Exponential Molecular Amplification by H2O2-Mediated Autocatalytic Deprotection of Boronic Ester Probes to Redox Cyclers

Article abstract of DOI:10.1002/chem.201900627

Herein, a new molecular autocatalytic reaction scheme based on a H₂O₂-mediated deprotection of a boronate ester probe into a redox cycling compound is described, generating an exponential signal gain in the presence of O₂ and a reducing agent or enzyme. For such a purpose, new chemosensing probes built around a naphthoquinone/naphthohydroquinone redox-active core, masked by a self-immolative boronic ester protecting group, were designed. With these probes, typical autocatalytic kinetic traces with characteristic lags and exponential phases were obtained by using either UV/Visible or fluorescence optical detection, or by using electrochemical monitoring. Detection of concentrations as low as 0.5 μm H₂O₂ and 0.5 nm of a naphthoquinone derivative were achieved in a relatively short time (<1 h). From kinetic analysis of the two cross-activated catalytic loops associated with the autocatalysis, the key parameters governing the autocatalytic reaction network were determined, indirectly showing that the analytical performances are currently limited by the slow nonspecific self-deprotection of boronate probes. Collectively, the present results demonstrate the potential of this new exponential molecular amplification strategy, which, owing to its generic nature and modularity, is quite promising for coupling to a wide range of bioassays involving H₂O₂ or redox cycling compounds, or for use as a new building block in the development of more complex chemical reaction networks.

Full text of DOI:10.1002/chem.201900627

A Journal of
Accepted Article
Title: Exponential Molecular Amplification by H2O2-mediated
Autocatalytic Deprotection of Boronic Ester Probes to Redox
Cyclers
Authors: Justine Pallu, Charlie Rabin, Geordie Creste, Mathieu Branca,
François Mavré, and Benoit Limoges
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201900627
Link to VoR: http://dx.doi.org/10.1002/chem.201900627
Supported by

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
Exponential Molecular Amplification by H2O2-mediated
Autocatalytic Deprotection of Boronic Ester Probes to Redox
Cyclers
Justine Pallu, Charlie Rabin, Geordie Creste, Mathieu Branca,* François Mavré,* Benoît Limoges*,[a]
Abstract: Herein, we describe a new molecular autocatalytic
reaction scheme based on a H2O2-mediated deprotection of a
boronate ester probe into a redox cycling compound, generating an
exponential signal gain in the presence of O2 and a reducing agent
or enzyme. For such a purpose, new chemosensing probes built
around a naphthoquinone/naphthohydroquinone redox-active core,
masked by a self-immolative boronic ester protecting group, were
designed. With these probes, typical autocatalytic kinetic traces with
characteristic lags and exponential phases were obtained using
either a UV-visible or fluorescence optical detection, or also using an
electrochemical monitoring. Detection of concentrations as low as
0.5 µM H2O2 and 0.5 nM of a naphthoquinone derivative were
achieved in a relatively short time (< 1 hr). From kinetic analysis of
the two cross-activated catalytic loops associated to the
autocatalysis, the key parameters governing the autocatalytic
reaction network were determined, indirectly showing that the
analytical performances are currently limited by the slow nonspecific
self-deprotection of boronate probes. Collectively, the present results
demonstrate the potential of this new exponential molecular
amplification strategy, which, due to its generic nature and
modularity, is quite promising for coupling to a wide range of
bioassays involving H2O2 or redox cycling compounds, or for being
used as a new building block in the development of more complex
chemical reaction networks.
amplifications originate from autocatalytic reaction schemes in
which one of the products catalyzes its own formation,6
a
process that is facilitated with nucleic acids thanks to the rich
panoply of biological reagents (e.g., polymerases, nucleases,
…) available to selectively replicate DNA or RNA,4, 5, 7 or even
self-replicate RNA through cleverly designed aptazyme.8
However, the autocatalytic concept of target replication or self-
replication is much challenging to transpose to non-nucleic acid
such as proteins or small organic molecules.9 This is particularly
true for proteins, even though the trypsinogen to trypsin
conversion may be considered as a self-replicative process.10
For small organic molecules, apart from the enzyme-based
ATP/ADP self-amplifying substrate cycle,11, 12 self-replication has
only been recently demonstrated through the emergence of a
range of new molecular-based autocatalytic concepts such as
the dendritic chain reactions,13-19 self-propagating thiolate-
disulfide exchanges,20, 21 activation of a supramolecular allosteric
catalyst,22 or photoinduced amplification of a radical chain
reaction.23 These different strategies have been established for
the detection of small analytes such as fluoride16, 18, 19 or chloride
20, 21
15, 17, 23
anions,22 thiols,14,
or H2O2.13,
Among these small
analytes, hydrogen peroxide is of particular interest because it
plays an important role in a variety of biological processes or
diseases, such as cell signaling, immune response, circadian
redox oscillations, Alzheimer diseases, and cancer.24, 25 Building
new signal amplification systems based on a H2O2-triggered
autocatalytic reaction is therefore of great interest not only for
analytical purposes, but also to better understand some
biological processes and to develop new molecular information
processing systems for synthetic biology.10, 12, 20, 26, 27
Introduction
Molecular and biomolecular amplifications are of upmost
importance in the transduction of (bio)chemical events into
sensitive output signals, with many implications in areas ranging
Herein, we describe
a new molecular-based autocatalytic
reaction network (or autocatalytic set 28) that leads to the
sensitive detection of H2O2 through an exponential signal gain
engendered by a H2O2-mediated autocatalytic deprotection of a
chemosensitive probe (pro-RCC) into a redox cycling compound
(RCC), as depicted in Scheme 1A. The main originality of this
reaction scheme is the coupling between two catalytic loops. A
first one leading to the release of a strongly reducing redox
cycling compound (RCCRed) from the reaction of H2O2 with the
pro-RCC probe, which then undergoes autoxidation in the
presence of dissolved O2 to give the oxidized form RCCOx and
H2O2, thereby closing a first catalytic loop (Loop 1) by further
reacting with a new molecule of pro-RCC (Scheme 1C). A
second one where RCCRed is regenerated from RCCOx by the
direct action of a reducing chemical agent (S) or indirectly
through a reducing enzyme (Enz) able to catalyze the electron
transfer between the substrate S and RCCOx (Scheme 1D). As
long as S and O2 are present in solution, H2O2 is uninterruptedly
produced by this second catalytic loop (Loop 2).
3
from cell signaling1 to in vitro diagnostic.2, Among the wide
variety of molecular amplification schemes, those based on non-
linear signal-gain are undeniably the most powerful in terms of
sensitivity and dynamic range. An emblematic example is the
exponential molecular amplification of nucleic acids by PCR, a
technique that was proven decisive in achieving the ultimate
specific quantification of a few target nucleic acid sequences in a
5
few tens of microliters.4,
Generally speaking, exponential
[a]
J. Pallu, Dr. C. Rabin, G. Creste, Dr. M. Branca, Dr. F. Mavré, Dr. B.
Limoges
Laboratoire d’Electrochimie Moléculaire, UMR 7591 CNRS
Université Paris Diderot, Sorbonne Paris Cité
15, rue Jean-Antoine de Baïf, F-75205 Paris Cedex 13, France
E-mail: limoges@univ-paris-diderot.fr ; francois.mavre@univ-paris-
diderot.fr ; mathieu.branca@univ-paris-diderot.fr
Supporting information for this article is given via a link at the end of
the document.
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
Scheme 1. (A) Principle of the H2O2-mediated autocatalytic deprotection of pro-RCC to RCC in the presence of O2 and a reducing substrate S, the reaction with S
being possibly catalysed by an enzyme Enz. (B) Boronate-based probes (pro-RCC) used in this work. (C) Principle of the H2O2-mediated catalytic deprotection of
pro-RCC to RCC. (D) Principle of the redox cycling of RCC.
The whole process (Scheme 1A), depending on the relative
rates of the two loops, can then produce an overstoichiometric
quantity of H2O2 from a single deprotection event, which is
equivalent to the following global reaction:
compound in order to obtain a pro-RCC probe capable to rapidly
and selectively deprotect in the presence of H2O2 and, at the
same time, to ensure an efficient masking of the RCC redox
activity, (ii) to identify a RCC with good redox cycling abilities
(i.e., by choosing a redox-active compound able to reversibly
and rapidly cycle in the presence of a reducing agent and
dissolved O2), (iii) to use a reducing chemical agent having a
sufficiently strong reducing power towards the electron acceptor
RCCOx and no cross-reactivity with O2, or, in the case of an
enzyme catalyzed reduction, to take advantage of an
oxidoreductase able to efficiently and selectively reduce RCCOx
but unable to generate O2- or H2O2 from O2, and last (iv) to find
reaction conditions that ensure a good chemical stability and
favorable reactivity of the assorted partners and intermediates.
Oxidative deprotection of aromatic boronates by H2O2 is a well-
known biorthogonal reaction to generate phenols,29 which,
thanks to its good chemoselectivity and capacity to operate
efficiently under mild or near-neutral aqueous media (either
k
H2O2
n H2O2
(1)
that becomes naturally autocatalytic when n > 1.
Alternatively, the proposed autocatalytic reaction network can be
viewed as two interconnected catalytic loops that mutually
activate (or catalyze) each other, wherein Loop 1 relies on H2O2
as a catalyst and produces RCC, while Loop 2 relies on RCC as
a redox catalyst and produces H2O2. The symmetric nature of
the reaction scheme therefore implies that this process can be
indifferently triggered by any of the cycling partners (i.e., H2O2 or
RCC). Once started, the reaction is expected to generate with
time an exponential accumulation of both H2O2 and RCC (as
well as P) until pro-RCC and/or S or O2 are eventually fully
consumed. This novel strategy conceptually differs from the
dendritic chain reactions previously reported,13-18 since it does
not require the design of a dendritic probe functionalized by
multiple self-immolative protecting groups, the latter of which is
often synthetically challenging. In the present work, the use of
rationally designed pro-RCCs (Scheme 1B) in an either all-
molecular or enzyme-aided format proves the relevance of this
conceptually new molecular amplification strategy, as well as its
versatility in assays involving either detection of H2O2 or
screening/detection of molecular compounds with redox cycling
abilities.
31,
directly30 or indirectly through self-immolative spacers18,
32),
has been demonstrated as a reaction of choice in the design of
a
myriad of boronic acid- or boronic ester-based H2O2-
chemosensing probes.30, 33-36
From there, the next criterion is to select a suitable RCC with
effective redox cycling properties. We thus postulated that redox
cyclers bearing a quinone center belonging to the family of
naphthoquinones (NQs) would be particularly suitable since they
were shown to readily redox cycle in the presence of O2 and a
reductant37 or a reductase,38, 39 a property at the origin of their
41, 42
wide utilization as potent therapeutic cytotoxic agents.40,
Indeed, their reduction potential is sufficiently low relative to
Results and Discussion
benzoquinones to normally ensure fast autoxidation by
44, 45
molecular oxygen,43,
and high enough compared to
1.
Design of the autocatalytic set
anthraquinones to undergo efficient chemical or enzymatic
reduction.
Having in mind the use of naphthoquinones as RCCs, we looked
for chemical and enzyme reducing agents that would efficiently
To demonstrate the feasibility of our approach, we first had to
consider the following criteria in the design of the autocatalytic
set: (i) to select an appropriate protection strategy of the RCC
2
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
react with the 1,4-naphthoquinone (1,4-NQ). Ascorbate has
In order to determine the deprotection rate of 1 by H2O2 we have
followed the reaction by fluorescence (the boronic probe 1
exhibiting a marked fluorescence at 465 nm in contrast to 1,4-
NQH2 or 1,4-NQ which do not fluoresce – see Figure S3A). The
time-course decreases of fluorescence emitted by probe 1 as a
function of different concentrations of H2O2 are reported in
Figure S3B. In agreement to the catalytic nature of Loop 1 all
fluorescence kinetic traces converge to a value close to zero
46-48
been first considered as a chemical reducing agent.37,
Alternatively, the thermostable FMN-dependent diaphorase (DI)
from the Bacillus stearothermophilus (EC 1.6.99.-) has been
envisioned as a reducing enzyme thanks to its ability to
efficiently catalyze the reduction of a wide range of NQs by
NADH. Interestingly, a rate constant as high as 109 M-1 s-1 has
been reported for the reduction of menadione by DI at pH 8.5,49
emphasizing the high reducing rate of this enzyme towards NQs.
according to
a monoexponential decay, indicating again
A further advantage of DI is its absence of cross-reactivity with
complete transformation of probe 1 into 1,4-NQ, even for
substoichiometric ratios of H2O2. Fitting the experimental curves
with a single exponential function allowed extracting kobs, which
once plotted as a function of [H2O2] leads to a linear relationship
(Figure S3C) from which the second order deprotection rate
constant k1 could be determined. A value of k1 = 3 ± 1 M-1s-1 was
obtained which is close to the deprotection rate constant of 5 ± 1
M-1s-1 that can also be recovered from analysis of the initial rates
of absorbance increase in Figure S2 (corresponding thus in this
case to an indirect evaluation of k1 from the growth rate of 1,4-
NQ). This value is typically within the range of kinetic rate
constants previously published for H2O2-mediated deprotection
of arylboronic esters at neutral or slightly basic pHs (i.e., rate
constants ranging from 0.07 to 4 M-1s-1).56-60
50
O2.49,
As an additional reducing enzyme, we have also
considered the pyrroloquinoline quinone-dependent glucose
dehydrogenase (PQQ-GDH) because of its high catalytic activity
towards glucose oxidation and concomitant reduction of a broad
range of quinones,51 and, as for DI, an absence of reactivity with
O2.
Finally, a slightly alkaline 0.1 M Tris buffer of pH 8.5 was chosen
because the oxidative deprotection of arylboronates by H2O2
and autoxidation of RCCred by O2 are both faster at higher pHs
-
(the nucleophile HO2 being the reactant involved in the
52
chemoselective deprotection of arylboronates,29,
and the
doubly deprotonated hydroquinone being the more reactive form
for O2 reduction owing to a higher driving force44). Moreover, this
buffer and pH value were previously shown optimal for the
catalytic reduction of a wide range of electron acceptors by DI in
the presence of NADH.49, 50 Besides, EDTA was added to the
reaction solution to scavenge and prevent any deleterious
effects of adventitious traces of heavy metal cations, the latter
being known to catalyze the oxidation of NQH2s by O2 and to
produce reactive oxygen species.53, 54
The first pro-RCC compound we have investigated is the probe
1 (pro-NQH2) shown in Schemes 1B and 2 (see Experimental
Section for chemical synthesis), wherein an aromatic boronate
protecting group has been installed, via a self-immolating spacer,
in place of one of the hydroxyl group of the 1,4-
naphthohydroquinone. Oxidative cleavage of the boronic ester
group in 1 by H2O2 is thus assumed to produce an intermediate
that rapidly self-immolate (through a 1,6-elimination) into the
desired 1,4-NQH2, which in the presence of O2 then leads to 1,4-
NQ and H2O2 (Scheme 2). Under the selected conditions, probe
1 was found relatively stable in solution (Figure S1), showing,
after fast hydrolysis of the pinacol ester (leading to a small drop
in absorbance during the first 10 minutes),55 no significant
absorbance change over at least 2 hrs incubation in the Tris
buffer. Moreover and as expected, after adding one equivalent
of H2O2, probe 1 is fully converted into 1,4-NQ in less than 2 hrs
(Figure S1). More interestingly, substoichiometric ratios of H2O2
can fully convert 1 into 1,4-NQ (Figure S2), highlighting the
catalytic nature of H2O2 in the chemical transformation of
compound 1. This behavior strongly supports the mechanism of
Loop 1 in Scheme 2. These results also show that probe 1 is an
a priori worthy candidate to explore the proposed concept of
exponential molecular amplification.
Scheme 2. Full autocatalytic reaction scheme involving probe 1 (the undesired
nonspecific self-deprotection of 1 is also included – dashed arrow).
It is important to note that the H2O2-mediated deprotection of 1 is
a two-step process wherein the second step of 1,6-elimination, if
not fast, can be rate-limiting.31 The fact that in Figure S3B the
overall curves fit well to a monoexponential decay and also that
in Figure S3C there is a perfect linear relationship between kobs
and [H2O2], even for substoichiometric [H2O2]/[1] ratios, strongly
supports a liberation rate of the spacer (i.e., k’1) that is never
rate-determining under our conditions. This would also mean
Separately, we have checked whether the stability of the
generated 1,4-NQ in solution is satisfying or not. Figure S1 show
that the UV-vis spectra of a freshly prepared Tris buffer solution
of 1,4-NQ evolves very slowly with time into a new one (as
attested by the isobestic points at  = 236 nm, 262 nm, 325 nm
and 361 nm), a transformation that appears only marginally
accelerated in the presence of H2O2.
that k’1 is most likely > 0.001 s-1, a value which agrees with that
61
reported for analogous self-immolative spacers.31,
The two
other assumptions we have implicitly made for the determination
of k1 is that the autoxidation rate of 1,4-NQH2 by O2 is fast
relative to the other reaction steps in Loop 1 and it leads to a
3
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
stoichiometric production of H2O2, which is reasonable given
what has been reported in a previous study of the 1,4-NQH2
autoxidation at neutral pH.45 In order to determine how fast is the
autoxidation of 1,4-NQH2 under our experimental conditions, we
performed cyclic voltammetric experiments under redox-
mediated catalysis (Figure S4). From analysis of the
electrocatalytic current response, an apparent rate constant of k2
= 3300 ± 300 M-1s-1 was determined. This value confirms that,
whatever the experimental conditions, the oxidation rate of 1,4-
NQH2 by O2 is fast enough to be none rate-limiting in Loop 1.
Another interesting information we can recover from
experiments in Figure S3 is the slow self-deprotection rate of 1
in the absence of H2O2, an undesired process which is assumed
to lead to the nonspecific release of 1,4-NQ but also possibly to
the less problematic release of the boronic group by
protodeboronation.62, 63 An apparent first-order rate constant of 4
× 10-5 s-1 was deduced.
Finally, taking into account the overall rate constants which
characterize the catalytic mechanism of Loop 1 (i.e., k1, k’1, k2,
k0), we were able from numerical simulations of kinetics to
reproduce quite well the experimental UV-Vis kinetic traces
characterizing the growth rate of 1,4-NQ from deprotection of
probe 1 (Figure S5), a result which finally allows for a rational
quantitative prediction of Loop 1. From an analytical point of
view, probe 1 can be directly used to reveal the presence of the
catalyst/trigger of Loop 1, i.e. H2O2. To assess the ability of
probe 1 to sensitively detect H2O2 under the catalytic conditions
of Loop 1, the calibration curve of H2O2 was plotted by reporting
the absorbance in Figure S2 for 1 hour reaction as a function of
H2O2 concentration. From this calibration plot, an H2O2 detection
limit of 1.5 µM can be estimated.
H2O2 to the micromolar range (at lower concentrations all tc
values converge to a same limit of 31 min – see Figure 1D).
Understanding the origin of this nonspecifically triggered
response is therefore of crucial importance to improve the
analytical performances of the method, an issue that is
reminiscent of exponential molecular amplifications.64 It is likely
that, in our case, the nonspecific response predominantly arises
from the moderate stability of 1 that slowly self-deprotects in
aqueous solution to cause the undesired leakage of 1,4-NQH2
(k0  10-5 s-1). Another plausible contribution is the slow cross
reactivity between O2 and Asc that can lead to the release of
-
O2 and so indirectly to the slow production of H2O2 by either
dismutation or cross-reactions with reducing species.65-67
With the aim to rationally predict the sigmoidal kinetic traces
obtained in Figure 1A and to better highlight the validity of the
autocatalytic reaction mechanism postulated in Scheme 2, we
completed our analysis of rate constants by an independent
kinetics characterization of Loop 2. The time-course of this redox
cycling reaction was monitored spectrophotometrically through
the absorbance decrease of Asc at 262 nm as a function of
different 1,4-NQ and Asc concentrations in an air-saturated Tris
buffer (pH 8.5). From analysis of the experimental kinetic traces,
a k3 rate constant value of 170 M-1 s-1 was estimated (see
Supporting Information for details). Compared to the apparent
autoxidation rate constant k2 of 3300 M-1 s-1 independently
determined by cyclic voltammetry, the magnitude of k3 here
definitely establishes that the redox cycling process is under
kinetic control of the reduction of 1,4-NQ by Asc.
An additional reaction that was necessary to evaluate as a
control but also for its potential influence on the nonspecific
response of the system is the autoxidation of ascorbate in an air-
saturated buffer. Monitoring the UV-vis absorbance change of
ascorbate solutions in Tris buffer (pH 8.5) allowed us to
demonstrate a slow zeroth order kinetics autoxidation of Asc
(Figure S8), leading to an apparent first order autoxidation rate
of 2 × 10-5 s-1 (a value which is in line to those previously
published65, 67). Assuming that the O2 concentration remains
constant (i.e., 0.25 mM at 1 atm and 25°C), a second order rate
constant of k’0 = 0.08 M-1s-1 could thus be estimated. This slow
autoxidation of ascorbate does not mean it necessarily leads to
a significant production of H2O2 and so to a large contribution to
the nonspecific triggering of the autocatalytic process. In order to
evaluate the degree to which this autoxidation may affect or not
the nonspecific response, Asc solutions (0.5 mM in Tris buffer)
were allowed to age between 15 min and 3 hrs before being
tested with probe 1 in the presence or absence of one
equivalent H2O2. The results we obtained (not shown)
demonstrate that, regardless of the ageing time, the
autocatalytic responses are identical to those recovered from a
freshly prepared Asc solution, thus definitively eliminating the
2.
Triggering the autocatalytic reaction with H2O2
Switching from a catalytic (Loop 1) to an autocatalytic (Loop 1 ×
Loop 2) reaction scheme requires only the addition of a reducing
agent in solution. As explained in the introduction, the reduction
of 1,4-NQ can be achieved either chemically (using ascorbate)
or enzymatically (using diaphorase/NADH or PQQ-GDH/glucose
as enzyme/substrate couples), two different strategies we have
investigated.
2.1. Autocatalytic set with ascorbate as reducing agent
Ascorbate (Asc) was first tested as a chemical reducing agent of
1,4-NQ. The reaction between
1 (50 µM) and different
concentrations of H2O2 (ranging from 0 to 150 µM) has thus
been characterized in the presence of an excess ascorbate (0.5
mM). The kinetic measurements were achieved with a kinetic
microplate reader for high throughput screening, wherein the
reaction solutions (200 µL working volume) distributed in the
microplate were left open to the air during the whole process.
Figure 1A shows a typical set of kinetics traces, the sigmoidal
shapes of which are experimental evidences for an autocatalytic
behavior.6 The autocatalytic nature of these kinetics is also
supported by the linearity (over more than one decade) found
between the threshold time tc (defined here as the time required
to reach 50% of amplitude signal change) and the H2O2
concentration in a semi-logarithmic scale (Figure 1D). In the
absence of H2O2, the prevalence of a nonspecifically triggered
autocatalytic reaction severely restricts the detection limit of
ascorbate autoxidation as
response.
a main source of nonspecific
Knowing the absolute values of all of the rate constants involved
in Loop 1 and Loop 2, it was then interesting to numerically
simulate the autocatalytic kinetics traces resulting from the
combination of these two cross-activated catalytic loops (see
Supporting Information for details). Remarkably, a quite good
correspondence between the simulated and experimental kinetic
traces was obtained using the following set of parameters:
4
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
A
B
3.5
3.0
2.5
2.0
1.0
0.8
0.6
0.4
0.2
0.0
ascorbate
NADH/DI


1.5
1.0
0.5
0.0
0
10 20 30 40 50 60 70
0
10
20
30
40
50
60
t (min)
t (min)
D
60
50
40
30
20
10
0
C
glucose/PQQ-GDH
0.20

0.15
0.10
glucose/PQQ-GDH
ascorbate
NADH/DI
-8 -8-8-8 -7-7
1230x4x01x101001100
10-6
10-5
10-4
0
10 20 30 40 50 60 70
[H2O2]0 (M)
t (min)
Figure 1. (A, B, C) UV-vis kinetic traces monitored at (A, C) 262 nm or (B) 340 nm in a 96-wells microtiter plate during the reaction generated from the mixing of
different concentrations of H2O2 (from left to right: 150, 50, 15, 5, 1.5, 0.5, 0.15, and 0 µM) to solutions containing 50 µM 1 and either (A) 0.5 mM Asc, (B) 10 nM
DI and 250 µM NADH, or (C) 10 nM PQQ-GDH and 1 mM glucose. All experiments were conducted in an air-saturated 0.1 M Tris buffer (pH 8.5) containing 10
µM EDTA. (D) Semi-logarithmic calibration curves obtained from the threshold times (tc) recovered from the crossing of the horizontal dashed line in A, B and C to
the kinetic traces (tc is defined here as the time for which the absorbance is decrease approximatively half to the maximal absorbance change amplitude). Errors
bars are standard deviations from triplicates.
k1 = 6 M-1 s-1, k’1 = 0.001 s-1, k2 = 3300 M-1 s-1, k0 = 10-5 s-1, k3 =
170 M-1 s-1 and k’0 = 0.08 M-1 s-1 (Figure 2), a result which finally
production rate of H2O2 by redox cycling. For this system,
besides the spontaneous self-deprotection of 1, the non-specific
response may also originate from residual traces of NADH
oxidase activity in DI (the DI supplier report an activity < 0.01
wt. %), or from NADH autoxidation even if the latter is very slow
compared to Asc (which is attested by the excellent stability of a
NADH solution in an air-saturated Tris buffer – Figure S8).68
corroborates the proposed mechanism and confirms
a
nonspecific response mainly governed by the slow spontaneous
self-deprotection of probe 1 to NQH2.
2.2. Autocatalytic set with enzymatic reduction
We next investigated the reaction of 1 (50 µM) with H2O2 in the
presence of 10 nM DI and 250 µM NADH. The consumption of
NADH was easily followed at 340 nm (Figure 1B). Again,
sigmoidal kinetics traces are observed, showing well-defined
lags and exponential phases. A good linearity is found between
tc and log([H2O2]0) over two decades from 1 to 100 µM (Figure
1D). Compared to ascorbate, the detection of H2O2 was
improved up to the submicromolar range (a LOD of 0.5 µM H2O2
can be estimated). Improvement of the detection limit mainly
results from a nonspecific blank response delayed to longer
times. More surprising is the somewhat lower analytical
sensitivity compared with ascorbate (the linear part of the
calibration curve in Figure 1D is indeed steeper for
NADH/diaphorase than for Asc, demonstrating thus a slower
response time for a given [H2O2]0 concentration). This lower
sensitivity is unexpected considering the higher reducing
efficiency of the NADH/DI couple compared to ascorbate. As we
will see later on, such an effect can be explained by an inhibition
of the autoxidation of 1,4-NQH2 by DI, leading thus to a slower
1.0
Experimental
Simulated
0.8
0.6
0.4
0.2
0.0

0
10 20 30 40 50 60 10 20 30 40 50 60
t (min)
t (min)
Figure 2. (Left) Same experimental absorbance kinetic traces than those
reported in Figure 1A but after normalization to the initial absorbance. The
H2O2 concentrations are (from left to right): 150, 50, 15, 5, 1.5, 0.5, 0.15 and 0
µM. (Right) Theoretical normalized kinetic traces obtained from numerical
simulation of the autocatalytic mechanism proposed in Scheme 2 and using
the following set of rate constants: k1 = 6 M-1s-1, k’1 = 0.001 s-1, k2 = 3300 M-1 s-
1, k0 = 10-5 s-1, k3 = 170 M-1 s-1 and k’0 = 0.08 M-1 s-1 (see Supporting
Information for details). The color code is the same than on the left graph.
5
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
In an attempt to disentangle these different possible
contributions, we have investigated the effect of a pre-incubation
period of 1 in the reaction buffer onto the nonspecific blank
response generated under autocatalysis conditions. It is clear
from the results in Figure S9 that a prolonged pre-incubation of 1
in the Tris buffer drastically shortens the lag phase associated to
the nonspecific response, suggesting again a major contribution
of the pro-RCC instability in this undesired process.
highest enzyme concentration of 30 nM. This behavior
witnesses an inhibitory effect of the enzyme at the early stage of
the amplification, an inhibition that most likely slows down the
autoxidation rate of 1,4-NQH2 by O2. This inhibitory effect,
particularly pronounced for DI, may explain the unexpectedly
lower sensitivity of the H2O2 calibration curve in Figure 1D when
using NADH/DI rather than ascorbate as a reducing agent.
Further investigations are required to clarify this effect.
As suggested earlier, PQQ-GDH is also a good candidate to
play the role of a reducing enzyme. However, glucose has the
inconvenient to be UV-vis transparent contrary to ascorbate and
NADH. It led us to monitor the reaction progress either through
the increase of the UV-vis absorbance at 262 nm, a wavelength
selective to the oxidized 1,4-NQ, or the decrease of the
fluorescence of 1 at 465 nm. Results in UV-vis are gathered in
Figure 1C and those for fluorescence are reported in Figure S10.
We may observe an exponential increase in absorbance change
that is synchronized to the exponential decrease in fluorescent
signal, witnessing an autocatalytic process wherein 1,4-NQ
exponentially accumulates concomitantly with the consumption
of pro-RCC. One may also observe that the analytical sensitivity
with this alternative reducing enzyme is not significantly different
from that of ascorbate (the linear part of the H2O2 calibration
curves are almost overlaid in Figure 1D). Moreover, the dynamic
range of the calibration curve is narrower than DI (less than two
decades, ranging from ca. 5 to 100 µM), mainly restricted by the
prevalence of a strong nonspecific response which reduces the
LOD to 1 µM H2O2.
Improving the different rates of the autocatalytic set (e.g. by
improving the rate constants of each of the two catalytic loops) is
a way to improve the analytical performances of the system. On
that regard, increasing the pH is supposed to have a positive
role on both loops (simultaneously favoring the reactivity
between the probe and H2O2 as well as the auto-oxidation of
RCCRed). We thus performed the reaction at a slightly higher pH
of 9.0 and as expected it had a positive impact on the kinetics of
the reaction (as exemplified with probe 1 and PQQ-GDH in
Figure S10, which shows a shift of tc towards shorter response
times). However no better performances could be reached,
meaning that slightly higher pH also concomitantly affects the
chemical stability of probe 1 and so the nonspecific response.
Another possibility to tune the autocatalytic reaction kinetics is to
adjust the enzyme concentration as this potentially alters the
kinetics of Loop 2. In principle, lowering the concentration of the
enzyme would decrease the regeneration rate of RCCRed and
therefore would impact the required overstoichiometric
production of H2O2 to reach an autocatalytic regime. Figures S11
and S12 clearly illustrate this effect on the non-specific
autocatalytic response involving either DI or PQQ-GDH (enzyme
concentrations were varied from 30 nM to 30 pM). Indeed, at
concentration lower than 1 nM DI or 3 nM PQQ-GDH, the lag
phase tends to disappear and the abrupt exponential
absorbance decrease is progressively converted to a linear
decrease (which is typical of a redox cycling process rate-
controlled by a zeroth order autoxidation kinetics), meaning that
the reaction enters a nearly linear kinetic regime for which the
value of n in eq. 1 is probably only slightly greater than 1.
Besides, for concentrations above 1 nM DI or 3 nM PQQ-GDH,
the higher the concentration, the steepest is the kinetic curve
during its exponential phase. What is somehow surprising for
these higher concentrations is the influence of the enzyme
concentration on the duration of the lag phase, especially for DI.
Counter-intuitively, the longer lag phase is observed for the
Independently of the optical detection methods, the autocatalytic
process was also investigated using electrochemistry, taking
advantage of the electroactivity of the released naphthoquinone.
Indeed, the redox activity of 1,4-NQ can be recovered from the
magnitude of the catalytic current response in cyclic voltammetry
associated to the redox-mediated reduction of O2 (see Figure S4
and Figure 3A below). To avoid the interference of H2O2 that is
electrocatalytically generated during a cyclic voltammetric scan,
the electrochemical measurement was performed outside of the
reaction vessel by periodically sampling a few tens of microliters
of the reaction mixture (using a miniaturized screen-printed
electrochemical cell). The resulting kinetic plots obtained for a
sample of 15 µM H2O2 and a negative control with no H2O2 are
reported in Figure 3B. As expected, the data show an
exponential growth of the electrochemical response, which
arises at an earlier tc value for the sample than for the negative
control. This result is in good correlation with that obtained using
an UV-vis detection mode (Figure 1B).
25
B
A
0
-10
-20
20
15
10
5
0
5 min
25 min
34 min
44 min
icat
0
0
20 40 60 80 100
-0.4
-0.2
0.0
t (min)
E (V vs. ECS)
Figure 3. (A) Cyclic voltammetric responses of a solution mix containing 15
µM H2O2, 50 µM 1, 10 nM DI and 250 µM NADH, recorded at different reaction
times (the time is reported on the graph). Scan rate: 50 mVs-1. (B)
Autocatalytic kinetic traces obtained by reporting the sampled catalytic current
responses in cyclic voltammetry as a function of the reaction time for a starting
H2O2 concentration of (blue dots) 15 µM and (red stars) 0 µM.
3.
Triggering the reaction with RCC through redox
cycling-based H2O2 production
A versatile function of this cross-catalytic reaction network lies in
its capacity to be triggered not only by H2O2, but also by any
RCC as illustrated by the reaction network depicted in Scheme 3
(the pro-RCC probe is here represented by the pro-NQH2 probe
1). This might be viewed as a drawback since any traces of RCC
contaminating the pro-RCC probe would lead to an autocatalytic
transformation of pro-RCC. However we can turn this as an
advantage for the detection of RCC traces or more generally of
any redox cycling compound that is a potent trigger of the
system by producing H2O2.
3.1. Screening the triggering capacity of different RCC
By definition, any RCC in the presence of O2 and a well-chosen
reducing agent can redox cycle and lead to the production of
hydrogen peroxide. In the presence of our pro-NQH2, H2O2
6
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
produced can further react and trigger the above described
autocatalytic process. This triggering capacity is also a mean to
assess the ability of a RCC to efficiently produce H2O2 by redox
cycling. It was thus investigated for a series of naphthoquinones
(0.2 µM of each NQ was added to a solution containing
substrate 1, DI and NADH). Note in that case the nature of the
NQ trigger can be different from the 1,4-NQ released by
deprotection of compound 1, and so the role of the NQ trigger is
only to start the reaction, while this is the 1,4-NQ that
exponentially accumulates over the long term.
Blank
2,3-Cl-1,4-NQ
1,2-NQ
2-A-3-Cl-1,4-NQ
2-MeO-1,4-NQ
2-OH-1,4-NQ
Menadione
1,4-NQ
1.0
0.8
0.6
0.4
0.2
0.0

0
20
40
60
80
t (min)
1.0
Menadione
0.8
0.6
0.4
0.2
0.0
1,2-NQ
1,4-NQ
2-OH-1,4-NQ
2-MeO-1,4-NQ
2-A-3-Cl-1,4-NQ
2,3-diCl-1,4-NQ
-0.6
-0.4
-0.2
E1/2 (V vs. SCE)
Scheme 3. Principle for the determination of the RCC triggering capacity using
1 as a pro-NQH2 probe.
Figure 4. (Top) UV-vis kinetic traces recorded at 340 nm (monitoring of NADH
consumption) in a 96-wells microtiter plate. Experiments were performed in an
air-saturated 0.1 M Tris buffer (pH 8.5) containing 10 µM EDTA, 50 µM 1, 10
nM DI, 250 µM NADH, and 0.2 µM of (orange) 1,4-NQ, (olive) 2-amino-3-
chloro-1,4-NQ, (red) 2,3-dichloro-1,4-NQ, (blue) 1,2-NQ, (magenta) 2-
methoxy-1,4-NQ, (purple) 2-hydroxy-1,4-NQ, (cyan) menadione, and (black)
no quinone. (Bottom) Threshold time (normalized to the nonspecific one
obtained from the blank) vs. apparent midpoint potential of NQs.
As shown in Figure 4 (top), the different NQs induced an
autocatalytic response that is faster than the blank recorded in
the absence of NQ, evidencing their triggering capacity. The
time response is also observed to vary significantly as a function
of the nature of RCC. The 2-amino-3-chloro- and 2,3-dichloro-
naphthoquinones provoked the fastest responses (tc  12-13
min), suggesting that they are more amenable to redox-cycle
rapidly and so to produce more H2O2 in a shorter time. For these
two molecules, redox cycling is so efficient that, during the lag
phase, NADH is well discernably linearly consumed (during the
first 5-7 minutes of the reaction) before the autocatalytic process
becomes kinetically significant and takes over. NQs can be
ranked according to their ability to redox-cycle more or less
rapidly, which under our experimental conditions (here in the
presence of DI) leads to the following ranking (from faster to
slower): 2-amino-3-chloro-1,4-NQ ≈ 2,3-dichloro-1,4-NQ > 2-
methoxy-1,4-NQ > 1,4-NQ ≈ 2-hydroxy-1,4-NQ > 1,2-NQ >
menadione. These results highlight the versatility and potentiality
of our approach since it can be used as a time-readout-based
method for the screening of redox cycling properties of any
compound.69-71 In order to see if there is a linear free-energy
relationship between the triggering properties of NQs and their
reduction potential, the tc value of each NQ has been reported
on the bottom graph of Figure 4 as a function of the apparent
midpoint potential of NQs (E1/2, independently determined by
cyclic voltammetry at pH 8.5 in the absence of O2). If we
assumed that, in the presence of NADH/DI, the autoxidation of
NQ by O2 is the rate-limiting step of Loop 2, one should thus
expect that the lower the E1/2 of the NQ/NQH2 couple (or more
rigorously the lower is the reducing potential of the NQ-/NQH2
-
couple relative to the reducing potential of O2/O2 43), the faster
the autoxidation reaction with O2 to produce H2O2. The graph
however shows that there is no obvious trend from the plot,
showing for instance with the 2,3-dichloro-1,4-NQ an
impressively fast response while its reducing potential is among
the highest of NQs. This behavior reflects a more complex
reality of the reduced naphthoquinones reactivity towards
molecular oxygen43 and also probably towards diaphorase. More
important is the fact that 2,3-dichloro-1,4-NQ is the best hit from
this screening experiment, allowing us to further consider it for
the design of an improved pro-RCC probe (vide infra).
3.2. Direct detection of RCC
In order to evaluate the lowest concentration of RCC leading to
a distinct autocatalytic kinetic trace from the nonspecific one,
2,3-dichloro-1,4-NQ was tested as a trigger over a range of
concentrations from 0.5 nM to 1.5 µM in the presence of 1, DI
and NADH. As shown in Figure 5, the lower the concentration of
2,3-dichloro-1,4-NQ, the longer is the time required to observe
the autocatalytic response. The tc values for concentrations
below 50 nM follow an approximately linear relationship with the
log scale of the initial RCCOx concentration (bottom graph in
Figure 5), and then tends to deviate from linearity at higher
7
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
concentrations before reaching a kinetic regime no longer
undergoes a first rapid hydrolysis of its pinacol ester followed by
a slow and relatively negligible nonspecific self-deprotection
over several hours (Figure S13, A and E), which is very
analogous to that obtained with probe 1. This pro-RCC also
readily reacts in the presence of one equivalent H2O2 (Figure
S13, C and E) at a deprotection rate that is similar to probe 1.
The substituted naphthoquinone also evolves slowly in the buffer
either in the presence or absence of one equivalent H2O2
(Figure S13, B, D and F), but the time scale over which these
transformations occur remains slow. Reaction of this new probe
with H2O2 in the presence of NADH/DI led to the autocatalytic
kinetic curves shown in Figure 6. As expected, complete NADH
consumption occurs in a much shorter time than probe 1,
indicating an effectively accelerated autocatalytic reaction that
results in an improved analytical sensitivity. Corollary, this faster
kinetics is also associated with a non-specific autocatalytic
response that arises earlier than previously. Consequently, even
though the response time is 4-fold shortened with probe 2, only
concentrations over the micromolar range are clearly discernible
from the blank response, which in terms of limit of detection is
finally worse than with probe 1. These results show that
accelerating the kinetics of Loop 2 is well adapted to shorten the
autocatalytic responses but unfortunately is not necessarily
appropriate to improve the detection limit. Such behavior can be
easily rationalized by the fact that the non-specific response
resulting from the leakage of 1,4-NQ in Loop 1 (i.e., related the
instability of the self-immolative protecting group installed on the
pro-RCC) is amplified as effectively as the 1,4-NQ specifically
released through reaction with H2O2. This also demonstrates
that such an issue cannot be solved by simply changing the
nature of the naphthyl core. We may therefore conclude that
improvement of the system will compulsorily necessitate the
design of alternative caging strategies, much stable chemically
and less prone to self-deprotection.
controlled by autocatalysis but simply by the catalytic Loop 1.
0.8
0.6

0.4
0.2
0.0
0
10 20 30 40 50 60 70
t (min)
60
50
40
30
20
10
0
-10
-1-11-1--11-1-11011
0 10
223.43x54.65.x61.5x.5x15x0x1x10x10110000
10-9 10-8 10-7 10-6
[RCCOx]0 (M)
Figure 5. (Top) UV-vis kinetic traces recorded at 340 nm (monitoring of NADH
consumption) in a 96-wells microtiter plate. All experiments were conducted in
an air-saturated 0.1 M Tris buffer (pH 8.5) containing 10 µM EDTA, 50 µM 1,
10 nM DI, 250 µM NADH and the following decreasing concentrations of 2,3-
dichloro-1,4-NQ (from left to right): 1500, 500, 150, 50, 15, 5, 1.5, and 0 nM.
(Bottom) Semi-logarithmic calibration plot obtained by reporting the threshold
time recovered from the crossing horizontal dashed line in A as a function of
2,3-dichloro-1,4-NQ concentration. Errors bars correspond to standard
deviations from triplicates.
This loss of linearity suggests again some mechanistic
complications which probably arise from competition reactions
between the different radical intermediates generated by the
system. Overall, subnanomolar concentrations of 2,3-dichloro-
1,4-NQ are easily detected with this method, a performance that
could not be reached through an only redox cycling molecular
amplification. Any event that could produce low amounts of this
RCC would therefore potentially be sensitively detected through
this approach.15 This performance also advocates the use of this
compound in the design of an improved ester boronic substrate.
3.3. Design of a new pro-RCC
Improving the different rate constants of the autocatalytic set
could help to improve the response of the system either by
shortening the assay time required for the detection of the
targeted H2O2 or RCC molecules, or by improving the
discrimination between the specific and nonspecific autocatalytic
responses. For such purpose, we have examined the possibility
to tune the reaction rates of the redox cycle (Loop 2) by
selecting a better redox cycler to design the probe. In particular,
we postulated that it would be advantageous to substitute the
masked 1,4-NQ in probe 1 by a 2,3-dichloro-1,4-NQ. We
therefore synthesized probe 2 (Scheme 1B, see Experimental
Section for chemical synthesis). In the absence of H2O2, 2
8
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
of detection limit, linear range and reproducibility (time vs.
absorbance). One must note that this strategy is conceptually
different from the previously reported autocatalytic dentritic self-
immolative chain reactions, which all require the challenging
design of a chemosensing probe capable to release by self-
immolation at least two equivalents of the trigger (directly or
indirectly).13-17 Here, no such dendritic chemosensing probe is
necessary because once deprotected by one equivalent of the
trigger (here H2O2), the probe leads to the release of a redox-
active catalyst which then induces an over production of the
trigger by redox cycling. The analytical performances of this new
exponential molecular amplification is also significantly improved
compared to the dendritic chain reaction,13 leading to a 10-fold
lower detection limit of H2O2 (i.e., 0.5 µM H2O2) in the presence
of the diaphorase/NADH reducing species.
1.2
1.0
0.8
0.6

0.4
0.2
0.0
0
10
20
30
t (min)
Our reported simple autocatalytic reaction network also presents
great versatility. It can be used to efficiently reveal the redox
cycling ability of any compound, necessitating only a RCC-
releasing probe such as 1 and a reducing agent. Rapid and
sensitive detection of nanomolar concentration of a RCC in a
mildly basic aqueous media is also demonstrated. The
nonspecific self-deprotection of the pro-RCC triggering
molecules currently limits the analytical performances of the
system. The design of more robust pro-RCC molecules is
therefore needed to reach improved analytical performances.
We believe that the present approach is particularly significant
for analytical purposes but also for cell imaging or drug delivery
applications, notably by taking advantage of probes or pro-drugs
able to deliver in vivo the reporting/active molecule in a selective
and autocatalytic manner. It could be also useful as a building
block in the development of more complex chemical reaction
networks such as those giving rise to complex spatiotemporal
responses, including oscillations and pattern formation.Fehler!
Textmarke nicht definiert.
60
50
40
30
20
10
0
probe 2
probe 1
-8 --8-88--8-8-878
0
10-6
[H2O2]0 (M)
10-5
10-4
-7
12.12354.30x45.x5.5x51x1xx1x01011010000
10
Figure 6. (Top) UV-vis kinetic traces recorded at 340 nm (monitoring of NADH
consumption) in a 96-wells microtiter plate. All experiments were conducted in
an air-saturated 0.1 M Tris buffer (pH 8.5) containing 10 µM EDTA, 50 µM 2 ,
10 nM DI, 250 µM NADH and the following H2O2 concentrations (from left to
right): (black) 150, (red) 50, (olive) 15, (magenta) 5, (purple) 1.5, (royal) 0.5,
(blue) 0.15, and (magenta) 0 µM. (Bottom) Semi-logarithmic H2O2 calibration
curve obtained by reporting (red dots) the threshold time recovered from the
crossing of the horizontal dashed line in A as a function of H2O2 concentration.
Errors bars correspond to standard deviations from triplicates. For comparison,
the data obtained with probe 1 (blue squares) under the same experimental
conditions are added (same data than in Figure 1D).
Experimental Section
1. Material and methods
NADH was purchased from TCI. Trizma® base (minimum 99.9% titration),
EDTA (99.999%), DMSO (anhydrous ≥ 99.9%) and PQQ were obtained
from Sigma-Aldrich. Hydrogen peroxide for analysis and D-Glucose were
acquired from Acros Organics, and hydrochloric acid (37 wt.%) was
Conclusions
In summary, the present work demonstrates the feasibility of
building a simple molecular autocatalytic reaction network based
on two cross-activated catalytic loops, necessitating only a
H2O2-sensitive probe and a reducing agent in a mild aqueous
media. The relevance of such autocatalytic reaction scheme is
exemplified with the triggering of the autocatalytic process by
H2O2, leading to the simple and sensitive detection of
submicromolar H2O2 concentrations. A very simple all-molecular
format is proposed by taking advantage of ascorbate as a
reducing agent and UV-vis absorbance as a simple read-out.
Alternatively, enzymes such as DI or PQQ-GDH may equally
participate in the reduction process of the redox cycling loop. So
far, the best results are obtained with DI, leading to a 0.5 µM
H2O2 detection limit and a linear range over nearly two decades.
An interesting feature of the system is the possibility to easily
switch from a catalytic (Loop 1) to an autocatalytic (Loop 1 ×
Loop 2) reaction scheme by simply adding a reducing agent in
solution. Comparison of the two systems (Figure S2 vs. Figure
1d) highlights the relevance of the autocatalytic regime in terms
purchased from VWR. Lyophilized diaphorase
I
from Bacillus
stearothermophilus (DI, EC 1.6.99.-) was obtained from Nipro (Japan).
Before use, the mother solution of DI was centrifuged (11000 × g, 30 min,
4°C) over a Nanosep membrane of 10 kDa molecular weight cut-off (Pall
Corporation) pre-equilibrated with PBS Buffer. The apoenzyme form of
PQQ-dependent glucose dehydrogenase (PQQ-GDH) was produced by
controlled expression in an E. Coli strain by N. Mano’s group (University
of Bordeaux, France) and reconstituted in its holo form following a
published procedure.51 Glucose Oxidase from Aspergillus Niger was
purchased from sigma.
Standard spectrophotometric characterizations were carried out with a
UV/Vis spectrophotometer Specord S600 from Analytical Jena (Jena,
Germany) and a Cary Elipse Fluorescence Spectrophotometer from
Agilent technologies (Santa Clara, CA, USA). A quartz cuvette with 1-cm
optical path-length was used for standard experiments. For the high
throughput kinetic experiments, a Spark® multimode microplate reader
from Tecan Trading AG (Männedorf, Switzerland) was used. In the case
of fluorescence kinetics, the microwells were excited at 340 nm with a
bandwidth of 20 nm (Tecan fluorescence filter ref 30113156) and the
emission of fluorescence was recorded at 465 nm with a bandwidth of 35
9
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
nm (Tecan fluorescence filter ref 30113196). Standard Nunc 96-well UV
transparent, clear flat-bottom microplates (Thermo Fisher) were used for
the absorbance measurements, while Tecan’s 96-well-black flat bottom
microplates were used for the fluorescence experiments. The reactions
were initiated by injecting 100 µL solutions of different H2O2
concentrations into 100 µL solutions containing the probe and any of the
other desired reactants. The final working solution volume in each
microwell was 200 µL. During the whole reaction, the microplates were
left open to the air.
which was directly used in the next step without further purification. 1H
NMR (400 MHz, CDCl3): = 8.09 (dd, J = 1.5 and 8.3 Hz, 1H), 7.75 (dd,
J = 1.5 and 8.3 Hz, 1H), 7.50 (ddd, J = 1.5, 6.8 and 8.3 Hz, 1H) 7.46 (ddd,
J = 1.5, 6.8 and 8.3 Hz, 1H), 6.97 (d, J = 8.2 Hz, 1H), 6.56 (d, J = 8.2 Hz,
1H), 5.73 (s, 1H), 2.44 (s, 3H).
1-acetoxy-4-(4-bromobenzyl)oxynaphthalene 5: To a solution of 4 (819
mg, 4.1 mmol) in DMF (11 mL) and under an argon atmosphere were
added K2CO3 (1.01 g, 6.1 mmol) and para-bromobenzyl bromide (840 mg,
4.1 mmol). The resulting mixture was stirred for 24 h at room temperature.
After completion of the reaction as indicated by TLC, the reaction
medium was quenched with water (41 mL) and the resulting solution was
extracted three times with ethyl acetate. The organic phases were
gathered, washed twice with water and once with brine. The organic
solution was then dried over Na2SO4, filtered and the solvent evaporated.
The crude product was purified by flash chromatography
(cyclohexane/ethyl acetate: 10:1) yielding 5 as a white powder (541 mg,
36%). 1H NMR (400 MHz, CDCl3)= 8.32 (dd, J = 1.5 and 7.5 Hz, 1H),
7.75 (dd, J = 1.5 and 7.5 Hz, 1H), 7.58-7.49(m, 4H), 7.36 (d, J = 8.4 Hz,
2H), 7.13 (d, J = 8.3 Hz, 1H), 6.78 (d, J = 8.3 Hz, 1H), 5.14 (s, 2H), 2.44
(s, 3H). 13C NMR (100 MHz, CDCl3): = 170.0 (C=O), 152.4 (Cq. Ar.),
140.5 (Cq. Ar.), 136.1(Cq. Ar.), 131.9 (CH. Ar.), 129.2 (CH. Ar.), 127.7 (Cq. Ar.),
127.2 (CH. Ar.), 126.5 (Cq. Ar.), 126.1(CH. Ar.), 122.7(CH. Ar.), 122.1 (Cq. Ar.),
121.1 (CH. Ar.), 117.8 (CH. Ar.), 104.5 (CH. Ar.), 69.8 (CH2), 21.1 (CH3).
The electrochemical experiments were carried out with a computer-
controlled AutoLab PGSTAT 12 potentiostat interfaced to a GPES 4.9
software (EcoChemie B.V. Utrecht, The Netherlands). A standard three-
electrode electrochemical cell was used to characterize by cyclic
voltammetry the catalytic reduction of O2 by 1,4-NQ, while miniaturized
three-electrode-based screen-printed electrochemical cells, similar to the
ones previously used by our group,59 were used for monitoring the
release of the 1,4-NQ during the course of an autocatalytic reaction.
The mother solutions of probes (50 mM) were prepared in DMSO and
stored in the fridge no more than 2-3 days. The ascorbate and NADH
solutions were prepared from solid powder just before used, while the
H2O2 solutions at different concentrations were prepared daily for each
experiment. Unless otherwise stated, a 0.1 M Tris buffer (prepared from
Trizma® base and HCl) of pH 8.5 containing 10 µM EDTA was
systematically used. For numerical simulations of kinetics, we used the
COPASI software (version 4.16).72
1-(4-bromobenzyl)-oxy-4-hydroxynaphthalene 6: To a solution of 5 (541
mg, 1.46 mmol) in ethanol (8.3 mL) at 0°C was slowly added 7.5 mL of a
10% wt. NaOH solution. The reaction medium was stirred for 3 h at 0°C
then 15 mL of water were added and the solution. The product was then
extracted three times with chloroform and the organic phases were
gathered, dried over Na2SO4, filtered and then concentrated. The crude
product was next purified by flash chromatography (cyclohexane/ethyl
acetate: 10:1) to yield 6 as a white product (262 mg, 55%). 1H NMR (400
MHz, CDCl3)= 8.29-8.24 (m, 1H), 8.15-8.10 (m, 1H), 7.56-7.48(m, 4H),
7.36 (d, J = 8.5 Hz, 2H), 6.68 (d, J = 8.2 Hz, 1H), 6.64 (d, J = 8.2 Hz, 1H),
5.11 (s, 2H), 5.03 (s, 1H). 13C NMR (100 MHz, CDCl3): = 148.8 (Cq. Ar.),
145.7 (Cq. Ar.), 136.6 (Cq. Ar.), 131.9 (CH. Ar.), 129.2 (CH. Ar.), 126.8 (Cq. Ar.),
126.2 (CH. Ar.), 126.2 (CH. Ar.), 125.6 (Cq. Ar.), 122.3 (CH. Ar.), 122.0 (Cq. Ar.),
121.7 (CH. Ar.), 108.0 (CH. Ar.), 105.5 (CH. Ar.), 70.1 (CH2).
2. Chemical synthesis
All chemicals for synthesis were purchased either from Sigma-Aldrich,
Fischer Scientific or TCI Europe and used without further purification
unless specified. Unless otherwise stated, reactions were conducted in
air. Flash chromatography was performed on Kieselgel 60 (35-70 µm)
silica gel. 1H and 13C NMR spectra were recorded on a Bruker AC 400
MHz and were measured using CDCl3, CD3OD, CD3CN, Acetone-d6,
DMSO-d6 or DMF-d7 as solvent. Multiplicities are reported as follows: s =
singlet, d = doublet, t = triplet, sept = septuplet, m = multiplet, bs = broad
singlet, bd = broad doublet. Chemical shifts are reported in  units to 0.01
ppm precision for 1H and 0.1 ppm for 13C using residual solvent as an
internal reference. Mass spectra were measured on a MAT95S Finnigan-
Thermo spectrometer at the Institut de Chimie Moléculaire et des
Matériaux, Université Paris-Sud.
1-(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolane)benzyl)oxy-4-
hydroxynaphthalene 1: In a dry microwave glass tube and under an
argon atmosphere were dissolved the compound 6 (100 mg, 0.30 mmol),
dry potassium acetate (89 mg, 0.91 mmol) and bis(pinacolato)diboron
(116 mg, 0.46 mmol) in 4 mL of anhydrous DMF. Argon was bubbled in
the reaction medium for 20 min then bis(acetonitrile)palladium dichloride
(11 mg, 0.04 mmol) and 1,1'-bis(diphenylphosphino)ferrocene (24 mg,
0.04 mmol) were added and the bubbling was continued for 10 min. The
reaction medium was then heated in a microwave at 120°C for 20 min
and was then diluted in ethyl acetate, filtered on celite and the solvent
was evaporated. The crude product was directly purified by flash
chromatography (hexane/ethyl acetate: 95/5) yielding 1 as a white
powder (89 mg, 78%). Samples of higher purity were obtained by
recrystallization of purified 1 by diffusion of pentane in diethyl ether. 1H
NMR (400 MHz, CDCl3)= 8.31-8.26 (m, 1H), 8.13-8.08 (m, 1H), 7.83
(d, J = 8.0 Hz, 2H), 7.54-7.46(m, 4H), 6.68 (d, J = 8.3 Hz, 1H), 6.66 (d, J
= 8.3 Hz, 1H), 5.20 (s, 2H), 4.92 (s, 1H), 1.34 (s, 12H). 13C NMR (100
MHz, CDCl3): = 148.9 (Cq. Ar.), 145.7 (Cq. Ar.), 140.8 (Cq. Ar.), 135.2 (CH.
Ar.), 126.8 (Cq. Ar.), 126.7 (CH. Ar.), 126.0 (CH. Ar.), 125.6 (Cq. Ar.), 122.4 (CH.
Ar.), 121.7 (CH. Ar.), 108.0 (CH. Ar.), 105.5 (CH. Ar.), 84.1 (Cq.), 70.7 (CH2),
25.1 (CH3). HRMS: m/z Calculated for C23H25BNaO4+: 399.1742 Found:
399.1730.
Probe 1 was synthesized according to the following reaction scheme:
1,4-diacetoxynaphthalene 3: In a dry Schlenk and under an argon
atmosphere, a mixture of 1,4-naphthoquinone (1 g, 6.3 mmol), acetic
anhydride (1.6 mL, (17.1 mmol), pyridine (1.6 mL) and zinc dust (4.6 g,
70 mmol) in degassed chloroform (24 mL) was heated at 70°C for 15 min.
The mixture was then cooled, filtered on celite, and the filtrate was
washed twice with a 0.5 N HCl solution, twice with water and once with
brine. The organic solution was then dried over Na2SO4, filtered and the
solvent evaporated yielding 3 as a white powder (1.09 g, 85%). 1H NMR
(400 MHz, CDCl3): = 7.86 (dd, J = 3.3 and 6.4 Hz, 2H), 7.53 (dd, J = 3.3
and 6.4 Hz, 2H), 7.24 (s, 2H), 2.44 (s, 6H).
Probe 2 was synthesized according to the following reaction scheme:
1-acetoxy-4-hydroxynaphthalene 4: This compound was synthesized
starting from 3 following a described procedure,73 yielding 4 as brown oil
10
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
(100 mg, 0.205 mmol) in degassed acetonitrile (2.2 mL). The reaction
medium was stirred at 50°C for 3h. Upon completion of the reaction
quenching with a saturated NH4Cl solution was carried out and the pH of
the solution was then adjusted to 2 by adding 1 N HCl. The product was
extracted three times with ether, the organic phases were gathered and
washer once with HCl 0.1 N. The organic solution was then dried over
Na2SO4, filtered and the solvent evaporated. The crude product was then
purified by gel filtration chromatography on Sephadex™ LH-20 using
methanol as eluent. The fractions containing pure product 2 were then
gathered and concentrated until a 2mL methanol solution was obtained.
200 µL of water was then added and the product was extracted several
times with n-hexane. Hexane extracts were gathered, dried on Na2SO4,
filtered and concentrated. The oily residue was then dissolved in a small
volume of ethanol before adding water to finally being lyophilized.
Compound 2 was obtained as a slightly magenta solid (36 mg, 39%). 1H
NMR (400 MHz, CDCl3): = 8.22-8.16 (m, 1H),= 8.04-7.98 (m, 1H),
7.87 (d, J = 8.0 Hz, 2H), 7.57 (d, J = 8.0 Hz, 2H), 7.54-7.48 (m, 2H), 5.99
(s, 1H), 5.09 (s, 2H), 1.35 (s, 12H). 13C NMR (100 MHz, CDCl3): =
145.7 (Cq. Ar.), 145.4 (Cq. Ar.), 140.0 (Cq. Ar.), 135.3 (CH. Ar.), 127.9 (Cq. Ar.),
127.9 (Cq. Ar.), 127.4 (CH. Ar.), 126.9 (CH. Ar.), 123.3 (Cq. Ar.), 122.7 (CH. Ar.),
122.2 (CH. Ar.), 122.1 (Cq. Ar.), 112.8 (Cq.), 84.1 (Cq.), 75.9 (CH2), 25.1
(CH3). HRMS: m/z Calculated for C23H23BCl2NaO4+: 467.0963 Found:
467.0949.
1-acetoxy-2,3-dichloro-4-hydroxynaphthalene 7: In a dry schlenk and
under an argon atmosphere,
a
mixture of 2,3-dichloro-1,4-
naphthoquinone (4g, 17.6 mmol), acetic anhydride (2 mL, 21.1 mmol),
pyridine (4.3 mL, 52.8 mmol) and zinc dust (11.5 g, 176 mmol) in
degased chloroform (95 mL) was heated at 70°C for 24 h. The mixture
was then cooled, filtered on celite, and the filtrate was washed twice with
a 0.5 N HCl solution, twice with water and once with brine. The organic
solution was then dried over Na2SO4, filtered and the solvent evaporated.
The crude product was then purified by flash chromatography
(cyclohexane 100% to cyclohexane/ethyl acetate: 4:1) yielding 7 as a
brown-red powder (3.6 g, 75%). 1H NMR (400 MHz, CDCl3)= 8.15
(ddd, J = 0.8, 1.6 and 8.3 Hz, 1H),= 7.74 (ddd, J = 0.8, 1.5 and 8.3 Hz,
1H), 7.58 (ddd, J = 1.5, 6.7 and 8.3 Hz, 1H), 7.54 (ddd, J = 1.5, 6.7 and
8.3 Hz, 1H), 6.21 (s, 1H), 2.53 (s, 3H).74
Acknowledgements
1-acetoxy-2,3-dichloro-4-(4-bromobenzyl)oxynaphthalene 8: To
a
solution of 7 (1.11 g, 4.1 mmol) in DMF (11mL) and under an argon
atmosphere were added K2CO3 (1.7 g, 12.3 mmol) and para-
bromobenzyl bromide (1.5 g, 6.0 mmol). The resulting mixture was stirred
overnight at room temperature. After completion of the reaction as
indicated by TLC, the reaction medium was quenched with water (41 mL)
and the resulting solution was extracted three times with ethyl acetate.
The organic phases were gathered, washed twice with water and once
with brine. The organic solution was then dried over Na2SO4, filtered and
the solvent evaporated. The crude product was purified by flash
chromatography (cyclohexane 100% to cyclohexane/ethyl acetate: 4:1)
yielding 8 as a white powder (1.24 g, 82%). Higher purity was reached
after recrystallization in hexane (846 mg, 68%). 1H NMR (400 MHz,
CDCl3)= 8.05-8.01 (m, 1H),= 7.79-7.75 (m, 1H), 7.59-7.50 (m, 4H),
7.45 (d, J = 8.4 Hz, 2H), 5.07 (s, 2H), 2.51 (s, 3H). 13C NMR (100 MHz,
CDCl3): = 168.1 (C=O), 150.0 (Cq. Ar.), 140.8 (Cq. Ar.), 135.5 (Cq. Ar.),
131.8 (CH. Ar.), 129.8 (CH. Ar.), 128.0 (CH. Ar.), 127.6 (CH. Ar.), 127.6 (Cq. Ar.),
126.8 (Cq. Ar.), 123.7 (Cq. Ar.), 123.2 (Cq. Ar.), 122.5 (Cq. Ar.), 122.4 (CH. Ar.),
121.5 (CH. Ar.), 75.1 (CH2), 20.5 (CH3).
This work was supported by Agence Nationale pour la Re-
cherche (ANR ECOSENS project– ANR-19-CE29-0022-0).
Keywords: autocatalysis • cross-activation • exponential
amplification • redox cycling • autoxidation • naphthoquinone •
hydrogen peroxide • molecular probe • chemosensor • aromatic
boronate
[1]
Q. Zhang, S. Bhattacharya, M. E. Andersen, Open Biology, 2013, 3,
130031.
[2]
[3]
[4]
[5]
S. Goggins, C. G. Frost, Analyst, 2016, 141, 3157–218.
P. Scrimin, L. J. Prins, Chem. Soc. Rev., 2011, 40, 4488–505.
J. Wilhelm, A. Pingoud, ChemBioChem, 2003, 4, 1120–8
M. Kubista, J. M. Andrade, M. Bengtsson, A. Forootan, J. Jonák, K.
Lind, R. Sindelka, R. Sjöback, B. Sjöolive, L. Strömbom, A. Ståhlberg, N. Zoric,
Mol. Aspects Med., 2006, 27, 95–125.
[6]
R. Plasson, A. Brandenburg, L. Jullien, H. Bersini, J. Phys. Chem. A,
2011, 115, 8073-85.
[7]
Y. Zhao, F. Chen, Q. Li, L. Wang, C. Fan, Chem. Rev., 2015, 115,
12491–545.
1-acetoxy-2,3-dichloro-4-(4-(4,4,5,5-tetramethyl-1,3,2-
[8]
[9]
B. J. Lam, G. F. Joyce, Nat. Biotechnol., 2009, 27, 288–92.
A. J. Bisnette, S. P. Fletcher, Angew. Chem. Int., 2013, 52, 12800–26.
dioxaborolane)benzyl)oxynaphthalene 9: In a dry Schlenk and under an
argon atmosphere were dissolved the compound 8 (800 mg, 1.8 mmol),
dry potassium acetate (535 mg, 5.5 mmol) and bis(pinacolato)diboron
(692 mg, 2.7 mmol) in 30 mL of anhydrous 1,4-dioxane. Argon was
bubbled in the reaction medium for 20 min then bis(acetonitrile)palladium
dichloride (66 mg, 0.3 mmol) and 1,1'-bis(diphenylphosphino)ferrocene
(141 mg, 0.3 mmol) were added and the bubbling was continued for 10
min. The reaction medium was then heated at 90°C for 2h and was then
diluted in ethyl acetate, filtered on celite and the solvent was evaporated.
The crude product was directly purified by flash chromatography
(hexane/ethyl acetate: 95:5) yielding 9 as a white powder which was
recrystallized in hexane to afford very pure sample (160 mg, 18%). 1H
NMR (400 MHz, CDCl3)= 8.09-8.04 (m, 1H), 7.88 (d, J = 8.0 Hz,
2H),= 7.78-7.74 (m, 1H), 7.58 (d, J = 8.0 Hz, 2H), 7.56-7.48 (m, 2H),
5.15 (s, 2H), 2.51 (s, 3H), 1.35 (s, 12H). 13C NMR (100 MHz, CDCl3): =
168.3 (C=O), 150.5 (Cq. Ar.), 140.8 (Cq. Ar.), 139.7 (Cq. Ar.), 135.3 (CH. Ar.),
128.1 (CH. Ar.), 127.9 (Cq. Ar.), 127.8 (CH. Ar.), 127.5 (CH. Ar.),127.0 (Cq. Ar.),
123.8 (Cq. Ar.), 123.4 (Cq. Ar.), 122.8 (CH. Ar.), 121.6 (CH. Ar.), 84.1 (Cq.), 76.0
(CH2), 27.1 (4 x CH3), 20.7 (CH3).
[10] S. N. Semenov, A. S. Y. Wong, R. M. van der Made, S. G. J. Postma, J.
Groen, H. W. H. van Roekel, T. F. A. de Greef, W. T. S. Huck, Nat. Chem.,
2015, 7, 160–5.
[11] E. Valero, R. Varon, F. Garcia-Carmona, Biochem. J., 2000, 350, 237–
43.
[12] T.-R. Chen, C.-F. Hsu, C.-L. Chen, H. A. Witek, P. L. Urban, ACS
Synthetic Biol., 2016, 5, 962–8.
[13] E. Sella, D. Shabat, J. Am. Chem. Soc., 2009, 131, 9934–6.
[14] E. Sella, R. Weinstain, R. Erez, N. Z. Burns, P. S. Baran, D. Shabat,
Chem. Commun., 2010, 46, 6575–7.
[15] K. Yeung, K. M. Schmid, S. T. Phillips, Chem. Commun., 2013, 49,
394–6.
[16] J.-A. Gu, V. Mani, S.-T. Huang, Analyst, 2015, 140, 346–52.
[17] T. Yoshii, S. Onogi, H. Shigemitsu, I. Hamachi, J. Am. Chem. Soc.,
2015, 137, 3360–5.
[18] M. E. Roth, O. Olive, S. Gnaim, D. Shabat, Chem. Rev., 2016, 116,
1309–52.
[19] X. Sun, S. D. Dahlhauser, E. V. Anslyn, J. Am. Chem. Soc., 2017, 139,
4635–8.
[20] S. N. Semenov, L. J. Kraft, A. Ainla, M. Zhao, M. Baghbanzadeh, V. E.
Campbell, K. Kang, J. M. Fox, G. M. Whitesides, Nature, 2016, 537, 656–60.
[21] X. Sun, E. V. Anslyn, Angew. Chem. Int. Ed., 2017, 56, 9522–6.
[22] N. C. Gianneschi, S. T. Nguyen, C. A. Mirkin, J. Am. Chem. Soc., 2005,
127, 1644–5.
[23] J. P. Goertz, I. M. White, Angew. Chem. Int. Ed., 2017, 56, 13411–5.
[24] S. G. Rhee, Exp. Mol. Med., 1999, 31, 53–9.
[25] H. Sies, Redox Biol., 2017, 11, 613–9.
2,3-dichloro-1-(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolane)benzyl)oxy-4-
hydroxynaphthalene 2: In a dry Schlenk and under an argon atmosphere,
tetramethylguanidine (104 µL, 1.03 mmol) was added to a solution of 9
11
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
[26] S. N. Semenov, A. J. Markvoort, T. F. A. de Greef, W. T. S. Huck,
Angew. Chem. Int. Ed., 2014, 53, 8066–9.
[27] H. W. H. van Roekel, B. J. H. M. Rosier, L. H. H. Meijer, P. A. J. Hilbers,
A. J. Markvoort, W. T. S. Huck, T. F. A. de Greef, Chem. Soc. Rev., 2015, 44,
7465–83.
[28] W. Hordijk, J. Theor. Biol., 2017, 435, 22–8.
[29] H. G. Kuivila, J. Am. Chem. Soc., 1954, 76, 870–4.
[30] E. W. Miller, A. E. Albers, A. Pralle, E. Y. Isacoff, C. J. Chang, J. Am.
Chem. Soc., 2005, 127, 16652–9.
[31] A. Alouane, R. Labruère, T. Le Saux, F. Schmidt, L. Jullien, Angew.
Chem. Int. Ed., 2015, 54, 7492–509.
[32] J. L. M. Jourden, K. B. Daniel, S. M. Cohen, Chem. Commun., 2011, 47,
7968–70.
[33] A. R. Lippert, G. C. van de Bittner, C. J. Chang, Acc. Chem. Res., 2011,
44, 793–804.
[34] K. Żamojć, M. Zdrowowicz, D. Jacewicz, D. Wyrzykowski, L.
Chmurzyński, Crit. Rev. Anal. Chem., 2016, 46, 171–200.
[35] K. B. Daniel, A. Agrawal, M. Manchester, S. M. Cohen, ChemBioChem,
2013, 14, 593–8.
[36] E. Sella, D. Shabat, Org. Biomol. Chem., 2013, 11, 5074–8.
[37] V. A. Roginsky, T. K. Barsukova, H. B. Stegmann, Chem Biol. Inter.,
1999, 121, 177–97.
[38] G. D. Buffinton, K. Öllinger, A. Brunmark, E. Cadenas, Biochem. J.,
1989, 257, 561–71.
[39] C. Giulivi, E. Cadenas, Biochem. J., 1994, 301, 21–30.
[40] K. W. Wellington, RSC Adv., 2015, 5, 20309–38.
[41] L-O. Klotz, X. Hou, C. Jacob, Molecules, 2014, 19, 14902–18.
[42] P. Rana, R. Naven, A. Narayanan, Y. Will, L. H. Jones, Med. Chem.
Commun., 2013, 4, 1175–80.
[43] Y. Song, G. R. Buettner, Free Rad. Biol. Med., 2010, 49, 919–62.
[44] R. Munday, Free Rad. Biol. Med., 1997, 22, 689–95.
[45] X. Yuan, C. J. Miller, A. N. Pham, T. D. Waite, Free Rad. Biol. Med.,
2014, 71, 291–302.
[46] Ž. Anusevicius, A. Ramanavicius, J. Šarlauskas, Chem. Papers, 1998,
52, 643–9.
[47] G. Silveira-Dorta, D. Manuel Monzón, F. Crisóstomo, T. S. Martín, V.
Martín, R. Carrillo, Chem. Comm., 2015, 51, 7027–30.
[48] Y. Shang, C. Chen, Y. Li, J. Zhao, T. Zhu, Environ. Sci. Techol., 2012,
46, 2935–42.
[49] B. Limoges, D. Marchal, F. Mavré, J.-M. Savéant, J. Am. Chem. Soc.,
2006, 128, 2084–92.
[50] Y. Ogino, K. Takagi, K. Kano, T. Ikeda, J. Electroanal. Chem., 1995,
396, 517–24.
[51] F. Durand, B. Limoges, N. Mano, F. Mavré, R. Miranda-Castro, J.-M.
Savéant, J. Am. Chem. Soc., 2011, 133, 12801–9.
[52] J. Zielonka, A. Sikora, M. Hardy, J. Joseph, B. P. Dranka, B.
Kalyanaraman, Chem. Res. Toxicol., 2012, 25, 1793–9.
[53] X. Yuan, A. N. Pham, C. J. Miller, T. D. Waite, Environ. Sci. Technol.,
2013, 47, 8355–64.
[54] G. R. Buettner, B. A. Jurkiewicz, Radiation Res., 1996, 145, 532–41.
[55] C. Achilli, A. Ciana, M. Fagnoni, C. Balduini, G. Minetti, Cent. Eur. J.
Chem., 2013, 11, 137–9.
[56] H. G. Kuivila, J. Am. Chem. Soc., 1954, 76, 870-4.
[57] A. Sikora, J. Zielonka, M. Lopez, J. Joseph, B. Kalyanaraman, Free
Radic. Biol. Med., 2009, 47, 1401–7.
[58] E. W. Miller, A. E. Albers, A. Pralle, E. Y. Isacoff, C. J. Chang, J. Am.
Chem. Soc., 2005, 127, 16652–9.
[59] L. Challier, R. Miranda-Castro, D. Marchal, V. Noël, F. Mavré, B.
Limoges, J. Am. Chem. Soc., 2013, 135, 14215–28.
[60] C.P. Lu, C.T. Lin, C.M. Chang, S.H. Wu, L.C. Lo, J. Agric. Food Chem.,
2011, 59, 11403–6.
[61] A. Alouane, R. Labruère, T. Le Saux, I. Aujard, S. Dubruille, F. Schmidt,
L. Jullien, Chem. Eur. J., 2013, 19, 11717–24.
[62] H. G. Kuivila, Jr. J. F. Reuwer, J. A. Mangravite, Canadian J. Chem.,
1963, 41, 3081–90.
[63] P. A. Cox, M. Reid, A. G. Leach, A. D. Campbell, E. J. King, G. C.
Lloyd-Jones, J. Am. Chem. Soc., 2017, 139, 13156–65.
[64] X. Sun, D. Shabat, S. T. Phillips, E. V. Anslyn, J. Phys. Org. Chem.,
2018, e3827.
[65] R. J. Wilson, A. E. Beezer, J. C. Mitchell, Thermochim. Acta., 1995,
264, 27–40.
[66] D. E. Cabelli, B. H. J. Bielski, J. Phys. Chem., 1983, 87, 1809–12.
[67] B. G. Ruiz, S. Roux, F. Courtois, C. Bonazzi, Food Res. Inter., 2018,
106, 901–8.
[68] T. V. Votyakova, I. J. Reynolds, Arch. Biochem. Biophys., 2004, 431,
138–44.
[69] K. M. Soares, N. Blackmon, T. Y. Shun, S. N. Shinde, H. K. Takyi, P.
Wipf, J. S. Lazo, P. A. Johnston, Assay Drug Dev Technol., 2010, 8, 152–74.
[70] P. A. Johnston, Curr. Opin. Chem. Biol., 2011, 15, 174–82.
[71] P. Rana, R. Naven, A. Narayanan, Y. Will, L. H. Jones,
MedChemComm, 2013, 4, 1175–80.
[72] S. Hoops, S. Sahle, R. Gauges, C. Lee, J. Pahle, N. Simus, M. Singhal,
L. Xu, P. Mendes, U. Kummer, Bioinformatics, 2006, 22, 3067–74.
[73] T. Kumamoto, N. Aoyama, S. Nakano, T. Ishikawa, Y. Narimatsu,
Tetrahedron-Asymmetry, 2001, 12, 791–95.
[74] S. Ohta, S. Hatakeyama, C. Aono, T. Koike, I. Tanimizu, H. Hirakawa,
Yakugaku Zasshi, 1992, 112, 585–91.
12
This article is protected by copyright. All rights reserved.

10.1002/chem.201900627
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 2:
FULL PAPER
Author(s), Corresponding Author(s)*
Page No. – Page No.
Title
Autocatalysis by cross-activation of two catalytic loops
13
This article is protected by copyright. All rights reserved.