-
Noname manuscript No.(will be inserted by the editor)
An analytical workflow for the molecular dissection
ofirreversibly modified fluorescent proteins
Vivien Berthelot · Vincent Steinmetz · Luis A.Alvarez · Chantal
Houée-Levin · FabienneMerola · Filippo Rusconi · Marie Erard
Received: date / Accepted: date
Abstract Owing to their ability to be genetically expressed in
live cells, fluores-cent proteins have become indispensable markers
in cellular and biochemical studies.These proteins can undergo a
number of covalent chemical modifications that may af-fect their
photophysical properties. Among others, such covalent modifications
maybe induced by reactive oxygen species (ROS), as generated along
a variety of bio-logical pathways, or under ionizing radiations. In
a previous report [1], we showedthat exposure of the cyan
fluorescent protein (ECFP) to •OH amounts that mimickthe conditions
of intracellular oxidative bursts (associated with intense ROS
produc-tion), led to observable changes in its photophysical
properties, yet in the absence ofany direct oxidation of the ECFP
chromophore. In this work, we analyse in depth theassociated
structural modifications occurring in the protein. Following the
quantifiedproduction of •OH, we devised a complete analytical
workflow, based on chromatog-raphy and mass spectrometry, that
allows us to fully characterize the oxidation events.While
methionine, tyrosine and phenylalanine are the only amino-acids
found to beoxidized, a semi-quantitative assessment of their
oxidation level shows that the pro-tein is preferentially oxidized
at eight residue positions. To account for the preferredoxidation
of a few, poorly accessible methionine residues, we propose a
multi-stepreaction pathway supported by pulsed radiolysis
experiments. The described exper-imental workflow is widely
generalizable to other fluorescent proteins, and opens
F.R. and M.E are co-last authors of this work.
V. Berthelot · V. Steinmetz · C. Houée-Levin · F. Merola · F.
Rusconi · M. ErardLaboratoire de chimie physique; UMR CNRS 8000;
Building 350,F-91405 Orsay Cedex, France,E-mail: marie.erard
@u-psud.fr;Tel. +33 1 69 15 30 14; Fax +33 1 69 15 61 88
F. Rusconi also at“Régulation et dynamique des génomes”; U
INSERM 565—UMR CNRS 7196; Muséum nationald’Histoire naturelle; 57,
rue Cuvier, Case Postale 26; F-75231 Paris Cedex 05, France,E-mail:
[email protected];Tel. +33 1 69 15 76 04; Fax +33 1 69 15
61 88
-
2 Vivien Berthelot et al.
the way to the identification of crucial covalent modifications
impacting their photo-physics.
Keywords protein oxidation · mass spectrometry · •OH radicals ·
cyan fluorescentprotein (ECFP) · γ & pulsed radiolysis
1 Introduction
Fluorescent proteins (called FPs, for short) have revolutionized
the mechanistic in-vestigations of cellular processes by live cell
imaging, flow cytometry and high-throughput bioassays [2]. These
FPs comprise 11 β sheets that form a barrel en-closing a coaxial
α-helix (Figure 1 a) which bears the chromophore resulting fromthe
cyclization of three consecutive amino-acids [65-66-67] (Figure 1
b). As such,these proteins have been used to build fully
gene-encoded sensors that allow one tomonitor a large set of
biological chemical events in live cells [3].
Engaging in chemical sensing through the use of FPs requires a
thorough knowl-edge of their photophysical responses to the
cellular micro-environment. While someattention has been paid to
the side-effects of either pH [4, 5, 6, 7], chloride ions [8],or
refractive index [9] on the photophysical properties of FPs, the
possible conse-quences of local production of reactive oxygen
species (ROS: O2•−, •OH, H2O2,HOCl) have seldom been addressed [10,
11, 12, 1]. Nevertheless, ROS have cru-cial roles in physiological
and pathological processes: they can be produced in largequantities
in cells, be potentially harmful and cause oxidative damage, the
so-calledoxidative stress [13, 14]. For example, the dynamics of
pathogen phagocytosis, a phe-nomenon correlated with high levels of
ROS production by enzymatic systems likeNADPH oxidase or
myeloperoxidase [15], was recently investigated with the helpof
FPs, [16, 17, 18, 19] while there has been a recent push to develop
members ofthe FP family as specific oxidation sensors [20, 21, 10,
11, 22, 18]. In some of thelatter cases [10, 11, 22, 18] as well as
in other reports [1, 12], the consequences ofROS exposure on the
photophysical properties of FPs have been reported and
showeddependency on either the nature of the oxidant and its
quantity, or the FP variant. Onthe other hand, the characterization
of ROS-induced chemical modifications of FPsmight prove fundamental
in the understanding of their photobleaching mechanisms.Indeed,
these photoreactions are thought to result from the uncontrolled
productionof oxidants (like ROS) from the singlet or triplet
fluorophore excited states [23]. Pho-tobleaching generates
variations of the fluorescence signal, known as photofatigue,thus
limiting the duration and the reliability of the measurements. This
photophys-ical limitation becomes critical when high illumination
densities are required, likein super-resolution optical microscopy
or single molecule applications [24, 25]. Sofar, very few studies
have been devoted to the identification of chemical modifica-tions
involved in photobleaching [23, 26] and both the pathways and the
targets aregenerally still unknown. In a previous study, we
analyzed in detail the photophysicalperturbations induced by
controlled amounts of hydroxyl (•OH) and superoxide an-ion (O2•−)
free radicals on the cyan fluorescent protein (ECFP) [1] that is
the mostwidely used donor in FRET-based imaging experiments in
combination with a yel-low partner (like Citrine) [3, 27]. We found
significant perturbations of the ECFP
-
Molecular dissection of modified fluorescent proteins 3
fluorescence (decreases in intensity and lifetime, without
changes in its excitationand emission spectra), which did not
result from direct chemical modification of thechromophore, but
rather from oxidations at other sites that remain to be
identified.
Understanding the causative relationships between the exposure
of FPs to ROSand the resulting modification of their photophysical
properties requires a thoroughstructural characterization of the
oxidized proteins. Due to the lack of appropriateanalytical tools,
these questions have not yet been addressed in detail up to
now.Crystallography might look like an appropriate approach because
fluorescent proteinsare highly compact and can be cristallized
easily. However, the expected multiplicityof the chemical
modifications limits the use of crystallography in this context.
Thebottom-up strategy that is often used in proteomics should hold
more promise. How-ever, the compactness of the FP barrel structure
makes these proteins refractory toconventional enzymatic
break-down, and thus previous mass spectrometric analyses(briefly
reviewed by Alvarez et al. (2009) [28]) had to resort to chemically
harsh pro-cedures with the risk of introducing unwanted chemical
modifications to the proteins.In a previous work, we devised a new
method aimed at digesting FPs in very mildconditions that opened
the way to the molecular dissection of chemically-modifiedFP
variants [28].
In the present report, we build on that previous work and
further enhance ouranalytical workflow to achieve the full
characterization of •OH-oxidized ECFP. Thisleads to the
identification of ECFP oxidized residues, along with a first
semi-quanti-tation of their oxidation events. We were able to spot
oxidation events, mostly re-stricted to a few specific residues
that might be responsible for the observed photo-physical changes.
In addition, pulsed radiolysis studies provided further insights
onthe •OH primary targets leading to these selective
oxidations.
2 Experimental section
2.1 Cyan fluorescent protein purification and endoAspN-based
digestion
ECFP refers to AvGFP F64L/S65T/Y66W/N146I/M153T/V163A/H231L
(Clontech).The production and the purification of His-tagged
recombinant ECFP were performedas described previously [1]. The
His-tag was cleaved away, leading to a cyan fluores-cent protein in
which the starting M residue is replaced by the GA dipeptide.
Theresidue numbering scheme relies on the chromophore TWG triplet
residues being atpositions 65-66-67 (see Figure 1 b). The ECFP
solution was dialysed against a 30 mMphosphate buffer at pH 7.5 for
irradiation experiments. The protein concentration wasassessed by
UV absorption of the chromophore [ε (λ434nm) = 30000 cm−1.M−1].
2.2 Radical production and radiolysis
A detailed description of the radiolysis procedures is provided
in Online Resources,section OR1. Briefly, the quantitative
production of •OH has been performed witheither stationary γ
radiolysis or high energy electron pulses. Their selection was
per-formed using the well-known method of scavengers in solutions
deoxygenated and
-
4 Vivien Berthelot et al.
saturated with N2O [29]. Additional pulsed radiolysis
experiments were performedusing •N3 radicals as oxidants. For the
irradiations, the ECFP concentration wasusually 5 µM unless
otherwise specified, at neutral pH. The doses applied were inthe
range 20–400 Gy (stationary radiolysis) or in the range 4–10 Gy
(pulsed ra-diolysis). For an easier understanding of the results,
the dose was replaced by theR = [•OH]/[ECFP] ratio, which is
proportional to the irradiation dose.
The pulsed radiolysis-elicited reaction was followed
spectrophotometrically be-tween λ350 nm and λ750 nm and the
difference absorption spectra of the reaction in-termediates were
reconstructed from the maxima of the absorbance variation in
theinvestigated wavelength range.
2.3 Structural modeling of residue accessibility
The ECFP structure at physiological pH is available from PDB
(ID: 2WSN) [30], butdoes not comprise the C-terminal tail of the
protein (sequence 230 TLGMDELYK 238).In order to correlate the mass
spectrometric results with the spatial localization of allof the
ECFP residues, we generated an automated model of ECFP using the
SWISS-MODEL workspace described in Online Resources, section
OR2.
We calculated the solvent-accessible surface percentage of each
residue usingthe GETAREA service provided by the Sealy Center for
Structural Biology at theUniversity of Texas Medical Branch,
Galveston, USA [31].
2.4 Chromatography
Desalting of the proteins and of peptide mixtures were performed
according to theprocedure described by [32], with the R2 Poros
polymeric resin (Applied Biosystems)as the reversed phase.
Peptides from endoAspN-digested ECFP [28] were separated by
reversed-phasehigh-performance liquid chromatography (RP-HPLC) on
an Åkta Purifier setup fromGE Healthcare, using a Stability S-C23
end-capped resin (Cil Cluzeau, Sainte-Foy-la-Grande, France). The
gradient was developped from 100 % buffer A [water, 0.035 %
tri-fluoroacetic acid (TFA)] to 100 % buffer B [95 % acetonitrile
in 0.035 % TFA] in60 min. Each fraction was manually collected as
the peak shape was monitored onthe Unicorn software chromatogram
view (GE Healthcare).
2.5 Mass spectrometry
Fluorescent proteins—irradiated or not—or peptides thereof were
analyzed by elec-trospray ionization (using either normal, micro-
or nano-spray) on three differentmass spectrometers : 1) a QStar
Pulsari hybrid quadrupole–time-of-flight (AppliedBiosystems) ; 2) a
7 T Apex Qe FT-ICR with the quadrupole-hexapole interface al-lowing
for m/z -based selection, ion accumulation and cooling before
injection in theICR cell (Bruker Daltonics); 3) an LTQ-Orbitrap
(Thermo Scientific) with a micro-spray source in line with a Dionex
Ultimate 3000 HPLC setup. In each case, a typical
-
Molecular dissection of modified fluorescent proteins 5
Fig. 1 a. Barrel structure of ECFP. b. ECFP sequence. Residues
that were found oxidized are denotedusing filled or semi-filled
circles (see text for details). The chromophore is typeset in bold
blue characters.
ionization protocol was used, with standard settings appropriate
for either a ≈ 30 kDaprotein or peptides. All the experiments were
performed in the positive ionizationmode. The typical amount of
protein used for one analysis was of 25 µL of a 10 µMsolution. The
peptidic solutions were typically in the 2–5 µM range. When
peptideanalyzes were performed by MALDI-TOF mass spectrometry, the
instrument was a4800 MALDI TOF/TOF Analyzer from AB Sciex and the
sample preparation wasaccording to the conventional dried droplet
procedure using α-cyano-4-hydroxycin-namic acid (ACHCA) as the
matrix. All gas-phase fragmentation spectra were ac-quired using
collision-induced dissociation (CID). The CID mass data were
analyzedto locate the oxidation events on m/z -selected ions.
Mass spectrometric data analysis was performed by first
exporting the data tosimple ASCII-formatted files so as to make use
of free and open source softwarefor all the analysis steps. Mass
spectrum visualization and analysis was performedby using the mMass
program (http://mmass.org) [33, 34]. Mass spectrometric
datasimulations and detailed analyses were performed using either
the massXpert pro-gram (http://massxpert.org) [35, 36] or GNU
polyXmass [37]. Data handling wasperformed by setting up a
relational database using the free and open source SQLitedatabase
(http://sqlite.org). All in-house programming was performed using
eitherC++ or Python as programming languages under a Debian
GNU/Linux computingplatform (http://debian.org).
-
6 Vivien Berthelot et al.
3 Results and discussion
3.1 Oxidation of the ECFP by •OH
After γ irradiation of purified ECFP with a delivered dose
yielding a R= [•OH]/[ECFP]ratio of 10, the protein was
desalted/concentrated and analyzed by ESI-MS. Figure 2shows the
spectra obtained for non-irradiated or irradiated ECFP.
Non-oxidized protein (panel A) produced a charge envelope with
peaks havingthe expected m/z ratio (e.g.: measured, m/z 841.21, z=
32; calculated, m/z 841.01).After oxidation with •OH (panel B),
that same peak almost disappeared, with theconcomitant build-up of
several new peaks (denoted with asterisks) separated fromeach other
by m/z 0.5, corresponding to a mass increment due to the binding of
oxy-gen atoms to the protein (∆M = 16 u). At least five resolved
oxidation levels couldbe observed. Although in our experimental
conditions, water γ radiolysis yields anamount of •H radicals that
is non-negligible ([•H]/[ECFP]≈ 1), we did not detectany of their
reaction products, like, for example, loss of H2S (∆M =−34 u) or
loss ofCH3SH, with ∆M =−48 u [38, 39, 40, 41]. Further, at the
level of the whole protein,no other •OH-induced modification was
observed, like, e.g. decarboxylation, or in-termolecular Cys- or
Tyr-based dimerization [for a review of ROS-induced
chemicalmodifications see [42]].
A dose-response experiment was performed whereby ECFP was
submitted to•OH oxidation with R ratios in the range 0–17. The
correlation between an increas-ing R ratio and the appearance of
ECFP oxidized variants, along with the concomi-tant decrease of the
mass peak corresponding to the unoxidized ECFP, is manifest,
asshown in panel C of Figure 2. With an R ratio of 17, the protein
was so heavily modi-fied that the obtained mass spectrum failed to
be informative: the molecular diversityassociated to the large
number of modified ECFP polypeptide triggers the classicalspectral
suppression phenomenon. On the other hand, for oxidation ratios of
≈ 17,we observed only a 15 % decrease in both the fluorescence
quantum yield and life-time of ECFP [1]. Such moderate concomitant
photophysical perturbations indicatethat many of these ECFP
oxidations actually have little or not impact on the
proteinphotophysical properties.
3.2 Characterization of the •OH-oxidized variants of ECFP
3.2.1 Direct analysis of the oxidized ECFP endoAspN-produced
peptidic mixture
Oxidized ECFP samples were subjected to an endoAspN digestion
according to theprotocol described by [28]. The peptidic mixture
was analyzed by ESI mass spec-trometry, by injection in the flow,
direct infusion or nano-spray. These mass spectro-metric analyses
failed to afford a correct sequence coverage, which contrasted
withthe excellent coverage for non-oxidized ECFP that we described
in an earlier report[28]. Most evident was the complete lack of
signal corresponding to the chromopep-tide (coordinates [36–75]),
comprising the three 65 TWG 67 residues that undergo aseries of
post-translational modifications to form the chromophore
(represented using
-
Molecular dissection of modified fluorescent proteins 7
Fig. 2 Mass spectrometry-based monitoring of the ECFP oxidation
with •OH. ECFP was either non-oxidized (a) or •OH-oxidized with R =
10 (b). The inset shows a zoomed view of the framed mass peak(z =
32). The non-oxidized protein was almost homogeneous, while the
•OH-oxidized protein showedseveral variants differing by the mass
of one or more oxygen atoms (∆M = 16 u). c. Oxidation progressionof
ECFP as observed upon increasing the R = [•OH]/[ECFP] ratio
(numbers next to the traces). The mostuseful R value for our
experiments was found to be in the range 4–10. Arrows point to the
mass peakcorresponding to the unmodified polypeptide.
-
8 Vivien Berthelot et al.
bold blue characters in Figure 1 b). This peptide is one of the
largest peptides ex-pected from an ECFP digestion by endoAspN, even
with partial cleavages occurringin other regions of the protein
sequence. We reasoned that this insufficient sequencecoverage was
due to a huge increase in the molecular complexity of the peptidic
mix-ture in the case of oxidized ECFP. Indeed, the whole set of
oxidation combinationsfor each oxidizable peptide could have made
the sample too complex to be succcess-fully analyzed without
suffering from spectral suppression. Furthermore, because
theoxidation of peptides increases their hydrophilicity, the
oxidized peptides could bedefavored in their desolvation/ionization
with respect to their unmodified counter-parts [43], thus leading
to a mass spectrometric signal loss that would mainly affectthe
subpopulation of peptides that was of greatest interest to this
study [44].
3.2.2 Chromatographic separation of ECFP peptides
In a first attempt to reduce mass spectral suppression, a
micro-chromatography ex-periment was performed as in [32], with a
step-gradient elution in two 20 % and60 % acetonitrile fractions.
Analysis of the fraction contents showed that a substantialamount
of mass spectrometric signal could be recovered. Indeed, the
chromopeptidewas detected in the 60 % acetonitrile fraction;
however, sequence coverage was stillpartial.
Fig. 3 Reversed-phase high performance liquid chromatography
separation of •OH-oxidized ECFP pep-tides following endoAspN
digestion. Absorbance detection was performed at λ214 nm and λ414
nm wave-lengths, for detection of the peptidic bond and of the
chromophore, respectively. The amount of proteininjected was 200
pmol. Each fraction was collected in a separate tube for further
analysis by mass spec-trometry.
-
Molecular dissection of modified fluorescent proteins 9
The peptidic mixture obtained upon endoAspN digestion of ECFP
was thus sub-jected to an HPLC separation using a reversed-phase
resin. Figure 3 shows the ob-tained chromatogram, with absorbance
detections at λ214 nm and λ414 nm. The latterwavelength corresponds
to the absorption band of the chromopeptide [4], thus allow-ing us
to monitor the fraction in which it eluted (at 56 min). All the
fractions were col-lected separately and later analyzed by ESI or
MALDI mass spectrometry to identifythe oxidized peptides. One
useful observation was that oxidized peptides did
almostsystematically elute in the fraction preceding the one
containing their non-oxidizedpeptide counterparts. This observation
was both expected (because the oxidized pep-tides are more
hydrophilic) and helpful in the analysis of the mass spectrometric
databecause one could predict in which fraction to search for
oxidized variants of anygiven peptide.
3.2.3 Full mapping of the oxidation sites in ECFP
The ECFP primary structure could be completely covered by our
mass spectrometricanalyses. In agreement with our preceding
findings (Figure 2), we focused our dataanalyzes on oxygen
additions (∆M= 16 u). Upon an entirely manual scrutiny of
massspectrometric data obtained for oxidized ECFP-derived peptides,
all the oxidationevents were mapped to a few positions in the
protein sequence. The residues found tobe oxidized are all either
aromatic or sulfur-containing residues [45, 46]. Two kindsof
residues could be singled out: residues oxidized with R =
[•OH]/[ECFP] ratios inthe range 4–10 and residues that could only
be observed as oxidized species uponintense irradiation (R ≈ 20).
In Table 1 and Figure 1 (panel b), we listed only thepeptides that
were found oxidized for R in the range 4–10. For each listed
peptide, theresidue detected as bearing the oxidation is specified,
along with its abundance and itssolvent accessibility. The
different residues belonging to the oxidized peptides
wereclassified into four categories: those that were found not
oxidized ( /0), barely oxidized(
-
10 Vivien Berthelot et al.
Table 1 ECFP oxidized peptides data for R = [•OH]/[ECFP] in the
range 4–10
ID Sequence Coordinates Oxid. residue / abundance SAA %1
GAVSKGEELFTGVVPILVEL [ –18] ? / -2 DVNGHKFSVSGEGEG [21–35] F27 /
0.73 DHMKQH [76–81] M78 / 0.9
4 DFFKSAMPEGYVQERTIFFK [82–101]
F83 / /0F84 / /0
M88 /Y92 / /0F99 / /0F100 / /0
0.50.200
58.30.2
5 DGNYKTRAEVKF [103–114] ? / -
6 EYNYISHNVYITA [142–154] ? /Y151 /
-59.6
7 DGPVLLPDNHYLSTQSALSK [190–209] Y200 / 33.0
8 DHMVLLEFVTAAGITLGMDELYK [216–238]
M218 /F223 / /0
M233 /Y237 /
040.258.4100
SAA: surface accessible area. The oxidized residues were
classified in four categories: those that werefound not oxidized (
/0), barely oxidized (
-
Molecular dissection of modified fluorescent proteins 11
Fig. 4 Mass spectrometric analysis of the •OH-oxidized ECFP
chromopeptide (R ≈ 20). Isotopic clusterswere found to correspond
to the chromopeptide, both in the non-oxidized form, at m/z
1120.83, z = 4 andin the singly-oxidized form at m/z 1124.85. The
chromopeptide was found to bear a disulfide bond.
oxidized residues hinted at indirect oxidation mechanisms that
we investigated nextby performing the pulsed radiolysis experiments
below.
4 Pulsed radiolysis
The first steps of the reaction of •OH with ECFP were
investigated by pulsed radi-olysis, where the time span for •OH
creation is short compared to the reaction time.This method affords
two useful data sets: the identification of reaction
intermediates(by their transient absorption spectra) along with
their formation and decay kinetics.
The absorbance kinetics traces recorded after the pulsed
production of •OH rad-icals were different depending on the
wavelength (Figure 5, panel a). At λ315 nm,the absorbance reached a
maximum value a few microseconds after the pulse andthen decayed.
The build-up phase is in agreement with a rate constant of (2.5±
0.5)1010 mol−1.L.s−1, which is consistent with that of the reaction
of •OH with pro-teins of the same size [49]. In the range λ390−410
nm, the absorbance reached itsmaximum ca. 50 µs after the pulse and
then decayed slowly in ≈ 100 ms (notshown). These kinetics, that
depend on wavelength, show that
several—chemicallydifferent—intermediates are produced in the first
tens of microsecond after the pulsed•OH production. Moreover, the
time-scale of the absorption build-up at λ390 nm didnot depend on
the [ECFP] nor on the [•OH] (Figure 5, panel a), indicating an
in-tramolecular process.
The time-resolved absorption spectrum of the ECFP solution
recorded 40 µs afterthe •OH pulse is shown in Figure 5, panel b,
trace •OH. The signal at λ320 nm might
-
12 Vivien Berthelot et al.
Fig. 5 a. Absorbance kinetic traces recorded after the pulse at
λ315 nm and λ390 nm. Experimental con-ditions were, for λ315 nm,
[ECFP] = 30 µM, [•OH] = 7 µM; for λ390 nm, curve 1: [ECFP] = 30
µM,[•OH] = 7 µM; for λ390 nm, curve 2: [ECFP] = 12 µM, [•OH] = 4.5
µM. b. Difference absorption spectrarecorded 40 µs and 500 µs after
the pulse, for •OH radicals and •N3 radicals, respectively.
Experimentalconditions were [ECFP] = 12 µM, [•OH] = 4.5 µM. The
experimental set-up time response was below0.1 µs.
-
Molecular dissection of modified fluorescent proteins 13
be attributed to addition of •OH to Tyr and Phe [50, 51]. The
bands in the λ390−410 nmrange could be attributed to ECFP-(TyrO•)
and/or to ECFP-(MetS∴X)•+ (X = O, Nor S) [52, 46, 45, 53] (Figure
5, panel b). To confirm the nature of these radicals, wecompared
this spectrum with the one obtained after an oxidation using azide
radicals(Figure 5, panel b, trace •N3), which are known to make
electron transfers with ty-rosine, thus yielding only the
ECFP-(TyrO•) radical, and to have low reactivity withmethionine
[54, 55]. The new spectrum appeared different from the former one,
withshifted maxima, confirming that TyrO-centered radicals are not
the only short-livedproducts of the ECFP oxidation and that
methionine-centered radicals are present.
In proteins, the tyrosine residue is usually thought to be the
“end point” of oxi-dation. For example, oxidations of tryptophane
or methionine residues would, in ourcase, end up with the formation
of ECFP-(TyrO•) by intramolecular electron transfer[56]. However,
in our experiments, we found that several methionine residues are
ulti-mately oxidized, despite the fact that, in some cases, these
residues are deeply buriedin the protein structure. In addition,
the reaction leading to the increase of absorbancein the λ390 nm
region in pulsed radiolysis experiments was of the first order,
indicatingan intramolecular reaction. We propose in Scheme 6
interpretations of these obser-vations. •OH radicals can react with
methionine or tyrosine residues, as observed inthe transient
absorption spectra shown in Figure 5 (steps a and b, respectively).
The•OH adduct on the tyrosine residue might either lead to DOPA
(step c) or undergoproton-catalyzed dehydration to the TyrO•
radical (step d), which in turn is repairedby a methionine residue
(step e), thus leading to methionine sulfoxide (step f). Inthis
protein, the distances between methionine and tyrosine residues can
be short(d [M78–Y200]= 8.2 Å, d [M218–Y143]= 4.6 Å and d
[M233–Y237]= 5.8 Å) orlonger (d [M88–Y39]= 20 Å (see Online
Resources Figure ??). In the former case,the electron transfer
should be fast, while in the latter case it can proceed by
hoppingthrough aromatic residues [57, 58], as illustrated in Online
Resources Figure ??. Inall cases, the electron transfer requires
that the reduction potential of the methionineresidue be lower than
that of tyrosine, which might indeed happen because the me-thionine
reduction potential is highly dependent on its immediate
environment [59].Furthermore, other studies on peptides or proteins
showed that many oxidations donot always end-up at tyrosine
residues [60, 61, 62, 63].
5 Conclusions
In this work, the described analytical workflow showed that
fluorescent protein oxi-dation was observed on methionine, tyrosine
and phenylalanine residues. One inter-esting finding is that the
three-dimensional modeling of the protein sets part of theoxidized
residues in the inner face of the barrel, hinting at indirect
oxidation path-ways involving one or more hops across the protein
structure. This observation is ofparticular interest in the context
of studies involving •OH-based oxidation of proteinsto determine
the three-dimensional structure of proteins or the topological
organiza-tion of protein assemblies [64, 65]. Our work could indeed
point at one difficulty inthis field, that might arise if the
ultimate oxidation state of any given residue were notthe sole
result of its mere accessibility to the solvent.
-
14 Vivien Berthelot et al.
Fig. 6 Reactions proposed to explain the obtained kinetics
traces
Another significant result, in combination with previous results
from our lab-oratory [1], is that the chemical modifications
identified in this report do elicit aperturbation of the
photophysical properties of the fluorescent protein, even if
thechromophore was not modified. This observation is consistent
with our previouslyproposed hypothesis that the oxidation of the
protein increases the barrel flexibilitythus enhancing
non-radiative deexcitation paths through excited-state
chromophoretorsion [1]. In addition, the present work shows that
these dynamical perturbationsmight be triggered only by a few
critical oxidations in the ECFP structure, all locatedat least 8 Å
away from the chromophore, thus excluding its direct quenching.
Acknowledgements V.B. received a doctoral fellowship from the
University of Paris-Sud, Orsay, France.L.A.A. was supported by a
fellowship from the French ministry of research (MESR). The authors
thankDr Philippe Maı̂tre (plateforme de spectrométrie de masse du
LCP, University of Orsay) for interestingdiscussions, Drs Lionel
Dubost and Arul Marie of the mass spectrometry facility of the
Muséum in Paris,France and Dr Jean-Pierre Le Caer of the mass
spectrometry facility of the ICSN in Gif-sur-Yvette, France.We
thank Dr Vincent Favaudon for the use of the pulsed radiolysis
set-up (Institut Curie, Orsay, France).We are indebted to the COST
CM1001 (non-enzymatic protein oxidation) action for fruitful
discussions.
-
Molecular dissection of modified fluorescent proteins 15
1 Online Resources
1.1 OR1 – Radical production and radiolysis
The quantitative production of •OH has been performed with
either stationary γ radi-olysis or high energy electron pulses. In
both cases, the selection of •OH or •N3 freeradicals was obtained
using the well-known method of scavengers [29]. For instance,in
N2O-saturated aqueous solutions, radiolysis creates •OH radicals
with a radiationchemical yield (G) equal to 0.55 µmol.J−1 [29]. In
N2O-saturated solutions and in thepresence of azide ions (1 mM
NaN3), azide radicals, •N3, are formed with a radiationchemical
yield equal to 0.55 µmol.J−1 [29].
Irradiations using γ rays were carried at room temperature using
the panoramic60[Co] source IL60PL Cis-Bio International (France) at
the University Paris-Sud (Or-say, France). The dose rate was
determined by Fricke dosimetry for each positionused in front of
the source and ranging from 2 to 4.5 Gy·min−1. The ECFP
concen-tration for irradiations was 5 µM at neutral pH (30 mM
phosphate buffer, pH 7.5,sample volume in the range 100–500 µL).
Samples were purged gently under agi-tation without bubbling with
N2O. The irradiation doses delivered were in the range20–400 Gy.
For an easier presentation and understanding of the results, the
dose wasreplaced by the concentration R = [radicals]/[ECFP] ratio,
which is proportional tothe dose where [radicals] = dose
·G(radical).
In pulsed radiolysis experiments, free radicals were generated
by delivering, intoan aqueous solution, a 800 ns pulse of high
energy electrons (≈ 4 MeV) from thelinear accelerator located at
the Curie Institute in Orsay, France [66]. The doses perpulse were
calibrated from the absorption of the thiocyanate radical SCN•−
obtainedby radiolysis of the thiocyanate ion solution in
N2O-saturated phosphate. These doseswere in the range 4–10 Gy. The
reaction was followed spectrophotometrically be-tween λ350 nm and
λ750 nm, in a 2 cm path length cuvette designed for pulse
radiolysisexperiments. The spectra of the intermediates were
reconstructed from the maximaof the absorbance variation in the
investigated wavelength range. The time responseof the experimental
setup is below 0.1 µs.
1.2 OR2 – Structural modeling of residue accessibility
ECFP structure at physiological pH is available from PDB (ID:
2WSN [30], but doesnot contain the N-terminal tail of the protein
(sequence 230TLGMDELYK238). Inorder to correlate the mass
spectrometric results with the spatial localization of allof the
ECFP residues, we generated an automated model of ECFP using the
SWISS-MODEL workspace [67, 68, 69] and the crystal structure for
the S65G, Y66G GFPvariant (PDB ID: 1QYO) [31] as a template that
contains the [T230-Y237] stretch has97.9 % homology with ECFP. We
then imported the resulting PDB file and alignedit with 2WSN with
the VMD software package [70] for use of the [T230-Y237]modeled
stretch in all the spatial analysis in this report in conjunction
with 2WSN.
-
16 Vivien Berthelot et al.
1.3 Plausible one-hop long-distance electron transfer between a
tyrosine andmethionine residues
Fig. 1 Modelling of the three-dimensional distance between the
exposed Tyr39 and the buried Met88. Theelectron transfer might
happen by following two routes: through Phe71 or Phe8.
References
1. Alvarez L, Levin CH, Merola F, Bizouarn T, Pasquier H, Baciou
L, RusconiF, and Erard M. Are the fluorescent properties of the
cyan fluorescent proteinsensitive to conditions of oxidative
stress? (2010) Photochem. Photobiol., 86(1):55 – 61.
2. Tsien RY. The green fluorescent protein. (1998) Annu. Rev.
Biochem, 67:509 –544.
3. Newman RH, Fosbrink MD, and Zhang J. Genetically encodable
fluorescentbiosensors for tracking signaling dynamics in living
cells. (2011) Chem. Rev.,111(5):3614 – 3666.
4. Fredj A, Pasquier H, Demachy I, Jonasson G, Levy B, Derrien
V, Bousmah Y,Manoussaris G, Wien F, Ridard J, Erard M, and Merola
F. The single T65S mu-tation generates brighter cyan fluorescent
proteins with increased photostabilityand pH insensitivity. (2012)
PLoS One, 7(11):e49149.
5. Villoing A, Ridhoir M, Cinquin B, Erard M, Alvarez L,
Vallverdu G, Pernot P,Grailhe R, Merola F, and Pasquier H. Complex
fluorescence of the cyan flu-orescent protein: Comparisons with the
H148D variant and consequences forquantitative cell imaging. (2008)
Biochemistry, 47(47):12483 – 12492.
6. Erard M, Fredj A, Pasquier H, Beltolngar DB, Bousmah Y,
Derrien V, VincentP, and Merola F. Minimum set of mutations needed
to optimize cyan fluorescentproteins for live cell imaging. (2013)
Mol Biosyst, 9(2):258 – 267.
7. Poëa-Guyon S, Pasquier H, Merola F, Morel N, and Erard M.
The enhancedcyan fluorescent protein: A sensitive pH sensor for
fluorescence lifetime imaging.(2013) Anal Bioanal Chem,
405(12):3983 – 3987.
-
Molecular dissection of modified fluorescent proteins 17
8. Griesbeck O, Baird GS, Campbell RE, Zacharias DA, and Tsien
RY. Reduc-ing the environmental sensitivity of yellow fluorescent
protein. Mechanism andapplications. (2001) J. Biol. Chem.,
276(31):29188 – 29194.
9. van Manen HJ, Verkuijlen P, Wittendorp P, Subramaniam V, van
den Berg, K T,Roos D, and Otto C. Refractive index sensing of green
fluorescent proteins inliving cells using fluorescence lifetime
imaging microscopy. (2008) Biophys. J.,94(8):L67 – 9.
10. Palazzolo AM, Suquet C, Konkel ME, and Hurst JK. Green
fluorescent protein-expressing Escherichia coli as a selective
probe for HOCl generation within neu-trophils. (2005) Biochemistry,
44(18):6910 – 6919.
11. Tsourkas A, Newton G, Perez JM, Basilion JP, and Weissleder
R. Detection ofperoxidase/H2O2-mediated oxidation with enhanced
yellow fluorescent protein.(2005) Anal. Chem., 77(9):2862 –
2867.
12. Bou-Abdallah F, Chasteen ND, and Lesser MP. Quenching of
superoxide radicalsby green fluorescent protein. (2006) Biochim.
Biophys. Acta, 1760(11):1690 –1695.
13. Brieger K, Schiavone S, Jr FJM, and Krause KH. Reactive
oxygen species: Fromhealth to disease. (2012) Swiss Med Wkly,
142:w13659.
14. Winterbourn CC. Reconciling the chemistry and biology of
reactive oxygenspecies. (2008) Nat Chem Biol, 4(5):278 – 286.
15. Winterbourn CC, Hampton MB, Livesey JH, and Kettle AJ.
Modeling the reac-tions of superoxide and myeloperoxidase in the
neutrophil phagosome: Implica-tions for microbial killing. (2006)
J. Biol. Chem., 281(52):39860 – 39869.
16. Hoppe AD and Swanson JA. Cdc42, Rac1, and Rac2 display
distinct patterns ofactivation during phagocytosis. (2004) Mol Biol
Cell, 15(8):3509 – 3519.
17. Ueyama T, Tatsuno T, Kawasaki T, Tsujibe S, Shirai Y,
Sumimoto H, Leto TL,and Saito N. A regulated adaptor function of
p40phox: Distinct p67phox mem-brane targeting by p40phox and by
p47phox. (2007) Mol Biol Cell, 18(2):441 –454.
18. Schwartz J, Leidal KG, Femling JK, Weiss JP, and Nauseef WM.
Neutrophilbleaching of GFP-expressing staphylococci: Probing the
intraphagosomal fateof individual bacteria. (2009) J Immunol,
183(4):2632 – 2641.
19. Tlili A, Erard M, Faure MC, Baudin X, Piolot T,
Dupre-Crochet S, and Nusse O.Stable accumulation of p67phox at the
phagosomal membrane and ROS produc-tion within the phagosome.
(2012) J Leukoc Biol, 91(1):83 – 95.
20. Belousov VV, Fradkov AF, Lukyanov KA, Staroverov DB,
Shakhbazov KS, Ter-skikh AV, and Lukyanov S. Genetically encoded
fluorescent indicator for intra-cellular hydrogen peroxide. (2006)
Nat Methods, 3(4):281 – 286.
21. Meyer AJ and Dick TP. Fluorescent protein-based redox
probes. (2010) AntioxidRedox Signal, 13(5):621 – 650.
22. Wang W, Fang H, Groom L, Cheng A, Zhang W, Liu J, Wang X, Li
K, Han P,Zheng M, Yin J, Wang W, Mattson MP, Kao JPY, Lakatta EG,
Sheu SS, OuyangK, Chen J, Dirksen RT, and Cheng H. Superoxide
flashes in single mitochondria.(2008) Cell, 134(2):279 – 290.
23. Bourgeois D, Regis-Faro A, and Adam V. Photoactivated
structural dynamics offluorescent proteins. (2012) Biochem. Soc.
Trans., 40(3):531 – 538.
-
18 Vivien Berthelot et al.
24. Lippincott-Schwartz J and Patterson GH. Photoactivatable
fluorescent proteinsfor diffraction-limited and super-resolution
imaging. (2009) Trends Cell Biol, 19(11):555 – 565.
25. Ha T and Tinnefeld P. Photophysics of fluorescent probes for
single-moleculebiophysics and super-resolution imaging. (2012)
Annu. Rev. Phys. Chem., 63:595 – 617.
26. van Thor JJ, Gensch T, Hellingwerf KJ, and Johnson LN.
Phototransformationof green fluorescent protein with UV and visible
light leads to decarboxylationof glutamate 222. (2002) Nat. Struct.
Biol., 9(1):37 – 41.
27. Piston DW and Kremers GJ. Fluorescent protein FRET: The
good, the bad andthe ugly. (2007) Trends Biochem. Sci., 32(9):407 –
414.
28. Alvarez LA, Merola F, Erard M, and Rusconi F. Mass
spectrometry-basedstructural dissection of fluorescent proteins.
(2009) Biochemistry, 48(18):3810– 3812.
29. Spinks JWT and Woods RJ. An introduction to radiation
chemistry. Wileyinterscience (1990).
30. Lelimousin M, Noirclerc-Savoye M, Lazareno-Saez C, Paetzold
B, Vot SL,Chazal R, Macheboeuf P, Field MJ, Bourgeois D, and Royant
A. Intrinsic dy-namics in ECFP and Cerulean control fluorescence
quantum yield. (2009) Bio-chemistry, 48(42):10038 – 10046.
31. Fraczkiewicz R and Braun W. Exact and Efficient Analytical
Calculation ofthe Accessible Surface Areas and Their Gradients for
Macromolecules (1998) J.Comp. Chem., 19:319 – 333.
32. Rusconi F, Schmitter JM, Rossier J, and le Maire M.
Chromatographic separationand sample preparation in one step for
MALDI mass spectrometric analysis ofsubpicomole amounts of
heterogeneous protein samples (1998) Anal. Chem., 70(14):3046 –
3052.
33. Strohalm M, Hassman M, Kosata B, and Kodicek M. MMass data
miner: Anopen source alternative for mass spectrometric data
analysis. (2008) Rapid Com-mun. Mass Spectrom., 22(6):905 –
908.
34. Strohalm M, Kavan D, Novak P, Volny M, and Havlicek V. MMass
3: A cross-platform software environment for precise analysis of
mass spectrometric data.(2010) Anal. Chem., 82(11):4648 – 4651.
35. Rusconi F and Belghazi M. Desktop prediction/analysis of
mass spectrometricdata in proteomic projects by using massXpert.
(2002) Bioinformatics, 18(4):644– 645.
36. Rusconi F. MassXpert 2: A cross-platform software
environment for polymerchemistry modelling and simulation/analysis
of mass spectrometric data. (2009)Bioinformatics, 25(20):2741 –
2742.
37. Rusconi F. GNU polyxmass: A software framework for mass
spectrometric sim-ulations of linear (bio-)polymeric analytes.
(2006) BMC Bioinformatics, 7:226.
38. Kadlcik V, Sicard-Roselli C, Houée-Levin C, Kodicek M,
Ferreri C, and Chat-gilialoglu C. Reductive modification of a
methionine residue in the amyloid-betapeptide. (2006) Angew. Chem.
Int. Ed. Engl., 45(16):2595 – 2598.
39. Mozziconacci O, Bobrowski K, Ferreri C, and Chatgilialoglu
C. Reactions ofhydrogen atoms with met-enkephalin and related
peptides. (2007) Chemistry, 13
-
Molecular dissection of modified fluorescent proteins 19
(7):2029 – 2033.40. Chatgilialoglu C, Ferreri C, Torreggiani A,
Salzano AM, Renzone G, and Scaloni
A. Radiation-induced reductive modifications of
sulfur-containing amino acidswithin peptides and proteins. (2011) J
Proteomics, 74(11):2264 – 2273.
41. Torreggiani A, Chatgilialoglu C, Ferreri C, Melchiorre M,
Atrian S, and Capdev-ila M. Non-enzymatic modifications in
metallothioneins connected to lipid mem-brane damages: Structural
and biomimetic studies under reductive radical stress.(2013) J
Proteomics.
42. Bachi A, Dalle-Donne I, and Scaloni A. Redox proteomics:
Chemical principles,methodological approaches and
biological/biomedical promises. (2013) Chem.Rev., 113(1):596 –
698.
43. Tang L and Kebarle P. Dependence of ion intensity in
electrospray mass spec-trometry on the concentration of the
analytes in the electrosprayed solution(1993) Anal. Chem.,
65(24):3654 – 3668.
44. Wilm M. Principles of electrospray ionization. (2011) Mol
Cell Proteomics, 10(7):M111 – 009407.
45. Houée-Levin. Radiation Chemistry of Proteins. Elsevier
(2001).46. Davies MJ and Dean RT. Oxford University Press
(1997).47. Blouquit Y, Duchambon P, Brun E, Marco S, Rusconi F, and
Sicard-Roselli C.
High sensitivity of human centrin 2 toward radiolytical
oxidation: C-terminaltyrosinyl residue as the main target. (2007)
Free Radic Biol Med, 43(2):216 –228.
48. Mozziconacci O, Mirkowski J, Rusconi F, Kciuk G, Wisniowski
PB, BobrowskiK, and Houée-Levin C. Methionine residue acts as a
prooxidant in the *OH-induced oxidation of enkephalins. (2012) J.
Phys. Chem. B, 116(41):12460 –12472.
49. Buxton GV, Greenstock CL, Helman WP, and Ross AB.
Critical-Review of RateConstants for Reactions of Hydrated
Electrons, Hydrogen-Atoms and HydroxylRadicals (OH/O) in
Aqueous-Solution (1988) J. Phys. Chem. Ref. Data, 17:513– 886.
50. Lichtin NN and Shafferman A. Reactions of phenylalanine
derivatives, methio-nine derivatives, and their mixtures in aqueous
solution with pulse-radiolyticallygenerated H atoms. (1974) Radiat.
Res., 60(3):432 – 440.
51. Solar S, Solar W, and Getoff N (1984) J. Phys. Chem.,
88(10):2091 – 2095.52. Holcman J, Bobrowski K, Schöneich C, and
Asmus KD. OH-induced oxidation
of cyclo-Met-Met. A search for a complexed OH-radical (1991)
Radiat. Phys.Chem., 37:473–478.
53. Houée-Levin C and Bobrowski K. The use of the methods of
radiolysis to explorethe mechanisms of free radical modifications
in proteins. (2013) J Proteomics,doi
10.1016/j.jprot.2013.02.014.
54. Land EJ and Prutz WA. Reaction of azide radicals with amino
acids and proteins.(1979) Int J Radiat Biol Relat Stud Phys Chem
Med, 36(1):75 – 83.
55. DeFelippis MR, Murthy CP, Faraggi M, and Klapper MH. Pulse
radiolytic mea-surement of redox potentials: The tyrosine and
tryptophan radicals. (1989) Bio-chemistry, 28(11):4847 – 4853.
-
20 Vivien Berthelot et al.
56. Prutz WA. Free radical transfer involving sulphur peptide
functions. PlenumPress New York (1990).
57. Lukacs A, Eker APM, Byrdin M, Brettel K, and Vos MH.
Electron hoppingthrough the 15 A triple tryptophan molecular wire
in DNA photolyase occurswithin 30 ps. (2008) J. Am. Chem. Soc.,
130(44):14394 – 14395.
58. Tarboush NA, Jensen LMR, Yukl ET, Geng J, Liu A, Wilmot CM,
and DavidsonVL. Mutagenesis of tryptophan199 suggests that hopping
is required for MauG-dependent tryptophan tryptophylquinone
biosynthesis. (2011) Proc. Natl. Acad.Sci. USA, 108(41):16956 –
16961.
59. Berges J, de Oliveira P, Fourre I, and Houée-Levin C. The
one-electron reductionpotential of methionine-containing peptides
depends on the sequence. (2012) J.Phys. Chem. B, 116(31):9352 –
9362.
60. Stuart-Audette M, Blouquit Y, Faraggi M, Sicard-Roselli C,
Houée-Levin C, andJolles P. Re-evaluation of intramolecular
long-range electron transfer betweentyrosine and tryptophan in
lysozymes. Evidence for the participation of otherresidues. (2003)
Eur. J. Biochem., 270(17):3565 – 3571.
61. Sicard-Roselli C, Lemaire S, Jacquot JP, Favaudon V,
Marchand C, and Houée-Levin C. Thioredoxin Ch1 of Chlamydomonas
reinhardtii displays an unusualresistance toward one-electron
oxidation. (2004) Eur. J. Biochem., 271(17):3481– 3487.
62. Kadlcik V, Sicard-Roselli C, Mattioli TA, Kodicek M, and
Houée-Levin C. One-electron oxidation of beta-amyloid peptide:
Sequence modulation of reactivity.(2004) Free Radic Biol Med,
37(6):881 – 891.
63. Mozziconacci O, Mirkowski J, Rusconi F, Pernot P, Bobrowski
K, and Houée-Levin C. Superoxide radical anions protect enkephalin
from oxidation if theamine group is blocked. (2007) Free Radic Biol
Med, 43(2):229 – 240.
64. Kiselar JG and Chance MR. Future directions of structural
mass spectrometryusing hydroxyl radical footprinting. (2010) J.
Mass Spectrom., 45(12):1373 –1382.
65. Vahidi S, Stocks BB, Liaghati-Mobarhan Y, and Konermann L.
Mapping pH-induced protein structural changes under equilibrium
conditions by pulsed oxida-tive labeling and mass spectrometry.
(2012) Anal. Chem., 84(21):9124 – 9130.
66. Favaudon V, Tourbez H, Houée-Levin C, and Lhoste JM. CO2.-
Radical inducedcleavage of disulfide bonds in proteins. A gamma-ray
and pulse radiolysis mech-anistic investigation. (1990)
Biochemistry, 29(49):10978 – 10989.
67. Arnold K, Bordoli L, Kopp J, and Schwede T. The SWISS-MODEL
workspace:A web-based environment for protein structure homology
modelling. (2006)Bioinformatics, 22(2):195 – 201.
68. Schwede T, Kopp J, Guex N, and Peitsch MC. SWISS-MODEL: An
automatedprotein homology-modeling server. (2003) Nucleic Acids
Res., 31(13):3381 –3385.
69. Guex N and Peitsch MC. SWISS-MODEL and the Swiss-PdbViewer:
An envi-ronment for comparative protein modeling. (1997)
Electrophoresis, 18(15):2714– 2723.
70. Humphrey W, Dalke A, and Schulten K. VMD: Visual molecular
dynamics.(1996) J Mol Graph, 14(1):33 – 8 – 27 – 8.