Main

Photochemical uncaging was developed by chemists for organic synthesis1,2,3 and later extended by biologists to living cells for the photorelease of bioactive molecules from biologically inert precursors4,5. The vast majority of caged compounds use either the ortho-nitrobenzyl2 or the 4,5-dimethoxy-2-nitrobenzyl3 (DMNB) chromophores (Table 1), in spite of their low photolytic efficiency6, as there has been no substantial, quantitative improvement in the photochemical properties of caging groups beyond these generically applicable chromophores. A practical consequence of this poor efficiency is that the brief bursts of light that are typically used for biological applications have a high photon flux, which can be demanding on cell viability, and bleach endogenous and exogenous chromophores. These drawbacks have limited the application of the uncaging technique for in vivo studies; the new caging chromophore we describe here was designed to address these shortcomings.

Table 1 Summary of the photochemical properties of widely used nitrobenzyl-caged compounds
Full size table

The extinction coefficient (ε) and the quantum yield (φ) are the two properties that define the photochemical efficiency of any chromophore. The former is an absolute value of a chromophore; the latter depends on the exact nature of the photolyzed bond and the chromophore itself. The four most widely used caged compounds (ATP5, IP3 (ref. 7), glutamate8, calcium9) with the ortho-nitrobenzyl chromophore have ε and φ in the range of 430–970 M−1 cm−1 and 0.14–0.63, respectively, whereas dimethoxynitrobenzyl (DMNB)-caged compounds (ATP10, cAMP11 and calcium12,13) have lower φ but higher ε. The product of these properties (ε·φ) for each caged compound defines the efficiency of its use of incident light. These data for the most widely used (and commercially available) caged compounds are summarized in Table 1.

We have synthesized a new, generically applicable caging chromophore, NDBF that is photolyzed 16–160 times more efficiently than other nitrobenzyl cages (Table 1). First application of this new chromophore was to caged Ca2+. We synthesized a photolabile, Ca2+-specific chelator (NDBF-EGTA) in 14 synthetic steps. Laser flash photolysis revealed that Ca2+ is released from the NDBF-EGTA:Ca2+ complex with rapid kinetics (rate of 20,000 s−1). NDBF had an ε of 18,400 M−1 cm-1 and a φ of photolysis of 0.7. Thus, NDBF-EGTA made efficient use of incident light. One and two-photon photolysis experiments in cardiac muscle cells confirmed the utility of the new cage. Because the NDBF protecting group uses well-established nitrobenzyl photochemistry, it should become the most effective generic caging chromophore to date.

Results

Synthesis of NDBF-EGTA

The synthesis of NDBF-EGTA (1) is outlined in Figure 1. The 3-nitro-2-ethyldibenzofuran (2) was brominated with N-bromosuccinimide, to give 3 in 82% yield. Conveniently, the oxidation of 3 to 4, and the bromination of 4 to give 5 were very clean reactions, with no column chromatography required. Ring closure of bromoketone 5 to epoxide 6 was affected by reduction of the ketone to the alcohol, which underwent in situ ring closure. We took this circuitous route to this key synthetic intermediate, as all the obvious, more direct routes, (for example, the regiospecific nitration of 2-acetodibenzofuanone) were unsuccessful. We regarded 2-(3-nitrodibenzofuran-2-yl)-oxirane (6) as the key synthetic intermediate in our synthesis of 1, because we had previously synthesized a photolabile EGTA from a similar epoxy intermediate14. Thus, conversion of 6 to 1 proceeded smoothly with an overall yield of 9% for the remaining six synthetic steps.

Figure 1: Synthesis of NDBF-EGTA.
figure 1

Reagents and conditions: (a) NBS, benzoylperoxide, CCl4, 4 h, reflux; (b) NaHCO3, DMSO, 2 h, 70 °C; (c) Br2, AcOH-CH2Cl2 (1:2), 8 h, ~21 °C; (d) (i) NaBH4, dioxane-MeOH (2:1), 30 min, ~21 °C; (ii) 2.5 N NaOH; (e) 2-(2-Chloroethoxy)ethanol, 2 h, 90 °C; (f) p-toluenesulphonyl anhydride, pyridine, DMAP, 30 min, ~21 °C; (g) NaN3, NaI, DMF, 3 h, 90 °C; (h) Ph3P, dioxane, 1 h, reflux; then NaOH (aqueous), 1.5 h, 90 °C; (i) ethyl bromoacetate, NaI, diisopropylethylamine, CH3CN, 12 h, reflux; (j) KOH, EtOH, 1 h, 60 °C.

Full size image

Absorption spectrum of NDBF-EGTA

The absorption spectrum of NDBF-EGTA is shown in Figure 2a. The extinction coefficient at the λmax of 330 nm is 18,400 M−1 cm−1. At 350 nm the extinction coefficient is 15,300 M−1 cm−1. NDBF-EGTA is highly soluble (200 mM) in aqueous buffer at neutral pH, and stable (2 years) in frozen solution at −80 °C.

Figure 2: Photophysical properties of NDBF-EGTA.
figure 2

(a) UV-visible absorption spectrum of a solution of NDBF-EGTA (0.10 mM) in aqueous buffer at pH 7.5 (40 mM Hepes, 100 mM KCl). (b) A single pulse from a frequency-doubled ruby laser (347 nm, 35 ns, 120 mJ) was used to photolyze a solution of NDBF-EGTA (0.2 mM) in the presence of 0.21 mM Ca2+ at pH 7.5 (40 mM Hepes) with 100 mM KCl at 20 °C. The fluorescence signal from 50 μM Ca-Green-5N was monitored: excitation 485 ± 10 nm, emission 570 ± 10 nm. The signal was fit to a single exponential expression (rate = 20,000 s−1), and the resultant values were used to calculate the solid curve overlaid on the trace.

Full size image

Ca2+ affinity of NDBF-EGTA

NDBF-EGTA has an apparent affinity for Ca2+ of 14 nM at pH 7.5 (Supplementary Fig. 1 online). This value decreases to about 100 nM at pH 7.2 (NDBF-EGTA has a Kd of 15 mM for Mg2+ at this pH; Supplementary Fig. 2 online) and increases to 5 nM at pH 7.8. We found that photolysis (19%) of a solution of NDBF-EGTA (0.5 mM) when [Ca2+]total of 0.55 mM was present changed the [Ca2+]free from 66 μM to 132 μM at pH 7.5. A simple model of the cation binding and photochemistry revealed that the photoproducts from NDBF-EGTA have an affinity for Ca2+of about 1 mM at pH 7.5. Therefore photolysis of NDBF-EGTA produces about a 140,000-fold change in the apparent affinity for Ca2+ at pH 7.5. This value is similar to NP-EGTA and DM-nitrophen, and is consistent with chelator backbone fragmentation.

Quantum yield of photolysis

We measured the φ of photolysis of NDBF-EGTA, by comparison with a known standard, 4-methoxy-7-nitroindolinyl (MNI)-glutamate15. The two compounds mixed so that their chromophores would absorb the same amount of light (MNI/NDBF ratio 3.7:1, total optical density 0.4). We determined the time course of photolysis of the two caged compounds by photolysis at 350 ± 5 nm. (We also photolyzed both cages separately to check that there was no interference from simultaneous photolysis.) High-performance liquid chromatography (HPLC) showed that NDBF-EGTA was photolyzed about 8 times faster than MNI-glutamate, implying a quantum yield of photolysis of 0.70 (± 0.03, s.d.; n = 8). Saturating [Ca2+] had no effect on this value.

Rate of Ca2+ release by the NDBF-EGTA:Ca2+ complex

We used laser flash photolysis techniques to measure the rate of Ca2+ release by the NDBF-EGTA:Ca2+ complex. Single pulses from a frequency-doubled ruby laser elicited rapid changes in [Ca2+] when monitored with a low affinity Ca2+ dye (Ca-Green-5N; Fig. 2b). The rate was 20,000 s−1 (± 210, s.d.; n = 11). We also used laser flash photolysis to detect transient absorbance signals from the expected aci-nitro intermediate that is part of the chelator fragmentation pathway (Supplementary Fig. 3 online).

Two-photon photolysis

We measured the two-photon cross-section of NDBF-EGTA by comparison with MNI-glutamate16. Two-photon photolysis of the mixture of caged compounds ([MNI-glutamate] = 1.4 mM, [NDBF-EGTA] = 0.1 mM) was accomplished by raster scanning the solution (10 μl, 1,024 pulses, 4 ms, 300 mW) in a cuvette on the stage of a two-photon microscope. HPLC showed that NDBF-EGTA was photolyzed 5 × faster than MNI-glutamate, indicating the new caging chromophore has a two-photon cross-section of about 0.60 GM (± 0.03, s.d.; n = 5). The presence of Ca2+ had no effect on this value.

We used confocal microscopy to measure the amount of Ca2+ uncaged by two-photon excitation of a solution of NDBF-EGTA, both in droplets and living cells. A pulse-train of increasing energy produced a series of rapid changes in [Ca2+] that depended on the square of the incident power17 (Fig. 3a). We have recently reported similar results for DM-nitrophen18. Direct comparison of the amount of Ca2+ released from the two cages by photolyzing solutions of the same [cage] and [Ca2+]resting (but different pH values, to ensure each chelator had the same initial Kd value of 5 nM; pH 7.2 for DM-nitrophen and pH 7.8 for NDBF-EGTA) revealed that NDBF-EGTA was sevenfold more effective at releasing Ca2+ than DM-nitrophen (Fig. 3b).

Figure 3: Two-photon photolysis of NDBF-EGTA in droplets and intact cardiac myocytes.
figure 3

(a) NDBF-EGTA:Ca2+ was photolyzed at 710 nm in a droplet (pH 7.5), using a mode-locked Titanium:sapphire laser. The shutter open-time was 50 ms per uncaging event. Laser-scanning confocal microscopy (in line-scan mode, 5 ms per line) was used to measure [Ca2+], with fluo-3 and standard data analysis methods6. The amplitude of the Ca2+ concentration signal versus the power incident in the image plane was fit with the equation. y = a + bxpow; the power-dependence had an exponent (pow) of 1.8. (b) Comparison of relative photolytic efficiencies of NDBF-EGTA (red circles) and DM-nitrophen (black squares) upon two-photon excitation at 710 nm. Each power-response plot includes ten different power levels, (in % of full power): 16, 22, 25, 31, 42, 53, 68, 84, 93 and 100, respectively. All traces were normalized to the highest two-photon power applied. The shutter open-time was 1 ms per uncaging event. (c) Line-scan images and averaged fluorescence traces recorded during diffraction-limited release of Ca2+ in single, intact, acutely isolated guinea pig cardiac myocytes. Changes in normalized fluorescence (F) are shown in pseudocolor as ΔF/F0. (d) Normalized amplitudes of Ca2+ signals (ΔF/F0) in cells recorded in the presence of caffeine to estimate the pure photolytic component (black dots), control after SR refilling (red squares) and after treatment with isoproterenol (blue triangles). All solutions had [cage] = 2 mM and [fluo-3] = 50 μM.

Full size image

We also tested two-photon photolysis of NDBF-EGTA in living cells. We loaded single guinea pig cardiac myocytes with NDBF-EGTA and fluo-3 via a patch pipette (we found that NDBF-EGTA and its photoproducts are biologically inert). We removed Ca2+ from the sarcoplasmic reticulum (SR) by brief treatment of the cell with caffeine, followed by a power train of two-photon laser pulses to estimate the photolytic component of the Ca2+ signals (Fig. 3c,d). After reloading the SR with four depolarizing pulses, we recorded another power relationship to reveal the amount of Ca2+-induced Ca2+ release (Fig. 3d). Treatment of the cell with isoproterenol enhanced the Ca2+-induced Ca2+ release (Fig. 3d).

Cardiac muscle contraction

We confirmed the efficiency of Ca2+ release from the NDBF-EGTA:Ca2+ complex by comparison with NP-EGTA. We have already used the latter for a detailed study of rapid activation kinetics of cardiac muscle fibers by UV-laser photolysis19. Therefore, we compared the tension transients in single guinea pig cardiac muscle fibers elicited by laser photolysis of both caged calciums in the same fiber (Fig. 4; NDBF-EGTA and NP-EGTA have the same affinity for Ca2+ at pH 7.5). Photolysis of NDBF-EGTA:Ca2+ complex produced almost maximal force at pH 7.5, using 70 mJ of energy from a frequency-doubled ruby laser (Fig. 4). The same energy and Ca2+ loading produced only few percent of maximal tension when NP-EGTA was used in a subsequent trail (Fig. 4). A second successive uncaging event produced a stronger contraction but more than two times the energy was used for this tension transient compared to that elicited by NDBF-EGTA photolysis.

Figure 4: Tension recordings of successive activations of a single trabecula by UV laser photolysis of caged Ca2+.
figure 4

Photolysis of NDBF-EGTA:Ca2+ complex (1 mM) produced almost maximal force using 70 mJ of energy (black line). The same laser energy and Ca2+-loading of 1 mM NP-EGTA produced only few percent of maximal tension under similar conditions (red line). A second, successive uncaging event produced a stronger contraction (blue line) but required more than twice the energy, 150 mJ.

Full size image

Discussion

The NDBF-EGTA chromophore described here has an extinction coefficient of 18,400 M−1 cm−1, about 19–43-fold and 3.3–4.3-fold larger than the popular ortho-nitrobenzyl and DMNB chromophores, respectively (Table 1). NDBF-EGTA is uncaged with a high quantum yield of 0.7, about the same as NPE-caged ATP and IP3, whereas most other widely used caged compounds have much lower quantum yields6,20,21 (Table 1). Thus, for UV photolysis NDBF-EGTA makes the better use (ε · φ) of excitation light intensity than any other nitrobenzyl caged compound.

Recently several groups have developed coumarin-caged compounds; all of which use the original photosolvolysis method1 to release the caged substrate (reviewed in ref. 20). Thus, the chemistry used by these cages is not generically applicable, as photosolvolysis only works for acidic function groups, in contrast to NDBF-based photochemistry. Specifically, coumarin photochemistry cannot be used to break ether14 or tertiary amino9,13 bonds like the nitroaromatic photochemistry we used (see Supplementary Fig. 4 online). It should be noted that even the best of these new cages is photolytically inferior to NDBF-EGTA (Supplementary Table 1 online). Notably, all the coumarin-caged compounds are also highly fluorescent at wavelengths that are used for imaging eGFP or YFP, so can severely interfere with the use of this technology. In contrast, like other nitroaromatic compounds9,22, NDBF-EGTA is not fluorescent.

We confirmed the photochemical superiority of NDBF-EGTA by direct comparison with NP-EGTA, using a bioassay of photolytic efficiency. We used both cages to initiate contraction of the same skinned cardiac muscle fiber under identical conditions. Photolysis of NDBF-EGTA produced almost full contraction, whereas NP-EGTA produced only a few percent of maximal tension (Fig. 4). These data show dramatically the comparative efficiency of photolysis of the old and new caging chromophores in a biological system that requires a large increase in [Ca2+]. NDBF-EGTA must still be used at the about same concentration as NP-EGTA, even though it is photochemically more efficient, as the chelator must be the dominant Ca2+ buffer before photolysis. But the large ε · φ of NDBF will allow substantially lower uncaging energies to be used in biological experiments to deliver quantitatively the same cellular stimulus, potentially greatly extending experimental scope resulting from prolonged cell viability.

One photosensitive Ca2+ chelator, azid-1 (refs. 23,24) is photolytically superior to NDBF-EGTA (Table 2). Its 'boutique' photochemistry involves rearrangement of an azidofura-like chromophore, whereas NDBF-EGTA uses generically applicable nitrobenzyl photochemistry. NDBF-EGTA has several other important advantages over azid-1: it has high affinity before photolysis and its photoproducts have a very low affinity for Ca2+ (therefore the ΔKd is 38–270-fold larger than that of azid-1). Furthermore, NDBF-EGTA:Ca2+ releases Ca2+ at least an order of magnitude faster than azid-1 photochemistry23, a property vital for diffraction-limited two-photon photolysis. So whereas azid-1 is photolytically superior to NDBF-EGTA (that is, higher ε · φ), the latter is chemically superior at actually releasing Ca2+ (that is, NDBF-EGTA has ΔKd of 140,000-fold, compared to 520-fold for azid-1).

Table 2 Summary of the properties of calcium cages
Full size table

The photochemical release of Ca2+ was consistent with two-photon excitation17 of NDBF-EGTA (Fig. 3a). We found that the two-photon cross-section of NDBF-EGTA was substantially larger than other nitrobenzyl caged compounds (0.6 GM versus 0.01 GM; Fig. 3b, and Tables 1 and 2). Thus, two-photon photolytic release of Ca2+ was also very efficient in acutely isolated cardiac myocytes; uncaging on the surface of the SR induced local calcium signals25 by Ca2+-induced Ca2+ release at pH 7.2 (Fig. 3c,d). Such conditions are somewhat unfavorable to Ca2+ release from NDBF-EGTA, as the chelator's affinity for Ca2+ is quite low at this pH (the Kd is 100 nM, and so it is only 50% loaded with Ca2+). Nevertheless, the robust Ca2+-induced Ca2+ release we observed indicates that NDBF-EGTA is sufficiently sensitive to two-photon excitation such that more than 50% of the total amount of cage can be rapidly photolyzed by within a diffraction-limited volume. These data suggest that the greatly superior two-photon cross-section of the NDBF chromophore means that the new cage should prove to be especially useful for in vivo two-photon uncaging studies, as it is in this situation that two-photon techniques show their real strength26,27.

The change in affinity for Ca2+ of NDBF-EGTA upon photolysis is very large (at pH 7.5 it is 140,000-fold). This property is critical for permitting chemically efficient release of calcium ions. For a 'normal' caged compound, like NPE-ATP5, the effector molecule is caged by covalent modification of substrate itself. So before photolysis, all of the ATP is caged. Upon irradiation, for every bond broken, a molecule of ATP is liberated and becomes available for receptor activation. This is not the case with caged Ca2+, as covalent bonds with free Ca2+ are not possible. Instead, Ca2+ is caged by chelation (that is, by an equilibrium reaction of ionic species). To cage Ca2+ most effectively we use high affinity chelators to bind as much Ca2+ as possible, but because the Kd is always finite, there must always be some free Ca2+, and free (unloaded) cage in solution. Therefore, to produce net release of the caged compound (that is, bound Ca2+), at least all of the unloaded cage must be photolyzed21. As the percentage [cage]free is set by its prephotolysis affinity, and the relative [cage]total and [Ca2+]total, the higher the affinity, the less net photolysis is required for effective Ca2+ release. To achieve chemically efficient release of Ca2+, our strategy has always been to design caged Ca2+ molecules with high prephotolysis affinities, which fragment in two upon photolysis21. NDBF-EGTA, DM-nitrophen13 and NP-EGTA9 all reflect this approach, and are much more chemically efficient than nitr-5 (ref. 12) or azid-1 (ref. 23) at releasing Ca2+ (Table 2).

The rate of Ca2+ release from the NDBF-EGTA:Ca2+ complex is fast. We used a low affinity Ca2+ dye to monitor the appearance of Ca2+. A single pulse of near-UV light produced a monoexponential signal with a rate of 20,000 s−1. Such fast Ca2+ release from caged Ca2+ is vital for many biological applications of this technology, as many Ca2+-activated processes are known to be rapid. Furthermore, if caged compounds are to be used for diffraction-limited, two-photon photolysis, the rate of uncaging must be substantially greater than about 3,300 s−1 (that is, half-time 0.3 ms), because this is the 'residence time' of excited molecules in the focal volume of the laser beam28. The Ca2+ release kinetics of NDBF-EGTA more than satisfies this criterion.

In summary, NDBF-EGTA processes a unique constellation of highly desirable physico-chemical properties that make it near ideal for use as a caged Ca2+ for both one and two-photon photolysis. Furthermore, because the NDBF cage uses the well-established, generically applicable nitrobenzyl photochemistry, this new chromophore can also be used to cage any functional group, promising ultra efficient release of a wide variety of molecules, both in cultured cells and in vivo, by either UV or two-photon photolysis.

Methods

Synthesis of NDBF-EGTA.

Full synthetic details are available in the Supplementary Methods online.

Physico-chemical characterization of NDBF-EGTA.

We measured the quantum yield, Ca2+ affinity, Mg2+ affinity and Ca2+ release rate in the same way as for NP-EGTA9,29 (Supplementary Methods online). The two-photon cross-section was measured in a manner similar the quantum yield, except we used a mode-locked titanium:sapphire laser (10W Verdi-Mira, Coherent) as the light source. We focused the 720-nm light through a low–numerical aperture (NA) lens (NA 4.0, 10 × magnification), via the scan-head of a two-photon microscope (Ultima, Prairie Technologies), into cuvette of low volume. Using multiple uncaging events (1,024), sufficient photolysis of a volume large enough for accurate analysis could be achieved (5–20% uncaging).

Two-photon photolysis in cardiac myocytes.

The experiments on isolated cells were carried out according to the guidelines of Swiss Animal Protection Law and with the permission of The State Veterinary Office, Bern, Switzerland. We isolated cardiac ventricular myocytes from adult male guinea pigs using established enzymatic methods18,15. The extracellular superfusion solution contained 140 mM NaCl, 5 mM KCl, 1.8 mM CaCl2, 1 mM CsCl, 0.5 mM BaCl2, 10 mM Hepes, 10 mM glucose (pH 7.4; adjusted with NaOH). Cells were continuously superfused using a custom-made rapid superfusion system (t1/2 < 500 ms). Where indicated, we added isoproterenol from a frozen stock (1 mM in 10% ascorbic acid). Applications of 20 mM caffeine were used to estimate the amount of photolysis from the cage. The pipette-filling solution contained 120 mM cesium-aspartate, 20 mM TEA-Cl, 2 mM NDBF-EGTA, 6 mM MgCl2, 1.5 mM CaCl2, 5 mM K2-ATP, 10 mM Hepes and 0.05 mM K5-fluo-3 (pH 7.2; adjusted with CsOH). We carried out all experiments were carried out at ∼21 °C. We performed patch clamp, laser scanning confocal microscopy, two-photon photolysis and data analysis as previously described18.

Cardiac muscle contraction.

We performed caged Ca2+ experiments on skinned cardiac muscle fibers as described previously19. The photolysis solutions contained 1 mM cage:Ca2+ complex, 100 mM TES, 37 mM HDTA, 10 mM creatine phosphate, 10 mM glutathione, 6.8 mM MgCl2, 5.5 mM ATP and 1 mg/ml creatine kinase, at an ionic strength of 200 mM. The pH of the photolysis solutions were set to a value of 7.5 for NDBF-EGTA and NP-EGTA, as both cages have a similar Ca2+ affinity at this pH, thus permitting us to determine the relative photolytic efficiency of the two Ca2+ cages from the evoked tension transients.

Additional methods.

Procedures for laser flash photolysis and time-resolved fluorescence experiments as well as determination of divalent cation affinities and quantum yields are described in Supplementary Methods.

Note: Supplementary information is available on the Nature Methods website.