Core refinement toward permeable β-secretase (BACE-1) inhibitors with low hERG activity

Article abstract of DOI:10.1021/jm3011349

By use of iterative design aided by predictive models for target affinity, brain permeability, and hERG activity, novel and diverse compounds based on cyclic amidine and guanidine cores were synthesized with the goal of finding BACE-1 inhibitors as a treatment for Alzheimer's disease. Since synthesis feasibility had low priority in the design of the cores, an extensive synthesis effort was needed to make the relevant compounds. Syntheses of these compounds are reported, together with physicochemical properties and structure-activity relationships based on in vitro data. Four crystal structures of diverse amidines binding in the active site are deposited and discussed. Inhibitors of BACE-1 with 3 μM to 32 nM potencies in cells are shown, together with data on in vivo brain exposure levels for four compounds. The results presented show the importance of the core structure for the profile of the final compounds.

Full text of DOI:10.1021/jm3011349

Subscriber access provided by FORDHAM UNIVERSITY
Article
Core Refinement towards Permeable ß-Secretase
BACE-1) Inhibitors with Low hERG Activity
(
Tobias Ginman, Jenny Viklund, Jonas Malmström, Jan Blid, Rikard Emond, Rickard Forsblom, Anh
Johansson, Annika Kers, Fredrik Lake, Fernando Sehgelmeble, Karin J Sterky, Margareta Bergh, Anders
Lindgren, Patrik Johansson, Fredrik Jeppsson, Johanna Fälting, Ylva Gravenfors, and Fredrik Rahm
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm3011349 • Publication Date (Web): 05 Nov 2012
Downloaded from http://pubs.acs.org on December 19, 2012
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 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Core Refinement towards Permeable βꢀSecretase
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
(BACEꢀ1) Inhibitors with Low hERG Activity
§
§
§
§
§
Tobias Ginman, Jenny Viklund, Jonas Malmström, Jan Blid, Rikard Emond, Rickard
§
§
§
§
§
Forsblom, Anh Johansson, Annika Kers, Fredrik Lake, Fernando Sehgelmeble, Karin J.
§
§
‡
#
∥
Sterky, Margareta Bergh, Anders Lindgren, Patrik Johansson, Fredrik Jeppsson, Johanna
#
§,
§
Fälting, Ylva Gravenfors * and Fredrik Rahm
§
#
∥
Department of Medicinal Chemistry; Department of Neuroscience; Department of DMPK;
‡
AstraZeneca R&D Södertälje, SEꢀ151 85, Södertälje, Sweden; Discovery Sciences;
AstraZeneca R&D Mölndal, SEꢀ431 83 Mölndal, Sweden
ABSTRACT
Using iterative design aided by predictive models for target affinity, brain permeability and
hERG activity, novel and diverse compounds based on cyclic amidine and guanidine cores were
synthesized with the goal of finding BACEꢀ1 inhibitors as a treatment for Alzheimer’s disease.
Since synthetic feasibility had low priority in the design of the cores, an extensive synthetic
effort was needed to make the relevant compounds. Syntheses of these compounds are reported,
together with physicochemical properties and structure activity relationships based on in vitro
data. Four crystal structures of diverse amidines binding in the active site are deposited and
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 83
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
discussed. Inhibitors of BACEꢀ1 with 3 ꢁM – 32 nM potencies in cells are shown, together with
data on in vivo brain exposure levels for four compounds. The results presented show the
importance of the core structure for the profile of the final compounds.
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
INTRODUCTION
Alzheimer’s disease (AD) is a progressive neurodegenerative disease, with a huge and growing
1
medical need. One of the pathological hallmarks of an AD brain is the presence of amyloid
2
plaques. The main constituent of these plaques is amyloidꢀβ (Αβ), which is produced by the
3
sequential proteolytic cleavage of amyloid precursor protein (APP). Betaꢀsite APP cleaving
enzyme (BACEꢀ1) is responsible for the first step in the processing of APP, followed by the
4ꢀ8
9
cleavage by the gammaꢀsecretase complex (γꢀsecretase). Early onset AD, as well as protection
1
0
from AD, is associated with genetic alterations in APP, thus implicating that the amyloid
pathway and alterations in the production of Aβ is important for disease. Taken together, this
indicates that BACEꢀ1 inhibition, to stop or reduce the production of Aβ, is an attractive
approach to find a treatment for Alzheimer’s disease. However, thirteen years after the cloning
and identification of BACEꢀ1, only a few BACEꢀ1 inhibitors have been tested in the early stages
1
1,12
of clinical trials.
In the search for BACEꢀ1 inhibitors, we have performed several lead generation screening
campaigns using various methods, where fragmentꢀbased lead generation has been the most
1
3
successful. The majority of the interesting structural motifs resulting from these efforts have
been based on amidine or guanidine core structures. These have been further developed into lead
series through scaffold hopping and optimization of the substituents (Figure 1).
2
ACS Paragon Plus Environment

Page 3 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
F
F
O
N
O
N
F
N
N
N
N
H N
H N
H N
2
N
2
H N
2
H N
2
2
N
N
N
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
13ꢀ21
Figure 1. Examples of five amidine or guanidine core structures developed into lead series.
Other groups have also presented amidine or guanidine lead series of BACEꢀ1 inhibitors
2
2ꢀ25
starting from hits identified from fragmentꢀbased or high throughput screens.
As previously
described, the amidine or guanidine moiety interacts efficiently with the catalytic aspartates of
BACEꢀ1 via a hydrogen bonding network. This strong interaction has allowed modifications of
2
6
other parts of the molecule to improve the overall property of the BACEꢀ1 inhibitor. Many of
these reported amidines and guanidines have, after extensive optimization, shown good potencies
2
2ꢀ25
in cellular assays and reduction of Aβ in plasma.
Recently, after overcoming problems with
2
0,21,27
limited brain exposure, some compounds have also shown reduction of Aβ in brain.
The
two hydrogen bond donors and the basic center on the amidine/guanidine core are however not
optimal for high brain exposure and many of the reported amidines suffer from Pꢀglycoprotein
28
(
Pꢀgp)ꢀmediated transport out of the brain, so called efflux. A low efflux has been shown to be
2
9
most important to achieve sufficient brain exposure in vivo. Other properties of the
amidine/guanidine inhibitors reported, which limit their exposure in brain, have been large size,
1
1
high plasma protein binding or high clearance.
In addition, the general properties of many of the described BACEꢀ1 inhibitors, featuring a
basic functionality and being quite lipophilic, coincides with properties correlated to inhibition of
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 83
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
30
the human etherꢀàꢀgoꢀgo related gene (hERG)ꢀencoded potassium ion channel. So far, only few
2
0,31
data points from hERG inhibition have been reported for BACEꢀ1 inhibitors.
This work describes our continued research in amidineꢀ and guanidineꢀbased BACEꢀ1
inhibitors. In light of the above mentioned challenges, we reasoned that the inherent properties of
the cores would be crucial for the profile of the final test compounds, and consequently we
focused on refining the amidineꢀcontaining core. Furthermore, we recognized the importance to
focus on optimization of the core early in the lead generation phase, since the selection of the
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
3
2
core is an important step, defining the framework for future lead optimization.
We thus set out to address target affinity, permeability and hERG activity in a multiꢀparameter
3
3,34
optimization fashion.
Initially, new core structures were designed from the generic amidine
structure in Figure 2. The design was aided by predictive chemistry, and a workflow based on
three in silico models was established to aid in the prioritization of a number of core structures
3
5
for synthesis. First, a novel tool based on quantum mechanical calculations was developed to
3
6
facilitate predictions of BACEꢀ1 affinity. Second, a predictive permeability tool was
35
developed, with suggested limits for pKa and lipophilicity. Third, to predict hERG activity, a
3
5
QSAR model based on pKa and lipophilicity and size was established.
The new core structures were designed iteratively, with the aim to achieve promising in silico
profiles predicting that the final test compounds would have a balanced profile regarding the
optimization parameters listed above. With the help of these predictive models, a
physicochemical space of interest, spanned by limits in pKa and lipophilicity, was suggested.
Unfortunately, these properties are inversely correlated when it comes to optimization of target
3
5
activity, permeability and hERG activity. In order to cover the interesting physicochemical
space, aiming to find cores with the most wellꢀbalanced set of properties, a number of cores were
4
ACS Paragon Plus Environment

Page 5 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
designed and selected for synthesis. More details on the selection workflow, predictions, and
analyses of how the inherent physicochemical properties of the core structures relate to in vitro
3
5
data for the test compounds, are described in a companion paper.
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 core structures were decorated with a selection of substituents from a preꢀdefined set (R1
and R2, Figure 2). These substituents were collected from a training set of previous inꢀhouse
series and literature compounds, and were chosen to facilitate coreꢀtoꢀcore comparisons of in
3
5
vitro results for leadꢀlike test compounds via matched pair analysis. In addition, the Rꢀgroups
were also selected for contributing equally to target potency, and carefully combined in order to
37
keep lipophilicity of the test compounds leadꢀlike, with cLogP between 1 and 3.5. Thus, we
envisaged quick assessments of whether the properties of the new cores were promising enough
for further lead optimization efforts.
A
A
R1
H N
N
2
R2
Figure 2. Amidine substructure with the “core” indicated on green background (A = any atom).
The R1ꢀ and R2ꢀsubstituents (and a phenyl group) added to each core were selected to allow for
3
5
match pair analyses. The R2ꢀsubstituent in the meta position is a common feature of known
1
9ꢀ20
BACEꢀ1 inhibitors for reaching into the S3 subꢀpocket of the active site.
Synthetic feasibility of the cores was intentionally given a lower selection priority. The
synthetic method developed for each core was used to synthesize just a few representative
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 83
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
compounds and the diversity of the cores made it difficult to use common intermediates. The
core structures representing the synthesized compounds are presented in Figure 3.
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
O
S
O
S
S
O
F
O
S
H N
N
H N
N
2
2
H N
N
H N
N
H
N
N
H N
N
H N
N
H N
N
2
2
2
2
2
2
A-1
A-2
A-3
B-1
B-2
B-3
B-4
B-Ref
O
N
S
N
O
S
N
N
N
N
N
N
H N
H N
H N
H N
H N
2
N
2
N
2
N
2
N
2
H N
2
N
H N
H N
H N
N
2
N
2
N
2
C-1
C-2
C-3
D-1
D-2
D-3
D-4
D-5
D-6
Figure 3. Core structures representing the compounds included in this report, divided into four
structural subgroups; A: 6ꢀmembered bicyclic cores, B: 6ꢀmembered monocyclic cores, C: 5ꢀ
membered bicyclic cores, D: 5ꢀmembered monocyclic cores. BꢀRef represents a compound that
23
have reached clinical Phase I studies. During this work, five of the cores (Aꢀ3, Bꢀ1, Bꢀ2, Bꢀ3
3
8,39,40
and Cꢀ1) were independently published by others.
SYNTHETIC CHEMISTRY
Within the iterative design cycles using the predictive in silico models, more than ten cores
35
were selected for synthesis. The synthesis of compounds representing cores of subtype A were
41
initiated prior to the setup of the predictive workflow. Herein we describe the successful
synthesis of compounds representing all of these cores, divided into their subgroups as described
in Figure 3.
6-Membered bicyclic amidines (core sub-type A). Our synthetic efforts started by preparing
4
1
core types Aꢀ1 and Aꢀ2. Attempts for improvements of this core subꢀtype, to decrease the
6
ACS Paragon Plus Environment

Page 7 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
20,21
lipophilicity and lower the pKa were inspired by previous work.
The synthesis of two
examples (8a and 8b) is described in detail in Scheme 1. The first step involved a condensation
of commercially available 1ꢀ(3ꢀbromophenyl)ethanone (1) and 2ꢀhydroxybenzamide (2) under
acidic conditions, continuously distilling off the formed water using a DeanꢀStark trap to form
the lactam 3. Thioamide 4 could then be obtained by treating lactam 3 with P S in pyridine.
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
5
Amidine 5 was then obtained by treating the thioamide 4 with ammonia in methanol in the
presence of a mild oxidant (tꢀBuOOH). However, the thiolactam 4 was somewhat prone to
hydrolysis which made the conversion to the amidine 5 hard to control, so a preferred alternative
route to amidine 5 was developed as outlined in Scheme 1. Thus, the lactam 3 was converted to
the chloroimidate 6 and the amidine 5 was then obtained in two steps, via the hydroxylamine 7.
The final compounds could then be obtained by a SuzukiꢀMiyaura coupling.
a
Scheme 1. Synthesis of core subꢀtypes Aꢀ1 and Aꢀ2
S
O
c
b
NH
NH2
O
Br
OH
2
a
O
O
4
NH2
O
Br
NH
N
O
Br
Br
1
3
5
Cl
O
O
N
HN
O
d
N
e
f
Br
Br
6
7
NH2
g
N
O
R
8
8
a R=5-pyrimidyl
b R=3-pyridyl
a
Reagents and conditions: (a) pꢀtoluenesulfonic acid monohydrate, toluene, reflux, DeanꢀStark.
b) P S , Pyridine, 120 °C (c) NH (7 M in MeOH), tꢀBuOOH (aq) (70%). (d) PCl , POCl , 50
(
2
5
3
5
3
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 83
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
°
C (e) MeONH Cl, DIPEA, DMF, 90 °C. (f) Zinc, AcOH, rt. (g) Pyrimidineꢀ5ꢀylboronic acid or
3
pyridineꢀ3ꢀylboronic acid, bis(triphenylphosphine)palladium(II)chloride, Na CO (1 M aq),
DME:H O (2:1), 80 °C.
2
3
2
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 evaluate the effect of the size of the nonꢀaromatic substituent, compound 9, with a slightly
larger R1ꢀgroup, was synthesized according to Scheme 1 starting from 1ꢀ(3ꢀ
bromophenyl)propanꢀ1ꢀone. It was also interesting to introduce an electron withdrawing fluoride
on the fused phenyl ring of the core to lower the pKa. The syntheses of two such examples (10a
and 10b) were performed according to Scheme 1.
NH2
F
NH2
N
N
O
O
R
N
N
10a R=5-pyrimidyl
10b R=3-pyridyl
9
One strategy that was in common for the different syntheses described below was that the
SuzukiꢀMiyaura reaction was preferably used in the last step to create multiple compounds from
a common intermediate. However, some amidines with high pKa values demanded high catalyst
loading to achieve complete conversion, probably due to catalyst poisoning. In this case, the
SuzukiꢀMiyaura reaction was performed on the lactam instead.
In the case of cores containing a thioether or a baseꢀsensitive amidine or lactam, the Suzukiꢀ
Miyaura reaction was run as the first step. Coupling of the acetophenone 1 with three various
boronic acids (Scheme 2), gave three intermediates (11a-c) that were commonly used in several
syntheses described below. The metaꢀbromide present in all starting materials was substituted for
an aryl ring using the corresponding boronic acid, a source of palladium(II), a base and the
8
ACS Paragon Plus Environment

Page 9 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
reaction was performed in an aqueous solvent at an elevated temperature. This method is referred
to as standard SuzukiꢀMiyaura conditions in the following text.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Scheme 2. Synthesis of compounds 11a-c
O
O
Br
Ar
a
1
11a Ar=5-pyrimidyl
1b Ar=3-methoxyphenyl
1c Ar=3-pyridyl
1
1
a
Reagents and conditions: (a) Standard SuzukiꢀMiyaura conditions using the corresponding
boronic acid, a source of palladium(II) and a base. The reaction was run in an aqueous solvent
mixture at 80 °C.
The endoꢀcyclic oxygen of the 2Hꢀbenzo[e][1,3]oxazinꢀ4ꢀamine core was replaced with sulfur
to generate a 2Hꢀbenzo[e][1,3]thiazinꢀ4ꢀamine (15) (core type A-3 in Figure 3). The synthesis is
described in detail in Scheme 3. The synthesis started from methyl 2ꢀmercaptobenzoate (12)
which was condensed with compound 11a and ammonia to form the lactam 13. The lactam was
then converted to the cyclic amidine 15 in two steps via the thiolactam 14.
a
Scheme 3. Synthesis of compound 15
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 83
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
O
S
S
S
O
a
NH
b
NH
c
SH
N
N
12
1
3
N
14
N
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
NH2
N
S
N
15
N
a
Reagents and conditions: (a) 11a, NH (7 M in MeOH), 150 °C. (b) P S , Pyridine, 120 °C. (c)
NH (7 M in MeOH), 90 °C.
3
2 5
3
6
-Membered monocyclic amidines (core sub-type B and B-ref). The synthesis of
compounds 19a-b started with the condensation of compound 1 with 3ꢀhydroxypropanamide
4
2
under acidic conditions, continuously distilling off the formed water using a DeanꢀStark trap to
form the lactam 16. The lactam was then converted to the cyclic amidine 18 in two steps via the
thiolactam 17. In the last step, the bromide (18) was coupled with an aryl ring using standard
SuzukiꢀMiyaura conditions to obtain compounds 19a-b (Scheme 4).
a
Scheme 4. Synthesis of core subꢀtype Bꢀ1
S
O
O
NH2
NH2
NH
b
NH
c
N
d
a
N
1
O
O
Br
O
Br
Br
R
1
6
17
1
1
9a R= 3-methoxyphenyl
9b R= 3-chlorophenyl
18
a
Reagents and conditions: (a) 3ꢀHydroxypropanamide, pTSA, toluene, reflux, DeanꢀStark . (b)
P S , HMDO, 1,2ꢀdichloroethane, 110 °C. (c) NH (7 M in MeOH), 40 °C. (d) 3ꢀ
2
5
3
10
ACS Paragon Plus Environment

Page 11 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
methoxyphenylboronic
acid
or
3ꢀchlorophenylboronic
acid,
1,1'ꢀ
bis(diphenylphosphino)ferroceneꢀpalladium(II)dichloride dichloromethane complex, cesium
carbonate, DME:H O:EtOH (6:3:1), 120 °C.
2
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
To make the corresponding sulfur analogs, methyl 3ꢀmercaptopropanoate (20) was converted
to the amide 21 by treatment with concentrated ammonium hydroxide (Scheme 5). The amide
was then refluxed in toluene together with either compound 11a or 11b under acidic conditions,
continuously removing the formed water by azeotrope distillation to form the lactams 22a and
22b, respectively. The respective lactams were then converted to the cyclic amidines 24a and
24b in two steps via the thiolactams 23a and 23b.
a
Scheme 5. Synthesis of core subꢀtype Bꢀ2
O
S
O
O
NH
a
b
c
H N
SH
O
SH
2
R
2
0
2
2
2a R=5-pyrimidyl
2b R=3-methoxyphenyl
21
S
S
NH2
NH
d
N
S
R
R
2
2
3a R=5-pyrimidyl
3b R=3-methoxyphenyl
24a R=5-pyrimidyl
24b R=3-methoxyphenyl
a
Reagents and conditions: (a) Ammonium hydroxide 25%, rt. (b) 11a or 11b, pTSA, toluene,
reflux, DeanꢀStark . (c) P S , HMDO, 1,2ꢀdichloroethane, 110 °C. (d) NH (7 M in MeOH), rt.
2
5
3
Moving the oxygen and sulfur one position further away from the carbon holding the R1ꢀgroup
changed the synthetic route completely. The synthesis of compound 30 started with ketone 11b
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 83
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
which was condensed by heating with potassium cyanide and ammonium carbonate in ethanol to
4
1
form the hydantoin 25 (Scheme 6). The hydantoin was then hydrolyzed by barium hydroxide at
4
1
high temperature to give the amino acid 26. The amino acid 26 was reduced with BH3
generated in situ from sodium borohydride and iodine) in tetrahydrofuran to form the amino
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
(
4
3
alcohol 27. The lactam 28 was formed by coupling the amine 27 with chloroacetyl chloride and
diisopropylethylamine in tetrahydrofuran and subsequent ring closure with aqueous potassium
3
9
hydroxide. The lactam 28 was then converted to the cyclic amidine 30 in two steps via the
thiolactam 29.
a
Scheme 6. Synthesis of compound 30
O
H
NH2
N
b
a
c
HO
HN
1
1b
O
O
O
O
2
5
26
O
S
NH2
d
NH
e
NH
HO
O
O
O
O
O
27
28
29
NH2
f
N
O
O
30
12
ACS Paragon Plus Environment

Page 13 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Reagents and conditions: (a) KCN, (NH ) CO , EtOH, 70 °C. (b) Ba(OH) , water, 150 °C. (c)
4
2
3
2
NaBH , I , THF, 0 °Cꢀreflux. (d) 1. Chloroacetyl chloride, DIPEA, THF, 0 °Cꢀrt, 2. KOH, water,
rtꢀ60 °C. (e) P S , HMDO, 100 °C. (f) NH (7 M in MeOH), 40 °C.
4
2
2
5
3
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 corresponding sulfur analog (core type B-1 in Figure 3) followed a similar route. The
synthesis of example 34 is described in detail in Scheme 7. Compound 27 was treated with
chlorosulfonic acid in diethyl ether to form compound 31. Residual reflux in a twoꢀphase system
4
4
of aqueous sodium hydroxide and toluene afforded the aziridine 32 by a Wenker reaction.
Treatment with methyl 2ꢀmercaptoacetate gave a ring expansion to the sixꢀmember lactam 33.
Finally, the lactam 33 was converted to the cyclic amidine 34 in two steps.
a
Scheme 7. Synthesis of compound 34
NH2
NH2
a
O
b
HO
O
S
O
O
HO
O
27
31
O
NH2
H
c
d, e
NH
N
N
S
O
S
O
O
3
2
33
34
a
Reagents and conditions: (a) Chlorosulfonic acid, Et O, rt. (b) NaOH (aq), toluene, reflux. (c)
2
Methyl 2ꢀmercaptoacetate, DIPEA, DMF, 60 °C. (d) P S , HMDO, 100 °C, (e) NH (7 M in
2
5
3
MeOH), 50 °C.
2
3
Compound 37, based on an already known core structure was synthesized to facilitate a
match pair analysis (B-ref). The synthesis of compound 37 is outlined in Scheme 8.
Acetophenone 11b was treated with vinylmagnesium bromide in 2ꢀmethyl tetrahydrofuran at ꢀ
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 83
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
o
7
8 C to give the tertiary alcohol 35. The alcohol was treated with thiourea under acidic
conditions to give compound 36 which was set up for an acidic ring closure to give the final
4
5
compound 37.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 8. Synthesis of compound 37
NH
OH
H N
S
b
2
a
1
1b
O
O
3
5
36
NH2
S
N
c
O
37
a
Reagents and conditions: (a) Vinylmagnesium bromide, 2ꢀMeꢀTHF, ꢀ78 °Cꢀrt. (b) Thiourea,
HCl, AcOH, 45 °C. (c) MeSO H, TFA, rt.
3
5-membered bicyclic amidines (core sub-type C). The synthesis of fused thiazole 41, an
analog of one of the reference compounds (Supporting information), is described in detail in
Scheme 9. In the first step, the commercially available compound 38 was converted to
sulfinimide 39, a stabilized imine, using 2ꢀmethylꢀ2ꢀpropanesulfinamide and titanium(IV)
ethoxide which was heated to reflux in 2ꢀmethyl tetrahydrofuran. The sulfinimide 39 was then
4
0
reacted with the C5ꢀanion of 2ꢀmethylꢀ1,3ꢀthiazoleꢀ4ꢀcarbonitrile. Upon removal of the
protecting group of the amine with hydrochloric acid, compound 40 was formed by spontaneous
ring closure. Finally, the bromide (40) was coupled with an aryl ring using standard Suzukiꢀ
Miyaura conditions to yield compound 41.
14
ACS Paragon Plus Environment

Page 15 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 9. Synthesis of compound 41
O
O
S
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
N
S
N
a
b
c
Br
Br
N
O
O
H N
2
3
8
3
9
Br
4
0
O
S
N
N
H N
2
N
N
41
a
Reagents and conditions: (a) 2ꢀMethylꢀ2ꢀpropanesulfinamide, titanium(IV) ethoxide, 2ꢀMeꢀ
THF, reflux. (b) 1. 2ꢀMethylꢀ1,3ꢀthiazoleꢀ4ꢀcarbonitrile, 2,2,6,6ꢀtetramethylpiperidineylꢀ
magnesium chloride lithium chloride (1 M in toluene/THF), THF, ꢀ25 °Cꢀrt. 2. HCl (1 M in
Et O), MeOH, rt. (c) Pyrimidineꢀ5ꢀylboronic acid, [1,1'ꢀbis(diphenylphosphino)ferrocene]ꢀ
2
palladium(II) chloride, K CO , DME:H O:EtOH (6:3:1), 100 °C.
2
3
2
A related structure having a fused pyrazole in the bicyclic core was also proposed. Having one
of its nitrogens at the point of fusion making it a cyclic guanidine, made the synthesis somewhat
different. The synthesis is described in detail in Scheme 10. Starting with ethyl 3ꢀmethylꢀ1Hꢀ
pyrazoleꢀ5ꢀcarboxylate (42), the ester was converted to the Weinreb amide 43 using
isopropylmagnesium chloride. The amide was then converted to the ketone 44 by reaction with
the monoꢀanion formed from 1,3ꢀdibromobenzene and nꢀbutyllithium at ꢀ78 °C in 2ꢀmethyl
tetrahydrofuran. The ketone 44 was reacted with 4ꢀmethoxyphenylmagnesium bromide with
copper(I) iodide as catalyst to form the tertiary alcohol 45. Heating the alcohol with thiourea in
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 83
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
acetic acid formed the cyclic guanidine 46. Finally, the bromide was replaced with an aryl ring
using standard SuzukiꢀMiyaura conditions to yield compound 47.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 10. Synthesis of compound 47
O
O
O
O
Br
a
b
c
O
N
N
N
N
N
N
N
H
H
H
43
44
4
2
O
O
Br
H
N
N
OH
N
N
N
N
N
d
e
H N
N
H N
N
2
2
Br
O
N
4
5
46
47
a
Reagents and conditions: (a) N,OꢀDimethylhydroxylamine hydrochloride, isoꢀ
propylmagnesium chloride (2 M in THF), 2ꢀMeꢀTHF, ꢀ30 °C; (b) 1,3ꢀDibromobenzene, nꢀ
Butyllithium (1.6 M in hexanes), 2ꢀMeꢀTHF, ꢀ78 to ꢀ30 °C; (c) 4ꢀMethoxyphenylmagnesium
bromide (1 M in THF), Copper(I) iodide, 2ꢀMeꢀTHF, 0 °Cꢀrt; (d) Thiourea, acetic acid, 100 °C;
(e) Pyrimidineꢀ5ꢀboronic acid, Bis(triphenylphosphine)palladium(II)chloride, Sodium carbonate
(1 M aq), DME:H O (2:1), 80 °C.
2
The synthesis of compound 54, also bicyclic but where the fused ring is aliphatic is outlined in
Scheme 11. The amide 49 was first prepared by treating ester 48 with ammonia in methanol at
elevated temperature. The piperidine was Bocꢀprotected to yield compound 50. The thioamide
was then prepared by treating compound 50 with phosphorus pentasulfide and
hexamethyldisiloxane. The piperidine was deprotected under acidic condition to obtain 52,
which was then treated with 3'ꢀbromoacetophenone (1) and ammonia to obtain the racemic
16
ACS Paragon Plus Environment

Page 17 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
bicyclic amidine 53. The final step was then a standard SuzukiꢀMiyaura reaction to obtain
diastereomers (R*,S*)ꢀ54 and (R*,R*)ꢀ54 after separation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Scheme 11. Synthesis of compounds (R*,S*)ꢀ54 and (R*,R*)ꢀ54
O
O
H
H
N
N
a
NH2
b
c
O
NH2
N
O
O
O
4
8
49
5
0
H N
2
d
e
N
N
NH2
+
NH2
N
Br
NH2
Cl
S
S
O
O
5
3
5
1
52
H N
H
H N
H
2
2
N
N
f
N
N
N
N
F
F
(R*, S*)ꢀ54
(R*, R*)ꢀ54
a
Reagents and conditions: a) 7 M NH in MeOH, 90 °C, 3 days, 99 %; b) DIPEA, THF:DMF
1:1), diꢀtertꢀbutyl dicarbonate, rt, on, quant.; c) P S , HMDO, THF, 110 °C, 30 min, 71 %; d) 5
M HCl in iPrOH, MeOH, rt on, 67 %; e) 3'ꢀbromoacetophenone (1), 7 M NH in MeOH, 50 °C,
3
(
2 5
3
on, 29 %; f) 5ꢀfluoropyridinꢀ3ꢀylboronic acid, cat. (PPh ) PdCl , sat. NaHCO , DME:H O (2:1),
3 2
2
3
2
80 °C, 1.5 h, (3R*,8aS*) 15 % and (3R*,8aR*) 23 %.
5
-membered monocyclic amidines (core sub-type D). In an effort to investigate smaller
cores, monocyclic 5ꢀmembered cores were synthesized. The synthesis of structure 58 is
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 83
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
described in detail in Scheme 12. In the first step, compound 11b was reacted with methyl 2ꢀ
mercaptoacetate (55) using ammonia and microwave heating to form the lactam 56. The lactam
was then converted to the cyclic amidine 58 in two steps via the thiolactam 57.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 12. Synthesis of compound 58
S
S
S
O
a
b
c
SH
N
N
N
O
O
S
H N
O
2
H
H
5
5
O
O
56
5
7
58
a
Reagents and conditions: (a) 11b, NH (7 M in MeOH), 150 °C. (b) P S , Pyridine, 120 °C.
3
2 5
(c) NH (7 M in MeOH), 60 °C.
3
However, this route had its limitations and only worked for acetophenones in a satisfactory
manner. Trying to use more hindered ketones like substituted benzophenones gave little or no
46
product. The problem was solved by using the preformed imine prior to the cyclization reaction
and running the reaction without solvent. This method was used in the synthesis of compound 64
as outlined in Scheme 13. Starting with 1ꢀbromoꢀ3ꢀiodobenzene (59) and pyridineꢀ3ꢀylboronic
acid under standard SuzukiꢀMiyaura conditions afforded the biaryl 60. The bromide (60) was
converted to the corresponding anion using nꢀbutyllithium at ꢀ78 °C in diethyl ether and was
reacted with 4ꢀmethoxybenzonitrile to form imine 61. The imine (61) was stirred neat with
methyl 2ꢀmercaptoacetate and dimethylamine hydrochloride at room temperature to form the
lactam 62, which was then converted to the cyclic amidine 64 in two steps via the thiolactam 63.
a
Scheme 13. Synthesis of compound 64
18
ACS Paragon Plus Environment

Page 19 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
NH
I
Br
a
N
Br
b
N
O
59
6
0
61
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
O
S
S
S
c
d
e
N
N
N
O
S
H N
2
H
H
N
N
N
62
63
64
a
Reagents and conditions: (a) Pyridineꢀ3ꢀylboronic acid, bis(triphenylphosphine)palladium(II)
chloride, Sodium carbonate (1 M aq), DME:H O (2:1), 80 °C. (b) nꢀBuLi, 4ꢀ
methoxybenzonitrile, Et O, ꢀ78 °Cꢀrt. (c) methyl 2ꢀmercaptoacetate, dimethylamine
2
2
hydrochloride, rt. (d) P S , Pyridine, 120 °C. (e) NH (7 M in MeOH), 60 °C.
2
5
3
Compound 65 was prepared following the same reaction sequence as described in Scheme 13,
using 3ꢀbromo benzonitrile, 3ꢀmethoxy phenylboronic acid and 4ꢀbromo pyridine.
N
S
N
H N
2
O
6
5
To synthesize the corresponding nitrogen analog 2,5ꢀdihydroꢀ1Hꢀimidazolꢀ4ꢀamine, a
monocyclic analog of 54, new chemistry was developed. The synthesis of compound 71 is
described in detail in Scheme 14. 2ꢀChloroacetamide (66) was reacted with ethylamine to make
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 83
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
amine 67. It was then protected as the benzyl carbamate 68 and the amide converted to the
thioamide 69 by reaction with phosphorus pentasulfide in dioxane by microwave heating. The
protecting group was removed with hydrogen bromide in acetic acid and intermediate 70
precipitated in diethyl ether. Compound 70 was reacted with compound 11c in methanol with 7
M ammonia to give the cyclic amidine 71. This core suffered from instability. To improve
stability, the nonꢀamidine nitrogen was acylated. The synthesis the acylated analog of 71,
compound 77 is outlined in Scheme 15. In the same fashion as for the tertiary amine 71, the
secondary amine 76 could be made from the aminothioacetamide 75, generated from compound
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
74, and ketone 11b. The amine could then be acylated with acetic anhydride. Residual treatment
with ammonia removed all acylation on the amidine itself to give the final compound 77.
a
Scheme 14. Synthesis of compound 71
O
O
O
O
O
a
b
c
H
Cl
N
N
H N
H N
H N
2
2
2
66
6
7
68
O
O
N
S
S
d
e
H
N
N
H N
N
H N
H N
2
2
2
69
70
N
71
20
ACS Paragon Plus Environment

Page 21 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Reagents and conditions: (a) Ethylamine (70% in water), rt. (b) NaHCO (aq), benzyl
3
chloroformate, dioxane, 0 °Cꢀrt (c) P S , dioxane, 95 °C. (d) HBr (33% in AcOH), rt. (e) 11c,
NH (7 M in MeOH), 60 °C.
2
5
3
a
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 15. Synthesis of compound 77
O
O
O
O
O
O
S
a
b
c
NH2
NH
NH
H N
H N
H N
2
2
2
7
2
74
73
O
H
N
N
S
N
N
H N
H N
2
2
d
e
NH2
H N
2
75
O
O
76
7
7
a
Reagents and conditions: (a) NaHCO (aq), benzyl chloroformate, dioxane, 0 °Cꢀrt (b) P S ,
dioxane, 95 °C. (c) HBr (33% in AcOH), rt. (d) bromide 11b, NH (7 M in MeOH), rt. (e) 1.
Ac O, TEA. 2. NH (7 M in MeOH).
3
2 5
3
2
3
To afford a cyclic amidine core without an additional heteroatom in the ring (D-4) compound
83 was synthesized, as described in detail in Scheme 16. 3ꢀBromobenzaldehyde (78) was reacted
47
in a Stetter reaction with sodium cyanide and acrylonitrile. The formed ketone 79 was reacted
with the anion prepared from 1ꢀbromoꢀ4ꢀmethoxybenzene and nꢀbutyllithium to form the tertiary
4
8
alcohol 80. This was now set up for an intramolecular Ritter reaction, which was performed in
methanesulfonic acid to form the lactam 81. The bromide (81) was reacted with a boronic acid
using standard SuzukiꢀMiyaura conditions to form the lactam 82 which was finally converted to
the cyclic amidine 83 in two steps.
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 83
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
a
Scheme 16. Synthesis of compound 83
O
O
OH
a
b
c
Br
Br
Br
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
H
N
N
7
8
7
9
80
O
O
O
O
d
e
O
N
O
N
H N
N
2
H
H
Br
N
N
8
1
N
N
8
2
83
a
Reagents and conditions: (a) NaCN, acrylonitrile, MeCN, rt. (b) 1ꢀBromoꢀ4ꢀmethoxybenzene,
nꢀBuLi, THF, ꢀ78 Cꢀ5 °C. (c) MsOH, 60 °C. (d) Pyrimidineꢀ5ꢀylboronic acid,
bis(triphenylphosphine)palladium(II) chloride, Sodium carbonate (1 M aq), DME:H O (2:1), 80
o
2
°
C. (e) 1. P S , Pyridine, 120 °C, 2. NH (7 M in MeOH), tꢀBuOOH (70%), rt.
2 5 3
To make the similar core with a methoxy group in the alphaꢀposition another synthetic route
was developed. The synthesis of compound 91 is outlined in Scheme 17. Compound 38 was first
converted to the oxime 84 with hydroxylamine hydrochloride. In order to activate the oxime, the
nitrogen was alkylated with a methylthiomethyl group, which also acted as a protecting group,
to form compound 85. Reaction with methyl acrylate in a [3+2] cycloaddition gave the
isoxazolidine 86. The protecting group was removed with sulfuric acid to form intermediate 87
which, upon treatment with molybdenum hexacarbonyl, afforded the reductive rearrangement to
the alpha hydroxy lactam 88. The hydroxy group was alkylated with methyl iodide in the
presence of silver oxide to give the alpha methoxy lactam 89. The bromide was reacted with a
22
ACS Paragon Plus Environment

Page 23 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
boronic acid using standard SuzukiꢀMiyaura conditions to obtain 90 which was finally converted
to the cyclic amidine 91 in two steps. The diastereomers of 91 were separated by chiral SFC.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Scheme 17. Synthesis of compound 91
S
HO
O
+
O
N
N
a
b
c
Br
Br
Br
O
O
O
38
8
4
85
O
O
O
O
O
O
O
S
N
d
NH
e
HO
f
Br
Br
N
O
H
O
O
8
6
Br
8
7
88
O
O
O
O
O
O
g
h
N
N
N
O
H
O
H
H N
2
Br
N
N
89
9
0
91
a
Reagents and conditions: (a) Hydroxylamine hydrochloride, pyridine, MeOH, 50 °C. (b)
DCC, DMSO, TFA, benzene, 0 °Cꢀrt. (c) methyl acrylate, 150 °C. (d) H SO , MeOH, 60 °C. (e)
Mo(CO) , MeCN, water, reflux. (f) MeI, Ag O, MeCN, 40 °C. (g) Pyridinꢀ3ꢀylboronic acid,
tetrakis(triphenylphosphine)palladium, K CO , dioxane, water, 90 °C. (h) 1. P S , pyridine, 115
C, 2. NH (7 M in MeOH), tꢀBuOOH (70% in water), rt.
2
4
6
2
2
3
2 5
°
3
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 83
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
To make the 5ꢀmembered cyclic guanidine core, D-6, the synthesis started from methyl 2ꢀ
aminoꢀ2ꢀ(3ꢀbromophenyl)propanoate (92), which was acylated with acetic anhydride to form
intermediate 93 which was ringꢀclosed in a Dieckmann condensation to afford the βꢀketo lactam
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
49
9
4 (Scheme 18). The lactam was decarboxylated with concentrated hydrochloric acid to give
the amino ketone 95 which was reacted with Fmocꢀisothiocyanate to give the protected thiourea
6. The protecting group was removed with ammonia (7 M in MeOH) to afford the free thiourea
9
which then ringꢀclosed to the cyclic thiourea. Addition of tertꢀbutyl hydroperoxide then gave the
cyclic guanidine 97. In the last step, the bromide was reacted with a boronic acid under standard
SuzukiꢀMiyaura conditions to yield the final compound 98. This compound was later
demonstrated to be unstable under acidic and neutral conditions.
a
Scheme 18. Synthesis of compound 98
O
NH
O
O
O
a
b
c
NH2
NH2
NH
O
O
O
O
Br
Br
Br
Br
9
2
93
94
95
H
N
H
N
N
N
Fmoc
d
e
f
O
H N
N
H N
N
S
2
2
Br
Br
O
96
9
7
98
a
Reagents and conditions: (a) Ac O, TEA, DCM, 0 °Cꢀrt (b) NaH, toluene, reflux (c) conc.
2
HCl, reflux. (d) Fmoc isothiocyanate, DCM, rt. (e) 1. NH (7 M in MeOH), rt, 2. tꢀBuOOH
3
(70%), rt. (f) 3ꢀmethoxyphenylboronic acid, bis(triphenylphosphine)palladium(II) chloride,
Sodium carbonate (1 M aq), DME:H O (2:1), 80 °C.
2
24
ACS Paragon Plus Environment

Page 25 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
RESULTS AND DISCUSSION
The reported compounds were evaluated for BACEꢀ1 inhibition in a time response
fluorescence resonance energy transfer (TRꢀFRET) protocol. The cellꢀbased assay for BACEꢀ1
inhibition used specific antibodies to monitor reduction of sAPPβ release from human neuronalꢀ
derived SHꢀSY5Y cells. The hERG activity in CHO cells was measured as a first indication for
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
5
0
51,52
potential cardiotoxicity. Experimental values for pKa were determined
and LogD was
5
3
estimated with reversed phase liquid chromatography (eLogD). Permeability in Cacoꢀ2 cells
was measured and classified as good (Papp >25), medium (5< Papp <25) or poor (Papp <5).
Susceptibility to Pꢀgp efflux (in MDCKꢀMDR1 cells), was also determined, and the efflux of the
compounds was classified as low (<2), medium (2ꢀ3) and high (>3). Finally, the results from
both permeability measurements were combined into an “overall permeability” classification of
good, medium or poor. Since both properties are important to have a chance for high brain
1
9,20,29
exposure in vivo, the least favourable result determined the categorization.
The data is divided into four tables according to the different core subꢀtypes (Tables 1ꢀ4). An
additional table, with data on previously published series, is also included in this report
(
Supporting information). These compounds have the same R1ꢀ and R2ꢀsubstituents as the
compounds synthesized here, and are used as additional references. With these substituents, we
hypothesized that the differences in the in vitro profiles of the compounds would largely be due
to the core structures, and thus the SAR discussions are focused to the cores.
Table 1. In vitro properties of core subꢀtype A
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 83
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
X
N
R
3
2
R
R
H N
2
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
BACEꢀ1,a
BACEꢀ1,
sAPPβ
IC50
(ꢀM)
a
b
hERG
IC50
e
TRꢀFRET
LE
FRET
c
d
P
app
ꢀ6
Efflux
Overall
perm.
X
R1
R2
R3 Comp
eLogD
pKa
f
g
IC50
(10 cm/s) class.
(
ꢀM)
(
ꢀM)
O
O
O
O
O
Me 5ꢀpyrimidyl
H
H
H
F
8a
8b
9
16
0.27
0.26
0.22
0.24
0.24
0.28
1.0
0.5
3.2
5.0
2.5
5.7
2.8
1.6
9.0
nd
1.2
1.8
1.8
1.6
2.3
1.7
8.1
8.2
8.4
6.9
7.0
8.3
31
17
nd
39
23
39
low
low
nd
G
M
nd
G
Me
Et
3ꢀpyridyl
25
5ꢀpyrimidyl
100
32
Me 5ꢀpyrimidyl
10a
10b
15
low
nd
Me
3ꢀpyridyl
F
32
G*
G
S
Me 5ꢀpyrimidyl
H
10
0.32
6.0
low
a
b
IC50 values are the means of at least two experiments. Measured in hERGꢀexpressing CHO cells using IonWorks
c
d
technology. Determined by reversed phase liquid chromatography. Determined by pressureꢀassisted capillary
e
f
electrophoresis. P is the measured permeability (A to B) through Cacoꢀ2 cells. MDCKꢀMDR1 efflux data were
app
g
generated using a transwell assay. G=good, M=medium, P=poor, *Efflux was not determined for this compound, so
the overall permeability classification is only based on Cacoꢀ2 Papp. nd; not determined.
Table 2. In vitro properties of core subꢀtype B
Y
X
N
z
2
R
H N
2
BACEꢀ1,
TRꢀFRET
IC50
BACEꢀ1,
b
e
a
a
hERG
P
app
LE
FRET
sAPPβ
c
d
ꢀ6
Efflux
Overall
X
Y
Z
R2
Comp
IC50
eLogD
pKa
(10
f
g
IC50
(ꢀM)
class.
perm.
(
ꢀM)
cm/s)
(ꢀM)
O
O
S
C
C
C
C
C
C
3ꢀMeOꢀphenyl
3ꢀchlorophenyl
5ꢀpyrimidyl
19a
19b
24a
10
6.3
1.0
0.31
0.34
0.41
0.63
0.79
nd
2.5
3.3
<0
9.3
9.3
9.1
13
2.1
21
low
nd
M
P*
G
6.15
>33
0.040
low
S
C
C
3ꢀMeOꢀphenyl
24b
0.63
0.38
0.063
11
2.5
9.3
5.2
nd
M*
26
ACS Paragon Plus Environment

Page 27 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
C
C
C
O
S
C
C
S
3ꢀMeOꢀphenyl
3ꢀMeOꢀphenyl
3ꢀMeOꢀphenyl
30
34
37
3.2
2.0
2.0
0.34
0.35
0.35
0.20
0.32
0.40
8.0
9.0
7.6
2.1
2.2
2.5
9.8
10.5
7.0
20
7.3
72
low
high
nd
M
P
C
G*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
IC50 values are the means of at least two experiments. Measured in hERGꢀexpressing CHO cells using IonWorks
c
d
technology. Determined by reversed phase liquid chromatography. Determined by pressureꢀassisted capillary
electrophoresis. P is the measured permeability (A to B) through Cacoꢀ2 cells. MDCKꢀMDR1 efflux data were
generated using a transwell assay. G=good, M=medium, P=poor, *Efflux was not determined for this compound, so
the overall permeability classification is only based on Cacoꢀ2 Papp. nd; not determined.
e
f
app
g
Table 3. In vitro properties of core subꢀtype C
BACEꢀ1,a
BACEꢀ1,
sAPPβ
IC50
(ꢀM)
a
b
hERG
IC50
e
TRꢀFRET
LE
FRET
c
d
P
app
ꢀ6
Efflux
class.
Overall
perm.
Structure
Comp
eLogD
pKa
8.0
f
g
IC50
(ꢀM)
(10 cm/s)
(
ꢀM)
N
H N
2
N
N
S
N
41
0.63
0.32
0.28
0.29
0.032
0.79
1.5
2
18
High
P
O
N
H N
2
N
N
N
N
4
7
8.6
2.3
5.4
30
Medium
M
O
H
2
N
H
N
N
N
(R*,S*)ꢀ
h
4
.0
0.31
0.27
0.50
nd
nd
1.5
2.0
nd
nd
nd
nd
nd
nd
nd
F
54
H N
H
2
N
N
N
(R*,R*)ꢀ
h
16
0.63
9.2
F
54
a
b
IC50 values are the means of at least two experiments. Measured in hERGꢀexpressing CHO cells using IonWorks
c
d
technology. Determined by reversed phase liquid chromatography. Determined by pressureꢀassisted capillary
electrophoresis. P is the measured permeability (A to B) through Cacoꢀ2 cells. MDCKꢀMDR1 efflux data were
generated using a transwell assay. G=good, M=medium, P=poor, During the assay condition, (R*,R*)ꢀ54 and
R*,S*)ꢀ54 epimerized. nd; not determined.
e
f
app
g
h
(
Table 4. In vitro properties of core subꢀtype D
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 83
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
3
1
R
X
R
2
N
R
H N
2
BACEꢀ1,
TRꢀFRET
IC50
BACEꢀ1,
sAPPβ
IC50
(ꢀM)
b
e
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
a
hERG
IC50
(ꢀM)
P
(10
app
LE
FRET
c
d
ꢀ6
Effluxf Overall
g
X
R1
R2
R3
Comp
eLogD
3.6
pKa
6.9
Class.
perm.
cm/s)
(
ꢀM)
3
ꢀMeOꢀ
S
S
S
Me
H
H
H
H
H
58
64
65
77
83
7.9
0.33
3.2
9.2
15
nd
M*
phenyl
4ꢀMeOꢀ
phenyl
3ꢀpyridyl
0.79
0.50
20
0.32
0.33
0.26
0.27
0.24
0.23
0.21
0.21
0.27
0.40
2.5
10
12
2.9
3.3
1.4
0.9
1.8
1.8
1.8
1.8
1.8
6.2
5.4
6.5
10
19
18
nd
M*
M
M
P*
P*
P
4
ꢀ
3ꢀMeOꢀ
phenyl
low
low
nd
pyridyl
Nꢀ
Acyl
3ꢀMeOꢀ
phenyl
Me
nd
>33
23
16
4ꢀMeOꢀ
Phenyl
5ꢀ
C
C
C
C
C
7.9
10
2.0
<0.3
7.2
nd
pyrimidyl
4
ꢀMeOꢀ
Phenyl
3ꢀpyridyl
3ꢀpyridyl
3ꢀpyridyl
3ꢀpyridyl
3ꢀpyridyl
OMe rac-91
0.32
0.40
0.79
1.3
2.4
1.5
2.5
2.5
3.2
8.9
nd
high
high
high
high
high
4
ꢀMeOꢀ
Phenyl
Iso1-
91
OMe
20
4
ꢀMeOꢀ
Phenyl
Iso2-
91
OMe
40
nd
nd
P
4
ꢀMeOꢀ
Phenyl
Iso3-
91
OMe
40
nd
nd
P
4ꢀMeOꢀ
Phenyl
Iso4-
91
C
OMe
3.2
0.079
nd
34
P
a
b
IC50 values are the means of at least two experiments. Measured in hERGꢀexpressing CHO cells using IonWorks
technology. Determined by reversed phase liquid chromatography. Determined by pressureꢀassisted capillary
electrophoresis. P is the measured permeability (A to B) through Cacoꢀ2 cells. MDCKꢀMDR1 efflux data were
generated using a transwell assay. G=good, M=medium, P=poor, *Efflux was not determined for this compound, so
c
d
e
f
app
g
the overall permeability classification is only based on Cacoꢀ2 Papp. nd; not determined.
Chemical stability. During the synthesis of core structure Cꢀ3 (54, Scheme 11), it was realized
that 54 slowly epimerized in the acidic media used when separating the diastereoisomers by
reversedꢀphase chromatography. No visible decay of the diastereomeric mixture was seen during
the acidic purification of 53 or during the basic conditions in the subsequent SuzukiꢀMiyaura
reaction to 54. However, treating 53 in acidic or basic media for prolonged time (more than one
28
ACS Paragon Plus Environment

Page 29 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
day) it was apparent that the core structure was not stable, but gave 3´ꢀbromoacetophenone and
amide 49 as decomposition products. From these stability findings we believe that the
epimerization occurred at the carbon holding the R1ꢀgroup and not at the alphaꢀcarbon to the
amidine moiety. Data for the two stereoisomers of 54 ((R*,S*)ꢀ54 and (R*,R*)ꢀ54) are reported
in Table 3 but in the cellular assay, with a long incubation time, epimerization occurred during
the experiment. Therefore, the two data points reported for the cellular assay represent in both
cases a mixture of the four stereoisomers. There was no sign of decomposition under these
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
Due to the above mentioned stability problem, all compounds representing new cores were
checked for their stability in DMSO as well as in buffer solutions at four different pHꢀvalues (1,
4.5, 7.4 and 9) over time. All of the cores were stable in DMSO and (unless discussed) the cores
were stable in the buffer used in the enzymatic and cell assays (pH 4.5 and 7.4 respectively) for
the time period needed to perform the experiment. The same is true for the conditions used in the
other assays included in this report.
As mentioned earlier, 71 suffered from instability problem showing reversibility back to the
acetophenone 11c even under the conditions used in the enzymatic assay. For this reason, no data
for compound 71 is reported here. By preparing 77, which was predicted to be more stable due to
the electron withdrawing character of the acyl group, stability increased to allow data to be
generated without compound decomposition. Data for compound 99 was excluded in this report
since it was shown to be unstable under acidic and neutral conditions.
It was also noticed that some amidines were slowly hydrolyzed to the lactam under basic (pH
9) conditions. This was first noticed for compounds of subꢀtype B, already in the preparation of
the compounds. Neutral conditions had to be used when purifying the compounds by preparative
29
ACS Paragon Plus Environment