Potential of 4-thiothymidine as a molecular probe for H2O2 in systems related to PhotoDynamic therapy: A structuristic and mechanistic insight by UV–visible and FTIR-ATR spectroscopies and by ElectroSpray ionization mass spectrometry

Article abstract of DOI:10.1016/j.molliq.2018.05.073

An investigation, based on the synergy between UV–Vis and FTIR-ATR spectroscopies and ElectroSpray Ionization Mass Spectrometry (ESI-MS), on the reactivity of 4-thiothymidine (S⁴-TdR) with H₂O₂, selected as a Reactive Oxyg

Full text of DOI:10.1016/j.molliq.2018.05.073

Accepted Manuscript
Potential of 4-thiothymidine as a molecular probe for H2O2 in
systems related to PhotoDynamic therapy: A structuristic and
mechanistic insight by UV–visible and FTIR-ATR spectroscopies
and by ElectroSpray ionization mass spectrometry
Vito Rizzi, Ilario Losito, Ramona Abbattista, Paola Fini, Angela
Agostiano, Tommaso R.I. Cataldi, Pinalysa Cosma
PII:
S0167-7322(18)30642-1
doi:10.1016/j.molliq.2018.05.073
MOLLIQ 9128
DOI:
Reference:
To appear in:
Journal of Molecular Liquids
Received date:
Revised date:
Accepted date:
5 February 2018
9 April 2018
16 May 2018
Please cite this article as: Vito Rizzi, Ilario Losito, Ramona Abbattista, Paola Fini, Angela
Agostiano, Tommaso R.I. Cataldi, Pinalysa Cosma , Potential of 4-thiothymidine as a
molecular probe for H2O2 in systems related to PhotoDynamic therapy: A structuristic
and mechanistic insight by UV–visible and FTIR-ATR spectroscopies and by ElectroSpray
ionization mass spectrometry. The address for the corresponding author was captured
as affiliation for all authors. Please check if appropriate. Molliq(2017), doi:10.1016/
j.molliq.2018.05.073
This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.

ACCEPTED MANUSCRIPT
Potential of 4-thiothymidine as a molecular probe for H O in systems related to
2
2
PhotoDynamic Therapy: a structuristic and mechanistic insight by UV-Visible
and FTIR-ATR spectroscopies and by ElectroSpray Ionization Mass
Spectrometry
a
a,b
a
a,c
a,c
Vito Rizzi , Ilario Losito , Ramona Abbattista , Paola Fini , Angela Agostiano , Tommaso R.I.
a,b
a,c
Cataldi and Pinalysa Cosma *
a
Università degli Studi “Aldo Moro” di Bari, Dip. Chimica, Via Orabona, 4- 70126 Bari, Italy
b
Centro Interdipartimentale SMART, Via Orabona, 4- 70126 Bari, Italy
c
Consiglio Nazionale delle Ricerche CNR-IPCF, UOS Bari, Via Orabona, 4- 70126 Bari, Italy
*
Corresponding Author: Pinalysa Cosma
Università degli Studi di Bari “Aldo Moro”
Dipartimento di Chimica,
Via Orabona 4, I-70126 Bari, ITALY
E-mail: pinalysa.cosma@uniba.it.

ACCEPTED MANUSCRIPT
ABSTRACT
An investigation, based on the synergy between UV‐ Vis and FTIR-ATR spectroscopies and
ElectroSpray Ionization Mass Spectrometry (ESI‐ MS), on the reactivity of 4‐ thiothymidine
4
(
S ‐ TdR) with H O , selected as a Reactive Oxygen Species (ROS) potentially related to Photo
2
2
Dynamic Therapy (PDT), was performed. The nucleoside main absorption band in the UV‐ Vis
4
range, centered at 337 nm, was monitored at several reaction times, confirming the S ‐ TdR
reactivity with a process leading to thymidine (TdR), as the main reaction product, detected through
its absorption at 270 nm. Actually, the generation of at least another by‐ product, absorbing at
4
slightly higher wavelengths compared to S ‐ TdR, was inferred from UV‐ Vis spectra. FTIR-ATR
3
and ESI‐ MS, MS/MS and MS measurements were then performed to unveil the nature of these
4
reaction products, evidencing the presence of a hydroxylated S ‐ TdR as important intermediate
reaction product. In particular, ESI-MS analyses suggested that the latter was generated as a couple
4
of isomeric compounds, with the OH linked either to the SH group arising from C =S
7
4
tautomerization or to the methyl carbon (C ) of S ‐ TdR. Interestingly, the generation of a covalent
4
dimer of S ‐ TdR, characterized by the presence of a disulfide (S‐ S) bond, was also clearly
inferred. The evolution of the cited products with reaction time was also be monitored using ESI-
4
MS, confirming the role of hydroxylated S ‐ TdR as an intermediate towards the generation of the
4
ultimate product, TdR. The described processes make S -TdR a promising candidate as a molecular
probe for the detection of H O in PDT-related systems.
2
2
KEYWORDS: 4‐ thiothymidine, ROS, PDT, ESI-MS, thymidine, chemical probe

ACCEPTED MANUSCRIPT
Introduction
Reactive Oxygen Species (ROS) are well known oxygen-related chemical species displaying a
1
,2
chemical reactivity much higher than that of molecular oxygen itself. Endogenously generated
ROS are important to life, being involved in several biological functions, yet their overproduction in
vivo might become deleterious for human health, with possible implications in extremely relevant
3
pathologies, like cancer, atherosclerosis, diabetes and neurodegenerative diseases. As an example,
the exposition to xenobiotics, infectious agents, polluting agent, UV light, smoke and radiation may
4
lead to an abnormal production of ROS. Interestingly, ROS overproduction can be also obtained
purposely, under specific operative conditions, by combining two non-toxic ingredients, i.e., an
appropriate chemical, acting as a photosensitizer (PS) and light, in the presence of oxygen. The
resulting ROS may be subsequently exploited for the selective destruction of pathogenic cells or
tissues, thus leading to an elegant light-activated approach for the treatment of hyperproliferative
5
,6
diseases and/or infections originated by microbial pathogens. In fact, photosensitized oxidations
have been the object of a remarkable interest for chemists and biologists in the last 30 years and
7
several novel therapies based on them have been proposed. Among the latter, Photo Dynamic
Therapy (PDT), currently feasible in several medical institutions around the world, is considered a
widely studied and promising technique for biomedical applications, meant as alternative or
7
adjuvant to classical therapies like radiotherapy, surgery and chemotherapy. The PDT efficiency
8
depends not only on the properties of the compound used as a photosensitizer, but also on the
9
8
characteristics of the irradiated area. As described in the literature, two main reaction pathways
could be followed by PSs during their photoactivation: Type I and Type II mechanisms. The former
is related to the generation, through appropriate reaction pathways, of radical species subsequently
•-
inducing the biomolecule oxidation, like the superoxide radical anion (O ) and the hydroxyl
2
•
radical ( OH); H O is also generated under these conditions. Among other possible ROS, the
2
2

ACCEPTED MANUSCRIPT
1
generation of singlet state oxygen ( O ), in turn formed through energy transfer processes from PS-
2
6
triplet state to molecular oxygen, is contemplated according to a Type II reaction. The ROS
detection and characterization can thus enable a better understanding of the reaction pathway
followed during a PDT process. At the same time, in the context of a pathology related to in vivo-
generated ROS, they may help also in relating a particular cellular signaling event to a specific
1
0
ROS, since ROS have been also shown to play a significant role in cellular signaling regulation.
In recent years, several analytical approaches have been proposed to face the complex analytical
challenge posed by the detection and characterization of ROS, both in vitro and in vivo. Among
4
,11-13
them, the use of fluorescent probes has been reported both for ROS
and for Reactive Nitrogen
, due to the high sensitivity achievable and the relatively simple manipulation.
Unfortunately, the low water solubility of several fluorescent probes can be a significant drawback
1
1,13
Species (RNS)
1
4
in applications to biological systems. Moreover, other factors should be considered in the design
of probes for biological systems, such as a low auto-oxidizability, high photostability and good cell
1
4
permeability.
Starting from these considerations, the search of different analytical approaches for ROS detection
1
was undertaken in our research group. In particular, the possibility of detecting O using molecular
2
,
probes like uric acid and 9,10-diphenyl-anthracene, and hydrogen peroxide and the superoxide
1
5
radical anion, using dichlorofluorescein and ferricytochrome c, respectively was explored. In fact,
these probes could be applied directly in biologically relevant systems and do not imply the use of
sophisticated methods and instrumentation. More recently, the sulphur-containing nucleoside 4-
4
14,16
thio(deoxy)thymidine (S -TdR), exhibiting a high solubility in water, has been taken in account.
A preliminary study based on spectroscopic techniques (UV-VIS, FTIR-ATR and NMR), in
synergy with Electrospray Ionization Mass Spectrometry (ESI-MS), emphasized the remarkable
stability of this species under visible light irradiation, due to its negligible absorbance in the visible
range. This finding assured the possibility of using it as a ROS probe even in PDT therapy-related
systems, involving the use of a visible light-absorbing photosensitizer, since potentially interfering

ACCEPTED MANUSCRIPT
1
4
4
auto-photosensitizing reactions could not occur. In a subsequent study, S -TdR proved to be an
excellent probe for singlet-state oxygen generated in situ through a Type II mechanism, involving
16
the Rose Bengal dye as photosensitizer . In particular, the generation of the (2’-deoxy)thymidine
4
(
TdR) and of a S-S bridged dimeric form of S -TdR (both resulting from the direct interaction
between the latter and singlet state oxygen) was inferred combining spectroscopic techniques and
ESI-MS, with the latter integrated by tandem MS (MS/MS) analyses. Starting from this
4
background, an investigation on the possibility of using S -TdR as an indirect molecular probe, in
aqueous environment, also for another relevant ROS, hydrogen peroxide, has been undertaken in
our laboratories and will be the object of the present paper. As in previous studies, the synergy
between UV-Vis and FTIR-ATR spectroscopies and the ESI-MS technique has been exploited to
4
study in detail the reactivity of H O with S -TdR. In the present case, the typical molecular
2
2
identification capabilities of ESI-MS and MS/MS have been enhanced using a high
resolution/accuracy mass spectrometer, based on the coupling between a quadrupole and an
Orbitrap mass analyzer.
Experimental
Chemicals
4
-thio(deoxy)thymidine and (2’-deoxy)thymidine were purchased from Carbosynth (Compton,
Berkshire, UK). LC-MS grade water and methanol, used as solvents for ESI-MS analyses, and
+
sodium chloride, used as an additive to promote Na -mediated cationization during the ESI process,
were purchased from Sigma-Aldrich (Milan, Italy). The same commercial source was also chosen
for hydrogen peroxide (30%). All the employed chemicals, having 99+% purity, were used as
4
received. S -TdR stock solutions were prepared in doubly distilled water or, in the case of ESI-MS
-3
analyses, in LC-MS grade water, at a concentration of 5.0  10 M and stored in the dark at -20°C
-3
-2
when not in use. Stock hydrogen peroxide solutions (final concentrations: 5.010 M or 1.010
M, according to the case) were prepared by dilution of the commercial product in doubly distilled

ACCEPTED MANUSCRIPT
-2
water. A sodium chloride 5.010 M solution was prepared in LC-MS water and, after proper
+
dilution, used as a source of Na ions, to be added as cationizing agents to reaction mixture aliquots
before ESI-MS analyses.
UV-Vis spectroscopy measurements
In order to investigate the reactivity of the nucleoside towards H O , solutions characterized by
2
2
4
different initial S -TdR/H O molar ratios were prepared in doubly distilled water, at pH 7, for UV-
2
2
Vis spectroscopy measurements. The effect of H O concentration was explored in the range
2
2
-3
-3
4
-3
1
.010 M – 250.010 M, keeping the S -TdR concentration fixed at 1.010 M. UV-Vis spectra
were recorded in a 200 – 450 nm range, at a 1 nm/s scan rate, using a Varian CARY 5 UV-Vis-NIR
spectrophotometer (Varian Inc., now Agilent Technologies Inc., Santa Clara, CA, USA). UV-Vis
spectra were acquired every 10 min or 5 min of reaction, according to the case, in a varying time
interval related to the speed of the oxidation process occurring in solution, but not longer than 120
min. The reaction was always let to occur at room temperature (25 °C). A Quartz cuvette with a 10
mm path length was adopted.
FTIR-ATR spectroscopy measurements
FTIR-ATR analyses were performed, after selected times, on reaction mixtures initially containing
4
-3
-3
-3
S -TdR and H O at concentrations 1.010 M and 5.010 or 250.010 M, respectively. FTIR-
2
2
-
1
-1
ATR spectra were recorded in a 600 – 4000 cm range, at a resolution of 4 cm , using a Fourier
Transform Infrared spectrometer 670-IR equipped with an ATR device (Varian Inc., now Agilent
Technologies Inc., Santa Clara, CA, USA). 32 scans were summed for each acquisition. Before
4
each analysis, an aliquot (50 μL) of the S -TdR reaction mixture was deposited on the surface of the
ATR device and the solvent was left to slowly evaporate before proceeding with the acquisition of
-3
spectra. For the sake of comparison, the FTIR-ATR spectrum related to a 1.010 M solution of
thymidine was also acquired.

ACCEPTED MANUSCRIPT
ElectroSpray-Ionization Mass Spectrometry (ESI-MS) measurements
-3
4
-3
Solutions containing initial concentration equal to 1.010 M, for S -TdR, and 5, 50, 80 or 5010
M for H O , according to the case, were prepared in LC-MS grade water and left at room
2
2
temperature for increasing times to enable the reaction between the two species, so that the
reactivity for different initial concentration ratios of the two reactants could be studied. Aliquots of
the reaction mixtures were withdrawn after 5, 10, 30, 60, 90 min and 120 min of reaction. After
+
each withdrawal an aliquot of the sodium chloride stock solution was added to reach a final Na
-4
concentration of 5.010 M. The sodium ion was introduced to promote the cationization of
thiothymidine and of its reaction products during the subsequent electrospray ionization, since
protonation had been previously found to lead to complex by-products, likely related to a partial
hydrolytic degradation of the nucleoside structure, thus complicating the investigation on
4
14,16
photochemical or chemical processes involving S -TdR.
Before proceeding with the analysis the
solutions were diluted 1:1 (v/v) with LC-MS grade methanol, in order to improve the
electronebulization efficiency during the ESI process. The resulting solutions were infused with a
syringe pump, at a 10 μL/min flow rate, into the ESI interface of the mass spectrometer. Two
different mass spectrometric instrumentations, both from Thermo Scientific (Waltham, MA, USA)
and equipped with an identical HESI (Heated ElectroSpray Ionization) interface, were adopted, in
positive ion mode, during the present investigation. In particular, a VelosPro spectrometer, based on
n
a double stage linear ion trap analyser, enabling MS (n > 2) measurements under low-energy
collisionally induced dissociation (CID) conditions, was adopted for the structural elucidation and
the monitoring of reaction products with time. The following values were adopted for the main
HESI interface and ion optics parameters of the VelosPro spectrometer: sheath gas flow rate, 30
(arbitrary units); auxiliary gas flow rate, 0 (arbitrary units); spray voltage, +6 kV; capillary
temperature, 290°C; S-Lens RF Level, 70 (arbitrary units). These values were obtained after
automatically optimizing those included in a default tuning file provided by Thermo Scientific for

ACCEPTED MANUSCRIPT
4
positive ion ESI-MS measurements at low flow rates, so that the main signal related to S -TdR was
maximized. ESI-MS full scan acquisitions were first performed, in a 50−1000 m/z range. ESI-
MS/MS acquisitions were subsequently obtained on targeted precursor ions (all singly charged),
chosen among those detected in the full scan spectra. In particular, the M + 0 isotopologue of each
precursor ion was isolated before fragmentation, using a 1 m/z units wide window centered on its
3
m/z ratio. The same approach was adopted also for MS measurements, performed on the most
relevant product ions resulting from MS/MS experiments. Collision energy was usually set to 35%
3
both for MS/MS and for MS measurements (note that, in this case, a 400% value corresponds to a
1
00 V excitation voltage). The control of the LC‐ MS instrumentation and the first processing of
data were performed by the Xcalibur software (version 2.2, SP1.48, Thermo Scientific). In order to
confirm the structural elucidation of reaction products achieved using the VelosPro spectrometer,
high accuracy/resolution Fourier transform MS measurements were also performed, on selected
reaction mixtures, using a Q-Exactive mass spectrometer (Thermo Scientific, Waltham, MA, USA),
an hybrid instrument based on the coupling between a quadrupole and an Orbitrap mass analyzer. In
order to guarantee the best possible consistency between the VelosPro and the Q-Exactive
instruments, the HESI interface of the latter was operated under the same conditions described
before. For the same reason, the S-Lens RF level was also identical to the previous one. FTMS
spectra were acquired in the m/z range 50-1000 after setting a mass resolving power of 140000
(measured at m/z 200), an Orbitrap fill-time of 200 ms and the Automatic Gain Control (AGC) level
5
at 510 . The m/z scale of the Q-Exactive spectrometer was periodically calibrated using a solution
containing caffeine, MRFA peptide and a mixture of fluorinated phosphazines (Ultramark),
provided by Thermo Scientific. As shown later, this procedure enabled the achievement of a mass
accuracy usually better than 5 ppm. MS/MS measurements were also performed using the High
Collisional Energy Dissociation (HCD) cell available with the Q-Exactive spectrometer. As
explained before for the VelosPro spectrometer, fragmentation spectra were obtained by isolating
only the M+0 isotopologue of each precursor ion of interest in the quadrupole analyzer of the Q-

ACCEPTED MANUSCRIPT
Exactive spectrometer, using a 1 m/z unit wide isolation window. A Normalized Collisional Energy
(NCE) comprised between 30 and 45 (arbitrary units), according to the precursor ion, was adopted
for fragmentation. The mass spectrometer was operated at a resolving power of 140000 also during
HCD-MS/MS acquisitions; in this case the Orbitrap fill‐ time was set to 100 ms and the AGC value
5
to 510 .
Results and discussion
UV-Vis spectroscopy analysis
4
Spectra referred to aqueous reaction mixtures (pH 7) characterized by different initial S -TdR/H O
2
2
molar ratios were acquired to infer preliminary information about the reactivity of the nucleoside
1
4,16
4
towards hydrogen peroxide. As reported earlier
, S -TdR (Figure 1) exhibits, in aqueous
solution, an intense absorption band in the UV region at around 337 nm, ascribed to a π-π*
transition having a charge transfer character, due to the presence of the chromophore C=S in the
S
7
4
NH
5
6
3
2
1
N
O
HO
O
H
H
H
H
H
OH
4
Figure 1: Chemical structure of S -TdR. The numeration of atoms
related to the pyrimidine ring, as used in the text, is reported.
4
chemical structure. It is worth to mention, that no significant reactivity of S -TdR could be inferred
from UV-Vis spectra, even for relatively high reaction times (120 min), when the initial
-3
concentration of H O was set below 1.010 M (data not shown). While, as shown in Figure 2A,
2
2

ACCEPTED MANUSCRIPT
-3
slight evidences for reactivity were obtained when the initial H O concentration was set at 1.010
2
2
M. Indeed, the absorbance at 337 nm decreased and a shoulder appeared at higher wavelength
values (ca. 370 nm). The observed red shift suggested the presence of a reaction product with a
more extended delocalization of π electrons. When the H O initial concentration was raised to
2
2
-3
4
5
.010 M (see Figure 2B) the S -TdR degradation was clearly more marked and a new band, with
1
4,16
maximum absorbance located at 270 nm, suggesting the formation of TdR, was observed.
-3
Further increasing the H O initial concentrations, like 50.0  10 M, the reaction resulted visibly
2
2
faster, thus spectra were acquired every 10 min, reducing the time interval. As shown in Figure 2C
4
the decrease of the main S -TdR absorption band and the contemporary increase of the shoulder at
higher wavelength were significant, already after 10 min of reaction. However, after 30 min the
signal at higher wavelengths began to decrease with the contemporaneous increase of a peak at 270
4
nm, suggesting that the shoulder of the S -TdR band could be related to an intermediate product
representing a precursor of TdR. This hypothesis was confirmed by spectra recorded at different
-3
reaction times, when the initial concentration of H O was increased to 100.010 M (Figure 2D).
2
2
-3
When a 250.010 M concentration was initially adopted for H O , the shoulder located at ca. 370
2
2
4
nm was so rapidly increased, that a convolution between the S -TdR band and the shoulder was
observed already after 10 min of reaction (Figure 2E). In this case, the TdR 270 nm band was the
only relevant spectral feature observed at wavelengths higher than 240 nm, already after 50 min of
reaction. Overall, the results inferred from the UV-Vis spectra reported in Figure 2 suggested that
consecutive reactions occurred under our experimental conditions, involving, as a first step, the
formation of the intermediate reaction product, mainly dependent on the hydrogen peroxide
concentration, while, as second one, the formation of TdR, that occurred, as a whole, delayed. In
17-19
accordance with mechanisms reported in the literature for thioketones or thiols oxidation,
the
main oxidative pathway leading to the formation of TdR could involve the generation, as an
4
intermediate reaction product, of a hydroxylated form of S -TdR, specifically, a sulfenic acid
4
formed through the hydroxylation of the S -TdR thiolic form, in turn resulting from a

ACCEPTED MANUSCRIPT
4
tautomerization at the level of the C =S bond. It is worth noting that the tautomerization would lead
4
3
4
to the presence of a C =C bond in the pyrimidine ring of S -TdR, thus improving the π electron
delocalization. This process would then support the hypothesis that the initially increasing shoulder,
4
observed at ca. 370 nm in the UV-Vis spectra of reaction mixtures, could be related to a S -TdR
2
0
derivative hydroxylated on the SH group resulting from a tautomerization process. Luo et al.
suggested that, under oxidative conditions, thiols hydroxylated on the SH group could undergo
dimerization to a product with a S-S linkage, by reacting with the non-hydroxylated form.
4
Unfortunately, a direct evidence for the generation of such a dimeric product in S -TdR – H O
2
2
reaction mixtures could not be obtained using UV-Vis spectroscopy, thus more detailed information
was searched for using FTIR-ATR spectroscopy and ESI-MS spectrometry.

ACCEPTED MANUSCRIPT
Figure 2: Comparison between UV-VIS absorption spectra obtained, at different reaction times
4
-3
(
1
see legends), for aqueous solutions (pH 7) initially containing S TdR 1.010 M and: (A) 1.0
-3
-3
-3
-3
-3
0 M, (B) 5.010 M, (C) 50.010 M, (D) 100.010 M and (E) 250.010 M of H O .
2
2
2.5
2.0
1.5
1.0
0.5
0.0
2.5
2.0
1.5
1.0
0.5
0.0
Time / min
Time / min
B
A
0
0
3
0
3
0
60
60
270 nm
2
00
250
300
350
400
450
200
250
300
350
400
450
Wavelength / nm
Wavelength / nm
2
.5
.0
.5
.0
2.5
2.0
1.5
1.0
Time / min
Time / min
0
0
2
1
1
1
2
3
4
0
0
0
0
10
20
30
40
360 nm
50
5
0
6
0
60

ACCEPTED MANUSCRIPT
FTIR-ATR spectroscopy measurements
FTIR-ATR measurements were focused on the possible identification of further products arising
4
from S -TdR upon reaction with H O . Starting from the results of UV-Vis measurements, two
2
2
4
-3
-3
reaction mixtures, initially containing S -TdR 1.010 M and H O at concentration levels 1.010
2
2
-3
M and 250.010 M, respectively, were explored. As a first step, a FTIR-ATR spectrum was
-3
obtained from a 1.010 M solution of the nucleoside in the absence of H O (see Figure 3A, black
2
2
line).
1.00
0.95
0.90
0.85
0.80
T %
-
1
2920 cm
-1
3387 cm
-1
-1
1623 cm
1685 cm
A
8
00
1200 1600 2800 3200 3600 4000
-1
Wavenumber / cm
-
1
1
0
0
0
0
0
.0
.9
.8
.7
3381 cm
T%
-1
1
655 cm
-1
-
1
1
437 cm
3
231 cm
-1
.6 1082 cm
B
.5
8
00
1200 1600 2800 3200 3600 4000
-1
Wavenumber / cm

ACCEPTED MANUSCRIPT
-1
Figure 3. Comparison between detailed views (wavenumber range: 550-4000 cm , with a break in the region
-
1
-3
1
800-2700 cm ) of the FTIR-ATR spectra obtained for: (A) solutions containing, respectively, 1.010 M of
4
-3
S -TdR (black line) and 1.010 M of TdR (gray line) at pH 7; (B) reaction mixtures arising from solutions
initially containing 1.010 M of S -TdR and, respectively, 5.010 M H O (gray line, obtained after 60
minutes of reaction) and 25010 M H O (black line, obtained after 20 minutes of reaction).
-3
4
-3
2
2
-3
2
2
4
The main features already observed for the S -TdR spectrum obtained at a pH close to 7 (see Refs.
-1
-1
1
4 and 16) were found; in particular, the prevailing bands were located at 1685 cm and 1623 cm ,
2
5
6
corresponding to the C =O and C =C bond stretching, respectively. Other detected bands were: (i)
-1
at 2920 cm , related to the C-H stretching of the CH OH group of deoxyribose and probably
2
4
-1
involving also vibrations related to the S TdR methyl group, and (ii) beyond 3000 cm , referred to
2
1
the O-H and N-H stretching in the pyrimidine and deoxy ribose rings. Not surprisingly, even after
6
5
0 min, only slight changes were observed for the reaction mixture initially containing H O
2 2
-3
4
.010 M (see Figure 3B, gray line), for which a low S TdR degradation rate was expected.
2
Actually, a transmittance decrease was observed for the band related to the C =O stretching (1685
-1
-1
cm ), compared to the one observed at 1623 cm . The comparison with the FTIR-ATR spectrum
-3
-1
obtained for TdR at 1.010 M concentration, dominated by a band at 1685 cm (see Figure 3A,
gray line), suggested that the slight accumulation of TdR in the reaction mixture could be
responsible for the observed effect. Moreover, a shift of the O-H/N-H stretching band from 3387
-1
-1
22-23
cm to 3381 cm was observed; starting from investigations already reported for sulfenic acids
,
4
this feature was considered a clue for the hydroxylation of the SH group of some S -TdR molecules
-3
in the reaction mixture. When the initial H O concentration was raised to 25010 M more
2
2
relevant variations of spectral features were observed (see Figure 3B, black line), as expected.
2
5
6
Indeed, bands related to C =O and C =C stretching were not resolved anymore and a unique band,
-1
centered at 1655 cm , was observed, likely due to the prevalence of the already discussed main
1
4-16
TdR-related band
.

ACCEPTED MANUSCRIPT
-1
-1
-1
Moreover, new relevant bands were detected at 1082 cm , 1437 cm and 3231 cm . The former
-1
21
one was close to the 1100 cm band already reported for R-SOH and/or RS=O species , thus
4
22,23
providing a further evidence about the hydroxylation occurring on the SH group of S -TdR.
The
band ascribed also to O-H stretching was significantly modified in shape and a clear contribution
-1
due to an absorption at 3231 cm was found, in accordance with studies dealing with the OH group
2
1
of RSOH compounds. At the same time, the band related to the bending of OH groups, at 1437
-1
cm , was remarkably enhanced, suggesting that an OH group not linked to a sulphur, but rather to a
carbon atom (reasonably, the methyl carbon) could be present on a further hydroxylated derivative
4
21
-1
of S -TdR. Finally, a steady transmittance decrease was observed beyond 900 cm , with a
-1
minimum reached just above 600 cm . This spectral feature has been previously ascribed to the C-S
1
4,16
and C-S-S-C stretching,
thus the generation of a dimer implying the presence of a S-S bridge
4
between two S -TdR molecules could be responsible for these spectral features. Overall, the
synergic use of UV-VIS and FTIR-ATR techniques confirmed the generation of TdR as the ultimate
4
product of the reaction between S -TdR and H O , more evidences for the presence of hydroxylated
2
2
forms and of a covalent dimer of the nucleoside in the reaction mixtures, were obtained. ESI-MS
analyses were thus performed to obtain a further support to the hypothesized reaction pathways.
ElectroSpray Ionization-Mass Spectrometry (ESI-MS) measurements
4
-3
As a first step of the ESI-MS investigation, reaction mixtures initially containing S -TdR 1.010
-3
M and H O 250.010 M were considered, since UV-Vis measurements had suggested a
2
2
significant accumulation of reaction products under these conditions (see Figure 1). Aliquots of the
mixture were thus withdrawn at different times, over a 90 min total reaction time, treated as
described in the Experimental section and then analyzed, under similar ESI and ion optics
conditions, using the VelosPro and the Q-Exactive spectrometers. For the sake of example, the high
resolution mass spectrum obtained using the Q-Exactive spectrometer after 90 min of reaction is
4
reported in Figure 4. As suggested by the corresponding UV-Vis spectrum, the concentration of S -

ACCEPTED MANUSCRIPT
TdR was significantly reduced after such a long time and only a weak signal related to its sodium
1
4,16
adduct, at m/z 281.0565, was found. In accordance with previous measurements,
performed
using a 3D ion trap mass spectrometer, a signal related to the sodium adduct of the nucleoside non
covalent dimer (m/z 539.1238), formed in the ESI source, was also detected.

ACCEPTED MANUSCRIPT
Figure 4. High resolution/accuracy ESI-MS spectrum obtained with the Q-Exactive quadrupole-Orbitrap mass
-3
4
-3
spectrometer for a 110 M solution of S -TdR 30 min after the addition of H O at 250.010 M concentration.
2
2
The reported assignments were based on a comparison with exact m/z ratios (see Table 1 for details). Note that the
m/z ratios or assignments reported for peaks that were found to be related to the same species are characterized by
the same color. See Table 1 and the text for the assignment of the three peaks detected at m/z values lower than 160.
+
[
TdR+Na]
2
65.0792
100
4
+
9
0
0
0
0
0
0
0
0
[HO-S -TdR+H]
2
75.0694
8
7
6
5
4
3
2
1
+
[
256 +H]
[
2
256 +Na]+
79.0949
2
57.1130
+
[
256 + 256 +Na]
5
35.2010
HO-S -TdR+Na]+
4
4
+
[
[S-S bridged S -TdR dimer + Na]
2
97.0514
+
537.1085
127.0504
41.0658
[TdR + 256 +Na]
521.1853
4
[
5
2 S -TdR+Na]+
1
39.1238
[
2 TdR+Na]+
1
59.0223
TdR+H]
5
07.1698
4
+
+
[S -TdR+Na]
[
0
2
81.0565
243.0975
0
1
00
150
200 250
300
350
400
450
500
550
600
650
700
750
m/z

ACCEPTED MANUSCRIPT
4
The detection of S -TdR mainly as sodium adduct, accompanied by low amounts of its non covalent
dimer sodium adduct and of its monomeric proton adduct was confirmed by the ESI-FTMS analysis
-3
of a nucleoside standard solution at 1.010 M concentration (data not shown). Several interesting
further signals were detected in the spectrum reported in Figure 4 and their accurate m/z ratios were
4
exploited to find possible assignments to S -TdR products. In accordance with UV-Vis data, TdR
was the main product, based on the spectral abundance of its sodium adduct (m/z 265.0792), and
4
even a small signal for its protonated adduct (m/z 243.0975) was detected. As observed for S -TdR,
1
4,16
both in our previous studies
and during ESI-MS measurements performed on its solution before
the addition of H O (data not shown), a fragment corresponding to the neutral loss of the
2
2
dehydrogenated form of deoxyribose (127.0504) from protonated TdR was generated in the ESI
source and thus detected in the MS spectrum. Additional peaks detected in the spectrum after 90
minutes of reaction clearly pointed towards the generation of further products, that could be
eventually responsible for the peculiar evolution observed in the UV-VIS spectra under identical
reaction conditions (see Figure 1). First, a peak was detected at m/z 297.0514, a value compatible
4
with the sodium adduct of a hydroxylated form of S -TdR, as suggested by the +16 Da nominal
mass shift occurred with respect to the unmodified form of the latter (indeed, 16 Da correspond to
the replacement of a H atom with a OH group). Interestingly, this derivative had been already
4
14
observed during an investigation on the S -TdR photodegradation. Once again, peaks related to
the corresponding protonated adduct (m/z 275.0694) and to the fragment arising from the already
described in-source loss of dehydro-deoxyribose from the latter (m/z 159.0223) were detected.
Interestingly, the dehydro-deoxyribose loss indicated that the OH group was introduced on the 4-
thiothymine ring, not on the deoxyribose one, a result that will be further discussed later. A weak
but important, from a mechanistic point of view, signal was also observed in the 90 min reaction
mixture spectrum, i.e., the one at m/z 537.1085, a value previously found in our laboratory during
4
4
photochemical experiments on S -TdR and assigned to already discussed covalent dimer of S -TdR,
1
4-16
characterized by the presence of a S-S bond.
As emphasized in Table 1, where the high

ACCEPTED MANUSCRIPT
4
resolution MS-based information on S -TdR and its reaction products has been summarized,
molecular formulas related to the structural assignments described so far were consistent with m/z
data within the mass accuracy expected for the Q-Exactive spectrometer, usually better than 1 ppm.
An additional set of signals was included in the spectrum obtained after 90 minutes of reaction,
namely those labeled with blue characters in Figure 4. Interestingly, two of them, found at m/z
values 257.1130 and 279.0949, seemed to correspond to the proton and sodium adduct,
respectively, of a compound having a nominal mass of 256 Da. As emphasized in the figure, the
sodium adducts of non covalent dimers involving the combination of this compound with itself
(535.2010) and with TdR (521.1853) were also observed, along with the typical product ion
resulting from the protonated form upon the in-source loss of dehydro-deoxyribose (141.0658).
4
These spectral features confirmed that all the cited signals were also related to a derivative of S -
TdR, yet its structural elucidation was not straightforward. The 257.1130 and 279.0949 m/z ratios
were thus used as input values in a search for a possible molecular formula based on the Molecular
Formula finder routine included in the ChemCalc portal, developed at the Institute of Chemical
Sciences and Engineering of the Ecole Polytechnique Federale de Lausanne (Switzerland) and
freely accessible at the Internet address: http://www.chemcalc.org/mf_finder. Along with the
accurate m/z value, the possible intervals for the numbers of atoms of elements included in the
molecular formula were introduced as input. The output consisted in a list of molecular formulas,
ranked in terms of accuracy (expressed in ppm units) with respect to the experimental m/z ratio,
4
Surprisingly, no formula including 10 carbon atoms, like the S -TdR one, was compatible with the
experimental m/z ratios (considering both protonation and sodiation) within the accuracy expected
for the Q-Exactive spectrometer. Indeed, the only reasonable formula for the neutral compound was
4
C H N O , containing 11 carbon atoms and formally obtainable from those related to neutral S -
11
16
2
5
TdR (C H N O S) and TdR (C H N O ) upon replacement, respectively, of a SH or OH group
1
0
14
2
4
10 14
2
5
with a CH O combination of atoms. Consequently, the involvement of methanol, added to the
3
reaction mixture aliquots just before proceeding to ESI-FTMS analysis, was hypothesized to

ACCEPTED MANUSCRIPT
explain the generation of the additional product. This hypothesis was proved by a specific
4
experiment. Two different aliquots of the same S -TdR/H O reaction mixture were withdrawn and
2
2
+
subjected to preliminary cationization with Na , as usual. Afterwards, one of the aliquots was
diluted with
Table 1. Summary of mass spectrometric data and structural information retrieved from the high
resolution/accuracy MS analysis of 4-thiothymidine and of the products of its reaction with H O , performed
2
2
with a Q-Exactive mass spectrometer.
Accurate m/z Exact m/z Accuracy
Molecular formula Compound
a
ratio
ratio
(ppm)
-0.71
-0.77
2
2
81.0565
59.0745
281.0567
259.0747
C H N NaO S
Sodiated 4-thiothymidine
Protonated 4-thiothymidine
1
0
14
2
4
C H N O S
1
0
15
2
4
1
5
43.0275
39.1238
143.0274
539.1241
+0.70
-0.56
C H N OS
5 7 2
Sodiated non covalent
dimer of 4-thiothymidine
C H N NaO S
8 2
2
0
28
4
Na+
S
S
N
N
N
O
N
O
O
O
HO
HO
OH
HO
537.1085
537.1084
+0.19
C H N NaO S
20 26 4 8 2
Sodiated
S-S
bridged
covalent dimer of 4-
thiothymidine
2
2
65.0792
43.0975
265.0795
243.0976
-1.13
-0.41
C H N NaO
5
Sodiated thymidine
1
0
14
2
C H N O
5
Protonated thymidine
1
0
15
2

ACCEPTED MANUSCRIPT
H+
O
NH
1
5
2
27.0504
07.1698
97.0514
143.0502
507.1698
297.0516
+1.40
0.00
C H N O
5 7 2 2
N
H
O
Sodiated non covalent
dimer of thymidine
C H N NaO
10
2
0
28
4
Sodiated
form
of
hydroxylated (on the 4-
-0.67
C H N NaO S
10 14 2 5
thiothymine
ring)
form
4-
of
thiothymidine
Protonated
hydroxylated (on the 4-
275.0694
275.0696
-0.73
C H N O S
10 15 2 5
thiothymine
ring)
4-
thiothymidine
Protonated
form
of
1
2
59.0223
79.0949
159.0223
279.0951
0.00
C H N O S
5 7 2 2
hydroxylated 4-thiothymine
Sodiated form of by-
b
product C H N O
5
1
1
16
2
-0.72
C H N NaO
11 16 2 5
Protonated form of by-
product C H N O
2
1
57.1130
41.0658
257.1132
141.0659
-0.78
-0.71
C H N O
11 17 2 5
1
1
16
2
5
Protonated form of by-
product C H N O after
C H N O
6
9
2
2
11 16
2
5
dehydro-deoxyribose loss

ACCEPTED MANUSCRIPT
Sodiated non covalent
dimer of by-product
C H N O
5
5
35.2010
21.1853
535.2010
521.1854
0.00
C H N NaO
22 32 4 10
1
1
16
2
5
Sodiated non covalent
dimer including by-product
C H N O and thymidine
-0.52
C H N NaO
24 39 4 10
1
1
16
2
5
a
6
Accuracy was calculated as follows: [(accurate m/z ratio – exact m/z ratio)/exact m/z ratio] * 10
b
See text and Section S1 of the Supplementary Material for details.
methanol in the usual 1:1 (v/v) ratio, the other was diluted with water in the same ratio.
Interestingly, signals related to the peculiar compound were found in the spectrum only in the first
case. Moreover, when the methanol-diluted solution was stored for 30 minutes and then re-analyzed
by ESI-MS, signals at m/z 279.0949 and 257.1130 were clearly increased, compared to the one of
sodiated TdR. These results suggested that methanol could react with one of the other species
related to 4-thiothymidine already present in the reaction mixture, leading to the additional by-
product. It is worth noting that no by-product was generated when methanol was used to dilute the
4
standard S -TdR or TdR aqueous solutions before ESI-MS analysis, thus none of these compounds
could be involved as the by-product precursor. As expected, the absence of methanol led to a
general decrease of signals in the ESI-MS spectrum, due to a more difficult generation of charged
droplets, the precursors of final ions in the ESI source. For this reason the possibility of discarding
the use of methanol for ESI-MS analyses was not considered as a reasonable option. Moreover, the
use of another ESI-friendly solvent, acetonitrile, instead of methanol, as organic modifier for the
ESI process was excluded, due to the risk of reaction with H O , a step of the so-called
2
2
Radziszewski reaction, leading to a Reactive Nitrogen Species (RNS), peroxyacetimidic acid, in
2
4
turn able to generate singlet state oxygen upon decomposition. In view of subsequent ESI-MS
4
analyses on the S -TdR/H O system, to be exploited for kinetic purposes, the time running between
2
2

ACCEPTED MANUSCRIPT
methanol addition and ESI-MS analysis of aliquots of reaction mixtures was kept as low as possible
(less than 2 minutes), in order to minimize the generation of the methanol-related by-product, and,
in turn, the alteration of the amounts of its possible precursors. Further information on the by-
product structure was searched for in a subsequent step of the investigation, dedicated to the use of
3
4
MS/MS and MS for the structural elucidation of S -TdR reaction products. It is worth noting that
the similarity, even in terms of relative abundances of different signals, between MS spectra
obtained on the same reaction mixture using the VelosPro and the Q-Exactive spectrometer was
3
assessed preliminarily. Afterwards both instruments could be used, in synergy, for MS/MS and MS
4
analyses of S -TdR and its reaction products.
3
ElectroSpray Ionization-Tandem/Sequential Mass Spectrometry (ESI-MS/MS, MS )
measurements
Tandem MS was first considered for the investigation on the products arising from the reaction
4
between S -TdR and H O . The already cited dehydro-deoxyribose loss was found to be a common
2
2
feature of CID- and HCD-MS/MS measurements, thus confirming the process occurring already in
4
the ESI source on protonated forms of S -TdR and TdR. The higher collisional energy available
during MS/MS acquisitions (compared to in-source fragmentation) made the process possible also
on the corresponding sodiated forms, not involved into any in-source fragmentation during previous
FTMS acquisitions (see Figure 4). Moreover, the neutral loss of dehydro-deoxyribose occurred
twice in the case of the m/z 537.1085 precursor ion, thus confirming its nature of dimeric product
4
related to S -TdR. As emphasized in Figure 4, the protonated form of the dimeric product was not
detected at all in the ESI-FTMS spectrum. This was reasonable, considering that the presence of
+
two deoxyribose rings increased the probability of Na adduct generation, being this cation
2
5
successfully co-ordinated by the deoxyribose OH groups. In order to draw more refined structural
4
3
information on the basic part of the S -TdR and TdR structures, MS measurements were
performed, with the VelosPro spectrometer, on the product ions resulting from the loss of dehydro-

ACCEPTED MANUSCRIPT
4
deoxyribose from sodium or proton adducts of S -TdR and TdR. Interestingly, no fragmentation at
all could be observed, even at high collisional energies, for product ions related to their sodium
4
3
adducts, i.e. ions having nominal m/z ratios equal to 165 (S -TdR) and 149 (TdR). Conversely, MS
spectra acquired on protonated forms of product ions resulting from dehydro-deoxyribose losses,
4
i.e. ions at nominal m/z 143 (S -TdR) and 127 (TdR) exhibited a non negligible fragmentation (see
Figure S1 of the Supplementary Material). It is worth noting that the same patterns were obtained
3
4
when MS analyses were performed on protonated forms of standard S -TdR and TdR and also
when MS/MS measurements were made on the in-source generated m/z 143 and 127 ions. As
apparent from Figure S1, tautomerism involving C=S and C=O groups and protonation played a
key role in the fragmentation, since they were both essential to enable neutral losses of H O and
2
H S. Protonation was also a required preliminary step to enable the loss of NH , which, as shown in
2
3
Figure S1, could potentially occur from both the NH groups of the precursor ions. The more
complex fragmentation pathway observed was the neutral loss of isocyanic acid (HN=C=O), since
it involved the breakage of the thiothymine or thymine ring, according to the case. Not surprisingly,
3
the resulting product ions (nominal m/z 100 and 84) exhibited very low abundances in the MS
4
3
spectra. The scenarios observed for hydroxylated S -TdR when MS/MS and MS measurements
were performed were more complex. Indeed, the usual dehydro-deoxyribose loss was only observed
from its sodium adduct, both under CID- and HCD-MS/MS conditions; the same occurred when the
HCD-MS/MS analysis was performed on its protonated adduct. The CID-MS/MS spectrum of the
latter (nominal m/z 275) exhibited a richer pattern: beside the signal related to the dehydro-
deoxyribose loss (nominal m/z 159), those related to neutral losses of H O (m/z 257), CO (m/z 247),
2
CH OH (m/z 243), H-N=C=O (m/z 232) and CS (m/z 231) were detected. Based on these findings
3
and literature reports on the reaction of ROS with sulphur-containing nuclebases (see, for example,
4
Ref. 18), the presence of two possible structures for hydroxylated S -TdR, namely one bearing the
OH group linked to the SH group resulting from C=S tautomerization, thus corresponding to a
sulfenic acid, and the other resulting from the formal replacement of a CH hydrogen atom with the
3

ACCEPTED MANUSCRIPT
OH group, was hypothesized. The possibility of getting a clue about the location of the OH group
3
on the 4-thiothymine ring was explored through fragmentation, under MS conditions, of the m/z
1
59 ion detected in MS/MS spectra, i.e. of the proton adduct of hydroxylated 4-thiothymine. As
shown in Figure S2 of the Supplementary Material, most product ions detected upon fragmentation
of the m/z 159 ion were compatible with both hypotheses, thus suggesting that both by-products
4
could be formed upon reaction of S -TdR with H O . The previously cited compound with
2
2
3
molecular formula C H N O was finally analyzed by MS/MS and MS ; as shown in Figure S3 of
1
1
16
2
5
the Supplementary Material, along with the usual loss of dehydro-deoxyribose (leading to the
product ion with m/z 141.0), neutral losses of water and methanol were inferred from the MS/MS
spectrum of the compound proton adduct (nominal m/z 257), obtained using the VelosPro
3
spectrometer. The same losses were observed in the MS spectrum obtained for the m/z 141 ion,
shown in the lower part of Figure S3. Notably, the water loss from the m/z 141 ion was increased
significantly, compared to that of methanol, because the process could occur also upon protonation
2
2
of the OH group linked to the C atom, resulting from tautomerization on the C =O group. The
neutral loss of methanol was consistent with the presence of a -OCH side group in one of the
3
structures compatible with the C H N O formula, shown in Figure S3, whereas the water loss,
1
1
16
2
5
observed already in the MS/MS spectrum, suggested the presence, in an alternative structure, of an
OH group linked to the additional carbon arising from methanol (see the structure drawn on the
right in the inset to the upper panel of Figure S3). As for the only signal with an even-numbered
3
m/z ratio, 84.0, detected in the MS spectrum, corresponding to a fragment with a single N atom,
due to the nitrogen rule, only one structure could be hypothesized for it and, as shown in Figure S3,
it was common to both the isomeric structures of the C H N O product. High resolution/accuracy
11
16
2
5
HCD-MS/MS measurements were also performed, using the Q-Exactive spectrometer, on the
sodiated adduct of the C H N O species. However, only the typical dehydro-deoxyribose neutral
1
1
16
2
5
loss was observed in this case. Nonetheless, the m/z value found for the unique product ion
detected, 163.0477, was in excellent agreement with the one predicted upon a dehydro-deoxyribose

ACCEPTED MANUSCRIPT
loss from the structures shown in Figure S3, i.e., 163.0478. As discussed in detail in Section S1 and
Figure S4 of the Supplementary Material, an exploration of the literature confirmed the possibility
that both the structures hypothesized for the C H N O compound could be formed upon addition
1
1
16
2
5
4
of methanol to S -TdR/H O reaction mixtures before ESI-MS analyses.
2
2
Time evolution of reaction products
Preliminary analyses carried out to draw structural information clearly evidenced that significant
4
variations of ESI-MS signals related to S -TdR and its products occurred when aliquots withdrawn
4
at different reaction times were analyzed. Among easily inferred features, a fast decrease of the S -
-3
TdR signal and a contemporary increase of the TdR one were observed when a 250×10 M initial
4
concentration of H O was adopted. A signal due to hydroxylated S -TdR and a weak, but not
2
2
negligible, signal due to the S-S bridged covalent dimer of the nucleoside were also observed. In
order to study these evolutions more systematically and appreciate the effect of H O concentration,
2
2
4
-3
reaction mixtures initially containing S -TdR at 110 M concentration and H O at four different
2
2
-3
concentrations, namely 5, 50, 80 and 25010 M, were considered. The reaction was let to proceed
for a maximum time of 90 min and aliquots were withdrawn after 5, 30, 60 and 90 min, subjected to
+
cationization with Na , as described in the experimental section, diluted 1:1 (v/v) with LC-MS
grade methanol and then analyzed by ESI-MS. For the reasons mentioned in the previous section,
the ESI-MS analysis was performed after no more than 2 min from the methanol addition, in order
4
to minimize the inevitable generation of the by-products due to the methanolysis of S -TdR
hydroxylated on the S atom (see Figure S4), a process clearly not related to the reaction with H O
2
2
4
but able to alter the actual response of hydroxylated S -TdR. ESI-MS analyses on reaction mixture
+
aliquots were performed, after Na -cationization and dilution with methanol, by direct infusion into
the VelosPro mass spectrometer at a 10 μL/min flow rate. The linear ion trap spectrometer was
adopted since it provided high sensitivity and, at the same time, suitable acquisition speed. Since all
reaction products had already been characterized before, the high mass resolution and accuracy

ACCEPTED MANUSCRIPT
provided by the Q-Exactive spectrometer were not required at this stage. Moreover, since the
4
maximum time span monitored for the reaction of S -TdR with H O (90 min) was not much higher
2
2
than the time reasonably required for an LC-MS run performed on a single reaction aliquot, the
4
chromatographic approach was not considered suitable to separate S -TdR and its reaction products
and monitor them at different reaction times. Actually, the further interaction, during a LC
4
separation, of a relevant reaction product like hydroxylated S -TdR with a typical LC-MS organic
solvent like methanol and the risk of interference from singlet state oxygen related to the eventual
use of acetonitrile as the organic modifier for the LC-MS experiment, would have determined a
significant alteration of the actual distribution of reaction products, when inferred from LC-MS
data. On the contrary, ESI-MS spectra resulting from the infusion of reaction aliquots appropriately
prepared for the analysis provided a fast representation of the distribution of products present in
each aliquots, as already emphasized by Figure 4. For each analyzed aliquot mass spectra were
averaged for a 2 min interval, then the spectral intensities related to the main isotopologues of each
reaction product were considered to calculate ESI-MS response ratios with respect to the intensity
4
of the m/z 281.0 ion, i.e., the one related to residual S -TdR. Such ratios could not be used to infer
the corresponding molar ratios, since the ESI yields for reaction products (apart from TdR) could
not be estimated, due to the lack of standard analogues. Nonetheless, the response ratios could be
conveniently used to make comparisons between different reaction times, for identical initial
concentrations of the reactants, and between different initial concentrations, for the same reaction
times. Indeed, these ratios helped in minimizing the effects of response fluctuations that might
occur between different ESI-MS measurements and thus make the use of absolute signal intensities
rather unreliable.

ACCEPTED MANUSCRIPT
6
1
1
1
0
9
8
7
6
5
4
3
2
1
0
-
3
5 x 10 M
-
3
5
4
3
2
1
0
5 x 10 M
-
3
5
0 x 10 M
-3
5
0 x 10 M
0
20
40
60
80
100
0
20
40
60
80
100
reaction time / min
reaction time / min
5,0
4,5
4,0
3,5
3,0
2,5
2,0
1,5
1,0
0,5
0,0
0
,12
,10
-3
-3
5
x 10 M
5 x 10 M
-
3
-3
0
5
0 x 10 M
50 x 10 M
0,08
,06
0,04
0
0
,02
,00
0
0
20
40
60
80
100
0
20
40
60
80
100
reaction time / min
reaction time / min
4
Figure 5. Plots of ESI-MS signal ratios, calculated with respect to the one of S -TdR (nominal m/z 281),
obtained for products detected (as sodium adducts) and characterized at different times in reaction mixtures
4
4
-3
-3
-3
4
initially containing S -TdR S -TdR 1.0 10 M and H O 5.010 and 5010 M: hydroxylated S -TdR
2
2
(
nominal m/z 297), its by-product with molecular formula C H N O (nominal m/z 279, see Figure S3 of
11 16 2 5
the Supplementary Material for its possible structures), TdR (nominal m/z 265) and the S-S bridged dimer of
4
S -TdR (nominal m/z 537). Error bars express standard deviations obtained for three replicates of each
reaction. See text for details.
The compensation of between-run signal fluctuations was confirmed by the good reproducibility
4
observed for product/S -TdR ESI-MS response ratios obtained for a specific time from three
4
replicates of the S -TdR – H O reaction for each combination of the initial conditions. Plots of
2
2
4
such ratios for the three reaction products characterized before, i.e., hydroxylated S -TdR, TdR and
4
the S-S bridged covalent dimer of S -TdR, are reported in Figures 5 and 6, each referred to a couple
of different initial H O concentrations.
2
2

ACCEPTED MANUSCRIPT
8
7
6
5
4
3
2
1
0
-
3
1
1
1
1
6
4
2
0
8
6
4
2
0
8
0 x 10 M
-3
80 x 10 M
-
3
2
50 x 10 M
-3
250 x 10 M
0
20
40
60
80
100
0
20
40
60
80
100
reaction time / min
reaction time / min
3
0
5
0
5
0
5
0
-3
-3
0
,4
8
0 x 10 M
80 x 10 M
-
3
-3
2
2
1
1
2
50 x 10 M
250 x 10 M
0,3
0,2
0
,1
,0
0
0
20
40
60
80
100
0
20
40
60
80
100
reaction time / min
reaction time / min
2
1
1
1
1
1
0
8
6
4
2
0
8
6
4
2
0
-
3
80 x 10 M
-
3
2
50 x 10 M
0
20
40
60
80
100
reaction time / min