Design and synthesis of bicyclic acetals as Beta Secretase (BACE1) inhibitors

Article abstract of DOI:10.1016/j.bmc.2017.03.030

Taking advantage of the structural similarity between aspartic proteases, small-molecule peptidomimetic inhibitors that already showed activity towards Secreted Aspartic Protease 2 as anti-Candida agents and HIV protease inhibitors were exploited as potential BACE1 inhibitors. A focused library of 6,8-dioxa-3-azabicyclo[3.2.1]-octane peptidomimetic scaffolds was synthesized and assayed towards BACE1 enzyme, resulting in the identification of a thiolactam-containing hit compound possessing IC₅₀ in the low micromolar range, and confirming the bicyclic acetal portion as a potential transition state analogue in the interaction with catalytic aspartic acid residues.

Full text of DOI:10.1016/j.bmc.2017.03.030

Accepted Manuscript
Design and synthesis of bicyclic acetals as Beta Secretase (BACE1) inhibitors
Riccardo Innocenti, Elena Lenci, Gloria Menchi, Alberto Pupi, Andrea
Trabocchi
PII:
S0968-0896(17)30528-X
http://dx.doi.org/10.1016/j.bmc.2017.03.030
BMC 13628
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
30 January 2017
10 March 2017
15 March 2017
Please cite this article as: Innocenti, R., Lenci, E., Menchi, G., Pupi, A., Trabocchi, A., Design and synthesis of
bicyclic acetals as Beta Secretase (BACE1) inhibitors, Bioorganic & Medicinal Chemistry (2017), doi: http://
dx.doi.org/10.1016/j.bmc.2017.03.030
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Graphical Abstract
Design and synthesis of bicyclic acetals as Beta Secretase
(BACE1) inhibitors
Riccardo Innocenti,a Elena Lenci,a Gloria Menchi,a,b Alberto
Pupi,b,c,d Andrea Trabocchia,b,*
a Department of Chemistry “Ugo Schiff”, University of
Florence, Via della Lastruccia 13, 50019 Sesto Fiorentino, Florence, Italy
b Interdepartmental Center for Preclinical Development of Molecular Imaging (CISPIM), University of
Florence, Viale Morgagni 85, 50134 Florence, Italy
c Department of Clinical and Experimental Biomedical Science “Mario Serio”, University of Florence, Largo
Brambilla 3, 50134 Florence, Italy
d Azienda Ospedaliera Universitaria Careggi (AOUC), Largo Brambilla 3, 50134 Florence, Italy

Bioorganic & Medicinal Chemistry
journal homepage: www.els evier.com
Design and synthesis of bicyclic acetals as Beta Secretase (BACE1) inhibitors
Riccardo Innocenti,a,§ Elena Lenci,a,§ Gloria Menchi,a,b Alberto Pupi,b,c,d Andrea Trabocchia,b,*
a Department of Chemistry “Ugo Schiff”, University of Florence, Via della Lastruccia 13, 50019 Sesto Fiorentino, Florence, Italy
b Interdepartmental Center for Preclinical Development of Molecular Imaging (CISPIM), University of Florence, Viale Morgagni 85, 50134 Florence, Italy
c Department of Clinical and Experimental Biomedical Science “Mario Serio”, University of Florence, Largo Brambilla 3, 50134 Florence, Italy
d Azienda Ospedaliera Universitaria Careggi (AOUC), Largo Brambilla 3, 50134 Florence, Italy
* Corresponding author. Phone, +39 055 4573507; fax, +39 055 4574913; E-mail, andrea.trabocchi@unifi.it.
§ These authors contributed equally to the work.
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
Taking advantage of the structural similarity between aspartic proteases, small-molecule
peptidomimetic inhibitors that already showed activity towards Secreted Aspartic Protease 2 as
anti-Candida agents and HIV protease inhibitors were exploited as potential BACE1 inhibitors.
A focused library of 6,8-dioxa-3-azabicyclo[3.2.1]-octane peptidomimetic scaffolds was
synthesized and assayed towards BACE1 enzyme, resulting in the identification of a thiolactam-
containing hit compound possessing IC50 in the low micromolar range, and confirming the
bicyclic acetal portion as a potential transition state analogue in the interaction with catalytic
aspartic acid residues.
Available online
Keywords:
Peptidomimetics
Amino acids
Enzyme inhibition
Alzheimer’s disease
Drug discovery
2009 Elsevier Ltd. All rights reserved.
1. Introduction
but about 5-10% ends at residue 42, which is a major player in
AD pathogenesis. Thus, both β and γ secretases are required for
According to the World Health Organization (WHO), about
35.6 millions people have been diagnosed for dementia in 2010,
and the number is projected to double every twenty years.1 The
overall cost for dementia worldwide has been estimated
approximately 604 billion dollars.2 Alzheimer’s Disease (AD) is
the most common cause of dementia in older people, consisting
of a chronic and progressive neurodegenerative disorder.3 The
classic clinical symptoms of the disease are memory loss and
difficulties with thinking, problem-solving or language. Actually,
the accessible therapies for such disease succeeded only in
slowing the cognitive decline due to AD. The pathological
hallmarks of AD include two types of lesions in the brain, the
extracellular accumulation of amyloid plaques composed of the
β-amyloid (Aβ) peptide, and intracellular neurofibrillary tangles
(NFT) resulting from the aggregation of hyperphosphorylated
microtubule-associated protein tau.4
the production of Aβ, and the inhibition or modulation of these
enzymes has been considered a key therapeutic approach for
reducing cerebral Aβ concentrations in patients with AD. The
important role of BACE1 in APP processing was evinced in vivo
from studies with knockout mice showing that both wild type and
mutant forms of APP failed to produce Aβ in the absence of
BACE1,7 further corroborating the importance of BACE1 as a
target for the therapeutic intervention of AD.
The first-peptide based BACE1 inhibitors used a transition-
state mimetic approach using a non-hydrolyzable peptide isostere
at the scissile site. Successively, other non-peptide inhibitors
based on diverse heterocyclic scaffolds were developed, all
addressing the catalytic aspartic acid diad through acid-base and
hydrogen-bond interactions.8 Some of these molecules possessing
high efficacy in vitro proved to reduce the secretion of Aβ in cell
cultures.9 Indeed, an important issue in the development of
BACE1 inhibitor drugs encompasses the difficulty in crossing the
blood-brain barrier (BBB), thus resulting in low potency in vivo,
and recent papers reported interesting improvements towards
brain penetration.10,11
The formation of Aβ is a sequential proteolytic process
consisting of the cleavage of the amyloid precursor protein (APP)
by the β- and γ-secretase enzymes. The β secretase, referred to as
Beta-site Amyloid precursor protein (APP) Cleaving Enzyme 1
(BACE1), is the enzyme that initiates Aβ production by cleaving
the extracellular domain of APP, generating a N-terminal Aβ
fragment and the membrane bound C-terminal fragment C99.5
Successively, intramembrane processing of the latter fragment by
γ-secretase yields Aβ.6 Most Aβ peptides terminate at residue 40
Several years ago our research group introduced a novel class
of bicyclic peptidomimetic scaffolds showing dipeptide isostere
potential.12 Such compounds are easily achieved from the
synthetic combination of suitable amino acids and tartaric acid or

carbohydrates derivatives, and show remarkable stability of the
acetal portion to acidic conditions.13 More recently, we identified
this class of peptidomimetic compounds as potential aspartic
protease inhibitors taking advantage of their role as dipeptide
isosteres, and of the bicyclic acetal portion as a potential
transition-state analogue in the interaction with key residues of
enzyme’s catalytic site. Among this library of peptidomimetics,
two hit candidates proved to be effective against drug-resistant C.
albicans strains in in vivo experiments.14-16 The structural analogy
of aspartic proteases expressed in different pathological systems
suggested the screening of this library of bicyclic compounds
towards other important therapeutic targets, such as the HIV
protease. Indeed, sub-micromolar hit compounds were identified
as HIV protease inhibitors,17 thus corroborating the capability of
these compounds as aspartic protease inhibitors by means of a
transition-state isosterism.
BACE1 (PDB: 4J0P) resulted in perfect alignment of the two
DTG and DSG aspartic protease motifs, showing a RMSD of
1.44 ang calculated on all alpha-carbon atoms (Figure 1, bottom).
Even in this case, the flap regions showed some degree of
overlap, although BACE1 showed flap residues being somewhat
distant to catalytic aspartic acid residues with respect to SAP2.
Specifically, the interactive hydrogen-bond donor Q134 of
BACE1 proved to be distant to the DTG/DSG protease motifs by
about 3.7 ang more than the corresponding G85 of SAP2, similar
to what was observed for the HIV protease. This evidence
showed that the 3D structural similarity was not as strict as
observed in the case of SAP2/HIV protease overlap,21 and was
considered an important rationale for the selection of candidate
bicyclic scaffolds addressing both the catalytic aspartic acids and
the flap region for the inhibition assays towards BACE1.
Accordingly, in this work we envisaged the application of
these bicyclic compounds against BACE1 as an additional
aspartic protease target. The analysis of the similarity of BACE1
with SAP2 from C. albicans and HIV-1 protease was carried out
through molecular modeling, and the key structural features of
the active site were investigated by a three-dimensional structural
superimposition of BACE1 with SAP2 and the HIV protease.
Also, we expanded the array of this class of bicyclic compounds
by the synthesis of novel bicyclic thiolactams, with aim to
explore the potential of the sulfur atom as a hydrogen-bond
acceptor in the key functional group at position 2 of these
scaffolds, and we applied them together with representative
library members in enzyme inhibition assays for the
identification of novel BACE1 inhibitors.
2. Results and Discussion
3D Structural alignment. BACE1 is a transmembrane
aspartic protease adopting a bilobal structure with the substrate
binding pocket between the N-terminal and C-terminal lobes of
the enzyme. It is closely related to the pepsin family of aspartic
proteases, and the catalytic domain of BACE is characterized by
the two canonical aspartic protease motifs of the sequence DTGS
and DSGT that both contribute to form the active site of the
enzyme.8,18 The flap region closes over the top of the cleft when
bound to the substrate/inhibitor and is characterized by Y132,
T133 and Q134.
We investigated the structural similarity between BACE1 and
both SAP2 and HIV protease, and the key structural features
around the active site characterized by the flap region and the two
DTGS and DSGT aspartic protease motifs by a three-dimensional
structural superimposition of the proteases using molecular
modeling (Figure 1). The Swiss PDB viewer (SPDBV) program
(4.0.1 version, Swiss Institute of Bioinformatics)19,20 was used to
superimpose three dimensional (3D) structures of the aspartic
proteases taking into account the alpha-carbon atoms of the
highly conserved DTG motifs.
Fig. 1. Top: superimposition of HIV-1 protease (PDB: 4HLA in green)
and BACE1 (PDB: 4J0P in cyan); down: superimposition of C.albicans
SAP2 (PDB: 1EAG in grey) and BACE1 (PDB: 4J0P in cyan). Both fittings
have been obtained with SPDBV. The red circles highlight the 3D overlap of
the aspartic protease residues around the flap regions.
Pairwise sequence alignment. Despite differences in both
size and overall amino acid sequence between BACE1, SAP2
from C. albicans, and the HIV-1 protease, significant homologies
exist between their active sites. In order to compare the primary
structures of these proteases, we performed a local sequence
alignment using the EMBOSS matcher algorithm,22 and high
similarity was found in the active site regions of both SAP2 and
the HIV-1 protease (Figure 2 and Table 1). Indeed, sequence
alignment of active sites containing the DTG motif disclosed a
similarity of 55% and identity of 30% between BACE1 N-
terminal domain (79-98 residues) and HIV-1 protease active site
region (11-30 residues). Higher similarity was found when
comparing the N-terminal regions of BACE1 (75-101 residues)
The superimposition of the HIV protease (PDB: 4HLA) and
BACE1 (PDB: 4J0P) resulted in the optimal alignment of the two
canonical aspartic protease motifs of the sequence DTG and
DSG, showing a RMSD of 1.48 ang calculated on all alpha-
carbon atoms (Figure 1, top). Although the flap regions proved to
overlap in general 3D structure, I50 of the HIV protease
overlapped with Y132 of BACE1 rather than Q134, which is the
corresponding hydrogen-bond donor in BACE1 interacting with
the substrate/inhibitor. Accordingly, Q134 proved to be distant
from the aspartic protease motifs of about 3.8 ang more than the
corresponding I50 as in the HIV protease. Similarly, the
superimposition of SAP2 of C. albicans (PDB: 1EAG) and

and SAP2 (14-40 residues), both containing the DTG sequence,
which showed 74.1% of similarity and 55.6% of identity. In
addition, the C-terminal regions of these two proteases showed
some degree of similarity, too, possessing 48.6% of similarity
and 18.9% of identity around the DSG catalytic motif.
ester lactams, the reaction of ester thiolactams 1-3 with primary
amines resulted in product degradation, possibly by ring opening
of the thiolactam ring.
Enzyme inhibition assays. The screening of compounds 1-21
(Table 2) at 10 µM conc. for enzyme inhibition was performed to
evaluate the modulation of BACE1 inhibition by the
Table 1. Local sequence similarities between BACE1 (4J0P)
and SAP2 (1EAG) / HIV-1 aspartic protease (4HLA) in the
active site region performed using EMBOSS matcher pairwise
sequence alignment.
stereochemical arrangement of the pharmacophoric elements of
the bicyclic compounds, and the different location of the amino
acid side-chain isostere. A fluorometric assay kit (SensoLyte
520 β-Secretase Assay Kit, Anaspec, USA) employing the
fluorescent substrate HiLyte Fluor™ 488-Glu-Val-Asn-Leu-Asp-
Ala-Glu-Phe-Lys (QXL™ 520)-OH was used, with excitation
and emission wavelengths of 490 and 520 nm, respectively.
Inhibition data (Table 2) showed lower inhibition potency as
compared to what observed for SAP2 and HIV protease,
suggesting that BACE1 possesses significant structural
differences in the catalytic site, such as the different spatial
relationship between the flap region and the two DTG and DSG
aspartic protease motifs, as previously observed.
%
%
Enzymes
Sequence
identity similarity
N-terminal BACE1 vs
HIV pr
79-98 vs 11-30
75-101 vs 14-40
287-323 vs 215-252
30
55
N-terminal BACE1 vs
SAP2
55.6
18.9
74.1
48.6
C-terminal BACE1 vs
SAP2
Lead SAP2 inhibitors 10 and 11, and the stereoisomers of 11
containing the leucinol appendage (cmpd 12-14) showed poor
inhibition, whereas lead HIV protease inhibitors 15 and 16
showed better profile, indicating good accomodation in the
BACE1 catalytic site. Similarly to what observed in previous
reports on inhibition data towards SAP2 and HIV protease,14,17
compounds lacking the carbonyl/thiocarbonyl group showed no
inhibition (Table 2, 21). Compound 4, possessing a C=S bond,
proved to be the most interesting inhibitor, and for this
compound a dose-response measurement using 0.003-30 µM as
the range of inhibitor concentrations was carried out, resulting in
a IC50 value of 2.6 µM (Figure 3).
Fig. 2. Sequence alignment between BACE1 (4J0P) and SAP2 (1EAG) /
HIV-1 aspartic protease (4HLA) in the active site region using EMBOSS
matcher pairwise sequence alignment.
Synthesis. The synthesis of bicyclic peptidomimetic
compounds was achieved using suitable amino acid and tartaric
acid derivatives through a coupling/cyclization approach, as
previously reported.12,13 Moreover, in order to expand the array of
bicyclic compounds with scaffold possessing the internal C=S
bond, a series thiolactams were conceived and synthesized from
the parent esters (Scheme 1).
The presence of the C=S bond in the bicyclic scaffold proved
to be relevant, as thioamides 1-8 showed some inhibition with
respect to other scaffolds, although with different degree,
suggesting a role of the sulfur atom in addressing the flap region
as a hydrogen-bonding acceptor. The morfoline appendage
proved beneficial (Table 2, compounds 1, 4, 7), although
contrasting results were achieved when changing the scaffold
(see inhibition data for 4-6, and for 5, 6 compared to 8, 9),
suggesting the importance of balancing contributions from such
appendage, the scaffold type and the C=O/C=S bond. Impaired
activity was observed as a function of the appendage at the
nitrogen atom and at the acetal position 5 of the scaffold, too, as
compounds 17-20 failed to show any inhibition. No significant
correlation was evinced with respect to the stereochemistry,
although scaffold B (Scheme 1) showed some preference, and the
matched combination of the decoration of such scaffold and the
presence of C=O/C=S bond contributed significantly to
inhibition, too (see inhibition data for compounds 5, 8, 15 and
16).
Molecular Docking. Autodock 4.0.125 was used to evaluate
the binding energies of minimum-energy conformations of
compound 4. Best-scoring conformations were clustered and
visually inspected for enzyme-ligand interactions. Specifically,
hydrogen-bonds to side chain oxygen atoms of the catalytic
aspartic acids (D93 and D289), the hydrophobic interactions in
the S1/S1’ pockets and hydrogen-bonding to the ‘flap region’
were taken into account.
Scheme 1. Synthetic approach for bicyclic acetal thiolactams
4-9.
Specifically, ester thiolactams 1-3 were prepared in good
yields starting from the corresponding bicyclic lactams I-III by
treatment with Lawesson’s reagent in dry toluene.23 Taking
advantage of the increased reactivity of the ester moiety at C-7
carbon atom,24 subsequent direct aminolysis under neat
Docking results of 4 showed this molecule addressing the S1
binding pocket, and the acetal bridge being oriented towards
catalytic D93 and D289, as expected (Figure 4).
conditions gave the corresponding amide thiolactams 4-9 after
overnight reaction at 45 °C. Contrary to the reactivity of bicyclic

Table 2. Inhibition data at 10 µM conc. of selected bicyclic compounds with diverse scaffold decoration and stereochemistry.
scaffold
%
scaffold
%
scaffold
%
Cmpd
Cmpd
Cmpd
Inhib.
Inhib.
Inhib.
O
O
Bn
OEt
OEt
N
N
O
O
1
8
15
O
O
16
16
36
0
46
32
Bn
N
N
N
S
O
O
Bn
Bn
2
9
16
N
S
3
4
5
6
7
10
11
12
13
14
17
18
19
20
21
15
61
5
10
8
0
0
0
0
0
29
0
O
H
O
N
Bn
N
9
0
0
O
Interestingly, the global-minimum conformer was
characterized by a hydrogen-bond between the thiocarbonyl
group at position 2 of the scaffold and Q134 amide proton on the
flap region of BACE1, confirming the typical binding orientation
of these scaffolds within aspartic acid proteases. A π-stacking
interactions between the N-benzyl group and F169/W176 side
chains was observed as a key hydrophobic contact. The
morpholine appendage proved to interact with R259 and Y189
establishing both polar and hydrophobic contacts. These results
indicated the absence of interactions within S3 subsite, which are
reported being essential for achieving high potency.26
Fig. 4 Compound 4 docked into the active site of BACE1 (PDB: 4J0P),
highlighting protein residues (magenta) that form key interactions: D93/D289
vs acetal bridge, F169/W176 vs N-benzyl group, Q134 vs C=S bond,
R259/Y189 vs morpholine group. Non-polar hydrogen atoms are omitted for
clarity.
Taking advantage of a library of bicyclic molecular scaffolds
possessing a bicyclic acetal portion as a potential transition-state
analogue, which already showed activity towards other aspartic
proteases such as SAP2 and HIV protease, we reasoned to test
them towards BACE1 in enzyme inhibition kinetic assays. The
analysis of the similarity of BACE1 with SAP2 from C. albicans
and HIV-1 aspartic proteases was carried out by molecular
modeling, and revealed a certain degree of structural analogy
within the catalytic site, although the flap regions proved to be
somewhat different in the three enzymes. Accordingly, a pool of
amide thiolactams, aimed to target the BACE1 flap region more
efficiently, were synthesized and assayed towards BACE1, thus
allowing for the identification of compound 4 with potency in the
Figure 3. Percentage of inhibition of BACE1 activity by 4, showing an
IC50 of 2.6 1.6 µM. Experiments were conducted in triplicate. Data are
presented as means SD from three independent experiments.
3. Conclusions
BACE1 is the enzyme that initiates Aβ production by cleaving
the extracellular domain of APP, and is recognised as a key
therapeutic approach for reducing cerebral Aβ concentrations in
patients with Alzheimer’s disease.

low micromolar range. Interestingly, other library members
containing the C=O bond did not show similar potency as in
previous studied aspartic proteases, confirming the existence of
significant structural differences between these enzymes within
the catalytic site.
Purification by flash chromatography eluiting with 3:1
hexane-ethyl acetate (Rf 0.45) gave the pure product 2 as a
yellow oil with a yield of 72%. [α]D23 = -101.4 (c 1.0, CHCl3). 1H
NMR (400 MHz, CDCl3) δ 7.49-7.16 (m, 8H), 7.06 (d, J = 7.7
Hz, 2H), 6.17 (d, J = 15.1 Hz, 1H), 5.62 (s, 1H), 5.50 (s, 1H),
4.79 (s, 1H), 4.48 (d, J = 15.1 Hz, 1H), 3.78 (s, 3H), 3.47 (dd, J =
10.8, 3.9 Hz, 1H), 3.23 (dd, J = 13.8, 3.9 Hz, 1H), 2.86 (dd, J =
13.7, 11.1 Hz, 1H). 13C NMR (50 MHz, CDCl3) δ 195.53, 168.99,
135.26, 134.61, 129.15, 129.05, 128.23, 127.85, 127.64, 127.45,
101.17, 83.43, 79.90, 63.09, 52.84, 52.57, 36.53. MS (ESI) m/z
(%): 406.17 [(M + Na)+, 100]. Anal. Calcd for C21H21NO4S
(383.46): C, 65.78; H, 5.52; N, 3.65. Found: C, 66.23; H, 5.79;
N, 3.49.
Molecular docking calculations on compound 4 within
BACE1 active site revealed an interesting binding mode,
confirming the property of this scaffold in targeting the catalytic
diad with the acetal moiety and interacting with the flap region
through C=S hydrogen-bond. This compound will need further
investigation and structure refinement in a hit-to-lead process in
order to achieve promising BACE1 inhibitors based on novel
molecular scaffolds. Specifically, potential toxicity issues of the
thioamide bond will be studied, and chemical optimization will
be be driven to improve bioactivity, without inhibiting the related
aspartic protease cathepsin D, and with suitable in vivo profile for
reducing cerebral Aβ concentrations in patients with Alzheimer’s
disease.
4.5. Methyl (1R,2R,5S,6S)-2,3-dibenzyl-4-thioxo-8-oxa-3-
azabicyclo[3.2.1]octane-6-carboxylate (3)
Purification by flash chromatography eluiting with 3:1
hexane-ethyl acetate (Rf 0.45) to give the pure product 3 as a
yellow oil with a yield of 69%. [α]D23 = +102.2 (c 1.0, CHCl3).
1H NMR (400 MHz, CDCl3) δ 7.44-7.18 (m, 8H), 7.06 (d, J = 7.1
Hz, 2H), 6.17 (d, J = 15.1 Hz, 1H), 5.62 (s, 1H), 5.51 (s, 1H),
4.80 (s, 1H), 4.48 (d, J = 15.1 Hz, 1H), 3.78 (s, 3H), 3.46 (dd, J =
7.3, 3.2 Hz, 1H), 3.23 (dd, J = 13.8, 4.0 Hz, 1H), 2.85 (dd, J =
13.7, 11.0 Hz, 1H) 13C NMR (50 MHz, CDCl3) δ 195.53, 168.99,
135.26, 134.61, 129.15, 129.05, 128.23, 127.85, 127.64, 127.45,
101.17, 83.43, 79.90, 63.09, 52.84, 52.57, 36.53. MS (ESI) m/z
(%): 406.25 [(M + Na)+, 100]. Anal. Calcd for C21H21NO4S
(383.46): C, 65.78; H, 5.52; N, 3.65. Found: C, 66.11; H, 5.69;
N, 3.51.
4. Experimental Section
4.1. General
Analytical grade solvents and commercially available reagents
were used without further purification. Reactions requiring an
inert atmosphere were carried out under a nitrogen atmosphere.
Dry toluene was distilled over Na/benzophenone. Flash column
chromatography (FCC) purifications were performed using
Merck silica gel (40-63 µm). TLC analyses were performed on
Merck silica gel 60 F254 plates. 1H NMR spectra were recorded
on a Varian Mercury 400 (1H: 400 MHz),13 C spectra were
recorded on a Varian Gemini 200 (13C: 50 MHz). The chemical
shifts (δ) and coupling constants (J) are expressed in parts per
million (ppm) and hertz (Hz), respectively. Optical rotations were
measured with JASCO DIP-360 digital polarimeter. Elemental
analyses were recorded on a Perkin Elmer 240 CHN Analyzer.
Mass spectra were carried out by direct inlet on a LCQ FleetTM
Ion Trap LC/MS system (Thermo Fisher Scientific) with an ESI
interface in the positive mode.
4.6. General procedure for the synthesis of amide thiolactams 4-
9
The reaction was carried out on a scale of 0.2-0.6 mmol. An
oven dried 25 mL round bottomed flask was equipped with a
magnetic stir bar and charged with the corresponding compound
1-3 (1 eq). The flask was cooled in an ice bath and the
appropriate amine was added (10 eq). The reaction mixture was
heated at 45 °C overnight. After adding ethyl acetate (40 mL),
the organic phase was washed with 1 M HCl (3 x 15 mL), satd.
NaHCO3 solution (3 x 15 mL) and brine (1 x 20 mL). The
organic phase was dried over Na2SO4, concentrated under a
reduced pressure and the crude product was purified by flash
chromatography.
4.2. General procedure for the synthesis of ester thiolactams 1-3
The reaction was carried out on a scale of 0.2-0.6 mmol. An
oven dried 25 mL round bottomed flask with two necks was
equipped with a magnetic stir bar and charged with the
Lawesson’s reagent (0.5 eq). A 0.2 M solution of the
corresponding bicyclic compound I-III (1 eq) in dry toluene was
added to the Lawesson’s reagent. The reaction mixture was
refluxed for two hours, then the solvent was removed under
reduced pressure and the crude product was purified by flash
column chromatography.
4.7. (1R,5R,7R)-3-benzyl-7-(morpholin-4-ylcarbonyl)-6,8-dioxa-
3-azabicyclo[3.2.1]octane-2-thione (4)
Obtained from 1 using morpholine as the amine. Purification
by flash chromatography eluiting with 2:3 hexane-ethyl acetate
(Rf 0.41) gave the pure product 4 as a white solid with a yield of
71%. [α]D23 = -36.9 (c 1.0, CHCl3). 1H NMR (400 MHz, CDCl3)
δ 7.45-7.15 (m, 5H), 5.95 (d, J = 2.3 Hz, 1H), 5.53 (s, 1H), 5.21
(d, J = 14.5 Hz, 1H), 5.07 (d, J = 14.5 Hz, 1H), 4.97 (s, 1H),
3.79-3.55 (m, 8H), 3.49 (dd, J = 13.5, 2.4 Hz, 1H), 3.25 (d, J =
13.5 Hz, 1H). 13C NMR (50 MHz, CDCl3) δ 196.08, 165.87,
134.14, 129.15, 128.43, 128.18, 99.96, 82.72, 79.36, 66.85,
66.61, 54.95, 53.69, 46.05, 42.72. MS (ESI) m/z (%): 371.22 [(M
+ Na)+, 100]. Anal. Calcd for C17H20N2O4S (348.42): C, 58.60; H,
5.79; N, 8.04. Found: C, 59.12; H, 5.98; N, 7.95.
4.3. Methyl (1S,5R,6R)-3-benzyl-4-thioxo-8-oxa-3-
azabicyclo[3.2.1]octane-6-carboxylate (1)
Purification by flash chromatography eluiting with 3:1
hexane-ethyl acetate (Rf 0.38) provided pure 1 as a brown solid
with a yield of 65%. [α]D23 = -74.4 (c 1.0, CHCl3). 1H NMR (400
MHz, CDCl3) δ 7.44-7.20 (m, 5H), 5.95 (d, J = 2.3 Hz, 1H), 5.50
(s, 1H), 5.14 (d, J = 2.6 Hz, 2H), 4.82 (s, 1H), 3.81 (s, 3H), 3.47
(dd, J = 13.5, 2.4 Hz, 1H), 3.25 (d, J = 13.5 Hz, 1H). 13C NMR
(50 MHz, CDCl3) δ 195.02, 168.83, 133.98, 128.87, 128.11,
127.83, 99.89, 82.92, 79.13, 77.80, 77.16, 76.52, 54.58, 53.37,
52.70. MS (ESI) m/z (%): 316.25 [(M + Na)+, 100]. Anal. Calcd
for C14H15NO4S (293.34): C, 57.32; H, 5.15; N, 4.77. Found: C,
57.72; H, 5.21; N, 4.68.
4.8. (1R.4S,5R,7R)-3.4-dibenzyl-7-(morpholin-4-ylcarbonyl)-
6,8-dioxa-3-azabicyclo[3.2.1]octane-2-thione (5)
Obtained from 2 using morpholine as the amine. Purification
by flash chromatography eluiting with 1:1 hexane-ethyl acetate
(Rf 0.47) to give the pure product 5 as a colorless oil with a yield
of 62%. [α]D23= -78.2 (c 1.0, CHCl3). 1H NMR (400 MHz,
4.4. Methyl (1S,2S,5R,6R)-2,3-dibenzyl-4-thioxo-8-oxa-3-
azabicyclo[3.2.1]octane-6-carboxylate (2)

CDCl3) δ 7.46-7.14 (m, 8H), 7.05 (d, J = 7.5 Hz, 2H), 6.12 (d, J
= 15.1 Hz, 1H), 5.63 (s, 1H), 5.50 (s, 1H), 4.94 (s, 1H), 4.53 (d, J
= 15.1 Hz, 1H), 3.74-3.43 (m, 9H), 3.22 (dd, J = 13.8, 3.8 Hz,
1H), 2.87 (dd, J = 13.3, 11.4 Hz, 1H), 13C NMR (50 MHz,
CDCl3) δ 196.29, 165.93, 135.32, 134.73, 129.27, 129.22,
128.32, 127.74, 127.51, 101.12, 83.10, 80.07, 66.85, 66.60,
63.37, 52.69, 46.01, 42.67, 36.6. MS (ESI) m/z (%): 461.39 [(M
+ Na)+, 100]. Anal. Calcd for C24H26N2O4S (438.54): C, 65.73; H,
5.98; N, 6.39. Found: C, 65.99; H, 6.03; N, 6.31.
with a yield of 62%. [α]D23 = +49.6 (c 1.0, CHCl3). 1H NMR (400
MHz, CDCl3) δ 7.43-7.17 (m, 8H), 7.05 (d, J=6.8 Hz, 1H), 6.10
(d, J = 15.1 Hz, 2H), 5.58 (s, 1H), 5.53 (s, 1H), 4.95 (s, 1H), 4.53
(d, J = 15.1 Hz, 1H), 4.14 (q, J = 7.1 Hz, 2H), 3.65-3.39 (m, 9H),
3.21 (dd, J = 13.8, 4.0 Hz, 1H), 2.85 (dd, J = 13.8, 11.0 Hz, 1H),
1.25 (t, J = 7.1 Hz, 3H) ), 13C NMR (50 MHz, CDCl3) δ 196.29,
166.04, 155.46, 135.33, 134.75, 129.26, 129.13, 128.38, 128.04,
127.79, 127.56, 101.16, 83.09, 80.20, 63.39, 61.93, 52.74, 45.32,
43.82, 43.50, 42.25, 36.73, 14.77. MS (ESI) m/z (%): 532.34 [(M
+ Na)+, 100]. Anal. Calcd for C27H31N3O5S (509.62): C, 63.63; H,
6.13; N, 8.25. Found: C, 63.89; H, 6.23; N, 8.10.
4.9. (1S.4R,5S,7S)-3.4-dibenzyl-7-(morpholin-4-ylcarbonyl)-6,8-
dioxa-3-azabicyclo[3.2.1]octane-2-thione (6)
4.13. Enzymatic BACE1 Assay
Obtained from 3 using morpholine as the amine. Purification
by flash chromatography eluiting with 1:1 hexane-ethyl acetate
(Rf 0.47) to give the pure product 6 as a colorless oil with a yield
of 69%. [α]D23 = +77.1 (c 1.0, CHCl3). 1H NMR (200 MHz,
CDCl3) δ 7.58 – 7.26 (m, 8H), 7.06 ( d, J=7.5 Hz, 2H), 6.12 (d, J
= 15.1 Hz, 1H), 5.63 (s, 1H), 5.50 (s, 1H), 4.94(s, 1H), 4.53 (d, J
= 15.1 Hz, 1H), 3.74-3.43 ( m, 9H), 3.23 (dd, J = 13.8, 3.9 Hz,
1H), 2.87 (dd, J = 13.6, 11.1 Hz, 1H), 13C NMR (50 MHz,
CDCl3) δ 196.29, 165.93, 135.32, 134.73, 129.27, 129.22,
128.32, 127.74, 127.51, 101.12, 83.10, 80.07, 66.85, 66.60,
63.37, 52.69, 46.01, 42.67, 36.6. MS (ESI) m/z (%): 461.27 [(M
+ Na)+, 100]. Anal. Calcd for C24H26N2O4S (438.54): C, 65.73; H,
5.98; N, 6.39. Found: C, 65.97; H, 6.06; N, 6.18.
Primary BACE1 enzymatic activity was assessed using a
fluorometric assay kit (SensoLyte 520 β-Secretase Assay Kit,
Anaspec, USA) employing the fluorescent substrate HiLyte
Fluor™ 488-Glu-Val-Asn-Leu-Asp-Ala-Glu-Phe-Lys (QXL™
520)-OH. All the measurements were performed in 96-well plates
with a BMG labtech OptimaStar microplate reader. The
inhibitors (10 µM) were pre-incubated with the BACE1 enzyme
(1 µg/mL) for 10 minutes at 25 °C before starting the reaction by
substrate addition. The fluorescence was monitored over 30
minutes (λex =490 nm, λem=520 nm) at 25 °C. The percentages of
inhibition for the test compounds were determined through the
equation (1-Vs/Vo)∗100, where Vs is the initial velocity in the
presence of the inhibitor and Vo is the initial velocity of the
unhibited reaction. The IC50 values were obtained by dose-
response measurements using 0.003-30 µM as the range of
inhibitor concentrations. All the experiments were performed in
triplicate and the collected data were analyzed using Graphpad
4.0 software package (Graphpad Prism, San Diego, CA).
4.10. Ethyl 4-{[(1R,5R,7R)-3-benzyl-2-thioxo-6,8-dioxa-3-
azabicyclo[3.2.1]oct-7-yl]carbonyl}piperazine-1-carboxylate (7)
Obtained from 1 using ethyl piperazine-1-carboxylate.
Purification by flash chromatography eluiting with 1:1 hexane-
ethyl acetate (Rf 0.23) to give the pure product 7 as a yellow
solid with a yield of 61%. [α]D23 = -41.5 (c 1.0, CHCl3). 1H NMR
(400 MHz, CDCl3) δ 7.44-7.18 (m, 5H), 5.92 (s, 1H), 5.50 (s,
1H), 5.22 (d, J = 14.5 Hz, 1H), 5.04 (d, J = 14.5 Hz, 1H), 4.97 (s,
1H), 4.15 (q, J = 7.1 Hz, 2H), 3.72-3.42 (m, 9H), 3.24 (d, J =
13.5 Hz, 1H), 1.26 (t, J = 7.2 Hz, 3H) 13C NMR (50 MHz,
CDCl3) δ 195.93, 165.84, 155.34, 134.05, 129.07, 128.36,
128.12, 99.89, 82.54, 79.37, 77.79, 77.16, 76.52, 61.80, 54.85,
53.59, 45.24, 43.70, 43.38, 42.14, 14.68. MS (ESI) m/z (%):
442.31 [(M + Na)+, 100]. Anal. Calcd for C20H25N3O5S (419.49):
C, 57.26; H, 6.01; N, 10.01. Found: C, 58.22; H, 6.10; N, 9.90.
4.14. Tridimensional structural alignment
The Swiss PDB viewer (SPDBV) program (4.0.1 version,
Swiss Institute of Bioinformatics)19,20 was used to superimpose
the three dimensional (3D) structures of HIV-1 aspartic protease
(PDB code: 4HLA) and Secreted aspartic protease 2 from C.
albicans (PDB code: 1EAG) with BACE1 (PDB code: 4J0P).
Alpha-carbon atoms of the highly conserved DTG motif (e.g.
Asp93, Thr94 and Gly95 for BACE1) were initially
superimposed using the “fit molecule from selection” option.
Then, using the “improve fit” option, SPDBV was asked to
minimize the root-mean square distance (RMSD) between the
corresponding atoms using a least square algorithm. The
calculation was extended to neighbors until the maximum
number of aligned atoms with the lowest RMSD was obtained.
4.11. Ethyl 4-{[(1R.4S,5R,7R)-3.4-dibenzyl-2-thioxo-6,8-dioxa-3-
azabicyclo[3.2.1]oct-7-yl]carbonyl}piperazine-1-carboxylate (8)
Obtained from 2 using ethyl piperazine-1-
carboxylate.Purification by flash chromatography eluiting with
1:1 hexane-ethyl acetate (Rf 0.36) to give the pure product 8 as a
white solid with a yield of 65%. [α]D23 = -51.0 (c 1.0, CHCl3). 1H
NMR (400 MHz, CDCl3) δ 7.45-7.17 (m, 8H), 7.06 (d, J = 6.9
Hz, 2H), 6.12 (d, J = 15.1 Hz, 1H), 5.62 (s, 1H), 5.51 (s, 1H),
4.95 (d, J = 5.2 Hz, 1H), 4.16 (q, J = 7.0 Hz, 2H), 3.71-3.30 (m,
9H), 3.23 (dd, J = 13.8, 3.9 Hz, 1H), 2.87 (dd, J = 13.7, 11.1 Hz,
1H), 1.27 (t, J = 7.0 Hz, 3H), 13C NMR (50 MHz, CDCl3) δ
196.29, 166.04, 155.46, 135.33, 134.75, 129.26, 129.13, 128.38,
128.04, 127.79, 127.56, 101.16, 83.09, 80.20, 63.39, 61.93,
52.74, 45.32, 43.82, 43.50, 42.25, 36.73, 14.77. MS (ESI) m/z
(%): 532.03 [(M + Na)+, 100]. Anal. Calcd for C27H31N3O5S
(509.62): C, 63.63; H, 6.13; N, 8.25. Found: C, 64.01; H, 6.28;
N, 8.01.
4.15. Pairwise sequence alignment
Primary sequences of BACE1 (PDB: 4J0P), HIV-1 (PDB:
4HLA) and SAP2 (PDB: 1EAG) aspartic proteases were
retrieved from the PDB in FASTA format. EMBOSS matcher,
EMBL-EBI,21 was used to carry out local alignments of the two
proteins, by identifying local similarities in the two protein
sequence using an algorithm based on Bill Pearson’s Lalign
application, version 2.0u4 (Feb. 1996). EBLOSUM-62
substitution matrix was used and penalties of 14 and 4 were
applied for gap opening and extension, respectively.
4.16. Docking Calculations
Automated docking studies were carried out using the
Lamarckian Genetic Algorithm (LGA) as a search engine
implemented in the Autodock 4.0.1 program.25 The
4.12. Ethyl 4-{[(1S.4R,5S,7S)-3.4-dibenzyl-2-thioxo-6,8-dioxa-3-
azabicyclo[3.2.1]oct-7-yl]carbonyl}piperazine-1-carboxylate (9)
AutoDockTools 1.4.5 (ADT) graphical interface27 was used to
prepare BACE1 and inhibitor PDBQT files. Coordinates of
compound 4 were generated using Spartan (version 5.147), and
Obtained from 3 using ethyl piperazine-1-carboxylate.
Purification by flash chromatography eluiting with 1:1 hexane-
ethyl acetate (Rf 0.36) to give the pure product 9 as a white solid

11. Mandal, M.; Wu, Y.; Misiaszek, J.; Li, G.; Buevich, A.; Caldwell,
J.P.; Liu, X.; Mazzola, R.D.; Orth, P.; Strickland, C.; Voigt, J.;
Wang, H.; Zhu, Z.; Chen, X.; Grzelak, M.; Hyde, L.A.; Kuvelkar,
R.; Leach, P.T.; Terracina, G.; Zhang, L.; Zhang, Q.; Michener,
M.S.; Smith, B.; Cox, K.; Grotz, D.; Favreau, L.; Mitra, K.;
Kazakevich, I.; McKittrick, B.A.; Greenlee, W.; Kennedy, M.E.;
Parker, E.M.; Cumming, J.N.; Stamford, A.W. J. Med. Chem.
2016, 59, 3231-3248 and ref. therein.
then energy-minimized through the AM1 semi-empirical method
to calculate the equilibrium geometry. The coordinates of
BACE1 protease were retrieved from the Protein Data Bank
(PDB code: 4J0P), and enzyme-inhibitor complex was unmerged
for achieving the free enzyme structure. Water molecules were
removed. Hydrogen atoms were added to the enzyme and
ligands, Gasteiger charges were added and non-polar hydrogens
were merged. Three-dimensional energy scoring grids of 0.375 Å
resolution and 40 Å×40 Å×40 Å dimensions were computed. The
center of the grid was set using the coordinated of one of the
oxygen atoms of D289. A total of 50 runs with a maximum of
2500000 energy evaluations were carried out for each ligand,
using the default parameters for LGA. Cluster analysis was
performed on docked results using a root-mean-square (rms)
tolerance of 1.5 Å. The analysis of the binding mode of the
docked conformations was carried out using the Autodock plugin
within PyMol software v0.99.28
12. Guarna, A; Guidi, A.; Machetti, F.; Menchi, G.; Occhiato, E.G.;
Scarpi, D.; Sisi, S.; Trabocchi, A. J. Org. Chem. 1999, 64, 7347-
7364.
13. Trabocchi, A.; Menchi, G.; Guarna, F.; Machetti, F.; Scarpi, D.;
Guarna, A. Synlett 2006, 331-353.
14. Trabocchi, A.; Mannino, C.; Machetti, F.; De Bernardis, F.;
Arancia, S.; Cauda, R.; Cassone, A., Guarna, A. J. Med. Chem.
2010, 53, 2502-2509.
15. Calugi, C.; Trabocchi, A.; De Bernardis, F.; Arancia, S.; Navarra,
P.; Cauda, R.; Cassone, A.; Guarna, A. Bioorg. Med. Chem. 2012,
20, 7206-7213.
16. De Bernardis, F.; Arancia, S.; Tringali, G.; Greco, M. C.;
Ragazzoni, E.; Calugi, C.; Trabocchi, A.; Sandini, S.; Graziani, S.;
Cauda, R.; Cassone, A.; Guarna, A.; Navarra, P. J. Pharm.
Pharmacol. 2014, 66, 1094-1101.
Acknowledgments
17. Calugi, C.; Guarna, A.; Trabocchi, A. Eur. J. Med. Chem. 2014,
84, 444-453.
Fondazione Cassa di Risparmio di Firenze is acknowledged
for the CRF grant (2015.0935) entitled ‘Sviluppo di nuovi
inibitori peptidomimetici di beta-secretasi 1 (BACE-1) per il
monitoraggio e trattamento di placche amiloidi nel morbo di
Alzheimer’.
18. Hong, L.; Koelsch, G.; Lin, X.; Wu, S.; Terzyan, S.; Ghosh, A.K.;
Zhang, X.C.; Tang, J. Science 2000, 290, 150-153.
19. Guex, N.; Peitsch, M.C. Electrophoresis 1997, 18, 2714-2723.
20. ExPAsy Proteomics Server. http://www.expasy.org
21. Calugi, C.; Guarna, A.; Trabocchi, A. J. Enzyme Inhib. Med.
Chem. 2013, 28, 936-943.
22. EMBOSS Matcher - Pairwise Sequence Alignment from the
EMBL-EBI tools website.
References and notes
http://www.ebi.ac.uk/Tools/psa/emboss_matcher/
23. Pederson, B.S.; Sheibye, S.; Nilsson, N.H.; Lawesson, S.-O. Bull.
Soc. Chim. Belg. 1978, 87, 223-228.
24. Machetti, F.; Bucelli, I.; Indiani, G.; Kappe, C.O.; Guarna, A. J.
Comb. Chem. 2007, 9, 454-461.
25. Morris, G.M.; Goodsell, D.S.; Halliday, R.S.; Huey, R.; Hart,
W.E.; Belew, R.K.; Olson, A.J. J. Comput. Chem. 1998, 19, 1639-
1662.
26. See, for example: Rombouts, F.J.; Tresadern, G.; Delgado, O.;
Martínez-Lamenca, C.; Van Gool, M.; García-Molina, A.; Alonso
de Diego, S.A.; Oehlrich, D.; Prokopcova, H.; Alonso, J.M.;
Austin, N.; Borghys, H.; Van Brandt, S.; Surkyn, M.; De Cleyn,
M.; Vos, A.; Alexander, R.; Macdonald, G.; Moechars, D.; Gijsen,
H.; Trabanco, A.A. J. Med. Chem. 2015, 58, 8216-8235.
27. Gillet, A.; Sanner, M.; Stoffler, D.; Olson, A. Structure 2005, 13,
483-491.
1. WHO (2012). Dementia, a Public Health Priority. Geneva: World
Health Organization.
2. Wilmo, A.P.; Martin, P. World Alzheimer Report 2010, The
Global Economic Impact of Dementia. London: Alzheimer’s
Disease International, 2010.
3. Selkoe, D.J. Physiol. Rev. 2001, 81, 741-766
4. Lee, V.M.; Balin, B.J.; Otvos, Jr, L.; Trojanowski, J.Q. Science
1991, 251, 675-678.
5. Vassar, R.; Kovacs, D.M.; Yan, R.; Wong, P.C. J. Neurosci. 2009,
29, 12787-12794.
6. Sisodia, S.S.; St George-Hyslop, P.H. Nat. Rev. Neurosci. 2002, 3,
281-290.
7. Luo, Y.; Bolon, B.; Kahn, S.; Bennett, B.D.; Babu-Khan, S.;
Denis, P.; Fan, W.; Kha, H.; Zhang, J.; Gong, Y.; Martin, L.;
Louis, J.C.; Yan, Q.; Richards, W.G.; Citron, M.;Vassar, R. Nat.
Neurosci. 2001, 4, 231-232.
28. DeLano, W.L. ‘The PyMOL Molecular Graphics System.’
8. Ghosh, A.K.; Osswald, H.L. Chem. Soc. Rev. 2014, 43, 6765-6813
9. John, V. Curr. Top. Med. Chem. 2006, 6, 569-578.
10. Butler, C.R.; Ogilvie, K.; Martinez-Alsina, L.; Barreiro, G.; Beck,
E.M.; Nolan, C.E.; Atchison, K.; Benvenuti, E.; Buzon, L.; Doran,
S.; Gonzales, C.; Helal, C.J.; Hou, X.; Hsu, M.H.; Johnson, E.F.;
Lapham, K.; Lanyon, L.; Parris, K.; O'Neill, B.T.; Riddell, D.;
Robshaw, A.; Vajdos, F.; Brodney, M.A. J. Med. Chem. 2017, 60,
386-402.
DeLano Scientific LLC, San Carlos, CA, USA
[http://www.pymol.org].
Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016