Caged compounds: photorelease technology for control of cellular chemistry and physiology

Abstract

Caged compounds are light-sensitive probes that functionally encapsulate biomolecules in an inactive form. Irradiation liberates the trapped molecule, permitting targeted perturbation of a biological process. Uncaging technology and fluorescence microscopy are 'optically orthogonal': the former allows control, and the latter, observation of cellular function. Used in conjunction with other technologies (for example, patch clamp and/or genetics), the light beam becomes a uniquely powerful tool to stimulate a selected biological target in space or time. Here I describe important examples of widely used caged compounds, their design features and synthesis, as well as practical details of how to use them with living cells.

The idea behind the caging technique is that a molecule of interest can be rendered biologically inert (or caged) by chemical modification with a photoremovable protecting group (Fig. 1). Illumination results in a concentration jump of the biologically active molecule that can bind to its cellular receptor, switching on (or off) the targeted process. Virtually every kind of signaling molecule or second messenger, of every size|[mdash]|from protons to proteins|[mdash]|has been caged¹.

Figure 1. General strategies caged by either multistep or direct caging.

The caging chromophore prevents receptor binding until it is cleaved by light. Second messengers and hormones can be caged by both strategies, but the illustration shows only two examples for simplicity.

Why are caged compounds so useful? A single component of cellular chemistry can control the function of a cell, and such cellular regulation can be temporally or spatially defined, intracellular or extracellular, and amplitude- or frequency-modulated². Photomanipulation of cellular chemistry using caged compounds provides a uniquely powerful means to interact with such cellular dynamics, as it can touch upon any one of the above dimensions. Thus, since light passes through cell membranes, uncaging can rapidly release a biomolecule in an intracellular compartment. This space is not readily accessible to many second messengers (for example, inositol-1,4,5-trisphosphate (IP₃), ATP, Ca²⁺, cAMP, cGMP) when they are applied to cells externally, as their charge makes them impermeable to the plasma membrane. Furthermore, uniform illumination results in release throughout the cytosol, or the release can be localized by focusing the uncaging beam on one part of a cell. Likewise, extracellular uncaging of neurotransmitters and hormones is tunable, allowing stimulation of many neurons simultaneously or of single synapses|[mdash]|by global or focused illumination, respectively. Light cannot only be directed, but also modulated in time and amplitude. Thus, uncaging can also be used to produce rapid, repetitive release of biomolecules or finely graded changes in the magnitude of stimulation.

Examples of important caged biomolecules or second messengers are calcium³, ⁴, ⁵, ⁶, neurotransmitters⁷, ⁸, ⁹, ¹⁰, inositols¹¹, ¹², nucleotides¹³, ¹⁴, peptides¹⁵, ¹⁶, enzymes¹⁷, ¹⁸, ¹⁹, mRNA²⁰ and DNA²¹. Apart from Ca²⁺, all these mole-cules are caged by covalent modification of one part of their structure with a photoremovable chromophore. Although chemists create most new caged compounds, many biologists have embraced the use of these powerful tools to answer biological questions. For example, calcium uncaging with molecules like NP-EGTA (Fig. 2) has been widely used to study many Ca²⁺-controlled processes. In particular, secretory processes in neuronal and non-neuronal cells have been extensively studied with caged calcium²². Rapid uncaging of glutamate using two-photon photolysis is particularly useful for the rational stimulation of visually identified synapses in complex tissue preparations such as acutely isolated brain slices from the hippocampus²³ (Fig. 2b). Finally, uncaging of mRNA in vivo in zebrafish is an excellent example of a biological application of uncaging in whole animals²⁰ (Fig. 2c). Before any of these experiments can be performed, however, it is necessary to cage the molecule of interest. The basic methods for the construction of caged compounds are discussed below. Although some methods require complex chemistry, others are simple enough for biologists to do themselves.

Figure 2. Structures and photochemistry of caged compounds.

(a) NPE-ATP¹³ cannot be hydrolyzed before uncaging as covalent attachment of the caging chromophore to the | [gamma] |-phosphate prevents enzymatic access. Photolysis breaks the bond between the benzylic carbon of the chromophore and an oxygen atom, liberating the caged ATP. NPE-ATP has been widely used to control molecular motors³¹. (b) MNI-Glu does not bind to postsynaptic glutamate-gated ion channels because of modification of the side-chain carboxylate with the caging chromophore¹⁰. Uncaging restores the | [gamma]|-carboxylate by donation of an oxygen atom from the nitro group. (c) The translational activity of Bhc-mRNA is latent because of chemical modification of multiple phosphates on the backbone (for simplicity only a single cage is shown). Irradiation initiates a photosolvolysis reaction that releases the caged mRNA²⁰. (d) NP-EGTA cages Ca²⁺ by high affinity binding5, photolysis breaks the Ca²⁺ coordination sphere in two, yielding low-affinity fragments that release the bound Ca²⁺.

Making caged compounds

Caged compounds are made using synthetic organic chemistry. Syntheses are usually multistep, but some caged compounds are made with one-step 'direct' caging (Fig. 1). Multistep syntheses are usually required because most natural products have many functional groups of equivalent reactivity.

As organic chemists have made a huge number of caged compounds, a full description of all of them is beyond the scope of this review (for chemistry reviews, see refs. ¹,²⁴). Furthermore, chemical applications of photorelease technology are not covered in this review (for example, Affymetrix production of genechips²⁵, ²⁶). In this section I outline syntheses of the caged compounds that have been widely used by biologists as well as new, conceptually important methods of synthesizing caged macromolecules.

General design guidelines

Caged compounds must be biologically inert before photolysis. This means the probe should be neither an agonist nor antagonist when applied at a useful concentration to the biological preparation. The rate of uncaging of the caged biomolecules needs to be faster than the process being studied, if kinetics is an essential part of the process being studied. The absolute speed requirement will vary depending on the application: neurotransmission is much faster than gene transcription, but both can be regulated by calcium. Finally, the higher the efficiency of uncaging, the easier it is to use a caged compound, but caged compounds with modest uncaging efficiencies (for example, 1-(ortho-nitrophenyl)-ethyl (NPE)-caged ATP and IP₃) have been used to great effect in many biological experiments. Lower uncaging efficiencies can, however, be deleterious for cell health in the case of UV light|[ndash]|mediated uncaging, or in the case of two-photon excitation, the rate of uncaging could be too slow to be useful (see Box 1 and Fig. 3 for a description of important optical characteristics of compounds and their effect on uncaging efficiency).

Box 1. Definition of Photochemical Properties of Caged Compounds.

The absorption of light

When we think of light as an electromagnetic wave, the key idea for light absorption by a chromophore is that the oscillating electric field of light interacts with an electron as if it were an oscillating dipole. Effective interaction requires a ‘resonance’ between the light wave and a molecular electronic energy gap (that is, the ground state and an excited state, hence absorption is strictly quantized). Electrons oscillate at about 10¹⁵ Hz, corresponding to wavelengths of about 200–700 nm, hence this is called the ‘photochemical range’ of light. If we think of light as a stream of particles, instead of an electromagnetic wave, then a photon becomes a ‘reagent’, and absorption is a bimolecular reaction that occurs by collision of particles. The propensity for this reaction is quantified in the extinction coefficient (ε) of a molecule (units: M^–1 cm^–1). The larger the ε, the more likely it is that a photon will be absorbed. Two laws relate ε to actual measurement of absorption: (i) the proportion of light absorbed is independent of initial intensity, I₀ (Lambert's law), and (ii) transmitted light intensity, I_t, is proportional to concentration (c) of absorbing molecules (Beer's law). Hence the optical density of a solution is log(I₀ / I_t), and ε= log(I₀ / I_t) × c × l, where l is the optical pathlength.

The use of absorbed light

The quantum yield of reaction (=) is the measure of how many excited state molecules lead to uncaged products (units: moles per Einstein, but normally given as a unitless number).

Two-photon excitation

Normally an excited state produced by the absorption of a single photonic quantum, as described above, occurs at a rate of about 10¹⁵ s^–1. If two photons of half the energy of such a quantum collide with a chromophore within 10^–18 s, then ‘virtually simultaneous’ absorption of two photons occurs that varies quadratically with the incident flux density. The probability of two-photon excitation (n) is dependent on several variables⁹⁷:

$$n\propto \frac{\delta {〈\mathrm{P}〉}^2}{{\tau }_{\mathrm{p}}{f}_{\mathrm{p}}^2}{\left(\frac{{\text{NA}}^2}{2\text{hc}\lambda }\right)}^2$$

(where δ in the two-photon cross section, P the incident power, τ the pulse-width, ƒ the repetition rate of the laser light, λ, the wavelength of excitation light, NA the numerical aperture of the microscope objective, and h, c are constants). One consequence of the nonlinear nature of two-photon is that excitation is dramatically localized the z direction. Thus, when used for uncaging, two-photon excitation may allow focal bursts of release with a volume of much less than 1 fL. This strict three-dimensional confinement is dependent upon rapid uncaging, as the residence time of an excited molecule with the two-photon volume is ∼0.3 ms (ref. ⁹⁸). Thus, two-photon excitation of a slowly released caged glutamate64 does not allow highly localized released but creates a ‘mist’ of uncaged glutamate (Fig. 4). But if glutamate is released quickly, uncaging can be close to the diffraction limit¹⁰,⁸⁸.

Figure 3. The degree of spatial confinement of two-photon uncaging depends on the rate of substrate release.

Caging small molecules and ions

Caged neurotransmitters

Caged glutamate is the most widely used caged neurotransmitter by biologists, and many syntheses of caged glutamate have been published using different chromophores and/or different caging strategies²⁷. Very few of these caged glutamates actually satisfy all the requirements of neuroscientists for photorelease in complex biological tissue such as acutely isolated brain slices (that is, rapid uncaging, photochemically efficient release, hydrolytic stability, biological inertness and, ideally, two-photon excitation sensitivity).

The |[alpha]|-carboxy-ortho-nitrobenzyl (CNB) protecting group was the first truly successful caging chromophore for neurotransmitters. The four-step synthesis of CNB-Glu⁹ has proved especially important for neurobiology. Glutamate is released quickly (half-time, 25 |[mu]|s) and efficiently to activate |[alpha]|-amino-5-hydroxy-3-methyl-4-isoxazole propionic acid (AMPA) receptors with their normal time course (that is, a rise of about 0.5 ms). CNB-GABA has also been synthesized²⁸, but it has been reported to be mildly antagonistic²⁹. Unfortunately, CNB-Glu is not sensitive to two-photon uncaging, so 4-methoxy-7-nitroindolinyl (MNI)-Glu has been developed to address this need¹⁰. Both caged compounds require careful purification after the final deprotection step, which is identical in both syntheses, because the presence of even the slightest amount of free glutamate can be toxic to acutely isolated brain slices. The synthesis of CNB- and MNI-caged amino acids involves several steps requiring silica gel column chromatography to remove organic side products. This makes the preparation of these compounds much more challenging for biologists to do themselves, but CNB-Glu and MNI-Glu are commercially available (Table 1).

Table 1. Properties of selected commercially available caged compounds.

Caged compound	Φ	ε(M–1 cm–1)	Φ × ε	Rate (s–1)	Stability
Calcium chelators
DM-nitrophena,b	0.18	4,300	774	3.8 × 104	Complete
NP-EGTAa	0.23	970	194	6.8 × 104	Complete
nitr-5b	0.012	5,500	66	2.5 × 103	Complete
diazo-2a	0.03	22,800	1,596	2.3 × 103	Complete
Neurotransmitters
CNB-Glua	0.14	500	70	4.8 × 104	Fair
CNB-GABAa	0.16	500	70	3.6 × 104	Fair
CNB-	0.8	430	344	1.7 × 104	Excellent
carbamoylcholinea
MNI-Gluc	0.085	4,300	366	∼105	Excellent
Phosphates
NPE-IP3a,b	0.65	430	280	225 and 280	Excellent
NPE-cAMPb	0.51	430	219	200	Fair
DMNPE-cAMPa	0.05	5,000	250	300	Poor
NPE-cADPribosea	0.11	430	271	18	Excellent
NPE-ATP-a,b	0.63	430	271	90	Excellent
DMNPE-ATP a	0.07	5,000	350	18	Fair
Fluorophores
bis-CMNB-fluoresceina	ND	2,000	ND	ND	Complete
DMNB-HPTS a	ND	∼5,000	ND	ND	Complete

^aFrom Invitrogen (Molecular Probes).

^bFrom Calbiochem.

^cFrom Tocris. ε, extinction coefficient; Φ, quantum yield. ND, not determined.

Caged nucleotides, nucleosides and inositols

Caged ATP¹³ and cAMP³⁰ were the first caged compounds to be synthesized and uncaged in living cells. Caged ATP was made by coupling NPE-caged phosphate to ADP and required three synthetic steps, whereas caged cAMP was caged directly using hyper-reactive diazo chemistry that allows selective caging of phosphates. The diazo approach has proved very useful for caging phosphates in a wide variety of molecules. ATP and IP₃ have both been caged in one step using diazo chemistry¹², ¹⁴. There have also been many studies published using NPE-caged ATP and IP₃ by biologists who make these caged compounds³¹, ³², ³³. There are two advantages to biologists in making their own caged ATP or IP₃: a large quantity can be made a reasonable price, and purity can be guaranteed.

Although NPE-caged compounds have proven extremely useful for biologists, they are not perfect caged compounds. Improvements in one or more of their properties (solubility, stability, rate of uncaging and two-photon cross-section) have been reported, however. For example, NPE-cAMP is not very soluble at pH 7, so a highly soluble version of this compound has been synthesized in nine synthetic steps³⁴, which has the added benefit of also being more hydrolytically stable. The rate of uncaging of NPE-ATP is a little slow for some studies (rate is 83 s^|[minus]|1), so many faster-uncaging caged ATP probes have been synthesized. Two interesting biological studies have been published showing that the rate of photolysis of NPE-ATP can be rate-limiting³⁵, ³⁶. NPE-IP₃ has a low two-photon cross-section (|[sim]| 0.001 GM), so a caged IP₃ with larger cross-section (0.035 GM) has been synthesized in seven synthetic steps, permitting localized two-photon uncaging of IP₃ in living cells³⁷. Finally, membrane-permeant ester derivatives of caged cAMP and IP₃ probes have been made using multistep syntheses that permit loading of these second messengers into intact cells³⁸, ³⁹, ⁴⁰.

Caged calcium

The inorganic cation calcium cannot form covalent bonds to caging groups the way organic molecules can, so a new caging strategy has been developed. Photolabile derivatives of known high-affinity calcium chelators (BAPTA, EDTA and EGTA) have been synthesized. These molecules decrease their affinity for calcium upon irradiation, thus uncaging some of the bound calcium (Fig. 1d). Ideally the Ca²⁺ cage should bind as much Ca²⁺ as possible, but as the dissociation constant is always finite, there must always be some free Ca²⁺ and free (unloaded) cage. This situation is quite different from that of other caged molecules. Thus, in the case of caged Ca²⁺, two additional properties define efficiency of caging, the affinity for Ca²⁺ before and after photolysis (Table 2). The higher the affinity of the chelator, the more it can be loaded with Ca²⁺ before [Ca²⁺]_free reaches an activating level, and the lower to photoproduct affinity, the more Ca²⁺ is released by photolysis (Table 2). To produce net release of bound Ca²⁺, all of the unloaded cage must be photolyzed, else it acts as a 'calcium sink' that will re-complex the photoreleased calcium (reviewed in ref. ⁴¹).

Table 2. Properties of calcium cages.

	Ca2+ Kd (nMa)	Product Kd (mM)	ΔKd	Φ	ε (M–1 cm–1)	Φ × ε	Rate calcium release (s–1)	2PCS (GM)
NDBF-EGTA	100	2.0	20,000	0.7	18,400	12,880	20,000	0.6
DMNPE-4	48	2.0	42,000	0.09	5,120	461	45,000	0.01
NP-EGTA	80	1.0	12,500	0.23	975	224	68,000	0.001
DM-nitrophen	5	3.0	60,0000	0.18	4,300	774	38,000	0.01
nitr-5	145	0.0063	54	0.012	5,500	66	2,500b	NDc
azid-1	230	0.12	520	1.0	33,000	33,000	500	1

^aMeasured at pH 7.2.

^bGiven is the rate of appearance of organic photo product; Ca²* is probably released at the same rate.

^cThis value has not been measured, but is probably 0.01 by analogy with DM-nitrophen. ε, extinction coefficient; Φ, quantum yield. 2PCS, two-photon cross-section. ND, not determined.

The general strategy of calcium uncaging has been applied in two different ways. One based on photochemical modification of the buffering capacity of BAPTA derivatives (nitr-5, (ref. ³) or azid-1 (ref. ⁴²)), and the other based on photochemical scission of the backbone of either EDTA (dimethoxy (DM)-nitrophen4) or EGTA (nitrophenyl (NP)-EGTA⁵). The compounds 1-(4,5-dimethoxy-2-nitrophenyl)-EGTA (DMNPE)-4 (ref. ⁶), NDBF-EGTA⁴³. DM-nitrophen, NP-EGTA and nitr-5 are all commercially available, and therefore have been used in hundreds of biological experiments⁴¹, ⁴⁴, ⁴⁵. The properties of all these calcium cages are summarized in Table 2.

Caging macromolecules

A variety of caged peptides and enzymes have been made. The former are usually small, inhibitory peptides that can be used to disrupt an important intracellular protein-protein interaction. The latter are macromolecular counterparts of standard caged compounds, in which the catalytic function of the enzyme is blocked by the caging chromophore. Such syntheses are often more challenging than those discussed above, as peptides seem to be inherently unstable. But the development and use of caged peptides is important because they provide another means of controlling the fate of a cell that is conceptually distinct from the uncaging of small molecules, especially given that many cellular processes are not regulated by essential cofactors but simply by protein-protein interactions. An alternative to such direct protein caging is to cage the translational or transcriptional machinery that produces such gene products. Both strategies have been extensively reviewed⁴⁶, so I outline only the basic concepts involved in caging macromolecules below.

Caging peptides and proteins

Proteins and peptides have been caged with commercially available reagents that covalently modify specific amino acid residues. G-actin was the first protein to be caged using this ‘shotgun’ approach17. Proteins such as PKA and cofilin have also been caged using a similar strategy¹⁹, ⁴⁷, ⁴⁸. There are several specific points to consider when planning the synthesis of a caged enzyme: (i) use of caged proteins is complicated by issues of residual activity of the caged protein (it is very difficult to have 100% caging, so a few percent residual activity may give rise to ambiguous results), (ii) recovery of activity after uncaging is problematic, especially when ‘shotgun’ caging is used as multiple sites must be uncaged for activation, and (iii) caged proteins must be microinjected into cells⁴⁶.

Two approaches to the synthesis of caged proteins using modern molecular techniques have been developed. The first uses unnatural amino acid mutagenesis. T4 lysozyme was caged as proof-of-principle of this idea¹⁸. Caged ion channels and phosphoproteins have been synthesized using this method, and such syntheses require the combination of a dazzling array of technical skills (reviewed in ref. 49). The second approach uses expressed protein ligation (reviewed in ref. 50) to assemble a caged semi-synthetic protein. For example, Smad2 is regulated by dephosphorylation of two C-terminal serine residues, and so a caged version of the protein was synthesized by solid phase synthesis of the caged phosphoserine portion of the enzyme, which was then coupled to recombinant Smad2-MH2 domain to yield a caged protein⁵¹.

The basic strategy for synthesizing caged peptides was delineated with the development of caged MLK_II inhibitors. Using solid-phase peptide synthesis, a caged tyrosine was incorporated into a small inhibitory peptide¹⁵. Other peptides have been caged using the same approach¹⁶, ⁵², ⁵³, ⁵⁴.

Peptides and enzymes are normally caged by some site-specific crucial modification with a standard small caging chromophore such as that first used with ATP. But it is often hard to know the position of a single crucial residue, or it is difficult to construct such uniquely caged compounds. An alternative means of disrupting protein-protein interactions is with shear bulk. An amyloidogeic human prion protein fragment was caged by adding a polycationic peptide to the N terminus via a photocleavable cross-linker⁵⁵. This constitutes a conceptually novel addition to the arsenal of caging strategies that could be adapted for many other macromolecules for which precise structure-function relationships are not known or perhaps for which important sites are hard to modify selectively with small blocking groups. A similar strategy has been applied, using cross-linked antibodies⁵⁶.

Caged mRNA and DNA

Turning on one gene in space-time is especially attractive for delineation of protein function in whole organisms, and this is where uncaging technology has found important applications. Genetic function can be controlled by caging large mRNA or DNA fragments.

The simplest way to cage mRNA has also proven to be the most useful biologically. Direct, multi-site caging of mRNA²⁰, ⁵⁷, ⁵⁸, ⁵⁹ or DNA²¹ with a reactive diazo coumarin chromophore, similar to the original caged cAMP synthesis³⁰, (Bhc; Fig. 2c) efficiently inactivates the molecules.

The techniques for caging essentially any biomolecule or second messenger have now been developed, so that protons⁶⁰, inorganic cations, gases⁶¹, small organic molecules and macromolecules⁴⁶, ⁶² can be caged. But given the complexity of the syntheses of most caged compounds, commercial availability constrains their use for most laboratories. As caged peptides and proteins are quite unstable, companies cannot keep large amounts in stock. This means that most biologists are restricted to the few caged compounds that are commercially available as a result of high demand (Table 1) or they must collaborate with academic laboratories that synthesize caged compounds.

Using caged compounds

This section aims to give biologists an idea how to think about the practicalities of using caged compounds and provides some biological examples illustrating the basic principles of caged compounds.

Photochemical guidelines

Chemical synthesis of a biologically inert caged compound is only the first step in developing a useful caged probe. There are many other chemical and photochemical properties that must be understood for the cages to be designed and used effectively (Box 1).

Chemical considerations

Organic synthesis is typically performed in nonaqueous solvents, but all caged compounds are used with living cells. So the final step of the synthesis usually renders the caged compound water-soluble. This essential property is often much more difficult to achieve than it sounds, for two reasons: (i) adding a hydrophobic caging chromophore to a moderately soluble natural product often makes it sparingly soluble at physiological pH, and (ii) for practical purposes, one often requires a high concentration of the caged compound as a stock solution. Some substrates are so highly soluble (for example, EGTA, IP₃) that adding the caging chromophore to these molecules makes little difference to their solubility, whereas others can be profoundly affected by caging (for example, serotonin, cAMP, GABA, DAG, sphingosine-1-phosphate). This problem is often ignored in the literature, but can be solved by adding additional (usually negative) charge to the caging chromophore, which unfortunately complicates the synthesis considerably³⁴.

Aqueous stability is also a vital property of caged compounds. Some chemical bonds are sensitive to aqueous hydrolysis (most importantly esters and to a lesser extent amides), whereas others are not (ethers, amines and carbamates). Poor aqueous stability results in the spontaneous hydrolytic release of the caged compound and has bedeviled the development of a good caged glutamate (ref. ²⁷ and Table 1). Caging the amine either directly or via the carbamate is the obvious solution to this crucial problem⁶³, but these caged glutamates are uncaged too slowly to be useful²⁷. Caging glutamate as an amide has provided a recent solution to this important dilemma¹⁰, ⁶⁴.

The rate of uncaging is an important property of most caged compounds. The actual rate of uncaging has been reported for only a few caged bioactive molecules. The fundamental problem with such rate determinations is that there are few indicators for biomolecules that respond with useful temporal speed. Thus whilst indicators for glutamate, IP₃, cAMP, cGMP and ATP do exist, only the kinetics of the ATP indicator have been characterized and therefore used to quantify the rate of uncaging¹⁴. Good indicators for cations (for example, Ca²⁺) exist, and have been used to characterize some caged calcium compounds. Chemists attempted to solve this difficult situation by studying the photochemistry of the caging chromophores. But this approach has turned out to be much more complicated than expected (reviewed in ref. ²⁴). The primary photochemical reaction of ortho-nitrobenzyl, the most widely used and only generically applicable type of caging chromophore, is wellstudied and involves a photochromic aci-nitro intermediate⁶⁵. It was thought that since the appearance of ATP from NPE-caged ATP was concomitant with the decay of its aci-nitro intermediate¹⁴, that the decay of this species could be used as an indicator of the release of other caged compounds (for example, IP₃ (ref. ¹²) and glutamate⁹). Rigorous study of the kinetics of model ortho-nitrobenzyl (ortho-NB)|[ndash]|caged compounds (caged ethers) has however shown this is not always the case⁶⁶. The situation is further complicated by a report that 4,5-dimethoxy-2-nitrobenzyl (DMNB)-caged ethers uncage much faster than simple ortho-NB|[ndash]|caged ethers⁶⁷. It seems for ortho-NB|[ndash]|caged phosphates and carboxylates, aci-nitro decay is rate-limiting, whereas for ortho-NB|[ndash]|caged ethers it is not, and for DMNB-caged ethers the situation is unclear. These data suggest that for nitroaromatic caged compounds the rate of appearance of each caged effector from photolysis of caged compounds must be determined.

The reaction mechanisms of two other caging chromophores have been thoroughly studied: coumarin-caged phosphates and para-hydrophenacetyl|[ndash]|caged acids (reviewed in ref. ²⁴). Both uncage via photosolvolysis⁶⁸, which is a rapid process (rates >10⁶ s^|[minus]|1), but it only works for certain functional groups, so unfortunately such chemistry is of limited applicability. Additionally, there are no commercially available molecules caged with these chromophores.

The efficiency of use of the incident uncaging light, the product of the extinction coefficient and quantum yield of reaction, expressed as |[epsi]| |[times]| |[phgr]|, is an important property of caged compounds and may constrain feasible applications of any caged compound. The size of the |[epsi]| can also constrain application of a caged compound. For practical purposes, optical densities of the cage in a cell should be less than 20%, else inhomogeneous uncaging will result from inner filtering effects across the cell during rapid photolysis. For example, if the concentration of the cage is 2 mM, then the extinction coefficient at 350 nm of the cage must be less than 10 mM^|[minus]|1 cm^|[minus]|1 for a 100 | [mu]|m pathlength. In chromaffin cells of about 20 |[mu]|m diameter, the concentration of DM-nitrophen often used is 10 mM (refs. ^{69, 70, 71, 72}). This solution has an optical density of about 0.09. If the same concentration of azid-1 (ref. ⁴²) was used, then the optical density would rise to 0.66. Rapid illumination in such a situation would produce a graded photolysis of the cage and an uneven release of Ca²⁺ across the cell because of the inner filtering effect by azid-1 as a result of its strong absorption. Thus, for most onephoton uncaging experiments, moderate extinction coefficients (500|[ndash]|5,000 M^|[minus]|1 cm^|[minus]|1) can be advantageous. But smaller cells or experiments that require low concentrations of substrate to be uncaged or have a slow time frame can tolerate caging chromophores with large extinction coefficients²⁰, ⁵⁹, ⁷³.

The compatibility of caging chromophores with other fluorescent probes is another consideration when designing an experiment. Nitroaromatic chromophores are not fluorescent⁵, and can be used in conjunction with a wide range of probes such as fura-2, fluo-4 and all GFPs²². There are potential problems of using fluorescent caging chromophores (7-diethylaminocoumarinyl-4-methyl (DEAC), 6-bromo-7-hydroxycoumarin-4-methyl (Bhc) and others) as these have high fluorescent quantum yields and emit light in the same region as many analytical fluorescent probes.

Light sources for uncaging

The majority of uncaging experiments use flash lamps (for example, pulsed xenon or mercury arc lamps focused with a parabolic mirror) or lasers as the light source⁷⁴, and there are advantages and disadvantages to both. Flash lamps are robust, reasonably cheap and can efficiently uncage compounds (up to 80%) in a 1-ms pulse⁶⁹. Their relatively long pulse-width permits multiple rounds of excitation of the same molecule, as the excited state lifetime of aromatic molecules is only a few nanoseconds; hence the chemical yield from flash-lamp excitation normally exceeds the quantum yield. Being pulsed, flash lamps also conveniently obviate the need of a shutter.

The near-UV lasers that are used for uncaging are much more expensive to purchase and maintain than flash lamps. Lasers can be pulsed or continuous wave light sources. Many confocal microscopes are supplied with a continuous wave Ar-Kr laser (output 354|[ndash]| 363 nm), which can be used for uncaging in conjunction with a shutter. Pulsed lasers have been used in many experiments with caged compounds. The ideal pulsed laser (frequencydoubled ruby, pulse-width 35 ns at 347 nm) is unfortunately no longer available commercially. The widely available frequency-tripled Nd-YAG laser has a very short pulsewidth (3 ns) that limits the chemical yield to the quantum yield for a single pulse and the slow repetition rate does not permit closely spaced (less than 1 ms) multiple pulses.

Compared to flash lamps or near-UV lasers, two-photon excitation offers some potential intrinsic advantages in terms of depth of uncaging in tissues and axial confinement. Such performance is dependent, however, on the characteristics of the caged compound (Box 1 and Fig. 3). For two-photon uncaging, solid-state mode locked Ti:sapphire lasers are used. These lasers are in effect pseudo-continuous, so must also be shuttered when used for photorelease experiments.

Shutters for light sources can be mechanical or optical. Optical shutters are most often supplied with commercial confocal or two-photon microscopes and laser launches. These shutters are either acousto-optical tunable filters (AOTF) or electro-optical modulators (EOM). If very fast, precise control of uncaging is required (submillisecond timing), then AOTF or EOM must be used. But if periods of >1 ms are adequate, then mechanical shutters work well.

Use with cells

Caged compounds need to be applied to cells in a controlled manner when hormones and transmitters are being used for extracellular uncaging or loaded into the cytosol if second messengers are being released. For quantitative understanding of the effects of uncaging, the amount released per light pulse needs to be measured.

Cell loading and extracellular application

Caged second messengers have been loaded into cells by various means. The most precise way to do this is via a patch pipette. A known concentration of caged compound can be dialyzed into the cell cytosol using this method, with a defined concentration of any other compound such as a Ca²⁺ dye. This technique is especially useful with caged Ca²⁺ compounds, as a Ca²⁺ dye and a known amount of Ca²⁺ typically accompany the probe⁶⁹, ⁷⁰, ⁷¹, ⁷². Caged compounds can be made cell permeable using the acetoxymethyl (AM) esterification technique popular with Ca²⁺ dyes. Cells are loaded by exposing them to a relatively low (< 5 |[mu]|M) extracellular concentration of the cell-permeable AM version of the compound. The caged compound is de-esterified by intracellular esterases. This loading method will lead to qualitative results because the amount of caged compound in the cells is unknown⁷⁵, ⁷⁶. These two approaches are essentially complementary in their advantages and disadvantages: dialyzing the cell gives a known concentration of caged compound but also robs the cytosol of its normal milieu, whereas the AM technique preserves the latter at the expense of precision.

Microinjection and the use of cell penetrating peptides (CPPs) offer alternatives to whole-cell patch clamp and AM esters. Microinjection, like patch clamp, is a skill that requires much practice to perfect. It is less precise than the patch clamp technique in defining the final concentration of caged compound, as the cells' cytosolic volume always varies. It does, however, have the advantages that the normal cytosolic milieu is maintained and many cells can be loaded in a few minutes¹⁹, ⁵¹, ⁵⁴. CPPs have been extensively studied as a means of transporting molecules into cells, as they are an especially attractive means of potentially selectively loading drugs into cells. The mechanisms of loading are far from being completely understood; endocytosis and simple plasma membrane permeation are two possible routes into cells, and it is sometimes thought that both mechanisms occur in parallel. This technique has much potential, but has only been used in one experiment with caged compounds so far⁵³.

Passive diffusion and detergent plasma membrane permeablization⁷⁷ are the remaining methods that have been used to load caged compounds into cells. The latter is quite disruptive to cells, and sometimes cannot be tolerated by them at all. Passive diffusion can only be used when the plasma membrane is actually permeable to the caged compound, and the compound is sufficiently water soluble as to be useful inside the cell³⁰; this is clearly a very fine balance. Almost no caged compounds actually satisfy these criteria. These methods require the caged compound to be continuously applied to cells at a high concentration in a relatively large volume, as in both techniques the cytosol is homeostatic with the extracellular fluid (unlike CPP and AM techniques, these passive loading methods do not concentrate caged compounds inside the cell).

Caged peptides and proteins usually require microinjection or loading into cells via a patch pipette (reviewed in ref. ⁴⁶). This requirement can be overcome by linking the caged peptide to a CPP. Inclusion of a fluorescent label can allow easy visualization of cellular uptake⁵³).

Extracellular application of caged hormones and neurotransmitters is far simpler than intracellular loading. The caged compound can be applied to the cell-bathing solution⁷, ⁹, ⁷⁸ or locally to one part of the preparation via a picospritzer¹⁰. The former method consumes large quantities of caged compounds, but probably defines the concentration more accurately than the latter method.

Quantification of uncaging

Quantification of the amount of photolysis in situ is often an important part of the use of photorelease technology. One solution to this challenging problem is to couple a fluorescence change to the uncaging event; this is really only feasible for large molecules such as peptides or proteins. An elegant realization of inherent quantification of uncaging has been realized with a method for visualizing 14-3-3 activation (that is, phosphorylation) by synthesis of a fluorescent NB-caged phosphopeptide that changes its fluorescence emission upon binding to 14-3-3 (ref. ⁵²).

The extent of uncaging of other types molecules may be measured by fluorescence imaging or from a calibrated bio-response, or estimated from the photochemical properties of the caged compound and the measured flux density in the image plane. Fluorescent Ca²⁺ dyes have been used extensively for quantification of Ca²⁺ uncaging in situ^{69, 70, 71, 72, 79, 80, 81, 82, 83}. Good indicators have not been made for other caged compounds, so the extent of uncaging of these probes must be estimated from a calibration of the flux density of the uncaging beam in the sample, along with the known photophysical properties of the caged compound (quantum yield, extinction coefficient and concentration). If uncaging is performed on the stage of a fluorescent microscope, a simple means of quantification of uncaging has been devised. NPE-caged ATP releases one proton for every molecule of ATP, so from the pH change in a droplet (not a cell) containing a known concentration of NPEATP, the amount of ATP liberated can be estimated, and provide a simple means to estimate the photon flux (only knowledge of the relative flux is required). The percentage photolysis of caged compound (X) actually used is then a ratio of the concentrations and properties of NPE-ATP and caged X, and the relative photon fluxes⁸⁴.

Features and advantages of uncaging

Uncaging has many useful features and advantages compared to other methods for changing the concentration of a molecule inside or on living cells. Specifically, uncaging can be: intracellular, extremely rapid, controlled in time or space, and quantitatively controlled and repeated. There have been many hundreds of studies published using caged compounds. I have selected a few of these to illustrate the advantages of uncaging (see Box 2 for tips on handling caged compounds). Of course, many of these experiments use more than one of the aspects mentioned above simultaneously.

Box 2. Handling Caged Compounds.

Commercially available caged compounds (Table 1) are stored as solids and are sold in small quantities (1–10 mg aliquots). The most convenient way to handle these is to dissolve the entire aliquot in water, dividing this solution into aliquots and storing them at –80°C until needed. Of the compounds listed in Table 1, only caged cAMP and fluorescein are not highly soluble in water, solutions of at least 100 mM can be made of the other compounds. Solutions of calcium cages are stable for many years frozen. MNI-Glu is more stable than CNB-Glu: CNB-Glu has a half-life at room temperature of about 27 h, whereas no hydrolysis of MNI-Glu can be detected after 8 h at room temperature (pH 7.4). Solutions of MNI-D-aspartate can be stored at 4°C for 2 d without detectable hydrolysis99. NPE-caged phosphates have also been found to be highly stable in solution. Exposure to white light should be kept to a minimum. In practice it has been found that caged compounds are not hypersensitive to low levels of room light, but the epi-illumination path of microscopes must be filtered appropriately. I use Roscolux number 10 yellow light filters for all white lights in my lab.

Speed of release

The rate of photorelease of the caged substrate from biologically inert precursor is generally much faster than rapid changing of the solution. This is especially true of complex biological preparations in which it is very difficult to displace the nonactivating solution with a new mixture containing the crucial activating component. In part, this is due to the tortuous nature of many biological preparations and their delicate nature, but also there are inevitable diffusional delays that arise from simple mass action. Use of caged compounds permits the intimate proximity of effector and receptor that bypasses these difficulties. Thus, calcium uncaging in single skeletal muscle fibers activates contraction about five times faster than the most rapid solution change⁸⁵. Multidisciplinary techniques have become so refined that NPE-ATP has been used to measure the rate of force generation by a single dynein molecule (Fig. 4a).

Figure 4. Examples of biological applications of caged compounds.

(a) The rate of force development generated by a single dynein molecule can be measured using an optical trap after photolysis of NPE-ATP⁹⁶. Uncaging of ATP develops 6 pN of force from one dynein arm moving rapidly. (b) Strategy for mapping of APMA-receptor density by two-photon uncaging of glutamate. (c) Scheme for uncaging of mRNA in zebrafish. mRNA encoding the protein of interest is caged, injected into zebrafish at the single-cell stage and uncaged in a discrete volume of the zebrafish at a selected time during development. (d) Quantitative, rapid uncaging of Ca²⁺ in the calyx of Held allows correlation of presynaptic [Ca²⁺] with evoked postsynaptic currents⁸².

Location of release

Another powerful aspect of uncaging technology is that release only occurs where light is incident. Thus, filling the cytosolic compartment of cells with caged compounds permits either global or local concentration changes of the caged compound to be effected simply by using whole-cell or subcellular illumination. For example, an important advantage of intracellular Ca²⁺ uncaging mentioned above was that since Ca²⁺ release was evenly distributed throughout the cytosol (that is, global uncaging), Ca²⁺ microfluorometry could be used to measure the [Ca²⁺], giving an estimate of the [Ca²⁺] at the plasma membrane from channel opening. Local (subcellular) uncaging of Ca²⁺ in longitudinally projecting motor neurons slows rapid growth, consistent with previous observations on pathfinding that growth cone stalling and axon retraction are associated with high-frequency [Ca²⁺] transients^{75, 86}. Similar types of experiments have been performed with caged IP₃ (refs. 84,87). A focused UV beam permits such local uncaging, especially in already small intracellular compartments, but extracellular confinement of uncaging cannot be assisted by cellular geometry, so experimenters have turned to the inherent three-dimensionality of two-photon excitation to achieve localized release of neurotransmitters in the extracellular space⁷⁸ such that diffraction-limited uncaging of glutamate is now possible (Fig. 4b). This technique is starting to find wide application^{10, 88, 89, 90, 91}.

Timing of release

Since uncaging releases its active component independently of the normal biological source, the timing of release can be dictated solely by experimenter and imposed upon the cell in any temporal pattern. Thus, uncaging is orthogonal to the metabolic history of the cell or animal. A striking example of this is the use of caged mRNA that was injected into zebrafish at the one-cell embryonic stage, then uncaged at least 24 h later, at a particular point during development (Fig. 4c).

Intracellular release

Since its inception, the uncaging technique has been used for intracellular release^{13, 30} as light passes through the plasma membrane, enabling concentration changes in the otherwise inaccessible intracellular compartment (Fig. 3c,d). This aspect of the uncaging method is probably the single most powerful advantage compared to other methods of changing solute concentrations. A classic and important example of intracellular uncaging was the photorelease of IP₃ in smooth and skeletal muscle when biological uncaging was still in its infancy¹¹. Loading of caged IP₃ into relaxed muscle, followed by rapid uncaging showed that only in smooth muscle could this second messenger mobilize intracellular Ca²⁺ fast enough to mimic physiological stimulations. This definitively settled a hotly debated question of the time. Intracellular uncaging of Ca²⁺ has also been widely used to study secretory processes in many cell types^{69, 70, 71, 72, 79, 80, 81, 82, 83}.

Quantitation of release

Since the 'calcium hypothesis' for neurotransmitter release⁹² was proposed, obtaining a quantitative relationship between presynaptic [Ca²⁺] and postsynaptic response has been something of a 'Holy Grail' for synaptic physiologists. A fundamental technical difficulty with obtaining the correlation has been the size of the Ca²⁺ microdomains that result from channel opening near the plasma membrane; namely, they are so small that they are beyond the resolution of the light microscope. Global uncaging of Ca²⁺ circumvented this problem, allowing precise correlations between synaptic [Ca²⁺] and postsynaptic current to be measured at a central synapse for the first time^{82, 83} (Fig. 4d).

Future directions

Photolysis of caged compounds is now a well-established technique for studying living cells. A vast array of biomolecules have been caged by chemists and used in biological experiments. The most important of these caged compounds are commercially available (Table 1). The great majority of applications of these compounds use regular near-UV light for uncaging. Recently, two-photon excitation of caged compounds has become practical using mode-locked, Ti:sapphire lasers. However, relatively few caged biomolecules (glutamate^{10, 93} and calcium⁴³) undergo useful, let alone exceptional two-photon excitation. Development of new chromophores with excellent two-photon cross-sections is one frontier for probe development. This is especially important if biologists want to apply uncaging to living animals as the complex biological tissue of the body is highly scattering to light, especially at short wavelengths (350|[ndash]|550 nm). A second frontier for probe development is photoreversible caged compounds. All the probes discussed in this review are activated unidirectionally. Recently techniques for photochemical switching of ion channel conduction⁹⁴, and the |[alpha]|-helical content of peptides have been developed⁹⁵. These studies are signposts for new directions for caged compounds.