Modifications of boronic ester pro-chelators triggered by hydrogen peroxide tune reactivity to inhibit metal-promoted oxidative stress

Article abstract of DOI:10.1039/b705199a

Several new analogs of salicylaldehyde isonicotinoyl hydrazone (SIH) and salicylaldehyde benzoyl hydrazone (SBH) that contain an aryl boronic ester (BSIH, BSBH) or acid (BASIH) in place of an aryl hydroxide have been synthesized and characterized as maske

Full text of DOI:10.1039/b705199a

View Article Online / Journal Homepage / Table of Contents for this issue
An international journal of inorganic chemistry
www.rsc.org/dalton
Number 43 | 21 November 2007 | Pages 4873–5092
ISSN 1477-9226
HOT ARTICLE
Katherine J. Franz et al.
HOT ARTICLE
Philip J. Blower and
Modifications of boronic ester
pro-chelators triggered by hydrogen
peroxide tune reactivity to inhibit
metal-promoted oxidative stress
Stephen J. Mather et al.
How do HYNIC-conjugated peptides
bind technetium? Insights from
LC-MS and stability studies

View Article Online
PAPER
www.rsc.org/dalton | Dalton Transactions
Modifications of boronic ester pro-chelators triggered by hydrogen peroxide
tune reactivity to inhibit metal-promoted oxidative stressࠤ
Louise K. Charkoudian, David M. Pham, Ashley M. Kwon, Abbey D. Vangeloff and Katherine J. Franz*
Received 5th April 2007, Accepted 19th June 2007
First published as an Advance Article on the web 19th September 2007
DOI: 10.1039/b705199a
Several new analogs of salicylaldehyde isonicotinoyl hydrazone (SIH) and salicylaldehyde benzoyl
hydrazone (SBH) that contain an aryl boronic ester (BSIH, BSBH) or acid (BASIH) in place of an aryl
hydroxide have been synthesized and characterized as masked metal ion chelators. These pro-chelators
show negligible interaction with iron(III), although the boronic acid versions exhibit some interaction
with copper(II), zinc(II) and nickel(II). Hydrogen peroxide oxidizes the aryl boronate to phenol, thus
converting the pro-chelators to tridentate ligands with high affinity metal binding properties. An X-ray
crystal structure of a bis-ligated iron(III) complex, [Fe(SBH(m-OMe)3)2]NO3, confirms the meridonal
binding mode of these ligands. Modifications of the aroyl ring of the chelators tune their iron affinity,
whereas modifications on the boron-containing ring of the pro-chelators attenuate their reaction rates
with hydrogen peroxide. Thus, the methoxy derivative pro-chelator (p-OMe)BASIH reacts with
hydrogen peroxide nearly 5 times faster than the chloro derivative (m-Cl)BASIH. Both the rate of
pro-chelator to chelator conversion as well as the metal binding affinity of the chelator influence the
overall ability of these molecules to inhibit hydroxyl radical formation catalyzed by iron or copper in
the presence of hydrogen peroxide and ascorbic acid. This pro-chelator strategy has the potential to
improve the efficacy of medicinal chelators for inhibiting metal-promoted oxidative stress.
is associated with retinal toxicity hypothesized to result from
Introduction
overall retinal iron deficiency.9
Oxidative stress is implicated in a wide variety of diseases, includ-
Our strategy to develop targeted iron chelators is to synthesize
ing but not limited to diabetes, atherosclerosis, aging, macular
“masked” chelators that have little to no affinity for metal ions
degeneration, and neurodegeneration.1–4 Depending on properties
until the mask is selectively removed by the presence of reactive
of their coordination environment, metal ions like copper and iron
oxygen species such as hydrogen peroxide (H2O2). The goal is to
can catalyze the production of highly reactive hydroxyl radicals
develop reagents that are activated only under disease conditions
(OH•)5 that damage lipids, proteins and nucleic acids. Metal-
to reveal high-affinity iron-binding ligands that compete for the
redox-active iron that is the source of OH• generation. Our first-
promoted oxidative stress may therefore be a critical component
generation pro-chelator, BSIH,18 contains a boronic ester in place
of a phenolic oxygen that is a key donor atom of salicylaldehyde
isonicotinoyl hydrazone (SIH), a well-studied high-affinity, mem-
brane permeable ligand that scavenges and incapacitates redox-
active iron.19–21 Aryl boronic acids and their esters are well known
to deboronate via oxidation of the carbon–boron bond initiated
by nucleophilic attack by H2O2, followed by aryl migration to
a borate intermediate that rapidly hydrolyzes to give the phenol
and borate ester or boric acid.22 This reactivity has made boronic
esters attractive targets for the development of highly selective
fluorescent probes for H2O2.23–27 Boronic acids do not appear to
have intrinsic toxicity issues, and the end product boric acid is
considered non-toxic to humans.28 These properties coupled with
their relative stability make them useful for a variety of biological
and medicinal applications.28
As shown in Scheme 1, H2O2 reacts selectively to convert
BSIH to SIH, which forms a stable bis-ligated complex with
Fe3+. In this report, we present the synthesis, characterization,
and hydroxyl radical inhibition capacities of six additional BSIH
analogs together with their unmasked chelator versions and
iron complexes. By varying the substituents on these analogs,
several properties of the pro-chelator/chelator combinations
can be tuned, including lipophilicity, iron binding affinity, and
in diseases where normal metal ion homeostasis is impaired or
where aberrant metal accumulation occurs.6 For example, the iron
load in the substantia nigra brain region of Parkinson’s disease
patients is 35% higher than that in healthy age-matched controls,7,8
and iron levels in maculas of those 65 or older are higher than
those under 65.9 Inhibiting iron-promoted oxidative stress by
inactivating catalytic iron is therefore a promising strategy for
treating such diseases.10–17
While chelating agents developed for iron-overload diseases
have been proposed for use in degenerative diseases, their intrinsic
affinity for other essential metals like zinc and copper may
have toxic consequences. In addition, their high affinity for iron
results in competition with essential iron-containing enzymes. For
example, treatment of macular degeneration with desferrioxamine
Department of Chemistry, Duke University, Durham, NC, 27708–0346,
USA. E-mail: katherine.franz@duke.edu; Fax: +1 919 660 1605; Tel: +1
919 660 1541
† Based on the presentation given at Dalton Discussion No. 10, 3rd–5th
September 2007, University of Durham, Durham, UK.
‡ CCDC reference numbers 643169–643174. For crystallographic data in
CIF or other electronic format see DOI: 10.1039/b705199a
§ Electronic supplementary information (ESI) available: Crystallographic
details, including full atom-numbering schemes, and NMR spectra. See
DOI: 10.1039/b705199a
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 5031–5042 | 5031
©

View Article Online
and aryl aldehydes in acidic aqueous or methanolic solution.
This class of aroyl hydrazone compounds has shown wide
utility as tridentate metal chelating agents.29 The pro-chelators
presented here are based on two families of aroyl hydrazone
ligands, those of salicylaldehyde isonicotinoyl hydrazone (SIH)
and salicylaldehyde benzoyl hydrazone (SBH). Our first generation
pro-chelator replaced the phenol of SIH with a boronic acid
pinacol ester and was given the abbreviation BSIH.18 Derivatives
in which a boronic acid moiety replaces the boronic ester
are denoted BASIH. Substitution on the B- or OH-containing
ring is designated by the substituent in parentheses before the
abbreviation. For example, the pro-chelators (p-OMe)BASIH and
(m-Cl)BASIH and their corresponding chelators (p-OMe)SIH and
(m-Cl)SIH in Fig. 1a. Derivatives with substituents on the aroyl
ring are based on the SBH ligand and are designated with the
substituent in parentheses after the abbreviation. These com-
pounds include the pro-chelators BSBH(m-OMe), BSBH(m-OEt)
and BSBH(OMe)3 and their corresponding chelators SBH(m-
OMe), SBH(m-OEt) and SBH(OMe)3, all shown in Fig. 1b.
Finally, benzaldehyde isonicotinoyl hydrazone (BIH) was also
prepared as a control compound that contains the hydrazone
functionality but lacks the key metal-binding phenol group
(Fig. 1c).
Scheme 1 Boronate-masked aroyl hydrazone pro-chelators “BL” are
weak metal-binding ligands, but are converted by H2O2 to chelators “L”
that bind Fe3+ to form bis-ligated [FeL2]+ complexes. X and Y are chloro,
methoxy, or ethoxy substituents, while Z is either N or CH.
rate of H2O2-dependent unmasking of the pro-chelator to the
chelator.
Results and discussion
Synthesis, nomenclature and structure of pro-chelators and
chelators
The compounds shown in Fig. 1 were synthesized by a straightfor-
ward Schiff base condensation reaction between aryl hydrazides
Fig. 1 Chemical structures and abbreviations of the pro-chelators (left-hand columns) and chelators (right-hand columns) used in this study. (a)
Derivatives of salicylaldehyde isonicotinoyl hydrazone (SIH) and their boronic ester (BSIH) or boronic acid (BASIH) prochelators with modifications on
the B/OH-containing ring, (b) derivatives of salicylaldehyde benzoyl hydrazone (SBH) and their boronic ester prochelators (BSBH) with modifications
on the aroyl ring, (c) benzaldehyde isonicotinoyl hydrazone.
5032 | Dalton Trans., 2007, 5031–5042
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Fig. 2 ORTEP structural diagrams showing 50% probability ellipsoids and partial atom numbering schemes for (a) the pro-chelator BSBH(m-OMe)
and (b) its chelating version SBH(m-OMe).
As with BSIH,18 the X-ray crystal structures of BSBH(m-OMe)
and BSBH(m-OEt) reveal that in the crystalline solid the B atom in
all cases is anti to the imine N1 atom and the configuration about
=
the C7 N1 bond is E. The E configuration is retained in the
structures of the corresponding chelators, but the phenolic OH
assumes a syn conformation. Fig. 2 shows the structures of the
BSBH(m-OMe) and SBH(m-OMe) pro-chelator/chelator pair; a
summary of X-ray diffraction parameters is found in Table 1 and
selected bond distances and angles are collected in Table 2. The
structures of m-ethoxy analogs BSBH(m-OEt) and SBH(m-OEt)
as well as the trimethoxy SBH(OMe)3 chelator are found in the
ESI.§
The syn conformation adopted by the chelators is stabilized
by a hydrogen bond between the phenolic hydrogen and the
˚
˚
imine nitrogen (dSBH(m-OMe) = 2.653(1) A; dSBH(m-OEt) = 2.619(1) A;
˚
dSBH(OMe)3 = 2.656(3) A). This conformation favorably arranges the
carbonyl O, imine N and phenol O on the same face for tridentate
metal binding.
As anticipated, the bis-ligated [FeL2]+ complexes contain a ferric
ion in distorted octahedral geometry with two SIH-type ligands
bound in meridonal configuration with tridentate coordination.
An example is shown in Fig. 3 for the SBH(OMe)3 derivative.
Selected bond lengths and angles are in Tables 2 and 3. Although
crystals have also been obtained for [Fe(SIH)2]+ and [Fe(SBH(m-
OEt))2]+ and preliminary diffraction data reveal similar structural
trends, [Fe(SBH(OMe)3)2]+ is the only derivative that has provided
high-quality data to date.
Fig. 3 ORTEP structural diagram showing 50% probability ellipsoids
and partial atom numbering scheme for [Fe(SBH(MeO)3)2]+. Each SBH
ligand coordinates Fe through its phenolate O (O5, O10), imine N (N1,
N3), and carbonyl O (O1, O6). The phenolate O is the one masked by B
in the pro-chelators. The nitrate counter anion is not shown.
A comparative analysis of 15 derivatives of SIH and SBH metal
complexes of the first and second row transition metals shows
that the positions of the three donor atoms shift on average
A Cambridge Structural Database search of aroyl hydrazones
returns several derivatized SIH, PIH and SBH analogs (where
PIH is pyridoxyl isonicotinoyl hydrazone) along with a diverse
number of metal complexes spanning from first row transition
metals to uranium. Among this rich array of complexes there
are surprisingly few iron examples, and these only include mono-
ligated complexes of the type [FeCl2(MeOH)(L)].30–32 The structure
of [Fe(SBH(OMe)3)2]+ in Fig. 3 therefore represents the first
structurally characterized example of the bis-ligated form, which
is the expected speciation in biological environments of pH 7.4.
Table 3 highlights the consistency in the iron coordination
geometry and bond lengths between the mono-ligated structures
and that of [Fe(SBH(OMe)3)2]NO3. The phenolate O–Fe distance
˚
only ∼0.04 A across the series. Furthermore, projection of a
set of 18 SIH and SBH ligand structures onto the Fe(SBH)
˚
coordinates shows that the free ligands need only shift ∼0.09 A to
accommodate iron binding (see ESI§). This analysis illustrates that
even in the absence of a coordinating metal center, derivatives of
SIH and SBH are conformationally pre-arranged with optimized
geometry for metal coordination.
Metal specificity
Initial assessment of BSIH and iron suggested that a strong Fe3+
complex only forms once the pro-chelator is converted to the SIH
chelator.18 In order to assess the interaction between the pro-
chelators and metals in more detail, UV-vis experiments were
conducted at concentrations that enable visualization of weak
˚
is the shortest donor atom-to-Fe distance in all cases, being 1.88 A
in the [Fe(SBH(OMe)3)2]NO3 example.
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 5031–5042 | 5033
©

View Article Online
Table 1 Summary of X-ray diffraction parameters
[Fe(SBH(OMe)3)2]NO3 · Pentanes
SBH(m-OMe)
SBH(m-OEt)
SBH(OMe)3
BSBH(m-OMe)
[BSBH(m-OEt)]2·H2O
Formula
FW
C39 H34 Fe N5 O13
836.56
11.7263(11)
28.622(3)
12.2079(13)
90
96.464(5)
90
4071.3(7)
4
C15 H14 N2 O3
270.28
11.4645(9)
13.6942(11)
8.5467(8)
90
92.025(4)
90
1341.0(2)
4
C16 H16 N2 O3
284.31
11.5979(3)
14.0704(3)
9.4947(3)
90
110.1330(10)
90
1454.74(7)
4
C17 H18 N2 O5
330.33
14.787(7)
13.198(6)
8.225(2)
90
101.262(5)
90
1574.2(11)
4
C21 H25 B N2 O4
380.24
17.8735(16)
12.7925(10)
9.0933(9)
90
90
90
2079.2(3)
4
C44 H56 B2 N4 O9
806.55
18.293(7)
8.639(3)
29.494(10)
90
100.504(5)
90
4583(3)
4
˚
a/A
˚
b/A
˚
c/A
a/◦
b/◦
c /◦
V/A
Z
3
˚
Cryst. syst.
Space group
T/K
Monoclinic
P 21/c
Monoclinic
P 21/c
Monoclinic
P 21/c
Monoclinic
P 21/c
173(2)
0.71073
1.394
Orthorhombic
P c a 21
298(2)
0.71073
1.215
Monoclinic
P 21/n
298(2)
173(2)
298(2)
298(2)
˚
k/A
0.71073
1.365
0.44
0.71073
1.339
0.095
0.71073
1.298
0.091
0.71073
1.179
0.081
q/g cm−3
l/mm−1
0.104
0.083
R1 (obsd data)
wR2 (all data,
F2 refinement)
0.731
0.2733
0.0415
0.1122
0.0392
0.1081
0.05
0.1092
0.0427
0.1185
0.0828
0.2631
Table 2 Selected bond lengths and bond angles of chelators, prochelators, and [Fe(SBH(OMe)3)2]NO3
[Fe(SBH(OMe)3)2]
NO3·Pentanesa
SBH(m-OMe)b
SBH(m-OEt)b
SBH(OMe)3
BSBH(m-OMe)c
[BSBH(m-OEt)]2·H2Oc
a
˚
Bond lengths/A
O1–C8
C1–X
N1–N2
N1–C7
N2–C8
1.266(5)
1.312(5)
1.376(5)
1.281(6)
1.325(5)
1.231(1)
1.354(2)
1.383(1)
1.276(2)
1.339(2)
1.224(1)
1.353(2)
1.374(1)
1.271(2)
1.352(2)
1.238(3)
1.349(3)
1.375(3)
1.277(3)
1.334(3)
1.231(3)
1.557(4)
1.381(2)
1.272(3)
1.358(3)
1.237(4)
1.562(6)
1.375(4)
1.273(5)
1.357(5)
Bond angles/◦
O1–C8–C9
O1–C8–N2
N2–C8–C9
N1–N2–C8
N2–N1–C7
N1–C7–C6
C1–C6–C7
C6–C1–X
122.3(4)
119.0(4)
118.7(4)
116.4(3)
118.3(3)
124.3(4)
122.9(4)
122.1(5)
121.6(1)
123.1(1)
115.3(1)
120.2(1)
115.5(1)
121.7(1)
122.8(1)
121.8(1)
121.7(1)
122.2(1)
116.1(1)
119.0(1)
118.4(1)
121.2(1)
121.9(1)
122.0(1)
122.5(2)
123.2(2)
114.3(2)
121.5(2)
114.2(2)
122.9(2)
122.5(2)
122.0(2)
121.3(2)
122.4(2)
116.4(2)
118.7(2)
115.8(2)
119.1(2)
117.2(2)
126.9(2)
121.0(3)
122.0(3)
117.0(3)
118.9(3)
115.5(3)
118.7(3)
121.9(3)
125.8(3)
a X = O5. b X = O3. c X = B1.
complexes with Cu2+, Zn2+, Ni2+ and Fe3+. Studies were performed
in methanol to ensure solubility of metal salts and chelators.
Fig. 4 shows the visible spectrum between 500–800 nm of a
methanolic solution containing 10 mM Fe(NO3)3 and 10 mM L,
where L is one of the pro-chelators BASIH or BSIH, their corre-
sponding chelator SIH, or the potentially bidentate hydrazone
BIH. In the absence of metal, none of the L compounds has
an absorbance feature in this wavelength range, and Fe(NO3)3
displays only a weak tail from a band centered around 360 nm.
Solutions containing Fe3+ and SIH are a deep red color and
exhibit a broad d–d absorbance band in this region. In contrast,
addition of BIH to Fe(NO3)3 results in only a very subtle color
change from yellow to pale orange, with a feature around 530 nm.
Replacing BIH by either BASIH or BSIH results in a slightly
more intense orange color and an increase in the shoulder around
530 nm. The spectra of the BASIH and BSIH solutions are nearly
indistinguishable, suggesting a similar mode of interaction of the
boronic acid and the boronic ester versions of the pro-chelators
with Fe3+. They are distinct from the control BIH, however,
suggesting that the boron-containing moiety may weakly interact
with the metal in addition to the O/N hydrazone feature that all
of these compounds share. The pyridine N on the aroyl ring can
also interact with metal ions in all of these cases.
When the 10 mM solutions are diluted to 100 lM, the features
described above for BSIH, BASIH and BIH are barely perceptible.
Furthermore, the pale orange color of the 10 mM solutions
disappears when phosphate buffer is added to the methanol
solutions, suggesting that phosphate is sufficient for competing
Fe3+ from these compounds. In contrast, the SIH complex remains
intact in phosphate buffer. This analysis suggests that the pro-
chelators are unlikely to compete in a biological setting for iron
binding until they are unmasked to reveal the chelating version.
5034 | Dalton Trans., 2007, 5031–5042
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Table 3 Comparison of selected bond lengths and bond angles for iron complexes
[FeCl2(S)L]
[FeL4 ]NO3
2
L1
L2
L3
Ligand 1
Ligand 2
˚
Bond lengths/A
=
O–Fe (C O)
N–Fe
2.068(2)
2.119(2)
1.874(2)
2.033(3)
2.135(4)
1.897(3)
2.012(6)
2.131(6)
1.905(6)
2.091(3)
2.116(3)
1.887(4)
2.071(3)
2.091(3)
1.884(3)
O–Fe (PhO−)
Bond angles/◦
=
(C O) O–Fe–N
O–Fe–O
74.4(1)
159.3(1)
85.5(1)
74.3(1)
158.2(1)
84.8(1)
75.5(2)
159.8(2)
84.3(2)
74.8(1)
159.4(1)
84.7(1)
167.1(1)
89.1(1)
97.0(2)
90.2(1)
90.6(1)
74.7(1)
159.6(1)
85.1(1)
N–Fe–O (PhO−)
N1–Fe–N3
O1–Fe–O6
O5–Fe1–O10
O1–Fe1–O10
O5–Fe1–O6
L1 = SBH (ref. 30); L2 = (p-MeO)SBH(m-OMe) (ref. 31); L3 = PIH (ref. 32); L4 = SBH(OMe3) (this work); S = MeOH for L1 and L2 and H2O for L3.
Fig. 4 Visible spectra of 10 mM methanol solutions of Fe3+ with 10 mM
Fig. 5 Visible spectra of 10 mM methanol solutions of CuSO4 with 10 mM
added BASIH, BIH, or BSIH. The d–d band of Cu2+ centered around
800 nm is most affected by the addition of BIH and BASIH.
added BASIH, BIH, BSIH, or SIH. Only SIH forms intensely colored
solutions with Fe3+
.
In contrast to the very weak interaction with Fe3+, BIH does
form a complex with Cu2+, as evidenced by precipitate formation
upon addition of BIH to a methanolic solution of CuSO4 and a
corresponding color change from very pale blue–green to more
intense green. The spectra in Fig. 5 show the shift of the Cu2+ d–d
band from 800 nm for CuSO4 to 700 nm for the mixture with
BIH, clearly indicating a change in coordination environment
around Cu2+. The carbonyl O and the imine N of hydrazones
are known to form bidentate complexes with Cu2+.33 When BIH is
replaced by BSIH, there is less of a color change and less precipitate
formation, although the solution does become cloudy, suggesting
some complexation. In the case of Cu2+, the spectra with BSIH
and BASIH are distinct. As shown in Fig. 5, the d–d band for
the Cu2+ plus BASIH solution is more similar to that of the BIH
solution than the BSIH solution. Not as much precipitate forms,
however, and solutions of Cu2+ and BASIH in methanol are a
bright aquamarine color as opposed to the green of the BIH
complex. As seen in the Fe3+ case, mixtures of Cu2+ with SIH
display the most intense colors among this series. The spectrum is
not shown in Fig. 5 since the complex completely precipitates from
solution. Complexes of SBH-type ligands with Cu2+ are known.34,35
Collectively, these results suggest that the N/O unit and/or the
pyridine N of these aroyl hydrazones indeed complex Cu2+, and
that the boronic acid and boronic ester moieties also participate
in these interactions, but to different extents. It appears from
these studies that the boronic acid version interacts more strongly
than the ester version. Studies with Ni2+ and Zn2+ support
the observation that boronic acid versions of the pro-chelators
are more prone to metal interactions than the boronic ester
counterparts. As shown in Fig. 6, the d–d band of NiSO4 in
methanol does not shift upon addition of BIH or BSIH, but
does shift considerably upon addition of BASIH. The shift in
absorbance coincides with considerable precipitate formation,
which is not observed with BIH or BSIH and Ni2+. As with
the other metal ions, an intense color change accompanies the
addition of SIH to Ni2+, with considerable precipitation of an
orange solid. In the case of Zn2+, mixtures with BSIH or BIH
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 5031–5042 | 5035
©

View Article Online
were prepared to ensure complete formation of FeL2, which was
monitored by its absorbance at 480 nm. In the case of SIH, we
found that a 1 : 3 : 10 ratio of Fe3+ : SIH : EDTA resulted in a
decrease of the initial [Fe(SIH)2]+ concentration to just under 50%,
as shown by the bar graph in Fig. 7. These same concentrations
and ratios were then used for the other analogs in order to establish
their relative affinity compared to SIH.
Fig. 6 Visible spectra of 10 mM methanol solutions of NiSO4 with 10 mM
added BASIH, BIH, or BSIH. The d–d band of Ni2+ is most affected by
the addition of BASIH.
remain clear, while those with BASIH become cloudy and with
SIH considerable precipitate forms.
These studies suggest that boronic ester analogs of this class of
pro-chelators will be preferable for future in vivo studies in order to
prevent unwanted interactions with metal ions prior to activation
to the chelator form.
Fig. 7 Percentage of the [FeL2]+ species (as monitored at 480 nm), that
persists in solutions containing 0.15 mM FeCl3, 0.45 mM chelator (L), and
1.5 mM EDTA (Fe : L : EDTA = 1 : 3 : 10) equilibrated overnight in 50 :
50 methanol–phosphate buffer (20 mM, with 500 mM NaCl, pH 7.4).
Relative stability of iron complexes
The thermodynamic stability of iron–chelator complexes is one
of several important properties that can influence the therapeutic
application of these molecules.14 Although pre-organized, hex-
adentate ligands generally have very high affinities for Fe3+, their
large size can limit their access across cellular membranes. The
smaller, tridentate chelators of the PIH/SIH family have molecular
weights under 500, but retain strong affinity for ferric iron. The
pre-organization of the 3 donor atoms described in the structural
analysis above is likely a contributing factor to the stability of these
complexes. A direct comparison of overall formation constants
is complicated when comparing ligands of varying pKa’s and
different binding stoichiometries, therefore it is often useful to
compare pFe (−log[Feaq3+]) values, which provide an assessment
of the amount of uncomplexed iron that would be available at
pH 7.4 with total ligand and iron concentrations of 10 lM and
1 lM, respectively.36 Variations of reported pFe values for SIH
and PIH ligands range from 23 to 50.37–39 For comparison, the pFe
of EDTA is 25.1, while that of desferrioxamine is 26.6.36 Although
the very high value for SIH is unrealistic since it is known not to
compete with desferrioxamine, the range gives an indication of the
effectiveness of this class of chelators. The wide discrepancy in their
pFe values arises from difficulties encountered with measuring
accurate stability constants for these compounds due to solubility
problems as well as hydrolysis of the ligands under the extreme
low and high pH values accessed during typical potentiometric
titrations.38–40
The data in Fig. 7 show that altering the electronic nature of the
phenol-containing ring with either an electron-donating methoxy
group or an electron-withdrawing chloro group has little influence
on the relative stability of the chelators. Both (p-OMe)SIH and (m-
Cl)SIH retain 40–45% of their iron when challenged with EDTA
under these conditions. In contrast, both of the SBH derivatives
investigated retain only about 10% of their signal, indicating that
they are weaker binders compared with the SIH analogs.
These data collectively indicate that the electron-donating pyri-
dine ring of SIH derivatives plays a significant role in stabilizing
high affinity Fe3+ complexes. The data also corroborate prior
studies by placing SIH higher than EDTA on a relative pFe scale.
Kinetics of pro-chelator to chelator conversion
Since the rates of conversion of the pro-chelators to their active
chelator versions will have important ramifications for their
potential biological applications, we compared the oxidation rates
of analogs BSIH, BASIH, BSBH(MeO)3, (p-OMe)BASIH, and
(m-Cl)BASIH under pseudo first-order conditions of excess H2O2.
In order to keep all of the pro-chelators and chelators in solution
during these reactions, all kinetic runs were carried out in a mixed
solvent system of 50% methanol, 50% 20 mM phosphate buffer at
pH 7.4. Reactions were monitored spectrophotometrically. Fig. 8
shows the conversion of 50 lM (p-OMe)BASIH to (p-OMe)SIH
with 1 mM H2O2 as a representative example of a typical kinetics
run. The isosbestic points at 270 and 330 nm indicate clean
conversion between the two species. As shown in Fig. 9, the
observed pseudo first-order rate constants (kobs) show a linear
In this study, we chose to estimate the relative binding ability
among our series of chelators by monitoring the competition
between FeL2 and EDTA in solution conditions of 20 mM
phosphate buffer at pH 7.4 with 50% methanol and 500 mM NaCl.
Methanol was added as a co-solvent to ensure that all species
stayed in solution. Initial solutions of Fe3+ and 3-fold excess ligand
5036 | Dalton Trans., 2007, 5031–5042
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
chelators were dissolved in 100 mM aqueous NaOH and frozen
for 48 h. No changes in the UV-vis spectrum were noted after the
solutions were defrosted and allowed to sit at room temperature
for at least 4 h, indicating their stability under these conditions.
Inhibition of hydroxyl radical formation
We have previously shown that the H2O2-activated pro-chelator
BSIH inhibits the production of OH• produced under Fenton
conditions nearly as effectively as the chelator SIH itself.18 In
order to test how the chemical modifications on the BSIH
framework influence this property, we subjected the new pro-
chelators (m-Cl)BASIH, (p-OMe)BASIH, and BSBH(m-OMe) to
the deoxyribose assay in order to compare them with BSIH and
SIH. In this assay, hydroxyl radicals generated by reaction of
reduced iron and H2O2 degrade 2-deoxyribose into by-products
that are monitored spectrophotometrically after reaction with
thiobarbituric acid (TBA).41 Chelating agents that prevent the
metal from reacting with H2O2 result in a decrease in absorption
of the TBA-reactive species at 532 nm. Quenchers that scavenge
hydroxyl radicals also decrease the amount of TBA-reactive
species, although usually at higher concentrations compared to
effective chelating agents.
Fig. 8 Conversion of 50 lM (p-OMe)BASIH to (p-OMe)SIH by oxida-
tion with 1 mM H2O2 in 50 : 50 MeOH–20 mM phosphate buffer, pH 7.4.
The spectra represent the first 30 min of the reaction.
The data shown in Fig. 10 are presented as the ratio of the
absorbance at 532 measured in the presence (A) or absence
(Ao) of chelator or pro-chelator. A/Ao values below 1 indicate
protection against deoxyribose degradation, whereas values above
1 suggest an enhancement of metal-mediated radical formation.
Both BSIH and SIH provide nearly maximal protection against
deoxyribose degradation under these conditions with pro-chelator
concentrations as low as 50 lM. The BSBH analog, however,
does not achieve this level of protection until 200 lM of the pro-
chelator has been added. In addition, the slight increase in A/Ao
above 1 suggests that low concentrations of an SBH chelator may
actually support iron in an environment that promotes Fenton
chemistry, for examplevia mono-chelated binding. The behavior of
the BSBH(m-OMe) observed in the deoxyribose assay is consistent
Fig. 9 Plots of kobs vs H2O2 concentration for the conversion of the listed
pro-chelators to their respective chelator version. The second-order rate
constant, k, was obtained from the slope of these lines: BSIH k = 0.053,
BASIH k = 0.044, BSBH(MeO)3 k = 0.049, (p-OMe)BASIH k = 0.077,
(m-Cl)BASIH k = 0.017 M−1 s−1
.
dependence on the concentration of H2O2; the second-order rate
constant (k) for each analog was obtained from the slope of this
line.
A comparison of the data shown in Fig. 9 reveals that changing
the boronic ester functionality of BSIH (k = 0.053 M−1 s−1)
to the boronic acid analog BASIH (k = 0.044 M−1 s−1) has
little influence on the rate of H2O2-dependent oxidation to the
phenol. Likewise, modifying the aroyl ring, as in the case of
BSBH(MeO)3 (k = 0.049 M−1 s−1), has little influence on the
conversion rate. Chemically modifying the boron-containing aryl
ring, however, causes significant changes to the rate. For example,
placing an electron-donating methoxy group para to the boronic
acid in (p-OMe)BASIH results in a 1.45-fold acceleration (k =
0.077 M−1 s−1). In contrast, the electron-withdrawing chloro
substituent in (m-Cl)BASIH slows the reaction down 3-fold (k =
0.017 M−1 s−1).
Fig. 10 Dependence of pro-chelator concentration on deoxyribose degra-
dation by hydroxyl radicals generated under Fenton conditions (200 lM
H2O2, 10 lM FeCl3, 2 mM ascorbic acid, 15 mM 2-deoxyribose in pH 7.4
phosphate buffer). A and Ao are the absorbance readings at 532 nm with
and without added chelator. Values of A/Ao < 1 signify protection against
deoxyribose degradation by OH•.
In order to test the stability of the pro-chelators in conditions
that may be used for future cell culture experiments, the pro-
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 5031–5042 | 5037
©

View Article Online
with the results described in the previous section that indicate SBH
analogs are weaker iron binders compared to the SIH analogs.
A comparison of (m-Cl)BASIH and (p-OMe)BASIH shows that
the rate of H2O2 activation does indeed correlate with the overall
inhibition of OH• formation. The p-OMe derivative provides the
fastest conversion from pro-chelator to chelator, and as shown
in Fig. 10 also provides the most protection against deoxyribose
degradation. In contrast, the m-Cl derivative is unmasked by H2O2
nearly 5 times slower than the p-OMe version, and after an hour
of being subjected to Fenton conditions in the deoxyribose assay,
this analog provides only modest protection. BSIH, which has
a rate constant in between these two derivatives, also shows an
intermediate level of protection. Given the rate constants obtained
from Fig. 9, only 2.4 lM of the (m-Cl)SIH would be present from
200 lM pro-chelator at the time the reactions are monitored. In
contrast, nearly 8 lM SIH and 11 lM (p-OMe)SIH would be
available. It is important to note that the reaction conditions used
for the kinetic analysis in Fig. 9 and the deoxyribose assay in
Fig. 10 are different, so these calculations only provide relative
estimates of predicted concentrations. Overall, the trend in the
protective effect of the pro-chelators mirrors their reactivity with
H2O2, with the more readily unmasked analogs providing the
maximum inhibition of OH• formation.
Fig. 11 Dependence of pro-chelator concentration on deoxyribose
degradation by hydroxyl radicals generated under copper-induced Fenton
conditions (100 lM H2O2, 10 lM CuSO4, 2 mM ascorbic acid, 15 mM
2-deoxyribose in pH 7.4 phosphate buffer). BSIH was also tested without
addition of H2O2 (“BSIH no H2O2”). A and Ao are the absorbance
readings at 532 nm with and without added chelator. Values of A/Ao <
1 signify protection against deoxyribose degradation by OH•. Bsal is
(2-formylphenyl)boronic acid pinacol ester, which was used as a control to
demonstrate that the reaction of a boronic ester with H2O2 alone does not
provide the protective effect observed for the chelators.
Copper is another biologically important metal ion that can
undergo Fenton-like reactions depending on its coordination
environment, and although its concentration in cells is lower than
that of iron, it is a more effective catalyst for OH• generation.1
SIH is known to inhibit copper-mediated OH• formation,42
a
conclusion verified in our studies for comparison purposes,
as shown by the deoxyribose degradation data presented in
Fig. 11. BSIH also shows a protective effect, although complete
suppression of TBA-reactive species is never fully achieved, even
in the presence of 40-fold excess chelator to copper, the highest
ratio tested. As shown in Fig. 11, desferrioxamine (DFO) provides
80–90% protection when present in concentrations equivalent or
higher than copper (10 lM). This effect is slightly better than
SIH, which provides about 75% protection above 20 lM, and
BSIH, which shows about 65% protection above 50 lM. The
boronate-masked salicylaldehyde Bsal ((2-formylphenyl)boronic
acid pinacol ester) does not show an effect on the A/Ao value until
the highest concentrations, indicating that consumption of H2O2
via reaction with the aryl boronic ester functionality is not the
source of the protective effect of the pro-chelator in this assay.
In the absence of added H2O2, mixtures of copper and ascorbic
acid in air are still capable of generating OH•. Whereas SIH and
DFO are effective chelators for inhibiting these reactions (data
not shown), BSIH is only effective at concentrations greater than
200 lM, as shown in Fig. 11. Since H2O2 is required to convert
BSIH to its tridentate chelator SIH,18 it is likely that this moderate
protective effect is due to complex formation between BSIH itself
and copper. As described above, BSIH–Cu complexes are weak,
therefore relatively high concentrations of the ligand are required
to form a complex that disfavors Fenton-type reactions.
Partition coefficients
A compound’s ability to permeate biological membranes via
passive diffusion is partly a function of its lipophilicity, which
is conveniently assessed by its partition coefficient, P, between
n-octanol and water.43 While several computer programs are
available to calculate partition coefficients based on molecular
fragments, such calculations for aroyl hydrazones can differ from
experimental values by as much as 3 log units.44 We therefore chose
to measure these values directly in an octanol–buffer system.
Table 4 lists the log P values assembled for several of the
compounds in Fig. 1 and their corresponding bis-ligated ferric
complexes. All of the compounds studied are moderately hy-
drophobic, with log P values ranging from 1.9 to 2.9. The pro-
chelators (BL) as a trend are less lipophilic than either their free
ligand (L) or their iron complex (FeL2). This trend is reversed for
the trimethoxy SBH derivative, where BSBH(OMe)3 has a log P
of 2.9, while its chelator and iron complex have log P values of 2.5.
The biological activity of iron chelators to mobilize iron from
cellular stores has been found to correlate to the lipophilicity
Table 4 Octanol/water partition coefficients (log P) of chelators (L), pro-
chelators (BL) and iron complexes [FeL2]+
log P (octanol–water)
L
(BL)
[FeL2]+
To test this hypothesis further, BIH was also tested in the
deoxyribose assay (in the presence of H2O2). As shown in Fig. 11,
its behavior tracks very similarly to that of BSIH without H2O2.
These data support the conclusion that the potentially bidentate
hydrazone functionality of this class of compounds interacts with
copper.
SIH
2.4
2.9
2.9
2.5
2.3a 1.9b
2.3
2.3
2.4
2.4
2.9
2.5
SBH(m-OMe)
SBH(m-OEt)
SBH(OMe)3
2.9
a BSIH. b BASIH.
5038 | Dalton Trans., 2007, 5031–5042
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
of both the free ligands and their iron complexes.44–48 For aroyl
hydrazones, ligands with log PL ∼2.8 and a log PFeL ∼3.1
show optimal iron mobilization activity.44 The parent BSIH pro-
chelator with log P = 2.3 falls slightly below this optimal range,
but the data in Table 4 show that modifications to the parent
compound can be made to increase the lipophilicity of the pro-
chelators.
Experimental
General considerations
Chemicals were obtained from Fisher Scientific or Acros Organics
and used without further purification unless otherwise noted;
(2-formylphenyl)boronic acid pinacol ester was purchased from
Combi-Blocks, Inc; (5-chloro-2-formylphenyl)boronic acid and
(4-methoxy-2-formylphenyl)boronic acid were purchased from
Matrix Scientific. SIH,50,51 BSIH, and [Fe(SIH)2]NO3 were pre-
pared as described previously.18 All solvents were reagent grade
and all aqueous solutions were prepared from nanopure water.
UV-vis spectra were recorded on a Phototonics Model 420 Fiber
Conclusions
In an effort to improve the utility of medicinal metal chelators, we
have introduced a pro-chelator strategy in which metal-binding
ligands are masked by protecting groups that are released under
conditions of oxidative stress to expose high-affinity ligands. Here,
we have introduced a series of pro-chelators in which modifications
to the aroyl hydrazone framework of the parent BSIH compound
tune their properties and reactivity. For example, replacing the
boronic ester masking group by a boronic acid reduces the
lipophilicity of the pro-chelator, but does not affect the rate of
H2O2-dependent conversion to the chelator. Neither the boronic
ester nor the boronic acid versions interact significantly with Fe3+,
although the boronic acid versions do form methanol-insoluble
precipitates with divalent metal ions such as Cu2+, Zn2+, and
Ni2+. The boronic ester versions are therefore likely to be more
favorable for future medicinal or biological applications, as it
will be desirable to avoid premature metal binding. In addition,
boronic acids readily bind vicinal diols such as carbohydrates,49
which could further complicate the availability of the chelator.
Replacing the pyridine ring of the SIH framework with an
aryl ring of SBH decreases the Fe3+ binding affinity of the
ligand, which in turn lowers its efficacy to inhibit iron-promoted
hydroxyl radical formation. The electron-donating nature of the
N atom in the pyridine ring provides more electron density on
the carbonyl O donor for improved metal binding. In terms of
metal affinity, this modification has more influence than changing
the electronic nature of the opposing phenol ring. As assessed by
EDTA competition studies, very little difference in relative iron
affinity was observed between the electron-donating p-OMe and
the electron-withdrawing m-Cl derivatives. Future studies will be
required to provide a more quantitative evaluation of binding
affinity.
Where these modifications do have significant impact is in the
rate of H2O2-dependent conversion from the pro-chelator to the
chelator. A substituent that is electron-donating with respect to
the boron accelerates this rate, while an electron-withdrawing
group slows it down. In moving from (p-OMe)BASIH to BASIH
to (m-Cl)BASIH, a nearly 5-fold difference in conversion rate is
observed. This range is highly advantageous for future biological
and medicinal studies, since the advantage of the pro-chelators
will be linked to their triggered activation only under conditions
of oxidative stress. Increasing the rate is clearly linked to improved
protection against metal-promoted oxidative stress. On the other
hand, analogs with slower rates will be more robust in avoiding
untimely chelator release by harmless background levels of H2O2.
How these chemical modifications affect in vivo function will be
a key step in validating the pro-chelator strategy for inhibiting
metal-promoted oxidative stress and is the subject of future
investigations.
Optic CCD Array UV-vis spectrophotometer. H and 13C NMR
1
spectra were recorded on a Varian Mercury 300, Inova 400 or
500 spectrometer; d values are in ppm and J values are in Hz.
IR Spectra were recorded on a Nicolet 360 FT-IR. Ionization
mass spectrometry (ESI-MS) was preformed on an Agilent 1100
Series LC/MSD trap spectrometer with a Daly conversion dynode
detector. Samples were infused via a Harvard apparatus syringe
pump at 33 lL min−1. Ionization was achieved in the positive- or
negative-ion mode by application of +5 or −5 kV at the entrance
to the capillary; the pressure of ◦the nebulizer gas was 20 psi. The
drying gas was heated to 325 C at a flow rate of 7 L min−1.
Full-scan mass spectra were recorded in the mass/charge (m/z)
range of 100–2000. Elemental analysis was performed by Desert
Analytics/TransWest Geochem, Tucson, AZ.
Synthesis
BASIH. Isonicotinic acid hydrazide (1 mmol, 0.137 g) was
dissolved in 2 mL of 0.1 M sodium acetate buffer (pH 4.5) and
added to a solution of (2-formylphenyl)boronic acid (1 mmol,
0.145 g) dissolved in 1 mL of methanol. The reaction was stirred
for 5 min at 100 ◦C and cooled over ice. The white precipitate
was collected via vacuum filtration, washed with water and dried
in vacuo to yield a white powder (215 mg, 80%). 1HNMR (CD3OD
400 MHz): d 7.34 (3H, m), 7.56 (1H, s), 7.84 (2H, d, J = 5.88),
8.42 (1H, s), 8.71(2H, d, J = 5.98). 13CNMR (CD3OD 101 MHz):
d 121.94, 128.57, 129.98, 131.23, 149.82, 150.75. ESI-MS: m/z
267.8 (M − H)−, 292 (M + Na)+. IR (neat, mmax/cm−1): 3204,
3056, 2360, 1547, 1379, 1357, 1338, 1195, 1168, 1145, 766, 693.
kmax(MeOH)/nm 302 (e/dm3 mol−1 cm−1 15 800); kmax(octanol) 258
(7800), 300 (9300).
(p-OMe)BASIH. A similar procedure to the one described
above for BASIH was followed, but replacing (2-formylphenyl)-
boronic acid with 4-methoxy-2-formylphenylboronic acid. A white
powder was isolated in 70% yield. 1HNMR (CD3OD 400 MHz): d
3.83 (3H, s), 6.98 (1H, dd, J = 8.14, 2.30), 7.28 (1H, d, J = 8.15),
7.36 (1H, d, J = 2.15), 7.84 (2H, d, J = 5.75), 8.41 (1H, s), 8.71
(2H, d, J = 5.08). 13CNMR (CD3OD 101 MHz): d 54.55, 112.77,
116.07, 121.95, 132.79, 149.84, 150.70. ESI-MS: m/z 322.0 (M +
Na)+, 297.8 (M − H)−. IR (neat, mmax/cm−1): 2204, 1647, 1552,
1358, 748, 618. kmax(MeOH)/nm 320 (e/dm3 mol−1 cm−1 19 700).
(p-OMe)SIH. A portion of (p-OMe)BASIH (0.5 mmol, 150
mg) was dissolved in 10 mL of MeOH. As H2O2 (11 mmol, 0.62 mL
50% H2O2) was added dropwise to the solution, the reaction
mixture turned clear yellow. After stirring for 2 h, solvent was
removed in vacuo to yield a yellow powder (123 mg, 91%). 1HNMR
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 5031–5042 | 5039
©

View Article Online
(CD3OD 400 MHz): d 3.74 (3H, s), 6.81 (1H, d, J = 8.94), 6.89
(1H, dd, J = 8.96, 3.02), 7.06 (1H, d, J = 2.99), 7.85 (2H, dd,
4.58, 1.57), 8.51 (1H, s), 8.71 (2H, d, J = 5.84). 13CNMR (CD3OD
101 MHz): d 55.14, 117.38, 118.31, 119.23, 122.15, 149.86, 152.95.
ESI-MS: m/z 269.9 (M − H)−.
J = 7.67), 7.45 (2 H, m), 7.53 (2 H, m), 8.65 (1 H, s), 11.29 (1 H, s),
12.08 (1 H, s). 13CNMR (DMSO, 101 MHz d ppm 55.28, 112.79,
116.34, 117.64, 118.6, 119.27, 119.75, 129.41, 129.66, 131.32,
134.09, 148.25, 157.38, 159.17, 162.47. ESI-MS m/z 268.9 (M −
H)−, 293.1 (M + Na)+. IR (neat, mmax/cm−1): 3450, 2244, 1668,
1283, 816, 754. kmax(octanol)/nm 288 (e/dm3 mol−1 cm−1 18 700),
332 (13 700).
(m-Cl)BASIH. A similar procedure to the one described above
for BASIH was followed, but replacing (2-formylphenyl)boronic
acid with 5-chloro-2-formylphenylboronic acid. A white powder
was isolated in 78% yield (236 mg). 1HNMR (CD3OD 400 MHz):
d 7.33 (1H, s), 7.40 (1H, d, J = 7.82), 7.60 (1H, d, J = 6.75), 7.84
(2H, d, J = 2.28), 8.40 (1H, s), 8.71 (2H, s). ESI-MS: m/z 325.9
(M + Na)+ 301.8 (M − H)−. IR (neat, mmax/cm−1): 3294, 2158,
2023, 1653, 1581, 1340, 1306, 1024, 704, 689. kmax(MeOH)/nm
302 (e/dm3 mol−1 cm−1 21 500).
BSBH(m-OEt). One equivalent of (2-formylphenyl)boronic
acid pinacol ester was added to a 2-mL solution of 3-ethoxybenz-
hydrazide (1 mmol, 185 mg) in ethanol and heated to 60 ◦C. After
stirring for 10 min, the solvent was removed in vacuo. The
white solid was washed with diethyl ether and dried in vacuo to
1
give 200 mg of a white powder (50% yield). HNMR: (DMSO,
500 MHz): d 1.35 (15H, s), 3.32 (1H, s), 4.09 (2H, q, J = 6.8), 7.14
(1H, d, J = 7.45), 7.43 (4H, m), 7.54 (1H, t, 7.38), 7.71 (1H, d, J =
7.31), 8.00 (1H, d, J = 7.74), 8.93 (1H, s), 11.93 (1H, s). 13C NMR
(DMSO 125 MHz): d 14.51, 24.51, 63.2, 83.83, 113.59, 117.47,
119.79, 125.49, 128.84, 129.55, 130.94, 134.96, 135.39, 139.57,
148.26, 158.32, 163.25. ESI-MS m/z 395.1 (M + H+). IR (neat,
mmax/cm−1): 3199, 2978, 1642, 1548, 1479, 1346, 1273, 1139, 1050,
965, 857, 806, 762, 655. kmax(MeOH)/nm 298 (e/dm3 mol−1 cm−1
20 000).
(m-Cl)SIH. A similar procedure to the one described above for
(p-OMe)SIH was followed, but replacing (p-OMe)BASIH with (m-
Cl)BASIH. A yellow powder was isolated in 89% yield (122 mg).
1HNMR (CD3OD 400 MHz): d 6.91 (2H, m), 7.46 (1H, d, J =
8.05), 7.85 (2H, d, J = 4.47), 8.51 (1H, s), 8.72 (1H, s). ESI-MS:
m/z 273.9 (M − H)−.
BIH. Equimolar quantities of isonicotinic acid hydrazide
(1 mmol, 0.137 g) and benzaldehyde (1 mmol, 0.102 mL) were
dissolved in 2 mL of 0.1 M sodium acetate buffer (pH 4.5); the
reaction was stirred for 5 min at 100 ◦C and cooled over ice. The
white precipitate was collected via vacuum filtration, washed with
water and dried in vacuo to yield 153 mg of a white precipitate (67%
yield). 1HNMR (CD3OD 400 MHz): d 7.41 (3H, m), 7.81 (1H, dd,
J = 6.68, 2.97), 7.86 (2H, dd, J = 4.51, 1.67), 8.33 (1H, s), 8.71
(2H, dd, J = 4.54, 1.64). 13CNMR (CD3OD 101 MHz): d 121.95,
127.75, 128.65, 130.72, 134.01, 141.22, 149.89, 150.67. ESI-MS
m/z 223.8 (M − H)−, 247.9 (M + Na)+. IR (neat, mmax/cm−1):
2357, 2172, 2000, 1684, 1564, 1285, 686. kmax(MeOH)/nm 301
(e/dm3 mol−1 cm−1 19 500).
SBH(m-OEt). A similar procedure to the one described above
for SBH(m-OMe) was followed on a 0.5 mmol scale, but using
3-ethoxybenzhydrazide. 93 mg of a white powder product was
isolated (65% yield). 1HNMR (DMSO, 400 MHz): d 1.35 (3H, t,
J = 6.8), 4.10 (2H, q, J = 6.8), 6.92 (2H, m), 7.14 (1H, d, J =
7.6), 7.16 (1H, t, J = 7.6), 7.46 (4H, m), 8.64 (1H, s), 11.28 (1H,
s), 12.05 (1H, s). 13CNMR (DMSO 101 MHz): d 14.51, 63.24,
113.3, 116.33, 117.93, 118.58, 119.25, 119.66, 129.42, 129.66,
131.3, 134.02, 148.22, 157.38, 158.42, 162.43. ESI-MS m/z 285.1
(M + H+), 307.2 (M + Na+). IR (neat, mmax/cm−1): 3191, 3056,
1649, 1599, 1548, 1481, 1357, 1293, 1238, 1137, 1082, 957, 816,
723, 680. kmax(MeOH)/nm 288 (e/dm3 mol−1 cm−1 22,200), 298
(21 200), 330 (16 000).
BSBH(m-OMe). One equivalent of (2-formylphenyl)boronic
acid pinacol ester was added via syringe to a 2-mL solution of 3-
methoxybenzhydrazide (1 mmol, 166 mg) in methanol and heated
to 60 ◦C. After stirring for 10 min, 1 mL of chilled diethyl ether was
added and the reaction was placed over ice. A white precipitate
was collected via vacuum filtration, washed with diethyl ether, and
dried in vacuo to provide 284 mg of a white powder (72% yield).
1HNMR (DMSO, 500 MHz d ppm 1.33 (12H, s), 3.83 (3H, s), 7.16
(1H, d, J = 7.39 Hz), 7.45 (4H, m), 7.54 (1H, t, J = 7.40 Hz), 7.71
(1H, d, J = 7.31 Hz), 8.01 (1H, d, J = 7.81 Hz), 8.94 (1H, s), 11.96
(1H, s) 13CNMR (DMSO 125 MHz): d 24.51, 55.25, 83.83, 113.06,
117.17, 119.87, 125.47, 128.85, 129.54, 130.93, 135.00, 135.38,
139.54, 148.31, 159.06, 163.25. ESI-MS m/z 381.1 (M + H)+. IR
(neat, mmax/cm−1): 3204, 3060, 2963, 1643, 1589, 1544, 1482, 1349,
1276, 1136, 1056, 965, 862, 805, 758, 695, 655. kmax(MeOH)/nm
298 (e/dm3 mol−1 cm−1 22 600).
BSBH(OMe)3. A similar procedure to the one described above
for BSBH(m-OMe) was followed, but using 3,4,5-trimethoxy-
benzhydrazide. A white powder (550 mg) was isolated in 81% yield.
1HNMR (DMSO, 500 MHz): d ppm 1.33 (12H, s), 3.72 (3H, s),
3.86 (6H, s), 7.19 (2H, s), 7.43 (1H, t, J = 5.51), 7.54 (1H, m),
7.71 (1H, d, J = 6.53), 7.98 (1H, d, J = 5.89), 8.92 (1H, s), 11.84
(1H, s). 13CNMR (DMSO 75 MHz): d 24.52, 56.03, 60.04, 83.84,
105.34, 125.62, 128.81, 130.92, 135.34, 139.47, 140.27, 148.35,
152.56, 163.02. ESI-MS m/z 441.1 (M + H)+. IR (neat, mmax/cm−1):
2965, 1651, 1584, 1501, 1453, 1414, 1324, 1269, 1228, 1142, 1123,
1062, 1003, 858, 777, 726. kmax(MeOH)/nm 300 (e/dm3 mol−1 cm−1
25 100).
SBH(OMe)3. A similar procedure to the one described above
for SBH(m-OMe) was followed, but using 3,4,5-trimethoxy-
benzhydrazide. 185 mg of a white powder was isolated (56% yield).
1HNMR (DMSO, 400 MHz): d ppm 3.73 (3H, s), 3.86 (6H, s), 6.92
(2H, t, J = 8.63), 7.29 (3H, m), 7.55 (1H, d, J = 7.38), 8.65 (1H, s),
11.24 (1H, s), 11.94 (1H, s). 13CNMR (DMSO, 101 MHz): d ppm
38.80, 39.01, 39.22, 39.43, 39.64, 39.85, 40.06, 56.02, 60.05, 105.14,
116.31, 118.67, 119.27, 127.79, 129.21, 131.29, 140.50, 147.83,
152.65, 157.31, 162.13. ESI-MS m/z 331.2 (M + H+), 353.2 (M +
Na+). IR (neat, cm−1): 2996, 1643, 1573, 1495, 1452, 1410, 1330,
SBH(m-OMe). One equivalent of salicylaldehyde was added
to a 2-mL solution of 3-methoxybenzhydrazide (0.5 mmol, 83 mg)
in methanol in a 10-mL round bottomed flask. The reaction was
stirred for 10 min at 60 ◦C and cooled over ice. White needle-like
crystals suitable for X-ray diffraction were collected via vacuum
filtration (98 mg, 74% yield). 1H NMR (DMSO, 400 MHz d ppm
3.83 (3 H, s), 6.93 (2 H, m), 7.17 (1 H, d, J = 7.86), 7.30 (1 H, t,
5040 | Dalton Trans., 2007, 5031–5042
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
1227, 1119, 996, 851, 678. kmax(MeOH)/nm 290 (e/dm3 mol−1 cm−1
Kinetics
24 500), 300 (25 500), 330 (19 300).
The rates of oxidation of the pro-chelators to their respective
chelator were investigated under pseudo first-order conditions of
excess H2O2. In a typical kinetic study, 3 mL of a 30 lM solution
of the pro-chelator in 50 : 50 (v/v) methanol–NaHPO4 buffer
(20 mM, pH = 7.4) were loaded into a quartz cuvette. Upon
addition of H2O2 (300 lM–10 mM) to the mixture, spectra were
taken at time increments such that a minimum of 30 spectra were
collected throughout a kinetic run and at least 50% conversion to
the chelator was observed spectrophotometrically. The negative
slope of the linear fit of ln(Abs) at the kmax characteristic of the
Fe complexes
The iron complexes of the SBH derivatives were prepared by
refluxing 2 equivalents of the ligand with one equivalent of
Fe(NO3)3·9H2O in methanol for one hour, as described previously
for [Fe(SIH)2](NO3).18
[Fe(SBH(m-OMe))2](NO3). ESI-MS m/z 594.0 (M+). IR
(neat, mmax/cm−1): 2978, 1602, 1549, 1433, 1387, 1321, 1303, 1244,
1205, 1149, 1089, 1032, 905, 821, 803, 731. kmax(MeOH)/nm 297
(e/dm3 mol−1 cm−1 24 000), 338 (12 500). Elemental Analysis calc.
for C30H26FeN5O9: C, 54.9; H, 4.0; N, 10.7; found: C, 54.7; H, 4.1;
N, 10.6%.
prochelator vs time plot provided the observed rate constant kobs
.
Equivalent values were obtained by using the increase in kmax of
the chelator. For each pro-chelator kobs was determined at five
different H2O2 concentrations. The slope of the linear fit of kobs vs
[H2O2] provided k (M−1 s−1).
[Fe(SBH(m-OEt))2](NO3). ESI-MS m/z 622.0 (M+). IR (neat,
mmax/cm−1): 2936, 1606, 1536, 1385, 1341, 1312, 1245, 1125, 990,
910, 834, 758. kmax(MeOH)/nm 296 (e/dm3 mol−1 cm−1 31 300),
338 (15 700). Elemental Analysis calc. for C32H30FeN5O9·H2O: C,
54.7; H, 4.6; N, 10.0; found: C, 55.2; H, 4.6; N, 9.8%.
2-Deoxyribose assay
The formation of hydroxyl radicals was measured by using a 2-
deoxyribose oxidative degradation assay.41 Because many standard
buffers such as HEPES and TRIS are hydroxyl radical scavengers,
all assays were performed in 50 mM NaH2PO4 buffered to pH 7.4.
Standard reaction conditions consisted of 1 mL buffered solutions
prepared by sequential addition of the following reagents at their
final concentrations: 0–400 lM chelator or pro-chelator, 10 lM
FeCl3 or CuSO4, 15 mM 2-deoxyribose, H2O2 (200 lM for iron
assays and 100 lM for copper assays) and 2 mM ascorbic acid.
The reactions were stirred for 60 min at 37 ◦C, then quenched with
1 mL of TBA and 1 mL of trichloroacetic acid. After heating to
100 ◦C for 20 min, the solutions were cooled to room temperature
and the absorbance at 532 nm recorded. The data are reported
as A/Ao, where A is the absorbance at a 532 nm at a specific
chelator concentration, and Ao is the absorbance at 532 nm
for the background reaction containing no added chelator. All
measurements were performed in triplicate, and error bars reflect
the standard deviation from triplicate runs.
[Fe(SBH(OMe)3)2](NO3). Black needle crystals suitable for X-
ray diffraction were grown in a 6 mm glass tube by slow diffusion
of pentane into a saturated solution of [Fe(SBH(OMe)3)2](NO3)
dissolved in chlorobenzene. ESI-MS m/z 714.1 (M+). IR (neat,
mmax/cm−1): 2927, 1605, 1543, 1501, 1386, 1339, 1311, 1242, 1124,
992, 832, 755. kmax(MeOH)/nm 300 (e/dm3 mol−1 cm−1 41 100),
338 (24 700). Elemental Analysis calc. for C34H34FeN5O13·H2O: C,
51.4; H, 4.3; N, 8.8; found: C, 51.2; H, 4.65; N, 8.75%.
Ligand–metal selectivity
The interaction of ligands BSIH, BASIH, BIH and SIH with
NiSO4, Fe(NO3)3, CuSO4 and Zn(OAc)2 were studied. In a typical
reaction, the metal (10 mM) was added to a solution of ligand
(10 mM) in methanol. The solutions were allowed to equilibrate
and any precipitation due to metal–ligand complexation was
noted. UV-vis spectra were collected for the reaction mixtures that
did not precipitate or for the supernatants of mixtures in which
precipitate formed.
Determination of partition coefficients
Solutions of 50–100 lM of compound were prepared in 50 mL
of 1-octanol saturated with 50 mM aqueous phosphate buffered
saline (PBS, pH = 7.4). The initial absorbance of this stock
solution was measured to provide Ao. Preliminary experiments
were conducted to verify that these absorbance values increased
linearly with concentration. Aliquots of this solution (x = 1–
5 mL) were diluted with 1-octanol saturated with PBS to give
a total volume of 5 mL, which were combined in 50 mL tubes with
45 mL of PBS saturated with 1-octanol. The partition mixtures
were allowed to equilibrate for 5 days before the absorbance of the
1-octanol layer was measured (Aoct). Using this partition volume
ratio of 9 : 1 (45 mL PBS/5 mL 1-octanol), the partition coefficient
P was calculated according to the following equation:45
Determination of relative conditional stability by EDTA
competition
Solutions containing a 1 : 3 ratio of FeCl3 (0.15 mM) to ligand L
(0.45 mM) were prepared in 10 mL of 50 : 50 (v/v) methanol :
NaHPO4 buffer (pH = 7.4) containing 500 mM NaCl. The Fe3+
was added from a methanol stock solution to the prepared ligand
solution to avoid precipitation of iron hydroxides at pH 7.4. An
initial absorbance reading at 480 nm was measured to indicate the
total amount of [FeL2]+ present. A 10-fold excess (with respect to
Fe) of EDTA (1.5 mM) was added, and the solutions were allowed
to equilibrate for > 12 h in the dark. A final absorbance reading
at 480 nm was measured to indicate the remaining [FeL2]+ species
was still present in the presence of competing EDTA (Fe–EDTA
species have no absorbance at 480 nm).
[octanol]
(45Aoct
(A0x − 5Aoct
)
P =
= 9
[aqueous phase]
)
This equation may be rearranged in terms of x so that the
partition coefficient may be obtained from the slope of the plot of
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 5031–5042 | 5041
©

View Article Online
x versus Aoct/Ao:
21 T. Simunek, C. Boer, R. A. Bouwman, R. Vlasblom, A. M. G.
Versteilen, M. Sterba, V. Gersl, R. Hrdina, P. Ponka, J. J. de Lange,
W. J. Paulus and R. J. P. Musters, J. Mol. Cell. Cardiol., 2005, 39,
345–354.
ꢀ
ꢁ
405
Aoct
A0
x =
+ 5
P
22 H. G. Kuivila and A. G. Armour, J. Am. Chem. Soc., 1957, 79, 5659–
5662.
Note added in proof
23 L. C. Lo and C. Y. Chu, Chem. Commun., 2003, 2728–2729.
24 A. E. Albers, V. S. Okreglak and C. J. Chang, J. Am. Chem. Soc., 2006,
128, 9640–9641.
25 M. C. Y. Chang, A. Pralle, E. Y. Isacoff and C. J. Chang, J. Am. Chem.
Soc., 2004, 126, 15392–15393.
26 E. W. Miller, A. E. Albers, A. Pralle, E. Y. Isacoff and C. J. Chang,
J. Am. Chem. Soc., 2005, 127, 16652–16659.
A separate report of a boronic acid pro-chelator has recently
appeared.i,ii
(i) Y. Wei and M. Guo, Angew. Chem., Int. Ed., 2007, 46, 4722–
4725.
27 E. W. Miller, O. Tulyathan, E. Y. Isacoff and C. J. Chang, Nat. Chem.
Biol., 2007, 3, 263–267.
(ii) Y. Wei and M. Guo, Angew. Chem., Int. Ed., 2007, 46, 6948.
28 W. Q. Yang, X. Gao, and B. H. Wang, in Boronic Acids, ed. D. G. Hall,
Wiley-VCH, Weinheim, 2005, pp. 481-512.
Acknowledgements
29 D. S. Kalinowski and D. R. Richardson, Pharmacol. Rev., 2005, 57,
547–583.
This work was supported by Duke University and the Parkinson’s
Disease Foundation.
30 A. A. Aruffo, T. B. Murphy, D. K. Johnson, N. J. Rose and V.
Schomaker, Acta Crystallogr., Sec. C, 1984, C40, 1164–1169.
31 L. H. Huo, S. Gao, J. W. Liu, J. Li, H. Zhao and J. G. Zhao, Acta
Crystallogr., Sect. E, 2004, E60, m673–m675.
32 T. B. Murphy, N. J. Rose, V. Schomaker and A. Aruffo, Inorg. Chim.
Acta, 1985, 108, 183–194.
References
1 B. Halliwell and J. M. C. Gutteridge, Free Radicals in Biology and
Medicine, 3rd edn, Oxford University Press, New York, 1999.
2 B. Halliwell, J. Neurochem., 2006, 97, 1634–1658.
33 A. Kapor, B. Ribar, V. M. Leovac, G. Argay, A. Kalman and S. Y.
Chundak, J. Coord. Chem., 1996, 38, 139–144.
3 L. Zecca, M. B. H. Youdim, P. Riederer, J. R. Conner and R. R.
Crichton, Nat. Rev. Neurosci., 2004, 5, 863–873.
4 K. J. Barnham, C. L. Masters and A. I. Bush, Nat. Rev. Drug Discovery,
2004, 3, 205–214.
34 E. W. Ainscough, A. M. Brodie, A. J. Dobbs, J. D. Ranford and J. M.
Waters, Inorg. Chim. Acta, 1998, 267, 27–38.
35 S. C. Chan, L. L. Koh, P. H. Leung, J. D. Ranford and K. Y. Sim, Inorg.
Chim. Acta, 1995, 236, 101–108.
5 J. L. Pierre and M. Fontecave, BioMetals, 1999, 12, 195–199.
6 M. Valko, H. Morris and M. T. D. Cronin, Curr. Med. Chem., 2005,
12, 1161–1208.
7 D. T. Dexter, F. R. Wells, A. J. Lees, F. Agid, Y. Agid, P. Jenner and
C. D. Marsden, J. Neurochem., 1989, 52, 1830–1836.
8 E. Sofic, W. Paulus, K. Jellinger, P. Riederer and M. B. H. Youdim,
J. Neurochem., 1991, 56, 978–982.
36 A.-M. Albrecth-Gary and A. L. Crumbliss, in Metal Ions Biological
Systems, ed. A. Sigel and H. Sigel, Marcel Dekker, New York, 1973,
vol. 35, pp. 239–327.
37 J. L. Buss, F. M. Torti and S. V. Torti, Curr. Med. Chem., 2003, 10,
1021–1034.
38 L. M. W. Vitolo, G. T. Hefter, B. W. Clare and J. Webb, Inorg. Chim.
Acta, 1990, 170, 171–176.
9 J. L. Dunaief, Invest. Ophthalmol. Visual Sci., 2006, 47, 4660–4664.
10 M. B. H. Youdim, G. Stephenson and D. Ben, Shachar, Ann. N. Y. Acad.
Sci., 2004, 1012, 306–325.
11 D. R. Richardson, Ann. N. Y. Acad. Sci., 2004, 1012, 326–341.
12 D. Kaur, F. Yantiri, S. Rajagopalan, J. Kumar, J. Q. Mo, R. Boon-
plueang, V. Viswanath, R. Jacobs, L. Yang, M. F. Beal, D. DiMonte,
I. Volitaskis, L. Ellerby, R. A. Cherny, A. I. Bush and J. K. Andersen,
Neuron, 2003, 37, 899–909.
39 J. L. Buss and P. Ponka, Biochim. Biophys. Acta, 2003, 1619, 177–186.
40 D. R. Richardson, L. M. W. Vitolo, G. T. Hefter, P. M. May, B. W.
Clare, J. Webb and P. Wilairat, Inorg. Chim. Acta, 1990, 170, 165–170.
41 B. Halliwell, J. M. C. Gutteridge and O. I. Aruoma, Anal. Biochem.,
1987, 165, 215–219.
42 M. Hermes-Lima, M. S. Goncalves and R. G. Andrade, Mol. Cell.
Biochem., 2001, 228, 73–82.
43 C. Hansch, Acc. Chem. Res., 1969, 2, 232–239.
44 J. T. Edward, F. L. Chubb and J. Sangster, Can. J. Physiol. Pharmacol.,
1997, 75, 1362–1368.
13 D. Ben, Shachar, N. Kahana, V. Kampel, A. Warshawsky and M. B. H.
Youdim, Neuropharmacology, 2004, 46, 254–263.
14 Z. D. Liu and R. C. Hider, Coord. Chem. Rev., 2002, 232, 151–171.
15 A. Dedeoglu, K. Cormier, S. Payton, K. A. Tseitlin, J. N. Kremsky, L.
Lai, X. H. Li, R. D. Moir, R. E. Tanzi, A. I. Bush, N. W. Kowall, J. T.
Rogers and X. D. Huang, Exp. Gerontol., 2004, 39, 1641–1649.
16 A. Yiakouvaki, J. Savovic´, A. Al-Qenaei, J. Dowden and C. Pourzand,
J. Invest. Dermatol., 2006, 126, 2287–2295.
45 P. V. Bernhardt, P. Chin, P. C. Sharpe, J. Y. C. Wang and D. R.
Richardson, J. Biol.Inorg. Chem., 2005, 10, 761–777.
46 B. L. Rai, L. S. Dekhordi, H. Khodr, Y. Jin, Z. Liu and R. C. Hider,
J. Med. Chem., 1998, 41, 3347–3359.
47 J. T. Edward, P. Ponka and D. R. Richardson, Biometals, 1995, 8, 209–
217.
48 J. T. Edward, Biometals, 1998, 11, 203–205.
49 J. Yan, G. Springsteen, S. Deeter and B. H. Wang, Tetrahedron, 2004,
60, 11205–11209.
17 A. Gaeta and R. C. Hider, Br. J. Pharmacol., 2005, 146, 1041–1059.
18 L. K. Charkoudian, D. M. Pham and K. J. Franz, J. Am. Chem. Soc.,
2006, 128, 12424–12425.
19 M. Horackova, P. Ponka and Z. Byczko, Cardiovasc. Res., 2000, 47,
50 J. T. Edward, M. Gauthier, F. L. Chubb and P. Ponka, J. Chem. Eng.
Data, 1988, 33, 538–540.
529–536.
20 D. R. Richardson and P. Ponka, J. Lab. Clin. Med., 1998, 131, 306–
315.
51 P. Ponka, J. Borova, J. Neuwirt and O. Fuchs, FEBS Lett., 1979, 97,
317–321.
5042 | Dalton Trans., 2007, 5031–5042
This journal is
The Royal Society of Chemistry 2007
©