Glycosylated tetrahydrosalens as multifunctional molecules for Alzheimer's therapy

Article abstract of DOI:10.1039/b902545f

The tetrahydrosalens N,N′-bis(2-hydroxybenzyl)-ethane-1,2-diamine (H₂L¹), N,N′-bis(2-hydroxybenzyl)-(-)-1,2- cyclohexane-(1R,2R)-diamine (H₂L²), N,N′-bis(2- hydroxybenzyl)-N,N′-dimethyl-ethane-1,2-diamine (H₂L ³), N,N′-bis(2-hydroxybenzyl)-N,N′-dibenzyl-ethane-1,2- diamine (H₂L⁴), and N,N′-bis(2-(4-tert-butyl) hydroxybenzyl)-ethane-1,2-diamine (H₂L⁵), as well as their prodrug glycosylated forms, GL^1-5, have been prepared and evaluated in vitro for their potential use as Alzheimer's disease (AD) therapeutics. Dysfunctional interactions of metal ions, especially those of Cu, Zn, and Fe, with the amyloid-β (Aβ) peptide are hypothesised to play an important role in the aetiology of AD, and disruption of these aberrant metal-peptide interactions via chelation therapy holds considerable promise as a therapeutic strategy. Tetrahydrosalens such as H₂L^1-5 have a significant affinity for metal ions, and thus should be able to compete with the Aβ peptide for Cu, Zn, and Fe in the brain. This activity was assayed in vitrovia a turbidity assay; H₂L¹ and H₂L ³ were found to attenuate Aβ_1-40 aggregation after exposure to Cu²⁺ and Zn²⁺. In addition, H ₂L^1-5 were determined to be potent antioxidants on the basis of an in vitro antioxidant assay. GL^1-5 were prepared as metal binding prodrugs; glycosylation is intended to prevent systemic metal binding, improve solubility, and enhance brain uptake. Enzymatic (β-glucosidase) deprotection of the carbohydrate moieties was facile, with the exception of GL⁴, demonstrating the general feasibility of this prodrug approach. Finally, a representative prodrug, GL³, was determined to be non-toxic over a large concentration range in a cell viability assay.

Full text of DOI:10.1039/b902545f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Glycosylated tetrahydrosalens as multifunctional molecules for
Alzheimer’s therapy†
Tim Storr,‡a Lauren E. Scott,a Meryn L. Bowen,a David E. Green,§a Katherine H. Thompson,a
Harvey J. Schugar*b and Chris Orvig*a
Received 6th February 2009, Accepted 16th February 2009
First published as an Advance Article on the web 5th March 2009
DOI: 10.1039/b902545f
The tetrahydrosalens N,N¢-bis(2-hydroxybenzyl)-ethane-1,2-diamine (H2L1), N,N¢-bis(2-
hydroxybenzyl)-(-)-1,2-cyclohexane-(1R,2R)-diamine (H2L2), N,N¢-bis(2-hydroxybenzyl)-N,N¢-
dimethyl-ethane-1,2-diamine (H2L3), N,N¢-bis(2-hydroxybenzyl)-N,N¢-dibenzyl-ethane-1,2-
diamine (H2L4), and N,N¢-bis(2-(4-tert-butyl)hydroxybenzyl)-ethane-1,2-diamine (H2L5), as well as
their prodrug glycosylated forms, GL1–5, have been prepared and evaluated in vitro for their potential
use as Alzheimer’s disease (AD) therapeutics. Dysfunctional interactions of metal ions, especially those
of Cu, Zn, and Fe, with the amyloid-b (Ab) peptide are hypothesised to play an important role in the
aetiology of AD, and disruption of these aberrant metal–peptide interactions via chelation therapy
holds considerable promise as a therapeutic strategy. Tetrahydrosalens such as H2L1–5 have a significant
affinity for metal ions, and thus should be able to compete with the Ab peptide for Cu, Zn, and Fe in
the brain. This activity was assayed in vitro via a turbidity assay; H2L1 and H2L3 were found to
attenuate Ab1–40 aggregation after exposure to Cu2+ and Zn2+. In addition, H2L1–5 were determined to be
potent antioxidants on the basis of an in vitro antioxidant assay. GL1–5 were prepared as metal binding
prodrugs; glycosylation is intended to prevent systemic metal binding, improve solubility, and enhance
brain uptake. Enzymatic (b-glucosidase) deprotection of the carbohydrate moieties was facile, with the
exception of GL4, demonstrating the general feasibility of this prodrug approach. Finally, a
representative prodrug, GL3, was determined to be non-toxic over a large concentration range in a cell
viability assay.
development of new effective therapies.4,5 AD is characterised by
Introduction
the progressive loss of neuronal cell populations, protein aggre-
gation, and extensive evidence of oxidative stress.6 Intracellular
deposits of hyperphosphorylated tau protein7 and extracellular
aggregates of the 39–42 residue amyloid-b (Ab) peptides8 are the
pathological hallmarks of AD, however the amyloid hypothesis
has long been a dominant theory to explain the aetiology of
AD,9,10 linking the Ab peptide to the neurodegeneration in the
disease. More specifically, the amyloid hypothesis postulates that
plaque Ab depositions, or partially aggregated soluble Ab, trigger
a neurotoxic cascade resulting in AD pathology.11 In this work,
dual function chelating antioxidant molecules designed to remove
and/or redistribute excess metals in the brain and reduce the
concomitant oxidative stress are investigated as potential AD
therapies.
Fenton processes involving redox active metal ions (of Cu,
Fe),1,6,12 possibly in concert with impaired cellular energy
metabolism, lead to increased oxidative stress, a major feature
of age-related neurodegenerative diseases. Iron levels in AD brain
tissue are elevated; amyloid plaque deposits contain remarkably
high concentrations of Cu and Zn,13–15 and have been described as
“metallic sinks”.16 The interaction of Ab with metal ions in vitro
leads to aggregation, with Zn2+ accelerating the deposition process
more rapidly, compared to Cu2+ and Fe3+, at pH 7.4.17,18 Further
in vitro studies have demonstrated that Cu2+ and Fe3+ potentiate
the neurotoxicity of Ab via redox-cycling and the production of
H2O2 in the presence of dioxygen.19 This H2O2 production can
Molecules designed to sequester excess metals in the brain and
aid in the removal of such metals are attractive therapeutic
targets in neurodegenerative disorders such as Parkinson’s disease,
Alzheimer’s disease (AD), Creutzfeldt–Jakob disease, and amy-
otrophic lateral sclerosis.1,2 Of these diseases, AD affects more than
24 million people worldwide and it is expected that this number
will double every 20 years to reach 81 million by 2040.3 Current
treatment options for AD are limited, and there is a need for the
aMedicinal Inorganic Chemistry Group, Department of Chemistry, Univer-
sity of British Columbia, 2036 Main Mall, Vancouver, British Columbia,
V6T 1Z1, Canada. E-mail: orvig@chem.ubc.ca; Fax: +1 604-822-2847;
Tel: +1 604-822-4449
bDepartment of Chemistry and Chemical Biology, Rutgers, The State
University of New Jersey, 610 Taylor Road, Piscataway, NJ, 08854-
8087, USA. E-mail: schugar@rutchem.rutgers.edu; Fax: +1 732-445-5312;
Tel: +1 732-445-2602
† Electronic supplementary information (ESI) available: Enzymatic cleav-
age of glycosylated tetrahydrosalen pro-ligands by b-glucosidase and
analysis by TLC and ESI-MS+; cytotoxicity assay of a representative
pro-ligand. See DOI: 10.1039/b902545f
‡ Present address: Department of Chemistry, Simon Fraser University,
Burnaby, British Columbia, V5A 1S6, Canada.
§ Present address: STTARR Innovation Centre, Radiation Medicine
Program, Princess Margaret Hospital, University Health Network, 101
College St, Toronto, Ontario, M5G 1L7, Canada; and Department
of Radiation Physics, Radiation Medicine Program, Princess Margaret
Hospital, University Health Network, 610 University Avenue, Toronto,
Ontario, M5G 2M9, Canada.
3034 | Dalton Trans., 2009, 3034–3043
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
be significantly attenuated by the presence of the chelators di-
ethylenetriaminepentaacetic acid (DTPA), triethylenetetraamine
(TETA), and desferrioxamine (DFO, Chart 1),19 highlighting
metal sequestration as a potential AD treatment strategy. Treat-
ment of the Tg2576 transgenic mouse (murine model for AD)
with clioquinol (Chart 1), an 8-hydroxyquinoline chelator, reduced
brain amyloid accumulation,20 and the effect was recently repro-
duced with the lipophilic chelator DP-109 (Chart 1).21 Separate
clinical studies of both the strong metal ion chelator DFO,22
and hydroxyquinoline derivatives23–25 have shown indications of
slowing neurological decline. However, long-term use of strong
chelators that are not tissue-specific, such as DFO, is likely to
affect the homeostasis of essential metal-requiring biomolecules
such as metalloenzymes.26 Specific, rather than systemic, chelation
of excess metals in the brain of AD patients is therefore highly
preferred.
Chart 2 Carbohydrate-containing multifunctional molecules for AD
therapy.
the potential of facilitated carbohydrate transport across the
BBB.
A series of tetrahydrosalen compounds (Chart 3) were investi-
gated in this work as the tetradentate N2O2 donor set is known
to have a moderate affinity for metal ions,30,34 and in addition,
the phenolic moieties of H2L1–5 have the potential to act as
antioxidants.35,36 These compounds differ from our previous work
on tetrahydrosalens30 in that O-glycosylation effectively masks the
chelator, thereby preventing systemic metal binding. The molecule
in this form is referred to as a prodrug as the primary function-
alities (metal binding or antioxidant activity) are masked; once
the carbohydrate moiety is removed the molecule is referred to
as a free pro-ligand or pro-chelator, able to bind metal ions in
solution. The development of multifunctional metal ion chelating
antioxidant molecules with attached carbohydrate moieties offers
a new strategy for Alzheimer’s treatment, beyond current therapies
which offer only symptomatic relief.
Chart 1 Metal ion chelators for AD therapy.
There is considerable promise in enhancing the targeting and ef-
ficacy of metal sequestration through pro-ligand design.27 A series
of pro-ligands, formed from their inactive boronic ester precursors
by H2O2 oxidation, was recently shown to inhibit Fe-promoted
OH∑ radical formation; a significant advance in the design of
selective metal passivating agents.28 Our research has focused on
the development of metal sequestration agents (Chart 2), using
carbohydrates to increase water solubility, minimise toxicity, and
improve targeting.29,30
A glycoconjugate approach has been used in an effort to enhance
the central nervous system (CNS) targeting of therapeutics such
as dopamine derivatives for Parkinson’s disease therapy.31 The
brain requires a significant amount of glucose to maintain
normal bodily functions (up to 30% of total body glucose
consumption32), and this demand is met by the high density of
hexose transporters (a family of membrane proteins designated
GLUT)33 at the blood-brain barrier (BBB). The GLUTs, including
GLUT-1, GLUT-3, and GLUT-4, at the BBB, also offer the
potential for transporter-facilitated drug delivery,33 for increased
brain access to drug molecules. Indeed, positive brain uptake
by a radiolabeled hydroxypyridinone glycoconjugate29 highlights
Chart 3 Tetrahydrosalen ligands H2L1–5, and carbohydrate-protected
analogues GL1–5
.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 3034–3043 | 3035
©

View Article Online
precipitate washed with EtOH and Et2O to afford the imine as
Experimental procedures
a light yellow solid. dH (300 MHz; MeOD): 8.85 (2H, s, imine
3
3
General methods
H), 7.86 (2H, d, Jc,d = 7.5, ArH), 7.43 (2H, dd, Jb,c = 7.2,
3Ja,b = 8.4, ArH), 7.23 (2H, d, Ja,b = 8.4, ArH), 7.08 (2H, dd,
3
All solvents and chemicals (Fisher, Aldrich, Lancaster) were
reagent grade and used without further purification unless
otherwise specified. Water was deionised, purified (Barnstead
D9802 and D9804 cartridges) and distilled with a Corning MP-1
Mega-Pure Still. Synthetic human Ab(1–40) was purchased from
3Jb,c = 7.2, Jc,d = 7.5, ArH), 4.89 (2H, d, J1,2 = 6.9, H-1), 3.86
3
3
3
2
(4H, m, NCH2CH2N), 3.73 (2H, dd, J5,6a = 5.4, J6a,6b = 11.7,
H-6a), 3.45 (2H, m, H-5), 3.44 (8H, m, H-2, H-3, H-4, H-6b).
MS (+ LSIMS) m/z (relative intensity) = 593 ([L + H]+, 100).
The imine was suspended in MeOH (7 mL) and NaBH4 (0.030 g,
0.79 mmol) was added. The reaction mixture was heated to reflux
and another equivalent of NaBH4 (0.028 g, 0.79 mmol) was
added after 3 h. The reaction mixture was cooled after 6 h, H2O
(6 mL) was added, and the solvent removed in vacuo. The residue
was then purified by size-exclusion chromatography on Sephadex
G-10 (MeOH eluent) to afford the product GL1 as a white solid
(0.234 g, 50% yield over two steps). dH (400 MHz; MeOD): 7.52
Advanced ChemTech (Louisville, KY). H and 13C{ H} NMR
spectra were recorded at room temperature on a Bruker AV-300
or an AV-400 instrument at 300.13 (75.48 for 13C NMR) or 400.13
(100.62 for 13C NMR) MHz, respectively. Ar = aromatic, and
J values are given in Hz. Mass spectra (+ ion) were obtained
on a Kratos Concept II H32Q instrument (Cs+, LSIMS), a
Macromass LCT (electrospray ionisation) instrument, or a Bruker
Biflex IV (MALDI) instrument. C, H, N analyses were performed
at UBC by M. Lakha (Carlo Erba analytical instrumentation).
UV-Visible data were collected using a Hewlett-Packard 8543
diode array spectrometer, or a Molecular Devices Thermo-
max microplate reader. Compounds H2L1,34 H2L2,37 H2L3,38
H2L4,39 and 2-(bromomethyl)phenyl-2,3,4,6-tetra-O-acetyl-b-D-
glucopyranoside40 were prepared by literature methods.
1
1
(4H, m, ArH), 7.41 (2H, d, 3Ja,b = 8.5, ArH), 7.21 (2H, dd, 3Jb,c
=
7.2, 3Jc,d = 7.6, ArH), 5.01 (2H, d, 3J1,2 = 7.7, H-1), 4.47 (2H, d,
2
2Jgem = 12.8, ArCH2), 4.39 (2H, d, Jgem = 12.8, ArCH2), 3.94
2
3
(2H, overlapping dd, J6a,6b = 11.8, H-6b), 3.75 (2H, dd, J5,6a
=
2
5.4, J6a,6b = 11.8, H-6a), 3.54 (12H, m, H-2, H-3, H-4, H-5,
NCH2CH2N). dC (100.62 MHz; MeOD): 156.39 (ArC), 129.90
(ArCH), 129.30 (ArC), 128.27, 121.91, 115.66 (ArCH), 102.71
(C1), 77.26, 76.45 (C3/C5), 73.62 (C2), 69.79 (C4), 60.87 (C6),
48.52, 47.96 (ArCH2, NCH2CH2N). Found: C, 54.5; H, 6.9; N,
4.5%; M (mass spectrum), 596. C28H40N2O12·H2O requires C,
54.7; H, 6.9; N, 4.6%; M, 596.
Synthesis
N ,N¢-Bis(2-(4-tert-butyl)hydroxybenzyl)-ethane-1,2-diamine
(H2L5). Ethylenediamine (0.203 g, 3.38 mmol) and 5-tert-butyl-
2-hydrozybenzaldehyde 1 (1.203 g, 6.76 mmol) in EtOH (30 mL)
were heated at reflux for 4 h followed by solvent removal in vacuo.
dH (400 MHz; DMSO-d6): 8.62 (2H, s, imine H), 7.60 (2H, d,
N,N¢-Bis(2-(phenyl b-D-glucopyranoside)benzyl)-(-)-1,2-cyclo-
hexane-(1R,2R)-diamine (GL2). The title compound GL2
(0.274 g, 71% over two steps) was prepared from (1R,2R)-(-)-1,2-
diaminocyclohexane (0.071 g, 0.62 mmol) and helicin 2 (0.351 g,
1.23 mmol) by a procedure analogous to that described for GL1.
4Jb,d = 2.5, ArH), 7.49 (2H, dd, Ja,b = 8.6, Jb,d = 2.5 ArH),
3
4
3
7.08 (2H, d, Ja,b = 8.6, ArH), 3.88 (4H, s, NCH2CH2N), 1.25
dH (300 MHz; MeOD): 7.28 (6H, m, ArH), 7.02 (2H, ddd, 4Ja,c
=
(18H, s, C(CH3)3). ES-MS m/z (relative intensity) = 403 ([M
+ Na]+, 100), 381 ([M + H]+, 10). This compound was reduced
in the next step without further purification. The imine was
dissolved in MeOH (30 mL) and the solution cooled in an ice-
bath. NaBH4 (0.256 g, 6.77 mmol) was added in portions and
then the reaction mixture was warmed to rt with stirring. The
solvent was removed after 5 h and the residue partitioned between
a saturated NaHCO3 (50 mL) solution and CH2Cl2 (30 mL)
and the aqueous layer extracted with CH2Cl2 (2 ¥ 20 mL). The
combined organic fractions were dried with Na2SO4, filtered, the
solvent was removed in vacuo, and the residue purified by silica
gel chromatography (95 : 5 CH2Cl2 : MeOH eluent) to afford the
product H2L5 as a light yellow solid (0.919 g, 71%). dH (300 MHz;
1.2, 3Jc,d = 7.9, 3Jb,c = 7.2, ArH), 4.95 (2H, d, 3J1,2 = 7.5, H-1), 3.90
(2H, dd, 3J5,6b = 2.1, 2J6a,6b = 12.0, H-6b), 3.84 (2H, d, 2Jgem = 12.6,
ArCH2), 3.78 (2H, d, 2Jgem = 12.6, ArCH2), 3.72 (2H, dd, 3J5,6a
=
5.7, 2J6a,6b = 12.0, H-6a), 3.46 (8H, m, H-2, H-3, H-4, H-5), 2.36
(2H, m, ring H), 2.02 (2H, m, ring H), 1.69 (2H, m, ring H), 1.19
(4H, m, ring H). dC (75.48 MHz; MeOD): 155.45 (ArC), 130.67,
129.73 (ArCH), 128.09 (ArC), 123.02, 115.52 (ArCH), 100.74
(C1), 76.03, 75.61 (C3/C5), 72.77 (C2), 69.34 (C4), 60.63 (C6),
58.43 (ring C), 44.62 (ArCH2) 29.47, 24.16 (ring CH2). Found: C,
57.1; H, 7.0; N, 4.2%; M (mass spectrum), 650. C32H46N2O12·H2O
requires C, 57.5; H, 7.2; N, 4.2%; M, 650.
3
4
N ,N¢-Bis(2-(phenyl b-D-glucopyranoside)benzyl)-N ,N¢-di-
methyl-ethane-1,2-diamine (GL3). Compound 340 (2.04 g,
3.95 mmol) and N,N¢-dimethylethylenediamine (0.158 g,
1.79 mmol) were dissolved in THF (30 mL); NEt3 (0.399 g,
3.95 mmol) was then added. The reaction mixture was stirred
for 6 h and then the solvent was removed in vacuo. The residue
was partitioned between a saturated NaHCO3 (15 mL) solution
and CH2Cl2 (15 mL) and the aqueous portion was extracted
with CH2Cl2 (2 ¥ 20 mL). The organic extracts were combined,
dried over Na2SO4, filtered, and evacuated. The crude material
was purified by silica gel chromatography (95 : 5 CH2Cl2 : MeOH
eluent) to afford the acetylated product as a white solid (0.890 g,
52%). dH (300 MHz; CDCl3): 7.45 (2H, d, 3Jc,d = 7.3, ArH), 7.19
CDCl3): 7.18 (2H, dd, Ja,b = 8.5, Jb,d = 2.5, ArH), 6.97 (1H,
4
3
d, Jb,d = 2.5, ArH), 6.76 (1H, d, Ja,b = 8.5, ArH), 3.99 (4H, s,
ArCH2), 2.86 (4H, s, NCH2CH2N), 1.26 (18H, s, C(CH3)3). dC
(75.48 MHz; CDCl3): 155.41 (ArC), 141.84 (ArC), 125.55, 125.19
(ArCH), 121.35 (ArC), 115.76 (ArCH), 50.76 (NCH2C6H4OH),
48.79 (NCH2CH2N), 34.21 (C(CH3)3), 31.52 (C(CH3)3). Found:
C, 74.0; H, 9.4; N, 7.1%; M (mass spectrum), 384. C24H36N2O2
requires C, 73.6; H, 9.5; N, 7.15%; M, 384.
N,N¢-Bis(2-(phenyl
b-D-glucopyranoside)benzyl)-ethane-1,2-
diamine (GL1). Salicylaldehyde-b-D-glucopyranoside (helicin) 2
(0.461 g, 1.62 mmol) and ethylenediamine (0.048 g, 0.81 mmol)
were added to EtOH (12 mL) and the suspension was heated
to reflux for 24 h. The reaction mixture was filtered and the
3
3
3
(2H, dd, Jb,c = 7.2, Ja,b = 8.1, ArH), 7.05 (2H, dd, Jc,d = 7.3,
This journal is The Royal Society of Chemistry 2009
3036 | Dalton Trans., 2009, 3034–3043
©

View Article Online
3
3Jb,c = 7.2, ArH), 6.99 (2H, d, Ja,b = 8.1, ArH), 5.30 (4H, m,
as a white solid (1.18 g, 53%). dH (300 MHz; CDCl3): 7.54 (2H,
dd, 4Jb,d = 1.5, 3Jc,d = 7.8, ArH), 7.20 (12H, m, ArH), 6.95 (4H,
m, ArH), 5.25 (4H, m, H-2, H-3), 5.13 (2H, m, H-4), 4.95 (1H, d,
H-2, H-3), 5.17 (2H, dd, 3J2,3 = 9.5, 3J4,5 = 9.1, H-4), 5.06 (1H, d,
3
2
3J1,2 = 7.6, H-1), 4.26 (2H, dd, J5,6a = 5.5, J6a,6b = 12.2, H-6a),
3
2
3
2
4.15 (2H, dd, J5,6b = 2.3, J6a,6b = 12.2, H-6b), 3.84 (2H, ddd,
3J1,2 = 7.8, H-1), 4.24 (2H, dd, J5,6a = 5.4, J6a,6b = 12.3, H-6a),
3J4,5 = 9.5, J5,6a = 5.5, J5,6b = 2.3, H-5), 3.51 (4H, s, ArCH2),
2.57 (4H, s, NCH2CH2N), 2.21 (6H, s, NCH3), 2.06 (6H, s,
COCH3), 2.05 (6H, s, COCH3), 2.04 (6H, s, COCH3), 2.03
(6H, s, COCH3). Found: C, 55.45; H, 6.1; N, 2.9%; M (mass
spectrum), 960. C46H60N2O20·H2O requires C, 55.4; H, 6.5; N,
2.8%; M, 960. The acetylated compound (0.610 g, 0.63 mmol)
was dissolved in dry MeOH (12 mL) and whilst stirring NaOMe
(0.137 g, 2.54 mmol) was added. Rexyn (H+ form) was added
after 2 h, and then the mixture was filtered and the solvent was
removed in vacuo. The crude material was purified by silica gel
chromatography (1 : 1 EtOAc : MeOH eluent) followed by size-
exclusion chromatography on Sephadex G-10 (MeOH eluent)
4.15 (2H, dd, J5,6b = 2.1, J6a,6b = 12.3, H-6b), 3.90 (2H, ddd,
3
3
3
2
3J4,5 = 9.5, 3J5,6a = 5.4, 3J5,6b = 2.1, H-5), 3.56 (2H, d, 2Jgem = 13.5,
2
NCH2C6H4OH), 3.51 (2H, d, Jgem = 15.4, NCH2C6H5), 3.44
2
2
(2H, d, Jgem = 13.5, NCH2C6H4OH), 3.41 (2H, d, Jgem = 15.4,
NCH2C6H5), 2.55 (4H, m, NCH2CH2N), 2.03 (6H, s, COCH3),
2.02 (6H, s, COCH3), 2.01 (6H, s, COCH3), 1.89 (6H, s, COCH3).
Found: C, 59.3; H, 5.9; N, 2.4%; M (mass spectrum), 1112.
C58H69N2O20·3H2O requires C, 59.6; H, 6.5; N, 2.4%; M, 1112.
The acetylated product was dissolved in dry MeOH (20 mL) and
NaOMe (0.231 g, 4.28 mmol) was added with stirring. Rexyn
(H+ form) was added after 2 h, the mixture was filtered and the
solvent removed under reduced pressure. The crude material
was purified by silica gel chromatography (4 : 1 MeOH : CH3CN
eluent) and the isolated material was then triturated with acetone
(15 mL) to afford the product GL4 as a white solid (0.378 g, 46%).
dH (300 MHz; MeOD): 7.40 (12H, m, ArH), 7.28 (4H, m, ArH),
7.09 (2H, ddd,4Ja,c = 1.2, 3Jb,c = 7.2, 3Jc,d = 7.5, ArH), 4.85 (1H,
to afford the product GL3 as a white solid (0.181 g, 46%). dH
(300 MHz; MeOD): 7.41 (6H, m, ArH), 7.14 (2H, dd, Jc,d
3
=
3
3
7.3, Jb,c = 7.2, ArH), 4.94 (2H, d, J1,2 = 7.4, H-1), 4.12 (2H,
d, 2Jgem = 12.7, ArCH2), 4.01 (2H, d, 2Jgem = 12.7, ArCH2), 3.92
(2H, dd, 3J5,6b = 1.8, 2J6a,6b = 12.0, H-6b), 3.73 (2H, dd, 3J5,6a = 5.4,
2J6a,6b = 12.0, H-6a), 3.50 (8H, m, H-2, H-3, H-4, H-5), 3.09 (4H, s,
NCH2CH2N), 3.09 (6H, s, NCH3). dC (75.48 MHz; MeOD):
158.52 (ArC), 133.85, 132.47 (ArCH), 124.89 (ArC), 124.73,
118.31 (ArCH), 104.20 (C1), 78.94, 78.42 (C3/C5), 75.48 (C2),
71.74 (C4), 62.84 (C6), 57.52 (ArCH2), 53.44 (NCH2CH2N), 42.26
(NCH3). Found: C, 56.0; H, 7.45; N, 4.1%; M (mass spectrum),
624. C30H44N2O12·H2O requires C, 56.1; H, 7.2; N, 4.4%; M, 624.
3
d, J1,2 = 7.8, H-1), 3.92 (10H, m, NCH2C6H4OH, NCH2C6H5,
H-6b), 3.73 (2H, dd, 3J5,6a = 5.4, 2J6a,6b = 12.0, H-6a), 3.46 (8H, m,
H-2, H-3, H-4, H-5), 3.01 (4H, s, NCH2CH2N). dC (75.48 MHz;
MeOD): 158.43 (ArC), 137.06 (ArC), 133.35, 131.76, 131.59,
130.27, 129.80 (ArCH), 126.65 (ArC), 124.52, 117.88 (ArCH),
104.17 (C1), 78.92, 78.36 (C3/C5), 75.41 (C2), 71.73 (C4),
62.92 (C6), 60.50 (NCH2C6H4OH), 54.71 (NCH2C6H5) 50.79
(NCH2CH2N). Found: C, 62.2; H, 6.7; N, 3.3%; M (mass
spectrum), 777. C42H52N2O12·2H2O requires C, 62.1; H, 6.9; N,
3.45%; M, 777.
2-Formylphenyl-2,3,4,6-tetra-O-acetyl-b-D-glucopyranoside (4).
Helicin 2 (1.006 g, 3.54 mmol) was dissolved in pyridine (2 mL)
and the mixture cooled in an ice-bath. To this solution, acetic
anhydride (2.913 g, 28.54 mmol) was added dropwise. The solvent
was removed under reduced pressure after 2 h and the residue
dissolved in EtOAc (30 mL) and extracted with H2O (3 ¥ 50 mL).
The organic layer was dried with Na2SO4, filtered, and the solvent
removed in vacuo. The crude material was re-crystallised from hot
EtOH (10 mL) to afford 4 as colourless crystals (0.717 g, 83%).
4-tert-Butyl-2-formylphenyl-2,3,4,6-tetra-O-acetyl-b-D-gluco-
pyranoside (6). a-D-Glucopyranosyl bromide tetraacetate 5
(2.382 g, 5.79 mmol) and 5-tert-butyl-2-hydrozybenzaldehyde 1
(2.061 g, 11.56 mmol) were dissolved in acetone (20 mL) and
NaOH (4 mL, 5% w/w) was added dropwise. The reaction was
stirred for 24 h and then H2O (10 mL) was added. The reaction
mixture was extracted with CH2Cl2 (3 ¥ 20 mL), the combined
organic extracts were dried over Na2SO4, filtered, and then the
solvent was removed in vacuo. The residue was purified by silica
gel chromatography (1 : 1 Et2O : pet. ether eluent) to afford the
product 6 as a white solid (0.971 g, 33%). dH (300 MHz; CDCl3):
dH (300 MHz; CDCl3): 10.35 (1H, s, CHO), 7.86 (1H, dd, 4Jb,d
1.6, 3Jc,d = 7.5, ArH), 7.56 (2H, ddd, 4Jb,d = 1.6, 3Jb,c = 7.2, 3Ja,b
=
=
3
3
8.1, ArH), 7.19 (2H, dd, Jb,c = 7.2, Jc,d = 7.5, ArH), 7.11 (2H,
3
d, Ja,b = 8.1, ArH), 5.36 (2H, m, H-3, H-4), 5.18 (2H, m, H-1,
H-4), 4.30 (1H, dd, 3J5,6a = 5.1, 2J6a,6b = 12.3, H-6a), 4.18 (1H, dd,
3J5,6b = 2.7, 2J6a,6b = 12.3, H-6b), 3.90 (1H, ddd, 3J4,5 = 10.1, 3J5,6a
=
10.34 (1H, s, CHO), 7.87 (1H, d, Jb,d = 2.6, ArH), 7.59 (1H,
4
5.1, 3J5,6b = 2.7, H-5), 2.07 (3H, s, COCH3), 2.06 (3H, s, COCH3),
2.06 (3H, s, COCH3), 2.05 (3H, s, COCH3). Found: C, 55.8; H,
5.4%; M (mass spectrum), 452. C21H24O11 requires C, 55.75; H,
5.35%; M, 452.
dd, Ja,b = 8.7, Jb,d = 2.6, ArH), 7.04 (1H, d, Ja,b = 8.7, ArH),
3
4
3
3
5.34 (2H, m, H-2, H-3), 5.19 (1H, m, H-4), 5.15 (1H, d, J1,2
=
3
2
7.5, H-1), 4.30 (2H, dd, J5,6a = 5.2, J6a,6b = 12.4, H-6a), 4.18
(2H, dd, 3J5,6b = 2.4, J6a,6b = 12.4, H-6b), 3.89 (2H, ddd, J4,5
=
2
3
3
3
9.6, J5,6a = 5.2, J5,6b = 2.4, H-5), 2.08 (3H, s, COCH3), 2.06
(3H, s, COCH3), 2.05 (3H, s, COCH3), 2.02 (3H, s, COCH3), 1.31
(18H, s, C(CH3)3). Found: C, 57.9; H, 6.4%; M (mass spectrum),
508. C25H32O11·0.5H2O requires C, 58.0; H, 6.4%; M, 508.
N,N¢-Bis(2-(phenyl
b-D-glucopyranoside)benzyl)-N,N¢-di-
(0.909 g,
benzyl-ethane-1,2-diamine (GL4). Compound
4
2.01 mmol) and N,N¢-dibenzylethylenediamine41 (0.256 g,
1.05 mmol) were dissolved in 1,2-dichloroethane (3 mL) and
sodium (tris)acetoxyborohydride (0.632 g, 3.00 mmol) was added
in portions. The reaction was stirred for 24 h and then quenched
with a saturated NaHCO3 (10 mL) solution. The reaction solution
was then extracted with EtOAc (3 ¥ 10 mL), the organic fractions
dried with Na2SO4, filtered, and the solvent removed in vacuo.
The crude material was purified by silica gel chromatography
(97.5 : 2.5 CH2Cl2 : MeOH eluent) to afford the acetylated product
N,N¢-Bis(2-(4-tert-butylphenyl-b-D-glucopyranoside)benzyl)-
ethane-1,2-diamine (GL5). 4-tert-Butyl-2-formylphenyl-2,3,4,6-
tetra-O-acetyl-b-D-glucopyranoside 6 (0.971 g, 1.91 mmol) and
ethylenediamine (0.057 g, 0.95 mmol) were mixed and the reaction
mixture was heated to reflux for 24 h. The precipitate was then
washed with EtOH and Et2O to afford the imine as a light yellow
solid. dH (400.13 MHz; DMSO-d6): 8.45 (2H, s, imine H), 7.80
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 3034–3043 | 3037
©

View Article Online
4
3
4
(2H, d, Jb,d = 2.5, ArH), 7.49 (2H, dd, Ja,b = 8.7, Jb,d = 2.5,
ArH), 7.08 (2H, d, Ja,b = 8.7, ArH), 5.46 (2H, d, J1,2 = 7.9,
flat-bottomed 96-well plates (Nalge-Nunc, Rochester, NY); each
test well was loaded first with buffer (100 uL), Ab and metal
ion solutions (25 mL each). After a short incubation time the pro-
ligand solution (50 mL ) was added; each turbidity trial was done in
quadruplicate to allow for statistical analysis of error. Positive and
negative controls were used to ensure the validity of the results. The
plate was loaded into a Molecular Devices Thermomax microplate
reader programmed to agitate the plate for 30 seconds before
measuring absorbance at 405 nm. To ensure that the test solutions
were not responsible for the observed absorbance at 405 nm, blank
readings were taken with all well components except Ab peptide;
the tetrahydrosalen compounds did exhibit significant absorbance
at 405 nm, and this absorbance was accounted for in both cases
with the use of such blanks. The absorbance values obtained from
the test wells (containing Ab) were normalised with respect to
the appropriate blank solutions. The amount of absorbance is
indicative of the amount of aggregated Ab in the well, which is
inversely proportional to the efficacy of the test compound in the
disaggregation assay.
3
3
3
3
H-1), 5.42 (2H, dd, J2,3 = 9.0, J3,4 = 9.1, H-3), 5.14 (2H, dd,
3J2,3 = 9.0, J1,2 = 7.9, H-2), 5.00 (2H, dd, J3,4 = 9.1, J4,5
=
=
3
3
3
3
2
9.5, H-4), 4.22 (2H, m, H-5), 4.19 (2H, dd, J5,6a = 5.2, J6a,6b
2
12.6, H-6a), 4.08 (2H, br d, J6a,6b = 12.6, H-6b), 3.79 (4H, m,
NCH2CH2N), 2.01 (6H, s, COCH3), 2.00 (6H, s, COCH3), 1.99
(6H, s, COCH3), 1.98 (6H, s, COCH3), 1.24 (18H, s, C(CH3)3).
Found: C, 59.9; H, 6.65; N, 2.7%; M (mass spectrum), 1040.
C36H56N2O12 requires C, 60.0; H, 6.6; N, 2.7%; M, 1040. The
imine (0.791 g, 0.67 mmol) was added to MeOH (15 mL), and
to this suspension NaBH4 (0.0251 g, 0.67 mmol) was added. The
mixture heated to reflux, cooled to room temperature after 3 h,
and stirred for a further 16 h before adding H2O (2 mL) to quench
the reaction. The solvent was removed and the crude material was
purified by silica gel chromatography (1 : 1 EtOAc : MeOH eluent)
to afford the product GL5 as a white solid (0.186 g, 39% over
two steps). dH (300 MHz; MeOD): 7.26 (4H, m, ArH), 7.13 (2H,
3
3
d, Ja,b = 8.7, ArH), 4.83 (2H, d, J1,2 = 7.4, H-1), 3.88 (2H, d,
2Jgem = 12.5, ArCH2), 3.86 (br d, 2J6a,6b = 12.0, H-6b), 3.65 (2H, dd,
3J5,6a = 5.0, 2J6a,6b = 12.0, H-6a), 3.60 (2H, d, 2Jgem = 12.5, ArCH2),
3.41 (8H, m, H-2, H-3, H-4, H-5), 2.70 (4H, s, NCH2CH2N), 1.27
(18H, s, (C(CH3)3)). dC (75.48 MHz;MeOD):155.80(ArC), 146.64
(ArC), 129.13 (ArC), 128.95, 126.86, 117.00 (ArCH), 103.97 (C1),
78.33, 78.17 (C3/C5), 75.02 (C2), 71.45 (C4), 62.66 (C6), 50.76
(NCH2C6H4OH ), 48.79 (NCH2CH2N), 35.13 (C(CH3)3), 32.08
(C(CH3)3). Found: C, 61.2; H, 8.2; N, 4.2%; M (mass spectrum),
708. C36H56N2O12 requires C, 61.0; H, 8.0; N, 3.95%; M, 708.
Trolox equivalent antioxidant capacity (TEAC) assay
The tetrahydrosalen compounds (H2L1–5) were tested for their
ability to scavenge free radicals using the TEAC assay.44 The
results were compared to antioxidant standards such as 6-
hydroxy-2,3,7,8-tetramethylchroman-2-carboxylic acid (Trolox),
( )-a-tocopherol, and butylated hydroxytoluene (BHT). A 2,2¢-
azinobis-(3-ethylbenzothiazoline-6-sulfonic acid (ABTS∑+) radical
cation decolourisation assay44 was used to determine relative
TEAC values. ABTS was dissolved in water (7 mM) and reacted
with potassium persulfate (2.45 mM) in the dark for 16 h. During
this time ABTS oxidises to the colored ABTS∑+ radical cation. The
ABTS∑+ solution was diluted with MeOH to an absorbance of 0.70
( 0.02) at 740 nm after equilibrating to 30 ◦C (Fisher Isotemp
circulating water bath). Stock solutions of the tetrahydrosalen
compounds in MeOH were diluted so that addition of 20 mL to
2 mL of ABTS∑+ solution caused a reduction of 20–80% in A740 nm
as a result of the reduction to ABTS. The final concentrations
of the pro-ligand ranged from 5–15 mM to obtain the necessary
absorbance change. After initial mixing, the A740 nm was measured
at 30 ◦C after 1, 3 and 6 minutes with each experiment performed
in triplicate. The percentage change of the absorbance at 740 nm
was calculated and plotted as a function of tetrahydrosalen
concentration. The slopes were then compared to the standard
Trolox, with its TEAC value normalised to 1.
Turbidity assay
Two representative tetrahydrosalen compounds H2L1 and H2L3
were tested for their ability to attenuate metal ion-promoted
aggregation of human amyloid peptide 1–40 (Ab1–40) fol-
lowing an established protocol.30,42,43 4-(2-Hydroxy-ethyl)-1-
piperazineethanesulfonic acid (HEPES) buffer was prepared by
diluting HEPES (4.76 g, 20.0 mmol) and NaCl (8.76 g, 150 mmol)
with water (~850 mL) near to the full volume of 1000 mL. Chelexꢀ
resin was added to the solution to rid the buffer of any residual
transition metal ions that may confound results. The Chelexꢀ-
R
R
containing buffer solution was incubated at 4 ◦C for 16 h. The
buffer was allowed to warm to room temperature, the pH adjusted
with 1 M NaOH (to pH 6.6 for Cu and pH 7.4 for Zn) and made
up to the 1000 mL mark. Separate buffers were prepared for the
Cu- and Zn-mediated experiments, but both contained the same
amount of HEPES, NaCl and Chelexꢀ. Chelexꢀ was removed by
filtration through a Millipore 0.22 mm acetate filter which also
ensured minimal contamination by other solids.
R
R
Enzymatic glycoside cleavage
Human amyloid-beta peptide (Ab1–40, 1 mg) was weighed and
diluted with water (1.155 mL); the mixture was sonicated for one
minute, allowedtostandfor 30seconds, sonicatedfor an additional
minute, and then filtered through a Whatman 0.2 mm nylon
syringe filter. Separate solutions of ethylenediaminetetraacetic
acid (EDTA, 14.5 mg), H2L1 (10.8 mg), and H2L3 (6.0 mg) were
prepared at 200 mM in the appropriate volume of HEPES buffer;
pH 6.6 for Cu(II) assay, pH 7.4 for Zn(II) assay. Zn(II) (200 mM, in
pH 7.4 HEPES) and Cu(II) (200 mM, in pH 6.6 HEPES) solutions
were prepared by diluting Aldrich Atomic Absorption Standard
Solutions. Aggregation trials were performed in clear polystyrene
Preliminary reactivities of all five glycosylated tetrahydrosalen
compounds with Agrobacterium sp. b-glucosidase45 (Abg) were
tested and monitored by TLC and +ES-MS. To 1 mL Eppendorf
tubes at 25 ◦C, glycosylated tetrahydrosalen compound (20 mL,
20 mM), phosphate buffer (50 mL, 75 mM, pH 7.4), and Abg (5 mL,
178 mg mL-1) were added and the resulting solution mixed. After
2 h the reactions were monitored by TLC (1 : 1 MeOH : EtOAc
eluent) using a UV lamp and charring of the TLC plates (5%
H2SO4 in EtOH) to follow the cleavage process. The reactions
were also followed by mass spectrometry (+ES-MS) to correlate
spectral changes with de-glycosylation.
3038 | Dalton Trans., 2009, 3034–3043
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Cell toxicity studies
hydrolytic decomposition. In addition, the increased flexibility
of the tetrahydrosalen compounds allows for the stabilisation of
multiple coordination geometries. Deprotonated H2L1 has been
extensively studied as a ligand for Fe(III),37 Sn(IV),50 Cr(III),37
Zn(II),34 Cu(II),34 Ni(II),34,37 and Co(II).34,37 Deprotonated H2L2
has been studied as a ligand for Cr(III)37 as well as for Sn(IV).50
Deprotonated H2L3 has been investigated as a ligand for model
Mo(VI/V) centers of Mo hydroxylases and related enzymes,38 as
well as for potential catalytic applications complexed to Al(III)39
and Ti(IV).51 Al(III) complexes of H2L4 have been investigated for
catalytic purposes.39
A representative glycosylated tetrahydrosalen compound (GL3)
was used in this study to ascertain preliminary toxicity values.
Human breast cancer cells were used (MDA-MB-435S)46 that had
previously been cultured and prepared with experimental details
described elsewhere.47 Cells were transferred to a 96-well plate by
addition of 100 mL of the cell solution (1 ¥ 104 cells) to 54 of the
wells. Growth medium (100 mL) was used as blanks in another 6
wells. The plate was incubated at 37 ◦C for 24 h and then solutions
(100 mL) of GL3 at 8 different concentrations (10–5000 mM) in
phosphate buffered saline (PBS) were added to 48 wells. The
remaining 6 wells containing cells served as the control; PBS was
added (100 mL) and the plate was incubated at 37 ◦C for 3 days. An
identical procedure was used for cisplatin in this study. Cell toxicity
was quantified by the MTT assay (MTT = 3-(4,5-dimethylthiazol-
2-yl)-2,5-diphenyltetrazolium bromide).48 The percentage cell via-
bility was calculated by dividing the average absorbance of the cells
treated with GL3 or cisplatin by that of the control. The percent
cell viability vs. drug concentration (logarithmic scale) was plotted
to determine the IC50 (drug concentration at which 50% of the cells
are viable relative to the control), with the estimated error derived
from the average of 6 trials.
The synthesis of the carbohydrate-protected tetrahydrosalen
compounds (GL1–5) proceeded by a variety of routes, either by re-
ductive amination, or N-alkylation of the diamines with the appro-
priate carbohydrate-protected phenols. All of the carbohydrate-
protected tetrahydrosalen compounds were isolated as their
b-anomers to afford potential substrates for the b-glucosidase en-
zymes. Initial attempts to directly glycosylate the tetrahydrosalen
compounds H2L1–5 using the Koenigs–Knorr reaction52 with either
silver carbonate or silver triflate and a-D-glucopyranosyl bromide
tetraacetate 5 were unsuccessful. Steric interference may inhibit
the formation of the bis-glycosylated tetrahydrosalen compounds
53
by this route. Glycoside bond formation utilising BF3·OEt2 and
pentaacetylglucose was also attempted, however this route was
not successful. Boron complexation by the tetrahydrosalen com-
pounds, analogous to reported boron–pyridinone complexes,54
could explain why BF3·OEt2 may not be an ideal coupling reagent.
Phenols were thus glycosylated prior to the formation of the
tetrahydrosalen framework to circumvent the lack of reactivity
discussed above. Using this general route, further structural
diversity could easily be obtained through the use of a series of
diamines.
Both GL1 and GL2 were synthesised by reaction of the
appropriate diamine with two equivalents of helicin 2 (Scheme 2).
The isolated imine was then reduced using NaBH4 in MeOH
to afford GL1 and GL2, in 50% and 71% yield, respectively,
over two steps. GL2 is a more rigid version of GL1 exhibiting
Results and discussion
Synthesis
The synthesis of H2L5 (Scheme 1) closely followed the lit-
erature preparations of H2L1–4; in each case the appropriate
salicylaldehyde was coupled with a diamine and either iso-
lated (as imine) or immediately reduced in situ (as iminium
ion) to afford the corresponding tetrahydrosalen compounds.
A variety of symmetric primary (ethylenediamine, (1R,2R)-
(-)-1,2-diaminocyclohexane) and secondary (N,N¢-dimethyl-
ethylenediamine, N,N¢-dibenzylethylenediamine) diamines were
investigated to afford a series of tetrahydrosalen compounds for
further study. It was postulated that the presence of the t-butyl
group para to the phenol moiety in H2L5 would enhance the
antioxidant potential of this compound.
Scheme 1 Synthesis of H2L5: (a) ethylenediamine, EtOH. (b) NaBH4,
MeOH; 71% yield for two steps.
Metal complexes of tetrahydrosalen compounds are not as com-
mon as are those of the corresponding salen analogues. This can
be attributed to the widespread interest in metal–salen complexes
for catalytic applications,49 as well as the increased difficulty in
synthesising the saturated analogues. However, unlike the imine
Scheme 2 Synthesis of GL1 and GL2: (a) diamine (ethylenediamine)
(GL1), (1R,2R)-(-)-1,2-diaminocyclohexane (GL2), EtOH. (b) NaBH4,
MeOH. Yield over two steps: 50% (GL1), 71% (GL2).
=
CH N linkages in salen compounds, the CH2–NH functionality
in tetrahydrosalen compounds is very stable with respect to
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 3034–3043 | 3039
©

View Article Online
a 1,2-diaminocyclohexane backbone. This rigidity is known to
increase metal complex stability. GL1 and GL2 were found to be
associated with one molecule of water in the solid state, a result
not unexpected as the glycosylated tetrahydrosalen compounds
discussed here are hygroscopic.
GL3 was synthesised as shown in Scheme 3. N-Alkylation of
N,N¢-dimethylethylenediamine with two equivalents of compound
3 in the presence of NEt3 and subsequent acetyl group removal
using NaOMe in MeOH afforded GL3 in moderate yield. Com-
pound 3 was seen as a useful building block for the attachment
of glycoside-protected phenols to suitable compounds such as
amines and other nucleophiles. GL3 could also be efficiently
synthesised by the reductive amination of helicin (2) with N,N¢-
dimethylethylenediamine. Initial attempts to synthesiseGL4 via the
route used for GL3 were only partially successful. The coupling
of N,N¢-dibenzylethylenediamine with compound 3 resulted in
the isolation of acetyl-protected GL4 in only 9% yield. However,
a reductive amination reaction between aldehyde 4 and N,N¢-
dibenzylethylenediamine, using sodium tris(acetoxy)borohydride,
and subsequent acetyl group removal afforded GL4 in moderate
yield (Scheme 4).
Scheme 5 Synthesis of GL5: (a) NaOH, acetone; 33% yield. (b) Ethylene-
diamine, EtOH. (c) NaBH4, MeOH; 39% yield over two steps.
a food preservatives, and ( )-a-tocopherol (vitamin E), found
naturally in foods. The tetrahydrosalen compounds H2L1–5 were
monitored for antioxidant activity by the TEAC assay.44 This assay
has been used as a simple method to quantify the antioxidant
activity of biological fluids, extracts, and pure compounds by
measuring the disappearance of the ABTS∑+ radical cation via UV-
visible spectroscopy. The ability of the tetrahydrosalen compounds
to quench the ABTS∑+ radical cation was compared to Trolox,
a more water soluble analogue of a-tocopherol, and the results
are shown in Fig. 1. H2L1–5 displayed a similar ability to quench
the ABTS∑+ radical cation and exhibited TEAC values that were
enhanced in comparison to ( )-a-tocopherol and BHT. It should
be noted however that each tetrahydrosalen compound contains
two phenolic moieties capable of quenching the ABTS∑+ radical
cation. Introduction of a t-butyl group para to the ring hydroxyl
moiety in H2L5 did not increase the antioxidant properties of this
analogue in comparison to the other compounds in the series.
H2L1–5 were not as potent antioxidants as our recently reported
glycosyl-functionalised N2O2 chelators,30 which are essentially
Scheme 3 Synthesis of GL3: (a) N,N¢-dimethylethylenediamine, NEt3,
THF; 52% yield. (b) NaOMe, MeOH; 46% yield.
Scheme
4
Synthesis of GL4: (a) N,N¢-dibenzylethylenediamine,
NaBH(OAc)3, CH2ClCH2Cl; 53% yield (b) NaOMe, MeOH; 46% yield.
The synthesis of GL5 is shown in Scheme 5. Coupling of 5-tert-
butyl-2-hydroxybenzaldehyde 1 and a-D-glucopyranosyl bromide
tetraacetate 5 using a published protocol,55 afforded compound
6 in 33% yield. Reaction of two equivalents of 6 with ethylenedi-
amine formed the bis-imine, which was then reduced with NaBH4.
Under these conditions the sugar acetyl groups were also removed
to afford GL5 in one step.
TEAC values of the tetrahydrosalen compounds
Phenols are known to react with lipid radicals,35,36 ultimately
acting as chain breaking antioxidants.36 Examples of common
phenolic antioxidants are butylated hydroxytoluene (BHT) and
butylated hydroxyanisole (BHA), synthetic compounds used as
Fig. 1 Trolox Equivalent Antioxidant Capacity (TEAC) values at 1, 3,
and 6 minutes for H2L1–5, ( )-a-tocopherol, and BHT. Error bars represent
SD above and below the mean TEAC value (determined in triplicate).
3040 | Dalton Trans., 2009, 3034–3043
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
protected hydroquinones; oxidation to the associated quinones
could be the reason for the enhanced antioxidant properties as
compared to H2L1–5. It is clear from this study that the tetrahy-
drosalen compounds have the potential to act as antioxidants.
reduced this absorption by varying extents; lower absorbance
of the resultant solution indicates greater disaggregation efficacy
for the pro-ligand of interest (Fig. 2). Ethylenediaminetetraacetic
acid (EDTA) was used as a positive control, as it is known to
attenuate metal ion–Ab aggregation.29 The effect of representative
pro-ligands on metal-induced Ab1–40 aggregation was assessed at
physiological pH (7.4) for Zn2+ and at lower pH (6.6) for Cu2+.
The interaction of the Ab peptide with Cu2+ is known to be pH-
dependent with no aggregation apparent at physiological pH.18
Enzymatic cleavage of the tetrahydrosalen glycosides
The tetrahydrosalen glycosides GL1–5 require that the glucose
protecting group is removed before metal binding can take place.
It was hypothesised that this could occur enzymatically in vivo
as there are a number of enzymes that cleave b-glycosides in
humans.56 Examples of such enzymes include a cytostolic broad
specificity b-glucosidase (EC 3.2.1.21), glycosylceramidase (EC
3.2.1.62), and the membrane-bound lactase-phlorizin hydrolase
(EC 3.2.1.62/108).56,57 As the human glucosidase enzymes are
difficult to obtain, and more than one enzyme could potentially act
as a site for glycoside cleavage, rat brain homogenate has been used
as a model system to determine overall b-glucosidase activity.58,59
Using this method, glucosidase activity of approximately 4–
4.5 nmol h-1 mg-1 brain homogenate has been determined in male
Wistar rats.58 Additionally, b-glucosidase activity was found to be
very low in plasma, signifying that carbohydrate prodrugs should
be stable in this medium.59
Initial screening of the glycosylated tetrahydrosalen compounds
GL1–5 was undertaken to determine if the glycoside moieties
could be enzymatically cleaved to afford the corresponding pro-
ligands H2L1–5. A broad specificity b-glucosidase, Agrobacterium
b-glucosidase (Abg),45 was used to test for glycoside cleavage and
the reactions were monitored by thin layer chromatography (TLC)
and mass spectrometry. Of the five glycosylated tetrahydrosalen
compounds, only GL4 did not act as a substrate for Abg.
In all other cases significant reactivity was observed by TLC
(Fig. S1, ESI†). The additional steric bulk of the benzyl groups,
compared to the other tetrahydrosalen compounds, could inhibit
the interaction of GL4 with Abg. Mass spectrometric (+ES-MS)
analysis of the diluted enzyme reactions correlated nicely with the
results from the TLC study (Fig. S2, ESI†). In all cases, except GL4,
mass peaks corresponding to the loss of either one or two glucose
moieties were present in the enzyme reactions. The reactions of
GL1 and GL2 with Abg were not complete at 2 h, as small
peaks corresponding to the glycosylated starting materials were
evident in addition to the expected deglycosylated compounds.
This preliminary study is quite encouraging and shows that four
of the five glycosylated prodrugs undergo glycoside cleavage in the
presence of the Abg enzyme to afford the desired tetrahydrosalen
pro-ligands.
Fig. 2 Tetrahydrosalen compounds H2L1 and H2L3 as inhibitors of metal
ion-promoted Ab1–40 aggregation (“M” = Zn2+, pH 7.4, black; or Cu2+
,
pH 6.6, gray), by absorbance measurement at 405 nm (A405). Error bars
represent SD above and below the mean (n ≥ 3). A low relative absorbance
indicates high disaggregation efficiency for the test compound.
In order to better understand the disaggregation activity of the
N2O2 ligand system we compared the metal binding affinity of
H2L1, H2GL1 (Chart 2; ref. 30), and EDTA at specific pH values
(Table 1). For a given complex equilibrium system and pH, the
concentration of the unchelated metal ion (referred to as pM =
-log[Munchelated]) represents a direct estimate of the ligand–metal
affinity taking into account all relevant equilibria.62,63 As stability
constants are a measure of the overall stability of a metal–ligand
species, they do not reflect the metal-binding affinity of ligands at a
specific pH value, and thus a direct comparison can be misleading.
The zinc(II) trials demonstrate that, while both tetrahydrosalen
compounds are less effective for Zn-aggregated Ab dissolution
as compared to the strong metal ion chelator EDTA, both H2L1
and H2L3 significantly reduce the amount of aggregated Ab in
the solution. The relative disaggregation activity parallels the
calculated pZn values in Table 1.
Ab aggregation inhibition studies
Table 1 Calculated pM (-log[M]unchelated; M = Zn2+, Cu2+) for a solution
containing a 1 : 1 mixture of indicated ligand and metal ion ([M2+
]
=
Zn2+ and Cu2+ are known to cause Ab aggregation in solution.18,27
A turbidity assay was performed to investigate the ability of repre-
sentative tetrahydrosalen pro-ligands (H2L1 and H2L3) to dissolve
such metal ion-aggregated Ab species in vitro. Synthetic human
Ab1–40 was dissolved in a buffered aqueous solution to which was
added the metal ion (Cu2+ or Zn2+). Metal ion addition caused Ab
aggregation, which, through formation of a turbid solution, was
measurable by absorption spectrophotometry; suspended solids
(such as aggregated Ab) scattered light and increased the apparent
absorbance. Subsequent addition of metal-binding pro-ligands
tot
[ligand]tot = 0.001 M)
pZn
pCu
Compound
pH 6.6
pH 6.6
pH 7.4
H2L1a
4.2
4.2
8.3
7.2
6.8
9.0
8.5
7.9
9.5
H2GL1b
EDTAc
a Calculated from ref. 34. b From ref. 30. c Values calculated from ref. 60,61.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 3034–3043 | 3041
©

View Article Online
30
The increased Cu-binding affinity of H2L1 over H2GL1,
at
ing into synthesis of radiolabeled derivatives for biodistribution
studies. Our initial cell studies show a representative glycosylated
tetrahydrosalen (GL3) to be non-toxic in vitro. The results reported
here suggest that GL1–5 have the potential to act as multifunctional
Alzheimer’s disease therapeutics.
pH 6.6 (Table 1) is reflected in the slightly enhanced disaggregation
activity of H2L1 in this assay. Interestingly, the relative disaggre-
gation activity of EDTA and H2L1 are approximately equal for
the Cu trials, despite the large difference in pM values (Table 1).
We speculate that the increased disaggregation activity of H2L1,
above that expected based on metal-binding affinity, could be
due to a number of factors including increased hydrophobicity
and/or intercalating ability of this tetrahydrosalen compound.64,65
These features may allow for improved interaction with the Cu–Ab
binding site, and subsequent disaggregation. Further experiments
will be required to understand the dominant processes involved,
but the current results highlight the potential benefit of molecules
capable of multiple complimentary interactions in inhibiting
metal–Ab aggregation.
Acknowledgements
The authors gratefully acknowledge the Canadian Institutes of
Health Research (CIHR) for a Proof of Principle grant. T. S.
acknowledges NSERC for a postgraduate scholarship; L. E. S.
recognises support from the Alzheimer Society of Canada and
the Institute of Aging (CIHR) for a doctoral training award, the
Province of British Columbia for a Pacific Century Scholarship,
and the University of British Columbia for a University Graduate
Fellowship. We thank Dr David Kennedy and Dr Elena Polishchuk
for their assistance with the cell toxicity assay, and Dr Dana Devine
and her lab for allowing us the use of their plate reader.
Cell toxicity studies
Human breast cancer cells46 were used in this study to determine
the biological activity (toxicity) of a representative glycosylated
tetrahydrosalen compound (GL3), in comparison to the cytotoxic
agent cisplatin (Fig. S3, ESI†). Ideally, GL3 would not be cytotoxic
in vitro over a large dosage range. An MTT assay, which is
an in vitro colourimetric method of determining cell viability,
was used here to assess for cell toxicity.47 The results of the
MTT assay clearly show that GL3 is not toxic in this cell line
at the concentrations studied (0.01–5 mM). Cisplatin was toxic,
exhibiting an IC50 value of 35 5 mM. A toxicity study of the
full series of tetrahydrosalen compounds, and their glycosylated
analogues, is warranted in a more relevant cell model.
Notes and references
1 K. J. Barnham and C. L. Masters, Nat. Rev. Drug Discovery, 2004, 3,
205.
2 P. M. Doraiswamy and A. E. Finefrock, Lancet Neurol., 2004, 3, 431.
3 C. P. Ferri, M. Prince, C. Brayne, H. Brodaty, L. Fratiglioni, M.
Ganguli, K. Hall, K. Hasegawa, H. Hendrie, Y. Huang, A. Jorm, C.
Mathers, P. R. Menezes, E. Rimmer and M. Scazufca, Lancet, 2005,
366, 2112.
4 I. Melnikova, Nat. Rev. Drug Discovery, 2007, 6, 341.
5 E. D. Roberson and L. Mucke, Science, 2006, 314, 781.
6 E. Gaggelli, H. Kozlowski, D. Valensin and G. Valensin, Chem. Rev.,
2006, 106, 1995.
7 M. P. Mazanetz and P. M. Fischer, Nat. Rev. Drug Discovery, 2007, 6,
464.
8 M. G. Spillantini, M. L. Schmidt, V. M.-Y. Lee, J. Q. Trojanowski, R.
Jakes and M. Goedert, Nature, 1997, 388, 839.
Summary
In this work a series of tetrahydrosalen compounds H2L1–5, and
their glycosylated analogues GL1–5, were synthesised and charac-
terised by a variety of methods. The synthetic routes reported here
are amenable to the future synthesis of derivatives with poten-
tially improved pharmacokinetic properties. The tetrahydrosalen
compounds displayed significant antioxidant ability in an ABTS∑+
radical cation competition assay. In addition, the glycosylated ana-
logues, with the exception of GL4, were determined to be substrates
for Abg, indicating the potential of the carbohydrate prodrug
strategy. H2L1–5 are expected to have a moderate affinity for Cu2+
and Zn2+,30,34 and will likely compete with the Ab peptide for these
metal ions.30 Two representative tetrahydrosalen compounds were
subjected to in vitro turbidity assays and demonstrated significant
ability to inhibit both Zn2+- and Cu2+-mediated aggregation of the
amyloid beta peptide. Interestingly, H2L1 was found to have an
activity comparable to that of EDTA in the Cu-mediated Ab1–40
disaggregation trials, despite the lower pCu of H2L1. This may
be attributable to a preferable association between the aromatic
elements of H2L1 and the amyloid-beta peptide in solution; in any
case, it highlights the importance of structure—beyond the metal-
chelating moiety—of any metal–protein attenuating compound
for inhibition of metal ion-mediated Ab aggregation.
9 J. Hardy and D. J. Selkoe, Science, 2002, 297, 353.
10 J. A. Hardy and G. A. Higgins, Science, 1992, 256, 184.
11 D. J. Selkoe, Physiol. Rev., 2001, 81, 741.
12 K. Hensley, N. Hall, R. Subramaniam, P. Cole, M. Harris, M. Aksenov,
M. Aksenova, S. P. Gabbita, J. F. Wu, J. M. Carney, M. A. Lovell and
W. R. Markesbery, J. Neurochem., 1995, 65, 2146.
13 M. A. Smith, P. L. R. Harris, L. M. Sayre and G. Perry, Proc. Natl.
Acad. Sci. USA, 1997, 94, 9866.
14 M. A. Lovell, J. D. Robertson, W. J. Teesdale, J. L. Campbell and W. R.
Markesbery, J. Neurol. Sci., 1998, 158, 47.
15 C. J. Frederickson, J.-Y. Koh and A. I. Bush, Nat. Rev. Neurosci., 2005,
6, 449.
16 C. S. Atwood, M. E. Obrenovich, T. Liu, H. Chan, G. Perry, M. A.
Smith and R. N. Martins, Brain Res. Rev., 2003, 43, 1.
17 A. I. Bush, W. H. Pettingell, G. Multhaup, M. d. Paradis, J.-P. Vonsattel,
J. F. Gusella, K. Beyreuther, C. L. Masters and R. E. Tanzi, Science,
1994, 265, 1464.
18 C. S. Atwood, R. D. Moir, X. Huang, R. C. Scarpa, M. E. Bacarra,
D. M. Romano, M. A. Hartshorn, R. E. Tanzi and A. I. Bush, J. Biol.
Chem., 1998, 273, 12817.
19 X. Huang, C. S. Atwood, M. A. Hartshorn, G. Multhaup, L. E.
Goldstein, R. C. Scarpa, M. P. Cuajungco, D. N. Gray, J. Lim,
R. D. Moir, R. E. Tanzi and A. I. Bush, Biochemistry, 1999, 38,
7609.
20 R. A. Cherny, C. S. Atwood, M. E. Xilinas, D. N. Gray, W. D. Jones,
C. A. McLean, K. J. Barnham, I. Volitakis, F. W. Fraser, Y. S. Kim,
X. D. Huang, L. E. Goldstein, R. D. Moir, J. T. Lim, K. Beyreuther, H.
Zheng, R. E. Tanzi, C. L. Masters and A. I. Bush, Neuron, 2001, 30,
665.
The moderate affinity of H2L1–5 for metal ions at physiological
pH may reduce the toxicity commonly associated with chelation
therapy, especially with the potential for increased tissue specificity
due to the attached carbohydrate functions. Investigation is ongo-
21 J.-Y. Lee, J. E. Friedman, I. Angel, A. Kozak and J.-Y. Koh, Neurobiol.
Aging, 2004, 25, 1315.
22 D. R. Crapper McLachlan, A. J. Dalton, T. P. Kruck, M. Y. Bell, W. L.
Smith, W. Kalow and D. F. Andrews, Lancet, 1991, 337, 1304.
3042 | Dalton Trans., 2009, 3034–3043
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
23 B. Regland, W. Lehmann, I. Abedini, K. Blennow, M. Jonsson, I.
Karlsson, M. Sjogren, A. Wallin, M. Xilinas and C. G. Gottfries,
Dement. Geriatr. Cogn. Disord., 2001, 12, 408.
43 X. Huang, C. S. Atwood, R. D. Moir, M. A. Hartshorn, J.-P. Vonsattel,
R. E. Tanzi and A. I. Bush, J. Biol. Chem., 1997, 272, 26464.
44 R. Re, N. Pellegrini, A. Proteggente, A. Pannala, M. Yang and C.
Rice-Evans, Free Radical Biol. Med., 1999, 26, 1231.
45 J. B. Kempton and S. G. Withers, Biochemistry, 1992, 31, 9961.
46 B. Brinkley, P. Beall, L. Wible, M. Mace, D. Turner and R. Cailleau,
Cancer Res., 1980, 40, 3118.
47 A. Wu, D. C. Kennedy, B. O. Patrick and B. R. James, Inorg. Chem.,
2003, 42, 7579.
48 T. Mosmann, J. Immunol. Methods, 1983, 65, 55.
49 L. Canali and D. C. Sherrington, Chem. Soc. Rev., 1999, 28, 85.
50 R. Garcia-Zarracino, J. Ramos-Quinones and H. Hopfl, J. Organomet.
Chem., 2002, 664, 188.
24 C. W. Ritchie, A. I. Bush, A. Mackinnon, S. Macfarlane, M. Mastwyk,
L. MacGregor, L. Kiers, R. Cherny, Q.-X. Li, A. Tmmer, D. Carrington,
C. Mavros, I. Volitakis, M. Xilinas, D. Ames, S. Davis, K. Beyreuther,
R. E. Tanzi and C. L. Masters, Arch. Neurol., 2003, 60, 1685.
25 P. A. Adlard, R. A. Cherny, D. I. Finkelstein, E. Gautier, E. Robb, M.
Cortes, I. Volitakis, X. Liu, J. P. Smith, K. Perez, K. Laughton, Q.-X.
Li, S. A. Charman, J. A. Nicolazzo, S. Wilkins, K. Deleva, T. Lynch,
G. Kok, C. W. Ritchie, R. E. Tanzi, R. Cappai, C. L. Masters, K. J.
Barnham and A. I. Bush, Neuron, 2008, 59, 43.
26 S. Singh, H. Khodr, M. I. Taylor and R. C. Hider, Biochem. Soc. Symp.,
1995, 61, 127.
27 A. Dedeoglu, K. Cormier, S. Payton, K. A. Tseitlin, J. N. Kremsky, L.
Lai, X. Li, R. D. Moir, R. E. Tanzi and A. I. Bush, Exp. Gerontol.,
2004, 39, 1641.
51 J. Balsells, P. J. Carroll and P. J. Walsh, Inorg. Chem., 2001, 40, 5568.
52 J. Conchie and G. A. Levvy, in Methods In Carbohydrate Chemistry,
ed. R. L. Whistler and M. L. Wolfrom, Academic Press, New York,
USA, 1963, vol. II, pp. 335.
28 L. K. Charkoudian, D. M. Pham, A. M. Kwon, A. D. Vangeloff and
K. J. Franz, Dalton Trans., 2007, 5031.
53 Y. Mikata, K. Yoneda, T. Tanase, I. Kinoshita, M. Doe, F. Nishida, K.
Mochida and S. Yano, Carbohydr. Res., 1998, 313, 175.
54 W. O. Nelson, C. Orvig, S. J. Rettig and J. Trotter, Can. J. Chem., 1988,
66, 132.
55 K. Kohata, H. Higashio, Y. Yamaguchi, M. Koketsu and T. Odashima,
Bull. Chem. Soc. Jpn., 1994, 67, 668.
56 P. M. Coutinho and B. Henrissat, Carbohydrate-Active Enzymes server
at http://afmb.cnrs-mrs.fr/CAZY/. Accessed February 2006.
57 H. Matern, H. Boermans, F. Lottspeich and S. Matern, J. Biol. Chem.,
2001, 276, 37929.
58 J. D. McCarter, M. J. Adam, N. G. Hartman and S. G. Withers,
Biochem. J., 1994, 301, 343.
29 H. Schugar, D. E. Green, M. L. Bowen, L. E. Scott, T. Storr, K.
Bo¨hmerle, F. Thomas, D. D. Allen, P. R. Lockman, M. Merkel, K. H.
Thompson and C. Orvig, Angew. Chem., Int. Ed., 2007, 46, 1549.
30 T. Storr, M. Merkel, G. X. Song-Zhao, L. E. Scott, D. E. Green, M. L.
Bowen, K. H. Thompson, B. O. Patrick, H. J. Schugar and C. Orvig,
J. Am. Chem. Soc., 2007, 129, 7453.
31 F. Bonina, C. Puglia, M. Rimoli, D. Melisi, G. Boatto, M. Nieddu, A.
Calignano, G. La Rana and P. de Caprariis, J. Drug Targeting, 2003,
11, 25.
32 J. C. La Manna and S. I. Harik, Brain Res., 1985, 326, 299 .
33 A. A. Qutub and C. A. Hunt, Brain Res. Rev., 2005, 49, 595.
34 W. Gruenwedel, Inorg. Chem., 1968, 7, 495.
35 G. W. Burton and K. U. Ingold, J. Am. Chem. Soc., 1981, 103, 6472.
36 F. Ponticelli, A. Trendafilova, M. Valoti, S. Saponara and G. Sgaragli,
Carbohydr. Res., 2001, 330, 459.
59 C. Fernandez, O. Nieto, E. Rivas, G. Montenegro, J. A. Fontenla and
A. Fernandez-Mayoralas, Carbohydr. Res., 2000, 327, 353.
60 A. E. Martell and R. M. Smith, Critical Stability Constants, Plenum
Press, New York, USA, vol. 1–6, 1974–1989.
61 A. E. Martell and R. D. Hancock, Metal Complexes in Aqueous
37 P. Baran, A. Bottcher and H. Elias, Z. Naturforsch., 1992, 47b, 1681.
38 C. J. Hinshaw, G. Peng, R. Singh, J. T. Spence, J. H. Enemark, M. Bruck,
J. Kristofzski, S. L. Merbs, R. B. Ortega and P. A. Wexler, Inorg. Chem.,
1989, 28, 4483.
Solutions, Plenum Press, New York, USA, 1996.
62 Y. Chiu and J. W. Canary, Inorg. Chem., 2003, 42, 5107.
63 C. J. Fahrni and T. V. O’Halloran, J. Am. Chem. Soc., 1999, 121,
11448.
39 P. Hormnirun, E. L. Marshall, V. C. Gibson, A. J. P. White and D. J.
Williams, J. Am. Chem. Soc., 2004, 126, 2688.
40 J. C. Briggs, A. H. Haines and R. J. K. Taylor, J. Chem. Soc., Perkin
Trans. 1, 1995, 27.
41 A. E. Frost and H. H. Freedman, J. Org. Chem., 1959, 24, 1905.
42 L. E. Scott, B. D. G. Page, B. O. Patrick and C. Orvig, Dalton Trans.,
2008, 6364.
64 R. A. Cherny, K. J. Barnham, T. Lynch, I. Volitakis, Q.-X. Li, C. A.
McLean, G. Multhaup, K. Beyreuther, R. E. Tanzi, C. L. Masters and
A. I. Bush, J. Struct. Biol., 2000, 130, 209.
65 C. Rodriguez-Rodriguez, N. Sanchez de Groot, A. Rimola, A. Alvarez-
Larena, V. Lloveras, J. Vidal-Gancedo, S. Ventura, J. Vendrell, M.
Sodupe and P. Gonzalez-Duarte, J. Am. Chem. Soc., 2009, 131, 1436–
1451.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 3034–3043 | 3043
©