Optical control of NMDA receptors with a

Full text

Turn on search term navigation

ARTICLE

Received 28 Aug 2014 | Accepted 13 Jul 2015 | Published 27 Aug 2015

Laura Laprell1,*, Emilienne Repak2,3,*, Vilius Franckevicius1, Felix Hartrampf1, Jan Terhag4, Michael Hollmann4,

Martin Sumser1, Nelson Rebola2,3, David A. DiGregorio2,3 & Dirk Trauner1

N-methyl-D-aspartate receptors (NMDARs) play a central role in synaptic plasticity, learning and memory, and are implicated in various neuronal disorders. We synthesized a diffusible photochromic glutamate analogue, azobenzene-triazole-glutamate (ATG), which is specic for NMDARs and functions as a photoswitchable agonist. ATG is inactive in its dark-adapted trans-isoform, but can be converted into its active cis-isoform using one-photon (near UV) or two-photon (740 nm) excitation. Irradiation with violet light photo-inactivates ATG within milliseconds, allowing agonist removal on the timescale of NMDAR deactivation. ATG is compatible with Ca2 imaging and can be used to optically mimic synaptic coincidence detection protocols. Thus, ATG can be used like traditional caged glutamate compounds, but with the added advantages of NMDAR specicity, low antagonism of GABAR-mediated currents, and precise temporal control of agonist delivery.

DOI: 10.1038/ncomms9076 OPEN

Optical control of NMDA receptors with a diffusible photoswitch

1 Department of Chemistry and Pharmacology, Ludwig-Maximilians-Universitat, Mnchen, and Center for Integrated Protein Science, Munich 81377, Germany. 2 Institut Pasteur, Unit of Dynamic Neuronal Imaging, 25 rue du Dr Roux, Paris Cedex 15 75724, France. 3 CNRS UMR 3571, Genes, Synapses, and Cognition, Institut Pasteur, 25 rue du Dr Roux, Paris Cedex 15 75724, France. 4 Ruhr-Universitat-Bochum, Department of Biochemistry, Bochum 44780, Germany. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to D.D.(email: mailto:[email protected]

Web End [email protected] ) or to D.T. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076

Ionotropic glutamate receptors mediate fast excitatory synaptic transmission and are ubiquitously expressed in the central nervous system. They can be separated into three major

classes: a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs), kainate receptors and N-methyl-D-aspartate receptors (NMDARs). The latter are involved in the induction of synaptic plasticity, the cellular correlate of learning and memory1. NMDARs, of which a functional tetrameric receptor structure has been reported lately2,3, have also been implicated in a variety of neurological diseases and dysfunctions including ischaemia-related cell death, Alzheimers, Huntingtons and Parkinsons Diseases, as well as schizophrenia and autism spectrum disorders1. NMDARs are heterotetramers whose subunit composition varies based upon brain region, maturation and synaptic activity. Subunit differences result in variations in receptor kinetics, which confer different computational properties on the receptors1,4. Such variations in kinetic behaviour of NMDAR subunits have been examined on extrasynaptic and recombinant receptors using outside-out patch clamp recordings and fast application of agonist5,6.

The study of glutamate receptors in their native environment has been facilitated by the development of optical tools, namely light-sensitive caged agonists and antagonists, which take advantage of the temporal and spatial precision that light provides. Compounds including caged glutamate, caged NMDA and caged MK-801 have proven very useful for nely tuned, non-invasive studies of NMDARs711. For example, MNI-glutamate uncaging in a diffraction-limited laser illumination volume enables synaptic-like activation of both AMPARs and NMDARs1215. Larger illumination volumes can be used to quantify activation and desensitization kinetics of synaptic receptors15. However, the study of deactivation of receptors must be done following rapid removal of agonist, a feat not possible with caged glutamate because of slow clearance by diffusion or uptake15. This is particularly challenging in functional networks within brain tissue.

Molecular photoswitches provide an interesting alternative to caged compounds since they can be rapidly and repeatedly switched on and off and do not generate photochemical byproducts. Over the past decade, several photoswitches have been developed including both photoswitchable tethered ligands and freely diffusible photochromic ligands (PCL) that allow for the optical control of transmembrane receptors and, by extension, neural systems16. GluAzo is a photochromic version of glutamate that functions as a PCL for the kainate receptors GluK1 and GluK217, and ATA-3 is a PCL selective for AMPARs18 (Fig. 1a). Both GluAzo and ATA-3 are active in the thermally relaxed, dark-adapted state, thus their use with networks of neurons is challenging as the whole preparation must be illuminated to prevent tonic activity.

We now report a unique photochromic agonist that targets a different family of glutamate receptors. This compound, azobenzene-triazole-glutamate (ATG), complements ATA-3 and GluAzo since it selectively activates NMDARs (Fig. 1b). In addition, it possesses an important functional advantage that distinguishes it from our previously developed PCLs: it is inactive in the dark-adapted trans-isoform, but quickly converts into its active cis-isoform when irradiated with ultra-violet (UV) light (370 nm). As such, it is not excitotoxic when applied to neural networks in the absence of light, and it becomes an agonist of NMDARs with millisecond precision when irradiated. In addition, we show that ATG is the rst diffusible PCL that can be precisely controlled with two-photon excitation (740 nm).

ResultsSynthesis and Photophysical Characterization of ATG. ATG was synthesized in a few steps from the known glutamate

derivative 4(R)-propargyl glutamate (1) using click chemistry (Fig. 1c). In brief, 1 was treated with azobenzene azide 2 in the presence of a copper catalyst to afford triazole 3. Global deprotection then yielded ATG. A thermally stable stilbene analogue of cis-ATG, termed cis-STG, was synthesized in a similar fashion by treating 1 with 4 and deprotecting the resultant triazole 5 (Fig. 1d). Details of the synthesis and full characterization can be found in Supplementary Fig. 10, Supplementary Note 1 and in the Supplementary Methods.

ATG behaves as a regular azobenzene that can be converted to its cis-isoform with UV-A light. Conversion into the thermodynamically favourable trans-isoform requires irradiation with violet light (Fig. 1 and Supplementary Fig. 1a). Thermal relaxation into the trans-isoform is very slow in physiological buffer solution in accordance with the regular azobenzene nature of ATG (Supplementary Fig. 1b).

Photopharmacology of ATG. To evaluate ATG as a photo-switchable agonist in neurons, we performed electrophysiological recordings in mouse layer 2/3 cortical neurons in acute coronal slices, while continuously perfusing 200 mM ATG in articial cerebrospinal uid (ACSF) (Fig. 2, Supplementary Fig. 2a,b). Using the whole-cell voltage-clamp conguration, we examined the spectral sensitivity of cis-ATG-evoked currents from 350 to 410 nm (Fig. 2a, Supplementary Fig. 2a). Maximal ATG-elicited currents were observed in response to 360 nm light,(3.75 mW mm 2; Supplementary Fig. 2c), whereas above 390 nm cis-ATG-mediated currents were negligible (o 10 pA). This

action spectrum corresponds to the maximal conversion to cis-ATG after 365 nm irradiation, as determined by UVVIS spectroscopy (Supplementary Fig. 1a). Switching back to trans-ATG was fastest using 425 nm light (toff 0.17 s0.03;

n 16, Fig. 2b and Supplementary Fig. 2b). According to

these data, ATG is cis-active, which is in sharp contrast to other photoswitchable glutamate receptor agonists previously developed17,18. Using our illumination system, the doseresponse curve of ATG indicated an EC50 value of 185 mM under 370 nm light (Fig. 2c). In whole cell current-clamp recordings of mouse cortical layer 2/3 neurons, ATG triggered action potential (AP) ring under 370 nm light (Fig. 2d). Using 420 nm light, AP-ring could be quickly silenced. Thus, ATG photoswitching can be used to control neural activity with light.

We examined cis-ATG-mediated currents in the presence of various antagonists of ionotropic glutamate receptors (iGluRs) to identify the molecular targets of ATG. Application of NBQX (25 mM), an AMPAR-selective antagonist, had no effect on light-evoked currents (Supplementary Fig. 3). By contrast, D-AP5 (40 mM), a competitive NMDAR-selective antagonist, and MK-801 (50 mM), a use-dependent pore blocker that preferentially acts on NMDARs, completely abolished cis-ATG-mediated AP ring and currents, respectively (Fig. 2e and Supplementary Fig. 3)19.

Using heterologous expression strategies we next examined whether cis-ATG can activate NMDA receptors containing different isoforms. Cis-ATG-mediated currents were not detectable in HEK cells expressing different subunit combinations. In Xenopus oocytes, however, the high level of receptor expression facilitated the detection of cis-ATG-mediated currents (Supplementary Fig. 4). To normalize for different expression levels we compared the steady-state amplitude of cis-ATG-mediated currents in oocytes expressing recombinant diheteromeric NMDARs with those evoked by saturating NMDA concentrations (1 mM, Supplementary Fig. 4). We found that photoactivation of ATG is able to activate all subunit combinations between GluN1-1a and either GluN2A, B, C or D. The observation that cis-ATG-mediated currents were smaller

2 NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076 ARTICLE

HOOC NH2

N N

NH2

HOOC COOH

trans-GluAzo trans-ATA

N N N

NH2

370 nm

415 nm

NH2

HOOC COOH

trans-ATG cis-ATG

N N

NHBoc

HOOC COOH

1 3

HOOC COOH

CuSO4, sodium ascorbate (92%)

HOOC COOH

NH2HCl

N N

HCl, EtOAc

(95%)

trans-ATG

N N

NHBoc

CuSO4, sodium ascorbate (82%)

NHBoc

HOOC COOH

N N

HCl, EtOAc

(80%)

NH2HCl

HOOC COOH

cis-STG

Figure 1 | Design and synthesis of ATG. (a) Structures of GluAzo, a photochromic agonist of kainate receptors, and ATA, a photochromic agonist of AMPA receptors in their respective trans isoform. (b) Structure and photophysical properties of ATG. The molecule consists of a photoswitchable azobenzene, a triazole and a glutamate moiety. The trans- and cis-conguration of ATG are shown. (c) Synthesis of the azobenzene ATG usingclick chemistry. (d) Synthesis of the stilbene cis-STG using click chemistry.

than those elicited by superfusion of NMDA (o5%), may be because of increased light absorption resulting in incomplete photoconversion in the oocyte setup.

To further demonstrate the selectivity of ATG for NMDARs, we recorded currentvoltage (IV) relationships comparing NMDA (200 mM) puff application with ATG photoswitching (Fig. 2f). Because of their magnesium sensitivity, NMDARs are

partially blocked at resting membrane potentials, imparting a J-shaped IV relationship20 (Fig. 2f, right, black), which we observed for both NMDA application and cis-ATG-mediated currents (Fig. 2f). As expected for non-selective cation channels, the reversal potential was close to 0 mV. In the absence of external Mg2 the IV relationship was found to be linear, as expected for NMDARs (Fig. 2f, right, blue).

NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076

a c

Normalized current

1.0

0.8

0.6

0.4

0.2

0.0

Toff (s)

1.2

0.8

0.4

0.0

Relative current

0.6

0.4

0.2

350

Wavelength (nm) Wavelength (nm)

370

390

410

400

450

500

550

6 6 4

2 2

4 100

(M) ATG

20 mV

1 s

420 nm light

370

AP-5 wash-in After wash-out

40 mV

10 s

1 4

2 5

20 mV

500 ms

1 4

2 5

250

(pA)

150

250 pA

500 ms

90 40 10

(mV)

150

NMDA cis-ATG cis-ATG - Mg2+

250

350

Figure 2 | Photopharmacology of ATG. (a) Action spectrum of ATG recorded in layer 2/3 cortical neurons in an acute slice preparation in presence of 200 mM ATG in ACSF. Current amplitude was measured after 5 s light stimulation with the respective wavelength and normalized to the maximal current amplitude at 360 nm. (b) Wavelength screening for toff kinetics of ATG-mediated currents between 400 and 560 nm light. Best toff kinetics were achieved at 400450 nm light. (c) Doseresponse relationship of ATG-mediated currents in cortical slice preparations. Concentrations from 1 to 500 mM were tested. The EC50 is 185 mM (black dashed line) and was calculated using the Hill-equation. (d) Current-clamp recording of a layer 2/3 cortical neuron.

Irradiation with 370 nm light (purple) induces robust action potential ring that is terminated by irradiation with 420 nm light (blue). (e) Washing in D-AP-5 (40 mM), an NMDA-specic antagonist, blocks the ATG-mediated light-dependent action potential ring. (f) Currentvoltage relationships indicative of NMDARs as targets for ATG. Black; currentvoltage relationship of puff-applied NMDA (200 mM) currents (n 12 cells). Red; currentvoltage

relationship of ATG-mediated currents under 370 nm light (n 10 cells). Blue; currentvoltage relationship of ATG-mediated currents in the absence of

Mg2 ions (n 10 cells). Error bars indicate s.e.m.

4 NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076 ARTICLE

a b

375 nm

Bath application

[cis-ATG]

Time

375 nm only

375 nm + 405 nm

405 nm

50 pA

2 s

2.5 ms 10 ms 50 ms

Local application

1,000

Half decay (ms)

100

0.5 s

ATG 375 nm onlyATG 375 nm 405 nm

MNI-glu + TBOA

ATG 375 nm 405 nm

ATG 375 nm

50 ms

Figure 3 | Dendritic NMDAR currents evoked by rapid laser-mediated photoswitching of ATG. (a) Schematic diagram showing putative sculpting of NMDAR gating by ATG photoswitching. A near-diffraction-limited spot of 375 nm light switches ATG to an active cis-conformation (top) that can activate the NMDAR transiently. When the 375 nm laser light is followed quickly by a brief 405 nm laser pulse focused over a larger volume, ATG is converted to the inactive trans-conformation (bottom), eliminating cis-ATG-mediated current more quickly than via diffusional clearance of cis-ATG. (b) Upper traces show light-evoked NMDAR currents recorded in CA1 pyramidal neurons while bath applying 200 mM ATG, in response to a 500 ms 375 nm laser pulse immediately followed by various durations of 405 nm laser pulses. Lower traces show smaller currents evoked by 405 nm pulses alone. Inset: confocal image of dendrite stimulated in these recordings. Purple dot indicates targeted point of ATG stimulation. Scale bar 3 mm (c) Normalized population averages of NMDAR currents evoked by 375 nm laser pulse (100 ms) alone (red; n 9 cells), or 375 nm followed by 405 nm laser pulse (50 ms; magenta; n 9 cells) when locally applying

ATG (100 mM) with a patch pipette. Blue trace represents uncaging-evoked NMDAR responses when locally applying MNI-glutamate (100 mM; n 5 cells).

Dotted line on the magenta trace in the inset indicates the double exponential decay function. (d) Bar graph shows half-decay of NMDAR currents from cells inc. Error bars indicate s.e.m. *Po0.05 for all three comparisons (Steel Dwass all pairs nonparametric multiple comparison test).

We further examined whether the thermally stable cis-ATG analogue (cis-STG, Fig. 1d), which does not photoswitch, has similar pharmacology and specicity for NMDARs. When puff applied, the stilbene cis-STG indeed elicited APs in mouse layer 2/3 cortical neurons (Supplementary Fig. 5a). The J-shaped I-V relationship (Supplementary Fig. 5b) indicates that cis-STG also targets NMDARs. As such, cis-STG represents a new structural class of agonist for these receptors.

One drawback of traditional caged glutamate compounds is that they are often antagonists of GABAAR-mediated synaptic currents21. We therefore tested the effect of trans-ATG on GABAAR-mediated inhibitory postsynaptic currents (IPSCs) in hippocampal CA1 pyramidal neurons. Using 400 mM bath application of ATG, we observed no detectable alteration of spontaneous IPSCs, but observed a 386% (n 10 cells) block of

evoked IPSCs and an increase in coefcient of variation, consistent with a presynaptic target (Supplementary Fig. 6). Nevertheless, this is less than the 50% block by RuBi-glutamate (300 mM21), the 55% block by CDNI-glutamate (400 mM22) and the 83% block by the commonly used MNI-glutamate (300 mM21).

ATG-mediated photoswitching of NMDAR gating. We next considered the possibility that ATG photoswitching could be used

to perform temporally precise agonism of NMDARs. We used fast digitally controlled diode lasers at 375 nm to switch to cis-ATG and 405 nm to preferentially switch to trans-ATG (Fig. 3a). The 375 nm laser light was focused to a near-diffraction-limited spot (full-width half-maximum (FWHM): 300 nm; Supplementary Fig. 7) and the 405 nm laser beam was adjusted to form a 4 mm spot (full-width half-maximum; Supplementary

Fig. 7). NMDAR currents were recorded from CA1 pyramidal neuron dendrites within 100 mm of the soma at a holding potential of 30 mV. In the rst set of experiments, 200 mM ATG was bath-applied (Fig. 3b). Laser illumination of 100 ms at 375 nm (B150 mW) was required to see signicant cis-ATG-mediated currents (average peak response: 8712 pA; n 6

cells). Illumination for 5 s with 375 nm laser light resulted in large currents (49456 pA, n 18 cells) that rose over the light pulse

duration (Supplementary Fig. 8a) and decayed over tens of seconds (Supplementary Fig. 8a,b). The reduction in cis-ATG-mediated current was maximal when using 50 ms illumination of 405 nm laser light (11 mW), but incomplete (563 % reduction, n 14 cells; Fig. 3b). The residual current was similar in ampli

tude to that when delivering a 50 ms 405 nm light pulse alone (P 0.35; Wilcoxon-matched pair signed rank test). The UVVIS

spectra indicate that irradiation with 405 nm light produces a

NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076

375 nm

405 nm

375 nm

405 nm

Wild-type GluN2A KO

740 nm

20 pA

400 ms

20 pA

375 nm only375 nm 405 nm

405 nm only

740 nm

2 s

160

140

120

100

20 pA

200 ms

Half-decay (ms)

WT synaptic (n = 13) WT ATG (n = 9)

KO synaptic (n = 10) KO ATG (n = 5)

20 pA

10 ms

WTATG

KOATG

WT synaptic

KO synaptic

300 ms

Figure 4 | Comparison of ATG photoswitching responses between wild-type and GluN2A KO animals. (a) Population averages of light-evoked currents from WT CA1 pyramidal cells in response to 375 nm (100 ms) only, 375 nm followed by 405 nm (50 ms), and 405 nm only when locally applying ATG (100 mM) with a patch pipette (n 9 cells). (b) Population

averages of photoswitching currents from GluN2A KO animals under same conditions as (a) (n 5 cells). (c) (left) Normalized currents in response to

375405 nm photoswitching from (a) and (b) and population averages of NMDAR EPSCs in wild-type (n 13 cells) and KO animals (n 10 cells).

Traces were aligned on their peaks and electrical artifacts from presynaptic stimulation have been blanked. Right: Bar graph of half-decays. Error bars indicate s.e.m. *Po0.05 and NS indicates comparisons that are not signicantly different (Steel Dwass all pairs nonparametric multiple comparison test).

Isochronal amplitude (pA)

1.0

0.8

0.6

0.4

0.2

0.0

Distance from spine (m)

Figure 5 | Localized two-photon activation of ATG. (a) Cis-ATG-mediated current evoked by two-photon illumination (1 ms, 740 nm) in a CA1 pyramidal cell while bath applying 400 mM ATG. (b) 2P-evoked cis-ATG-mediated currents with illumination spot parked at 0.5, 2 and 4 mm away from spine head. Illumination duration was 1 ms, and the wavelength set at 740 nm. (inset) Enlarged view of distance-dependent ATG evoked responses. Box over traces illustrates the time window over which spatial dependence was estimated for isochronal amplitude plots in (c) (Scale bar 2 mm). This was chosen to correspond to the time point at which the largest current reached 75% of its amplitude. (c) Normalized isochronal plots for six cells, with the average in black (half-width half-maximum 2.0 mm,

red dotted lines). Error bars indicate s.e.m.

small but measurable amount of cis-ATG (Supplementary Fig. 1a). This could explain why 405 nm illumination reduces current following 375 nm illumination, but evokes current when not preceded by 375 nm illumination (Fig. 3b).

When we examined the spatial dependence of cis-ATG-mediated currents, we saw little change in the peak amplitude even if the 375 nm laser was positioned up to 100 mm away from the dendrite. This nding, along with the slow rise and decay times, led us to the hypothesis that we had a very large effective photoactivation volume because of ATG photoswitching out-of-focusan effect observed for bath application of caged compounds23. To further improve the kinetics of the NMDAR response, we tested local perfusion of ATG at 100 mM. Under these conditions, the decay of light-evoked cis-ATG-mediated currents was accelerated over sixfold, and we observed a halving of the peak current at only 5 mm away from the dendrite (Supplementary Fig. 8d). Local perfusion also permitted the use of higher concentrations of ATG to improve the activation time of NMDAR currents to o100 ms, similar to that of the widely used caged compound MNI-glutamate at similar concentrations (Supplementary Fig. 9a). Additionally, local application of ATG

resulted in an enhanced fractional reduction of cis-ATG-mediated currents upon 405 nm light illumination (Fig. 3c and Supplementary Fig. 9b), up to 773% reduction, 100 mM (n 9

cells). Finally, and most importantly, local application of ATG resulted in faster decay of NMDAR currents following 405 nm illumination (ATGoff): the half-decay of 334 ms (n 9 cells;

Fig. 3d) was over 12 times faster than NMDAR current decays recorded in response to MNI-glutamate uncaging (42272 ms, n 5 cells, P 0.01). Interestingly, the weighted decay time

constant of ATGoff (tweighted of 10242 ms; n 9 cells; see

example t in Fig. 3c inset) was intermediate to the decay values for recombinant receptors containing either GluN1/GluN2A (tweighted 29 ms) and GluN1/GluN2B (t

weighted

193 ms)24.

6 NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076 ARTICLE

+TTX and felodipine

20 mV

500 ms

ATG-induced calcium transients

F/F

0.2

0.1

500 ms

0 TTX TTX+Fel

F/F

3.6 s

0.0 s 6.9 s

30%

Figure 6 | Calcium imaging using ATG in acute hippocampal slices. (a) Cis-ATG-mediated (200 mM) electrical signals in ACSF (left) and in the presence of 40 mM felodipine (Fel) and 1 mM TTX (right), elicited with 370 nm light and terminated with 420 nm (250 ms light pulse for each wavelength).

(b) Calcium transients from responsive cells in the eld of view corresponding to cis-ATG-mediated recording presented in (a). Bar graph: quantication of calcium transients (ATG TTX: n 18 experiments and ATG TTX felodipine: n 10 experiments). *Po0.05, Wilcoxon rank-sum test. Error bars

indicate s.e.m. (c) Changes in uorescence (DF/F) at different time points of the calcium transient; prior to light stimulation, immediately after illumination and after returning to basal calcium levels.

We therefore considered the possibility that the rapid ATGoff

decay could be used to estimate relative contributions of GluN2A and GluN2B to dendritic NMDAR activation.

In mature CA1 pyramidal neurons (4P21), NMDARs are mostly comprised of tri-heteromers of GluN1, GluN2A and GluN2B4,25 whose kinetics are dominated by the rapid deactivation of GluN2A26,27. We therefore compared the decay of ATGoff in WT and GluN2A KO mice. To get a better estimate of the ATGoff kinetics without contamination of 405 nm-induced cis-ATG-mediated currents, we subtracted a scaled curve t to the 405 nm-induced current of each individual cell (Fig. 4a,b). The half-decay of subtracted ATGoff currents was 304 ms (n 9

cells) similar to the decay of NMDAR EPSCs recorded at the same age (403, n 13 cells; P 0.18; Steel Dwass all pairs

nonparametric multiple comparison test; Fig. 4c). To determine whether the decay of ATGoff currents was inuenced by GluN2A expression, we examined ATGoff decays in GluN2A KO mice (P29-P55). Surprisingly, the decay of ATGoff remained largely unaltered (KO half-decay: 315 ms, n 5, P 0.99 Steel Dwass

all pairs nonparametric multiple comparison test), but the evoked synaptic responses were nearly three times slower (KO half-decay: 1137 ms, n 10, P 0.0004, Steel Dwass all pairs

nonparametric multiple comparison test; Fig. 4c), conrming the kinetic inuence of GluN2A expression. Thus, the ATGoff

decay is insensitive to subunit composition of NMDARs that are known to alter channel deactivation28.

Two-photon activation of ATG. One strategy to achieve very localized photoactivation is to use two-photon (2P) excitation.

We found that, during bath application of 400 mM ATG, femtosecond pulsed-laser illumination at 725740 nm evoked detectable cis-ATG-mediated currents when using illumination durations as brief as 250 ms, which is much more efcient than one-photon activation of ATG (Fig. 5a), and similar to illumination durations required to evoke NMDAR currents using MNI-glutamate2932. One millisecond duration pulses (740 nm) produced an average cis-ATG-mediated current of 487 pA (rise time (10-90%) 444 ms, n 16 spines), larger than

the published values for single spine activation using caged-glutamate2931. Nevertheless, spatial dependence of the amplitude of 2P-evoked cis-ATG-mediated currents indicated a local activation within 2 mm of the spine head (Fig. 5b,c).

Combining ATG photoactivation with Ca2 imaging. The calcium permeability of NMDARs links their activity to post-synaptic biochemical alterations, such as spine morphology and glutamate receptor expression, which are associated with synaptic plasticity20. Since ATG specically acts on NMDARs, we examined the possibility of imaging its effect with a Ca2 -sensitive uorescent dye (Fig. 6). To this end, we incubated hippocampal slices with Quest Fluo-8-AM, a membrane-permeable Ca2 indicator that is compatible with the activation wavelength of ATG because of its excitation wavelength of 490 nm (Fig. 6b). We applied tetrodotoxin citrate (TTX) (1 mM) and felodipine (40 mM), which prevents opening of voltage-gated Ca2 channels to limit the response to

NMDAR-mediated calcium entry (Fig. 6 right half). Indeed, upon irradiation with 370 nm in the presence of ATG,

NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076

20 mV

2 s

10 ms

50 ms

100 ms

370 nm

Relative number of spikes

1.0

0.8

0.6

0.4

0.2

20 mV

2 s

***

20 mV

Pre100

Pre50

Pre10

Post100

Post50

Post10

Zero

370 nm

PRE

ZERO

POST

Figure 7 | Coincidence detection using ATG in layer 2/3 cortical neurons. Coincidence detection of cis-ATG mediated current (200 mM) paired with antidromic stimulation. (a) Antidromic stimulation (black bars) of the postsynaptic cell 10 ms before, during and 10 ms after the light stimulation (purple trace). (b) As in (a), but with 50 ms intervals. (c) As in (a), but with 100 ms intervals. (d) Quantication of coincidence detection. Relative number of spikes compared with condition ZERO, when both stimuli were applied together (n 11 cells). Statistics were calculated using the Wilcoxon

rank-sum test (*Po0.05,**Po0.01, ***Po0.001).

we observed light-evoked cis-ATG-mediated Ca2 transients (Fig. 6b; DF/F 203%, n 18) which decreased in amplitude

in the presence of blockers (DF/F 133%, n 10). We

attributed the remaining Ca2 transient to the inux through ATG-mediated NMDAR opening.

Mimicking synaptic coincidence detection with ATG. One of the most intriguing properties of NMDARs is their voltage-sensitive Mg2 block at resting membrane potential.

Accordingly, the presence of a neurotransmitter and concomitant postsynaptic depolarization are required to activate the ion channel, rendering NMDARs coincidence detectors of pre- and postsynaptic activity33. To test whether cis-ATG-mediated NMDAR activation can replace the synaptic stimulation necessary for coincidence detection, we designed a stimulus protocol, which couples antidromic stimulation (electrical stimulus of the axon hillock; Fig. 7, black line) of the postsynaptic cell with light activation (Fig. 7, purple line) of NMDARs at dened intervals. The electric stimuli were given at subthreshold intensities and did not induce AP ring by themselves. By varying the interval between the two stimuli, we showed that only coincident activation with light and current injection led to suprathreshold signals that generated APs. If the stimuli did not coincide or the delay between stimuli was too long, the number of spikes generated was signicantly reduced (Fig. 7d).

DiscussionTo circumvent the limitations of traditional irreversible caged agonists, we synthesized a novel photo-reversible glutamate receptor agonist, ATG. This compound has functional features that make it unique and useful to neurobiological research: It is inactive in the dark, selectively targets NMDARs, can be quickly activated by irradiation with 360375 nm light (or 725740 nm

light, 2P activation) and quickly deactivated with 405460 nm light. Using rapid laser illumination, we demonstrated the unique photoswitching between a cis-active and trans-inactive compound, which enabled activation and deactivation of NMDARs on the timescale of their intrinsic gating properties. We also demonstrated that ATG is amenable to combination with other optical techniques, namely Ca2 imaging. Finally, we showed that ATG could be used to mimic synaptic activation in coincidence detection protocols.

In the absence of detailed structural data and molecular dynamics calculations, it is difcult to explain why cis-ATG is the active form and why it is selective for NMDARs. Structure-activity relationship studies indicate that the pharmacological space available for ATG is relatively narrow, which makes it a challenge to develop red-shifted derivatives (D. Trauner, unpublished results). As the physiological activity of the stilbene analogue cis-STG indicates, the diazene unit (N N bond) is not

essential for ligand binding and is probably not engaged in hydrogen-bonding interactions to the ligand-binding domain. This is also apparent in a recent X-ray structure of GluAzo bound to the GluK2 ligand binding domain34. Attempts to red-shift the action spectrum of ATG by substituting it with strongly electron-donating substituents, such as a diethylamino group, in the 40-position have so far yielded inactive compounds, indicating that substituents in this position clash with the ligand-binding domain in its closed state. Other ways to red-shift, for example, with heterocyclic azobenzenes, can be imagined and are under active investigation. Finally, cis-ATG and cis-STG represent a new class of agonists (and potentially antagonists) for ionotropic glutamate receptors that can be rapidly assembled using click-chemistry.

Herein, we show that ATG photoswitching and light-dependent NMDAR activation can be achieved with a monochromator, LEDs or laser light source. In principle, simple LEDs or standard light sources used in uorescence imaging could also

8 NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076 ARTICLE

be used. It should be taken into account, however, that the kinetics of photoswitching are strongly dependent on the applied light intensities. ATG exhibits a high molar extinction coefcient (48,778 cm1M1 at 330 nm). Although such a feature enables efcient light absorption, it also reduces the light intensity at the focal point as a consequence of Lambert-Beers law23. We circumvented this limitation by performing local perfusion of ATG to limit the optical path length in which photoswitching occurred. Two-photon excitation also provides a solution to the problem of out-of-focus light absorption while in parallel improving depth penetration in scattering tissue and providing intrinsic localization, properties well-suited for activation of single spines with a PCL. Deactivation at longer wavelengths can still be performed by large 1P illumination areas, which is advantageous for switching any molecules that have diffused out of the diffraction-limited illumination volume. Although the 2P cross-section has not been estimated, the ability of brief laser illumination durations (o1 ms) to evoke synaptic-like NMDAR current amplitudes could render ATG a useful complement to MNI-glutamate. Very recently, MAG2p a derivative of glutamate that can be activated with 2P absorption, has been introduced35,36. This compound, however, is covalently attached to the receptor and not freely diffusible like ATG.

Because diffraction-limited illumination volumes are much larger than a single synaptic vesicle fusion event, the photolysis of caged neurotransmitter in diffraction-limited spots (by one- or two-photon excitation) is unlikely to mimic the rapid decay of neurotransmitter clearance15. PCL compounds, however, are ideally suited to overcome this diffusion-limited problem because photoisomerization of azobenzenes can occur on a picosecond timescale, sufcient for photo-inactivation of an agonist before it diffuses away. It remains to be determined if the switching speed is similar when cis-ATG is bound to the receptor. Nevertheless, we were able to demonstrate that ATG photoswitching can accelerate NMDAR current decays which, when using classical caged agonists, would have been limited by diffusional clearance. This requires that the inactive trans-ATG have an afnity low enough to be effectively a non-agonist. Consistent with the relative low-afnity of trans-ATG, the ATGoff decays are nearly thirty times faster than the decay of cis-ATG currents. Trans-ATG also appears to unbind NMDARs more rapidly than glutamate, as the half-decay is over twelve times faster than that achieved by glutamate uncaging (Fig. 3). We demonstrate, for the rst time, the fundamental ability to switch off channel agonism, thereby speeding the decay of NMDAR currents.

We were surprised that the ATGoff decay was not altered in GluN2A KO mice, especially since faster unbinding rates of glutamate are thought to underlie the faster deactivation rates of GluN2A-containing channels34. One possible interpretation of these results is that the afnity of trans-ATG is so low that photoconversion reveals a ligand-independent closing transition. To conrm this, future experiments would be required, especially to conrm rapid switching of cis-ATG when bound to the NMDAR.

Our results show that ATG is a powerful tool that permits precise temporal control of NMDAR gating not otherwise achieved with state-of-the-art caged compounds. They extend the reach of photopharmacology to an important subtype of glutamate receptors and demonstrate that photoswitchable neurotransmitters that are inactive in the dark can be synthesized and used to precisely control receptor activation in their native environment.

Methods

UVVis spectra. ATG was dissolved to a concentration of 50 mM in buffer containing (in mM) 138 NaCl, 1.5 KCl, 2.5 CaCl2, 1.2 MgCl2, 10 Glucose and 5

HEPES, adjusted to pH 7.4. UVVIS spectra were taken in a 100 ml cuvette with the switching light (monochromator) introduced through a glass bre from the top of the cuvette, perpendicular to the light path of the spectrometer (Varian, Cary 50). The kinetics of the trans- to cis-conversion were recorded at the maximal absorption wavelength of trans-ATG (330 nm). To achieve fast switching rates, we used high power LEDs at 365 and 460 nm (Prizmatix) for trans-cis and cis-trans isomerization, respectively.

Cortical slice preparation and external solutions. Cortical coronal slices were prepared from C57Bl6JRj mice (postnatal day 1015, both male and female animals were used without known experimenter bias). Following decapitation, the brain was rapidly removed and transferred to an ice-cold saline solution composed of (in mM) 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 0.5 CaCl2, 7 MgCl2, 25 glucose, 75 sucrose saturated with carbogen (95% O2/5% CO2). Slices (300 mm thick) were made using a Campden vibratome 7,000 smz-2 (NPI Electronic). Slices were incubated at 34 C for 30 min in ACSF composed of (in mM) 125 NaCl, 2.5 KCl,1.25 NaH2PO4, 26 NaHCO3, 2 CaCl2, 1 MgCl2, 20 glucose saturated with carbogen (95% O2 and 5% CO2). After incubation, slices were stored at room temperature from 30 min to ve hours before being recorded. Experiments were carried out at room temperature. Unless stated otherwise, ATG was added from a 200 mM dimethyl sulfoxide stock to the ACSF to yield a nal concentration of 200 mM. The

ACSF was heated to 40 C to improve the solubility of the ATG stock. The solution was not ltered because ATG adheres to lter materials.

For the identication of target receptors and calcium imaging experiments, the iGluR antagonists NBQX (25 mM) and D-AP-5 (40 mM), and the channel blockers

TTX (1 mM) and felodipine (40 mM) were bath-applied, whereas MK-801 (50 mM) (all from Abcam) was loaded into the patch pipette. NMDA (1 mM, Sigma-Aldrich) and cis-STG (200 mM) were puff-applied through a glass pipette using a pressure ejection system (PDES, NPI Electronic). For voltage-clamp recordings, TTX was added to the ACSF.

Patch clamp recordings of cortical layer 2/3 neurons. Pyramidal neurons were patched using re-polished glass electrodes with a resistance of 69 MO. Current-clamp recordings were carried out using the following intracellular solution (in mM): 140 K-gluconate, 10 HEPES, 12 KCl, 4 NaCl, 4 Mg-ATP,0.4 Na2-GTP. For whole-cell voltage-clamp recordings, we used (in mM) 110 Cs-gluconate, 15 NaCl, 10 HEPES, 5 TEA, 0.16 EGTA, 4 Mg-ATP, 0.4 Na2-GTP. Recordings were made with an EPC 10 USB amplier, controlled by the Patchmaster software (HEKA). Data was ltered at 2.910 kHz and digitized at 50 kHz. Holding potential was corrected for a 14 mV liquid junction potential. Cells were rejected if leak currents were 4200 pA or series resistance 425 MO.

Data was analysed using the Patchers Power Tools (MPI Gttingen) and routines written in IgorPro (Wavemetrics).

For antidromic stimulation, glass electrodes (5 MO) lled with ACSF were placed within 20 mm of the axon hillock and the stimulus pulse was applied through an isolated stimulation unit (A-M Systems). The stimulation intensity was set to be subthreshold. The temporal pattern of the antidromic and ATG light stimuli were controlled through the Patchmaster software (HEKA).

Hippocampal slice preparation and external solutions. Hippocampal coronal slices were prepared from C57Bl6JRj mice (Janvier Labs) (postnatal day 15-55, both male and female animals were used without known experimenter bias) following decapitation and rapid removal of the brain37. A Leica VT1200S vibratome was used to make 250 mm thick slices while the brain was immersed in an ice-cold saline solution composed of (in mM) 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3,0.5 CaCl2, 8 MgCl2, 25 glucose, 230 sucrose and 0.5 ascorbic acid saturated with 95% O2/5% CO2. After 30 min incubation at 33 C in solution composed of(in mM) 125 NaCl, 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 2 CaCl2, 1 MgCl2,25 glucose and 0.5 ascorbic acid saturated with 95% O2/5% CO2, slices were stored at room temperature from 30 min to 5 h in the same solution before being recorded. NMDAR currents were recorded in the presence of (in mM) 10 SR95531 (Abcam Biochemicals) to block GABARs, 0.3 strychnine (Sigma-Aldrich) to block glycine receptors, 5 2,3-Dioxo-6-nitro-1,2,3,4-tetrahydrobenzo[f]quinoxaline-7-sulfonamide (NBQX) (Abcam Biochemicals) to block AMPARs, and 1 TTX (Abcam Biochemicals) to minimize spontaneous activity. 50 mM D-serine (Sigma-Aldrich) was also included in the bath solution to saturate the co-agonist binding site of NMDAR. GABAR currents were isolated through the addition of (in mM) 0.3 strychnine, 5 NBQX, 10 D-(-)-2-Amino-5-phosphonopentanoic acid (Abcam Biochemicals) and 20 7-chlorokynurenic acid (Abcam Biochemicals). ATG or 4-methoxy-7-nitroindolinyl-caged L-glutamate (MNI-glutamate; Tocris Bioscience) was perfused locally using a 36 mm tip-diameter patch pipette. The

ATG and MNI-glutamate perfusion solutions contained (in mM) 110 NaCl,2.5 KCl, 2 NaHCO3, 1.25 NaH2PO4, 30 HEPES, 10 Glucose, 2 CaCl2, 1 MgCl2,0.05 Alexa Fluor 488 (Life Technologies), 0.01 SR95531, 0.0003 strychnine,0.005 NBQX, 0.001 TTX, 0.05 D-serine, and where noted, 0.050.1 D, L-threo-b-Benzyloxyaspartic acid (Tocris Bioscience, Bristol, UK). Alexa Fluor 488 was used to visualize the perfusion and ensure its regularity over the course of the experiment. The pH of the nal perfusion solution was adjusted to 7.3 after dilution. As ATG is not readily soluble in water, a stock solution (10 mM) was prepared in 0.1N NaOH.

NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076

Patch clamp recordings in hippocampal CA1 pyramidal neurons. Whole-cell voltage-clamp was performed from visually identied hippocampal CA1 pyramidal cells at 32 C using an Axopatch 700B, Molecular Devices. Fire-polished patch electrodes had a tip resistance of 4-6 MO, and contained a Cs-methanesulfonate-based internal solution composed of (in mM) 105 CsCH3O3S, 10 EGTA, 3 CaCl2,4 MgCl2, 10 Hepes, 4 NaCl, 4 NaATP, 0.4 NaGTP and 5 phosphocreatine for recording NMDAR currents. To record GABAAR currents, we used a similar solution except that we did not use NaCl and we added (in mM) 70 CsCl and 35 CsCH3O3S. 40-100 mM Alexa Fluor 594 was also added to the solution to visualize dendritic morphology during whole-cell experiments for all experiments except GABAAR recordings. Extracellular stimulation of EPSCs and IPSCs was performed at 0.1 Hz using a constant voltage stimulator (2060 V, 50 ms, Digitimer) and a4 MO resistance pipette. We recorded IPSCs for 10 min in ACSF, then 20 min in ACSF plus 400 mM ATG. Currents were ltered at 410 kHz then digitized at 100 kHz (NI PCI-6052E, National Instruments) using the software Neuromatic (www.neuromatic.thinkrandom.com/). Ofine, traces were ltered between410 kHz. Holding potentials were corrected for a 7 mV liquid junction potential

(JPCalcW), and then set at 30 mV, except for GABAAR currents which were

recorded at 70 mV.

Photoswitching of ATG and calcium imaging. A Poly-V monochromator(FEI Systems) controlled through the Patchmaster software was used to toggle between trans- and cis-ATG at varying wavelengths (370 and 420 nm unless otherwise indicated). During calcium imaging experiments, the monochromator was controlled by the Live Analysis software (FEI). The calcium indicator Quest-Fluo-8-AM (50 mM MoBiTec) was added to the ACSF, and acute hippocampal slices were incubated for at least 20 min at 37 C to facilitate uptake of the indicator. Calcium changes were recorded at 480 nm, digitized at 10 Hz, background corrected and the DF/F0 ratio was calculated using IgorPro routines and ImageJ (NIH).

Laser-mediated photoswitching and uncaging. Visually guided patch experiments were performed using differential interference contrast or Dodt contrast. Dendrites were visualized using either a confocal (592 nm excitation of Alexa Fluor 594) or a 2P imaging system (810 nm, Ultima, Prairie Technologies). We coupled a 405 nm diode laser (Model PhoxX 405-120, Omicron Laserage) and a 375 nm diode laser (Model PhoxX 375, Omicron Laserage) into the photoactivation galvanometers of the scanhead (for rapid spot positioning) using single-mode optical bres (Part QPMJ-A3S,A3S-400-3/125-3-2-1, Oz Optics). For one-photon photoswitching, we used 150 mW (bath application) or 375 mW (local application)

of 375 nm and 10.5 mW of 405 nm, measured after the objective. For two-photon photoactivation, a pulsed Ti:Sapphire laser (Chameleon Ultra II, Coherent) tuned to either 725 nm or 740 nm was directed into a photoactivation path. Photo-activation was performed on proximal dendrites within 100 mm of the soma to minimize possible effects of dendritic ltering. Illumination spots were typically placed o0.5 mm from the tip of the spine head, except when longer duration pulses were used (4500 ms), in which case the spot was placed 1 mm away.

In vitro transcription and preparation of cRNA. For expression in Xenopus laevis oocytes, clones (derived from Rattus norvegicus) of GluN1-1a (genebank accession number: U08261), GluN2A (AF001423), GluN2B (U11419.1), GluN2C (U08259.1) and GluN2D (U08260.1), each in the X. laevis oocyte expression vector pSGEM, were transcribed to cRNA in vitro using the T7 mMESSAGE mMACHINE Kit (Ambion) according to the protocol provided. Transcribed cRNA was isolated with a spin-column kit (Clean & Concentrator 25, Zymo) and cRNA integrity was checked via denaturing agarose gel electrophoresis. RNA concentration was determined photometrically with a NanoPhotometer (Implen) and the concentrations of all samples were adjusted to 200 ng ml 1with nuclease-free water.

Expression and two-electrode recordings in Xenopus oocytes. Frog oocytes were surgically removed from the ovaries of X. laevis (Nasco) anesthetized with ethyl 3-aminobenzoate methanesulfonate (2.3 g l 1; Sigma). The lumps of oocytes were incubated with 300 U ml 1 (10 mg ml 1) collagenase type I (Worthington

Biochemicals) for 3 h at 21 C in Ca2 -free Barths solution (in mM) 88 NaCl,1.1 KCl, 2.4 NaHCO3, 0.8 MgSO4, 15 HEPES, pH adjusted to 7.6 with NaOH) with slow agitation to remove the follicular cell layer, and then washed extensively with Barths solution (in mM) 88 NaCl, 1.1 KCl, 2.4 NaHCO3, 0.8 MgSO4, 0.4 CaCl2,0.3 Ca(NO)3 15 HEPES, pH adjusted to 7.6 with NaOH). Oocytes were maintained in Barths solution supplemented with 100 mg ml 1 gentamycin, 40 mg ml 1 streptomycin and 63 mg ml 1 penicillin. Intact oocytes of stages V or VI were selected and cRNA was injected with a Nanoliter 2010 injector (WPI) within 8 h after surgery. For expression of GluN1/GluN2 heteromers, 20 nl (4 ng) of cRNA for each subunit were injected. Electrophysiological recordings were carried out5 days after injection. Two-electrode voltage clamping was performed using a TurboTec-10CX amplier (npi electronic) controlled by Pulse software (HEKA). For photoswitching experiments, the recording chamber was illuminated using LEDs (365 and 460 nm, Prizmatix) coupled to a light guide, which was placed directly above the oocyte. LEDs were controlled via the TTL outputs of the

ADC/DAC (ITC-16, Instrutech). Borosilicate glass capillaries (Harvard Instruments) were pulled to resistances of 0.11 MO with a vertical puller (PIP5, HEKA) and lled with 3 M KCl.

Oocytes were clamped at 70 mV. All recordings were performed in

continuously superfused with Barium Ringer (BaR, in mM) 115 NaCl, 2.5 KCl,1.8 BaCl2, 10 HEPES-NaOH, pH 7.2). When a stable holding current was attained, the recording protocol was started. To further prevent opening of calcium-induced chloride channels, niumic acid (NFA, 250 mM) was added to the BaR. All agonist solutions contained 10 mM glycine. In addition, the BaR used for recording photo-currents was supplemented with ATG (200 mM). For recording of cis-ATG mediated currents, the microscope light was switched-off and the photoswitching recording sequence was started. A 5 s pulse of blue light (460 nm) was followed by 5 s of UV light (365 nm) and, again, 5 s of blue light. These protocols were carried out rst in the absence of ATG to control for possible artifacts, and then repeated in the presence of 200 mM ATG. For the recording of NMDA-induced currents, 1 mM NMDA in BaR supplemented with 10 mM glycine was perfused. After steady-state currents were achieved, the NMDA was washed out until a stable baseline was reached. The application of 1 mM NMDA was repeated until two consecutive applications resulted in similar steady-state amplitudes. For analysis, light-induced ATG-independent currents were subtracted from cis-ATG-mediated currents. The steady-state amplitude of the corrected cis-ATG-mediated currents were then normalized to steady-state amplitude of NMDA-induced currents.

Data analysis. Data analysis was performed using the Neuromatic analysis package and custom routines within the IgorPro environment (Wavemetrics). Cells were rejected from analysis if the leak current was 4 100 pA at 30 mV holding

potential or the series resistance was over 15 MO, except for the GluN2A KO animals in which larger leak currents were accepted. Cis-ATG-mediated current peak amplitudes from one-photon photoswitching were measured over a 1 ms window around the peak. Error bars are presented as the means.e.m. unless otherwise indicated. Current decays were estimated from a t with a double exponential equation y A1exp{ (x x0)}/t1} A2exp{ (x x0}/t2}. The weighted decay time constant (tweighted) was calculated as A

A A

AA A

t2.

Isochronal amplitudes represent averages of the currents in a 200 ms window centred at the time point at which the on-spine response had reached 75% of its peak value.

To better estimate the NMDAR current decay following 405 nm illumination without contamination from partial cis-activation of ATG, we performed a subtraction protocol. 405 nm-induced currents were t to an empirical function that describes the rising phase and dual exponential decay38:

I t

A1 1 exp

t t0

trise

n A2exp

t t0

tdecay1

The ts of each individual cell were scaled by eye to maximally overlay the slow decay component of 405 nm only and 375/405 nm-induced currents. We justied such a scaling because the population traces overlapped (Fig. 4a,b). The scaled ts for each cell were then subtracted from the associated average current trace induced by 375 nm illumination followed by 405 nm illumination. Cells were only analysed if three or more recordings were performed under each illumination condition (375 nm only, 375 nm followed by 405 nm, 405 nm only).

GABAR spontaneous events (Supplementary Fig. 6) were detected using an event detection algorithm within the Neuromatic analysis software based on initial 10 pA threshold detection. Events were rened using two additional search features: (1) event onset detection using a 1.5 ms backward sliding window to ensure the point of threshold detection was 7 s.d. of background noise; and

(2) event peak detection using a 0.2 ms forward sliding window to ensure the event peaked within 2 ms of the time point of event threshold detection with a current greater than that of the point of event threshold by at least 2 s.d. of background

noise. Parameters were determined empirically, but were held constant for all analysis.

Peak amplitudes of IPSCs (Supplementary Fig. 6) were estimated from a 200 ms window centred at the time point of the peak value of the average of all events for a particular condition (number of events were typically between 50 and 300 per cell). Each synaptic current peak amplitude was corrected for series resistance error, as well as for a slight (B5%) decay in peak amplitude of IPSCs over the duration of the recording, which was not due to ATG activity (estimated from sham experiments, n 5 cells). The coefcient of variation of the GABAR IPSCs was calculated as

s s

1 A2

exp

t t0

tdecay2

p peak where s represents the s.d. over a 200 ms window.

Background variance was measured as the standard deviation over a 200 ms window immediately before the stimulation pulse.

Statistical analyses requiring multiple comparisons were rst examined with a nonparametric one-way analysis of variance (KruskalWallis) followed by nonparametric tests between specic values. We considered comparisons to be signicantly different if P-values were o0.05.

10 NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9076 ARTICLE

References

1. Paoletti, P., Bellone, C. & Zhou, Q. NMDA receptor subunit diversity: impact on receptor properties, synaptic plasticity and disease. Nat. Rev. Neurosci. 14, 383400 (2013).

2. Karakas, E. & Furukawa, H. Crystal structure of a heterotetrameric NMDA receptor ion channel. Science 344, 992997 (2014).

3. Lee, C. H. et al. NMDA receptor structures reveal subunit arrangement and pore architecture. Nature 511, 191197 (2014).

4. Tovar, K. R., McGinley, M. J. & Westbrook, G. L. Triheteromeric NMDA receptors at hippocampal synapses. J. Neurosci. 33, 91509160 (2013).

5. Lester, R. A., Clements, J. D., Westbrook, G. L. & Jahr, C. E. Channel kinetics determine the time course of NMDA receptor-mediated synaptic currents. Nature 346, 565567 (1990).

6. Banke, T. G. & Traynelis, S. F. Activation of NR1/NR2B NMDA receptors. Nat. Neurosci. 6, 144152 (2003).

7. Gee, K. R., Niu, L., Schaper, K., Jayaraman, V. & Hess, G. P. Synthesis and photochemistry of a photolabile precursor of N-methyl-D-aspartate (NMDA) that is photolyzed in the microsecond time region and is suitable for chemical kinetic investigations of the NMDA receptor. Biochemistry 38, 31403147 (1999).

8. Maier, W., Corrie, J. E., Papageorgiou, G., Laube, B. & Grewer, C. Comparative analysis of inhibitory effects of caged ligands for the NMDA receptor.J. Neurosci. Methods 142, 19 (2005).9. Rodriguez-Moreno, A. et al. Presynaptic induction and expression of timing-dependent long-term depression demonstrated by compartment-specic photorelease of a use-dependent NMDA receptor antagonist. J. Neurosci. 31, 85648569 (2011).

10. Kramer, R. H., Fortin, D. L. & Trauner, D. New photochemical tools for controlling neuronal activity. Curr. Opin. Neurobiol. 19, 544552 (2009).

11. Thompson, S. M. Flashy science: controlling neural function with light.J. Neurosci. 25, 1035810365 (2005).12. Matsuzaki, M. et al. Dendritic spine geometry is critical for AMPA receptor expression in hippocampal CA1 pyramidal neurons. Nat. Neurosci. 4, 10861092 (2001).

13. Carter, A. G. & Sabatini, B. L. State-dependent calcium signaling in dendritic spines of striatal medium spiny neurons. Neuron 44, 483493 (2004).

14. Bagal, A. A., Kao, J. P. Y., Tang, C. M. & Thompson, S. M. Long-term potentiation of exogenous glutamate responses at single dendritic spines. Proc. Natl Acad. Sci. USA 102, 1443414439 (2005).

15. DiGregorio, D. A., Rothman, J. S., Nielsen, T. A. & Silver, R. A. Desensitization properties of AMPA receptors at the cerebellar mossy ber granule cell synapse.J. Neurosci. 27, 83448357 (2007).16. Fehrentz, T., Schonberger, M. & Trauner, D. Optochemical genetics. Angew. Chem. Int. Ed. Engl. 50, 1215612182 (2011).

17. Volgraf, M. et al. Reversibly caged glutamate: a photochromic agonist of ionotropic glutamate receptors. J. Am. Chem. Soc. 129, 260261 (2007).

18. Stawski, P., Sumser, M. & Trauner, D. A photochromic agonist of AMPA receptors. Angew. Chem. Int. Ed. Engl. 51, 57485751 (2012).

19. Coan, E. J., Saywood, W. & Collingridge, G. L. MK-801 blocks NMDA receptor-mediated synaptic transmission and long term potentiation in rat hippocampal slices. Neurosci. Lett. 80, 111114 (1987).

20. Dingledine, R. B. K., Bowie, D. & Traynelis, S. F. The glutamate receptor ion channels. Pharmacol. Rev. 51, 761 (1999).

21. Fino, E., Araya, R., Peterka, D. S., Salierno, M., Etchenique, R. & Yuste, R. RuBi-glutamate: two-photon and visible-light photoactivation of neurons and dendritic spines. Front. Neural Circuits 3, 2 doi: 10.3389/neuro.04.002.2009 (2009).

22. Amatrudo, J. M. et al. Wavelength-selective one- and two-photon uncaging of GABA. ACS Chem. Neurosci. 5, 6470 (2014).

23. Trigo, F. F., Corrie, J. E. T. & Ogden, D. Laser photolysis of caged compounds at 405nm: photochemical advantages, localisation, phototoxicity and methods for calibration. J. Neurosci. Methods 180, 921 (2009).

24. Bidoret, C., Ayon, A., Barbour, B. & Casado, M. Presynaptic NR2A-containing NMDA receptors implement a high-pass lter synaptic plasticity rule. Proc. Natl Acad. Sci. USA 106, 1412614131 (2009).

25. Gray, J. A. et al. Distinct modes of AMPA receptor suppression at developing synapses by GluN2A and GluN2B: single-cell NMDA receptor subunit deletion in vivo. Neuron 71, 10851101 (2011).

26. Hansen, K. B., Ogden, K. K., Yuan, H. & Traynelis, S. F. Distinct functional and pharmacological properties of triheteromeric GluN1/GluN2A/GluN2B NMDA receptors. Neuron 81, 10841096 (2014).

27. Stroebel, D., Carvalho, S., Grand, T., Zhu, S. & Paoletti, P. Controlling NMDA receptor subunit composition using ectopic retention signals. J. Neurosci. 34, 1663016636 (2014).

28. Makara, J. K. & Magee, J. C. Variable dendritic integration in hippocampal CA3 pyramidal neurons. Neuron 80, 14381450 (2013).

29. Sobczyk, A. & Svoboda, K. Activity-dependent plasticity of the NMDA-receptor fractional Ca2 current. Neuron 53, 1724 (2007).

30. Busetto, G., Higley, M. J. & Sabatini, B. L. Developmental presence and disappearance of postsynaptically silent synapses on dendritic spines of rat layer 2/3 pyramidal neurons. J. Physiol. 586, 15191527 (2008).

31. Noguchi, J., Matsuzaki, M., Ellis-Davies, G. C. R. & Kasai, H. Spine-neck geometry determines NMDA receptor-dependent Ca2 signaling in dendrites.

Neuron 46, 609622 (2005).32. Nevian, T. Single spine Ca2 signals evoked by coincident EPSPs and

backpropagating action potentials in spiny stellate cells of layer 4 in the juvenile rat somatosensory barrel cortex. J. Neurosci. 24, 16891699 (2004).33. Reiter, A., Skerra, A., Trauner, D. & Schiefner, A. A photoswitchable neurotransmitter analogue bound to its receptor. Biochemistry 52, 89728974 (2013).

34. Erreger, K., Dravid, S. M., Banke, T. G., Wyllie, D. J. & Traynelis, S. F. Subunit-specic gating controls rat NR1/NR2A and NR1/NR2B NMDA channel kinetics and synaptic signalling proles. J. Physiol. 563, 345358 (2005).

35. Izquierdo-Serra, M. et al. Two-photon neuronal and astrocytic stimulation with azobenzene-based photoswitches. J. Am. Chem. Soc. 136, 86938701 (2014).

36. Carroll, E. C. et al. Two-photon brightness of azobenzene photoswitches designed for glutamate receptor optogenetics. Proc. Natl Acad. Sci. USA 112, E776E785 (2015).

37. Bischofberger, J., Engel, D., Li, L., Geiger, J. R. P. & Jonas, P. Patch-clamp recording from mossy ber terminals in hippocampal slices. Nat. Protoc. 1, 20752081 (2006).

38. Nielsen, T. A., DiGregorio, D. A. & Silver, R. A. Modulation of glutamate mobility reveals the mechanism underlying slow-rising AMPAR EPSCs and the diffusion coefcient in the synaptic cleft. Neuron. 42, 757771 (2004).

Acknowledgements

This study was supported by an ERC Advanced Grant (No. 268795 to D.T.), the Centre National de la Recherche Scientique through the Actions Thmatiques et Initiatives sur Programme, Fondation Fyssen, Fondation pour la Recherche Mdicale, Fdration pour la Recherche sur le Cerveau, Agence Nationale de la Recherche (ANR-07-NEUR-008 and ANR-2010-BLANC-1411), NeRF Equipment Mi-lourd 2010 and Ecole des Neurosciences de Paris to D.D. E.R. was supported by the Ecole des Neurosciences de Paris and the Fondation de la Recherche Mdicale (FDT20130928054). We thank Pierre Paoletti and his laboratory for providing us with the GluN2A KO animals. We thank Gal Moneron for helpful advice on optical design. We also thank Florian Rckerl and GM for their help in constructing and aligning the uncaging system used in the 2P photoactivation experiments.

Author contributions

V.F. and F.H. designed and synthesized the compounds. L.L. carried out electro-physiology for basic characterization of ATG and cis-STG, calcium imaging experiments and coincidence detection in collaboration with M.S. E.R. performed single dendrite ATG photoswitching and control synaptic experiments. E.R. and N.R. performed synaptic experiment on GluN2A KOs. N.R. performed two-photon activation experiments in collaboration with E.R. J.T. performed oocyte experiments in collaboration with M.S. and L.L. which were directed by M.H. and L.L. All authors contributed to writing the manuscript. D.T. and D.D. directed the study.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions

Web End =reprintsandpermissions/

How to cite this article: Laprell, L. et al. Optical control of NMDA receptors with a diffusible photoswitch. Nat. Commun. 6:8076 doi: 10.1038/ncomms9076 (2015).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

NATURE COMMUNICATIONS | 6:8076 | DOI: 10.1038/ncomms9076 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 11

Word count: 9542

Show less

Abstract

Translate

N-methyl-D-aspartate receptors (NMDARs) play a central role in synaptic plasticity, learning and memory, and are implicated in various neuronal disorders. We synthesized a diffusible photochromic glutamate analogue, azobenzene-triazole-glutamate (ATG), which is specific for NMDARs and functions as a photoswitchable agonist. ATG is inactive in its dark-adapted trans-isoform, but can be converted into its active cis-isoform using one-photon (near UV) or two-photon (740 nm) excitation. Irradiation with violet light photo-inactivates ATG within milliseconds, allowing agonist removal on the timescale of NMDAR deactivation. ATG is compatible with Ca²⁺ imaging and can be used to optically mimic synaptic coincidence detection protocols. Thus, ATG can be used like traditional caged glutamate compounds, but with the added advantages of NMDAR specificity, low antagonism of GABAR-mediated currents, and precise temporal control of agonist delivery.

Details

Title

Optical control of NMDA receptors with a diffusible photoswitch

Author

Laprell, Laura; Repak, Emilienne; Franckevicius, Vilius; Hartrampf, Felix; Terhag, Jan; Hollmann, Michael; Sumser, Martin; Rebola, Nelson; Digregorio, David A; Trauner, Dirk

Pages

8076

Publication year

2015

Publication date

Aug 2015

Publisher

Nature Publishing Group

e-ISSN

20411723

Source type

Scholarly Journal

Language of publication

English

DOI

https://doi.org/10.1038/ncomms9076

ProQuest document ID

1707848400

Optical control of NMDA receptors with a diffusible photoswitch

Jump to:

Full text

Abstract

Details

Suggested sources