Optically Selective Two-photon Uncaging of Glutamate at 900 nm

Abstract

We have synthesized a 7-diethylaminocoumarin (DEAC) derivative that allows wavelength selective, two-photon uncaging at 900 nm versus 720 nm. This new caging chromophore, called DEAC450, has an extended π-electron moeity at the 3-position that shifts the absorption spectrum maximum of DEAC from 375 nm to 450 nm. Two-photon excitation at 900 nm was more than 60-fold greater than at 720 nm. Two-photon uncaging of DEAC450-Glu at 900 nm at spine heads on pyramidal neurons in acutely isolated brain slices generated postsynaptic responses that were similar to spontaneous postsynaptic excitatory miniature currents, whereas significantly higher energies at 720 nm evoked no currents. Since many nitroaromatic caged compounds are two-photon active at 720 nm, optically selective uncaging of DEAC450-caged biomolecules at 900 nm may allow facile two-color optical interrogation of bimodal signaling pathways in living tissue with high resolution for the first time.

The use and recording of color in a scientific context is now a fundamental part of what we do in biomedical research¹. Thus, it is difficult to imagine using confocal fluorescent imaging if it was still monochrome². Fortunately many technological advances have been combined to allow us to use fluorescence imaging to monitor many aspects of neuronal activity in real time in spectrally separate channels³. In contrast, our ability to manipulate cell function with comparable chromatic diversity lags behind imaging and seriously limits our ability to study multiple signaling pathways simultaneously⁴.

Neuroscience is a field in which optical actuation of cell function has been widely used. For example, four methods have been developed for photocontrol of neuronal membrane potential: (1) neurotransmitter uncaging to activate endogenous ligand-gated ion channels^5,6; (2) chemical modification of mutated ion channels with optical switches⁷; (3) photochemical stimulation of genetically targeted alien ion channels⁸; and (4) excitation of genetically delivered photoregulated ion pumps and channels^9–12. Each of these methods has advantages and disadvantages. The first is powerful because it directly activates native receptors, so is useful for understanding the details of cellular physiology in vitro¹³. A striking feature of methods 2 and 4 is the wavelength selectivity that is inherent to or engineered into chromophores such that two colors of light can be used orthogonally for different purposes.

Starting in 1978, hundreds of biological studies have been reported using nitrobenzylcaged compounds using photolysis at short wavelengths of light, i.e. in the 350–400 nm range for one-photon^14–19 and 720–740 nm for two-photon^18,20,21 photolysis. Caging chromophores that absorb at longer wavelengths than these compounds have only been recently developed, so these have been applied to relatively few biological questions. In particular, several substituted organic, laser-dye based 7-aminocoumarins²² and organicinorganic hybrid chromophores based on the ruthenium- bipyridyl^23,24 (RuBi) scaffold are effectively photolyzed at wavelengths longer than 400 nm for one-photon or 740 nm for two-photon photolysis. Caged neurotransmitters using such chromophores are important additions to the optical arsenal available to neurobiologists, but their absorption spectra lack pronounced minima at short wavelengths (Figure 1). Here we introduce a new caged glutamate compound, called DEAC450-Glu (Figure 1a) that is relatively photoinactive at short wavelengths (e.g. 720 nm) and undergoes maximal two-photon excitation at 900 nm. This significant bathochromic shift thus extends the color palette of two-photon photolysis to a region that it is optically complementary to many other caged compounds.

Figure 1.

Synthesis and spectral properties of DEAC450-Glu. (a) Reagents and conditions: (a) tert-Butyldimethylsilylchloride, imidazole. (b) N-bromosuccinimide, NaOAc. (c) tert-Butylacrylic acid, Pd(OAc)₂, LiCl, NaHCO₃, Bu₄NCl. (d) TFA. (e) Di-tert-butylaspartate, EDC. (f) TBAF. (g) tert-Butyl-BOC-L-glutamate, EDC followed by TFA. (b) Absorption spectra of CDNI-Glu (cyan), N-DCAC-GABA (DEAC core in pink) and DEAC450-Glu (red). (c) Absorption spectra of RuBi-Glu (green), ANBP-GABA (orange) and DEAC450-Glu (red).

The synthesis of DEAC450-Glu (Figure 1a) started by protection of the known DEAC-alcohol²⁵ to give 1. The coumarin 3-position was functionalized with NBS²⁶ to give 2, followed by Heck coupling²⁷ of tert-butylacrylic acid with 2 to give 3. Deprotection of the tert-butyl to acid 4, was followed by carbodiimide coupling of di-tert-butyl-D-Asp to give 5. Selective removal of the silyl group gave alcohol 6, which was coupled to acid side chain of L-glutamate to give fully protected DEAC450-Glu. Finally, the remaining protecting groups were removed and the product purified by HPLC to give 7 (DEAC450-Glu). DEAC450-Glu was found to be soluble (up to 7.5 mM) and quite stable at pH 7.4. Solutions showed no hydrolysis over 60 days at −20°C or at RT for 5 h, and at 37°C only 2% was hydrolyzed in 2 h. Irradiation of DEAC450-Glu and RuBi-Glu²⁴ at 473 nm at pH 7.4 revealed that the former was photolyzed three times faster, corresponding to a quantum yield of photolysis of 0.39. DEAC450-alcohol was found, by HPLC analysis of the reaction mixture, to be cleanly released. The new chromophore has an extinction coefficient of 43,000. These data taken together show that DEAC450-Glu is photochemically one of the most efficient caged Glu probes that has been developed (Table 1). Importantly, the absorption maximum at 450 nm is significantly red-shifted compared to simple DEAC derivatives such as N-DCAC-GABA²⁸ (Figure 1b). Furthermore, the absorption minimum of DEAC450 is at the maximum for the CDNI^28,29 chromophore (Figure 1b). The relative absorption at the λ_min is 9% of the λ_max; no such distinct minima exist for other caged compounds (Figure 1c) that have been recently shown to be two-photon sensitive^24,30,31. This relative difference in linear absorption is further enhanced in the two-photon domain; the DEAC450 chromophore is >60x more fluorescent at 900 nm than 720 nm. This significant difference in two-photon excitation lead us to test comparative uncaging of DEAC450-Glu on single spine heads in acutely isolated brain slices at these wavelengths.

Table 1.

Summary of the properties of caged glutamate probes.

Cage	ε/M−1cm−1 (λmax, nm)	QY	ε.QY	2PuCS/GM (λ, nm)
MNI2,44	4,300 (330)	0.085	357	0.06 (740)
RuBi24	5,600 (450)	0.13	728	0.14 (800)
PMNB45	9,900 (317)	0.1	990	0.45 (800)
CDNI35	6,400 (330)	0.5	3,200	0.06 (720)
PNEB46	9,900 (317)	0.1	990	3.3 (740)
DEAC450	43,000 (450)	0.39	16,800	0.5* (900)

Symbols and abbreviations. ε, extinction coefficient; λ_max, absorption maximum; QY, quantum yield; 2PuCS, two-photon uncaging cross section;

^*estimated from Fig. 2.

Two-photon photolysis of a solution of DEAC450-Glu (local perfusion from a nearby pipette at 0.25 mM) at 900 nm (10 mW, 0.5 ms) induced excitatory post-synaptic currents at an isolated spine head (Figure 2a) that were similar in size and duration to spontaneous miniature excitatory postsynaptic potentials³². A comparable amount (144%) of energy at 720 nm evoked no response (Figure 2a). This latter energy dose can be used to photolyze nitroindolinyl-caged neurotransmitters such as MNI-Glu^32–35 (Figure 2b) and CDNI-GABA²⁹. Note that DEAC cages and other fluorophores do not interfere with nitroaromatic photolysis at 720 nm^18,28. Power response curves implied that DEAC450-Glu was uncaged by two-photon excitation at 900 nm (Figure 2c), just as MNI-Glu and NDBF-EGTA at 720 nm^32,36. Similar to other caged Glu and GABA probes^24,29,30, DEAC450-Glu had off-target pharmacological side effects upon GABA-A receptor currents. We found the EC₅₀ for blocking evoked GABA-A receptor cur- rents on layer 2/3 pyramidal neurons was about 33 µM. In comparison, two commercially available caged neurotransmitters, MNI-Glu and RuBi-Glu, had EC₅₀ values of 105 µM and 7.7 µM, respectively (Supporting Information).

Figure 2.

Optically selective two-photon uncaging of glutamate at 900 nm. Caged neurotransmitters were topically applied to pyramidal cells through a puffer pipette positioned just above the surface of an acutely isolated mouse brain slice. Excitation was with a mode-locked Ti:sapphire laser tuned to the specified wavelength. Power was measured at the exit of the microscope objective. Currents were measured at the cell soma using whole-cell voltage-clamp recordings. (a) Two-photon fluorescence image of dendritic segment on a pyramidal neuron filled with Alexa-594 (top panel). DEAC450-Glu, applied at 0.25 mM, was irradiated with 900 nm or 720 nm light for a period 0.5 ms. The photolysis beam was positioned (yellow dot) next to a spine head and the postsynaptic current induced by 10 mW at 900 nm was similar to that evoked by synaptic release. Irradiation at the same position with 12 mW of 720-nm light evoked no current. (b) MNI-Glu, applied at 10 mM, was irradiated with 720 nm light for a period of 0.5 ms and evoked a postsynaptic current that was to similar synaptic release. Importantly, nitroindolinyl-caged glutamate is 38-fold less two-photon active at 830 nm²⁸. (c) Powerdosage curve of evoked postsynaptic current showed a two-photon excitation effect for the postsynaptic current for DEAC450-Glu at 900 nm (solid points). A much weaker response was evoked at 720 nm (open circles).

DEAC450-Glu has a uniquely powerful set of properties when compared to the many other caged glutamates. Table 1 shows a summary of the properties of a range of widely used and recently developed caged glutamate probes. It can be seen that DEAC450-Glu is highly efficient for uncaging with visible light. For example, in comparison to the two commercially available probes^24,32 in Table 1, DEAC450-Glu is about 23x more active than RuBi-Glu at 450 nm, and about 329x more active than MNI-Glu at 405 nm. The latter wavelength corresponds to excitation with a purple laser that is widely deployed on confocal microscopes³⁷. The 11-fold difference in excitation at 350 versus 450 nm may permit two-color uncaging experiments in the linear excitation domain, when DEAC450 is paired with regular nitrobenzyl caged compounds. However, we believe the real strength of our new caging chromophore for two-color uncaging may be seen when using two-photon excitation. In this modality, the relative ability to excite DEAC450 at 900 nm versus 720 nm is > 60x. Such optical selectivity allowed us to induce currents at single spine heads in acutely isolated brain slices, that were similar in size and kinetics to synaptic events³², by uncaging at 900 nm (Figure 2a). Importantly, we found that a higher energy dosage at the shorter wavelength (720 nm) evoked no significant currents (Figure 2a). This short wavelength is the one that has been widely used for uncaging nitroindolinyl-caged transmitters (e.g. MNI and CDNI compounds^29,32,35, Table 1). Thus, DEAC450 and such chromophores could form a near-perfect pair of optically complementary cages for two-color, two-photon uncaging. Several other reports of two-color uncaging have appeared^38–42, however all these approaches are constrained in some important way. Some require the complete photolysis of the longer wavelength before applying short wavelength uncaging³⁸- ⁴⁰. Others require the relative concentrations of the caged molecules to be set such that compounds have hugely different concentration ratios⁴¹. Finally, some chromophores require uncaging with light that is not compatible with modern microscope glass or amenable to single synapse stimulation⁴². In contrast, our new caging chromophore, DEAC450, enables, for the first time, two-photon uncaging of a biomolecule at long wavelengths (900 nm) with almost complete optically selectivity versus shorter wavelengths (720 nm). For synaptic physiology, this technological advance is an important break through, as it may allow the study of how excitatory and inhibitory transmitters sculpt dendritic integration with single synapse resolution⁴³.