Introduction

As the main input nucleus of the basal ganglia, the dorsal striatum guides goal-directed and habit learning by plastically adjusting how it processes cortical inputs1,2. Most of the relevant processes are shaped, at least in part, by the monoamine serotonin (5-HT) – this includes long term depression at corticostriatal synapses3, lateral inhibition between spiny projection neurons4 and release of another monoamine (dopamine) from nigrostriatal projections5. As such, regulation of 5-HT signaling is crucial to striatal function6.

Dysregulated striatal 5-HT signaling has been implicated in numerous neurological, psychiatric and addictive disorders; strikingly, pathological changes in striatal 5-HT signaling may occur in tandem with changes in both dopamine and acetylcholine7,8,9,10. This may not be a coincidence. It has become apparent that monoamines and acetylcholine (ACh) can intimately shape each other’s release in the striatum: for example, dopamine regulates the excitability of striatal cholinergic interneurons11 (CINs), and ACh release from CINs directly facilitates dopamine release by engaging nicotinic ACh receptors (nAChRs) on nigrostriatal axons12,13,14,15,16. Whether such an interplay extends to 5-HT is unclear: it is well-established that 5-HT directly regulates the activity of CINs (and presumably ACh release) in a region-specific manner17,18, but evidence for direct functional control of 5-HT release by ACh is limited19,20,21. The main source of striatal 5-HT is a long range projection from the dorsal raphe nucleus22,23,24,25, and serotonergic axons within the striatum express nAChRs. We therefore sought to determine if CIN-derived ACh can directly facilitate local 5-HT release, in a nAChR-dependent way.

Here, we show that synchronous activation of CINs triggers local 5-HT release in the dorsal, but not ventral, striatum via nAChRs expressed on serotonergic axons. We further show that CIN–5-HT coupling expands the spatial extent of 5-HT release and is selectively amplified in the Sapap3-/- mouse model of obsessive–compulsive disorder (OCD)–like behavior.

Results

Local nAChR-dependent 5-HT release in the dorsal – but not ventral – striatum

We stereotaxically inoculated the dorsal striatum of mice with an adeno-associated virus (AAV) harboring the genetically-encoded green fluorescent 5-HT sensor, GRAB-5HT26, and when the sensor was expressed, we prepared acute brain slices through the dorsal striatum (Fig. 1a). Intrastriatal electrical stimulation induced a robust increase in GRAB-5HT fluorescence, which decayed over 10 s of seconds, as measured using two-photon laser scanning microscopy (2PLSM) (Fig. 1b) (see Methods). Application of the non-selective nAChR antagonist mecamylamine (10 µM) significantly reduced the amplitude of the striatal GRAB-5HT response (averaged over the region-of-interest, see Methods; Fig. 1a–c), suggesting that, as is the case for dopamine (DA)12,13,14,15,16,27, nAChR activation also facilitates striatal 5-HT release [N = 3 mice, n = 9 slices, t16 = –5.57, P = 4.25·10–5, linear-mixed effect model (LMEM)].

Fig. 1: Nicotinic acetylcholine receptors (nAChRs) elevate and extend the spatial range of serotonin (5-HT) release evoked by intrastriatal electrical stimulation in dorsal striatum (DS).
Fig. 1: Nicotinic acetylcholine receptors (nAChRs) elevate and extend the spatial range of serotonin (5-HT) release evoked by intrastriatal electrical stimulation in dorsal striatum (DS).
Full size image

a Schematic of adeno-associated virus (AAV) injection into the DS leading to GRAB-5HT expression (left). Representative field-of-view (FoV) of a region of DS expressing GRAB-5HT imaged with two-photon laser scanning microscopy (2PLSM, right). b GRAB-5HT response to electrical stimulation (10 pulses at 10 Hz, 2 ms), before (black) and after (magenta) application of 10 µM mecamylamine (mec). c Distribution of peak GRAB-5HT responses before and after application of 10 µM mec (N = 3 mice, n = 9 slices, t16 = –5.57, P = 4.25∙10–5, LMEM). Boxplots indicate: range (whiskers), median (thick bar) and 25 and 75th percentiles (box). d Same as (a) for the ventral striatum. e Same as (b) for the ventral striatum. f Same as (c) for the data in (e) (N = 4 mice, n = 23 slices, LMEM). g Schematic of electrical stimulation of sagittal slices through DS indicates the location of the sequence of 100 µm wide rectangular ROIs used to measure the peak GRAB-5HT response as function of distance from the stimulating electrode. (left). Representative traces of GRAB-5HT responses averaged over individual ROIs (0, 200, 500 µm), before (black) and after (magenta) application of 10 µM mec (right). h Lorentzian fit of maximal GRAB-5HT responses as a function of distance from the stimulating electrode (g), before (black) and after (magenta) application of 10 µM mec (N = 3 mice, n = 9 slices, t158 = –2.68, P = 0.0081, LMEM). Horizontal lines indicate means, vertical lines indicate 95% confidence intervals, and circles represent individual datapoints. ** P < 0.01; *** P < 0.001; n.s. – not significant. ACSF – artificial cerebrospinal fluid; LMEM – linear mixed-effects model.

Bath application of DA (10 µM) had no effect on GRAB-5HT fluorescence in slices (Supplementary Fig. 1a–c), confirming that changes in sensor fluorescence were not the result of nonselective detection of DA26. We further confirmed the specificity of GRAB-5HT fluorescence in our experiments by demonstrating that 1) RS 23597-190 (35 µM; a selective antagonist of the 5-HT2C receptor, on which the GRAB-5HT design is based) eliminated evoked fluorescence (N = 3 mice, n = 6 slices, t10 = –3.45, P = 0.006, LMEM) (Supplementary Fig. 2a, b) and 2) citalopram (30 µM), a selective 5-HT reuptake inhibitor (SSRI), significantly reduced the amplitude of the 5-HT release (N = 4 mice, n = 12 slices, t22 = –4.64, P = 1.3·10–4, LMEM), likely due to presynaptic inhibition of release by 5-HT1B auto-receptors18,28,29,30,31,32, and significantly slowed the temporal decay of evoked fluorescence (N = 4 mice, n = 12 slices, t22 = –4.29, P = 3·10–4, LMEM), neither of which could be attributed to the passage of time (Supplementary Fig. 2c–h).

Compared to the dorsal striatum, the ventral striatum receives substantially denser 5-HT innervation33,34. We therefore tested whether nAChRs also facilitate 5-HT release in ventral striatum. Surprisingly, we found that while electrical stimulation elicited robust increases in GRAB-5HT fluorescence in the ventral striatum, their amplitude was not reduced by mecamylamine (Fig. 1d–fN = 4 mice, n = 23 slices, t40 = 0.399, P = 0.692, LMEM), suggesting that nAChRs do not contribute to 5-HT release in the ventral striatum.

nAChRs extend the spatial range of striatal 5-HT release

We and others have shown that activation of nAChRs on nigrostriatal axon terminals not only increases the amount of DA release, but also nearly doubles the spatial range of its release12,13. We therefore measured the spatial extent of GRAB-5HT responses by quantifying evoked fluorescence at 100 µm-wide increments from the stimulating electrode and fitting a Lorentzian function to the maximal responses (N = 3 mice, n = 9 slices) as a function of distance (Fig. 1g, h). The Lorentzian fell off with a length scale of 183 µm under control conditions. Mecamylamine application not only reduced the overall amplitude of the response (t158 = –5.8, P = 3.55·10–8, LMEM), but also significantly shortened the length scale of the Lorentzian to 126 µm (t158 = 2.68, P = 0.0081, LMEM), indicating that nAChRs increase the spatial extent of striatal 5-HT release by approximately 45% (Fig. 1h).

Optogenetic activation of CINs saturates striatal nAChR-dependent 5-HT release

We observed that intrastriatal electrical stimulation in the dorsal striatum promotes nAChR-sensitive release of 5-HT, but can endogenous ACh released from CINs drive nAChR-dependent release of striatal 5-HT on its own? To this end, we inoculated the dorsal striatum of ChAT-Cre mice (which express Cre recombinase under the choline acetyltransferase promoter in CINs) with a mixture of the AAVs harboring GRAB-5HT (broadly labeling the dorsal striatum) and DIO-channelrhodopsin-2 (ChR2)-mCherry (specifically labeling CINs) (see Methods) (Fig. 2a; Supplementary Fig. 3a). The two-photon laser scanning per se did not alter the electrophysiological behavior of CINs (N = 5 mice, n = 8 CINs, t26 = –1.35, P = 0.18, LMEM) (Supplementary Fig. 3d–i). A brief (1 ms) 470 nm (32 mW) LED stimulation, which is sufficient to drive a time locked action potential in recorded CINs (Supplementary Fig. 3f), resulted in a significant increase in GRAB-5HT fluorescence, with a slow decay lasting tens of seconds (Fig. 2b). Evoked changes in GRAB-5HT fluorescence were not an optical artifact of blue light stimulation, as: 1) under our experimental conditions, 470 nm LED simulation on its own did not alter GRAB-5HT fluorescence (Supplementary Fig. 1d, e), and 2) CIN-dependent 5-HT release could be replicated by optogenetically activating CINs using the red-shifted opsin [using 1 ms, 590 nm (7.3 mW) LED pulses in dorsal striatal brain slices from ChAT-cre mice inoculated with a mixture of GRAB-5HT and cre-dependent ChrimsonR-tdTomato] (Supplementary Fig. 4). Thus, synchronous activation of CINs is sufficient to evoke 5-HT release in the dorsal striatum. The GRAB-5HT signal was completely abolished by 10 μM mecamylamine or dihydro-β-erythroidine (DHβE), suggesting that CIN-evoked 5-HT release is mediated by α4β2 type nAChRs [ChR2: N = 3 mice, n = 9 slices with mecamylamine, N = 1 mouse, n = 3 slices with DHβE (pooled) t22 = –7.9., P = 7.3·10–8, LMEM; Chrimson: N = 2 mice, n = 11 slices (all with mecamylamine), t20 = –7.18, P = 5.9·10–7, LMEM] (Fig. 2b, c; Supplementary Fig. 4b, c).

Fig. 2: Optogenetic activation of cholinergic interneurons (CIN) leads to nAChR-dependent 5-HT release in dorsal striatum (DS).
Fig. 2: Optogenetic activation of cholinergic interneurons (CIN) leads to nAChR-dependent 5-HT release in dorsal striatum (DS).
Full size image

a Schematic of AAV injection into the DS leading to non-selective expression of GRAB-5HT and selective expression of channelrhodopsin-2 (ChR2)-mCherry in CINs in ChAT-cre mice (left). 2PLSM image of a representative region of DS expressing GRAB-5HT (green) overlaid with a Z-stacked image of CINs expressing ChR2-mCherry (red). b GRAB-5HT response to a brief (1 ms long) optogenetic stimulation (470 nm), before (black) and after (magenta) application of 10 µM mecamylamine (mec). Vertical cyan line indicates a brief period of blanking of the stimulation artifact (see Methods). c Distribution of peak GRAB-5HT responses before (black) and after (magenta) application of nAChR blockers (mec and DHβE data pooled, N = 4 mice, n = 12 slices, t22 = –7.9, P = 7.3·10–8, LMEM). d nAChR-dependent 5-HT release in the DS (left) and the ventral striatum (right, gray background) evoked by electrical (black) and optogenetic (red) stimulation [ventral vs. dorsal: N = 7 mice, n = 21 slices, t26 = –4.5, P = 1.2·10–4 (electrical), N = 11 mice, n = 28 slices,, t32 = –5.24, P = 9.9·10–6 (optogenetic), LMEM]. e Same as (b), with the additional blue trace indicating the response after the application of a cocktail of glutamatergic, GABAergic and muscarinic receptor blockers. f Same as (c) for the data in (e). g Distribution of response decay kinetics before (black) and after (blue) application of cocktail of antagonists (N = 3 mice, n = 8 slices, mec vs. ACSF: t21 = –6.02, P = 5.69·10-6, cocktail vs. mec: t21 = –5.42, P = 2.19·10-5, LMEM). Boxplots indicate: range (whiskers), median (thick bar) and 25 and 75th percentiles (box). ** P < 0.01; *** P < 0.001; n.s – not significant. ACSF – artificial cerebrospinal fluid; LMEM – linear mixed-effects model.

Strikingly, the increase in GRAB-5HT fluorescence induced by synchronous optogenetic activation CINs is similar to the nAChR-dependent component of the GRAB-5HT signal evoked by electrical stimulation (Fig. 2d) (N = 9 mice, n = 25 slices, t23 = 0.01, P = 0.99, LMEM), suggesting that CINs are likely the sole source of ACh that promotes release of 5-HT in the dorsal striatum. Consistent with the results presented above using electrical stimulation (Fig. 1d–f), optogenetic stimulation of CINs failed to elicit detectable 5-HT release in ventral striatum (Supplementary Fig. 5) (N = 7 mice, n = 25 slices, t45 = 0.95 P = 0.35, LMEM).

Is CIN control of 5-HT release direct, or are the key ACh receptors on striatal loci other than serotonergic axons? To rule out the involvement of postsynaptic nAChRs on GABAergic interneurons35, presynaptic nAChRs on glutamatergic inputs36 or muscarinic ACh receptors in general37,38 we repeated the above experiment in the presence of GABA, glutamate and muscarinic ACh receptor antagonists (10 µM SR 95531, 2 µM CPG 55845, 10 µM DNQX, 50 µM D-AP5 and 10 µM atropine). This antagonist cocktail had no significant effect on either the amplitude (N = 3 mice, n = 8 slices, t21 = –0.59, P = 0.56, LMEM) or the temporal decay (t21 = –1.98, P = 0.061, LMEM) of the CIN-evoked 5-HT release, which was eliminated by subsequent blockade of nAChRs (Fig. 2e–g). Together, these data suggest that ACh acts directly on nAChRs located on 5-HT fibers to mediate 5-HT release [t21 = –6.02, P = 5.69·10-6 (ACSF vs. mec), t21 = –5.42, P = 2.19·10-5 (cocktail vs. mec), LMEM]. Nevertheless, to determine whether nAChRs amplify a putative co-release of 5-HT from DA axons39, or any other mechanism by which excitation of striatal DA axons may indirectly cause the release of striatal 5-HT, we inoculated 1) the substantia nigra pars compacta (SNc) of TH-Cre mice (which express Cre-recombinase under the tyrosine hydroxylase promoter) with an AAV harboring DIO-ChR2-mCherry and 2) the dorsal striatum with an AAV harboring GRAB-5HT (Supplementary Fig. 6a–d). Direct optogenetic activation of DA fibers induced a small but significant increase in GRAB-5HT. The increase, however, was only about a fifth of the size of the GRAB-5HT signal generated by synchronous optogenetic activation of CINs (Fig. 2c, e, Supplementary Fig. 4) and was insensitive to mecamylamine, indicating that DA axons do not contribute to the nAChR-dependent release of striatal 5-HT (Supplementary Fig. 6e, f).

nAChR-dependent monoamine release is elevated in a mouse model of OCD

If CINs induce the release of 5-HT under physiological conditions, then in a pathophysiologically hypercholinergic striatum this mechanism should be amplified. We recently reported that Sapap3-/- mice, which display OCD-like motor behaviors such as compulsive grooming, exhibit a hypercholinergic phenotype in the striatum: the density of CINs is higher, their autonomous firing rate is higher, and they release more ACh in response to electrical stimulation in acute striatal slices40. We therefore asked whether cholinergic modulation of 5-HT release is amplified in Sapap3-/- mice. Indeed, intrastriatal electrical stimulation in the dorsal striatum induced significantly larger GRAB-5HT fluorescence signals in the dorsal striatum of Sapap3-/- mice, and this difference disappeared in the presence of mecamylamine (Fig. 3). Supporting the conclusion that elevated 5-HT release is the result of increased CIN-mediated ACh release, only the mecamylamine-sensitive component of the GRAB-5HT signal was elevated in Sapap3-/- mice (Fig. 3c), suggesting that basal striatal 5-HT release capacity is unaltered. Corroborating that ACh facilitates 5-HT release by a mechanism similar to the one by which it facilitates DA release, intrastriatal electrical stimulation also induced significantly higher nAChR-sensitive (but not nAChR-insensitive) striatal DA release in Sapap3-/- mice (Supplementary Fig. 7). Overall, this is consistent with the observation that 5-HT turnover (but not baseline levels) is elevated in Sapap3-/- mice41.

Fig. 3: Increased cholinergic-mediated striatal serotonin release in Sapap3 KO mice.
Fig. 3: Increased cholinergic-mediated striatal serotonin release in Sapap3 KO mice.
Full size image

a Representative line scan traces of GRAB-5HT fluorescence in the dorsal striatum of wild type (WT) (top, black) and knockout (KO) Sapap3 (bottom, blue) mice, before and after bath application of mecamylamine (mec, magenta). b Distribution of peak GRAB-5HT responses before (black – WT, cyan - KO) and after (magenta) application of mecamylamine in WT (left, N = 5 mice, n = 9 slices t16 = –3.5054, P = 0.002, LMEM) and KO (right, N = 6 mice, n = 12 slices t22 = –6.066, P = 4.2∙ 10-6 LMEM). c Quantification of mec-sensitive (left, t19 = 2.89, P = 0.009) and -insensitive (right, t19 = 9.5∙ 10-5 P = 0.999, LMEM) components of GRAB-5HT signal in WT and KO mice. Boxplots indicate: range (whiskers), median (thick bar) and 25 and 75th percentiles (box). * P < 0.05; ** P < 0.01; *** P < 0.001; n.s – not significant. ACSF – artificial cerebrospinal fluid.

Discussion

While the presence of nAChRs on 5-HT axons in the striatum was previously reported19,20,21, we provide a functional demonstration that activation of CINs can induce local 5-HT release, a phenomenon that extends our understanding of reciprocal ACh-monoamine regulation to include 5-HT12,13,15,16,17,18,27. Specifically, as previously shown for DA12,13, ACh released by CINs can directly induce, amplify and extend the spatial footprint of striatal 5-HT release via activation of axonal nAChRs. The present study employed convergent pharmacological and circuit-based evidence to demonstrate CIN-dependent 5-HT release; future studies may utilize additional approaches, such as selectively knocking down nAChR subunit expression in serotonergic dorsal raphe neurons or testing 5-HT-insensitive GRAB variants, to further corroborate and expand upon these findings.

While we observed CIN-dependent 5-HT release in the dorsal striatum, this mechanism was absent in the ventral striatum. This finding appears counterintuitive, given that 5-HT innervation density is greater in the ventral striatum33,34. One interpretation is that this difference reflects the lower density of CINs in the ventral striatum42,43, suggesting that CIN density, rather than 5-HT axonal density, may be the key determinant of this neuromodulatory interaction. Alternatively, the 5-HT fibers that innervate the dorsal and ventral striatum may have different nAChR expression profiles, leading to differential sensitivity to ACh24,44. Other region-specific aspects of CIN function and 5-HT signaling may further amplify regional differences. In the dorsal striatum, 5-HT increases CIN excitability and firing rate, whereas in the ventral striatum 5-HT does the opposite17,18,45. This differential modulation, combined with regional heterogeneities in CIN intrinsic function, may synergize to shape the release dynamics of both ACh46,47 and 5-HT.

We have provided an ex vivo demonstration that activation of CINs leads to local nAChR-dependent 5-HT release. How may this mechanism be engaged in vivo? In the case of DA, where release is augmented by the same mechanism that we report for 5-HT, this has turned out to be a difficult question to answer – multiple labs over 3 decades have reported in vitro nAChR-dependent DA release39,48,49 driven by CINs12,13,14,15,16, but attempts to demonstrate how this mechanism is engaged in vivo15,50 have been met by skepticism or constrained success51,52,53. Very recently, however, this long-standing impasse has been broken by the identification of specific behavioral conditions under which CINs are recruited to augment DA release in vivo54. Thus, context is critical52,55,56. This is likely to be the case for modulation of 5-HT release in vivo as well – based on the work examining CIN-dependent DA release in vivo, it is unlikely we would stumble upon a measurable CIN-dependent 5-HT release in the absence of an appropriately defined behavioral or circuit-state context. One potential clue comes from the observation that local transfusion of nicotine, albeit independent of CIN activity, has been shown to be necessary and sufficient to drive striatal 5-HT release, but only in the presence of a stressful stimulus (foot shock)21. The precise activity of CINs that would be required to evoke 5-HT in this context is unclear, and may not be on the same sub-second timescale of the coupling we describe. Thus, future work is needed to determine the behavioral context in which nAChR activation leads to time-locked local release of 5-HT, but we show here that the requisite mechanism is in place and selectively restricted to specific striatal subregions. Notably, CINs are recruited synchronously by salient events, usually by way of activation of their thalamic input57,58,59, so the CIN synchrony required to drive local 5-HT release is likely attained in vivo.

Local coupling between ACh and 5-HT release is exaggerated in the striatum of Sapap3-/- mice, which exhibit pathologically elevated ACh release40, suggesting that nicotinic amplification of monoamine release may be a general property of hypercholinergic states in the striatum. Notably, it is only the nAChR-dependent component of locally induced monoamine release that is increased in Sapap3-/- mice (Fig. 3c and Supplementary Fig. 7c). Moreover, in the case of 5-HT, the nAChR-dependent component becomes even larger than the non-nAChR-dependent component, indicating that local hijacking of 5-HT fibers may become the main medium of 5-HT release in Sapap3-/- mice. If and how pathological modulation of 5-HT release is further shaped by other striatal circuit components (i.e., GABAergic interneurons and astrocytes) or varies across species60 remains to be determined.

What does this imply for the etiology of OCD-like behaviors? The first-line pharmacological treatment for OCD is selective serotonin reuptake inhibitors (SSRIs); while the mechanism of action is typically thought to involve increasing 5-HT tone, conclusive proof is lacking and some evidence suggests that chronic SSRI treatment may suppress 5-HT synthesis10,61. The answer to this question is beyond the scope of our current study, but our data offer a mechanism for monoamine dysregulation in a mouse model of compulsive behavior.

Hypercholinergic states arise in the striatum in other neurological disorders, notably Parkinson’s disease (PD)62, raising the possibility that exaggerated recruitment of 5-HT release by CINs may contribute to PD pathophysiology. In particular, after the loss of dopaminergic innervation to the striatum, 5-HT fibers proliferate and their terminals convert levodopa into DA, resulting in the non-canonical release of DA from 5-HT terminals63,64. Given the growing evidence implicating CINs in levodopa induced dyskinesia (LID)65,66,67, it is tempting to speculate that abnormal CIN activity could amplify this non-canonical release of DA leading to dyskinetic symptoms64. On a final speculative note, evidence suggests that both dopamine and 5-HT may set behavioral states that shape learning and disease68. Our data suggest a mechanism by which CINs may facilitate interactions between such monoamine-dependent states.

Methods

Animals

Experimental procedures adhered to and received prior written approval from the Institutional Animal Care and Use Committees (IACUC) of the Hebrew University of Jerusalem (HUJI, protocol number MD-20-16113-3) and Stony Brook University (SBU). Experiments were carried out in the following mouse strains: 1. C57BL/6 J mice (Strain #:000664; Jackson Laboratories, Bar Harbor, ME, USA); 2. ChAT-IRES-Cre (∆neo) transgenic mice (stock number 031661; Jackson Laboratories, Bar Harbor, ME, USA); 3. TH-Cre 1 (Strain #:008601; Jackson Laboratories, Bar Harbor, ME, USA, RRID:IMSR_JAX:008601); and 4. Sapap3 conditional knock-in mice40. All Sapap3-/- mice lacked both copies of the Sapap3 gene. Two-to-seven-month-old C57BL/6 J and transgenic mice were used for experiments. Sapap3-/- mice were 3-7 months old, when they exhibited compulsive grooming. Sex was not considered as a factor in the design of the experiments, and so mice of both sexes were used and their results pooled. All mice were housed under a 12 h light/dark cycle with food and water ad libitum. Ambient temperatures and humidity were 22 °C ± 2 °C and 50 ± 10%, respectively.

Stereotaxic surgeries

At HUJI

Mice were deeply anesthetized with isoflurane in a non-rebreathing system (2.5% induction, 1–1.5% maintenance) and placed in a stereotaxic frame (model 940, Kopf Instruments, Tujunga, CA, USA). Temperature was maintained at 35 °C with a heating pad, antibiotic eye ointment was applied to prevent corneal drying, and animals were hydrated with a bolus of injectable saline (5 ml/kg) mixed with an analgesic (5 mg/kg meloxicam). To calibrate specific injection coordinates, the distance between bregma and lambda bones was measured and stereotaxic placement of the mice was readjusted to achieve maximum accuracy. A small hole was bored into the skull with a micro drill bit and a glass pipette with AAVs attached to Nanoject III system (Drummond) was slowly inserted at the relevant coordinates under aseptic conditions. The coordinates of dorsal striatum (DS) injection were: +0.5 mm anteroposterior (AP); +/- 2.3 mm mediolateral (ML); −2.8 mm dorsoventral (DV), and for the SNc: -3.1 mm anteroposterior (AP); +/- 1.2 mm mediolateral (ML); −4.2 mm dorsoventral (DV) relative to bregma using a flat skull position. The coordinates of the nucleus accumbens (NAc) core injection were: +1.5 mm anteroposterior (AP); +/- 0.9 mm mediolateral (ML); −4.1 mm dorsoventral (DV). To minimize backflow, solution was slowly injected (1 nl/second) and the pipette was left in place for 5–7 min before slowly retracting it. For 5-HT sensor experiments, a total amount of 500 nl of an adeno-associated virus harboring the green-fluorescent GRAB-5HT3.5 construct (AAV-9hSyn-5HT3.5; 2.20 × 1013 vg/ml; WZ Biosciences Inc. Lot No. 20201210), recently renamed g5-HT2h21, was injected into DS or 300 nl to the NAc. For optogenetic stimulations of CINs a total amount of 650 nl adeno-associated virus cocktail of GRAB-5HT3.5 construct and ChR2 construct (AAV5/ Ef1a-DIO-hchR2(H134R)-mCherry; 4× 1012 vg/ml; UNC vector core; Lot# AV4313-2A) or Chrimson construct (pAAV/Syn-FLEX-rc[ChrimsonR-tdTomato]; 12× 1012 vg/ml; Addgene #62723 – AAV5; Lot# v122159) was injected into DS of ChAT-IRES-Cre mice; for optogenetic stimulation of dopamine axons 500 nl of GRAB-5HT3.5 was injected into the DS and 250 nl of the ChR2 construct in to the SNc of TH-Cre mice.

At SBU (experiments involving Sapap3 -/- mice)

Mice were induced and maintained under anesthesia via inhalation of vaporized isoflurane (3% induction, 1–2% for maintenance). Prior to any surgical procedures, sedation was confirmed via toe pinch, then animals were placed in a stereotaxic apparatus, the scalp was shaved and disinfected with 70% ethanol and Betadine solution, and eyes were covered with protective gel (Puralube Ophthalmic Ointment, Dechra Veterinary Products). A 1 μL Hamilton Neuros micro-volume syringe was used for unilateral DS injections (0.5-62 mm AP; −2.2-3 mm ML; −2.8 mm DV) and manipulated via a motorized stereotaxic injector (Quintessential Stereotaxic Injector QSI, Stoelting Co.) guided by Angle Two software (Leica Biosystems). 500 nl of AAV9-hSyn-5HT3.5 (≥1x1013vg/ml) or AAV9-hSyn-dLight1.1 (≥1x1013vg/ml) were injected over 5 min and a further 5 min was allowed for diffusion before slowly withdrawing the needle. All mice were administered analgesic (Meloxicam, 5 mg/kg, subcutaneous injection) immediately after surgery and monitored daily for 72 h.

Acute slice preparation

At HUJI

After prolonged incubation time (at least 3 weeks), inoculated mice were deeply anesthetized with ketamine (200 mg/kg)–xylazine (23.32 mg/kg) and perfused transcardially with ice-cold modified artificial cerebrospinal fluid (ACSF) bubbled with 95% O2–5% CO2, and containing (in mM): 2.5 KCl, 26 NaHCO3, 1.25 Na2HPO4, 0.5 CaCl2, 10 MgSO4, 0.4 ascorbic acid, 10 glucose, and 210 sucrose. The brain was removed and sagittal slices sectioned at a thickness of 275 µm were obtained in ice-cold modified ACSF. Slices were then submerged in ACSF, bubbled with 95% O2–5% CO2, and containing (in mM): 2.5 KCl, 126 NaCl, 26 NaHCO3, 1.25 Na2HPO4, 2 CaCl2, 2 MgSO4, and 10 glucose, and stored at room temperature for at least 1 h prior to recording.

At SBU (experiments involving Sapap3 -/- mice)

2-5 weeks after surgery, acute coronal slices (275 μm) containing DS were obtained from mice following anesthetization with ketamine/xylazine (100 mg/kg / 7 mg/kg) and transcardial perfusion with ice-cold artificial cerebral spinal fluid (ACSF) containing (in mM): 124 NaCl, 3 KCl, 1 CaCl2, 1.5 MgCl2, 26 NaHCO3, 1 NaH2PO4, and 14 glucose. Slices were cut using a VT-1000 S vibratome (Leica Microsystems, Buffalo Grove, IL) and transferred to a holding chamber where they were incubated at 32 °C for 45 min in ASCF containing (in mM) 2 CaCl2 and 1 MgCl2, then kept at room temperature until recording. During recordings, slices were perfused with ACSF heated to 32 °C. All solutions were continuously bubbled with carbogen (95% O2 and 5% CO2).

Slice visualization and 2PLSM imaging

At HUJI

Slices were transferred to the recording chamber of Femto2D-resonant scanner multiphoton system (Femtonics Ltd., Budapest, Hungary) and perfused with oxygenated ACSF at 32 °C. A ×16, 0.8 NA water immersion objective was used to examine the slice using oblique illumination.

2PLSM monoamine imaging and electrical / optogenetic stimulation

The 2PLSM excitation source was a Chameleon Vision 2 tunable pulsed laser system (680–1080 nm; Coherent Laser Group, Santa Clara, CA). Optical imaging of GRAB-5HT3.5 signals was performed by using a 920-nm excitation beam (10 mW after the objective), cholinergic mCherry imaging was performed by using a 1060 nm excitation beam. The fluorescence emission was detected and collected with gated GaAsP photomultiplier tubes (PMTs) for high sensitivity detection of fluorescence photons as part of the Femto2D-resonant scanner. Optical and electrophysiological data were obtained using the software package MESc (Femtonics, Budapest, Hungary), which also integrates the control of all hardware units in the microscope, opotogenetics & electrical stimulator. The software automates and synchronizes the imaging signals and stimulations (electrical/optical). Data were extracted from the MESc package (Femtonics) to personal computers using custom-made code in MATLAB (MathWorks, Natick, MA, USA) code.

Fields-of-view (FoV) of approximately 307 µm × 307 µm were selected and imaged at 31 Hz. Scans were performed using 0.6 μm pixels. Regions-of-interest (ROI) where fluorescence was highly expressed in the slice were identified automatically via custom MATLAB code binarizing the image by pixel intensities. The system is also equipped with full-field 470 nm and 590 nm LED illumination, which was used to stimulate the CINs in the ChAT-IRES-Cre mice transfected with DIO-hchR2(H132R)-mCherry, or with FLEX-rc[ChrimsonR-tdTomato], respectively. We empirically chose a LED stimulus amplitude (32 mW and 7.3 mW, respectively, after the objective) that reliably gave measurable GRAB-5HT responses, and a pulse duration of 1 ms as this sufficed to reliably trigger a single action potential in CINs (Supplementary Fig. 3f). Five milliseconds before the LED pulse, the gated PMTs were turned off for a duration of 50 ms. Nevertheless, despite removal of the voltage from the dynode of the PMT the massive flux of blue photons creates a huge photocurrent on the GaAsP photocathode which subsequently saturate the sensor’s electronics leading to a light artifact in the signal lasting approximately 250 ms, evident even in non-expressing slices from wild type mice. Therefore, the first 258 ms (8 samples at 31 Hz) after any LED stimulation were removed from the data of all optogenetic experiments, which precludes observing the initial rise of the GRAB signals. When using ChrimsonR-tdTomato (Supplementary Fig. 4), for consistency, the same procedures were carried out. Electrical stimulation was carried out by a parallel bipolar Platinum-Iridium electrode with diameter of 125 µm and spacing of 500 µm (FHC, PBSA0575). The magnitude of the stimulus was controlled through a stimulus isolator (ISO-Flex, MicroProbes) while the frequency and duration were controlled by computer software (MESc, Femtonics). A total of 10 pulses (10 Hz, 2 ms duration, 3 mA) were delivered to elicit a maximal response (Supplementary Fig. 8)13. To obtain sufficient signal-to-noise-ratio each optogenetic experiment was repeated 9 times, and each electric stimulation experiment was repeated 4 times with a 60 second inter-stimulus-interval. Traces were averaged over all repetitions and smoothed with a Gaussian filter with a width of approximately 400 ms. Fluorescent changes over time (∆F/F0) were calculated as \(\triangle F/{F}_{0}=(F-{F}_{0})/{F}_{0}\), where F is the maximal fluorescent value within 5 seconds of the stimulation recorded while evoking electrical / optogenetic stimulations; and where F0 denotes the averaged baseline fluorescence (1-3 s long, as appropriate for each experiment) before the stimulation.

Slopes for ACSF, Cocktail and Citalopram conditions were extracted by a linear fit from the peak amplitude to the end of the decay duration. In Fig. 2b, d and Supplementary Fig. 2a, amplitude and slopes were reported after subtraction of the slope of the mecamylamine condition to account for photobleaching using the DETREND command in MATLAB.

Time lapse imaging

One-second-long time lapse imaging (Supplementary Fig. 1) were taken with ~30 seconds interval. Each point was calculated as the mean fluorescence intensity in the ROI (method explained above) over entire recording session (parameter mentioned above). 10 µM DA or 10 µM 5-HT were diluted in ACSF (see above) and perfused onto recording chamber for ~ 15 minutes.

Estimation of GRAB-5HT spatial scale

In Fig.1h, ∆F/F0 data points were fitted with a Lorentzian function \(a/[1+{(x/\lambda )}^{2}]\) to extract the length scale λ of the fluorescence signals’ decay with distance. Individual mice were considered independent samples.

Whole cell recording and electrophysiology

For whole-cell current-clamp recordings from CINs, the pipette contained (in mM): 135.5 KCH3SO4, 5 KCl, 2.5 NaCl, 5 Na-phosphocreatine, 10 HEPES, 0.2 EGTA, 0.21 Na2GTP, and 2 Mg1.5ATP (pH = 7.3 with KOH, 280–290 mOsm/kg). CINs were identified by expression of red mCherry molecules along with their morphology, spontaneous firing and by the presence of a voltage-sag (due to their prominent HCN currents).

At SBU (Slice 2PLSM 5-HT and dopamine sensor imaging in experiments involving Sapap3 -/- mice)

Recordings were made using an Ultima Laser Scanning Microscope system (Bruker Nano, Inc, Middleton, WI) equipped with an Olympus 60X/1.0 water-dipping lens (Olympus, Center Valley, PA) and Chameleon Ultra II laser (Coherent, Inc., Santa Clara, CA). To evoke 5-HT or dopamine release, a bipolar stimulating microelectrode (FHC Inc., Bowdoin, ME) powered by an ISO-Flex Stimulus Isolator (Microprobes for Life Science) was embedded into the surface of the slice just outside the 60x objective field-of-view (200μm2 FOV area). 2-photon laser line scans in the shape of a spiral (920 nm, 10μs/pixel dwell, 21.2-83 ms/line) covering the entire FOV were acquired while delivering electrical stimuli (1 s at 10 Hz, 2 ms pulse width, 3 mA for 5-HT; single 1 ms pulse for dopamine). Line scans began 2 s before stimulation and continued for 57 s (5-HT) or 8 s (dopamine) after stimulation ended. 10 µM mec or DHβE were then perfused onto the slice for 10 min and the line scan protocol was repeated. For baseline and post-drug recordings, three 60 s recordings were averaged. ΔF/F0 was calculated as above.

Immunohistochemistry and confocal imaging

After viral injection (see stereotaxic surgeries), mice were deeply anesthetized with ketamine-xylazine followed by cold perfusion to the heart of PBS and 4% PFA. The brain was removed and kept overnight at 4 °C in 4% PFA. The next day, the brain was washed in PBS (three times, 20 min each) before thin 50 µm coronal slices were made using a vibratome (Leica VT1000S). Slices were briefly washed with CAS-BLOCK (Life Technologies) before being incubated in CAS-BLOCK (300 µl) overnight with Goat anti Vesicular-Acetylcholine-transporter antibodies (ABN100MI, Thermo Fisher, LOT: 3974859,1:200). The slices were then washed in PBS (three times, 20 min each) and incubated with CAS-BLOCK and secondary antibodies (Abcam ab6566, LOT: GRI159765-18, 1:200) for 3 hr. Antifade Mounting Medium (VECTASHIELD) was applied to prevent slice bleaching.

Confocal imaging was acquired by Spinning Disk confocal microscope (Nikon) using Yokogawa W1 Spinning Disk camera and a 50 µm pinhole and a 20x dry objective with 0.75 N. A (CFI PLAN APO VC). Excitation laser used were 488 nm, 561 nm, 638 nm to excite GRAB-5HT, mCherry, anti VAChT, respectively. Optical and data were obtained using the software package NIS (Nikon), which also integrates the control of all hardware units in the microscope. Data was extracted from the ND2 package to personal computers and visualized with Fiji software. Low-magnification images were acquired using Nikon SMZ-25 fluorescent stereoscope equipped with ×1 and ×2 objectives.

Statistical analysis

Due to the nested design of this experiment, wherein we sampled multiple slices from individual mice, we fit the data with a linear mixed-effects model (LMEM), where ∆F/F0 from each slice was modeled as a sum of a fixed effect product (e.g., drug vs. ACSF) plus uncorrelated random effects due to each individual mouse. When fitting the Lorentzian, the model was of a product of two fixed effect, which are the treatment and the distance. To fit the LMEMs to the data, we used the MATLAB (Mathworks) FITLME command. We report the number of mice, the number of slices, Student’s tdof statistic (dof = degrees of freedom) and two-sided P values. The null hypotheses were rejected if the P value was below 0.05. For the bootstrap analysis of spike latency, recordings were obtained from 9 cells across 5 animals. The latency of the first spike from the onset of each recording was measured and compared against a bootstrap distribution generated by sampling spike times from the first 5 seconds of the recording 100,000 times. LED-evoked spike latencies and spontaneous first-spike latencies were each tested against this bootstrap distribution (Supplementary Fig. 3g).

Reagents used

Mecamylamine hydrochloride (mec), Sigma-Aldrich, Lot # 019M4108V CAS: 826-39-1; Atropine sulfate salt monohydrate, Sigma-Aldrich, Lot # BCBH8339V CAS No.: 5908-99-6; SR 95531 hydrobromide (Gabazine), Hello Bio, CAS: 104104-50-9; DNQX, Tocris, CAS: 2379-57-9; D-AP5, Hello Bio, CAS:79055-68-8; CGP 55845 hydrochloride, Hello Bio, CAS: 149184-22-5; Dihydro-β-erythroidine hydrobromide, Tocris, CAS: 29734-68-7; Citalopram hydrobromide (HB2142), Hello Bio, CAS [59729-32-7]; Dopamine hydrochloride (DA), Sigma-Aldrich, CAS No.: 62-31-7; RS 23597-190, Tocris, CAS No.: 149719-06-2. Serotonin (5HT), Sigma-Aldrich, CAS No.: 153-98-0.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.