Glutathione-mediated activation of a disulfide containing Fe3+ complex

Article abstract of DOI:10.1016/j.ica.2019.03.006

Metal complexes containing ligands that can be cleaved using biological reductants are possible alternatives to traditional photocaged metal ions. This strategy has been demonstrated in this proof-of-concept study with a disulfide containing Fe³⁺

Full text of DOI:10.1016/j.ica.2019.03.006

Inorganica Chimica Acta 490 (2019) 139–143
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Research paper
3
+
Glutathione-mediated activation of a disulfide containing Fe
complex
⁎
Audrey G. Fikes, Kanchan Aggarwal, Emily L. Que
Department of Chemistry, The University of Texas at Austin, 105 E. 24th Street Stop A5300, Austin, TX 78712, United States
A R T I C L E I N F O
A B S T R A C T
Keywords:
Metal complexes containing ligands that can be cleaved using biological reductants are possible alternatives to
traditional photocaged metal ions. This strategy has been demonstrated in this proof-of-concept study with a
3
+
Fe
complex
Caged metal complex
Disulfide
3+
3+
disulfide containing Fe
complex, FeL1Cl. The loss of tight Fe
complexation due to disulfide reduction is
observed through a decrease in a ligand to metal charge transfer band in its UV–vis spectrum. The mechanism of
this reaction was investigated using mass spectrometry and stopped flow kinetics. A coumarin fluorescence assay
was used to determine the ability of this complex to catalyze Fenton chemistry and produce hydroxyl radical.
Hydroxyl radical production by FeL1Cl was low but in the presence of the reductant glutathione, hydroxyl
radical production is increased, suggesting that reduction of the disulfide bond by glutathione uncages the
reactivity of the Fe center in this complex.
Glutathione
Fenton chemistry
ROS production
1
. Introduction
Caged metal ions are useful tools for studying cellular metal ion
ligand backbone that can be cleaved via methods other than irradiation.
Disulfide bonds are an attractive option and have been used extensively
as redox controlled triggers for intracellular cargo release [19,20].
Disulfides are easily reduced to thiols, especially in a cellular environ-
ment due largely to the high concentrations of glutathione (GSH) which
is present at mM levels in the cell [20]. In concurrence with negligible
levels of GSH outside the cell, it is easy to envision a disulfide-con-
taining caged metal complex that will be selectively reduced inside a
cell and increase the lability of the metal ion due to the subsequent
decrease in binding affinity through the chelate effect.
homeostasis and have been used to investigate a range of metals in-
2
+
2+
2+/3+
2+
cluding Ca , Zn , Fe
, and Cu [1–7]. Complexes designed for
controlled intracellular release of metal ions have historically taken
advantage of photolabile ligands that attenuate metal ion binding af-
finity upon irradiation [4,5]. Two common strategies include photo-
induced decrease in ligand donor atom strength [1,3,6,8] and photo-
induced ligand cleavage resulting in a decrease in the chelate effect
[
9–11]. Photocaged metals have been used successfully to probe the
Several disulfide containing metal complexes have been made over
the past 40 years [21]. Most transition metal containing disulfide
complexes have been characterized primarily using X-ray crystal-
lography [22–25] and have not been investigated for their reactivity
with biological reductants. Previously reported complexes have also
been synthesized as potential catalysts [26–29] while others have been
made to investigate the interactions between thiols, disulfides, and
biologically relevant metal ions due to their importance in cellular
environments [30,31]. However, the use of disulfide metal complexes
as caged metal ions for controlling metal ion reactivity and lability has
not been explored. To this end, we have investigated a disulfide con-
2
+
role of Ca , including its role in neurotransmitter release [12], vas-
cular dilation and constriction by astrocytes [13], and glutamate re-
lease in astrocytes [14]. However, there are limitations to relying on
light as the needed stimulus for uncaging metal ions, including low
quantum yields [2,3,7], slow rate of release [2], small changes in metal
ion binding affinity [3,15], and often the need to use UV light, which
3
+
can damage and kill cells. In regard to Fe , a few photocaged com-
3
+
plexes have been reported. Naturally occurring photoactive Fe
siderophore complexes have been investigated [16,17] and have in-
3
+
spired the creation of a siderophore-based trinuclear Fe
complex,
2
+
3+
which releases Fe upon irradiation [18]. The photocage FerriCast has
taining Fe complex, FeL1Cl, for its reactivity and its possible use for
3
+
3+
also been used to successfully modulate ligand binding affinity for Fe
after UV irradiation [6].
selective de-caging of Fe
GSH (Scheme 1).
in the presence of high concentrations of
An alternative strategy is the inclusion of a labile bond within the
Abbreviations: GSH, glutathione; LMCT, ligand to metal charge transfer; DMSO, dimethyl sulfoxide; MeOH, methanol; DTT, dithiothreitol; HRMS, high resolution
mass spectrometry; 7-HC, 7-hydroxycoumarin
⁎
Corresponding author.
E-mail address: emilyque@cm.utexas.edu (E.L. Que).
https://doi.org/10.1016/j.ica.2019.03.006
Received 23 January 2019; Received in revised form 4 March 2019; Accepted 4 March 2019
Available online 05 March 2019
0020-1693/ © 2019 Elsevier B.V. All rights reserved.

A.G. Fikes, et al.
Inorganica Chimica Acta 490 (2019) 139–143
presence of GSH, suggesting that Fe3+ is no longer bound to the parent
complex. We note that in the absence of GSH, FeL1Cl is readily de-
tected under the same HRMS conditions. The free ligand, L1, is detected
both before (Figs. S13 and S14) and after addition of GSH to FeL1Cl
(
Figs. S15 and S16). Detection of L1 in the control solution containing
3+
only FeL1Cl is likely due to loss of Fe during the ionization process.
After addition of GSH, since no intact FeL1Cl is detected, detection of
L1 is assumed to be due to continued disulfide exchange leading to the
3
+
reformation of L1 without any Fe
bound. This is supported by the
observation that there is no reappearance of the LMCT and the corre-
3
+
sponding deep purple color, indicating that Fe
bound to the ligand.
is no longer tightly
The kinetics of the disulfide cleavage were also investigated. The
reaction was predicted to be first order in both reactants and second
order overall, as shown in Eq. (1). Under pseudo first order conditions
with excess GSH, the rate law was expected to become Eq. (2).
Scheme 1. GSH activation of a disulfide containing Fe3+ complex.
2
. Results
r = k [GSH][FeL1Cl]
(1)
r = kobs [FeL1Cl]
(2)
The disulfide containing complex FeL1Cl was synthesized as re-
ported in literature [24]. The ligand was synthesized via condensation
of 2,2′-diaminodiphenyl disulfide with salicylaldehyde. The absorbance
spectrum of the complex is characterized by an intense phenolato to
The data shows, however, that in the presence of excess GSH, the
reaction is second order in FeL1Cl (Fig. 2B). This can be explained
through the proposed mechanism in Eqs. (3) and (4) and Fig. 2D. Initial
disulfide cleavage by GSH results in the creation of a free thiol, LSH,
from the complex. This free thiol is available to reduce intact complexes
of FeL1Cl. Because of this, the observed rate law is second order in
FeL1Cl and first order in GSH (Eq. (5)). When carried out with excess
GSH, the rate law is simplified to Eq. (6).
3
+
Fe
ligand to metal charge transfer (LMCT) band around 546 nm
−
1
1
−1
−1
(
ɛ = 3900 M cm
)
in
DMSO
[27]
and
562 nm
−
(
ɛ = 2860 M cm ) in CH
2
Cl
2
[32]. The magnetic moment was de-
3
+
termined to be 6.2 µ
B
indicating a high spin Fe complex. The ligand
has an irreversible reduction at −1.91 V corresponding to the imine
while the disulfide redox couple has an E1/2 = −1.26 V (Fig. S1A)
(3)
3
+
FeL1Cl + GSH
FeLSSGCl + LSH
which is typical of aromatic disulfides [33]. The complex has an Fe
/
2
+
Fe redox couple at −1.48 V. The reduction potential for the imine is
shifted to −0.721 V while the E1/2 for the disulfide is shifted to
FeL1Cl + LSH
L1 + LSH + FeLSCl
(4)
(5)
(6)
2
r
k [GSH][FeL1Cl]
−
0.575 V (Fig. S1B). The positive shift in the reduction potential of
both the imine and disulfide are indicative of coordination of the ni-
2
r
kobs [FeL1Cl]
3
+
trogen and sulfur, respectively, to the Fe . The resulting polarization
of the bonds makes both moieties easier to reduce as indicated by the
change in reduction potentials.
As expected, the reaction between FeL1Cl and GSH was first order
in GSH over a wide range of GSH concentrations as shown by the plot of
3
+
k
k
vs the initial GSH concentration (Fig. S5C). Unexpectedly, a plot of
To determine the feasibility of GSH mediated Fe release from the
initial FeL1Cl complex via disulfide reduction, absorbance of the LMCT
was monitored upon addition of GSH. As expected, a decrease in the
obs
obs vs initial FeL1Cl concentration shows that kobs decreases with in-
creasing FeL1Cl concentration in a linear fashion as opposed to in-
3
+
creasing in a quadratic manner (Fig. 2C). This is most likely due to
LMCT is seen with GSH addition, indicating decomplexation of Fe
3
3
+
+
interference in absorbance caused by complexes of weakly bound Fe
with the resulting products from the reaction. This includes an Fe
(
Fig. 1A, Fig. S2). A similar decrease in the LMCT is also seen upon
addition of other reductants, including DTT and sodium dithionite (Fig.
S3). These observed changes differ from those of the reaction of ligand
L1 with GSH (Fig. 1B), in which changes are only observed below
complex formed with a ligand-GSH adduct, FeLSSGCl (Fig. 2D), which
is observed by HRMS after the reaction of FeL1Cl with GSH (Fig. S19)
3
+
and another Fe
complex that forms with two intact L1 ligands, ob-
4
50 nm. This suggests that the decrease in LMCT is a result of a change
served in small amounts via HRMS both before and after reduction with
in the coordination of the metal center, likely due to loss of phenolate-
3
+
GSH (Table S1, Figs. S17 and S18). Mild interference in absorbance
Fe coordination bonds. The products of the reduction were evaluated
using high resolution mass spectrometry (HRMS) (Table S1). Two
species containing GSH adducts were detected indicating that loss of
3
+
between 500 and 600 nm can also be caused by complexation of Fe
with GSH itself [34]. When we modeled this proposed mechanism
(KinTek Explorer software [35]) including absorbance interference and
considering Eq. (3) as faster than Eq. (4), the same trend is seen in that
the kobs decreases linearly with increasing FeL1Cl concentration (Fig.
S6B).
3
+
Fe
coordination is due to cleavage of the ligand disulfide by GSH
(
Figs. S11 and S12). Importantly, no FeL1Cl is detected by HRMS in the
In order to demonstrate that the intact FeL1Cl complex fully se-
3
+
questers Fe , the complex’s ability to catalyze Fenton chemistry was
investigated. Hydroxyl radicals react with coumarin to produce 7-hy-
droxycoumarin (7-HC), which fluoresces at 400 nm [36–38]. This re-
3
+
action with coumarin takes place readily in the presence of labile Fe
and H
2
O . Comparison of the relative integrated fluorescence of 7-HC
2
formed under different reaction conditions can demonstrate the ability
of FeL1Cl to produce ROS only after reaction with GSH. The ability of
FeCl and FeL1Cl to catalyze hydroxyl radical formation in the pre-
3
Fig. 1. UV–vis spectra of (A) 100 µM FeL1Cl in DMSO after addition of 1 equiv.
sence and absence of GSH was compared (Fig. 3). FeCl
high levels of 7-HC fluorescence, which are not significantly affected by
the presence of equimolar GSH or L1, as mixture of FeCl and L1 at
room temperature does not lead to the formation of FeL1Cl. When
alone leads to
3
of GSH and (B) 100 μM L1 in DMSO after addition of 10 equiv. of GSH. A
decrease in the LMCT band at 546 nm for FeL1Cl, which is not seen for L1,
3
indicates loss tightly bound Fe3
+
.
140

A.G. Fikes, et al.
Inorganica Chimica Acta 490 (2019) 139–143
Fig. 2. (A) The change in absorbance at 562 nm over time for the reaction of FeL1Cl (80–105 µM) with 1 mM GSH in MeOH. (B) A plot of 1/A562nm vs time for the
initial 15 s of the reaction between constant GSH and varying FeL1Cl concentrations; the linear relationship demonstrates that the reaction is second order in FeL1Cl.
The slopes of the lines give the kobs values. (C) A decrease in kobs values with increasing [FeL1Cl] is observed, indicating interference at the observed wavelength
(
562 nm) due to products from the reaction. As the concentration of FeL1Cl increases, the concentration of interfering products increases leading to the unexpected
relationship between kobs and [FeL1Cl]. (D) Proposed mechanism that explains the decreasing trend in kobs and the second order dependence on FeL1Cl. The first step
involves the reaction of FeL1Cl with GSH leading to the formation of FeLSSGCl, which is proposed to be one of the possible species absorbing weakly at 562 nm and
causing interference. In step two, the free thiol LSH, from step one, reacts rapidly with remaining FeL1Cl leading to an overall reaction mechanism that appears to be
second order in FeL1Cl.
disulfide bond by GSH uncages the Fe3+ reactivity and restores its
ability to catalyze formation of ROS in the presence of H
2
O . These
2
results demonstrate this proof of concept study that disulfide containing
ligands can be used to make caged metal complexes that effectively
mask a metal ion’s reactivity until exposure to a reducing environment
uncages the metal, thus restoring reactivity.
3
. Conclusions
In conclusion, a disulfide-containing Fe3+ complex has been studied
3
+
as an alternative to photocaged metals for controlling Fe
reactivity
through ligand centered reduction. This proof-of-concept study de-
monstrates the feasibility of disulfide dependent caged metal complexes
that circumvents common problems associated with photolabile li-
gands. Current efforts are geared towards the development of aqueous-
compatible complexes for biological studies of the interplay of metal
ion homeostasis and oxidative stress.
4
. Experimental
Fig. 3. Normalized fluorescence of 7-hydroxycoumarin produced from the re-
action of coumarin with hydroxyl radical in the presence of H . No difference
4.1. Materials
2 2
O
in fluorescence signal is observed between the control containing only cou-
All chemicals and solvents were purchased from Alfa Aesar and
3
+
marin and the sample with FeL1Cl, indicating that the Fe is effectively caged.
1
13
Fisher Scientific and used as received. H and C NMR spectrum were
acquired on a 400 MHz Agilent MR spectrometer. NMR samples were
In the presence of GSH, the fluorescence signal increases due to the uncaging of
3
+
the Fe
which is then free to catalyze hydroxyl radical production.
prepared in DMSO‑d
6
and chemical shifts are reported in ppm.
Fluorescence levels observed after addition of GSH to FeL1Cl are similar to
those produced by FeCl . Data represents mean values (n = 3) ± the standard
deviation ( p < 0.0001).
Electrospray Ionization Mass Spectrometry (ESI-MS) was performed by
the Mass Spectrometry Facility of the Department of Chemistry at UT
Austin. UV–vis spectroscopic studies were preformed using an Agilent
Cary 60 UV–vis Spectrophotometer. Kinetic measurements were made
using a Hi-Tech Scientific SFA-20 Rapid Kinetics Accessory in tandem
with an Agilent 8453 UV–vis Spectrophotometer. Cyclic voltammetry
was carried out using a CHI 660D electrochemical workstation from the
UT Austin Center for Electrochemistry. Fluorescence measurements
were made using an Agilent Cary Eclipse Fluorescence
Spectrophotometer.
3
*
**
FeL1Cl, H
2
O , and coumarin were combined, similar levels of fluores-
2
cence were observed as in the control reaction between coumarin and
3
+
H
2
O in the absence of any Fe source. These results indicate that the
2
complex is unable to catalyze hydroxyl radical formation due to its fully
occupied coordination sphere. Upon incubation of FeL1Cl with GSH,
the fluorescence levels increased, suggesting that cleavage of the
141

A.G. Fikes, et al.
Inorganica Chimica Acta 490 (2019) 139–143
4.2. Synthesis of FeL1Cl [24,32]
4.7. Kinetics
2
,2′-Diaminodiphenyl disulfide (500.4 mg, 2.015 mmol) was dis-
Kinetic experiments were carried out in MeOH at 26 °C with an
solved in 12 mL MeOH. Salicylaldehyde (0.211 mL, 2.015 mmol) was
dissolved in 1.5 mL MeOH and added to the first solution. Lithium
excess amount of GSH. FeL1Cl has a strong LMCT at 562 nm, which
3
+
disappears upon disulfide reduction and concomitant Fe
release.
(
14.0 mg, 2.017 mmol) was dissolved in 5 mL MeOH and added to the
Thus the change in absorbance at 562 nm was monitored over time.
Second order dependence on FeL1Cl was determined by holding [GSH]
concentration constant (1.0 mM) and varying initial [FeL1Cl] con-
centration between 80 and 105 μM. The pseudo second-order kobs va-
lues were determined by plotting 1/A and obtaining the slopes of the
resulting linear plots. A graph of kobs values vs initial [FeL1Cl] gives a
line with a negative slope indicating formation of a product that has
weak interfering absorbance around 562 nm (see Kinetics Modeling
below.) First order dependence on GSH was determined by keeping
ligand solution and stirred at room temperature. A 3 mL MeOH solution
of FeCl (163.4 mg, 1.007 mmol) was added dropwise to the solution.
3
The solution was refluxed for 30 min and allowed to cool to room
temperature. A dark solid precipitated out of solution. Filtering and
washing with hot MeOH gave pure product with a yield of 61%.
Product formation was confirmed through UV–vis spectroscopy and
−
1
−1
+
HRMS. ε562nm in DCM: 2860 M cm
[32]. HR ESI-MS (ESI ,
+
MeOH): calculated for [M-Cl] 510.0154, found 510.0155.
[
FeL1Cl] constant (100 μM) and varying the initial [GSH] from 50 μM
to 1.4 mM. The kobs values were obtained by plotting 1/A. A graph of
obs vs initial [GSH] gives a straight line indicating first order depen-
4
.3. Synthesis of L1 [28]
k
dence on GSH. The slope of the line gives a third order rate constant of
A modified literature procedure was used to synthesize L1 [28].
,2′-diaminodiphenyl disulfide (439.6 mg, 1.770 mmol) and salicy-
−
2 −1
2
.0 M
s
.
2
laldehyde (0.463 mL, 4.421 mmol) were combined in 60 mL EtOH. A
yellow solid immediately began precipitating. The mixture was stirred
for 15 min at room temperature. The mixture was refrigerated for three
hours and then filtered and washed with cold EtOH to give a yield of
4
.8. Kinetics modeling with KinTek explorer
The simulations for the possible mechanism for FeL1Cl reduction by
1
GSH were done using KinTek Explorer software [35]. The model used
was:
9
0%. H NMR (400 MHz, DMSO‑d
6
) δ 12.56 (s, 1H), 9.02 (s, 1H), 7.71
(
dd, J = 7.7, 1.8 Hz, 1H), 7.56 (dd, J = 7.9, 1.5 Hz, 1H), 7.50 (dd,
J = 7.9, 1.4 Hz, 1H), 7.46 (ddd, J = 8.1, 7.3, 1.7 Hz, 1H), 7.36 (td,
J = 7.6, 1.5 Hz, 1H), 7.29 (td, J = 7.6, 1.4 Hz, 1H), 7.04 – 6.97 (m, 2H).
(7)
(8)
FeL1Cl + GSH = FeLSSGCl + LSH
1
3
LSH + FeL1Cl = L1
C NMR (DMSO‑d
6
, 400 MHz): δ 164.1, 160.6, 146.5, 134.4, 133.2,
+
1
30.8, 128.6, 128.4, 126.6, 119.9, 119.0, 117.2. LR ESI-MS (ESI
,
The forward rate for each equation were set so that k > k and the
1
2
+
MeOH): calculated for [M+H] 457.1, found 457.1.
backward rates were set to small values as both reactions are irrever-
sible. The output was set to:
4
.4. UV–vis spectroscopy
a FeL1Cl + b FeLSSGCl
where a and b are the extinction coefficients for the absorbance of
FeL1Cl and FeLSSGCl. They were set in such a way that a was higher
than b. The concentrations of GSH and FeL1Cl were kept similar to the
experimental conditions. The data generated for total absorbance versus
time under the set conditions was fitted to the following exponential
equation:
To study the reaction of FeL1Cl with different reductants, the
change in the UV–vis spectrum of 100 μM FeL1Cl was monitored while
GSH, DTT, and sodium dithionite were titrated into the solution, re-
spectively. The studies were carried out in DMSO and MeOH. Data was
collected with an Agilent Technologies Cary 60 UV–vis spectrometer.
Stock solutions of FeL1Cl (1 mM) were prepared in DMSO or MeOH.
GSH stock solution (100 mM) was made using 1:1 water:DMSO. Stock
solutions of DTT (100 mM) were made in DMSO or MeOH. Stock so-
(
kt)
Y= Ae
+ B
where A is the amplitude, k is the rate constant, and B is a constant. The
rate was plotted with respect to concentration of variable species. It
should be noted that simulations were run to confirm the proposed
mechanism based on the experimental data and not used to calculate
lutions of Na
2
S
2
O (100 mM) were made in water. Control studies were
4
done with equivalent amounts of water containing no reductant to
demonstrate that changes in the UV–vis spectrum are due to the reac-
tion of FeL1Cl with the reductant and not hydrolysis of FeL1Cl due to
the presence of water.
the simulated kobs values, for several reasons. The absolute values of k
1
and k are unknown as are the extinction coefficients of the interfering
2
species. In addition to this, the kinetics model in the simulations are
fixed throughout the time period, while under experimental conditions,
it might change with time as the relative concentration of each species
varies. For this reason, simulated kobs values do not necessarily match
those obtained experimentally.
4
.5. Cyclic voltammetry
Electrochemical data was collected with a CHI 660D electro-
chemical workstation. Measurements of a 2 mM solution of FeL1Cl and
L1 were recorded at 100 mV/s in a glovebox. A three-electrode cell was
+
used, with a platinum electrode as the working electrode, an Ag/Ag
4
.9. Coumarin hydroxylation fluorescence assay
non-aqueous electrode as a reference electrode (a 10 mM solution of
+
AgNO
3
in DMF was used as a Ag source), and a platinum wire aux-
In order to determine that ability of FeL1Cl to catalyze hydrogen
iliary counter electrode. Bn
4
NBF (0.1 M) was used as the electrolyte
4
peroxide induced ROS production in the presence and absence of GSH,
a coumarin hydroxylation assay was used [38]. Coumarin reacts with
hydroxyl radical to yield 7-hydroxycoumarin, which fluoresces around
400 nm in MeOH. Comparison of the integrated fluorescence of reaction
and the spectra were calibrated versus ferrocene.
4
.6. Mass spectrometry
3
+
solutions containing coumarin, H
2
O
2
, and Fe
indicates the relative
High resolution ESI-FIA-MS and ESI-LC-MS data was collected on an
amount of hydroxyl radicals produced. Reactions containing FeL1Cl in
Agilent Q-TOF LC/MS. Reactions were carried out in methanol with an
initial FeL1Cl concentration of 100 µM and equimolar amounts of GSH.
Samples were analyzed within one hour of GSH addition. Control
samples contained only 100 µM FeL1Cl and no GSH.
the absence and presence of GSH were compared to reactions con-
taining FeCl
and FeCl
monstrate the necessity of the intact FeL1Cl complex to mitigate
3
with and without GSH. Control reactions containing L1
3
(both with and without GSH) were also carried out to de-
142

A.G. Fikes, et al.
Inorganica Chimica Acta 490 (2019) 139–143
hydroxy radical production. Reactions were carried out in degassed
MeOH in an anaerobic chamber to prevent quenching of the hydroxyl
radical by oxygen. All stock solutions were made in an anaerobic
chamber using degassed solvents. Stock solutions of coumarin (5 mM),
[13] S.J. Mulligan, B.A. MacVicar, Calcium transients in astrocyte endfeet cause cere-
brovascular constrictions, Nature 431 (2004) 195.
[
[
14] V. Parpura, P.G. Haydon, Physiological astrocytic calcium levels stimulate gluta-
mate release to modulate adjacent neurons, Proc. Natl. Acad. Sci. U.S.A. 97 (15)
(2000) 8629–8634.
15] C. Gwizdala, C.V. Singh, T.R. Friss, J.C. MacDonald, S.C. Burdette, Quantifying
factors that influence metal ion release in photocaged complexes using ZinCast
derivatives, Dalton Trans. 41 (26) (2012) 8162–8174.
FeCl
3
(5 mM), FeL1Cl (1 mM), and H
2
O (50 mM) were made MeOH. A
2
1
00 mM solution of GSH was made in water. This solution was used to
make a 5 mM stock solution of GSH in MeOH. A solution of L1
12.584 mM) was made in EtOAc. This solution was used to make the
final stock solution of L1 (135.14 μM) in MeOH. The final concentra-
tions of coumarin and H in all the samples were 1 mM each. The
final concentrations of either FeCl or FeL1Cl was 100 μM. The final
[16] K. Barbeau, E.L. Rue, K.W. Bruland, A. Butler, Photochemical cycling of iron in the
surface ocean mediated by microbial iron(iii)-binding ligands, Nature 413 (6854)
(
(
2001) 409–413.
[
17] F.C. Küpper, C.J. Carrano, J.-U. Kuhn, A. Butler, Photoreactivity of iron
O
2 2
(
III)–aerobactin: photoproduct structure and iron(III) coordination, Inorg. Chem. 45
(15) (2006) 6028–6033.
3
[
18] H. Sayre, K. Milos, M.J. Goldcamp, C.A. Schroll, J.A. Krause, M.J. Baldwin, Mixed-
donor, α-hydroxy acid-containing chelates for binding and light-triggered release of
iron, Inorg. Chem. 49 (10) (2010) 4433–4439.
concentration of GSH, when present, was 100 μM. The final con-
centration of L1, when present, was 100 μM. Fluorescence measure-
ments of the reaction solutions were taken after 16 h of incubation. The
excitation wavelength was 344 nm (slit width 20.) Total fluorescence
was measured by integrating the fluorescence spectra from 375 nm to
[19] L. Brülisauer, M.A. Gauthier, J.-C. Leroux, Disulfide-containing parenteral delivery
systems and their redox-biological fate, J. Controlled Release 195 (Supplement C)
(
2014) 147–154.
[
20] M.H. Lee, Z. Yang, C.W. Lim, Y.H. Lee, S. Dongbang, C. Kang, J.S. Kim, Disulfide-
cleavage-triggered chemosensors and their biological applications, Chem. Rev. 113
6
40 nm. All samples were performed in triplicate.
(
7) (2013) 5071–5109.
[
[
21] S. Moosun, S.J. Laulloo, M.G. Bhowon, Metal complexes of diaryl and 2,2′-dipyridyl
disulfides, J. Sulfur Chem. 33 (6) (2012) 661–691.
Acknowledgements
22] P.E. Riley, K. Seff, Crystal and molecular structure of chloro(bis{2-[(2-pyr-
idylmethyl)amino]ethyl} disulfide)nickel(II) perchlorate, Inorg. Chem. 11 (12)
This work was funded by a grant from the Welch Foundation (F-
(
1972) 2993–2999.
1
883) and start-up funds from the University of Texas at Austin. The
[
23] M.M. Kadooka, L.G. Warner, K. Seff, Crystal and molecular structure of dichloro[bis
(2-pyridyl) disulfide]cobalt(II), Inorg. Chem. 15 (4) (1976) 812–816.
authors thank Prof. Aimin Liu and Christopher I. Davis from the
University of Texas at San Antonio for use of their Hi-Tech Scientific
SFA-20 Rapid Kinetics Accessory and Agilent 8453 UV-vis
Spectrophotometer for stopped flow kinetic studies. We also acknowl-
edge the use of a Bruker AVIII HD500 with Prodgiy liquid nitrogen
cryoprobe supported by NIH grant 1 S10OD021508.
[24] J.A. Bertrand, J.L. Breece, Preparation and structure of chloro(bis{salicylidenimi-
nephenyl}disulfido)iron(III), Inorg. Chim. Acta 8 (1974) 267–272.
[
25] A. Elmali, Y. Elerman, Crystal structure of chloro[bis(5-bromo-salicylidenimine-
phenyl)disulfide]-iron(III) complex, Anal. Sci. 18 (12) (2002) 1399–1400.
[26] M.G. Bhowon, S. Jhaumeer-Laulloo, N. Soukhee, A. Allibacus, V. Shiboo, Synthesis,
catalytic and antibacterial activity of 2-aminophenyldisulphide, J. Coord. Chem. 60
(
12) (2007) 1335–1343.
[
27] N. Roy, S. Sproules, T. Weyhermüller, K. Wieghardt, Trivalent iron and ruthenium
complexes with a redox noninnocent (2-mercaptophenylimino)-methyl-4,6-di-tert-
butylphenolate(2-) ligand, Inorg. Chem. 48 (8) (2009) 3783–3791.
Appendix A. Supplementary data
[
28] A. Donzelli, I. Metushi, P.G. Potvin, Titanium(IV) complexes of disulfide-linked
schiff bases, Inorg. Chem. 51 (9) (2012) 5138–5145.
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.ica.2019.03.006.
[29] G. Prakash, R. Ramachandran, M. Nirmala, P. Viswanathamurthi, W. Linert,
Catalytic role of ruthenium(III) complexes bearing cystamine-based disulfide li-
gands on the N-alkylation of o-substituted anilines, Monatsh. Chem. 145 (12)
References
(
2014) 1903–1912.
[
30] R. Uma, M. Palaniandavar, Copper(II)-disulphide interaction in copper complexes
containing salicylaldimine and pyridylaldimine ligands: synthesis, spectra and
redox behaviour, Trans. Met. Chem. 18 (6) (1993) 629–634.
[
[
1] R.Y. Tsien, R.S. Zucker, Control of cytoplasmic calcium with photolabile tetra-
carboxylate 2-nitrobenzhydrol chelators, Biophys. J. 50 (5) (1986) 843–853.
2] S.R. Adams, J.P.Y. Kao, G. Grynkiewicz, A. Minta, R.Y. Tsien, Biologically useful
[
31] H.F. Hsu, C.L. Su, N.O. Gopal, C.C. Wu, W.C. Chu, Y.F. Tsai, Y.H. Chang, Y.H. Liu,
T.S. Kuo, S.C. Ke, Redox chemistry in the reaction of oxovanadium(V) with thiolate-
containing ligands: the isolation and characterization of non-oxo vanadium(IV)
complexes containing disulfide and thioether groups, Eur. J. Inorg. Chem. 2006 (6)
(2006) 1161–1167.
2
+
chelators that release Ca upon illumination, J. Am. Chem. Soc. 110 (10) (1988)
3
212–3220.
[
[
[
[
[
[
3] C. Gwizdala, D.P. Kennedy, S.C. Burdette, ZinCast-1: a photochemically active
2+
chelator for Zn , Chem. Commun. 45 (2009) 6967–6969.
4] K.L. Ciesienski, K.J. Franz, Keys for unlocking photolabile metal-containing cages,
Angew. Chem. 50 (4) (2011) 814–824.
[
32] J.W. Pyrz, X. Pan, D. Britton, L. Que, Models for catechol dioxygenases. Structure of
bromobis[2-(2'-hydroxyphenyl)benzothiazolato]iron(III) derived from the bro-
moiron(III) complex of 2,2'-bis((salicylideneamino)phenyl) disulfide, Inorg. Chem.
30 (18) (1991) 3461–3464.
2
+
5] C. Gwizdala, S.C. Burdette, Following the Ca roadmap to photocaged complexes
2
+
for Zn and beyond, Curr. Opin. Chem. Biol. 17 (2) (2013) 137–142.
[
[
[
33] T.M. Chang, E. Tomat, Disulfide/thiol switches in thiosemicarbazone ligands for
redox-directed iron chelation, Dalton Trans. 42 (22) (2013) 7846–7849.
34] M.Y. Hamed, J. Silver, M.T. Wilsonx, Studies of the reactions of ferric iron with
glutathione and some related thiols, Inorg. Chim. Acta 78 (1983) 1–11.
35] K.A. Johnson, Z.B. Simpson, T. Blom, Global kinetic explorer: a new computer
program for dynamic simulation and fitting of kinetic data, Anal. Biochem. 387 (1)
6] D.P. Kennedy, C.D. Incarvito, S.C. Burdette, FerriCast: a macrocyclic photocage for
3
+
Fe , Inorg. Chem. 49 (3) (2010) 916–923.
7] D.P. Kennedy, D.C. Brown, S.C. Burdette, Probing nitrobenzhydrol uncaging me-
chanisms using FerriCast, Org. Lett. 12 (20) (2010) 4486–4489.
8] J. Cui, R.A. Gropeanu, D.R. Stevens, J. Rettig, A.D. Campo, New photolabile
2
+
BAPTA-based Ca cages with improved photorelease, J. Am. Chem. Soc. 134 (18)
(
2009) 20–29.
(
2012) 7733–7740.
[
36] G. Louit, S. Foley, J. Cabillic, H. Coffigny, F. Taran, A. Valleix, J.P. Renault, S. Pin,
The reaction of coumarin with the OH radical revisited: hydroxylation product
analysis determined by fluorescence and chromatography, Radiat. Phys. Chem. 72
(2) (2005) 119–124.
[
9] J.H. Kaplan, G.C. Ellis-Davies, Photolabile chelators for the rapid photorelease of
divalent cations, Proc. Natl. Acad. Sci. U.S.A. 85 (17) (1988) 6571–6575.
[
[
[
10] H.M.D. Bandara, D.P. Kennedy, E. Akin, C.D. Incarvito, S.C. Burdette, Photoinduced
2+
release of Zn with ZinCleav-1: a nitrobenzyl-based caged complex, Inorg. Chem.
[
[
37] Z.-R. Lin, L. Zhao, Y.-H. Dong, Quantitative characterization of hydroxyl radical
generation in a goethite-catalyzed Fenton-like reaction, Chemosphere 141 (2015)
4
8 (17) (2009) 8445–8455.
11] K.L. Ciesienski, K.L. Haas, M.G. Dickens, Y.T. Tesema, K.J. Franz, A photolabile
ligand for light-activated release of caged copper, J. Am. Chem. Soc. 130 (37)
7
–12.
38] C.B.C. Rutely, F. Jean-M, Z.T. Walter, D.-B. Xochitl, S. Mika, Towards reliable
quantification of hydroxyl radicals in the Fenton reaction using chemical probes,
RSC Adv. 8 (10) (2018) 5321–5330.
(
2008) 12246–12247.
12] K. Wadel, E. Neher, T. Sakaba, The coupling between synaptic vesicles and Ca2+
channels determines fast neurotransmitter release, Neuron 53 (4) (2007) 563–575.
143