Nmoc-DBHQ, a new caged molecule for modulating sarcoplasmic/endoplasmic reticulum Ca2+ ATPase activity with light flashes.

We report the synthesis and characterization of O[o-nitromandelyloxycarbonyl]-2,5-di(tert-butyl)hydroquinone (Nmoc-DBHQ), a new “caged” reagent for photoreleasing DBHQ, a membrane-permeant, reversible inhibitor of the sarcoplasmic/endoplasmic reticulum Ca2+ ATPase (SERCA). The Nmoc group is a new caging group developed for the current application. Photolysis of Nmoc-DBHQ proceeds with t1/2 = 126 ± 2 μs, and t1/2 for subsequent release of DBHQ is estimated to be ∼5 ms. Nmoc-DBHQ thus allows rapid and reversible modulation of SERCA activity in living cells. Through its acetoxymethyl ester, Nmoc-DBHQ can be loaded into cells easily by incubation. We demonstrate the use of Nmoc-DBHQ for photomodulating SERCA activity in fibroblasts and vagal sensory neurons. We further demonstrate the utility of pulsed DBHQ photorelease for probing and manipulating dynamic phenomena such as [Ca2+] oscillations in fibroblasts.

Intracellular concentration and distribution of the ubiquitous second messenger Ca 2ϩ is tightly controlled by a number of pathways (1). The interaction of the pathways that mobilize and regulate free Ca 2ϩ levels can result in highly complex and dynamic signaling patterns, such as Ca 2ϩ oscillations and waves (1)(2)(3). Pulsed perturbation of the concentrations of various second messengers, achieved by flash photolysis of caged inositol-1,4,5-trisphosphate, diacylglycerol, and Ca 2ϩ (4 -8), has yielded highly specific mechanistic information about these dynamic phenomena (5,9). Although the role of second messengers themselves in dynamic signaling phenomena has been studied by photorelease techniques, the contribution of pathways that regulate second messenger levels remains unexplored.
The family of sarcoplasmic/endoplasmic reticulum Ca 2ϩ ATPases (SERCA) 1 that sequester Ca 2ϩ into the sarcoplasmic reticulum and endoplasmic reticulum are important regulators of cytosolic free Ca 2ϩ levels (10). We reasoned that the effect of these pumps on Ca 2ϩ oscillations and waves could be eluci-dated by the development of a method for the pulsed modulation of their activity. This could be accomplished by the preparation of a caged, reversible SERCA inhibitor. Of the known SERCA inhibitors (11), 2,5-di(tert-butyl)-1,4-hydroquinone (DBHQ, 1) (12,13) is ideally suited for the development of a caged SERCA modulator because it is a structurally simple and reversible inhibitor that is commercially available in large quantities.

EXPERIMENTAL PROCEDURES
Synthesis-Reagents and solvents were ACS or high pressure liquid chromatography grade and were used as received from Aldrich or Fisher. Dimethylformamide and dichloromethane were stored over 3 Å molecular sieves. All oxygen-and water-sensitive reactions were performed under dry argon atmosphere. For water-sensitive reactions, glassware was dried at 130°C for at least 3 h and cooled under a stream of argon or in a desiccator prior to use. Silica gel 60 (230 -400 mesh, Merck) was used for column chromatography. Melting points were recorded on a Melt-temp II (Laboratory Devices) apparatus coupled to an Omega (Omega Engineering) HH23 digital thermometer and are uncorrected. The structures of all purified products were established by NMR spectral analysis. Spectra were recorded on a General Electric QE-300 (300 MHz) NMR spectrometer. Samples were dissolved in CDCl 3 (0.03% tetramethylsilane) unless otherwise stated and were referenced to tetramethylsilane. Samples in solvents other than CDCl 3 were referenced to the residual solvent peak. Resonances are reported in the following format: NMR (solvent): chemical shift in ppm downfield from tetramethylsilane, multiplicity (s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; b, broad), spin-spin coupling constant if appropriate, and integrated number of protons. High resolution mass spectrometric analysis was performed at the University of Maryland (College Park, MD) on a model VG707E spectrometer (VG Analytical).
2-Hydroxy-2- [2-nitrophenyl]acetonitrile (4)-Sodium bisulfite (8.26 g, 79 mmol) was added to a suspension of 2-nitobenzaldehyde (10 g, 66 mmol) in water (60 ml). After the reaction mixture had stirred for 10 min, it was cooled in ice water. A solution of potassium cyanide (5.16 g, 79 mmol; dissolved in 30 ml of water) was added dropwise. The mixture was stirred for 30 min, warmed to room temperature, and filtered. The solid was washed with water and air dried to give 10 g (85%) of 4. The material could be recrystallized from methanol/water. mp. 89 -91°C. 1  2-Hydroxy-2- [2-nitrophenyl]acetic Acid (5)-Cyanohydrin 4 (9.5 g, 48 mmol) was refluxed in concentrated hydrochloric acid (100 ml) for 2.5 h. The solution was cooled to room temperature and extracted with ethyl acetate. The extract was dried (MgSO 4 ) and evaporated, leaving 12.98 g of crude acid, which had acetic acid as an impurity. This was taken on to the methyl ester without further purification.
Flash Photolysis Kinetics-A 660 M solution of the sodium salt of Nmoc-DBHQ was prepared in 150 mM NaCl buffered at pH 7.2 with 10 mM phosphate. The stirred solution was photolyzed with 308 nm, 100-mJ, 10-ns pulsed emission from a XeCl excimer laser (Questek 2110) while the absorbance of the solution at 440 nm was measured. By monitoring the appearance and decay of the absorbance due to the aci-nitro intermediate generated by photolysis, the kinetics of uncaging could be examined (16,17). In some runs, the experimental solution was continuously purged with nitrogen gas, although purging produced no observable difference in the kinetic behavior of the system.
Cell Culture-Fisher rat embryo fibroblasts of the cell line REF52 were cultured on 25-mm diameter glass coverslips as described previously (18). Acutely dissociated neurons from the nodose ganglia of ferret were prepared and plated onto 25-mm diameter glass coverslips coated with poly-D-lysine as described previously (19).
Loading Cells with the Caged Reagent and the Fluorescent Indicator Fluo-3-Cells were coloaded with indicator and caged reagent by incubation for 60 -80 min at room temperature with medium containing 2-5 M fluo-3/AM, 20 -40 M Nmoc-DBHQ/AM, and less than 0.015% (w/v) of the surfactant Pluronic F-127 to enhance the aqueous solubility of the AM esters. Fluo-3/AM, Nmoc-DBHQ/AM, and Pluronic F-127 were kept frozen as, respectively, 5 mM, 40 mM, and 15% (w/v) stock solutions in dry dimethyl sulfoxide until use. For loading, Dulbecco's modified Eagle's medium buffered with HEPES at pH 7.4 was used for REF52 cells and Leibovitz L-15 medium containing 10% (v/v) fetal bovine serum and buffered at pH 7.4 was used for nodose neurons. After loading, coverslips bearing cells were washed with fresh medium and mounted in culture dishes charged with 4 ml of experimental medium. Experiments with REF52 cells were conducted in Hanks' balanced salt solution, whereas HEPES-buffered Locke solution was used for nodose neurons.
Microfluorometry and Photorelease-Cells were positioned on the stage of an inverted microscope (Diaphot, Nikon) equipped with a dichroic mirror that transmits at wavelengths longer than 500 nm but has high reflectance in the UV and 450 -500 nm range so that photolysis and epifluorescence measurements could be performed simultaneously. A spectrofluorometer (CM1T10I, SPEX Industries) fitted with a 450-W xenon arc lamp was coupled to the microscope epifluorescence port through a fiberoptic cable and was operated in the microfluorometry mode. Fluo-3 was excited by 500 nm light (4 nm bandwidth) from a monochromator, whereas its emission, collected through the microscope objective (UV-F, ϫ40, N.A. 1.3, Nikon), after passing through a 530 nm bandpass filter, was sampled by a photomultiplier tube, the output from which was digitized and stored for subsequent analysis. Instrument control, data acquisition and analysis were performed through DM3000CM software (SPEX Industries) running on a dedicated personal computer.
Output from a 50-W mercury arc lamp (HBO50; Osram), filtered through 2-mm UG-1 glass to isolate UV light in the 300 -400 nm range, was used for flash photolysis conducted in parallel with microfluorometry. A 400 nm long pass dichroic mirror was placed so that it was colinearly aligned with the epifluorescence port and the dichroic mirror underneath the microscope objective and served to guide the photolysis beam into the optical path of the microscope. The photolytic light traversed a path of approximately 38 cm, as measured from the mercury arc to the sample above the objective. An electromechanical shutter (Vincent Associates, Rochester, NY) was used to control UV light exposure. During photolysis, fluorescence data collection was interrupted by a second shutter to prevent damage to the photomultiplier.
Biochemical Reagents-Cell culture reagents and media were obtained from Life Technologies, Inc. Fluo-3/AM and Pluronic F-127 were from Molecular Probes (Eugene, OR). R 8 -vasopressin and gramicidin D were from Sigma.

RESULTS AND DISCUSSION
The three most commonly used SERCA inhibitors are thapsigargin (20,21), cyclopiazonic acid (22), and DBHQ (12,13). Because its inhibitory action is irreversible (23), thapsigargin is not a suitable target for a caged reagent to be used for reversible photomodulation of SERCA activity. Cyclopiazonic acid has a relatively complex molecular structure and, being a biosynthetic product of fungal origin, is available only in small quantities at high expense, which makes it an unattractive starting material for organic synthesis. In contrast, DBHQ is structurally simple, incorporating only one type of reactive functional group for caging purposes, and is commercially available in large quantities. These advantages, together with its reversibility, made DBHQ our preferred target for caging.
The great majority of photoreleasable compounds have used caging groups structurally based on the 2-nitrobenzyl system (24). Although the simple parent 2-nitrobenzyl moiety is a common caging group, it was not appropriate for caging DBHQ; preliminary experiments indicated that UV irradiation of cells bathed in medium containing 2-nitrobenzyl alcohol resulted in irreversible inhibition of the SERCA pump. Because photolysis of any 2-nitrobenzyl-caged compound is expected to generate the same photochemical byproducts, we inferred that the byproducts of photolyzing a 2-nitrobenzyl-caged DBHQ would not be inert.
Hess and co-workers have shown that the ␣-carboxy-nitrobenzyl group is useful for caging neuroactive amino acids (25)(26)(27)(28)(29). Photodeprotection was shown to proceed rapidly and with high quantum yield (28). We recognized that the carboxylate on this caging group would reduce the reactivity of the photochemical byproduct. Furthermore, the presence of the carboxylate offers the added advantages of increasing the water solubility of the caged compound and allowing for the preparation of a caged AM ester, which could be passively loaded into cells.
Caging DBHQ directly with the ␣-carboxy-nitrobenzyl group requires the formation of a benzyl ether, which model studies indicated was problematic. For example, reaction of DBHQ with 2-nitrobenzyl chloride in the presence of K 2 CO 3 yielded numerous compounds that were difficult to isolate and characterize. Reasoning that the difficulties encountered in benzyl ether formation were at least partially the result of the sterically congested environment surrounding the phenolic hydroxyl groups of DBHQ, we postulated that an efficacious caging reaction would need to proceed through a different mechanism. The Nmoc group was designed as a photocleavable caging group that would combine the desirable qualities of the ␣-carboxy-nitrobenzyl group with a caging reaction that proceeds via carbonyl substitution. Irradiation of Nmoc-DBHQ (2) with UV light would result in the formation of DBHQ-bicarbonate (3), which would rapidly decompose under physiological conditions to DBHQ (1) and carbon dioxide as shown in Scheme 1. The photochemical side product, 2-(2-nitrosophenyl)glyoxylate (3a) is the same as that generated by photolysis of ␣-carboxy-nitrobenzyl-caged molecules (25)(26)(27)(28)(29), for which no adverse biological effects have ever been reported. Carbon dioxide, liberated by decarboxylation of DBHQ-bicarbonate (3), is a normal product of metabolism and would thus also be innocuous.
Nmoc-DBHQ was prepared as outlined in Scheme 2. 2-Nitrobenzaldehyde was converted to the cyanohydrin (4), which was hydrolyzed to o-nitromandelic acid (5) by refluxing in concentrated hydrochloric acid. Methyl o-nitromandelate (6) was prepared by Fisher esterification of 5. Caging DBHQ necessitated the preparation of an activated o-nitromandelyl-oxycarbonyl group. The oxycarbonyl imidazole derivative (7) was considered superior to the chloroformate because of its stability and ease of preparation. Treatment of alcohol 6 with carbonyl diimidazole gave 7 in high yield as a stable crystalline solid. The desired carbonate (8) was formed when a solution of 7 and DBHQ in N,N-dimethylformamide was heated with a catalytic amount of 4-dimethylaminopyridine. Saponification of the methyl ester (8) followed by acidification gave Nmoc-DBHQ (2). To facilitate loading of the caged reagent into cells, the AM ester of Nmoc-DBHQ (9) was prepared. Neutralization of 2 with one equivalent of sodium hydroxide, followed by esterifi-cation with bromomethyl acetate in the presence of tetrabutylammonium iodide gave Nmoc-DBHQ/AM (9).
To demonstrate the photoreactivity of Nmoc-DBHQ, a series The spectra shown were recorded after the stirred sample had been exposed for 0, 10, 20, 40, 80, 160, 320, 640, and 1290 s to 365 nm light at an intensity of 5.7 ϫ 10 Ϫ8 einstein⅐cm Ϫ2 ⅐s Ϫ1 , as determined by ferrioxalate actinometry. B, exponential time course of photolysis. Graph of ⌬A 290 , the change in absorbance of the sample at 290 nm (taken from the spectra shown in A), as a function of time of photolysis by light of known intensity, I 0 . The data are well fit by a single exponential with characteristic time constant () of 311 Ϯ 17 s. The quantum yield of photolysis of Nmoc-DBHQ was thus determined to be Q ϭ 2.303I 0 ⑀ 365 ϭ 0.10 (6,14). SCHEME 1. Photolysis of Nmoc-DBHQ.

SCHEME 2. Synthesis of Nmoc-DBHQ and Its AM Ester.
of UV-visible absorption spectra were acquired from a solution of the sodium salt of Nmoc-DBHQ that was being photolyzed with 365 nm light (Fig. 1A). The spectroscopic changes resulting from photolysis are consistent with those expected from the classic o-nitrobenzyl rearrangement. In particular, photolysis causes an increase in absorbance at longer wavelengths, a result of the long wavelength absorption by the highly conjugated byproduct (3a in Scheme 1). Because the spectra show good isosbestic points through the course of photolysis, we were able to determine the quantum efficiency of photolysis of Nmoc-DBHQ by analyzing the absorbance changes as a function of time (6,14). Spectroscopic changes during photolysis followed an exponential time course, as shown in Fig. 1B. The quantum efficiency of photolysis, Q, was thus determined to be 0.10 (6,14).
The kinetics for photolytic removal of the caging group were examined by monitoring the transient absorbance changes characteristic of the short-lived aci-nitro intermediate generated during photolysis of o-nitrobenzyl moieties (30 -33). The time course for the decay of the transient absorbance following laser pulse photolysis of Nmoc-DBHQ is shown in Fig. 2. The decay is dominated by an exponential component with a lifetime () of 182 Ϯ 1 s. 2 These results imply that the photochemical cleavage reaction is essentially complete in approximately 550 s (i.e. in about three lifetimes).
Photocleavage of the nitromandelyl moiety leaves the carbonate monoester of DBHQ (3 in Scheme 1), which must lose CO 2 to liberate DBHQ. Although the rate of decarboxylation, k dec , cannot be measured in the present system, it can be estimated from published studies of similar reactions. The decarboxylation rate (k dec ) of carbonate monoesters has been shown to follow the relationship: log(k dec ) ϭ 15.1 Ϫ 1.16pK a , where pK a is the pK a of the hydroxyl group that is esterified to the carbonate (34,35). Because the pK a of DBHQ is estimated to be 11.21 (36), the rate constant for decomposition of the carbonate monoester of DBHQ is expected to be k dec Ϸ 130 s Ϫ1 (t1 ⁄2 Ϸ 5.3 ms) at 25°C. We demonstrate in subsequent biological experiments that DBHQ photorelease is sufficiently rapid for probing Ca 2ϩ signaling dynamics.
Biological efficacy of Nmoc-DBHQ was tested in living cells. Loading cells simultaneously with Nmoc-DBHQ and the fluorescent Ca 2ϩ indicator, fluo-3 allowed the effects of photoreleasing DBHQ on intracellular Ca 2ϩ dynamics to be monitored through the Ca 2ϩ -sensitive fluorescence of fluo-3. Results of DBHQ photorelease experiments, performed in rat embryo fibroblasts (REF52 cell line) as well as in acutely isolated nodose neurons of adult ferret, are shown in Fig. 3. As expected, the data in Fig. 3 show that photorelease of DBHQ within cells lead to rapid, dose-dependent transient increases in [Ca 2ϩ ] i . The resting cytosolic [Ca 2ϩ ] is maintained through a dynamic balance of active pumping processes that remove Ca 2ϩ from the cytosol and passive leaks that introduce Ca 2ϩ into the cytosol. When photoreleased DBHQ disrupts the pump-leak balance by inhibiting the SERCA pumps, an increase in [Ca 2ϩ ] i is observed. The [Ca 2ϩ ] i rise is transient because SERCA inhibition by DBHQ is reversible and because free DBHQ is a small, uncharged, membrane-permeant molecule that is cleared from the cell by diffusion. Once DBHQ diffuses out of the cell and is Arrowheads and associated numbers indicate, respectively, the times and durations of photolysis light flashes. Corresponding dips in the experimental traces are due to temporary interruption of fluorescence data acquisition by an electromechanical shutter to avoid exposure of the fluorescence photomultiplier tube to leakage from the high intensity photolytic flash. Each plus signs (ϩ) on an experimental trace marks the first intensity datum collected after a photorelease episode. Note that the trace does not return to the pre-photolysis baseline because the byproduct nitrosoketone absorbs more strongly than the starting material at wavelength longer than ϳ360 nm (compare preand post-photolysis spectra in Fig. 1A).
lost to the bathing medium, inhibition of the SERCA pumps ceases, and the resting [Ca 2ϩ ] i is re-established.
Comparing the results from rat fibroblast and ferret nodose neuron (Fig. 3, A and B, respectively) shows that the kinetics of onset of the response to DBHQ photorelease are similar for the two cell types (rise times of 7.8 Ϯ 3.4 and 6.2 Ϯ 2.4 s, respectively). In contrast, decay of the response to photorelease appears more protracted in the nodose neuron. Whereas t1 ⁄2 for recovery from photorelease is of the order of 10 -15 s in the fibroblast, in the nodose neuron the value increases to 40 -50 s. The different kinetics of recovery likely reflect differences in the surface-to-volume ratios of the two cell types. Whereas the fibroblast is an adherent cell in monolayer culture with flat, extended morphology, and is at most a few microns in thickness, the acutely dissociated nodose neuron is spherical, with a diameter of ϳ55 m (19). Clearance of photoreleased DBHQ from the cell interior by diffusion is thus expected to be slower for the nodose neuron than for the fibroblast.
For demonstrating its usefulness in manipulating dynamic cellular phenomena, Nmoc-DBHQ was used to perturb [Ca 2ϩ ] i oscillations. Highly regular calcium oscillations were initiated in REF52 fibroblasts by activation of the inositol trisphosphate signaling pathway coupled with depolarization (37), as shown in Fig. 4A. The effect of transient modulation of SERCA activity by photoreleasing DBHQ is shown in Fig. 4. It can be seen from Fig. 4B that pulsed inhibition of the SERCA pump produces distinct changes in the timing of [Ca 2ϩ ] oscillations. 3 Specifically, transient SERCA inhibition by DBHQ photorelease causes a delay in the occurrence of the next endogenous oscillation peak. Alternatively stated, DBHQ photorelease during an oscillation cycle increases the period of that cycle. The extent of period lengthening is in proportion to the amount of photorelease, with short light flashes producing only slight delays, whereas long flashes can cause significant period lengthening (Fig. 4B). Such quantitative information should be useful for constructing, as well as validating, biophysical models of intracellular Ca 2ϩ dynamics (see for example Refs. 5 and 9).
In summary, we have synthesized and characterized Nmoc-DBHQ, a caged inhibitor of the SERCA pump. Nmoc-DBHQ allows rapid and reversible modulation of the SERCA activity in living cells and should be useful for probing systems whose rapidly varying Ca 2ϩ dynamics make study inaccessible through conventional techniques of reagent delivery. 3 The perturbation effects are the result of photorelease and are not observed when light flashes are delivered in the absence of caged reagent, as has been demonstrated previously (5,9,37). ] i were elicited in a REF52 fibroblast and then perturbed by pulses of DBHQ photorelease. The oscillations were initiated by bath application of 500 nM R 8 -vasopressin, followed by 500 nM gramicidin D. B, expanded plot of a portion of the trace in A to show more clearly the effect of DBHQ photorelease on the timing of the oscillations. Arrowheads and associated numbers indicate, respectively, the times and durations of photolytic light flashes. Cell loading with Nmoc-DBHQ and fluo-3 was as described under "Experimental Methods." Experiment was conducted in Hanks' balanced salt solution at room temperature.