Abstract
Norepinephrine (NE) and epinephrine (Epi), two key biogenic monoamine neurotransmitters, are involved in a wide range of physiological processes. However, their precise dynamics and regulation remain poorly characterized, in part due to limitations of available techniques for measuring these molecules in vivo. Here, we developed a family of GPCR Activation-Based NE/Epi (GRABNE) sensors with a 230% peak ΔF/F0 response to NE, good photostability, nanomolar-to-micromolar sensitivities, sub-second rapid kinetics, high specificity to NE vs. dopamine. Viral- or transgenic- mediated expression of GRABNE sensors were able to detect electrical-stimulation evoked NE release in the locus coeruleus (LC) of mouse brain slices, looming-evoked NE release in the midbrain of live zebrafish, as well as optogenetically and behaviorally triggered NE release in the LC and hypothalamus of freely moving mice. Thus, GRABNE sensors are a robust tool for rapid and specific monitoring of in vivo NE/Epi transmission in both physiological and pathological processes.
Author Contributions
Y. L conceived and supervised the project. J.F., M.J., H.Wang, A.D., and Z.W. performed experiments related to sensor development, optimization, and characterization in culture HEK cells, culture neurons and brain slices. Y.Z., P.Z. and J.J.Z designed and performed experiments using Sindbis virus in slices. C.Z., W.C., and J.D. designed and performed experiments on transgenic fish. J.L., J.Zhou, H.Wu, J.,Zou, S.A.H., G.C., and D.L. designed and performed experiments in behaving mice. All authors contributed to data interpretation and data analysis. Y. L and J.F. wrote the manuscript with input from M.J., J.L., and D.L. and help from other authors.
Acknowledgements
This work was supported by the National Basic Research Program of China (973 Program; grant 2015CB856402), the General Program of National Natural Science Foundation of China (project 31671118), the NIH BRAIN Initiative grant U01NS103558, the Junior Thousand Talents Program of China, the grants from the Peking-Tsinghua Center for Life Sciences, and the State Key Laboratory of Membrane Biology at Peking University School of Life Sciences to Y. L; the Key Research Program of Frontier Sciences (QYZDY-SSW-SMC028) of Chinese Academy of Sciences, and Shanghai Science and Technology Committee (18JC1410100) to J.D.; the NIH grants R01MH101377 and R21HD090563 and an Irma T. Hirschl Career Scientist Award to D.L.; and the Intramural Research Program of the NIH/NIEHS of the United States (1ZIAES103310) to G.C.
We thank Yi Rao for sharing the two-photon microscope and Xiaoguang Lei for the platform support of the Opera Phenix high-content screening system at PKU-CLS. We thank the Core Facilities at the School of Life Sciences, Peking University for technical assistance. We thank Bryan L. Roth and Nevin A. Lambert for sharing stable cell lines and plasmids. We thank Yue Sun, Sunlei Pan, Lun Yang, Haohong Li for inputs on sensors’ characterization and application. We thank Yanhua Huang, Liqun Luo and Mickey London for valuable feedback of the manuscript.
Introduction
Both norepinephrine (NE) and epinephrine (Epi) are key monoamine neurotransmitters in the central nervous systems and peripheral organs of vertebrate organisms. These transmitters play an important role in a plethora of physiological processes, allowing the organism to cope with its ever-changing internal and external environment. In the brain, NE is synthesized primarily in the locus coeruleus (LC), a small yet powerful nucleus located in the pons. Noradrenergic LC neurons project throughout the brain and exert a wide range of effects, including processing sensory information (Berridge and Waterhouse, 2003), regulating the sleep-wake/arousal state (Berridge et al., 2012), and mediating attentional function (Bast et al., 2018). Blocking noradrenergic transmission causes impaired cognition and arousal and is closely correlated with a variety of psychiatric conditions and neurodegenerative diseases, including stress (Chrousos, 2009), anxiety (Goddard et al., 2010), depression (Moret and Briley, 2011), attention-deficit hyperactivity disorder (ADHD) (Berridge and Spencer, 2016), and Parkinson’s disease (PD) (Espay et al., 2014). In the sympathetic nervous system, both NE and Epi play a role in regulating heart function (Brodde et al., 2001) and blood pressure (Zimmerman, 1981).
Despite their clear importance in a wide range of physiological processes, the spatial and temporal dynamics of NE and Epi in complex organs (e.g. the vertebrate brain) are poorly understood at the in vivo level due to limitations associated with current detection methods. Classic detection methods such as microdialysis-coupled biochemical analysis (Bito et al., 1966; Justice, 1993; Watson et al., 2006) have low temporal resolution, requiring a relatively long time (typically 5 min/collection) and complex sampling procedures, thereby limiting the ability to accurately measure the dynamics of noradrenergic activity in the physiological state (Chefer et al., 2009). Recent improvements in microdialysis—in particular, the introduction of the nano-LC-microdialysis method (Lee et al., 2008; Olive et al., 2000)—have significantly increased detection sensitivity; however, this approach is still limited by a relatively slow sampling rate (on the order of several minutes). On the other hand, electrochemical detection techniques based on measuring currents generated by the oxidation of NE/Epi (Bruns, 2004; Park et al., 2009; Robinson et al., 2008; Zhou and Misler, 1995) provide nanomolar sensitivity and millisecond temporal resolution; however, their inability to distinguish NE and Epi from other monoamine neurotransmitters— particularly dopamine (Robinson et al., 2003)—presents a significant physiological limitation with respect to measuring noradrenergic/adrenergic transmission both in ex vivo tissue preparations and in vivo. In addition, both microdialysis-based and electrochemical techniques are designed to detect volume-averaged NE/Epi levels in the extracellular fluid and therefore cannot provide cell type–specific or subcellular information.
Real-time imaging of NE dynamics would provide an ideal means to non-invasively track NE with high spatiotemporal resolution. A recent innovation in real-time imaging, the cell-based reporters known as CNiFERs (Muller et al., 2014), converts an extracellular NE signal into an intracellular calcium signal that can be measured using highly sensitive fluorescence imaging. However, CNiFERs require implantation of exogenous cells and can report only volume transmission of NE/Epi. By contrast, genetically encoded sensors, in theory, circumvent the above-mentioned limitations to provide fast, clear, non-invasive, cell type–specific reporting of NE/Epi dynamics. In practice, all genetically encoded NE sensors developed to date have poor signal-to-noise ratio and narrow dynamic range (e.g., a <10% change in FRET ratio under optimal conditions) (Nakanishi et al., 2006; Vilardaga et al., 2003; Wang et al., 2018b), thus limiting their applicability, particularly in in vivo applications.
To overcome these limitations, we developed a series of genetically encoded single-wavelength fluorescent GRABNE sensors with rapid kinetics, a ΔF/F0 dynamic range of ∽200%, and EGFP-comparable spectra, brightness, and photostability. Here, we showcase the wide applicability of our GRABNE sensors using a number of in vitro and in vivo preparations. In every application tested, the GRABNE sensors readily reported robust, chemical-specific NE signals. Thus, our GRABNE sensors provide a powerful imaging-based probe for measuring the cell-specific regulation of adrenergic/noradrenergic transmission under a wide range of physiological and pathological conditions.
Results
Development and characterization of GRABNE sensors
Inspired by the structure (Rasmussen et al., 2011a; Rasmussen et al., 2011b) and working mechanism (Chung et al., 2011; Manglik et al., 2015; Nygaard et al., 2013) of the β2 adrenergic G protein–coupled receptor (GPCR), we exploited the conformational change between the fifth and sixth transmembrane domains (TM5 and TM6, respectively) upon ligand binding to modulate the brightness of an attached fluorescent protein. Building upon the successful strategy of generating GPCR activation-based sensors for acetylcholine (GACh) (Jing et al., 2018) and dopamine (GRABDA) (Sun et al., 2018), we first systematically screened human adrenergic receptors as a possible scaffold. We inserted circular permutated EGFP (cpEGFP) into the third intracellular loop domain (ICL3) of three α-adrenergic receptors (α1DR, α2AR, and α2BR) and two β-adrenergic receptors (β2R and β3R) (Fig. 1A). Among these five constructs, we found that α2AR-cpEGFP had the best membrane trafficking, indicated by its high colocalization with membrane-targeted RFP (Fig. S1); we therefore selected this construct as the scaffold for further screening.
The length of the linker surrounding the cpEGFP moiety inserted in G-GECO (Zhao et al., 2011), GCaMP (Akerboom et al., 2012), GACh (Jing et al., 2018), and GRABDA (Sun et al., 2018) can affect the fluorescence response of cpEGFP-based indicators. Thus, as the next step, we systematically truncated the linker which starts with the entire flexible ICL3 of α2AR surrounding cpEGFP (Fig. 1B). We initially screened 275 linker-length variant proteins and identified a sensor (GRABNE0.5m) with a modest response to NE (Fig. 1B, right). From this scaffold, we performed a random mutation screening of seven amino acids (AAs) in close proximity to the cpEGFP moiety; two of these AAs are on the N-terminal side of cpEGFP, and the remaining five are on the C-terminal side of cpEGFP (Fig. 1C). From approximately 200 mutant versions of GRABNE0.5m, we found that GRABNE1m—which contains a glycine-to-threonine mutation at position C1—provided the best performance with respect to ΔF/F0 and brightness (Fig. 1C, middle and right).
Next, we expressed GRABNE1m in HEK293T cells and applied NE in a range of concentrations. NE induced a fluorescence change in GRABNE1m-expressing cells in a dose-dependent manner, with an EC50 of 0.93 μM and a maximum ΔF/F0 of approximately 230% in response to a saturating concentration of NE (100 μM) (Fig. 1D, middle and right). We also introduced mutations in α2AR in order to increase its sensitivity at detecting NE. We found that a single T6.34K point mutation (Ren et al., 1993)—which is close to the highly conserved E6.30 site—resulted in a 10-fold increase in sensitivity (EC50 ∽83 nM) to NE compared with GRABNE1m; this sensor, which we call GRABNE1h, has a maximum ΔF/F0 of ∽130% in response to 100 μM NE. As a control, we also generated GRABNEmut, which has the mutation S5.46A at the putative ligand-binding pocket and therefore is unable to bind NE (Fig. 1D); this control sensor has similar brightness and membrane trafficking (Fig. S1 and S2A), but does not respond to NE even at 100 μM (Fig. 1D, middle and right).
To examine whether our GRABNE sensors can capture the rapid dynamic properties of NE signaling, including its release, recycling, and degradation, we bathed GRABNE1h-expressing HEK293T cells in a solution containing NPEC-caged NE; a focused spot of 405-nm light was applied to locally uncage NE by photolysis (Fig. 2A). Transient photolysis induced a robust increase in fluorescence in GRABNE1h-expressing cells (mean on time constant 137ms, single exponential fit), which was blocked by application of the α2-adrenergic receptor antagonist yohimbine (Fig. 2B,C). To characterize both the on and off rates (τon and τoff, respectively) of the GRABNE sensors, we locally applied various compounds to GRABNE-expressing cells using rapid perfusion and measured the fluorescence response using high-speed line scanning (Fig. 2D,E). The average delay intrinsic to the perfusion system (measured by fitting the fluorescence increase in the co-applied red fluorescent dye Alexa 568) was 34 ms (Fig. 2F). Fitting the fluorescence change in each sensor with a single exponential function yielded an average τon of 72 and 36 ms for GRABNE1m and GRABNE1h, respectively, and an average τoff of 680 and 1890 ms for GRABNE1m and GRABNE1h, respectively (Fig. 2E,F). The faster on-rate and slower off-rate of GRABNE1h compared to GRABNE1m is consistent with its relatively higher affinity for NE.
High ligand specificity is an essential requirement for tools designed to detect structurally similar monoamine-based molecules. Importantly, our GRABNE sensors, which are based on α2AR, respond to both NE and Epi, but do not respond to other neurotransmitters (Fig. 2G). The sensors also respond to the α2AR agonist brimonidine but not the β2-adrenergic receptor agonist isoprenaline, which indicates receptor-subtype specificity. Moreover, the NE-induced fluorescence increase in GRABNE-expressing cells was blocked by the α-adrenergic receptor antagonist yohimbine, but not the β-adrenergic receptor antagonist ICI 118,551. Additionally, because NE and DA are structurally similar yet functionally distinct, we characterized how our GRABNE sensors respond to various concentrations of DA and NE. Wild-type α2AR has an 85-fold higher affinity for NE versus DA (Fig. 2H, right); in contrast, GRABNE1m has a 350-fold higher affinity for NE, whereas GRABNE1h was similar to the wild-type receptor, with a 37-fold higher affinity for NE (Fig. 2H). In contrast, fast-scan cyclic voltammetry (FSCV) was unable to differentiate between NE and DA, producing a nearly identical response to similar concentrations of NE and DA (Fig. 2I) (Robinson et al., 2003). To test the photostability of our NE sensors, we continuously illuminated GRABNE-expressing HEK293T cells using either 1-photon (confocal) or 2-photon laser microscopy and found that the GRABNE sensors are more photostable than EGFP under both conditions (Fig. S2C). Taken together, these data suggest that the GRABNE sensors can be used to measure the dynamic properties of noradrenergic activity with high specificity for NE over other neurotransmitters.
Next, we examined whether our GRABNE sensors can trigger GPCR-mediated downstream signaling pathways. First, we bathed GRABNE1m-expressing cells in a saturating concentration of NE for 2 h, but found no significant internalization of GRABNE1m (Fig. 2J). Similarly, we found that both GRABNE1m and GRABNE1h lack β-arrestin–mediated signaling, even at the highest concentration of NE tested (Fig. 2K), suggesting that the GRABNE sensors are not coupled to β-arrestin signaling. In addition, GRABNE1m and GRABNE1h had drastically reduced downstream Gi coupling compared to wild-type α2AR, which was measured using a Gi-coupling–dependent luciferase complementation assay (Fig. 2L) (Wan et al., 2018). We also found that G protein activation by GRABNE1m measured using the highly sensitive TGFα shedding was reduced by about 100-fold compared to the wild-type receptor (Fig. S2B) (Inoue et al., 2012). Finally, blocking G protein activation by treating cells with pertussis toxin (Fig. 2M) had no effect on the fluorescence response of either GRABNE1m or GRABNE1h, indicating that the fluorescence response of GRABNE sensors does not require G protein coupling (Rasmussen et al., 2011a). Taken together, these data indicate that GRABNE sensors can be used to report NE concentration without inadvertently engaging GPCR downstream signaling.
Characterization of GRABNE sensors in cultured neurons
The expression, trafficking, and response of proteins can differ considerably between neurons and cell lines (Marvin et al., 2013; Zou et al., 2014). Therefore, to characterize the performance of GRABNE sensors in neurons, we co-expressed GRABNE together with several neuronal markers in cultured cortical neurons. Both GRABNE1m and GRABNEmut trafficked to the cell membrane and co-localized with the membrane-targeted marker RFP-CAAX (Fig. 3A,B). Upon bath-application of a saturating concentration of NE, GRABNE1m and GRABNE1h had a peak ΔF/F0 of approximately 230% and 150%, respectively, whereas GRABNEmut had no response (Fig. 3D,E); these results are similar to our results obtained with HEK293T cells. Moreover, the NE-induced response in GRABNE1m-expressing cells was similar among various subcellular compartments identified by co-expressing GRABNE1m with either the axonal marker synaptophysin (SYP) or the dendritic marker PSD95 suggesting that GRABNE sensors enable the detection of NE throughout the neurons (Fig. 3C). Both GRABNE1m- and GRABNE1h-expressing neurons had a dose-dependent fluorescence increase in response to NE, with mean EC50 values of 1.9 μM and 93 nM, respectively (Fig. 3F). Consistent with high selectivity for NE, GRABNE1m and GRABNE1h have a 1000-fold and 7-fold higher affinity, respectively, for NE versus DA (Fig. 3F). Moreover, GRABNE1m responded specifically to NE and Epi, but did not respond to several other neurotransmitters and ligands, including the β2-adrenergic receptor agonist isoprenaline, histamine, dopamine, and serotonin (Fig. 3G). Finally, culturing GRABNE1m-expressing neurons in 100 μM NE for one hour did not cause internalization of the sensor, and the fluorescence increase was both stable for the entire hour and blocked completely by the α2-adrenergic receptor antagonist yohimbine (Fig. 3H,I). Thus, our GRABNE sensors have the necessary affinity and specificity to faithfully measure noradrenergic signaling in neurons.
Characterization of GRABNE sensors in both cultured and acute brain slices
To further test the GRABNE sensors in vitro, we expressed GRABNE1m and GRABNE1h in cultured hippocampal slices using a Sindbis virus expression system (Fig. S3A). In both GRABNE1m-expressing CA1 neurons and GRABNE1h-expressing CA1 neurons, exogenous application of NE in ACSF—but not ACSF alone—evoked a robust increase in fluorescence (Fig. S3B-D). In contrast, NE had no detectable effect on GRABNEmut-expressing neurons (Fig. S3C,D). Application of several α-adrenergic receptor agonists, including epinephrine and brimonidine, also generated a fluorescence increase in GRABNE1m-expressing neurons (Fig. S3C,F), consistent with data obtained using cultured cells. The rise and decay kinetics of the change in fluorescence were second-order, which reflects the integration of the time required to puff the drugs onto the cells and the sensor’s response kinetics (Fig. S3E,G). We also prepared acute hippocampal slices in which GRABNE1h was expressed using an adeno-associated virus (AAV); in this acute slice preparation, the GRABNE1h-expressing hippocampal neurons are innervated by noradrenergic fibers, which was confirmed by post-hoc staining using an antibody against dopamine beta hydroxylase (Fig. S3H,I). Application of electrical stimuli at 20 Hz for 1 s elicited a robust increase in GRABNE1h fluorescence, and this increase was blocked by the application of yohimbine (Fig. S3J). Consistent with our results obtained using cultured slices, exogenous application of various α-adrenergic receptor agonists, including NE, Epi, and brimonidine— but not the β-adrenergic receptor agonist isoprenaline—evoked a fluorescence increase in GRABNE1h-expressing neurons, and this response was blocked by yohimbine, but not by the β-adrenergic receptor antagonist ICI 118,551 (Fig. S3K).
Next, we examined whether our GRABNE sensors can be used to monitor the dynamics of endogenous NE. We expressed GRABNE1m in the locus coeruleus (LC), which contains the majority of adrenergic neurons within the brain (Fig. 4A). Two weeks after AAV injection, we prepared acute brain slices and observed GRABNE1m expression in the membrane of LC neurons using two-photon microscopy (Fig. 4A). We then used electrical stimuli to evoke the release of endogenous NE in the LC in the acute slices. Applying one or two stimuli did not produce a detectable fluorescence increase in GRABNE1m-expressing neurons; in contrast, applying 10 or more stimuli at 20 Hz caused a progressively stronger response (Fig. 4B). Application of the voltage-activated potassium channel blocker 4-aminopyridine, which increases Ca2+ influx during the action potential, significantly increased the fluorescence response, whereas application of Cd2+ to block calcium channels abolished the stimulation-induced fluorescence increase (Fig. 4C), consistent with presynaptic NE release being mediated by Ca2+ influx. We also performed line-scanning experiments in order to track the kinetics of NE release (Fig. 4D, left). A brief electrical stimulation induced a rapid fluorescence response with a mean τon and τoff of 37 ms and 600 ms, respectively (Fig. 4D, middle and right). Taken together, these data indicate that GRABNE1m can be used to monitor the release of endogenous NE in real time.
NE released into the synapse is recycled back into the presynaptic terminal by the norepinephrine transporter (NET). We therefore tested the sensitivity of GRABNE1m to NET blockade using desipramine. In the presence of desipramine, electrical stimuli caused a larger fluorescence response in GRABNE1m-expressing neurons compared to ACSF alone (Fig. 4E). Moreover, desipramine significantly slowed the τoff of the fluorescence signal, consistent with reduced reuptake of extracellular NE into the presynaptic terminal. To rule out the possibility that the change in the fluorescence response was caused by a change in synaptic modulation over time, we applied repetitive electrical stimuli at 5-min intervals to GRABNE1m-expressing neurons and found that the stimulation-evoked response was stable for up to 40 min (Fig. 4F). Finally, we examined the specificity of the stimulation-induced response. Compared with a robust response in control conditions, the α-adrenergic antagonist yohimbine blocked the response; moreover, no response was elicited in LC neurons expressing GRABNEmut or in LC neurons expressing a dopamine version of the sensor (GRABDA1m) (Fig. 4G). In contrast, cells expressing GRABDA1m responded robustly to the application of DA, and the GRABNE1m and GRABDA1m responses were abolished by yohimbine and the dopamine receptor antagonist haloperidol, respectively (Fig. 4H). Taken together, these data indicate that GRABNE1m is both sensitive and specific for detecting endogenous noradrenergic activity in LC neurons.
GRABNE1m detects both exogenous NE application and endogenous NE release in awake zebrafish
Zebrafish is both a genetically accessible vertebrate species and an optically transparent organism, thus serving as a suitable model for in vivo imaging. We generated the transgenic zebrafish line Tg(HuC:NE1m), which pan-neuronally expresses the GRABNE1m sensor. Pan-neuronal expression was confirmed by GRABNE1m fluorescence on the cell membrane of neurons throughout the brain (Fig. 5A). Bath application of 50 μM NE—but not DA at the same concentration—elicited a robust increase in fluorescence intensity that was blocked completely by the subsequent application of 50 μM yohimbine (Fig. 5B-D). In addition, a separate zebrafish line expressing GRABNEmut did not respond to NE (Fig. 5C,D). Taken together, these data indicate that GRABNE1m can be used to measure NE in an in vivo model.
Next, we investigated whether GRABNE1m can be used to measure the dynamics of endogenous noradrenergic activity induced by a visual looming stimulus, which triggers a robust escape response in zebrafish. We applied repetitive looming stimuli while using confocal imaging to measure the fluorescence of GRABNE1m-expressing neurites in the optic tectum (Fig. 5E). Each looming stimulus induced a time-locked increase in GRABNE1m fluorescence, which was blocked by bath application of yohimbine but was unaffected by the β-adrenergic receptor antagonist ICI 118,551 (Fig. 5F,G). In contrast, the same looming stimuli had no effect in animals expressing GRABNEmut (Fig. 5F,G). In addition, adding desipramine to block NE reuptake slowed the decay of the fluorescence signal (Fig. 5H). By sparse expression of GRABNE1m in individual neurons in zebrafish larvae via transient transfection, we were also able to record robust signals corresponding to NE release at single-cell resolution in response to repetitive looming stimuli (Fig. 5I-K), confirming that our GRABNE sensors can be used to sense NE release at a single-cell level with high spatiotemporal resolution.
GRABNE1m detects optogenetically evoked NE release in freely moving mice
Having demonstrated the proof-of-concept in a relatively simple in vivo vertebrate system, we next examined whether the GRABNE sensors can be used to monitor the noradrenergic activity in the mammalian brain by virally expressing GRABNE1m (non-Cre dependent) together with the optogenetic actuator C1V1 (Cre-dependent) in the LC of Th-Cre mice (Fig. 6A). Optogenetic stimulation of LC NE neurons using 561 nm laser pulses reliably evoked an increase in GRABNE1m fluorescence in fiber photometry recording of freely moving mice. Moreover, Intraperitoneal (i.p.) injection of desipramine produced a slow progressive increase in basal GRABNE1m fluorescence (consistent with an increase in extracellular NE levels) and caused an increase in the magnitude and decay time of the light-activated responses. I.p. injection of yohimbine abolished both the increase in basal GRABNE1m fluorescence and the light-evoked responses (Fig. 6B-D). In contrast, treating mice with either GBR 12909 (a selective blocker of dopamine transporters) or eticlopiride (a specific D2R antagonist) had no effect on the light-evoked responses in GRABNE1m fluorescence (Fig. 6C-E). To further test the selectivity of GRABNE1m between NE and dopamine, we co-expressed GRABNE1m and DIO-C1V1 both in the LC and in the substantia nigra pars compacta (SNc) of Th-Cre mice (Fig. 6F). In these mice, optogenetic stimulation of dopamine neurons in the SNc did not cause any changes in the GRABNE1m fluorescence in the SNc. In contrast, stimulating NE neurons in the LC produced a clear increase in GRABNE1m fluorescence (Fig. 6F, G). These results confirm that the increase of GRABNE1m fluorescence reflects the release of endogenous NE from noradrenergic neurons in the LC.
Using GRABNE1m to track endogenous NE dynamics in the mouse hypothalamus during freely moving behaviors
In the brain, the hypothalamus mediates a variety of innate behaviors essential for survival, including feeding, aggression, mating, parenting, and defense (Hashikawa et al., 2016; Sokolowski and Corbin, 2012; Yang and Shah, 2016). The hypothalamus receives extensive noradrenergic projections (Moore and Bloom, 1979; Schwarz and Luo, 2015; Schwarz et al., 2015) and expresses an abundance of α2-adrenergic receptors (Leibowitz, 1970; Leibowitz et al., 1982). Microdialysis studies found that the hypothalamus is among the brain regions that contains high level of NE during stress (McQuade and Stanford, 2000; Pacak et al., 1995; Shekhar et al., 2002; Tanaka, 1999). To better understand the dynamics of NE signaling in the hypothalamus under stress, we virally expressed hSyn-GRABNE1m in the lateral hypothalamus of C57BL/6 mice. Three weeks after virus injection, we performed fiber photometry recordings of GRABNE1m fluorescence during a variety of stressful and non-stressful behaviors in freely moving mice (Fig. 7).
During forced swim test and tail suspension test (both of which were stressful), we observed a significant increase in GRABNE1m fluorescence. During forced swim test, the fluorescence signal increased continuously regardless of the animal’s movements and started to decrease only after the animal was removed from the water (Fig. 7C1-E1). During the 60-s tail suspension test, the signal began to rise when the animal was first pursued by the experimenter’s hand, increased continuously while the animal was suspended by the tail, and decreased rapidly back to baseline levels when the animal was returned to its home cage (Fig. 7C2-E2). Additionally, when a human hand was placed in front of the animal, we observed a small and transient increase in GRABNE1m fluorescence (Fig. 7C3-E3). In contrast, the presence of a non-aggressive mouse of either the same or the opposite sex or close social interaction with the conspecific (7C4-E4, C5-E5) caused no significant change in GRABNE1m fluorescence. Lastly, neither sniffing nor eating a food attractant—in this case, peanut butter—had an effect on GRABNE1m fluorescence (Fig. 7C6-E6). These data provide evidence that noradrenergic activity in the lateral hypothalamus occurs primarily under stressful conditions.
Finally, to confirm that the GRABNE1m sensor indeed detects changes in NE concentration instead of other monoamine neurotransmitters, such as dopamine, we injected mice with a specific NET inhibitor atomoxetine (3 mg/kg i.p.) to inhibit the reuptake of NE. Although atomoxetine had no effect on the peak change in GRABNE1m fluorescence during the tail suspension test, it significantly slowed the return to baseline levels after each tail suspension (Fig. 7F1-I1); in contrast, treating mice with the α-adrenergic receptor antagonist yohimbine (2 mg/kg) both decreased the peak change in GRABNE1m fluorescence and significantly accelerated the return to baseline (Fig. 7F1-I1). Treating mice with either the selective DAT inhibitor GBR 12909 (10 mg/kg, i.p.) or the D2 receptor antagonist sulpiride (50 mg/kg, i.p.) had no effect on the peak change in GRABNE1m fluorescence or the time to return to baseline (Fig. 7F2-I2). In summary, these data demonstrate that our GRABNE sensors are suitable for monitoring endogenous noradrenergic activity in real time, with high spatiotemporal precision, during freely moving behavior in mammals.
Discussion
Here, we report the development and validation of GRABNE1m and GRABNE1h, two genetically encoded norepinephrine/epinephrine sensors that can be used both in vitro and in vivo to monitor noradrenergic activity with high temporal and spatial resolution, high ligand specificity, and cell type specificity. In mouse acute brain slices, our GRABNE sensors detected NE release from the LC in response to electrical stimulation. In zebrafish, the GRABNE sensors reported looming-induced NE release with single-cell resolution. In mice, the GRABNE sensors reported the time-locked release of NE in the LC triggered by optogenetic stimulation, as well as changes in hypothalamic NE levels during a variety of stress-related behaviors.
Compared to existing methods for detecting NE, our GRABNE sensors have several distinct advantages. First, NE has been difficult to distinguish from DA in vivo (e.g. by fast-scan cyclic voltammetry) (Park et al., 2009; Robinson et al., 2003), largely because of their structural similarities with only one hydroxyl group difference. Our GRABNE sensors have extremely high specificity for NE over other neurotransmitters and chemical modulators, including DA (Figs. 2H, 3F). GRABNE1m has a roughly 1000-fold higher affinity for NE over DA when expressed in neurons, even better than the 85-fold difference of the wild-type α2-adrenergic receptor. Thus, our GRABNE sensors provide new opportunities to probe the dynamics of noradrenergic activity with high specificity, which is particularly valuable when studying the many brain regions that receive overlapping dopaminergic and noradrenergic inputs. One thing to note is that GRABNE sensors are engineered from the α2a receptor, which may not be suitable for pharmacological investigation of α2a receptor related regulations.
Second, our GRABNE sensors have extremely high sensitivity for NE. Specifically, the EC50 for NE approaches sub-micromolar levels, with a 200%—or higher—increase in fluorescence intensity upon binding NE. By comparison, recently published FRET-based NE indicators produce a signal change of ≤10% under optimal conditions (Wang et al., 2018a; Wang et al., 2018b). Thus, GRABNE sensors have much improved characteristics to monitor endogenous in vivo NE dynamics. Third, GRABNE sensors have brightness and photostability properties that rival EGFP, which permits stable recordings across extended experimental sessions. Fourth, because they provide sub-second response kinetics and are genetically encoded, our GRABNE sensors can non-invasively report noradrenergic activity in vivo with single-cell resolution and a high recording rate (∽30 Hz). Finally, because the GRABNE sensors can traffic to various surface membranes, including the cell body, dendrites, and axons, and because they perform equally well in these membrane compartments, they can provide subcellular spatial resolution, which is essential for understanding compartmental NE signaling in vivo.
Ligand binding to endogenous GPCRs drives G-protein activation and receptor internalization. If present in GRABNE sensors, these responses could interfere with endogenous signaling fidelity and disrupt normal neuronal activity. To assess this risk, we characterized the downstream coupling of our GRABNE sensors with both G protein–independent and G protein–dependent pathways. Importantly, the introduction of the cpEGFP moiety in the GRABNE sensors resulted in non-detectable engagement of arrestin-mediated desensitization/internalization, which ensures more consistent surface expression of the sensor and that the GRABNE sensors do not inadvertently activate arrestin-dependent signaling. With respect to G protein–dependent signaling, we found that although physiological levels of NE robustly induce a change in GRABNE1m fluorescence, they do not engage downstream G protein signaling (Fig. 2J-M).
Noradrenergic projections throughout the brain originate almost exclusively from the LC, and NE release plays a role in a wide range of behaviors, including cognition and the regulation of arousal, attention, and alertness (Berridge and Waterhouse, 2003; Li et al., 2018; Schwarz et al., 2015). In this respect, it is interesting to note that our in vivo experiments revealed that GRABNE sensors can reliably report looming-evoked NE release in the optic tectum of live zebrafish. Moreover, our fiber photometry recordings of GRABNE sensors in the hypothalamus of freely behaving mice revealed specific changes in noradrenergic activity under stressful conditions (e.g., a tail lift or forced swimming), whereas non-stressful conditions such as feeding and social interaction did not appear to alter noradrenergic activity. These data are generally consistent with previous data obtained using microdialysis to measure NE (McQuade and Stanford, 2000; Pacak et al., 1995; Shekhar et al., 2002; Tanaka, 1999). Importantly, however, our approach yielded a more temporally precise measurement of noradrenergic activity with the promise of higher spatial and cell-type specificity.
NE circuits of the LC receive heterogeneous inputs from a broad range of brain regions and send heterogeneous outputs to many brain regions (Schwarz et al., 2015). Congruously, altered noradrenergic activity has been associated with a broad range of brain disorders and conditions, including ADHD, PD, depression, and anxiety (Marien et al., 2004). The complexity of these disorders may, in part, reflect the complexities of noradrenergic circuits and signals, which previous tools have been unable to fully dissect. Thus, understanding the regulation and impact of noradrenergic activity during complex behavior demands technological advances, such as the GRABNE sensors we present here. Deploying these in concert with other cell-specific tools for reporting (Jing et al., 2018; Patriarchi et al., 2018; Sun et al., 2018) and manipulating neurotransmitter levels (Fenno et al., 2011; Urban and Roth, 2015) should increase our understanding of the circuits and mechanisms that underlie brain functions in both health and diseases.
Experimental model and subject details
Primary cultures
Rat cortical neurons were prepared from postnatal day 0 (P0) Sprague-Dawley rat pups (both male and female, randomly selected; Beijing Vital River). In brief, cortical neurons were dissociated from dissected P0 rat brains in 0.25% Trypsin-EDTA (Gibco), plated on 12-mm glass coverslips coated with poly-D-lysine (Sigma-Aldrich), and cultured at 37°C in 5% CO2 in neurobasal medium (Gibco) containing 2% B-27 supplement, 1% GlutaMax, and 1% penicillin-streptomycin (Gibco).
Cell lines
HEK293T cells were obtained from ATCC (CRL-3216) and verified based on their morphology under the microscope and by their growth curve. Stable cell lines expressing the wild-type α2-adrenergic receptor or various GRABNE sensors were constructed by co-transfecting cells with the pPiggyBac plasmid carrying target genes with Tn5 transposase into a stable HEK293T-based cell line expressing chimeric Gαq/i and AP-TGFα (Inoue et al., 2012). Cells that stably expressed the target genes were selected by treating with 2 mg/ml Puromycin (Sigma) after reaching 100% confluence. The HTLA cells used for the TANGO Assay stably express a tTA-dependent luciferase reporter and a β-arrestin2-TEV fusion gene and were a gift from Bryan L. Roth (Kroeze et al., 2015). All cell lines were cultured at 37°C in 5% CO2 in DMEM (Gibco) supplemented with 10% (v/v) fetal bovine serum (Gibco) and 1% penicillin-streptomycin (Gibco).
Mice
All procedures regarding animals were approved by the respective Animal Care and Use Committees at Peking University, New York University, University of Southern California and the US National Institutes of Health, and were performed in compliance with the US National Institutes of Health guidelines for the care and use of laboratory animals. Wild-type Sprague-Dawley rat pups (P0) were used to prepare cultured cortical neurons. Wild-type C57BL/6 and Th-Cre mice (MMRRC_031029-UCD, obtained from MMRRC) were used to prepare the acute brain slices and for the in vivo mouse experiments. Experimental Th-Cre mice were produced by breeding Th-Cre hemizygous BAC transgenic mice with C57BL/6J mice. All animals were housed in the animal facility and were family-housed or pair-housed in a temperature-controlled room with a 12hr-12h light-dark cycle (10 pm to 10 am light) with food and water provided ad libidum. All in vivo mouse experiments were performed using 2-12-month-old mice of both sexes.
Zebrafish
The background strain for these experiments is the albino strain slc45a2b4. To generate transgenic zebrafish, Both the pTol2-HuC:GRABNE1m plasmid and Tol2 mRNA were co-injected into single-cell stage zebrafish eggs, and the founders of HuC:NE1m were screened. HuC:NEmut transgenic fish were generated as described above using the pTol2-HuC:GRABNEmut plasmid. Adult fish and larvae were maintained on a 14h-10h light-dark cycle at 28°C. All experimental larvae were raised to 6-8 days post-fertilization (dpf) in 10% Hank’s solution, which consisted of (in mM): 140 NaCl, 5.4 KCl, 0.25 Na2HPO4, 0.44 KH2PO4, 1.3 CaCl2, 1.0 MgSO4, and 4.2 NaHCO3 (pH 7.2). Larval zebrafish do not undergo sex differentiation prior to 1 month post-fertilization (Singleman and Holtzman, 2014).
Method details
Molecular cloning
The molecular clones used in this study were generated by Gibson Assembly using DNA fragments amplified using primers (Thermo Fisher Scientific) with 25-bp overlap. The Gibson Assembly cloning enzymes consisted of T5-exonuclease (New England Biolabs), Phusion DNA polymerase (Thermo Fisher Scientific), and Taq ligase (iCloning). Sanger sequencing was performed using the sequencing platform at the School of Life Sciences of Peking University in order to verify the sequence of all clones. All cDNAs encoding the candidate GRABNE sensors were cloned into the pDisplay vector (Invitrogen) with an upstream IgK leader sequence and a downstream IRES-mCherry-CAAX cassette to label the cell membrane. The cDNAs of select adrenergic receptor candidates were amplified from the human GPCR cDNA library (hORFeome database 8.1), and cpEGFP from GCaMP6s was inserted into the third intracellular loop (ICL3). The insertion sites for the GRABNE sensors were screened by truncating the ICL3 of the α2-adrenergic receptor at the 10-amino acid (AA) level, followed by fine-tuning at the 1-AA level. Coupling linkers were randomized by PCR amplification using randomized NNB codons in target sites. Other cDNAs used to express the GRABNE sensors in neurons were cloned into the pAAV vector using the human synapsin promoter (hSyn) or TRE promoter. pAAV-CAG-tTA was used to drive expression of the TRE promoter. The plasmids carrying compartmental markers were cloned by fusing EGFP-CAAX, RFP-CAAX (mScarlet), KDELR-EGFP, PSD95-RFP, and synaptophysin-RFP into the pDest vector. To characterize signaling downstream of the GRABNE sensors, we cloned the sensors and the wild-type α2-adrenergic receptor into the pTango and pPiggyBac vector, respectively. GRABNE1m-SmBit and α2AR-SmBit constructs were derived from β2AR-SmBit (Wan et al., 2018) using a BamHI site incorporated upstream of the GGSG linker. LgBit-mGsi was a gift from Nevin A. Lambert.
Expression of GRABNE sensors in cultured cells and in vivo
The GRABNE sensors were characterized in HEK293T cells and cultured rat cortical neurons, with the exception of the TANGO assay and TGFα shedding assay. HEK293T cells were passaged with Trypsin-EDTA (0.25%, phenol red; Gibco) and plated on 12-mm size 0 glass coverslips in 24-well plates and grown to ∽70% confluence for transfection. HEK293T cells were transfected by incubating cells with a mixture containing 1 μg of DNA and 3 μg of PEI for 6 h. Imaging was performed 24-48 h after transfection. Cells expressing GRABNE sensors for screening were plated on 96-well plates (PerkinElmer).
Cultured neurons were transfected using the calcium phosphate method at 7-9 DIV. In brief, the neurons were incubated for 2 h in a mixture containing 125 mM CaCl2, HBS (pH 7.04), and 1.5 μg DNAh. The DNA-Ca3(PO4)2 precipitate was then removed from the cells by washing twice with warm HBS (pH 6.80). Cells were imaged 48 h after transfection.
For in vivo expression, the mice were anesthetized by an i.p. injection of 2,2,2-tribromoethanol (Avetin, 500 mg/kg body weight, Sigma-Aldrich), and then placed in a stereotaxic frame for injection of AAVs using a Nanoliter 2000 Injector (WPI) or Nanoject II (Drummond Scientific) microsyringe pump. For the experiments shown in Figures 4 and 6, the AAVs containing hSyn-GRABNE1m/NE1mut/DA1m and Ef1a-DIO-C1V1-YFP were injected into the LC (AP: −5.45 mm relative to Bregma; ML: ±1.25 mm relative to Bregma; and DV: -2.25 mm from the brain surface) or SNc (AP: -3.1 mm relative to Bregma; ML: ±1.5 mm relative to Bregma; and DV: -3.8 mm from the brain surface) of wild-type or Th-Cre mice. For the experiments shown in Figure 7, 100 nl of AAV9-hSyn-GRABNE1m (Vigene, 1×1013 titer genomic copies per ml) were unilaterally into the hypothalamus (AP: -1.7 mm relative to Bregma; ML: 0.90 mm relative to Bregma; and DV: -6.05 mm from the brain surface) of wild-type (C57BL/6) mice at a rate of 10 nl/min.
Fluorescence imaging of HEK293T cells and cultured neurons
HEK293T cells and cultured neurons expressing GRABNE sensors were screened using an Opera Phenix high-content imaging system (PerkinElmer) and imaged using an inverted Ti-E A1 confocal microscope (Nikon). A 60x/1.15 NA water-immersion objective was mounted on the Opera Phenix and used to screen GRABNE sensors with a 488-nm laser and a 561-nm laser. A 525/50 nm and a 600/30 nm emission filter were used to collect the GFP and RFP signals, respectively. HEK293T cells expressing GRABNE sensors were first bathed in Tyrode’s solution and imaged before and after addition of the indicated drugs at the indicated concentrations. The change in fluorescence intensity of the GRABNE sensors was calculated using the change in the GFP/RFP ratio. For confocal microscopy, the microscope was equipped with a 40x/1.35 NA oil-immersion objective, a 488-nm laser, and a 561-nm laser. A 525/50 nm and a 595/50 nm emission filter were used to collect the GFP and RFP signals, respectively. GRABNE-expressing HEK293T cells and neurons were perfused with Tyrode’s solutions containing the drug of interest in the imaging chamber. The photostability of GRABNE sensors and EGFP was measured using a confocal microscope (for 1-photon illumination) equipped with a 488-nm laser at a power setting of ∽350 μW, and using a FV1000MPE 2-photon microscope (Olympus, 2-photon illumination) equipped with a 920-nm laser at a power setting of ∽27.5 mW. The illuminated region was the entire HEK293T cell expressing the target protein, with an area of ∽200 μm2. Photolysis of NPEC-caged-NE (Tocris) was performed by combining fast scanning with a 76-ms pulse of 405-nm laser illumination by a confocal microscope.
TANGO assay
NE at various concentrations (ranging from 0.1 nM to 100 μM) was applied to α2AR-expressing or NE1m-/NE1h-expressing HTLA cells (Kroeze et al., 2015). The cells were then cultured for 12 hours to allow expression of the luciferase gene. Furimazine (NanoLuc Luciferase Assay, Promega) was then applied to a final concentration of 5 μM, and luminescence was measured using a VICTOR X5 multilabel plate reader (PerkinElmer).
TGFα shedding assay
Stable cell lines expressing Gαi-AP-TGFα together with the wild-type α2AR or GRABNE sensors were plated in a 96-well plate and treated by the addition of 10 μl of a 10x solution of NE in each well, yielding a final NE concentration ranging from 0.1 nM to 100 μM. Absorbance at 405 nm was read using a VICTOR X5 multilabel plate reader (PerkinElmer). TGFα release was calculated as described previously (Inoue et al., 2012). Relative levels of G protein activation were calculated as the TGFα release of GRABNE sensors normalized to the release mediated by wild-type α2AR.
FSCV
Fast-scan cyclic voltammetry was performed using 7-μm carbon fiber microelectrodes. Voltammograms were measured with a triangular potential waveform from -0.4 V to +1.1 V at a scan rate of 400 V/s and a 100-ms interval. The carbon fiber microelectrode was held at -0.4 V between scans. Voltammograms measured in the presence of various different drugs in Tyrode’s solution were generated using the average of 200 scans followed by the subtraction of the average of 200 background scans. Currents were recorded using the Pinnacle tethered FSCV system (Pinnacle Technology). Pseudocolor plots were generated using Pinnacle FSCV software. The data were processed using Excel (Microsoft) and plotted using Origin Pro (OriginLab).
Luciferase complementation assay
The luciferase complementation assay was performed as previously described (Wan et al., 2018). In brief, ∽48h after transfection the cells were washed with PBS, harvested by trituration, and transferred to opaque 96-well plates containing diluted NE solutions. Furimazine (Nano-Glo; 1:1000; Promega) was added to each well immediately prior to performing the measurements with Nluc.
Fluorescence imaging of GRABNE in brain slices
Fluorescence imaging of acute brain slices was performed as previously described (Sun et al., 2018). In brief, the animals were anesthetized with Avertin, and acute brain slices containing the LC region or the hippocampus region were prepared in cold slicing buffer containing (in mM): 110 choline-Cl, 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 7 MgCl2, 25 glucose, and 2 CaCl2. Slices were allowed to recover at 35°C in oxygenated Ringers solution containing (in mM): 125 NaCl, 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 1.3 MgCl2, 25 glucose, and 2 CaCl2 for at least 40 minutes before experiments. An Olympus FV1000MPE two-photon microscope equipped with a 40x/0.80 NA water-immersion objective and a mode-locked Mai Tai Ti:Sapphire laser (Spectra-Physics) tuned to 920 nm were used for imaging the slices. For electrical stimulation, a concentric electrode (model #CBAEC75, FHC) was positioned near the LC region, and the imaging and stimuli were synchronized using an Arduino board controlled using a custom-written program. The imaging speed was set at 0.148 s/frame with 128 x 96 pixels in each frame. The stimulation voltage was set at ∽6 V, and the duration of each stimulation was typically 1 ms. Drugs were either delivered via the perfusion system or directly bath-applied in the imaging chamber.
For immunostaining of brain sections, GRABNE-expressing mice were anesthetized with Avetin, and the heart was perfused with 0.9% NaCl followed by 4% paraformaldehyde (PFA). The brain was then removed, placed in 4% PFA for 4 h, and then cryoprotected in 30% (w/v) sucrose for 24 h. The brain was embedded in tissue-freezing medium, and 50-µm thick coronal sections were cut using a Leica CM1900 cryostat (Leica, Germany). A chicken anti-GFP antibody (1:500, Abcam, #ab13970) was used to label GRABNE, and a rabbit anti-DBH antibody (1:50, Abcam, #ab209487) was used to label adrenergic terminals in the hippocampus. Alexa-488-conjugated goat-anti-chicken and Alexa-555-conjugated goat-anti-rabbit secondary antibodies were used as the secondary antibody, and the nuclei were counterstained with DAPI. The sections were imaged using a confocal microscope (Nikon).
Fluorescence imaging of zebrafish
Tg(HuC:GRAB-NE1m) zebrafish larvae were imaged by using an upright confocal microscope (Olympus FV1000, Japan) equipped with a 20x water-dipping objective (0.95 NA). The larvae were first paralyzed with α-bungarotoxin (100 μg/ml, Sigma), mounted dorsal side up in 1.5% low melting-point agarose (Sigma), and then perfused with an extracellular solution consisting of (in mM) 134 NaCl, 2.9 KCl, 4 CaCl2, 10 HEPES, and 10 glucose (290 mOsmol/L, pH 7.8). Images were acquired at 1-2 Hz with a view field of 800 × 800 pixels and a voxel size was 0.62 × 0.62 × 2.0 μm3 (x × y × z). To detect the sensor’s response to exogenous NE, 50 μM L-(-)-norepinephrine (+)-bitartrate salt monohydrate (Sigma) in 5 μM L-ascorbic acid and 50 μM yohimbine hydrochloride (TOCRIS) were sequentially applied to the bath. To detect endogenous NE release, visual looming stimuli, which mimic approaching objects or predators (Yao et al., 2016) were projected to the larvae under a red background. Each trial lasted 5 s, and 5 trials were performed in a block, with a 90-s interval between trials. To examine the specificity of responses, ICI 118,551 hydrochloride (50 μM, Sigma) and desipramine hydrochloride (50 μM, Sigma) were applied. Looming stimuli in transiently transfected HuC:GRABNE1m zebrafish were measured at single-cell resolution by using the same conditions described above.
Fiber photometry recording in freely moving mice during optical stimulation
In the all-optic experiments shown in Figure 6, multimode optical fiber probes (105/125 µm core/cladding) were implanted into the LC (AP: -5.45 mm relative to Bregma; ML: ±0.85 mm relative to Bregma; and DV: -3.5 mm from the brain surface) and the SNc (AP: -3.1 mm relative to Bregma; ML: ±1.5 mm relative to Bregma; and DV: -3.85 mm from the brain surface) in mice four weeks after viral injection. Fiber photometry recording in the LC and/or SNc was performed using a 473-nm laser with an output power of 25 µW measured at the end of the fiber. The measured emission spectra were fitted using a linear unmixing algorithm (https://www.niehs.nih.gov/research/atniehs/labs/ln/pi/iv/tools/index.cfm). The coefficients generated by the unmixing algorithm were used to represent the fluorescence intensities of various fluorophores (Meng et al., 2018). To evoke C1V1-mediated NE/DA release, pulse trains (10-ms pulses at 20 Hz for 1 s) were delivered to the LC/SNc using a 561-nm laser with an output power of 9.9 mW measured at the end of the fiber.
Fiber photometry recording in mice during behavioral testing
For the experiments in Figure 7, a fiber photometry recording set-up was generated and used as previously described (Falkner et al., 2016). GRABNE1m was injected into the lateral hypothalamus (Bregma AP: -1.7mm; ML: 0.90 mm DV: -4.80 mm) of C57BL/6 mice in a volume of 100 nl containing AAV9-hSyn-GRABNE1m (Vigene, 1×1013 titer genomic copies per ml) at 10 nl/min. A 400-µm optic fiber (Thorlabs, BFH48-400) housed in a ceramic ferrule (Thorlabs, SFLC440-10) was implanted 0.2 mm above the injection site. The virus was left to incubate for three weeks. Prior to fiber photometry recording, a ferrule sleeve was used to connect a matching optic fiber to the implanted fiber. For recordings, a 400Hz sinusoidal blue LED light (30 µW; M470F1 driven by an LEDD1B driver; both from Thorlabs) was bandpass-filtered (passing band: 472 ± 15 nm, Semrock, FF02-472/30-25) and delivered to the brain in order to excite GRABNE1m. The emission light passed through the same optic fiber, through a bandpass filter (passing band: 534 ± 25 nm, Semrock, FF01-535/50), and into a Femtowatt Silicon Photoreceiver, which recorded the GRABNE1m emission using an RZ5 real-time processor (Tucker-Davis Technologies). The 400-Hz signals from the photoreceiver were extracted in real time using a custom program (Tucker-Davis Technologies) and used to reflect the intensity of the GRABNE1m fluorescence signal.
Behavioral assays
All behavioral tests were performed at least one hour after the onset of the dark cycle. For the tail suspension test, each mouse was gripped by the tail and lifted off the bottom of its cage six times for 60 s each, with at least one minute between each lift. For the forced swim test, the mouse was gently placed in a 1000-ml conical flask containing lukewarm water and removed after 4-6 minutes. After removal from the water, the mouse was gently dried with paper towels and placed in the home cage on a heating pad. For conspecific assays, an adult C57BL/6 group-housed mouse of either sex was placed inside the test mouse’s cage for 10 minutes. No sexual behavior or aggressive behavior was observed during the interaction. For the food assay, ∽4g of peanut butter was placed in the cap of a 15-ml plastic tube and placed inside of the test mouse’s cage for 10 minutes. During that period, the test mouse was free to explore, sniff, and eat the peanut butter. All videos were acquired at 25 frames per second and manually annotated frame-by-frame using a custom MATLAB program (Lin et al., 2011). “Contact” with the social stimulus refers to the period in which the test mouse sniffed or was sniffed by the intruder. “Contact” with the peanut butter refers to the period in which the test mouse sniffed or ate the peanut butter. “Lift” refers to the period in which the experimenter gripped the mouse’s tail and lifted the mouse into the air.
Quantification and statistical analysis
For the imaging experiments using cultured HEK293T cells, primary neurons, and brain slices, images were first imported to ImageJ software (National Institutes of Health) for fluorescence intensity readouts, and then analyzed using MATLAB (MathWorks) with a custom-written script or Origin Pro (OriginLab). The fluorescence response traces in the brain slices shown in Figure 4 were processed with 3x binning and then plotted.
Time-lapse images of the zebrafish were analyzed using Fiji to acquire the fluorescence intensity in the region of interest (ROI) in each frame. A custom-written MATLAB program was then used to calculate the change in fluorescence intensity (ΔF/F0) as follows: ΔF/F0=(Ft-F0)/F0, where Ft was the fluorescence intensity at time t and F0 was the average fluorescence intensity during the entire time window. Statistical analyses were performed using GraphPad Prism 6 and Origin Pro (OriginLab).
For the fiber photometry data shown in Figure 7, the MATLAB function “msbackadj” with a moving window of 25% of the total recording duration was first applied to obtain the instantaneous baseline signal (Fbaseline). The instantaneous ΔF/F was calculated as (Fraw – Fbaseline)/Fbaseline, and a peri-stimulus histogram (PSTH) was calculated by aligning the ΔF/F signal of each trial to the onset of the behavior of interest. The response elicited during a behavior was calculated as the average ΔF/F during all trials of a given behavior. The response between behavioral periods was calculated as the average ΔF/F between two behavioral episodes excluding 4 s immediately before the behavior’s onset, as some uncontrolled and/or unintended events (e.g., chasing the animal before the tail suspension test) may have occurred during that period. The baseline signal was calculated as the average ΔF/F 100 s prior to the start of the behavioral test. The peak response after each drug injection was calculated as the average maximum ΔF/F during all tail suspension trials. The decay time was calculated as the average time required to reach half of the peak response.
Except where indicated otherwise, group differences were analyzed using the Student’s t-test, Wilcoxon matched-pairs signed rank test, Shapiro-Wilk normality test, one-way ANOVA test, or Friedman’s test. Except where indicated otherwise, all summary data are presented as the mean ± SEM.
Data and software availability
The custom MATLAB programs using in this study will be provided upon request to the corresponding author.
Footnotes
Declaration of Interests
The authors declare competing financial interests. J.F., M.J., H.Wang, and Y. L have filed patent applications whose value might be affected by this publication.
Author Contributions
Y. L conceived and supervised the project. J.F., M.J., H.Wang, A.D., and Z.W. performed experiments related to sensor development, optimization, and characterization in culture HEK cells, culture neurons and brain slices. Y.Z., P.Z. and J.J.Z designed and performed experiments using Sindbis virus in slices. C.Z., W.C., and J.D. designed and performed experiments on transgenic fish. J.L., J.Zhou, H.Wu, J.,Zou, S.A.H., G.C., and D.L. designed and performed experiments in behaving mice. All authors contributed to data interpretation and data analysis. Y. L and J.F. wrote the manuscript with input from M.J., J.L., and D.L. and help from other authors.
Author Contributions
Y. L conceived and supervised the project. J.F., M.J., H.Wang, A.D., and Z.W. performed experiments related to sensor development, optimization, and characterization in culture HEK cells, culture neurons and brain slices. Y.Z., P.Z. and J.J.Z designed and performed experiments using Sindbis virus in slices. C.Z., W.C., and J.D. designed and performed experiments on transgenic fish. J.L., J.Zhou, H.Wu, J.,Zou, S.A.H., G.C., and D.L. designed and performed experiments in behaving mice. All authors contributed to data interpretation and data analysis. Y. L and J.F. wrote the manuscript with input from M.J., J.L., and D.L. and help from other authors.