Besting vitamin E: Sidechain substitution is key to the reactivity of naphthyridinol antioxidants in lipid bilayers

Article abstract of DOI:10.1021/ja309153x

A series of naphthyridinol analogs of α-tocopherol (α-TOH, right) with varying sidechain substitution was synthesized to determine how systematic changes in the lipophilicity of these potent antioxidants impact their radical-trapping activities in lipid b

Full text of DOI:10.1021/ja309153x

Article
pubs.acs.org/JACS
Besting Vitamin E: Sidechain Substitution is Key to the Reactivity of
Naphthyridinol Antioxidants in Lipid Bilayers
Bo Li,† Jitendra R. Harjani,† Nicholas S. Cormier,† Hasam Madarati,‡ Jeffrey Atkinson,‡ Gonzalo Cosa,§
and Derek A. Pratt*,†
†Department of Chemistry, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada
‡Department of Chemistry and Centre for Biotechnology, Brock University, St. Catharines, Ontario L2S 3A1, Canada
§Department of Chemistry and Centre for Self-Assembled Chemical Structures, McGill University, 801 Sherbrooke St. West,
Montreal, Quebec H3A 0B8 Canada
S
* Supporting Information
ABSTRACT: A series of naphthyridinol analogs of α-
tocopherol (α-TOH, right) with varying sidechain substitution
was synthesized to determine how systematic changes in the
lipophilicity of these potent antioxidants impact their radical-
trapping activities in lipid bilayers, regenerability by water-soluble reductants, and binding to human tocopherol transport protein
(TTP). The activities of the naphthyridinols were assayed in phosphatidylcholine unilamellar liposomes using a recently
developed high-throughput assay that employs a boron dipyrromethene conjugate of α-TOH (H2B-PMHC) that undergoes
fluorescence enhancement upon oxidation. The naphthyridinols afforded a dose-dependent protection of H2B-PMHC consistent
with unprecedented peroxyl radical-trapping activity in lipid bilayers. While sidechain length and/or branching had no effect on
their apparent reactivity, it dramatically impacted reaction stoichiometry, with more lipophilic compounds trapping two peroxyl
radicals and more hydrophilic compounds trapping significantly less than one. It is suggested that the less lipophilic compounds
autoxidize rapidly in the aqueous phase and that preferential partitioning of the more lipophilic compounds to the bilayer
protects them from autoxidation. The cooperativity of a lipophilic naphthyridinol with water-soluble reducing agents was also
studied in liposomes using H2B-PMHC and revealed superior regenerability by each of ascorbate, N-acetylcysteine, and urate
when compared to α-TOH. Binding assays with human TTP, a key determinant of the bioavailability of the tocopherols, reveal
that the naphthyiridinols can be very good ligands for the protein. In fact, naphthyridinols with sidechains of eight or more
carbons had affinities for TTP which were similar to, and in one case 10-fold better than, α-TOH.
sequence by reaction with chain-carrying peroxyl radicals (eqs 5
and 6, Scheme 1).9 α-TOH, the most biologically active
congener of vitamin E (kinh = 3.2 × 106 M−1 s−1 at 30 °C),9 is
the major lipid-soluble radical-trapping antioxidant in human
plasma and circulating lipoproteins. It is well established as one
of the most reactive radical-trapping antioxidants in vitro and is
commonly used as the standard against which others are
evaluated.10 Given its high reactivity as an inhibitor of lipid
peroxidation and the implication of this process in the
pathogenesis of degenerative disease, a staggering number of
clinical trials employing α-TOH in a therapeutic and/or
preventive capacity have been conducted over the past 30
years (the NIH clinical trials database currently lists well over
300 relating to cancer and cardiovascular and neurodegener-
ative diseases alone). The results have been largely disappoint-
ing, leading researchers to question whether oxidative damage
has a causal or consequential role in the onset and development
of degenerative disease.11,12 Perhaps a more appropriate
question is: Is α-TOH the best compound to study, or is it
simply convenient to do so?
INTRODUCTION
■
The oxidative modification of proteins and DNA as a result of
polyunsaturated lipid peroxidation has been implicated in the
onset of essentially all degenerative diseases, including
atherosclerosis,1,2 neurodegenerative disease,3,4 and cancer.5,6
Peroxidation of polyunsaturated lipids proceeds by the
prototypical free radical chain reaction (eqs 1−4, Scheme
1)7,8 and can be inhibited by radical-trapping antioxidants, such
as α-tocopherol (α-TOH, Chart 1), which interrupt the chain
Scheme 1. Mechanism of Lipid Peroxidation and Its
Inhibition by Phenolic Antioxidants (e.g., α-TOH)
Received: September 20, 2012
Published: December 31, 2012
© 2012 American Chemical Society
1394
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
unknown whether they will be regenerated by co-antioxidants
in the aqueous medium surrounding the lipid bilayers. Indeed,
the cooperativity displayed between water-soluble antioxidants,
such as ascorbate (vitamin C), and α-TOH has long been
thought to be key to the activity of both compounds in
vivo.19−21 Furthermore, the impact of changes in the lipophilic
sidechain of tocopherols (or any tocopherol-like ligands) on
their affinity for the TTP has never been systematically
surveyed, and it remains to be determined whether the simpler
naphthyridinol 3 will bind at all, given the removal of the
quaternary carbon (whose stereochemistry is known to be
important in binding the tocopherols)22−24 and the relocation
of the lipophilic sidechain to a position α to the aromatic ring.
Herein we describe the preparation of a small library of
naphthyridinols with sidechains of varying length and
branching, the results of high-throughput competitive kinetic
studies of their reactivities in phosphatidylcholine liposomes
under various conditions, which make use of a fluorescence-
based microplate assay (recently developed by one of us),30 and
the determination of their abilities to bind to recombinant
human TTP. Together, these data provide a clear picture of the
vital role of the sidechain on the reactivity and potential
bioavailability of naphthyridinol antioxidants and set the stage
for experiments in vivo.
Chart 1. Highly Reactive Radical-Trapping Antioxidants
Some time ago we and our collaborators demonstrated that
incorporation of nitrogen atoms in the aromatic rings of
phenolic antioxidants at the 3 and/or 5 positions relative to the
phenolic hydroxyl greatly stabilizes the compounds to
autoxidation.13,15 As a result, the phenol-like 3-pyridinol and
5-pyrimidinol compounds can be substituted with very strong
electron-donating (e.g., N,N-diakylamino) groups to weaken
the O−H bond14 that must be broken in the transfer of the
labile H-atom to peroxyl radicals (as in eq 5, Scheme 1). For
example, pyrimidinol 1 and pyridinol 2 are stable in aerated
solutions and react two- and five-fold faster with peroxyl
radicals in organic solution than α-TOH.13,15,25 The bicyclic
pyridinol compounds 3a and 4 are even more reactive, with rate
constants measuring 28- and 88-fold, respectively, that
determined for α-TOH under the same conditions.15,25 While
4 has limited stability in aerated solutions, 3a can be more
easily manipulated, suggesting that it is a better compromise
between activity and stability. Its lipophilic analog 3b16 and the
tocopherol analog 5 (which was dubbed N-tocopherol)17 are
both excellent inhibitors of cholesterol ester oxidation in
human low-density lipoprotein (LDL), and the latter was found
to bind to recombinant human tocopherol transport protein
(TTP) at least as well as α-TOH.17 Despite these exciting
results, further experimentation with 5 has been all but
impossible due to its lengthy synthesis, which required 17
chemical steps. Since 3b has similar activity to 5 in LDL, but is
more synthetically accessible, it is a much better candidate for
further studies.
It is well-known that the localization and mobility of
tocopherols and their analogs can play at least as important a
role in their antioxidant activity as their inherent chemical
reactivity upon moving from homogeneous organic solvents to
the heterogeneous environment of a lipid bilayer.18 While we
have worked hard over the years to optimize the chemical
reactivity of the phenolic headgroup toward peroxyl radicals in
homogeneous solution in our development of 3 (and 5), we
have yet to undertake a detailed study of their reactivities in
lipid bilayers and to determine how changes in their physical
properties will impact their activities. Furthermore, since the
pyridinoxyl radicals derived from 3 (and 5) are more stable
than the tocopheroxyl radical derived from α-TOH,15,17 it is
RESULTS
■
Synthesis of Naphthyridinols with Different Sidechain
Substitution. In previous work, we (and others) have accessed
the tetrahydronaphthyridinol core structure of 3 (and 5)
starting from either 2-amino-4,6-lutidine15,26 or pyridoxine
(vitamin B6).27 The former approach (6 steps) featured two
very low-yielding reactions: the (initial) ring annulation (23%)
and the (final) installation of the pyridinol moiety via a low-
yielding lithiation/oxidation sequence (25%). This served as
motivation for the development of the latter approach (11
steps), which carried the pyridinol moiety through the
synthesis, and although longer, did not suffer a single low-
yielding step. In parallel, we found that a Cu-mediated
alkoxylation/hydrogenolysis sequence was a much better
option than the lithiation/oxidation approach.28 In addition,
Hecht’s group provided a concise route to the tetrahydronaph-
thyridinol core from 2-piperidinonea shorter, higher-yielding
approach than our earlier annulation strategy.29 This prompted
us to consider a combination of these two developments to
prepare our small library.
Since we desired a variety of substitution at the amine
nitrogen, we wanted to carry out the alkylation in the final step.
Therefore, we elected to benzylate the piperidone to protect
this nitrogen through the annulation and aryl substitution steps,
after which it could be removed and the amine alkylated with
different sidechains. The target compounds are shown as 12a−
h, the oxalic acid addition salts of naphthyridinols (3). We
chose increments of four carbons for the linear sidechains (n-
butyl, n-octyl, n-dodecyl, n-hexadecyl) and increments of five
carbons for the branched sidechains, including one, two, and
three of the isoprene units found in the phytyl sidechain of α-
TOH, in addition to the derivative lacking a sidechain. This
range of sidechain lengths provides a range of over 7 orders of
magnitude in predicted values of logP (see Supporting
Information for values). The eight compounds were prepared
as in Scheme 2.
Briefly, 2-piperidinone was first protected with a benzyl
moiety to provide 6, which was ketalized with dimethyl sulfate/
1395
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
quenched by photoinduced electron transfer from the electron-
rich PMHC group, whereas upon reaction of the phenol with
peroxyl radicals (eq 7), the fluorescence of the probe is turned
‘on’ since the resultant chromanone is not sufficiently electron
rich to reduce the BODIPY moiety upon excitation.31,32 The
probe was recently used in a high-throughput approach utilizing
a microplate reader to monitor the antioxidant status (α-TOH
and PMHC) in liposomes of different phospholipid composi-
tion. The probe, the fluorescence assay, and the data analysis
provide a new method to obtain, in a rapid parallel format,
relative antioxidant activities in phospholipid membranes.30
Representative results of oxidations of solutions of H2B-
PMHC-supplemented liposomes containing different concen-
trations of added antioxidant are shown in Figure 1. These
results were obtained from oxidations mediated by the
hydrophilic azo-compound 2,2′-azobis(amidinopropane) dihy-
drochloride (AAPH) (2.7 mM), which yields water-soluble
peroxyl radicals with a reported rate of Ri = 2ekd[AAPH] = 3.1
× 10−9 M s−1 under these conditions.33 Representative data
corresponding to oxidations inhibited by the highly lipophilic
derivative 12e the more hydrophilic derivative 12b are shown
in Figure 1A,C, respectively, while data for oxidations inhibited
by α-TOH and its truncated analog PMHC are shown for
comparison in Figure 1B,D, respectively (data for the other
naphthyridinols are included in the Supporting Information).
Five concentrations of each antioxidant were examined from
which a clear dose−response relationship is evident. That is,
with the addition of increasing amounts of antioxidant, the time
required to achieve the maximum rate of fluorescence increase
was extended proportionately. The ‘inhibited period’, where the
oxidation of H2B-PMHC is inhibited (or retarded) by the
added antioxidant, is analogous to the ‘inhibited period’ that is
observed in inhibited autoxidations of hydrocarbons,34 which
are usually monitored by either O2 consumption or the
formation of product hydroperoxidesmuch more laborious
experiments not amenable to high-throughput study. Through-
out, we have determined the length of the inhibited period (τ)
from the intersection of the lines of best fit to the maximum
rate of increase in fluorescence intensity (due to oxidation of
the chromanol moiety of H2B-PMHC) and the slower
subsequent increase in fluorescence intensity (due to follow-
up reactions of the fluorophore), as this is the time at which all
of the added antioxidant as well as the phenolic moiety of the
probe has reacted (see Supporting Information).57 The
inhibited period provides important insight into the reactivities
of the antioxidant under investigation, as its duration reflects
the relative stoichiometry of the reaction of the antioxidant with
peroxyl radicals, since the probe is present in a constant amount
in all experiments.30
Scheme 2. Synthesis of Naphthyridinol Antioxidants
sodium methoxide and then condensed with 4-aminopent-3-en-
2-one to give 7. The naphthyridine was halogenated with N-
bromosuccinimide and then methoxylated to provide 9. While
we had initially attempted the alkoxylation with benzyl alcohol,
as we had in the preparation of 1 and 2,28 we found much lower
yields for benzyloxylation of 8 presumably due to its greater
electron richness, unless we pushed the reactions with large
concentrations of CuI and Cs2CO3. The benzyl-protecting
group was then removed by hydrogenolysis, and the key
intermediate 10 was alkylated with various alkyl bromides to
afford the O-methylated naphthyridinols 11b−h. Removal of
the O-methyl group using the dimethylsulfide complex of
boron tribromide (BBr3·Me2S) followed by addition of oxalic
acid provided the ammonium oxalate salts of the naphthyr-
idinols 12a−h. Unlike the free naphthyridinols 3, these
derivatives are indefinitely stable at room temperature in an
aerobic atmosphere.
Peroxyl Radical Trapping by Naphthyridinols in
Liposomes. The peroxyl radical-trapping activities of the
naphthyridinols were assayed in unilamellar liposomes prepared
from egg phosphatidylcholine using the fluorogenic H2B-
PMHC (eq 7) probe (λex = 485 nm; λem = 520 nm).30 H2B-
The inhibited periods observed in Figure 1 are very similar
for the same concentration of 12e, α-TOH, and PMHC (Figure
1A,B,D, respectively), becoming longer with increasing
antioxidant concentration. In contrast, the inhibited periods
observed for 12b (Figure 1C) are comparably shorter and do
not increase to the same extent with increasing antioxidant
concentration. This is representative of the trends observed
with all of the naphthyridinols (see Supporting Information).
The more lipophilic derivatives (C10H21 or longer) have similar
inhibited periods to α-TOH and PMHC, and the more
hydrophilic compounds (C5H11 or shorter) have much shorter
inhibited periods (less than half at higher concentrations), and
the octylated derivative is roughly intermediate between them.
PMHC features a boron dipyrromethene (BODIPY) fluo-
rophore conjugated to 2,2,5,7,8-pentamethyl-6-hydroxychro-
man (PMHC), an analog of α-TOH which lacks the lipophilic
phytyl sidechain. The function of the probe is straightforward:
In the ‘off’ state, the fluorescence of the BODIPY moiety is
1396
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
Figure 1. Representative fluorescence (at 520 nm) intensity−time profiles from AAPH-mediated (2.7 mM) oxidations of egg phosphatidylcholine
liposomes (1 mM in PBS buffer, pH 7.4) containing 0.15 μM H2B-PMHC and increasing concentrations (1.5 μM, cyan; 3.0 μM, blue; 4.5 μM,
green; 6.0 μM, red; and 7.5 μM, black) of 12e (A), α-TOH (B), 12b (C), and PMHC (D).
In addition to relative stoichiometry, the data provide insight
into the relative rates for the reactions of peroxyl radicals with
the different antioxidants vs the fluorogenic H2B-PMHC, since
the rate of fluorescence increase during the inhibited period
represents the competition of the added antioxidant and H2B-
PMHC for peroxyl radicals. Focusing on this part of the data, it
is clear that α-TOH is by far the least reactive compound
(Figure 1B), as it barely retards H2B-PMHC oxidation, except
perhaps in the first few minutes of the experiment when the
highest antioxidant concentration is used (7.5 μM). PMHC is
more reactive (Figure 1D), as is clear from the more obvious
inhibited periods. A kinetic analysis based on the initial rates of
H2B-PMHC oxidation in the presence and absence of added
antioxidant has been carried out to provide an expression useful
for the quantification of the relative rate constants in this
competition (eq 8):30
Interestingly, the naphthyridinols 12e and 12b (Figure 1A,C,
respectively) clearly display much faster rates of reaction with
peroxyl radicals than both α-TOH and PMHC, as they
completely suppress H2B-PMHC oxidation throughout the
inhibited periods. In fact, the same is true of all of the other
substituted naphthyridinols (see Supporting Information).
Based on the lack of any emission enhancement during the
inhibited periods observed in the presence of the naphthyr-
idinols, the relative rate constants for H-atom transfer from the
naphthyridinols to peroxyl radicals cannot be derived from eq
8. However, a lower bounds can be estimated of ≥30 times the
rate constant for H2B-PMHC (which has essentially the same
rate constant for reaction with peroxyl radicals as does α-
TOH). Analogous results were obtained when experiments
were carried out using the hydrophobic azo-compound 2,2′-
azobis(4-methoxy-2,4-dimethylvaleronitrile) (MeOAMVN)
(see Supporting Information).35 The inhibited periods for all
of the naphthyridinols as well as α-TOH and PMHC are
summarized in Figure 2 for the five concentrations that were
studied, alongside the corresponding data for the oxidations
with lipophilic peroxyl radicals. Presented in this way it is clear
that increasing the concentration of the less lipophilic
naphthyridinols has very little effect on the observed inhibited
periods, while increasing the concentration of the more
H2B‐PMHC
⎛
⎜
⎝
⎞
⎟
⎠
I∞ − It
k
inh
⎛
⎞
⎟
t
τ
⎜
−ln
= −
ln 1 −
unknown
⎝
⎠
I∞ − I0
k
(8)
inh
Thus, from a plot of −ln[(I∞ − It)/(I∞ − I0)] vs −ln(1 − t/τ),
the relative rate constant can be determined (see Supporting
Information). This analysis indicates a ∼10-fold increase in the
rate constant for the reaction of PMHC and hydrophilic
radicals compared to α-TOH under these conditions.30
1397
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
Figure 3. Representative fluorescence (at 520 nm) intensity−time
profiles from oxidations of egg phosphatidylcholine liposomes (1 mM
in PBS buffer, pH 7.4) containing 0.15 μM H2B-PMHC with 2.7 mM
AAPH in the presence of 7.5 μM of 12b (blue, pH 5.8; red, pH 7.4) or
PMHC (green, pH 5.8; black, pH 7.4).
hydrophilic radicals and 1.5-fold when the oxidation was carried
out with hydrophobic radicals (see Supporting Information).
Peroxyl Radical Trapping by Pyri(mi)dinols in Lip-
osomes. To provide further structure−reactivity data to help
understand the origin of the shorter inhibited periods for the
less hydrophobic naphthyridinols, analogous liposome oxida-
tions inhibited by the less reactive pyrimidinol 1 and pyridinol
2 were carried out with both hydrophilic and lipophilic peroxyl
radicals (see Supporting Information for the fluorescence
intensity−time traces). The inhibited periods, determined as a
function of antioxidant concentration, are presented in Figure
4, alongside those for naphthyridinol 12a and PMHC for
comparison. Clearly, the naphthyridinol has the shortest
inhibition periods, followed by the pyridinol 2, the pyrimidinol
1, and finally PMHC. The inhibited periods follow the trend in
the standard potentials (E°) of each of the compounds, which
were determined by cyclic voltammetry (see Supporting
Information) and are given alongside the structures in Figure 4.
Cooperativity with Water-Soluble Antioxidants. In
order to evaluate the interaction of the naphthyridinols with
water-soluble reductants, oxidations of H2B-PMHC-loaded
liposomes supplemented with a relatively small and constant
concentration (1.5 μM) of the most lipophilic (hexadecylated)
naphthyridinol 12e were carried out in buffered solutions to
which were added increasing amounts of water-soluble
antioxidants (ascorbate, N-acetylcysteine or urate; 1.5−15
μM). Representative results are shown in Figure 5. The
oxidations were carried out using only the lipid-soluble radical
initiator MeOAMVN, since the water-soluble antioxidants react
directly with hydrophilic peroxyl radicals generated from the
water-soluble initiator AAPH and therefore do not provide any
information on potential cooperativity between them and the
naphthyridinols. In addition to these series of experiments,
oxidations of liposomes to which no lipophilic antioxidant was
added were carried out in parallel in order to account for either
the reaction of the water-soluble reductants with lipophilic
peroxyl radicals or the probe-derived phenoxyl radical (formed
from reaction of H2B-PMHC with lipophilic peroxyl radicals) at
Figure 2. Inhibited periods (averages of at least three measurements)
observed for naphthyridinols 12a−h, PMHC, and α-TOH as a
function of antioxidant concentration when egg phosphatidylcholine
liposomes were oxidized with hydrophilic (A) or lipophilic (B) peroxyl
radicals.
lipophilic naphthyridinols, PMHC, or α-TOH leads to a
substantial increase in inhibited periods.
In an attempt to shed light on the origin of the shorter
inhibited periods for the less hydrophobic naphthyridinols,
additional oxidations were carried out in a buffer of different
pH. In parallel, both AAPH- and MeOAMVN-mediated
oxidations were carried out at pH 5.8 and 7.4 with two
different concentrations of either naphthyridinol 12b or PMHC
(4.5 and 7.5 μM). Representative results are shown in Figure 3
for the AAPH-mediated oxidations with 7.5 μM of either 12b
or PMHC (others are included in the Supporting Information).
It is important to note that the change in pH had no effect on
the rate of H2B-PMHC oxidation either alone (not shown) or
in the presence of PMHC. However, a marked increase in the
inhibition period was clear for at more acidic pH. The increase
was almost 2-fold when the oxidation was carried out with
1398
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
Information). Smaller, but reproducible, differences were also
observed upon addition of a small (constant) amount of 12e to
the oxidations in the presence of increasing ascorbate
concentration, in that the inhibited period was elongated and
the rate of inhibited oxidation of the probe was nil when both
12e and ascorbate were present (compare Figure 5A,D), as is
the case for 12e alone.
The inhibited periods can be plotted as a function of
coantioxidant concentration in the presence of naphthyridinol
12e (see Supporting Information) from which it is clear that
the inhibited periods increase linearly with coantioxidant (i.e.,
water-soluble reductant) concentration. Furthermore, the
magnitudes of the correlations indicate that ascorbate extends
the inhibited periods more effectively (11.8 min/μM) than
both N-acetylcysteine (7.6 min/μM) and urate (6.8 min/μM).
All three correlations have smaller magnitudes than that
between the inhibited period and the concentration of 12e
alone (18.2 min/μM).
Binding to the TTP. The binding of the naphthyridinols
12b−h to the TTP was assayed using the fluorescent
tocopherol analogue NBD-Toc.36 The fluorescence of NBD-
Toc (λex = 470 nm; λem = 535 nm) diminishes upon its
displacement from TTP’s binding site by competitive ligands
due to intramolecular quenching of the NBD fluorophore by
the benzochromanol moiety. From the fluorescent titration of
TTP with NBD-Toc, a dissociation constant (Kd) of 45 15
nM is obtained. By way of comparison, a radio ligand binding
assay with 3H-α-TOH gives a Kd of 25 nM for α-TOH binding
to TTP, which demonstrates the validity of the fluorescence
assay. Effective dissociation constants (Kd,eff) for the various
naphthyridinols are given in Table 1 as is the value obtained for
α-TOH, which is provided for comparison.
DISCUSSION
■
Excitement surrounding the promise of radical-trapping
antioxidants for degenerative disease prevention has receded
somewhat in recent years due in large part to the disappointing
results of clinical trials with α-TOH, ascorbate, and β-carotene,
among other less-trumpeted natural products.11,12 However, it
must be pointed out that these compounds have well-
documented shortcomings as radical-trapping antioxidants.
For instance, in the absence of water-soluble reductants (or
reduced coenzyme Q10), α-TOH can mediate the oxidation of
polyunsaturated lipids in circulating LDLs,37−39 via the
thermodynamically favorable chain transfer reaction:
Figure 4. Inhibited periods (averages of at least three measurements)
observed for pyrimidinol 1, pyridinol 2, naphthyridinol 12a, and
PMHC as a function of antioxidant concentration when egg
phosphatidylcholine liposomes were oxidized with hydrophilic (A)
or lipophilic (B) peroxyl radicals. Standard potentials (vs NHE at 298
K in CH3CN) are given in the legend.
the aqueous/lipid interface. The results of these control
experiments reveal that neither N-acetylcysteine nor urate are
particularly effective at inhibiting H2B-PMHC oxidation under
these conditions (see Supporting Information). In contrast,
ascorbate is quite effective (Figure 5D), yielding a clear
inhibited period, which increases with increasing concentration
of ascorbate. The addition of a small (constant) amount of 12e
to the oxidations carried out in the presence of increasing
amounts of either N-acetylcysteine and urate (Figures 5B and
5C) led to a complete suppression in the rate of probe
oxidation similar to what is observed when 12e is used alone in
increasing amounts under these conditions (see Supporting
(9)
α‐TO• + L−H → α‐TOH + L•
Hence, under some (physiologically relevant) conditions, α-
TOH can mediate the peroxidation of lipids, rather than inhibit
the process. Ascorbate, which is water-soluble, does not inhibit
lipid peroxidation directly, but can do so only in a cooperative
manner with a lipid-soluble radical-trapping antioxidant such as
α-TOH.21 Furthermore, since ascorbate is such a powerful
reductant, it can act as a prooxidant through reactions with
product hydroperoxides40 or O2.41 The problems with β-
carotene are even more significant; it can trap peroxyl radicals
at low partial pressures of O2, but acts as a prooxidant at normal
1399
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
Figure 5. Representative fluorescence (at 520 nm) intensity−time profiles from MeOAMVN-mediated (0.68 mM) oxidations of egg
phosphatidylcholine liposomes (1 mM in PBS buffer, pH 7.4) containing 0.15 μM H2B-PMHC and 1.5 μM of naphthyridinol 12e in the presence of
various concentrations (0 μM, black; 1.5 μM, red; 3.0 μM, green; 7.5 μM, blue; and 15 μM, cyan) of (A) ascorbate, (B) N-acetylcysteine, and (C)
urate. Results of corresponding experiments lacking 12e are shown for ascorbate in (D), whereas those for N-acetylcysteine and urate, which show
no interaction between them and H2B-PMHC, are included in the Supporting Information.
Given the foregoing, the ideal peroxyl radical-trapping
antioxidant must, at a minimum, be a readily accessible
lipophilic compound that: (1) is more reactive to peroxyl
radicals than α-TOH; (2) is less reactive to chain transfer than
α-TOH; (3) maintains these reactivity differences when
embedded in lipid bilayers; (4) is at least as regenerable by
phase-separated (i.e., water-soluble) reductants as α-TOH; and
(5) has good bioavailability and limited metabolism in the liver,
which likely requires that it have high affinity for the tocopherol
transport protein.. If such a compound were available for in vivo
study, its use would almost certainly help in shedding definitive
light on whether lipid peroxidation has a causal or
consequential role in degenerative disease pathogenesis.
Our work to date on the naphthyridinols reveals that they
meet the first two of these criteria. The rate constant for the
reaction of the naphthyridinols 3 with peroxyl radicals is ∼30-
fold greater than that measured for α-TOH in homogeneous
organic solution and likely takes place with a negligible Ea, since
log k = 7.9 is very close to the expected log A = 8 for this
reaction.43 The O−H BDE of the naphthyridinol 3a (75.2 kcal/
mol)15 is ∼2 kcal/mol lower than the O−H BDE of α-TOH
Table 1. Effective Dissociation Constants (Kd,eff) for
Naphthyridinol Binding to Recombinant hTTP Measured by
Competition with Fluorogenic NBD-Toc (1 μM) in SET
Buffer, pH 7.4 at 20°C
compound
Kd,eff (μM)
a
12b
12c
nc
3.7 0.4
1.3 0.2
1.0 0.2
12d
12e
a
12f
nc
12g
1.0 0.1
0.1 0.1
1.1 0.2
12h
α-TOH
a
nc = noncompetitive with NBD-Toc.
partial pressures..42 In light of these chemical facts, one has to
wonder if it is wise to downplay the idea that radical-trapping
antioxidants may have a preventive role to play in vivo on the
basis of the clinical data collected to date.
1400
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
(77.2 kcal/mol),44 implying that the chain transfer reaction of
the naphthyridinol-derived aryloxyl radical and a polyunsatu-
rated lipid can be expected to be thermoneutral at best (the bis-
allylic C−H BDE is ∼76 kcal/mol).45 Indeed, naphthyridinols
3b16 and 517 did not appear to mediate lipid peroxidation to a
significant extent in our preliminary studies of LDL oxidation
under conditions where α-TOH did so. While these results
have been highly encouraging, we had yet to address the other
three criteria identified above. The recent development of the
chromanol-BODIPY conjugates as indicators of the antioxidant
status (and indirectly the extent of lipid peroxidation) in lipid
bilayers,30−32 in combination with some synthetic advances,28,29
prompted us to quantify how effective the naphthyridinols are
at trapping peroxyl radicals in lipid bilayers and to understand
the role of sidechain substitution on their reactivity and
regenerability in these systems.
Oxidations of H2B-PMHC-supplemented liposomes with
either hydrophilic or lipophilic peroxyl radicals were completely
suppressed by the naphthyridinols, regardless of the length or
branching of the sidechain. This is in contrast to the results
obtained with α-TOH and PMHC, which reveal an obvious
sidechain dependence in that the latter is much more efficient
at preventing probe oxidation than the former. The relative rate
constants can be determined from the initial rates of
fluorescence intensity increase and reveal a difference of
roughly 1 order of magnitude in the case of hydrophilic
radicals and ∼3-fold in the case of lipophilic radicals.30 The
former can be explained simply by the greater access of PMHC
to hydrophilic peroxyl radicals owing to the lack of a lipophilic
sidechain,57 while the latter may be approaching the
reproducibility of the measurement. While it could be expected
that a corresponding difference should exist for the
naphthyridinols of different sidechain substitution, it could
not be observed using the H2B-PMHC probe, presumably
because it is not sufficiently reactive to compete for peroxyl
radicals with any of the naphthyridinols. As a result, only an
estimate of the lower bounds of the reactivity of the
naphthyridinols toward peroxyl radicals of at least 30-fold
(relative to the probe) can be derived from these experiments.
Since the reactivity of the probe toward peroxyl radicals is
essentially the same as that of α-TOH,30 this provides a direct
comparison with the reactivity of α-TOH as well. The much
greater activity of the naphthyridinols is consistent with their
much higher inherent chemical reactivities (vide supra), which
may be further bolstered by the greater polarity of the
naphthyridinol moiety compared to the benzochromanol
moiety found in α-TOH and PMHC. While the naphthyridinol
is not expected to be either deprotonated (pKa of the O−H of
2 is 10.1)46 or protonated (pKa of pyridinium derived from 2 is
less than 6)46 to a significant extent at pH 7.4, the two nitrogen
atoms are expected to increase the polarity and H-bond
accepting ability of the naphthyridinol compared to the
benzochromanol.
nonradical products (i.e., eqs 5 and 6).9,21 The fact that the
inhibited periods of the less lipophilic naphthyridinols approach
one-third of those observed for PMHC and α-TOH (at the
upper-end of the concentrations we studied) implies that only a
corresponding fraction of peroxyl radicals are trapped by these
derivatives.
The reduced stoichiometries of the more hydrophilic
naphthyridinols can be explained by their direct reaction with
O2 (eq 10), which deplete the antioxidant. Since this reaction
can be expected to be faster in aqueous solution, which better
supports charge development in the electron transfer reaction,
it should be more apparent in the reactions of the more
hydrophilic compounds which can partition there.27 Since the
oxidation of phenols in aqueous solution is greatly accelerated
upon deprotonation of the phenol to yield its corresponding
phenoxide,47 follows that this pre-equilibrium should contribute
to the rate. Indeed, results of liposomal oxidations carried out at
slightly depressed pH (of 5.8), which would shift any
unfavorable, but kinetically relevant, pre-equilibrium and slow
the rate of competing autoxidation, support the latter
mechanism. Indeed, while there was no change in the inhibited
periods observed for PMHC (which has a higher pKa and is less
oxidizable, vide infra), there was a significant increase in the
inhibited periods observed for the naphthyridinol (12b) at the
lower (more acidic) pH.
Further evidence for the contribution of naphthyridinol
autoxidation to the shorter inhibited periods observed for the
less lipophilic derivatives was provided from the results of
liposome oxidations that were carried out on in the presence of
pyrimidinol 1 and pyridinol 2. While these compounds have log
P values in the same range as naphthyridinol 12a and PMHC
(see Supporting Information), they have redox potentials which
span the range between 12a (E° = 0.25 V) and PMHC (E° =
0.98 V) at E° = 0.71 and 0.47 V, respectively. Indeed, the
inhibited periods for each of the five concentrations we studied
follow the trend PMHC > pyrimidinol 1 > pyridinol 2 >
naphthyridinol 12a, which is clearly consistent with the trends
in E°; hence, the greater the one-electron oxidizability of the
compound, the lesser the inhibited period. Therefore, it would
appear that the longer inhibited periods for oxidations inhibited
by the more lipophilic compounds arise simply because their
partitioning to the lipid region protects them from
autoxidation, which proceeds much faster in the aqueous
phase. Consistent with this, additional experiments carried out
with lipophilic analogs of pyrimidinol 1 and pyridinol 2
(compounds 13 and 14, respectively, see Supporting
Information for details) show significantly longer inhibited
periods in corresponding liposome oxidations.
Although liposomal oxidations of the H2B-PMHC probe did
not reveal any differences in the apparent kinetics of the
reactions of the naphthyridinols of different sidechain
substitution with peroxyl radicals, the differences in the
inhibited periods are consistent with significantly different
reaction stoichiometries. This, again, is in contrast with α-TOH
and PMHC, which gave approximately equal inhibited periods
for the various concentrations that were studied, and is fully
consistent with previous work, which has demonstrated that
both of these phenols react with two peroxyl radicals to form
The ability of ascorbate to recycle α-TOH via reduction of
the α-tocopheroxyl radical (α-TO•) at the interfacial region of
the bilayer is believed to be key to the lipophilic peroxyl radical-
1401
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
trapping antioxidant activity of both compounds in vivo.
Consistent with this, oxidations of liposome-embedded H2B-
PMHC with lipophilic peroxyl radicals could be inhibited in a
dose-dependent manner by added ascorbate, presumably due to
scavenging of the probe-derived aryloxyl radical (eq 11). Upon
alone, the inhibited periods were almost independent of the
sidechain length. For example, where the inhibited periods
decreased from 155 to 125 to 62 min as the sidechain of the
naphthyridinol was shortened from n-hexadecyl to n-octyl to n-
butyl in oxidations inhibited by 7.5 μM of 12e, 12c, and 12b,
respectively, when any of the three were used at 1.5 μM along
with 6 μM of ascorbate, the inhibited periods were all within a
few minutes of each other (104 5). At first glance, it would
appear that ascorbate is protecting the more hydrophilic
naphthyridinol from autoxidation. However, this would deplete
the ascorbate, which would give rise to shorter inhibited
periods. However, careful consideration of the data obtained in
the absence of ascorbate reveals that under these conditions
and at a concentration of 1.5 μM, autoxidation of the
hydrophilic naphthyridinols is not yet competitive with peroxyl
radical trapping (see Supporting Information). Therefore, since
the maximum concentration of 12b cannot exceed 1.5 μM
throughout the experiment regardless of ascorbate concen-
tration, autoxidation is not a problem, and the dependence of
the inhibited period on sidechain length is minimized.
Analogous experiments were carried out with N-acetylcys-
teine and urate. Glutathione, as well as its biosynthetic
precursor cysteine, are very poor co-antioxidants with α-
TOH, and urate has not demonstrated any cooperativity at
all.51 Consistent with this, both N-acetylcysteine and urate were
not effective scavengers of the H2B-PMHC-derived aryloxyl
radical in liposomal oxidations. However, when a small amount
of lipophilic naphthyridinol was incorporated, as was the case
with ascorbate, probe oxidation was completely suppressed, and
the inhibited periods increased in a dose-dependent fashion
with increasing NAC or urate concentration. In this case, the
magnitude of the correlation of the length of the inhibited
period and coantioxidant concentration was only ∼42 and 38%
(NAC and urate, respectively) of what is observed with the
naphthyridinol alone, as compared to 65% for ascorbate (cf.
Figure 6). This is consistent with the fact that N-acetylcysteine
and urate are expected to be only one-electron reductants (eq
12), so maximum regeneration would yield a correlation of only
incorporation of a small, constant amount (1.5 μM) of lipid-
soluble naphthyridinol (e.g., the hexadecylated derivative 12e)
into the liposomes, probe oxidation was completely suppressed,
and the inhibited period could be increased in a dose-
dependent manner with increasing ascorbate concentration. In
fact, the data were almost indistinguishable from those obtained
from oxidations carried out in the presence of increasing
concentration of the naphthyridinol 12e alone.This provides
the first evidence that the naphthyridinols can indeed be
recycled with ascorbate as is the case for α-TOH, despite the
fact that the reaction is less exothermic by 2 kcal/mol (vide
supra). However, it should be pointed out that the magnitude
of the correlation of the inhibited period and coantioxidant
concentration was only ∼65% of what is observed for the
naphthyridinol alone (cf. Figure 6). Since the naphthyridinol
Figure 6. Crystallographic positions of amino acid sidechains in hTTP
in the vicinity of its ligand, α-TOH.52
traps 2 peroxyl radicals, as does PMHC and α-TOH, this means
that each molecule of ascorbate is supplying less than 2
reducing equivalents to regenerate the naphthyridinol (∼0.65 ×
2 = 1.3). Ingold,21 Barclay,48 and others have found that while
ascorbate can indeed regenerate α-TOH, this process often
proceeds without complete fidelity; a fact that has been
explained by the intervention of other competing reactions,
such as oxidation of ascorbate or the ascorbyl radical anion by
O2 in competition with either the disproportionation of the
ascorbyl radical anion or its reaction with an aryloxyl (such as
α-TO•), thereby effectively wasting reducing equivalents.
When analogous experiments were carried out with small
amounts (1.5 μM) of the other naphthyridinols (see
Supporting Information), we found the same proportional
increase in inhibited period with added ascorbate. However,
unlike liposome oxidations inhibited by the naphthyridinol
half that possible for ascorbate, which is a two-electron
reductant.We can only speculate that the increased polarity of
the naphthyridinol is responsible for its greater regenerability
by NAC and urate, since the thermodynamics are less favorable
than the corresponding reactions for α-TOH (vide supra).
Secretion of α-TOH from hepatic cells for delivery to
peripheral tissues and circulating lipoproteins is facilitated by
the tocopherol transport protein (TTP), which displays a clear
preference for α-TOH (Kd = 25 nM) over the other tocopherol
congeners (β-, γ-, and δ-TOH), with relative 1/Kd’s determined
to be 1:0.2:0.09:0.04 for α:β:γ:δ, respectively.24 The three-
dimensional structure of recombinant human TTP with α-
TOH bound has been determined to a resolution of 1.5 Å52
and clearly shows that the binding pocket has evolved to bind
α-TOH, with close contacts all along the periphery of the fully
methylated phenolic ring (Figure 6). Although it is known that
1402
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
replacement of the 16-carbon sidechain in α-TOH with a single
carbon carboxylic acid moiety (to give the water-soluble
tocopherol analog, trolox) increases the Kd over 40-fold,24 there
has never been a systematic study of sidechain length on
binding to the TTP.
no detectable oxidation of H2B-PMHC took place in the
presence of any of the naphthyridinols under conditions where
either α-TOH or its truncated analog PMHC were effective
only in retarding the rate of oxidation, indicating an
unprecedented peroxyl radical-trapping activity in lipid bilayers
and suggesting that the same difference in reactivity may be
expected in vivo. While sidechain length and/or branching did
not have an effect on the apparent reactivity of the
naphthyridinols to either hydrophilic or lipophilic peroxyl
radicals, it had a dramatic effect on their stoichiometry, with
more lipophilic compounds trapping about two peroxyl radicals
and more hydrophilic compounds trapping significantly fewer
than one. It is suggested that the more hydrophilic compounds
autoxidize in the aqueous phase and that the preferential
partitioning of the more lipophilic compounds to the lipid
phase protects them from this deleterious pro-oxidative
reaction. Studies at more acidic pH, as well as with hydrophilic
and hydrophobic pairs of less electron-rich pyridinols, and
related pyrimidinols support this assertion. The cooperativity of
the most lipophilic naphthyridinol with water-soluble reducing
agents was also studied in liposomes using H2B-PMHC.
Despite the fact that the naphthyridinol-derived radical is more
stable than the radical derived from α-TOH, it appeared to
have better regenerability by each of ascorbate, N-acetylcysteine
and urate, suggesting that this may be expected in vivo. It is
suggested that the greater polarity of the naphthyridinol moiety
makes regeneration more efficient than for α-TOH. Binding
assays with recombinant human TTP, a key determinant of the
bioavailability of the tocopherols, reveal that these compounds
have very high affinities. In fact, naphthyridinols with sidechains
of eight or more carbons had similar affinities (and in one case,
10-fold better) for the protein than α-TOH, suggesting that
similar bioavailabilities may be expected in vivo.
In previous work, we demonstrated that replacement of the
benzochromanol core of α-TOH with the tetrahydronaphthyr-
idinol core in 5 did not impair its binding to human TTP
(hTTP).17 In fact, it bound (marginally) better, a result we
ascribed to the potential for a strong H-bond between the
tertiary amine and Ser136. From the outset of the current
study, we were concerned that the attachment of the lipophilic
sidechain to the tertiary amine (in lieu of the methyl group in
5) and the corresponding removal of the quaternary center,
known to be a key determinant in why synthetic (racemic) and
natural (R)-α-TOH differ significantly in their bioavailabil-
ities,24 would disrupt binding. However, when we examined the
binding of our new compounds to recombinant hTTP, we were
gratified to learn that these changes did not significantly affect
binding. While the butyl- and isopentyl-substituted naphthyr-
idinols were noncompetitive with the fluorogenic NBD-Toc
probe for hTTP’s active site (implying very poor affinity) and
the octyl-substituted derivative was only a fair competitor, all
other derivatives had very similar or better binding than α-
TOH itself (cf. Table 1). The most lipophilic linear alkyl-
substituted naphthyridinol (hexadecylated 12e) had essentially
equivalent binding to hTTP when compared to α-TOH (Kd,eff
= 1.0
0.2 vs 1.1
0.2 μM), whereas the most lipophilic
branched alkyl-substituted compound (the farnesol-derived,
C15H31-substituted 12h) is roughly 10-fold better (Kd,eff = 0.1
0.1 μM). The latter is the best ligand for hTTP reported to date
and presumably demonstrates superior binding to the linear
analog because of advantageous interactions between the
sidechain methyl substituents and the binding pocket residues
evolved to bind the phytyl tail of the tocopherols.52 Of course,
whether superior binding is actually a desirable characteristic of
these analogs for in vivo studies is unclear, since the ligand must
eventually be released to peripheral tissues and circulating
lipoproteins in order to serve its purpose as a radical-trapping
antioxidant, and the mechanism of ligand release from TTP
remains unknown.53
Although the high affinity of the naphthyridinol derivatives
for the TTP is likely critical to ensure their appropriate
systemic distribution in animal models, it may also have
implications on any potential cytotoxicity of these compounds.
The strong binding of α-TOH to TTP is believed to contribute
to its slower metabolism compared to the other members of the
vitamin E family, which are rapidly metabolized by hepatic
CYP4F2 in order to improve their water-solubility and facilitate
excretion.54−56 Thus, the high affinity of the naphthyridinols to
TTP may also help protect them from metabolism in the liver.
EXPERIMENTAL SECTION
Synthesis. Complete details are provided in the Supporting
Information.
■
Liposome Oxidations. Liposome Preparation. Egg phosphati-
dylcholine was weighed (75 mg) in a dry vial and dissolved in a
minimum volume of chloroform. The solvent was then evaporated
under argon to yield a thin film on the vial wall. The film was left
under vacuum to remove any remaining solvent for 1 h. The lipid film
was then hydrated with 4.83 mL of a 10 mM phosphate buffered-saline
(PBS) solution containing 150 mM NaCl (pH 7.4), yielding a 20 mM
lipid suspension. The lipid suspension was subjected to 10 freeze−
thaw−sonication cycles, where each cycle involved storing the vial with
the solution in dry ice for 4 min, thawing at rt for 4 min, followed by 4
min of sonication. The lipid suspension was then extruded 20−25
times using a mini extruder equipped with a 100 nm polycarbonate
membrane.
Inhibited Oxidations. To individual 21.4 μL aliquots of the 20 mM
liposome solution were added increasing amounts (5, 10, 15, 20, and
25 μL, respectively) of a solution of the test antioxidant in either
aqueous acetonitrile (129 μM) or 5 μL of a solution of H2B-PMHC in
acetonitrile (12.9 μM). Each resultant solution was then diluted to 400
μL with PBS, from which 280 μL of each was loaded into a well of a
96-well microplate. The solution was equilibrated to 37 °C for 5 min,
after which 20 μL of a solution of azo compound (40.5 mM in 2,2′-
azobis-(2-amidinopropane)monohydrochloride (AAPH) in PBS or
10.1 mM in 2,2′-azobis-(4-methoxy-2,4-dimethylvaleronitrile) (MeO-
AMVN) in acetonitrile) was added to each well using the reagent
dispenser of the microplate reader. The fluorescence was then
monitored for 6 h at 50 s time intervals (λex = 485 nm; λem = 520 nm).
The final solutions in each well were 1 mM in lipids, 0.15 μM in H2B-
PMHC, 2.7 mM in AAPH, or 0.68 mM in MeOAMVN and either 1.5,
3.0, 4.5, 6.0, or 7.5 μM in antioxidant. Each liposome contained, on
CONCLUSIONS
■
Herein we have presented a series of naphthyridinol analogs of
α-TOH with varying sidechain substitution, which were
synthesized in an expeditious manner in order to determine
how systematic changes in the lipophilicity of these potent
antioxidants impact both their radical-trapping activities in lipid
bilayers, regenerability by water-soluble reductants and binding
to the human TTP. Liposomes supplemented with the different
naphthyridinols were oxidized using either hydrophilic or
lipophilic peroxyl radicals and consistently revealed a dose-
dependent protection of the fluorogenic H2B-PMHC. In fact,
1403
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
average, 15 fluorophores with an egg phosphatidylcholine/fluorophore
molar ratio of 6700:1. Under these conditions, no fluorescence self-
quenching was expected to occur within the liposome bilayer.30
Control experiments wherein the liposome solutions were purified by
size exclusion chromatography following supplementation with H2B-
PMHC and lipophilic antioxidants (e.g., α-TOH, 12e) revealed that
they are indeed incorporated into the liposomes using the above
procedure (see Supporting Information), as the fluorescence profiles
are essentially indistinguishable.
Co-Inhibited Oxidations. To individual 21.4 μL aliquots of the 20
mM liposome solution were added 5 μL of a solution of 12e in
aqueous acetonitrile (129 μM) or simply 5 μL of acetonitrile only,
followed by 5 μL of a solution of H2B-PMHC in acetonitrile (12.9
μM). To the aliquots of the two sets of samples were then added
increasing amounts (0, 5, 10, 25, and 50 μL) of solutions of water-
soluble reducing agent (ascorbate, urate, or N-acetylcysteine) in PBS
(129 μM). Each resultant solution was then diluted to 400 μL with
PBS, from which 280 μL was loaded into a well of a 96-well
microplate. The microplate was equilibrated to 37 °C for 5 min, after
which 20 μL of a solution of MeO-AMVN in acetonitrile (10.1 mM)
was added to each well using the reagent dispenser of the microplate
reader. The fluorescence was then monitored for 6 h at 50 s time
intervals (λex = 485 nm; λem = 520 nm). The final solutions in each
well were 1 mM in lipids, 0.15 μM in H2B-PMHC, 0.68 mM in
MeOAMVN, 0 or 1.5 in 12e μM and either 0, 1.5, 3.0, 7.5, or 15 μM
in water-soluble reducing agent.
Voltammetry. Standard potentials for the oxidation of 1, 2, 12a,
and PMHC were determined from cyclic voltammagrams measured
using a three-electrode cell equipped with a glassy carbon working
electrode, a platinum counter electrode, and an Ag/AgNO3 reference
electrode. Voltammagrams were obtained at a scan rate of 100 mV/s
in dry acetonitrile using Bu4N·PF6 (0.1 M) as electrolyte at 25 °C. All
potentials are reported vs the normal hydrogen electrode (NHE) via
reference to the ferrocene/ferrocenium couple.
hTTP Binding Affinity by Competition Studies. Recombinant
hTTP (0.35 μM) was incubated with 1.0 μM NBD-Toc in SET buffer
(250 mM sucrose, 1 mM EDTA, 50 mM Tris, pH 7.4) at
approximately 20 °C. After reaching equilibrium, the maximum
fluorescence was recorded, and a small aliquot of competitor dissolved
in EtOH was added. Approximately 10 min was required for the
competitor to reach equilibrium again before the new fluorescence
value was recorded. The competitor aliquots were added in increasing
concentrations, and the total volume of the competitor added did not
go above 1% of the total volume of the solution in the cuvette. Data
were plotted using the Prism software package (Prism GraphPad v4.0)
and analyzed using nonlinear regression for a one-site competition
model, from which effective dissociation constants, Kd,eff, were
calculated for competing ligands 12c−e, 12g, and 12h as well as α-
TOH.
ACKNOWLEDGMENTS
■
We are grateful to Katerina Krumova for helpful discussions
and to the Natural Sciences and Engineering Research Council
of Canada and the Canada Foundation for Innovation for
grants to D.A.P., J.A., and G.C. D.A.P. also acknowledges the
support of the Canada Research Chairs program.
REFERENCES
■
(1) Steinberg, D.; Parthasarathy, S.; Carew, T. E.; Khoo, J. C.;
Witztum, J. L. New Engl. J. Med. 1989, 320, 915−924.
(2) Witztum, J. L.; Steinberg, D. J. Clin. Invest. 1991, 88, 1785−1792.
(3) Simonian, N. A.; Coyle, J. T. Annu. Rev. Pharmacol. Toxicol. 1996,
36, 83−106.
(4) Barnham, K. J.; Masters, C. L.; Bush, A. I. Nat. Rev. Drug
Discovery 2004, 3, 205−214.
(5) Marnett, L. Carcinogenesis 2000, 21, 361−370.
(6) Benz, C. C.; Yau, C. Nat. Rev. Cancer 2008, 8, 875−879.
(7) Porter, N. A. Acc. Chem. Res. 1986, 19, 262−268.
(8) Pratt, D. A.; Tallman, K. A.; Porter, N. A. Acc. Chem. Res. 2011,
44, 458−467.
(9) Burton, G. W.; Ingold, K. U. Acc. Chem. Res. 1986, 19, 194−201.
(10) Valgimigli, L.; Pratt, D. A. In Encyclopedia of Radicals in
Chemistry, Biology and Materials; John Wiley & Sons, Ltd: Chichester,
U.K., 2012.
(11) For a good discussion, see: Lonn, M. E.; Dennis, J. M.; Stocker,
R. Free Radical Biol. Med. 2012, 53, 863−884 and references cited
therein.
̈
(12) Steinhubl, S. Am. J. Cardiol. 2008, 101, S14−S19.
(13) Pratt, D. A.; DiLabio, G. A.; Brigati, G.; Pedulli, G. F.;
Valgimigli, L. J. Am. Chem. Soc. 2001, 123, 4625−4626.
(14) Pratt, D. A.; DiLabio, G. A.; Mulder, P.; Ingold, K. U. Acc. Chem.
Res. 2004, 37, 334−340.
(15) Wijtmans, M.; Pratt, D. A.; Valgimigli, L.; DiLabio, G. A.;
Pedulli, G. F.; Porter, N. A. Angew. Chem., Int. Ed. 2003, 42, 4370−
4373.
(16) Kim, H.-Y.; Pratt, D. A.; Seal, J. R.; Wijtmans, M.; Porter, N. A.
J. Med. Chem. 2005, 48, 6787−6789.
(17) Nam, T.-G.; Rector, C. L.; Kim, H.-Y.; Sonnen, A. F. P.; Meyer,
R.; Nau, W. M.; Atkinson, J.; Rintoul, J.; Pratt, D. A.; Porter, N. A. J.
Am. Chem. Soc. 2007, 129, 10211−10219.
(18) Niki, E.; Noguchi, N. Acc. Chem. Res. 2004, 37, 45−51.
(19) Golumbic, C.; Mattill, H. A. J. Am. Chem. Soc. 1941, 63, 1279−
1280.
(20) Frei, B.; England, L.; Ames, B. N. Proc. Natl. Acad. Sci. U.S.A.
1989, 86, 6377.
(21) Doba, T.; Burton, G. W.; Ingold, K. U. Biochim. Biophys. Acta
1985, 835, 298−303.
(22) Traber, M. G.; Burton, G. W.; Ingold, K. U.; Kayden, H. J. J.
Lipid Res. 1990, 31, 675−685.
(23) Burton, G.; Traber, M.; Acuff, R.; Walters, D.; Kayden, H.;
Hughes, L.; Ingold, K. Am. J. Clin. Nutr. 1998, 67, 669−684.
(24) Panagabko, C.; Morley, S.; Hernandez, M.; Cassolato, P.;
Gordon, H.; Parsons, R.; Manor, D.; Atkinson, J. Biochemistry 2003,
42, 6467−6474.
(25) Valgimigli, L.; Brigati, G.; Pedulli, G. F.; DiLabio, G. A.;
Mastragostino, M.; Arbizzani, C.; Pratt, D. A. Chem. Eur. J. 2003, 9,
4997−5010.
ASSOCIATED CONTENT
■
S
* Supporting Information
Complete synthetic details and compound characterization,
supplemental fluorescence intensity−time profiles referred to in
the text but not shown, representative cyclic voltammagrams
and binding curves used to determine effective dissociation
constants in the TTP assays. This material is available free of
charge via the Internet at http://pubs.acs.org/.
(26) Wijtmans, M.; Pratt, D. A.; Brinkhorst, J.; Serwa, R.; Valgimigli,
L.; Pedulli, G. F.; Porter, N. A. J. Org. Chem. 2004, 69, 9215−9223.
(27) We (and others) speculated that similar reactions lead to
differences in the stoichiometries observed in autoxidations of
multilamellar phosphatidylcholine liposomes inhibited by water-
soluble vitamin B12-derived pyridinol antioxidants in previous work.
See: Serwa, R.; Nam, T.-G.; Valgimigli, L.; Culbertson, S.; Rector, C.
L.; Jeong, B.-S.; Pratt, D. A.; Porter, N. A. Chem.Eur. J. 2010, 16,
14106−14114.
AUTHOR INFORMATION
■
Corresponding Author
dpratt@uottawa.ca
Notes
(28) Nara, S. J.; Jha, M.; Brinkhorst, J.; Zemanek, T. J.; Pratt, D. A. J.
Org. Chem. 2008, 73, 9326−9333.
The authors declare no competing financial interest.
1404
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405

Journal of the American Chemical Society
Article
(29) Lu, J.; Cai, X.; Hecht, S. M. Org. Lett. 2010, 12, 5189−5191.
(30) Krumova, K.; Friedland, S.; Cosa, G. J. Am. Chem. Soc. 2012,
134, 10102−10113.
(31) Oleynik, P.; Ishihara, Y.; Cosa, G. J. Am. Chem. Soc. 2007, 129,
1842−1843.
(32) Krumova, K.; Oleynik, P.; Karam, P.; Cosa, G. J. Org. Chem.
2009, 74, 3641−3651.
(33) Barclay, L. R. C.; Locke, S. J.; MacNeil, J. M.; VanKessel, J.;
Burton, G. W.; Ingold, K. U. J. Am. Chem. Soc. 1984, 106, 2479−2481.
(34) Burton, G. W.; Ingold, K. U. J. Am. Chem. Soc. 1981, 103, 6472−
6477.
(35) Noguchi, N.; Yamashita, H.; Gotoh, N.; Yamamoto, Y.;
Numano, R.; Niki, E. Free Radical Biol. Med. 1998, 24, 259−268.
(36) Nava, P.; Cecchini, M.; Chirico, S.; Gordon, H.; Morley, S.;
Manor, D.; Atkinson, J. Bioorg. Med. Chem. 2006, 14, 3721−3736.
(37) Bowry, V. W.; Ingold, K. U. Acc. Chem. Res. 1999, 32, 27−34.
(38) Bowry, V. W.; Stocker, R. J. Am. Chem. Soc. 1993, 115, 6029−
6044.
(39) Ingold, K. U.; Bowry, V. W.; Walling, C.; Stocker, R. Proc. Natl.
Acad. Sci. U.S.A. 1993, 90, 45−49.
(40) Lee, S. H.; Oe, T.; Blair, I. A. Science 2001, 292, 2083−2086.
(41) Taqui Khan, M. M.; Martell, A. E. J. Am. Chem. Soc. 1967, 89,
4176−4185.
(42) Burton, G.; Ingold, K. Science 1984, 224, 569−573.
(43) Benson, S. W. Thermochemical Kinetics; John Wiley & Sons,
1976.
(44) Lucarini, M.; Pedulli, G. F.; Cipollone, M. J. Org. Chem. 1994,
59, 5063−5070.
(45) Trenwith, A. B. J. Chem. Soc. Faraday Trans. 1 1982, 78, 3131.
(46) Nam, T.-G.; Nara, S. J.; Zagol-Ikapitte, I.; Cooper, T.;
Valgimigli, L.; Oates, J. A.; Porter, N. A.; Boutaud, O.; Pratt, D. A.
Org. Biomol. Chem. 2009, 7, 5103.
(47) Weinberg, D. R.; Gagliardi, C. J.; Hull, J. F.; Murphy, C. F.;
Kent, C. A.; Westlake, B. C.; Paul, A.; Ess, D. H.; McCafferty, D. G.;
Meyer, T. J. Chem. Rev. 2012, 112, 4016−4093.
(48) Barclay, L. R. C.; Locke, S. J.; MacNeil, J. M. Can. J. Chem. 1983,
61, 1288−1290.
(49) Barclay, L. J. Biol. Chem. 1988, 263, 16138−16142.
(50) Motoyama, T.; Miki, M.; Mino, M.; Takahashi, M.; Niki, E.
Arch. Biochem. Biophys. 1989, 270, 655−661.
(51) Niki, E.; Saito, M.; Yoshikawa, Y.; Yamamoto, Y.; Kamiya, Y.
Bull. Chem. Soc. Jpn. 1986, 59, 471−477.
(52) Min, C. K.; Kovall, R. A.; Hendrickson, W. A. Proc. Natl. Acad.
Sci. U.S.A. 2003, 100, 14713−14718.
(53) Morley, S.; Cecchini, M.; Zhang, W.; Virgulti, A.; Noy, N.;
Atkinson, J.; Manor, D. J. Biol. Chem. 2008, 283, 17797.
(54) Birringer, M.; Drogan, D.; Brigelius-Flohe, R. Free Radical Biol.
Med. 2001, 31, 226−232.
(55) Sontag, T. J.; Parker, R. S. J. Biol. Chem. 2002, 277, 25290−
25296.
(56) Sontag, T. J.; Parker, R. S. J. Lipid Res. 2007, 48, 1090−1098.
(57) Fluorescence measurements suggest that α-tocopherol is in a
largely aggregated state in phospholipid membranes (of rat liver
microsomes), which may contribute directly to a slower rate of
reaction with peroxyl radicals as well as reduced mobility between
layers of the bilayer. See Kagan, V. E.; Serbinova, E. A.; Bakalova, R. A.;
Stoytchev, Ts. S.; Erin, A. N.; Prillipko, L. L.; Evstigneeva, R. P.
Biochem. Pharmacol. (Amsterdam, Neth.) 1990, 40, 2403−2413.
1405
dx.doi.org/10.1021/ja309153x | J. Am. Chem. Soc. 2013, 135, 1394−1405