Distinguishing alternative reaction pathways by single-molecule fluorescence spectroscopy

Article abstract of DOI:10.1002/anie.201300100

Focus on chemical transitions: Epoxidation of a double bond in conjugation to a fluorescent dye was studied at single-molecule level. Direct observation of oxirane formation, indicated as a spectral shift from substrate to product state, revealed an alternative reaction pathway for the epoxidation reaction. Copyright

Full text of DOI:10.1002/anie.201300100

Angewandte
Chemie
DOI: 10.1002/anie.201300100
Single-Molecule Chemistry
Distinguishing Alternative Reaction Pathways by Single-Molecule
Fluorescence Spectroscopy**
Arina Rybina, Carolin Lang, Marcel Wirtz, Kristin Grußmayer, Anton Kurz, Frank Maier,
Alexander Schmitt, Oliver Trapp, Gregor Jung, and Dirk-Peter Herten*
Over the past two decades, single-molecule fluorescence
spectroscopy (SMFS) has been employed to resolve hetero-
geneities in solid-state physics, biophysics, and for localization
studies with high spatial resolution.[1–7] Although the notation
“molecule” is inherently connected to chemistry, applications
of SMFS to reaction dynamics were rare, despite the early
report of observed self-sensitized photo-oxidation of single
terrylene molecules hosted in a p-terphenyl crystal.[8] So far,
single-molecule methods are mostly used to investigate
catalytic turnover of enzymes.[9–11] Only recently, SMFS has
been applied to study heterogeneous catalysis. The cleavage
of fluorogenic esters was used to investigate and map the
spatial distribution of active catalytic sites for ester hydrolysis
on crystal particles[12,13] and to characterize the catalytic
behavior of individual enzymes.[14] With a similar approach,
catalytic redox reaction of fluorogenic substrates have been
studied at the surface of gold nanoparticles[15] and titanosili-
cate zeolites.[16] Aiming at progress in studying heterogeneous
organometallic reactions, the kinetics of ligand exchange at
platinum has been studied on the molecular level.[17] Recently,
the high spatial localization accuracy of SMFS allowed
homogeneous catalysis to be distinguished from surface
reactions.[18] All of the mentioned single-molecule studies
share the same experimental approach in which individual
chemical reactions are indicated by the occurrence of single
fluorescent spots that are, for example, due to the conversion
of ubiquitous leuko-dyes or quenched substrates into brightly
fluorescing product molecules. Here, light emission appears
during the last event in a succession of elementary reaction
steps. We have shown earlier that a reversible complexation
of metal ions can be characterized with SMFS by immobiliz-
ing a fluorescently labeled ligand.[19] Binding of a metal ion
leads to specific fluorescence quenching of the ligand such
that the forward chemical process (complexation) and the
reverse reaction (dissociation) can be quantitatively studied
by recording fluorescence trajectories of single ligand mole-
cules. A similar approach was used to observe the reversible
redox reaction of immobilized perylene diimide dyes.[20] In the
present work we expand this concept to the more general case
of irreversible reactions. To observe the initial and final states,
we had to make sure that the reactive group was part of the
chromophoric structure.[21] More specifically, we were inter-
ested in following the conversion of a single substrate
molecule along its reaction pathway to the final product
during the well-known epoxidation reaction on a double bond
with m-chloroperbenzoic acid (mCPBA).[22]
Here, the course of the reaction is embedded by the
disappearance of the fluorescent substrate and the appear-
ance of the fluorescent product. The experimentally most
convenient and reliable indicator of a completed reaction is
the change of the emission color.[8,23] Apart from the design of
an appropriate probe molecule for epoxidation, where we
could rely on previous experiments with BODIPY dyes,[23,24]
other challenges have to be overcome.
For maintenance of single-molecule conditions during the
whole reaction, the fluorescent substrate is less abundant by
several orders of magnitude than in experiments with
fluorogenic substrates, but the concentration of the co-
reagent prevails in a large excess. Continuous observation
until reactive collision takes place can be achieved using less
mobile or even immobile substrates. Finally, the substrate
must fluoresce until the reaction proceeds hence demanding
a photostable fluorophore without addition of stabilizers. In
the following, we present a system which fulfills the men-
tioned requirements.
[*] A. Rybina, K. Grußmayer, Dr. A. Kurz, Dr. D.-P. Herten
Cellnetworks Cluster & Physikalisch-Chemisches Institut
Universitꢀt Heidelberg
Im Neuenheimer Feld 267, 69120 Heidelberg (Germany)
E-mail: dirk-peter.herten@urz.uni-hd.de
Dr. C. Lang, F. Maier, Prof. O. Trapp
Organisch-Chemisches Institut, Universitꢀt Heidelberg
Im Neuenheimer Feld 270, 69120 Heidelberg (Germany)
To probe the epoxidation reaction with SMFS, we
designed the fluorescent probe oxyallyl BDP 1 (1-methyl-
(E)-3-(2-allyloxy)styryl-4,4’-difluoro-bora-3a,4a-diaza-(s)-
indacene) consisting of a BODIPY core expanded by an
oxyallyl styryl unit (Figure 1a). Its synthesis followed the
conventional Knoevenagel-like condensation.[25,26] The pre-
viously introduced styryl-BODIPY system has already shown
promising results in epoxidation reactions,[24] while here the
additional tagging with an oxyallylic residue serves as linker
for covalent surface binding (Figure 2a). This latter moiety
does not expand the chromophoric p-system, resulting in
almost indistinguishable spectral properties compared to the
M. Wirtz, Dr. A. Schmitt, Prof. G. Jung
Biophysikalische Chemie, Universitꢀt des Saarlandes
Campus Geb. B2.2, 66123 Saarbrꢁcken (Germany)
[**] We thank the Deutsche Forschungsgemeinschaft (DFG) for their
financial support (EXC81, SFB623). We also acknowledge Stephen
Hashmi (Heidelberg University) for fruitful discussions. Volker
Huch is gratefully acknowledged for X-ray crystallography. Michael
Schwering and Dominik Brox have continuously supported the
project with their expertise in microscopy.
Supporting information for this article, including details of reagents
used, instruments, and analytical data, including spectroscopic
characterization, is available on the WWW under http://dx.doi.org/
10.1002/anie.201300100.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Figure 2. Proposed reaction of immobilized 3 and characteristic signa-
tures of the different reactions and reaction pathways observed in
time-resolved single-molecule experiments. a) Reaction process of the
epoxidation of immobilized 3 to 4. b) Representation of the three
different conversions observed in single-molecule experiments.
c)–e) Single-molecule traces of substrate (570–615 nm, red line) and
product emission (500–525 nm, green line) recorded at 7.4 Hz show
specific features that are characteristic for substrate conversion via
a dim intermediate state (c), immediate conversion (d), and photo-
bleaching of 3 (e).
Figure 1. Epoxidation of 1 with a 200-fold excess of mCPBA yields 2
with significantly changed emission properties from 1. a) Crystal
structure of the substrate oxyallyl BDP 1. b,c) Epoxidation of 1 (red
line) with mCPBA to the product 2 (green line) can be followed as
a spectral shift in fluorescence emission. The green and red shaded
areas signify the wavelength range of the emission filters used in
single-molecule experiments for product and substrate detection,
respectively. Emission spectra were recorded after addition of a 200-
fold excess of mCPBA (ca. 0.2 mm) in CH2Cl2 over 6 h with an
excitation wavelength of 470 nm.
methylpolysiloxane) by hydrosilylation of the allylic double
bond.[27–29] The resulting labeled polymer was immobilized on
a glass cover slide. We tested different solvents for their
suitability in SMFS and found dichloromethane (CH2Cl2) to
be the most suitable for the epoxidation reaction, whereas
toluene and ethylacetate showed a rather high fluorescence
background. Reactions of individual substrate molecules 3
with mCPBA in CH2Cl2 to oxidized product 4 were recorded
on a TIRF microscope equipped with simultaneous laser
excitation at 532 and 488 nm (Figure 2a,b). Dual-color
detection in the wavelength ranges of 570–615 nm for the
yellow and 500–525 nm for the green fluorescence (Figure 1c)
was realized by an emCCD camera. The emCCD was
operated at a frame rate of 7.4 Hz; that is, 135 ms per frame
to optimize the signal-to-noise ratio and at the same time to
minimize photo-induced blinking and bleaching. The immo-
formerly studied system. Furthermore, the missing conjuga-
tion ensures that surface binding minimizes influences on the
electronic properties of the substrate.
Epoxidation of the yellow fluorescent BODIPY moiety
(lem = 572 nm) with mCPBA in excess leads to formation of 2
with green fluorescence (lem = 518 nm) owing to shortening of
the chromophoric unit (Figure 1b).[23,24] The blue shift of the
emission by roughly 50 nm as a result of the transformation
provides the required color change by which the initial
substrate 1 and the reaction product 2 can be unambiguously
assigned in a two-channel detection process. The overall
performance remains unchanged when the dye is immobi-
lized.
For single-molecule experiments, 1 was first covalently
attached to polysiloxane polymer HMPS (hydridomethyldi-
2
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
bilized probes exhibit a photostability sufficient for extended
observation periods (up to one minute; Figure 2e). Moreover,
the marginal hydrogenation of the double bond during
preparation leading to initially green fluorescent spots was
used to establish perfect alignment of the two detection
channels.
Epoxidation was initiated by addition of mCPBA
(0.2 mm) in CH2Cl2. Immediately after sealing with a second
cover slide, the sample was placed on the microscope for
image acquisition (ca. 2 min). In total, we acquired 1595
single-molecule traces, which were manually inspected and
categorized in three relevant classes (Figure 2b). The traces in
Figure 2c–e show the background corrected fluorescence
emission recorded in the substrate (red line) and the product
channel (green line). More than 90% of single-molecule
traces showed a loss of substrate emission in a single frame
without change to product emission. We attribute this
behavior to photobleaching of individual substrate molecules
or oxidative destruction of the chromophoric system as a side-
reaction. The latter is also observed in ensemble experiments
(Figure 1c). However, a relatively high percentage of traces
(115 traces, that is, ca. 7%) exhibit a clear spectral shift of the
emission from substrate to product channel (Figure 1c). Most
of the transformations occur in direct manner while few of
them exhibit a dark intermediate state.
Figure 3. a) Normalized probability histogram of the lag time Dt that
is fitted best by a double exponential model with time constants of
120 ms (75%) and 1.4 s (25%). b) Reaction process of the concerted
epoxidation mechanism as proposed by Bartlett.[22]
Again, we observed that product emission is always
terminated in a single frame owing to photobleaching (Fig-
ure 2c,d), clearly indicating the single-molecule nature of the
investigated system. We emphasize at this point the well-
known benefits of single-molecule techniques for investigat-
ing chemical reactions in which significant trajectories can be
selected specifically. Design of our experiments and spectral
detection ranges strongly suggest that the spectral shift
observed in single-molecule experiments signifies conversion
of the double bond in 3 to an epoxide in 4 (Figure 2c,d). To
support this interpretation, we carried out control experi-
ments with immobilized 3 in the absence of mCPBA
(Supporting Information, Figure S17). Here, most of the
traces indicate photobleaching of the immobilized substrate,
but only a minor part of the traces (< 0.4% out of > 1200
traces) show a similar spectral shift. The more than 20-fold
increased probability is a clear indication that the observed
spectral shift in presence of mCPBA occurs owing to
epoxidation of the exocylic double bond, leaving the chro-
mophoric BODIPY core intact, in good agreement with
previous findings.[23,24]
Epoxidation by mCPBA is assumed to proceed most
probably in a concerted reaction[22] via a single transition state
within a few picoseconds and thus far below the time
resolution of the image acquisition. Most trajectories
obeyed this expected immediate conversion from yellow to
green fluorescence. However, long-lasting dim states were
frequently observed during conversion (Figure 2c). To char-
acterize these transitions we collected the lag time Dt
(Figure 2c inset) of the dim states in a histogram. Figure 3a
shows a semilogarithmic plot of the probability densities,
which we obtained by normalization of the histogram
(Supporting Information, Figures S14, S15).
The simplest model that fitted the data with equally
distributed fit residuals was a bi-exponential function that
yielded two time constants of 120 ms and 1.4 s and amplitudes
of 75% and 25%, respectively. This corroborates that the two
time constants reflect two distinguishable reaction pathways,
yielding the same product 4. The more abundant reaction
path, which can be attributed to the lifetime below our
instrumental response time, is compatible with the widely
accepted concerted reaction mechanism (Figure 3b).
The intermediate state with the lifetime of roughly 1 s
demands, however, an alternative explanation. Experiments
at higher excitation intensities do not support any contribu-
tion of a light-driven process (Supporting Information, Fig-
ure S16). We also examined whether the lag time Dt depends
on the reaction time starting from the initiation of the
experiment and found that subsequently taken movies show
very similar probability density distributions for Dt (Support-
ing Information, Figure S15). Thus, the lifetime of the
intermediate state does not depend on the overall progress
of the reaction in the ensemble. It is therefore unlikely that
reaction pathways are caused by side-products, which would
accumulate with time. Control experiments did not provide
any indication of further, equally abundant fluorescent side-
products (Supporting Information, Figure S4).
The shorter pathway is below the time resolution of our
instrument and corresponds to the well-known concerted
mechanism occurring on picosecond timescale. The slower
pathway strongly indicates a transient intermediate formed
upon epoxidation of the conjugated double bond with
mCPBA. Although our current data allows distinguishing
the two pathways, we can only speculate on its nature.
We could exclude light-driven dim states by additional
experiments at higher excitation intensity, which gave similar
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
[6] S. Manley, J. M. Gillette, H. Shroff, H. Hess, E. Betzig, J.
Lippincott-Schwartz, Nat. Methods 2008, 5, 155 – 157.
[7] I. L. C. Buurmans, B. M. Weckhuysen, Nat. Chem. 2012, 4, 873 –
886.
[8] T. Christ, F. Kulzer, P. Bordat, T. Baschꢀ, Angew. Chem. 2001,
113, 4323 – 4326; Angew. Chem. Int. Ed. 2001, 40, 4192 – 4195.
[9] H. P. Lu, L. Xun, X. S. Xie, Science 1998, 282, 1877 – 1882.
[10] Y. Jiang, N. R. Douglas, N. R. Conley, E. J. Miller, J. Frydman,
W. E. Moerner, Proc. Nat. Acad. Sci. USA 2011, 108, 16962 –
16967.
[11] L. C. Tabares, D. Kostrz, A. Elmalk, A. Andreoni, C. Dennison,
T. J. Aartsma, G. W. Canters, Chem. Eur. J. 2011, 17, 12015 –
12019.
[12] M. B. J. Roeffaers, B. F. Sels, H. Uji-i, F. C. De Schryver, P. A.
Jacobs, D. E. De Vos, J. Hofkens, Nature 2006, 439, 572 – 575.
[13] M. B. J. Roeffaers, G. De Cremer, J. Libeert, R. Ameloot, P.
Dedecker, A.-J. Bons, M. Bꢁckins, J. A. Martens, B. F. Sels, D. E.
De Vos, J. Hofkens, Angew. Chem. 2009, 121, 9449 – 9453;
Angew. Chem. Int. Ed. 2009, 48, 9285 – 9289.
[14] K. Velonia, O. Flomenbom, D. Loos, S. Masuo, M. Cotlet, Y.
Engelborghs, J. Hofkens, A. E. Rowan, J. Klafter, R. J. M. Nolte,
F. C. de Schryver, Angew. Chem. 2005, 117, 566 – 570; Angew.
Chem. Int. Ed. 2005, 44, 560 – 564.
[15] C. Peng, W. Xu, X. Zhou, D. Panda, A. Kalininskiy, Chem. Phys.
Lett. 2009, 470, 151 – 157.
[16] T. Tachikawa, S. Yamashita, T. Majima, Angew. Chem. 2010, 122,
442 – 445; Angew. Chem. Int. Ed. 2010, 49, 432 – 435.
[17] S. M. Canham, J. Y. Bass, O. Navarro, S. Lim, N. Neeladri, S. A.
Blum, Organometallics 2008, 27, 2172 – 2175.
[18] N. M. Esfandiari, S. A. Blum, J. Am. Chem. Soc. 2011, 133,
18145 – 18147.
results. Thus, possible explanations for the intermediate state
could be alternative reaction pathways. A radical intermedi-
ate is improbable as those are usually not associated with
fluorescence emission and might not explain the weak
emission we observe. Another option, a possible fluorescence
quenching caused by formation of a p-complex of the
substrate with mCPBA, could not be confirmed by titration
experiments with the inert derivative m-chlorobenzoic acid
(Supporting Information, Figure S18). Alternatively, an
extended mechanism involving protonated epoxide as inter-
mediate has been proposed and discussed already by different
authors in the 1990s.[30] This pathway was also supported
recently by quantum chemical calculations.[31] The transient
nature of the proposed intermediate, however, imposes
practical limitations for its experimental verification so far
and might be fully understood with support of theoretical
calculations estimating the energetic profiles of reaction
pathways and intermediates.
In conclusion, we visualized the irreversible conversion of
a substrate into a product by means of single-molecule
detection. We provide a versatile experimental setup com-
prising probe design, immobilization and experimental con-
ditions for single-molecule experiments that have high
potential to study the wealth of organic reactions at double
bonds. Interpretation of our experiments is based on the
correlation between the spectral shifts observed in bulk and at
single-molecule level indicative for a reaction of the double
bond. Reaction details will be fully understood with addi-
tional chemical analysis. We further envision design of similar
fluorescent probes where different specific reactions occur-
ring can be directly observed and studied by means of single-
molecule techniques. Such a method has high potential,
opening new perspectives on observing, understanding, and
ultimately controlling chemical reactions.[32] Beyond that,
direct observation of the stochastic nature of chemical
reactions provides a new way for perceiving, conceiving,
and realizing chemistry.
[19] A. Kiel, J. Kovacs, A. Mokhir, R. Krꢂmer, D. P. Herten, Angew.
Chem. 2007, 119, 3427 – 3430; Angew. Chem. Int. Ed. 2007, 46,
3363 – 3366.
[20] R. Ameloot, M. Roeffaers, M. Baruah, G. De Cremer, B. Sels, D.
De Vos, J. Hofkens, Photochem. Photobiol. Sci 2009, 8, 453 – 456.
[21] G. Jung, A. Schmitt, M. Jacob, B. Hinkeldey, Ann. N. Y. Acad.
Sci. 2008, 1130, 131 – 137.
[22] P. D. Bartlett, Rec. Chem. Prog. 1950, 11, 47 – 51.
[23] G. De Cremer, M. B. J. Roeffaers, E. Bartholomeeusen, K. Lin,
P. Dedecker, P. P. Pescarmona, P. A. Jacobs, D. E. De Vos, J.
Hofkens, B. F. Sels, Angew. Chem. 2010, 122, 920 – 923; Angew.
Chem. Int. Ed. 2010, 49, 908 – 911.
[24] A. Schmitt, B. Hinkeldey, B. Hçtzer, G. Jung, J. Phys. Org. Chem.
2009, 22, 1233 – 1238.
Received: January 6, 2013
Revised: February 25, 2013
Published online: && &&, &&&&
[25] A. Loudet, K. Burgess, Chem. Rev. 2007, 107, 4891 – 4932.
[26] G. Ulrich, R. Ziessel, A. Harriman, Angew. Chem. 2008, 120,
1202 – 1219; Angew. Chem. Int. Ed. 2008, 47, 1184 – 1201.
[27] H. Cousin, O. Trapp, V. Peulon-Agasse, X. Pannecoucke, L.
Banspach, G. Trapp, Z. Jiang, J. C. Combret, V. Schurig, Eur. J.
Org. Chem. 2003, 3273 – 3287.
[28] C. Lang, U. Gꢂrtner, O. Trapp, Chem. Commun. 2011, 47, 391 –
393.
[29] M. J. Spallek, G. Storch, O. Trapp, Eur. J. Org. Chem. 2012,
3929 – 3945.
[30] R. D. Bach, C. Canepa, J. E. Winter, P. E. Blanchette, J. Org.
Chem. 1997, 62, 5191 – 5197.
[31] H. Shi, Z. Zhang, Y. Wang, J. Mol. Catal. A 2005, 238, 13 – 25.
[32] M. B. J. Roeffaers, G. De Cremer, H. Uji-i, B. Muls, B. F. Sels,
P. A. Jacobs, F. C. De Schryver, D. E. De Vos, J. Hofkens, Proc.
Nat. Acad. Sci. USA 2007, 104, 12603 – 12609.
Keywords: epoxidation · fluorescence probes ·
.
reaction dynamics · single-molecule spectroscopy
[1] S. van de Linde, M. Heilemann, M. Sauer, Ann. Rev. Phys.
Chem. 2012, 63, 519 – 540.
[2] M. A. Thompson, D. L. Matthew, W. E. Moerner, Annu. Rev.
Biophys. 2012, 41, 321 – 342.
[3] D. Wçll, E. Braeken, A. Deres, F. C. De Schryver, H. Uji-i, J.
Hofkens, Chem. Soc. Rev. 2009, 38, 313 – 328.
[4] G. De Cremer, B. F. Sels, D. E. de Vos, J. Hofkens, M. B. J.
Roeffaers, Chem. Soc. Rev. 2010, 39, 4703 – 4717.
[5] F. Kulzer, T. Xia, M. Orrit, Angew. Chem. 2010, 122, 866 – 879;
Angew. Chem. Int. Ed. 2010, 49, 854 – 866.
4
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Single-Molecule Chemistry
Focus on chemical transitions: Epoxida-
tion of a double bond in conjugation to
a fluorescent dye was studied at single-
molecule level. Direct observation of
oxirane formation, indicated as a spectral
shift from substrate to product state,
revealed an alternative reaction pathway
for the epoxidation reaction.
A. Rybina, C. Lang, M. Wirtz,
K. Grußmayer, A. Kurz, F. Maier,
A. Schmitt, O. Trapp, G. Jung,
D.-P. Herten*
&&&& — &&&&
Distinguishing Alternative Reaction
Pathways by Single-Molecule
Fluorescence Spectroscopy
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!