Peptidomimetic β-Secretase Inhibitors Comprising a Sequence of Amyloid-β Peptide for Alzheimer's Disease

Article abstract of DOI:10.1021/acs.jmedchem.5b00658

Alzheimer's disease is a grave social problem in an aging population. A major problem is the passage of drugs through the blood-brain barrier. This work tests the hypothesis that the conjugation of peptidomimetic β-secretase inhibitors with a fragment of amyloid-β peptide facilitates entrance into the central nervous system. HVR-3 (compound 4), one of the conjugation products, was found to be as potent as OM00-3, a known peptidomimetic inhibitor, 4-fold more selective toward β-secretase 1 in relation to β-secretase 2 and 3-fold more resistant to in vitro metabolization in human serum. Its intravenous administration to mice and Wistar rats generated an active metabolite recovered from the rodent's brains.

Full text of DOI:10.1021/acs.jmedchem.5b00658

Subscriber access provided by UNIV OF MISSISSIPPI
Article
Peptidomimetic #-secretase inhibitors comprising a
sequence of Amyloid-# peptide for Alzheimer's disease
Helder Vila Real, Helena Coelho, João Rocha, Adelaide Fernandes, Maria Rita
Ventura, Christopher David Maycock, Olga Iranzo, and Ana Luísa Simplício
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b00658 • Publication Date (Web): 10 Jun 2015
Downloaded from http://pubs.acs.org on June 22, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
Peptidomimetic βꢀsecretase inhibitors comprising a sequence
of Amyloidꢀβ peptide for Alzheimer's disease
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
4
Helder VilaꢀReal 1,2, Helena Coelho 1,2,3, João Rocha , Adelaide Fernandes , M. Rita
Ventura 2, Christopher D. Maycock 2,5, Olga Iranzo 2,6, Ana L. Simplício 1,2 *
1 Instituto de Biologia Experimental e Tecnológica, Av. da República, Estação Agronómica
Nacional, 2780ꢀ157 Oeiras, Portugal
2
Instituto de Tecnologia Química e Biológica, Av. da República, Estação Agronómica
Nacional, 2780ꢀ157 Oeiras, Portugal
3
Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa, Quinta da Torre,
Campus Universitário, 2829ꢀ516 Caparica, Portugal.
4
Faculdade de Farmácia, Universidade de Lisboa, Av. Prof. Gama Pinto, 1649ꢀ003
Lisboa, Portugal.
5
Faculdade de Ciências, Universidade de Lisboa, Campo Grande, 1749ꢀ016 Lisboa,
Portugal.
6
Present address: Institut des Sciences Moléculaires de Marseille, Aix Marseille
Université, Centrale Marseille, iSm2 UMR 7313, 13397 Marseille, France.
*Corresponding author: Ana Simplício; Eꢀmail: anas@itqb.unl.pt; Telephone: +351
214469782; Address: Av. da República, Estação Agronómica Nacional 2780ꢀ157 Oeiras,
Portugal.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 39
1
2
3
4
5
Abstract
6
7
8
9
Alzheimer’s disease is a grave social problem in an ageing population. A major problem is
the passage of drugs through the bloodꢀbrain barrier. This work tests the hypothesis that
the conjugation of peptidomimetic βꢀsecretase inhibitors with a fragment of Amyloidꢀβ
peptide facilitates entrance into the Central Nervous System.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HVRꢀ3 (compound 4), one of the conjugation products, was found to be as potent as
OM00ꢀ3, a known peptidomimetic inhibitor, fourꢀfold more selective towards βꢀsecretase 1
in relation to βꢀsecretase 2 and threeꢀfold more resistant to in vitro metabolization in
Human serum. Its intravenous administration to mice and Wistar rats generated an active
metabolite recovered from the rodent’s brains.
Keywords: Amyloidꢀβ peptide, Alzheimer's disease, bloodꢀbrain barrier, Central Nervous
System, HVRꢀ3, OM00ꢀ3, peptidomimetic βꢀsecretase inhibitor.
2
ACS Paragon Plus Environment

Page 3 of 39
Journal of Medicinal Chemistry
1
2
3
Introduction
4
5
6
7
8
9
Alzheimer’s disease (AD), a progressive neurodegenerative disease of the Central
Nervous System (CNS), is generally characterized by a slow and inflexible progression of
dementia associated with cognitive and memory decline, speech loss and personality
changes. Its hallmark pathological lesions occur in the brain memory and cognition regions,
consisting on neurofibrillary tangles and amyloid plaques.1 These plaques are composed of
Amyloidꢀβ peptide (Aβ), a highly insoluble peptide with tendency to oligomerize and
aggregate, which is produced from the sequential cleavage of the Amyloidꢀβ precursor
protein (APP) by βꢀsecretase 1 (BACEꢀ1) and γꢀsecretase.2 In addition, neurofibrillary
tangles, which are formed by the accumulation of abnormal filaments of tau protein, have
been reported to be induced by Aβ, trough a mechanism of hyperphosphorylation, an
initiating step for tangles occurrence.3 Considering this, agents capable of reducing the
production of Aβ, such as BACEꢀ1 inhibitors, may block or reduce the occurrence of
amyloid plaques and neurofibrillary tangles and consequently preventing AD progression.4
The first reported BACEꢀ1 inhibitors were peptidomimetics composed of a transitionꢀstate
analogue at the scissile peptide bond.5ꢀ7 The Tang and Ghosh group developed compound
1 (OM00ꢀ3) (Figure 1),6 one of the yet most potent BACEꢀ1 inhibitors reported and only
composed of natural amino acids plus a synthetic hydroxyethylene LeucineꢀAlanine
dipeptide transitionꢀstate isostere. Compound 1 is an optimized compound for the inhibition
of BACEꢀ1.8 Despite the high potency shown in vitro towards BACEꢀ1, this kind of inhibitor
has however been shown to be ineffective in the treatment of AD due to the lack of the
required pharmacokinetic properties, particularly the ability to cross the bloodꢀbrain barrier
(BBB). Large size and high hydrophilicity were considered the two main issues preventing
passive BBB crossing and therefore attention was pointed towards the development of
nonꢀpeptidic inhibitors,9,10 which being smaller and more hydrophobic, have better
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 39
1
2
3
4
5
6
7
8
9
conditions to permeate the BBB. Unfortunately, their therapeutic potential in the CNS has
been limited by Pꢀglycoprotein efflux. Also, smaller size contributes to nonspecificity and
therefore to increased toxicity, contrary to peptides that are generally considered to be
safe. Currently there is no effective treatment capable of modifying the progression of AD
and available drugs only act on symptomatic improvement. The development of drugs
capable of blocking or delaying the disease progression remains a challenge due to the
inability of BACEꢀ1 inhibitors to cross the BBB.10
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In this work we report a new approach to overcoming the low permeability of
peptidomimetic BACEꢀ1 inhibitors across the BBB. This approach is based on the
conjugation of a carrier peptide with a peptidmometic BACEꢀ1 inhibitor in order to facilitate
its transcytosis across the BBB.11 Compound 1 will be used as a starting compound of the
8
rational design, since it was reported to be a high potent peptidomimetic inhibitor , but
12
simultaneously incapable to penetrate cells and to cross the plasma membrane , being
12
widely accepted to be too large to transverse the BBB . The transcytosis route has the
advantage of not limiting the compounds size for BBB permeation,12 and has been targeted
for example through conjugation of peptidomimetic inhibitors with a fragment of Tat protein
from HIVꢀ1,12 as well as with the Antennapedia peptide, penetratin.13
The innovation reported in the present work is the use of, Aβ 18–23', as a carrier peptide
for BACEꢀ1 inhibitors through the BBB. The potential usefulness of this strategy is based
on the property of Aβ to recognize the Receptor for Advanced Glycation Endproducts
(RAGE) enabling entrance into the CNS across the BBB, through receptor mediatedꢀ
transcytosis.14,15 Aβ 18–23' is a hexapeptide sequence corresponding to the binding region
of Aβ to RAGE.16 Factors that may contribute for the success of using this carrier peptide in
AD include the following: (a) RAGE is overexpressed in AD both in the BBB and neurons,17
potentially enabling a higher BBB crossing efficiency and internalization within the
4
ACS Paragon Plus Environment

Page 5 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
endossomes of neurons, where the drug will act, through RAGEꢀmediated endocytosis as
reported for Aβ;18 (b) Aβ 18–23' acts as a competitive inhibitor of the entry of Aβ peptide
into cells expressing RAGE,16 which potentially helps to avoid reꢀentry of Aβ into the CNS,
as supported by the vascular theory for the development of AD;19 (c) RAGE has
advantages in comparison with more ubiquitous receptors such as transferrin receptor
(TfR) or lipoprotein receptorꢀrelated protein (LRP) in which, competition with natural ligands
may contribute to adverse effects.20 In the present study we specifically tested the
hypothesis of conjugating compound 1,6 with Aβ 18–23',16 in order to target the CNS, thus
leading to a potential drug for AD treatment.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
* RAGE ꢀ Receptor for Advanced Glycation Endproducts
TfR ꢀ transferrin receptor
LRP ꢀ lipoprotein receptorꢀrelated protein
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 39
1
2
3
Results and discussion
4
5
6
7
8
9
The rational design of peptidomimetic BACEꢀ1 inhibitors was based on the conjugation of 1
with Aβ 18–23' and on the results obtained from several tests including: BACEꢀ1 inhibition,
selectivity in relation to BACEꢀ2, metabolization and toxicological profile and also a
pharmacokinetic study in order to evaluate distribution into the brain.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Influence of conjugation with Aβ 18–23' on BACEꢀ1 inhibition potency
The strategy used to design compound 2 (HVRꢀ1) (Table 1) is outlined in Figure 1 and
involved truncation of the three residue sequence at the Cꢀterminal of 1 before conjugation
with Aβ 18–23', in order to obtain the shortest sequence possible. This involved minimal
disturbance to the original structure since the three residue sequence at the Nꢀterminal of
Aβ 18–23' only differs by one amino acid from the truncated sequence. This lead to a
peptide of eleven residues, while the mere addition of Aβ 18–23' to 1 would give a
sequence of fourteen residues. Therefore, the glutamic acid in position P3' of 1 becomes a
phenylalanine in position 7 of 2. This change should not affect binding to BACEꢀ1 since the
preference indexes of both residues in BACEꢀ1 substrates are identical,8 but in order to
evaluate to what extent the conjugation of 1 with Aβ 18–23' affects the inhibition of BACEꢀ
1, compound 2 was tested for BACEꢀ1 inhibition.
As illustrated in Figure 2, conjugation affects the potency of the inhibitor to only a very
small extent (Ki (compound 2) = 2.0 ± 0.1 nM vs. Ki (compound 1) = 1.2 ± 0.4 nM, mean value ±
standard deviation, n = 3, p ≤ 0.05, independent twoꢀtailed Student’s tꢀtest). The value
reported here for the Ki of 1 is fourꢀfold higher than the previously reported value of 0.31
nM.8 This is due to the fact that a permanently active form of BACEꢀ1 was used before,
while in the present study a proꢀBACEꢀ1 was employed possessing a pro domain in
equilibrium with an opened and closed form, leading to a reduction of the enzyme activity.21
6
ACS Paragon Plus Environment

Page 7 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
For a similar inhibitor, OM99ꢀ2, a sevenꢀfold potency decrease was observed when proꢀ
BACEꢀ1 (Ki = 9.8 nM) was used instead of active BACEꢀ1 (Ki = 1.4 nM).21,22 Moreover, the
potency decrease observed for 2 (about twoꢀfold) in comparison with 1, is less than that
reported for the conjugation of OM99ꢀ2 with a carrier peptide of the Tat protein of HIVꢀ1 or
when 1 was conjugated with a nonapeptide of arginines, where the inhibition potency over
the active form of BACEꢀ1 decreased twenty eight (39 nM) and sixꢀfold (1.7 nM),
respectively, when compared to the reported value of 0.31 nM.8
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Effect of the exchange of a phenylalanine by a tyrosine at position 8
BACEꢀ2, the closest homologous enzyme to BACEꢀ1, has received great attention by
researchers due to the need to obtain a BACEꢀ1 selective drug to avoid possible
secondary effects such as diabetes.23 The disease may arise from BACEꢀ2 inhibition
during the production of insulin by pancreatic βꢀcells.24 In addition, BACEꢀ2 contributes to
the degradation of Aβ,25 meaning that its inhibition may lead to the progression of AD.
The results reported for the preference of amino acids at the eight subsites of BACEꢀ1 and
BACEꢀ2 substrates reveal that the preference index for phenylalanine at position P4'
(Figure 1) in the case of BACEꢀ2 is very high when compared to the null preference index
for tyrosine,23 while it is similar to the one for tyrosine in the case of BACEꢀ1.8 This means
that an exchange of a phenylalanine by a tyrosine in 2 should have little effect on the
inhibition of BACEꢀ1 while significantly reduce affinity for BACEꢀ2. Also, this change should
have minimal effect on RAGE binding since tyrosine, like phenylalanine, is a hydrophobic
aromatic amino acid, which is a requisite for RAGE binding. Gospodarska et al. reported
that the Aβ 18–23’ sequence involving a hydrophobic stretch, flanked by two negatively
charged residues at the Cꢀterminal is crucial for RAGE recognition.16 Based on these facts,
we hypothesised that the exchange of the phenylalanine for a tyrosine at position P4',
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 39
1
2
3
4
5
corresponding to position 8 of 2, affording compound 3 (HVRꢀ2) (Table 1) should increase
the selectivity for BACEꢀ1 in relation to BACEꢀ2.
6
7
8
9
As outlined in Figure 3, compound 3 shows a ratio between Kiapp BACEꢀ2 and Ki BACEꢀ1 of 7.5,
which is higher than that of 2 (3.7) and of 1 (2.5). The design of 3 lead therefore to an
inhibitor twoꢀfold more selective than 2, which in its turn was already more selective than 1,
indicating that conjugation with the carrier peptide also contributes to increase the
selectivity for BACEꢀ1 in relation to BACEꢀ2.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Influence of Nꢀterminal acylation on the metabolization profile
NꢀAcylation and Cꢀamidation are regularly used in nature to obtain endꢀprotected
neuropeptides with improved stability and activity.26 Small Lꢀamino acid peptides with free
Nꢀ and Cꢀterminals suffer fast metabolization by exopeptidases, sometimes within a few
minutes as in the case of somatostatin, but its halfꢀlife was improved from 3 to 400 min
after Nꢀterminal acylation.27 Also, Nꢀterminal acylation may improve the cell penetration
and passage though barriers such as the BBB due to increased lipophilicity.26 Considering
this in order to increase the inhibitor resistance against metabolization, 3 was acetylated at
the Nꢀterminal rendering compound 4 (HVRꢀ3) (Table 1).
A comparative metabolic stability study of 1, 3 and 4 was carried out through in vitro
incubation in Human serum and mice brain homogenate. Halfꢀlives were determined from
the exponential metabolization profiles shown in Figure 4. Nꢀacylation contributed
significantly to stability as the halfꢀlives of 4 increased threeꢀfold in serum and twoꢀfold in
brain homogenate, compared to those of 3. The carrier peptide Aβ 18–23' caused however
some detrimental effect in the stability of 1 in brain homogenate since 1 is twoꢀfold more
stable than 3. The metabolization halfꢀlives of 4 in serum (6.0 hours) and brain
homogenate (3.4 hours) indicate potential for its use as a drug. Some previously reported
8
ACS Paragon Plus Environment

Page 9 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
peptides tested for the treatment of Alzheimer's disease only last a few minutes in mice
brain homogenate even after Nꢀterminal protection.28
6
7
8
9
The potency of 4 over BACEꢀ1 is not statistically different from that of 1 having a Ki of 2.0 ±
0.6 (mean value ± standard deviation, n = 3, p > 0.05, independent twoꢀtailed Student’s tꢀ
test). Finally, among the studied compounds, 4 was also the most selective for BACEꢀ1 in
relation to BACEꢀ2 showing a ratio of 9.2 between Kiapp BACEꢀ2 and Ki BACEꢀ1 (Figure 3), being
fourꢀfold more selective than 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Evaluation of the cytotoxic potential in Cacoꢀ2 cells
Cacoꢀ2 cells were used as a preliminary assay to study compounds toxicity.29 The
cytotoxicity assay evaluated the dependence of Cacoꢀ2 cell viability with the inhibitor
concentration, of compounds: 1, 2, 3 and 4 during incubation periods of 4 and 24 hours.
Cellular viability was maintained above 80% within a range of 25 nM to 50 ꢀM for all
inhibitors, even after 24 hours demonstrating that the compounds are not toxic up to a
concentration magnitude that exceeds their inhibition constant of more than 25 thousand
fold.
In vivo pharmacokinetic study for evaluation of compound 4 delivery into brain
An in vivo pharmacokinetic study was performed in order to evaluate the capacity of the
best inhibitor in vitro, compound 4, to reach the brain in mice. Both serum and brain
samples were analyzed after intravenous administration of 4. In serum samples, compound
4 was not detected by HPLCꢀFLU half an hour after intravenous administration. This
indicates that 4 is rapidly distributed, metabolized or eliminated in mice. Accordingly,
compound 4 with a peak of 1326.5 amu and a retention time of 23.6 minutes could not be
detected in brain samples; however, a metabolite with a peak of 1285.4 amu (Figure 5) and
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 39
1
2
3
4
5
6
7
8
9
a very close retention time of 22.9 minutes was detected in brain samples by HPLCꢀFLU
and LCꢀMS even two hours after administration. The fact that this metabolite doesn't
appear in serum samples dismisses any potential contamination of brain samples by the
blood.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In order to identify the metabolite and according to the mass found it was hypothesized that
it could be compound 5 (HVRꢀ4) (table 1) as the molecular ion found (1285.4 amu) is
consistent with the calculated exact mass for 5 [M+H]+ (1285.6 amu). The MS2
fragmentation of the molecular ion: 1285.4 amu, afforded the fragmentation spectrum
shown in Figure 5 and the main m/z peaks were assigned based on the expected ions
given by mMass software for compound 5,30: MꢀH2O (expected = 1267.6, found = 1267.4),
Mꢀ2H2O (expected = 1249.6, found = 1249.4), b10 (expected = 1152.5, found = 1152.5),
z10ꢀH2O (expected = 1121.5, found = 1121.7), b9 (expected = 1023.5, found = 1023.3), b8
(expected = 952.4, found = 952.4), a8 (expected = 924.5, found = 924.4), b7 (expected =
789.3, found = 789.4), a7ꢀH2O (expected = 743.4, found = 743.2), b6 (expected = 642.4,
found =642.4), b6ꢀH2O (expected = 624.4, found = 624.3), z5ꢀH2O (expected = 610.2,
found = 610.2) and b5 (expected = 543.3, found = 543.2). The correspondence between
the observed m/z peaks (Figure 5) with the expected ions given by mMass suggests
compound 5 as the metabolite found in the brain samples.
In order to further confirm the chemical structure of the metabolite as 5, compound 3 was
hydrolyzed in order to generate 5 by chemical deamination of the Cꢀterminal. The Cꢀ
terminal amide of 3 is the only primary amide present making it the most prone to suffer
hydrolysis under moderate basic conditions (Figure 6). The hydrolysis product was
analyzed by HPLCꢀFLU and LCꢀMS, having the some retention time of 22.9 min (Figure 7)
as the metabolite(Figure 5) and the major molecular ion found (1285.6 amu) (Figure 8) is
consistent with the calculated exact mass for 5 [M+H]+ (1285.6 amu). A final confirmation
10
ACS Paragon Plus Environment

Page 11 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
of the metabolite structure was provided by the correspondence of the MS2 fragmentation
spectrum of the metabolite (Figure 5) with that of the product of hydrolysis (Figure 8),
where all the main peaks identified in the product of hydrolysis (Figure 8) were also
identified in the fragmentation spectrum of the metabolite, even including peak b10ꢀH20
(1134.6) not highlighted in Figure 5 but still present.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In a parallel experiment used to detect 4 by MALDIꢀTOF/TOF in several brain regions of
Wistar rats, 4 also suffered biotransformation into 5. In agreement with previous
experiments in mice, 4 was not detected in the rat cortex after two hours, but 5 was. On the
other hand, both 4 and 5 were detected in serum samples of Wistar rats after two hours,
raising the hypothesis that disposition of 4 is faster in mice than in Wistar rats.
During the biotransformation of 4 into 5, the peptide suffered hydrolytic reactions at both
the Nꢀ and Cꢀterminal amides. Despite being generally considered robust amine
derivatives,31 amides do suffer metabolization. As shown on Figure 4, 4 is in fact twoꢀfold
more stable than 5 in Human serum, possibly because the latter is more susceptible to the
attack of serum exopeptidases. On the other hand the opposite happened in mice brain
homogenate (Figure 4), where 5 is more stable than 4. In the brain the action of amidases
acting on terminal amides seem to be more relevant than exopeptidases. Primary amides,
such as the Cꢀterminal amide of 4, are easily converted into the corresponding acid by
amidases,32 which are nonspecific hydrolytic enzymes mainly found in the liver.33
Amidases may also be found in other organs including kidney and brain, where Nꢀ
deacylation of acetophenetidines has been observed.34
The modifications suffered by 4 when converted to 5 are slight, and resulted in an active
metabolite as potent as 4 (1.7 ± 0.2 nM, mean value ± standard deviation, n = 3, p > 0.05,
independent twoꢀtailed Student’s tꢀtest). The high potency of 5 was expected since it has a
very similar structure to 3 (the only difference between 5 and 3 is in the Cꢀterminal,
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 39
1
2
3
4
5
6
7
8
9
possessing a carboxylic acid instead of an amide) that is as potent as 4. The benefits of
having an amide at the Cꢀterminal is dual, namely the increase in the resistance against
metabolization in serum and the absence of a negative charge that could disturb RAGE
binding and consequently, BBB crossing.16 Since 5 was found in the brain, this means that
either RAGE binding is not affected or that deamidation is slow enough to allow prior
distribution to the brain. Moreover the presence of the carboxylic acid at the Cꢀterminal also
didn’t affect BACEꢀ1 inhibition due to the long distance in relation to the binding site (Figure
1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Looking from a different perspective, the chemical modifications induced by the
biotransformation of 4 can be even more important for therapy than the initial compound
itself. The primary amine formed at the Nꢀterminal and the carboxylic acid at the Cꢀterminal
increase the molecule hydrophilicity, by the generation of a positive and a negative charge,
respectively, at physiological pH. These chemical changes will contribute to impede 5
clearance from the CNS into the bloodstream leading to a higher residence time in the
CNS since a more hydrophilic molecule will have more difficulty diffusing across the BBB
into the bloodstream.26 This hypothesis gains even more support by the fact that 5 was
found at approximately 20 nM in mice brain up to one hour after administration, which is
within the therapeutic concentration range, while 4 was not detected after half an hour post
administration.
12
ACS Paragon Plus Environment

Page 13 of 39
Journal of Medicinal Chemistry
1
2
3
Conclusions
4
5
6
7
8
9
The conjugation of 1 with Aβ 18–23' giving compounds 2, 3 and 4 shows little or no effect
on the BACEꢀ1 inhibition potential and at the same time contributes to increased selectivity
in relation to BACEꢀ2, with 4 exhibiting fourꢀfold more selective than 1 for BACEꢀ1 in
relation to BACEꢀ2. Compound 4 was also demonstrated to be more resistant against in
vitro metabolization in Human serum (threeꢀfold).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Preliminary data in Cacoꢀ2 cells indicates that compounds 1, 2, 3 and 4 are potentially non
toxic in concentrations that are therapeutically relevant, which complies with the general
idea of peptides being low toxicity drugs.
Compound 4 was not detected in serum, nor in mice brain after intravenous administration,
but 5, an active metabolite of 4, was found in mice brain and Wistar rat cortex after two
hours post administration. Peptide 4 suffered two biotransformation steps elicited by
amidases consisting of an Nꢀdeacylation and a Cꢀdeamination. This transformation
afforded an active metabolite, 5 having increased hydrophilicity that may contribute to a
larger residence time in the CNS thus increasing its therapeutic index.
We conclude that the design of compound 4, a peptidomimetic BACEꢀ1 inhibitor based on
the conjugation of 1 with Aβ 18–23', was successful in targeting the CNS. Altogether these
results corroborate the potential of 4 to reach CNS as a prodrug for the treatment of
Alzheimer's disease. Future developments include the use of AD mice models to determine
the efficacy of 4 as a prodrug.
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 39
1
2
3
Experimental section
4
5
6
7
Peptide synthesis and purification
8
9
Peptides were synthesized by manual standard Solid Phase Peptide Synthesis procedure
using standard protocols,35 starting from Fmocꢀprotected amino acids (Novabiochem)
activated with N,N,N′,N′ꢀTetramethylꢀOꢀ(1Hꢀbenzotriazolꢀ1ꢀyl)uronium hexafluorophosphate
(HBTU) and 1ꢀhydroxybenzotriazole (HOBt) using N,Nꢀdiisopropylethylamine (DIEA) in
dimethylformamide. A Rink amide resin (MBHA resin [100ꢀ200 mesh], 0.59 mmol/g
substitution, Novabiochem) was used for peptide synthesis yielding a Cꢀterminal amide
upon resin cleavage. Exceptions were 1 and the fluorogenic product (McaꢀSEVNLꢀCOOH)
of the hydrolysis of McaꢀSEVNLDAEFKꢀDNP by BACEꢀ1 for which two Wang resins were
used to generate a Cꢀterminal carboxylic acid instead; these were respectively: FmocꢀPheꢀ
Wang resin ([100ꢀ200 mesh], 0.61 mmol/g substitution, Novabiochem) and FmocꢀLeuꢀ
Wang resin ([100ꢀ200 mesh], 0.64 mmol/g substitution, Novabiochem). Capping of the Nꢀ
terminal was accomplished for 4 (Table 1) with 50 equivalents of acetic anhydride and 25
equivalents of DIEA in dimethylformamide during 1 hour, after Fmoc deprotection of the
last added amino acid with 20% piperidine in dimethylformamide. In the synthesis of Mcaꢀ
SEVNLꢀCOOH, 7ꢀmethoxycoumarin (Mca) (Novabiochem) was added to the Nꢀterminal of
SEVNLꢀCOOH activated by Oꢀ(7ꢀazabenzotriazolꢀ1ꢀyl)ꢀ1,1,3,3ꢀ tetramethylꢀuronium
hexafluorophosphate (HATU) and HOBt in dimethylformamide. The FmocꢀLeu*Ala isostere
was synthesized as previously described,36 and was incorporated into peptides: 1, 2, 3 and
4 using the same standard protocols as for natural amino acids.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Peptides were purified by preparative reverseꢀphase HPLC and lyophylized.
Chromatographic conditions were as follow: column C12 (Jupiter Proteo, Phenomenex)
250 mm x 21.2 mm, 4 ꢀm; eluent (A) water/TFA (99.9:0.1, V/V), (B) acetonitrile/water/TFA
14
ACS Paragon Plus Environment

Page 15 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(90:9.9:0.1, V/V). The flow rate was 10 mL/min and the column was kept at room
temperature. Reꢀanalysis by analytical reverseꢀphase HPLC and Electrospray Mass
Spectometry (ESIꢀMS analysis) is in accordance with a homogenous product of 98% purity
or higher. Molecular structures and molecular masses of the synthetic peptides inhibitors of
BACEꢀ1 are outlined in Table 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Assay for BACEꢀ1 and BACEꢀ2 activities
The kinetic experiments were preformed in 20 mM sodium acetate buffer, pH 4.5 at 37ºC in
15% dimethyl sulfoxide (DMSO), 1% triton Xꢀ100 (SigmaꢀAldrich), 250 nM of Mcaꢀ
SEVNLDAEFKꢀDNP (Bachem) as substrate and 2 U/mL of BACEꢀ1 (human, recombinant,
21267 U/mg, purity > 90%, SigmaꢀAldrich) or 10 U/mL of BACEꢀ2 (human, recombinant,
40000 U/mg, purity > 80%, Enzo). The initial rates of substrate hydrolysis were confirmed
to be directly proportional to the enzyme concentrations in a range of 1 ꢀ 5 U/mL for BACEꢀ
1 and 5 ꢀ 20 U/mL for BACEꢀ2. Initial rates were determined below conversion of 5% of
substrate, after stopping the reaction with 25% of a 2.5 M sodium acetate solution after 2
and 1 hour for BACEꢀ1 and BACEꢀ2, respectively. The hydrolysis of the fluorogenic
substrate was quantified by reverse phase HPLC with fluorescence detection (HPLCꢀFLU)
using an excitation wavelength of 323 nm and an emission wavelength of 382 nm (Merck
Hitachi, Elite LaChrom). A C18 column (Gemini, Phenomenex) 150 mm, 4.60 mm, 5 ꢁm
was used at 30ºC with a linear gradient from (A) water/TFA (99.9:0.1, V/V), to (B)
acetonitrile/water/TFA (90:9.9:0.1, V/V) at 1 mL/min. The program started at 90% A for two
min, graded to 70% A during ten min and finally to 0% A during two min. The retention time
of the reaction product (McaꢀSEVNLꢀCOOH) was 11 minutes and quantification was
performed by external calibration, using a peptide synthesized and purified as in the
Peptide synthesis and purification section.
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 39
1
2
3
4
5
6
Determination of apparent tightꢀbinding inhibition constants
Apparent tightꢀbinding inhibition constants, Kiapp, were determined from the fitting of
experimental doseꢀresponse data to the Morrison equation (Equation 1), where v is the
initial rate observed at each inhibitor concentration, [I], v0 is the initial rate observed without
inhibitor and [E] is the enzyme concentration.37,38 Enzyme as well as substrate
concentrations were kept constant with increasing inhibitor concentration, ranging from
0.04 nM to 283 nM, according to the potency of the inhibitor. The Morrison equation fit to
experimental data was carried out through nonꢀlinear regression by minimising the residual
sum of squares between the experimental data points of the relative activity vs the
logarithm of the inhibitor concentration and those estimated by the model, using Solver
addꢀin from Microsoft Excel 2003, considering the following options: Newton method; 100
iterations, precision of 10−6, 5 % of tolerance and 1x10−4 convergence.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
[E
]
−
[I
]
− Kiapp
+
([
E
]
−
[E]
[
I
]
− Kiapp
)
+ 4[
E
Kiapp
]
v = v0
2
Equation 1
Determination of the inhibition constants for BACEꢀ1
Inhibition constants, Ki, were determined for BACEꢀ1, from the direct application of the
ChengꢀPrusoff equation (Equation 2),39 and using the KM value reported of 4.5 ꢀM for the
McaꢀSEVNLDAEFKꢀDNP substrate.21 For BACEꢀ2, the KM has not yet been reported and
therefore Ki was not determined.
Kiapp
Ki =


[ ]
S
1+



KM

Equation 2
16
ACS Paragon Plus Environment

Page 17 of 39
Journal of Medicinal Chemistry
1
2
3
In vitro metabolic stability assay
4
5
6
7
Compounds were prepared at 50 ꢀM in phosphateꢀbuffered saline (PBS) with 2% DMSO.
10 ꢀL of the solution was added to of Human serum (40 ꢀL) or to 10% of mice brain
8
9
homogenate in PBS with 0.5% Triton Xꢀ100 (40 ꢀL). All compounds were tested using the
same biological samples of Human serum and mice brain homogenate to assure identical
deactivation conditions. The solutions were incubated during different time periods at 37ºC,
after which metabolization was stopped by adding one volume of acetonitrile. Precipitated
serum and brain proteins were removed by centrifugation at 14000 g for 10 min. Tests
were made confirming that acetonitrile does not precipitate the synthetic peptides. The
remaining peptide in the supernatant was quantified by HPLCꢀFLU (Merck Hitachi, Elite
LaChrom), (exλ = 255 nm and emλ = 285 nm, to detect compounds containing
phenylalanine, and exλ = 275 nm and emλ = 307 nm whenever tyrosine is present).
Chromatographic conditions were as described for determination of enzyme activities. A
calibration curve was built for each compound and the decay in the concentration of the
peptide in the incubated samples was followed till approximately 50% decay.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The metabolization kinetics of the synthetic peptides followed a monophasic behavior in
vitro.The fit of a monophasic deactivation model to experimental data was carried out
through nonꢀlinear regression by minimising the residual sum of squares between the
experimental data points of the relative quantity of intact product vs. time and those
estimated by the model, using Solver addꢀin from Microsoft Excel 2003, considering the
following options: Newton method; 100 iterations, precision of 10−6, 5 % of tolerance and
1x10−4 convergence. The halfꢀlife, t1/2, was determined trough interpolation.
Cytotoxicity assay
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 39
1
2
3
4
5
6
7
8
9
Human colon carcinoma cells (Cacoꢀ2) from the American Type Culture Collection (ATCC
HTBꢀ37), between passage number 26 to 41, were routinely cultured in Tꢀ75 flasks (BD
Biosciences) using high glucose (4.5 g/L) Dulbecco’s Modified Eagle Medium (DMEM),
supplemented with 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) nonꢀessential
aminoacids (Gibco; Grand Island, USA), at 37 °C in a humidified atmosphere containing
5% (v/v) CO2. Once a week, before reaching confluence, cells were cultured using
Dulbecco’s phosphate buffered saline (DPBS) and 0.25% (w/v) TrypsinꢀEDTA (Gibco). The
cytotoxicity assay was performed using the Promega CellTiter 96 cellular viability assay,
MTS, to determine the number of viable cells in culture. Confluent Cacoꢀ2 cells were
incubated with increasing concentrations of testing compounds from 2.5 nM to 50 ꢀM in 0.5
% FBS supplemented DMEM, for 4 hours and 24 hours exposure periods. After the
incubation period the medium was removed and 100 ꢀL of MTS mixture was added to
each well. Cells were incubated with MTS for 4 hours at 37ºC with 5% CO2 at fully
humidified atmosphere. After the incubation period the absorbance of formazan was read
at 490 nm using a plate reader (BioTekTM Power Wave XS). The results were determined
as a percentage of the cellular viability in relation to the control composed of cells with
0.5% FBS supplemented DMEM.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Distribution study in animals
Female mice (n = 12, 8 – 12 weeks, weighing 22 – 37 g) and male Wistar rats (n = 8, 8 –
12 weeks, weighing 182 – 201 g) were purchased from Harlan Ibérica, Barcelona, Spain.
All animals received a standard diet and water ad libitum. Experiments were conducted
according to the Home Office Guidance in the Operation of Animals (Scientific Procedures)
Act 1986, published by Her Majesty’s Stationary Office, London, UK and the Institutional
18
ACS Paragon Plus Environment

Page 19 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Animal Research Committee Guide for the Care and Use of Laboratory Animals published
by the US National Institutes of Health (NIH Publication No. 85ꢀ23, revised 1996), as well
as to the currently adopted EC regulations (2010/63/EU). Finally, the studies are in
compliance with the ARRIVE Guidelines for Reporting Animal Research’ summarized at
www.nc3rs.org.uk. Before experiments animals were fasted for 24h.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
PBS was used to prepare a solution of 4 for intravenous administration at a dose of 0.53
mg/kg. Experiments were done in triplicate. Mice and Wistar rats were sacrificed at 0.5, 1
and 2 hours after dosing and were previously anesthetized with sodium pentobarbital (6
mg/kg i.p.) for collection of blood by cardiac puncture and brain tissue was perfused by
intraventricular injection of 10 mL of chilled saline solution to remove residual blood.
Animal samples preparation
Serum was isolated from the blood samples by centrifugation at 657 g for 10 min, at 4ºC.
Serum proteins were precipitated with one volume of acetonitrile and were removed by
centrifugation at 14000 g for 10 min. Brain was promptly removed and washed with PBS
and blotted dry. Meninges were carefully removed. In the case of Wistar rats only cortex
and hippocampus were collected while for mice the whole brain was used. The brain
samples were homogenized with 2 volumes (V/w) of PBS/acetonitrile/Triton Xꢀ100
(50/50/0.5) using firstly a pestle and then a pipette and finally a sonicator. Compound 3
was added as internal standard to best mimic the recovery of compound 4. The pellet was
removed centrifuging at 14000 g for 10 min. Samples were purified by preparative reverseꢀ
phase HPLC Using a C18 column (Jupiter, Phenomenex) 250 x 21.2 mm, 5 ꢀm.
Chromatographic conditions were as follow: eluent (A) water/TFA (99.9:0.1, V/V), (B)
acetonitrile/water/TFA (90:9.9:0.1, V/V). The linear gradient was started at 80 % A, graded
to 68% for 2 min and then was kept at 68% for 17 min finally grading to 0% A for 5 min.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 39
1
2
3
4
5
6
7
8
The flow rate was 10 mL/min and the column was at room temperature. Collected fractions
(100 ml) were lyophilized, reconstituted in 200 ꢀL of PBS and stored at –20ºC prior
analysis.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Analysis of animal samples
Mice samples were analyzed for detection of the compounds (4 and their metabolites) by
HPLCꢀFLU using an excitation wavelength of 275 nm and an emission wavelength of 307
nm, as well as by LCꢀESIꢀMS (LTQ, ThermoFinigan) in the positive mode. Mass range was
measured from 250ꢀ2000 amu. The ESI source conditions were adjusted as follows:
source capillary operating at 5 kV and source temperature at 300 ºC. A fullꢀscan MS
method in positive mode was used for sample analysis of the molecular ions: [M+H]+,
[M+Na]+ of 4 as well as its metabolites. The chromatographic step was performed in a C8
column (LiChroCART, Merck) 100 mm x 4.6 mm, 5 ꢁm at 30 ºC and 1 mL/min, under the
following
conditions:
eluent
(A)
water/formic
acid
(99.9:0.1,
V/V),
(B)
acetonitrile/water/formic acid (90:9.9:0.1, V/V). Water and acetonitrile were of LCꢀMS
grade. The linear gradient was initially at 80 % A and graded to 65 % eluent A during 27
min, and to 0 % eluent A for another minute.
For identification of 4 and its metabolite in Wistar rat brain, samples were analysed using a
4800plus MALDIꢀTOF/TOF (AB Sciex) mass spectrometer in the positive reflector MS and
MS/MS modes with 3200 laser shots per spectrum and data was collected using the 4000
Series Explorer Software v.3.5.3 (Applied Biosystems). The maximum precursor mass
tolerance (MS) was 50 ppm and the maximum fragment mass tolerance (MS/MS) was 0.3
Da. 0.6 ꢁL of each sample was directly spotted on a MALDI plate and 0.6 ꢁL αꢀCyanoꢀ4ꢀ
hydroxycinnamic acid (CHCA) matrix (LaserBio Labs; 5mg/ml in 50% (V/V) acetonitrile with
5% (v/v) formic acid) was added.
20
ACS Paragon Plus Environment

Page 21 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
For confirmation purposes of the metabolite found in vivo, 5 was produced from 3: 200 ꢀM
of 3 were hydrolyzed in a 500 mM aqueous solution of Sodium Hydroxide overnight.
Reaction was stopped at pH = 7 with acetic acid and the product 5 was obtained in a 52%
yield after purification by preparative reverseꢀphase HPLC and lyophylization, using the
chromatographic conditions previously described for peptides. Reꢀanalysis by analytical
reverseꢀphase HPLCꢀUV, LCꢀMS, MS/MS and MALDIꢀTOF/TOF is in accordance with a
homogenous product. The molecular structure and molecular mass of the product of
hydrolysis 5 is outlined in Table 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
*HBTU ꢀ N,N,N′,N′ꢀTetramethylꢀOꢀ(1Hꢀbenzotriazolꢀ1ꢀyl)uronium hexafluorophosphate
HOBt ꢀ 1ꢀhydroxybenzotriazole
DIEA ꢀ N,Nꢀdiisopropylethylamine
Mca ꢀ 7ꢀmethoxycoumarin
DNP ꢀ 2,4ꢀdinitophenol
FLU ꢀ Fluorescence
Kiapp ꢀ Apparent tightꢀbinding inhibition constants
DMEM ꢀ Dulbecco’s Modified Eagle Medium
FBS ꢀ fetal bovine serum
DPBS ꢀ Dulbecco’s phosphate buffered saline
MTS
tetrazolium
CHCA ꢀ αꢀCyanoꢀ4ꢀhydroxycinnamic acid
ꢀ
3ꢀ(4,5ꢀdimethylthiazolꢀ2ꢀyl)ꢀ5ꢀ(3ꢀcarboxymethoxyphenyl)ꢀ2ꢀ(4ꢀsulfophenyl)ꢀ2Hꢀ
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 39
1
2
3
Acknowledgements
4
5
6
7
8
9
We wish to acknowledge the financial support provided by Fundação para a Ciência e a
Tecnologia through the Strategic Project: PEstꢀOE/EQB/LA0004/2013, the Postdoc grant
SFRH/BPD/82097/2011 and its support for the NMR spectrometers as part of The National
NMR Facility (RECI/BBBꢀBQB/0230/2012) and the mass spectrometers as part of the
Rede Nacional de Espectrometria da Massa (RNEM). We also would like to acknowledge
the UniMS – Mass Spectrometry Unit and particularly Elisabete Pires, ITQB/IBET (Oeiras,
Portugal) for mass spectrometry analysis. Finally, we also acknowledge Liliana Bernardino
and Ana Clara Cristovão from UBI (Covilhã, Portugal), for their proactive help in the tests
using Wistar rats.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
22
ACS Paragon Plus Environment

Page 23 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Tables
Table 1:
M.M. / g.molꢀ1
Ki
Kiapp
Kiapp
#
Structure
(exact mass /
amu)
BACEꢀ2 . Ki
BACEꢀ1
BACEꢀ2
10
11
12
13
14
15
ꢀ1
/ nM
/ nM
BACEꢀ1
O
O
936.1
1.2
3.0
O
O
H
N
H
N
H2N
1
2
2.5
3.7
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
N
N
N
CO2H
H
H
H
O
H
O
O
O
OH
OH
OH
(935.5)
± 0.4
± 0.3
CO2H
CO2H
CO2H
CO2H
O
O
O
O
1268.4
2.0
7.3
O
O
H
N
H
H
N
H
N
H2N
N
N
N
H
N
H
N
H
N
NH2
CO2H
H
O
H
O
(1267.6)
± 0.1
± 0.9
O
CO2H
CO2H
OH
O
CO2H
O
O
O
1284.4
1.9
14.2
O
O
H
N
H
N
H
N
H
N
H2N
3
4
5
7.5
9.2
7.5
N
H
N
H
N
N
H
N
NH2
CO2H
H
O
H
O
O
(1283.6)
± 0.2
± 2.2
CO2H
CO2H
OH
O
CO2H
O
O
1326.5
2.0
18.3
O
O
O
H
H
H
N
H
N
H
N
N
N
N
N
N
N
N
NH2
CO2H
H
H
H
O
H
H
O
O
OH
O
O
(1325.6)
± 0.6
± 0.5
CO2H
CO2H
OH
CO2H
O
O
O
1285.4
1.7
12.7
O
O
O
H
H
N
H
N
H
N
H2N
N
N
H
N
H
N
N
N
OH
CO2H
H
O
H
H
O
OH
O
O
(1284.6)
± 0.2
± 1.9
CO2H
CO2H
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 39
1
2
3
Figures
4
5
6
7
8
9
E
L
D
L
A
V
E
F
F
F
*
P4
P3
P2
P1
P1’
P2’
P3’
P4’
20
8
1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
V
V
F
F
A
A
E
D
18
6
19
7
21
9
22
23
Aβ 18–23’
E
D
E
L
D
L
A
*
10
11
1
2
3
4
5
2
RAGE binding
BACEꢀ1 inhibition
Figure 1:
1.2
1
1
2
0.8
0.6
0.4
0.2
0
ꢀ1.5
ꢀ0.5
0.5
1.5
2.5
Log[Inhibitor] / nM
Figure 2:
24
ACS Paragon Plus Environment

Page 25 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
1.2
1
1.2
1
8
9
0.8
0.6
0.4
0.2
0
0.8
0.6
0.4
0.2
0
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.5
3.7
ꢀ1.5
ꢀ0.5
0.5
1.5
2.5
ꢀ1.5
ꢀ0.5
0.5
1.5
2.5
Log[Inhibitor 1] / nM
Log[Inhibitor 2] / nM
1.2
1.2
BACEꢀ1
BACEꢀ2
1
1
0.8
0.8
0.6
7.5
0.4
0.6
9.2
0.4
0.2
0.2
0
0
ꢀ1.5
ꢀ0.5
0.5
1.5
2.5
ꢀ1.5
ꢀ0.5
0.5
1.5
2.5
Log[Inhibitor 4] / nM
Log[Inhibitor 3] / nM
Figure 3:
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 39
1
2
3
4
5
6
7
8
1
1
0.8
0.6
0.4
0.2
0
Brain homogenate
Serum
1
1
3
4
5
3
0.8
4
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
0.6
0.4
0.2
0
0
2
4
6
0
2
4
6
t / h
t / h
Figure 4:
26
ACS Paragon Plus Environment

Page 27 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
amostra_7 #680 RT: 22.81 AV: 1 NL: 5.39E1
F: ITMS + c ESI Full ms2 1285.00@35.00 [ 350.00ꢀ2000.00]
[MꢀH2O]+
1267.37
OH
O
z10
a7
z5 b7
CO2H
50
O
O
O
O
H
O
H
H
H
N
H2N
N
N
N
H
45
40
35
30
25
20
15
10
5
N
N
N
N
N
OH
CO2H
H
H
H
H
O
OH
O
O
O
CO2H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CO2H
a8
b5
b8
b6
5
b9 b10
5
b6
642.35
a7ꢀH2O
[Mꢀ2H2O]+
1121.71b101249.36
z10ꢀH2O
743.22
b6ꢀH2O
624.37
z5ꢀH2O
610.22
1152.45
b9
1023.31
b8
b7
952.36
b5
a8
789.39
830.45
1222.16
924.39
543.16
961.70 1035.63
382.23
400
718.39
700
1353.93
1300
500
600
800
900
1000
1100
1200
m/z
Internal
standard
[M+H]+
1285.39
1200
1000
800
1313.81
600
1302.50
1291.27
1294.34
400
1279.53
1280
1315.61
1332.19
1339.64
1347.34
1350
1327.65
200
1290
1300
1310
1320
1330
1340
m/z
Figure 5:
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 39
1
OH
O
2
3
4
5
6
7
8
9
CO2H
O
O
O
O
O
H
N
H
N
H
N
H
N
H2N
NaOH(aq)
N
N
N
H
N
H
N
NH2
CO2H
H
H
H
O
OH
O
O
O
CO2H
3
CO2H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Na+
O
OH
O
O
O
O
O
O
O
H
N
H
N
H
N
H
N
H2N
Na+
+ NH3(g)
N
H
N
H
N
H
N
H
N
O
O
H
O
Na+
OH
O
O
O
O
Na+
5
Na+
O
O
O
O
Figure 6:
28
ACS Paragon Plus Environment

Page 29 of 39
Journal of Medicinal Chemistry
1
2
3
4
5
6
FL
FL
Compound 5
bace27021506400x
Compound 5
bace27021506400y
Retention Time
Retention Time
140
120
100
140
120
100
80
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
80
60
60
40
20
0
40
20
Compound 4 + 5
0
ꢀ20
ꢀ40
ꢀ60
-20
-40
-60
Compound 5
Compound 4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Minutes
Figure 7:
29
ACS Paragon Plus Environment