Structure-Based Design of Selective β-Site Amyloid Precursor Protein Cleaving Enzyme 1 (BACE1) Inhibitors: Targeting the Flap to Gain Selectivity over BACE2

Article abstract of DOI:10.1021/acs.jmedchem.9b00309

BACE1 inhibitors hold potential as agents in disease-modifying treatment for Alzheimer's disease. BACE2 cleaves the melanocyte protein PMEL in pigment cells of the skin and eye, generating melanin pigments. This role of BACE2 implies that nonselective and

Full text of DOI:10.1021/acs.jmedchem.9b00309

Subscriber access provided by Bibliothèque de l'Université Paris-Sud
Article
Structure-Based Design of Selective #-Site Amyloid
Precursor Protein Cleaving Enzyme 1 (BACE1) Inhibitors:
Targeting the Flap to Gain Selectivity over BACE2
Kazuki Fujimoto, Eriko Matsuoka, Naoya Asada, Genta Tadano, Takahiko Yamamoto, Kenji Nakahara, Kouki
Fuchino, Hisanori Ito, Naoki Kanegawa, Diederik Moechars, Harrie J.M. Gijsen, and Ken-ichi Kusakabe
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00309 • Publication Date (Web): 25 Apr 2019
Downloaded from http://pubs.acs.org on April 25, 2019
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 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 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.
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 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structure-Based Design of Selective β-Site Amyloid
Precursor Protein Cleaving Enzyme 1 (BACE1)
Inhibitors: Targeting the Flap to Gain Selectivity
over BACE2
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
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
†
,#
†,#
†,#
†,#
Kazuki Fujimoto, Eriko Matsuoka, Naoya Asada, Genta Tadano, Takahiko
Yamamoto,†.┴ Kenji Nakahara, Kouki Fuchino, Hisanori Ito, Naoki Kanegawa, Diederik
†
†
†
‡
║
§
*,†
Moechars, Harrie J.M. Gijsen, and Ken-ichi Kusakabe
†
‡
Discovery Research Laboratory for Core Therapeutic Areas, Research Laboratory for
Development, Shionogi Pharmaceutical Research Center, 1-1 Futaba-cho 3-chome,
Toyonaka, Osaka 561-0825, Japan
§
║
Medicinal Chemistry, Neuroscience Biology, Janssen Research & Development,
Turnhoutseweg 30, B-2340 Beerse, Belgium
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 68
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
BACE1 inhibitors hold potential as agents in disease-modifying treatment for Alzheimer’s
disease. BACE2 cleaves the melanocyte protein PMEL in pigment cells of the skin and eye,
generating melanin pigments. This role of BACE2 implies that non-selective and chronic
inhibition of BACE1 may cause side effects derived from BACE2. Herein, we describe the
discovery of potent and selective BACE1 inhibitors using structure-based drug design. We
targeted the flap region, where the shape and flexibility differ between these enzymes.
Analysis of the cocrystal structures of an initial lead 8 prompted us to incorporate spirocycles
followed by its fine-tuning, culminating in highly selective compounds 22. The structures 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
2
2 bound to BACE1 and BACE2 revealed that a relatively high energetic penalty in the flap of
the 22-bound BACE2 structure may cause a loss in BACE2 potency, thereby leading to its high
selectivity. These findings and insights should contribute to responding to the challenges in
exploring selective BACE1 inhibitors.
INTRODUCTION
Alzheimer’s disease (AD) is the most common form of dementia. In 2016, an estimated 47
1
million people were living with dementia. Of them, 60 to 80% were estimated to have AD,
2
and 10% of people aged 65 and older were suffering from the disease. The number of
patients with AD will escalate rapidly, as the figure is expected to double every 20 years to
1
reach 131 million by 2050. AD ultimately leads to death within 3 to 9 years after diagnosis
ACS Paragon Plus Environment
2

Page 3 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
and is the sixth-leading cause of death in the United States. Additionally, there is a huge
1
,2
impact on the quality of life for both the AD patients and their family caregivers. The
current therapeutic agents, such as the acetylcholinesterase inhibitors and the N-methyl-D-
aspartate receptor antagonist, only possess modest symptomatic effects and do not halt or
slow the disease progression. Also, the treatment gains obtained from the drugs are often lost
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
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
within a few years. Therefore, there is a great unmet medical need for a disease-modifying
drug that can benefit patients and caregivers by improving clinical outcomes.3
The accumulation of extracellular deposits of amyloid plaques composed of amyloid β-
peptides (Aβ) with 38 to 43 amino acids is a hallmark of AD. Histopathological and genetic
evidence underlie the amyloid cascade hypothesis that AD may be caused by deposition of
4
Aβ. The β-site amyloid precursor protein cleaving enzyme 1 (BACE1), also known as β-
5
secretase, initiates the production of Aβ and hence is a prime target for AD. Since the
6
discovery of BACE1 in 1999, the pharmaceutical industry and academia have pursued orally
available and brain penetrant BACE1 inhibitors with an acceptable safety profile. The initially
reported BACE1 inhibitors, such as peptidomimetics and hydroxyethylamines, struggled to
achieve balanced in vitro potency and favorable pharmacokinetic profiles, due to their large
7
molecular weight and high polar surface area. The identification of small molecule amidine-
8
,9
based inhibitors addressed the issues observed in the peptidomimetic analogs. Leveraging
structure-based design and tuning physicochemical properties culminated in the discovery of
1
0
several clinical BACE1 inhibitors, such as verubecestat (1, MK-8931), elenbecestat (2, E-
1
1
12
13
2
609), lanabecestat (3, AZ-3293), umibecestat (CNP520), and atabecestat (JNJ-
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 68
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
4
5
4861911). Unfortunetaly, 1 and 3 failed due to lack of efficacy in their phase III clinical
trials,10d,12b and the development of atabecestat was stopped due to significant elevation of
liver enzymes.14b
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
4

Page 5 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Chart 1. Representative BACE1 inhibitors in Clinical Development
O
F
NH2
H N
2
N
H N
2
S
F
S
O
N
F
N
H
H
N
H
N
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
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
N
N
N
N
N
N
O
O
O
O
F
F
1 (verubecestat)
2 (elenbecestat)
3 (lanabecestat)
The Dominantly Inherited Alzheimer Network (DIAN) study, which enrolled participants at
risk for carrying a mutation for autosomal dominant AD, reveals that deposition of Aβ,
measured by positron-emission tomography (PET), begins approximately 15 years before the
age of expected symptom onset, followed by increased tau protein in the CSF and
15
hippocampal atrophy leading to cognitive and clinical changes. As in the DIAN study, a
similar trend of Aβ deposition, brain atrophy, and cognitive decline is observed in patients
1
6
with sporadic AD. Large phase III studies for 1 and anti-Aβ antibodies involving patients
with prodromal or mild-to-moderate AD failed due to no improvement in clinical
1
7
outcomes, however, this may have been due to the intervention with these anti-Aβ
therapies being too late to change the clinical outcomes. Thus, the evidence indicates that
anti-Aβ drugs such as BACE1 inhibitors need to be given in the early stage of the disease,
1
8
most likely at the presymptomatic stages, to prevent the development of AD. This
consideration points to the need for chronic administration for more than 20 years, thereby
requiring safer profiles for these anti-Aβ drugs including BACE1 inhibitors.
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 68
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
BACE2 is a close homolog of BACE1 (52% sequence identity and 68% sequence similarity)
1
9
and is primarily expressed in the peripheral tissues but at low levels in the brain. The
enzyme cleaves the proproliferative type 1 transmembrane protein TMEM27 involved in 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
20
regulation of β cell mass. BACE2 was also found to process the pigment cell-specific
melanocyte protein (PMEL), which is specifically expressed in pigment cells of the eye and
-
/-
21
skin. Indeed, BACE2 mice exhibited a silvery depigmented coat. Furthermore, chronic
9i
administration of BACE1 inhibitors, such as 1 or NB-360, resulted in depigmentation of the
coats of rats and rabbits.10b,22 These findings for BACE2 may indicate its potential safety risk
associated with chronic and non-selective inhibition of BACE1 inhibitors.
Development of selective BACE1 inhibitors remains a significant challenge, given the
homology between BACE1 and BACE2. Indeed, the clinical compounds 1, 2, and 3 have
suboptimal BACE1 selectivity of 1.0-, 12- and 3.1-fold over BACE2 in our FRET biochemical
assay. Although a number of crystal structures of inhibitor-bound BACE1 have been
7
b
23
reported, only a few high-resolution crystal structures are available for BACE2. Bernard
and Gsell et al. paved the way to obtaining high quality inhibitor-bound BACE2 structures
by the use of surface mutagenesis or crystallization helpers such as camelid heavy chain
antibody (also known as V H or Nanobody or Xaperone) or human Fyn kinase derived SH3
H
domain (Fynomer).23b Comparative analysis of inhibitor-bound BACE1 and BACE2
structures revealed differences in conformational flexibility in the 10s loop and the flap
2
3
region, where BACE2 stabilized these closed conformations, whereas BACE1 could
ACS Paragon Plus Environment
6

Page 7 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
4
accommodate the open conformation. Thus, these conformational differences could make a
large binding site available in BACE1 relative to that in BACE2, thereby offering the design
strategy of fine-tuning the size of substituents occupying the two sites. At the outset of this
work, pioneering work from Wyeth had identified a highly selective inhibitor 4 (Chart 2),
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
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
5
where one of the ethyl groups engages with the flap (Chart 3). In parallel with our efforts,
several pharmaceutical companies reported selective BACE1 inhibitors, such as compounds 5
26
(
PF-6751979), 6, and 7, in patent and scientific literature, which appear to utilize the 10s
loop or the flap to gain selectivity, given the crystal structures of the analogues bound to
BACE1. Herein, we describe our medicinal chemistry efforts of exploring next generation
BACE1 inhibitors with high selectivity over BACE2, starting from Compound J (8) (Chart
2
7
3
). Leveraging the cocrystal structures of 4 and 8 led to a successful strategy of targeting the
flap followed by incorporating spirocycles at the 5-position on the thiazine, culminating in
the highly selective BACE1 inhibitors 21 and 22. Comparative analysis of the cocrystal
structures of 22 bound to BACE1 and BACE2 provided insight into the selectivity.
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 68
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
Chart 2. Representative Selective BACE1 inhibitors
H N
2
H N
S
2
N
N
H
N
N
S
O
N
N
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
HN
O
O
N
O
F
F
4 (Wyeth)
5
(Pfizer)
O
N
H N
2
S
H N
2
S
N
N
N
O
O
N
H
F
N
H
N
N
F
N
N
N
O
O
F
F
F
6
(Eli Lilly)
7 (Amgen)
Chart 3. Lead Compounds
H N
2
N
N
H2N
S
Cl
N
N
O
5
N
H
N
N
N
Flap
O
F
Flap
4 (Wyeth)
BACE1 IC : 44 nM
8 (Compound J)
BACE1 IC : 4.6 nM
50
50
BACE2 IC : >10 M
BACE2 IC : 8.7 nM
5
0
50
Selectivity: >227x
Selectivity: 1.9x
Cell A IC : 138 nM
Cell A IC : 0.93 nM
5
0
50
RESULTS AND DISCUSSION
Inhibitor Design. As the first step of our efforts, we established the binding mode of
compound 8 by soaking it with apo-BACE1 crystals to facilitate medicinal chemistry
ACS Paragon Plus Environment
8

Page 9 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
7
design. Figure 1A highlighted non-conserved residues around 8, consisting of key residues
of BACE1, such as Pro70, Gln12, and Arg128 located in the flap, 10s loop, and S2’ pocket,
2
5
respectively. At the onset of this study, compound 4 was known as a selective BACE1
inhibitor with a BACE1 selectivity of >227-fold in our FRET biochemical assay (BACE1 IC50
= 44 nM, BACE2 IC = >10 μM), though it suffered from suboptimal cellular potency (IC =
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
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
5
0
50
1
38 nM; Chart 3). The significant selectivity was rationalized by the favorable interaction of
25
the ethyl group on the pyridine with the flap. Indeed, a difference in the conformation of
the flap was observed between BACE1 and BACE2. As mentioned above, the flap in apo
2
4
BACE1 tends to adopt an open conformation, whereas that in apo BACE2 favors a closed
2
3
one. Because one of the non-conserved residues in the flap in BACE1 is Pro70, replaced by
2
3
Lys86 in BACE2, this proline residue may affect the flexibility. Unlike the apo BACE1
structures, the crystal structure of 8 bound to BACE1 adopts a closed conformation in the
flap (Figure 1B). Further analysis of the cocrystal structure of 8 suggested that introduction of
substituents at the 5-position with the stereochemistry in Figure 1B on the thiazine ring
could provide a favorable interaction with Tyr71 in the flap. Therefore, we reasoned that
targeting the flap using the cellular potent lead 8 could improve selectivity while retaining
cellular potency.
The 10s loop also adopts distinct conformations between BACE1 and BACE2, providing a
smaller S3 pocket in BACE2 than in BACE1.23,24 Thus, utilizing this pocket offers another
opportunity to design BACE1 selective inhibitors. Indeed, Hilpert et al. at Roche explored
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 68
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
8
amide substituents in their oxazine series positioned in the S3 pocket. Whereas
incorporation of larger substituents at this position achieved increased selectivity, an
enzyme-to-cell potency shift was observed in these compounds, resulting in suboptimal
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
28
cellular activity. With this point in mind, we optimized the amide moiety in 8 interacting
with the 10s loop in the S3 pocket, prior to exploring substituents at the 5-positon.
Figure 1. (A) A schematic representation of non-conserved residues between BACE1 and
BACE2 including key interactions. (B) Crystal structure of 8 bound to a flap-closed form of
BACE1. Substituents projecting from the 5-positon to reach the flap.
Optimization of the Amide Substituent in 8. The optimization began with adjusting the size
of the chloro group in 8 to strike a balance between potency and selectivity. Replacement of
the chloro with a fluorine (9) reduced both BACE1 and BACE2 potency. Compound 10 with
ACS Paragon Plus Environment
1
0

Page 11 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a cyano group as a larger substituent had a similar potency to 8 and improved selectivity over
BACE2 (6.6-fold). A variety of pyrazine analogs with different R groups (11-14) was then
pursued. The difluoromethyl 11 displayed a higher BACE2 IC value of 102 nM, although
5
0
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
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
BACE1 potency was diminished as well. Compound 12, possessing a methoxy group,
imparted an increase in BACE1 potency, whereas the BACE2 activity was retained relative to
the difluoro counterpart 11, resulting in increased BACE1 selectivity. Introduction of a
bulkier trifluoromethoxy group (13) retained BACE1 activity and, as expected, provided
further improvement in selectivity (44-fold). However, as observed in the previous work at
2
8
Roche, such modification caused an enzyme-to-cell shift, leading to a moderate cellular
IC of 13 nM. A balanced profile of BACE1 selectivity and cellular potency was seen when
5
0
introducing a fluromethoxy group at the R position, leading to compound 14. Although the
selectivity of 14 was still moderate (16-fold), this compound served as a starting point to
explore substituents at the 5-positon on the thiazine ring targeting the flap.
ACS Paragon Plus Environment
1
1

Journal of Medicinal Chemistry
Page 12 of 68
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
Table 1. Potency and Selectivity of the Amide Analogs
H N
2
S
F
H
N
R
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
O
Biochemical IC50 (nM)a,b
Selectivity
Cellular A
IC50 (nM)
compd
R
(BACE2/BACE1)
a,c
BACE1
4.6
BACE2
8.7
Cl
N
N
8
9
1.9x
1.1x
0.93
F
13
14
16
NC
F
N
10
3.2
20
21
6.6x
5.1x
0.93
7.2
11
F
N
N
102
N
N
O
O
O
12
11
85
316
78
7.7x
44x
16x
5.4
13
F3C
F
N
N
1
3
4
7.2
4.9
N
N
1
1.8
a
b
Values represent the mean values of at least two independent experiments. Biochemical
c
fluorescence resonance energy transfer (FRET) assay. IC determined by measuring the levels
5
0
of secreted Aβ42 in a human neuroblastoma SKNBE2 cell line expressing the wild-type
amyloid precursor protein (hAPP695) via a sandwich αlisa assay.
ACS Paragon Plus Environment
1
2

Page 13 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Initial Exploration of Substituents at the 5-Position of the Thiazine. As discussed above, the
cocrystal structure of 8 in BACE1 prompted the introduction of substituents at the 5-position
with the (R) configuration as shown in Figure 1B. Addition of a methoxy group provided
compound 15, which reduced BACE2 potency resulting in improved selectivity of 30-fold
relative to the non-substituted 14. The cocrystal structure of the initial lead 8 suggested that
larger substituents, such a phenyl group with (R)-configuration (16), could not be
accommodated due to steric hindrance from the flap. To confirm the steric constraint around
this region, both (R)- and (S)-phenyl 16 and 17 were synthesized. Contrary to our
expectation, compound 16 increased potency in both BACE1 and BACE2, leading to the
same level of selectivity as 14. The phenyl thiazine 17 with the opposite configuration had a
comparable BACE1 potency to 14, whereas selectivity was decreased due to slightly
increased BACE2 activity relative to the (R)-phenyl 16. To gain insight into this unexpected
finding, the cocrystal structures of 16 and the diastereomer counterpart 17 bound to BACE1
were solved.
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
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
1
3

Journal of Medicinal Chemistry
Page 14 of 68
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
Table 2. Initial Exploration of the 5-Substituted Thiazines
H N
2
S
F
F
O
N
H
N
N
N
R
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
Biochemical IC50 (nM)a,b
Selectivity
BACE2/BACE1)
Cellular A
IC50 (nM)
compd
R
H
(
a,b
BACE1
4.9
BACE2
78
14
16x
30x
1.8
15
8.3
1.7
251
28
4.5
2.3
(
R)
O
16
16x
(
(
R)
S)
17
3.3
13
3.9x
2.0
a
b
Values represent the mean values of at least two independent experiments. Biochemical
c
fluorescence resonance energy transfer (FRET) assay. IC determined by measuring the levels
5
0
of secreted Aβ42 in a human neuroblastoma SKNBE2 cell line expressing the wild-type
amyloid precursor protein (hAPP695) via a sandwich αlisa assay.
X-ray Analysis of 16 and 17 Leading to the Design of Spiro-Thiazines. The crystal structure
of 16 confirms that the flap adopts an open conformation, unlike that of 8 bound to BACE1
(Figure 2A). The hydrogen bond between the flap residues Tyr71 and Trp76 is disrupted,
thereby moving away the Tyr71 side chain to generate a new binding pocket that
accomodates the phenyl group in 16. In contrast, 17 complexed with BACE1 shows a pseudo
flap-closed conformation where the side chain in Tyr71 has a good alignment with that of 8
ACS Paragon Plus Environment
1
4

Page 15 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
bound to BACE1 (close conformation), though the hydrogenbond between Tyr71 and Trp76
is not present (a weak water-mediated hydrogen bond is observed), and the phenyl in 17
points toward the solvent region (Figure 2B), explaining why compound 17 exhibited slightly
reduced potency in BACE1. We postulated that a similar change may have occurred in
BACE2. Therefore, using the cocrystal structure of 16, we looked for an alternative design of
substituents at this position, where the analysis prompted us to incorporate spirocycles that
could provide an optimal vector to the flap. Alternatively, replacement of the phenyl in 16
with bulkier substituents, such as saturated rings, was predicted to form favorable
interactions with the flap to gain selectivity.
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
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
Figure 2. (A) Crystal structure of 16
bound to BACE1. (B) 17 bound to BACE1.
Identification of Selective BACE1 Inhibitors 21 and 22. To test the design hypothesis, a
saturated ring of a tetrahydro-2H-pyran (18) was introduced at the 5-position (Table 3). The
ACS Paragon Plus Environment
1
5

Journal of Medicinal Chemistry
Page 16 of 68
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
compound 18 showed a potency comparable to the phenyl-substituted 16 and improved
selectivity over BACE2. To our delight, the corresponding spirocycle 19 provided a further
gain in selectivity. The use of a smaller ring size, as in 20, resulted in a decreased selectivity
as expected, whereas compound 21 with a difluorocyclohexyl ring displayed significantly
reduced BACE2 potency, achieving a 104-fold selectivity. Further gain in selectivity was
realized when the difluoromethylene in 21 was replaced with a sulfonyl group. The sulfonyl
compound 22 demonstrated a remarkable 550-fold selectivity in the FRET assay, although
this change was associated with reduced cellular potency (Table 3). This result was attributed
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
-
6
to low lipophilicity (LogD = 0.62) resulting in a poor permeability of 1.6 × 10 cm/s (Table
). On the other hand, compounds 19 and 21 with good cellular potency displayed improved
permeability (Table 4).
4
ACS Paragon Plus Environment
1
6

Page 17 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. Optimization of the 5-Substituted Thiazines
H N
2
S
F
F
O
N
H
N
L
N
N
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
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
Biochemical IC50 (nM)a,b
Selectivity
Cellular A
IC50 (nM)
compd
L
(BACE2/BACE1)
a,b
BACE1
3.2
BACE2
74
18
23x
42x
2.0
O
19
15
631
6.8
O
F
2
0
1
5.5
16
129
23x
1.2
4.3
F
2
1660
104x
F
F
22
3.8
2090
550x
13
O
S
O
a
b
Values represent the mean values of at least two independent experiments. Biochemical
c
fluorescence resonance energy transfer (FRET) assay. IC determined by measuring the levels
of secreted Aβ42 in a human neuroblastoma SKNBE2 cell line expressing the wild-type
amyloid precursor protein (hAPP695) via a sandwich αlisa assay.
5
0
ACS Paragon Plus Environment
1
7

Journal of Medicinal Chemistry
Page 18 of 68
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
Table 4. Binding Affinity, Metabolic Stability, P-gp Efflux, and Physicochemical Profiles of 19,
21, and 22
a
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
K (nM)
i
Binding selectivity
BACE1 BACE2 (BACE2/BACE1)
HLM/MLMb P-gp ERc Pappd Log De
Compd
19
21
22
6.5
4.0
1.9
479
2000
1740
74x
500x
916x
51 / 5.0
27 / 4.5
55 / 17
11
2.4
20
21
7.1
1.6
1.2
2.3
0.62
a
The binding affinity (K ) was determined in a competitive radioligand binding assay using
i
3
the tritiated non-selective BACE1/BACE 2 inhibitor [ H]-JNJ-962 (JNJ-962: BACE1 K = 0.69
i
nM, BACE2 K = 0.29 nM). Values represent the mean values of at least two determinations.
i
b
% Remaining after 30 min incubation with human liver (HLM) and mouse liver microsomes
MLM). Efflux ratio measured in LLC-PK1 cells transfected with human MDR1. Permeability
coefficient measured in LLC-PK1 cells (×10 cm/s). LogD determined in 1-octanol/phosphate
c
d
(
-
6
k
buffer at pH 7.4.
To further investigate the BACE1 selectivity over BACE2 observed in compounds 19, 21, and
2
2, the binding affinities (K ) to the respective enzymes were determined in a competitive
i
radio ligand binding assay using a tritiated non-selective BACE1/BACE2 inhibitor (JNJ-
2
9
9
62), which had high affinities for both BACE1 and BACE2 (BACE1 K = 0.69 nM, BACE2
i
K = 0.29 nM). As shown in Table 4, compound 19 had a binding selectivity of 74-fold,
i
similar to the selectivity observed in the FRET assay. The spiro compounds 21 and 22
achieved a pronounced binding selectivity of 500-fold and 916-fold, respectively. On the
basis of the outstanding selectivity observed for 21 and 22, these compounds were profiled in
ACS Paragon Plus Environment
1
8

Page 19 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a mouse pharmacokinetics (PK) and pharmacodynamics (PD) model to investigate their
effect on Aβ reduction in the brain.
Prior to the PK/PD study in mouse, metabolic stability and P-gp mediated efflux were
profiled including compound 19 (Table 4). Unfortunately, all the compounds exhibited poor
metabolic stability in mouse (<20% remaining after 30 min incubation in microsomes),
whereas they were more metabolically stable in human. A high P-gp efflux ratio of 20 was
realized with the sulfonyl compound 22, whereas 21, possessing the difluoromethylene
instead of the sulfonyl, showed improved P-gp efflux; the tetrahydropyran 19 also had a high
P-gp efflux. This result indicated that polar substituents, such as sulfone and ether moieties,
could be involved in an interaction with P-gp, resulting in increased P-gp efflux. Although
increasing permeability is often a successful strategy for reducing P-gp efflux, it does not
account for the P-gp efflux ratios observed in these compounds.
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
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
Mouse PK/PD Study of 21 and 22. Considering the low metabolic stability observed for 21
and 22, Male ICR (Crlj:CD) mice were orally administered a high dose of 30 mg/kg (n = 4),
and the total Aβ levels were measured in the brain at 2, 4, and 6 h post dose. Unfortunately,
no significant Aβ reduction was observed for the sulfonyl-substituted 22. The poor efficacy
was explained by the low brain levels at all the time points derived from its low permeability
and high P-gp efflux, and consequently it exhibited a brain-to-plasma ratio (K ) of 0.010. On
p
the other hand, compound 21 with improved P-gp efflux and permeability demonstrated
robust Aβ reduction at the same dose (Figure 3). Maximum total Aβ reduction of 61% was
ACS Paragon Plus Environment
1
9

Journal of Medicinal Chemistry
Page 20 of 68
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
observed at 4 h post dose at a brain level of 626 ng/mL with an improved K of 0.55. The
p
improved physicochemical properties and reduced P-gp efflux translated into the significantly
improved brain levels, which led to the pronounced central activity.
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
Figure 3. Reduction of total Aβ in the brain of male ICR mice (n = 4) after oral administration
of 21 as a solution of 20% 2-hydroxypropyl-β-cyclodextrin (HPBCD) at 30 mg/kg.
Crystal Structures of 22 Bound to BACE1 and BACE2. To gain insight into the pronounced
BACE1 selectivity over BACE2 observed in 22, the cocrystal structures of this compound
bound to both the enzymes were determined (Figure 4). Soaking of apo BACE1 crystals
generated the 22-bound BACE1 structure with a resolution of 2.6 Å. Like the 16-bound
BACE1 structure, the flap forms an open conformation, where the hydrogen bond
interaction between the phenolic OH in Tyr71 and the indole NH in Trp76 is disrupted
ACS Paragon Plus Environment
2
0

Page 21 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
Figure 4A). A slight movement of the flap over that in apo BACE1 (PDB ID: 1w50)24a is
observed with a distance of 1.5 Å based on the α carbon in Tyr71 (Figure 4B), whereas no
significant movement is seen between the 16- and 22-bound BACE1 structures. Unlike the
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
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
6-bound BACE1 structure, the spirocyle in 22 interacts well with the side chain in Tyr71,
confirming our design hypothesis.
The binary complex of BACE2 and a crystallization helper of Xaperone XA4813 was
30
cocrystallized with 22 with a resolution of 1.5 Å, according to a paper published from
Roche.23b As expected, the 22-bound BACE2 structure displays a flap-open conformation,
consistent with the corresponding BACE1 structure (Figure 4C). Notably, the binding of 22
to BACE2 causes a significant movement of the flap by 3.2 Å, relative to its position in an apo
BACE2 (PDB ID: 3zkg) (Figure 4D).23b Thus, a larger movement of the flap over the apo
enzyme is confirmed in the BACE2 (3.2 Å vs 1.5 Å). More importantly, the difference in
mean B-factors (temperature factor) between all of the complex and the flap of the 22-bound
2
2
2
BACE2 is up to 18.3 Å (B-factor of the complex = 32.0 Å ; B-factor of the flap = 50.3 Å ),
2
2
whereas in the 22-bound BACE1, it falls to 2.0 Å (B-factor of the complex = 45.9 Å ; B-factor
2
of the flap = 47.9 Å ). This indicates that the 22-bound BACE2 was destabilized relative to
the 22-bound BACE1, which may result in decreased potency in BACE2 but a retained
BACE1 activity, providing an explanation for the increase in selectivity observed in
compound 22. We believe that these structural insights as well as the cocrystal structures of
ACS Paragon Plus Environment
2
1

Journal of Medicinal Chemistry
Page 22 of 68
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 bound to BACE1 and BACE2 could aid the design of novel selective BACE1 inhibitors
targeting the flap.
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
Figure 4. Crystal structure of 22 bound to BACE1 and BACE2. (A) A view of the flap region
in the 22-bound BACE1 structure. (B) Superposition of the flap in the 22-bound BACE1 and
apo BACE1 structure (PDB ID: 1w50). The color in the flap indicates B-factors (Bule: low B-
factors; Orange: high B-factors). (C) A view of the flap region in the 22-bound BACE2
structure. (D) Superposition of the flap in the 22-bound BACE2 and apo BACE2 structure
(PDB ID: 3zkg). The color means the same as B).
ACS Paragon Plus Environment
2
2

Page 23 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CHEMISTRY
The compounds 8−14 bearing various amide moieties were prepared from aniline intermediate
3 (Scheme 1). The aniline 23 and the carboxylic acids used for synthesis of 11, 13 and 14 were
2
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
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
prepared according to known procedures.27,28,31,32 Amide formation was accomplished without
protection of the amidine moiety by adding 1.0 equiv of HCl over 23, furnishing the thiazines
8
−14.
Scheme 1. Synthesis of Thiazines 8−14a
H N
S
F
H N
S
F
2
2
a
N
H
N
H N
2
R
N
O
23
8: R = 5-chloro-pyridin-2-yl
9: R = 5-fluoro-pyridine-2-yl
0: R = 5-cyano-pyridine-2-yl
1: R = 5-difluoromethyl-pyrazine-2-yl
2: R = 5-methoxy-pyrazine-2-yl
3: R = 5-(2,2,2-trifluoroethoxy)pyrazine-2-yl
4: R = 5-fluoromethoxy-pyrazine-2-yl
1
1
1
1
1
a
Reagents and conditions: (a) RCO H, EDC∙HCl, aq HCl, THF or MeOH, 0 °C or 0 °C to rt,
2
4
9−100%.
The synthesis of the methoxy thiazine 15 started with (R)-tert-butylsulfinylketimine 25,27b,33
readily available from the corresponding ketone (Scheme 2). The lithium enolate of methyl 2-
methoxyacetate was transformed to the corresponding titanium enolate using ClTi(Oi-Pr)3,
which was reacted with the sulfinyl ketimine 25 to give the ester 26.27c The ester group in 26
ACS Paragon Plus Environment
2
3

Journal of Medicinal Chemistry
Page 24 of 68
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
was reduced to the corresponding alcohol with LiBH followed by cleavage of the sulfineamide
4
group and subsequent thiourea formation using benzoyl isothiocyanate (BzNCS) to yield the
thiourea 27. Thiazine cyclization was accomplished with Ghosez’s reagent34 to give the
thiazine 28. Following deprotection of the benzoyl group in 28 with hydrazine monohydrate,
thiazine 29 was nitrated with fuming nitric acid, and subsequent reduction of the nitro group
and amide formation provided the desired thiazine 15.
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
Scheme 2. Synthesis of Thiazine 15a
O
S
S
S
N
F
O
NH
BzHN
NH
a
b
c
CO Me
2
OH
OMe
OMe
F
F
25
26
27
BzHN
S
H N
2
S
H N
2
S
F
F
O
d
e
N
N
N
H
N
OMe
OMe
N
N
OMe
O
F
F
28
29
15
aReagents and conditions: (a) n-BuLi, diisopropylamine, methyl 2-methoxyacetate, ClTi(Oi-
Pr) , THF, −78 °C, 72%; (b) (i) LiBH , MeOH, THF, 0 °C, (ii) 4 M HCl in 1,4-dioxane, MeOH,
3
4
rt, (iii) BzNCS, DCM, rt, 82% over 3 steps; (c) Ghosez’s reagent, DCM, rt, 90%; (d) hydrazine
monohydrate, MeOH, THF, rt, 93%; (e) (i) HNO , H SO , TFA, −20 °C, (ii) Fe, NH Cl, EtOH,
3
2
4
4
toluene, H O, 80 °C, (iii) 5-(fluoromethoxy)pyrazine-2-carboxylic acid, EDC∙HCl, aq HCl,
2
MeOH, 0 °C to rt, 58% over 3 steps.
ACS Paragon Plus Environment
2
4

Page 25 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
As shown in Scheme 3, synthesis of thiazines 16 and 17 was achieved through a modified
3
5
route to Scheme 2. Reaction of (R)-tert-butylsulfinylketimine 30 and the titanium enolate
of ethyl 2-phenylacetate yielded ester 31 as an inseparable 2:1 diastereomeric mixture. The
tert-butylsulfinamide group in 31 was removed with HCl, and protection of the amine with
Boc group provided 32 and 33, which were separable by silicagel column chromatography.
The ester groups in 32 and 33 were reduced to the corresponding alcohols, followed by
deprotection of the Boc groups and thiourea formation to afford 34 and 35, respectively.
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
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
4
Thiazine formation using Ghosez’s reagent provided 36 and 37, of which the benzoyl
groups were deprotected, followed by introduction of an azide group with sodium azide and
subsequent reduction of the azide group to provide the corresponding anilines. Finally,
amide coupling of the anilines afforded the target compounds 16 and 17. The absolute and
relative configurations were assigned based on the crystal structures of 16 and 17 bound to
BACE1.
ACS Paragon Plus Environment
2
5

Journal of Medicinal Chemistry
Page 26 of 68
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
Scheme 3. Synthesis of Thiazines 16 and 17a
O
S
S
N
F
O
NH
NHBoc
a
b
c
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
Br
Br
CO Et
Br
CO Et
2
2
Ph
R1
F
F
1
30
31
32: R =(R)-Ph
3: R =(S)-Ph
1
3
S
BzHN
S
F
H N
S
2
F
O
d
e
BzHN
NH
N
N
H
N
Br
Br
R1
N
N
R1
OH
R1
O
1
F
F
1
1
3
4: R =(R)-Ph
36: R =(R)-Ph
16: R =(R)-Ph
1
1
1
35: R =(S)-Ph
37: R =(S)-Ph
17: R =(S)-Ph
aReagents and conditions: (a) n-BuLi, diisopropylamine, ethyl 2-phenylacetate, ClTi(Oi-Pr)3,
THF, −78 °C, 91%, dr = 2:1; (b) (i) 4 M HCl in 1,4-dioxane, MeOH, rt, (ii) Boc O, MeOH, 60 °C,
2
5
7% and 28% (over 2 steps); (c) (i) LiBH , MeOH, THF, 0 °C, (ii) TFA, DCM, rt, (iii) BzNCS,
4
DCM, rt, 78 and 87% over 3 steps; (d) Ghosez’s reagent, 0 °C to rt, 90% and 81%; (e) (i) DBU,
1
2
MeOH, 60 °C, (ii) NaN , CuI, (1R,2R)-N ,N -dimethylcyclohexane-1,2-diamine, sodium L-
3
ascorbate, EtOH, H O, 100 °C, (iii) Fe, NH Cl, EtOH, H O , THF, 60 °C or Fe, NH Cl, EtOH,
2
4
2
4
H O, THF, 60 °C, (iv) 5-(fluoromethoxy)pyrazine-2-carboxylic acid, EDC∙HCl, aq HCl, MeOH,
2
0 °C, 46% and 55% over 4 steps.
As shown in Scheme 4, tetrahydro-2H-pyran substituted thiazine 18 was synthesized from
the ketimine 25. Reaction of 25 with ethyl 2-(tetrahydro-2H-pyran-4-yl)acetate yielded the
ester 38 as an inseparable diasteromeric mixture. Cleavage of the sulfineamide group and
subsequent Boc protection provided 39. The ester group in 39 was reduced to the
corresponding alcohol, of which the hydroxyl intramolecularly reacted with the Boc group
to give carbamate 40. The carbamate 40 was protected with a Boc group to give 41. Ring
ACS Paragon Plus Environment
2
6

Page 27 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
opening of the carbamate 41 with K CO in MeOH, deprotection of the Boc group, and
2
2
subsequent thioureation provided 42. Thiazine formation of the thiourea 42 was
accomplished with N,N-diethylaminosulfur trifluoride (DAST) to give 43.27d Boc protection
of 43 allowed isolation of 44 and 45 as single isomers by silicagel column chromatography.
Removal of the benzoyl group was conducted by hydrolysis using K CO in MeOH following
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
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
3
protection of the aminde NH with a Boc group. The resulting Boc group was deprotected
with TFA to afford 46, which was converted into the target compound 18 according to the
established procedure as described in Schemes 1 and 3. The stereochemistry of 18 was
confirmed by NOESY correlations using the intermediates 44 and 45 (Figures S1 and S2, see
Supporting Information).
ACS Paragon Plus Environment
2
7

Journal of Medicinal Chemistry
Page 28 of 68
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
Scheme 4. Synthesis of Thiazines 18a
O
S
S
O
NH
BocHN
F
O
O
O
F
N
F
a
b
c
d
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
CO Et
2
OEt
HN
O
F
O
O
25
3
8
39
40
S
O
O
BzHN
NH
BzHN
S
BzBocN
S
e
f
g
BocN
OH
N
N
O
O
O
F
F
F
F
4
1
42O
43
44: 5-(R)
5: 5-(S)
4
H N
2
S
F
H N
2
S
F
H N
2
S
F
O
h
i
j
N
N
N
H
N
H N
2
N
N
O
O
O
O
F
46
47
18
aReagents and conditions: (a) n-BuLi, diisopropylamine, ethyl 2-(tetrahydro-2H-pyran-4-
yl)acetate, ClTi(Oi-Pr) , THF, −78 °C, 43%, dr = 2.9:1; (b) (i) 4 M HCl in 1,4-dioxane, MeOH,
3
rt, (ii) Boc O, THF, 60°C , 88% over 2 steps; (c) LiBH , MeOH, THF, 50 °C, 24%; (d) Boc O,
2
4
2
DCM, THF, 60 °C, 73%; (e) (i) K CO , MeOH, rt, (ii) TFA, DCM, rt, (iii) BzNCS, DCM, rt, 49%
2
3
over 3 steps; (f) DAST, DCM, 0 °C, 73%; (g) Boc O, THF, rt, 47% and 26%; (h) (i) K CO ,
MeOH, rt, (ii) TFA, DCM, rt, 91% over 2 steps; (i) (i) HNO , H SO , TFA, −20 °C, (ii) Fe, NH Cl,
2
2
3
3
2
4
4
toluene, H O, 80 °C, 74% over 2 steps; (j) 5-(fluoromethoxy)pyrazine-2-carboxylic acid,
2
EDC∙HCl, aq HCl, MeOH, rt, 76%.
The synthesis of thiazine 19 started with sulfinyl ketimine 25 (Scheme 5). Addition of the
titanium enolate of ethyl tetrahydro-2H-pyran-4-carboxylate to the ketimine 25, cleavage of
the sulfinaminde, and subsequent Boc protection afforded 49. The subsequent steps to the
ACS Paragon Plus Environment
2
8

Page 29 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
target 19 were conducted according to the procedures described in Scheme 4. Similar to the
syntheses of compounds 18 and 19, the spiro thiazines 20 and 21 were synthesized according
to Schemes 4 and 5 as shown in Schemes 6 and 7, respectively.
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
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
Scheme 5. Synthesis of Thiazines 19a
O
S
S
O
O
N
F
O
NH
F
O
O
BocHN
F
O
O
a
b
c
HN
OMe
S
OMe
O
F
50
25
48
49
O
O
F
BzHN
S
N
BzHN
NH
d
e
f
g
BocN
OH
O
O
F
O
F
51
52
53
H N
2
S
F
H N
2
S
F
H N
2
S
F
O
h
i
N
N
N
H
N
N
H N
2
N
O
O
O
O
F
54
55
19
aReagents and conditions: (a) n-BuLi, diisopropylamine, ethyl tetrahydro-2H-pyran-4-
carboxylate, ClTi(Oi-Pr) , THF, −78 °C, 75%; (b) (i) 4 M HCl in 1,4-dioxane, MeOH, rt, (ii)
3
Boc O, MeOH, 60 °C, quant. over 2 steps; (c) LiBH , MeOH, THF, 50 °C, 70%; (d) Boc O,
2
4
2
DMAP, THF, 50 °C, 84%; (e) (i) K CO , MeOH, rt, (ii) TFA, DCM, rt, (iii) BzNCS, DCM, 0 °C
2
3
to rt, 74% over 3 steps; (f) DAST, DCM, 0, 57%; (g) hydrazine monohydrate, MeOH, THF, rt ,
7
2
8
1%; (h) (i) HNO , H SO , TFA, −20 °C to 0°C, (ii) Fe, NH Cl, toluene, H O, 80 °C, 78% over
steps; (i) 5-(fluoromethoxy)pyrazine-2-carboxylic acid, EDC∙HCl, aq HCl, MeOH, 0 °C to rt,
1%.
3
2
4
4
2
ACS Paragon Plus Environment
2
9