Supplementary Figure 1 Cholinergic neurons projecting to the VTA are concentrated in the caudal mesopontine region. (a) Schematic showing the sites of retrograde tracer injections in the VTA: cholera toxin b (CTb) was injected in the rostral VTA and red retrobeads (RB-R) in the caudal VTA to identify the topography of the projections from the PPN and LDT. (b) ChAT-immunopositive neurons contained CTb and/or RB-R in both PPN and LDT (arrows; the PPN neuron is triple-labeled and the LDT neuron in doublelabeled). (c) Pattern of retrograde labeling in relation to the expression of ChAT at three mediolateral levels of the PPN/LDT in a representative animal. (d) Similar numbers of VTA-projecting cholinergic neurons were detected at different mediolateral levels. A larger proportion of non-cholinergic VTA-projecting neurons was observed in the most lateral sections of the PPN. (e) The number of cholinergic neurons containing one of the tracers was normalized against the total number of cholinergic neurons and expressed as a percentage of the total. This was quantified at three rostrocaudal 900 μm-segments of the PPN and in the LDT. The number of cholinergic neurons projecting to the VTA is higher in the caudal PPN segments and the LDT. This contrasts with the low number of ChAT+/RB-R+ neurons projecting to the caudal VTA located in the rostral PPN. Bars represent SEM (n = 4 rats).
Supplementary Figure 2 Specificity of transduction for cholinergic neurons in different brain areas. We tested the specificity of the transduction by injecting AAV2-EF1a-DIO-YFP in other brain regions in a different set of animals (n=3 rats), including the cerebellum, ventromedial thalamus, superior colliculus and striatum as well as the PPN and LDT (n = 5 rats). Positive (i.e. YFP-expressing) somata were only observed in the striatum, the PPN and LDT, and no retrograde transduction of cholinergic neurons was detected. The presence of two fluorescent markers in the cholinergic neurons of the PPN (AAV2-EF1a-DIO- YFP) and LDT (AAV5-EF1a-DIO-mCherry) of ChAT::Cre+ rats was used to evaluate the specificity for the transduction, in combination with immunofluorescent detection of ChAT. See Supplementary Table 2 for values.
Supplementary Figure 3 Lack of forebrain cholinergic innervation of the VTA. Coronal (a-e) or sagittal (f) sections showing the VTA following injections of AAV2-DIO-EF1a-eYFP into the medial septum (a), vertical limb of the diagonal band of Broca (b), horizontal limb of the diagonal band of Broca (c), nucleus basalis of Meynert (d), medial habenula (e) and the parabigeminal nucleus (f) (n = 3 per group). None of the transductions in these areas produced cholinergic axonal labeling in the VTA. The borders of the VTA were delimited by TH immunostaining (in red). Scale bar: 500 µm.
Supplementary Figure 4 Mediolateral distribution of cholinergic axons in the VTA. (a) The axons used in the analyses in Fig. 1c and 2c were traced and digitized, as shown in this representative sections (n = 3 rats each). (b) The axonal density was normalized to detect areas of preferential innervation from either the PPN or the LDT. Whereas in the lateral and middle VTA we observed a similar pattern, in the medial VTA some differences were observed: PPN tends to innervate the most rostral regions of the VTA, where LDT innervation is low, but then drops in the most caudal regions, where LDT innervation is higher. This coincides with the low numbers of retrogradely labeled cholinergic neurons in the PPN when the injection was located in the caudal VTA. (c) Quantification of the total length cholinergic axons in the PPN and the LDT (t = 1.29, two-tailed t-test, P = 0.266). Bars represent SEM. * P < 0.05.
Supplementary Figure 5 Analysis of the response of a representative neuron to the laser stimulation. (a) Three responses to the laser were obtained for this neuron. Spike events were extracted and merged into a single spike train (black trace). (b) Cumulative distribution function (CDF) of the three individual spike trains together with that of the merged spike train (black dots) were computed by increasing the CDF one unit each time a spike occurred (note that in order to maintain the same scale, the CDF of the merged spike train must be divided by the number of spike trains merged). Regression slope of the CDF was computed at each spike time by using a local (18 neighboring spikes) linear regression analysis (see grey dots on the black trace and dashed red lines for an example at three different time points, red dots). (c) Estimated instantaneous firing rate (regression slope) for the three
individual trains and for the merged spike train where red dots correspond to the regression points marked in panel b. (d) Smoothed version and z-scored version of the instantaneous firing rate shown in panel c (for merged trace only). Horizontal lines correspond to the mean (full) and 5 th and 95 th percentile (dashed) values of the firing rate during pre-stimulus period (-10 to 0 s). Red trace demarks the response period of the neuron (i.e., a significant P < 0.05 increase in the firing rate).
Supplementary Figure 6 Controls for stimulation of cholinergic axons in the VTA. (a) Following injection of AAV2-DIO-EF1a-eYFP into the PPN/LDT in ChAT::cre rats, individual TH+ neurons (n = 5 neurons) were recorded in vivo during blue light delivery and subsequently labeled with neurobiotin (b). The same protocol was followed as that used in figures 1 and 2. No change in firing was observed in neurons transduced with YFP only. Scale bar: 20 µm.
Supplementary Figure 7 Electrical stimulation of the PPN produces short-latency responses in DA neurons. (a) A neuron that was recorded and labeled by the juxtacellular method and identified as dopaminergic by the expression of immunoreactivity for TH. (b) Raster plot and peri-stimulus time histogram (PSTH) of the neuron shown in a, following electrical stimulation of the PPN (bin size 1 ms). (b ) A representative spike following the stimulus (s; 0.5 ms duration, 0.5-0.8 ma, 0.5 Hz). Consistent with previous reports, short-latency action potentials were elicited in the DA neurons within a few milliseconds of the stimulus
being delivered in the PPN; this contrasts with the slow modulatory effect of the optogenetic activation of PPN cholinergic afferents. (c) Short-latency responses were consistent across the DA neuronal population (n = 12 neurons in 5 rats). (d, e, e ) Non-DA neurons (i.e. TH-immunonegative, d) were also sampled (n = 4 neurons in 5 rats). PSTH (bin size 2 ms) showing a long-lasting inhibition (average time of inhibition: 30ms +/- 8.6ms, e). Responses in non-da neurons were more heterogeneous.
Supplementary Figure 8 Nicotinic modulation of VTA neurons. (a) Effect of carbachol puffing on VTA DA neurons (n = 9 neurons) that were labeled and identified as TH+. (b) Example of VTA neurons showing classical dopaminergic features with low spontaneous firing and/or presence of Ih current. (c) Carbachol puffs (200 µm) provoke excitatory postsynaptic potentials (EPSPs) that are blocked by mecamylamine (n = 9 neurons; 5 µm) but not by DhβE (1 µm) or methylcaconitine (500 nm). (d) Carbachol-induced EPSPs were not blocked by bicuculline (10 µm), CNQX (10 µm) or APV (10 µm) (n = 4 neurons). (e) The mecamylamine effect on EPSP size was significantly different to the control condition (t = 7.281, two-tailed t-test, P = 0.000342). (f) Effect of carbachol puffing on a VTA non-da neuron, confirmed as TH- by immunofluorescence, and showing
classical non-da neuron features such as higher firing rate and lack of Ih current. (g) Carbachol puffing provokes robust EPSPs that are blocked by bath application of MLA (500 nm) but not by DHβE (1 µm) or mecamylamine (5 µm). Scale bar in fluorescent images: 20 µm. Acetylcholine has been reported to induce strong somatic excitation of dopamine neurons in the VTA (Eddine, R., et al. (2015) Sci Rep 5, 8184) and its effects consist of a presynaptic component, modulating glutamate release (Schilstrom, B., et al. (1998) Neuroscience 82, 781 789), and a postsynaptic component, involving nicotinic and muscarinic mechanisms (Gronier, B., et al. (2000) Psychopharmacology 147, 347-355; Mameli-Engvall, M., et al. (2006) Neuron 50, 911-921; Mansvelder, H.D., et al. (2002) Neuron 33, 905-919). Furthermore, muscarinic agonists increase the frequency of spontaneous action potentials in DA neurons (Lacey, M.G., et al. (1990) Pharmacol Exp Ther 253, 395-400; Scroggs, R.S., et al. (2001) J Neurophysiol 86, 2966-2972) and induce glutamate release concomitantly with a decrease in GABA release in the VTA (Grillner, P., et al. (2000) Neuroscience 96, 299-307; Mansvelder, H.D., et al. (2000) Neuron 27, 349-357). Our results are in agreement with other studies showing that mecamylamine was able to abolish the effects mediated by cholinergic agonists in DA neurons (Zhang, L., et al. (2005) J Physiol 586, 469-481) or following place preference associated with carbachol injections in the VTA (Ikemoto, S. & Wise R.A., et al. (2002) J Neurosci 22, 9895-9904). While our experiments seem to indicate an effect predominantly mediated by nicotinic receptors, muscarinic receptors are highly expressed in the VTA and are also likely to contribute to the modulation of DA neurons, as shown previously (Forster, G.L. and Blaha C.D. (2000) Eur J Neurosci 12, 3596 3604).
Supplementary Figure 9 Expression of vesicular acetylcholine transporter (VAChT) in YFP-labeled axons. Axons expressing YFP in the VTA from both PPN- and LDT-transduced animals were tested for the presence of vesicular transporters by immunofluorescence. VAChT immunofluorescence was detected in the majority of axonal varicosities expressing YFP (arrowheads), as has also been reported in the striatum (Dautan et al (2014) J Neurosci 34, 4509-4518). In contrast, the vesicular glutamate transporter 2 (VGluT2) was never observed to co-localize with YFP in axonal varicosities, even though we detected large numbers of VGluT2-positive varicosities within the VTA. Thus, in agreement with in situ hybridization data (Wang and Morales (2009) Eur J Neurosci 29,340-358) the axons of cholinergic neurons from the brainstem lack VGluT2.
Supplementary Figure 10 Individual response of DA neurons to aversive (pinch) and laser stimulation. Representative traces of two TH+ neurons recorded in vivo during a hind-paw pinch (red square) and laser stimulation (blue square) of PPN (a) or LDT (b) cholinergic axons in the VTA. Scale bar: 50µm.
Supplementary Figure 11 Behavioral responses to the optogenetic activation of PPN and LDT axons are blocked by cholinergic receptor antagonists. (a) Optogenetic activation of cholinergic axons in the VTA of PPN (n = 12 rats), LDT (n = 6 rats) or wild-type groups (n = 8 rats) failed to produce stimulation-locked effects on locomotor activity in the presence of cholinergic antagonists (i.p. 0.3 ml final volume, MLA: 6 mg/kg; DHβE: 3 mg/kg; atropine: 0.5 mg/kg and mecamylamine: 1.0 mg/kg). Locomotor activity in all groups was similar (2-way ANOVA; stimulation effect: F (2,19) = 0.589, P = 0.565; group effect: F (2,19) = 0.140, P = 0.871). (b) Cumulative distance travelled during laser stimulation (10 Hz, 50 ms, 8 s duration, every 2 min) of cholinergic axons in wild-type (green), PPN (gray) or LDT (red) groups following a systemic injection of a cocktail of cholinergic antagonists (1-way ANOVA: F (2, 21) = 1.722, P = 0.205). (c) Number of lever presses per minute during acquisition of the operant task for wild-type (green, n = 10), PPN (gray, n = 12) and LDT (red, n = 10) animals. No significant differences were observed (2-way ANOVA; day effect F (4,26) = 4.336, P = 0.08; group effect: F (2,29) = 0.174, P = 0.842; interaction F (8,54) = 1.038, P = 0.426).
PPN LDT VTA-projecting ChAT(+) neurons VTA-projecting ChAT(-) neurons % ChAT(+) neurons retrogradely labeled 197 144 692 133 25.9 27.3 CTb/ChAT(+) 125 84 RB-R/ChAT(+) 72 60 Table S1. Retrograde cell count in PPN and LDT. Two retrograde tracers were injected in the VTA, cholera toxin b (CTb) in its rostral portion and red retrobeads (RB- R) in its caudal portion (n = 4). The total number of retrogradely labeled neurons for each tracer was quantified in sagittal sections at three mediolateral levels. Immunohistochemistry for ChAT was combined with tracer detection to identify the cholinergic projection neurons. With the exception of the percentage of cholinergic neurons labeled, all numbers represent sums of neurons in all animals.
PPN (n = 5) LDT (n = 4) % cholinergic neurons transduced within the injection diameter % of transduced neurons immunopositive for ChAT 66 ± 6.78 58.2 ± 4.89 91.4 ± 0.45 91 ± 0.31 Table S2. Analysis of the transduction specificity of the viral vector (Means ± SEM).