Here we report that the concerted action of a nucleoside triphos-phate and Mg2+ on rapeseed 2-Cys Prx reversibly impairs the peroxidase activity and promotes the formation of high molecu
Trang 1of rapeseed 2-Cys peroxiredoxin
Martin Aran1, Daniel Caporaletti1, Alejandro M Senn1, Marı´a T Tellez de In˜on2, Marı´a R Girotti1, Andrea S Llera1and Ricardo A Wolosiuk1
1 Instituto Leloir, IIBBA-CONICET, Universidad de Buenos Aires, Argentina
2 INGEBI-CONICET, Buenos Aires, Argentina
Rather than viewing reactive oxygen species (ROS) as
toxic by-products of aerobic metabolism we now
know them to be members of signaling networks
that modulate important physiological processes [1,2]
Germane to the homeostatic regulation of ROS
con-centrations, a large group of peroxidases devoid of
selenium- and heme-prosthetic groups, the
peroxi-redoxins (Prx) (EC 1.11.1.15), catalyze the reduction
of hydroperoxides and peroxinitrite [3–6] The number
of subfamilies in this ubiquitous family of proteins
varies depending on the classification criteria used
but, in all cases, one subfamily encompasses
polypep-tides in which there is strict conservation of two
cyste-ine residues – the 2-Cys Prx [7–9] The typical 2-Cys
Prx, found in prokaryotes and eukaryotes, is a head-to-tail arranged homodimer in which one of the con-served cysteines of the polypeptide is linked via an intercatenary disulfide bond to the complementary cysteine of the paired subunit Crucial to the peroxi-dase activity is the cysteine residue located at the N-terminal region, ‘the peroxidatic cysteine’, whose sulfhydryl group (-Cys-SH) turns into sulfenic acid (-Cys-SOH) after reacting with a hydroperoxide (ROOH) The sulfenate subsequently reacts with the cysteine located in the C-terminal region of the paired polypeptide, ‘the resolving cysteine’, forming a second intermolecular disulfide linkage (-Cys-S-S-Cys-) Com-pleting the peroxidatic cycle, one of the disulfides is
Keywords
2-Cys peroxiredoxin; ATP binding;
autophosphorylation; sulfinic-phosphoryl
anhydride; sulfonic-phosphoryl anhydride
Correspondence
R A Wolosiuk, Instituto Leloir, Patricias
Argentinas 435, C1405BWE Buenos Aires,
Argentina
Fax: +54 11 5238 7501
Tel: +54 11 5238 7500
E-mail: rwolosiuk@leloir.org.ar
(Received 12 November 2007, revised
14 January 2008, accepted 16 January
2008)
doi:10.1111/j.1742-4658.2008.06299.x
2-Cys peroxiredoxins (2-Cys Prx) are ubiquitous thiol-containing
peroxidas-es that have been implicated in antioxidant defense and signal transduction Although their biochemical features have been extensively studied, little is known about the mechanisms that link the redox activity and non-redox processes Here we report that the concerted action of a nucleoside triphos-phate and Mg2+ on rapeseed 2-Cys Prx reversibly impairs the peroxidase activity and promotes the formation of high molecular mass species Using protein intrinsic fluorescence in the analysis of site-directed mutants, we demonstrate that ATP quenches the emission intensity of Trp179, a residue close to the conserved Cys175 More importantly, we found that ATP facilitates the autophosphorylation of 2-Cys Prx when the protein is succes-sively reduced with thiol-bearing compounds and oxidized with hydroper-oxides or quinones MS analyses reveal that 2-Cys Prx incorporates the phosphoryl group into the Cys175 residue yielding the sulfinic-phosphoryl [Prx-(Cys175)-SO2PO3 )] and the sulfonic-phosphoryl [Prx-(Cys175)-SO3
PO3 )] anhydrides Hence, the functional coupling between ATP and 2-Cys Prx gives novel insights into not only the removal of reactive oxygen species, but also mechanisms that link the energy status of the cell and the oxidation of cysteine residues
Abbreviations
2-Cys Prx, 2-Cys peroxiredoxin; ANS, 8-anilinonaphthalene-1-sulfonate; ROS, reactive oxygen species.
Trang 2cleaved back to thiols by the concerted action of
cel-lular reductants and protein-disulfide oxido-reductases
Often regarded as a peroxidase, the additional
func-tion as a molecular chaperone, identified in human
and yeast 2-Cys Prx [10,11], exhibits a marked
sensi-tivity to a variety of compounds and conditions, such
as reductants and ROS An imbalance between these
dual functions is probably associated with many
human pathologies, such as thyroid tumors, breast
and lung cancer, Alzheimer’s disease and
neurodegen-erative disorders [12]
By contrast to the extensive literature on the redox
modulation of different enzyme activities in
chloro-plasts and non-photosynthetic systems, mainly via
thiol–disulfide exchange [13–16], systematic efforts to
examine the opposite action of non-redox processes on
redox reactions are scarce Most biochemical research
on 2-Cys Prx has studied the interplay between the
stimuli and the abundance of reductants and oxidants
(thermodynamic control), whereas the catalytic
fea-tures (kinetic control) have been less clear Therefore,
the reactions leading to ROS generation and
detoxifi-cation have been elucidated, but little is known about
how oxidative stress is linked to non-redox processes
in the signaling networks that modulate cellular
func-tions Studies addressing this issue have found
signi-ficant changes in the quaternary structure and dual
functions when human 2-Cys Prx is phosphorylated on
Thr90 by cyclin-dependent protein kinases, preferably
CDK1 (formerly Cdc2) [17,18] A putative
intermedi-ate at the peroxidatic cysteine (-Cys-S(=O)-O-PO3 ))
has recently been suggested in the multiple-step
pro-cess underlying sulfiredoxin-mediated reduction of
2-Cys Prx-SO2H, however, experimental evidence is
lack-ing [19–21] As with many proteins, phosphorylation
of 2-Cys Prx via these two mechanisms requires the
participation of additional catalysts, i.e protein kinase
and sulfiredoxin Despite numerous studies showing
the close association between ATP and chaperone
activity [22], with the exception of serving as the
phos-phoryl donor for CDK1 and sulfiredoxin, the direct
interaction of a nucleotide with 2-Cys Prx has not been
previously addressed Here, we report that the
con-certed action of a nucleoside triphosphate and Mg2+
on rapeseed 2-Cys Prx impairs the peroxidase activity
More importantly, MS studies show that the successive
action of a reductant and an oxidant makes the
pro-tein a recipient of the phosphoryl moiety in sulfonic
and sulfinic acid forms of Cys175 Hence, ATP has
emerged as both a novel regulator of 2-Cys Prx
func-tions and the phosphoryl donor for the
autophospho-rylation at the resolving cysteine
Results
The concerted action of nucleotides and Mg2+ modulates the peroxidase activity of 2-Cys Prx
A thorough inspection of the available X-ray structure
of human 2-Cys Prx (PDB entry: 1QMV) revealed that the covalently linked homodimer creates a large inter-nal cavity comprising segments of two polypeptides (-Leu42–Pro46, Arg129–Ile133, Gln141–Asn146, Gly151–Arg159-) and (-Pro54–Ile57, Cys175–Gly178-) Because the size of the cavity (0.515 nm3) in silico eas-ily docks a molecule of ATP (0.33 nm3), it was impor-tant to know whether the functions of 2-Cys Prx were sensitive to insertion of the nucleotide As shown in Fig 1A, a nucleoside triphosphate in concert with
Mg2+lowered the peroxidase activity in a dose-depen-dent manner, purine nucleotides being more potent than pyrimidine derivatives In particular, the response
of the peroxidase activity to increasing concentrations
of ATP exhibited three well-defined stages: (a) monot-onous decay (I0.5= 0.25 mm), (b) stabilization at half
of the maximal activity from 0.9 to 1.2 mm, and (c) a sharp decrease to undetectable levels beyond 1.5 mm (I0.5= 1.40 mm) Interestingly, inhibition mediated by the other purine nucleotide, GTP, was significantly similar in the first two stages, but lacked the third Following these initial experiments, we investigated whether other phosphorylated compounds and bivalent cations exhibited similar capacity In the presence of
2 mm Mg2+, the rate of H2O2 removal was inhibited
by 60, 5 and 5% when it was assayed with 2 mm ADP, AMP or orthophosphate, respectively (not shown) By contrast, Mg2+ was the most efficient cation in assisting nucleotide-dependent inhibition (100%), whereas Ca2+ (92%), Mn2+ (85%) and Zn2+ (65%) exhibited a varying capacity when the peroxi-dase activity was assayed in the presence of fixed con-centrations of both ATP (3 mm) and bivalent cations (3 mm), (Fig 1B) At this stage, the lack of peroxidase activity might be attributed to an irreversible change in 2-Cys Prx triggered by the binding of ligands Against this possibility, totally inactive 2-Cys Prx, caused by incubation with 3 mm ATP and 3 mm Mg2+, immedi-ately recovered 75% of its original activity upon chela-tion of Mg2+ by the addition of 5 mm EGTA (Fig 1C) Clearly, the capacity of ATP to lower the peroxidase activity in the absence of exogenous com-ponents revealed a mechanism that is substantially different from the inhibition brought about by the phosphorylation of human 2-Cys Prx mediated by the CDK1–cyclin B complex [17]
Trang 3In addition to the peroxidase activity that functions
in the cellular defense against ROS, yeast and human
2-Cys Prx are molecular chaperones [10,11] Therefore,
it was important to examine 2-Cys Prx beyond a single
activity and establish whether the regulation described
above previously had wider implications We found
that the rapeseed orthologue efficiently prevents the
thermal aggregation of citrate synthase indicating that
the chaperone activity is likely to be a general function
of typical 2-Cys Prx (Fig 1D) Remarkably,
incorpo-ration of increasing amounts of Mg2+into the
incuba-tion milieu led to a concomitant reducincuba-tion in the
chaperone capacity which, at variance with the
peroxi-dase activity, was not affected by the presence of
2.5 mm ATP These data provide the first evidence
that the dual functions of 2-Cys Prx can be differen-tially regulated by ATP and Mg2+
The interaction with ATP modifies structural features of 2-Cys Prx
Given the essential role of ATP in the peroxidase activity, we evaluated changes in the structure of 2-Cys Prx brought about by the concerted action of the nucleotide and the bivalent cation To accurately determine the molecular mass of our 2-Cys Prx prepa-rations, static light-scattering measurements were per-formed, because this spectroscopic technique allows direct estimation of the species in solution [23] In the absence of perturbants, the predominant form of
A
C
B
D
Fig 1 Effect of nucleotides ⁄ Me 2+ on the functions of 2-Cys Prx (A) Concerted action of nucleotides ⁄ Mg 2+ on the peroxidase activity The reaction, carried out in a solution supplemented with 3 l M 2-Cys Prx, 2 m M MgCl2and the indicated nucleoside triphosphate, was started by the addition of 0.13 m M H 2 O 2 and the remnant of reduced dithiothreitol was estimated with Ellman’s reagent after 15 min Data from seven independent experiments were averaged and standard deviations were calculated Control activity: 4.3 nmol H2O2 reduced per min (B) Effect of ATP ⁄ Me 2+ on peroxidase activity The assay was essentially similar to (A), except that the concentrations of ATP and the Me 2+
(Mg 2+ , Ca 2+ , Mn 2+ , Zn 2+ ) were both fixed at 3 m M (C) EGTA-mediated reversal of (ATP ⁄ Mg 2+ )-dependent inhibition of peroxidase activity 2-Cys Prx (3 l M ) was incubated for 3 min with 2 m M ATP and 2 m M MgCl 2 After the addition of EGTA to a final concentration of 5 m M , the protein solution was further incubated for 5 min and the peroxidase activity was assayed as in (A) (D) Effect of ATP and Mg 2+ on the chap-erone activity 2-Cys Prx (5 l M ) was incubated in 25 m M Tris ⁄ HCl (pH 8) containing, as indicated, different concentrations of MgCl 2 and 2.5 m M ATP After 10 min at 25 C followed by 10 min at 45 C, the assay was started by the addition of citrate synthase and measured as described in Experimental procedures [25].
Trang 4rapeseed 2-Cys Prx (polypeptide: 22 316 kDa) had a
molecular mass of 260 kDa indicating that it was
essentially similar to counterparts from other sources,
wherein covalently linked dimers (a2) associate
non-covalently, forming doughnut-shaped decamers (a2)5
(not shown) [24,25] By contrast, size-exclusion
chro-matography revealed that protein dissolved in
solu-tions containing 3 mm ATP and 3 mm Mg2+exhibited
a 2550 kDa higher-order assembly that returned
imme-diately to the decameric state upon the removal of
ATP or Mg2+ It is noteworthy that low
concentra-tions of two well-known intracellular components
converted the rather stable decamer to higher-order
assemblies whose molecular mass approached that of
the dodecahedron [(a2)5]12observed recently in electron
microscopy preparations of the erythrocyte orthologue
treated with polyethylene glycol [26]
Data on the inhibition of peroxidase activity along
with the reversible oligomerization of 2-Cys Prx were
consistent with a specific binding of ATP⁄ Mg to the
protein In line with this prediction, positive and
nega-tive differences in absorbance appeared following
incu-bation of 2-Cys Prx with ATP in the absence and
presence of Mg2+, respectively (supplementary
Fig S1) Although these experiments confirmed an
interaction between the nucleotide⁄ Me2+ couple and
the protein, the differential response could not be
attributed specifically to any of the interacting species
Therefore, we turned our attention to fluorescence
emission spectroscopy which provides information
about the polarity of local environments surrounding
either extrinsic probes that bind to proteins or intrinsic
fluorophores buried in the protein interior In a first
set of experiments, we relied on a biophysical probe commonly used to study the characteristics of protein surfaces, 8-anilinonaphthalene-1-sulfonate (ANS), which, as expected, exhibited an emission maximum wavelength at 512 nm that was not modified by the presence of 3 mm ATP or 3 mm Mg2+ (Fig 2A) At variance, reflecting the affinity of this extrinsic probe towards exposed protein hydrophobic surfaces, 2-Cys Prx led to a marked enhancement of the emis-sion intensity with a concurrent displacement of the
A
C
B
Fig 2 Effect of the binding of ATP ⁄ Mg 2+ to 2-Cys Prx on the
fluo-rescence emission of extrinsic and intrinsic chromophores (A)
Sen-sitivity of the extrinsic probe ANS Binding of ANS to 2-Cys Prx
was performed for 10 min at 25 C in solutions of 20 m M Tris ⁄ HCl
(pH 7.8) containing 75 l M ANS (e 350nm : 5000 M )1Æcm)1), and, as
indicated, 10 l M 2-Cys Prx, 3 m M ATP and 3 m M Mg 2+ Protein
solutions were excited at 370 nm and emission spectra were
scanned from 410 to 600 nm The spectral bandwidths were 5 nm.
(B) Quenching of tryptophan fluorescence Equilibrium fluorescence
measurements were conducted increasing the concentrations of
ATP or ADP, as indicated, while keeping constant the concentration
of 2-Cys Prx (2 l M ) and Mg2+(2 m M ) After correction for the inner
filter effect, data were fitted to the saturation curve equation using
nonlinear least-squares regression analyses The difference in
fluo-rescence (DF) between 2-Cys Prx (F o ) and 2-Cys Prx-ATP-Mg2+
complex (F) at 340 nm was plotted according to Lehrer [28] (inset).
(C) Quenching of emission intensity in W88F and W179F 2-Cys Prx.
Fluorescence measurements were performed as described in (B),
except that W88F and W179F mutants replaced for the wild-type
2-Cys Prx.
Trang 5spectrum to a maximum at 480 nm At this stage, the
addition of 3 mm ATP and 3 mm Mg2+ did not shift
the maximum emission wavelength, but progressively
increased the emission intensity, indicating that the
nucleotide and the bivalent cation significantly
enhanced the proportion of protein hydrophobic
patches
Although these experiments were informative
regard-ing the ability of 2-Cys Prx to interact with ATP, it
was imperative to determine the nucleotide binding
site This information could be gained from the
intrin-sic fluorescence because the constituent polypeptide
held two conserved tryptophan residues that exhibited
a maximum emission wavelength centered at 343 nm,
suggesting a rather polar environment around the
in-dol side chains (supplementary Fig S2A) [27]
Unfor-tunately, the concentration of nucleotides in these
experiments never exceeded 0.2 mm because the intense
inner filter effect caused by the purine ring impaired
the excitation of tryptophan residues Despite this
limi-tation, if ATP perturbs the environment of Trp88 or
Trp179 to some extent, the fluorescence emission
should show a shift in maximum wavelength or a
decrease in intensity when the nucleotide changes the
conformation of the protein or collides with the
fluoro-phore, respectively Incorporation of ATP⁄ Mg did not
shift the maximum wavelength but caused a marked
quenching of the emission intensity that was much less
pronounced with ADP⁄ Mg (Fig 2B) Stern–Volmer
analyses showed a pronounced downward curvature as
result of a heterogeneous response of intrinsic
fluoro-phores towards the quencher In this context, if the
deviation from linearity reflected a fluorophore
inac-cessible to the nucleotide, the Stern–Volmer relationship
should become linear using the expression developed by
Lehrer [Fo⁄ (Fo)F) = 1 ⁄ fa+ (1⁄ {faÆKSVÆ[Q]})] [28,29]
As shown in Fig 2B (inset), the straight line was
con-gruent with a unique tryptophan residue of 2-Cys Prx
accessible to ATP⁄ Mg (fa= 0.26; KSV= 9.7· 10)3Æ
m)1) To unambiguously define the indol ring sensitive
to ATP⁄ Mg, we examined the intrinsic emission
fluo-rescence in variants of 2-Cys Prx where Trp88 and
Trp179 were replaced conservatively by phenylalanine
via site-directed mutagenesis The results in Fig 2C
clearly illustrate that the marked reduction in emission
intensity caused by the quencher in W88F 2-Cys Prx
was similar to its wild-type counterpart, whereas the
W179F variant was insensitive to ATP⁄ Mg These
findings demonstrated that the ATP binds to 2-Cys
Prx close to Trp179 and, as a consequence, to the
resolving Cys175 In this study, two complementary
experiments indicated that the conservative
replace-ment of tryptophan residues did not lead to gross
alterations in the structure of 2-Cys Prx First, the emission spectrum of W88F 2-Cys Prx was similar to its wild-type counterpart (kmax= 343 nm), whereas that of the W179F variant was slightly blue-shifted (kmax= 338 nm) (supplementary Fig S2A) Second, the catalytic capacity was not affected because neither the basal nor the ATP-inhibited peroxidase activities were appreciably different from wild-type 2-Cys Prx (Fig S2B)
The sequential action of reductants and oxidants predisposes 2-Cys Prx to autophosphorylation
In considering whether the interaction with ATP⁄ Mg proceeded further to the specific phosphorylation of 2-Cys Prx, we noted that ten serine, one threonine and two tyrosine residues appeared as putative sites for phosphorylation (program netphos2.0, Expassy) Therefore, we conducted a phosphorylation assay in which our preparation of rapeseed 2-Cys Prx was first treated with reductants and oxidants, then sub-sequently incubated with [c32P]ATP⁄ Mg2+ and finally subjected to non-reducing SDS⁄ PAGE (Fig 3A)
In this successive in vitro reductionfi oxidation of 2-Cys Prx, we were compelled to use high concentra-tions of cumene hydroperoxide in the oxidation step because high concentrations of dithiothreitol, required for the complete and fast cleavage of disulfide bonds, remained in the solution To our surprise, a 23 kDa-labeled band appeared when the recombinant protein was (a) incubated successively with 10 mm dithiothrei-tol, 10 mm cumene hydroperoxide and [c32P]ATP⁄
Mg2+, (b) subjected to non-reducing SDS⁄ PAGE, and (c) characterized by Ponceau Red staining and autora-diography (Fig 3B) Although not shown, four control experiments carried out under comparable conditions were consistent with the specific covalent binding of the phosphoryl moiety to 2-Cys Prx First, 32P-labeled bands did not appear in the autoradiography when chloroplast thioredoxin-m, chloroplast fructose-1,6-bisphosphatase or a-lactalbumin were used in place
of 2-Cys Prx Second, the autophosphorylation of 2-Cys Prx could not be attributed to artifacts linked to the unspecific binding of the nucleotide, as neither the presence of ADP, AMP or GTP, nor pulse and chase experiments with 3 mm nonradioactive ATP affected the incorporation of the 32P-label into the protein Third, supporting the formation of a covalent link as opposed to a protein highly resistant to SDS denatur-ation [30], the radioactive label remained linked to 2-Cys Prx after boiling or digestion with trypsin but was completely stripped from the protein by incubation with alkaline phosphatase Fourth, the requirement for
Trang 6MgCl2 was neither replaced nor affected by CaCl2 or
MnCl2 Moreover, the requirement for the sequence
reductionfi oxidation was further supported by
experiments in which a compound generally used for
cleaving disulfide bonds (2-mercaptoethanol) partially
substituted for dithiothreitol and hydroperoxides
(H2O2, t-butyl hydroperoxide) and two structurally dif-ferent quinones (2-hydroxy-1,4-naphthoquinone and 1,4-dihydroxyanthraquinone) were as efficient as cum-ene hydroperoxide (supplementary Fig S3)
The above quenching of Trp179 fluorescence by ATP was of particular interest in characterizing the autophosphorylation because the evolutionary conser-vation of this residue in the 2-Cys Prx subfamily is unknown We therefore examined the ability of W88F and W179F 2-Cys Prx to incorporate the 32P-label after successive incubations with dithiothreitol and cumene hydroperoxide As shown in Fig 3C, the for-mer variant was indistinguishable from wild-type 2-Cys Prx, whereas the latter was not functional Near Trp179, the resolving cysteine is an additional con-served residue that can be predicted to interact with ATP Supporting this view, we estimated in modeling work on 2-Cys Prx that the nitrogen atom in the indol ring of Trp179 is located 1.571 and 0.401 nm from the sulfur atoms of the peroxidatic and resolving cysteines, respectively [31] Taken together, the close proximity
to Trp179 and the requirement for sequential reduc-tionfi oxidation raised the possibility that Cys175 was actively involved in incorporation of the phos-phoryl moiety Consistent with this, Fig 3C shows that a serine in place of Cys53 and Cys175 retained and abrogated, respectively, the ability to incorporate the 32P-label into 2-Cys Prx Notably, this active par-ticipation of the resolving cysteine in the autophospho-rylation uncovered a new function that departed markedly from the known role in the peroxidase activity
Surprisingly, autophosphorylation of C53S 2-Cys Prx did not require successive incubation with dithiothreitol and the hydroperoxide but it was extre-mely sensitive to the addition of dithiothreitol (Fig 3D) Given that the sulfur atom in the cysteine residues of proteins can adopt various oxidation numbers, our preparations of C53S 2-Cys Prx may have contained some proportion of spontaneously oxidized Cys175 To evaluate this possibility, C53S 2-Cys Prx was digested with trypsin and the peptides were examined by MS A peak at m⁄ z 2800.36 exhibited the expected mass of the intrapeptide span-ning from residue 160 to residue 184 [-RflT160 LQAL-QYVQENPDEVCPAGWKPGEK184flS-], wherein the sulfur atom of Cys175 was totally reduced (Fig 4)
Of note, the presence of the sulfhydryl group at Cys175 was confirmed in parallel experiments in which MS studies were conducted with the adduct formed between C53S 2-Cys Prx and iodoacetate (not shown) Because the less intense peak at
m⁄ z 2832.36 was consistent with the addition of two
A
B
Fig 3 Autophosphorylation of 2-Cys Prx (A) Experimental outline.
(B) Requirement of reductants and oxidants 2-Cys Prx was (a)
incubated successively with 10 m M dithiothreitol (DTT), 10 m M
cumene hydroperoxide (CuOOH) and [c 32 P]ATP, (b) subjected to
non-reducing SDS⁄ PAGE, and (c) transferred to nitrocellulose
membranes for protein estimation and autoradiography, as
described in Experimental procedures (C) Role of conserved
tryptophan and cysteine residues W88F, W179F, C53S and
C175S 2-Cys Prx were incubated, as indicated, with 10 m M
dithiothreitol, 10 m M cumene hydroperoxide and [c 32 P]ATP prior to
non-reducing SDS ⁄ PAGE and autoradiography, as outlined in (A).
(D) Autophosphorylation of C53S 2-Cys Prx C53S 2-Cys Prx was
incubated for 10 min only in the presence and absence of 10 m M
dithiothreitol prior to the addition of [c 32 P]ATP, non-reducing
SDS ⁄ PAGE and autoradiography.
Trang 7oxygen atoms to the respective 160–184 tryptic
pep-tide, we further analyzed the sequence of informative
ions to confirm the presence of a sulfinic group at
Cys175 Accordingly, fragment ions from y1 to y9
showed the expected mass for residues spanning
from Lys184 to Pro176, whereas trapped ions
beyond y10 exhibited a mass shift of 32 The
unequivocal assignment of two oxygen atoms to
the Cys175 residue of C53S 2-Cys Prx revealed the
unsuspected formation of oxyacid groups at sulfur
atoms of the resolving cysteine
ATP phosphorylates the sulfinic and sulfonic
forms of the Cys175 residue
Although the Tyr166 residue in the 160–184 peptide
appeared in silico to be one site for the incorporation
of a phosphoryl group (netphos 2.0, Expassy), we
found in separate experiments that
autophosphoryla-tion of the Y166F mutant was similar to wild-type
2-Cys Prx Given the absence of another putative site,
we approached the localization of the phosphoryl
moi-ety by examining the mass spectra of proteolytic digests
obtained from 2-Cys Prx treated successively with
dithiothreitolfi cumene hydroperoxidefi ATP To
locate any modification in the sequence of the 160–184 peptide, we relied on not only the difference in masses (m⁄ z value), but also the y-series, the complementary b-series and the coincidence of stretches assigned to identical signals from different experiments In these analyses, the peaks at m⁄ z 2800.36 and 2832.36 reflected the expected mass of the 160–184 peptide holding at Cys175 a sulfhydryl group and two addi-tional oxygen atoms, respectively (Fig 5A) As illus-trated for the former signal and in consonance with above results (see Fig 4), the y- and b-ion series obtained for selected trapped ions confirmed that the sulfur atom of the resolving Cys175 bore a sulfhydryl group From the repertoire of less intense signals but with low noise levels, two novel species at m⁄ z 2934.36 and 2950.35 were particularly attractive because the masses matched the monosodium adducts [M + Na]+
of the phosphorylated 160–184 peptide bearing sulfinic and sulfonic groups, respectively (Fig 5B) [32–34] As illustrated for the latter signal, sequence informative y-ions from m⁄ z 0 to 970 were identical to those obtained in the spectra of m⁄ z 2800.36 (Fig 5A) and 2832.36 (see Fig 4), thus proving that they originated from the 160–184 peptide But more importantly, the absence of ions from y10 to y19 and the presence of
Fig 4 MS ⁄ MS spectra of the 160–184 tryptic peptide from C53S 2-Cys Prx Expanded view of peaks at m ⁄ z 2800.36 and 2832.35 and the fragmentation of the peak at m ⁄ z 2832.35 2-Cys Prx was digested with trypsin and prepared for MALDI-TOF MS as described in Experi-mental procedures Data were first collected, smoothed and calculated the centroid using the software FLEXANALYSIS , and then plotted in GRAPHPAD PRISM All labeled peaks were at least three times above background The amino acid sequence of the 160–184 tryptic peptide bearing the sulfinic group is displayed above the spectrum The fragmentation patterns that generate ions y and b are illustrated along the peptide sequence wherein (*) are fragment ions bearing –SO 2 H.
Trang 8shifted ions from y10* to y17* revealed that the
sul-fonic form of Cys175 held the monosodium adduct of
one phosphoryl group (-SO3PO3 )) thereby providing
the first direct evidence for the phosphorylation of an
oxyacid group at a cysteine residue The diagnostic
value of MS profiles regarding mainly the selected
peaks was confirmed in a total of 17 independent
spectra obtained with different instruments and samples
Discussion
Over the last decade it has become apparent that 2-Cys Prx is a key component of signal transduction
A
B
Fig 5 MS⁄ MS spectra of the 160–184 tryptic peptide from 2-Cys Prx Phosphorylated 2-Cys Prx was digested with trypsin and prepared for MALDI-TOF MS as described in Experimental Procedures Data were examined as described in Fig 4 The amino acid sequence of the 160–184 tryptic peptide bearing unphosphorylated and phosphorylated cysteines are displayed above the spectrum The fragmentation pat-terns that generate ions y and b are illustrated along the peptide sequence wherein (*) are fragment ions bearing –SO3PO3HNa (A) Expanded view of peaks at m ⁄ z 2800.36 and 2832.35 and the fragmentation of the peak at m ⁄ z 2800.36 (B) Expanded view of peaks at
m ⁄ z 2934.36 and 2950.35 and the fragmentation of the peak at m ⁄ z 2950.35.
Trang 9pathways, ultimately controlling proteins involved in
diverse cellular processes, such as cell proliferation,
differentiation, apoptosis and photosynthesis [25,35–
38] This study is the first to demonstrate that the
activities associated with 2-Cys Prx are regulated
directly by mechanisms sensitive to nucleotides and
bivalent cations in which the concerted action of
both compounds reversibly impairs the peroxidase
activity, whereas only Mg2+ lowers the chaperone
capacity [10,11] In addition to the differential
regula-tion of the dual funcregula-tions, inhibiregula-tion of the
peroxi-dase activity is highly specific because, of the
nucleotides presented here, purine derivatives are
markedly more effective than pyrimidine bases Given
that nucleotides do not participate directly in the
reduction of hydroperoxides, it follows that the
observed loss of activity is almost certainly due to a
local effect on the structure of the protein (see
below) These findings are important for
understand-ing the fundamental question of how 2-Cys Prx
uti-lizes non-redox compounds to regulate the associated
functions and, in so doing, to cope with situations of
oxidative stress This extremely rapid and reversible
association with low molecular mass compounds
devoid of redox capacity may have wide applicability
because we recently reported that 2-Cys Prx in
con-certed action with fructose-1,6-bisphosphate and
Ca2+ stimulates the activity of chloroplast
fructose-1,6-bisphosphatase [25]
2-Cys Prx is an obligate homodimer (a2) whose
con-version to doughnut-shaped (a2)5 decamer is redox
sensitive [24] Apropos, oxidants drive the human and
yeast orthologues from lower molecular mass forms to
higher molecular mass complexes and, in so doing,
impair the peroxidase activity and enhance the
chaper-one capacity [10,11] Although the transition of
2-Cys Prx among oligomers with different molecular
masses may be conceptually adequate for the
regula-tion of associated funcregula-tions [24,27], the unprecedented
ATP-mediated oligomerization is beyond the scope of
this study and it will be reported elsewhere However,
spectroscopic studies of 2-Cys Prx variants clearly
dis-cerned the role of ATP UV-differential
spectropho-tometry and fluorescence emission of the extrinsic
probe ANS initially revealed that the protein interacts
directly with ATP, and further exploration of the
intrinsic fluorescence emission in site-directed mutants
unambiguously assigned the binding site close to
Trp179 In the crystal structure of human 2-Cys Prx,
this region encloses a cavity large enough to hold
nu-cleotides in which the tryptophan residue homologous
to rapeseed Trp179 is located far from the peroxidatic
Cys53 and close to the resolving Cys175 [31] Given
that (a) the mechanism of peroxidase activity includes the formation of an intercatenary disulfide bond link-ing the peroxidatic cysteine with the resolvlink-ing counter-part [3] and (b) ATP locates near the latter (this study), it is reasonable to suggest that the reversible binding of ATP⁄ Mg2+ halts the catalytic cycle via ste-ric perturbation of the resolving cysteine However, we can not exclude the possibility that the reduction of hydroperoxides is inhibited by an allosteric effect of ATP⁄ Mg2+ on the peroxidatic cysteine Although fur-ther studies are required to clarify this issue, our data definitively identify the region surrounding the resolv-ing cysteine of typical 2-Cys Prx as the target for nucleotides
The main outcome of our study is, however, the importance of oxyacid groups at the resolving Cys175 for the in vitro autophosphorylation of 2-Cys Prx A combination of evidence from the lack of a similar capacity in other proteins to the behavior of site-direc-ted mutants clearly dismiss the possibility that trace quantities of contaminating bacterial kinases may co-purify with the recombinant protein [38] The finding that the successive addition of a reductant and an oxi-dant promotes incorporation of the c-phosphoryl moi-ety of ATP indicates that, like other events mediated
by 2-Cys Prx, autophosphorylation depends on a spe-cific redox state The 23 kDa subunit contains two cysteines conserved throughout evolution, and analyses
of site-directed mutants show that Cys175 holds the unique reactive thiol involved in autophosphorylation Moreover, MS detection of over-oxidized sulfur atoms
at the resolving cysteine led us to conclude that the sulfonic and sulfinic forms are necessary for linking the phosphoryl moiety to the protein Further exami-nation of the oxidative step reveals that the autophos-phorylation proceeds in redox environments milder than those induced by harsh oxidants Indeed, the midpoint redox potentials of 2-hydroxy-1,4-naphtho-quinone (Em7=)0.15 V) and 1,4-dihydroxy-9,10-anthraquinone (Em7=)0.18 V) are much lower than
H2O2(Em7=)1.76 V) which is usually used in studies
of ROS [1,2,4] These data uncover the capacity of the rapeseed resolving Cys175 for the oxidation to sulfinic acid, a process that clearly departs from similar sulfur chemistry at the peroxidatic cysteine [19,20,39,40] Two lines of research have examined the phos-phorylation of 2-Cys Prx First, it has been shown that several cyclin-dependent protein kinases promote
in vitro the specific phosphorylation of human 2-Cys Prx at a threonine residue homologous to Thr91 in the rapeseed orthologue [17] Second, the finding that the thiol of mammalian sulfiredoxin [Srx-SH] recruits the c-phosphoryl moiety of ATP
Trang 10yielding a thiophosphate [Srx-S-PO3 )] led to the
proposal that sulfiredoxin subsequently transfers the
phosphoryl group to the sulfinic form of the
peroxid-atic cysteine in human PrxI [-Cys-S(=O)-OH] [19–21]
At this stage, the sulfinic–phosphoric mixed anhydride
[-Cys-S(=O)-O-PO3 )] would be cleaved by a thiol
reductant [R-S-H] yielding a disulfide-S-monoxide
[-Cys-S(=O)-S-R] that would be finally reduced back
to thiol [-Cys-SH] In this context, the strategy of our
phosphorylation of 2-Cys Prx diverges markedly from
previous studies in two important aspects: neither
requires a complementary catalyst, like
cyclin-depen-dent kinases or sulfiredoxin, nor proceeds via Thr91
or the peroxidatic cysteine Indeed, our data provide
entry into a previously unsuspected mechanism by
which the successive reductionfioxidation of 2-Cys
Prx generates oxyacid groups at Cys175 for the
subse-quent formation of the sulfinic-phosphoryl
[-(Cys175)-SO2PO3 )] and sulfonic-phosphoryl [-(Cys175)-SO3
PO3 )] anhydrides (Scheme 1) Related to this, the
mechanism by which dithiothreitol alone dramatically abrogates the autophosphorylation remains unknown Does the reductant impair the process (a) before incorporation of the phosphoryl moiety by remov-ing the oxyacid groups or (b) after formation of -(Cys175)-S(=O)1–2-O-PO3) by cleaving the mixed anhydride? The answer to these questions will reveal whether the oxyacid group itself at Cys175 or the sulfi(o)nic-phosphoric anhydride are endowed with an unusual reactivity to reductants
Although many studies have concentrated on the events underlying phosphorylation of the peroxidatic cysteine of 2-Cys Prx [19–21], we know of none that addressed the resolving cysteine In almost all typical 2-Cys Prx, the function of the latter residue was hith-ertho confined to participating in the formation of an intercatenary disulfide bond with the sulfenic acid of the peroxidatic cysteine Moreover, in line with the current paradigm on the mechanism for the reduction
of hydroperoxides [8], the resolving cysteines of try-paredoxin peroxidase and AhpC from Trypanosoma brucei brucei and Salmonella typhymurium have been identified as targets in the reduction of the disulfide bond for the reactivation of peroxidase activity [41,42] Against this background, we put forward a new scenario wherein ATP interacts actively with 2-Cys Prx and, in so doing, modifies the quaternary structure and associated functions Moreover, the unusual phosphorylation of Cys175 oxyacid groups brings together the redox chemistry of the sulfur atom and the phosphorylating capacity of ATP, thereby providing a versatile mechanism wherein Cys175 appears as dual sensor able to perceive changes in the redox and energy status of the cell By virtue of the flexibility of using redox and nonredox chemistries at
a single cysteine residue, the possibilities to process a wide spectrum of stimuli into different cellular responses greatly extend the prevalent view circum-scribed to redox transformations of sulfhydryl groups [43]
Experimental procedures
Materials Recombinant rapeseed 2-Cys Prx was prepared as described previously [25] Biochemicals were purchased from Sigma-Aldrich (St Louis, MO, USA)
Construction of 2-Cys Prx mutants C-terminal hexahistidine-tagged variants of 2-Cys Prx were generated by the PCR megaprimer method using, in the
Scheme 1 Reaction scheme for the autophosphorylation of
2-Cys Prx A reductant cleaves the intercatenary disulfide bond
yielding the reduced form of the sulfur atom at Cys175 (reaction 1).
The subsequent oxidation transforms the sulfhydryl group into the
sulfinic and sulfonic species (reaction 2) The reactivities of these
groups and the close proximity to the ATP binding site facilitate the
incorporation of the c–phosphoryl moiety (reaction 3) linking in
con-sequence redox and nonredox chemistries Eventually, a
phospha-tase facilitates the hydrolysis of the phosphoryl group and returns
Cys175 to the oxidized state.