A fluorescent sensor for real-time measurement of extracellular oxytocin dynamics in the brain

Oxytocin (OT), a hypothalamic neuropeptide that acts as a neuromodulator in the brain, orchestrates a variety of animal behaviors. However, the relationship between brain OT dynamics and complex animal behaviors remains largely elusive, partly because of the lack of a suitable technique for its real-time recording in vivo. Here, we describe MTRIAOT, a G-protein-coupled receptor-based green fluorescent OT sensor that has a large dynamic range, suitable affinity, ligand specificity for OT orthologs, minimal effects on downstream signaling and long-term fluorescence stability. By combining viral gene delivery and fiber photometry-mediated fluorescence measurements, we demonstrate the utility of MTRIAOT for real-time detection of brain OT dynamics in living mice. MTRIAOT-mediated measurements indicate variability of OT dynamics depending on the behavioral context and physical condition of an animal. MTRIAOT will likely enable the analysis of OT dynamics in a variety of physiological and pathological processes.

O T, a neuropeptide produced by neurons in the paraventricular nucleus (PVN) and supraoptic nucleus of the hypothalamus, regulates many physiological processes in both peripheral tissues and the central nervous system. OT neurons primarily send axons to the posterior pituitary, where OT is released into the peripheral circulation as a peripheral hormone that promotes childbirth and lactation. In addition, axons of OT neurons project to numerous other brain regions, where they release OT to regulate diverse physiological functions, such as sensory processing, feeding control, social cognition and emotion 1 . Impairment of OT signaling in the brain may underlie cognitive and emotional dysfunction associated with neurodevelopmental disorders (for example, autism spectrum disorders and schizophrenia) and brain aging 2,3 . Given that the timescales of OT-related physiological and pathological phenomena are diverse (from seconds, minutes, hours and days to potentially much longer periods), the dynamics of OT in the brain may be variable depending on behavioral patterns, types of stimuli and physical conditions of animals.
In addition to its roles as an endogenous ligand, OT has emerged as a potential therapeutic agent for psychiatric disorders based on the finding that administration of exogenous OT enhances positive emotions in humans 4 . Therefore, OT administered peripherally, such as via intranasal or intravenous routes, was originally believed to reach the brain and exert therapeutic effects. However, the potency of exogenous OT administration is contraversial [5][6][7] , and a recent study in humans mostly invalidated initial observations of the effects of exogenous OT administration 8 . Therefore, whether OT administered via a peripheral route can efficiently reach the brain through the nose-brain pathway and/or blood-brain barrier remains unclear 5 .
In this context, techniques permitting detection of brain OT dynamics are needed. However, currently available methods, such as microdialysis [9][10][11] and a reporter gene-based assay (that is, iTango) 12 , have limitations, especially in terms of temporal resolution. Recently, fluorescent sensors consisting of a G-protein-coupled receptor (GPCR) with a fluorescent protein replacing the amino acids in its third intracellular loop (IL3), have been engineered as promising tools for real-time detection of neurotransmitters and neuromodulators such as dopamine, acetylcholine, norepinephrine and adenosine [13][14][15][16][17][18] . Inspired by these strategies, we developed a sensitive fluorescent OT sensor named MTRIA OT , which is composed of a medaka OT receptor (OTR) and a circularly permutated green fluorescent protein (cpGFP)-based fluorescent module named MTRIA (Modular fluorescence unit fused with TRansmembrane region-to-IntrAcellular loop linkers). Using fiber photometry-mediated fluorescence recording, we demonstrate that MTRIA OT can report a variety of OT dynamics in the mouse brain, including artificially evoked OT signals, endogenous OT responses during natural behaviors and altered endogenous OT dynamics by chemical and physical perturbations.

Results
Development of a sensitive fluorescent oxytocin sensor. We initially conducted a preliminary screen for an OTR that showed good targeting to the plasma membrane (PM), a key property for PM-targeted fluorescent sensors 15,19,20 . We chose medaka OTR (meOTR) as a scaffold for our fluorescent sensor because this receptor showed the best targeting to the PM among candidate OTRs from six vertebrate species (human, mouse, chicken, snake, frog and medaka) in human embryonic kidney 293T (HEK293T) cells (Fig. 1a-c and Extended Data Fig. 1a). After optimization of the insertion site of cpGFP in the IL3 of meOTR (Extended Data Fig. 1b-g and Supplementary Note 1), we screened the mutant sensors in HEK293T cells using the following three steps ( Fig. 1d and Supplementary Note 1). We sequentially performed mutagenesis in linkers in the N-terminal and C-terminal regions surrounding the cpGFP, neighboring regions of the linkers ranging from the transmembrane helix to the intracellular loops (TM-to-loop) and residues in cpGFP (Fig. 1e). Through these three screening steps, we obtained OT-1.0, OT-2.0 and OT-3.0, respectively. OT-3.0, the final product of the screening, had an approximately 720% ΔF/F 0 fluorescence response upon stimulation with 100 nM OT, which A fluorescent sensor for real-time measurement of extracellular oxytocin dynamics in the brain Daisuke Ino 1,2 ✉ , Yudai Tanaka 1,3 , Hiroshi Hibino 2 and Masaaki Nishiyama 1 Oxytocin (OT), a hypothalamic neuropeptide that acts as a neuromodulator in the brain, orchestrates a variety of animal behaviors. However, the relationship between brain OT dynamics and complex animal behaviors remains largely elusive, partly because of the lack of a suitable technique for its real-time recording in vivo. Here, we describe MTRIA OT , a G-protein-coupled receptor-based green fluorescent OT sensor that has a large dynamic range, suitable affinity, ligand specificity for OT orthologs, minimal effects on downstream signaling and long-term fluorescence stability. By combining viral gene delivery and fiber photometry-mediated fluorescence measurements, we demonstrate the utility of MTRIA OT for real-time detection of brain OT dynamics in living mice. MTRIA OT -mediated measurements indicate variability of OT dynamics depending on the behavioral context and physical condition of an animal. MTRIA OT will likely enable the analysis of OT dynamics in a variety of physiological and pathological processes.
was suppressed by the OTR antagonist L-368,899 (L-36) (Fig. 1f,g). We also generated an OT-insensitive sensor (OT-3.0-mut) that contained the Y206A mutation, which abolished its ligand-binding capacity 21 (Fig. 1f,g). Characterization of the dose-dependent fluorescence responses of OT-1.0 to OT-3.0 in HEK293T cells (Fig. 1h) showed that our screening improved the dynamic range of fluorescence responses (F max ) with little change to the half-maximal effective concentration (EC 50 ) value. We designated the fluorescent module of OT-3.0 as MTRIA (Extended Data Fig. 2) and renamed OT-3.0 as MTRIA OT .
We then further characterized the ligand specificity of MTRIA OT . Compared to its sensitivity to OT, the sensor was similarly sensitive to isotocin (an OT analog in fish), much less sensitive to vasopressin orthologs (vasopressin and vasotocin) and inotocin (an OT and vasopressin ortholog in insects), and insensitive to nematocin (an OT and vasopressin ortholog in nematodes; Extended Data Fig. 3a,b).   We also examined the basic properties of MTRIA OT , such as its coupling capacity with downstream effectors, long-term fluorescence stability, kinetics and pH sensitivity. We used a Ca 2+ imaging experiment and a split-luciferase complementation assay to assess the coupling of MTRIA OT with G αq protein and β-arrestin signaling pathways. MTRIA OT had no detectable effects on either of these effectors, whereas wild-type meOTR was able to couple with them (Extended Data Fig. 3c,d). Furthermore, we did not detect internalization of MTRIA OT from the PM when we chronically exposed MTRIA OT -expressing HEK293T cells to 100 nM OT (Extended Data Fig. 3e). We conducted a kinetic analysis of MTRIA OT using a local puff of OT and L-36, which yielded an on-rate and an off-rate of approximately 1.2 s and 26 s, respectively (Extended Data Fig. 3f). Finally, changes in the extracellular pH level ranging from 6.6 to 8.2 minimally affected basal and OT-induced MTRIA OT fluorescence (Extended Data Fig. 3g), indicating little pH dependence of MTRIA OT within the physiological extracellular pH range. We also examined the performance of MTRIA OT in cultured rat hippocampal neurons. When expressed in primary neurons, MTRIA OT showed good PM localization both in soma and neurites (Extended Data Fig. 4a). Application of OT induced a robust increase in the fluorescence intensity of MTRIA OT , and subsequent application of L-36 suppressed its fluorescence to a level comparable to basal fluorescence (Extended Data Fig. 4b,c). The EC 50 value of MTRIA OT measured in primary neurons was 20.2 nM (Extended Data Fig. 4d), almost the same value as that measured in HEK293T cells. Kinetic time constants measured in primary neurons were faster (on-rate of ~0.83 s, off-rate of ~9.5 s; Extended Data Fig. 4e,f) than those observed in HEK293T cells, similar to a previously reported GPCR-based fluorescent sensor 18 . Taken together, these results indicate that our MTRIA OT sensor is equipped with enough sensitivity, specificity and stability to accurately detect extracellular OT dynamics.

Validation of MTRIA OT in vivo.
Having validated the basic properties of MTRIA OT in cultured cells, we next tested whether our OT sensor is applicable to experiments in the brains of living mice. Using an adeno-associated virus (AAV), we expressed MTRIA OT in the anterior olfactory nucleus (AON), a major target site of OT in the brain with high OTR expression levels 22,23 . We then conducted fiber photometry recordings through an implanted cannula placed above the injection site ( Fig. 2a and Extended Data Fig. 5). First, we examined the responses of MTRIA OT after intracerebroventricular infusion of 10 µl saline containing various amounts of OT (0, 0.002, 0.02, 0.2, 2 and 20 µg) in anesthetized mice. When we serially applied increasing amounts of OT through a stainless-steel cannula implanted into the lateral ventricle (Fig. 2a), we observed robust fluorescence increases upon stimulation with OT at doses of 0.2 µg and above (Fig. 2b,c), comparable doses to those which are required to trigger OT-dependent animal behaviors [24][25][26] . This finding indicates that MTRIA OT is capable of real-time detection of extracellular OT in living brains.
Peripheral administration of less than 20 µg of OT is reportedly sufficient to affect animal behaviors 27,28 ; therefore, we next evaluated whether OT levels in the AON increased after application of exogenous OT from two distinct peripheral administration routes: intranasal and intraperitoneal. Neither intranasal nor intraperitoneal administration of 20 µg of OT induced detectable fluorescence responses of MTRIA OT in anesthetized mice (Fig. 2d,e). Taken together with the finding that MTRIA OT had robust responses to intracerebroventricular administration of OT solution at a dose of 0.2 µg (Fig. 2b,c), our data indicate that less than 1% of peripherally administered OT can reach the AON.
We next examined whether MTRIA OT can detect optically evoked OT release by combining fiber photometry recording and optogenetic stimulation. To this end, we virally expressed ChRmine-mSca, a fusion protein of red-shifted channelrhodopsin ChRmine 29 and mScarlet (mSca), in OT-expressing PVN neurons and MTRIA OT in AON cells (Fig. 2f). For control experiments, we expressed mSca instead of ChRmine-mSca in OT neurons. Our histological analysis verified that AAV vectors harboring the OT promoter (OTp) resulted in the expression of gene products specifically in OT neurons in the PVN (Fig. 2g). We measured MTRIA OT responses in the AON upon optogenetic stimulation of OT neurons with trains of red-light pulses (647-nm wavelength, 10-ms pulse width, 20-Hz frequency) lasting for 30 s, and parameters were comparable to those previously used to evoke optogenetic OT release 23,30,31 . We found that MTRIA OT signals in the AON of ChRmine-expressing mice were progressively enhanced with increasing power of the stimulation laser (from 0, 5 and 12.5 to 25 mW) whereas those in control mice showed no detectable changes upon stimulation at any laser intensity (Fig. 2h,i). To estimate the kinetic properties of MTRIA OT in vivo, we determined the time constants of OT signals induced with 25 mW of light, which were stable enough for a single-exponential fitting. The mean rise and decay times were 7.8 s and 14.6 s, respectively (Fig. 2j). Taken together, these observations indicate that MTRIA OT can be used to measure endogenous OT release in vivo with high sensitivity and rapid kinetics.
Monitoring oxytocin dynamics in the brains of freely behaving mice. Having shown that MTRIA OT is functional in the mouse brain, we next examined whether MTRIA OT can be used to assess endogenous OT dynamics in the brains of behaving mice. OT levels in cerebrospinal fluid reportedly show daily fluctuations 11,32 ; however, the precise temporal profile has not been fully determined, largely because of the limited sampling rate of conventional approaches. Therefore, we recorded MTRIA OT responses during daily behavior. We virally expressed MTRIA OT in the AON and measured its fluorescence responses using fiber photometry (Fig. 3a). We found that transient increases of OT signals were repeated at approximately 2-h intervals in the AON of mice freely behaving in a cage with food and water supplied ad libitum ( Fig. 3b-d). We named this ultradian OT rhythm 'OT oscillation' . OT oscillations were absent in simultaneously recorded reference signals (405-nm excited MTRIA OT ) and in signals recorded using control sensors (470-nm and 405-nm excited MTRIA OT -mut; Fig. 3b), excluding the involvement of artifacts derived from movements and/or autofluorescence. We selectively transduced tetanus toxin light chain (TeLC), which prevents vesicular transmitter release 33 , in OT-expressing neurons in the PVN using an AAV vector harboring the OT promoter (Extended Data Fig. 6). Expression of TeLC suppressed the increase in OT compared with the control (Fig. 3e-h), confirming that OT oscillation is dependent on OT release from oxytocinergic neurons in the PVN. These results demonstrate the utility of MTRIA OT to capture ultradian OT dynamics in the brain.
Taking advantage of the rapid kinetics of MTRIA OT , we also tested whether the sensor can detect faster phenomena. Interaction with a conspecific animal results in prompt activation of OT neurons 34 ; therefore, we examined social interaction-induced MTRIA OT responses in the AON using fiber photometry (Fig. 4a). Following an encounter with a conspecific mouse enclosed in a wire cage, the fluorescence signal of MTRIA OT started to gradually increase, which was characterized by a rise time constant of ~1 min. We also verified that there was no apparent increase in the signal when stimulated with a toy mouse ( Fig. 4b-d). These results indicate that MTRIA OT can be used to detect OT responses with a minute-order time constant. We next explored faster OT responses. OT levels are thought to increase in response to a wide variety of stressful stimuli 35 . A major target of stress-induced OT signals is the central nucleus of the amygdala 30,35 ; therefore, we virally expressed MTRIA OT around this region and measured OT responses after a tail lift as a rapid stress stimulus (Fig. 4e). After lifting the tail of a mouse, we observed a rapid rise in MTRIA OT signal. Such robust responses were absent in simultaneously recorded 405-nm-excited reference signals, as well as 470-nm-excited and 405-nm-excited signals recorded without stimulation, indicating artifacts are negligible (Fig. 4f,g). The mean time to reach peak signal value was approximately 6 s ( Fig. 4h), demonstrating the utility of MTRIA OT to detect OT responses that rapidly appear within a few seconds. Taken together, the above results show the potential of MTRIA OT for analyzing OT responses during various timescales.
Examination of potential factors affecting brain oxytocin dynamics. We next explored potential factors affecting patterns of OT dynamics in the brain. General anesthetics exert substantial effects on neurotransmission 36,37 ; therefore, we evaluated the impact of two anesthetics on OT levels in the AON using fiber photometry recording. First, we examined the effect of a mixture of dexmedetomidine, butorphanol and midazolam (mix-anes). When mix-anes was intraperitoneally administered, the fluorescence signal of MTRIA OT fell to a level below the baseline (Fig. 5a,b). The undershot signal of MTRIA OT was reversed to a level comparable to baseline after administration of atipamezole, a drug that has an antagonistic effect on mix-anes ( Fig. 5a,b). We also examined the effect of isoflurane, an inhalation anesthetic, on MTRIA OT responses by sequentially changing the dose from 1% to 4% to 0%. Consistent with the above result, the level of MTRIA OT fluorescence became lower as the dose of the anesthetic increased (Fig. 5c,d). We confirmed that mix-anes and isoflurane did not affect basal or OT-evoked fluorescent signals in HEK293T cells (Extended Data Fig. 7), excluding the possibility of direct suppressive effects of these anesthetics on the fluorescence signal of the sensor. Together, these results demonstrate the suppressive effect of anesthesia on brain OT levels.
Feeding is associated with OT release 38 ; therefore, we examined the impact of fasting stress on brain OT dynamics. After about half a day of food deprivation, patterns of OT oscillation gradually became disturbed and the oscillatory signals undershot to a level below the initial baseline (Fig. 5e,f). We named this phenomenon 'OT turbulence' (Fig. 5e). After refeeding, OT turbulence halted and normal f, representative traces of MtrIA Ot activities recorded from mice either expressing mSca or coexpressing teLC and mSca in the PVN. g,h, Summary of the peak number every hour (g) and peak z-score (h; n = 3 mice in mSca and n = 4 mice in teLC-P2A-mSca). Statistics used were an unpaired two-tailed t-test (P = 8.2 × 10 −4 in g and P = 1.6 × 10 −4 in h). Graphs represent the mean ± s.e.m. (c, g and h). ***P < 0.001, *P < 0.05, NS (c, g and h).
OT oscillation quickly recovered (Fig. 5e,f). Considering the role of OT as an appetite suppressant 39 , OT turbulence may manage hunger stress during starvation. Aging is associated with a decline in the neuroendocrine system 3,40 ; therefore, we next analyzed differences in AON OT dynamics in mice at varying ages (~2 months, ~6 months, ~1 year and ~2.5 years). We verified that AAV-mediated transduction of the sensor was successful in old animals (Extended Data Fig. 8a). Fiber photometrymediated fluorescence measurements revealed the occurrence of oscillatory OT responses in all groups, but the frequency of OT  is shaded to indicate the period of anesthesia after mix-anes administration (dark blue) and period after release by the antagonist, atipamezole (pink). b, Summary of minimum z-score values before mix-anes administration, after administration of mix-anes and following subsequent administration of atipamezole (n = 4 mice in before and mix-anes, n = 3 mice in atipamezole). c, Schematic of the recording and representative trace of MtrIA Ot activity showing the impact of isoflurane on Ot signal. the background of the trace is shaded to indicate the period of 1% isoflurane administration (gray), of 4% isoflurane administration (dark blue), and after release from the anesthesia (0%, pink). d, Summary of the tail values of z-scores during the above three states (n = 5 mice). e, representative trace of MtrIA Ot activity showing the impact of food deprivation on Ot oscillation. the background of the trace is shaded to indicate the period of food deprivation. the period of Ot turbulence is colored dark blue and the peaks of undershot signal are indicated by arrowheads. f, Summary of minimum z-score values before, during and after food deprivation (n = 3 mice). g, representative traces of MtrIA Ot fluorescence signals in mice at the indicated age (2 months, 6 months, 1 year or 2.5 years). h, Summary of the peak number every hour (top) and peak z-score (bottom; n = 3 mice). Graphs represent the mean ± s.e.m. (b, d, f and h). ***P < 0.001, *P < 0.05 and NS unless otherwise stated (b, d, f and h). Statistics (b, d, f and h) are summarized in Supplementary Note 3.
transients became slower in older animals, although there were no detectable changes in amplitude (Fig. 5g,h). It may be possible that the reduced frequency of OT oscillation results from the vulnerability of tissue in older animals to the surgery required for fiber photometry recording. However, odor-induced Ca 2+ responses in the AON were robust even in 1-year-old mice (Extended Data Fig. 8b), supporting the validity of our fiber photometry recordings in old mice. These results indicate that an altered frequency of OT oscillation may underlie aging-associated decline of brain function.
Developing other G-protein-coupled receptor-based sensors using MTRIA. MTRIA was able to successfully produce a sensitive optical readout of OTR activation; therefore, we examined whether MTRIA was also able to detect ligand-binding-induced conformational changes of other GPCRs (Fig. 6a). We cloned 184 receptors for 46 ligands derived from human, mouse, zebrafish or medaka and conjugated MTRIA to a region ranging from position 5.62 of TM5 to position 6.36 of TM6 in the receptors (Extended Data  Fig. 2a). To our surprise, almost 30% of the engineered sensors (54 of 184 proteins) showed a marked fluorescence increase (>50% ΔF/F 0 ) upon stimulation with high concentrations of their specific ligand (Fig. 6b). In Extended Data Fig. 9 and Supplementary Table 1, we named the 24 sensors that showed the largest fluorescence response among the sensors sharing the same ligand MTRIA sensors (for example, a sensor for dopamine was called MTRIA DA ). We hope that this approach, the MTRIA system, will help accelerate the engineering of various GPCR-based sensors.

Discussion
In this study, we described the development and application of a fluorescent sensor for OT, MTRIA OT . MTRIA OT -mediated in vivo fluorescence recording can report artificially evoked OT responses as well as endogenously controlled OT signals with a fast temporal resolution.
Although the impacts of exogenous OT administration on mood and emotions are expected to lead to treatments for psychiatric disorders, direct evidence for an increase in brain OT levels following peripheral OT administration remains limited. However, neither intranasal nor intraperitoneal administration of a relatively large amount had a substantial effect on OT levels in the AON. Further validation is needed to clarify whether this is the case for other brain regions.
Our in vivo measurements using MTRIA OT revealed OT oscillation, ultradian OT rhythms with intervals of approximately 2 h in the brains of adult mice. These type of OT dynamics have not been reported using conventional approaches, such as microdialysis, presumably because of the limited sampling rate and sensitivity. Future efforts are needed to clarify the regulatory mechanism and physiological significance of OT oscillation. Given that hormones associated with appetite and metabolism can stimulate OT release 38,41 , the rhythmic activities of brain OT signals may be autonomically controlled via interoceptive signals associated with feeding behaviors. Our observations of disturbance in the patterns of OT oscillation by food deprivation and aging support this idea.
An issue to be addressed in the future is the quantitative measurement of OT concentrations in the living brain. This will help to directly compare data acquired from different measurement sites, in different animals, or from different experimental setups. Development of next-generation OT sensors dependent on changes in quantitative fluorescence parameters (for example, fluorescence lifetime and efficiency of Förster resonance energy transfer) 42,43 will offer an opportunity for real-time monitoring of absolute OT quantities in the living brain.
Overall, our MTRIA OT -mediated measurements provide information about OT dynamics in the brain. Moreover, our analyses indicate that patterns of OT dynamics in the brain can be affected by anesthesia, food deprivation and aging. Therefore, it is important to carefully consider the experimental conditions and state of participants when interpreting OT measurement data in the brain. Indeed, context-dependent and participant-dependent inconsistencies regarding the effects of OT in human clinical trials 44 may be related to such parameters. The use of MTRIA OT will allow us to extend our knowledge of brain OT dynamics and its association with complex behaviors.

online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/ s41592-022-01597-x.

Methods
DNA constructs. DNA encoding GPCRs, a red fluorescent Ca 2+ sensor jRGECO1a 45 and PM-targeted mScarlet 46 (mSca mem ) were amplified by PCR and then subcloned into the pCIS vector, a CAG promoter-driven expression vector 47 . Template DNA for GPCRs was prepared using the following four approaches. First, cDNA libraries of a mouse (Mus musculus) brain, zebrafish (Danio rerio) brain and medaka (Oryzias latipes) brain were prepared using an RNeasy Kit (Qiagen) and QuantiTect Reverse Transcription Kit (Qiagen). Second, genomic DNA of human (Homo sapiens) and mouse were isolated from HEK293T cells and a mouse brain, respectively. Third, gene fragments of OTRs derived from chicken (Gallus gallus), snake (Notechis scutatus) and frog (Xenopus laevis) were synthesized by Integrated DNA Technologies. Finally, for cloning of human dopamine receptor DRD1, human muscarinic receptor CHRM3 and human adrenergic receptor ADRA2A, plasmids encoding dLight1.1, GACh2.0 and GRAB NE 1m were used as templates, respectively [14][15][16] . GPCRs used in Fig. 6 and the primer pairs for their cloning are listed in Supplementary Table 2. To assess localization of OTRs in HEK293T cells, a human influenza hemagglutinin (HA)-tag was added to the N-terminal regions of OTRs using PCR. For the construction of green fluorescent transmitter sensors, cpGFPs derived from GCaMP6s were inserted into the IL3 regions of GPCRs using either overlap extension PCR 48 or the Gibson assembly protocol 49 . For mutant sensor screening, site-directed mutagenesis was performed using primers containing mutated nucleotide(s). In particular, randomized site-directed mutagenesis was achieved using primers containing randomized codons (NNN). For the construction of plasmid DNA in the β-arrestin recruitment assay, fragments of NanoLuc-SmBit and LgBit-were amplified from pNL1.1 (Promega). Subsequently, they were fused with the C terminus of the OTR and the N terminus of rat β-arrestin 50  Adeno-associated virus production. HEK293T cells were simultaneously transfected with pHelper, XR8 and a transfer plasmid using PEI MAX transfection reagent, except for production of a viral vector for jGCaMP8s where XR9 was used instead of XR8. After overnight incubation, the medium was changed. Supernatant was collected at 48-h and 96-h time points, after medium exchange. Viral particles were purified using a polyethylene glycol-mediated precipitation method 51 . Purified viral particles were then concentrated using an ultrafiltration membrane unit (Amicon Ultra; EMD Millipore). Virus titers were determined by quantitative PCR using a pair of primers for the WPRE sequence (forward, actgtgtttgctgacgcaac; reverse, agcgaaagtcccggaaag). Viral vectors with titers > 10 12 gc per ml were used for fiber photometry measurements.

Immunostaining of transfected cells.
For immunostaining analysis of the cellular localization of OTRs, HEK293T cells co-transfected with an HA-tagged OTR and mSca mem (10:1 ratio) were fixed in PBS (FUJIFILM Wako) containing 4% (wt/vol) paraformaldehyde for 10 min at room temperature. Next, cells were permeabilized and blocked for 30 min at room temperature in blocking solution (PBS containing 0.2% TritonX-100 and 5% normal goat serum (FUJIFILM Wako)) and then incubated with an anti-HA antibody (rabbit; MBL; 1:2,000 dilution) for 30 min at room temperature. After washing with PBS containing 0.1% TritonX-100 (PBST), samples were incubated with Alexa 488-conjugated anti-rabbit IgG antibody (goat; Jackson ImmunoResearch) for 30 min at room temperature. After three washes with PBST, stained cells in PBS were analyzed by microscopy.  Fig. 2g and Extended Data Figs. 3f,g, 4 and 6-8 were obtained using SpinSR10, a spinning-disk confocal microscope system (Olympus) equipped with lasers (at 405, 488, 561 and 637 nm; Coherent), a scanner unit (CSU-W1 SoRa; Yokogawa) and a complementary metal oxide semiconductor camera (Orca Flash 4.0, Hamamatsu Photonics). The system was controlled by cellSens software (Olympus). An objective lens (UPlanSApo ×10, NA 0.40; UPlanSApo ×20, NA 0.75; or UPlanSApo ×30, NA 1.05 Sil; Olympus) was used. For live-imaging experiments, data were acquired either every 100 ms (for kinetic analysis in Extended Data Figs. 3f and 4e,f) or every 5 s (for other experiments). In these experiments, laser power was 200 µW or less at the front aperture of the objective lens, and no apparent photobleaching was observed with these conditions. Therefore, we did not conduct photobleaching correction. To analyze the cellular localization of fluorescent sensors, z-stack images were acquired at 0.5-µm steps. For the screening of fluorescent sensors, transfected cells seeded onto glass-bottomed chamber slides were soaked in 200 µl of artificial extracellular solution (ECS; 150 mM NaCl, 4 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES and 5.6 mM glucose; pH 7.4). Cells were then stimulated with 200 µl of drug-containing ECS, in which the drug concentration was twice the final concentration, administered through a pipette. To analyze the ligand dose-response curves of fluorescent sensors, time constants of fluorescent responses, intracellular Ca 2+ responses, pH sensitivity and long-term signal stability, transfected cells were seeded onto a collagen-coated glass-bottomed dish and continuously perfused with ECS. Cells were then rapidly stimulated with drug-containing solution using a custom-made perfusion system equipped with solenoid valves. For the measurement of kinetic time constants, the tip of the perfusion system was placed as close to the imaged cells as possible. Data were processed using Fiji software 52 and Excel (Microsoft), and fitting of fluorescence traces was conducted using Python modules. After conducting background subtraction, fluorescence signal (F) on a single cell was obtained. Signals were normalized to ΔF/F 0 defined by (F − F 0 )/F 0 where F 0 is the fluorescence value just before stimulation. Fluorescence traces were not smoothed. Values of ΔF/F 0 in Figs. 1e,h and 6b, and Extended Data Figs. 3b,g, 4c,d, 7b,d and 9 were quantified as the tail values at the end of drug stimulation. Values of peak ΔF/F 0 in Fig. 1g and Extended Data Fig. 3c were quantified as the maximum values during drug stimulation. Parameters of dose-response curves (F max and EC 50 in Fig. 1h β-arrestin recruitment assay. The recruitment of β-arrestin to meOTR and MTRIA OT was assessed using a split-luciferase complementation assay. HEK293T cells that were seeded onto an opaque 96-well plastic plate were co-transfected with LgBit-arrestin and either meOTR-SmBit or MTRIA OT -SmBit using PEI MAX reagent. Twenty-four hours after transfection, cells were washed with HBSS and stimulated with HBSS containing 100 nM OT. After the addition of a luciferase substrate (ONE-Glo Luciferase Assay System, Promega), luminescence was measured using a plate reader (Infinite 200 Pro F Plex; Tecan). were purchased from Nacalai. Epinephrine was purchased from Daiichi-Sankyo and norepinephrine (NE) was purchased from Alfresa Pharma Corporation. Nematocin was synthesized by Cosmo Bio. The final concentrations of applied ligands in Fig. 6b were as follows; 100 µM: DA, NE, ACh, His, 5-HT, ATP, UDP and Ado; 1 µM: Enk, C3a, Ana, LPA, PAF, Pgd, Pge, Pgf and Pgi; and 100 nM: Mtn, Agt, Avp, Bdk, Cck, Crh, ET-1, Gal, Ghr, Msh, Mch, Nmb, NmU, NpFF, Npy, Nts, Ncp, Ox, Prlh, Pth, Sst, SP, Nkb, Uts, C5a, Thr, S1P, Ltb and Lte.

Reagents
Animal surgery. Animal experiments were reviewed and approved by Institutional Animal Use and Care Committees of Kanazawa University and Osaka University. All animal procedures were conducted in accordance with the guidelines of Kanazawa University and Osaka University. C57BL/6J or C57BL/6N female mice at 6-8 weeks, 6 months, 1 year or 2.5 years of age were purchased from CLEA Japan. Mice were kept in cages at 23 ± 1.5 °C with 45% ± 15% humidity under a regular 12-h dark/light cycle with ad libitum access to food and water. Stereotaxic surgery was performed under anesthesia with isoflurane (FUJIFILM Wako). The depth of anesthesia was assessed using the tail pinch method. Body temperature was maintained using a heating pad. One microliter of AAV suspension within a glass micropipette was injected into the corresponding site(s); left medial AON , an opioid analgesic, were administered after the surgery. At 2 weeks or more following AAV injection, mice were used for fiber photometry experiments. On the measurement day, the cannula for fiber photometry recording was coupled to a patch cable (M79L01; Thorlabs) via an interconnector (ADAF1 or ADAF2; Thorlabs). For optogenetic experiments, the cannula for optogenetic stimulation was additionally connected to a red laser (OBIS LX, 647 nm, 100 mW; Coherent) through a patch cable (MAF3L1; Thorlabs) equipped with an interconnector (ADAF1).
Fiber photometry measurements. The fiber photometry setup (Extended Data Fig. 5a) was constructed as previously described, with a minor modification 53 . Excitation light from a light-emitting diode (LED) was directed into a patch cable (M79L01) through an objective lens (CFI Plan Fluor ×20 lens; Nikon), and the emission light was projected onto the sensor of a scientific complementary metal oxide semiconductor camera (Zyla 4.2 P; Andor Technology) after passing through a dichroic mirror (Di01-R405/488/561/635-25×36; Semrock) and emission filter (YIF-BA510-550S; SIGMA KOKI). To acquire an OT-dependent signal and isosbestic signal, two different LED light sources-a 470-nm light (light from M470F3 filtered with FB470-10, 4 µW at the fiber tip; Thorlabs) and 405-nm light (light from M405FP1 filtered with FB410-10, 4 µW at the fiber tip; Thorlabs)were alternatively switched. Data were acquired at 2-s intervals with an LED light exposure of 0.3 s except for measurements of fast responses (Fig. 2f-j and 4e-h), where data were acquired at 0.5-s intervals with an LED light exposure of 0.08 s. In these experimental conditions, no apparent photobleaching was observed. Therefore, we did not conduct photobleaching correction. Device control and data acquisition were conducted using a custom-made LabVIEW program (National Instruments). Experiments were conducted within a cage (270-mm width, 440-mm length and 187-mm height), and the environment around the cage was captured every 1 s using an overhead camera (BFS-U3-I3Y3-C; FLIR Systems). To visualize the behaviors of mice and timing of simulations in the dark environment, light from an infrared LED array (AE-LED56V2; Akizuki) was used. Unless stated otherwise, mice were kept in freely behaving conditions and fed standard pellet and water ad libitum during measurements. For recording of social stimulation-induced OT responses (Fig. 4a-d), a cylindrical wire cage (10-cm diameter and 10-cm height) containing either a mouse or a toy was exposed to a mouse at the indicated times. Similarly, for recording acute stress-induced OT responses (Fig. 4e-h), the tail of a mouse was lifted for 10 s at the indicated times. To assess OT responses upon exogenous OT administration (Fig. 2a-e) and the impact of mix-anes (Fig. 5a,b), mice were anesthetized by intraperitoneal administration of a mixture of 0.375 mg per kg body weight dexmedetomidine hydrochloride, 2 mg per kg body weight midazolam and 2.5 mg per kg body weight butorphanol tartrate. For recovery from the above anesthesia, 0.75 mg per kg body weight atipamezole hydrochloride was administrated intraperitoneally. Before starting optogenetic experiments (Fig. 2f-j) and assessment of the impact of isoflurane (Fig. 5c,d), mice were placed on a heat pad and maintained with 1% isoflurane for at least for 20 min to stabilize the fluorescence signal of MTRIA OT . Data were processed using Fiji software and Excel (Microsoft), and kinetic parameters were calculated using Python modules. After conducting background subtraction, the fluorescence signal (F) was normalized to z-scores defined by (F − F mean )/F sd , where F mean and F sd are the mean and standard deviation of resting fluorescence signals at baseline, respectively. Resting fluorescence signals were taken for the indicated period as follows: 5 s (Fig. 4f-h), 10 s (Fig. 2h-j), 1 min ( Fig. 4b-d), 5 min (Fig. 5c,d and Extended Data Fig. 8b) and 10 min (Figs. 2b-e, 3 and 5e-h). Data were not smoothed except for the fluorescence traces in optogenetic experiments (Fig. 2h-j). For analysis of fluorescence signals in optogenetic experiments, data were filtered using a moving average filter with a kernel (1/5, 1/5, 1/5, 1/5, 1/5) to improve the quality of single-exponential fitting. Time constants (τ on and τ off in Fig. 2j and Rise τ in Fig. 3d) were calculated by fitting either the rise phase or the decay phase with a single-exponential function. Time to peak in Fig. 4h was obtained as a period from the beginning of tail lift to the time point of maximal z-score. AUC values in Figs. 2i and 4c,g were calculated by integrating the values of z-score for 30 s, and 5 min and 10 s, respectively. Z-scores in Figs. 2c and 5d were quantified as the tail values at the end of the indicated stimulation. Peak z-scores in Fig. 2e and Extended Data Fig. 8b were quantified as the peak values following stimulation. The remaining peak z-scores (Figs. 3c,h and 5h) and minimum z-scores (Fig. 5a,f) were obtained as 10-min average values around the peak and the minimum of transient OT responses, respectively. The numbers of peaks in Figs. 3g and 5h were calculated as an occurrence of transient signal with an amplitude exceeding a z-score of 3.

Immunohistochemistry.
Immunohistochemical experiments were performed to confirm sensor expression and cannula insertion sites after fiber photometry measurements and to assess the specificity of OTp-mediated gene transduction. The brains of anesthetized mice were rapidly removed after decapitation, immersed in PBS containing 4% (wt/vol) paraformaldehyde overnight at 4 °C, and then incubated in PBS containing 20% (wt/vol) sucrose overnight at 4 °C. After embedding in Optimal Cutting Temperature Compound (Sakura Finetek), coronal sections of brains were cut at either a 50-µm (for anti-GFP antibody staining) or a 30-µm (for anti-OT antibody staining) thickness using a cryostat (CM1950; Leica Microsystems). For immunostaining, sections were permeabilized in blocking solution for 30 min at room temperature and then incubated with either anti-GFP antibody (rabbit; MBL; 1:1000 dilution) or anti-OT antibody (rabbit; Abcam; 1:200 dilution) overnight at 4 °C. After three washes with PBST, sections were incubated with Alexa 488-conjugated anti-rabbit IgG antibody (goat; Jackson ImmunoResearch; 1:2,000 dilution) for 40 min at room temperature. After three washes with PBST and three washes with PBS, samples were mounted in polyvinyl alcohol-based mounting medium containing 1 µg ml −1 DAPI (FUJIFILM Wako) and 2.5% (wt/vol) 1,4-diazabicyclo[2.2.2]octane (Sigma) as an anti-fade reagent. Stained sections were then imaged using a spinning-disk confocal microscope system (either Dragonfly 301 or SpinSR10). Mice with misplaced cannula were excluded from analysis.

Statistics and reproducibility.
All summary data are expressed as either the mean ± s.e.m. or mean ± s.d. For comparisons of two groups, we used two-tailed Student's t-tests. One-way ANOVA was performed when more than two groups were compared, followed by Tukey-Kramer, Bonferroni's or Dunnett's post hoc test. Throughout the study, P < 0.05 was considered statistically significant. Data distribution was assumed to be normal, and variance was similar between groups that were statistically compared. No statistical methods were used to predetermine sample sizes, but our sample sizes were similar to those generally used in the field [13][14][15][16][17] . Regarding the representative data in Figs. 1e, 2g and 4e and Extended Data Figs. 4a,b, 5b, 6 and 8a, we repeated the experiments two or more times and obtained similar results.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
DNA sequences of the sensors developed in this study have been deposited at DNA Data Bank of Japan (accession nos. LC720900-LC720924). Plasmid DNAs are available from either Addgene (plasmid nos. 184594-184620) or the corresponding author. The raw data for fluorescence recordings are not suitable for distribution through public repositories due to the large file size and are available from the corresponding author upon reasonable request. Source data are provided with this paper. Fig. 1 | optimization of a fluorescent oT sensor. a, representative images of HeK293t cells co-expressing an HA-tagged Otr (green) and a mSca mem (magenta). Normalized fluorescence intensities on the dotted lines are shown on the right. b, Schematic representation of cpGFP insertion into the IL3 of Otr. c, Sequence alignment of regions of the IL3 of human dopamine receptor D1 (hDrD1), a scaffold of a dopamine sensor (dLight1), and meOtr. d, representative images of HeK293t cells co-expressing the indicated meOtr-cpGFP chimera (green) and mSca mem (magenta). e, Pearson correlation coefficients comparing meOtr-cpGFP chimeras and mSca mem are summarized as mean ± SeM (n = 13, 12, 12, 13, 12, 13, 13, 11, 13, 13, 13, 12, 13,

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted

Software and code
Policy information about availability of computer code Data collection Fusion software (Andor, ver. 2.1) and cellSense (Olympus, ver. 2.2) was used for data acquisition in spinning-disk confocal microscopic analyses.Custom-designed LabVIEW software (National Instruments, ver. 2018 or 2020) was used for data acquisition in fiber photometry analyses.

Data analysis
ImageJ software (NIH, ver. 1.52) and python modules (numpy, scipy.optimize.curve_fit) were used for image analysis and data processing. R (ver. 4.05) and Excel (Microsoft, 2016 or 2019) were used for statistical analysis and data plotting.
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A description of any restrictions on data availability -For clinical datasets or third party data, please ensure that the statement adheres to our policy Plasmid DNAs are available either from Addgene (plasmid #184594−184620) or the corresponding author. Source data are provided with this paper. All other raw data can be made available upon reasonable request.
2 nature portfolio | reporting summary

March 2021
Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design
All studies must disclose on these points even when the disclosure is negative.

Sample size
No statistical methods were used to predetermine the sample sizes, but our sample sizes were similar to those generally used in the field ( Data exclusions For fiber photometry analysis, mice with misplacement of the cannula were excluded from the analysis.

Replication
All attempts at replication were successful. All experiments were repeated at least twice.
Randomization Cells and animals were randomly assigned into control or experimental groups.

Blinding
Data collection and analysis were not performed in a blinded manner because the experimental conditions were obvious to the researchers. However, standardized procedures for data collection and analysis were used to prevent human bias.

Reporting for specific materials, systems and methods
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.

Authentication
The cell line used has not been authenticated, but its morphology was similar to that reported by the supplier (https:// www.atcc.org/products/crl-3216).

Mycoplasma contamination
Cell lines used were not tested for mycoplasma contamination.
Commonly misidentified lines (See ICLAC register) No commonly misidentified cell lines were used.