Isoflurane Inhibits Dopaminergic Synaptic Vesicle Exocytosis Coupled to CaV2.1 and CaV2.2 in Rat Midbrain Neurons

Abstract Volatile anesthetics affect neuronal signaling by poorly understood mechanisms. Activation of central dopaminergic pathways has been implicated in emergence from general anesthesia. The volatile anesthetic isoflurane differentially inhibits glutamatergic and GABAergic synaptic vesicle (SV) exocytosis by reducing presynaptic Ca2+ influx without affecting the Ca2+-exocytosis relationship, but its effects on dopaminergic exocytosis are unclear. We tested the hypothesis that isoflurane inhibits exocytosis in dopaminergic neurons. We used electrical stimulation or depolarization by elevated extracellular KCl to evoke exocytosis measured by quantitative live-cell fluorescence imaging in cultured rat ventral tegmental area neurons. Using trains of electrically evoked action potentials (APs), isoflurane inhibited exocytosis in dopaminergic neurons to a greater extent (30 ± 4% inhibition; p < 0.0001) than in non-dopaminergic neurons (15 ± 5% inhibition; p = 0.014). Isoflurane also inhibited exocytosis evoked by elevated KCl in dopaminergic neurons (35 ± 6% inhibition; p = 0.0007), but not in non-dopaminergic neurons (2 ± 4% inhibition). Pharmacological isolation of presynaptic Ca2+ channel subtypes showed that isoflurane inhibited KCl-evoked exocytosis mediated exclusively by either CaV2.1 (P/Q-type Ca2+ channels; 30 ± 5% inhibition; p = 0.0002) or by CaV2.2 (N-type Ca2+ channels; 35 ± 11% inhibition; p = 0.015). Additionally, isoflurane inhibited single AP-evoked Ca2+ influx by 41 ± 3% and single AP-evoked exocytosis by 34 ± 6%. Comparable reductions in exocytosis and Ca2+ influx were produced by lowering extracellular [Ca2+]. Thus, isoflurane inhibits exocytosis from dopaminergic neurons by a mechanism distinct from that in non-dopaminergic neurons involving reduced Ca2+ entry through CaV2.1 and/or CaV2.2.


Introduction
General anesthetics are essential medicines that induce a reversible state of amnesia, unconsciousness, and immobility in the face of intensely painful stimuli. Despite their widespread use in modern medicine, their mechanisms of action are not well understood (Hemmings et al., 2005b). The amnestic, hypnotic, and immobilizing effects of anesthetics differ in dose dependence, neuroanatomical regions involved, and molecular targets consistent with multiple mechanisms working in parallel to produce the state of anesthetic-induced unresponsiveness . However, general anesthesia can produce serious adverse side effects, including cardiovascular, respiratory, and cognitive dysfunction. It is therefore critical to identify the anesthetic mechanisms relevant for both their on-target and off-target actions, with the ultimate goals of designing safer and more selective anesthetics and of using currently available anesthetics in a rational mechanism-based manner to maximize therapeutic ratio.
Volatile anesthetics such as isoflurane modulate synaptic and extrasynaptic neurotransmission through multiple postsynaptic targets, primarily by potentiating inhibitory GABA A receptors and depressing excitatory glutamatergic transmission via ionotropic glutamate receptors (Rudolph and Antkowiak, 2004). However, the GABA A receptor antagonist bicuculline does not antagonize isoflurane-induced immobility, indicating a role for other targets in this effect (Zhang et al., 2004). The presynaptic effects of volatile anesthetics are not as well characterized as their postsynaptic effects due to the small sizes of nerve terminals and technical limitations of conventional electrophysiological techniques in recording presynaptically. Nevertheless, considerable neurochemical and neurophysiological evidence indicates that volatile anesthetics directly inhibit neurotransmitter release (Hemmings et al., 2005a,b).
Synaptic vesicle (SV) exocytosis is tightly coupled to the amount of Ca 2ϩ entering the presynaptic bouton (Wu et al., 2004), which is determined primarily by presynaptic voltage-gated ion channels (Na ϩ , Ca 2ϩ , and K ϩ channels) and modulatory receptors. Isoflurane depresses action potential (AP) amplitude in axons and boutons, which results in downstream reductions in Ca 2ϩ influx and neurotransmitter release (Wu et al., 2004;Hemmings et al., 2005a;Ouyang and Hemmings, 2005). Isoflurane also inhibits neurotransmitter release from isolated nerve terminals with greater potency from glutamatergic than from GABAergic terminals Hemmings, 2003, 2006), consistent with neurotransmitter-specific presynaptic anesthetic mechanisms. The cellular and molecular bases of this synaptic selectivity are unclear.
Voltage-gated Ca 2ϩ channels play an essential role in neurotransmission by mediating Ca 2ϩ influx that is closely coupled to exocytosis. Presynaptic Ca 2ϩ channels are possible targets for inhibition of neurotransmitter release by volatile anesthetics, and are also involved in producing myocardial depression and vasodilation leading to significant cardiovascular side effects (Lynch et al., 1981;Bosnjak et al., 1991). Synaptic transmission at most central nervous system synapses is mediated by multiple Ca 2ϩ channel subtypes that are closely coupled to SV exocytosis, most prominently Ca V 2.1 (P/Q-type Ca 2ϩ channels) and Ca V 2.2 (N-type Ca 2ϩ channels; Wheeler et al., 1994;Wu et al., 1999). The degree of Ca V 2.1 and Ca V 2.2 involvement differ not only between different neuron classes (Murakami et al., 2002;Evans and Zamponi, 2006) but also between nerve terminals of the same afferent axon (Reid et al., 1997;Ariel et al., 2013). Reports of the effects of volatile anesthetics on specific Ca 2ϩ channel subtypes are inconsistent (Hall et al., 1994;Study, 1994;White et al., 2005;Joksovic et al., 2009) such that the extent to which inhibition of Ca 2ϩ channels contributes to inhibition of SV exocytosis remains unclear.
Recent work suggests that presynaptic Ca 2ϩ channels are not the principal targets involved in the inhibition of glutamate and GABA release by volatile anesthetics (Westphalen et al., 2013). Despite their central roles in wakefulness and arousal (Monti and Monti, 2007), few studies have investigated anesthetic effects on aminergic neurons, which have distinct mechanisms of transmitter release (Liu et al., 2018). Electrical stimulation of dopaminergic neurons in the rat ventral tegmental area (VTA), one of the principal midbrain dopaminergic nuclei (Barrot, 2014), induces emergence from isoflurane anesthesia in rats . We sought to clarify the neurotransmitter selectivity and presynaptic targets of volatile anesthetics by investigating the effects of isoflurane, a representative halogenated ether anesthetic, on SV exocytosis from central dopaminergic neurons to test the hypothesis that isoflurane inhibits dopamine release by a mechanism distinct from that involved in nondopaminergic neurons.

Reagents and solutions
Isoflurane was obtained from Abbott, -conotoxin GVIA and -agatoxin IVA from Alomone Labs, and 6-cyano-7nitroquinoxaline-2,3-dione (CNQX) and (2R)-amino-5phosphonovaleric acid (AP5) from Tocris. All other reagents were from Sigma-Aldrich. Rat vMAT2-pHluorin was kindly provided by Robert Edwards (University of California, San Francisco, CA), and mouse VAMP-mCherry was from [Timothy Ryan]. Tyrode's solution (119 mM NaCl, 2.5 mM KCl, 2 mM CaCl 2 , 2 mM, MgCl 2 , 25 mM HEPES, and 30 mM glucose; pH 7.4) was used as the standard buffer in all experiments. The glutamate receptor antagonists CNQX (10 M) and AP5 (50 M) were added to Tyrode's solution to block postsynaptic excitatory synaptic transmission. For single AP studies, Tyrode's solution contained 4 mM CaCl 2 . Saturated stock solutions of isoflurane were prepared and diluted to 0.7 mM [two times the minimum alveolar concentration (2 MAC)] in Tyrode's solution and into gas-tight glass syringes for focal perfusion onto imaged neurons through a 150-m diameter polytetrafluoroethylene tube in the imaging chamber. Accounting for 10 -20% loss, the predicted final concentration of isoflurane was 0.64 mM, which corresponds to 2 MAC in rats, a clinically relevant concentration equivalent to two time the ED 50 , corrected to the experimental temperature of 30°C (Franks and Lieb, 1993). Isoflurane was applied for 5 min before imaging to allow uptake and equilibration. At the conclusion of each experiment, a sample was taken from the chamber for analysis of delivered isoflurane concentration using a Shimadzu GC-2010 Plus gas chromatography with external standard calibration (Ratnakumari and Hemmings, 1998).

Cell culture
Experiments were conducted according to protocols approved by the [Weill Cornell Medicine] Institutional Animal Care and Use Committee and conformed to National Institutes of Health Guidelines for the Care and Use of Animals. Glial monolayers were prepared from cerebral cortex as feeder layers for primary VTA neuron cultures from Sprague Dawley postnatal day 1 male and female rats (Charles River Laboratories) as described previously (Mena et al., 1997). After 7 d in vitro (DIV), neurons were transfected with vMAT2-pHluorin or VAMP-mCherry using a DNA-calcium phosphate coprecipitation protocol (Goetze et al., 2004;Jiang and Chen, 2006) modified to ensure low density transfection so that images could be obtained from a single neuron. Data were acquired from only one neuron per coverslip to avoid the contaminating and potentially irreversible effects of each drug treatment. Each experimental group contained coverslips from two to four different batches of primary neuron cultures to minimize artifacts due to differing culture conditions.

Imaging SV exocytosis
Live-cell epifluorescence imaging employed a Zeiss Axio Observer microscope with images acquired using an Andor iXon ϩ CCD camera (model DU-897E-BV) and APs were evoked with 1-ms current pulses delivered via platinum-iridium electrodes. Depolarization with elevated K ϩ Tyrode's solution (50 mM KCl substituted for 50 mM NaCl and buffered to pH 7.4) was used to evoke SV exocytosis independent of Na v involvement (57). Elevated K ϩ Tyrode's solution was applied onto imaged neurons using a pressurized injector (PDES System, ALA) for 4 s at 29 l/s as the chamber was continuously perfused with Tyrode's solution with or without added drugs. Fluores-cence data were acquired as described, and total pool (TP) of SVs was identified by perfusion with Tyrode's solution containing 50 mM NH 4 Cl (substituted for 50 mM NaCl and buffered to pH 7.4).

Imaging calcium influx
VAMP-mCherry, a red fluorescent protein fused to VAMP (vesicle associated membrane protein), was used to identify synaptic boutons for Ca 2ϩ imaging experiments. Transfected neurons were loaded with 7 M Fluo-5F AM, incubated for 10 min at 30°C, and washed by superfusion with Tyrode's solution for 15 min. Neurons were stimulated with a single AP 5 times at 2-min intervals during superfusion with Tyrode's solution containing 2 mM Ca 2ϩ with or without 2 MAC isoflurane.

Immunocytochemistry
Post hoc immunolabelling with mouse anti-tyrosine hydroxylase (TH) monoclonal antibody (MAB318, Millipore) was used to identify dopaminergic neurons following live cell imaging. Fixed neurons were immunolabelled with either a 1:1000 dilution of Alexa Fluor 594 goat antimouse (for SV exocytosis experiments using vMAT2-pHluorin) or Alexa Fluor 488 goat anti-mouse (for Ca 2ϩ imaging experiments). Imaged neurons were identified by coordinates on the coverslips and photographed.

Image and statistical analysis
Fluorescence data were analyzed in ImageJ (http:// rsb.info.nih.gov/ij) with a custom plug-in (http://rsb.info. nih.gov/ij/plugins/time-series.html). Transfected boutons were selected as regions of interest (ROIs) based on their response to 50 mM NH 4 Cl for SV exocytosis experiments or labeling with VAMP-mCherry for Ca 2ϩ measurements. Each bouton was subjected to a signal-to-noise ratio (SNR) calculation based on its response to the first control electrical stimulation, and ⌬F was calculated as the difference of the average intensities between F peak and F baseline . Fluorescence intensity changes for Ca 2ϩ measurements were normalized to baseline as ⌬F/F: (F peak -F baseline )/F baseline . Boutons with SNR Ͼ 5 were used in the analysis. Data are expressed as mean Ϯ SD. To allow expression of inhibition or potentiation, drug effects are shown as a percentage of either TP or control response. Statistical significance was determined by paired or unpaired two-tailed or one-tailed Student's t tests and by paired or unpaired one-way ANOVA with Tukey's post hoc test, with p Ͻ 0.05 considered significant. Normality was assayed using the Shapiro-Wilk normality test. All statistical data are displayed in Table 1. Statistical analysis and graph preparation used GraphPad Prism v7.05 (GraphPad Software, Inc.).

Results
We used high resolution microscopy to quantify exocytosis at dopaminergic nerve terminals by the fluorescence change of pH-sensitive pHluorin fused to the luminal domain of the vesicular mononamine transporter vMAT2 (Anantharam et al., 2010;Pan and Ryan, 2012). Cultured rat midbrain neurons transfected with vMAT2-pHluorin were stimulated with trains of 100 APs at 10 Hz to elicit SV New Research exocytosis ( Fig. 1A). Increases in fluorescence (⌬F) following stimulation indicate alkalization of intravesicular pHluorin due to SV exocytosis. The difference between baseline and stimulus-evoked peak fluorescence reflects the amount of SV exocytosis; quenching of fluorescence in the post-stimulus period indicates SV endocytosis and re-acidification (Sankaranarayanan et al., 2000;Atluri and Ryan, 2006;Fig. 1B). The biosensor vMAT2-pHluorin reliably measured SV exocytosis over time with minimal decay in signal over the course of three control stimulations (stimulation 1 ϭ 6.7 Ϯ 0.9% of TP; stimulation 2 ϭ 6.9 Ϯ 0.9% of TP; stimulation 3 ϭ 6.8 Ϯ 0.9% of TP; n ϭ 8; p ϭ 0.93). Fluorescence data from each cell were normalized to the total SV pool defined by perfusion with 50 mM NH 4 Cl, which alkalizes the acidic SV interior and unquenches pHlourin fluorescence of the entire SV pool (Fig. 1A,B). Both dopaminergic and non-dopaminergic neurons can be transfected by vMAT2-pHluorin; dopami-nergic neurons were positively identified by post hoc immunolabeling with mouse anti-TH (Fig. 1C).

Isoflurane effect is Na ؉ channel independent in dopaminergic neurons
Ca 2ϩ -dependent SV exocytosis evoked by elevated extracellular KCl occurs by sustained depolarization that is independent of Na V activation: it is insensitive to the specific Na v blocker tetrodotoxin (TTX), in contrast to phasic AP-evoked SV exocytosis, which is completely blocked by TTX (Westphalen and Hemmings, 2003). Superfusion of TTX abolished SV exocytosis evoked electrically to -0.5 Ϯ 0.2% of TP, yet had no effect on the response to elevated KCl (control with KCl ϭ 12.9 Ϯ 1.8% of TP, KCl with TTX ϭ 13.3 Ϯ 1.7% of TP; p ϭ 0.92; n ϭ 6; Fig. 3) d . We compared the effects of isoflurane on SV exocytosis evoked by elevated KCl, which directly activates Ca 2ϩ channels, to SV exocytosis evoked electrically to determine whether isoflurane acts upstream or downstream of Ca 2ϩ entry (Fig. 4A). We used a KCl concentra-  tion that evoked similar peak SV exocytosis compared to that obtained with the 100 AP stimulus train. Isoflurane inhibited elevated KCl-evoked exocytosis in dopaminergic neurons from 11.4 Ϯ 1.3% of TP in control to 7.8 Ϯ 1.5% of TP with isoflurane (35 Ϯ 6% inhibition; p ϭ 0.0007; n ϭ 8; Fig. 4B,D) e . In contrast, isoflurane did not inhibit KCl-evoked SV exocytosis in non-dopaminergic neurons: 10.4 Ϯ 0.9% of TP in control and 10.3 Ϯ 1.1% of TP with isoflurane (2 Ϯ 4% inhibition; p ϭ 0.72; n ϭ 6; Fig.  4C) f . Thus, isoflurane inhibited SV exocytosis in dopaminergic neurons by an Na V -independent pathway, in contrast to its Na v -dependent inhibition of SV exocytosis in non-dopaminergic neurons (Wu et al., 2004;Westphalen et al., 2013;Baumgart et al., 2015).

Ca 2؉ channel subtypes mediating exocytosis in dopaminergic neurons
The role of specific Ca 2ϩ channel subtypes in isoflurane inhibition of SV exocytosis was studied using the subtype-specific neurotoxin -conotoxin GVIA to block Ca V 2.2 or -agatoxin IVA to block Ca V 2.1 (Fig. 5A). Conotoxin alone inhibited SV exocytosis by 43 Ϯ 3% (n ϭ 9) in dopaminergic neurons and by 68 Ϯ 3% (n ϭ 5) in nondopaminergic neurons, consistent with a greater contribution of Ca V 2.2 in non-dopaminergic neurons than in dopaminergic neurons (p ϭ 0.015; Fig. 5B) g . The Ca V 2.1 blocker agatoxin alone inhibited SV exocytosis by 83 Ϯ 5% (n ϭ 7) in dopaminergic neurons and by 63 Ϯ 11% (n ϭ 6) in non-dopaminergic neurons, confirming a greater contribution of Ca V 2.1 to SV exocytosis in dopaminergic neurons (conotoxin inhibition ϭ 43 Ϯ 3%; agatoxin inhibition ϭ 83 Ϯ 5%; p Ͻ 0.0001) h . In non-dopaminergic neurons, there was no significant difference in inhibition by conotoxin or agatoxin (p ϭ 0.99) i , indicating similar contributions by both Ca V 2.1 and Ca V 2.2. There was no effect of the L-type Ca 2ϩ channel inhibitor nimodipine (10 M, Nimo) on SV exocytosis from dopaminergic or nondopaminergic neurons (DA Ctrl ϭ 7.8 Ϯ 0.4% vs DA ϩ Nimo ϭ 7.2 Ϯ 0.5% of TP, n ϭ 7, p ϭ 0.139; non-DA Ctrl ϭ 10.1 Ϯ 1.8% vs non-DA ϩ Nimo ϭ 9.8 Ϯ 1.9% of TP, n ϭ 5, p ϭ 0.122). SV exocytosis in both dopaminergic and non-dopaminergic neurons was mediated exclusively by Ca V 2.1 and Ca V 2.2 since conotoxin and agatoxin together completely blocked SV exocytosis (Fig. 5B). A bouton by bouton analysis from all recorded neurons examined the effect of conotoxin (Fig. 5C) and agatoxin (Fig. 5D) in dopaminergic neurons and showed that the effects of the toxins correlated to the averaged effects (Fig. 5B).

Isoflurane inhibits exocytosis by reducing Ca 2؉ entry
Isoflurane inhibits SV exocytosis in glutamatergic and GABAergic hippocampal neurons by reducing Ca 2ϩ influx without affecting Ca 2ϩ sensitivity indicated by the relationship between intracellular Ca 2ϩ concentration and SV exocytosis (Baumgart et al., 2015). We examined this in dopaminergic neurons by comparing the effects of reduced extracellular Ca 2ϩ concentration ([Ca 2ϩ ] e ) on SV exocytosis and Ca 2ϩ influx using single AP stimuli to determine Ca 2ϩ sensitivity (Fig. 7A,C). Isoflurane inhibited single AP-evoked SV exocytosis by 34 Ϯ 6% (0.89 Ϯ 0.19% of TP for 4 mM [Ca 2ϩ ] e control vs 0.58 Ϯ 0.12% of TP with isoflurane; p ϭ 0.005) m . This reduction in exocytosis was mimicked by reducing [Ca 2ϩ ] e from 4 mM to 2 mM in the absence of isoflurane. Exocytosis in 2 mM [Ca 2ϩ ] e was 0.51 Ϯ 0.13% of TP (44 Ϯ 5% reduction vs 4 mM [Ca 2ϩ ] e ), which was not significantly different from the inhibition by isoflurane in 4 mM [Ca 2ϩ ] e (p ϭ 0.63; n ϭ 6; Fig. 7B) n . A comparable effect was observed using the Ca 2ϩ indicator Fluo-5F to measure changes in intracellular [Ca 2ϩ ] (Fig. 7D-G

Discussion
Isoflurane inhibited SV exocytosis from cultured dopaminergic neurons by reducing Ca 2ϩ entry though both Ca V 2.1 and Ca V 2.2 by a mechanism that is independent of Na ϩ channel activation (Fig. 8). This is in contrast to the predominant Na ϩ channel-dependent mechanism observed for release of glutamate or GABA in non-dopaminergic cortical and hippocampal neurons (Westphalen et al., 2010;Baumgart et al., 2015). These findings reveal important neurotransmitter-selective differences in the presynaptic mechanisms of isoflurane, a clinically essential volatile anesthetic. These differences provide a pharmacological rationale for developing novel anesthetics targeting specific anesthetic endpoints mediated by a single neurotransmitter system, for example dopaminergic control of emergence from unconsciousness.

Relationship between Ca 2؉ channel subtypes and exocytosis
The Ca 2ϩ channel subtypes present in dopaminergic neuron somata identified using whole-cell voltage-clamp recordings do not necessarily reflect the presynaptic Ca 2ϩ channels involved in SV exocytosis. In rat dopaminergic midbrain neurons, somatic Ca 2ϩ currents are inhibited by the L-type channel blocker nimodipine (by 28%), the Ca V 2.2 blocker conotoxin (by 22%), and the Ca V 2.1 blocker agatoxin (by 37%; Cardozo and Bean, 1995). However, synaptic boutons in rat dopaminergic neurons are too small for such direct electrophysiological recording of presynaptic Ca 2ϩ currents. We used high-resolution live-cell imaging to measure SV exocytosis and Ca 2ϩ influx, employing specific Ca 2ϩ channel toxins to deter-mine contributions of the major presynaptic Ca 2ϩ channel subtypes. SV exocytosis in rat dopaminergic VTA neurons was mediated exclusively by Ca V 2.1 and Ca V 2.2, with Ca V 2.1 predominating. Using KCl-induced depolarization to evoke SV exocytosis independent of Na ϩ channel involvement, isoflurane inhibited exocytosis mediated by either Ca V 2.1 or Ca V 2.2 to a similar degree, suggesting a lack of subtype selectivity. Alternatively, the effects of isoflurane are mediated via an unknown mechanism distinct to dopaminergic VTA neurons. One possibility is that K ϩ channel activity alters the resting membrane potential and therefore modifies open probability of Ca V 2.1 and Ca V 2.2.
Isoflurane inhibited SV exocytosis in dopaminergic neurons evoked by electrical stimulation of APs or elevated KCl depolarization, in contrast to non-dopaminergic neurons in which isoflurane inhibited AP-evoked but not KCl-evoked exocytosis. AP-evoked exocytosis requires activation of Na ϩ channels to sufficiently depolarize boutons and activate the presynaptic Ca 2ϩ channels linked to exocytosis. Clamped depolarization by elevated KCl is less physiologic than repetitive electrical stimulation by causing sustained depolarization (Tibbs et al., 1989), which could alter Ca 2ϩ channel relationship to exocytosis. This is suggested by the finding that with electrical stimulation conotoxin inhibited exocytosis in dopaminergic neurons by ϳ40% and agatoxin inhibited by ϳ80%, while with KCl-evoked depolarization both conotoxin and agatoxin inhibited exocytosis by 20 -30%. The linkage between SV exocytosis to critical presynaptic Ca 2ϩ channels in dopaminergic neurons is selectively sensitive to isoflurane, since isoflurane inhibition of KClevoked exocytosis was observed in dopaminergic, but not in non-dopaminergic, neurons. This fundamental neurotransmitter-specific difference in the relationship be- tween Ca 2ϩ channels and exocytosis results in profound neurotransmitter-specific differences in anesthetic sensitivity with potential neuropharmacological implications. This neurochemical difference is preserved in nigrostriatal dopaminergic nerve terminals prepared from rat striatum in which isoflurane also inhibits dopamine release via a Na ϩ channelindependent mechanism, an action that might contribute to the motor effects of anesthetics (Westphalen et al., 2013). The cellular and molecular attributes underlying this selective anesthetic pharmacology in dopaminergic neurons are unknown and await further characterization of the neurobiology of dopaminergic compared to non-dopaminergic neurons.

Mechanisms of dopamine SV exocytosis
Monoamine neurotransmitters such as dopamine are packaged into both small SVs and large dense core vesicles (LDCVs) for release. The biosensor vMAT2-pHluorin labels both small SVs and LDCVs (Fei et al., 2008); however, only small SVs localize to presynaptic active zones of synaptic boutons for exocytosis, while LDCVs engage primarily in extrasynaptic exocytosis (Thureson-Klein, 1983;Südhof and Rizo, 2012). Moreover, the kinetics of SV exocytosis from small SVs and LDCVs are distinct: small SVs fuse within 1 ms of Ca 2ϩ channel opening (Cohen et al., 1991), while LDCV fusion is 100-fold slower and therefore less tightly regulated to AP stimulation (Almers, 1990;Martin, 1994). Based on these characteristics, the SV exocytosis measured by the vMAT2-pHluorin method is primarily from small SVs (Pothos et al., 2000;Leenders et al., 2002). Anesthetic effects on asynchronous neuronal LDCV exocytosis might involve distinct mechanisms.
Neuroendocrine cells such as adrenal chromaffin cells or PC12 cells are frequently used to study catecholaminergic SV exocytosis but exhibit release mostly from LD-CVs (Voets et al., 2001). This is an important distinction as the subcellular and molecular organization of the secretory machinery differ between dopaminergic neurons and neuroendocrine cells, which makes the latter poor models for midbrain neurons. For example, neuroendocrine cells do not have active release zones co-localized with Ca 2ϩ microdomains, and the functional linkage of Ca 2ϩ channels to release sites is not as tight as in neurons (Wu et al., 2009). Moreover, the Ca 2ϩ channel subtypes linked to SV exocytosis differ between small SVs and LDCVs, with L-type channels closely linked to LDCV exocytosis (Park and Kim, 2009). In contrast, we found dopaminergic SV exocytosis to be independent of L-type Ca 2ϩ channels.

Differences between dopaminergic and nondopaminergic neurons
When comparing the effects of isoflurane on dopaminergic and non-dopaminergic neurons it is important to consider that vMAT2 is not endogenously expressed in non-dopaminergic neurons (Yoo et al., 2016). However, vMAT2-pHluorin is still effective in measuring SV exocytosis in non-dopaminergic neurons due to ectopic expression. Transfection of vMAT2-pHluorin involves overexpression of vMAT2, which can increase quantal size, but this does not interfere with its use as an indicator of SV fusion (Pothos et al., 2000;Erickson et al., 2006).
In contrast to dopaminergic neurons, elevated KClevoked SV exocytosis from non-dopaminergic neurons was insensitive to isoflurane. This is consistent with previous observations of a Na V -dependent/Ca v -independent mechanism for rat glutamatergic and GABAergic hippocampal neurons (Hemmings et al., 2005a;, despite their expression of both Ca V 2.1 and Ca V 2.2 (Qian and Noebels, 2001;Kamp et al., 2012). There are other important differences between dopaminergic and non-dopaminergic neurons that could explain their differential sensitivities to anesthetics. Nondopaminergic neurons from the VTA are primarily GABAergic, some of which are capable of co-releasing glutamate (Carr and Sesack, 2000;Creed et al., 2014;Barker et al., 2016). In cultured rat hippocampal neurons, isoflurane continued by paired one-way ANOVA, Tukey's post hoc test. C, Schematic of the Ca 2ϩ influx protocol with Fluo-5F (yellow bar) followed by washout and  In nondopaminergic (non-DA) neurons, isoflurane inhibits SV exocytosis by inhibiting Na ϩ channels to reduce nerve terminal excitability. In dopaminergic (DA) neurons, isoflurane can also reduce SV exocytosis by a Na ϩ channel-independent mechanism mediated by inhibition of Ca V 2.1 and Ca V 2.2. The fluorescence of vMAT2-pHluorin is quenched by the acidic (ϳpH 5.5) SV interior. On exocytosis and fusion with the plasma membrane, vMAT2-pHluorin fluoresces on exposure to the pH neutral (ϳpH 7.4) synaptic cleft. inhibits SV exocytosis from glutamatergic boutons more potently than from GABAergic boutons due to a greater reduction in presynaptic Ca 2ϩ influx (Baumgart et al., 2015), indicating that GABAergic neurons are less sensitive to isoflurane. This neuronal phenotypic difference in anesthetic sensitivity is consistent with the data presented here, as isoflurane more potently inhibited electrically-evoked SV exocytosis in dopaminergic neurons than in non-dopaminergic VTA neurons. It is likely that the non-dopaminergic VTA neurons were primarily GABAergic given their abundance in this nucleus. These neurotransmitter-specific differences in presynaptic sensitivity to isoflurane are likely due to differential presynaptic expression of specific ion channel subtypes (Johnson et al., 2017) with different anesthetic sensitivities that contribute to SV exocytosis.
Considering their numerous possible subunit compositions and splice variants, different ion channel subtypes and variants could contribute to presynaptic Ca 2ϩ entry and SV release in different boutons (Meir et al., 1999). Differences in ion channel expression and degree of functional linkage (tight or loose) between Ca 2ϩ entry and SV exocytosis also exist between various neuronal phenotypes (Eggermann et al., 2011). Differences in expression of various presynaptic Ca 2ϩ binding proteins (e.g., calmodulin, present in all neurons, or calbindin, selectively expressed in some dopaminergic neurons) might also determine observed nerve terminal-specific differences in anesthetic sensitivity (Pan and Ryan, 2012). Further studies are necessary to determine the molecular specializations that underlie these presynaptic differences in anesthetic sensitivity.

Role of dopaminergic neurons in general anesthesia
Mammalian dopaminergic neurons are located primarily in the substantia nigra pars compacta and the VTA. They project widely to forebrain regions including the dorsal striatum and nucleus accumbens, where dopamine release is essential to motor function and motivated behaviors, respectively (Wise, 2004). Electrical stimulation of the VTA, but not of the substantia nigra, facilitates emergence from isoflurane anesthesia in adult rats Taylor et al., 2016). This effect is mediated by D1 dopamine receptors since selective pharmacological activation of D1, but not D2, receptors induces emergence from isoflurane anesthesia (Taylor et al., 2013). Moreover, the dopamine transporter (DAT) inhibitors methylphenidate and dextro-amphetamine restore conscious behaviors in rats anesthetized with isoflurane, propofol, or sevoflurane Kenny et al., 2015). Our findings that isoflurane inhibits SV exocytosis from midbrain dopamine neurons provides mechanistic support for the hypothesis that reduced dopamine signaling is associated with isoflurane-induced unconsciousness through reduced D1 receptor activation due to reduced dopamine exocytosis.

Conclusions
Isoflurane inhibits SV exocytosis from rat dopaminergic neurons through direct inhibition of Ca 2ϩ influx mediated by Ca V 2.1 and/or Ca V 2.2 independent of Na ϩ channel involvement. This neurotransmitter-selective presynaptic mechanism provides a molecular substrate for the role for dopaminergic VTA neurons in modulating isofluraneinduced unconsciousness. Improved understanding of these mechanisms is critical to elucidating how general anesthetics work to optimize their safe use and develop more specific drugs with fewer adverse effects.