Application of fragment-based lead generation to the discovery of novel, cyclic amidine β-secretase inhibitors with nanomolar potency, cellular activity, and high ligand efficiency

Article abstract of DOI:10.1021/jm070829p

Fragment-based lead generation has led to the discovery of a novel series of cyclic amidine-based inhibitors of β-secretase (BACE-1). Initial fragment hits with an isocytosine core having millimolar potency were identified via NMR affinity screening. Stru

Full text of DOI:10.1021/jm070829p

5912
J. Med. Chem. 2007, 50, 5912-5925
Application of Fragment-Based Lead Generation to the Discovery of Novel, Cyclic Amidine
â-Secretase Inhibitors with Nanomolar Potency, Cellular Activity, and High Ligand Efficiency§
Philip D. Edwards,*,† Jeffrey S. Albert,† Mark Sylvester,† David Aharony,† Donald Andisik,† Owen Callaghan,+
James B. Campbell,† Robin A. Carr,+ Gianni Chessari,+ Miles Congreve,+ Martyn Frederickson,+ Rutger H. A. Folmer,‡
Stefan Geschwindner,‡ Gerard Koether,† Karin Kolmodin,‡ Jennifer Krumrine,† Russell C. Mauger,† Christopher W. Murray,+
Lise-Lotte Olsson,‡ Sahil Patel,+ Nate Spear,† and Gaochao Tian†
CNS DiscoVery Research, AstraZeneca Pharmaceuticals LP, 1800 Concord Pike, PO Box 15437, Wilmington, Delaware 19850-5437,
Global Structural Chemistry, AstraZeneca, S-431 83 Mo¨lndal, Sweden, and Astex Therapeutics Ltd, 436 Cambridge Science Park,
Milton Road, Cambridge CB4 0QA, United Kingdom
ReceiVed July 9, 2007
Fragment-based lead generation has led to the discovery of a novel series of cyclic amidine-based inhibitors
of â-secretase (BACE-1). Initial fragment hits with an isocytosine core having millimolar potency were
identified via NMR affinity screening. Structure-guided evolution of these fragments using X-ray
crystallography together with potency determination using surface plasmon resonance and functional enzyme
inhibition assays afforded micromolar inhibitors. Similarity searching around the isocytosine core led to the
identification of a related series of inhibitors, the dihydroisocytosines. By leveraging the knowledge of the
ligand-BACE-1 recognition features generated from the isocytosines, the dihydroisocytosines were efficiently
optimized to submicromolar potency. Compound 29, with an IC50 of 80 nM, a ligand efficiency of 0.37,
and cellular activity of 470 nM, emerged as the lead structure for future optimization.
Introduction
The most widely held hypothesis for the pathogenesis of
Alzheimer’s disease is that mismetabolism of amyloid precursor
protein (APP) generates â-amyloid (Aâ) fibrils that aggregate
to form neurofibrillary plaques, one of the morphological
hallmarks of AD.4 These Aâ fibrils and plaques interact directly
with neurons initiating neurodegeneration leading to dementia.
Two proteases are responsible for the proteolytic processing of
APP: â-secretase and γ-secretase.5 â-Secretase (BACE-1) is a
membrane-associated aspartyl protease and cleaves the membrane-
bound APP at an extramembrane site.6-9 Subsequent to â-secre-
tase cleavage, γ-secretase, a member of a unique class of
intramembrane aspartyl proteases cleaves the C99 fragment
generated from APP by BACE-1 at an intramembrane site,
generating Αâ and releasing it from the membrane. Since
â-secretase cleavage of BACE-1 is the rate-controlling step
in APP processing to Αâ, it is considered a leading target
for treating AD via control of Αâ release and deposition,
and a major effort has been directed to identify â-secretase
inhibitors.
Alzheimer’s disease (ADa) is a devastating disease affecting
approximately 4.5 million individuals in the US.1 AD primarily
affects individuals over 60, and as the baby boomers age, the
percentage of the population afflicted with AD will increase,
with recent estimates suggesting that by 2050 as many as 16
million Americans will be afflicted.1 In addition to the individual
suffering and financial cost of AD, the impact on the families
of those with AD is also significant. One in 10 Americans report
having a family member with AD, and 7 out of 10 AD patients
live at home.2
Consequently, the need to develop drugs to treat AD is a
high priority in both academia and the pharmaceutical industry,
with most major pharmaceutical companies having active AD
programs. Currently, there are only two classes of drugs
approved for treating AD.3 Several acetylcholine esterase
inhibitors have been on the market for a number of years and
improve cognition by increasing synaptic levels of acetylcholine.
Recently, a new drug has been approved, memantine, which is
believed to improve cognition by controlling levels of glutamate
via antagonism of the NMDA receptor. However, these drugs
only treat the symptomology of AD and do not address the
underlying neuropathology.
Over the past decade, a large number of peptidic and pseudo-
peptidic BACE-1 inhibitors have been reported.10-15 The vast
majority of these possess a hydroxyl group that displaces an
active-site water molecule and forms hydrogen bonding interac-
tions with the catalytic aspartates.16-25 A seminal report by
Ghosh and Tang described the X-ray crystal structure of a large
peptidic inhibitor (OM99-2, 1, Figure 1) complexed to BACE-
126 and paved the way for structure-based design approaches
to the development of BACE-1 inhibitors. While OM99-2 is a
very potent BACE-1 inhibitor, its large size makes it unattractive
for drug development, especially for CNS diseases for which
the drug needs to penetrate the blood-brain barrier.27,28 Ligand
efficiency (LE) has been recently introduced as a means of
assessing a compound’s potency relative to its size.29 In this
report, we define LE as the free energy of binding divided by
the heavy atom count (HAC), LE ) -∆G/HAC ≈ -RT
ln(IC50)/HAC. If we consider as a target for a drug candidate a
molecular weight of <500 and an IC50 <10 nM, it can readily
§ Coordinates for the â-secretase complexes with compounds 8c, 24, and
27 have been deposited in the Protein Data Bank under accession codes
2va5, 2va6, and 2va7, together with the corresponding structure factor files.
* To whom correspondence should be addressed. Phone: 302-886-8371;
fax 302-886-5382; e-mail: philip.edwards@astrazeneca.com.
† AstraZeneca CNS Discovery Research.
‡ AstraZeneca Global Structural Chemistry.
+ Astex Therapeutics Ltd.
a Abbreviations. Aâ, beta-amyloid; ACN, acetonitrile; AD, Alzheimer’s
disease; APP, amyloid precursor protein; BACE-1, â-site APP cleaving
enzyme; n-BuLi; n-butyllithium; CDI, 1,1′-carbonyldiimidazole; DMF,
dimethylformamide; DCM, dichloromethane; EtOAc, ethyl acetate; EtOH,
ethanol; FRIT, fragment hit; HCl, hydrochloric acid; LE, ligand efficiency;
MeOH, methanol; MeI, methyl iodide; NaCl, sodium chloride; NaH, sodium
hydride; SFC, supercritical fluid chromatography; TEA, triethylamine; TFA,
trifluoroacetic acid; THF, tetrahydrofuran.
10.1021/jm070829p CCC: $37.00 © 2007 American Chemical Society
Published on Web 11/07/2007

Cyclic Amidine â-Secretase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 5913
Figure 1. Examples of previously reported peptidic (1) and nonpeptidic (2-4) BACE-1 inhibitors, and initial fragment hits 5-8.
be derived that an LE of >0.3 is desired. By comparison, the
peptidic inhibitor OM99-2 (MW ) 893, Ki ) 1.6 nM) has an
LE of only 0.19.30
Chemistry
The synthesis of the isocytosine analogues is outlined in
Scheme 1. Condensation of a â-ketoester 9 with guanidine
afforded the 6-substituted isocytosines 8.41 Methylation of the
N3 nitrogen was effected with MeI.42 In the case of the
isocytosines, methylation occurred predominantly on the N3
nitrogen. Minor amounts of other methylated products were
readily removed by chromatography. Preparation of the biphenyl
analogues 11 was achieved via Suzuki coupling of bromide 10e.
Preparation of the dichlorophenyl dihydroisocytosine deriva-
tives entailed condensation of the R,â-unsaturated esters 12 with
guanidine43 (Scheme 2). Protection of the exocyclic amino group
of 13d as the N,N-dimethylformamidine,44 methylation with MeI
followed by deprotection with methanolic ammonia44 afforded
the N3-methyldihydroisocytosine 16. In contrast to the isocy-
tosines, the dihydroisocytosines required the protection step to
avoid significant amounts of undesired methylation products.
Application of the above route to the construction of other
6-substituted dihydroisocytosines proved to be limited because
of the generation of large amounts of unwanted mono- and
polymethylated side products during the methylation step.
Consequently, we developed a synthesis in which the N3-methyl
group was regiospecifically incorporated prior to construction
of the dihydroisocytosine ring system (Scheme 3). The lynchpin
for this route was the R,â-unsaturated N-methylcyanamide 20,45
wherein the N-methyl substituent is ultimately transformed into
the N3-methyl group. Michael addition of methoxybenzylamine
to 20, followed by intramolecular cyclization yielded the desired
While the vast majority of reported BACE-1 inhibitors are
peptidic or pseudopeptidic, a number of reports have emerged
revealing BACE-1 inhibitors with little or no peptidic chara-
cter.31-35 Particularly noteworthy are two recent reports of
nonpeptidic BACE-1 inhibitors that possess novel functionality
for interacting with aspartyl proteases. Cole et al. have described
a series of acyl guanidines (2) in which the guanidine group
interacts with the catalytic aspartates.36 Rajapakse et al. have
reported a series of oxadiazole BACE-1 inhibitors (3) in which
the nitrogens of the oxadiazole ring form hydrogen bonds to
the flap region of BACE-1.37 These inhibitors represent
significant advances in the development of nonpeptidic, drug-
like BACE-1 inhibitors and possess moderate to good LE.
Previously, we revealed the discovery of a novel series of
2-aminopyridine BACE-1 inhibitors, 4, using high throughput
crystallographic screening.38,39 In an accompanying paper in this
journal issue, we reported the identification and characterization
of a related series of isocytosine-based inhibitors, 5, using NMR
and other biophysical techniques.40 Similarity searching of our
corporate compound collection led to screening of compounds
6-8a, with 8a emerging as the lead isocytosine fragment hit.
Herein, we describe the optimization of the isocytosine inhibitors
and disclose the discovery of a novel series of dihydroisocy-
tosines that possess nanomolar potency, excellent LE, and
cellular activity.

5914 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Scheme 1. Synthesis of Isocytosines 8, 10-11a
Edwards et al.
binding interactions; an enzyme inhibition assay is required to
accomplish this. Without such a functional assay, it cannot be
assured that any new binding interactions actually increase the
affinity of the ligand for the enzyme. We used a fluorescence
resonance energy transfer (FRET) assay as our standard assay
for determination of functional inhibition. However, because
of the observed propensity for nonspecific inhibition of BACE-1
at high ligand concentrations together with solubility limitations
of many ligands under the FRET assay conditions, we found
the FRET assay often to be unreliable for determining potencies
>100 µM. To assist in the determination of the rank order
potency of fragment hits (FRITs) with affinities between 1000
and 100 µM, we employed surface plasmon resonance (SPR)
spectroscopy, a technique which allows determination of binding
constants as well as identification of compounds with noncom-
petitive modes of inhibition. The third assay we utilized was
an electrochemoluminescence-based assay and is referred to as
the IGEN assay. This assay was used for compounds having
intrinsic color or fluorescence that interfered with the FRET
colorimetric readout. Because of the fact that the various assays
were brought on stream at different times and that the use of a
particular assay was dependent upon compound properties and
potency ranges, only a limited number of compounds in this
report have inhibition values reported for all three assays.
However, as can be seen in Table 1, where a compound has
inhibition data from all three assays, the agreement was quite
good, with a difference of generally less than two-fold.
Consequently, we were confident in assessing the relative
potency between compounds even when the inhibition values
were from different assays, especially since the goal during this
phase of the project was not to identify structural changes that
resulted in minor improvements in potency but rather to drive
potency increases of low affinity FRITs to submicromolar
potency and to identify preferred core scaffolds. Details of the
assays are included in the Experimental Section.
a Reagents and conditions: (a) NH2C()NH)NH2‚H2CO3, EtOH; (b)
CH3I, K2CO3, DMF; (c) RB(OH)2, Cs2CO3, PdCl2(PPh3)2.
Scheme 2. Synthesis of Dihydroisocytosines 13, 16a
Crystallography. A key component of our FRIT evolution
strategy was the determination of the structures of our inhibitors
bound to BACE-1. We previously described the method for the
production of BACE-1 protein crystals suitable for soaking
experiments.47 Compounds were soaked into the BACE-1
crystals. AutoSolve was used to fit and score fragments into
the Fo-Fc electron density maps that were calculated after initial
automatic refinement against an unligated BACE-1 structure.
Protein ligand structures of bound compounds were then
subjected to further refinement steps. Table 2 lists the crystal-
lographic data for the three structures reported herein.
Results and Discussion
a Reagents and conditions: (a) NH2C()NH)NH2‚HCl, NaOMe, EtOH
or NMP; (b) (CH3)2NCH(OCH3)2, DMF; (c) CH3I, K2CO3, DMF; (d) 7N
NH3/MeOH.
We embarked on our search for drug-like BACE-1 inhibitors
by implementing more traditional lead generation approaches
such as high-throughput screening, rational design of pseudopep-
tidic transition-state isosteres, and virtual screening using the
published crystal structure of BACE-1 complexed with OM99-
2.26 While we did identify pseudopeptidic inhibitors with
potencies of <10 µM, they lacked defined SAR and we were
unable to maintain this level of potency in compounds with good
LE (>0.25). We encountered a different issue with hits derived
from HTS and virtual screening. Utilizing NMR to assess the
mechanism of inhibition, we discovered that many of the HTS
and virtual screening hits were noncompetitive inhibitors, and
thermal unfolding studies in the presence of those hits suggested
that some of these owed their inhibition to destabilization of
the BACE-1 tertiary structure. Consequently, we did not
consider any of the hits from HTS or virtual screening
sufficiently attractive and decided not to pursue them.
N3-methyl-substituted intermediate 21. Deprotection of the N1-
methoxybenzyl group46 afforded intermediate 22. Hydrogenation
to remove bromide of 22a afforded the 6-phenyl analogue 23.
As with the isocytosines, Suzuki coupling of 22a,b afforded
the biphenyl analogues 24-27. Separation of 27 into its
enantiomers 28 and 29 was achieved by chiral SFC.
Assays. A key component of our fragment-based lead
generation strategy was the utilization of X-ray crystal structures
of fragments bound to BACE-1 to evolve the fragments to more
potent compounds. Crystal structures of low affinity fragments
are an invaluable aid for progressing them by allowing the
determination of binding orientation, the degree of subsite
occupancy, and polar interactions. However, these structures
cannot quantitatively assess the functional relevance of any

Cyclic Amidine â-Secretase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 5915
Scheme 3. Synthesis of Dihydroisocytosines 23-29a
a Reagents and conditions: (a) (CH3O)2P(O)CH2CO2tBu, n-BuLi, THF; (b) 50% TFA/DCM; (c) oxalyl chloride, DMF (cat.), DCM; (d) N-methylcyanamide
prepared from BrCN, NH2CH3, Na2CO3, THF; (e) 4-MeOBnNH2, DMF; (f) (NH4)2Ce(NO3)6, acetonitrile, H2O; (g) H2, Pd(OH)2, TEA, MeOH; (h) RB(OH)2,
Pd(PPh3)2Cl2, K3PO4, DME, H2O, EtOH, for 25 no RB(OH)2; (i) chiral HPLC resolution.
Table 1. Activity Data for Isocytosines and Dihydroisocytosinesa
SPR % Inhib
1 mM
SPR % Inhib
500 µM
SPR
IC50 (µM)
FRET
IC50 (µM)
IGEN
IC50 (µM)
cell assay
IC50 (µM)
LE based
on FRET
compd
5
6
28
62
7
42
8a
69
8b
8c
NV
89
86
130
0.28
8d
8e
NT
NT
8f
0
10a
10b
10c
10d
11a
11b
11c
13a
13b
13c
13d
16
23
24
25
26
27
28
29
72
180
28
16
220
36
77
150
27
46
0.29
0.29
0.28
0.27
69
94
29
30
24
51
5.7
190
120
5.3
90
99
5.9
0.25
0.29
91
20
69
140
480
79
190
0.34
42
93
82
0.33
0.39
0.49
0.36
0.34
0.33
0.35
0.23
0.37
60
7.5
6.1
2.0
0.38
34
1.6
0.20
35
3.5
3.8
0.59
3.8
0.28
0.67
1.7
0.64
0.08
0.47
a NV ) no value obtained; NT ) not tested.
The lack of success employing these traditional approaches
led us to investigate fragment-based lead generation. As
described in the accompanying article, we implemented 1D
NMR screening of fragment libraries.40 While a number of
fragments were discovered that competitively inhibited BACE-
1, the lead structure that emerged from this effort was the
6-propylisocytosine 5. In the surface plasmon resonance (SPR)
assay, compound 5 had an inhibition of 28% at 1 mM (Table
1). Screening of related structures from our corporate compound
collection identified three additional isocytosines, 6-8a. Com-

5916 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Edwards et al.
Table 2. Crystallographic Data Collection and Refinement Statistics
8c
24
27
Data Collection
X-ray source
resolution (Å)
SRS 14.1, λ ) 0.977 Å
2.75
ESRF ID14.2, λ ) 0.933 Å
2.5
ESRF ID14.3, λ ) 0.931 Å
2.2
no. unique reflections
completeness (%)a
average multiplicity
13125
98.8(99.9)
2.4
17850
99.9(100)
2.5
25942
99.9(100)
2.5
a,b
Rmerge
12.1(46)
6.9(39.6)
6.6(41.8)
Refinement
Rcryst
Rfree
22.8
31.9
0.012
1.4
45.7
54.7
30.0
22.4
28.9
0.013
1.5
47.5
45.1
37.7
22.3
28.0
0.015
1.5
41.5
26.4
40.1
rmsd bond lengths (Å)
rmsd bond angles (deg)
average B-factor protein (Å2)
average B-factor ligand (Å2)
average B-factor solvent (Å2)
b
a Numbers in parentheses indicate the highest shell values. Rmerge ) ΣhΣi|I(h,i)-<I> (h)|ΣhSi<I>(h); I(h,i) is the scaled intensity of the ith observation
of reflection h, and <I> (h) is the mean value. Summation is over all measurements. Rcryst ) Σhkl,work||Fobs| -k|Fcalc||/Σhkl|FobsI, where Fobs and Fcalc are the
observed and calculated structure factors, k is a weighting factor and work denotes the working set of 95% of the reflections used in the refinement. Rfree
) Σhkl,test||Fobs| - k|Fcalc||/Σhkl|Fobs|, where Fobs and Fcalc are the observed and calculated structure factors, k is a weighting factor, and test denotes the test
set of 5% of the reflections used in cross validation of the refinement. λ refers to wavelength, Rmsd to root-mean-square deviations.
Our initial efforts with the isocytosine series focused on
improving potency by optimizing the 6-subsituent for binding
in the S1/S3 subsites, which form a contiguous, relatively
hydrophobic binding pocket. Lengthening or shorting of the
6-position linker reduced activity relative to the two-carbon ethyl
group 8a (7, 8f). Consequently, the rest of our studies in the
isocytosine series used the ethyl linker. The napthyl analogue
8b was prepared to more fully fill the S1/S3 subpockets.
However, we were not able to generate an IC50 value because
of the compound’s poor solubility. To increase solubility, and
potentially exploit a hydrogen bond interaction with the enzyme,
the indole derivative 8c was prepared, which had an SPR IC50
of 86 µM, and 130 µM in the FRET assay. The crystal structure
of BACE-1 complexed with 8c indicated that a hydrogen bond
was formed between the indole NH and the carbonyl oxygen
atom of Gly-230 (Figure 3). The N1 nitrogen atom and exocyclic
NH2 group of the isocytosine core forms hydrogen bonds with
the catalytic aspartates as anticipated from the structures of the
aminopyridines in BACE-138,39 and of 8a in endothiapepsin.40
Figure 2. (a) Interactions of isocytosine with endothiapepsin. (b) Novel
An additional observation from the structure of 8c complexed
with BACE-1 was that the N3 NH group does not form any
cyclic amidine-based aspartyl protease pharmacophore.
pound 8a was the most potent, with an inhibition at 1 mM of
69% in the SPR assay. We were able to determine the binding
mode of 8a from the X-ray structure of its complex with
endothiapepsin.40 We used endothiapepsin as a surrogate for
BACE-1 since at the time we initiated these studies we did not
have access to suitable crystals of BACE-1. The interactions
of the isocytosine core with the catalytic aspartates are illustrated
in Figure 2a.
Ultimately, a collaboration was established with Astex
Therapeutics who had successfully crystallized apo-BACE-1 in
a crystal form that allowed for fragment soaking. Astex had
independently identified a class of 2-aminopyridine BACE-1
inhibitors via their high throughput crystallography technology.38
While the aminopyridines and isocytosines are distinct scaffolds
with differing physicochemical properties, they shared two
important enzyme recognition features: a cyclic amidine motif
that interacts with the catalytic aspartates, and a hydrophobic
substituent in the 6-position that binds in the S1 subsite of
BACE-1. It was quickly recognized that these binding elements
represented a novel, nonpeptidic aspartyl protease pharmacoph-
ore, depicted in Figure 2b, and we therefore focused our efforts
on exploiting this novel finding. Optimization of the aminopy-
ridine series has been previously described.39
interactions with the enzyme and is pointed toward the S′
2
subsite. This finding led us to investigate substitutions on the
N3 nitrogen atom. We found that a simple methyl substituent
modestly increased potency. Thus, the N-methyl phenethyl
analogue 10a had a FRET IC50 of 220 µM while the IC50 was
77 µM for the N3 methylated indolyl analogue 10c (vs 130 µM
for 8c). N-Methylation also allowed determination of an IC50
for the 6-napthethyl-substituted isocytosine 10b. With an SPR
IC50 of 28 µM, and 36 µM in the FRET assay, 10b was the
most potent isocytosine identified at this point.
Examination of the BACE-1 crystal structures of 8c (Figure
3) and the aminopyridines39 suggested that the bicyclic substit-
uents were not fully occupying the S1/S3 subsite. This led us to
incorporate biphenyl-based substituents (10d, 11a-c, Table 1).
Because of intrinsic fluorescence, the IC50 of 11a could not be
determined in the FRET assay. Compound 11c, with IC50 values
of 5.7, 5.9, 5.3 µM (SPR, FRET, IGEN assays, respectively),
was the most potent isocytosine identified.
As part of our exploration of the isocytosine series, we
screened related 2-amino heterocycles from our corporate
compound collection. The most promising hit we identified was
the dihydroisocytosine 13a, which had an SPR inhibition of 20%

Cyclic Amidine â-Secretase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 5917
Figure 3. Structure of 8c complexed with BACE-1. All pictures in this section were generated using PyMOL Molecular Graphics System: DeLano,
W.L. The PyMOL Molecular Graphics System; DeLano Scientific: Palo Alto, CA.
Figure 4. Energy difference between the pseudoequatorial conforma-
tion of 13a (A) and the bound pseudoaxial conformation (B), and the
energy difference between the conformations of the 6-methyl analogue
13b. Initial conformations were generated in an MMFF conformational
search. AM1 geometry optimization of the lowest energy pseudoaxial
and pseudoequatorial conformations preceded HF/6-316* energy
calculations. Aqueous solvation energies are from the SM5.4 model.
Figure 5. Structure of 24 complexed with BACE-1.
dihydroisocytosine 16 and the 3-methyl-6-phenyl dihydroiso-
cytosine 23 both had single digit micromolar potency and
All calculations were performed in Spartan ‘02.
demonstrated a >10-fold increase in activity relative to their
at 500 µM (Table 1). Incorporation of a 6-dichlorophenyl group
increased potency, with 13c having an SPR IC50 of 480 µM.
While a crystal structure of 13a was obtained, the resolution of
the structure was low. However, the resolution was sufficient
to reveal that the bound conformation of 13a had the 6-phenyl
group oriented in the less favorable pseudoaxial conformation
(Figure 4). Calculations indicated that the psuedo-axial confor-
mation was 1.4 kcal higher in energy than the pseudoequatorial
conformation in water (1.7 kcal/mol in gas phase). This led us
to incorporate a 6-methyl substituent in an attempt to increase
the relative stability of the pseudoaxial conformation, thus
reducing the energetic penalty for adoption of the higher energy
pseudoaxial bound conformation and thereby increase potency.
The calculated energy difference between the pseudoaxial and
pseudoequatorial conformations is less than 0.1 kcal/mol for
the 6-methyl analogue in water and slightly favored in the gas
phase (-0.53 kcal/mol). We were also concerned about possible
metabolic oxidation of the 6-H dihydroisocytosine to the
corresponding isocytosine, a process that would be blocked by
the 6-methyl substituent. Incorporation of a 6-methyl group into
13a afforded 13b, which had IC50 values of 140 µM (SPR) and
190 µM (FRET), while the 6-methyl dichlorophenyl analogue
13d had an IC50 of 82 µM (FRET).
nonmethylated congeners. In addition, each demonstrated
functional cellular activity with IC50 values of approximately 4
µM (Table 1). Incorporation of the biphenyl methoxy motif
afforded 24, which with a FRET IC50 of 380 nM and an IC50
of 590 nM in the cell-based assay was the first submicromolar
BACE-1 inhibitor we identified. Thus, by leveraging the
knowledge gained from the aminopyridine and isocytosine
structure-activity studies, we were able to rapidly progress the
dihydroisocytosines through targeted synthesis of only seven
compounds to sub-micromolar potency. The ligand efficiency
of the dihydroisocytosines was significantly better than that of
the isocytosines, as was their potency and solubility. As a result,
the dihydroisocytosines emerged as our leading series.
Examination of the crystal structure of 24 revealed that the
oxygen atom of the 3′-methoxy substituent displaced a water
molecule present in the apo-BACE-1 structure and formed a
hydrogen bond with the hydroxyl group of Ser-229 (Figure 5).
In addition, it appeared that the phenyl group directly attached
to the dihydroisocytosine core did not optimally fill the S1 pocket
when compared to the ethyl-linked phenyl group in the isocyto-
sines such as 11 (Figure 3). It was not clear that use of an ethyl
linker would improve the potency of the dihydroisocytosines
because of the penalty from increasing rotational entropy that
this modification would introduce. To investigate the phenethyl
linker in the dihydroisocytosines, we prepared compounds 25-
As anticipated from the SAR of the isocytosines, an N3-methyl
group further increased potency. The 3-methyl-6-dichlorophenyl

5918 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Edwards et al.
Figure 6. (a) Structure of 27 complexed with BACE-1. (b) The overlap of 24 (yellow) with 27 (cyan).
LC/MS spectrometer. The heating of reactions with microwaves
was performed on a Personal Chemistry Smith Synthesizer unit
(monomodal, 2.45 GHz, 300 W max.) for the times indicated.
Analytical supercritical fluid chromatography (SFC) was performed
on a Berger Instruments SFC Analytix Platform and preparative
SFC chromatography performed on the Berger Multigram II
Preparative-SFC. Standard silica gel flash chromatography was
employed as a method for purification of intermediates with the
indicated solvent mixtures. Alternately, intermediates were also
purified on an Isco CombiFlash Sq 16x instrument using prepack-
aged disposable RediSep SiO2 stationary phase columns (4, 12, 40,
120 g sizes) with gradient elution at 5-125 mL/min of selected
bisolvent mixture, UV detection (190-760 nm range) or timed
collection, 0.1 mm flow cell path length.
All solvents and reagents were obtained from commercial sources
and used without further purification unless otherwise noted.
Compounds 5-7 were obtained from the AstraZeneca corporate
compound collection as was the initial screening sample for 8a.
Compound 8a was resynthesized for conformation of structure,
purity, and activity as described below. HPLC purification condi-
tions can be found in the Supporting Information.
2-Amino-6-phenethyl-3H-pyrimidin-4-one (8a). To a solution
of 9a (2.36 g, 10.7 mmol) in ethanol (36 mL) was added guanidine
carbonate (0.97 g, 5.4 mmol), and the reaction was heated under
reflux overnight. The reaction was allowed to cool, and the resulting
solid was collected by filtration and rinsed with ethanol (10 mL).
The solid was dried under reduced pressure to give the desired
product as a white solid (1.47 g, 64%). 1H NMR (300 MHz, DMSO-
d6) δ 2.53 (t, J ) 9.4 Hz, 2H), 2.85 (t, J ) 7.9 Hz, 2H), 5.39 (s,
1H), 6.46 (s, 2H), 7.30 (m, 5H), 10.57 (s, 1H); MS (ES+) m/z 216
[M + H]+; HRMS (TOF) m/z calcd for C12H13N3O [M + H]+,
216.1131; found, 216.1137.
27. While compound 25 was less potent than the corresponding
phenyl-linked analogue 23, the biphenyl methoxy analogue 27
was slightly more potent than 24. Compound 27 was also active
in the cell-based assay with an IC50 of 640 nM. The structure
of the complex of 27 with BACE-1 illustrates how well 27 fills
the S1/S3 subsites (Figure 6a). The overlap of the bound
structures of 27 and 24 reveals that the biphenyl group of
27 extends a little deeper into the S1/S3 subsite, and that the
dihydroisocytosine core is tilted toward the S′ subsites while
maintaining the key interactions with the catalytic aspartates
and Ser-229.
Chromatographic separation of the enantiomers of 27 afforded
the two pure enantiomers, 28 and 29. The absolute stereochem-
ical assignment was based on the activity of the compounds
and analogy with the stereochemistry of the bound conformation
in the crystal structure of 27. Thus, the inactive isomer, 28,
was assigned the S configuration, while the active isomer, 29,
was assigned the R configuration. With a FRET IC50 of 80 nM,
29 was the most potent compound we have identified to date,
was active in the cell-based assay, and possessed very high
ligand efficiency (0.37).
Conclusion
Using NMR screening, we identified a novel series of cyclic
amidines with millimolar potency. The application of several
different biophysical methodologies allowed evolution of these
weak fragment hits to micromolar inhibitors of BACE-1.
Screening of related structures resulted in the discovery of the
dihydroisocytosines, a series that afforded compounds with high
potency and cellular activity. Translation of SAR from the iso-
cytosines allowed us to rapidly and efficiently increase potency
in the dihydroisocytosine series. Ligand efficiency was used to
guide the optimization of inhibitor series and assess the intrinsic
binding avidity of core structures. The cyclic amidine motif re-
presents a novel aspartyl protease core and should find applica-
tion to the design and development of inhibitors of other aspartyl
proteases such as renin and HIV protease. Future publications
from our group will detail the evolution of this pharmacophore
to new core structures with greater intrinsic potency.
2-Amino-6-[2-(1H-indol-6-yl)ethyl]pyrimidin-4(3H)-one (8c).
To 9c (approximately 65% pure with over-reduced product as the
major contaminant) (20 g, 77 mmol) was added ethanol (160 mL)
under argon, and to this was added guanidine carbonate (9.0 g, 50
mmol). The reaction was heated under reflux overnight and then
concentrated until approximately 50 mL of ethanol remained. To
this was added water (50 mL), and the mixture was stirred for 3 h.
The resulting solid was collected by filtration, washed with water
(approximately 50 mL), and then dried under reduced pressure at
60 °C overnight to give the title compound as a yellow solid (7.9
g, 62%). Another batch of the title compound slowly crystallized
1
from the filtrate to afford a white solid (2.4 g, 19%). H NMR
Experimental Section
(300 MHz, DMSO-d6) δ 2.56 (t, J ) 7.9 Hz, 2H), 2.93 (t, J ) 7.9
Hz, 2H), 5.36 (s, 1H), 6.34 (d, J ) 2.9 Hz, 1H), 6.62 (s, 2H), 6.85
(d, J ) 8.0 Hz, 1H), 7.18 (s, 1H), 7.23 (d, J ) 3.1 Hz, 1H), 7.41
(d, J ) 8.1 Hz, 1H), 10.90 (s, 1H); MS (APCI+) m/z 255 [M +
H]+; HRMS (TOF) m/z calcd for C14H14N4O [M + H]+, 255.1240;
found, 255.1241.
Chemistry. Proton magnetic resonance (1H NMR) spectra were
recorded on a Bruker Avance DPX 300 MHz spectrometer, and
the chemical shifts are reported in parts-per-million (δ) from a
tetramethylsilane internal standard. High-resolution mass spectra
were recorded on an Agilent Technologies 6210 Time-of-Flight

Cyclic Amidine â-Secretase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 5919
2-Amino-6-[2-(3-bromophenyl)ethyl]pyrimidin-4(3H)-one (8e).
To a solution of 9e (9.72 g, 32 mmol) in ethanol (120 mL) was
added guanidine carbonate (2.9 g, 16 mmol), and the reaction was
heated under reflux overnight. The reaction was allowed to cool
and the resulting solid collected by filtration and washed with
ethanol (20 mL). The solid was dried under reduced pressure to
then the 3-(3-bromophenyl)propionic acid/CDI solution was added
dropwise to the mixture of MgCl2, potassium ethyl malonate, and
TEA. The reaction was stirred overnight and then heated at 90 °C
for 3 h. It was then allowed to cool to room temperature, filtered,
and rinsed with acetonitrile (3 × 100 mL). The combined filtrates
were concentrated under reduced pressure and then partitioned
between methylene chloride and water. The product was extracted
into the methylene chloride layer, washed with 10% aqueous citric
acid solution, dried over sodium sulfate, filtered, and concentrated
under reduced pressure to give the title compound which was used
without purification (9.72 g, 75%). MS (APCI+) m/z 299 [M +
H]+.
1
give the title compound as a white solid (6.8 g, 71%). H NMR
(300 MHz, DMSO-d6) δ 2.53 (t, J ) 8.1 Hz, 2H), 2.86 (t, J ) 7.8
Hz, 2H), 5.39 (s, 1H), 6.46 (s, 2H), 7.23 (m, 2H), 7.37 (m, 1H),
7.42 (s, 1H), 10.58 (s, 1H); MS (APCI+) m/z 294 [M + H]+.
Ethyl 5-(2-Naphthyl)-3-oxopentanoate (9b). Diisopropylamine
(8.7 mL, 62 mmol) was dissolved in THF (100 mL) under nitrogen
and cooled to 0 °C, and to this was added N-butyllithium (1.6 M,
3.8 mL, 65 mmol). The resulting solution was then added to ethyl
acetoacetate (3.8 mL, 30 mmol) and allowed to stir at 0 °C for
approximately 25 min. Next, a solution of 2-(bromomethyl)-
naphthalene (6.6 g, 30 mmol) in THF (90 mL) was added over
approximately 45 min. The reaction was allowed to stir for 3 h at
0 °C and then quenched with a mixture of concentrated HCl (5.2
mL), water (14 mL), and diethyl ether (40 mL). The mixture was
stirred for 20 min and then partitioned between diethyl ether (300
mL) and water (150 mL). The layers were separated, and the
aqueous phase was extracted with diethyl ether (150 mL). The
combined organics were washed approximately 10 times with water
(10 × 100 mL), dried over sodium sulfate, and concentrated under
reduced pressure. The resulting material was purified by silica gel
chromatography using hexanes/DCM (8:2, 1:1, 2:8) step gradient
to give the title compound as a light yellow liquid (2.83 g, 36%).
1H NMR (300 MHz, CDCl3) δ 1.24 (t, J ) 7.1 Hz, 3H), 2.97 (m,
2H), 3.08 (m, 2H), 3.43 (s, 2H), 4.17 (q, J ) 7.1 Hz, 2H), 7.31 (m,
1H), 7.44 (m, 2H), 7.62 (s, 1H), 7.78 (m, 3H); MS (APCI+) m/z
293 [M + Na]+.
2-Amino-3-methyl-6-phenethyl-3H-pyrimidin-4-one (10a). To
a stirred solution of 8a (410 mg, 1.89 mmol) in DMF (36 mL)
were added potassium carbonate (260 mg, 1.89 mmol) and
iodomethane (0.12 mL, 1.89 mmol). The reaction was allowed to
stir for 3 days and then added to a large volume of water. The
material was extracted into diethyl ether (2 × 60 mL), dried over
sodium sulfate, filtered, and concentrated under reduced pressure.
Because of residual DMF, additional water was added with the
resulting white solid collected by filtration and dried overnight under
reduced pressure to give the title compound as a white solid (82
mg, 19%). 1H NMR (300 MHz, DMSO-d6) δ 2.53 (t, J ) 9.9 Hz,
8H), 2.86 (t, J ) 8.0 Hz, 5H), 3.21 (s, 6H), 5.50 (s, 2H), 7.05 (s,
3H), 7.22 (m, 9H); MS (APCI+) m/z 230 [M + H]+; HRMS (TOF)
m/z calcd for C13H15N3O [M + H]+, 230.1288; found, 230.1292.
2-Amino-6-[2-(3-bromophenyl)ethyl]-3-methylpyrimidin-4(3H)-
one (10e). To a stirred solution of 8e (6.7 g, 23 mmol) in DMF
(410 mL) were added potassium carbonate (2.8 g, 20 mmol) and
iodomethane (1.3 mL, 20 mmol). The reaction was allowed to stir
for 3 days, and then another portion of potassium carbonate (0.94
g, 7 mmol) and iodomethane (0.43 mL, 7 mmol) was added. The
reaction was again allowed to stir overnight and then added to a
large volume of water (approximately 8 L). The material was
extracted into diethyl ether (6 × 200 mL), and the resulting solution
was concentrated under reduced pressure. A portion of the resulting
solid (3.0 g) was stirred in methylene chloride (260 mL) and then
filtered to give the desired product as a white solid (2.2 g, 92%).
1H NMR (300 MHz, DMSO-d6): δ 2.54 (t, 2H, J ) 8.5 Hz), 2.87
(t, 2H, J ) 7.7 Hz), 3.22 (s, 3H), 5.50 (s, 1H), 7.07 (s, 2H), 7.24
(m, 2H), 7.37 (m, 1H), 7.43 (s, 1H); MS (APCI+) m/z 308 [M +
H]+; HRMS (TOF) m/z calcd for C13H14BrN3O [M + H]+,
308.0398; found, 308.0348.
2-Amino-3-methyl-6-[2-(3-pyridin-3-ylphenyl)ethyl]-3H-pyri-
midin-4-one (11a). To 10e (TFA salt) (0.050 g, 0.12 mmol) was
added a solution of dimethoxyethane/water/ethanol (7:3:2 ratio, 2
mL) followed by pyridine-3-boronic acid (0.020 g, 0.16 mmol),
cesium carbonate (0.12 g, 0.36 mmol), and dichlorobis(triph-
enylphosohine)palladium(II) (5 mg, 0.006 mmol), and the mixture
was heated by microwave at 150 °C for 15 min. Insoluble materials
were removed by filtration, and the solution was purified by Method
E to afford the desired product as a white TFA salt (80 mg, 40%).
1H NMR (300 MHz, DMSO-d6) δ 2.80 (t, J ) 7.7 Hz, 2H), 2.99
(t, J ) 7.7 Hz, 2H), 3.26 (s, 3H), 5.86 (s, 1H), 7.36 (d, J ) 7.6 Hz,
1H), 7.47 (t, J ) 7.6 Hz, 1H), 7.73-7.59 (m, 3H), 8.30 (d, J ) 8.0
Hz, 1H), 8.72-8.63 (m, 3H), 8.99 (s, 1H); MS (APCI+) m/z 307
[M + H]+; HRMS (TOF) m/z calcd for C18H18N4O[M + H]+,
307.1559; found, 307.1598.
Ethyl 5-(1H-Indol-6-yl)-3-oxopentanoate (9c). 6-Formylindole
(15.0 g, 103 mmol) was dissolved in dry THF (410 mL) under
argon, to this was added [3-(ethoxycarbonyl)-2-oxopropyl]triph-
enylphosphonium chloride (66.0 g, 155 mmol), and the reaction
was cooled to 5 °C. Sodium hydride (60%, 6.50 g, 412 mmol) was
then added in portions over 10 min, the cooling bath removed, and
the reaction allowed to stir overnight. Another portion of sodium
hydride (60%, 4.10 g, 102 mmol) was added, the reaction allowed
to stir for 2 h, and then another portion of [3-(ethoxycarbonyl)-2-
oxopropyl]triphenylphosphonium chloride (22 g, 51 mmol) added.
The reaction was again allowed to stir overnight and cooled to 5
°C, and to this were added saturated aqueous ammonium chloride
(200 mL) and water (100 mL). Ethyl acetate (100 mL) was added
and the product extracted into the organic phase. The organic phase
was dried over sodium sulfate, filtered, concentrated under reduced
pressure, and purified by silica gel chromatography (30% ethyl
acetate/hexanes) to give ethyl (4E)-5-(1H-indol-6-yl)-3-oxopent-
1
4-enoate48 as a white solid (20.6 g, 78%). H NMR (300 MHz,
DMSO-d6) δ 1.20 (t, J ) 7.1 Hz, 3H), 3.85 (s, 2H), 4.12 (q, J )
7.1 Hz, 2H), 6.49 (s, 1H), 6.83 (d, J ) 16.1 Hz, 1H), 7.39 (d, J )
8.3 Hz, 1H), 7.49 (m, 1H), 7.59 (d, J ) 8.3 Hz, 1H), 7.72 (s, 1H),
7.79 (d, J ) 16.1 Hz, 1H), 11.38 (s, 1H); MS (ES-) m/z 256 [M
- H]-.
Ethyl (4E)-5-(1H-indol-6-yl)-3-oxopent-4-enoate (20.6 g, 80
mmol) was dissolved in ethanol (160 mL) under argon, the solvent
was degassed with argon, and 10% Pd/C (4.25 g) was added. The
mixture was placed under 1 atm of hydrogen and stirred vigorously
for 2 h. The mixture was filtered through Celite, washed with
ethanol, and concentrated under reduced pressure to give (20 g, 77
mmol) of a white solid consisting of a 65:35 mixture of product
and over-reduced material which was used without further purifica-
tion. MS (ES-) m/z 258 [M - H]-.
Ethyl 5-(3-Bromophenyl)-3-oxopentanoate (9e). To a round-
bottom flask was added magnesium chloride (10.4 g, 109 mmol),
acetonitrile (580 mL), potassium malonate (15.6 g, 92.0 mmol),
and TEA (19.5 mL, 140 mmol). Separately, 3-(3-bromophenyl)-
propionic acid (10 g, 44 mmol) was dissolved in acetonitrile (200
mL), and to this was added 1,1′-carbonyldiimidazole (CDI) (7.8 g,
48 mmol). Both were allowed to stir for approximately 2.5 h, and
3-(3,4-Dichlorophenyl)but-2-enoic Acid Ethyl Ester (12d). To
a -78 °C stirred solution of triethyl phosphonoacetate (11.5 mL,
58.2 mmol) in THF (100 mL) was added n-BuLi in hexanes (1.6
N, 38 mL, 61 mmol), and the reaction was stirred at -78 °C for
10 min. To this mixture was added a solution of 3,4-dichloroac-
etophenone (10.0 g, 52.9 mmol) in THF (10 mL), and the reaction
was allowed to warm to room temperature and stirred for 18 h.
The solvent was removed under reduced pressure to yield a yellow
solid. To this was added 400 mL 1:3 Et2O: hexanes, and the solids
were triturated for 1 h. The resulting precipitate was removed by
filtering through Celite, and the filtrate was collected and concen-
trated under reduced pressure to give the title compound as a crude
orange oil (12.14 g, 88%). This material was carried directly into

5920 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Edwards et al.
the next reaction without further purification. 1H NMR (300 MHz,
DMSO-d6) δ 1.24 (t, J ) 6.9 Hz, 3H), 2.48 (s, 3H), 6.22 (s, 1H),
7.50 (d, J ) 2.4 Hz, 1H), 7.84 (d, J ) 2.1 Hz, 2H).
yellow oil was purified on silica gel (0.5 kg) in a large filter funnel
eluting with 1 L each of a 5% step gradient 0-100% hexanes/
methylene chloride to afford the title compound as a clear oil (18.5
1
g, 58%). H NMR (300 MHz, DMSO-d6) δ 2.10 (s, 3H), 2.79 (s,
2-Amino-6-(3,4-dichlorophenyl)-6-methyl-5,6-dihydro-3H-py-
rimidin-4-one (13d). To a solution of 12d (100 mg, 0.386 mmol)
in 2.0 mL of NMP were added guanidine hydrochloride (147 mg,
1.54 mmol) and sodium methoxide (62.0 mg, 1.62 mmol), and the
reaction was subjected to microwaves at 200 °C for 10 min. The
solids were filtered from the reaction, and the filtrate was used
directly for purification using Method C. The combined purified
fractions were lyophilized to give the title compound as a white
4H), 7.22-7.24 (m, 2H), 7.35-7.39 (m, 1H), 7.43 (s, 1H).
3-(3-Bromophenyl)but-2-enoic Acid tert-Butyl Ester (18a). To
a -78 °C stirred solution of tert-butyldimethyl phosphonoacetate
(21.9 mL, 0.111 mol) in THF (150 mL) was added n-BuLi in
hexanes (1.6 M, 72 mL, 0.12 mol), and the reaction was stirred at
-78 °C for 10 min. To this mixture were added 17a and
3′-bromoacetophenone (Aldrich) (13.4 mL, 0.100 mol), and the
reaction was allowed to warm to room temperature and was stirred
for 18 h. The THF was removed under reduced pressure to yield a
yellow solid. To this was added hexanes (300 mL), and the solids
were triturated for 1 h. The mixture was filtered through Celite
and the filtrate concentrated under reduced pressure to give the
title compound as a yellow oil (28.9 g, 97%). This was carried
1
TFA salt (49 mg, 33%). H NMR (300 MHz, DMSO-d6-/TFA-d)
δ 1.65 (s, 3H), 3.17 (m, 1H), 3.36 (d, J ) 16.5 Hz, 1H), 7.42 (d,
J ) 8.4 Hz, 1H), 7.70 (m, 2H); MS (+ES) m/z 272 [M + H]+;
HRMS (TOF) m/z calcd for C11H11Cl2N3O [M + H]+, 272.0352;
found, 272.0354.
N′-[4-(3,4-Dichlorophenyl)-4-methyl-6-oxo-1,4,5,6-tetrahydro-
pyrimidin-2-yl]-N,N-dimethylformamidine (14). To a stirred
solution of 13d (0.25 g, 0.94 mmol) in DMF (5 mL) was added
dimethylformamide dimethylacetal (0.160 mL, 1.17 mmol), and the
reaction was stirred for 2 h. The DMF was removed under reduced
pressure to yield a pale yellow oil. The oil was purified by ether
trituration (2 × 20 mL) to give the title compound as a colorless
oil (0.30 g, 99%). 1H NMR (300 MHz, DMSO-d6) δ 1.67 (s, 3H),
3.11-3.17 (br s, 4H), 3.20-3.29 (br s, 4H), 7.47 (d, J ) 8.4 Hz,
1H), 7.69 (d, J ) 8.4 Hz, 1H), 7.76 (s, 1H), 8.56 (s, 1H); MS (ES
+) m/z 327 [M + H]+.
1
directly into the next reaction. H NMR (300 MHz, DMSO-d6) δ
1.20 (s, 3H), 1.47 (s, 6H), 2.10 (d, J ) 1.4 Hz, 0.7H), 2.44 (d, J )
1.3 Hz, 1.3H), 5.87 (d, J ) 1.5 Hz, 0.3H), 6.05 (d, J ) 1.3 Hz,
0.7H), 7.19-7.38 (m, 2H), 7.49-7.59 (m, 2H), 7.71 (t, J ) 1.8
Hz, 1H).
5-(3-Bromophenyl)-3-methylpent-2-enoic Acid tert-Butyl Es-
ter (18b). To a -78 °C stirred solution of tert-butyldimethyl
phosphonoacetate (19.5 mL, 89.61 mmol) in tetrahydrofuran (100
mL) was added n-butyllithium in hexanes (1.6 M, 35.8 mL, 89.61
mmol), and the reaction was stirred at -78 °C for 25 min. To this
mixture was added 17b (18.5 g, 81.5 mmol), the reaction was
allowed to warm to room temperature and stirred for 1.5 h, and
the solvent removed under reduced pressure. The resulting solid
was triturated with hexanes (300 mL) for 1 h and filtered through
Celite and the filtrate concentrated under reduced pressure to give
a single isomer of the title compound as a clear oil that was used
in the next reaction without further purification. 1H NMR (300 MHz,
DMSO-d6) δ 1.42 (d, J ) 3.2 Hz, 9H), 1.86 (d, J ) 1.3 Hz, 1H),
2.12 (d, J ) 1.1 Hz, 2H), 2.37-2.43 (m, 2H), 2.73-2.78 (m, 2H),
5.58-5.60 (m, 1H), 7.24 (d, J ) 5.2 Hz, 2H), 7.35-7.40 (m, 1H),
7.45 (d, J ) 5.4 Hz, 1H).
N′-[4-(3,4-Dichlorophenyl)-1,4-dimethyl-6-oxo-1,4,5,6-tetrahy-
dropyrimidin-2-yl]-N,N-dimethylformamidine (15). To a stirred
solution of 14 (0.31 g, 0.94 mmol) and potassium carbonate (0.140
g, 1.03 mmol) in DMF (70 mL) was added iodomethane (0.060
mL, 1.0 mmol), and the reaction was stirred for 18 h. To this
mixture were added additional potassium carbonate (0.140 g, 1.03
mmol) and iodomethane (0.060 mL, 1.0 mmol), and the mixture
was stirred an additional 18 h. The THF was removed under reduced
pressure to yield a cloudy yellow oil. To this was added hexanes
(250 mL), and the reaction was stirred for 10 min. The resulting
precipitate was filtered and the filtrate concentrated under reduced
pressure. The crude compound was purified using flash chroma-
tography (silica gel, 5:95 ethyl acetate: hexanes) to give the title
3-(3-Bromophenyl)but-2-enoyl Chloride (19a). A solution of
18a (28.9 g, 0.097 mol) in trifluroracetic acid:methylene chloride
(1:1, 300 mL) was stirred at room temperature for 15 min, and the
solvents were removed under reduced pressure. The crude yellow
solid was triturated in hexanes (400 mL), filtered, and dried under
reduced pressure to afford a single isomer of 3-(3-bromophenyl)-
1
compound as a colorless oil (160 mg, 50%). H NMR (300 MHz,
DMSO-d6) δ 1.67 (s, 3H), 3.11 (s, 3H), 3.19 (s, 3H), 3.22 (d, J )
16.5 Hz, 1H), 3.36 (s, 3H), 3.57 (d, J ) 16.5 Hz, 1H), 7.49 (d, J
) 8.3 Hz, 1H), 7.65 (d, J ) 8.4 Hz, 1H), 7.82 (s, 1H), 8.65 (s,
1H); MS (ES +) m/z 341 [M + H]+.
1
but-2-enoic acid as a white solid (8.87 g, 38%). H NMR (300
MHz, DMSO-d6) δ 2.46 (s, 3H); 6.11 (s, 1H); 7.37 (t, J ) 7.8 Hz,
1H); 7.53 (m, 2H); 7.72 (t, J ) 1.5 Hz, 1H). The filtrate was
removed of solvent under reduced pressure, resulting in a second
impure batch of mixed isomers, which was dried under high vacuum
over night (15 g). The dry material was triturated with hexanes
(500 mL, 0.12 M) at room temperature for 30 min. The precipitate
was filtered to give a very pale yellow solid consisting of mixed
2-Amino-6-(3,4-dichlorophenyl)-3,6-dimethyl-5,6-dihydro-3H-
pyrimidin-4-one (16). To a solution of 15 (0.16 g, 0.47 mmol) in
MeOH (15 mL) was added 7 N methanolic ammonia (3.0 mL, 21
mmol), and the reaction was heated to 60 °C for 3 h. The MeOH
was removed under reduced pressure to yield an amber syrup. To
this was added acetonitrile:water:TFA (75:25:0.1, 4 mL), and the
resulting mixture was filtered. The filtrate was purified using
Method C. The combined purified fractions were lyophilized to
1
isomers (10.2 g, 44%). H NMR (300 MHz, DMSO-d6/TFA-d) δ
1
2.13 (s, 0.6H), 2.49 (s, 0.4), 5.94 (s, 0.2H), 6.14 (s, 0.8 H), 7.23-
7.34 (br m, 2H), 7.43-7.66 (br m, 1.2H), 7.72 (s, 0.8H). The filtrate
was concentrated, and the resulting waxy yellow solid was triturated
with hexanes (50 mL) to give a white solid (1.4 g, 6%). 1H NMR
(300 MHz, DMSO-d6/TFA-d) δ 2.13 (s, 2.4H), 2.49 (s, 0.6), 5.94
(s, 0.8H), 6.14 (s, 0.2H), 7.23-7.34 (br m, 2H), 7.43-7.66 (br m,
1.8H), 7.72 (s, 0.2H).
give the title compound as a white TFA salt (30 mg, 22%). H
NMR (300 MHz, DMSO-d6) δ 1.65 (s, 3H), 3.12 (s, 3H), 3.19 (d,
J ) 16.5 Hz, 1H), 3.50 (d, J ) 16.5 Hz, 1H), 7.41 (d, J ) 8.4 Hz,
1H), 7.67 (d, J ) 8.4 Hz, 1H), 7.72 (s, 1H), 8.80 (br s, 2H), 10.46
(br s, 1H); MS (ES+) m/z 286 [M + H]+; HRMS (TOF) m/z calcd
for C12H13Cl2N3O[M + H]+, 286.0508; found, 286.0515.
4-(3-Bromophenyl)butan-2-one (17b).49 To 2,4-pentanedione
(15.8 mL, 0.154 mol) in ethanol (150 mL) was added potassium
carbonate (19.4 g, 0.140 mol) followed by 3-bromobenzyl bromide
(35.0 g, 0.140 mol). The reaction was refluxed for 18 h and was
cooled to room temperature, and the salts were removed by
filtration. The filtrate was concentrated under reduced pressure and
the resulting yellow solid partitioned between ethyl acetate and
water (200 mL/100 mL). The ethyl acetate layer was washed twice
with aqueous HCl (1 N, 100 mL), once with saturated aqueous
sodium bicarbonate (100 mL), and once with saturated aqueous
sodium chloride (100 mL), dried over sodium sulfate, and filtered,
and the filtrate concentrated under reduced pressure. The resulting
To a suspension of 3-(3-bromophenyl)but-2-enoic acid (mixed
isomers) (16.2 g, 67.3 mmol) in 250 mL of methylene chloride
was added oxalyl chloride (7.10 mL, 80.8 mmol) followed by DMF
(260 µL, 3.37 mmol), and the reaction was stirred at room
temperature. After 2 h the solvent was removed under reduced
pressure to give the title compound as a light pink solid (17.0 g,
97%). 1H NMR (300 MHz, CDCl3) δ 2.22 (s, 2.4H); 2.51 (s, 0.6H);
6.25 (s, 0.2H); 6.44 (s, 0.8H); 7.32 (m, 1H); 7.42 (d, J ) 6.9 Hz,
1H); 7.56 (d, J ) 6.9 Hz, 1H); 7.58 (s, 1H).
5-(3-Bromophenyl)-3-methylpent-2-enoyl Chloride (19b). To
18b was added a solution of trifluroracetic acid:dichloromethane

Cyclic Amidine â-Secretase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 5921
(1:1, 100 mL), and the reaction was stirred at room temperature
for 15 min. The solvents were removed under reduced pressure,
and the resulting solid was triturated with hexanes (100 mL). The
mixture was filtered and the solid dried under reduced pressure to
give 5-(3-bromophenyl)-3-methylpent-2-enoic acid as a white solid
(18.48 g, 84%) yield over two steps. 1H NMR (300 MHz, DMSO-
d6) δ 1.88 (d, J ) 1.3 Hz, 1H), 2.14 (d, J ) 1.1 Hz, 2H), 2.43 (t,
J ) 8.1 Hz, 2H), 2.71-2.80 (m, 2H), 5.62 (d, J ) 1.1 Hz, .5H),
5.66 (s, 5H), 7.25 (d, J ) 4.5 Hz, 2H), 7.35-7.42 (m, 1H), 7.47
(s, 1H).
added 4-methoxybenzylamine (12.8 mL, 97.7 mmol). After 4 h,
the solvent was removed under reduced pressure, and the crude
compound was purified by flash chromatography on a 6′′ × 8′′
silica gel column eluting with 5% MeOH/DCM to give the title
1
compound as an amber gum (20.4 g, 94%). H NMR (300 MHz,
DMSO-d6/TFA-d) δ 1.34 (s, 3H), 1.93 (m, 2H), 2.55-2.64 (m,
2H), 3.05 (dd, J ) 24.0, 16.8 Hz, 2H), 3.27 (s, 3H), 3.80 (s, 3H),
4.87 (dd, J ) 36.4, 18.5 Hz, 2H), 6.97 (d, J ) 8.6 Hz, 1H), 7.14-
7.34 (m, 6H), 7.38 (d, J ) 1.6 Hz, 1H); MS (APCI+) m/z 444 [M
+ 1]+.
To a 0 °C stirred solution of 5-(3-bromophenyl)-3-methylpent-
2-enoic acid (18.5 g, 68.7 mmol) in dichloromethane (100 mL)
was added oxalyl chloride (7.2 mL, 82.10 mmol) followed by N,N-
dimethylformamide (0.266 mL, 3.43 mmol), and the reaction was
stirred at 0 °C for 1 h and allowed to warm to room temperature.
After 1 h the solvent was removed under reduced pressure to give
the title compound as a cloudy oil, which was immediately carried
forward into the next reaction (21.78 g, 110%).
3-(3-bromophenyl)-N-cyano-N-methylbut-2-enamide (20a).
To a solution of cyanogen bromide (4.24 g, 40.0 mmol) (TOXIC!)
in 100 mL of THF at -60 °C was added sodium carbonate (6.36
g, 60.0 mmol) followed by dropwise addition of a solution of 2.0
M methylamine solution in THF (20 mL 40 mmol). The bath
temperature was kept below -20 °C for 2 h. The reaction was
filtered cold through a pad of Celite under nitrogen. A solution of
19a (single isomer) (5.19 g, 20.0 mmol) in 100 mL of THF was
added to the filtrate. To this mixture was added N,N-diisopropyl-
ethylamine (4.2 mL, 24 mmol), the reaction was stirred at room
temperature for 2 h, and the solvents were removed under reduced
pressure. The crude compound was purified using flash chroma-
tography on silica gel eluting with DCM to give the title compound
as an off-white solid (4.29 g, 75%). 1H NMR (300 MHz, DMSO-
d6) δ 2.44 (s, 3H); 3.22 (s, 3H); 6.65 (s, 1H); 7.42 (t, J ) 7.8 Hz,
1H); 7.58 (d, J ) 8.4 Hz, 1H); 7.65 (d, J ) 7.8 Hz, 1H); 7.76 (t,
J ) 1.8 Hz, 1H).
2-Amino-6-(3-bromophenyl)-3,6-dimethyl-5,6-dihydro-3H-py-
rimidin-4-one (22a). To a solution of 21a (15.5 g, 37.1 mmol) in
150 mL of acetonitrile was added 50 mL of water followed by
ceric ammonium nitrate (61.1 g, 111 mmol), and the reaction stirred
for 18 h. Celite (32 g) was added followed by sodium bicarbonate
(31.2 g, 371 mmol), and reaction was stirred for 1 h. Additional
Celite (15 g) was added and the mixture stirred for an additional 1
h. The reaction was filtered through Celite and the filtrate
concentrated under reduced pressure. A crude purification was
performed using silica gel eluting with 15:85:0.1 MeOH:DCM:
acetic acid. The resulting orange solid was triturated with methanol
to give the first batch of the title compound. The solvents were
removed from the filtrate under reduced pressure, and the resulting
orange solid was triturated with ethanol to give a second batch of
the title compound. The batches were combined to give the title
compound as an off-white solid (8.75 g, 79%). 1H NMR (300 MHz,
DMSO-d6/TFA-d) δ 1.64 (s, 3H), 3.14 (s, 3H), 3.19 (d, J ) 16.5
Hz, 1H), 3.49 (d, J ) 16.2 Hz, 1H), 7.39 (m, 2H), 7.55 (m, 1H),
7.67 (s, 1H): MS (APCI+) m/z 296 [M + H]+;
6-(3-Bromophenethyl)-2-amino-5,6-dihydro-3,6-dimethylpy-
rimidin-4(3H)-one (22b). To a solution of 21b (20.3 g, 45.6 mmol)
in acetonitrile (132 mL) was added water (33 mL) followed by
ceric ammonium nitrate (75.0 g, 137 mmol), and the reaction was
stirred for 18 h. Celite (40 g) was added followed by sodium
bicarbonate (38.0 g, 452 mmol), and the reaction was stirred for 1
h. Additional Celite (20 g) was added, the reaction filtered through
Celite washing with acetonitrile, and the filtrate concentrated under
reduced pressure. Ethanol was added to the residue, the solids were
removed by filtration, and solvent was removed under reduced
pressure. The resulting solids were triturated with diethyl ether,
and the solid material was dried under reduced pressure. This
material was partitioned between ethyl acetate (500 mL) and a
saturated sodium chloride/water solution (1:1, 500 mL). The organic
layer was washed with saturated sodium chloride and dried over
sodium sulfate, and the solvents were removed under reduced
pressure. The crude material was purified by flash chromatography
on a 6 in. × 8 in. silica gel column eluting with dichloromethane:
methanol:acetic acid (90:10:0.1 to afford the title compound as a
tan solid (13.3 g, 90%). 1H NMR (300 MHz, DMSO-d6/TFA-d) δ
1.33 (s, 3H), 1.86 (m, 2H), 2.64 (t, J ) 8.5 Hz, 2H), 2.79 (d, J )
16.4 Hz, 1H), 2.94 (d, J ) 16.4 Hz, 1H), 3.20 (s, 3H), 7.26 (d, J
) 5.0 Hz, 2H), 7.40 (td, J ) 4.5, 1.9 Hz, 1H), 7.49 (s, 1H); MS
(APCI+) m/z 324 [M + 1]+.
2-Amino-3,6-dimethyl-6-phenyl-5,6-dihydro-3H-pyrimidin-4-
one (23). To a solution of 22a (100 mg, 0.338 mmol) and
triethylamine (103 mg, 1.01 mmol) in MeOH (34 mL) was added
palladium hydroxide (40 mg). The mixture was charged with
hydrogen (50 psi) and shaken on a Parr shaker for 2 h. The reaction
was filtered through Celite and the filtrate concentrated under
reduced pressure. The crude compound was purified using Method
C to afford the title compound as a white TFA salt (8 mg, 10%).
1H NMR (300 MHz, DMSO-d6/TFA-d) δ 1.63 (s, 3H), 3.08 (s,
3H), 3.18 (d, J ) 16.5 Hz, 1H), 3.42 (d, J ) 16.5 Hz, 1H), 7.42
(m, 5H); MS (APCI+) m/z 218 [M + H]+; HRMS (TOF) m/z calcd
for C12H15N3O [M + H]+, 218.1288; found, 218.1293.
2-Amino-6-(3′-methoxybiphenyl-3-yl)-3,6-dimethyl-5,6-dihy-
dro-3H-pyrimidin-4-one (24). To a solution of 22a (47.0 mg, 0.132
mmol) in 1.5 mL of 7:3:2 1,2-dimethoxyethane:water:ethanol were
added cesium carbonate (129 mg, 0.396 mmol), 3-methoxyphe-
nylboronic acid (26.0 mg, 0.172 mmol), and dichlorobis(triph-
enylphosphine)palladium(II) (4.6 mg, 0.0065 mmol). The reaction
5-(3-Bromophenyl)-N-cyano-N,3-dimethylpent-2-enamide (20b).
To a -78 °C stirred solution of cyanogen bromide (14.6 g, 0.137
mmol) (TOXIC!) in tetrahydrofuran (100 mL) was added sodium
carbonate (21.8 g, 0.206 mmol) followed by dropwise addition of
a solution of methylamine in THF (2.00 M, 68.7 mL, 0.137 mol).
The mixture was stirred at -78 °C for 1.5 h. The reaction was
filtered cold through Celite under a blanket of nitrogen and a
solution of 19b (19.75 g (theoretical), 0.0686 mmol theoretical) in
THF (100 mL) was added to the filtrate. To this mixture was added
N,N-diisopropylethylamine (14.4 mL, 82.0 mol), reaction stirred
at room temperature for 2 h, and the solvents were removed under
reduced pressure. The crude compound was purified on silica gel
(0.5 kg) in a large filter funnel eluting with dichloromethane to
1
give the title compound as a tan solid (16.6 g, 79%). H NMR
(300 MHz, DMSO-d6) δ 1.99 (d, J ) 1.3 Hz, 1H), 2.14 (d, J ) 1.1
Hz, 2H), 2.52 (dd, J ) 3.8, 2.0 Hz, 2H), 2.79 (t, J ) 7.7 Hz, 2H),
3.13 (s, 3H), 6.22 (d, J ) 6.6 Hz, 1H), 7.24-7.27 (m, 2H), 7.37-
7.41 (m, 1H), 7.49 (s, 1H); MS (APCI+) m/z 307 [M + 1]+.
6-(3-Bromophenyl)-1-(4-methoxybenzylamino)-3,6-dimethyl-
5,6-dihydro-3H-pyrimidin-4-one (21a) To a solution of 20a
prepared from multiple batches with both isomers (12.8 g, 45.7
mmol) in 50 mL of DMF was added 4-methoxybenzylamine (14.9
mL, 114 mmol). After 4 h the solvent was removed under reduced
pressure. The crude compound was purified by flash chromatog-
raphy on silica gel eluting with DCM, 2.5% MeOH/DCM, 5%
MeOH/DCM followed by a second flash chromatography on silica
gel eluting with Et2O, EtOAc, 5% MeOH/EtOAc, 10% MeOH/
EtOAc to afford the title compound as an off white-solid (15.4 g,
81%). 1H NMR (300 MHz, DMSO-d6/TFA-d) δ 1.65 (s, 3H), 3.20
(s, 3H), 3.30 (d, J ) 16.5 Hz, 1H), 3.58 (d, J ) 16.8 Hz, 1H), 3.78
(s, 3H), 4.97 (dd, J ) 4.8 Hz, 2H), 6.96 (d, J ) 8.7 Hz, 2H), 7.34
(m, 4H), 7.57 (m, 2H); MS (APCI+) m/z 416 [M + H]+; LC
Method A: tR ) 1.80 min.
6-(3-Bromophenethyl)-1-(4-methoxybenzyl)-tetrahydro-2-imino-
3,6-dimethylpyrimidin-4(1H)-one (21b). To a stirred solution of
20b (15.0 g, 48.8 mmol) in N,N-dimethylformamide (50 mL) was

5922 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Edwards et al.
was heated by microwave for 15 min at 150 °C after which the
solvents were removed under a stream of nitrogen. To the resulting
brown gum was added ACN:water:TFA (75:25:0.1, 2.0 mL), and
the precipitate was removed. The filtrate was purified using Method
C to afford the title compound as a white TFA salt (25 mg, 43%).
1H NMR (300 MHz, DMSO-d6/TFA-d) δ 1.71 (s, 3H), 3.13 (s,
3H), 3.21 (d, J ) 16.5 Hz, 1H), 3.59 (d, J ) 16.2 Hz, 1H), 3.85 (s,
3H), 6.98 (d, J ) 3.9 Hz, 1H), 7.23 (m, 2H), 7.41 (m, 2H), 7.51 (t,
J ) 7.8 Hz, 1H), 7.64 (d, J ) 7.5 Hz, 1H), 7.71 (s, 1H); MS
(APCI+) m/z 324 [M + H]+; HRMS (TOF) m/z calcd for
C19H21N3O2 [M + H]+, 324.1707; found, 324.1718.
2-Amino-6-[2-(3′-methoxybiphenyl-3-yl)ethyl]-3,6-dimethyl-
5,6-dihydro-3H-pyrimidin-4-one (27). This material was prepared
according to the method described for 24 using 22b and 3-meth-
oxyphenylboronic acid to afford the title compound as a white TFA
salt (65 mg, 45%). 1H NMR (300 MHz, DMSO-d6/TFA-d) δ 1.37
(s, 3H), 1.94 (m, 2H), 2.71 (m, 2H), 2.82 (d, J ) 16.4 Hz, 1H),
2.98 (d, J ) 16.4 Hz, 1H), 3.21 (s, 3H), 3.84 (s, 3H), 6.95 (dd, J
) 8.0, 2.2 Hz, 1H), 7.19 (d, J ) 2.2 Hz, 1H), 7.23 (t, J ) 6.9 Hz,
2H), 7.38 (t, J ) 7.9 Hz, 2H), 7.51 (m, 2H); MS (ES+) m/z 352
[M + H]+; HRMS (TOF) m/z calcd for C21H25N3O-2 [M + H]+,
352.2020; found, 352.2019.
Synchrotron Radiation Facility, Grenoble, France and Synchrotron
Radiation Source, Daresbury, U.K. All data sets were processed
with MOSFLM50 and scaled using SCALA.51 Structure solution
and initial refinement of the protein-ligand complexes were carried
out on our automated scripts using a modified apo-BACE-1
structure (PDB id code 1w50) as a starting model. Bound ligands
were automatically identified and fitted into Fo - Fc electron density
using AutoSolve52 and further refined using automated scripts
followed by rounds of rebuilding in AstexViewer253 and refinement
using Refmac.54 Data collection and refinement statistics for selected
crystal structures are presented in Table 2.
Biology. This section describes the various assays used to assess
the inhibition of BACE-1 by the compound reported in this paper.
In several of the assays up to 5% DMSO was used to ensure
complete solubility of the test compound. Appropriate control
experiments demonstrated that the levels of DMSO used had no
detrimental effects on the inhibition values generated.
Preparation of BACE-1. The BACE-1 used in most assays was
expressed in Escherichia coli and purified as previously described.40
The stock solution of BACE-1 was 1 mg/mL in 20 mM Tris, 150
mM NaCl, and 1 mM DTT at pH 8.2. The BACE-1 used for the
SPR assay of compounds 5-7 was prepared according to the
procedure of Lullau et al.55
(S)-2-Amino-6-[2-(3′-methoxybiphenyl-3-yl)ethyl]-3,6-dimethyl-
5,6-dihydro-3H-pyrimidin-4-one (28). To 27, 165 mg, was added
methanol, and the enantiomers were separated on a preparative
supercritical fluid chromatography system using Method F. The
solvent was removed from the product fractions using a Genevac
evaporator to give a waxy solid. This solid was then purified using
Method C. The combined purified fractions were lyophilized to
BACE-1 SPR Assay. Sensor Chip Preparation: BACE-1 was
assayed on a Biacore3000 instrument by attaching either a peptidic
transition state isostere (TSI) or a scrambled version of the peptidic
TSI to the surface of a BIAcore CM5 sensor chip. The peptides,
KFES-statine-ETIAEVENV and KTEEISEVN-statine-VAEF, were
purchased from the Synpep Corporation and BioSource, respec-
tively. The scrambled peptide KFES-statine-ETIAEVENV was
coupled to channel 1, and the TSI inhibitor KTEEISEVN-statine-
VAEF was coupled to channel 2 of the same chip using the
following protocol. The two peptides were dissolved at 0.2 mg/
mL in 20 mM sodium acetate pH 4.5, and then the solutions were
centrifuged at 14K rpm to remove any particulates. Carboxyl groups
on the dextran layer of the chip were activated by injecting a 1:1
mixture of 0.5 M N-ethyl-N′-(3-dimethylaminopropyl)carbodiimide
(EDC) and 0.5 M N-hydroxysuccinimide (NHS) at 5 µL/min for 7
min. The stock solution of the control peptide was injected in
channel 1 for 7 min at 5 µL/min, and the remaining carboxyl groups
were blocked by injecting 1 M ethanolamine for 7 min at 5 µL/
min. The TSI peptide was coupled to channel 2 in the same manner.
Assay Protocol: The running buffer for the BACE-1 SPR assay
consisted of 25 mM sodium acetate, 200 mM NaCl, 0.005% SP20,
and 5% DMSO at pH 4.5. The stock solution of BACE-1 was
diluted to 0.5 µM in running buffer minus DMSO. The diluted
BACE-1 was mixed with DMSO or compound dissolved in DMSO
to achieve a final DMSO concentration of 5%. The mixing was
done in a 96-well plate that was then sealed and maintained at 4
°C for 1 h. The mixture was then injected over channels 1 and 2 of
the CM5 BIAcore chip at a rate of 20 µL/min, and the signal was
recorded in response units (RU). Specific binding of BACE-1 to
the TSI inhibitor was monitored on channel 2, while any binding
to channel 1 was nonspecific and was subtracted from the channel
2 response. BACE-1 binding to the chip was quantified by meas-
uring the rate of the increase in response 20 s after injection. The
DMSO control was defined as 100% binding, and the effect of the
inhibitor was reported as percent inhibition of the DMSO control.
FRET Assay Protocol. The enzymatic activity of BACE-1 was
measured with an internally quenched fluorescent substrate, E(e-
dans)ISEVNLDAEFK(dabcyl) purchased from BioSource. Cleavage
of the substrate resulted in a fluorescent product that was measured
on a Wallac Victor II plate reader with an excitation wavelength
of 360 nm and an emission wavelength of 485 nm. A 4 mM stock
solution of substrate was prepared in 45 mM Tris, 45 mM borate,
and 1 mM EDTA at pH 8.2. BACE-1 was diluted 1:70 in 40 mM
MES pH 5.0, and substrate was diluted to 30 µM in the same buffer.
Enzyme and substrate stock solutions were kept on ice until the
assay was initiated. The assay was done in a 384-well plate with a
final reaction volume of 20 µL. Enzyme (9 µL) was added to the
plate, and then 1 µL of compound in DMSO was added and
1
give the title compound as a white TFA salt (28.9 mg, 18%). H
NMR (300 MHz, DMSO-d6/TFA-d) δ 1.36 (s, 3H), 1.93 (t, J )
8.5 Hz, 2H), 2.70 (t, J ) 8.9 Hz, 2H), 2.82 (d, J ) 16.4 Hz, 1H),
2.97 (d, J ) 16.4 Hz, 1H), 3.20 (s, 3H), 3.84 (s, 3H), 6.95 (dd, J
) 8.1, 2.0 Hz, 1H), 7.18 (d, J ) 2.2 Hz, 2H), 7.23 (t, J ) 6.9 Hz,
3H), 7.38 (t, J ) 7.9 Hz, 2H), 7.51 (t, J ) 7.7 Hz, 2H); MS
(APCI+) m/z 352 [M + H]+; SFC Method G, tR ) 5.18 min, >99%
pure, >99% ee; HRMS (TOF) m/z calcd for C21H25N3O2 [M +
H]+, 352.2020; found, 352.2023.
(R)-2-Amino-6-[2-(3′-methoxybiphenyl-3-yl)ethyl]-3,6-dimethyl-
5,6-dihydro-3H-pyrimidin-4-one (29). This enantiomer was iso-
lated from the chiral purification of 28 and afforded the title
1
compound as a white TFA salt (36.2 mg, 22%). H NMR (300
MHz, DMSO-d6/TFA-d) δ 1.36 (s, 3H), 1.93 (t, J ) 8.5 Hz, 2H),
2.71 (t, J ) 14.7 Hz, 2H), 2.82 (d, J ) 16.4 Hz, 1H), 2.97 (d, J )
16.4 Hz, 1H), 3.20 (s, 3H), 3.84 (s, 3H), 6.95 (dd, J ) 8.1, 2.0 Hz,
1H), 7.18 (d, J ) 2.2 Hz, 1H), 7.23 (t, J ) 6.9 Hz, 2H), 7.38 (t, J
) 7.9 Hz, 2H), 7.51 (t, J ) 7.7 Hz, 2H); MS (APCI+) m/z 352 [M
+ H]+; SFC Method G, tR ) 4.36 min, >99% pure, >99% ee;
HRMS (TOF) m/z calcd for C21H25N3O-2 [M + H]+, 352.2020;
found, 352.2025.
Crystallography. Recombinant human BACE-1 residues 14-
453 was produced in bacteria as inclusion bodies and refolded using
a method described previously.47 Crystals of apo-BACE-1 suitable
for ligand soaking were obtained by a procedure described by Patel
et al.47 In brief, BACE-1 was buffer exchanged into 20 mM Tris
(pH 8.2), 150 mM NaCl, 1 mM DTT and concentrated to 8 mg/
mL. DMSO (3% v/v) was added to the protein prior to crystal-
lization. Apo-BACE-1 crystals were grown using the hanging drop
vapor diffusion method at 20 °C. Protein was mixed with an equal
volume of mother liquor containing 20-22.5% (w/v) PEG 5000
monomethylethyl (MME), 200 mM sodium citrate (pH 6.6), and
200 mM ammonium iodide. Crystals were cryoprotected for data
collection by brief immersion in 30% PEG 5000 MME, 100 mM
sodium citrate (pH 6.6), 200 mM ammonium iodide, and 20% (v/
v) glycerol. For inhibitor soaking, 0.5-0.1 M DMSO stock solutions
were made. A 10-fold dilution of the compound stock solution in
a stabilization solution (33% (w/v) PEG 5000 monomethylethyl
(MME), 110 mM sodium citrate (pH 6.6), and 220 mM ammonium
iodide was made. Crystals were added to the soaking solution for
up to 6 h prior to flash freezing and data collection. Data were
collected at a number of synchrotron beam lines at the European

Cyclic Amidine â-Secretase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 5923
preincubated for 5 min. Substrate (10 µL) was added, and the
reaction proceeded in the dark for 1 h at room temperature. The
final dilution of enzyme was 1:140; the final concentration of
substrate was 15 µM (Km of 25 µM). BACE-1 activity in the
presence of DMSO alone defined the 100% activity level, and 0%
activity was defined by completely inhibiting BACE-1 with 1 µM
of a TSI inhibitor (KTEEISEVN-sta-VAEF).
medium with DMSO at a final concentration of 1%. After the plate
was incubated at 37 °C for 24 h, 100 µL of cell medium was
transferred to a round-bottom 96-well plate to quantify Aâ1-40
levels. The cell culture plates were saved for ATP assay as described
below. To each well of the round-bottom plate, 50 µL of detection
solution containing 0.2 mg/mL RaAb40 (BioSource) and 0.25 µg/
mL biotin-4G8 (Signet) (prepared in Dulbecco’s phosphate-buffered
saline (DPBS) with 0.5%BSA and 0.5% Tween-20) was added and
incubated at 4 °C for >7 h. Then a 50 µL solution (prepared in the
same buffer as above) containing 0.062 mg/mL Ru-GAR and 0.125
mg/mL Dynabeads were added per well. The solutions in the round-
bottom plate were shaken at room temperature on a plate shaker
for 1 h, and the plates were then measured for electrochemilumi-
nescence counts in a Bioveris M8 Analyzer. Aâ1-40 standard
curves were obtained with two-fold serial dilution of an Aâ1-40
stock solution of known concentration in the same cell culture
medium used in cell-based assays. ATP Assay: As indicated above,
after transferring 100 µL of medium from cell culture plates for
Aâ1-40 detection, the plates, which still contained cells, were
saved for cytotoxicity assays by using the assay kit (ViaLight Plus)
from Cambrex BioScience that measures total cellular ATP. Briefly,
to each well of the plates, 50 µL of cell lysis reagent was added.
The plates were incubated at room temperature for 10 min. Two
minutes following addition of 100 µL reconstituted ViaLight Plus
reagent for ATP measurement, the luminescence of each well was
measured in an LJL Analyst or Wallac Victor II plate reader.
IGEN Assay. Assay Protocol: BACE-1 activity was also
quantified with a nonfluorescent substrate, and the product of the
reaction was measured with a technology developed by Bioveris.
The nonfluorescent assay was run as follows. BACE-1 was diluted
1:100 in 40 mM MES pH 5.0, and a 2 mM stock solution of Biotin-
TEEISEVNLDAEFRHDS (purchased from Synpep Corporation)
was diluted to 12 µM in the same buffer. DMSO stock solutions
of compounds or DMSO alone were diluted to the desired
concentration in 40 mM MES pH 5.0. Compound in DMSO (3
µL) was added to the plate followed by 27 µL of BACE-1, and
then the mixture preincubated for 5 min. The reaction was started
by adding 30 µL of substrate. The final dilution of enzyme was
1:200; the final concentration of substrate was 6 µM (Km is 150
µM). After a 20 min reaction at room temperature, the reaction
was stopped by removing 10 µL of the reaction mix and diluting
it 1:25 in 0.20 M Tris pH 8.0. Addition of DMSO alone defined
100% BACE-1 activity while 1 µM KTEEISEVN-sta-VAEF was
used to define 0% activity. Product Measurement with Bioveris
Technology: The biotin-labeled product was captured with strepta-
vidin-coated Dynabeads and then quantified with a custom-made
rabbit polyclonal antibody (Zymed) made to recognize the Swedish
mutant neoepitope formed after BACE-1 cleavage. The Bioveris
technology is based on the electrochemiluminescence of ruthenium-
based proprietary labels, which in this case was covalently attached
to a goat anti-rabbit antibody (Ru-GAR) purchased from Jackson
ImmunoResearch Labs. The Ru-GAR was made according to the
protocol suggest by the manufacturer. Briefly, the Ru label and
antibody were mixed at a 10:1 challenge ratio and then incubated
for 1 h. The reaction was quenched with 2 M glycine and the excess
label removed by gel filtration over a NAP 5 column from
Amersham Biosciences. Quantification was done as described
below. All antibodies and the streptavidin-coated Dynabeads were
diluted into Dulbecco-PBS containing 0.5% bovine serum albumin
and 0.5% Tween-20. The product was quantified by adding 50 µL
of a 1:5000 dilution of the neoepitope antibody to 50 µL of the
1:25 dilution of the BACE-1 reaction mixture. Then, 100 µL of
Dulbecco-PBS (0.5% BSA, 0.5% Tween-20) containing 0.2 mg/
mL IGEN beads and a 1:5000 dilution of the Ru-GAR antibody
were added. The final dilution of neoepitope antibody was 1:20 000,
the final dilution of Ru-GAR was 1:10 000, and the final concentra-
tion of beads was 0.1 mg/mL. The mixture is read on a Bioveris
M8 Analyzer after a 2 h incubation at room temperature.
Acknowledgment. We wish to thank Dr. Christof Angst for
his support and encouragement during the course of this work,
and Dr. Todd Brugel for helpful suggestions in the preparation
of this manuscript.
Supporting Information Available: Detailed description of
HPLC methods, HPLC purity analysis of final compounds, and
experimental procedures for compounds 8b, 8d, 8f, 9a, 9d-f, 10b-
d, 11b,c, 12c, 13a-c, 25, and 26. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) Herbert, L. E.; Scherr, P. A.; Bienias, J. L.; Bennett, D. A.; Evans,
D. A. Alzheimer Disease in the U.S. population: prevalence estimates
using the 2000 census. Arch. Neurol. 2003, 60, 1119-1122.
(2) Losing a million minds: confronting the tragedy of Alzheimer’s
Disease and other dementias. U.S. Congress Office of Technology
Assessment; U.S. Goverment Printing Office: Washington, DC,
1987; p 14.
(3) Alzheimer’s Disease medications fact sheet, NIH Publication No.03-
3431; Alzheimer’s Disease Education and Referral Center: Silver
Spring, MD, 2006; pp 1-6.
(4) Hardy, J.; Selkoe, D. J. The amyloid hypothesis of Alzheimer’s
disease: progress and problems on the road to therapeutics. Science
2002, 297, 353-356.
BACE-1 Cellular Assay. Generation of HEK-Fc33-1 Cell
Line: The cDNA encoding full length BACE-1 was fused in frame
with a three amino acid linker (Ala-Val-Thr) as described by Vassar
et al.9 to the Fc portion of the human IgG1 starting at amino acid
104. The BACE-1-Fc construct was then cloned into a GFP/pGEN-
IRES-neoK vector (a proprietary vector of AstraZeneca) for protein
expression in mammalian cells. The expression vector was stably
transfected into HEK-293 cells using the calcium phosphate method.
Colonies were selected with 250 µg/mL of geneticin sulfate (G-
418). Limited dilution cloning was performed to generate homo-
geneous cell lines. Clones were characterized by levels of APP
expression and Aâ1-40 secreted in the conditioned media using
an ELISA assay. Aâ1-40 secretion of BACE-1/Fc clone Fc33-1
was moderate. Cell Culture: HEK293 cells stably expressing human
BACE-1 (HEK-Fc33-1) were grown at 37 °C in Dulbecco’s
modified Eagle’s medium (DMEM) containing 10% heat-inhibited
FBS, 0.5 mg/mL antibiotic-antimycotic solution, and 0.05 mg/mL
G-418. Aâ1-40 Release Assay: Cells were harvested when
between 80 and 90% confluent. An amount of 100 µL of cells at
a cell density of 1.5 × 106/mL were added to a white 96-well cell
culture plate with clear flat bottom, or a clear, flat-bottom 96-well
cell culture plate, containing 100 µL of inhibitor in cell culture
(5) Sinha, S.; Lieberburg, I. Cellular mechanisms of â-amyloid production
and secretion. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 11049-11053.
(6) Hussain, I.; Powell, D.; Howlett, D. R.; Tew, D. 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. Identification of a Novel Aspartic Protease (Asp 2) as
â-Secretase. Mol. Cell Neurosci. 1999, 14, 419-427.
(7) Lin, X.; Koelsch, G.; Wu, S.; Downs, D.; Dashti, A.; Tang, J. Human
aspartic protease memapsin 2 cleaves the â-secretase site of â-amy-
loid precursor protein. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 1456-
1460.
(8) Sinha, 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, G.; Tung, J.; Schenk, D.; Seubert, P.;
Suomensaari, S. M.; Wang, S.; Walker, D.; Zhao, J.; McConlogue,
L.; John, V. Purification and cloning of amyloid precursor protein
â-secretase from human brain. Nature 1999, 402, 537-540.
(9) 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.; Fisher, 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. â-Secretase cleavage of Alzhe-
imer’s amyloid precursor protein by the transmembrane aspartic
protease BACE. Science 1999, 286, 735-741.

5924 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Edwards et al.
(10) Cumming, J. N.; Iserloh, U.; Kennedy, M. E. Design and development
of BACE-1 inhibitors. Curr. Opin. Drug DiscoVery DeV. 2004, 7,
536-556.
(11) John, V.; Beck, J. P.; Bienkowski, M. J.; Sinha, S.; Heinrikson, R.
L. Human â-Secretase (BACE) and BACE Inhibitors. J. Med. Chem.
2003, 46, 4625-4630.
(12) Durham, T. B.; Shepherd, T. A. Progress toward the discovery and
development of efficacious BACE inhibitors. Curr. Opin. Drug
DiscoVery DeV. 2006, 9, 776-791.
(13) Baxter, E. W.; Reitz, A. B. BACE inhibitors for the treatment of
Alzheimer’s disease. Annu. Rep. Med. Chem. 2005, 40, 35-48.
(14) Schmidt, B.; Baumann, S.; Braun, H. A.; Larbig, G. Inhibitors and
modulators of â- and γ-secretase. Curr. Top. Med. Chem. 2006, 6,
377-392.
(15) Guo, T.; Hobbs, D. W. Development of BACE-1 inhibitors for
Alzheimer’s disease. Curr. Med. Chem. 2006, 13, 1811-1829.
(16) Chen, S.-H.; Lamar, J.; Guo, D.; Kohn, T.; Yang, H.-C.; McGee, J.;
Timm, D.; Erickson, J.; Yip, Y.; May, P.; McCarthy, J. P3 cap
modified Phe*-Ala series BACE inhibitors. Bioorg. Med. Chem. Lett.
2004, 14, 245-250.
(17) Lamar, J.; Hu, J.; Bueno, A. B.; Yang, H.-C.; Guo, D.; Copp, J. D.;
McGee, J.; Gitter, B.; Timm, D.; May, P.; McCarthy, J.; Chen, S.-
H. Phe*-Ala-based pentapeptide mimetics are BACE inhibitors: P2
and P3 SAR. Bioorg. Med. Chem. Lett. 2004, 14, 239-243.
(18) Hu, B.; Fan, K. Y.; Bridges, K.; Chopra, R.; Lovering, F.; Cole, D.;
Zhou, P.; Ellingboe, J.; Jin, G.; Cowling, R.; Bard, J. Synthesis and
SAR of bis-statine based peptides as BACE 1 inhibitors. Bioorg.
Med. Chem. Lett. 2004, 14, 3457-3460.
(19) Ghosh, A. K.; Devasamudram, T.; Hong, L.; DeZutter, C.; Xu, X.;
Weerasena, V.; Koelsch, G.; Bilcer, G.; Tang, J. Structure-based
design of cycloamide-urethane-derived novel inhibitors of human
brain memapsin 2 (â-secretase). Bioorg. Med. Chem. Lett. 2005, 15,
15-20.
(20) Ghosh, A. K.; Bilcer, G.; Harwood, C.; Kawahama, R.; Shin, D.;
Hussain, K. A.; Hong, L.; Loy, J. A.; Nguyen, C.; Koelsch, G.;
Ermolieff, J.; Tang, J. Structure-based design: potent inhibitors of
human brain memapsin 2 (â-secretase). J. Med. Chem. 2001, 44,
2865-2868.
(21) Hom, R. K.; Fang, L. Y.; Mamo, S.; Tung, J. S.; Guinn, A. C.;
Walker, D. E.; Davis, D. L.; Gailunas, A. F.; Thorsett, E. D.; Sinha,
S.; Knops, J. E.; Jewett, N. E.; Anderson, J. P.; John, V. Design and
synthesis of statine-based cell-permeable peptidomimetic inhibitors
of human â-secretase. J. Med. Chem. 2003, 46, 1799-1802.
(22) Hom, R. K.; Gailunas, A. F.; Mamo, S.; Fang, L. Y.; Tung, J. S.;
Walker, D. E.; Davis, D.; Thorsett, E. D.; Jewett, N. E.; Moon, J.
B.; John, V. Design and synthesis of hydroxyethylene-based pepti-
domimetic inhibitors of human â-secretase. J. Med. Chem. 2004, 47,
158-164.
(23) Yang, W.; Lu, W.; Lu, Y.; Zhong, M.; Sun, J.; Thomas, A. E.;
Wilkinson, J. M.; Fucini, R. V.; Lam, M.; Randal, M.; Shi, X.-P.;
Jacobs, J. W.; McDowell, R. S.; Gordon, E. M.; Ballinger, M. D.
Aminoethylenes: a tetrahedral intermediate isostere yielding potent
inhibitors of the aspartyl protease BACE-1. J. Med. Chem. 2006,
49, 839-842.
(24) Brady, S. F.; Singh, S.; Crouthamel, M.-C.; Holloway, M. K.; Coburn,
C. A.; Garsky, V. M.; Bogusky, M.; Pennington, M. W.; Vacca, J.
P.; Hazuda, D.; Lai, M.-T. Rational design and synthesis of selective
BACE-1 inhibitors. Bioorg. Med. Chem. Lett. 2004, 14, 601-604.
(25) Hanessian, S.; Yun, H.; Hou, Y.; Yang, G.; Bayrakdarian, M.;
Therrien, E.; Moitessier, N.; Roggo, S.; Veenstra, S.; Tintelnot-
Blomley, M.; Rondeau, J.-M.; Ostermeier, C.; Strauss, A.; Ramage,
P.; Paganetti, P.; Neumann, U.; Betschart, C. Structure-based design,
synthesis, and memapsin 2 (BACE) inhibitory activity of carbocyclic
and heterocyclic peptidomimetics. J. Med. Chem. 2005, 48, 5175-
5190.
(26) Hong, L.; Koelsch, G.; Lin, X.; Wu, S.; Terzyan, S.; Ghosh, A. K.;
Zhang, X. C.; Tang, J. Structure of the protease domain of memapsin
2 (â-secretase) complexed with inhibitor. Science 2000, 290, 150-
153.
(27) Clark, D. E. Rapid calculation of polar molecular surface area and
its application to the prediction of transport phenomena. 2. Prediction
of blood-brain barrier penetration. J. Pharm. Sci. 1999, 88, 815-
821.
(28) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J.
Experimental and computational approaches to estimate solubility
and permeability in drug discovery and development settings. AdV.
Drug DeliVery ReV. 2001, 46, 3-26.
from an IC50 with a LE derived from a Ki is not strictly valid.
However the differences are small. Since Ki values are generally lower
than IC50 values, use of a Ki generally underestimates a LE relative
to one derived from an IC50.
(31) Stachel, S. J.; Coburn, C. A.; Steele, T. G.; Jones, K. G.; Loutzenhiser,
E. F.; Gregro, A. R.; Rajapakse, H. A.; Lai, M.-T.; Crouthamel, M.-
C.; Xu, M.; Tugusheva, K.; Lineberger, J. E.; Pietrak, B. L.; Espeseth,
A. S.; Shi, X.-P.; Chen-Dodson, E.; Holloway, M. K.; Munshi, S.;
Simon, A. J.; Kuo, L.; Vacca, J. P. Structure-based design of potent
and selective cell-permeable inhibitors of human â-secretase (BACE-
1). J. Med. Chem. 2004, 47, 6447-6450.
(32) Coburn, C. A.; Stachel, S. J.; Li, Y.-M.; Rush, D. M.; Steele, T. G.;
Chen-Dodson, E.; Holloway, M. K.; Xu, M.; Huang, Q.; Lai, M.-T.;
DiMuzio, J.; Crouthamel, M.-C.; Shi, X.-P.; Sardana, V.; Chen, Z.;
Munshi, S.; Kuo, L.; Makara, G. M.; Annis, D. A.; Tadikonda, P.
K.; Nash, H. M.; Vacca, J. P. Identification of a small molecule
nonpeptide active site â-secretase inhibitor that displays a nontra-
ditional binding mode for aspartyl proteases. J. Med. Chem. 2004,
47, 6117-6119.
(33) Garino, C.; Pietrancosta, N.; Laras, Y.; Moret, V.; Rolland, A.;
Quelever, G.; Kraus, J.-L. BACE-1 inhibitory activities of new
substitutedphenyl-piperazinecoupledtovariousheterocycles: Chromene,
coumarin and quinoline. Bioorg. Med. Chem. Lett. 2006, 16, 1995-
1999.
(34) Huang, D.; Luethi, U.; Kolb, P.; Cecchini, M.; Barberis, A.; Caflisch,
A. In silico discovery of â-secretase inhibitors. J. Am. Chem. Soc.
2006, 128, 5436-5443.
(35) Huang, D.; Luethi, U.; Kolb, P.; Edler, K.; Cecchini, M.; Audetat,
S.; Barberis, A.; Caflisch, A. Discovery of cell-permeable nonpeptide
inhibitors of â-secretase by high-throughput docking and continuum
electrostatics calculations. J. Med. Chem. 2005, 48, 5108-5111.
(36) Cole, D. C.; Manas, E. S.; Stock, J. R.; Condon, J. S.; Jennings, L.
D.; Aulabaugh, A.; Chopra, R.; Cowling, R.; Ellingboe, J. W.; Fan,
K. Y.; Harrison, B. L.; Hu, Y.; Jacobsen, S.; Jin, G.; Lin, L.; Lovering,
F. E.; Malamas, M. S.; Stahl, M. L.; Strand, J.; Sukhdeo, M. N.;
Svenson, K.; Turner, M. J.; Wagner, E.; Wu, J.; Zhou, P.; Bard, J.
Acylguanidines as small-molecule â-secretase inhibitors. J. Med.
Chem. 2006, 49, 6158-6161.
(37) Rajapakse, H. A.; Nantermet, P. G.; Selnick, H. G.; Munshi, S.;
McGaughey, G. B.; Lindsley, S. R.; Young, M. B.; Lai, M.-T.;
Espeseth, A. S.; Shi, X.-P.; Colussi, D.; Pietrak, B.; Crouthamel, M.-
C.; Tugusheva, K.; Huang, Q.; Xu, M.; Simon, A. J.; Kuo, L.;
Hazuda, D. J.; Graham, S.; Vacca, J. P. Discovery of oxadiazoyl
tertiary carbinamine inhibitors of â-secretase (BACE-1). J. Med.
Chem. 2006, 49, 7270-7273.
(38) Murray, C. W.; Callaghan, O.; Chessari, G.; Cleasby, A.; Congreve,
M.; Frederickson, M.; Hartshorn, M. J.; McMenamin, R.; Patel, S.;
Wallis, N. Application of fragment screening to â-secretase. J. Med.
Chem. 2007, 50, 1116-1123.
(39) Congreve, M.; Aharony, D.; Albert, J.; Callaghan, O.; Campbell, J.;
Carr, R. A. E.; Chessari, G.; Cowan, S.; Edwards, P. D.; Frederickson,
M.; McMenamin, R.; Murray, C. W.; Patel, s.; Wallis, N. Application
of fragment screening to the discovery of aminopyridines as inhibitors
of â-secretase. J. Med. Chem. 2007, 50, 1124-1132.
(40) Geschwindner, S.; Olsson, L.-L.; Deinum, J.; Albert, J.; Edwards,
P. D.; de Beer, T.; Folmer, R. H. A. Discovery of novel warheads
against â-secretase through fragment-based lead generation. J. Med.
Chem. 2007, 50, 5903-5911.
(41) Hyde, K. A.; Acton, E. M.; Baker, B. R.; Goodman, L. Potential
anticancer agents. LXXI. Diaminopyrimidines and diamino-s-triazines
related to daraprim. J. Org. Chem. 1962, 27, 1717-1722.
(42) Alker, D.; Campbell, S. F.; Cross, P. E.; Burges, R. A.; Carter, A.
J.; Gardiner, D. G. Long-acting dihydropyridine calcium antagonists.
3. Synthesis and structure-activity relationships for a series of
2-[(heterocyclylmethoxy)methyl] derivatives. J. Med. Chem. 1989,
32, 2381-2388.
(43) Burtner, R. R. 2-Amino-6-aryl-5,6-dihydro-4-hydroxypyrimidines. US
Patent US2748120, 1956.
(44) Vincent, S.; Grenier, S.; Valleix, A.; Salesse, C.; Lebeau, L.;
Mioskowski, C. Synthesis of enzymically stable analogues of GDP
for binding studies with transducin, the G-protein of the visual
photoreceptor. J. Org. Chem. 1998, 63, 7244-7257.
(45) Cockerill, A. F.; Deacon, A.; Harrison, R. G.; Osborne, D. J.; Prime,
D. M.; Ross, W. J.; Todd, A.; Verge, J. P. An improved synthesis of
2-amino-1,3-oxazoles under basic catalysis. Synthesis 1976, 591-
593.
(46) Bull, S. D.; Davies, S. G.; Garner, A. C.; O’Shea, M. D. Conjugate
additions of organocuprates to a 3-methylene-6-isopropyldiketopip-
erazine acceptor for the asymmetric synthesis of homochiral R-amino
acids. J. Chem. Soc., Perkin Trans. 1 2001, 3281-3287.
(47) Patel, S.; Vuillard, L.; Cleasby, A.; Murray, C. W.; Yon, J. Apo and
inhibitor complex structures of BACE (â-secretase). J. Mol. Biol.
2004, 343, 407-416.
(29) Hopkins, Andrew, L.; Groom, Colin, R.; Alex, A. Ligand ef-
ficiency: a useful metric for lead selection. Drug. DiscoVery Today
2004, 9, 430-431.
(30) Since IC50 values are dependent upon experimental conditions and
are not a kinetic constant, the comparison of a LE value derived

Cyclic Amidine â-Secretase Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 5925
(48) Brown, F. J.; Matassa, V. G. Preparation of heterocyclic carboxamides
as leukotriene antagonists. US Patent Cont.-in-part of U.S. Ser. No.
181,455, abandoned, 1990.
(49) Boatman, S.; Harris, T. M.; Hauser, C. R. Synthesis of ketones of
the type CH3COCH2R from acetylacetone and halides with ethanolic
potassium carbonate. An alkylation-cleavage process. J. Org. Chem.
1965, 30, 3321-3324.
(50) Leslie, A. G. W.; Brick, P.; Wonacott, A. MOSFLM. Daresbury Lab.
Inf. Quart. Protein Crystallogr. 1986, 18, 33-39.
(51) Bailey, S. The CCP4 suite: programs for protein crystallography.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 1994, D50, 760-763.
(52) Blundell, T. L.; Abell, C.; Cleasby, A.; Hartshorn, M. J.; Tickle, I.
J.; Parasini, E.; Jhoti, H. High-throughput X-ray crystallography for
drug discovery. Spec. Publ. - R. Soc. Chem. 2002, 279, 53-59.
(53) Hartshorn, M. J. AstexViewerTM: a visualization aid for structure-
based drug design. J. Comput.-Aided Mol. Des. 2003, 16, 871-881.
(54) Winn, M. D.; Isupov, M. N.; Murshudov, G. N. Use of TLS
parameters to model anisotropic displacements in macromolecular
refinement. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2001, D57,
122-133.
(55) Luellau, E.; Kanttinen, A.; Hassel, J.; Berg, M.; Haag-Alvarsson,
A.; Cederbrant, K.; Greenberg, B.; Fenge, C.; Schweikart, F.
Comparison of batch and perfusion culture in combination with pilot-
scale expanded bed purification for the production of soluble
recombinant â-secretase. Biotechnol. Prog. 2003, 19, 37-44.
JM070829P