Fluorescent Nitrones for the Study of ROS Formation with Subcellular Resolution

Reactive oxygen species (ROS) are the prize we are paying for an energy-efficient life under oxygen. They play a role in many diseases such as atherosclerosis, hypertension, ischemiareperfusion injury, inflammation, type-2 diabetes, certain neurodegenerative diseases, even cancer and, certainly, aging (Kohen & Nyska, 2002). Among these ROS are several radicals such as the hydroxyl, peroxy, alkoxy as well as the superoxide anion radical (Boveris, 1977). Direct measurement of the radicals is hampered by their short half lifes in the range of nanoto microseconds. They can, however, be trapped by addition to nitrones leading to relatively stable nitroxides with t1⁄2 of minutes (Janzen, 1971). Respiring mitochondria are a major source of ROS, particularly of the superoxide anion radical, which is formed in complexes I and III (Cadenas & Davies, 2000; Droge, 2002; Inoue et al., 2003; Turrens, 2003).


Introduction
Reactive oxygen species (ROS) are the prize we are paying for an energy-efficient life under oxygen.They play a role in many diseases such as atherosclerosis, hypertension, ischemiareperfusion injury, inflammation, type-2 diabetes, certain neurodegenerative diseases, even cancer and, certainly, aging (Kohen & Nyska, 2002).Among these ROS are several radicals such as the hydroxyl, peroxy, alkoxy as well as the superoxide anion radical (Boveris, 1977).Direct measurement of the radicals is hampered by their short half lifes in the range of nanoto microseconds.They can, however, be trapped by addition to nitrones leading to relatively stable nitroxides with t½ of minutes (Janzen, 1971).Respiring mitochondria are a major source of ROS, particularly of the superoxide anion radical, which is formed in complexes I and III (Cadenas & Davies, 2000;Dröge, 2002;Inoue et al., 2003;Turrens, 2003).
A couple of years ago we have described the synthesis and first application of a fluorescent nitrone composed of tert-butyl-nitrone and a p-nitro-stilbene moiety, which can be used for the detection of ROS with subcellular resolution.Short-lived radicals add to the nitrone under formation of a nitroxide radical which then quenches the fluorescence of the p-nitrostilbene (Hauck et al., 2009).Similar double labels have previously been synthesized or at least suggested, partially in collaboration with us.Likhtenshtein and Hideg suggested in the eighties to couple nitroxides to a fluorophore which will become fluorescent only after reduction of the nitroxide moiety in viable biological systems (Bystryak et al., 1986).Hideg used fluorescent pyrrolines which can be oxidized to nitroxides mainly by singlet oxygen (Kalai et al., 1998).Rosen and collaborators investigated a fluorescent nitrone much like ours composed of nitrobenzene instead of nitrostilbene.The authors did not observe a decrease in the fluorescence upon reaction of the nitrone with α-hydroxyethyl radicals which was probably due to an only very small concentration of the nitroxide formed in this reaction (Pou et al., 1995).To the best of our knowledge, this work has not been continued.Somewhat different approaches for the detection and trapping of ROS were taken by Bottle et al., 2003, Heyne et al., 2007and Blough et al., 1988.Here we describe some extended studies with the p-nitrostilbene-tert-butyl-nitrone (1) under a variety of physiological conditions i.e., employing inhibitors of components of the respiratory chain, the F1-F0-ATP synthase and the membrane potential as well as similar studies with a corresponding coumaryl-styryl-tert-butyl-nitrone derivative (2) and finally, a third compound based on 4-pyrrolidine-1,8-naphthylimido-methylbenzene as fluorophore (3).

p-Nitrostilbene-tert-butyl-nitrone, 1
Fig. 1 shows the reaction of the previously employed fluorescent nitrone 1 with the hydroxyl radical under formation of the corresponding nitroxide.The hydroxyl radical was generated by the Fenton reaction and the product extracted from the aqueous phase with degassed ethyl acetate yielding an EPR spectrum composed of 6 lines typical for hydroxyl radical adducts (Hauck et al., 2009).
When cultures of various cell lines, i.e., HeLa, COS 7 and CHO, were incubated with 1 and subsequently washed, it had almost exclusively accumulated in mitochondria as shown by co-staining with tetramethylrhodamine ethylester (TMRE, Farkas et al., 1989) and confocal laser spectroscopy (Fig. 2).Without any further addition the fluorescence slowly decreased to almost none within about 20 min.However, generation of hydroxyl radicals by the Fenton reaction (Walling, 1975) reduced this time to 20 sec (Fig. 3, Hauck et al., 2009).Similar results were obtained in presence of the complex I and III inhibitors, rotenone and antimycin A, but within a timeframe of 40 to 60 sec.Under these conditions both complexes are known to produce substantial amounts of the superoxide anion radical (Dlaskova et al., 2008;Han et al., 2001).In these early studies a quasi-confocal microscope was used (Axiovert 440, equipped with an ApoTome, Carl Zeiss, Jena) which has broad bandwidth filters, only.Thus, not allowing for monitoring small shifts in the emission spectra of 1 upon addition of the radical as had been observed in cell-free controls.The accumulation of 1 in mitochondria came unexpected but is possibly due to an effect previously observed with so-called Skulachev ions, derivatives of ubiquinone or plastoquinone with long hydrophobic side chains composed of up to 10 isoprene units and a positively charged triphenylphosphonium group at the end, e.g., SkQ1 as introduced by Vladimir Skulachev.The positive charge is not localized but spread over the aromatic rings, thus allowing for membrane permeation due to the already negative potential of cells with respect to their outside and even more so of mitochondria, the only human organelle with a negative potential with respect to the cytoplasm.Once inside, they are trapped by the opposite potential outside (Severin et al., 2010).Similar effects have been observed with certain amphiphilic dipolar compounds to which 1 could belong (M.V. Skulachev, personal communication).The fluorescence half-life of 1 in mitochondria was studied under a variety of conditions.Data are summarized in Table I together with those from the coumaryl derivative 2. But why look for another double label? 1 has an absorption maximum at 383 nm which is well separated from the emission at 568 nm.However, only recently have confocal laser microscopes been equipped with lasers of 405 nm, most commercial instruments operate at 480 nm and above, rather outside the absorption range of 1.Moreover, stilbenes as fluorophores pose yet another problem, cis-trans isomerization upon irradiation leading to a substantial shift in emission wavelength which is accompanied by photobleaching and also recovery rendering interpretation of data more complex (Fig. 6).
Unexpectedly, the absorbance and fluorescence properties of 1 and 2 do not differ very much as shown in Fig. 5, A & B. Unfortunately, 2 turned out to be far more toxic to cells in comparison with 1. Growing adherent MCF-7 cells, a human breast adenocarcinoma cell line, were exposed to these compounds for three days.After fixation of viable cells with trichloroacetic acid, the damaged cells were removed by washing and cell proteins of the remaining cells were stained with sulforhodamine and measured photometrically at 570 nm (Skehan et al., 1990).
In a rough estimate assuming sigmoidal behavior in a semi-logarithmic plot the half maximal dosis, IC50, of 1 was about 300 µM as compared to 40 µM for 2. Simultaneous irradiation at 366 nm for 3 min further reduced this value to 30 µM.The known toxicity of coumarin may play a role here (Fig. 7, A & B; Oodyke, 1974).Therefore spin-trapping experiments in the presence of inhibitors were primarily carried out with 1.However, 2 may be employed as well by using lower concentrations of 10 to 20 µM.The effect of rotenone as inhibitor of complex I of the respiratory chain on the formation of the superoxide anion radical formation in HeLa cells (Hauck et al., 2009) was followed with 2 (Fig. 8).The initial fluorescence and after 100 or 90 sec, respectively, is shown in the absence and presence of rotenone.The time course of this decay as measured in 0.5 sec intervals is shown as well.Clearly, inhibition of complex I significantly increases ROS formation.The figure also reveals that the intracellular distribution of 2 is not as enriched in mitochondria as had been found with 1.Evidence that quench did not result, or at least not largely, from photobleaching came from experiments in which the shutter was closed during measurements.However, some non-specific redox reactions cannot be excluded.

Detailed studies with p-nitrostilbene-tert-butyl-nitrone, 1
Spin-trapping of mainly superoxide anion radicals formed in mitochondria under various conditions was followed in HeLa cells as previously (Hauck et al., 2009) and as well for better comparison with data employing 2. Fig. 9 shows the fluorescence decay after 100 sec in the absence of any inhibitor.All subsequent data are based on this control .Results are summarized in Fig. 16.
The inhibitors of complexes I, rotenone, and complex III, antimycin A, have the strongest effect, particularly in combination.KCN is known to inhibit complex IV (Leavesley et al., 2010) and oligomycin blocks ATP synthesis by binding to the oligomycin sensitivity conferring protein, OSCP, of the F0 moiety of F1F0-ATP synthase (Xu et al., 2000).The latter leads to an increased membrane potential and drives the respiratory chain backwards.In the absence of sufficient substrates for NADH this is known to produce superoxide anion radicals in complex I as in complex III via reaction of oxygen with the coenzyme Q semiquinone radical (Muller et al., 2004).
Rather interesting is the effect of the protonophore CCCP (carbonyl-cyanide m-chlorophenyl hydrazone) which reduces the membrane potential (Nieminen et al., 1994) and thus, apparently also reduces ROS formation.The half-life of the fluorescence in its presence is almost doubled as compared to the control corroborating the opposite effect of oligomycin.In order to allow for excitation at longer wavelengths a third compound was synthesized according to the scheme in Fig. 17   but fluorescence recovered within four to five seconds and then decayed completely over the next 15 min.What could cause fluorescence recovery?The half-life of nitroxides as Tempone or Tempamine in viable systems is in the order of 30 min at most due to the reducing milieu in the cell (Berliner, 1991).Hence, the fluorescence being quenched by the radical could come back.There is however, also the possibility that differences are due to the experimental setup, i.e., a continuous flow device for the medium at the ApoTome to which the inhibitor was added, whereas in these experiments a concentrated solution of antimycin A etc. was manually injected directly into the medium surrounding the adherent cells.

4-Bromobenzaldehyde-dimethyl-acetal
4-Bromobenzaldehyde (94.9 mmol) was dissolved in 120 ml of dry methanol under an argon atmosphere, 96 µl titanium chloride was added and the solution stirred for 20 minutes.After addition of 240 µl trietylamine and 15 minutes of stirring, water was added and the product was extracted with diethyl ether.The organic phase was dried and the acetal isolated by evaporation of the solvent. [
Fluorescent Nitrones for the Study of ROS Formation with Subcellular Resolution 363

UV/VIS and fluorescence spectroscopy
500 µM stock solutions of the spin traps in dimethyl sulfoxide containing 1 % C12E9 were employed.For fluorescence spectra the solutions were added to 25 mM phosphate buffer, pH = 7.2, with 1 % Triton-X 100 to a final concentration of 1 µM.To simulate photobleaching in the fluorescence microscope, the solutions were irradiated for three minutes with a blue LED light source (λ = 366 nm) and compared to samples without prior irradiation.

Cytotoxicity
Cytotoxicity of the spin traps was determined by the sulforhodamine assay.170 µl of a cell suspension of 8000 cells/ml of Hela or MCF-7 cells were added to 20 ml of RPMI medium and 1 ml each was pipetted into a 24 wells plate and incubated for 48 h at 37 °C under 5 % CO2.After this time, the cells were incubated for another 72 h with RPMI medium containing 0.5 % DMSO and the spin trap at six concentrations varying from 0 to 200 µM.The reaction was stopped by addition of 100 µl of 50 % trichloroacetic acid for 1 h at 4 °C and subsequently washed four times with cold water and then dried for 24 h at room temperature followed by addition of 250 µl of sulforhodamine B solution (0.4 % in 1 % acetic acid; Sigma, Taufkirchen).After 30 min wells were washed three times with cold water and another three times with 1 % acetic acid, dried for 24 h at room temperature and extinction measured at 570 nm after addition of 1 ml of 10 mM Tris base solution, pH 10.0.
In case of 3 concentrations were varied from 0 to 5 µM in RPMI medium containing 0.5 % DMSO and 0.5 % ethanol in addition.

Fluorescence microscopy measurements
These were carried out using a quasi-confocal microscope (Axiovert 440 equipped with an ApoTome, Carl Zeiss, Jena) as previously described (Hauck et al., 2009).For corresponding measurements with 3 a Nikon Eclipse E 600 confocal microscope equipped with a Hamamatsu ORCA-ER camera was employed.After 20 minutes of incubation with 1 µM spin trap and 0.5 % DMSO the cells were washed three times with RPMI medium and the coverslip was mounted on a chamber and put under the microscope.Imaging was achieved by laser excitation at 488 nm.
To determine the half-life of fluorescence, a representative cell was defined as region of interest (ROI) and the evolution of average intensity of the ROI was investigated in the presence and absence of various inhibitors of components of mitochondrial proteins.

Figure 1 .Fluorescent
Figure 1.Structural formula of the p-nitrostilbene-tert-butyl-nitrone 1 and its reaction with the hydroxyl radical

Figure 4 .Figure 5 .
Figure 4. Synthetic scheme for the synthesis of 2

Figure 6 .
Figure 6.Photobleaching of 2 (CSN, A) und 1 (NSN, B) upon irradiation at 310 nm for 1 min.The recovery of 1 with time is shown in panel C.

Figure 7 .Fluorescent
Figure 7. Cytotoxicity of 1 (A) and 2 (B) as determined in MCF-7 cells after a three days exposure.B, lower trace upon additional irradiation at 366 nm.

Figure 8 .
Figure 8. False color representation of the fluorescence decay of 2 upon trapping of ROS in HeLa cells in the absence (upper panels) and presence of 10 µM rotenone (lower panels), 0 and 100 sec after addition of the inhibitor as well as the time course of these changes (bottom panel).