1. Art and Diseño

Journal of Inorganic Biochemistry 70 (1998) 91±98
Study of the zinc-binding properties of glutathione by di€erential
pulse polarography and multivariate curve resolution
M.S. Dõaz-Cruz, J. Mendieta, A. Monjonell, R. Tauler, M. Esteban
*
Departament de Quõmica Analõtica, Facultat de Quõmica, Universitat de Barcelona, Avda. Diagonal 647, 08028-Barcelona, Spain
Received 14 July 1997; received in revised form 14 January 1998; accepted 23 January 1998
Abstract
The complexation between Zn2‡ and glutathione (GSH), in borate bu€er, is studied by Di€erential Pulse Polarography (DPP).
Data obtained at pH 8.5 from a titration of Zn2‡ by GSH, i.e. at di€erent GSH-to-Zn2‡ concentration ratios have been analyzed by a
recently proposed multivariate curve resolution method. DPP signals obtained within a wide pH range, but at ®xed 1.25:1 GSH-toZn2‡ concentration ratios, have been resolved individually at each pH value. Results from both sets of experiments can be explained
with a complexation model involving the formation of a predominant complex, with a stoichiometry of 2:2 GSH : Zn2‡ .This result
is in contrast with those for the Cd2‡ ±GSH systems, where two complexes, 2:1 and 2:2 GSH : Cd2‡ , were detected. A scheme for the
global electrode process is given. Ó 1998 Elsevier Science Inc. All rights reserved.
Keywords: Zinc; Glutathione; Polarography; Voltammetry; Factor analysis; Peak ®tting; Curve resolution
1. Introduction
The binding of metal ions by proteins and peptides is
of fundamental interest due to the importance of metal
ions in biological systems. Metals may be part of the active sites of enzymes, stabilise the macromolecular structure of proteins and a€ect enzymes or membranes to
control cell metabolism. Due to their high anity to sulphur, the interaction of heavy metals with sulfhydryl
containing amino acids and proteins plays a major role
in their environmental and biochemical behaviour [1±6].
For instance, data on at least 34 enzymes suggest that
Zn(II) may be involved as a necessary cofactor. Several
metabolically important reactions are catalysed by these
enzymes, including hydrolysis, hydration, oxidation±reduction and group-transfer reactions.
Voltammetric techniques have been widely used for
the study of the interactions between metal ions and a
diversity of ligands [7], including either simple monomeric or macromolecular ligands [8±11]. Di€erential
Abbreviations: ALS, alternating least squares; DPP, di€erential pulse
polarography; EFA, evolving factor analysis; FA, factor analysis;
GSH, glutathione (c-L -glutamyl-L -cysteinylglycine); MCR, multivariate curve resolution; NMR, nuclear magnetic resonance; SMDE,
static mercury drop electrode; SVD, singular value decomposition
*
Corresponding author. Tel.: +34 93 4021277; fax: +34 93 4021233;
e-mail: [email protected].
0162-0134/98/$19.00 Ó 1998 Elsevier Science Inc. All rights reserved.
PII: S 0 1 6 2 - 0 1 3 4 ( 9 8 ) 1 0 0 0 3 - X
pulse polarography (DPP) is useful due to its very low
detection limits, being complementary to spectrophotometry, potentiometry and nuclear magnetic resonance
(NMR). But, in general, the DPP behaviour of such systems containing biologically and/or environmentally relevant ligands is rather involved, with several signals that
are changing continuously in their position, morphology
or relative peak size, when the pH or the metal-to-ligand
ratio changes.
Multivariate Curve Resolution (MCR) has been proposed as a tool to get as much information as possible
from voltammetric data sets that a priori are very dicult, if not impossible, to interpret [12]. This procedure
is based on the use of a family of computational and statistical techniques based on Factor Analysis (FA) and
concerned with the resolution of the sources of variation
in an experimental data set.
The reliability of the proposed method has been recently checked through the DPP study of the complexation of Cd2‡ by glutathione (GSH) [13]. GSH is the
major low-molecular-weight thiol compound of the living plants or animal cells, and it is involved in many aspects of metabolism [14,15], including removal of
hydroperoxides, protection from ionising radiation,
maintenance of the sulfhydryl status of proteins and aiding detoxi®cation and/or excretion. Considering the
abundance of GSH in biological systems and its high af®nity for the essential Zn(II), the complexation is impor-
92
M.S. Dõaz-Cruz et al. / J. Inorg. Biochem. 70 (1998) 91±98
tant in the biological chemistry of Zn(II). For instance,
it has been determined by 1 H NMR that Zn(II) is coordinated by GSH in intact human erythrocytes [16] and
this is of interest since Zn(II)-induced protection of chinese hamster cells and three derived cell lines from the
toxic e€ects of melphalan, an alkylating agent used in
cancer chemotherapy.
The MCR method allowed the detection of the formation of 2:2 and 1:2 Cd2‡ : GSH complexes which
were not possible to detect by the visual inspection
and the direct analysis of DPP data [13].
In this paper, the procedure is extended to the analysis of the Zn2‡ ±GSH system, which yields DPP signals,
at very negative potentials, almost totally overlapped. It
is possible to detect the formation of two complexes, as
well as to propose an electrodic process for their reduction and, with the help of a molecular modelling program, a probable structure for the complexes formed.
2. Experimental section
2.1. Chemicals
Reduced GSH (>99% iodometric purity con®rmed
by both elemental analysis and HPLC with spectrophotometric detection at 220 nm), Na2 B4 O7 á10H2 O,
Zn(NO3 )2 á4H2 O, HNO3 and NaOH (Titrisol) solution
were obtained from Merck (analytical grade). The samples were usually dissolved in 0.13 M borate bu€er solution at the required pH. Ultrapure water was obtained
from a Milli-Q plus 185 System (Millipore).
2.2. Instrumentation
An Autolab System (EcoChemie) attached to a Metrohm 663 VA Stand (Metrohm) and a personal computer
was used to perform the DPP measurements. For the automatic additions of the NaOH and HNO3 solutions the
system was also connected to a Metrohm 665 Dosimat.
The manual additions of chemicals were made with
Socorex Swiss micropipetes. All the experiments were
carried out in a glass cell at room temperature under puri®ed nitrogen atmosphere. The software package used
was the GPES3 (EcoChemie) for the acquisition of the
data, and Sigma Plot, Table Curve 2D (both from Jandel
Scienti®c), and MATLAB 4.2 (Math Works Inc.) in the
data treatment. For all experiments, the working, reference and auxiliary electrodes were the static mercury
drop electrode (SMDE), with a drop area of 0.40 mm2 ,
Ag/AgCl, KCl (3 M), and glassy carbon, respectively.
Pulse duration of 40 ms, pulse amplitude of 25 mV, drop
time of 0.8 s and a scan rate of 6 mV/s were used.
2.3. Polarographic measurements
Di€erent bu€er solutions were tested in order to
choose the best medium for working at very negative potentials (ca. )1.3 V), as is the case when Zn-complexes
are involved. From those tested, borate bu€er (0.13
M) was the most idoneous, allowing a linear relationship
peak current vs. Zn2‡ concentration inside a wide range
(0±60 lM), at pH 8.5.
2.4. Experiments at ®xed pH and variable GSH-to-Zn2‡
concentration ratios
10 ml of 0.13 M borate bu€er solution at the choosen
pH are placed into the polarographic cell and deaerated
with pure nitrogen (1 atm) during 20 min before starting
the experiment. Then, the DPP scan is recorded. An aliquot of the standard Zn2‡ solution, in the same borate
bu€er medium, is manually added with a Dosimat to get
a 1 ´ 10ÿ5 M Zn2‡ solution. After deaerating for 1 min
with mechanical stirring, a new DPP polarogram is recorded. Then, diverse aliquots of a standard GSH solution are manually added, in order to get di€erent GSHto-Zn2‡ concentration ratios, from 0 to 12. All these solutions were deareated and mechanically stirred for 1 min
after each addition and before recording the respective
DPP polarograms. To avoid the presence of oxygen in
the cell, pure nitrogen was passed over the surface of the
solution during the measurement step of the experiments.
2.5. Data treatment
The method for data treatment is based on factor
analysis and multivariate curve resolution [12,13,17].
In a previous work about the GSH±Cd(II) system, the
¯owchart of the procedure for the data treatment was
described in detail [13]. The experimental DPP signals
obtained at di€erent GSH-to-Zn2‡ concentration ratios
were smoothed (fast Fourier ®ltering) and background
corrected before the matrix form arrangement. Then,
singular value decomposition (SVD) [18] was applied
to get the number of components contributing to the
DPP response. After that, evolving factor analysis
(EFA) [19,20] and alternating least-squares (ALS) [17]
were applied to get the individual polarograms and component distribution. Di€erent constraints were taken
into account during the ALS optimization such as nonnegativity and peak shape. Peak shape constraint is
based on the ®tting of the experimental DPP signal
shape to a parametric equation that provides the theoretical DPP peak shape [12].
As it was mentioned in the study of the GSH±Cd system [13], due to the nonbilinear [21] structure of the data
matrix obtained when pH is changed, the MCR method
could not be applied in this case [13,17]. The analysis of
these pH-dependent data at ®xed GSH-to-Zn2‡ concentration ratio was performed by deconvolution of each individual polarogram by nonlinear curve ®tting, assuming
that each polarogram is the sum of some peaks previously de®ned by the parametric equation mentioned above.
2.6. Molecular modelling
In order to explore if the hypothetical structure proposed from electrochemical data is reasonable, Desktop
Molecular Modeller (Oxford University Press) was used.
M.S. Dõaz-Cruz et al. / J. Inorg. Biochem. 70 (1998) 91±98
The program allows to ®nd the complex structure with
the minimum free energy, knowing the distances and angles of every expected bond with the metal. The program
calculates the partial charge of the atoms, except for ionized carboxylic acid groups in which the additional negative charge was equally distributed between the two
oxygen atoms.
3. Results and discussion
3.1. Experiments at ®xed pH and variable GSH-to-Zn2‡
concentration ratios
The polarographic behaviour of the Zn2‡ ±GSH system is rather complex. Fig. 1 summarizes the results obtained at pH 8.5 for several GSH-to-Zn2‡ concentration
ratios.
In absence of GSH in the solution, only one peak
with its maximum at ca. )1.10 V appears, which corresponds to the DPP signal of the reduction of the electrochemically labile metal under such conditions (free, plus
that probably complexed with borate and hydroxo anions).
After the addition of GSH, the current at the potential corresponding to the reduction of the labile metal
ion decreases. At the same time, an increment in the currents corresponding to the more negative potentials (in
the range between )1.15 and )1.35 V) is observed. After
further additions of GSH the voltammograms evolve in
the same way, but even at high GSH-to-Zn2‡ concentra-
93
tion ratios ([GSH]/[Zn2‡ ] ˆ 12) the voltammograms
present a main peak with a maximum at ca. )1.3 V
and a shoulder at ca. )1.15 V.
The direct interpretation of these results is not possible because of the changing size and shape of the peaks.
Therefore, the use of some multivariate analysis method,
such as that previously described [13], is highly recommendable to interpret the system response.
The ®rst step in the multivariate method is to obtain
the number of di€erent components in the system that
contribute to the DPP signal. In this work, the most
probable number of components necessary to explain
the data variation was obtained by the inspection of
the singular values of the data matrix. This is represented in Fig. 2. As it can be seen, three singular values are
higher than the rest, and the fourth singular value is at
the background noise level, suggesting that three components are needed to explain the experimental data.
This number is also checked running an ALS optimization considering only two factors (data not shown). The
lack of ®t obtained in these conditions (8.16%) was higher than the background noise level. Moreover, the residuals were accumulated in the potential region between
the two peaks, suggesting the lack of a third component.
At this level of the description of results, it must be remarked that component means mathematical component, and it does not mean necessarily chemical
component. This point will be discussed further in detail.
Once the number of components has been estimated,
EFA [19,20] provides information about how the system
evolves along the experiment. With this estimation, ALS
Fig. 1. Di€erential pulse polarograms of the Zn2‡ ±GSH system for di€erent concentration ratios at 2 ´ 10ÿ5 mol/l GSH and pH ˆ 8.5. (Mesh has
been plotted from the experimental data matrix after the background current subtraction and the smoothing of individual polarograms by fast Fourier transform ®ltering.)
94
M.S. Dõaz-Cruz et al. / J. Inorg. Biochem. 70 (1998) 91±98
Fig. 2. Singular values of the experimental data matrix obtained for
the Zn2‡ ±GSH system (2 ´ 10ÿ5 mol/l GSH and pH ˆ 8.5).
optimization [17] recovers an optimal set of concentration pro®les and individual polarographic responses.
In a ®rst ALS optimization run, only the nonnegativity constraint for concentration and voltammograms
was applied. The individual normalized polarographic
responses and concentration pro®les obtained are shown
in Fig. 3(a) and (b), respectively. The percentage of unexplained variance is low (lack of ®t of 3.79%). Therefore, it is concluded that experimental data are
satisfactorily explained by the model shown in Fig. 3(a)
and (b).
The contribution of the mathematical component I
(Fig. 3(b)) to the global signal decreases when the
GSH concentration increases. Because of its shape and
characteristic peak potential, the normalized individual
polarographic response corresponding to this component (I in Fig. 3(a)) is identi®ed as the peak corresponding to the reduction of free Zn2‡ . As a consequence,
component I corresponds to free Zn2‡ ion.
After the addition of GSH, a second component (II
in Fig. 3) appears and its contribution to the global signal increases until a GSH-to-Zn2‡ concentration ratio of
ca. 1. At this ratio its contribution reaches a plateau until a ratio of 4, and then decreases slowly until reaches
zero at high GSH-to-Zn2‡ concentration ratios (ca.
[GSH]/[Zn2‡ ] ˆ 10).
The third component (III in Fig. 3) appears at
[GSH]/[Zn2‡ ] ˆ 1, just when component II reaches the
plateau. The contribution of III increases until [GSH]/
[Zn2‡ ] ˆ 10, when it seems to reach the saturation.
The individual normalized polarographic responses
corresponding to components II and III appear at more
negative potentials than that of component I, which induces to think that they are related to di€erently complexed Zn2‡ ions. Moreover, in contrast with the
signal corresponding to the free Zn2‡ (component I in
Fig. 3(a)), polarograms of components II and III in
Fig. 3(a) do not present the expected peak shape for a
Fig. 3. ALS optimization of the experimental data set with nonnegativity constraint: (a) Individual normalized polarograms; (b) relative contribution of each component to the global signal.
typical polarographic response. Component II is a double peak with a local maximum at ca. )1.16 V and other
maximum at ca. )1.26 V. Moreover, this polarogram
overlaps almost completely with the signal corresponding to the component III. A polarographic response
with the same features that are described here for the
component II was previously found in the GSH±Cd2‡
system [12,13]. By the use of a constraint implemented
in the ALS algorithm to take into account the theoretically expected peak-shaped polarographic response, this
double peak was resolved into two independent peaks
which were interpreted as due to the electrochemical reduction of two metal ions in a binuclear cluster.
In order to apply the same approach to the GSH±
Zn2‡ system, a second ALS optimization run, using nonnegativity and peak shape constraints was performed.
The individual polarographic responses and contribution pro®les are shown in Fig. 4(a) and (b), respectively. Despite the use of the shape constraint, the
percentage of unexplained variance does not increase excessively, yielding a lack of ®t of only 4.39%.
The normalized individual polarograms (Fig. 4(a))
obtained in these conditions present the theoretical
M.S. Dõaz-Cruz et al. / J. Inorg. Biochem. 70 (1998) 91±98
95
Fig. 5. Dependence of the peak potential on pH for a 1.25 GSH-toZn2‡ concentration ratio (3.125 ´ 10ÿ5 mol/l GSH and 2.5 ´ 10ÿ5
mol/l Zn2‡ solution).
This behaviour is di€erent to that observed for the
GSH±Cd2‡ system (see Fig. 5 in Ref. [13]), in which four
components were detected. In that case, data were interpreted on the basis on a sequential formation of two
complexes. The complex 2:1 GSH : Cd appears from
the beginning, but the second complex (2:2 GSH : Cd)
only exists once the GSH-to-Cd2‡ ratio exceeds 0.5 [13].
Fig. 4. ALS optimization of the experimental data set with nonnegativity and signal shape constraints: (a) Individual normalized polarograms; (b) relative contribution of each component to the global
signal.
shape for polarographic peaks. Although the three
peaks are partially overlapped, their maxima are well de®ned. Similarly to the GSH±Cd2‡ system [12,13] these
two peaks (II and III) can be also considered as the electrochemical responses due to the reduction of Zn2‡ in
two di€erent chemical environments (i.e. coordinated
to ligands of di€erent nature).
Then, the term component must be understood in
these measurements as reducible metal ion, and not as
a chemical species. This interpretation is the only one
which yields to a reasonable interpretation of the experimental data. Only in the case of free Zn2‡ ion, the reducible metal ion and the species are the same.
Taking into account this interpretation of the polarographic responses, the pro®les obtained in these conditions (Fig. 4(b)) show that the signals corresponding to
two complexed Zn2‡ (components II and III) are present
even at very high GSH-to-Zn2‡ concentration ratios.
This fact suggests that the predominant GSH±Zn2‡
complex in the concentration range studied has a 2:2
stoichiometry and contains a binuclear cluster of high
stability.
3.2. Experiments at variable pH: In¯uence of pH on the
peak potential (Ep )
The binding of Zn2‡ by GSH has been the aim of a
great number of studies [22±27], mostly concerning pH
titration experiments. Due to the multiple binding sites
in GSH, and lack of information at the molecular level
provided by pH titration methods, there is a lack of
agreement concerning to the model for the complexation
equilibria.
It is well known that the potential at which an electrode process takes place is in¯uenced, in a predictable
manner, by the pH of the solution, as it is derived from
the Nernst equation [13]. The slope value of the plot E
vs. pH allows the determination of the ratio between
the number of protons involved in the reduction process
and the number of electrons transferred in it
E ˆ Econst ÿ …2:303mRT =nF †pH;
where m and n correspond to the number of protons and
electrons involved in the electrode reaction, respectively,
and E is a characteristic potential such as the half-wave
potential (E1=2 ) in direct current polarography, or the
peak potential (Ep ) in DPP or other voltammetric techniques giving peak-shaped signals.
The evolution of the peak potentials (Ep ), obtained at
the GSH-to-Zn2‡ concentration ratio equal to 1.25, as a
function of pH is shown in Fig. 5.
96
M.S. Dõaz-Cruz et al. / J. Inorg. Biochem. 70 (1998) 91±98
The signal corresponding to the reduction of the labile Zn2‡ (curve I), changes very slightly until pH ca.
8. From this pH the peak potential shifts to more negative values, probably due to the formation of hydroxocomplexes because a similar behaviour has been observed for free Zn2‡ in the absence of GSH (data not
shown).
The peaks II and III, corresponding to the reduction
of the complexed Zn2‡ ions, appear at the same pH region and present practically the same pH dependence,
until a pH value of ca. 9.5, with a slope value close to
0.059 V/log unit, which denotes an electrodic process involving the same number of protons and electrons. The
small deviations with respect to the expected theoretical
value (0.059 V/log unit), mainly for the peak II at
pH > 9.5, can be due to the coexistence of other minoritary species with diverse degrees of protonation.
From these results it is reasonable to conclude that
two cysteines and two other deprotonable groups, most
probably the two amino terminals, must be involved in
the complexation of two Zn2‡ ions by two molecules
of GSH; this means the formation of only one complex
with a stoichiometry of 2:2 GSH : Zn, in the presence of
an excess of metal ion, and that peaks II and III correspond to the reduction of two di€erently bound Zn2‡
ions.
At this stage, it is possible to propose the most probable electrode reduction process for the 2:2 GSH : Zn
complex based on the above exposed conclusions.
Fig. 6 summarizes the main features of the process.
For all GSH-to-Zn2‡ concentration ratios, the 2:2
GSH : Zn complex is formed with two di€erently bound
Zn2‡ ions. This complex contains a binuclear cluster
linked by bridging cysteine thiolate ligands. The two a-
amino groups and probably the two glycil carboxylic acids are also involved in the tetrahedral coordination
sphere of the two Zn.
The electrode process takes place in two steps. The
®rst step, denoted by SA in Fig. 6, is associated to the reduction of the Zn2‡ coordinated to the two binding
sulfhydryl groups and the two amino groups of the glutamyl residues (curve II in Fig. 5). As a consequence of
the ®rst step, the second one, SB , corresponds to the reduction at more negative potentials of the Zn bonded to
the two sulfhydryl groups and the two glycine carboxylic
acids (curve III in Fig. 5).
The results obtained combining all the experiments
performed and analysed in this work seem to conclude
the presence of only the Zn2 (GSH)2 complex, containing
a binuclear Zn2‡ cluster.The complex structure derived
from our data agrees with that proposed by Martin
and Edsall [26] from the analysis of pH titration data.
These authors propose the formation of complexes in
which the cysteinyl or glutamyl group of one glutathione
molecule and the cysteinyl or glutamyl group of another
one are bound to the same metal, resulting in polynuclear complexes.
In order to explore if the proposed cluster is stereochemically compatible with the tripeptide structure,
Desktop Molecular Modeller software was used [28].
Additional information about distances and angles of
every expected bond with the metal was obtained from
the literature for similar binuclear clusters [29], and
these data were implemented inside the database of the
program.
Fig. 7 shows the minimum total energy of the structure for the Zn2 (GSH)2 complex. The coordination
sphere of the metal ions has been completed with the
Fig. 6. Proposed electrode reaction mechanism for the reduction process of the 2:2 GSH : Zn2‡ complex.
M.S. Dõaz-Cruz et al. / J. Inorg. Biochem. 70 (1998) 91±98
97
Fig. 7. Proposed structure model for the 2:2 GSH : Zn2‡ complex.
two glycine carboxylic groups. The total energy of the
structure at the end of the minimization was 45.1
Kcal/mol. The energy contributions due to the bonds,
angles, and torsions were not so high. Furthermore,
the contributions due to the charge and Van der Waals
forces were negative, helping to the stability of the structure.
A similar binuclear cluster structure was obtained for
the Cd2 (GSH)2 complex, involving the two a-amino
glutamyl ends, but this was only supported by the
pH variation experiments forming the complex at
pH ˆ 10 and then decreasing the pH towards more acidic values [13]. When these kinds of experiments were carried out with the Zn2‡ ±GSH system, the E vs. pH plot
was identical to that in Fig. 5 (data not shown), indicating that there is not di€erence among forming the
Zn2 (GSH)2 complex at basic pH, or going from acidic
to basic pH.
The di€erences between these quite similar complexes
of Zn2‡ and Cd2‡ can be explained in terms of the Pearson's theory [30] of soft/hard acids and bases. Zn2‡ is in
the borderline between hard and soft metal ion, while
Cd2‡ is soft. The a-amino glutamyl end is a hard ligand,
so it binds easier to Zn2‡ than to Cd2‡ . Due to this, the
formation of the cluster proposed for the 2:2 Zn : GSH
complex is impossible at more acidic pH than in the case
of Cd2‡ complex [13].
In conclusion, in our opinion the application of electrochemical techniques in conjunction with powerful
multivariate analysis methods is an interesting approach
which merits continued developments and rigorous testing with complex systems, in order to check its advantages and limitations. It could be a useful complement
to the more powerful and sophisticated techniques such
as 1 H or 13 C-NMR coupled to mass spectrometry. A
very interesting advantage of the electrochemical approach lies in the very low concentrations needed
(lM) which allows to perform studies even at high pH
values where classical structural techniques, such as
circular dichroism or NMR, working at mM level
cannot be used because of the precipitation of some
species.
Acknowledgements
The authors gratefully acknowledge ®nancial support
from the Ministerio de Educaci
on y Cultura (DGICYT,
Projects PB93-1055 and PB93-0744), and from the Direcci
o General de Universitats de la Generalitat de Catalunya (GRQ93-1028). We also acknowledge Prof. A.
Sorribas of the Departament de Ciencies Mediques Basiques, Universitat de Lleida, and Dr. A. Herraez of the
Departamento de Bioquõmica y Biologõa Molecular,
Universidad de Alcala de Henares, for the facilities of
using Jandel Scienti®c software and DeskTop Molecular
Modeller, respectively. J. Mendieta and M.S. Dõaz-Cruz
also acknowledges ®nancial support from the Programa
de Acciones para la Incorporaci
on de Doctores y
Tecn
ologos from the Ministerio de Educaci
on y Cultura
and from a doctoral Grant of the Universitat de Barcelona, respectively.
References
[1] B.L. Vallee, D.D. Ulmer, Ann. Rev. Biochem. 41 (1972) 91.
[2] B.L. Vallee, W.E.C. Wacker, in: H. Neurath (Ed.), The Proteins,
vol. 5, Academic Press, New York, 1970.
[3] A.J. Carty, Amer. Chem. Soc. Symp. Ser. 82 (1978) 339.
[4] J.P. Greenstein, M. Winitz, Chemistry of the Amino Acids, vol. 2,
Wiley, New York, 1961, p. 569.
[5] F.R.N. Gurd, P.E. Wilcox, Advan. Protein Chem. 11 (1956) 311.
[6] M. Friedman, Protein-Metal Interactions, Plenum Press, New
York, 1974.
[7] J. Bu‚e, Complexation reactions in aquatic systems: An analytical approach, Ellis Horwood, Chichester, 1988, pp. 467.
[8] M. Esteban, C. Ari~
no, J.M. Dõaz-Cruz, E. Casassas, Trends Anal.
Chem. 12 (1993) 276 and references cited therein.
[9] M.A.G.T. van den Hoop, H.P. van Leeuwen, Anal. Chim. Acta
273 (1993) 275 and references cited therein.
[10] J.M. Dõaz-Cruz, C. Ari~
no, M. Esteban, E. Casassas, Anal. Chim.
Acta 273 (1993) 289 and references cited therein.
[11] A.M. Nadal, C. Ari~
no, M. Esteban, E. Casassas, Electroanalysis
4 (1992) 757.
[12] J. Mendieta, M.S. Dõaz-Cruz, R. Tauler, M. Esteban, Anal.
Biochem. 240 (1996) 134.
98
M.S. Dõaz-Cruz et al. / J. Inorg. Biochem. 70 (1998) 91±98
[13] M.S. Dõaz-Cruz, J. Mendieta, R. Tauler, M. Esteban, J. Inorg.
Biochem. 66 (1997) 29.
[14] H. Sies, A. Wendel (Eds.), Functions of Glutathione in Liver and
Kidney, Springer, Berlin, 1978.
[15] I.M. Arias, W.B. Jakoby (Eds). Glutathione: Metabolism and
Function, Raven, New York, 1976.
[16] D.L. Rabenstein, A.A. Isab, FEBS Lett. 121 (1980) 61.
[17] J.M. Dõaz-Cruz, R. Tauler, B.S. Grabaric, M. Esteban, E.
Casassas, J. Electroanal. Chem. 393 (1995) 7 and references cited
therein.
[18] G.H. Golub, Ch.F. van Loan, Matrix Computation, The John
Hopkins University Press, Baltimore, MD, 1989.
[19] H. Gampp, M. Maeder, Ch. Meyer, A.D. Zuberb
uhler, Talanta
32 (1985) 1133.
[20] H.R. Keller, D.L. Massart, Anal. Chim. Acta 246 (1991) 279.
[21] K.S. Booksh, B.R. Kowalski, Anal. Chem. 66 (1994) 782A.
[22] N.C. Li, O. Gawron, G. Bascuas, J. Amer. Chem. Soc. 76 (1954)
225.
[23] D.D. Perrin, A.E. Watt, Biochim. Biophys. Acta. 230 (1971) 96.
[24] A.M. Corrie, M.D. Walker, D.R. Williams, J. Chem. Soc.,
Dalton Trans. (1976) 1012.
[25] N.C. Li, R.A. Manning, J. Amer. Chem. Soc. 77 (1955) 5255.
[26] R.B. Martin, J.T. Edsall, J. Amer. Chem. Soc. 81 (1959) 4044.
[27] M.L.D. Touche, D.R. Williams, J. Chem. Soc., Dalton Trans.
(1976) 1355.
[28] Desktop Molecular Modeller, version 3.0, Oxford University
Press, Oxford, 1994.
[29] M.Y. Chow, X.M. Chen, T.C.W. Mak., J. Chem. Soc., Dalton
Trans. (1993) 3413.
[30] R.G. Pearson, J. Amer. Chem. Soc. 85 (1963) 3533.