl)acrylamide (2.5 g, 5.04 mmol, 61.2 % yield), as a light yellow solid. MS (ESI
1
positive ion) m/z: 496.0, 498.0 (M+, M+2). H NMR (400 MHz, Chloroform-d) δ 7.78 – 7.61
(m, 5H), 7.36 (d, J = 8.5 Hz, 2H), 6.89 (d, J = 8.5 Hz, 2H), 6.38 (d, J = 15.1 Hz, 1H), 5.60 (s,
1
H), 5.10 (s, 1H), 4.74 (s, 2H), 3.84 (s, 3H), 3.44 – 3.38 (m, 2H), 1.52 – 1.48 (m, 2H), 0.98 (s,
9H).
Step 5: Synthesis of 3-(6-bromo-2-((4-methoxybenzyl)amino)quinolin-3-yl)-N-(3,3-
dimethylbutyl)propanamide: To a solution of (E)-3-(6-bromo-2-((4-
methoxybenzyl)amino)quinolin-3-yl)-N-(3,3-dimethylbutyl)acrylamide (1 g, 2.014 mmol, 1.0
equiv) in ethylacetate (20 mL), was added 10 % platinum on carbon (100 mg, 10 wt%) at room
temperature. After addition, the reaction mixture was stirred under hydrogen pressure (5 psi) at
room temperature for 4 h. After completion (monitored by LCMS), the reaction mixture was
passed through a celite bed and washed with ethyl acetate (100 mL). The filtrate was
concentrated in vacuo to yield 3-(6-bromo-2-((4-methoxybenzyl)amino)quinolin-3-yl)-N-(3,3-
dimethylbutyl)propanamide (1 g, 2.006 mmol, 100 % yield) as a light yellow solid. MS (ESI
1
positive ion) m/z: 498.0, 500.0 (M+, M+2). H NMR (400 MHz, Chloroform-d) δ 7.68 (d, J =
2
.1 Hz, 1H), 7.63 – 7.56 (m, 2H), 7.50 (s, 1H), 7.39 (d, J = 8.5 Hz, 2H), 6.89 (d, J = 8.6 Hz, 2H),
ACS Paragon Plus Environment

Page 29 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
.69 (s, 1H), 5.32 (s, 1H), 4.75 (d, J = 5.1 Hz, 2H), 3.81 (s, 3H), 3.24 – 3.18 (m, 2H), 2.93 (t, J =
.1 Hz, 2H), 2.49 (t, J = 7.1 Hz, 2H), 1.33 – 1.28 (m, 2H), 0.90 (s, 9H).
7
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 6: Synthesis of 3-(2-amino-6-bromoquinolin-3-yl)-N-(3,3-dimethylbutyl)propanamide:
To a solution of 3-(6-bromo-2-((4-methoxybenzyl)amino)quinolin-3-yl)-N-(3,3-
dimethylbutyl)propanamide (500 mg, 1.003 mmol, 1.0 equiv) in trifluroacetic acid (10 mL), was
added anisole (1096 µl, 10.03 mmol, 10 equiv) at room temperature. After addition, the reaction
mixture was stirred at 50 °C for 3 h. After completion of the reaction (monitored by TLC), the
reaction mixture was concentrated and then diluted with ethyl acetate (40 mL). The organic layer
was washed with 10% NaHCO solution (10 mL) and water (5 mL), dried over Na SO and
3
2
4
finally concentrated in vacuo to yield the crude 3-(2-amino-6-bromoquinolin-3-yl)-N-(3,3-
dimethylbutyl)propanamide (300 mg, 0.793 mmol, 79 % yield) as a light yellow solid. MS (ESI
1
positive ion) m/z: 378.0 (M+). H NMR (400 MHz, DMSO-d ) δ 7.83 – 7.80 (m, 2H), 7.79 (s,
6
1H), 7.52 (d, J = 8.8 Hz, 1H), 7.37 (d, J = 8.8 Hz, 1H), 6.49 (s, 2H), 3.06 – 3.00 (m, 2H), 2.80 (t,
J = 7.2 Hz, 2H), 2.43 (t, J = 7.2 Hz, 2H), 1.27 – 1.22 (m, 2H), 0.87 (s, 9H).
Step 7: Synthesis of 3-(2-amino-6-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)quinolin-3-
yl)-N-(3,3-dimethylbutyl)propanamide: To a solution of 3-(2-amino-6-bromoquinolin-3-yl)-N-
(3,3-dimethylbutyl)propanamide (300 mg, 0.793 mmol, 1.0 equiv) in 1,4-dioxane (6 mL), was
added potassium acetate (233 mg, 2.379 mmol, 3.0 equiv) and bis(pinacolato)diboron (242 mg,
0
.952 mmol, 1.2 equiv). The mixture was then degassed with nitrogen, and PdCl (dppf)-CH Cl
2 2 2
(32.4 mg, 0.040 mmol, 0.05 equiv) was added at room temperature. After addition, the reaction
mixture was stirred at 85 °C for 4 h. After completion of the reaction (monitored by TLC), the
ACS Paragon Plus Environment