The oxidation of 6-hydroxydopamine in aqueous solution. Part 3. Kinetics and mechanism of the oxidation with iron(III)

Article abstract of DOI:10.1039/b007203f

The kinetics of the oxidation of 6-hydroxydopamine [5-(2-aminoethyl)benzene-1,2,4-triol, protonated form H₃LH⁺] by iron(III) under anaerobic conditions are presented. A complex mechanism whereby the o- (oQ), p- (pQ), and triketoquinones (tQ) are formed via parallel inner- and outer-sphere electron transfer mechanisms has been established. The outer-sphere mechanism is particularly fast (nearly diffusion limiting) and predominates. By following the dependence of the rate on ionic strength it has been shown that a deprotonated form of 6-hydroxydopamine reacts via an outer-sphere reaction with all species of iron. Like the other catecholamines [3,4-dihydroxy-1-(2-amino-ethyl)benzenes], but to a much smaller extent, complex formation occurs by FeOH²⁺ reacting with the fully protonated form of 6-hydroxydopamine. Three different semiquinones are initially produced; two of them, the triketo- and p-semiquinones, are tautomers. The o- and triketo-semiquinones react quickly with another iron atom to form their respective quinones. The p-semiquinone, however, is seemingly stable, partly reacting with more iron and partly disproportionating to form pQ and reforming 6-hydroxydopamine. At pHs above 2.5. pQ and oQ are in equilibrium via a deprotonated quinone Q^-. The biological implications of this mechanism are discussed.

Full text of DOI:10.1039/b007203f

View Article Online / Journal Homepage / Table of Contents for this issue
1
2
The oxidation of 6-hydroxydopamine in aqueous solution. Part 3.
Kinetics and mechanism of the oxidation with iron(III)†
Guy N. L. Jameson*‡ and Wolfgang Linert
2
Institute of Inorganic Chemistry, Technical University of Vienna, Getreidemarkt 9/153, A-1060,
Vienna, Austria. E-mail: gjameson@physics.emory.edu; wlinert@mail.zserv.tuwien.ac.at
Received (in Cambridge, UK) 5th September 2000, Accepted 15th February 2001
First published as an Advance Article on the web 20th March 2001
ϩ
The kinetics of the oxidation of 6-hydroxydopamine [5-(2-aminoethyl)benzene-1,2,4-triol, protonated form H LH ]
3
by iron() under anaerobic conditions are presented. A complex mechanism whereby the o- (oQ), p- (pQ), and triketo-
quinones (tQ) are formed via parallel inner- and outer-sphere electron transfer mechanisms has been established.
The outer-sphere mechanism is particularly fast (nearly diffusion limiting) and predominates. By following the
dependence of the rate on ionic strength it has been shown that a deprotonated form of 6-hydroxydopamine reacts
via an outer-sphere reaction with all species of iron. Like the other catecholamines [3,4-dihydroxy-1-(2-amino-
2ϩ
ethyl)benzenes], but to a much smaller extent, complex formation occurs by FeOH reacting with the fully
protonated form of 6-hydroxydopamine. Three different semiquinones are initially produced; two of them, the
triketo- and p-semiquinones, are tautomers. The o- and triketo-semiquinones react quickly with another iron atom
to form their respective quinones. The p-semiquinone, however, is seemingly stable, partly reacting with more iron
and partly disproportionating to form pQ and reforming 6-hydroxydopamine. At pHs above 2.5, pQ and oQ are
Ϫ
in equilibrium via a deprotonated quinone Q . The biological implications of this mechanism are discussed.
Introduction
The reduction of iron() by 6-hydroxydopamine (protonated
ϩ
form, H LH §) is of considerable biological interest because
3
it provides a possible mechanism for the release of iron from
1
ferritin during the progression of Parkinson’s disease.
However, the potentially simpler but important kinetics of the
reaction of 6-hydroxydopamine with aqueous iron() have
until now not been investigated.
2,3
In general, catecholamines react only at low pH (0–3) with
iron(). At higher pHs bis- and tris-complexes are formed,
which are relatively stable towards internal electron transfer,
probably due to the catecholamine acting as a non-innocent
ligand. The mono-chelate complexes, which absorb at around
7
00 nm, protonate to form mono-dentate complexes, which all
undergo internal electron transfer to form the respective semi-
quinone and release an iron(). These semiquinones are
unstable and react immediately with another iron(), see
Scheme 1. Note that excess iron() reverses all these reactions
to re-produce the green complex.
Scheme 1 The oxidation of dopamine at low pH by iron() via
complex formation.
4
Noradrenaline, however, is seen to react via two parallel
pathways, the major one involving complex formation, and the
other involving direct electron transfer. (The latter is only
observed when noradrenaline is present in relatively low con-
centrations.) Similarly, at low pH, there is no evidence for a
complex between 6-hydroxydopamine and iron(). It is there-
fore believed that a similar mechanism is operating.
The kinetics of 6-hydroxydopamine oxidation are, however,
particularly difficult to interpret because of the complicated
product speciation. Depending on the pH, up to four different
quinones are produced by the two, one-electron transfers to
iron()—the p- (pQ), o- (oQ), triketo- (tQ) and the deproton-
ated-quinone (Q ) (see Parts 1 and 2).
Ϫ
6
-Hydroxydopamine and its various quinones absorb in the
UV–Vis region of the electronic spectrum making them suitable
for study by stopped-flow spectroscopy. Unfortunately, 6-
hydroxydopamine is photolysed when exposed to light of a
wavelength less than approximately 300 nm producing
unknown products that also absorb in this region. However,
as was shown in Part 2, because 6-hydroxydopamine is in
excess this reaction could be corrected for by subtraction of
an appropriate linear absorption–time curve. Another compli-
cation is that pQ and tQ have strongly overlapping absorptions
with an envelope maximum at 265 nm. Both species were there-
fore followed together until a pH of 2.5, at which point only
pQ is formed. The previous part reports the molar absorbances
of all the species and a study of the product speciation. This
†
Rate constants and absorbance data are available as supplementary
data. For direct electronic access see http://www.rsc.org/suppdata/p2/
b0/b007203f/
‡
Present address: Department of Physics, Rollins Research Center,
Emory University, Atlanta, GA 30322, USA.
For convenience, –OH protons are written to the left, and the –NH3
ϩ
§
proton to the right of L.
DOI: 10.1039/b007203f
J. Chem. Soc., Perkin Trans. 2, 2001, 569–575
This journal is © The Royal Society of Chemistry 2001
569

View Article Online
Fig. 1 The structures and optical properties of the species investigated. (Note that the maxima quoted for pQ and tQ refer to the maximum of the
absorption envelope and not those of the individual quinones, see Part 2.)
information, some of which is presented in Fig. 1, is utilised in
the present pH-dependent kinetic studies.
Experimental
(
i) Chemicals
The 6-hydroxydopamine used came from two sources, Sigma-
Aldrich and Fluka. Iron() was in the form of Fe(NO ) ؒ9H O
3
3
2
as supplied by Fluka (pro analysi).
(
ii) pH Meter
A Radiometer pH meter (pHM84) was used with pH electrodes
from Cole-Palmer, Radiometer, and Hamilton. Initial cali-
bration was carried out using two buffers (pH 2.00 and 7.00)
obtained from either Riedle-de Haën or Merck. Further
calibration of the results was achieved by using the formula
5
established by Gorton. The formula, pHcorr = pH Ϫ {0.131/
(
0.984pH)} was verified by titrating various concentrations of
nitric acid in 0.1 M solutions of KNO3.
(
iii) Stopped-flow
A stopped-flow system from Applied Photo-Physics was used
SX-17MV) with PEEK valves and flow system to minimise
(
the entry of oxygen during runs. The flow system was prepared
at the beginning of each day by purging for one hour with a
Fig. 2 A. The formation of oQ is first-order. ([L] = 1.52 mM;
Fe]added = 0.255 mM; pH = 1.02.) B. The formation of (pQ ϩ tQ) is
0
.25 M sodium dithionite solution.
[
Data collection and initial manipulation were carried out on
only first-order over about one half-life under these conditions.
an Acorn 5000 computer using Applied Photo-Physics’ own
software. Further manipulation of the data was carried out
employing Excel on a PC.
([L] = 0.379 mM; [Fe]added = 0.0409 mM; pH = 1.13.)
6
-Hydroxydopamine was always in at least a five-fold excess
over iron() in order to achieve pseudo-(first)-order kinetics.
The range of concentrations used was: for 6-hydroxydopamine
0
.30–4.00 mM; and for iron() 0.02–0.80 mM. The ionic
strength was maintained at 0.1 M with KNO (except when
3
ionic strength dependencies were measured). All solutions were
purged with argon and then transferred to the stopped-flow
apparatus in gas-tight Hamilton syringes in order to exclude
oxygen.
Results and discussion
Fig. 3 log(kobs/[L] ) vs. pH where k is the initial first-order rate
Kinetics at low pH
T
obs
constant (typical results). The fitted curve relies on data for oQ alone
Ϫ
As stated above, at low pH no complex was observed. The
formation of oQ was found to obey an accurate first-order
relationship (i.e. d[oQ]/dt = kobs[oQ]), see Fig. 2. Although
the sum of pQ and tQ did not follow such a rate law over the
complete time-scale, it was found that the first part of the
absorption–time curve could also be fitted to a first-order rate
law (see Fig. 2). (The deviation towards the end of the reaction
is consistent with a second-order rate law, see below.)
at low pH (see text). ᭺ = oQ; ᮀ = (pQ ϩ tQ); ᭝ = Q ; ᭹ = L.
this is the only species that is accurately first-order over the
whole time-scale and can be followed without the interference
of photolysis—see Part 2.
It can be seen that all species (the appearance of oQ and
(pQ ϩ tQ); and the disappearance of 6-hydroxydopamine) have
the same initial rate at low pH. This shows that they are, in
fact, all following the disappearance of iron i.e. that this is
the rate-determining step). At low pH (1–2) the reaction is
A plot of log(kobs/[L] ) vs. pH for all species, including
T
Ϫ
Q
and 6-hydroxydopamine itself, is given in Fig. 3. (The
Ϫ
2,3
formation of the deprotonated quinone, Q is discussed below.)
reciprocal first-order in protons. For most catecholamines
The fitted curve relies on data for oQ alone at low pH because
this proton dependence is caused by the iron speciation and was
5
70
J. Chem. Soc., Perkin Trans. 2, 2001, 569–575

View Article Online
Fig. 4 The ionic strength dependence of k at two pHs. ᭺ = oQ,
obs
Ϫ
ᮀ
= (pQ ϩ tQ), ᭝ = Q — = pH 1.07, – – – = pH 3.50, ؒؒؒؒؒ = theoretical
slope observed for dopamine.
Fig.
5
Evidence for the existence of a complex between 6-
hydroxydopamine and iron(). ([L] = 1.06 mM; [Fe]added = 0.185 mM;
pH = 3.53.)
2ϩ
explained by [Fe(OH)] reacting exclusively with the fully
protonated ligand, thus enabling an accurate value for the
2ϩ
FeOH
Ϫ1
protonation constant of [Fe(OH)] (K
= 660 M ) to be
Limiting value of the rate
obtained. When this relationship was used to fit the data
obtained in this study, however, a good fit was only possible
over a very short pH range. An alternative explanation was
therefore sought.
A protonation constant|| can be expressed as the ratio of the
rates of the protonation and deprotonation reactions e.g. the
last protonation constant of 6-hydroxydopamine, K is given by
4
eqn. (3).
Ionic strength dependence
ϩ
[
H LH ]
3
kϩ
8.88
(
3)
K =
4
=
ϩ
= 10
The relationship shown in eqn. (1) exists between rate constants
[
H LH][H ] kϪ
2
6
2
i
and ionic strength in which I = ¹Σz m ; z is the charge number
i
i
¯
²
Ϫ1 Ϫ¹

²
of an ion i and m is its molality;¶ A = 0.509 (mol kg )
for
i
Protonation reactions are almost always fast and in fact
11 Ϫ1 Ϫ1
aqueous solutions.
nearly diffusion limited i.e. k ≈ 10
M
s . This means that
ϩ
because the final protonation of 6-hydroxydopamine is very
k
8.88
Ϫ1
2 Ϫ1
rxn
¹
high (K = 10 M ) k must be very slow i.e. k ≈ 10 s . This
4
log ͩ ͪ= 2Az z I

²
(1)
Ϫ
Ϫ
0
rxn
ϩ Ϫ
k
has important consequences because it means that any reaction
ϩ
of H LH must be limited by the rate of deprotonation of H LH .
2
3
0
¹
If, therefore, log (krxn/krxn) is plotted vs. I ² then the slope will
The formation of the deprotonated 6-hydroxydopamine,
give a measure of the product of the charges of the reacting
species. The gradient of such a plot for dopamine, which reacts
H LH, can therefore be treated as a Bodenstein intermediate,
where krxn is the sum of all the rate constants referring to the
2
2
solely via a complex, was found to be 2. This is effectively the
theoretical slope (2.04) for the reaction between a doubly
charged species with a singly charged one. The data were there-
disappearance of iron, giving eqn. (4). This leads to eqn. (5).
d[H LH]
2
ϩ
2ϩ
ϩ
= 0 = k [H LH ] Ϫ
3
fore consistent with a reaction between [Fe(OH)] and H LH .
Ϫ
2
dt
However, 6-hydroxydopamine is found to have a near zero
ionic strength dependence at pHs of 1.07 and 3.53 (see Fig. 4)
and this is consistent with a reaction between all forms of iron
ϩ
k [H LH][H ] Ϫ k [H LH][Fe] (4)
ϩ
2
rxn
2
T
ϩ
[
H LH ]
3
with the deprotonated ligand (H LH) which has an overall
[H LH] =
(5)
2
2
ϩ
{
K [H ] ϩ (k [Fe] /k )}
4 rxn T
charge of zero [eqn. (2)].
Ϫ
ϩ
At all pH values below about 7, [H LH ] = [L] and hence
eqn. (6) applies. The factor 2 appears because the formation of
one quinone requires two iron() ions.
d[Fe]
3
T
(
2)
Ϫ
= krxn[Fe] [H LH]
T 2
dt
This is, of course, assuming that the rate constants are nearly
3ϩ
Reaction rate =
the same for both iron species involved (i.e. both Fe and
2ϩ
2krxn[Fe]
T
[L]
T
[
Fe(OH)] ).
(6)
2
krxn[H LH][Fe] =
2 T
ϩ
The slight positive dependence implies that, whereas nor-
{K
4
[H ] ϩ (krxn[Fe]
T
/k )}
Ϫ
4
adrenaline reacts mainly by complex formation with a small
contribution of an outer-sphere reaction, 6-hydroxydopamine
reacts mainly by the outer-sphere reaction with a small
contribution from that proceeding via complex formation.
Furthermore, the slope measured at a pH of 3.53 is slightly
more positive than at a pH of 1.07 (0.35 compared with
ϩ
It can furthermore be seen that when [H ] is very low (i.e. pH
ϩ
is high) K [H ] Ӷ (k [Fe] /k ) so that eqn. (7) is true. Note
that this means that the rate is also eventually zero-order in
iron.
4
rxn
T
Ϫ
obs
k
(
7)
=
2kϪ
0
.081), and this adds to the expectation that the reaction via
[
L]T
ϩ
lim[H ]→0
complex formation will increase in importance as the pH is
increased. This is confirmed when both the concentrations and
the pH are high and a small amount of a complex between
iron() and 6-hydroxydopamine (which then very rapidly
disappears) can be observed at 720 nm (see Fig. 5).
Fig. 3 (see also Fig. 6) indeed shows that kobs/[L]T for pQ
at high pH (paradoxically, Q
does increase above this value for reasons explained below).
Ϫ2 Ϫ1
Ϫ
reaches a limit of 2200 M
s
Ϫ2 Ϫ1
This value of 1100 M
s
seems a little high because it implies
¶
Note that molalities should in fact be used, but the error introduced
by using molarities is not large. However, deviations should be expected
at high values of I, because the Debye–Hückel limiting law is no longer
valid.
|| Protonation constants are used throughout this work in line with com-
pilations of data, and to be consistent with metal complex formation
constants.
J. Chem. Soc., Perkin Trans. 2, 2001, 569–575
571

View Article Online
Ϫ
d[Q ]
1
d[oQ]
dt
1
d[pQ]
dt
(
12)
=
×
=
×
ϩ
oQ
ϩ
pQ
dt
K [H ]
K [H ]
oQ
d[oQ]
dt
K
d[pQ]
dt
(
13)
=
×
pQ
K
Over this pH range, oQ cannot be measured reliably because
Ϫ
the broad band at 480 nm (corresponding to Q ) swamps its
small contribution at 390 nm and so only pQ data can be used.
Ϫ
At equilibrium the rate of appearance of the sum of pQ and Q
now equals half the rate of disappearance of iron (the rate-
determining step), eqn. (14).
Ϫ
d[Fe]T
dt
d[pQ] d[Q ]
Ά dt
Fig. 6 Plot of kobs/[L]T vs. pH showing the limiting rate for oxidation
Ϫ
= 2
ϩ
Ϫ
·
to pQ and deviation from this by the Q (see text) (᭺ = oQ;
dt
Ϫ
ᮀ
= (pQ ϩ tQ); ᭝ = Q ).
d[pQ]
dt
1
=
2
1 ϩ
1
1
Ϫ1 Ϫ1
Ά
pQ
ϩ
·
that k is 8.34 × 10
M
s
and this is faster than the accepted
K [H ]
ϩ
7
rate of transfer of a proton in water. This could, however, be
explained by the first deprotonation constant referring to the
deprotonation of one of the two o-hydroxy groups, thus provid-
2
(k ϩ k )[Fe] [L]
1
1
2
T
T
(
14)
=
1 ϩ
ϩ
Ά
pQ
ϩ
·
{K [H ] ϩ (k ϩ k )[Fe] /k }
K [H ]
4
1
2
T
Ϫ
ing a statistical factor of two. k is, therefore, effectively about
Ϫ
Ϫ1
Ϫ1
ϩ
5
50 M s . The product of the deprotonation is almost
At pH > 3.5, K [H ] Ӷ (k ϩ k )[Fe] /k . Hence eqn. (15)
4
1
2
T
Ϫ
certainly internally hydrogen bonded thereby making k more
may be derived.
Ϫ
favourable but even allowing for these effects the value remains
inexplicably high.
At low pH the expression for the rate of reaction [eqn. (6)] is
given in eqn. (8).
d[Fe]T
dt
1
(
15)
Ϫ
= 2k [L] 1 ϩ
T
Ϫ
Ά
pQ
ϩ
·
K [H ]
Ϫ
This shows that now the rate of formation of Q can
paradoxically exceed the limit set by deprotonation of 6-
hydroxydopamine by virtue of the deprotonation reactions of
the quinones. All the data can now be plotted with appropriate
theoretical lines—see Fig. 6.
2
krxn[Fe] [L]
T
T
(
8)
Reaction rate = 2krxn[H LH][Fe] =
2
T
ϩ
K [H ]
4
ϩ
This is reciprocal first-order in [H ] as is in fact observed.
ϩ
A plot of kobs/[L] vs. 1/[H] thus enables k to be calculated. A
T
rxn
9
Ϫ1 Ϫ1
value of 1.34 ± 0.05 × 10 M
s
was obtained. Note that the
Second-order component of the reaction
same result is obtained from the intercept in Fig. 3.
Ϫ
As shown above, the appearance of neither pQ nor Q is first-
Over a relatively short intermediate pH range (which,
naturally, also depends upon the initial iron() concentration)
the rate with respect to [Fe]T is neither first-order nor zero-
order. But, for reactions carried out under pseudo-order
conditions, the rate is independent of the concentration of the
reactant not in excess whether it has a first- or a zero-order
order over the whole time-scale. In fact, towards the end of the
reaction, they both follow a rate equation corresponding to a
second-order reaction (see Fig. 7).
The extent of this second-order contribution depends upon
both the relative concentration of iron() and on the pH.
At low pH and at higher iron concentrations the curve
exhibits first-order behaviour over the measured time-scale. At
higher pHs and lower iron concentrations it is mainly second-
order.
dependence. Thus the values of kobs/[L] are continuous with
T
respect to change in pH. (Following the disappearance of iron
directly would, of course, show a change in order but this is
unfortunately impossible using this technique.)
It is known that semiquinones can undergo the following
8
disproportionation equilibrium (where sq = semiquinone;
Q = quinone; L = fully reduced molecule), eqn. (16).
Deprotonation of the quinones
It can be seen from Figs. 3 and 6 that the rate of formation
Ϫ
of Q , unlike that of pQ, and contrary to expectations, does
sq ϩ sq
Q ϩ L
(16)
not seem to be limited by the rate of deprotonation of
the 6-hydroxydopamine. However, this paradox is readily
explained because pQ, oQ and Q are in equilibrium above
a pH of about 2.5 [eqns. (9) and (10), rearranging gives
eqn. (11)].
This second-order contribution could, therefore, be the
combination of two p-semiquinones to produce pQ and a 6-
hydroxydopamine, presumably by hydrogen atom abstraction
Ϫ
(
see Scheme 2). This is supported by the observation that
Ϫ
pQ and Q are still being formed long after the formation of
oQ has stopped, (which implies that all the iron() has
been consumed). One can also conclude from this that the
p-semiquinone is particularly stable in contrast to the
o-semiquinone, which reacts immediately with another iron().
Further evidence for the existence of this reaction is provided
by the fact that, at very low concentrations of iron, an induc-
tion curve is observed followed by second-order kinetics
(see Fig. 7). In other words the p-semiquinone is initially
formed very rapidly but because of a lack of iron(), pQ
is subsequently formed exclusively by means of the dispropor-
tionation reaction given in Scheme 2.
[
Ϫ
oQ]
Ϫ
ϩ
oQ
pQ
(9)
(10)
(11)
Q ϩ H
oQ
K
K
=
=
=
ϩ
[
Q ][H ]
[
Ϫ
pQ]
Ϫ
ϩ
Q ϩ H
pQ
ϩ
[
Q ][H ]
[
oQ]
oQ
[pQ]
pQ
Ϫ
[
Q ] =
ϩ
ϩ
K [H ] K [H ]
Differentiating yields eqn. (12), which can be rearranged to
eqn. (13).
The proposed reaction scheme is given in eqns. (17)–(19).
5
72
J. Chem. Soc., Perkin Trans. 2, 2001, 569–575

View Article Online
Scheme 2 Plausible reaction scheme for the disproportionation of the p-semiquinone.
d[psq]
2
=
Ϫk [psq]
(24)
3
dt
and integrated from [psq] to [psq] and t to t to give eqns.
0
0
(
25)–(27). This is then inserted into the rate equation and the
[
psq]
t
d[psq]
(25)
(26)
(27)
(28)
=
Ϫk3 dt
∫
2
∫
[
psq]
t
[
psq]
0
0
1
1
Ϫ
= k (t Ϫ t )
3
0
[
psq] [psq]0
[
psq]0
[
psq] =
{
1 ϩ k [psq] (t Ϫ t )}
3
0
0
[
pQ]
t
2
k3[psq]0
dt
d[pQ] =
∫
∫
2
2
{1 ϩ k [psq] (t Ϫ t )}
3 0 0
[
pQ]
0
t
0
resulting expression integrated to give eqn. (28) which can be
rewritten as eqn. (29) or (30).
[
pQ] Ϫ [pQ] =
0
2
Ϫ
k3[psq]
1
1
Fig. 7 A. Second-order contribution to the rate of formation of Q
0
(29)
(30)
Ϫ
Ά
·
corresponding to disproportionation of the semiquinone. ([L] = 0.662
mM; [Fe]added = 91.2 µM; pH = 5.00.) B. Plot showing induction period
followed by second-order kinetics for low iron of (pQ ϩ tQ).
2
k3[psq]0 k [psq] {1 ϩ k [psq] (t Ϫ t )}
3
0
3
0
0
[
psq]0
[psq]0
(
[L] = 0.311 mM; [Fe]added = 2.05 µM; pH = 0.98.) (The derivation of the
[
pQ] =
ϩ [pQ] Ϫ
0
theoretical curves is given later in the text. The number of data points
has been reduced from 1000 for clarity.)
2
2{1 ϩ k [psq] (t Ϫ t )}
3 0 0
Therefore, [pQ] was plotted against 1/2{1 ϩ k [psq] (t Ϫ t )},
k
2
3
0
0
Fe() ϩ L
Fe() ϩ psq
psq ϩ psq
Fe() ϩ psq
(17)
(18)
(19)
and the values of k , t , and [psq] varied until the best straight
3
0
o
line was fitted. (Absorptions were converted to concentrations
Ϫ
fast (k)
(i.e. [pQ] or [Q ]) by using the extinction coefficients previously
Fe() ϩ pQ
pQ ϩ L
obtained in Part 2.) Note that when t is defined, so also is
0
[
pQ] , and the intercept can be used to check that the fit
0
k
3
is internally consistent. Furthermore, t was increased until
0
the value for k did not vary any further. That is because as the
3
extent of reaction increased so the assumption that all the iron
has been used up is more valid. Therefore only the second-order
reaction is taking place. An example of the fit obtainable is
given in the insets in Fig. 7.
From this the rate equations [eqns. (20)–(23)] can be obtained.
d[Fe()]
dt
d[Fe()]
dt
Ϫ
= [Fe()]{k [L] ϩ k[psq]} =
(20)
(21)
2
Values for the second-order rate constant have been obtained
Ϫ
over a wide range of pHs (at higher pHs the formation of Q
can be followed) and these are shown in Fig. 8.
d[L]
2
Ϫ
= k [Fe()][L] Ϫ ¹k [psq]
2 3
¯
The observed rate constant is seen to be relatively constant
but then starts to rise slowly above a pH of about 6, consistent
with a log K value of 6.8 although this value is, of course, only
approximate. Although the two protonated semiquinones react
fairly slowly at low pH, when the semiquinone deprotonates the
reaction rate increases rapidly, presumably because hydrogen
atom abstraction is facilitated. The log K value of about 6.8
therefore refers to the deprotonation of the semiquinone and it
is significant that it lies between that of the 6-hydroxydopamine
(log K = 8.88) and the quinone (log K = 4.13).
dt
²
d[psq]
dt
2
=
k [Fe()][L] Ϫ k[Fe()][psq] Ϫ k [psq] (22)
2
3
d[pQ]
dt
2
=
k[Fe()][psq] Ϫ ¹k [psq]
(23)
3
¯
²
Towards the end of the reaction (t = t ) all the iron has been
0
consumed (i.e. [Fe()]→0) and the semi-quinone (psq) has built
This deprotonated p-semiquinone will presumably have the
structure shown below.
up to a concentration [psq] , eqn. (24). This can be rearranged
0
J. Chem. Soc., Perkin Trans. 2, 2001, 569–575
573

View Article Online
Scheme 3 The final reaction scheme.
There are two possible ways to explain the enhanced reac-
tivity of this species. Either the hydrogen bridge changes the
Ϫ
reactivity of the –O
group or, conceivably, the bridging
᝽
proton could be more readily removed as an atom. However,
this latter explanation would require the interaction of two
negatively charged centres.
Other approaches
Fig. 8 pH dependence of the second-order rate constants for the dis-
proportionation of the semiquinones. The superimposed curve assumes
During the course of this work, two other lines of approach to
the problem of solving the second-order reaction were used
with some limited success.
Ϫ1 Ϫ1
a log K value of 6.8 and a maximum rate of 950 M
s . ᮀ = pQ,
Ϫ
᭝
= Q .
Firstly, the scheme (assuming oQ was not formed) was
modelled using the non-linear fitting program Scientist from
Micromath. Although this succeeded in producing curves of
the correct general form observed (i.e. very similar to the curves
illustrated in Figs. 2 and 7) the program (or computer) could
not cope with the numerical solutions of the differential
equations required in order to fit actual data with theory.
solved analytically. Therefore, the method finally adopted was
that outlined above, despite the fact that it was only possible to
deal with either end of the reaction. Even this fails as far as
initial rate determinations are concerned when the iron concen-
trations are low and a long induction period occurs (Fig. 7).
The overall reaction scheme
9
10
Secondly, it was seen that Chien and later Pearson et al.
had solved some of these very complicated reaction schemes by
the use of Bessel functions. Again this line of approach had a
certain success but to fit the overall scheme proved too difficult
because the parallel reaction to form the o-quinone cannot be
ignored and, unfortunately, a scheme including this cannot be
At low pH it has been shown in Part 2 that tQ is the main
component of the 265 nm band. At these pHs there is little sign
of a second-order reaction occurring indicating that this, like
oQ, is formed by rapid attack of a second iron() on the
semiquinone.
5
74
J. Chem. Soc., Perkin Trans. 2, 2001, 569–575

View Article Online
Table 1 The rate constants obtained (see Scheme 3)
addition of a third hydroxy group to dopamine radically alters
its properties explaining why, alone among the catecholamines,
6-hydroxydopamine is able to interact with the iron() stored
in ferritin. (Our investigations into this reaction will be
published elsewhere.)
Ϫ1
Ϫ1 Ϫ1
Ϫ1 Ϫ1
Ϫ1 Ϫ1
Ϫ1 Ϫ1
k /s
kϩ/M
s
k1/M
s
k2/M
s
k3/M
s
Ϫ
11
8
8
5
50
4.2 × 10
4.6 × 10
8.8 × 10
pH dependent
The problem of a pH dependence on the amount of pQ
Acknowledgements
formed at low pH has been solved by allowing a protonation
equilibrium between the triketo- and p-semiquinones. This was
necessary because there is no equilibrium between the quinones
themselves over the time-scales employed. This was established
by NMR studies (see Part 1 in this series) and discussed fully
when drawing up the speciation diagram in Part 2. The com-
pleted reaction scheme incorporating all these features is
depicted in Scheme 3.
The authors wish to thank the “Fonds zur Förderung der
wissenschaftlichen Forschung” (FWF) (Project No. 11218-
CHE), and the Austrian Federal Ministry of Science and
Transport (Project No. GZ 70.023/2-Pr/4/97). This project was
also supported by the “Österreichische Nationalbank” (Project
No. 5556). We would further like to thankfully acknowledge
the support of this project via the EU COST project D8
(
Chemistry of Metals in Medicine) and the TMR project
The rate of disappearance of iron(), k = (k ϩ k ), has
rxn
Ϫ1 Ϫ1
1
2
9
(TOSS) supported by the European Community under the
Contract number ERB-FMRX-CT98-0199.
been calculated to be 1.34 ± 0.05 × 10 M s . It has now been
established, by NMR and the speciation investigations, that the
ratio of the sum of tQ and pQ produced at low pH to oQ is
1
.94. It therefore follows that from this that k = 1.94k .
2 1
References
Hence the rate of formation of oQ from the reaction of all
species of iron() with the deprotonated 6-hydroxydopamine,
k , is 4.6 × 10 M
1
2
3
W. Linert, E. Herlinger, R. F. Jameson, E. Kienzl, K. Jellinger and
M. B. H. Youdim, Biochem. Biophys. Acta, 1996, 1316, 160.
U. El-Ayaan, E. Herlinger, R. F. Jameson and W. Linert, J. Chem.
Soc., Dalton Trans., 1997, 2813.
W. Linert, R. F. Jameson and E. Herlinger, Inorg. Chim. Acta, 1991,
187, 239; W. Linert, E. Herlinger and R. F. Jameson, J. Chem. Soc.,
Perkin Trans. 2, 1993, 2435.
8
Ϫ1 Ϫ1
8
Ϫ1 Ϫ1
s
and that for pQ, k = 8.8 × 10 M
s
1
2
(
Table 1).
Biological implications and conclusions
It can be clearly seen that 6-hydroxydopamine reacts with
iron() in an extremely complicated way. The mechanism is not
completely unique, however, and comparisons with the other
catecholamines, and noradrenaline in particular, can be drawn.
Dopa, dopamine, and adrenaline form catecholate complexes
which undergo internal electron transfer. Noradrenaline, on
the other hand, was found to react by parallel inner- and outer-
sphere mechanisms but mainly via complex formation. A
relation between reduction potential and reactivity should be
possible but the redox potentials are difficult to measure and
almost certainly the one-electron redox potentials are more
important and as yet unmeasured.
The ability of 6-hydroxydopamine to react via an outer-
sphere mechanism allows very fast rates to be achieved. This
reaction via direct electron transfer (“outer-sphere” electron
transfer) means that 6-hydroxydopamine can react at all pHs
with iron(), whereas dopamine forms bis- and tris-complexes
at physiological pHs that are stable towards electron exchange.
This seems to be an important part of its chemistry and maybe
explains why it is so dangerous to biological systems. The
4
U. El-Ayaan, R. F. Jameson and W. Linert, J. Chem. Soc., Dalton
Trans., 1998, 1315.
5 J. E. Gorton, PhD Thesis, University of St. Andrews, Scotland,
968.
6
1
P. W. Atkins, Physical Chemistry, 4th edn., Oxford University Press,
Oxford, 1990, p. 860.
H. Strehlow, Rapid Reactions in Solution, VCH, Weinheim, 1992,
p. 108; D. A. McQuarrie and J. D. Simon, Physical Chemistry,
2,3
4
7
A
Molecular Approach, University Science Books, Sausalito,
California, 1997, p. 1070.
8
J. K. Dohrmann and B. Bergmann, J. Phys. Chem., 1995, 99, 1218;
A. E. Alegría, M. López and N. Guevara, J. Chem. Soc., Faraday
Trans., 1996, 92, 4965; V. A. Roginsky, L. M. Pisarenko, W. Bors,
C. Michel and M. Saran, J. Chem. Soc., Faraday Trans., 1998, 94, 13,
1
835; V. A. Roginsky, L. M. Pisarenko, W. Bors and C. Michel,
J. Chem. Soc., Perkin Trans. 2, 1999, 871.
9 J.-Y. Chien, J. Am. Chem. Soc., 1948, 70, 2256.
0 R. G. Pearson, L. C. King and S. H. Langer, J. Am. Chem. Soc.,
1
1
1
1
951, 73, 4149.
1 Part 1. G. N. L. Jameson, A. B. Kudryarstev and W. Linert, J. Chem.
Soc., Perkin Trans. 2, 2001 (DOI: 10.1039/b007912j).
2 Part 2. G. N. L. Jameson and W. Linert, J. Chem. Soc., Perkin
Trans. 2, 2001 (DOI: 10.1039/b007162p).
J. Chem. Soc., Perkin Trans. 2, 2001, 569–575
575