The design of multifunctional antioxidants against the damaging ingredients of oxidative stress

Article abstract of DOI:10.1080/10426500801898200

Oxidative stress is a biochemical condition associated with a sharp increase in intracellular concentrations of a range of oxidative stressors, including reactive oxygen species, reactive nitrogen species and labile metal ions. It is associated with a wide range of human disorders, such as inflammatory diseases, neurodegenerative diseases, glaucoma, and cancer. Equally importantly, oxidative stress is pronounced in older people, which makes it an important matter in an ageing Society. Not surprisingly, antioxidants have become a major focus of modern drug development. While natural antioxidants, such as phenolic aromatic compounds, vitamin C, vitamin E and curcumin have shown promising results, the development of effective synthetic antioxidants is often problematic. We have recently proposed the rational design of multifunctional antioxidants which target oxidative stressors in a more comprehensive manner. Such compounds may, for instance, combine catalytic sites with metal binding sites. Here we present the synthesis of representative molecules with combined catalytic and metal binding properties. The apparent 'antioxidant' activities have been studied in vitro and, for the most promising compound, have been confirmed in cultured skin cells exposed to UVA radiation.

Full text of DOI:10.1080/10426500801898200

This article was downloaded by: [York University Libraries]
On: 22 November 2014, At: 08:03
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK
Phosphorus, Sulfur, and Silicon
and the Related Elements
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/gpss20
The Design of Multifunctional
Antioxidants Against the
Damaging Ingredients of
Oxidative Stress
Susanne Mecklenburg a , Catriona A. Collins b ,
Mandy Döring a , Torsten Burkholz a , Muhammad
Abbas a , Fiona H. Fry b , Charareh Pourzand c &
Claus Jacob a
a Division of Bioorganic Chemistry, School of
Pharmacy , Saarland University , Saarbruecken,
D-66041, Germany
b School of Biosciences , University of Exeter ,
Exeter, EX4 4 QD, United Kingdom
c Department of Pharmacy and Pharmacology ,
University of Bath , Claverton Down, Bath, BA2 7AY,
United Kingdom
Published online: 29 Oct 2008.
To cite this article: Susanne Mecklenburg , Catriona A. Collins , Mandy Döring ,
Torsten Burkholz , Muhammad Abbas , Fiona H. Fry , Charareh Pourzand & Claus Jacob
(2008) The Design of Multifunctional Antioxidants Against the Damaging Ingredients
of Oxidative Stress, Phosphorus, Sulfur, and Silicon and the Related Elements, 183:4,
863-888, DOI: 10.1080/10426500801898200
To link to this article: http://dx.doi.org/10.1080/10426500801898200
PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness,
or suitability for any purpose of the Content. Any opinions and views
expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the
Content should not be relied upon and should be independently verified with
primary sources of information. Taylor and Francis shall not be liable for any
losses, actions, claims, proceedings, demands, costs, expenses, damages,
and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the
Content.
This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan,
sub-licensing, systematic supply, or distribution in any form to anyone is
expressly forbidden. Terms & Conditions of access and use can be found at
http://www.tandfonline.com/page/terms-and-conditions

Phosphorus, Sulfur, and Silicon, 183:863–888, 2008
Copyright © Taylor & Francis Group, LLC
ISSN: 1042-6507 print / 1563-5325 online
DOI: 10.1080/10426500801898200
The Design of Multifunctional Antioxidants Against
the Damaging Ingredients of Oxidative Stress
Susanne Mecklenburg,1 Catriona A. Collins,2
1
1
1
¨
Mandy Doring, Torsten Burkholz, Muhammad Abbas,
Fiona H. Fry,2 Charareh Pourzand,3 and Claus Jacob1
1Division of Bioorganic Chemistry, School of Pharmacy, Saarland
University, D-66041 Saarbruecken, Germany
2School of Biosciences, University of Exeter, Exeter EX4 4 QD,
United Kingdom
3Department of Pharmacy and Pharmacology, University of Bath,
Claverton Down, Bath BA2 7AY, United Kingdom
Oxidative stress is a biochemical condition associated with a sharp increase in intra-
cellular concentrations of a range of oxidative stressors, including reactive oxygen
species, reactive nitrogen species and labile metal ions. It is associated with a wide
range of human disorders, such as inflammatory diseases, neurodegenerative dis-
eases, glaucoma, and cancer. Equally importantly, oxidative stress is pronounced
in older people, which makes it an important matter in an ageing Society. Not sur-
prisingly, antioxidants have become a major focus of modern drug development.
While natural antioxidants, such as phenolic aromatic compounds, vitamin C, vi-
tamin E and curcumin have shown promising results, the development of effective
synthetic antioxidants is often problematic. We have recently proposed the rational
design of multifunctional antioxidants which target oxidative stressors in a more
comprehensive manner. Such compounds may, for instance, combine catalytic sites
with metal binding sites. Here we present the synthesis of representative molecules
with combined catalytic and metal binding properties. The apparent ‘antioxidant’
The authors are indebted to the following colleagues in the UK and Germany for their
tireless support and advice with synthesis, electrochemistry, complex chemistry and, of
course, working with FEK4 and FCP7 cells: Rex Tyrrell (University of Bath), Jennifer
Littlechild (University of Exeter), James Tucker (University of Birmingham), Kaspar
Hegetschweiler (Saarland University) and Awais Anwar (Saarland University).
The authors would also like to acknowledge financial support from Saarland Uni-
versity, the Ministry of Culture of the Saarland, Fresenius Medical Care (stipend T.B.),
DAART (UK) (stipend for C.A.C.), EATL (for F.H.F.) and the University of Exeter. This
financial support has not caused any conflict of interest to any of the participating au-
thors.
Address correspondence to Claus Jacob, Division of Bioorganic Chemistry, School of
Pharmacy, Universitaet des Saarlandes, Campus B 2.1, P O Box 151150 D-66041, Saar-
bruecken, Germany. E-mail: c.jacob@mx.uni-saarland.de
863

864
S. Mecklenburg et al.
activities have been studied in vitro and, for the most promising compound, have
been confirmed in cultured skin cells exposed to UVA radiation.
Keywords Antioxidants; catalysis; metal binding; radiation damage; selenium
INTRODUCTION
Oxidative stress (OS) is a biochemical condition associated with a sig-
nificant increase of oxidative stressors, such as reactive oxygen species
(ROS), peroxynitrite and free radicals.1 It can have various causes,
among them an impaired antioxidant cellular defense, and manifests
itself in a range of human disorders. The latter may, for instance, in-
clude various auto-inflammatory diseases, Rheumatoid Arthritis, neu-
rodegenerative diseases (including Alzheimer’s disease and Parkinson’s
disease), glaucoma, diabetes, and cancer.
Importantly, OS is not only a hallmark of many human illnesses, but
also occurs in humans as part of the normal ageing process.1 Although
OS does not immediately cause illnesses in older people, it nonethe-
less poses a certain kind of “burden” for them. It may, for instance,
ultimately lead to various disorders associated with OS, most notably
perhaps inflammatory diseases and cancer. As a result, OS has become
an important center point of research into the causes and possible treat-
ments of OS-related diseases.
While OS itself is seldom the ultimate cause of these diseases, it can
be seen as a point of intervention, either preventive or therapeutically.
Within this context, antioxidants provide a certain protection from OS.
Research in this group of compounds has increased dramatically dur-
ing the last decade, and antioxidants such as vitamin C, α-tocopherol,
curcumin and various natural phenolic aromatic compounds (e.g., cat-
echins, quercetin derivatives), have been associated with disease pre-
ventive and possibly therapeutic properties.
Among the most interesting antioxidants are catalytic antioxidants,
such as the antioxidant enzymes superoxide dismutase (Cu,Zn-SOD
and Mn-SOD), catalase (CAT), glutathione peroxidase (GPx) and per-
oxiredoxins (Prx).1,2 Compared to simple reducing agents, such as glu-
tathione (GSH), these natural catalysts have several significant ad-
vantages. First, they react rapidly with oxidative stressors, such as the
superoxide radical anion and H2O2. In many cases, the rate constants
of these catalytic reactions are orders of magnitude higher than the
ones of un-catalyzed interactions, e.g., the reduction of H2O2 by GSH
takes hours in the absence of GPx, and seconds in its presence. Sec-
ondly, the “chemistry” of these catalysts is tightly controlled, i.e., they
produce harmless side-products when compared to some of the reaction

Multifunctional Antioxidants
865
products of other antioxidants. And, perhaps most importantly, cata-
lysts are recycled, and at minute concentrations can perform the same
ROS reduction reaction thousands of times, compared to just once by
one-shot antioxidants, which need to react in stoichiometric amounts.3
Not surprisingly, the development of synthetic antioxidants, which
mimic the catalytic activity of these enzymes, has recently been at the
forefront of antioxidant development. Among the antioxidant enzymes
in question, the selenium enzyme GPx provides a particularly inter-
esting template for mimics.4 Selenium (and tellurium) compounds are
more stable chemically and metabolically when compared to metal-
complexes mimicking the SOD and CAT enzymes. At the same time,
comparably simple Se compounds are able to catalyze the reaction of
H2O2 with GSH to H2O and GSSG effectively. In contrast, potential
Prx-mimics are more difficult to devise, since the catalytic sulfur in
these enzymes forms a transient sulfenic acid (RSOH), which is fine in
the active site of the enzymes, but too reactive in most small molecules,
where it readily reacts with thiols to form a disulfide and water.
Nonetheless, the notion that one antioxidant enzyme mimic alone
may mount an effective antioxidant defense is hampered by the fact
that OS is a multi-stressor event.1 By simply targeting one of the many
oxidative stressors, other stress events will largely go unaffected and
continue to damage the cell. We have therefore recently proposed the
use of multifunctional antioxidants to impair oxidative stressors more
widely (Figure 1).5 At the time, two oxidative stressors, H2O2 and labile,
redox active metal ions, such as Cu+/2+ and Fe2+/3+ were chosen as pos-
sible targets. Both stressors form key junctions at OS redox cascades:
H2O2 is the follow-on product of O•− and a direct precursor of the highly
2
reactive hydroxyl radical (HO•). It is also a substrate of myeloperoxi-
dase and hence a precursor of hypochlorous acid, HOCl. Labile copper
and iron ions, which may have been liberated from proteins by ROS,
catalyze the reduction of H2O2 to HO• in Fenton-type reactions. Hy-
droxyl radicals, in turn, may attack metalloproteins and liberate more
copper and iron ions, entering a vicious, oxidizing cycle involving H2O2
and copper/iron ions.1
As part of our previous studies, selenium derivatives of pyridine, ani-
line and quinoline, i.e., 2,2ꢀ-diselane-1,2-diyldipyridine, 2,2ꢀ-diselane-
1,2-diyldianiline, 8,8ꢀ-diselane-1,2-diyldiquinoline, respectively, were
synthesized (Figure 2).5 These compounds were chosen as represen-
tatives of different classes of nitrogen/selenium compounds which may
bind metals via the N and Se atom. They were subsequently evaluated
for catalytic, metal binding and antioxidant activity in a range of in
vitro assays and fibroblast (FEK4 and FCP7) cell culture. In essence,
2,2ꢀ-diselane-1,2-diyldianiline exhibited the best protection against cell

866
S. Mecklenburg et al.
FIGURE 1 The cascade of reactive oxygen species (ROS) leading to cell damage
and cell death. This cascade of ROS offers several points of intervention, among
them catalysts with SOD-like and GPx-like activity and removal of labile copper
and iron ions to prevent HO• formation. Together, these points of intervention
provide the basis for the design of multifunctional agents discussed here. Other
approaches, such as inhibition of SOD by 2-methoxyestradiol or inhibition of
myeloperoxidase, have also been discussed in the literature, yet are not free
of possible complications (such as O•2− or H2O2 build-up, respectively).1,31 In-
hibitions of Nox and Duox are also possible. Please note that this sketch of
biochemical processes is necessarily incomplete. Reduction of H2O2 to HO• by
metal ions may be catalytic in the presence of intracellular reducing agents.
damage induced by UVA-radiation.5 This compound was more active in
this cell culture model when compared to the benchmark catalytic an-
tioxidant ebselen. It also exhibited a comparably low electrochemical
potential and interacted with Cu2+ ions, as shown by Differential Pulse
Polarography and UV/VIS spectrophotometry.
Nonetheless, this compound primarily exerted its effects in the fi-
broblast cells by raising the total level of GPx-like activity. In contrast,
there was little evidence of interaction with the labile iron pool, as mea-
sured by using the Calcein loading assay. In this respect, the weakly
coordinating nitrogen/selenium-ligands were obviously inferior to the
classical metal chelator, desferrioxamine (DFO), which is currently used
to treat metal-related aspects of OS. In turn, the employment of consid-
erably higher concentrations of Se-compounds, which may ultimately
affect the labile iron pool, was unrealistic.
In order to address this apparent problem, a set of new com-
pounds has now been chosen to combine GPx-like catalytic activity

Multifunctional Antioxidants
867
FIGURE 2 Multifunctional agents tested previously. While 2,2ꢀ-diselane-
1,2-diyldianiline, and to a lesser extent 2,2ꢀ-diselane-1,2-diyldipyridine, pro-
tect cultured FCP7 fibroblast cells against UVA radiation, 8,8ꢀ-diselane-1,
2-diyldiquinoline acts as photosensitizer.5 Mechanistic studies in cell culture
indicate that the protective effect associated with these compounds is mostly
the result of GPx-like activity.
with enhanced metal binding properties. Here, we briefly describe the
synthesis of three representative compounds, their electrochemical be-
havior, and interaction with metal ions. We also demonstrate a rather
promising antioxidant activity for one of the compounds in cultured
FEK4 cells under UVA radiation-induced OS.
RESULTS
Figure 3 shows the three multifunctional compounds (1–3) selected for
this study. These compounds were chosen because they combine a cat-
alytic Se atom with a metal binding site. The catalytic site is similar
to the one found in other, conventional GPx-mimics, with selenium at-
tached to one aromatic residue for increased chemical and metabolic
stability. The metal binding sites of these compounds are based on well-
known copper/iron binding sites, i.e., tacn (1,4,7-triazacyclononane), cy-
clen (1,4,7,10-tetraazacyclododecane) and protoporphyrin IX. The latter
is particularly interesting, since it not only lends itself to the seques-
tration of adventitious metal ions, but may also provide the precursor

868
S. Mecklenburg et al.
for metalloporphyrins and their associated, often catalytic activity (e.g.
as SOD-mimics).6−8
Synthesis of Compounds
Compounds 1–3 were synthesized successfully, albeit at rather mod-
est yields. The individual synthetic pathways are shown in Figure 3a
to Figure 3c. All three compounds were not previously reported in the
1
literature. Analytical data obtained for these compounds, such as H
NMR, 13C NMR and LC-MS confirmed the structures proposed, whilst
HPLC ascertained the purity of the compounds required for biochemical
and cell culture studies. Within this context, the synthesis of compounds
1 and 2 was based on the coupling of a selenium-containing precursor
on the well-known macrocycles tacn and cyclen. In case of 2, two al-
ternative avenues were followed (Figure 3b), both of which resulted in
the desired product at comparable yields. In contrast, the synthesis of
3 was based on rather straightforward acid/amine coupling procedures
between protoporphyrin IX and the Se-containing amine precursor (Fig-
ure 3c). The two coupling methods employed both led to the formation
of 3—with the ethylchloroformate method the milder and in the end
more promising avenue.
Interaction of Compounds with Metal Ions
Interactions with metal ions (Cu2+, Fe2+, Fe3+, and Zn2+) were evalu-
ated using Cyclic Voltammetry (for 1 and 2), and because of appropriate
spectroscopic properties, with UV/VIS spectrophotometry (for 3).
Cyclic Voltammetry was used to investigate the influence of com-
pounds on the redox behavior of Cu2+ ions, which are particularly
FIGURE 3 Synthesis of compounds 1–3. Details of the synthetic pathways are
provided in the text. The sulfur analogue of 3 was synthesized according to the
same method. (Continued)

Multifunctional Antioxidants
869
FIGURE 3 Continued
damaging when acting as Fenton-type catalysts during OS. Figure 4
illustrates the interactions of 1 and 2 with Cu2+ ions. Under the ex-
perimental conditions chosen, the copper ion exhibits a reduction peak
at Epc = −180 mV and an oxidation signal Epa = +330 mV vs. the
Ag/AgCl reference electrode (SSE). The one electron transfer between

870
S. Mecklenburg et al.
FIGURE 3 Continued
Cu2+ and Cu+ is quasi-reversible (ꢀE = 510 mV), with a formal redox
potential of E0ꢀ = +75 mV. Addition of 1 and 2 results in a depression
of both, the oxidation and reduction signal of the copper ion, which
disappear at a ligand to metal ratio of approximately 1:1 (Figure 4a

Multifunctional Antioxidants
871
FIGURE 4 Electrochemical changes of copper oxidation/reduction upon addi-
tion of 1 and 2 (Figure 4a and 4b, respectively). The addition of both compounds
resulted in a significant decrease of peak currents. Ratios of compound to Cu2+
were 0, 0.33, 0.5 and 1.0 for (1) to (4). Experimental details are given in the
text. Similar results were also obtained when DFO was used. A similar effect
was observed for the electrochemical behavior of iron ions.

872
S. Mecklenburg et al.
for 1, Figure 4b for 2). There were also slight shifts in Epa and Epc
upon addition of ligand, but these changes were within the margins of
experimental error. In order to confirm that metal binding effects were
responsible for the disappearance of the reduction and oxidation waves,
DFO was used as a positive control. This compound also led to a com-
plete depression of the oxidation and reduction peaks of the Cu2+/Cu+
redox pair. Similar results were obtained with ironions: The reduction
peak of Fe3+, which was observable under the conditions used at Epa
= −400 mV, decreased significantly upon addition of 1, 2 and DFO.9
While compound 3 was too insoluble to be used in electrochemical
experiments (concentrations of around 100 μM are required to produce
reliable results by Cyclic Voltammetry), its spectrophotometric proper-
ties could be used to assess metal binding and exchange.
As expected, the protoporphyrin-containing molecule was able to in-
teract with copper, iron and zinc ions. Binding of Cu2+, Fe2+, and Fe3+
resulted in significant changes to the compound’s UV/VIS absorption
spectrum, especially in the region between 450 and 700 nm (Figure 5).
Importantly, all metals were added as chlorides, hence ruling out any
specific effects due to different counter anions, such as ligand binding.
In short, addition of Cu2+ resulted in a decrease of absorbance bands
at 407 nm (Soret band), 504, 541, 576, 628, and 669 nm. In turn, there
were distinct signals for the copper complex at 532, 569, and 602 nm
(Figure 5a).10 The interaction with Fe2+ resulted in a slight shift and
decrease of the Soret band from 407 nm to 400 nm. As for copper, the
smaller bands at 504, 541, 576, 628, and 669 nm all decreased (or were
shifted) with new transitions at 535, 561, and 608 nm. Addition of Fe3+
slightly decreased and shifted the Soret band to 411 nm, resulted in
the characteristic decreases/shifts of absorbancies between 500 and
700 nm and, like Fe2+, led to characteristic signals at 535, 560, and
606 nm. In contrast, Zn2+ ions, which are known to bind strongly to
protoporphyrin IX, had only a marginal effect on the electronic spec-
trum of 3 with a slight decrease and shift in the Soret band from 407 to
411 nm.
Similar results were obtained for protoporpyrin IX itself, which was
used as a control and to find out if modification of the protoporphyrin
with the Se-moiety causes any significant changes in the macrocycle’s
behavior toward metal ions. The similarity of results for 3, its protopor-
phyrin precursor, and its sulfur-analogue (which was tested alongside
3) indicates that this is clearly not the case.
Since there were significant differences between the spectra of 3 in
the presence of copper and iron on the one hand, and zinc on the other,
it was possible to qualitatively ascertain metal preferences of 3. The
behavior of the metal-loaded forms of 3 in the presence of equimolar

Multifunctional Antioxidants
873
FIGURE 5 Interaction of 3 with metal ions as monitored by UV/VIS spec-
trophotometry. Spectra were recorded in CHCl3 at 20◦C for biochemically rele-
vant Cu2+, Fe2+, Fe3+ and Zn2+ ions and are shown for Cu2+ and Fe2+ (Figure
5a and 5b, respectively). The absorption spectrum of 3 (10 μM) is shown as
broken lines ( ) before and as solid lines (—-) after addition of metal ions (100
μM). Addition of Cu2+ and Fe2+ ions (as CuCl2 and FeCl2 in methanol, respec-
tively) resulted in significant changes in the spectrum, which are detailed in
the text. The addition of 100 μM Fe3+ (as FeCl3) resulted in changes almost
identical to the ones observed for Fe2+. Addition of Zn2+ (as ZnCl2) did not alter
the spectrum of 3 significantly.

874
S. Mecklenburg et al.
amounts and 3-fold excess of other metals was therefore monitored.
The behavior of Cu2+ and Fe2+/Fe3+ in the presence of Zn2+ was of
particular interest: Is 3 able to uptake a redox-active, damaging metal
ion (copper or iron) and simultaneous release/ignore a more benign,
antioxidant metal (zinc)?
In short, addition of equimolar amounts or a 3-fold excess of Cu2+
,
Fe2+, and Fe3+ over Zn2+ to the zinc-form of 3 resulted in spectra charac-
teristic of the Cu2+, Fe2+ and Fe3+ forms of 3, while addition of equimo-
lar amounts or a 3-fold excess of Zn2+ over Cu2+ or Fe2+ to the copper- or
iron-form of 3 did not change the spectra characteristic of the copper-
or iron complex. The results for Cu2+ and Fe2+ are shown in Figure 6.
Taken together, they indicate that Cu2+ and Fe2+ ions are indeed taken
up by 3 in the presence of Zn2+, an effect also observed for Fe3+ (spec-
trum not shown). Please note that these results are of a qualitative
nature, performed in an organic solvent and clearly must be followed
by a full complexometric study in the future.
FIGURE 6 Metal exchange at the metal binding site of 3. Spectra were
recorded as described for Figure 5. The zinc complex (broken lines (- - - -)) was
formed by addition of 100 μM ZnCl2 to 10 μM of 3. Addition of 300 μM Cu2+ or
Fe2+ resulted in spectral changes indicative of the formation of the Cu2+ and
r r r
Fe2+ complex (solid (—) and dotted line (
), respectively). The same
changes were observed for 100 μM Cu2+ or Fe2+. This qualitative experiment
demonstrates that Cu2+ and Fe2+ are able to compete with/substitute for Zn2+
The resulting “copper (iron) for zinc” behavior has significant biochemical im-
plications as described in the text.
.

Multifunctional Antioxidants
875
The Catalytic Site
The redox properties of 1–3 were investigated using electrochemistry
(Cyclic Voltammetry), the thiophenol (PhSH) and the metallothionein
(MT) assays.5,11 Cyclic Voltammetry provides initial information about
oxidation and reduction processes associated with the selenium atom,
such as the position of the oxidation peak Epa. The PhSH assay is
indicative of the reaction of H2O2 with GSH. The MT assay measures
zinc release from this zinc-sulfur protein in the presence of a peroxide,
namely tBuOOH.
Compounds 1 and 2 exhibited a Se-oxidation peak (Epa) at around
+1200 mV vs. SSE. The lack of a corresponding reduction signal is
evidence of an irreversible oxidation of the Se-atom to a selenium cation,
which reacts further in aqueous solution. The Epa values for 1 and 2
are rather positive and point towards electrochemically active, but not
particularly reducing selenium-compounds.12
The rather modest ‘activity’ of the selenium redox center was con-
firmed in the PhSH and MT assays.5 In the PhSH assay, 1–3 only
slightly enhanced the rate of PhSH oxidation in the presence of H2O2
(1.9-fold increase by 1, 2.5-fold by 2, approx. 1.1-fold increase by 3 from
8 nMs−1 for the non-catalyzed reaction, respectively). Compound 1 was
also tested in the MT assay, where it enhanced zinc release from MT in
the presence of tBuOOH within 60 min from 15 to 23%. In comparison,
selenols tested previously led to rates of up to 600 nMs−1 in the PhSH
assay and almost 80% zinc release in the MT assay.
The in vitro studies identified 1 and 2 as the more promising com-
pounds, i.e., as (modest) reducing agents/GPx-like catalysts with good
metal binding properties and a reasonable solubility in aqueous media.
The efficiency of 3 remained somewhat vague due to low solubility in
aqueous media. Since tacn has previously been associated with antiox-
idant activity, it was decided to test 1 in FEK4 fibroblast cell culture
initially.
Antioxidant Activity in Fibroblast Cell Culture
In order to determine the antioxidant capacity of 1, FEK4 cells were
grown and exposed to 500 kJ m−2 UVA radiation for 18 h in the ab-
sence and presence of 1 and tacn as Se-free control.13 The subsequently
performed MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide) cell survival assay indicates that 50 μM of 1 is able to prevent
virtually all OS damage caused by UVA radiation, while tacn itself
is only modestly active at this concentration (Figure 7). Importantly,
neither tacn nor 1 were themselves cytotoxic at the concentrations

876
S. Mecklenburg et al.
†
100
80
60
40
20
0
†
*
Control
Control Control
+ tacn
UVA
UVA
UVA
+ tacn
+ 1
+ 1
FIGURE 7 Protective effects of 1 (50 μM) on FEK4 fibroblast cells exposed to
500 kJm−2 UVA radiation. Cells were incubated with 1 or tacn 18 h previous to
exposure. Cell viability was determined 4 h after irradiation using the standard
MTT assay.5,13,14 Tacn (50 μM), which reflects the metal binding chemistry of 1,
was used for comparison. Data are means SD of 3 samples from 3 independent
experiments. ∗Significantly different from non-irradiated control; †Significantly
different from UVA-irradiated control, P < 0.05.
used. Indeed, separate studies measuring the cytotoxicity of compounds
confirmed that 1 was non-toxic (i.e., cell viability > 90%) in FEK4 cells
up to 100 μM, while the benchmark compound ebselen already showed
toxic effects at 15 μM concentrations (measured in FCP7).
In order to find out if the protective effects of 1 are due to either
GPx-like catalysis, metal binding, or both, various cell-based assays
were performed, and, where appropriate, the results obtained for 1 were
compared to the ones for ebselen, tacn or DFO as representatives of
catalytic and metal binding activity, respectively.
The protective effect of tacn in FEK4 cells against UVA radiation,
as measured in the MTT assay, was smaller compared to the one of 1:
While 1 provided virtual complete protection (39% absolute increase in
cell viability), tacn only partially protected the cells (15% increase in

Multifunctional Antioxidants
877
viability). These results point towards a significant, yet not exclusive
share of the Se moiety in the protective effect of 1.
This initial notion was supported by an investigation of the adeno-
sine triphosphate (ATP) levels in cells exposed to 500 kJ m−2 UVA
radiation.14 While ATP levels in cells without protection dropped by
90% upon radiation, they remained at 62% of the control value in the
presence of 50 μM of 1. DFO (at 100 μM) showed a similar protection,
keeping ATP levels at 58% of the control. At 50 μM, 1 also led to a 2.2-
fold increase in GPx-like activity measured in FEK4 cells. Although this
effect was considerably less than the one observed for ebselen (4.0-fold
increase at 5 μM), it confirmed that 1 can contribute to the catalytic
antioxidant activity in the cell, albeit at a modest level - which is in
accordance with the results obtained in the PhSH and MT assay.
In summary, 1 provided strong protection of FEK4 cells against 500
kJ m−2 UVA radiation at concentrations not toxic to these cells and by
a mechanism probably combining elements of metal binding (like DFO)
and modest GPx-like catalysis (like ebselen).
DISCUSSION
The results obtained in this study support the notion that multifunc-
tional compounds combining selenium and metal binding (here: macro-
cycle) chemistry may act as rather effective antioxidants against the
oxidative damage induced by UVA radiation. Since the most promising
compound, 1, also has low toxicity in the cells tested, i.e., can be used in
reasonable concentrations, this finding is interesting in itself and may
be used for the future development of antioxidants in the area of skin
protection, for instance in sunscreens.
This study has also provided a number of in vitro and cell culture
results, which are worth considering in more detail. First of all, the
catalytic and metal binding properties associated with 1 in vitro and
in FEK4 cells are somewhat surprising. Rather than providing high
catalytic activity, 1 is only modestly active in the PhSH and MT as-
says, which is also reflected in cell culture, where GPx-like activity
is only increased 2.2-fold. Although this can be explained mechanisti-
cally by considering the catalytic cycle (Figure 8), it is surprising from
a biochemist’s point-of-view: Apparently, high catalytic activity alone
does not equate the perfect antioxidant. This notion is supported by
the fact that some excellent GPx-like catalysts are actually cytotoxic in
the presence of H2O2, probably due to random use of intracellular thiol
substrates.15−19 In contrast, more modest catalysts may behave less
aggressively in cells, yet may still provide some additional GPx-like
activity to protect the cells.

878
S. Mecklenburg et al.
FIGURE 8 Chemical basis of protective and regulatory effects associated with
1 in cell culture. The Se-based redox center is able to catalytically remove H2O2
in the presence of a thiol, most likely GSH. At the same time, the tacn site is
able to interact with metal ions, such as copper and iron (possibly exchanging
these metal ions for others, such as zinc). In addition to these events, GPx-like
catalysis and metal binding may also trigger intracellular redox signaling (e.g.,
via changes in GSH and GSSG levels), yield a copper- or iron-based SOD mimic
(most likely for 3) and provide an antioxidant response via zinc release. Please
note: The image of the copper-complex is for illustration only. Copper is likely
to bind to additional ligands, including H2O, Cl− or a second tacn molecule to
achieve tetra- or hexa-coordination. The precise stoichiometry of metal binding
in vivo has not yet been studied; it will depend on a number of factors and may
ultimately vary.6,7

Multifunctional Antioxidants
879
A similar notion applies to metal binding. Extraordinarily strong
metal chelating agents are known to be toxic in cells due to their ability
to remove metal ions from proteins and enzymes. In contrast, modest
metal binding agents provide protection against labile metal ions with-
out ‘freeloading’ on protein bound metal ions.20 As such, tacn provides
some protection against UVA induced OS which is almost certainly due
to sequestration of copper and/or iron ions. This effect is significantly
enhanced, however, in the presence of the Se-moiety in 1. In short, the
combination of modest catalysis with modest metal binding provides
good protection, while stronger agents may ultimately turn out to be
toxic to the cell.
One of the aims of this study was the design and synthesis of chem-
ically simple, yet biochemically quite active agents. The design of com-
pounds was guided by the biochemical principle of catalytic H2O2 re-
moval on the one hand and sequestration of labile copper (and iron)
ions to counteract Fenton-type events on the other. The synthesis of
the resulting compounds 1–3 could be carried out in a few steps and by
using comparatively simple chemical procedures, albeit in comparably
low yield. The latter was due to inherent problems associated with Se
chemistry, the requirement for highly purified samples, and the initial
lack of optimization of the synthetic method, which will be performed
as part of a follow-up study.
A second aim of the study was a comparison of mono- with dise-
lenides, such as the compounds evaluated by us previously (Figure 2).
The switch from diselenides to monoselenides results in a rather differ-
ent redox chemistry (Figure 8), which manifests itself in a range of in
vitro findings. Diselenides are reduced to selenol(ate) by reversible two-
electron transfer, which generally occurs at mercury (or other metal)
electrodes around −700 to −800 mV vs. SSE. In contrast, monoselenides
are irreversibly oxidized by one-electron transfer at around +800 to
+1200 mV, depending on the nature of substituents attached to the
Se atom.11,21 In this context, 1 and 2 show an oxidation behavior typ-
ical of Se oxidation, but with rather high Epa values. This value is
generally lower for diarylselenides, whereby the aromatic system en-
dows Se with a lower Epa value and the compound with increased
chemical and metabolic stability. Future addition of more aromatic
groups may therefore lower the Epa value of compounds such as 1
and 2 and increase their activity. On the other hand, such aromatic
systems tend to decrease solubility in water, which may ultimately be
counter-productive.
Together with the redox cycle characteristic for monoselenides, the
Epa values of 1 and 2 may also explain their lower activity in the PhSH
and MT assays when compared to the diselenides shown in Figure 2.5 To

880
S. Mecklenburg et al.
counter-balance this loss of GPx-like activity, metal binding is enhanced
in 1–3 by the presence of macrocyclic moieties. The electrochemical
studies in the presence of Cu2+ and Fe2+ show that 1 and 2 have pro-
nounced effects on the oxidation and reduction of these transition metal
ions in vitro. The changes observed are similar to the ones seen for DFO,
implying that 1 and 2 can remove labile copper and iron ions from their
respective pools. Complexation of these metal ions is essential for the
antioxidant activity of DFO.22 It also explains the protective properties
of tacn shown in Figure 7 and seems to contribute to the activity of 1 in
FEK4 cells.
Interestingly, our in vitro findings also translate into cell culture.
The 2.2-fold increase in GPx-like activity in FEK4 cells, for instance,
is likely to be directly due to the catalytic activity of 1, although a
more complex chain of events, such as Se-cysteine synthesis and GPx
production stimulated by 1 cannot be ruled out completely at this stage.
Similarly, the protection exerted by 1 against UVA-radiation, which is
comparable to DFO, yet superior to tacn, points towards a combination
of GPx-like catalysis and metal binding as the most likely mode(s) of
protective, antioxidant action.
It is also interesting to note that a recent study has employed copper-
containing tacn derivatives as antioxidants with an apparent SOD-like
activity.6 A metal-free (or zinc-containing) tacn derivative may there-
fore not only act as metal sequestering agent, but also as a “smart”
precursor of a metal complex, which is formed in the presence of labile
copper ions and then provides antioxidant SOD-like activity. Compound
1 may therefore provide a lead structure for the development of further
multifunctional antioxidants, not least because it also has shown low
toxicity in cell culture.
Future studies may well expand the synthetic aspects of this work to
include more complicated chemical structures, which are, in principle,
possible. For instance, Singh and colleagues have recently synthesized
larger Se-containing macrocycles, which provide an additional motif for
the design of multifunctional antioxidants.23 Similarly, the synthesis of
protoporphyrin based Se-compounds needs to be expanded. Although
our first attempts to assess these compounds in vitro have been marred
by solubility problems, concentrations required in vivo are considerably
lower than the ones required in our in vitro assays. Ultimately, more
soluble derivatives of 3 may possess several advantages. They are easy
to synthesize, are based on a natural moiety, contain two Se centers,
strongly interact with copper and iron ions, provide the basis for a bio-
chemically advantageous copper/iron for zinc metal exchange and may
form useful SOD mimics upon copper/iron uptake (see above).

Multifunctional Antioxidants
881
The basic metal exchange experiments, albeit providing only very
rough and qualitative estimates, also bode well: It may be possible to
use macrocyclic compounds to exchange labile copper and iron ions for
labile zinc ions, with dramatic and highly beneficial effects on the intra-
cellular redox environment. Furthermore, protoporphyrin compounds
are likely to feature promising transport properties due to their am-
phiphilic character and may act as photosensitizers in the metal-free
form.23 These aspects are currently under investigation.
In conclusion, we have obtained a first set of initial, promising re-
sults using a combination of Se-based catalysis and macrocycle-based
metal interactions. This approach now clearly leads itself to further,
more extensive studies combining advanced synthetic chemistry with
comprehensive, in depth biochemical and cell culture studies.
EXPERIMENTAL
Materials
Chemical reagents were purchased from Sigma-Aldrich-Fluka (Darm-
stadt, Germany) and used without further purification unless stated
otherwise. 3-(N-morpholino)propanesulfonic acid (MOPS, 99%) was ob-
tained from ACROS (Geel, Belgium) and Annexin V from Roche (Lewes,
UK). Cell culture materials were from Life Technologies (Paisley,
Scotland) except fetal calf serum (FCS), which was obtained from
PAA Laboratories (Teddington, UK) and phosphate buffered saline
(PBS), which was from Oxoid (Basingstoke, UK). Apoglow kit reagents
were from Lumintech (Nottingham, UK). Cd,Zn-MT was purchased
from Sigma and Zn7-MT was reconstituted and purified according
to a standard method.24 Deionised, MilliQ water (resistance ≥18
Mꢁ cm) was used for the electrochemical, in vitro and cell culture
experiments.
For chemical synthesis, dry solvents were essential for the synthesis
of compounds, i.e., acetonitrile, THF and CH2Cl2 were refluxed with cal-
cium hydride and freshly distilled before use. Dry DMF was purchased
from Fluka. All reactions were carried out under Argon (4.6) using stan-
dard Schlenk techniques. Silica gel (Macherey-Nagel, 50–200 μm) was
used for column chromatography.
1H NMR spectra were recorded at 500 MHz, 13C NMR spectra at
125 MHz on a Bruker Avance 500 spectrometer. Chemical shifts are
reported in δ (ppm), expressed relative to the solvent signal at 7.26
ppm (CDCl3, 1H NMR) and at 77.16 ppm (CDCl3, 13C NMR).

882
S. Mecklenburg et al.
For LC-MS a RP C18 NUCLEODURꢀR 100-3 (125 × 3 mm) column
(Macherey-Nagel GmbH, Dueren, Germany) was used as stationary
phase and MS analysis was performed on a TSQ Quantum mass spec-
trometer equipped with an ESI source and a triple quadrupole mass
detector (Thermo Finnigan, San Jose, CA). A gradient of methanol
(containing 0.1% formic acid) in water (containing 0.1% formic acid)
was used, which varied from 5% to 100% over 13 min at a flow rate of
400 μl min−1
.
Synthesis of compounds 1–3
1-[3-(phenylseleno)propyl]-1,4,7-triazacyclononane
(compound 1, Figure 3a)
1,4,7-Triazacyclononane (tacn, 137.1 mg, 1.06 mmol) was dissolved
in 30 ml of dry acetonitrile and K2CO3 (138.2 mg, 1.0 mmol) was added
at room temperature. A solution of 1-bromo-3-phenylselenopropane25
(183.6 mg, 0.66 mmol) in 20 ml of dry acetonitrile was then added drop-
wise over 1.5 h. The reaction was kept at room temperature and stirred
overnight. The solvent was then evaporated. The residue was taken up
in MeOH and CH2Cl2, the resulting solution was filtered and the sol-
vent evaporated to yield the product as light yellow viscous oil (210.1
mg, 0.64 mmol) in 97.5% yield.
C15H25N3Se (Mr = 326.34 g mol−1). 1H NMR (500 MHz, CDCl3):
δ = 1.81 (quint, 2 H, CH2CH CH2, 3 J = 6.9 Hz), 2.88 (m, 18 H, CH
2
HH
2
bridge, CH tacn, NH), 7.20 (m, 3 H, CH phenyl), 7.45 (m, 2 H, CH
2
phenyl).
13C NMR (125 MHz, CDCl3): δ = 25.2, 34.0, 42.1, 48.9, 49.2, 57.8
(CH2 bridge, CH2 tacn), 126.9, 129.1, 130.4, 132.7 (CH phenyl, Cq).
LC-MS: tR, m/z (%) = 5.7 min, 257.0 (100) [PhSeC3H6NC2H4
NH+H]+, 216.0 (45) [PhSeC3H6NH2+H]+.
1,7-bis[3-(phenylseleno)propyl]-1,4,7,10-
tetraazacyclododecane (Compound 2,
Figure 3b)
Compound 2 was synthesized by two alternative methods, one in-
volving nucleophilic substitution on a brominated precursor, the other
Schiff base formation with subsequent reduction (see Figure 3b).
The first method was analogous to the synthetic method for 1. To a
solution of 1,4,7,10-tetraazacyclododecane (cyclen, 172.3 mg, 1.0 mmol)
and K2CO3 (138.2 mg, 1.0 mmol) in dry acetonitrile was added 1-bromo-
3-phenylselenopropane (139.0 mg, 0.5 mmol) dropwise at room tem-
perature over 1.5 h. The solution was stirred at room temperature

Multifunctional Antioxidants
883
overnight and filtered. The solvent was subsequently evaporated to
yield a light yellow oil as crude product, which was purified by prepar-
ative HPLC using a gradient of 25% to 100% of methanol (contain-
ing 0.1% formic acid) in water (containing 0.1% formic acid). Retention
time of product: tR = 20–24 min. The solvent was then removed under
reduced pressure to yield a white solid. In order to avoid formic acid
salts of the product, the latter was dissolved in 30 ml CH2Cl2, washed
with 10 ml conc. NH3 solution and after that with water until the pH of
the washings was 7.0. The organic layer was separated and dried over
Na2SO4. CH2Cl2 was evaporated to give the desired product as lightly
yellowish, highly viscous oil (42.2 mg, 0.074 mmol) in 29.8% yield.
Alternatively, cyclen (172.3 mg, 1.0 mmol) was dissolved in 25 ml
of dry THF. A solution of 3-(phenylseleno)propanal26 (426.3 mg, 2.0
mmol) in 18 ml was added dropwise at room temperature over a period
of 10 min. After 15 min, sodium triacetoxyborohydride (593.4 mg, 2.8
mmol) was added and the reaction was stirred at room temperature
overnight. The solvent was evaporated under reduced pressure and the
crude product was purified by preparative HPLC as described above to
give the desired product as lightly yellowish, viscous oil (73.7 mg, 0.13
mmol) in 13.0% yield.
C26H40N4Se2 (Mr = 566.54 g mol−1). 1H NMR (500 MHz, CDCl3):
δ = 1.84 (quint, 4 H, CH2CH CH2, 3 J = 6.9 Hz), 2.54 (m, 14 H, CH
2
HH
2
3
tacn, CH bridge, NH), 2.63 (t, 8 H, CH tacn, J = 4.7 Hz), 2.88 (t,
2
2
HH
3
4H, CH bridge, J = 6.9 Hz), 7.22 (m, 6 H, CH phenyl), 7.46 (m, 4
2
HH
H, CH phenyl).
13C NMR (125 MHz, CDCl3): δ = 25.8, 27.4, 45.6, 52.2, 54.7 (CH2
bridge, CH2 cyclen), 126.7, 129.1, 130.6, 132.6 (CH phenyl, Cq).
LC-MS: tR, m/z (%) = 7.2 min, 569.1 (100) [M+H]+.
3,3ꢀ-(3,7,12,17-Tetramethyl-8,13-divinylporphyrin-2,18-
diyl)bis{N-[3-(phenylseleno)propyl]propanamide} (Compound
3, Figure 3c)
Compound 3 was synthesized from protoporphyrin IX and an appro-
priate selenium precursor, 3-(phenylseleno)propan-1-amine, by two al-
ternative methods, one involving the acid chloride reaction with amine,
the other a coupling of the acid with amine using ethylchloroformate
(Figure 3c).
As part of the first method, a suspension of protoporphyrin IX di-
sodium salt (106.0 mg, 0.18 mmol) in 10 ml dry CH2Cl2 was treated
with a large excess oxalylchloride (CO2Cl2, 1 ml). The resulting green
suspension was stirred at room temperature in the dark for 2 h. The
solvent and excess CO2Cl2 were removed under reduced pressure and

884
S. Mecklenburg et al.
the residue containing crude protoporphyrin bis(acid)chloride was re-
dissolved in 15 ml CH2Cl2 and used without further purification. It was
added dropwise to a solution of 3-(phenylseleno)propan-1-amine27,28
(275.0 mg, 1.28 mmol) and triethylamine (4 ml) in 55 ml dry DMF over
30 min. The reaction was then stirred at room temperature overnight.
The solvent was removed under reduced pressure and the resulting
brown residue was dissolved in 50 ml of CH2Cl2 and washed three
times with saturated brine. The organic layer was dried over Na2SO4,
filtered and the solvent removed under reduced pressure to give a dark
black solid. The impure product was cleaned by three consecutive silica
gel column chromatography purifications, using solvents CH2Cl2 and
MeOH (95:5). The product, 3, was obtained as a purple solid (25.6 mg,
0.03 mmol) in a yield of 16.7%.
C52H56N6O2Se2 (Mr = 954.96 g mol−1). 1H NMR (500 MHz, CDCl3):
δ = 1.38–1.47 (m, 4 H, CH ), 2.19–2.22, 2.24–2.27 (each t, 2 H, CH ,
2
2
3 J = 7.0 Hz), 2.92–2.96 (t, 4 H, CH , 3 J = 7.6 Hz) 2.98–3.05 (m,
HH
2
HH
4 H, COCH ), 3.37, 3.40, 3.45, 3.55 (each s, 3 H, CH ), 4.16–4.20 (m,
2
3
4 H, COCH2CH ), 6.06–6.15, 6.20–6.31 (each m, 2 H, CH vinyl), 6.74,
2
2
6.82 (each s, 1 H, CONH), 6.95–7.06 (m, 8 H, CH phenyl), 7.52–7.54,
7.70–7.72 (each m, 1 H, CH phenyl), 7.97–8.03, 8.11–8.17 (each m, 1 H,
CH vinyl), 9.71, 9.76, 9.79, 9.86 (each s, 1 H, CH methin).
LC-MS: tR, m/z (%) = 15.7 min, 957.2 (100) [M+H]+.
The comparably low yield of the acid chloride to amine coupling may
have been due to the rather aggressive chemicals used. As an alter-
native, coupling with ethylchloroformate was investigated. Protopor-
phyrin IX (111.4 mg, 0.20 mmol) was dissolved in 5 ml dry chloroform
and cooled to 0◦C. N-methylmorpholine (45.9 mg, 0.45 mmol) was added
and the mixture was stirred at 0 ◦C for 15 min. Ethylchloroformate (52.4
mg, 0.48 mmol) was added and the reaction mixture was stirred at 0
◦C for another 30 min. At this point, 3-(phenylseleno)propane-1-amine
(99.8 mg, 0.47 mmol) was added and the reaction was stirred at 0 ◦C for
1 h, followed by stirring at room temperature overnight. The solvent
was removed under reduced pressure and the resulting brown residue
was purified once by column chromatography as described above, to
obtain 3 as a purple metallic solid (117.6 mg, 0.12 mmol) in a yield of
62.1%. Analytical data was in accordance with the one obtained by the
CO2Cl2 coupling method.
Electrochemistry
Cyclic Voltammograms were recorded at a BAS 100W electrochemical
workstation to evaluate the redox behavior of the Se-compounds and

Multifunctional Antioxidants
885
their interactions with Cu2+ ions. In all electrochemical experiments,
recordings were taken on a glassy carbon working electrode (which was
cleaned with Al2O3 and H2O between each recording), using an Ag/AgCl
electrode (SSE) as reference and a platinum spiral as counter elec-
trode. Unless stated otherwise, Cyclic Voltammograms were recorded
at 20◦C with a scan rate of 500 mV s−1. Three full cycles were recorded
as part of each scan, and all experiments were performed in tripli-
cate. Ferrocene was used as a reference to control/adjust the potential
scale.
Oxidation of the Se-redox center was studied between −500 mV and
+1300 mV vs. SSE in potassium phosphate (KPi) buffer (pH 7.4), con-
taining 33% methanol to ensure sufficient solubility of the compounds
tested. In accordance with our previous studies, the interaction with
Cu2+ ions was studied in 10 mM MOPS buffer (pH 7.4) to avoid precip-
itation of metals/ligand or significant metal-buffer interactions.5 The
buffer was treated with Chelex 100 prior to use to remove any traces of
interfering metal ions. The potential range was between −350 mV and
+550 mV vs. SSE.
UV/VIS Spectrophotometry and UV-Based Assays
for Catalytic Activity
UV/VIS spectra were recorded on a Cary 50 Bio spectrophotometer
from Varian Inc.. Unless stated otherwise, quartz cuvettes were used
throughout. All recordings were taken at 20oC and repeated at least
three times.
The thiophenol (PhSH) assay, which is indicative of catalysts enhanc-
ing the rate of reaction of H2O2 with PhSH, was employed to assess the
catalytic activity of compounds. This assay is somewhat easier to use
when compared to the coupled GPx-assays or the MT assay. It avoids
solubility problems often associated with aromatic Se-compounds by
operating in methanol and provides a fast and reliable method for as-
sessing (catalyzed) thiol oxidation events. In short, thiophenol (1 mM
final concentration) was dissolved in methanol and oxidation of the thiol
was initiated by addition of 2 mM H2O2. The formation of PhSSPh was
followed at 305 mn and 20oC for 60 min and the initial rates of reaction
(1 to 10 min) were compared for reactions in the absence and pres-
ence of Se-compounds (100 μM each).5 Ebselen was used as a positive
control.
The more complicated MT assay, which measures the oxidative re-
lease of Zn2+ ions from the zinc/sulfur protein MT and subsequent
uptake of the metal ion by the chromophoric dye 4-(2-pyridylazo)-
resorcinol (PAR), was used for the most promising compound, 1.29

886
S. Mecklenburg et al.
The experimental procedure was identical to the one described in the
literature.
FEK4 Cell Culture
The human skin fibroblast cells FEK4 and FCP7 were derived from
newborn foreskin explants.13,14 Cells were cultured routinely in Earle’s
minimal essential medium (EMEM) supplemented with 2 mM L-
glutamine, 50 u/ml penicillin, 50 μg/ml streptomycin, 0.2% sodium bi-
carbonate and 15% FCS and incubated in a humidified atmosphere at
37◦C with 5% CO2. For experiments, fibroblast monolayers were grown
in dishes/coverslips for 3–4 days in order to reach 80% confluency. FEK4
cells were used between passages 12 to 14 and FCP7 cells between pas-
sages 4 to 11.
Selenium compounds were dissolved in DMSO and added to the cell
medium at the required concentrations 18 h prior to experiments. The
concentration of DMSO was kept constant at 0.1% of the medium in
order to avoid cellular effects of the vehicle itself. Control cells were
pretreated with DMSO alone.
For the irradiation step, the medium was removed from the cells
and retained as a condition medium. Cells were washed with PBS and
covered in 2 ml of warm PBS containing Ca2+ and Mg2+ ions (0.01%
CaCl2 and MgCl2). Monolayers of cells were irradiated at 25oC using a
broad spectrum Sellas 4kW UVA lamp (Sellas, Germany) with signifi-
cant emission in the range of 350–400 nm. UVA doses were measured
using an IL1700 radiometer (International Light, Newburyport, MA).
The UVA doses used were 250 kJ m−2 (moderate dose) and 500 kJ m−2
(high dose), which are equivalent to the amount the surface of the skin
would be exposed to over 70 or 140 min, respectively, at noon on a cloud-
less summer day, at a northern latitude of 30 to 35◦.
After irradiation, cells were washed with PBS, the condition medium
was added back and cells were incubated at 37◦C for 4 h. Control cells
were treated in the same manner, except that they were not irradi-
ated and were instead stored in dark conditions at room temperature
during the irradiation process. Cell viability was measured with the
MTT colorimetric assay at 550 mm using a MR5000 microplate reader
(Dynatech Laboratories, West Sussex, UK).
The GPx activity in FEK4 cells was measured according to the
method by Flohe´ and Gunzler, whereby NADPH consumption at 340
nm is used as a measure of the rate of glutathione disulfide (GSSG) for-
mation due to the GPx reaction (the assay is initiated by the addition
t
of GSH, BuOOH and NADPH).30 ATP was measured using the com-
mercially available Apoglow adenylate nucleotide ratio assay.14

Multifunctional Antioxidants
887
REFERENCES AND NOTES
[1] B. Halliwell and J. M. C. Gutteridge, Free Radicals in Biology and Medicine, 4th ed.
(Oxford University Press, Oxford, 2007).
[2] Z. A. Wood, E. Schroder, J.R. Harris, and L.B. Poole, TIBS, 28, 32–40 (2003).
[3] Many of these catalysts do, of course, require a reductant as co-substrate to function,
with the notable exception of SOD and CAT which redox-disproportionate O•2− and
H2O2, respectively. In the case of GPx and Prx, GSH and reduced thioredoxin (Trx)
act as respective co-substrates. While this dependency may be seen as a potential
disadvantage, it also provides a certain element of biochemical feedback and control:
Once GSH or Trx levels are depleted in the cell, these antioxidant enzymes switch
off, the cellular defense capitulates and the cell undergoes (controlled) cell death.
This chain of events has recently been studied to considerable extent for the Prx
enzymes.
[4] G. Mugesh and H. B. Singh, Chem. Soc. Rev., 29, 347–357 (2000).
[5] C. A. Collins, F. H. Fry, A. L. Holme, A. Yiakouvaki, A. Al-Qenaei, C. Pourzand, and
C. Jacob, Org. Biomol. Chem., 3, 1541–1546 (2005).
[6] Q. X. Li, Q. H. Luo, Y. Z. Li, and M. C. Shen, Dalton Trans., 2329–2335 (2004).
[7] X. Liang and P. J. Sadler, Chem. Soc. Rev., 33, 246–266 (2004).
[8] K. Pong, Expert. Opin. Biol. Ther., 3, 127–139 (2003).
[9] The oxidation wave of Fe2+ was not clearly observable due to low concentrations
of the metal ion in solution, which in turn was limited by the solubility of 1
and 2.
[10] Changes in the UV region between 260 nm and 360 nm are not due to metal binding
to protoporphyrin but are associated with the copper and iron ions themselves.
[11] G. I. Giles, F. H. Fry, K. M. Tasker, A. L. Holme, C. Peers, K. N. Green, L. O. Klotz,
H. Sies, and C. Jacob, Org. Biomol. Chem., 1, 4317–4322 (2003).
[12] Compound 3 was poorly soluble in aqueous media and no identifiable oxidation
current could be observed.
[13] C. Pourzand, R. D. Watkin, J. E. Brown, and R. M. Tyrrell, Proc. Natl. Acad. Sci.
USA, 96, 6751–6756 (1999).
[14] J. L. Zhong, A. Yiakouvaki, P. Holley, R. M. Tyrrell, and C. Pourzand, J. Invest.
Dermatol., 123, 771–780 (2004).
[15] N. M. Giles, N. J. Gutowski, G. I. Giles, and C. Jacob, FEBS Lett., 535, 179–182
(2003).
[16] N. M. Giles, G. I. Giles, J. E. Holley, N. J. Gutowski, and C. Jacob, Biochem. Phar-
macol., 66, 2021–2028 (2003).
[17] G. I. Giles, N. M. Giles, C. A. Collins, K. Holt, F. H. Fry, P. A. S. Lowden, N. J.
Gutowski and C. Jacob, Chem. Commun., 2030–2031 (2003).
[18] F. H. Fry, A. L. Holme, N. M. Giles, G. I. Giles, C. Collins, K. Holt, S. Pariagh, T.
Gelbrich, M. B. Hursthouse, N. J. Gutowski, and C. Jacob, Org. Biomol. Chem., 3,
2579–2587 (2005).
[19] F. H. Fry and C. Jacob, Current Pharm. Des., 12, 4479–4499 (2006).
[20] Z. D. Liu and R. C. Hider, Coord. Chem. Rev., 232, 151–171 (2002).
[21] G. I. Giles, K. M. Tasker, R. J. K. Johnson, C. Jacob, C. Peers, and K. N. Green, Chem.
Commun., 2490–2491 (2001).
[22] G. S. Tilbrook and R. C. Hider, Metal Ions in Biological Systems, 35, 691–730 (1998).
[23] S. Panda, S. S. Zade, H. B. Singh, and R. J. Butcher, Eur. J. Inorg. Chem., 172–184
(2006).
[24] M. Vasˇa´k, Methods Enzymol., 205, 41–44 (1991).
[25] J. E. Baldwin, R. M. Adlington, and J. Robertson, Tetrahedron, 45, 909–922 (1989).

888
S. Mecklenburg et al.
[26] S. Harusawa, S. Tomii, C. Takehisa, H. Ohishi, R. Yoneda, and T. Kurihara, Chem.
Pharm. Bull., 40, 2279–2282 (1992).
[27] A. Khanna, A. Bala, and B. L. Khandelwal, J. Organomet. Chem., 494, 199–204
(1995).
[28] S. V. Amosova, N. A. Makhaeva, A. V. Martynov, V. A. Potapov, B. R. Steele, and I. D.
Kostas, Synthesis, 1641–1648 (2005).
[29] C. Jacob, W. Maret, and B. L. Vallee, Proc. Natl. Acad. Sci. U. S. A., 96, 1910–1914
(1999).
[30] L. Flohe and W. A. Gunzler, Methods Enzymol., 105, 114–121 (1984).
[31] E. O. Hileman, J. Liu, M. Albitar, M. J. Keating, and P. Huang, Cancer Chemother
Pharmacol, 53, 209–219 (2004).