Synaptotagmin-1 is a Ca²⁺ sensor for somatodendritic dopamine release

SUMMARY

Modes of somatodendritic transmission range from rapid synaptic signaling to protracted regulation over distance. Somatodendritic dopamine secretion in the midbrain leads to D2 receptor-induced modulation of dopamine neurons on the timescale of seconds. Temporally imprecise release mechanisms are often presumed to be at play, and previous work indeed suggested roles for slow Ca²⁺ sensors. We here use mouse genetics and whole-cell electrophysiology to establish that the fast Ca²⁺ sensor synaptotagmin-1 (Syt-1) is important for somatodendritic dopamine release. Syt-1 ablation from dopamine neurons strongly reduces stimulus-evoked D2 receptor-mediated inhibitory postsynaptic currents (D2-IPSCs) in the midbrain. D2-IPSCs evoked by paired stimuli exhibit less depression, and high-frequency trains restore dopamine release. Spontaneous somatodendritic dopamine secretion is independent of Syt-1, supporting that its exocytotic mechanisms differ from evoked release. We conclude that somatodendritic dopamine transmission relies on the fast Ca²⁺ sensor Syt-1, leading to synchronous release in response to the initial stimulus.

In brief

Lebowitz et al. show that synaptotagmin-1 is important for somatodendritic dopamine release triggered by single action potentials but not for spontaneous release or release evoked by high-frequency trains. Hence, multiple Ca²⁺ sensors mediate actionpotential-evoked somatodendritic release in the midbrain with crucial roles for the fast, synchronous sensor synaptotagmin-1.

Graphical Abstract

INTRODUCTION

Neuronal somata and dendrites are equipped with machinery to detect, transduce, and process neurotransmitter signals. However, many neurons also release neurotransmitters and neuromodulators through vesicular exocytosis from these compartments.^1,2 Somatodendritic exocytosis is involved in processes ranging from synaptic-like local signaling to paracrine regulation over long distances.^3–5 Distinct release profiles may be established by varying individual components of the exocytotic pathway, for example the transmitter-containing vesicles, the protein machinery that triggers exocytosis, or the structural organization of secretory sites at the somatodendritic membrane. The molecular organization and function of the various somatodendritic release pathways may be shared or specific depending on the neurotransmitter, necessitating detailed investigation of each transmitter.

An important form of somatodendritic exocytosis is that of dopamine in the midbrain, which underlies a transmission mode that controls the excitability of dopamine neurons.^6–10 We here focused on dissecting the Ca²⁺-triggering mechanisms of somatodendritic dopamine release. At conventional synapses, release has been classified by its temporal relationship to action potential firing. Synchronous release is completed within milliseconds of presynaptic depolarization and is mediated by the fast Ca²⁺ sensors synaptotagmin-1 (Syt-1), Syt-2, and Syt-9.^11–15 Asynchronous release relies on the buildup of residual Ca²⁺, persists for more than tens of milliseconds, and is triggered by additional Ca²⁺ sensors that include Syt-7.^16–20 Spontaneous release is independent of action potential firing, and its reliance on Ca²⁺ entry has remained uncertain, yet Ca²⁺ sensors including Syt-1 are important regulators of this form of release.^15,20–24

In somatodendritic dopamine transmission, activation of D2 G protein-coupled receptors on dopamine neurons elicits an inhibitory postsynaptic current (D2-IPSC) that is carried by K⁺ ions through the activation of G protein-gated inwardly rectifying K⁺ (GIRK) channels.^7,9,10 A prevailing model has been that somatodendritic dopamine exocytosis uses distinct protein machinery compared with axonal dopamine release.^25,26 Specifically, while synchronous dopamine release from axons relies on Syt-1 and is rapid and precise,^27–29 somatodendritic release may have a different Ca²⁺ dependence, may rely on slower Ca²⁺ sensors, and may depend on Syt-7 and Syt-4.^30–33 This is consistent with some, but not all, aspects of somatodendritic dopamine signaling. For example, D2 receptor activation necessitates ~100 μM dopamine,³⁴ suggesting the need for synchronous exocytosis such that the extracellular dopamine concentration increases rapidly. In agreement, knockout of RIM, a protein that co-organizes release-ready vesicles with Ca²⁺ entry to accelerate release,^35–41 strongly impairs evoked somatodendritic dopamine transmission.⁴² These data suggest rapid somatodendritic dopamine release mechanisms, but a suitable fast Ca²⁺ sensor has not been identified with confidence.

Here, we find that the fast Ca²⁺ sensor Syt-1 is important for evoked somatodendritic dopamine release. Conditional knockout of Syt-1 from dopamine neurons strongly reduced D2-IPSCs elicited by single stimuli, likely due to a decrease in vesicular release probability. This deficit was restored by viral re-expression of Syt-1. High-frequency stimulus trains induced robust D2-IPSCs after Syt-1 ablation, indicating that an additional release pathway is activated. Lastly, Syt-1 knockout did not impair spontaneous D2-IPSCs and neither did lowering the extracellular Ca²⁺ concentration. We conclude that Syt-1 is a Ca²⁺ sensor for synchronous somatodendritic dopamine release. Additional sensors must be present in these mice to mediate transmission during repetitive activity after Syt-1 deletion, likely via asynchronous release. Spontaneous somatodendritic dopamine release is independent of Ca²⁺ and of the fast Ca²⁺ sensor Syt-1.

RESULTS

Syt-1 knockout strongly impairs somatodendritic dopamine release evoked by single stimuli

Of the known Ca²⁺ sensors, Syt-1 and Syt-7 are expressed in dopamine neurons.^43,44 We recently established that Syt-1 is a fast Ca²⁺ sensor for triggering axonal dopamine release.²⁹ Based on these data and on the properties of somatodendritic dopamine transmission, we hypothesized that Syt-1 triggers somatodendritic dopamine secretion. We performed electrophysiological recordings in brain slices (Figure 1A)^10,42 of mice with genetic ablation of Syt-1 selectively in dopamine neurons. We generated these mice though intercrossing of conditional, floxed Syt-1 mice⁴⁵ with DAT^IRES—Cre mice⁴⁶ (Figure 1B) as described before.^29,47 In the following experiments, conditional knockout (cKO) mice for Syt-1 (Syt-1 cKO^DA) are mice with two floxed alleles for Syt-1 and a heterozygote DAT^IRES—Cre allele, and Syt-1 control mice are littermate or age-matched mice from the same crossings with two wild type alleles for Syt-1 and a heterozygote DAT^IRES—Cre allele. The basal electrical properties of dopamine neurons, including the tonic firing rate, series resistance, and capacitance, were not altered in brain slices of Syt-1 cKO^DA mice (Figure S1).

Figure 1. Somatodendritic dopamine release is strongly decreased in Syt-1 cKO^DA mice.

(A and B) Schematic of recording (A) in the substantia nigra pars compacta (SNc) and of Syt-1 cKO^DA mice (B) generated as described before.²⁹ Exon numbers refer to coding exons.

(C–E) Example traces (C) and quantification of amplitude (D) and charge (E, area under the curve) of D2-IPSCs evoked by a single stimulus. Syt-1 control 20 cells/6 mice, and Syt-1 cKO^DA 23 cells/6 mice.

(F–H) Example traces of D2-IPSCs evoked by two electrical stimuli with a 1 s interval (F), and quantification of the first peak amplitude (G) and ratio of the amplitude of the second to the first D2-IPSC (H, paired pulse ratio). Syt-1 control 17 cells/7 mice, and Syt-1 cKO^DA 10 cells/4 mice. Data are mean ± SEM; ***p < 0.005 as determined by Mann Whitney rank-sum tests (D, E, and G) or an unpaired Student’s t test (H). For basal electrical properties, see Figure S1; for D2-IPSC rise times, see Figure S2.

We determined if evoked somatodendritic dopamine release relies on Syt-1 (Figures 1 and S2). Release was monitored via D2-IPSCs, and we focused on the substantia nigra pars compacta (SNc) because axon collaterals are likely absent in the SNc, different from the ventral tegmental area (VTA).^48–50 Single electrical stimuli were applied through a stimulation electrode that was positioned ~50–100 μm away from a dopamine neuron recorded in whole-cell voltage clamp and held at −55 mV (Figure 1A). The average D2-IPSC amplitude and charge transfer evoked with a single stimulus were decreased by ~80% in Syt-1 cKO^DA mice, establishing that evoked somatodendritic dopamine release relies on Syt-1 (Figures 1C–1E). We next quantified the ratio of D2-IPSCs in response to two consecutive stimuli (paired pulse ratios) at 1 Hz (Figures 1F–1H). Previous work has shown that with paired stimuli, the second D2-IPSC is strongly depressed, similar to axonal dopamine release measured with different approaches.^27,51,52 In slices of Syt-1 control mice, the ratio was 0.5 and was increased to 0.8 in Syt-1 cKO^DA mice (Figures 1G and 1H). In synaptic physiology, paired pulse ratios are often used to estimate vesicular release probability, and they are inversely correlated with it.⁵³ Following this model, somatodendritic dopamine release has a high vesicular release probability that is mediated by the presence of Syt-1. Altogether, these experiments establish that evoked somatodendritic dopamine release is strongly decreased after Syt-1 ablation and suggest that Syt-1 is an important Ca²⁺ sensor for it.

Syt-1 re-expression restores somatodendritic dopamine release in Syt-1 cKO^DA mice

It is possible that Syt-1 is needed early in dopamine neuron development for maturation of the somatodendritic release machinery but that it does not directly act as a Ca²⁺ sensor for secretion. We sought to address this possibility by restoring somatodendritic release through re-expression of Syt-1 in adult Syt-1 cKO^DA mice. We constructed an adeno-associated virus (AAV) for Cre-dependent co-expression of Syt-1 and the fast channelrhodopsin variant oChIEF-citrine⁵⁴ using a 2A-based strategy (Figure 2A). The virus was produced as serotype 9 (AAV9) and was injected into the midbrain of Syt-1 cKO^DA mice to generate Syt-1 cKO^DA + Syt-1 mice. Syt-1 control mice were injected with a virus for Cre-dependent oChIEF-citrine expression (Figure 2A).

Figure 2. Syt-1 re-expression restores somatodendritic dopamine release in Syt-1 cKO^DA mice.

(A) Strategy of rescue experiments.

(B–D) Example traces (B) and quantification of amplitude (C) and paired pulse ratio (D) of optogenetically induced D2-oIPSCs (two stimuli at 0.4 Hz) in control mice expressing oChIEF-citrine (Syt-1 control) and Syt-1 cKO^DA mice co-expressing oChIEF-citrine and Syt-1 (Syt-1 cKO^DA + Syt-1). Syt-1 control 18 cells/6 mice, and Syt-1 cKO^DA + Syt-1 13 cells/5 mice.

(E–G) As in (B)–(D) but for pairs of electrical stimuli (at 1 Hz) in the same cells; n as in (B)–(D).

Data are mean ± SEM; no significant differences were detected with Mann Whitney rank-sum tests (C, D, and F) or an unpaired Student’s t test (G).

After 20 to 30 days of expression, we performed slice recordings in SNc areas with strong oChIEF-citrine fluorescence. To verify that the recorded neuron received inputs from dopamine neurons expressing the virus, two pulses of blue light were applied 2.5 s apart, and the optically induced D2 receptor-mediated IPSCs (D2-oIPSCs) were quantified (Figures 2B–2D). D2-oIPSCs were readily detected in both conditions, and the paired pulse ratio was indistinguishable. Hence, Syt-1 re-expression might restore somatodendritic dopamine release in Syt-1 cKO^DA mice. To test this, we measured electrically evoked D2-IPSCs (Figures 2E–2G) from the same cells used to record D2-oIPSCs (Figures 2B–2D). Robust D2-IPSCs were present in both conditions, and the paired pulse ratio was indistinguishable between Syt-1 control and Syt-1 cKO^DA + Syt-1 mice. In these analyses, we only included neurons with a D2-oIPSC equal to or greater than the D2-IPSC generated by a single electrical stimulus to select for cells with robust viral expression in the inputs. Altogether, the data indicate that Syt-1 re-expression restores most release in Syt-1 cKO^DA mice and support that Syt-1 operates as a Ca²⁺ sensor for somatodendritic dopamine secretion.

High-frequency trains induce robust somatodendritic release in Syt-1 cKO^DA mice

At conventional synapses, stimulus trains induce asynchronous glutamate or GABA release after Syt-1 knockout, and this asynchronous release is often sufficient to maintain synaptic charge transfer.^55,56 A release component that is Syt-1 independent was also detected for axonal dopamine release after Syt-1 ablation, but its detection with amperometry in striatal slices required blocking dopamine reuptake or strong depolarization via KCl.²⁹ In vivo microdialysis, a method that samples over long time periods and large striatal space, revealed robust in vivo dopamine levels in Syt-1 cKO^DA mice.²⁹ These observations might suggest that asynchronous release is sufficient to maintain dopamine secretion in response to repetitive stimulation after removing the fast Ca²⁺ sensor.

We probed whether repetitive stimulation induced somatodendritic dopamine release in the absence of Syt-1. First, we evoked D2-IPSCs with 5 pulses at 40 Hz in Syt-1 cKO^DA and control mice (Figures 3A–3D). These stimuli induced robust D2-IPSCs in Syt-1 cKO^DA slices, revealing that short trains restore a substantial amount of dopamine release. When we compared the relative increase of the 5-pulse D2-IPSC (Figure 3B) with the single-pulse D2-IPSC (Figure 1D, recorded from the same cells), we found a significantly greater increase in Syt-1 cKO^DA mice (Figure 3D). In Syt-1 control neurons, the 5-pulse D2-IPSC was ~3-fold greater than the single-pulse D2-IPSC, consistent with the depression observed in response to paired stimuli (Figure 1G). In Syt-1 cKO^DA neurons, the 5-pulse D2-IPSC was increased ~9-fold over the single-pulse D2-IPSC, establishing a strong enhancement. These data indicate that repetitive stimulation, likely via buildup of residual Ca²⁺, triggers release through engaging additional Ca²⁺ sensors, at least in the Syt-1 cKO^DA mice.

Figure 3. Somatodendritic dopamine release in response to stimulus trains.

(A–D) Example traces (A) and quantification of amplitude (B) and charge (C, area under the curve) of D2-IPSCs evoked with a train of five electrical stimuli at 40 Hz, and ratio of the 5- to 1-stimulus D2-IPSC (D, single-stimulus amplitudes are from Figure 1C and were recorded in the same cells). Syt-1 control 20 cells/6 mice, and Syt-1 cKO^DA 23 cells/6 mice.

(E–H) Example traces of D2-IPSCs evoked by 5 stimuli at 1 Hz followed by a 5-stimulus train at 100 Hz (E), and quantification of the summed 5-stimulus D2-IPSC charge at 1 Hz (F), and the 5-stimulus amplitude (G) and charge (H) at 100 Hz. Syt-1 control 8 cells/3 mice, and Syt-1 cKO^DA 6 cells/2 mice.

Data are mean ± SEM; *p < 0.05, ***p < 0.005 as determined by Mann Whitney rank-sum tests (B–D) or unpaired Student’s t tests (F–H).

The D2-IPSC can be depressed with 1-Hz stimulation, and it builds up rapidly when the stimulation frequency is switched to 100 Hz following this depression.⁵¹ We investigated this by comparing Syt-1 cKO^DA with Syt-1 control mice (Figures 3E–3H). The overall charge during a 5-stimulus 1-Hz train was moderately decreased in Syt-1 cKO^DA slices compared with controls (Figure 3F). The D2-IPSCs evoked by 100-Hz 5-stimulus trains that followed the fifth stimulus at 1 Hz were comparable between Syt-1 cKO^DA and control slices (Figures 3G and 3H), similar to the 5-stimulus D2-IPSC at 40 Hz (Figures 3A–3D). These data support that a release pathway that is less dependent on Syt-1 is activated at high stimulus frequencies. Another fast, synchronous Ca²⁺ sensor is unlikely to mediate these D2-IPSCs because this sensor would be activated by single stimuli, which is largely excluded by the strong impairment of single-pulse D2-IPSCs (Figure 1). Instead, a slower, higher-affinity sensor is likely at play, possibly Syt-7 via mediating asynchronous release.³¹

Spontaneous somatodendritic dopamine release is Syt-1 independent

Midbrain dopamine neurons have spontaneous somatodendritic release that persists in the absence of action potential firing.^42,57 This is similar to conventional synapses, where miniature release events occur, are suppressed by the fast Ca²⁺ sensor Syt-1, and may be regulated by alternate Ca²⁺ sensors.^15,20–24 The exocytotic pathway for spontaneous somatodendritic dopamine release has not been well characterized, but it is vesicular and occurs at fusion sites that do not require the release site scaffold RIM.^42,57

With the goal to improve detection of spontaneous D2-IPSCs, we overexpressed D2 receptors in the midbrain with AAVs (Figure 4A). We first tested whether evoked release is altered by D2 receptor overexpression (Figure S3). In both control and Syt-1 cKO^DA neurons, evoked D2-IPSCs and their paired pulse ratios were similar to the ones without overexpression (Figures 1, S3). D2 receptor overexpression greatly facilitated the detection of spontaneous events compared with previously published studies (Figure 4B).^42,57 We measured spontaneous D2-IPSCs at baseline, in low extracellular Ca²⁺, and after blockade of action potential firing. In Syt-1 cKO^DA neurons, the frequencies and amplitudes of spontaneous D2-IPSCs were similar to controls (Figures 4B–4D). In both genotypes, reducing extracellular Ca²⁺ from 2.4 to 0.5 mM did not decrease event frequency or amplitude, nor did blockade of sodium channels with 1 μM TTX. We conclude that spontaneous somatodendritic dopamine release is independent of Syt-1. This is surprising because Syt-1 removal typically increases spontaneous event frequency at regular synapses.^15,22,24,58 These results underscore that evoked and spontaneous somatodendritic fusion events may not use the same release machinery⁴² and bolster the model that distinct secretory pathways mediate evoked and spontaneous somatodendritic dopamine release.

Figure 4. Spontaneous somatodendritic dopamine release is not detectably impaired by Syt-1 cKO^DA.

(A) Strategy for AAV-mediated overexpression of D2 receptors (human, short version [hD2R_S]) in the midbrain.

(B–D) Example traces (B) and quantification of frequency (C) and amplitude (D) of spontaneous D2-IPSCs recorded in brain slices after viral expression of D2 receptors in basal recording conditions, after reducing extracellular Ca²⁺ from 2.4 to 0.5 mM, or following blockade of sodium channels with 1 μM TTX. Basal: Syt-1 control 17 cells/7 mice, and Syt-1 cKO^DA 18 cells/5 mice; 0.5 mM Ca²⁺: Syt-1 control 9 cells/4 mice, and Syt-1 cKO^DA 10 cells/5 mice; 1 μM TTX: Syt-1 control 7 cells/6 mice, and Syt-1 cKO^DA 7 cells/4 mice.

Data are mean ± SEM; no significant differences were detected with two-way ANOVA. For evoked D2-IPSCs after D2 receptor overexpression, see Figure S3.

DISCUSSION

We assessed Ca²⁺-sensing mechanisms for the triggering of somatodendritic dopamine release and found that the fast Ca²⁺ sensor Syt-1 is important. Genetic ablation of Syt-1 in dopamine neurons strongly decreased D2-IPSCs evoked by a single stimulus, and viral re-expression of Syt-1 restored D2-IPSCs. D2-mediated transmission depresses during repetitive stimulation, supporting the model of a high initial release probability. After knockout of Syt-1, release during high-frequency stimulus trains was nearly normal, suggesting that Syt-1-independent release mechanisms respond to Ca²⁺ buildup. Spontaneous D2-IPSCs were unaffected by ablation of Syt-1, adding to a growing body of evidence indicating that spontaneous and evoked dopamine release are mediated by separate exocytotic pathways.

Comparison with previous work on somatodendritic and axonal release

Our work relates to earlier and recent studies on somatodendritic dopamine release. Based on the distinct Ca²⁺ dependence of somatodendritic and axonal dopamine release, and relying on knockdown experiments in cultured dopamine neurons, it was proposed that Syt-4 and Syt-7 are involved.^32,59 Indeed, constitutive double knockout of Syt-4 and Syt-7, but not knockout of Syt-7 alone, impaired somatodendritic dopamine release measured by voltammetry in brain slices.³⁰ It is noteworthy that vertebrate Syt-4 is unlikely a Ca²⁺ sensor as it lacks key Ca²⁺-binding sites.^60,61 Two recent studies have confirmed roles for Syt-7 in somatodendritic release. One study found that the strong depression of D2-IPSCs during stimulus trains was alleviated by constitutive knockout of Syt-7.⁶² The other³¹ used Syt-7 antibodies or constitutive Syt-7 knockout to assess D2-IPSCs that may arise from autocrine signaling in which a dopamine neuron senses its own released dopamine.⁶³ Both manipulations impaired this form of transmission during short stimulus trains. Furthermore, Syt-1 antibodies applied through the recording pipette impaired this autocrine transmission in response to a single stimulus, but it remained uncertain whether the antibody blocked Syt-1 function and whether Syt-1 controls the more commonly detected non-autocrine D2-IPSCs that we assessed here. Finally, Syt-1 knockout mice were also studied with voltammetry using either optogenetic or electrical stimulation, and mixed effects were observed with uncertain contributions of Syt-1 to somatodendritic release depending on the stimulation method.⁶⁴ Altogether, these recent studies support roles for Syt-7 in somatodendritic release. However, they did not allow for a decisive identification of the necessity and identity of the synchronous Ca²⁺ sensor for somatodendritic release.

The effect of Syt-1 deletion on evoked somatodendritic dopamine release (Figures 1, 2, and 3) resembles release deficits seen at classical synapses. At hippocampal synapses, Syt-1 ablation strongly impairs synchronous release, but the charge transfer during stimulus trains is maintained, and the resulting transmission is sufficient to support at least some aspects of behavior, for example memory acquisition.^55,56 The reliance on a fast Ca²⁺ sensor for midbrain dopamine transmission is counterintuitive since the speed of information transfer of this transmission mode is limited by G protein signaling, making it much slower than fast synaptic transmission.^9,65,66 A possible explanation is that synchronous dopamine release is needed to establish the high concentration of dopamine, ~100 μM, that is necessary for D2 receptor activation.^34,67

Distinct release pathways for multiple modes of somatodendritic dopamine transmission

Our work defines the Ca²⁺-triggering mechanism for synchronous somatodendritic dopamine release in response to the initial stimulus. The data also show that during stimulus trains, dopamine release recovers in the absence of Syt-1. Based on these findings, and building on previous work, we propose that somatodendritic dopamine release occurs through three exocytotic pathways. The first pathway is fast and synchronous and relies on the Ca²⁺ sensor Syt-1 (Figures 1 and 2) and the release site organizer RIM.⁴² The second pathway is important during repetitive activity, for example burst firing, and it does not depend on Syt-1 (Figure 3). Secretion of this second pathway occurs at release sites organized by RIM, as RIM ablation abolishes it.⁴² Some studies suggest that Syt-7 is involved.^30–32,62 This transmission mode is likely mediated by asynchronous release in response to build up of Ca²⁺ during repetitive activity, consistent with roles of Syt-7, a Ca²⁺ sensor for asynchronous release at synapses.^16,17,20 Finally, previous work^42,57 and our data (Figure 4) uncover the presence of a spontaneous dopamine release pathway that is independent of action potential firing and is not sensitive to lowering extracellular Ca²⁺ or to removing Syt-1 or RIM.

Together, these studies reveal the distinct involvement of specific proteins, Syt-1, Syt-7, and RIM, in synchronous, asynchronous, and spontaneous somatodendritic dopamine release. It is interesting to assess how these distinct molecular requirements could reflect distinct cell biological pathways for secretion. While Syt-1 is a key component of synaptic vesicles,^68,69 Syt-7 is not localized on synaptic vesicles but instead at the target membrane or on different intracellular membrane compartments.^70,71 The identity of the intracellular compartment for somatodendritic dopamine secretion remains uncertain. Early studies established that somatodendritic dopamine release depends on SNARE proteins^25,72 and on Ca²⁺,^{10,32,33,59,73,74} indicating that exocytosis underlies this form of release. Classical work further suggested that tubulo-vesicular structures, rather than small synaptic vesicles, may be involved.^75–77 Together, these studies and our work raise the possibility that multiple membrane compartments mediate somatodendritic dopamine release: small synaptic vesicles with Syt-1 may account for the initial synchronous component, while release in response to high-frequency trains, and possibly spontaneous release, may be supported by tubulo-vesicles or other compartments. This model may also explain why the asynchronous component is difficult to detect in experiments assessing axonal release.²⁹ It is possible that axonal release is dominated by small clear vesicles, which densely occupy the dopamine axon.^78–80

Limitations of the study

Our work reveals important roles for Syt-1 in somatodendritic dopamine release. One limitation to the interpretation of our work is that morphological characterization of the underlying signaling structure remains difficult. Release site architecture and release-receptor assemblies have not been visualized, the intracellular membrane compartments that undergo exocytosis remain uncertain, and it was not possible to determine whether exogenously expressed proteins, for example D2 receptors or Syt-1, localize like the endogenous counterparts. In the VTA, but less likely in the SNc, axon collaterals might be present as well. Hence, models on the molecular and cellular organization of this release mode can currently only be inferred from functional experiments like the one presented here, characterizing transmission deficits after gene knockout. Localizing Syt-1 and other proteins for somatodendritic transmission remains a future goal. In addition, the D2-IPSC measurements used here differ from axonal release measurements with electrochemical detection and from measurements of synaptic transmission monitored via ionotropic receptors. It is difficult to directly compare effect magnitudes across methods, and it remains uncertain whether classical concepts of synaptic transmission, for example on the inverse relationship of paired pulse ratios and vesicular release probability, fully apply. Nevertheless, our experiments establish that Syt-1 is critical for somatodendritic dopamine release evoked by single action potentials.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead contact, Pascal S. Kaeser (kaeser@hms.harvard.edu).

Materials availability

Plasmids generated for this study will be shared within limits of respective material transfer agreements. AAVs are exhaustible and will be shared as long as they are available and within limits of respective material transfer agreements. Mouse mutant alleles are publicly available as outlined in the key resources table.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
(+)-MK 801 maleate	Hello Bio	Cat# HB0004
NBQX	Hello Bio	Cat# HB0442
CGP55845 hydrochloride	Hello Bio	Cat# HB0960
Picrotoxin	Hello Bio	Cat# HB0506
Tetrodotoxin citrate	Tocris	Ca# 1069
Experimental models: Cell lines
HEK293T cells	ATCC	Cat#: CRL-3216 RRID: CVCL_0063
Experimental models: Organisms/strains
Mouse: Syt-1 floxed; generated from C57BL/6Ntac-Syt1tm1a(EUCOMM)Wtsi/WtsiCnrm	Zhou et al.45	EMMA ID EM:06829, RRID:IMSR_EM:06829; the identifier refers to the line before flp recombination
Mouse: B6.SJL-Slc6a3tm1.1(cre)Bkmm/J; DATIRES–Cre	Backman et al.46	JAX 006660, RRID: IMSR_JAX:006660
Recombinant DNA
pAAV-hSyn-FLEX-oChIEF-citrine-P2A-Syt-1	This paper; Kaeser lab, plasmid was used to generate AAVs.	lab plasmid code (LPC): p869
pAAV-hSyn-FLEX-oChIEF-citrine	Addgene, Lin et al.,54 plasmid was used to generate AAVs.	Plasmid# 50973, RRID:Addgene 50973; LPC: p901
pFB-CAG-SNAP-hD2Rshort-P2A-GFP	This paper; ViroVek, plasmid was used to generate AAVs.	LPC: p1038
Software and algorithms
AxoGraph	AxoGraph	RRID:SCR_014284; https://www.axograph.com
LabChart	AD Instruments	https://www.adinstruments.com/products/labchart
Prism 9	GraphPad	RRID: SCR_002798; https://www.graphpad.com

Data and code availability

• Data reported in this paper will be shared by the lead contact upon request.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animals

Conditional knockout mice for Syt-1 in dopamine neurons were generated through mouse breeding as described before,²⁹ crossing Syt-1^floxed mice⁴⁵ with DAT^IRES—Cre mice,⁴⁶ and striatal phenotypes were described.^29,47 Syt-1^floxed mice were generated from a gene targeting experiment (C57BL/6Ntac-^{Syt1tm1a(EUCOMM)Wtsi}/WtsiCnrm) for coding exon 2 (if non-coding exons are included in the numbering, the corresponding exon number is exon 6),^45,81 and further characterized in.^82–84 In all experiments, Syt-1 cKO^DA mice are mice homozygote for Syt-1 floxed (Syt-1^f/f) alleles and with a heterozygote allele for DAT^IRES—Cre; Syt-1 control mice are littermate mice or age-matched mice from the same breedings with homozygote Syt-1 wild type alleles (Syt-1^+/+) and a heterozygote allele for DAT^IRES—Cre. Male and female adult mice (25–45 weeks old) were used for all experiments. Genotype comparisons were performed by an experimentalist blind to genotype, and unblinding of genotypes was done after experiments and analyses were completed. All animal experiments were performed according to protocols approved by the respective Animal Care and Use Committee at Harvard University and at Oregon Health and Science University.

Cell lines

HEK293T cells were purchased from ATCC (CRL-3216, RRID: CVCL_0063, an immortalized, authenticated cell line of female origin). The cells were expanded and stored as stock vials in liquid nitrogen. They were thawed for use and were grown in Dulbecco’s Modified Eagle Medium (DMEM) with 10% fetal bovine serum (Atlas Biologicals F-0500-D) and 1% Penicillin-Streptomycin. The cells were grown in a tissue culture incubator set to 37°C and 5% CO₂, and they were passaged every 2 to 3 days at a ratio of 1:5 to 1:7. Each batch of cells was replaced after ~20 passages with cells thawed from a fresh stock vial.

METHOD DETAILS

Molecular biology and production of AAV viruses

Adeno-associated viruses (AAVs) were either generated in the lab or by ViroVek (Hayward, CA). All AAVs were of the AAV9 serotype and genomic titers ranged from 1.43 × 10¹² to 2.15 × 10¹³ viral genome copies/mL. Cre-dependent AAVs were injected in the substantia nigra for expression of oChIEF-citrine with or without Syt-1. For expression oChIEF-citrine alone, pAAV-hSyn-FLEX-oChIEF-citrine⁵⁴ (RRID:Addgene_50973, LPC: p901) was used; the following protein is encoded (numbered in subscript according to amino acid sequences of GenBank KF437396.1 for oChIEF and GenBank MK301208.1 for citrine): _oChIEF,1MVSR … YESS₃₅₉-LE-_citrine,1MVSKG … ELYK₂₃₉. For co-expression of oChIEF and synaptotagmin-1, we constructed an AAV vector, pAAV-hSyn-FLEX-oChIEF-citrine-P2A-Syt-1 (LPC: p869) producing citrine-tagged oChIEF and rat synaptotagmin-1. The following two proteins are encoded (numbered in subscript according to amino acid sequences of GenBank KF437396.1 for oChIEF, GenBank MK301208.1 for citrine, and NCBI Ref Seq NM_001033680.2 for Syt-1): _oChIEF,1MVSR … YESS₃₅₉-LEG-_citrine,2VSKG … ELYK₂₃₉-GAPGSGATNFSLLKQAGDVEENPG and PAPL-_Syt-1,2VSAS … AVKK₄₂₁. AAVs were generated in HEK293T cells in the lab using calcium phosphate transfection. Cells were collected and lysed 72 h after transfection. AAV viral particles were extracted and purified from the 40% layer after iodixanol gradient ultracentrifugation. An AAV encoding the short version of the human D2 receptor (hD2R_S) was generated by ViroVek (Hayward, CA). A plasmid encoding an SNAP-tagged (at the N-terminus) hD2R_S and a cytosolic GFP (via a P2A sequence) was generated and provided to ViroVek. This plasmid was subcloned to generate pFB-CAG-SNAP-hD2Rshort-P2A-GFP (LPC: p1038). The following two proteins are encoded (numbered in subscript according to amino acid sequences of GenBank KM043782.1 for the SNAPtag, NCBI Ref Seq NM_016574.4 for hD2Rshort, and GenBank MF405194.1 for GFP): MKTIIALSYIFCLVFADYKDDDDA-_SNAPtag,172MDKD … PGLG₃₅₃-_hD2Rshort,1MDPL … ILHC₄₁₄-SGSGATNFSLLKQAGDVEENPG and P-_GFP,1MVSK … ELYK₂₃₉. After verification, the plasmid was used by ViroVek for AAV production.

Stereotaxic viral delivery

Mice (postnatal days 56–115) were anesthetized and placed in a stereotaxic frame for midbrain injections of AAV viruses. Anesthesia was maintained throughout surgery with isoflurane. 120 nL of AAV was injected bilaterally into the ventral midbrain at a rate of 0.1 μL/min using an UltraMicroPump3 microinjection syringe pump (World Precision Instruments, Sarasota FL) controlled by a Wizard 211 digital readout system (ANILAM Inc., Jamestown NY). Injection coordinates were (in mm): 2.3 posterior from bregma; ± 1.3 lateral from midline; 4.5 below pia. Post-operative analgesia was provided following approved protocols. Experiments with AAV-mediated D2 receptor expression were performed 10 to 20 days after surgery. Experiments with oChIEF-citrine expression (and for the rescue condition Syt-1 co-expression) were performed 20 to 30 days after surgery.

Brain slice preparation

Mice (25–45 weeks old) were anesthetized with isoflurane and decapitated, and the brain was collected in 32–35°C warm Krebs buffer containing (in mM): 126 NaCl, 2.5 KCl, 1.2 MgCl₂, 2.4 CaCl₂, 1.4 NaH₂PO₄, 25 NaHCO₃, and 11 Dextrose. The Krebs buffer used for extraction, cutting, and recovery contained 10 μM MK-801 to prevent NMDA-mediated excitotoxic damage. 222 μm thick horizontal slices containing the midbrain were cut using a vibrating microtome in warm Krebs buffer bubbled with 95% O₂/5% CO₂ and slices were allowed to recover in the same buffer at 30°C for ≥30 min before recordings. Prior to recordings, slices were hemisected along the midline to record from each side separately. For recording, the hemisected slice was transferred to the recording chamber, which was continuously perfused at 2 to 3 mL/min with bubbled Krebs buffer at 34°C. All experiments were completed within 7 h of slicing.

Electrophysiological recordings in brain slices

SNc dopamine neurons were identified by their morphology and by their localization lateral to the medial terminal nucleus of the accessory optic system. D2-IPSCs were recorded following previously established protocols.^10,42,57 Specifically, experiments were conducted in Krebs buffer with the addition of NBQX (600 nM), CGP55845 (300 nM), and picrotoxin (100 μM). Recordings were made using glass pipettes (1.0–2.5 MΩ) filled with an internal solution containing (in mM): 100 K-methanesulphonate, 20 NaCl, 1.5 MgCl₂, 10 HEPES (K), 2 ATP, 0.2 GTP, 10 phosphocreatine, and 10 BAPTA (4K). The recorded cells exhibited ~1–5 Hz spontaneous firing as assessed in cell-attached configuration, and firing rates were recorded and analyzed in a subset of experiments in cell attached mode with an amplitude threshold. After break-in, cells were voltage-clamped at −55 mV and resistance and capacitance were monitored; cells with a series resistance ≥12 MΩ were discarded. A focal monopolar or bipolar stimulation electrode was used to stimulate the cells with a current of ~50–250 μA once per min. The typical distance of the stimulation electrode to the recorded cell was 50–100 μm. For assessment of evoked release, the stimulation intensity was set such that a D2-IPSC was detected in response to five stimuli at 40 Hz, and only cells with a detectable 5-stimulus D2-IPSC were included. The number and frequency of electrical stimuli varied, and the stimulation protocol is described in each figure; if multiple stimulus protocols were used for a given cell, the order was randomized but stimulation intensity was kept identical across protocols. For optogenetic stimulation, a 470 nm LED (Thorlabs) was used to deliver paired flashes of blue light (2 mW, 25 ms, 0.4 Hz) through a 40× water-immersion objective directly over the recorded cell. As with electrical stimulation, paired light flashes were delivered once per minute. All data reported for stimulated D2-IPSCs reflect the average of three consecutive sweeps. Analyses of spontaneous D2-IPSCs were constrained to the intervals between stimuli or after stimulus protocols were completed. These recordings were performed in the same solutions as described above for most cells; for a subset of cells, 1 μM TTX was added or the extracellular solution contained 0.5 mM CaCl₂ and 5 mM MgCl₂ instead of 2.4 CaCl₂ and 1.2 MgCl₂. Data acquisition was performed with AxoGraph (Berkley, CA) and recordings were monitored with LabChart (AD instruments, Colorado Springs, CO). D2-IPSC amplitudes were measured as the average of a 20 ms window centered around the peak outward current determined in AxoGraph. The charge transfer was quantified as the area under the curve starting at stimulation for 1.5 s, at which time the D2-IPSC returned to baseline. Paired pulse ratios were calculated as the ratio of the second D2-IPSC peak to the first. In rescue experiments, we used optogenetic activation to assess the D2-oIPSC for each recorded cell. We then used electrical stimulation to assess rescue and only included cells for which the D2-oIPSC amplitude was equal to or greater than that of the D2-IPSC generated by a single electrical stimulus to select for cells with robust viral expression in the inputs. Spontaneous D2-IPSCs were identified using a sliding template in AxoGraph as described previously.^42,57 The template was based on the shape of the D2-IPSC and used the following parameters: a 50 ms baseline, a rise time (10–90%) of 50 ms, and a decay t of 200 ms. The threshold for event detection was 3.7 times the standard deviation of baseline noise for events that fit the waveform of the sliding template, and events had to be separated by a minimum of 300 ms. For all experiments examining spontaneous D2-IPSCs and several experiments examining evoked D2-IPSCs, sulpiride was added at the end of the recording period to inhibit D2 receptors, which led to a block of spontaneous and evoked D2-IPSCs, respectively, as established before.^10,57 Data acquisition and analyses were conducted by experimenters blind to genotype.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data are displayed as mean ± SEM with individual values as circles, and statistical significance is denoted as *p < 0.05, **p < 0.01, and ***p < 0.005 in the figures. Representative traces except for those of spontaneous D2-IPSCs are the average of three consecutive sweeps. Statistical testing was performed in GraphPad Prism 9 (GraphPad, San Diego, CA). For statistical analyses, each cell was considered an observation because the technical nature of these recordings precluded getting sufficient numbers of cells per mouse for averaging per mouse. Datasets were tested for normality with a Shapiro-Wilk test. For comparison of two groups with a single variable, normally distributed data were assessed with unpaired Student’s t-tests, and non-normally distributed data with Mann-Whitney rank sum tests. For spontaneous D2-IPSC analyses, a two-way ANOVA was used. Sample sizes, statistical tests used, and significance are reported for all data in the corresponding figures and figure legends.

Highlights.

• Somatodendritic dopamine release is assessed in Syt-1 knockout mice via D2-IPSCs

• Single action potential-evoked D2-IPSCs strongly depend on the fast Ca²⁺ sensor Syt-1

• High-frequency trains induce similar D2-IPSCs in Syt-1 knockout and control mice

• Spontaneous somatodendritic dopamine release is independent of Syt-1