New pyrazolyl and thienyl aminohydantoins as potent BACE1 inhibitors: Exploring the S2′ region

Article abstract of DOI:10.1016/j.bmcl.2011.07.057

The proteolytic enzyme β-secretase (BACE1) plays a central role in the synthesis of the pathogenic β-amyloid in Alzheimer's disease. SAR studies of the S2′ region of the BACE1 ligand binding pocket with pyrazolyl and thienyl P2′ side chains are reported. These analogs exhibit low nanomolar potency for BACE1, and demonstrate >50- to 100-fold selectivity for the structurally related aspartyl proteases BACE2 and cathepsin D. Small groups attached at the nitrogen of the P2′ pyrazolyl moiety, together with the P3 pyrimidine nucleus projecting into the S3 region of the binding pocket, are critical components to ligand's potency and selectivity. P2′ thiophene side chain analogs are highly potent BACE1 inhibitors with excellent selectivity against cathepsin D, but only modest selectivity against BACE2. The cell-based activity of these new analogs tracked well with their increased molecular binding with EC₅₀ values of 0.07-0.2 μM in the ELISA assay for the most potent analogs.

Full text of DOI:10.1016/j.bmcl.2011.07.057

Bioorganic & Medicinal Chemistry Letters 21 (2011) 5164–5170
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
New pyrazolyl and thienyl aminohydantoins as potent BACE1 inhibitors:
Exploring the S20 region q
Michael S. Malamas a, , Jim Erdei a, Iwan Gunawan a, Keith Barnes b, Yu Hui b, Matthew Johnson b,
⇑
Albert Robichaud a, , Ping Zhou a, Yinfa Yan a, William Solvibile a, Jim Turner a, Kristi Yi Fan a,à
,
Rajiv Chopra a,§, Jonathan Bard a, Menelas N. Pangalos a,||
a Pfizer Neuroscience Princeton, 865 Ridge Road, Monmouth Junction, NJ 08852, USA
b Albany Molecular Research, Albany, NY 12203, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The proteolytic enzyme b-secretase (BACE1) plays a central role in the synthesis of the pathogenic b-amy-
loid in Alzheimer’s disease. SAR studies of the S20 region of the BACE1 ligand binding pocket with pyraz-
olyl and thienyl P20 side chains are reported. These analogs exhibit low nanomolar potency for BACE1,
and demonstrate >50- to 100-fold selectivity for the structurally related aspartyl proteases BACE2 and
cathepsin D. Small groups attached at the nitrogen of the P20 pyrazolyl moiety, together with the P3
pyrimidine nucleus projecting into the S3 region of the binding pocket, are critical components to ligand’s
potency and selectivity. P20 thiophene side chain analogs are highly potent BACE1 inhibitors with excel-
lent selectivity against cathepsin D, but only modest selectivity against BACE2. The cell-based activity of
Received 6 June 2011
Revised 13 July 2011
Accepted 14 July 2011
Available online 23 July 2011
Keywords:
Alzheimer’s disease (AD)
BACE1 inhibitors
Beta-amyloid peptide (Ab)
Aminohydantoins
these new analogs tracked well with their increased molecular binding with EC50 values of 0.07–0.2
in the ELISA assay for the most potent analogs.
lM
Ó 2011 Elsevier Ltd. All rights reserved.
Alzheimer’s Disease (AD) is a progressive, degenerative disease
of the brain and recognized as the leading cause of dementia in the
ageing population. The prognosis of AD is poor and there is a great
need for medical intervention of this disease. Currently, the
approved medical therapies consist of cholinesterase inhibitors
tangles.3,4 b-Amyloid deposits are predominately aggregates of
the Ab peptides (Ab, 39–43 residues) resulting from the endoprote-
olysis of the amyloid precursor protein (APP).5,6 These peptide
fragments result from the sequential cleavage of APP, first at the
N-terminus by b-secretase enzyme (b-site APP cleaving enzyme,
BACE1),7–13,2,14–18 followed at the C-terminus by one or more
and N-methyl
D-aspartate (NMDA) antagonists, that reduce the
symptomatology of the disease in the initial phase, but do not
appear to be capable of curing or stopping its progression.1,2 The
pathological hallmarks of AD include the aggregation and extracel-
lular deposition of b-amyloid peptide (Ab), which leads to plaque
formation, and the abnormal hyperphosphorylation of tau protein,
which leads to the intracellular formation of neurofibrillary
c
-secretase complexes (intramembrane aspartyl proteases),10 as
part of the b-amyloidogenic pathway. Although the cause of AD
remains unknown, a large body of evidence is beginning to accu-
mulate that highlights the central role of Ab in the pathogenesis
of the disease.11–13,2 Thus, processes that limit Ab production and
deposition by preventing formation, inhibiting aggregation, and/
or enhancing clearance may offer effective treatments for AD. Since
b-secretase mediated cleavage of APP is the first and rate-limiting
step of the amyloidogenic processing pathway, BACE1 inhibition is
considered a prominent therapeutic target for treating AD by
diminishing Ab peptide formation in AD patients.
Abbreviations: AD, Alzheimer’s disease; Ab, b-amyloid peptide; APP, b-amyloid
precursor protein; BACE, b-site APP cleaving enzyme; FRET, fluorescence resonance
energy transfer; ELISA, enzyme-linked immune sandwich assay; CHO, Chinese
hamster ovary.
q
Atomic coordinates of the BACE1 crystal structure with compounds 28 (3S7L)
Recently, we have disclosed the discovery of small molecule
aminohydantoins as potent BACE1 inhibitors.14a–d In early SAR
investigations, we quickly learned that the truncation of the tetra-
hydropyrimidine portion of the high-throughput screening hit 1
(Fig. 1) led to the more compact aminohydantoin 2, which showed
and 63 (3S7M) have been deposited in the Protein Data Bank, Research Collabo-
ratory for Structural Bioinformatics, Rutgers University, New Brunswick, New
Jersey.
⇑
Corresponding author.
E-mail address: malamas.michael@gmail.com (M.S. Malamas).
Present address: Lundbeck Research, 215 College Road Paramus, New Jersey.
Present address: Vitae Pharmaceuticals, 502 West Office Center DriveFort
a 10-fold enhancement in potency (IC50 = 3.4 l
M).14d Soon after-
à
ward, we carried out detailed structure–activity relationship stud-
ies around compound 2 that ultimately led to compounds 3a–c
(Fig. 1),14a which resulted in 1900-fold improvement of ligand
Washington, PA.
§
Present address: Novartis, 250 Massachusetts Avenue, Cambridge, MA.
Present address: AstraZeneca, Alderley Park, SK10 4TG, UK.
||
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.07.057

M. S. Malamas et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5164–5170
H2N
5165
Me
H2N
H2N
Me
N
N
X
N
N
N
O
3a-b
N
SAR
truncate
N
N
O
N
3c
OMe
S2'
N
Me
1(HTS Hit)
BACE1: IC50 = 38 uM
BACE2: IC50 = 38.3 uM
2
3b (X=F) 3c
3a(X=H)
IC50 = 3.4 uM
IC50 = 0.77 uM
IC50 >100uM
0.02
0.78
2.60
0.03
1.59
14.3
IC50 nM=0.04
IC50 nM=0.77
IC50 nM=12.5
Cathepsin D: IC50 >100uM
Figure 1. Initial optimization of HTS Hit 1.
Br
R3
R3
I
R3
I
a
b
7
affinity (IC50 = 20 nM; 3b vs 1) for the BACE1 enzyme. In addition,
derivative 3b demonstrated about 40- and 100-fold selectivity
against the closely related aspartyl proteases BACE2 and cathepsin
D, respectively. X-ray crystallographic studies14a,b and molecular
modeling calculations15 have indicated that interactions of the
P20 side chain of the ligand with residues of the S20 region are
key contributors to the ligand’s potency and selectivity. These
observations were attributed; (i) to the difference of the protein
amino acid sequences of the ‘flap’ loops of the BACE1 and BACE2
proteins (Pro70 in BACE1 is replaced with Lys70 in BACE2) at close
proximity to the bound ligand; and (ii) the apparent hydrogen-
bond interaction between the para-methoxy group of the phenyl
moiety of 3b projecting into the S20 pocket and residue Trp76 of
the BACE1 enzyme backbone (X-ray crystallographic observa-
tions14a). The amino acid differences of the S20 micro-region of
BACE1 and BACE2, together with the dynamic motion of the ‘flap’
loop upon ligand engagement apparently influence the ligand/pro-
tein contacts and impede on the ligand’s affinity for the BACE2 site,
thus resulting in increased selectivity.
The physiological functions of BACE2 have yet to be fully char-
acterized, and although BACE1/BACE2 double knockout mice have
been reported to be viable16 selective inhibitors against BACE2 are
highly desirable to avoid potential side effects in human clinical
trials. More importantly, selectivity against cathepsin D was
deemed to be essential to the clinical development of BACE1 inhib-
itors, since cathepsin D plays a critical role in cellular processes and
functions as a positive mediator in apoptosis.17
N
R5
N
R5
N
R5
c
N
H
N
N
H
R1
4
5
6a R1= alkyl
6b: R1 = SEM
R3, R5 = H, alkyl
R3
O
R5
Br
N
R1
R1
d
N
N
N
Br
R5
O
9
R3
8
H2N
Me
H2N
Me
N
N
R3
N
N
R3
N
f, R2-B(OH)2
or h (R1 = SEM), i
O
R2
e
N
O
Br
N
R5
N
R5
R1
R1
11a: R1 = alkyl
11b: R1 = SEM
11c: R1 = H
10
Scheme 1. Reagents: (a) I2, CAN, CH3CN; (b) NaH, R1-halide or SEM-Cl, THF; (c)
Pd(PPh3)2Cl2, CuI, Et3N, DMF; (d) KMnO4, NaHCO3, MgSO4, acetone, water; (e) 1-
methylguanidine, Na2CO3, H2O, dioxane, EtOH; (f) aryl(heteroaryl)-boronic acid,
Ph(PPh3)4, dioxane, water; (h) 4 N HCl, MeOH; (i) NaH, R-halide.
presence of magnesium sulfate and sodium bicarbonate furnished
diketones 9. These 1,2-di-substituted diketones 9 were converted
to the aminohydantoins 10 upon treatment with 1-methylguani-
dine in the presence of potassium carbonate in dioxane and ethanol
(1:1). Palladium-catalyzed cross coupling reaction of aminohydan-
toins 10 with any number of heteroaryl boronic acids (Suzuki cou-
pling19) or heteroaryl trialykl/triaryl stannanes (Stille coupling20) in
the presence of Pd(0) or Pd(II) catalysts under a variety of well
known conditions produced the desired products 11a. Deprotection
of the SEM group of 11b with hydrochloric acid (4 N) in methanol
afforded pyrazole 11c, which was then alkylated as before to fur-
nish products 11a.
To further expand the scope and breadth of the SAR studies of
the previously described aminohydantoins,14 in this Letter, we re-
port the introduction of pyrazolyl and thienyl P20 side chains on
the parent scaffold 3 in order to further explore the SAR around
the S20 pocket of the enzyme. As described, our intent was to cap-
italize on the residue differences between the BACE1 and BACE2
enzymes and the direct contact(s) of the ligand with residues in
this micro-region and investigate the feasibility of rationally
designing new potent and selective BACE1 ligands.
The compounds needed to delineate the SAR for this study were
prepared according to synthetic Schemes 1–3. The described pyra-
zoles were prepared according to Scheme 1. Pyrazole 4 was treated
with iodine and ammonium hexanitratocerium (IV) to afford 4-iod-
opyrazole 5, which was alkylated with various alky halides in the
presence of sodium hydride to furnish 6a. Alternatively, the pyra-
zole nitrogen was protected with 2-(trimethylsilyl)ethoxymethyl
chloride (SEM-chloride) and sodium hydride (6b), thus allowing
the introduction of N-substituents at a later stage of the synthesis
to broaden the scope and breadth of the SAR investigation. Sono-
gashira coupling18 of pyrazoles 6a–b with 3-bromophenyl acety-
lene 7 in the presence of bis(triphenylphosphino)palladium (II)
chloride, copper iodine and triethylamine afforded acetylenes 8.
Oxidation of acetylenes 8 with potassium permanganate in the
The described thiophenes were prepared according to Schemes
2 and 3. In Scheme 2, two routes were used to prepare diketones
14. In route a, Sonogashira coupling of thiophenes 12 with 3-
bromophenyl acetylene in the presence of bis(triphenylphosphi-
no)palladium(II) chloride, copper iodide and triethylamine afforded
acetylenes 13. Oxidation of 13 with potassium permanganate in the
presence of sodium bicarbonate and magnesium sulfate afforded
diketones 14. In route b, commercially available benzyl bromides
were converted to phosphonium salts 16 upon treatment with
triphenylphosphine. These phosphonium salts 16 were initially
treated with a suitable base (n-butyl lithium, or potassium tert-
butoxide) and then with the requisite chlorides 15b to furnish
ylides 17. Oxidation of ylides 17 with potassium permanganate in
the presence of sodium bicarbonate and magnesium sulfate

5166
M. S. Malamas et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5164–5170
Br
R1
S
R1
Br
I
O
b
S
Br
R2
R1
a
S
R2
R2
Route a
O
R1, R2 = H, alkyl
13
12
A
14
R1
Br
X
PPh3 Br
16
PPh3
O
Route b
b
O
S
Br
R2
R1
R2
d
S
15a: X =OH
15b: X = Cl
17
c
H2N
Me
H2N
Me
N
N
R2-B(OH)2
f
N
O
R2
e
N
O
R2
Br
A
R2
S
R1
S
R1
18
19
Scheme 2. Reagents: (a) Pd(PPh3)2Cl2, CuI, Et3N, DMF; (b) KMnO4, NaHCO3, MgSO4, acetone, water; (c) SOCl2, benzene; (d) n-BuLi, toluene; (e) 1-methylguanidine, Na2CO3,
H2O, dioxane, EtOH; (f) aryl(heteroaryl)-boronic acid, Pd2(dba)3, PPh3, Na2CO3, DMF, water.
Cellular potency of advanced compounds was determined via a
Br
Br
cell-based Ab inhibition (Ab40 or Ab42) in an enzyme-linked im-
mune sandwich assay (ELISA) in Chinese Hamster Ovary (CHO)
cells, recombinantly expressing human wild-type APP (CHO-
wt).22 Potential compound toxicity was assessed via mitochondrial
function using MTS readout (MTS kit from Promega).
a
b
c
S
S
S
O
O
O
O
O
21
22
20
In this report, we have undertaken studies to methodically ex-
plore the SAR around the S20 region of the binding pocket by intro-
ducing pyrazolyl and thienyl P20 side chains on the parent scaffold
3. As stated, our goal was to capitalize on the residue differences
between the BACE1 and BACE2 enzymes at the S20 mirco-region
of the binding pocket to build ligand selectivity and explore the di-
rect contact(s) of the ligand with residue Trp76 present in this re-
gion to rationally design new potent ligands. During the course of
our SAR studies, we have maintained the P3 groups (pyrimidine
and 2-F pyridine) of the ligand, which we have previously14 opti-
mized for potency. Introduction of the N-methyl pyrazolyl group
on the parent scaffold (3b, Fig. 1) resulted in about threefold loss
of potency (27 vs 3c; Table 1). The larger ethyl analog 28 was equi-
potent to the methyl analog 27. Introduction of a fluoro-substitu-
ent at the S1 phenyl of 28 afforded analog 29, which was about
twofold more potent than 28. The trifluoroethyl analog 30 was
equipotent for BACE1 to that of 28, and it was about sixfold more
selective against BACE2. This finding was not unexpected, since
introduction of electronegative groups (i.e., CF3) that project into
S20 region of the binding pocket have contributed to selectivity
enhancement against BACE2, as we have previously disclosed.14b
Increasing the size of the N-alkyl substituent (31–35) has resulted
in about 2- to 5-fold enhancement in potency. The n-butyl analog
35 was the most potent analog with an IC50 value of 8 nM. The
bulkier cyclopentyl and ethylcyclohexyl analogs 36 and 37 were
also potent. The increased hydrophobic character of these new
analogs (32–37) has resulted in loss of selectivity against BACE2.
Molecular modeling calculations have suggested that the longer al-
kyl groups project deep into the S20 micro-region and make addi-
tional van der Waals contacts with hydrophobic residues, which
are most likely key contributors to the enhanced affinity for both
BACE1 and BACE2. All the N-alkyl analogs were selective against
cathepsin D, indicative of the distinct make-up of the cathepsin
D at this region, by comparing the amino acid sequences of the
two enzymes. In order to support our molecular modeling calcula-
tions and verify the orientation of the ligand within the BACE1
binding pocket, compound 28 was co-crystallized with BACE1. As
I
Br
e, f, g
d
S
O
O
O
S
24
23
H2N
H2N
CH3
CH3
O
N
N
N
O
h
N
Ar
Ar
O
S
S
O
O
26
25
Scheme 3. Reagents: (a) Br2, AlCl3, CH2Cl2; (b) ethylene glycol, p-TsOH; (c) n-BuLi,
Et2O, I2; (d) 3-Br phenyl acetylene, Pd(PPh3)2Cl2, CuI, Et3N, DMF; (e) KMnO4,
NaHCO3, MgSO4, H2O; (f) 1-mehtylguanidineꢀHCl, Na2CO3, dioxane, EtOH; (g) Ar-
B(OH)2, Pd(PPh3)2Cl2, Na2CO3, DME, H2O; (h) 1 N HCl, Et2O.
produced diketones 14. These 1,2-di-substituted diketones 14 were
converted to the final products 19, as before. In Scheme 3, Friedel–
Crafts bromination (Br2, AlCl3) of thiophene 20 afforded 4-bromo-
thiophene 21, which was then protected as the ketal 22 with ethyl-
ene glycol and para-toluenesulfonic acid. Metal–halogen exchange
of 22 with N-butyl lithium followed by quenching with iodine affor-
ded iodo-thiophene 23. Intermediate 23 was converted first to acet-
ylene 24 and then to aminohydantoin 25, as before. Acid hydrolysis
of the ketal of 25 with hydrochloric acid in ethyl ether afforded the
final product 26. Detailed experimental protocols of the compounds
described in this Letter have been reported.21
With the necessary tools needed to fully investigate the SAR, the
diverse array of compounds were profiled for their potency at the
target enzyme, BACE1, as well as the closely related enzymes that
were the focus of this investigation. The primary screening assays
utilized for the program were homogenous, continuous fluores-
cence resonance energy transfer (FRET) protocols, representing
competitive inhibition for BACE1, BACE2, and cathepsin D.21 The
data are reported in Tables 1–4.

M. S. Malamas et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5164–5170
5167
Table 1
Pyrimidinyl-pyrazole analogs
Me
O
H2N
N
N
N
N
N
R2
N
R1
a
a
a
a
Compd
R1
R2
BACE1IC50
0.10
(l
M)
LEb (BACE1)
BACE2 IC50
(
lM)
Cathepsin D IC50 (lM)
ELISA EC50 (lM)
27
28
29
30
31
32
33
34
35
36
37
CH3
CH2CH3
CH2CH3
CH2CF3
H
H
F
F
F
H
H
H
H
H
H
0.020
0.020
0.020
0.016
0.018
0.019
0.019
0.018
0.021
0.018
0.017
0.86
1.47 0.4
2.37
8.79
6.65
0.13
0.02
0.03
0.15
54.4
75.8 1.3
28.5
38.6
28.3
11.3
7.4
4.4 1.5
1.7 0.3
0.6 0.07
0.9 0.1
1.6 0.4
0.8 0.15
0.5 0.05
0.3 0.03
0.7 0.1
1.5 0.27
0.3 0.03
0.08 0.02
0.04
0.12
0.08
0.03
0.03
0.02
0.008
0.07
CH(CH3)2
CH2CH(CH3)2
CH2CH2CH(CH3)2
CH2CH2CH(CH3)3
CH2CH2CH2CH3
Cyclopentyl
CH2CH2cyclohexyl
3.8
11.9
17.3
4.4
0.17
0.02
0.02+ 0.001
a
IC50 and EC50 values are the means of at least two experiments SD. Values without SD are for a single determination only.
LE is the ligand efficiency based on the expressing the binding affinity as pIC50 = ꢁlogIC50 and using the MW as reference, or LE = pIC50/MW.
b
Table 2
Pyridinyl-pyrazole analogs
Me
O
H2N
N
N
F
N
N
R2
N
R1
a
a
a
a
Compd
R1
R2
BACE1 IC50
(lM)
BACE2 IC50
(
lM)
Cathepsin D IC50
(lM)
ELISA EC50
(l
M)
38
CH2CH3
CH2CH3
CH2CH3
CH2CH3
CH2CH2F
CH2CF3
CH(CH3)2
CH2CH(CH3)2
CH2CH2CH(CH3)2
CH2CH2CH(CH3)3
CH2CH(CH2CH3)2
CHCH2CH2CH3
Cyclopentyl
H
H
H
F
H
H
F
H
H
H
H
H
H
0.03
0.01
1.96
0.03
0.05
0.03 0.006
0.04
0.02 0.001
0.016 0.001
0.02 0.003
0.02 0.001
0.01
0.45
0.21
7.39
0.87
0.65
0.5 0.02
3.12
0.05
0.018
0.02
12.98
6.69
27% @ 100
4.03
13.12
4.79 0.3
4.85
1.47
0.83
0.36
1.42
0.49 0.18
0.26 0.07
>2.5
(S)-38
(R)-38
39
40
(S)-41
42
43
44
45
46
lM
0.42 0.16
1.17 0.28
0.43 0.07
0.5 0.01
0.14 0.03
0.16 0.01
0.11 0.002
0.16 0.03
0.2 0.03
0.41 0.09
0.036
0.05
0.12
47
48
1.91
1.79
0.02
a
IC50 and EC50 values are the means of at least two experiments SD. Values without SD are for a single determination only.
shown in Figure 2, the pyrimidine moiety projects deep into S3
pocket and makes a water-bridge contact with Ser229 through
the buried water found near the catalytic site of the enzyme, sim-
ilarly to the previously reported X-ray crystallographic data of the
parent scaffold 3.14a In addition, the crystal structure shows that
the pyrazolyl moiety projects directly toward the S20 pocket in
the same region of the P20 phenyl analog of the parent compound
(Fig. 2). The pyrazole ring nitrogen does not interact with Trp71;
however, the van der Waals contacts between the ligand and the
enzyme at the S20 region can possibly explain the ligand’s in-
creased potency of the N-alkyl analogs. The ‘flap’ loop region of
the protein backbone orients at a closed position similar to the
apo structure, in contrast to the parent scaffold, where the P20
substituted phenyl moiety pushes the ‘flap’ into an open position
(Fig. 2).
3- to 5-fold more potent than the pyrimidine analogs. Neverthe-
less, the pyridine analogs demonstrated an unexpectedly decrease
in selectivity against both the highly conserved BACE2 and cathep-
sin D enzymes. Retrospective examination of the amino acid se-
quences of the enzymes, it reveals that there are two amino acid
differences in the S3 region between BACE1 and BACE2 (BACE1
Asn295 is replaced by Leu in BACE2, and BACE1 Gln11 in replaced
by Arg in BACE2). In addition, the S3 pocket of the cathepsin D en-
zyme is smaller to that of BACE1 and overlaps, in part, with the re-
gion occupied by the conserved S3-water of the BACE1 enzyme
(not shown). Even though both the pyrimidine and pyridine groups
can be considered equal in size, their distinct electronic make-up
together with the residue differences at the S3 region of the en-
zymes could account for the ligand’s affinity divergence.
Furthermore, an investigation of the chiral preference of the
racemic mixtures of this class of compounds, once separated,
showed a clear preference for the S-enantiomer [see (S)-38 vs
To further expand the scope of the SAR, the pyridine analogs
(38–48; Table 2) were prepared. The pyridine analogs were about

5168
M. S. Malamas et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5164–5170
Table 3
Pyrimidinyl-thiophene analogs
Me
O
H2N
N
N
N
N
R1
S
R2
a
a
a
a
Compd
R1
R2
BACE1 IC50
(
lM)
BACE2 IC50
(l
M)
Cathepsin D IC50
(lM)
ELISA EC50 (lM)
49
50
51
52
53
54
55
56
H
CH3
H
CH2CH3
CH2CH3
Cl
CH3
COCH2CH3
H
H
CH3
H
CH3
H
0.07 0.01
0.14
0.37
0.08 0.03
0.18
0.08 0.02
0.47
0.68
0.78
0.61
0.56
0.49
0.62
1.15
0.07
26.4
30.5
38.65
23.8
11.8
29.92
31.65
13.05
0.47 0.19
0.55 0.05
1.31 0.21
0.37 0.02
0.23 0.07
0.34 0.04
2.23 0.7
Cl
H
0.015 0.01
0.07 0.02
a
IC50 and EC50 values are the means of at least two experiments SD. Values without SD are for a single determination only.
Table 4
Pyridinyl-thiophene analogs
Me
O
H2N
N
N
R3
R4
R5'
S
R2'
a
a
a
a
0
0
Compd
R3
R4
R2
R5
BACE1 IC50
(lM)
BACE2 IC50
(lM)
Cathepsin D IC50
(
lM)
ELISA EC50 (lM)
57
58
59
60
61
62
63
(S)-63
(R)-63
3-Pyridine
2-F-3-Pyridine
3-Pyridine
2-F-3-Pyridine
3-OMe-5-pyridine
2-Pyrazine
2-F-3-Pyridine
2-F-3-Pyridine
2-F-3-Pyridine
H
H
F
H
H
H
H
H
H
H
H
H
0.06 0.01
0.04 0.01
0.09
0.03
0.04 0.02
0.09
0.19 0.12
0.23
0.74
0.41
0.02 0.01
0.44
0.07 0.02
0.05
5.45
22.0 2.4
8.0
12.8
2.2
18.6
33.0
4.1 0.4
2.5
0.42 0.14
0.15 0.03
0.37 0.16
0.2 0.07
0.22 0.04
0.45 0.22
0.03 0.006
0.01 0.003
2.9 0.47
H
H
H
H
H
COEt
COEt
COEt
F
H
H
H
H
H
0.015 0.002
0.01 0.01
47% @ 2.5
l
M
78.4
a
IC50 and EC50 values are the means of at least two experiments SD. Values without SD are for a single determination only.
(R)-38]. This was apparent as the crystal structures and modeling
studies bore this out. For the less active enantiomer, one could
envision keeping the respective P1–P3 and P20 side-chains in their
designed pockets, but in doing so, the aminohydantoin would be
flipped into an unfavorable conformation and pressed into the
backbone of the active site of the enzyme. An alternate binding
pose of ligand S-38 within the binding pocket, by flipping the
respective P1–P3 and P20 side-chains while retaining the interac-
tions of the amino-imidazole moiety with the catalytic aspartic
acids, would result in much less contacts and steric interactions
conflicts between the ligand and the enzyme, and could reflect
the 196-fold potency loss for the R-38 enantiomer.
Ser229
wat
Asp228
Asp32
S3
Thr232
wat
S1
The cell-based activity of the N-alkyl analogs also tracked well
with their increased molecular binding with EC50 values of 0.1–
S2’
0.5
l
M in the ELISA assay for the most potent analogs.
closed position
FLAP
Focusing our attention once again back to the interaction of the
parent scaffold 3b with residue Trp76 of the S20 region,14 we intro-
duced a P20 thienyl nucleus in lieu of the pyrazolyl moiety in an at-
tempt to restore the hydrogen-bond interaction between ligand
and protein. To this end, we first have prepared analogs carrying
a P3 pyrimidine group. Thiophene 49 (Table 3) was equipotent to
N-ethyl pyrazole 28. Introduction of electron donating (alkyls) or
electron withdrawing (halogens) substituents at the 20 and 50-posi-
tion of the thiophene nucleus (50–55) have marginally affected the
ligand’s potency. The 50-substituted analogs were about 2- to
Trp76
Tyr71
Figure 2. Crystal structures of BACE1 complexed with 28 (shown in green) and 3a
(shown in yellow) are overlayed. Tyr71 sits at the closed position in complex with
28, compared to the open orientation in complex with 3a. Key hydrogen bonding
interactions between ligand and protein at the catalytic aspartic acids Asp32 and
Asp228, Trp76 of the S02 region and the S3 conserved water are highlighted with
blue dashed lines. The pyrazolyl ring nitrogen does not form hydrogen bonding
with Trp76.

M. S. Malamas et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5164–5170
5169
3-fold more potent than the analogous 20-substituted analogs.
Introduction of a ketone functionality at the 50-position of the thi-
ophene ring has resulted in about fivefold enhancement of the li-
gand’s potency (56 vs 52). To further expand the scope of these
findings, we have prepared analogs carrying a P3 pyridine moiety
(57–61; Table 4). We have previously shown that the P3 pyridine
enhanced the ligand’s affinity by about 2- to 3-fold. In agreement
with the previous findings, the prepared thiophene-pyridines
(57–61) were about 3-fold more potent than the pyrimidine ana-
logs. The 2-pyrazine analog 62 was similar in potency to that of
pyridine 57. Introduction of a ketone functionality at the 50-posi-
tion of the thiophene moiety (entry 63) has resulted in about
threefold enhancement of the ligand’s affinity. Molecular modeling
calculations (Fig. 3) have suggested that a hydrogen-bond interac-
tion between the carbonyl’s lone pair electrons and the NH hydro-
gen of Trp76 could possibly explain the potency increase of the
keto-analogs (56 vs 49, 63 vs 58). In order to support our molecular
modeling calculations and verify the orientation of the thiophene
ligand within the BACE1 binding pocket, compound 57 was co-
crystallized with BACE1. As shown in Figure 3, the pyridine moiety
projects deep into S3 pocket and makes a water-bridge contact
with Ser229 through the buried water found near the catalytic site
of the enzyme. In addition, the crystal structure shows that the li-
gand makes several other key hydrogen bonding interactions with
the catalytic aspartic acids (Asp32 & Asp228), and a rather weak
hydrogen bonding interaction with Trp76 in the S20 pocket be-
tween the thiophene sulfur and the NH of Trp76. The orientation
of the thiophene ligand in the BACE1 binding pocket is similar to
the S20 phenyl to that of the parent compound 3b.14a The thiophene
analogs also have shown high chiral preference [(S)-63 vs (R)-63].
To explain the enhanced potency of the 5-ketone thiophene
analogs 56 and 63, molecular modeling calculations (Fig. 4) have
suggested that a hydrogen-bond interaction between the car-
bonyl’s lone pair electrons and the NH hydrogen of Trp76 could
possibly explain the potency increase of these analogs (56 vs 49,
63 vs 58).
The thiophene analogs were highly selective against cathepsin
D, while they were only modestly selective against BACE2. The
cell-based activity of the thiophene analogs tracked well with their
increased molecular binding with EC50 values of 0.07–0.2 lM in
the ELISA assay for the most potent analogs. Despite the robust
cell-based activity of the pyrazole and thiophene analogs, in vivo
assessment showed no significant reduction of brain Ab40, perhaps
due to their limited brain permeability (AUCbrain/AUCplasma ꢂ0.1).
In summary, we have methodically explored the S20 region of
the BACE1 binding pocket, by replacing the P20 aryl group of the
parent compound 3 with pyrazolyl and thienyl P20 side chain moi-
eties. The P20 pyrazole SAR studies have pointed to two key obser-
vations; (i) small groups attached at the nitrogen of the P20
pyrazolyl moiety and (ii) the presence of the P3 pyrimidine nucleus
projecting into the S3 region of the binging pocket are critical com-
ponents to the ligand’s selectivity against BACE 2 and cathepsin D.
The SAR studies of the P20 thiophene ligands have produced highly
potent BACE 1 inhibitors with excellent selectivity against cathep-
sin D, but only modest selectivity against BACE2. A keto-function-
ality at the 50-positon of the thiophene nucleus has augmented
about 3- to 5-fold the potency of the ligand, which is likely due
to a hydrogen-bond interaction between the 50-carbonyl oxygen
of the keto-analogs and the NH of the Trp76 residue of the S20
sub-pocket, as suggested by molecular modeling calculations. The
cell-based activity of these new analogs tracked well with their in-
Figure 3. Crystal structure of BACE1 complexed with 57. Tyr71 (FLAP) sits at the
closed position. Key hydrogen bonding interactions between ligand and protein at
the catalytic aspartic acids Asp32 & Asp228, and the S3 conserved water are
highlighted with yellow dashed lines.
creased molecular binding with EC50 values of 0.07–0.2 lM in the
ELISA assay for the most potent analogs. In vivo assessment of pyr-
azole and thiophene analogs showed no significant reduction of
brain Ab40, perhaps due to their limited brain permeability (AUC-
brain/AUCplasma ꢂ0.1).
References and notes
1. Enz, A.; Amstutz, R.; Boddeke, H.; Gmelin, G.; Malanowski, J. Am. J. Geriatr.
Pharmacother. 2005, 3, 39.
2. Arendt, T. Neuroscience 2001, 102, 723.
3. Dickson, D. W. J. Neuropathol. Exp. Neurol. 1997, 56, 321.
4. (a) Selkoe, D. J. Science 1997, 275, 630; (b) Hardy, J.; Selkoe, D. J. Science 2002,
297, 353.
5. Vassar, R.; Citron, M. Neuron 2000, 27, 416.
6. Selkoe, D. J. Physiol. Rev. 2001, 81, 741.
7. Hussain, I.; Powell, D.; Howlett, D. R.; Tew, T. G.; Meek, T. D.; Chapman, C.;
Gloger, I. S.; Murphy, K. E.; Southan, C. D.; Ryan, D. M.; Smith, T. S.; Simmons, D.
L.; Walsh, F. S.; Dingwall, C.; Christie, G. Mol. Cell. Neurosci. 1999, 14, 419.
8. Lin, X.; Koelsch, G.; Wu, S.; Downs, D.; Dashti, A.; Tang, J. Proc. Natl. Acad. Sci.
U.S.A. 2000, 97, 1456.
9. Sinsa, S.; Anderson, J. P.; Barbour, R.; Basi, G. S.; Caccavello, R.; Davis, D.; Doan,
M.; Dovey, H. F.; Frigon, N.; Hong, J.; Jacobson-Croak, K.; Jewett, N.; Keim, P.;
Knops, J.; Lieberburg, I.; Power, M.; Tan, H.; Tatsuno, J.; Tung, J.; Schenk, D.;
Seubert, P.; Suomensaari, S. M.; Wang, S.; Walker, D.; Zhao, J.; McColongue, L.;
John, V. Nature 1999, 402, 537.
Figure 4. Molecular modeling calculations suggest that the 1-propanone group of
63 extends into S20 region and sits in close proximity to residue Trp76. A hydrogen-
bonding interaction between the carbonyl oxygen and the NH of the Trp76 can
possibly explain the near 3- to 5-fold increase in potency of the 1-propanone
bearing analogs.

5170
M. S. Malamas et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5164–5170
10. Vassar, R.; Bennett, B. D.; Babu-Khan, S.; Kahn, S.; Mendiaz, E. A.; Denis, P.;
Teplow, D. B.; Ross, S.; Amarante, P.; Loeloff, R.; Luo, Y.; Fischer, S.; Fuller, J.;
Edenson, S.; Lile, J.; Jarosinski, M. A.; Biere, A. L.; Curran, E.; Burgess, T.; Louis, J.-
C.; Collins, F.; Treanor, J.; Rogers, G.; Citron, M. Science 1999, 286, 735.
11. Selkoe, D. J. Ann. Intern. Med. 2004, 140, 627.
15. Docking calculations were performed using the Schrodinger software package
Maestro (www.schrodinger.com). After the binding site model was validated
by redocking of the X-ray ligands, docking of analogs was performed using the
GLIDE or CombiGlide with SP (standard precision) scoring. Conformational
analysis was performed using Macromodel with the OPLS-AA force fields, and a
GBSA solvation model implemented in Maestro (www.schrodinger.com). The
catalytic aspartic acids utilized in the modeling were treated as dianionic,
while the N1 of imidazolin-4-one at ligands was protonated.
16. Dominguez, D.; Tournoy, J.; Hartmann, D.; Huth, T.; Cryns, K.; Deforce, S.;
Serneels, L.; Camacho, I. E.; Marjaux, E.; Craessaerts, K.; Roebroek, A. J. M.;
Schwake, M.; D’Hooge, R.; Bach, P.; Kalinke, U.; Moechars, D.; Alzheimer, C.;
Reiss, K.; Paul Saftig, P.; De Strooper, B. J. Biol. Chem. 2005, 280, 30797.
17. Kagedal, K.; Johansson, U.; Llinger, K. O. FASEB 2001, 1592.
18. Sonogashira, K.; Tohda, Y.; Hagihara, N. Tetrahedron Lett. 1975, 50, 4467.
19. Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457.
20. Stille, J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508.
21. Malamas, M.; Zhou, P,; Fobare, W.; Solvibile, W.; Gunawan, I.; Erdei, J.; Yan, Y.;
Andrae, P.; Quagliato, D. U.S. Patent, 0004786, 2007.
22. Wu, P.; Brand, L. Anal. Biochem. 1994, 218, 1.
12. Alloul, K.; Sauriol, L.; Kennedy, W.; Laurier, C.; Tessier, G.; Novosel, S.;
Contandriopoulos, A. Arch. Gerontol. Geratr. 1998, 27, 189.
13. (a) Enz, A.; Amstutz, R.; Boddeke, H.; Gmelin, G.; Malanowski, J. Prog. Brain Res.
1993, 98, 431; (b) Fillit, H.; Hill, J. Am. J. Geriatr. Pharmacother. 2005, 3, 39.
14. (a) Malamas, M. S.; Erdei, J.; Gunawan, I.; Turner, J.; Hu, Y.; Wagner, E.; Fan, K.;
Chopra, R.; Olland, A.; Bard, J.; Jacobsen, S.; Magolda, R.; Pangalos, M.;
Robichaud, A. J. J. Med. Chem. 2010, 53, 1146; (b) Malamas, M. S.; Barnes, K.;
Johnson, M.; Hui, Y.; Zhou, P.; Turner, J.; Hu, Y.; Wagner, E.; Fan, K.; Chopra, R.;
Olland, A.; Bard, J.; Pangalos, M.; Reinhart, P.; Robichaud, A. J. Bioorg. Med.
Chem. 2010, 18, 630; (c) Malamas, M. S.; Robichaud, A.; Erdei, J.; Quagliato, D.;
Solvibile, W.; Zhou, P.; Morris, K.; Turner, J.; Wagner, E.; Fan, F.; Jacobsen, S.;
Reinhart, P.; Pangalos, M.; Riddell, D. Bioorg. Med. Chem. Lett. 2010, 20, 6597;
(d) Malamas, M. S.; Erdei, J.; Gunawan, I.; Barnes, K.; Johnson, M.; Hui, Y.;
Turner, J.; Hu, Y.; Erik Wagner, E.; Kristi Fan, K.; Olland, A.; Bard, J.; Robichaud,
A. J. Med. Chem. 2009, 52, 6314.