Myoglobin modification by enzyme-generated dopamine reactive species

Article abstract of DOI:10.1002/chem.200801014

The generation of reactive quinone species (DAQ) from oxidation of dopamine (DA) is involved in neurodegenerative pathologies like Parkinson's disease (A. Borta, G. U. Hoeglinger, J. Neurochem. 2007, 100, 587-595). The oxidation of DA to DAQ can occur either in a single two-electron process or in two consecutive one-electron steps, through semiquinone radicals, giving rise to different patterns of reactions. The former type of reaction can be promoted by tyrosinase, the latter by peroxidases in the presence of H₂O ₂, which can be formed under oxidative stress conditions. Both enzymes were employed for the characterization of the thiol-catechol adducts formed by reaction of DA and cysteine or glutathione, and for the identification of specific amino acid residues modified by DAQs in two representative target proteins, human and horse heart myoglobin. Our results indicate that the cysteinyl-DA adducts are formed from the same quinone intermediate independently of the mechanism of DA oxidation, and that the hallmark of a radical mechanism is the formation of the cystine dimer. The reactivity of quinone species also controls the DA-promoted derivatization of histidine residues in proteins. However, for the modification of the cysteine residue in human myoglobin, a radical intramolecular mechanism has been proposed, in which the protein acts both as the catalyst and target of the reaction. Most importantly, the modification of myoglobins through DAQ linkages, and in particular by DA oligomers, has dramatic effects on their stability, as it induces protein unfolding and incorporation into insoluble melanic precipitates.

Full text of DOI:10.1002/chem.200801014

FULL PAPER
DOI: 10.1002/chem.200801014
Myoglobin Modification by Enzyme-Generated Dopamine Reactive Species
Stefania Nicolis, Matteo Zucchelli, Enrico Monzani, and Luigi Casella*[a]
Dedicated to Professor Jan Reedijk on occasion of his 65th birthday
Abstract: The generation of reactive
quinone species (DAQ) from oxidation
of dopamine (DA) is involved in neu-
rodegenerative pathologies like Parkin-
sonꢀs disease (A. Borta, G. U. Hçglin-
ger, J. Neurochem. 2007, 100, 587–595).
The oxidation of DA to DAQ can
occur either in a single two-electron
process or in two consecutive one-elec-
tron steps, through semiquinone radi-
cals, giving rise to different patterns of
reactions. The former type of reaction
can be promoted by tyrosinase, the
latter by peroxidases in the presence of
H2O2, which can be formed under oxi-
dative stress conditions. Both enzymes
were employed for the characterization
of the thiol–catechol adducts formed
by reaction of DA and cysteine or glu-
tathione, and for the identification of
specific amino acid residues modified
by DAQs in two representative target
proteins, human and horse heart myo-
globin. Our results indicate that the
cysteinyl–DA adducts are formed from
the same quinone intermediate inde-
pendently of the mechanism of DA ox-
idation, and that the hallmark of a rad-
ical mechanism is the formation of the
cystine dimer. The reactivity of qui-
none species also controls the DA-pro-
moted derivatization of histidine resi-
dues in proteins. However, for the
modification of the cysteine residue in
human myoglobin, a radical intramo-
lecular mechanism has been proposed,
in which the protein acts both as the
catalyst and target of the reaction.
Most importantly, the modification of
myoglobins through DAQ linkages,
and in particular by DA oligomers, has
dramatic effects on their stability, as it
induces protein unfolding and incorpo-
ration into insoluble melanic precipi-
tates.
Keywords: dopamines · melanins ·
myoglobin · protein modifications ·
quinones
Introduction
minergic neurons of the substantia nigra (SN), reduced anti-
oxidant capacity due to glutathione (GSH) deficiency,[6–8]
and by an increase in iron concentration in the SN.[9–11]
In the brain, tyrosinase (Ty) may enzymatically oxidize
excess amounts of DA to form melanin, thus preventing the
slow progression of cell damage induced by the reactive
oxygen species (ROS) generated by the competing autoxida-
tion of DA.[12–14] Two types of melanin pigments are pro-
duced in mammals, the normal black-to-brown eumelanins
and the reddish-brown pheomelanins. Both eumelanin and
pheomelanin are generated from dopaquinone through a
series of redox reactions, but pheomelanin also needs the
presence of thiols (such as cysteine or GSH), as the units of
pheomelanins (benzothiazines) are formed from addition re-
actions of cysteine to dopaquinone.[15] Furthermore, the qui-
none formed upon oxidation of DA, dopamine quinone
(DAQ), forms an addition product with cysteine (cysteinyl–
DA).
Catecholamines, such as the neurotransmitter dopamine
(DA) and its precursor l-dopa, and oxidative stress are mu-
tually implicated in neurodegeneration. DA-induced neuro-
toxicity is in particular involved in the pathogenesis of Par-
kinsonꢀs disease (PD), a degenerative neurological disorder
characterized by hypokinesia, rigidity, and tremor.[1–4] Oxida-
tive stress has long been linked to the neuronal cell death
associated with various neurodegenerative pathologies, such
as Alzheimerꢀs disease, PD, and amyotrophic lateral sclero-
sis, but it is still unclear whether oxidative stress is a major
cause or merely a consequence of neurodegeneration.[5] PD,
in particular, is characterized by high levels of DA in dopa-
[a] Dr. S. Nicolis, Dr. M. Zucchelli, Dr. E. Monzani, Prof. L. Casella
Dipartimento di Chimica Generale
Università di Pavia
Via Taramelli 12, 27100 Pavia (Italy)
Fax : (+39)0382-528544
E-mail: bioinorg@unipv.it
The potential relevance of the oxidation products of DA
and cysteinyl–DA to PD prompted several groups to charac-
terize their structure, the metabolic pathways, and their tox-
icity.[16–24] The initial products from the reactions of cysteine
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801014.
Chem. Eur. J. 2008, 14, 8661 – 8673
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8661

with catecholamine quinones generated with a variety of
reactive oxygen and nitrogen species has been extensively
studied previously by our group, and this allows us to apply
an established protocol for protein fragmentation and analy-
sis of the modification sites.[43,44]
[21]
methods (electrochemically,[20] enzymatically with Ty/O2
Cꢀ
or peroxidase/H2O2,[22] with a system generating O2 and
H2O2,[23] or with FeII/H2O2[24]) have been shown to be the 2-
S- and 5-S-cysteinyl–catecholamine (always the predominant
product), and the 2-S,5-S-dicysteinyl–catecholamine conju-
gates.
Results
The cytotoxicity of catechols has been ascribed to cova-
lent binding of the quinones to various proteins, giving rise
to pathological modifications.[25–33] DAQ modification of
target proteins can apparently induce changes in their struc-
tures and properties, aggregation, precipitation, and in some
instances promote cell death. Actually, the accumulation of
aberrant or misfolded proteins, protofibril formation, and
deposits in Lewy bodies in the brain is one important fea-
ture in PD.[12,34,35] Quinones are reactive towards a variety of
nucleophiles, including the amino group of DA or l-dopa
and the side chains of many amino acids. Most amino acids
add relatively slowly to quinones, except for cysteine, which
not only favorably competes with intracyclization of the
amino group in DAQ, but adds to this quinone three orders
of magnitude faster than the side chain amino group of
other amino acids.[36] The reactivity of various quinones with
the cysteine residues of bovine serum albumin, alcohol de-
hydrogenase, isocitrate dehydrogenase,[37,38] and the human
DA transporter[39] have been reported.
The oxidation of DA to DAQ is a two-electron process
and it may occur either in a single step or in two consecutive
one-electron oxidations. In the latter case, a reactive semi-
quinone radical intermediate is formed, which may give rise
to a different pattern of reactions with respect to DAQ. In
this context, it is interesting that a simpler pattern of gluta-
thionyl–catechin adducts was obtained by using peroxidase
and H2O2 with respect to tyrosinase and O2.[40] Since a varie-
ty of oxidants of biological relevance are able to oxidize
DA, it would be important to assess whether the various ox-
idants may lead to different products in DA reactions. Here
we report on the analysis of the thiol–catechol adducts ob-
tained when lactoperoxidase (LPO), in the presence of
H2O2, or Ty are used for the oxidation of DA. The kinetic
analysis of the peroxidase-catalyzed oxidation of DA and
thiols leads to the proposal of a reaction pathway for the
generation of the thiol–catechol conjugates. In the case of
tyrosinase the activity is limited to DA, whereas thiols like
cysteine are inhibitors of the enzyme.[41] The reactive species
generated by the two enzyme systems have been also stud-
ied for their capability to modify amino acid residues in pro-
teins. To this end, two representative globular proteins were
used as targets, human myoglobin (HMb) and horse heart
myoglobin (hhMb), with the aim of investigating the com-
petitive occurrence of modification at cysteinyl, histidyl, or
lysyl residues, since only HMb contains a cysteine residue
(Cys110). Mbs have been chosen as target proteins for their
relatively small size, compact structure, and the extensively
accumulated biochemical and biophysical characterization
data, which includes a large number of X-ray crystal struc-
tures.[42] In addition, the modification of these proteins by
Catalytic activity of LPO and the Mbs in catechol oxidation:
LPO, hhMb, and HMb, in the presence of hydrogen perox-
ide, catalyze the oxidation of DA and l-dopa to the corre-
sponding quinones, which rapidly evolve to cyclic com-
pounds (aminochrome and dopachrome, respectively) with a
characteristic absorption band at l=476 nm. The reaction
rates depend on both the H2O2 and catechol concentration.
Under H2O2 saturating conditions, the protein activity de-
pends on the catechol concentration through Michaelis–
Menten-type behavior. From data analysis, the following ki-
netic parameters were obtained: the catalytic rate constant
(kcat), the Michaelis–Menten constant (KM), and kcat/KM; the
data are collected in Table 1. As expected, LPO exhibits
Table 1. Steady-state kinetic parameters for the LPO-, hhMb-, and
HMb-dependent oxidation of DA and l-dopa by H2O2 in 200 mm phos-
phate buffer (pH 7.5) at 258C.[a]
Protein
Substrate
KM [mm]
kcat [sꢀ1
]
kcat/KM [mꢀ1 sꢀ1
(5.6ꢂ0.6)”105
]
LPO
LPO
hhMb
hhMb
HMb
HMb
DA
l-dopa
DA
l-dopa
DA
l-dopa
3.6ꢂ0.5
n.d.
2010ꢂ90
n.d.
ACHTREUNG
(1.0ꢂ0.2)”104
46ꢂ5
1.42ꢂ0.05
0.98ꢂ0.11
ꢁ5.1
31ꢂ3
44ꢂ5
22ꢂ5
ꢁ270
19.4ꢂ0.9
570ꢂ30
1.06ꢂ0.07
0.60ꢂ0.01
[a] n.d.=not determined.
higher catalytic activity than hhMb and HMb. HMb is much
more effective than hhMb in discriminating between DA
and l-dopa. This behavior depends entirely upon a strong
difference in the KM values (ꢁ270 and ꢁ1 mm for DA and
l-dopa, respectively). As a result, the catalytic efficiency of
HMb at low catechol concentration (expressed by the pa-
rameter kcat/KM) is also significantly higher in the oxidation
of l-dopa with respect to the oxidation of DA.
To understand if this difference in selectivity between the
Mbs were connected to the mode of protein–substrate inter-
1
action, H NMR spectroscopy relaxation time measurements
were performed for both DA and l-dopa in the presence of
hhMb or HMb. In these experiments, the paramagnetic con-
tribution to relaxation by the high-spin Fe3+ center of the
protein can be exploited to get an estimate of the distances
of the protons of bound catechol from the iron atom.[45] The
data show that the effect of the Mbs on the relaxation of the
catechol protons is negligible: assuming an electron relaxa-
tion time (ts) of 5”10ꢀ11 s,[45] the iron–proton distances for
the Mb-bound catechols was evaluated to be above 11 Š in
all cases. This suggests that DA and l-dopa cannot enter
into the distal cavity of the protein and approach the heme;
therefore, substrate discrimination likely depends on specific
8662
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8661 – 8673

Myoglobin Modification
FULL PAPER
electrostatic or polar interactions of the catechols with resi-
dues on the protein surface.
compound II species with catechol is by far more efficient
than that with N-acetyl-l-cysteine.
The reactivity of the compound II derivative of the Mbs
was studied for hhMb, as it is known that HMb in the pres-
ence of H2O2 (and in the absence of substrates) does not
form a stable compound II but undergoes homodimer for-
mation through cross-linking between two cysteine resi-
dues.[46] Therefore, HMb is not suitable for mechanistic anal-
Reaction of LPO and hhMb compound II derivatives with
N-acetylcysteine and DA: The reaction of heme proteins
with hydrogen peroxide yields an unstable compound I in-
termediate that rapidly evolves to an appreciably longer-
lived compound II intermediate, the reduction of which con-
stitutes the rate-limiting step of the peroxidase catalytic
ysis. In the reduction of hhMb compound II to the met
IV
ꢀ
ꢀ
cycle. The reduction of the compound II intermediate LPO
form, the spectral changes for hhMb Fe =O reduction by
IV
III
ꢀ
Fe =O to the resting, met form LPO Fe by N-acetyl-l-
cysteine or DA was followed spectrophotometrically through
the shift of the Soret band under pseudo-first-order condi-
tions; with N-acetyl-l-cysteine (cys), the observed rate con-
stant (kobs) was linearly dependent on the substrate concen-
tration (Figure 1) according to Equation (1):
either DA or N-acetyl-l-cysteine with time followed first-
order behavior. The observed rate constants depend on the
substrate concentration with a saturating behavior, and can
be described by Equation (2):
kmax½SŠ
ð2Þ
kobs
¼
þ b
KD þ ½SŠ
kobs ¼ kII½cysŠ þ b
ð1Þ
in which KD represents the dissociation constant of the
IV
ꢀ
hhMb Fe =O/substrate (S) (i.e., DA or N-acetyl-l-cys-
teine) complex, and kmax is the first-order decay constant of
the ferryl complex to the met form. Also in this case, to
take into account the self-decay of the compound II inter-
mediate, the b constant was introduced into Equation (2).
Fitting of the experimental data to Equation (2) gave the ki-
netic parameters reported in Table 2.
Regarding the reduction of hhMb compound II by DA,
the high efficiency of the reaction precludes the possibility
of obtaining rate data employing [DA]>8 mm (above which
the reaction finishes in the mixing time), thus making the es-
timate of kmax unreliable. Therefore, the only reliable param-
IV
ꢀ
eter obtained from the reduction of hhMb Fe =O by DA is
the ratio kmax/KD =32mꢀ1 sꢀ1, which is very similar to the kcat
/
KM value obtained from steady-state studies (Table 1).
Moreover, from the comparison of the curves and the kinet-
ic parameters obtained with dopamine and N-acetyl-l-cys-
teine (reported in Figure 2 and Table 2, respectively), it
emerges that also for myoglobin (as for LPO) the reaction
of the compound II species with catechol is by far more effi-
cient than that with cysteine.
Figure 1. Dependence of the pseudo-first-order rate constant for reduc-
III
ꢀ
tion of LPO compound II (3.7 mm) to LPO Fe on N-acetylcysteine con-
centration in 0.2m phosphate buffer (pH 7.5) at 258C. The inset shows, as
an example, the absorbance changes (l=412–436 nm) with time in the
experiment with N-acetylcysteine (0.015 mm).
in which the nonzero value for the intercept on the y axis
IV
ꢀ
(b) is due to self-reduction of LPO Fe =O to the met
Identification of cysteinyl–DA and glutathionyl–DA conju-
gates: The Cys–dopa conjugates were isolated and charac-
terized by reacting l-dopa with Cys in the presence of tyro-
sinase and O2,[21] which oxidize the catechol to quinone in a
two-electron process. The adducts formed were 2-S-cystein-
yl–l-dopa, 5-S-cysteinyl–l-dopa (most abundant), and 2-S-5-
S-dicysteinyl–l-dopa. This product pattern is similar to that
form. Fitting of the experimental data to Equation (1) gave
the rate constant kII =(268ꢂ7)mꢀ1 sꢀ1 for the reduction of
LPO compound II with N-acetyl-l-cysteine.
In the presence of DA, the reduction of LPO compoun-
d II is very fast; with the amount of DA required by the
pseudo-first-order conditions, the reaction was over in less
than 1 s. Therefore, only an esti-
mate of the rate constant kII
IV
IV
ꢀ
ꢀ
could be obtained, approxi-
Table 2. Kinetic parameters for LPO Fe =O and hhMb Fe =O reduction by DA and N-acetylcysteine in
200 mm phosphate buffer (pH 7.5) at 258C.
mately 2.5”105 mꢀ1 sꢀ1, that is,
in the range of the kcat/KM pa- Protein
Substrate
KD [mm]
kmax [sꢀ1
]
kmax/KD [mꢀ1 sꢀ1
]
kII [mꢀ1 sꢀ1
]
ꢁ2.5”105
rameter obtained from steady-
LPO
LPO
hhMb
hhMb
DA
–
–
–
–
–
–
N-acetyl-l-cysteine
DA
N-acetyl-l-cysteine
268ꢂ7
state studies (Table 1). A com-
parison of the kII values shows
that the reaction of the LPO
19ꢂ9
0.6ꢂ0.2
32ꢂ4
–
–
<2
0.025ꢂ0.003
ꢁ15
Chem. Eur. J. 2008, 14, 8661 – 8673
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8663

L. Casella et al.
Figure 2. Dependence of the pseudo-first-order rate constant for reduc-
Figure 3. HPLC elution profile, with absorbance reading at l=220 nm,
for the LPO (8”10ꢀ8 m) catalyzed reaction of N-acetylcysteine (20 mm)
with DA (30 mm) in the presence of H2O2 (5 mm) in 200 mm phosphate
buffer, pH 7.5, at room temperature. The assignment of the peaks is
shown.
III
ꢀ
*
tion of hhMb compound II (3.7 mm) to hhMb Fe on DA ( ) or N-ace-
*
tylcysteine ( ) concentration in 0.2m phosphate buffer (pH 7.5) at 258C.
obtained by generating DAQ electrochemically from dopa-
mine,[20] whereas the 5-S-isomer was the only detectable
adduct when DA was oxidized by a system generating O2
cysteine (data not shown). Thus, the only marker for DA
semiquinone formation is the cystine dimer, whereas the
Cys–DA adducts are formed from the same quinone inter-
mediate independently of the mechanism of DA oxidation.
With the aim of modeling the reactivity of the Cys residue
in a peptide environment, the tripeptide glutathione (GSH)
was taken into consideration. When GSH was reacted with
LPO-generated dopamine semiquinone and quinone species,
the mixture of DA conjugates contained 2-S-glutathionyl–
DA, 5-S-glutathionyl–DA (see Figure S3 in the Supporting
Cꢀ
and H2O2 (as a model of oxidative stress).[23] With these last
two methods, catechol oxidation may occur through a one-
electron oxidation process leading to a transient dopamine
semiquinone radical, which then dismutates in solution to
DAQ and DA. To investigate the reactivity of the DA semi-
quinone radical, N-acetyl-l-cysteine was employed, in which
the protected amino group prevents the cyclization of cys-
teinyl–DA conjugates to generate benzothiazines.[47] The
semiquinone radical was generated by LPO in the presence
of H2O2, which is very efficient in this reaction.[48,49] The
products of the enzymatic reaction were separated by using
HPLC and characterized by using MS and NMR spectrosco-
py (with 1H and 1H-1H DQF-COSY spectra). Besides the
peak of unreacted DA, four major compounds were isolat-
ed, corresponding to 2-S-N-acetyl-l-cysteinyl–DA, 5-S-N-
acetyl-l-cysteinyl–DA (the predominant product), 2-S-5-S-
di-N-acetyl-l-cysteinyl–DA, and N-acetyl-l-cystine; no sig-
nificant amount of 6-S-N-acetyl-l-cysteinyl–DA was detect-
ed. An analogous mixture of products was reported for the
horseradish peroxidase catalyzed oxidation of l-dopa in the
presence of Cys.[22] Figure 3 shows the HPLC profile, togeth-
er with the assignment of the peaks to the various products.
The spectroscopic characterization for all the products is re-
ported in Table S1 in the Supporting Information; for the
cysteinyl–DA conjugates, the data are similar to those re-
ported previously by Xu et al.[20]
1
Information for its H-1H DQF-COSY spectrum), 2-S-5-S-
diglutathionyl–DA, and the glutathionyl dimer (Table S2 in
the Supporting Information), which are analogous to the
products derived from the reaction of N-acetyl-l-cysteine,
but also 6-S-glutathionyl–DA, which was obtained in a sig-
nificant amount to allow its NMR spectroscopic and MS
characterization. Also Ito et al.[50] reported, for l-dopa con-
jugates obtained by Ty oxidation, that the yield of 6-S-gluta-
thionyl–l-dopa is much higher than that of the 6-S-cystein-
yl–l-dopa adduct. The complete spectroscopic characteriza-
tion of the 2-S- and 6-S-glutathionyl–dopamine conjugates is
reported here for the first time, whereas the other adducts
have been previously characterized.[51]
Also in this case, for the reaction of GSH with DA in the
presence of Ty, under the same experimental conditions as
with LPO/H2O2, the same products with similar relative
yields were obtained, except for the absence of the gluta-
thionyl dimer. Thus, the latter compound is in vitro a
marker of the occurrence of dopamine semiquinone radi-
cals.
A product pattern of Cys–DA adducts similar to that ob-
tained with LPO, but with lower yields, was obtained with
the Mbs, indicating a similar mechanism for the Mb and per-
oxidase-mediated formation of the adducts. Also, when the
reaction of N-acetyl-l-cysteine with DA was carried out in
the presence of Ty, under the same experimental conditions
as before but without H2O2, the products and relative yields
were similar, except for N-acetyl-l-cystine, which in this
case was formed only in trace amounts by autoxidation of
Histidinyl–DA and lysyl–DA conjugates: Besides cysteine,
histidine can also form addition products with quinones, but
with lower efficiency (the reaction of DAQ with N-acetyl-
cysteine is estimated to be at least 106 times faster than that
of N-acetylhistidine) and a different regioselectivity.[52,53] By
operating under the same conditions as those described pre-
8664
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8661 – 8673

Myoglobin Modification
FULL PAPER
Table 3. Percentage derivatization of Mb residues (Cys110 in HMb,
His81 and/or His82 in hhMb derivatives, respectively) upon modification
of the proteins with DA in the presence of the catalytic systems indicat-
ed, operating in 200 mm phosphate buffer (pH 7.5). The percentage deri-
vatization of each modified residue was calculated from the HPLC–MS/
MS traces by considering the ratio of the ion current of the modified pep-
tide relative to the total ion current of the same peptide (in both the un-
modified and modified form).
viously for cysteine in the presence of LPO and H2O2, and
also with different concentrations of the reagents, both histi-
dine and N-a-acetylhistidine gave only the DA oxidation
products, which indicated that the reaction of imidazole
with the quinone is too slow to compete with the internal
cyclization of DAQ and the consequent oligomerization.
The same absence of nucleophilic reactivity towards
enzyme-generated DAQ species, at least under the relatively
mild conditions employed here, was observed for lysine.
These results agree with those obtained by Whitehead
et al.[39] regarding the reactivity of the electrochemically
generated DA quinone with cysteine, lysine, and histidine
(at pH 7.4); also in that case DAQ was found to react only
with cysteine. Products from the reaction of histidine with
quinones were obtained when N-acetyldopamine was em-
ployed,[53] in which the amino protecting group prevents the
faster DAQ cyclization reaction.[36]
Protein target
Catalytic system
Cys110
His81/His82
+151 amu [%]
+302 amu [%]
HMb
HMb
HMb
hhMb
hhMb
hhMb
HMb/H2O2
LPO/H2O2
Ty/O2
hhMb/H2O2
LPO/H2O2
Ty/O2
30
5
0
0
0
ꢁ3
ꢁ2
ꢁ2
0
[a]
[a]
[a]
[a] Cys110 is absent in hhMb.
fied peptide. The MS/MS spectrum of this modified peptide
is shown in Figure 4. In the case of hhMb, the modification
is observed at the histidine residues His81 and/or His82 in
the 79–87 peptide, with an increment of 302 amu with re-
Modification of HMb and hhMb by reactive DA quinones:
Although histidine (His) and lysine (Lys) as free amino
acids exhibit very low reactivity against DAQ and dopamine
semiquinone, their reactivity in a protein could be different
due to the influence of the protein environment. To assess
this point, the derivatization of Cys, His, and Lys residues in
the Mbs by the reactive quinone species generated from DA
was investigated. The modifications undergone by HMb and
hhMb were analyzed by using HPLC–ESI-MS/MS upon re-
acting DA (1 mm) and Mb (6”10ꢀ5 m) under various condi-
tions: 1) in the presence of H2O2 alone (0.3 mm), in which
the Mbs act both as a source and a target of the quinones;
2) in the presence of H2O2 (0.3 mm) and LPO (8”10ꢀ8 m);
and 3) in the presence of Ty (8”10ꢀ8 m). In the second and
third cases, the high reactivity of LPO[54] and Ty[55,56] with
DA make the contribution of the Mbs self-promoted deriva-
tization negligible. In these experiments, the reactants (H2O2
and DA) were added to the solutions divided into small ali-
quots; in this way, the concentration of the reactive quinone
or semiquinone species was always kept very low, with the
aim of simulating a pathophysiological condition.[57,58] More-
over, it is worth noting that no derivatization was observed
for the fraction of HMb or hhMb left in solution when these
proteins were reacted with higher concentrations of H2O2
(up to 3 mm concentration); this may indicate that under
forcing conditions the proteins were incorporated into the
melanic precipitate that is formed in the reactions (see
below).
Figure 4. MS/MS spectrum of the m/z 1033.1 peak (mass of 2064.3 amu)
assigned to the 103–118 peptide of HMb in a double-charged state con-
taining the Cys110–DA adduct. The assignment of the y and b ion series
are shown. Above the spectrum, the sequence of the 103–118 peptide is
shown with the modified residue in bold and with the summary of the y
and b ions found in the spectrum.
The modifications undergone by the proteins were identi-
fied by tandem MS studies on the polypeptide fragments re-
sulting from tryptic digestion of the apomyoglobin (apoMb)
derivatives. The data are reported in Table 3, together with
the percentage of derivatization obtained from the integra-
tion of the peaks in the chromatograms with extracted ion
current (EIC). The HPLC–ESI-MS/MS analysis with the
SEQUEST algorithm showed the presence of two types of
derivatization in the Mbs. In HMb, the Cys110 residue in
the 103–118 peptide was found modified with a DA moiety,
with a mass increase of 151 amu with respect to the unmodi-
spect to the molecular weight of the unmodified peptide. In
the latter case, the MS/MS data are consistent both with the
addition of one DA molecule to each His or a DA dimer to
a single His.
It has been reported that the oxidation of DA by perox-
idase/H2O2 (but not by Ty/O2) could also lead to the neuro-
toxin 6-hydroxydopamine, the toxicity of which is related to
the susceptibility to nucleophilic attack by the protein resi-
dues to the corresponding quinone.[59] Nevertheless, in our
Chem. Eur. J. 2008, 14, 8661 – 8673
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8665

L. Casella et al.
conditions we never observed this type of protein derivatiza-
tion, which would correspond to the presence of a modified
HMb or hhMb peptide with a mass increment of 167 amu
with respect to the unmodified peptide.
modified peptide. The MS/MS spectra are consistent with
the addition of five DA units at the His81/His82 cluster and
four DA units at the His93 residue (see Figure S4 in the
Supporting Information for its MS/MS spectrum). It is likely
that other multiderivatized sites exist in the protein–melanic
deposits, even though their detection may be complicated.
The greater reactivity of Cys relative to His residues ap-
pears both from the absence of His modification in HMb
and by the much larger amount of Cys110 derivatization in
HMb with respect to the His81/His82 derivatization in
hhMb. Interestingly, His modification is not observed in
HMb, possibly because Cys110 in HMb protects His resi-
dues from the reaction with the quinone species. Regarding
the amount of Cys110–DA formed in the various experi-
ments, the endogenous derivatization promoted by HMb ap-
pears to be the most efficient mechanism, as Cys110–DA is
formed in much higher yield (30%) than in the presence of
LPO as external catalyst (5%). On the other hand, at least
under the mild conditions employed here, Ty does not
appear to promote the formation of a DA adduct of the Cys
residue to any appreciable extent, possibly because the com-
peting melanization process is faster.
Guanidine hydrochloride (Gdn-HCl) denaturation assay:
The unfolding midpoint [Gdn-HCl]0 and thermodynamic pa-
rameters obtained for Gdn-HCl-induced denaturation of
native Mbs and the Mbs resulting from covalent DA modifi-
cation are reported in Table 4. In the case of hhMb, the deri-
Table 4. Thermodynamic parameters for unfolding induced by Gdn-HCl
(DGANꢀU and ꢀm) and Gdn-HCl concentration causing 50% denaturation
([Gdn-HCl]0) of native and modified Mbs (labeled Mbs*, and obtained
by reaction of the proteins with 1 mm DA and 0.3 mm H2O2), in 200 mm
phosphate buffer (pH 6.0) at 258C.
Protein
DGANꢀU [kcalmolꢀ1
]
ꢀm [kcalmolꢀ1 mꢀ1
]
A
hhMb
hhMb*
HMb
6.17ꢂ0.09
5.54ꢂ0.04
5.17ꢂ0.06
n.d.[a]
6.05ꢂ0.09
5.31ꢂ0.04
5.04ꢂ0.06
n.d.[a]
1.028ꢂ0.003
1.039ꢂ0.002
1.030ꢂ0.003
n.d.[a]
The reactivity of the His81/His82 residues in hhMb to-
wards DAQ appears to be different from the above-dis-
cussed Cys reactivity in HMb, since a low (<5% yield) and
comparable extent of derivatization was obtained in all the
conditions (hhMb/H2O2, LPO/H2O2, or Ty/O2).
HMb*
[a] The experimental data of protein absorbance versus [Gdn-HCl] show
a complex pseudosigmoidal behavior that cannot be fitted with the equa-
tion involving the denaturation of a single protein.
Tandem MS analysis of the protein/melanic precipitate: The
formation of a dark brown precipitate was observed in the
reaction mixtures obtained upon modification of both HMb
and hhMb in the presence of reactive DAQ species, whatev-
er their source (Mb/H2O2, LPO/H2O2, or Ty/O2). The
amount of insoluble material increased by dialyzing the
apoMb derivatives before subjecting them to tryptic frag-
mentation. The dark color of the precipitate, both before
and after heme extraction from the Mbs, indicates the pres-
ence of a melanin-type DA polymer. Upon melanin forma-
tion, the oligomers of reactive quinones may covalently link
the proteins, thus producing insoluble melanin–protein con-
jugates. For this reason, the precipitates collected from the
reaction mixtures were also subjected to peptic, tryptic, or
consecutive peptic–tryptic digestion, and the resulting poly-
peptide fragments were analyzed by using HPLC–ESI-MS/
MS. For both the HMb and hhMb derivatives, the peptides
resulting from each proteolytic treatment cover the com-
plete Mbs sequence and, together with the unmodified pep-
tides, also the modification at Cys110 (+151 amu) in HMb
and at His81/His82 (+302 amu) in hhMb were observed.
Numerous attempts were performed with the aim of identi-
fying other possible derivatization sites; in particular, the
SEQUEST algorithm was applied to the search of peptide
fragments containing histidine or lysine residues modified
with up to eight DA moieties per protein molecule. Actual-
ly, the hhMb/melanin precipitates obtained both in the ab-
sence and in the presence of LPO, and proteolyzed with
pepsin followed by trypsin, clearly revealed the presence of
the 80–96 peptide with a mass increment of 1359 amu (i.e.,
nine DA moieties) with respect to the corresponding un-
vatization of the His81/His82 residues with two DA mole-
cules, even at low levels (ꢁ3%, Table 3), appreciably affects
the stability of the protein, lowering both the free energy
change for conversion of native to unfolded protein in the
absence of denaturant (DGANꢀU) and the solvent-exposed sur-
face area (ꢀm). These results indicate that the modification
of polar residues (i.e., histidine) make the protein less sensi-
tive to a hydrophilic denaturant like Gdn-HCl, as confirmed
also by the higher Gdn-HCl concentration causing 50% de-
naturation ([Gdn-HCl]0) obtained in the case of modified
hhMb with respect to the native protein (Table 4).
A more considerable effect on protein stability is induced
by DA derivatization of the Cys110 residue in HMb. Actual-
ly, the presence of a significant fraction of modified protein
in the case of HMb reacted with DA/H2O2 (ꢁ30%, Table 3)
significantly alters the curve of absorbance versus [Gdn-
HCl]. The graph does not show a single sigmoid and there-
fore could not be fitted with the equation corresponding to
a simple two-state model employed for native HMb.
Discussion
Mechanism of the catalytic formation of cysteinyl–DA con-
jugates: The high reactivity of tyrosinase in the oxidation of
catechols like DA and l-dopa[55,56] requires the formation of
the quinone followed by the nucleophilic addition of the
ꢀ
SH group for the cysteinyl–catechol conjugates, obtained
from the reaction of catechols with N-acetylcysteine (or
GSH). In the case of peroxidases, the situation is complicat-
8666
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8661 – 8673

Myoglobin Modification
FULL PAPER
Scheme 1. Mechanism for the LPO/Mb-promoted formation of cystine (path a) and monocysteinyl–DA conjugates (path b).
ed since the enzyme-active species can react with both the
thiol and the catechol and with the latter substrate both the
quinone and semiquinone species are formed in the solu-
tion. The kinetic studies on the activity of the LPO or Mb
intermediates and the analysis of the composition of the
product mixture obtained from the reaction of DA with N-
acetylcysteine, or GSH, in the presence of LPO or Mb/
H2O2, in comparison with those obtained with Ty, supports
the reaction mechanism reported in Scheme 1.
the cysteinyl thiol group on the electron-deficient quinone
leads to the monocysteinyl adducts (2-S-Cys–DA, 5-S-Cys–
DA, or 2-S-GSH–DA, 5-S-GSH–DA, and 6-S-GSH–DA).
Conversely, the cystine dimer is generated by a coupling re-
action of two cysteinyl radicals.
Another possible pathway for the generation of the cys-
teinyl–dopamine adducts is the coupling between semiqui-
none and cysteinyl radicals. Nevertheless, such a reaction
can be excluded by considering the similar product composi-
tion for the Cys–DA adducts obtained by one-electron oxi-
dation (catalyzed by LPO or Mb) and two-electron oxida-
tion (catalyzed by Ty). The product distribution is controlled
by the reactivity of DAQ with cysteine, both for the LPO/
Mb-promoted and the Ty-promoted reactions, and not by
the reactivity of the semiquinone species. The formation of
the cystine dimer in the former reaction is significant, since
it can be considered the hallmark of a radical mechanism,
which is also that supposed to occur under oxidative stress
conditions. Regarding the regioselectivity of the nucleophilic
attack, the preference for the C5 position of the quinone
ring has been explained to result from an intramolecular
base-catalyzed Michael 1,6-addition, based on the formation
of a hydrogen bond between the carbonyl group at the C4
position and the thiol proton.[61,62]
In particular, as the second reduction step in the perox-
idase catalytic cycle[60], involving the compound II inter-
mediate, is generally the slow step, we determined the rate
IV
ꢀ
constants for the reduction of the LPO Fe =O intermedi-
ate by DA and N-acetylcysteine. The rate constants (kII) ob-
tained for the reaction of N-acetylcysteine and DA, 268 and
2.5”105 mꢀ1 sꢀ1, respectively, indicate that when both of
these substrates are present in solution, the peroxidase will
strongly prefer the reaction with catechol, which will be oxi-
C
dized to the semiquinone radical (DA ). As an example,
IV
ꢀ
with [DA]=18 mm, the LPO Fe =O species is converted
into the native form in less than 1 s, whereas with cysteine
at a similar concentration (15 mm) the reaction requires 400–
500 s (inset in Figure 1). In the case of myoglobin, the pref-
erence for the oxidation of DA with respect to cysteine is
not so striking, but at the concentration of the two sub-
strates employed in the formation of the cysteinyl–DA con-
jugates ([DA]=30 mm and [N-acetylcysteine]=20 mm), it is
The dicysteinyl–DA conjugates (2-S-5-S-diCys/GSH–DA)
can be obtained by reaction of cysteine with the 5-S-Cys–
DAQ adduct (or 2-S-Cys–DAQ, see below) generated from
the reaction between the DA and 5-S-Cys–DA semiqui-
nones, according to Scheme 2. The latter species is derived
from the direct one-electron oxidation of 5-S-Cys–DA by
possible to evaluate from the data in Table 2 that only a
IV
ꢀ
minor fraction (ꢁ6%) of hhMb Fe =O is reduced by cys-
teine rather than by DA. The higher DA reactivity could
also be evaluated from the extrapolation of the curves in
Figure 2 at high substrate concentration.
C
LPO/Mb compounds I and II (path c), or by the DA -medi-
ated oxidation of 5-S-Cys–DA (path d). Actually, dopamine
and cysteinyl–dopamine compete for the reaction with LPO/
Mb reactive species, the former being favored by its higher
concentration (path d) and the latter by its lower redox po-
tential[20] (path c).
C
The DA radical generated from the one-electron oxida-
tion of dopamine by the LPO or Mb compound I and II,
can either disproportionate, thus generating DAQ (path b,
Scheme 1), or oxidize a cysteine molecule to a cysteinyl rad-
ical (path a). In the former case, the nucleophilic attack of
Chem. Eur. J. 2008, 14, 8661 – 8673
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8667

L. Casella et al.
C
Scheme 2. Mechanism for the LPO/Mb-promoted formation of the dicysteinyl–DA conjugate through protein-active species (path c) and DA (path d).
Scheme 2 shows the formation of 2-S-5-S-diCys–DA start-
ing from 5-S-Cys–DA, since this is the predominant mono-
cysteinyl adduct in the reaction mixture, but the dicysteinyl
conjugate could also be obtained from 2-S-Cys–DA accord-
ing to an analogous reaction pathway.
confirmed by our results, since Cys110 was modified to a sig-
nificant extent in the HMb self-promoted reaction (Table 3).
It is interesting to note that with the enzymes used to pro-
mote DA oxidation, LPO is less efficient than HMb itself in
the DA derivatization of cysteine, and that Ty does not pro-
mote Cys110 derivatization to any detectable extent. This
could indicate a different reactivity of the HMb cysteine res-
idue in the one-electron and two-electron oxidation path-
ways. Although the three-dimensional structure of wild-type
HMb is not available, a reference structure for localizing the
residues subjected to modifications is the crystal structure of
the HMb mutant Lys45Arg/Cys110Ala (Figure 5).[69] In par-
The similarity in product pattern observed when GSH is
used instead of N-acetylcysteine indicates that Schemes 1
and 2 can be extended to cysteine-containing peptides such
as GSH and proteins, and also explains the generation of
the mono- and diglutathionyl conjugates (2-S-glutathionyl–
dopamine, 5-S-glutathionyl–dopamine, 6-S-glutathionyl–dop-
amine, and 2-S-5-S-diglutathionyl–dopamine, respectively).
It is worth noting that a key role has been attributed to
GSH in neuron degeneration; its translocation from glial
cells (where it is largely synthesized)[63] into the cytoplasm
of dopaminergic cell bodies in the SN, increases in response
to chronic brain insult.[16,64] Interestingly, the glutathionyl–
DA adducts, resulting from the scavenging of DAQ by
GSH, are hydrolyzed to cysteinyl–DA adducts by intra-
neuronal peptidase enzymes.[57,65,66] The possibility that
GSH, rather than cysteine, is the precursor of cysteinyl–DA
is supported by the higher GSH concentration in the brain
with respect to cysteine.[57,67] A neuroprotective role of GSH
has been also suggested since it can act as a DAQ scaveng-
er.[68]
Figure 5. Structure of the Lys45Arg/Cys110Ala mutant of HMb.[69] The
disposition of the side chains of Cys110, Hys81, and His82 present in the
wild-type protein are shown.
Myoglobin modification by DA reactive species—the initial
steps: Under oxidative stress conditions, where relatively
high levels of hydrogen peroxide are produced, activation of
heme proteins and in particular peroxidases can occur. In
the presence of catecholamines, such as DA, the consequent
DAQ reaction with target proteins can induce changes in
their structures and properties, with pathological implica-
tions.[25–33]
The reactive species formed upon oxidation of DA can
diffuse into the solution but also react with the protein re-
sponsible for their formation. The Mbs are good targets for
the investigation of the capability of the DA-generated reac-
tive species of modifying amino acid residues of the heme
protein generating them since, due to their limited reactivity,
they act both as the catalyst and substrate. In particular,
HMb contains several histidines and lysines, and a single
cysteine, which is expected to be the most reactive residue
towards quinone species.[52] The high Cys reactivity has been
ticular, Cys110 is not directly exposed to the protein surface,
therefore its modification probably indicates that the active
C
species, that is, DA or DAQ, diffuses from outside the distal
cavity (where DA binds to the Mbs according to our NMR
spectroscopy relaxation-time results) inside the protein. Ac-
cording to the data reported in Table 3, the diffusion of the
semiquinone species to the site of derivatization is more ef-
ficient, in particular when generated by the same HMb pro-
tein molecule that also contains the target residue of the de-
rivatization.
Another possibility is that the Cys110 residue is oxidized
to the cysteine radical upon reaction of HMb with
H2O2.[44,70] In this case, Cys110 will be formed by the HMb
C
active species analogue of compound I of peroxidases. This
8668
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8661 – 8673

Myoglobin Modification
FULL PAPER
yields an HMb compound II-like species that in turn reacts
with a DA molecule oxidizing it to semiquinone. The cou-
nucleophilic attack, we could extend the results obtained
from the free amino acid–catechol adduct to the protein
system by considering the C6 position of the DA aromatic
ring[53] as the main derivatization site. The reactivity of His
residues of hhMb towards DAQ
C
C
pling between DA and Cys110 radicals gives rise to the ob-
C
served HMb–DA adduct (Scheme 3). The DA radical may
species differs from that of Cys
in HMb in that a similar extent
of His81/His82 modification is
obtained in the former case
with all of the DAQ-generating
protein
systems
(Table 3),
Scheme 3. Radical mechanism for the HMb self-promoted formation of the Cys110–DA adduct.
which indicates an intermolecu-
lar mechanism of derivatization.
approach the thiol radical from the active site of HMb or
from the bulk of the solution in an intermolecular process.
This hypothesis outlines a completely different mechanism
for the HMb Cys110–DA adduct formation with respect to
that formed with free cysteine. In this case, in fact, the reac-
tion occurs through radical coupling, whereas the previous
mechanism consists of a nucleophilic addition of the thiol to
the DA–quinone. Some indication of the actual mechanism
could be deduced from the regiochemistry of the Cys–DA
covalent linking in the protein. The MS/MS fragmentation
pattern of the DA-modified peptide 103–118 shown in
Figure 4 does not allow one to distinguish between the pos-
sible isomers (2-S, 5-S, or 6-S). However, the presence of a
single peptide containing the DA modification on Cys110 in
the HPLC chromatogram seems to indicate that only one of
the possible Cys–DA isomers is present as the major or only
species in the modified protein, possibly at C5 of the aro-
matic ring, which is the principal site of attack with the free
amino acid with both nucleophilic or radical mechanisms.[20]
The reduced extent of modification of Cys110 produced
by LPO/H2O2 (Table 3) may be related to the fraction of hy-
drogen peroxide that, instead of giving rise to compound I
of the peroxidase, reacts with HMb to give the thiol radical.
In fact, although the rate constant for the formation of LPO
compound I is several orders of magnitude larger[71,72] than
that of Mb,[73] the concentration of the latter protein in solu-
tion is much larger (thousands fold). The relatively high
yield observed for Cys110 self-modification by HMb may be
seen as a protecting effect of the protein when it acts in
“side reactions” such as that generating the undesirable
DAQ reactive species. The scavenging effect of Cys110 is
also in agreement with the lack of His modification in HMb.
With hhMb it is possible to analyze the reactivity of histi-
dine residues, as cysteine is absent. Whereas upon treatment
of free histidine with the DAQ-generating systems no hysti-
dinyl–DA conjugates were formed, hhMb DA derivatization
occurred at H81 and/or H82 residues even under mild con-
ditions. This indicates that the hhMb environment can influ-
ence the reactivity of histidines, and enhance it. The effect
could be derived from a lowering of the pKa of the imida-
zole residue, which is induced by the protein residues in the
local environment through their (partial) positive charge
and polarity; this is suggested by the fact that modification
occurs at a His cluster. Regarding the regiochemistry of the
Thus, in this process the reactive DAQ species is generated
by a protein molecule different from that containing the his-
tidine targets of modification.
Myoglobin modification by DA reactive species—protein–
melanic conjugates: Modification of proteins through DAQ
linkages has dramatic effects on their stability, as it may
induce unfolding and protein precipitation. Even at early
stages, the reaction of quinone species with protein residues
depresses the thermodynamic stability of the protein
(Table 4), but further reactions lead to more dramatic conse-
quences. We systematically observed that the modification
of HMb or hhMb by reactive DAQ species occurs with the
concomitant formation of a dark melanic precipitate. Also,
the oligomers formed as intermediates in the process of mel-
anin formation are reactive species that may react with pro-
tein residues. The observation of Mb peptides in the tryptic
fragments from treatment of the protein–melanic precipitate
indicates that Mb molecules are incorporated into the pre-
cipitate. The inability to control these side reactions and the
precipitation of the protein prevents the full detection of the
protein residues modified by DAQ species. In addition, an
extensive modification of protein residues masks the recog-
nition sites to the proteolytic enzymes, thus preventing frag-
mentation of the protein and thorough analysis of the modi-
fication sites. Actually, the HPLC–MS/MS analysis always
showed the His81/His82-modified peptide in the hhMb–mel-
anin precipitates; it was also found left in solution. Since
this modification by itself does not induce the protein pre-
cipitation (as deduced from the unfolding studies), we can
presume that what causes hhMb precipitation is the cross-
linking between the protein and DAQ oligomers. In fact,
proteolytic digestion of hhMb–melanin precipitate showed
the peculiar modification of the His81/His82 cluster with
five DA units and the His93 residue with a 4-DA oligomer.
We can speculate that the derivatization of the His81/His82
cluster promotes the unfolding of the protein and this facili-
tates the attack of dopamine oligomers to other protein resi-
dues and causes the incorporation of the protein into the
precipitate. In this respect, it is worth noting that His93 is
the heme proximal ligand in the native protein and there-
fore it can become accessible only upon unfolding of hhMb.
Scheme 4 describes a possible mechanism for the forma-
tion of adducts between histidine residues and DA oligo-
Chem. Eur. J. 2008, 14, 8661 – 8673
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8669

L. Casella et al.
Scheme 4. Proposed mechanism for the formation of the adducts of histidine–DA oligomers in hhMb.
mers. The process starts with the initial addition of DAQ to
a histidine residue. Since His–DA, like Cys–DA,[20] is more
readily oxidized to quinone than DA itself, the protein-
bound DA can be an end product but also a potential elec-
tron-donor agent, as similarly proposed by Akagawa et al.
for the precipitation of the Cys–DA modified proteins.[13]
For hhMb and other proteins lacking cysteine it is necessary
to evoke the reaction between histidine residues and DA
oligomers as most likely responsible for the formation of
protein–melanic conjugates. It should be also appreciated
that DAQ undergoes a fast intramolecular cyclization and
that this reaction reduces its half-life in solution and com-
petes with the addition of an external nucleophilic group.
But in the oxidized form of His–DA the cyclization is pre-
vented, as the imidazole moiety blocks the C6 position of
the quinone ring, thus allowing the addition of the amino
group of another DA molecule. Therefore, in the mechanism
proposed in Scheme 4, the formation of the growing DA oli-
8670
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8661 – 8673

Myoglobin Modification
FULL PAPER
the formation of the 2,2’-azinobis(3-ethylbenzothizoline-6-sulfonic acid)
(ABTS) radical cation according to a standard enzymatic method.[76] Lac-
toperoxidase was purified from bovine milk as previously described;[77]
recombinant human Mb was expressed and purified as previously report-
ed.[44] In all the experiments the Mbs were utilized in their met form. All
spectrophotometric measurements were performed on a Hewlett Packard
HP 8452A diode array spectrophotometer.
gomeric chain occurs in several steps, in which the oxidation
of the catechol species bound to the protein yields quinone
derivatives that are unable to give internal cyclization. The
oxidation is then followed by the addition of a DA mole-
cule, instead of DAQ, to the modified protein. This type of
nucleophilic addition by DA amino groups to the quinone
species is in agreement with the constant mass increment of
151 amu observed in tandem MS studies, which corresponds
to the addition of entire DA units. In the mechanism report-
ed in Scheme 4, both the oxidation of the catechol ring of
the DA molecule directly connected to the histidine residue
and the terminal DA, according to path a and path b, re-
spectively, have been considered. The terminal arrows re-
ported in the scheme indicate that the DA oligomerization
process can proceed with the oxidation of every catechol
moiety present in the growing protein–melanic conjugate.
Furthermore, besides the reaction of the DA amino group,
the nucleophilic addition of DA hydroxyl groups to the qui-
none form of the growing adduct, which would lead to the
formation of ether bonds, also cannot be excluded as a fur-
ther modification pathway.
Kinetic studies of catechol oxidation: The kinetic experiments were car-
ried out in 200 mm phosphate buffer, pH 7.5, using a quartzcuvette with
path length of 1 cm, kept at (25.0ꢂ0.1)8C by using a thermostat, and
equipped with a magnetic stirrer. The initial solution containing the pro-
tein (LPO, hhMb, or HMb) and variable substrate (DA or l-dopa) con-
centrations (final volume 1600 mL) was obtained by mixing solutions of
appropriate concentration of the reagents in the buffer. The reaction was
started by the addition of the H2O2 solution and was followed during the
initial 10–15 s by monitoring the absorbance change at l=476 nm for
both the substrates (the lmax of both dopaminechrome (DAC) and dopa-
chrome (dopaC)). The conversion of the rate data from absorbance per
second into molarity per second was done by using the extinction coeffi-
cients of DAC and dopaC at l=476 nm (e=3300 and 3600mꢀ1 cmꢀ1, re-
spectively).[78] The kinetic parameters were obtained by fitting the plots
of experimental rates at different catechol concentrations to the Michae-
lis–Menten equation.
For each substrate, the rate dependence on the reactant concentrations
was studied through two steps: 1) finding a suitable [H2O2] that maximiz-
es the rate but avoids unwanted excess of the oxidant, and then using
this [H2O2] for step 2) in which the dependence of the rate versus [cate-
chol] was studied. The LPO and hhMb (or HMb) concentrations were
0.01 and 1 mm, respectively, while the concentrations of the other reac-
tants were as follows: with LPO and DA, [H2O2]=0.53 mm, [DA]=5–
50 mm; with LPO and l-dopa, [H2O2]=0.53 mm, [l-dopa]=2–40 mm;
with hhMb and DA, [H2O2]=36 mm, [DA]=5–250 mm; with hhMb and
l-dopa, [H2O2]=36 mm, [l-dopa]=2–33 mm; with HMb and DA,
[H2O2]=136 mm, [DA]=0.7–30 mm; with HMb and l-dopa, [H2O2]=
136 mm, [l-dopa]=0.6–25 mm.
Conclusion
The present study has shown that, notwithstanding the high
reactivity of the catechol-derived reactive species (semiqui-
nones and quinones), only specific amino acid residues are
modified in the proteins. Their surface exposure and the
local environment in the protein seems to be a key factor
ruling the reactivity. The extent of protein modification also
depends upon the system generating the reactive quinone
species, and whether the mechanism is of radical type and
involves intra- or intermolecular processes. Most important-
ly, when the derivatization becomes extensive it induces pro-
tein precipitation and its incorporation into the melanic pre-
cipitate. We are currently trying to set up new protocols for
the fragmentation and analysis of the insoluble protein–mel-
anic conjugates, as this type of investigation represents a
fundamental step in the understanding of the nature, com-
position, and mechanism of formation of neuromelanins and
other protective neuronal pigments.[74] We are currently ex-
tending the investigation of the effect of dopamine modifi-
cation to another protein of the globin family, human neuro-
globin, a protein that is expressed in the brain, and contains
three cysteine residues, but the function of which is still un-
clear.[75]
The reaction rates observed for the noncatalytic reaction, that is, in the
absence of the protein, or without hydrogen peroxide are completely
negligible.
IV
IV
ꢀ
ꢀ
Reactions of LPO Fe =O and hhMb Fe =O with N-acetylcysteine or
DA: The LPO and hhMb compound II intermediates were prepared by
incubating the proteins (3.7 mm in 200 mm phosphate buffer, pH 7.5) with
H2O2 (2 equiv) for about 15 min, until the Soret band shifted from 412 to
430 nm in the case of LPO, and from 410 to 420 nm in the case of hhMb,
IV
ꢀ
and stabilized at the final wavelengths. The reduction of LPO Fe =O to
LPO Fe , and hhMb Fe =O to hhMb FeIII, after the addition of the
reducing substrate, was followed spectrophotometrically by recording the
variation of absorbance at 412 and 436 nm, and at 410 and 428 nm, re-
III
IV
ꢀ
ꢀ
ꢀ
spectively, at (25.0ꢂ0.1)8C.
IV
ꢀ
The reaction of LPO Fe =O with DA was carried out by adding the
substrate at 18 mm final concentration; in the case of N-acetylcysteine,
different substrate concentrations (from 0.015 to 4.7 mm) were employed.
IV
ꢀ
The reactions of hhMb Fe =O with both DA and N-acetylcysteine were
studied by employing different substrate concentrations, from 0.15 to
7.8 mm for DA, and from 0.31 to 10 mm for N-acetylcysteine.
NMR spectroscopy relaxation measurements: The effect of the addition
of variable amounts of hhMb or HMb (0–200 mm) on the T1 relaxation
time for the protons of both DA (40 mm) and l-dopa (40 mm), in deuter-
ated 0.2m sodium phosphate buffer pD 7.5, was determined at 258C with
a Bruker AVANCE 400 NMR spectrometer operating at 400.13 MHz
proton resonance, by using the standard inversion recovery method.[79] To
eliminate interference by metal impurities, a small amount of ethylene-
diaminetetraacetic acid (EDTA) was added to the solutions.
Experimental Section
Reagents: All buffer solutions were prepared with deionized Milli-Q
water. Hydrogen peroxide (30% solution), dopamine hydrochloride, l-
dopa, N-acetyl-l-cysteine, N-a-acetyl-l-histidine, l-histidine, N-a-acetyl-
l-lysine, reduced l-glutathione, guanidine hydrochloride, trypsin, pepsin,
horse heart Mb, and mushroom tyrosinase were obtained from Sigma.
The other reagents were obtained at the best grade available. The con-
centration of hydrogen peroxide solutions was controlled by monitoring
HPLC analysis of amino acid–DA and GSH–DA conjugates: The prod-
uct mixtures derived from the reaction of amino acids (N-acetyl-l-cys-
teine, N-a-acetyl-l-histidine, l-histidine, and N-a-acetyl-l-lysine) and
GSH with the DA–oxidation products generated by LPO/H2O2, Mb/
H2O2, or Ty/O2 were obtained by allowing the reaction to occur for
10 min at room temperature in 200 mm phosphate buffer pH 7.5. The
Chem. Eur. J. 2008, 14, 8661 – 8673
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8671

L. Casella et al.
concentration of the reactants were as follows: 1) using N-acetyl-l-cys-
teine or GSH as nucleophiles, with LPO: [nucleophile]=20 mm, [DA]=
30 mm, [LPO]=8”10ꢀ8 m, [H2O2]=5 mm (divided into 5 aliquots of 1 mm
each, added every 5 min); with Mbs: [nucleophile]=20 mm, [DA]=
30 mm, [Mb]=2”10ꢀ6 m, [H2O2]=5 mm (divided into 5 aliquots of 1 mm
each, added every 5 min); with Ty: [nucleophile]=20 mm, [DA]=30 mm,
[Ty]=1”10ꢀ8 m; 2) in the case of N-a-acetyl-l-histidine, l-histidine, or N-
a-acetyl-l-lysine several runs were performed using DA (30 mm), and
changing the reagent concentrations as follows: [nucleophile]=20–
40 mm, [LPO]=8”10ꢀ8 m, [H2O2]=1–25 mm (divided into 5 aliquots, with
additions every 5 min). The products were analyzed by HPLC by using a
Jasco MD-1510 instrument with spectrophotometric diode array detec-
tion equipped with a Supelcosil LC18 reverse-phase semipreparative
column (5 mm, 250”10 mm). Elution was carried out starting with 0.1%
trifluoroacetic acid (TFA) in water for 5 min, followed by a linear gradi-
ent, in 55 min, to 20% acetonitrile in water containing 0.1% TFA. The
flow rate was 5 mLminꢀ1. Spectrophotometric detection of the eluate
was performed in the range 200–600 nm. The retention times of the prod-
ucts obtained with N-acetyl-l-cysteine and GSH are reported in Ta-
bles S1 and S2 in the Supporting Information, respectively; no derivatiza-
tion of N-a-acetyl-l-histidine, histidine, or N-a-acetyl-l-lysine was ob-
served.
flow rate of 0.2 mLminꢀ1; elution started with 98% solvent A for 5 min,
followed by a linear gradient from 98 to 55% A in 65 min. MS/MS spec-
tra obtained by collision-induced dissociation (CID) were performed
with an isolation width of 2 Th (m/z); the activation amplitude was
around 35% of the ejection radiofrequency (RF) amplitude of the instru-
ment. For the analysis of protein fragments derived from Mb derivatives,
the mass spectrometer was set such that one full MS scan was followed
by using ZoomScan mode and an MS/MS scan on the most intense ion
from the MS spectrum. To identify the modified residues, the acquired
MS/MS spectra were automatically searched against a protein database
for human or horse heart Mb using the SEQUEST algorithm incorporat-
ed into Bioworks 3.1 (ThermoFinnigan, San Jose, CA (USA)).
Guanidine hydrochloride denaturation assay: The stability to denatura-
tion of the modified HMb and hhMb derivatives, which were obtained by
reaction of the proteins (ꢁ5 mm), with 1 mm DA and 0.3 mm H2O2, and,
for comparison purposes, of the unmodified proteins, were determined in
200 mm phosphate buffer (pH 6.0) by monitoring the absorbance varia-
tion of the Soret band of the proteins upon addition of increasing
amounts (up to 1.7m final concentration) of a 8m guanidine hydrochlo-
ride solution in the same buffer. Data were corrected for dilution by the
Gdn–HCl addition. The Gdn–HCl concentration at 50% unfolding of the
proteins and the thermodynamic parameters for denaturation were evalu-
ated from the absorbance versus denaturant concentration curves accord-
ing to a standard method.[81]
MS and NMR spectroscopic characterization of cysteinyl and glutathion-
yl derivatives of DA: The products eluted from HPLC were collected,
taken to dryness, and analyzed by NMR spectroscopy and MS. 1H NMR
spectra of the compounds dissolved in D2O were recorded at 258C on a
Bruker AVANCE 400 spectrometer, operating at 9.37 T; 1H resonances
1
1
were assigned through the H and H-1H DQF-COSY spectra.
Acknowledgements
Electrospray ionization MS spectra were acquired by using an LCQ
ADV MAX ion-trap mass spectrometer equipped with an ESI ion source
and controlled by Xcalibur software 1.3 (Thermo-Finnigan, San Jose,
CA, USA). ESI experiments were carried out in positive-ion mode under
the following constant instrumental conditions: source voltage 5.0 kV, ca-
pillary voltage 46 V, capillary temperature 2108C, tube lens voltage 55 V.
This work was supported by funds from PRIN and FIRB projects of the
Italian MIUR. The University of Pavia (through a FAR project) and
CIRCMSB are also gratefully acknowledged for support.
[1] A. Borta, G. U. Hçglinger, J. Neurochem. 2007, 100, 587–595.
[2] D. Nyholm, Parkinsonism Relat. Disord. 2007, 13, S13–S17.
[3] C. F. Orr, D. B. Rowe, G. M. Halliday, Prog. Neurobiol. 2002, 68,
325–340.
[4] M. G. Tansey, M. K. McCoy, T. C. Frank-Cannon, Exp. Neurol. 2007,
208, 1–25.
[5] J. K. Andersen, Nat. Rev. Neurosci. 2004, 5, S18-S25.
[6] T. L. Perry, D. V. Godin, S. Hansen, Neurosci. Lett. 1982, 33, 305–
310.
[7] T. L. Perry, V. W. Yong, Neurosci. Lett. 1986, 67, 269–274.
[8] R. K. Pearce, A. Owen, S. Daniel, P. Jenner, C. D. Marsden, J.
Neural Transm. 1997, 104, 661–677.
[9] P. Riederer, E. Sofic, W. D. Rausch, B. Schmidt, G. P. Reynolds, K.
Jellinger, M. B. Youdim, J. Neurochem. 1989, 52, 515–520.
[10] E. Sofic, W. Paulus, K. Jellinger, P. Riederer, M. B. Youdim, J. Neu-
rochem. 1991, 56, 978–982.
[11] K. A. Jellinger, E. Kienzl, G. Rumpelmaier, W. Paulus, P. Riederer,
H. Stachelberger, M. B. Youdim, D. Ben-Scachar, Adv. Neurol. 1993,
60, 267–272.
[12] M. Asanuma, I. Miyazaki, N. Ogawa, Neurotoxic. Res. 2003, 5, 165–
176.
[13] M. Akagawa, Y. Ishii, T. Ishii, T. Shibata, M. Yotsu-Yamashita, K.
Suyama, K. Uchida, Biochemistry 2006, 45, 15120–15128.
[14] D. G. Graham, S. M. Tiffany, W. R. Bell, Jr., W. F. Gutknecht, Mol.
Pharmacol. 1978, 14, 644–653.
Modification of the Mbs by quinone species and tandem MS analysis of
protein fragments: Samples of modified HMb and hhMb were prepared
by adding DA (1 mm) and the following reagents to the protein solutions
(6”10ꢀ5 m, in 50 mm phosphate buffer, pH 7.5): a) for the Mb/H2O2 cata-
lytic system: 0.3 mm H2O2; b) for the LPO/H2O2 catalytic system: 8”
10ꢀ8 m LPO, 0.3 mm H2O2; c) for the Ty/O2 catalytic system: 1”10ꢀ8 m Ty.
The solutions of DA, H2O2, and Ty were added in 5 aliquots; after the
last addition, the proteins were allowed to react at room temperature for
10 min.
For the analysis of protein fragments, the modified HMb and hhMb de-
rivatives (and, for comparison purposes, the unmodified proteins) (about
1 mg) were transformed into the apoproteins by the standard hydrochlo-
ric acid/2-butanone method.[80] After overnight dialysis against 20 mm am-
monium bicarbonate buffer at pH 8.0, the melanic precipitates were sepa-
rated by centrifugation from the apoMb derivatives, washed three times
with water (to retain only the covalent melanin–protein conjugates in the
precipitate), acidified to pH 3 with HCl, washed again with water, and
subsequently resuspended in 20 mm ammonium bicarbonate buffer
(pH 8.0) for proteolysis with trypsin, and 20 mm phosphate buffer
(pH 2.5) for proteolysis with pepsin, respectively. All the digestions were
performed at 378C with the protease (trypsin or pepsin) 1:50 (w/w) for
3 h and overnight for the apoproteins and melanic residues, respectively.
The samples of HMb derivatives were further reacted with 5 mm dithio-
threitol for 15 min at 378C prior to the HPLC–MS/MS analysis, since
Cys-containing peptides generally couple during digestion. The reduction
of the disulfide bonds was necessary for the quantification of the unreact-
ed cysteine residues in HMb derivatives.
[15] S. Ito, Pigment Cell Res. 2003, 16, 230–236.
[16] F. Zhang, G. Dryhurst, J. Med. Chem. 1994, 37, 1084–1098.
[17] F. Zhang, G. Dryhurst, Bioorg. Chem. 1995, 23, 193–216.
[18] X.-M. Shen, G. Dryhurst, Chem. Res. Toxicol. 1996, 9, 751–763.
[19] H. Li, G. Dryhurst, J. Neurochem. 1997, 69, 1530–1541.
[20] R. Xu, X. Huang, K. J. Kramer, M. D. Hawley, Bioorg. Chem. 1996,
24, 110–126.
[21] S. Ito, G. Prota, Experientia 1977, 33, 1118–1119.
[22] S. Ito, K. Fujita, Biochim. Biophys. Acta Gen. Subj. 1981, 672, 151–
157.
LC–MS and LC–MS/MS data were obtained by using the LCQ ADV
MAX ion-trap mass spectrometer with the instrumental conditions re-
ported in the previous section. The system was run in automated LC–
MS/MS mode and using a Surveyor HPLC system (Thermo Finnigan,
San Jose, CA, USA) equipped with a BioBasic C18 column (5 mm, 150”
2.1 mm). The elution was performed by using 0.1% HCOOH in distilled
water (solvent A) and 0.1% HCOOH in acetonitrile (solvent B), with a
8672
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8661 – 8673

Myoglobin Modification
FULL PAPER
[23] J. P. E. Spencer, P. Jenner, S. E. Daniel, A. J. Lees, D. C. Marsden, B.
Halliwell, J. Neurochem. 1998, 71, 2112–2122.
[24] A. Palumbo, M. dꢀIschia, G. Misuraca, L. De Martino, G. Prota, Bio-
chim. Biophys. Acta Gen. Subj. 1995, 1245, 255–261.
[25] A. H. Stokes, T. G. Hastings, K. E. Vrana, J. Neurosci. Res. 1999, 55,
659–665.
[26] Y. Xu, A. H. Stokes, R. Roskoski, Jr., K. E. Vrana, J. Neurosci. Res.
1998, 54, 691–697.
[27] M. J. LaVoie, B. L. Ostaszewski, A. Weihofen, M. G. Schlossmacher,
D. J. Selkoe, Nat. Med. 2005, 11, 1214–1221.
[55] J. C. Espin, R. Varon, L. G. Fenoll, M. A. Gilabert, P. A. Garcia-
Ruiz, J. Tudela, F. Garcia-Canovas, Eur. J. Biochem. 2000, 267,
1270–1279.
[56] A. Granata, E. Monzani, L. Bubacco, L. Casella, Chem. Eur. J.
2006, 12, 2504–2514.
[57] N.-E. Andꢂn, K. Fuxe, B. Hamberger, T. Hçkfelt, Acta Physiol.
Scand. 1966, 67, 306–312.
[58] A. van der Vliet, J. P. Eiserich, B. Halliwell, C. E. Cross, J. Biol.
Chem. 1997, 272, 7617–7625.
[59] A. Napolitano, O. Crescenzi, A. Pezzella, G. Prota, J. Med. Chem.
1995, 38, 917–922.
[60] H. B. Dunford, Heme Peroxidases, Wiley-VCH, New York, USA,
1999.
[28] M. J. LaVoie, T. G. Hastings, J. Neurosci. 1999, 19, 1484–1491.
[29] T. G. Hastings, D. A. Lewis, M. J. Zigmond, Proc. Natl. Acad. Sci.
USA 1996, 93, 1956–1961.
[30] D. M. Kuhn, R. E. Arthur, Jr., D. M. Thomas, L. A. Elferink, J. Neu-
rochem. 1999, 73, 1309–1317.
[61] X. Huang, R. Xu, M. D. Hawley, T. L. Hopkins, K. J. Kramer, Arch.
Biochem. Biophys. 1998, 352, 19–30.
[31] D. M. Kuhn, R. Arthur, Jr., J. Neurosci. 1998, 18, 7111–7117.
[32] S. B. Berman, T. G. Hastings, J. Neurochem. 1997, 69, 1185–1195.
[33] S. B. Berman, M. J. Zigmond, T. G. Hastings, J. Neurochem. 1996,
67, 593–600.
[62] G. N. L. Jameson, J. Zhang, R. F. Jameson, W. Linert, Org. Biomol.
Chem. 2004, 2, 777–782.
[63] S. P. Raps, J. C. K. Lai, L. Hertz, A. J. L. Cooper, Brain Res. 1989,
493, 398–401.
[34] K. A. Conway, J.-C. Rochet, R. M. Bieganski, P. T. Lansbury, Jr., Sci-
ence 2001, 294, 1346–1349.
[64] J. Sian, D. T. Dexter, P. Jenner, D. C. Marsden, Br. J. Pharmacol.
1992, 107, 428P.
[35] E. Bossy-Wetzel, R. Schwarzenbacher, S. A. Lipton, Nat. Med. 2004,
10, S2–S9.
[65] E. Rosengren, E. Linder-Eliasson, A. Carlsson, J. Neural Transm.
1985, 63, 247–253.
[36] D. C. S. Tse, R. L. McCreery, R. N. Adams, J. Med. Chem. 1976, 19,
37–40.
[66] B. Fornstedt, E. Rosengren, A. Carlsson, Neuropharmacology 1986,
25, 451–454.
[37] T. Kato, S. Ito, K. Fujita, Biochim. Biophys. Acta Gen. Subj. 1986,
881, 415–421.
[38] S. Ito, T. Kato, K. Fujita, Biochem. Pharmacol. 1988, 37, 1707–1710.
[39] R. E. Whitehead, J. V. Ferrer, J. A. Javitch, J. B. Justice, J. Neuro-
chem. 2001, 76, 1242–1251.
[40] M. Y. Moridani, H. Scobie, P. Salehi, P. J. OꢀBrien, Chem. Res. Toxi-
col. 2001, 14, 841–848.
[41] J. R. Jara, P. Aroca, F. Solano, J. H. Martinez, J. A. Lozano, Biochim.
Biophys. Acta Gen. Subj. 1988, 967, 296–303.
[67] M. K. Gaitonde, Biochem. J. 1967, 104, 627–633.
[68] X.-M. Shen, G. Dryhurst, Tetrahedron 2001, 57, 393–405.
[69] S. R. Hubbard, W. A. Hendrickson, D. G. Lambright, S. G. Boxer, J.
Mol. Biol. 1990, 213, 215–218.
[70] P. K. Witting, D. J. Douglas, A. G. Mauk, J. Biol. Chem. 2001, 276,
3991–3998.
[71] H. Anni, T. Yonetani, Met. Ions Biol. Syst. 1992, 28, 219–241.
[72] P. J. Ohlsson, T. Yonetani, S. Wold, Biochim. Biophys. Acta Protein
Struct. Mol. Enzymol. 1986, 874, 160–166.
[42] M. Brunori, D. Bourgeois, B. Vallone, J. Struct. Biol. 2004, 147, 223–
234.
[43] S. Nicolis, E. Monzani, R. Roncone, L. Gianelli, L. Casella, Chem.
Eur. J. 2004, 10, 2281–2290.
[44] S. Nicolis, A. Pennati, E. Perani, E. Monzani, A. M. Sanangelantoni,
L. Casella, Chem. Eur. J. 2006, 12, 749–757.
[73] E. Monzani, G. Alzuet, L. Casella, C. Redaelli, C. Bassani, A. M.
Sanangelantoni, M. Gullotti, L. De Gioia, L. Santagostini, F. Chille-
mi, Biochemistry 2000, 39, 9571–9582.
[74] F. A. Zucca, G. Giaveri, M. Gallorini, A. Albertini, M. Toscani, G.
Pezzoli, R. Lucius, H. Wilms, D. Sulzer, S. Ito, K. Wakamatsu, L.
Zecca, Pigment Cell Res. 2004, 17, 610–617.
[45] C. Redaelli, E. Monzani, L. Santagostini, L. Casella, A. M. Sanange-
lantoni, R. Pierattelli, L. Banci, ChemBioChem 2002, 3, 226–233.
[46] P. K. Witting, D. J. Douglas, A. G. Mauk, J. Biol. Chem. 2000, 275,
20391–20398.
[47] H. Ozeki, S. Ito, K. Wakamatsu, I. Ishiguro, Biochim. Biophys. Acta
Gen. Subj. 1997, 1336, 539–548.
[75] T. Burmester, B. Weich, S. Reinhardt, T. Hankeln, Nature 2000, 407,
520–523.
[76] L. Casella, L. De Gioia, G. Frontoso Silvestri, E. Monzani, C. Re-
daelli, R. Roncone, L. Santagostini, J. Inorg. Biochem. 2000, 79, 31–
39.
[77] E. Monzani, A. L. Gatti, A. Profumo, L. Casella, M. Gullotti, Bio-
chemistry 1997, 36, 1918–1926.
[78] J. R. Ros, J. N. Rodrꢃguez-Lꢄpez, F. Garcꢃa-Cꢅnovas, Biochim. Bio-
phys. Acta Protein Struct. Mol. Enzymol. 1994, 1204, 33–42.
[79] R. L. Vold, J. S. Waugh, M. P. Klein, D. E. Phelps, J. Chem. Phys.
1968, 48, 3831–3832.
[48] D. Metodiewa, K. Reszka, H. B. Dunford, Arch. Biochem. Biophys.
1989, 274, 601–608.
[49] E. Ghibaudi, E. Laurenti, C. Pacchiardo, G. Suriano, N. Moguilev-
sky, R. P. Ferrari, J. Inorg. Biochem. 2003, 94, 146–154.
[50] S. Ito, A. Palumbo, G. Prota, Experientia 1985, 41, 960–961.
[51] F. Zhang, G. Dryhurst, J. Electroanal. Chem. 1995, 398, 117–128.
[52] R. Xu, X. Huang, T. D. Morgan, O. Prakash, Arch. Biochem. Bio-
phys. 1996, 329, 56–64.
[80] E. Antonini, M. Brunori, Hemoglobin and Myoglobin in Their Reac-
tions with Ligands, North-Holland Publishing Company, Amster-
dam, The Netherlands, 1971.
[53] X. Huang, R. Xu, M. D. Hawley, K. J. Kramer, Bioorg. Chem. 1997,
25, 179–202.
[81] R. Roncone, E. Monzani, S. Labꢆ, A. M. Sanangelantoni, L. Casella,
J. Biol. Inorg. Chem. 2005, 10, 11–24.
[54] E. Monzani, R. Roncone, L. Casella, M. Galliano, W. H. Koppenol,
Eur. J. Biochem. 2004, 271, 895–906.
Received: May 27, 2008
Published online: August 7, 2008
Chem. Eur. J. 2008, 14, 8661 – 8673
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8673