Caged AG10: new tools for spatially predefined mitochondrial uncoupling

Abstract

The study of mitochondria and mitochondrial Ca²⁺ signalling in localised regions is hampered by the lack of tools that can uncouple the mitochondrial membrane potential (ΔΨ_m) in a spatially predefined manner. Although there are a number of existing mitochondrial uncouplers, these compounds are necessarily membrane permeant and therefore exert their actions in a spatially unselective manner. Herein, we report the synthesis of the first caged (photolabile protected) mitochondrial uncouplers, based on the tyrphostin AG10. We have analysed the laser photolysis of these compounds, using ¹H NMR and HPLC, and demonstrate that the major product of caged AG10 photolysis is AG10. It is shown that photolysis within single smooth muscle cells causes a collapse of ΔΨ_m consistent with photorelease of AG10. Furthermore, the effect of the photoreleased AG10 is localised to a subcellular region proximal to the site of photolysis, demonstrating for the first time spatially predefined mitochondrial uncoupling.

---

Introduction

Mitochondria, by generating adenosine triphosphate (ATP), modulating intracellular Ca²⁺ signalling, regulating amino acid and lipid synthesis and initiating cell death, play a central role in controlling cell activities. Mitochondria coordinate these activities in highly localised regions to regulate the overall activity and performance of the cell. For example, mitochondria regulate Ca²⁺ release from intracellular stores, in localised regions of the cell, via an interaction with individual Ca²⁺ release channels¹ and control sub-plasma membrane [Ca²⁺] to manage exocytosis in nerve terminals² and adrenal chromaffin cells.³ Yet relatively little is known about how mitochondria acting in localised, restricted, regions control the overall performance of the cell.

A limitation in advancing understanding of the contribution of mitochondria to signalling and metabolism is an inability to inhibit their activity in restricted regions of the cell. Hitherto, it has been possible only to examine their role in various cell activities by inhibiting mitochondrial activity (e.g. pharmacologically) throughout the entirety of a cell. However, pharmacological disruption of the entire mitochondrial complement of a cell has a broad spectrum of effects, lacks spatial discrimination and provides limited information regarding the localised biological activities of the organelle. Herein, we describe a method for local inhibition of mitochondria in restricted regions of the cell.

Mitochondrial ATP production and uptake of Ca²⁺ are each driven by the large electrical potential [ΔΨ_m, about −150 to −180 mV]⁴ that exists across the inner mitochondrial membrane. Certain chemicals, with extensive conjugated π-bond systems and acidic protons, may dissipate proton and charge gradients across biological membranes. These chemicals act as protonophores as they can move across membranes either as protonated acids or as the deprotonated conjugated bases. By cycling across a membrane they increase proton conductance, which, in mitochondria, uncouples proton gradient formation from ATP production and causes ΔΨ_m to collapse. Protonophores such as carbonyl cyanide 3-chlorophenylhydrazone (CCCP), carbonyl cyanide 4-trifluoromethoxy phenylhydrazone (FCCP) or 2,4-dinitrophenol (DNP; see ESI‡ for chemical structures) are commonly used to dissipate ΔΨ_m and thus inhibit mitochondrial functions such as Ca²⁺ uptake. However, as these compounds are lipophilic and hence membrane permeant, it is difficult to apply them to specific predefined, subcellular regions.

Photolabile protected or “caged” compounds are biologically active molecules that have been rendered inert by the introduction of a photolabile protecting group to an important functionality. Light-evoked removal of the caging group releases the active compound and restores the pharmacological activity of the molecule. As only light is required to remove the caging group, this technique is non-invasive and thus particularly applicable to biological systems. In addition, temporal and spatial control which exceeds that of traditional pharmacological administration is afforded.^5–7 A number of biological molecules and systems have been studied successfully following the development of caged molecular probes e.g.D-myo-inositol 1,4,5-trisphosphate (InsP₃),^8,9glutamate,^10–13γ-aminobutyric acid (GABA),¹⁴Ca^2+15–22 and TRPV1^23–29 leading to insights into the distribution of receptors and kinetics of their activation.^5,7

One approach to study how mitochondria acting in localised regions control the overall performance of the cell is to photolyse a caged protonophore. Photolysis can be achieved locally in small, restricted regions of the cell, to inhibit mitochondrial activity only at the site of photolysis; however, there are presently no caged protonophores. Light- and phototsensitiser-mediated inactivation of mitochondrial function has been reported previously, the mechanism of which is likely due to singlet oxygen formation and cellular damage, including mitochondrial uncoupling;³⁰ however, there are no reports of inactivation of a sub-population of mitochondria within individual cells. We have chosen to investigate caging of the tyrphostin AG10 (2);³¹ caging of the important phenol group is viable from a chemical perspective and will remove the ability of the compound to act as a protonophore. Herein, the syntheses of two 4,5-dimethoxy-2-nitrobenzyl (DMNB)-caged AG10 derivatives (3 and 5) are described. The photolysis of these compounds was analysed using ¹H NMR and HPLC. The caged AG10 derivatives were then introduced into freshly-isolated smooth muscle cells and localised, subcellular photolysis achieved. Fluorescent imaging of ΔΨ_m showed that photolysis of the caged AG10 derivatives caused a localised ΔΨ_m depolarisation that remained largely confined to the site of photolysis even up to 20 min after photolysis. ΔΨ_m depolarisation occurred neither by the photolysis light alone in the absence of caged AG10 nor in the presence of caged AG10 without exposure to photolysis light. Photolysis of caged AG10 in a cell that had been treated with a cocktail of radical scavengers indicated that ΔΨ_m depolarisation was also not a result of free-radical production during the uncaging procedure. These caged AG10 derivatives will be useful for determining the influence that mitochondria acting in specific, small subcellular regions have over localised cellular processes.

Results

The synthesis of AG10 and caged derivatives

The synthesis of AG10 (2) commenced from 4-hydroxybenzaldehyde (1, Scheme 1). Piperidine-catalysed Knoevenagel-like condensation of 4-hydroxybenzaldehyde with malononitrile afforded AG10 in good yield. The synthesis of DMNB-caged AG10 (5) directly from AG10 was not possible (see ESI‡ for further details). However, reaction of 4-hydroxybenzaldehyde and 4,5-dimethoxy-2-nitrobenzyl bromide with K₂CO₃ in dimethylformamide (DMF)§ at 40 °C afforded a reasonable (51%) yield of the DMNB benzaldehyde derivative (4, Scheme 1). Condensation of 4 with malononitrile, in the presence of piperidine in ethanol under reflux, provided the desired DMNB-caged AG10 (5) in 70% yield (Scheme 1).

Scheme 1 The synthesis of AG10 (2), DMNB-caged AG10 (5) and CDMNB-caged AG10 (3).

In order to synthesise the carbonate-linked DMNB-caged AG10 derivative (CDMNB-caged AG10, 3), DMNB alcohol was reacted with phosgene to afford DMNB chloroformate. This compound was then reacted with AG10 in the presence of three equivalents of triethylamine (Scheme 1) to furnish the desired compound (3) in modest yield. Compound 3 is reasonably stable and was amenable to purification by silica gel column chromatography.

In an alternative strategy, 4-hydroxybenzaldehyde was treated with DMNB chloroformate in the presence of triethylamine, affording the desired DMNB carbonate derivative in good yield (88%, see ESI‡ for details). However, it proved impossible to condense the benzaldehyde derivative with malononitrile without destroying the carbonate linkage, presumably as a result of piperidine attacking the carbonatecarbonyl. Attempts to perform the Knoevenagel condensation under alternative conditions also failed.

To determine the optimum wavelength for photolysis of 3 and 5, the UV/Vis spectra of these compounds were obtained. In addition, the UV/Vis spectrum of AG10 was also taken (see ESI‡ for spectra).

Laser photolysis of the caged AG10 derivatives released AG10

Before the compounds were used for in vitro studies, their characteristics when subjected to laser photolysis were investigated. Both the DMNB-caged and the CDMNB-caged AG10 derivatives underwent photolysis when irradiated by a 950 mW laser with a wavelength of 355 nm. As it is possible for a number of photolytic degradation products to form, it was important to establish that AG10 was being produced once photolysis occurred, as expected. Photolysis experiments were conducted as outlined below and the products were analysed by ¹H NMR and HPLC. ¹H NMR studies indicated that the major photolytic products of DMNB-caged and CDMNB-caged AG10 were AG10 and an aldehyde. We assume that the aldehyde formed is the nitrosoaldehyde derived from the caging group, which has previously been shown to be a product of photolysis.⁶

Analysis of samples photolysed for 1–5 min showed the disappearance of the caged compound and the appearance of AG10 (see ESI‡ for a representative example). Inclusion of a known concentration of hexamethyldisiloxane as a standard allowed quantitative analysis of the photolysis by ¹H NMR. To determine whether ¹H NMR was sensitive enough to detect low levels of photolysis, the samples were also analysed by HPLC using a previously measured calibration curve. The DMNB-caged AG10 derivative was uncaged more effectively than the CDMNB-caged derivative when photolysed with a 355 nm laser (see ESI‡ ). Photolysis of the DMNB-caged derivative 5 for 5 min gave rise to 35% of the total material being AG10, when analysed by ¹H NMR, and 46% being AG10, when analysed by HPLC. Under the same conditions, photolysis of the CDMNB-caged AG10 derivative 3, only 26% (¹H NMR) to 29% (HPLC) of the total material was observed to be AG10. It is generally accepted that HPLC analysis is more sensitive than ¹H NMR analysis when quantitative data are required. More AG10 (produced by photolysis) was detected by HPLC than was detected by ¹H NMR and hence greater values for the percentage of AG10 present were observed. However, the general trends in amount of photolysis were similar for both methods of detection.

AG10 acts as an uncoupler to depolarise mitochondria

Freshly-isolated, single smooth muscle cells were loaded with the fluorescent dye TMRE (10 nM) to visualise the mitochondrial membrane potential (ΔΨ_m). In control, individual mitochondria appeared as punctate TMRE staining throughout the cell (Fig. 1Aii). For the purpose of image clarity on reproduction, individual images were filtered with an open, edge-detection filter (“top-hat” filter in Metamorph software) to enhance image contrast (the “raw” fluorescence image in Fig. 1Ci is shown filtered in Fig. 1Cii); however, all TMRE fluorescence values shown are from unfiltered data. Free, uncaged AG10 (100 μM, added to the bathing solution) caused a loss of TMRE fluorescence, indicating ΔΨ_m depolarisation (Fig. 1B and C). Subsequent addition of the commonly used uncoupler CCCP did not cause any further change in TMRE fluorescence. Thus synthetic AG10 did depolarise mitochondria as reported.³¹

Fig. 1 AG10 rapidly caused mitochondrial depolarisation. Freshly-isolated, single smooth muscle cells (brightfield, Ai, 10 μm scale bar) were loaded with the ΔΨ_m-sensitive dye TMRE (10 nM, Aii and C). TMRE fluorescence was predominantly localised to mitochondrial structures and was emphasised for display only by localised edge-detection analysis using a “top-hat” filter (Ci – before, and Cii – after applying the filter). AG10 (100 μM, added to the bathing solution, B) caused a loss of TMRE fluorescence, indicating ΔΨ_m depolarisation (fluorescence values from the regions shown in Aii are plotted in B). Subsequently, the uncoupler CCCP (2 μM) did not cause any further change in TMRE fluorescence.

Intracellular localised photolysis of DMNB-caged AG10 evokes a sustained, localised ΔΨ_m depolarisation

DMNB-caged AG10 (25 μM) was introduced into the cytosol of single smooth muscle cells using the access afforded by the whole-cell patch clamp electrode. A period of photolysis (4 min flashlamp fired at its maximum cycle rate of 0.1 Hz) of DMNB-caged AG10 depolarised ΔΨ_m as revealed by the localised loss of TMRE fluorescence (Fig. 2, example shown is from n = 13 cells). In control, individual mitochondria can be seen as punctate TMRE staining throughout the cell (Fig. 2Aii and iii). After photolysis of AG10, TMRE fluorescence was reduced only at the site of photolysis (Fig. 2Aiv) demonstrating ΔΨ_m depolarisation at that site.

Fig. 2 Localised mitochondrial depolarisation following photolytic release of AG10 in a small region of the cell. Isolated smooth muscle cells were loaded with TMRE (10 nM) then voltage-clamped in the whole-cell configuration. Ai: brightfield image of a cell (see patch electrode, bottom left) with the site of localised photolysis highlighted. Aii–v: TMRE fluorescence was reduced only in the region of photolytic release of AG10 (compare Aiii, before, with Aiv, immediately after photolysis; red region in v showed a decrease in TMRE fluorescence whereas regions above and below this did not, vi) and did not spread throughout the cell over the following 20 min (Av). Note that the cell contracted slightly over the 20 min period, causing a slight fluctuation in fluorescence, and also that TMRE fluorescence increased in regions neighbouring the site of AG10 release, presumably due to dye relocation to these sites from the regions of ΔΨ_m depolarisation. B: no loss of TMRE fluorescence was observed in a cell (brightfield, i) exposed to photolysis light in the absence of caged AG10 (compare TMRE before, ii, and after, iii, photolysis in the central region of the cell, red region in iv and corresponding trace, v). C: in the absence of UV light, caged AG10 did not affect TMRE fluorescence (shown before, ii, immediately after, iv, and 20 min after, vi, whole-cell patching a smooth muscle cell, i, with an electrode containing DMNB-caged AG10, 25 μM, iii). TMRE fluorescence values for the three regions drawn (i) were normalised to initial values (ΔF/F₀, vii). D: the photolysis reaction that is occurring to release AG10 (2).

The region of ΔΨ_m depolarisation neither spread through the cell nor did mitochondria repolarise over the next 20 min (Fig. 2Av and vii). This result suggests that, on photolytic release, AG10 incorporates into nearby intracellular membranes—causing ΔΨ_m depolarisation there—and does not equilibrate throughout the cell, presumably due to its hydrophobic nature. As a control, in the absence of caged AG10, repetitive exposure to UV flash light produced no reduction of TMRE fluorescence (Fig. 2B). TMRE fluorescence was also unaffected by caged AG10 in the absence of UV light (Fig. 2C).

To rule out the possibility of the mitochondrial depolarisation being caused by indiscriminate cellular damage effected by the by-product of the caging group, 4,5-dimethoxy-2-nitrobenzaldehyde was applied to smooth muscle cells (25 μM), by dialysis into the cytosol from the filling solution of a whole-cell patch pipette. In both the absence and presence of UV light this compound had no effect on the cells (data not shown).

The sustained (20 min) nature of the loss of TMRE fluorescence (Fig. 2) was not due to an inability of the dye to measure ΔΨ_m repolarisation. TMRE fluorescence reports ΔΨ_m depolarisation and repolarisation events in individual mitochondria and the entire mitochondrial complement of the cell. To demonstrate this, spontaneous ΔΨ_m depolarisation of individual mitochondria was induced by increasing the intensity of the fluorescence illumination light four-fold (by removing a neutral density 4 filter from the excitation lightpath) and increasing the concentration of TMRE (from 10 to 25 nM; Fig. 3A–C) as characterised previously;³² ΔΨ_m repolarisation occurred within periods as short as 2 s (40 s shown in Fig. 3B). Thus the longer period of ΔΨ_m depolarisation evoked by release of AG10 is not due to an inability of TMRE to report ΔΨ_m recovery.

Fig. 3 Transient ΔΨ_m depolarisation in individual mitochondria can be observed in single smooth muscle cells loaded with TMRE. A: mitochondrial TMRE fluorescence for approximately half of one intact smooth muscle cell is shown; two selected subregions are shown at enlarged scale and fluorescence intensity of four apparently individual, neighbouring mitochondria were measured (regions shown circled in four colours that relate to the four coloured traces in graphs Bii and Cii). Bi and Ci: selected frames at times indicated show localised regions of TMRE fluorescence fluctuation (red arrows). Bii and Cii: fluorescence intensity (F) of individual regions of interest of corresponding colour (shown in panel A) normalised to initial fluorescence values (F₀) show that in both cases the regions circled in red transiently lose fluorescence (depolarise) and then regain it (repolarise).

CDMNB-caged AG10 evokes more rapid ΔΨ_m depolarisation than DMNB-caged AG10

Initial studies using CDMNB-caged AG10 suggest that this compound may be of more use than DMNB-caged AG10 in photolysis studies (Fig. 4). The increased solubility of CDMNB-caged AG10 in DMSO and water allowed a greater concentration of this compound to be added to the intracellular pipette solution (62 μM versus 25 μM). Consequently, more significant and rapid ΔΨ_m depolarisation was observed when CDMNB-caged AG10 (3) was employed.

Fig. 4 CDMNB-caged AG10 photolysis evoked ΔΨ_m depolarisation. Isolated smooth muscle cells were loaded with TMRE (10 nM) and voltage-clamped in the whole-cell configuration. CDMNB-caged AG10 (62 μM) was introduced to the cytosolvia the pipette solution. Ai: brightfield image of a cell (see patch electrode, top left) with the site of localised photolysis (‘flash site’) highlighted. TMRE fluorescence (Aii–ix) was predominantly localised to mitochondria and was reduced only in the region of photolytic release of CDMNB-caged AG10. The images (Aii–ix) correspond to the time points indicated in B. TMRE fluorescence measurements (B) of the regions in Aii show that fluorescence decreased in only the region of photolysis (red). The increase in TMRE fluorescence in neighbouring regions (blue and green) probably arises from the redistribution of TMRE out of the depolarised mitochondria in the flash site. The increased solubility of CDMNB-AG10, when compared to DMNB-AG10, allowed a higher concentration of the former to be used and resulted in a more rapid depolarisation of ΔΨ_m. C: the photolysis reaction that is occurring to release AG10 (2).

Localised ΔΨ_m depolarisation by caged AG10 photorelease is not prevented by antioxidants

Exposure to intense light (e.g. the photolysis light) in the presence of mitochondrially-loaded fluorescent dyes may depolarise mitochondrial ΔΨ_mvia the production of free radicals or other oxidative species.^32–34 To determine whether or not oxidative species contribute to the ΔΨ_m depolarisation observed after photolytic release, AG10 was released in two separate regions of the same cell. In this experiment, the first photolysis event occurred at one end of the cell (Fig. 5Aiv) in the absence, and the second, at the other end of the same cell, in the presence of a cocktail of antioxidants [ascorbic acid (1 μM), catalase (250 units per mL), (R)-Trolox methyl ether (Trolox, 1 mM) and 2,2,6,6-tetramethylpiperidin-1-yloxy (TEMPO, 500 μM)] (Fig. 5Av). This mixture has been shown to inhibit light-evoked oxidative ΔΨ_m depolarisation.^32,33 Localised ΔΨ_m depolarisation by AG10, provided by localised photolytic release, was unchanged by the presence of the antioxidants (Fig. 5, n = 3). The antioxidants showed no effect on the cells when applied on their own.

Fig. 5 Mitochondrial depolarisation was not inhibited by antioxidants. Isolated single smooth muscle cells were loaded with TMRE (10 nM), voltage-clamped in the whole-cell configuration and DMNB-caged AG10 (25 μM) introduced to the cytosolvia the pipette solution. AG10 was focally released at two distinct sites within the cell (see photolysis sites marked on brightfield image of cell, Ai). First, AG10 was released at site 1 (red region drawn over TMRE fluorescence, Aii) and a localised ΔΨ_m depolarisation was observed in this region (Bi; compare Aiii – before with Aiv – after). The extracellular bathing solution was then supplemented with a mixture of antioxidants as described in the text. The photolysis site moved to region 2 (2nd site in Ai). AG10 was then released at site 2 and caused ΔΨ_m depolarisation here also (Bii, compare lower part of cell in Aiv with Av – after second AG10 release). No ΔΨ_m depolarisation was observed in two control regions—the upper and central regions of the cell (blue and green regions in Aii and correspondingly coloured traces in B). Indeed a transient increase in TMRE fluorescence was observed, probably due to dye redistribution from the AG10-depolarised regions. Raw TMRE fluorescence is shown in panel Aii whereas panels Aiii–v show TMRE fluorescence after applying a ‘top-hat’ filter.

Discussion

Here we have described the synthesis of two caged mitochondrial uncouplers (DMNB-caged AG10 and CDMNB-caged AG10) that when photolysed intracellularly in single smooth muscle cells evoke localised ΔΨ_m depolarisation. The synthesis of AG10 was achieved by condensation of 4-hydroxybenayldehyde (1) with malononitrile. The synthesis of DMNB-caged AG10 (5) directly from AG10 proved to be impossible (see ESI‡ for details). To overcome this problem, 4-hydroxybenzaldehyde was first reacted with DMNB bromide and the condensation reaction with malononitrile then performed. As 4-hydroxybenzaldehyde has a less extensive conjugated π-system than AG10, it was envisaged that this compound would be nucleophilic enough to react with DMNB bromide. This proved to be the case and the synthesis of the DMNB-caged aldehyde4 and subsequent condensation with malononitrile, to give the desired DMNB-caged AG10 derivative 5, proceeded smoothly.

The addition of a carbonate linker between the biologically active compound and the caging group can alter the wavelengths at which the compound absorbs,²⁸ allowing photolysis at a biologically less damaging wavelength, or an improved quantum yield.⁶ Therefore, it was decided to synthesise the DMNB-caged AG10 derivative with a carbonate linker 3. Initial studies treating AG10 with phosgene and subsequently 4,5-dimethoxy-2-nitrobenzyl alcohol did not yield the desired compound. Following the previous success of introducing the caging group onto 4-hydroxybenzaldehyde, this strategy was employed again. The carbonate-linked DMNB-caged 4-hydroxybenzaldehyde derivative was furnished in good yield. However, the carbonate linkage was susceptible to cleavage under the conditions employed for the condensation with malononitrile (see ESI‡ for details). Therefore, the preformation of the DMNB chloroformate was carried out and subsequent treatment of AG10 with this compound afforded the desired caged (3) compound in a moderate yield.

The UV/Vis spectra of AG10, DMNB-caged AG10 and CDMNB-caged AG10 were obtained (see ESI‡ for details). The UV/Vis spectrum of AG10 (0.01 M) in methanol showed two peaks, one with λ = 353 nm and a second with λ = 423 nm. We speculated that this second peak might be as a result of some AG10 being deprotonated in the methanol solution. To test this hypothesis, the UV/Vis spectrum of AG10 was taken in both acidic (0.01 M in MeOH, HCl added to attain pH 2) and basic conditions (0.01 M in methanol, DBU added to attain pH 12). It was observed that under acidic conditions the peak at 423 nm completely disappeared and only the peak at λ = 353 nm, with an extinction coefficient (ε) of 33 915, was present. Conversely, under basic conditions the peak at 353 nm completely disappeared and a peak with λ = 416 nm was present. These results seem to confirm our hypothesis that the peak at 423 nm is due to deprotonated AG10. DMNB-caged AG10 (0.01 M in DMSO) was observed to have λ_max of 351 nm and an extinction coefficient (ε) of 43 803. The CDMNB-caged AG10 (0.01 M in DMSO) had λ_max of 316 nm and an extinction coefficient (ε) of 12 057.

The λ_max values and extinction coefficients for DMNB-caged AG10 and CDMNB-caged AG10 explain the difference in photolysis characteristics of these compounds when irradiated at 355 nm. Not only is the λ_max of 3 shifted away from the irradiating wavelength of 355 nm, but ε is reduced, meaning that the irradiating light will be absorbed less efficiently resulting in less efficient photolysis.

Using single smooth muscle cells, it was first demonstrated that the synthetic AG10 (2) behaved as a mitochondrial uncoupler (Fig. 1). This experiment also demonstrated the lack of spatial discrimination obtained when AG10 is applied to the cells in a standard manner (i.e. by addition to the extracellular solution). Subsequently, it was shown that photolysis of DMNB-caged AG10 (5, Fig. 2) caused localised mitochondrial depolarisation. It was observed that this depolarisation was limited to the area of photolysis and it is likely that this localised effect of the photoreleased AG10 is due to the lipophilic nature of this compound. AG10 may be associating with the lipidmitochondrial membrane, hence limiting its diffusion. The lack of spatial discrimination observed when AG10 is applied to the extracellular bathing solution indicates that the compound diffuses within the solution and then enters the cell, causing global depolarisation. Exhaustive photolysis can lead to the depolarisation occurring slightly outside of the area of photolysis, indicating that the depolarisation is caused by the released AG10 and not a combination of caged AG10 and light (i.e. cellular damage). The sustained nature of the depolarisation may be attributed to the high local concentration of AG10 after photolysis. This sustained depolarisation is also observed when AG10 is applied to cells via the bathing solution. Control experiments showed that irradiation of single smooth muscle cells in the absence of caged AG10 or application of caged AG10 without photolysis has no effect on ΔΨ_m (Fig. 2B and C). In both the absence and presence of UV light, 4,5-dimethoxy-2-nitrobenzaldehyde had no effect on the smooth muscle cells. This result indicates that the by-product of the caging group is not producing a cellular effect within the timescale of our experiments and hence the effects observed are caused by photorelease of AG10 and its subsequent action, rather than cellular damage. Initial studies employing CDMNB-caged AG10 (3) demonstrated that, despite this compound being less effectively photolysed than DMNB-caged AG10 (5), it was more soluble in DMSO–water than 5 (at least 2.5-fold more soluble). It was therefore possible to apply a higher concentration of CDMNB-caged AG10 (62 μM versus 25 μM of DMNB-caged AG10) to cells and hence more rapid and effective mitochondrial uncoupling was observed (Fig. 4). Finally, we demonstrated that the effects on ΔΨ_m caused by photolysis of DMNB-caged AG10 were not inhibited in the presence of a cocktail of antioxidants. This result indicates that the depolarisation of the mitochondria was caused by AG10, and not non-specific damage due to the generation of other free radicals or oxidative species (Fig. 5). ΔΨ_m depolarisation caused by the release of AG10 is therefore distinct from light-evoked photodynamic oxidative damage that is used to induce cell death in cancerous cells and which perturbs multiple signalling pathways.^35,36AG10, on the other hand, is an ideal tool to locally depolarise ΔΨ_m without damaging other cellular processes.

Conclusions

In summary, we have reported the synthesis and characterisation of two novel caged AG10 derivatives. We have shown that the photolysis of these compounds leads to the release of AG10. Localised photorelease of AG10, furthermore, leads to mitochondrial uncoupling only in regions proximal to the photolysis site. Despite the less efficient photolysis of the CDMNB-caged AG10, this compound is more suited to use in biological systems, due to its improved solubility in DMSO–water solutions. The novel caged AG10 derivatives that we have synthesised have allowed, for the first time, mitochondrial uncoupling in a spatially predefined manner. As such, we expect that these compounds will be of use to those interested in the localised influence of mitochondrial function, such as over Ca²⁺ signalling processes. Localised influence of mitochondria on Ca²⁺ channels, particularly on the ER or SR, has been implicated to explain the affect that mitochondria have on the activity of these channels; however, this has not been shown directly. Caged AG10 will allow directed, localised ΔΨ_m depolarisation (hence removing the driving force for mitochondrial Ca²⁺ uptake) to clarify the role of localised mitochondrial activity on various Ca²⁺ channels.

Experimental

Synthetic chemistry

Full synthetic experimental procedures and characterisation data for the compounds reported can be found in the ESI‡ .

General laser photolysis procedure for ¹H NMR and HPLC studies

A solution of the caged AG10 derivative in DMSO (3 mM) was prepared. Care was taken to protect the solution from exposure to background UV irradiation. UV/Visabsorption spectra were obtained using a spectrophotometer (Lambda 950, Perkin Elmer) to provide an indication of which wavelengths were most likely to evoke photolysis (see ESI‡ for details).

During photolysis experiments, aliquots (1 mL) of the solution were placed into UV transmitting cuvettes (101-QS, Hellma) and exposed to radiation from the given laser source for fixed time durations. Light for photolysis was obtained using an optical parametric oscillator (Panther EX OPO, Continuum) that was pumped at 355 nm by a frequency tripled Nd:YAG laser (Surelite, Continuum). The pulses produced by the OPO had a duration of ∼4 ns at a repetition frequency of 10 Hz. The pulse energy varied depending on the wavelength in use; however, most photolyses employed a wavelength of 355 nm, which typically had a pulse energy of ∼100 mJ. Care was taken to ensure that the laser beam passed through the centre of the solution in the cuvette. After exposure to laser radiation the solution was analysed using ¹H NMR and/or HPLC (see ESI‡ for conditions).

Preparation of freshly dissociated cells

Single smooth muscle cells were enzymatically dissociated from the colon of male guinea pigs (∼600 g) as described previously.³⁷ Animals were humanely killed in accordance with the guidelines of the Animal (Scientific Procedures) Act UK 1986. All experiments and loading of cells with fluorescent dyes were carried out at 22 ± 1 °C.

Electrophysiology

Cells were patch-clamped using conventional tight seal whole-cell recording in order to introduce the caged compounds and maintain the plasma membrane potential (−70 mV) as previously described.^32,38 The extracellular solution contained (mM): Na glutamate 80, NaCl 40, tetraethylammonium chloride 20, MgCl₂ 1.1, CaCl₂ 3, Hepes 10, glucose 30 (pH 7.4 with NaOH) plus tetramethylrhodamine ethyl ester perchlorate (TMRE, 10 nM). The pipette solution contained (mM): Cs₂SO₄ 85, CsCl 20, MgCl₂ 1, Hepes 30, MgATP 3, sodium pyruvate 2.5, malic acid 2.5, NaH₂PO₄ 1, creatine phosphate 5, NaGTP 0.5 and the caged AG10 derivative (0.025 DMNB-caged AG10 or 0.062 CDMNB-caged AG10). The time from patching the cell to photolysis was typically 15 min. The high concentration of Hepes was to ensure adequate control of pH during ΔΨ_m depolarisation, pyruvate and malic acid to maintain mitochondrial activity and phosphocreatine and ATP to maintain [ATP] during experiments.

Imaging

Cells were loaded with the ΔΨ_m-sensitive dye TMRE (10 nM) and wortmannin (10 μM; to prevent contraction) for 30 min. TMRE is a lipophilic, cationic, fluorescent dye that accumulates within mitochondria according to their ΔΨ_m in a Nernstian fashion.^39,40 Above a certain concentration (e.g. ∼100 nM in the cells used in this study) TMRE self-quenches within mitochondria to reduce the fluorescence emanating from the indicator.⁴¹ In this study a sub-quenching concentration of TMRE (10 nM) was used to monitor ΔΨ_m of individual mitochondria. Cells were held in voltage-clamp to minimise changes in cytoplasmic [TMRE] due to changes in plasma membrane potential. Timelapse two-dimensional images of ΔΨ_m were obtained using a wide-field digital imaging system.^32,38 Single cells were excited at 560 nm (Polychrome IV monochromator, TILL Photonics, Martinsried, Germany) and light passed via a fibre optic guide through a 553 nm bandpass filter (bandpass 20 nm), a field stop diaphragm, a ND4 neutral density filter and reflected off a long-pass dichroic mirror (transmissive in the range 577–640 nm; Chroma, Rockingham, VT, USA) and then through an oil immersion objective (×40 UV 1.3 NA; Nikon UK, Surrey, UK). Emitted light was guided through a 594 nm barrier filter (bandpass 18 nm) to an intensified, cooled, frame transfer CCD camera (Pentamax Gen IV; Roper Scientific, Trenton, NJ, USA) controlled by MetaFluor software (Universal Imaging, Downingtown, PA, USA). Full frame images (150 × 150 pixels), with a pixel size of 563 nm at the cell, were acquired sequentially with an exposure period of 50 ms and acquisition rate of ∼5 Hz.

Localised flash photolysis

The output of a xenon flashlamp (Rapp Optoelecktronic, fired at 0.1 Hz), used to photolyse caged AG10, was focused and merged into the excitation light path via the fibre optic bundle and long-pass dichroic mirror at the lens part of the microscope’s epi-illumination attachment. The diameter of the fibre optic together with the lens magnification determined the area of photolysis (spot size ∼20 μm diameter).

Data analysis

Images were analysed using Metamorph 6.2 (Universal Imaging). TMRE fluorescence was measured over regions of interest drawn around individual mitochondria, subcellular regions or the whole-cell as described in the text. Each of the fluorescence signals from mitochondria and whole cells were normalised to starting baseline values (taken as 1) and the gradual decline in the signals due to dye photobleaching was corrected for by subtracting a linear extrapolation of bleaching from each region (thus making the baseline fluorescence values 0). Fluorescence signals were expressed as ratios (F/F₀ or ΔF/F₀) of fluorescence counts (F) relative to baseline values before stimulation (F₀). Images were further processed for visual display only by filtering with an 8 pixel diameter ‘top-hat’ open filter within Metamorph.

Drugs and chemicals

TMRE was purchased from Invitrogen. An initial sample of DMNB-caged AG10 was synthesised commercially by SiChem GmbH. All other drugs and reagents were purchased from Sigma.

Acknowledgements

JGM and SC were funded by Wellcome Trust (078054/Z/05/Z) and British Heart Foundation (PG/08/066), SJC, NA and MNS-H were funded by the BBSRC and the Leverhulme Trust. We thank the EPSRC National Mass Spectrometry Service (Swansea) for mass spectrometry data.

Notes and references

1. G. Csordas, C. Renken, P. Varnai, L. Walter, D. Weaver, K. F. Buttle, T. Balla, C. A. Mannella and G. Hajnoczky, J. Cell Biol., 2006, 174, 915–921.
2. J. D. Talbot, G. David and E. F. Barrett, J. Neurophysiol., 2003, 90, 491–502.
3. A. G. Garcia, A. M. Garcia-De-Diego, L. Gandia, R. Borges and J. Garcia-Sancho, Physiol. Rev., 2006, 86, 1093–1131.
4. D. G. Nicholls and M. Crompton, FEBS Lett., 1980, 111, 261–268.
5. G. Mayer and A. Heckel, Angew. Chem., Int. Ed., 2006, 45, 4900–4921.
6. C. G. Bochet, J. Chem. Soc., Perkin Trans. 1, 2002, 125–142.
7. P. Gorostiza and E. Isacoff, Mol. BioSyst., 2007, 3, 686–704.
8. J. W. Walker, J. Feeney and D. R. Trentham, Biochemistry, 1989, 28, 3272–3280.
9. S. J. Conway and G. J. Miller, Nat. Prod. Rep., 2007, 24, 687–707.
10. W. Maier, J. E. T. Corrie, G. Papageorgiou, B. Laube and C. Grewer, J. Neurosci. Methods, 2005, 142, 1–9.
11. G. Papageorgiou, D. Ogden and J. E. T. Corrie, J. Org. Chem., 2004, 69, 7228–7233.
12. M. Canepari, G. Papageorgiou, J. E. T. Corrie and D. Ogden, Eur. J. Neurosci., 2000, 12, 40.
13. G. Papageorgiou, D. C. Ogden, A. Barth and J. E. T. Corrie, J. Am. Chem. Soc., 1999, 121, 6503–6504.
14. G. Papageorgiou and J. E. T. Corrie, Tetrahedron, 2007, 63, 9668–9676.
15. G. C. R. Ellis-Davies, Nat. Methods, 2007, 4, 619–628.
16. G. C. R. Ellis-Davies, Biophotonics, Part A, 2003, 360, 226–238.
17. S. R. Adams, V. LevRam and R. Y. Tsien, Chem. Biol., 1997, 4, 867–878.
18. J. B. Shear, E. B. Brown, S. R. Adams, R. Y. Tsien and W. W. Webb, Biophys. J., 1996, 70, MP334.
19. J. H. Kaplan and G. C. R. Ellis-Davies, Proc. Natl. Acad. Sci. U. S. A., 1988, 85, 6571–6575.
20. J. H. Kaplan and G. Ellis-Davies, Biophys. J., 1988, 53, A36.
21. G. C. R. Ellis-Davies and J. H. Kaplan, J. Org. Chem., 1988, 53, 1966–1969.
22. C. C. Ashley, R. J. Barsotti, M. A. Ferenczi, T. J. Lea, I. P. Mulligan and R. Y. Tsien, J. Physiol., 1987, 390, P144.
23. S. J. Conway, Chem. Soc. Rev., 2008, 37, 1530–1545.
24. D. Gilbert, K. Funk, B. Dekowski, R. Lechler, S. Keller, F. Mohrlen, S. Frings and V. Hagen, ChemBioChem, 2007, 8, 89–97.
25. J. Zhao, T. D. Gover, S. Muralidharan, D. A. Auston, D. Weinreich and J. P. Y. Kao, Biochemistry, 2006, 45, 4915–4926.
26. J. L. Carr, K. N. Wease, M. P. Van Ryssen, S. Paterson, B. Agate, K. A. Gallagher, C. T. A. Brown, R. H. Scott and S. J. Conway, Bioorg. Med. Chem. Lett., 2006, 16, 208–212.
27. B. V. Zemelman, N. Nesnas, G. A. Lee and G. Miesenbock, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 1352–1357.
28. A. R. Katritzky, Y. J. Xu, A. V. Vakulenko, A. L. Wilcox and K. R. Bley, J. Org. Chem., 2003, 68, 9100–9104.
29. V. Hagen, B. Dekowski, N. Kotzur, R. Lechler, B. Wiesner, B. Briand and M. Beyermann, Chem.–Eur. J., 2008, 14, 1621–1627.
30. J. S. Modica-Napolitanoo, J. L. Joyal, G. Ara, A. R. Oseroff and J. R. Aprille, Cancer Res., 1990, 50, 7876–7881.
31. S. P. Soltoff, J. Biol. Chem., 2004, 279, 10910–10918.
32. S. Chalmers and J. G. McCarron, J. Cell Sci., 2008, 121, 75–85.
33. J. Jacobson and M. R. Duchen, J. Cell Sci., 2002, 115, 1175–1188.
34. J. Huser, C. E. Rechenmacher and L. A. Blatter, Biophys. J., 1998, 74, 2129–2137.
35. R. Hilf, J. Bioenerg. Biomembr., 2007, 39, 85–89.
36. N. L. Oleinick, R. L. Morris and T. Belichenko, Photochem. Photobiol. Sci., 2002, 1, 1–21.
37. J. G. McCarron and T. C. Muir, J. Physiol., 1999, 516, 149–161.
38. J. G. McCarron, D. MacMillan, K. N. Bradley, S. Chalmers and T. C. Muir, J. Biol. Chem., 2004, 279, 8417–8427.
39. M. R. Duchen, A. Leyssens and M. Crompton, J. Cell Biol., 1998, 142, 975–988.
40. R. C. Scaduto and L. W. Grotyohann, Biophys. J., 1999, 76, 469–477.
41. M. W. Ward, A. C. Rego, B. G. Frenguelli and D. G. Nicholls, J. Neurosci., 2000, 20, 7208–7219.

---

Footnotes

† Dedicated to Professor Andrew B. Holmes FRS, on the occasion of his 65th birthday.

‡ Electronic supplementary information (ESI) available: Additional synthetic schemes; representative ¹H NMR and HPLC analyses of photolysis; full experimental and characterisation data for the synthetic chemistry; UV/Vis spectra for the caged compounds and AG10. See DOI: 10.1039/b820415m

§ The abbreviations used are: DMF, dimethylformamide; HPLC, high pressure liquid chromatography; ΔΨ_m, mitochondrial membrane potential; NMR, nuclear magnetic resonance; TMRE, tetramethyl rhodamine ethyl ester perchlorate.

---