Glucosylpolyphenols as Inhibitors of Aβ-Induced Fyn Kinase Activation and Tau Phosphorylation: Synthesis, Membrane Permeability, and Exploratory Target Assessment within the Scope of Type 2 Diabetes and Alzheimer's Disease

Article abstract of DOI:10.1021/acs.jmedchem.0c00841

Despite the rapidly increasing number of patients suffering from type 2 diabetes, Alzheimer's disease, and diabetes-induced dementia, there are no disease-modifying therapies that are able to prevent or block disease progress. In this work, we investigate the potential of nature-inspired glucosylpolyphenols against relevant targets, including islet amyloid polypeptide, glucosidases, and cholinesterases. Moreover, with the premise of Fyn kinase as a paradigm-shifting target in Alzheimer's drug discovery, we explore glucosylpolyphenols as blockers of Aβ-induced Fyn kinase activation while looking into downstream effects leading to Tau hyperphosphorylation. Several compounds inhibit Aβ-induced Fyn kinase activation and decrease pTau levels at 10 μM concentration, particularly the per-O-methylated glucosylacetophloroglucinol and the 4-glucosylcatechol dibenzoate, the latter inhibiting also butyrylcholinesterase and β-glucosidase. Both compounds are nontoxic with ideal pharmacokinetic properties for further development. This work ultimately highlights the multitarget nature, fine structural tuning capacity, and valuable therapeutic significance of glucosylpolyphenols in the context of these metabolic and neurodegenerative disorders.

Full text of DOI:10.1021/acs.jmedchem.0c00841

Subscriber access provided by UNIV OF DURHAM
Article
Glucosylpolyphenols as inhibitors of A#-induced Fyn kinase
activation and Tau phosphorylation: synthesis, membrane
permeability, and exploratory target assessment within
the scope of type 2 diabetes and Alzheimer’s disease
Ana Marta de Matos, M. Teresa Blázquez-Sánchez, Andreia Bento-Oliveira, Rodrigo de
Almeida, Rafael Santana Nunes, Pedro LOpes, Miguel Machuqueiro, Joana Cristovão,
Cláudio M. Gomes, Cleide S. Souza, Imane G. El Idrissi, Nicola Antonio Colabufo, Ana Diniz,
Filipa Marcelo, M. Conceição Oliveira, Óscar López, José G. Fernández-Bolaños, Philipp
Dätwyler, Beat Ernst, Ke Ning, Claire Garwood, Beining Chen, and Amelia Pilar Rauter
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00841 • Publication Date (Web): 22 Sep 2020
Downloaded from pubs.acs.org on September 26, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 66
Glucosylpolyphenols as inhibitors of A-induced Fyn kinase activation and Tau
phosphorylation: synthesis, membrane permeability, and exploratory target assessment
within the scope of type 2 diabetes and Alzheimer’s disease
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
,§
1,§
1
1
1,2
Ana M. de Matos, M. Teresa Blázquez-Sánchez, Andreia Bento-Oliveira, Rodrigo F. M. de Almeida, Rafael Nunes, Pedro
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1,2
1,2
2
2
3
4
E. M. Lopes, Miguel Machuqueiro, Joana S. Cristóvão, Cláudio M. Gomes, Cleide S. Souza, Imane G. El Idrissi, Nicola A.
Colabufo, Ana Diniz, Filipa Marcelo, M. Conceição Oliveira, Óscar López, José G. Fernandez-Bolaños, Philipp Dätwyler,
Beat Ernst, Ke Ning, Claire Garwood, Beining Chen, and Amélia P. Rauter
4
5
5
6
7
7
8
8
9
9
3,*
1,*
1
Centro de Química Estrutural, Faculdade de Ciências, Universidade de Lisboa, Campo Grande, 1749-016 Lisboa, Portugal; 2Biosystems & Integrative Sciences
3
Institute, Faculdade de Ciências, Universidade de Lisboa, Campo Grande, 1749-016 Lisboa, Portugal; Department of Chemistry, The University of Sheffield,
Dainton Building, Brook Hill, Sheffield S3 7HF, United Kingdom; Dipartimento di Farmacia-Scienze del Farmaco, Università degli Studi di Bari “A. Moro”, Via
Orabona, 4, 70125, Bari, Italy; UCIBIO, REQUIMTE, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal; Mass
Spectrometry Facility at CQE, Insituto Superior Técnico, Av. Rovisco Pais, 1049-001, Lisboa, Portugal; Departamento de Química Orgánica, Facultad de Química,
Universidad de Sevilla, Apartado 1203, E-41071 Sevilla, Spain; Department of Pharmaceutical Sciences, University of Basel, Klingelbergstrasse 50, CH-4056
Basel, Switzerland; Department of Neuroscience, Sheffield Institute for Translational Neuroscience, The University of Sheffield, Sheffield S10 2HQ, United
4
5
6
7
8
9
Kingdom.
Abstract
Despite the rapidly increasing number of patients suffering from type 2 diabetes, Alzheimer’s disease, and diabetes-
induced dementia, there are no disease-modifying therapies able to prevent or block disease progress. In this work, we
investigate the potential of nature-inspired glucosylpolyphenols against relevant targets, including islet amyloid
polypeptide, glucosidases and cholinesterases. Moreover, with the premise of Fyn kinase as a paradigm-shifting target
in Alzheimer’s drug discovery, we explore glucosylpolyphenols as blockers of Aβ-induced Fyn kinase activation, while
looking into downstream effects leading to Tau hyperphosphorylation. Several compounds inhibit Aβ-induced Fyn
kinase activation and decrease pTau levels at 10 μM concentration, particularly the per-O-methylated
glucosylacetophloroglucinol and the 4-glucosylcatechol dibenzoate, the latter inhibiting also butyrylcholinesterase and

-glucosidase. Both compounds are non-toxic with ideal pharmacokinetic properties for further development. This work
ultimately highlights the multitarget nature, fine structural tuning capacity and valuable therapeutic significance of
glucosylpolyphenols in the context of these metabolic and neurodegenerative disorders.
Keywords: Glucosylpolyphenols, Fyn kinase, hIAPP, Tau phosphorylation, type 2 diabetes, Alzheimer’s disease,
diabetes-induced dementia
ACS Paragon Plus Environment

Page 3 of 66
Journal of Medicinal Chemistry
Introduction
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
More than 463 million adults are currently suffering from type 2 diabetes (T2D) worldwide, and up to 73% of them are
likely to be diagnosed with dementia, including Alzheimer’s disease (AD). T2D – the non-insulin dependent type of
diabetes – primarily arises from the ingestion of high-fat diets and lack of physical exercise, which leads to
hyperinsulinemia, dyslipidemia, insulin resistance and, ultimately, hyperglycemia. In turn, AD is characterized for the
presence of extracellular deposits of amyloid beta (Aβ) in the senile plaques and for intracellular neurofibrillary tangles
induced by deposits of hyperphosphorylated Tau protein, accompanied by synaptic dysfunction resulting in neuronal
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
C
death. Recent reports indicate that the cellular prion protein (PrP ) located in the neuronal cell surface works as a high-
affinity binding partner of Aβ oligomers (Aβos), leading to the activation of Fyn kinase, which triggers a cell signaling
3
pathway culminating in Tau hyperphosphorylation. Indeed, Fyn activity was found to be increased in AD brain by
C 4,5
exposure of neurons to Aβos via PrP . Moreover, genetic deletion of Fyn prevents Aβos-induced cell death in
hippocampus and Fyn inhibition restores synapse density and memory function in transgenic mice.6,7 Interestingly, Fyn
inhibition, deficiency, or genetic knockout were found to have increased glucose disposal due to increased insulin
sensitivity, improved fatty acid oxidation, with decreased visceral adipose tissue inflammation.8-10 Hence, the inhibition
of Fyn activity is also a relevant approach in the treatment of diabetes-induced dementia (DID), the so-called “type 3
diabetes”.
Other pathophysiological mechanisms are known to be present in both T2D and AD, namely peripheral and brain insulin
2
resistance and insulin-degrading enzyme (IDE) downregulation leading to increased brain Aβ levels. Furthermore,
cross-seeding events between the brain-penetrant islet amyloid polypeptide (IAPP) and Aβ have also been reported,
being likely to exacerbate the cognitive decline observed in patients suffering from both conditions.11,12 With the lack
of therapeutic alternatives able to block disease progression in both cases, we were interested in finding new molecular
entities able to tackle several molecular targets common to AD and DID with disease-modifying effects. For this
purpose, we turned to nature for inspiration. Polyphenols have been widely reported in the literature for their vast
therapeutic potential, with described antidiabetic, anti-inflammatory and neuroprotective effects.2,13-16 Polyphenol
glucosides (O-glucosyl polyphenols) and glucosylpolyphenols17 (C-glucosyl polyphenols, frequently named as
polyphenol C-glucosides) however, have improved palatability, oral bioavailability due to increased solubility, as well
as enhanced biological activity when compared to the corresponding aglycones, including improved amyloid-
remodeling effects.16,18-21 Importantly, C-glycosyl polyphenols are not liable to chemical and enzymatic hydrolysis, as
sugar is linked to the polyphenol by a C-C bond, and have been described to show higher antidiabetic effects with
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 66
improved target selectivity; for instance, the glucosyldihydrochalcone analogue of the glucoside phlorizin is selective
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
towards SGLT-2 vs. SGLT-1 transporters, while phlorizin is not.22-24
For all the above mentioned reasons, we were interested in exploring the potential multitarget bioactivity of
glucosylpolyphenols based on the structure of 8-β-D-glucosylgenistein (1, Figure 1), a natural glucosylisoflavone
previously reported by our group as a new and potent antidiabetic compound with potential against Aβ1−42-induced
neurotoxicity.25 This compound was found to inhibit IAPP aggregation and to interact with Aβ(1-42) polypeptide
through the same binding mode, involving the sugar moiety, H-6 of ring A, and the aromatic protons of ring B. Yet, we
did not have information as to whereas one or more phenol moieties were beneficial for activity, or even if the molecular
planarity of the aglycone was a crucial feature for the binding epitope and anti-aggregating activity of this compound.
Moreover, C-glucosyl polyphenols derived from acetophloroglucinol or hydroquinone have been reported in the
literature for having antidiabetic effects.26,27 On the basis of this information, we were interested in synthesizing
simplified analogues of 1 with a different hydroxylation pattern in ring A, maintaining the sugar β-C linkage found in
the original compound (Figure 1). To keep rings A and B linked by a three-bond spacer moiety for mimicking 1, we
planned on inserting benzoate moieties in glucosylhydroquinone (a) and glucosylcatechol derivatives (b), or ketone
moieties in glucosylphloroglucinol derivatives (c). Moreover, due to the extremely polar nature of the lead compound,
we were also interested in generating more lipophilic analogues of the natural scaffold with higher chances of crossing
the blood-brain barrier (BBB), namely by O-methyl protection of sugar hydroxy groups. The major goal was to explore
the therapeutic potential and physicochemical properties of compound 1, while comparing them to those of the newly
synthesized analogues and elucidating, whenever possible, structural requirements for bioactivity against multiple
targets involved in T2D and AD, including IAPP, Fyn kinase activation, Tau hyperphosphorylation, as well as
glucosidase and cholinesterase enzymes. Ultimately, we were interested in investigating the therapeutic potential of
glucosylpolyphenols against T2D and AD, while identifying new lead molecules for further pharmaceutical
development in the context of these pathologies.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
C-glycosylation is a key and particularly challenging synthetic step in our strategy. Several methods for C-glycosylation
are currently known, including nucleophilic attack of aromatic Grignard reagents to glycosyl halides,28 the use of
29
30,31
lactones and lithiated compounds, catalysis by transition metals or samarium diiodide,
intermolecular free radical
32
33
reactions and intramolecular aglycone delivery through the Fries-type rearrangement. The latter approach covers the
34
35
strategy first developed by Suzuki et al. and Kometani et al., and consists of a Lewis acid-catalyzed rearrangement
of a phenol glycoside to a C-glycosyl derivative, known as the Fries-type rearrangement. It has been exploited by various
authors up to the present days and successfully applied to the synthesis of flavonoid C-glycosides and of other complex
ACS Paragon Plus Environment

Page 5 of 66
Journal of Medicinal Chemistry
natural products.25, 36-38 In this sense, another goal for this work was to explore the feasibility of C-glycosylation by
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
using different glycosyl donors and acceptors, while studying their impact in the efficacy of the Fries-type rearrangement.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Rationale behind the synthesis of simplified analogues of 8-β-D-glucosylgenistein (1). R = H or Me; R'= H or Bz.
Results
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 66
Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
A. C-Glucosylation
3
9
For the generation of glucosylpolyphenols, we employed either a permethylated glucopyranoside (3, Scheme 1) or
per-benzylated glucosyl donors25,40 (4-6). Polyphenols containing their hydroxy groups in meta, para and ortho
orientation were used as acceptors in a series of C-glycosylation reactions, and the differences in their reactivity were
attentively explored. 2-Naphthol was also used to generate a C-glucosyl analogue with two fused planar rings, in order
to mimic rings A and C in the original structure.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
OH
O
OMe
O
OMe
HO
HO
MeO
MeO
MeO
MeO
a
b
O
H
OMe
OMe
OH
OH
OMe
3, 91%
OMe
2
7-11, 37%-66%
c,d
OBn
OBn
OBn
BnO
BnO
BnO
BnO
BnO
BnO
O
H
f
g
O
O
OH
OR
OH
OBn
OBn
OBn
4
, 58%
15-16, 8%-43%
5
6
R = C(NC)CCl3 (quant. yield)
R = Ac (quant. yield)
e
(derived from: hydroquinone
-naphtol)
2
OBn
BnO
O
H
BnO
OH
OBn
12-14, 6%-57%
derived from: catechol
phloroglucinol
acetophloroglucinol)
(
Scheme 1. Preparation of glucosyl donors and protected C-glucosyl phenols. Reagents and conditions: a) DMF, NaH, MeI, 0 ºC, 3 h; b) dry MeCN,
polyphenol, drierite, -78 ºC → r.t., TMSOTf, 18 h – 48 h; c) DMF, NaH, BnBr, 0 ºC → r.t., 20 h; d) AcOH, H SO , reflux, 36 h; e)
dichloromethane/MeCN, drierite, -78 ºC → r.t. or 40 ºC, TMSOTf, 8 h – 64 h; f) for compound 5: dichloromethane, molecular sieves 3Å, CCl CN,
ºC, 1 h; for compound 6: pyridine, DMAP, 0 ºC → r.t., Ac O, 2.5 h; g) for compound 15: dichloromethane/MeCN, drierite, -78 ºC → r.t.,
BF ‧Et O, 40 h; for compound 16: dichloromethane, molecular sieves 3Å, 0 ºC → r.t., TMSOTf, 20 h.
2
4
3
0
2
3
2
Precursors and conditions leading to the higher yields are presented in Tables 1 and 2. In the case of catechol and
hydroquinone, when using benzyl protected sugar donors, glycosylation yields were drastically lower when compared
to reactions either with phloroglucinol or trihydroxyacetophenone as sugar acceptors. In the first two cases, different
solvent proportions, anomeric protecting group and promoter equivalents were tried attempting to optimize the reaction
ACS Paragon Plus Environment

Page 7 of 66
Journal of Medicinal Chemistry
efficacy; yet, after much experimentation no significant improvements could be observed. Moreover, no significant
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
differences were found when trying to improve the efficacy of hydroquinone and catechol C-glycosylation using either
.
TMSOTf or BF Et O. Notwithstanding, for the first time, per-O-methyl-β-glucosylated polyphenols have been accessed
3
2
in good yields by using TMSOTf as promoter and fully O-methylated methyl glucoside as the glycosyl donor. This
methodology constitutes an advantage when compared to other approaches by saving reaction steps in the generation of
donors with good leaving groups.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 1. C-Glucosylation of polyphenols carried out with TMSOTf as promoter
Glycosyl donor
Glycosyl donor
OMe
OBn
Isolated
Isolated
Phenol
MeO
O
BnO
O
Yield (%)
Yield (%)
MeO
OMe
BnO
OR
OMe
OBn
OBn
O
BnO
OMe
OBn
MeO
6
OMe
BnO
H
O
63
Catechol
MeO
H
(
R = H)
ortho- Hydroxylation pattern
OH
12
OH
OH
7
OH
OMe
O
OBn
O
MeO
BnO
OMe
OBn
42
MeO
HO
BnO
HO
H
H
5
4
3
3
5
7
Phloroglucinol
OH
OH
(
R = H)
meta- Hydroxylation pattern
8
OH
13
OH
OBn
O
OMe
O
BnO
MeO
OBn
OMe
57
MeO
HO
BnO
HO
H
H
Trihydroxyacetophenone
OH
O
OH
O
(R = H)
meta- Hydroxylation pattern
9
OH
1
4
OH
OBn
O
OMe
O
MeO
BnO
OMe
OBn
8
MeO
HO
BnO
HO
Hydroquinone
H
H
(R = Ac)
para- Hydroxylation pattern
10
OH
15*
OH
OMe
O
OBn
O
MeO
BnO
OMe
OBn
MeO
HO
66
BnO
HO
43
3
R = CNHCCl )
H
H
2
-Naphthol
(
1
1
16
*
Compound 15 was obtained using BF
3 2
·Et O as promoter
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Methyl protected glucosyl donor gave, by reaction with all the acceptors tested, C-glucosyl polyphenols as the major
products (7-11, Table 1). Interestingly, with benzyl protected glucosyl donors only glucosylphloroglucinol 13, 3-
glucosyl-2,4,6-trihydroxyacetophenone 14 and 1-glucosylnaphthalen-2-ol 16 were formed in moderate yields as the
electron-donating effects of their aglycones was strong enough to promote C-glucosylation. On the other hand, catechol
and hydroquinone gave C-glucosyl derivatives in very low yield (Table 1), even after increasing reaction time, changing
solvent proportion, promoter and/or polyphenol molar proportion, and temperature (Table 2).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 2. Comparison of experimental conditions used in the C-glucosylation of hydroquinone and catechol with benzyl protected sugar donors.
Compound
Nr.
Sugar Donor
Nr.
Isolated
Yield
Polyphenol
Solvent
Promotor
Temperature
Time
Catechol
DCM/MeCN
(5:1)
TMSOTf
100 mol %
12
4
6
5
5
5
5
5
4
6
- 78 ºC → 40 ºC
- 78 ºC → 40 ºC
- 78 ºC → 40 ºC
- 78 ºC → rt
64 h
60 h
21 h
40 h
40 h
24 h
72 h
96 h
40 h
6%
2%
6%
2%
6%
6%
1%
7%
8%
1
50 mol %
Catechol
50 mmol %
DCM/MeCN
(5:1)
3 2
BF ‧Et O
100 mol%
1
1
1
1
2
5
5
5
1
Hydroquinone
00 mol%
DCM/MeCN
(1:1)
TMSOTf
50 mol %
2
Hydroquinone
50 mol%
DCM/MeCN
(1:1)
TMSOTf
50 mol %
1
Hydroquinone
50 mol%
DCM/MeCN
(5:1)
TMSOTf
50 mol %
- 78 ºC → 40 ºC
- 78 ºC → 40 ºC
- 78 ºC → 82 ºC
- 78 ºC → 40 ºC
- 78 ºC → 40 ºC
1
Hydroquinone
50 mol%
DCM/MeCN
(2:1)
TMSOTf
50 mol %
15
1
Hydroquinone
200 mol%
MeCN
TMSOTf
100 mol %
1
5
Hydroquinone
DCM/MeCN
(5:1)
3 2
BF ‧Et O
15
1
50 mol%
100 mol%
Hydroquinone
50 mol%
DCM/MeCN
(5:1)
3 2
BF ‧Et O
1
5
1
100 mol%
DCM – dichloromethane
Notably, after careful analysis of the NMR spectra we observed that the para-isomers are formed in the synthesis of
catechol C-glucosides 7 and 12, thus indicating that the Lewis acid-promoted Friedel–Crafts-type C-glycosylation is the
favored reaction mechanism, prevalent over the Fries-type rearrangement described for unprotected phenols. While the
synthesis of D-rhamnosyl41 and D-glucosyl42,43 aromatic derivatives has been previously described with protected
ACS Paragon Plus Environment

Page 9 of 66
Journal of Medicinal Chemistry
phenols, to the best of our knowledge, this is the first report of exceptions to the Fries-type rearrangement in the C-
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
glycosylation of unprotected phenols.
B. O-Acylation
A benzoyl group was regioselectively introduced in glucosylhydroquinone derivatives 10 and 15 to afford analogues of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
on the basis of a para hydroxylation pattern (a, Figure 1). Using imidazole, DMAP and benzoyl chloride, the desired
ester derivatives 17 and 19 were obtained as the major products in good yield, together with their dibenzoate analogues
8 and 20 (Scheme 2). Further deprotection of benzyl protected derivatives through catalytic hydrogenation gave the
1
corresponding deprotected compounds 21 and 22. For comparison purposes, compounds 14 and 16 were also
debenzylated to afford compounds 23 and 24, respectively (vd. experimental section).
OR
OR
OR
RO
RO
RO
OR
OR
O
OR
O
O
O
O
RO
HO
RO
HO
RO
O
H
H
H
a
O
OH
O
O
1
0 R = Me
17 R = Me, 65%
19 R = Bn, 53%
18 R = Me, 22%
20 R = Bn, 38%
15 R = Bn
R=Bn
R=Bn
b
b
OH
OH
HO
HO
OH
O
OH
O
O
HO
HO
HO
H
H
O
O
O
O
O
21, 96%
22, 90%
Scheme 2. Preparation of glucosylhydroquinone benzoates. Reagents and conditions: a) Dichloromethane, imidazole, DMAP, BzCl, 0 ºC → r.t.,
0 h – 120 h, b) EtOAc, Pd/C, H , r.t., 16 h – 22 h (R=Bn).
6
2
The observed regioselectivity of these O-acylation reactions may be related with stereochemical hindrance and eventual
hydrogen bonding between the free hydroxy group and the sugar, thus enhancing the relative reactivity of the remaining
phenol hydroxy group towards esterification. Accordingly, regioselective esterification was not observed with
glucosylcatechol derivatives 7 and 12 (structure type b, Figure 1, Table 2). Instead, by applying the same experimental
procedure, an inseparable mixture of mono-benzoylated compounds was obtained, which supports this hypothesis. For
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 66
comparison of bioactivity, the dibenzoate catechol analogues of compounds 18 and 22 were also synthesized (vd.
experimental section, compounds 25 and 26, respectively).
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Moreover, the hydroquinone and catechol per-O-benzyl glycosides 27α,β and 31 (Scheme 3), obtained as major
products under the C-glucosylation reaction conditions (Table 2), were also benzoylated and deprotected to afford the
corresponding α-glycosides 29 and 33 as major products in excellent overall yield.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
OBn
OBn
OBn
BnO
BnO
BnO
BnO
BnO
BnO
OH
a
O
O
O
O
O
O
O
O
O
O
+
94%
OBn
27/
OBn
OBn
28
28
b
b
OH
OH
HO
HO
HO
HO
O
O
O
O
O
OH
OH
O
O
O
2
9, 71%
30, 10%
OBn
O
OBn
O
OH
O
BnO
BnO
BnO
BnO
HO
HO
a
b
O
O
O
OH
O
O
O
O
OBn
OBn
OH
31
32, 82%
33, 83%
Scheme 3. Preparation of O-glucosyl hydroquinone and O-glucosyl catechol benzoates. Reagents and conditions: a) Dichloromethane, imidazole,
2
DMAP, BzCl, 0 ºC → r.t., 60 h – 120 h, b) EtOAc, Pd/C, H , r.t., 16 h – 22 h.
C. C-Acylation
The glucosylphloroglucinol 13 was originally chosen as the precursor of the planned analogue of compound 1 with the
meta hydroxylation pattern (c, Figure 1). Provided that this trihydroxybenzene is an extremely electron-rich aromatic
system, we were expecting a very straightforward Friedel-Crafts-type acylation to occur with phenylacetyl chloride in
the presence of a Lewis acid. After much experimentation employing a number of Lewis acids (e.g. BF ·Et O, TMSOTf
3
2
FeCl , TfOH) and several different conditions without any success, we hypothesized that the sugar moiety could be
3
reducing the reactivity of the aromatic ring or even being degraded in the course of these reactions. The initial C-
acylation of the phenol residue followed by C-glycosylation turned out to be the best option to address this issue. Due
to the dual reactivity of unprotected polyphenols towards electrophiles, hydroxy groups and, as in this case, highly
ACS Paragon Plus Environment

Page 11 of 66
Journal of Medicinal Chemistry
activated nucleophilic carbons, the control of O-/C-acylation was not an easy task. While an equimolecular amount or
an excess of TfOH in absence of solvent generated the di-C-acylated product, the use of TfOH 2% in MeCN rendered
a mixture of the O-/C-acylated products in a ratio of ca. 1/0.3 (Scheme 4). Then, using an excess of TfOH, which acted
both as the solvent and catalyst, compound 35, obtained in 25% yield from trihydroxybenzene, was rearranged into the
C-acylated analogue 34 in 39% yield, which was subsequently C-glycosylated to afford compound 36 in 33% yield.
After catalytic hydrogenation the final analogue 37 was isolated in 68% isolated yield.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
OBn
BnO
OBn
O
HO
OH
BnO
HO
HO
OH
a
c
H
OH
OH
OH
O
OH
36, 33%
O
34, 7%
d
a
OH
HO
OH
HO
OH
O
HO
O
HO
HO
b
H
OH
O
O
OH 35, 25%
OH
O
3
4, 39%
OH
37 (68%)
OBn
O
BnO
OBn
BnO
HO
H
OH
OH
13
Scheme 4. Preparation of compound 37. Reagents and conditions: a) phenylacetyl chloride, 2% TfOH/MeCN, 0⁰C → r.t., overnight. 34, 7%; 35,
5% b) TfOH, 100 ºC, 2 h, 39%; c) TMSOTf, dichloromethane/MeCN, compound 4, drierite, - 40 ºC → r.t., overnight, 33%; d) MeOH/EtOAc,
Pd/C, H , r.t., 3 h, 68%.
2
2
Computational Studies, Epitope Mapping, and Bioactivity Assays
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 66
A. DFT calculations and molecular interactions of rationally designed analogues with hIAPP
by STD-NMR
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
IAPP is co-secreted with insulin by pancreatic -cells. In prediabetes, insulin resistance leads to a compensatory
hypersecretion of insulin and IAPP, leading to its aggregation and deposition in the pancreas in the form of cytotoxic
amyloid oligomers and fibrils. Along with disease progression, this accumulation will lead to the loss and dysfunction
of -cells, which justifies why patients with advanced T2D are no longer able to produce insulin, despite being insulin
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
resistant. Hence, IAPP is an important therapeutic target in T2D, particularly in the prevention of pancreatic dysfunction
arising from aberrant insulin secretion. In this context the interaction of 1 against hIAPP was previously unveiled by
saturation-transfer difference (STD) NMR techniques, being also shown, by Atomic Force Microscopy, the ability of
2
3
this compound to inhibit hIAPP aggregation into amyloid oligomers and fibrils. Based on these findings, we were
interested in assessing if the rationally designed analogues 21 and 37 (aimed at mimicking the original scaffold) would
exhibit the same level of interaction with hIAPP and if the binding epitope would be maintained in the absence of the
central fused ring system. Being more easily accessed in fewer synthetic steps, both 21 and 37 have increased molecular
flexibility when compared to the lead compound 1. DFT calculations [PBE0/6-311G** (H O)] show that low-energy
2
conformations of compounds 21 and 37 are superimposable with compound 1 (Figure 2), namely with its anti-conformer
(defined by an antigeometry for the H1''-C1''-C8-C7 torsion angle), which is the preferentially adopted conformation of
1
in the presence of A(1-42) oligomers,25 suggesting that these molecules are able to mimic the original spatial
orientation of the sugar moiety relatively to rings A and B (see Figures S1-S2 and further details in the Supporting
Information).
Figure 2. DFT-calculated structure of anti-1 (in green), which is the preferentially adopted conformation in the presence of Aβ(1−42) oligomers,25
2
superimposed to the lowest energy conformations identified at the PBE0/6-311G** (H O) level of theory for compounds (A) 21 and (B) 37 (in
grey, red and white), obtained by root mean square (RMS)-fitting using all ring A carbon atoms of each compound.
ACS Paragon Plus Environment

Page 13 of 66
Journal of Medicinal Chemistry
1
2
3
The STD-derived binding epitope obtained for compounds 1, 21 and 37 against hIAPP by STD-NMR (Figure 3 and
Figures S3-S4) suggests that molecular planarity is not a structural requirement for binding and the absence of the central
fused ring system in compounds 21 and 37 does not disrupt the interaction of these compounds with hIAPP. As in the
case of compound 1, the highest STD intensities correspond to the protons of the aromatic core of compounds 21 and
37 (% STD > 80%), when compared to those detected for the glucosyl group (% STD < 40%).
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. STD-derived epitope mapping obtained for compounds 1,25 21 and 37 with hIAPP oligomers.
These experiments show that the binding affinity of the antidiabetic lead 1 is not related to the molecular planarity of
the isoflavone core. Being accessed in only 5 synthetic steps (instead of the 9 needed for the synthesis of the lead
molecule 1), compounds 21 and 37 exhibit a clear binding against hIAPP. Given the reported anti-amyloidogenic
25
properties of 1 against hIAPP, these results encourage further studies of these two simpler analogues to evaluate their
potential for the prevention of IAPP-induced pancreatic failure.
C
B. Inhibition of PrP -A oligomers interaction
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 66
In the last few years, the failure of several clinical trials targeting soluble and fibrillar A by monoclonal antibodies
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
have motivated the scientific community to work in the diversification of therapeutic targets for AD. One possible
4
4
strategy is to focus on the downstream effects of A rather than on its accumulation and aggregation. Soluble Aos
C
were shown to bind to PrP on the neuronal cell surface initiating a cascade through activation of Fyn kinase. Indeed, it
is possible to monitor the activation of Src family kinases (SFKs) such as Fyn kinase by measuring the expression of
phosphospecific epitopes, as previously reported.3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Furthermore, it is commonly assumed that formation of Aβ fibrils and plaque deposits is a crucial event in the
45
pathogenesis of AD. However, there is accumulating evidence that soluble oligomers are the most cytotoxic form of
Aβ although it is still unclear which size and morphology of the aggregates exert neurotoxicity. As with most of the
C
identified Aβ receptors, PrP was found to bind Aβos with much higher affinity than monomeric Aβ (mAβ). In this
work, Natural Aβos, a kind gift from Sheffield Institute for Translational Neuroscience – SITraN (UK), were used.
These were derived from Chinese Hamster Ovary cells (7PA2 cells) stably transfected with cDNA encoding APP751,
an amyloid precursor protein that contains the Val717Phe familial Alzheimer’s disease mutation, as previously
described.46 The Aβos solution contains between 12,000–14,000 pg/mL of total Aβos as measured by ELISA. This
concentration is comparable to that of Aβ peptides detected in human cerebrospinal fluid. The Aβos prepared represent
a heterogeneous population of monomers, dimers, trimers, tetramers, higher state soluble oligomers and other cellular
4
4
proteins as previously reported by Western blotting without further purification. The Aβos preparation using the same
4
6
protocol has been applied in the same way by other groups. The same batch of the recombinant soluble Aβos was used
for all experiments described in the manuscript to minimize impact of experimental variations caused by the
heterogeneous preparation of the Aβos. 1,000 pg/mL of natural Aβos were used to treat HEK 293 cells,
immunocytochemistry (ICC) was performed to detect cellular prion protein, then the slides where imaged with Confocal
Microscope Leica TCS SP5 II objective 63X oil form Leica Microsystems (Figure 4, A). In order to validate the observed
C
binding between PrP and Aβos we performed a PRNP knockdown by using the commercially available kit ON-
C
TARGETplus Human PRNP (5621) siRNA – SMARTpool. Because only the PrP on the cell surface fraction is
involved in the interaction with Aβos, the knockdown was combined with acute cleavage promoted by phospholipase
C (PLC). A live cell staining and imaging was performed, and cells were analyzed by Flow Cytometry: Fluorescence-
Activated Cell Sorting (FACS). As controls were used untreated cells and cells treated with ON-TARGETplus Non-
targeting siRNA Pool (scrambled siRNA). The result was a protein expression reduction by more than 80% (Figure 4,
C
B). It is also interesting to note that phospholipase C (PLC) can cleave PrP on the surface and improve the effects of
C
knockdown further, i.e. further reduces the PrP on the cell surface.
ACS Paragon Plus Environment

Page 15 of 66
Journal of Medicinal Chemistry
We were able to test the Aos binding to the prion protein in both HEK 293 cell line with endogenous or “high” PrPC
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
C
expression (Figure 4, C2) and “low” PrP expression through siRNA knockdown (Figure 4, C3). The two populations
3
were treated with the same concentration (1 x 10 pg/mL) of Aosfor 2 hours. Cells were then washed and stained with
anti-Aos antibodies and imaged by ImageXpress Micro Widefield High Content Screening System (Figure 4, C). It is
C
C
clearly seen that the binding of Aos to cell surface is PrP dependent, i.e. Aos binds to PrP on the cell surface.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
Figure 4. A. Immunocytochemistry (ICC) images of HEK 293 cells treated with natural Aβos (1 x 10 pg/mL). Pictures captured
with Leica TCS SP5 II B. Flow cytometry analysis (FACS) of transfected HEK 293 cells with PRNP siRNA against cellular prion
C
protein (PrP ). Results are expressed as the mean ± standard error mean (SEM); n = 3. Significant differences between control are
indicated with **** p ≤ 0.0001. C. Immunocytochemistry (ICC) analysis by ImageXpress. 1. Negative control represented by HEK
cells not transfected, treated with Aos and stained with only the secondary antibody AF488. 2. Aos binding to the prion protein
C
C
in HEK 293 cell line with “high” PrP expression. 3. Aos binding to the prion protein in HEK 293 cell line with “low” PrP
expression following knockdown performed by PRNP siRNA.
Compound screening in HEK 293 cell lines, previously treated with fresh natural Aos, showed compounds
C
interfering with the PrP -Aos binding (Figure 5).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
C
Figure 5. Screening for compounds able to induce a PrP -NAβos binding inhibition. All compounds were tested at 10 μM as final
concentration. Results are expressed as the mean ± standard error mean (SEM); n = 3. Significant differences between control are
C
indicated with **** p ≤ 0.0001. The PrP -NA1-42 binding (%) after treatment with the compounds is also indicated.
C. Inhibition of A-induced Fyn activation
®
The Opera High Content Screening System was used in this section, as it is applied to test drugs capable of reversing
the altered phenotype observed in AD such as Fyn activation.
Figure 6 shows that the level of Fyn activation of hiPSC derived neural progenitor cells from healthy donors increased
upon treatment with A, i.e. pFyn production is increased. However, we observed that the level of A-induced Fyn
activation was reduced back to normal control values in the presence of the commercial Fyn kinase inhibitor PP1, an
inhibitor of Src family tyrosine kinases Lck, Fyn, Hck, and Src. Moreover, it shows that compounds 8 and 9 (simple
per-O-methylglucosylphenols), 18 (per-O-methylglucosylhydroquinone dibenzoate), 21 (rationally designed
glucosylhydroquinone monobenzoate), 25 and 26 (both glucosylcathecol dibenzoate derivatives), 23 and 24 (fully
unprotected glucosylacetophloroglucinol and glucosylnaphthalene-2-ol) were able to significantly reduce A-induced
Fyn activation at 10 M. Moreover, these C-glucosyl polyphenols are indeed more active than aglycone genistein.
Fyn kinase plays an important role in the physiology of neuronal cells by regulating cell proliferation and differentiation
during the development of the CNS. This enzyme is also involved in signalling transduction pathways that
regulate survival, metabolism and neuronal migration.47 Considering that Fyn inhibition below the physiological levels
(basal levels) could be deleterious for the homeostasis of the cells, we decided to investigate the effects of the compounds
ACS Paragon Plus Environment

Page 17 of 66
Journal of Medicinal Chemistry
on the basal levels of pFyn. Thus, neuronal progenitor cells were treated with the compounds without addition of A to
determine whether the effects observed are independent of A treatment, and it was confirmed that tested compounds
and PP1 alone do not reduce the basal levels of pFyn (Figure 6 C).
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
A rather diverse selection of compounds was able to produce the desired effects, ranging from per-O-methyl and
polyhydroxy forms. Curiously, the natural compound that served as the inspiration for this study (1) was only able to
cause a non-significant reduction in A-induced Fyn activation. Yet, the rationally designed and more flexible
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
hydroquinone monobenzoate (21) exhibited significant differences when compared to A alone. In fact, chemical
modifications made in the original scaffold towards simpler versions of compound 1 without ring B (e.g. in compounds
8
, 9, and 23) were generally more beneficial for the desired activity. On the other hand, no conclusions could be drawn
regarding the advantages or disadvantages of sugars decorated with per-O-methyl groups, as no correlation between
structure and activity could be found regarding this matter. A good example is the presence and absence of these groups
in the two most complex hits found in this assay, compounds 25 and 26, respectively.
We also evaluated the activity of Fyn kinase in the presence of some compounds by ADP-Glo™ kinase assay, a
luminescent ADP detection assay (Figure 7). This assay provides a homogeneous and high-throughput screening method
to measure kinase activity by quantifying the amount of ADP produced during a kinase reaction.
As presented in Figure 7, PP1 was able to reduce the Fyn kinase activity at different concentrations from 1 to 50 µM,
as expected. Furthermore, from the evaluated compounds, only 8 and 10 were able to act as Fyn kinase inhibitors,
denoting that they may have an added therapeutic value against DID given the recognized role of Fyn kinase activity in
insulin sensitivity and lipid utilization.8-10 The fact that compounds 8 and 10, but not 9, were able to inhibit Fyn activity
indicates that in per-O-methyl sugar-containing structures, the acetyl moiety is detrimental for activity, and the para-
and ortho- hydroxylation pattern of the polyphenol is not relevant.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 19 of 66
Journal of Medicinal Chemistry
1
2
Figure 6. Effect of glucosylphenols in A-induced Fyn activation (A, B); effect of glucosylphenols on the basal levels of pFyn, in the absence of
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
®
A (C). The indirect activation of Fyn kinase was measured by immunofluorescence using Opera High Content Screening System (A). Cells
were exposed to 10 µM of compounds in association with A. The results were normalized against the control group, which was considered as
100%. (B and C) Percentage of number of pFyn + spots in each treatment group. Results are expressed as the mean ± standard error mean (SEM);
n = 3. Significant differences between control are indicated with # (p ≤ 0.05) and * (p < 0.05) when compared to A treatment (* p < 0.05) or **
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(p < 0.01) or *** (p < 0.001).
Figure 7. Effect of glucosylpolyphenols and the polyphenol glucoside 29 in the inhibition of Fyn kinase activity measured by the ADP-Glo™
kinase assay. Results are expressed as the mean ± SEM; n = 3. Significant differences between control are indicated with * (p <0.05) or ** (p <
0
.01) when compared with Aß treatment.
It is also important to note that, as tested by a thioflavin-T (ThT) fluorescence assay with A(1-42) (see Figure S5 in
Supporting Information), compounds of this series do not significantly inhibit A(1-42) aggregation per se, which
suggests that the inhibition of A-induced Fyn kinase activation is unlikely to occur exclusively via direct interaction
with A. Most importantly, these results indicate that these compounds are not PAINS acting via autoxidation of
catechol/hydroquinone and subsequent covalent binding to proteins, contrarily to quercetin, a well-known PAIN
4
8
compound used as positive control in this assay.
To estimate eventual behavior of compounds 9, 23, and 26 as pan-assay interference compounds (PAINS), in particular
as membrane PAINS, we have evaluated their potential using a computational protocol. The potential of mean force
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 66
(
PMF) for translocating a hydrophobic probe across a POPC bilayer loaded with compounds 9, 23, and 26 (10%
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
mol/mol) and their calculated membrane permeabilities, are shown in Figure S6 and Table S1 in Supporting Information.
Membrane PAINS, even mild ones such as resveratrol,78 make the membrane significantly more permeable to
hydrophobic compounds. In contrast, none of our compounds led to a significant increase in membrane permeability,
thus indicating that they do not act as membrane PAINS. In addition, we submitted their structure to the badapple online
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
9
service, and the resulting promiscuity indicators also confirm that these compounds will unlikely act as PAINS.
D. Aggregation Studies
With the formation of aggregates, the concentration of free monomers in solution decreases, while the number and/or
size of particles in suspension increase, and, consequently, so does light scattering. On the other hand, aggregate
formation might also induce changes in vibrational progression and the appearance of exciton bands, which are readily
detected in the electronic absorption spectra through changes in the spectral envelope, such as emergence of new bands,
band broadening and variation of the absorbance at λmax, which, if there is no aggregation and other interferences, should
have a linear relation with the concentration of the molecule.51
Aggregating and non-aggregating compounds have been successfully identified using static light scattering and/or
electronic absorption spectroscopy to detect such alterations caused by aggregation. For instance, quercetin52 and
miconazole50,53 were found to aggregate, while fluconazole and ketoconazole are non-aggregating molecules.50
Taking these findings into consideration, static light scattering, and electronic absorption spectroscopy were used to
assess compound aggregation behavior. Only compounds with interesting bioactivity were selected for these
experiments, namely compounds 8, 9, 10, 18, 21, 23, 24, 25, 26 and 33. The compounds under study were compared
with ketoconazole, a known non-aggregating molecule acting as the negative control, and with quercetin, a promiscuous
aggregator, used as the positive control.
Light scattering intensity for compounds 8, 9, 10, 23, 24 and 33 (Figure 8A and B), 21 and 26 (Figure 8C and D) was
similar or weaker than for ketoconazole, for concentrations ranging from 10 µM up to 100 µM. Moreover, the values
for those compounds were significantly lower than the light scattering intensity measured for quercetin. These results
indicate that those eight compounds do not aggregate in this concentration range. Moreover, by comparing the
normalized absorption spectra for each compound at different concentrations (Supporting Information, Figures S7 and
S8), no alterations were observed in the absorption spectra of those eight compounds, both in terms of energy, vibrational
progression or number of bands, also pointing to the absence of aggregation for these compounds. On the other hand, at
ACS Paragon Plus Environment

Page 21 of 66
Journal of Medicinal Chemistry
high concentrations the absorption spectra of the positive control, quercetin, suffers drastic changes (Figure S7 -
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
54
quercetin). First, the typical band of the monomeric species, with maximum at ca. 385 nm, suffers a blue shift to
3
30 nm and becomes broader. This is caused by the loss of the double bond character due to rotation of the 2-1' bond
5
4
out of plane, and consequently the loss of the planar conformation. Also, new bands are visible at ca. 375 nm that
indicate the presence of extended conjugation, through catechol-catechol bonds. A new band is also visible for the
highest concentration of 100 µM between 245 and 270 nm, which, when compared with the absorption spectra of the
different ionization states of the molecule55,56 may indicate an increase of the non-protonated quercetin species. All
these changes are related to the aggregation of the compound. In fact, the pKa of a compound in an aggregate (e.g.,
micellar) environment is different from the one of the monomeric species in solution, shifting the ionization
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
57
58
equilibrium. If any of the compounds tested were aggregating, changes in the absorption spectra would be readily
detected, which was not the case.
For compounds 18 and 25, solutions with only 1.25% and 2.5% DMSO were visibly turbid, especially for 100 µM,
which is an indication of the low aqueous solubility of these compounds that might be due to their high lipophilicity.
With a value as high as 5% of DMSO, the solutions with higher compound concentrations (50 µM and 100 µM) were
still turbid. However, this was not the case at 10 µM and, as can be observed in Figure 8 for this concentration (at which
the cellular studies were conducted), the light scattering intensity is lower than for the non-aggregator ketoconazole.
This indicates that at this concentration these two compounds are not aggregating.
Finally, a linear relationship was confirmed between the concentration and the peak absorbance for the lower energy
band of each compound and for the non-aggregating ketoconazole (Figures S9 and S10), while for the promiscuous
quercetin such relation does not follow a linear behavior (Figure S9 - Quercetin).
In summary, compounds 8, 9, 10, 21, 23, 24, 26 and 33 are not promiscuous aggregators in the concentration range
tested, which encompasses all the concentrations used for the other assays. Our results for compounds 18 and 25 show
that at the concentration of 10 μM no aggregation was detected but at high concentrations, the herein presented inhibition
constants should be considered only as estimates and interpreted with caution.
The aggregation studies confirm that bioactivities herein reported are not due to non-specific effects resulting from the
formation of compound aggregates and are thus the result of bona fide specific compound activity.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 8. Static light scattering intensity at 550 nm and 90°, for compounds 8, 9, 10, 23, 24, 33 (A and B), 21, 26 (C and D), 18 and 25 (E and F),
and the respective controls: Ketoconazole (Ket) and Quercetin (Quer) at 10, 50 and 100 µM. Samples were dissolved in 10 mM PBS (with 100 mM
NaCl, pH 7.4) and 1.25% (A and B), 2.5% (C and D) or 5% (E and F) DMSO. The values are the mean ± S.D. of at least two independent
experiments. The graphics without Quer (B and D) are for a better depiction of the behavior of low scattering compounds. Graphic F is a zoom-in
of E for a better observation of what is happening for the lowest concentration of the compounds. The lines are merely to guide the eye.
E. Inhibition of A-induced Tau phosphorylation
Intraneuronal Neurofibrillary tangles (NFTs) of paired helical filaments (PHF)s are a histopathological hallmark of
Alzheimer’s disease (AD). This NFTs are formed of hyperphosphorylated Tau. Tau is hyperphosphorylated in AD
brain at multiple sites including at residues Thr181.59-62 In order to assess if the compounds are indeed able to
accomplish the desired downstream effects by reducing A-induced Tau pathology, we performed a high-content
image screening (HCS) for phosphorylated Tau (pTau), at Thr181 as recognized by the antibody AT270, using
ACS Paragon Plus Environment

Page 23 of 66
Journal of Medicinal Chemistry
compounds that previously revealed to inhibit A-induced Fyn activation. Our data (Figure 9) revealed that cortical
neurons exposed to A have increased pTau levels when compared to DMSO controls. On the other hand, neurons
treated with A in addition to 10 µM of compounds 9, 10, 18, 23, 25, 26, 29, and genistein significantly reduced
the levels of pTau when compared to the A controls. Even though there was a reduction of pTau in cells treated
with compounds 8 and 33, this reduction was found not to be statistically significant. From all tested compounds,
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
9
, 18, 23, 25 and 26 were able to reduce A-induced Fyn activation, with concomitant decrease in A-induced pTau.
Figure 9. Effect of compounds against hyperphosphorylation of Tau induced by Aβ. Neurons treated with Aβ oligomers were evaluated against
®
pTau (AT270). Tau hyperphosphorylation was measured by immunofluorescence using Opera High Content Screening System. Cells were
exposed to 10 µM of each compound in association with Aβ for 4 days. Results were normalized against the control group considered as 100%.
The values are expressed as the mean ± SEM; n = 3. Significant differences between control are indicated with # (p ≤ 0.05) and * (p <0.05), ** (p
<
0.01) or *** (p < 0.001) when compared with Aβ treatment.
F. Cytotoxicity in neuronal cells derived from hiPSCs
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 66
In order to confirm that the synthesized compounds are not cytotoxic at relevant concentrations, we have differentiated
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
63
hiPSC cells derived from health control MIFF1 into neural cells. We observed that, after 20 days of differentiation,
these cells express specific neural progenitor markers such as Nestin. NPC cells were treated with each compound for
2
4 hours, and none presented any signs of cytotoxicity at 10 µM (Figure 10). Furthermore, compounds 23, 26, and 29
were not cytotoxic in concentrations up to 100 µM, while 9 is safe to administer up to a 50 µM concentration (data not
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
shown).
Figure 10. Cytotoxicity of C-glucosyl phenols and glucosides 29, 33 in neuronal cells derived from hiPSCs. Cell viability was measured in an
MTT assay. Cells were exposed to 10 µM of each compound for 24 h. Results were normalized relative to a control group considered as 100%.
The values are expressed as the mean ± SEM; n = 3. Significant differences between control are indicated with * (p <0.05).
G. Glycosidase and cholinesterase inhibitory activity screening
Postprandial glycemia control is key in managing T2D clinical manifestations. This control can be achieved through the
6
4
inhibition of intestinal glucosidases, in particular α-glucosidase. These enzymes catalyze the hydrolysis of complex
carbohydrates present in the gut into simple sugars able to be absorbed into the blood stream and thus contribute to the
65
increase of glycemia levels. Since we had previously elucidated the powerful α-glucosidase inhibitory activity of the
ethyl acetate extract of Genista tenera where compound 1 is the major component (97.6% for the extract vs. 82.2% for
ACS Paragon Plus Environment

Page 25 of 66
Journal of Medicinal Chemistry
the commercial drug acarbose), we were interested in finding out if it was due to the presence of the lead C-glucosyl
isoflavone.66 However, 1 was found to have only modest activity, with 14% inhibition at 100 μM (Table 3). This
compound was a slightly better β-glucosidase inhibitor, being able to decrease its activity in 23% at the same
concentration. Notably, these activities are cumulative with the antihyperglycemic effects of 1 observed in Wistar rats,
since treatment was administered i.p.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Genistein, on the other hand, is a powerful α-glucosidase uncompetitive inhibitor (84% inhibition at 100 μM; K = 12
ib
±
2 µM) and a moderate β-glucosidase competitive inhibitor (44% inhibition at 100 μM; K = 66 ± 13 µM), indicating
ia
that the presence of the C-C linked sugar moiety at C-8 is, in this case, detrimental to activity. Remarkably, the catechol
glucoside 33 was found to be the best glucosidase inhibitor amongst the synthesized analogues, with an excellent α-
glucosidase competitive inhibitor activity (74% inhibition at 100 µM; K = 39 ± 4 µM), and a modest β-glucosidase
ia
inhibitor activity (13% inhibition at 100 µM). Apart from this compound, only three others were able to concomitantly
inhibit both glucosidases: the hydroquinone derivatives 17 and 29, and the naphthalen-2-ol derivative 24.
Table 3. Glycosidase and cholinesterase (AChE and BuChE) inhibitory efficacy of compound 1 and analogues at 100 μM.
β-Glucosidase
AChE
BuChE
α-Glucosidase
Compound
Inhibition
Inhibition
Inhibition
Inhibition
OH
HO
OH
O
HO
HO
H
O
14%
23%
26%
n.i.
OH
O
OH
1
OMe
O
MeO
OMe
MeO
H
n.i.
14%
15%
16%
OH
OH
7
OMe
O
MeO
OMe
MeO
HO
H
OH
n.i.
15%
n.i.
10%
OH
8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 66
OMe
O
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
MeO
OMe
MeO
HO
H
OH
O
n.i.
n.i.
n.i.
16%
OH
9
OMe
O
MeO
OMe
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MeO
HO
H
n.i.
n.i.
n.i.
21%
23%
12%
OH
10
OMe
O
MeO
OMe
MeO
HO
H
n.i.
16%
n.i.
11
OMe
MeO
OMe
O
O
MeO
HO
H
11%
24%
14%
O
17
OMe
O
MeO
OMe
O
MeO
O
H
n.i.
n.i.
n.i.
18%
17%
18%
19%
16%
O
O
18
OH
HO
OH
O
HO
HO
H
n.i.
n.i.
O
O
21
OH
HO
OH
O
O
HO
O
H
n.i.
21%
O
O
22
OH
HO
OH
O
HO
HO
H
OH
O
n.i.
24%
n.i.
15%
OH
23
ACS Paragon Plus Environment

Page 27 of 66
Journal of Medicinal Chemistry
OH
1
2
3
4
5
6
7
8
9
HO
OH
O
HO
HO
H
12%
23%
n.i.
10%
2
4
OMe
MeO
OMe
O
O
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MeO
H
O
n.i.
27%
n.i.
12%
O
O
25
OH
HO
HO
OH
O
O
H
O
n.i.
17%
n.i.
39%
O
O
2
6
OH
O
HO
HO
OH
O
1
8%
4%
19%
n.i.
10%
O
O
29
OH
HO
7
OH
O
HO
O
O
Competitive inhibition
13%
10%
17%
O
K
ia = 39 ± 4 µM
33
OH
HO
OH
O
HO
HO
H
OH
O
n.i.
n.i.
n.i.
17%
OH
37
HO
O
O
8
4%
44%
Competitive inhibition
1
2%
41%
OH
Uncompetitive inhibition
OH
K
ib = 12 ± 2 µM
Kia = 66 ± 13 µM
Genistein
Kia - inhibition constant of the inhibitor binding the free enzyme; Kib - inhibition constant of the inhibitor binding the enzyme-substrate
Complex; n.i. = no inhibition; n.d. = not determined.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 66
Acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) are two well-characterized therapeutic targets in AD
owing to their ability to catalyse the hydrolysis of the neurotransmitter acetylcholine, which is responsible for the
cognitive functionality, and whose level is particularly low in AD patients. Three of the so far four FDA-approved drugs
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
for AD consist of selective or dual cholinesterase inhibitors, including donepezil, galantamine and rivastigmine. The
inhibition of AChE and BuChE correlate with lower Aβ levels, decreased Aβ aggregation, improved learning and
memory.67-70 BuChE is considered to play a minor role in the regulation in acetylcholine levels in healthy brains;
however, the levels of this enzyme progressively increase in AD, whereas those of AChE decline or remain unchanged.71
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Not so well studied and divulged is the role of butyrylcholinesterase in the etiology of T2D. However, elevated AChE,
but especially serum BuChE activity have been correlated with insulin resistance, increased adiposity and abnormal
serum lipid profile, being regarded as a risk factor for T2D.72-75 Thus, these two enzymes may be regarded as additional
therapeutic targets for DID.
Similar to what was described for α-glucosidase, the ethyl acetate extract of Genista tenera was capable of inhibiting
6
6
this enzyme (77.0% at 130 μg/mL). Hence, we were interested in assessing whether the anticholinergic activity of the
extract was due to the presence of 1 as a major component. This compound was however able to inhibit AChE only by
2
6% at 100 μM (43 μg/mL) and, in this assay, genistein presented merely half of the inhibitory capacity of 1 (Table 3).
On the contrary, genistein was a much stronger BuChE inhibitor than 1, displaying 41% inhibition at 100 μM. From the
synthesized analogues of 1, only compounds 7, 17, 18 and 33 were active against AChE, while roughly all presented a
BuChE inhibition capacity of, at least, 10%. Compounds 10, 11, 22 and 26 were able to inhibit BuChE in over 20% at
1
00 μM, from which compound 26 stands out with 39% inhibition.
H. Membrane permeability assays
Compounds were tested in a parallel artificial membrane permeability assay (PAMPA) in order to measure and
rationalize their potential to cross membrane barriers. Testosterone was used as the positive control in this assay. It is
important to note that this assay merely looks into the ability of compounds to passively diffuse through cell membranes.
Being glycosides, it is possible that the sugar moiety acts as a shuttle for their passage into the brain through GLUT-1
transporters highly expressed in the blood-brain barrier (BBB), as previously reported for similar molecules.76
To complete our analysis, the partition coefficient at physiological pH (log D ) was also determined for most
7
.4
compounds. Ideally, log D values should be located between 1 and 4 for a good compromise between solubility and
ACS Paragon Plus Environment

Page 29 of 66
Journal of Medicinal Chemistry
7
7
membrane permeability allowing oral availability, good cell permeation and low metabolic susceptibility. Results are
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
presented in Table 4.
e
Table 4. Calculated partition coefficient (c log P), effective permeability (log P ) and partition coefficient at pH 7.4 (log D7.4) of the synthesized
compounds and genistein
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compound Nr.
clogPa,b
log Pe
log D7.4
1
7
-0.17
- 4.63 ± 0.15
- 5.33 ± 0.08
-0.1 ± 0.1
1.1 ± 0.1
0.74
0
.58
8
9
- 5.24 ± 0.17
- 4.74 ± 0.02
- 5.52 ± 0.07
- 4.39 ± 0.04
1.6 ± 0.2
2.3 ± 0.3
n.d.
1.06
0
1
.75
.96
1
0
11
17
2.7 ± 0.2
Membrane retention
over 80%
2.70
3.95
3.2 ± 0.1
1
2
8
1
Equilibrated
- 6.35 ± 0.12
> 2.5
< 0.5
0
1
.60
.93
22
- 5.18 ± 0.61
2.0 ± 0.2
n.d.
-1.23
2
2
3
4
Below detection limit
- 6.41 ± 0.24
-0.44
3.81
n.d.
Partial membrane
retention
2
5
> 2.5
1
.95
2
2
3
3
6
9
3
7
- 5.06 ± 0.08
Below detection limit
- 5.85 ± 0.54
n.d.
0.59
1.0 ± 0.1
0.1 ± 0.3
0.58
0.13
n.d.
n.d.
Genistein
2.45
- 4.49 ± 0.04
3.3 ± 0.2
2
.99
Testosterone
- 4.42 ± 0.09
---
a
b
n.d. = not determined; n.d. - not determined; Calculated using ALOGPS 2.1. Based on c logP values, 1, 23 and 24
ACS Paragon Plus Environment