The Oxidation State of a Protein Observed Molecule-by-Molecule by sparkunder23


									The Oxidation State of a Protein Observed
Ralf Schmauder,[a] Fabio Librizzi,[b] Gerard W. Canters,[c] Thomas Schmidt,*[a] and
Thijs J. Aartsma[a]

We report the observation of the redox state of the blue copper          quenched by the absorption of the copper center of azurin in its
protein azurin on the single-molecule level. The fluorescence of a       oxidized state. In the reduced state, absorption is negligible, and
small fluorophore attached to the protein is modulated by the            thus no quenching occurs. We report on single-molecule meas-
change in absorption of the copper center via fluorescence reso-         urements, both in solution by using fluorescence correlation spec-
nance energy transfer (FRET). In our model system, the fluores-          troscopy (FCS) combined with fluorescence intensity distribution
cence label Cy5 was coupled to azurin from Pseudomonas aeru-             analysis (FIDA), and on surfaces by using wide-field fluorescence
ginosa via cysteine K27C. The Cy5 fluorescence was partially             microscopy.


Understanding the states and the related kinetic parameters of              Here we describe how this initial ensemble-type experiment
redox enzymes will help to unravel the mechanisms and func-              has been extended to the single-molecule level. Two ap-
tions of such proteins in living organisms. Despite the vast             proaches were used in analysis of single-molecule behavior:
amount of knowledge already available from bulk measure-                 The first approach involved monitoring the fluorescence fluctu-
ments,[1] single-molecule techniques will provide novel insight          ation caused by individual fluorophores diffusing through the
into biological redox processes by their ability to resolve sub-         confocal volume of a microscope. The data were evaluated in
populations of proteins without the need for external synchro-           terms of autocorrelation curves (in fluorescence correlation
nization, to add spatial and temporal selectivity, and to provide        spectroscopy, FCS[11, 12]) and by fluorescence intensity distribu-
ultimate sensitivity.[2]                                                 tion analysis (FIDA[13, 14]). The second approach employed fluo-
   Here we concentrate on one class of redox proteins, the               rescence wide-field microscopy, in which the fluorescence of
blue copper proteins. Azurin from Pseudomonas aeruginosa                 surface-adsorbed molecules is monitored by a low-noise CCD
was used as model system. This protein serves as a typical ex-           camera. The two methodologies yielded complementary infor-
ample of the many redox proteins in which cofactors change               mation about concentrations, diffusion behavior, distribution
their absorption characteristics on reduction and oxidation.             of species, and the photophysics of individual molecules.
Azurin, a 14 kDa protein which was extensively studied in the
past, has a single copper ion as its redox-active center. Origi-
nally, it was thought to be involved in the nitrate respiratory          Experimental Section
chain of the organism,[3, 4] but recent findings have made it            Protein engineering, purification, labeling, and bulk experiments
likely that azurin fulfills a role in oxidative stress response.[5] In   are described in detail elsewhere.[6, 15] Briefly, Cy5-maleimide (Amer-
its oxidized (Cu2 + ) form the protein has an absorption in the          sham Biosciences, Freiburg, Germany) was coupled to an engi-
550–650 nm range (e = 5600 mÀ1 cmÀ1), which corresponds to a             neered cysteine (K27C) of either copper or zinc azurin. Excess free
p–p* transition involving mainly the dx2Ày2 orbital of Cu2 + and         dye was removed by size-exclusion chromatography with a 5 kDa
a 3p orbital of the sulfur atom of Cys112. This absorption dis-          cutoff (centri-spin10, Princeton Separations, Adelphia, NJ, USA).
                                                                         Measurements were performed in phosphate-buffered saline (PBS:
appears when the Cu site is reduced (Cu + ), because the molec-
ular orbital in the d10 configuration of Cu + is filled.                 [a] R. Schmauder, Prof. Dr. T. Schmidt, Prof. Dr. T. J. Aartsma
   Previously, we demonstrated a novel fluorescence-based de-                Physics of Life Processes
tection scheme for this transition.[6] The methodology is based              Leiden Institute of Physics, Leiden University
                                                                             Niels Bohrweg 2, 2333 CA Leiden (The Netherlands)
on the redox-state-dependent change in absorption of the
                                                                             Fax: (+ 31) 715-275-819
nonfluorescent copper center, which quenches, by fluores-                    E-mail:
cence resonance energy transfer (FRET), the fluorescence of a            [b] Dr. F. Librizzi
small organic fluorophore conjugated to the protein. The                     Istituto Nazionale per la Fisica della Materia and
change in absorption of the copper center modulates the                      Dipartimento di Scienze Fisiche e Astronomiche
                                                                             Via Archirafi 36, 90123 Palermo (Italy)
spectral overlap and thus results in a change in Fçrster radius
                                                                         [c] Prof. Dr. G. W. Canters
and energy transfer efficiency.[7, 8] Note that no conformational
                                                                             Leiden Institute of Chemistry—Gorlaeus Laboratory
change is involved in the process,[9] as is evident from the crys-           Leiden University, P.O. Box 9504, 2300 RA Leiden (The Netherlands)
tal structure. A similar scheme was recently demonstrated in a               Supporting information for this article is available on the WWW under
chemical system.[10]                                                or from the author.

ChemPhysChem 2005, 6, 1381 – 1386     DOI: 10.1002/cphc.200400628            2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim                     1381
                                                                                                                                         T. Schmidt et al.

150 mm NaCl, 10 mm Na2HPO4/NaH2PO4, pH 7.4). To avoid unspecif-               where.[18] Briefly, samples were mounted on an inverted micro-
ic adsorption 0.01 vol % Tween20 (Sigma-Aldrich, Steinheim, Ger-              scope equipped with an Apochromate 100 ” /1.4 objective (Zeiss,
many) was added in the FCS/FIDA experiments. Experiments with-                Jena, Germany) and illuminated with 625 nm light of a dye laser fo-
out Tween20 gave qualitatively comparable results, but were com-              cused in the back focal plane of the objective. Use of an appropri-
plicated by adsorption of protein to the sample carrier. Oxidation            ate filter combination (DCLP650 (Chroma Technology, Brattleborro,
was performed with K3[Fe(CN)6], and reduction by adding dithio-               VT, USA), RG645-6 (Schott, Mainz, Germany), 10SW-750 (Newport,
threitol (DTT; both from Sigma-Aldrich, Steinheim, Germany). Stock            Irvine, CA, USA) permitted detection of single Cy5-azurin molecules
solutions of 100 mm were freshly prepared in H2O. Stock and ap-               on a liquid-nitrogen-cooled, back-illuminated CCD-camera system
propriately diluted solutions of both reagents were used within               (Princton Instruments, Trenton, NJ, USA). Low concentrations (ca.
48 h. Bulk experiments were performed on a L55 fluorimeter                    1 nm) of protein were unspecifically adsorbed on methanol-
(Perkin Elmer, USA) in cuvettes with 5 mm beam path. Excitation               washed cover glasses (Menzel, Germany). After a short incubation
was set to 645 nm, emission detection to 665 nm and 5 nm slits                the cover slips were washed with PBS. Imaging conditions were
were used.                                                                    typically 5 ms illumination time and 2 kW cmÀ2 intensity. Imaging
                                                                              frequency was on the order of 10 Hz. Per condition, series of 50–
Fluorescence correlation spectroscopy (FCS) was performed on a                100 images were acquired at five different areas of the sample.
commercial Confocor2 system (Zeiss, Jena, Germany; described by               The background was subtracted from the recorded images, and
Weisshart et al.[16]) by using a helium–neon laser at a wavelength            subsequently fluorescence peaks were fitted to Gaussian functions.
of 633 nm and an intensity of 30 kW cmÀ2 for excitation. Correla-             From the signal intensity and its estimated error of individual
tion curves were analyzed with the software supplied using a 3D               peaks, a probability density function (PDF) of the single-molecule
diffusion model and an exponential decay to approximate photo-                signal was calculated. The advantage of a PDF over a histogram is
physics contributions. Fitting yielded diffusion time and mean                that binning artifacts are avoided.
number of particles in the observation volume, as well as the trip-
let time and fraction and a parameter describing the axial/longitu-
dinal ratio of the assumed three-dimensional Gauss-shaped confo-
cal volume. Calibration with unconjugated Cy5 (assuming D =                   Results and Discussion
3.16 ” 10À10 m2 sÀ1, Zeiss manual) yielded an observation volume of
0.38 fL and a beam diameter of 230 nm.                                        Cy5, a stable and bright red cyanine dye, was covalently at-
Fluorescence intensity distribution analysis (FIDA) is a methodol-
                                                                              tached to cysteine K27C engineered into the blue copper pro-
ogy which permits the simultaneous determination of concentra-                tein azurin (Figure 1 A). The distance between Cy5 and the
tions and specific brightness values of different fluorescent species         photoactive copper center in azurin is about 3.5 nm. In this po-
in solution, starting from the distribution of the number of photon           sition the dye is far away from the hydrophobic patch[9] on the
counts. It was performed on the raw
data by means of a computer program
on the basis of the approach introduced
by Kask et al.[13] Distributions of the
number of photon counts were ob-
tained with a counting interval of 40 ms.
The regularization of the brightness dis-
tributions was achieved by a maximum-
entropy-based method.[17] In addition,
instead of a simple predefined grid of
brightness values, a set of Gaussian
curves was used. The number of Gaussi-
ans was equal to the number of points
in the brightness grid, and the width of
each Gaussian was equal to their spac-
ing. This procedure does not involve
more fitting parameters, and its benefit
is not in terms of accuracy of the fitting,
but in terms of a local smoothing of the
obtained brightness distributions. A grid
of 80 brightness values was used, with a
span from 0.3 to 300 kHz, equally
spaced on a logarithmic scale. When
tested on solutions of fluorescent dye of
known concentrations, the program
gave a single brightness peak as expect- Figure 1. A) Schematic representation of azurin (adapted from 4AZU.pdb).[9] The attachment site for the flu-
ed (data not shown). Fitting of the re- orophore (Cys27) and the copper atom are shown in space-filled format. B) Absorption and emission spectra
ported data by means of single compo- of Cy5 (black: absorption, gray: emission) are compared to the absorption spectra of azurin in the oxidized
nents instead of brightness distributions state (dashed). The extinction coefficient of azurin is magnified tenfold. The emission spectrum is normal-
gave similar results.                         ized to fit the graph. C) Bulk fluorescence measurements. The concentration of labeled protein was 40 nm.
                                               Increasing amounts of oxidizing (K3[Fe(CN)6]; gray, open symbols) or reducing (DTT; black, filled symbols)
Wide-field microscopy was performed            agents were added to copper-containing (squares, solid line) and zinc-containing azurin (circles, dashed
on a home-built setup, described else-         line). The initial increase in the case of reduction is due to reaction of the oxygen in the solution.

1382                2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim                           ChemPhysChem 2005, 6, 1381 – 1386
Single-Molecule Redox Chemistry of Azurin

protein surface and from the copper center, so that unspecific
interactions with the protein surface or electron transfer to the
dye is unlikely. In the oxidized state the copper center acts as
an energy acceptor and quenches the fluorescence of the Cy5
donor molecule. This quenching is defined by a dipole–dipole-
governed resonant energy transfer process characterized by a
Fçrster radius[7, 8] of R0 = 3.8 nm (see the spectra in Figure 1 B).
Given the distance and the value of R0 an energy-transfer effi-
ciency of about 50 % is predicted. In the reduced form the ab-
sorbance of the copper center and hence the energy-transfer
efficiency is negligible, which in turn results in a significant in-
crease in the dye’s fluorescence. A previous study showed that
this two-state quenching/dequenching behavior can be used
as sensitive readout of the redox state of azurin in a bulk ex-
periment.[6] Figure 1 C summarizes those previous results. A
copper–azurin Cy5 conjugate (Cy5-azurin; squares, solid lines)
and a Cy5-azurin containing a redox-inactive zinc center (Cy5-
azurinZn ; circles, dashed lines) were observed in buffer solu-
tions with increasing concentration of oxidizing (K3[Fe(CN)6];
open symbols, gray lines) and reducing (DTT; full symbols,
black lines) agents. As predicted from the energy-transfer
mechanism the fluorescence of Cy5-azurin is quenched by
50 % on oxidation and dequenched on reduction (solid lines in
Figure 1 C), whereas no change in the fluorescence signal was
observed for either the inactive Cy5-azurinZn (dashed lines in
Figure 1 C) or free Cy5 dye (data not shown).
                                                                        Figure 2. A) Correlation curves of Cy5 (normalized at the first data point,
                                                                        t = 0.2 ms) for different concentrations of K3[Fe(CN)6]: 1 mm (black,
Fluorescence Correlation Spectroscopy                                   G(0) = 8.9), 100 mm (dark gray, dash–dotted, G(0) = 8.0), 10 mm (gray, dotted,
                                                                        G(0) = 7.0), 1 mm (light gray, dotted, G(0) = 6.8), 100 nm (black, dotted,
Fluorescence correlation spectroscopy (FCS) was used to fur-            G(0) = 7.0), and 10 nm (gray, dashed, G(0) = 7.3, last four lines almost coin-
ther characterize the observed changes in the behavior of Cy5-          cide). A solution of 1 mm (light gray dotted line) was used for all subsequent
azurin in dependence on its oxidation state. Addition of very           FCS/FIDA experiments. B) Correlation curves of Cy5-azurin under oxidizing
                                                                        and reducing conditions. Correlation curves were normalized with the fitted
high concentrations (@ 10 mm) of the oxidizing reagent caused           mean number of molecules N. Solid line: copper form, N = 1.7; dashed line:
a marked shift towards short-time correlations (< 1 ms) and a           zinc form, N = 1.1; Black: 2.5 mm DTT added; light gray: 1 mm K3[Fe(CN)6]
loss of fluorescence for all samples measured, including free           added. For comparison, the correlation curve of free Cy5 is shown (gray,
Cy5 (Figure 2 A). This is in contrast to what was observed              dotted).

under low-light conditions in a fluorimeter (Figure 1 C). A likely
explanation for this observation is a possible oxygen sensitivity
of the cis–trans isomerization of Cy5, which occurs in the dye’s        1.8 nm and a diffusion constant of 1.25 Æ 0.1 ” 10À10 m2 sÀ1. On
excited state and is characteristic of the photophysics of this         oxidation or reduction azurin does not change its structure
dye.[19] Our interpretation is further supported by the observa-        and thus its hydrodynamic radius remains the same. Our find-
tion that oxygen-scavenger systems largely eliminate photo-             ing here corroborate the crystallographic findings,[9] as the dif-
bleaching and photoblinking of Cy5.[20] All results described in        fusion times found were independent of the oxidation state
the following were obtained with a maximum concentration of             (Figure 2 B).
1 mm of the oxidizing reagent, which still ensures almost com-
plete oxidation of the protein (1–10 nm) but is one order of
                                                                        Fluorescence Intensity Distribution Analysis
magnitude below the onset of competing photophysical ef-
fects.                                                                  To gain further information on the oxidation/reduction state of
   In the presence of 1 mm of the oxidizing reagent the correla-        individual proteins associated with this process, the fluores-
tion curves of both Cy5-azurin and Cy5-azurinZn were almost             cence signals obtained in the FCS experiments were further
indistinguishable from correlation curves measured with                 analyzed by fluorescence intensity distribution analysis (FIDA),
excess reducing agent (2 mm). The correlation times (t = 98 Æ 5         as developed by Kask et al.[13] Figure 3 A and B show the distri-
and 88 Æ 3 ms for Cy5-azurin and Cy5-azurinZn, respectively) cor-       butions of the number of photon counts of 20 s long measure-
respond to diffusion coefficients of approximately (1.35 Æ 0.1) ”       ments (40 ms counting intervals) of Cy5-azurin and of Cy5-azur-
10À10 and (1.5 Æ 0.1) ” 10À10 m2 sÀ1. These values are fully ration-    inZn, respectively. On average 1.8 Cy5-azurin and 1.1 Cy5-azur-
alized by the diffusion coefficient of a 14 kDa globular protein        inZn molecules where present in the detected volume of 0.4 fL.
such as azurin, characterized by an hydrodynamic radius of              Whereas the distributions of the number of photon counts

ChemPhysChem 2005, 6, 1381 – 1386              2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim                      1383
                                                                                                                                           T. Schmidt et al.

                                                                                             tion or heterogeneities in the labeling chemistry.
                                                                                             Hence we suggest that the interaction between the
                                                                                             protein and the fluorophore must partially change its
                                                                                             photophysics, and this is likely related to the two iso-
                                                                                             merization forms of Cy5.

                                                                                             Wide-Field Single-Molecule Imaging
                                                                                             We performed complementary experiments on im-
                                                                                             mobilized proteins using wide-field single-molecule
                                                                                             imaging.[21] Cy5-azurin and Cy5-azurinZn were ad-
                                                                                             sorbed on glass substrates and imaged with a low-
                                                                                             noise CCD camera. The samples were subsequently
                                                                                             treated in a cyclic protocol with the oxidizing agent,
                                                                                             washed, and treated with the reducing agent,
                                                                                             washed, and so forth. An example of the raw image
                                                                                             data is shown in Figure 4. Individual molecules were
                                                                                             simultaneously observed within an area of 10 ”
                                                                                             10 mm. Following the fluorescence signal of one of
                                                                                             the molecules in time (inset) revealed a single-step
                                                                                             photobleaching event after 63 frames, characteristic
Figure 3. FIDA data for Cy5-labeled copper azurin (A and C) and zinc azurin (B and D). A)    of an individual molecule.
and B) Distributions of the number of photon counts (symbols) and corresponding fits            In turn the sample was exposed to multiple oxida-
(lines) from the FIDA analysis. C) and D) Brightness distributions for alternating condi-    tion/reduction cycles. The change in fluorescence of
tions: reducing conditions (black line/full symbol), oxidizing conditions (gray line/open    three individual molecules (Figure 5 A) was followed
symbol), followed by reducing conditions at 2.5 mm DTT (gray/black symbols, dashed
line). The uncertainty in the brightness distributions, estimated by performing FIDA anal-   while the buffer conditions where changed (at
ysis on different data sets relative to the same samples, is on the order of spreading of    frames 23, 53, and 85, respectively). Switching of the
the distributions shown in panel D) (see Supporting Information).                            intensity with changes in the buffer conditions, as

strongly depend on oxidation conditions for the Cy5-azurin
sample (Figure 3 A), the Cy5-azurinZn sample (Figure 3 B)
showed no oxidation/reduction-dependent behavior.
   The distributions of the number of photon counts (Fig-
ure 3 A and B) were tested for the existence of various molecu-
lar species, as characterized by their molecular brightness
values. Analysis was performed by means of an unrestricted
FIDA.[13] Next to the background signal at 1 kHz (not shown in
Figure 3 C and D) two fluorescent species were found at
brightness values of about 45 and 70 kHz, respectively. For
Cy5-azurin under reducing conditions the two species were
about equally abundant (39 and 61 %), whereas under oxidiz-
ing conditions the fraction of the brighter species decreased to
5 % (Figure 3 C). In addition a slight overall loss in peak bright-
ness of 10 % was observed. The reaction was found to be re-
versible, as was apparent from cycling experiments (reduced:
black curve; oxidized: gray curve; re-reduced: dashed curve in
Figures 3 C and 3 D). The average brightness in the sample was                    Figure 4. Fluorescence image of Cy5-azurin adsorbed on a cover glass. The
                                                                                  image size was 10 ” 10 mm2, and the illumination time was set to 5 ms. The
more than 30 % lower in the oxidized state, a decrease compa-
                                                                                  inset shows the intensity trace of an individual molecule through several
rable to that found in bulk experiments (Figure 1 C). For Cy5-                    images. Note the single-step photobleaching event at frame 63.
azurinZn neither a change in relative abundance (63 and 37 %)
nor a shift in brightness was observed (Figure 3 D).
   Importantly, two species of different brightness were not ob-                  well as single-step photobleaching events in which the fluores-
served for the free dye (data not shown). In addition, the bi-                    cence signal dropped to background (the lower two trajecto-
modal pattern in the brightness distribution under reducing                       ries), were clearly observable. In total we analyzed 594 of such
conditions was also observed for Cy5-azurinZn. Therefore, the                     time traces for both Cy5-azurin and Cy5-azurinZn. From those
bimodal pattern is likely not caused by dye-detergent interac-                    traces the signal-to-noise (SN) ratio, as given by the signal di-

1384                 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim                            ChemPhysChem 2005, 6, 1381 – 1386
Single-Molecule Redox Chemistry of Azurin

                                                                                                                     We have demonstrated that the
                                                                                                                     detection of a protein’s redox-
                                                                                                                     state by fluorescence techniques
                                                                                                                     can be carried to the single-mol-
                                                                                                                     ecule level. For this we em-
                                                                                                                     ployed a novel approach in
                                                                                                                     which, in contrast to existing
                                                                                                                     studies, no fluorescent substrate
                                                                                                                     or product,[22] no fluorescent co-
                                                                                                                     factor,[23] and no conformational
                                                                                                                     change[24] is required. A change
                                                                                                                     in absorption of the protein
                                                                                                                     redox center is sufficient for our
                                                                                                                        As an example we chose a
                                                                                                                     typical class of redox-active pro-
                                                                                                                     teins, the blue copper proteins.
                                                                                                                     The readout is based on quench-
Figure 5. A) Intensity traces of different molecules through an image series. The oxidized sample is reduced (at     ing of an artificial fluorescence
frame 26), oxidized (frame 52), and subsequently reduced again (frame 84). B) Probability distribution function of
                                                                                                                     label by Fçrster energy transfer
the single-molecule signal for reduced Cy5-azurin (2926 molecules, black line), which was subsequently oxidized
(1825 molecules, gray line) and re-reduced (2189 molecules, black dashed line). C) AzurinZn data: reducing condi-    to the active site of the protein.
tions (1582 molecules, black line), oxidizing conditions (812 molecules, gray line), and re-reducing conditions (535 The quenching is modulated by
molecules, black dashed line).                                                                                       the dependence of the absorp-
                                                                                                                     tion properties of the active
                                                                                                                     center on its redox state. By
vided by its standard deviation, can be compared to the                              means of two complementary single-molecule techniques we
switching ratio R between the two oxidation states of azurin to                      were able to establish parameters for a sensitive and stable
define a robust estimator for the visibility of switching. By this                   readout of this process. To our surprise we found that Cy5-
definition the SN is between 3 and 6 for the traces shown,                           azurin consists of two species which differ in their molecular
whereas R is between 1.5 and 2. Hence, the visibility as defined                     brightness, that is, their photophysical behavior. Whether this
by the ratio SN/R is between 1.5 and 3, clearly sufficient for the                   is induced by isomerization of the label or is due to two dis-
reliable detection of single-molecule redox events.                                  tinct binding conformations of the label to the protein remains
   Analogous to the FCS data we further evaluated the single-                        to be investigated.
molecule imaging data in terms of fluorescence intensity distri-                        Our new approach should also be viewed in context with
butions.[21] Probability density functions of the detected signals                   earlier single-molecule approaches for the detection of
were calculated for more than 500 individual molecules for                           changes in protein activity. In previous reports the reaction
both Cy5-azurin and Cy5-azurinZn and compared for the differ-                        state of individual proteins which were associated with confor-
ent oxidation/reduction conditions (Figure 5 B and C). The dis-                      mational changes of the protein were monitored by FRET be-
tributions were characterized by a single maximum at signal                          tween two artificial fluorescence labels.[24–27] In another elegant
intensities of 285 and 225 counts/5 ms for Cy5-azurin and Cy5-                       study by Lu et al.[23] the change in fluorescence of a protein’s
azurin , respectively. The two populations, which were re-                           cofactor was used to follow enzyme activity. Here we have
solved in the FIDA analysis, are visible as asymmetries/should-                      shown that single-molecule methodology can be generally ap-
ers at 200 and 350 count/5 ms in the distributions shown in                          plied to a large class of proteins exhibiting changes in absorp-
Figure 5 B and C, respectively. As already seen in the FIDA anal-                    tion on changes in the (electro)chemical state, which where
ysis, reversible switching of the intensity distribution from 285                    not accessible at the single-molecule level so far. Our approach
to 150 count/5 ms was observed in the case of Cy5-azurin. The                        has some distinct advantages: 1) the change in state must not
switching was followed through several reduction/oxidation                           necessarily be accompanied by a change in structure, 2) the
cycles, whereas switching was absent for Cy5-azurinZn. Note                          small and fairly photostable organic dye introduced for detec-
that there was no significant fraction of inactive molecule in                       tion adds additional optical selectivity relative to the unavoida-
the Cy5-azurin sample; such molecules would result in a                              ble background of natural fluorophores in complex biological
double peak in the probability density functions, where only                         systems, 3) inherent spatial sensitivity, either due to the FRET
the ratio of the peaks is shifted on oxidation or reduction.                         process on the nanometer length scale or the spatial super-res-
                                                                                     olution on the tens of nanometer length scale obtained for
                                                                                     wide-field imaging, and 4) the broad and complementary time
                                                                                     window of the two techniques applied allows observation of

ChemPhysChem 2005, 6, 1381 – 1386                2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim                   1385
                                                                                                                                                 T. Schmidt et al.

dynamic processes between 100 ns and several minutes. The                          [8] T. Forster, Ann. Phys. 1948, 2, 55 – 75.
last two points are presumably of major biological interest as                     [9] H. Nar, A. Messerschmidt, R. Huber, M. Vandekamp, G. W. Canters, J.
                                                                                       Mol. Biol. 1991, 218, 427 – 447.
they will permit the analysis of the local (ca. 40 nm) oxidation                 [10] T. Fukaminato, T. Sasaki, T. Kawai, N. Tamai, M. Irie, J. Am. Chem. Soc.
conditions within a cell,[28, 29] and to further unravel the molecu-                   2004, 126, 14 843 – 14 849.
lar mechanisms associated with functions of proteins, which so                   [11] D. Magde, E. L. Elson, W. W. Webb, Phys. Rev. Lett. 1972, 29, 705 – 708.
far were inaccessible at the single-molecule level. Experiments                  [12] M. Eigen, R. Rigler, Proc. Natl. Acad. Sci. USA 1994, 91, 5740 – 5747.
                                                                                 [13] P. Kask, K. Palo, D. Ullmann, K. Gall, Proc. Natl. Acad. Sci. USA 1999, 96,
to explore these and further applications are currently under                          13 756 – 13 761.
way.                                                                             [14] Y. Chen, J. D. Muller, P. T. C. So, E. Gratton, Biophys. Chem. Biophys. J.
                                                                                       1999, 77, 553 – 567.
                                                                                 [15] I. M. C. van Amsterdam, M. Ubbink, G. W. Canters, Inorg. Chim. Acta
Acknowledgements                                                                       2002, 331, 296 – 302.
                                                                                 [16] K. Weisshart, V. Jungel, S. J. Briddon, Curr. Pharm. Biotechnol. 2004, 5,
                                                                                       135 – 154.
We thank Ellen de Waal and Thyra de Jongh for expert advice on
                                                                                 [17] W. H. Press, S. A. Teukolsky, W. T. Vetterling, B. P. Flannery, Numerical Rec-
protein purification and sample handling. Dr. G. A. Blab and Dr.                       ipes in C: The Art of Scientific Computing, 2nd ed., Cambridge University
C. Rischel are acknowledged for helpful discussions. We are grate-                     Press, Cambridge, UK, 1992.
ful to Prof H. Spaink for giving access to the Confocor2 system.                 [18] T. Schmidt, G. J. Schutz, W. Baumgartner, H. J. Gruber, H. Schindler, J.
                                                                                       Phys. Chem. 1995, 99, 17 662 – 17 668.
This work is part of the research programs “Physical Biology” and
                                                                                 [19] J. Widengren, P. Schwille, J. Phys. Chem. A 2000, 104, 6416 – 6428.
“Biomolecular Physics” of the Stichting voor Fundamenteel On-                    [20] Y. Harada, K. Sakurada, T. Aoki, D. D. Thomas, T. Yanagida, J. Mol. Biol.
derzoek der Materie (FOM) and the Stichting voor Aard en Lev-                          1990, 216, 49 – 68.
enswetenschappen (ALW), financially supported by the Neder-                      [21] T. Schmidt, G. J. Schutz, W. Baumgartner, H. J. Gruber, H. Schindler, Proc.
                                                                                       Natl. Acad. Sci. USA 1996, 93, 2926 – 2929.
landse Organisatie voor Wetenschappelijk Onderzoek (NWO).
                                                                                 [22] L. Edman, R. Rigler, Proc. Natl. Acad. Sci. USA 2000, 97, 8266 – 8271.
                                                                                 [23] H. P. Lu, L. Y. Xun, X. S. Xie, Science 1998, 282, 1877 – 1882.
                                                                                 [24] Y. Chen, D. H. Hu, E. R. Vorpagel, H. P. Lu, J. Phys. Chem. B 2003, 107,
Keywords: fluorescence correlation spectroscopy · FRET
                                                                                       7947 – 7956.
(fluorescence resonant energy transfer) · metalloproteins ·                      [25] T. J. Ha, A. Y. Ting, J. Liang, W. B. Caldwell, A. A. Deniz, D. S. Chemla, P. G.
redox chemistry · single-molecule studies                                              Schultz, S. Weiss, Proc. Natl. Acad. Sci. USA 1999, 96, 893 – 898.
                                                                                 [26] M. Borsch, M. Diez, B. Zimmermann, R. Reuter, P. Graber, FEBS Lett.
 [1] L. Stryer in Biochemistry, Freeman, New York 1997, pp. 145 – 440.                 2002, 527, 147 – 152.
 [2] S. Weiss, Science 1999, 283, 1676 – 1683.                                   [27] T. Funatsu, Y. Harada, M. Tokunaga, K. Saito, T. Yanagida, Nature 1995,
 [3] T. Horio, J. Biochem. 1958, 45, 267 – 279.                                        374, 555 – 559.
 [4] T. Horio, T. Higashi, H. Matsubara, K. Kusai, M. Nakai, K. Okunuki, Bio-    [28] G. T. Hanson, R. Aggeler, D. Oglesbee, M. Cannon, R. A. Capaldi, R. Y.
     chim. Biophys. Acta 1958, 29, 297 – 302.                                          Tsien, S. J. Remington, J. Biol. Chem. 2004, 279, 13 044 – 13 053.
 [5] E. Vijgenboom, J. E. Busch, G. W. Canters, Microbiology (Reading, U.K.)     [29] C. T. Dooley, T. M. Dore, G. T. Hanson, W. C. Jackson, S. J. Remington, R. Y.
     1997, 143, 2853 – 2863.                                                           Tsien, J. Biol. Chem. 2004, 279, 22 284 – 22 293.
 [6] R. Schmauder, S. Alagaratnam, C. Chan, T. Schmidt, G. W. Canters, T. J.
     Aartsma, unpublished results.
 [7] J. R. Lacowicz in Principles of Fluorescence Spectroscopy, Kluwer/Plenum,   Received: December 16, 2004
     New York, 1999, pp. 367 – 391.                                              Revised: April 6, 2005

1386                 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim                                ChemPhysChem 2005, 6, 1381 – 1386

To top