(A) Sections through the ventral midbrain of VGLUT2-EGFP BAC transgenic mice were immunostained for EGFP (green) and TH (red). The region of the VTA in the white box (above) is magnified in the images below. Many TH⁺ neurons do not express EGFP and several EGFP⁺ neurons do not contain TH, but a subset of TH⁺ cells do also express EGFP. Size bar above indicates 200 µm, below 50 µm. (B) Immunostaining of wild type mouse brain (+/+, left) shows VGLUT2 immunoreactivity absent from the unconditional knockout (Δ/Δ). (C) Immunofluorescence for YFP (green) and TH (red) in coronal brain sections from the VTA of a cKO mouse (with cre recombinase under control of the DAT gene promoter both inactivating VGLUT2 and driving expression of YFP from the (Gt)Rosa26Sor floxed-stop reporter) shows that the vast majority of YFP⁺ neurons co-express TH, demonstrating the specificity of recombination. The merged image below shows a magnification of the boxed areas above. Size bar, 200 µm. See also figure S1.

(A,B) Synaptic vesicles in 140 mM choline gluconate/10 mM K gluconate/10 mM Hepes, pH 7.4 were acidified by the sequential addition of ATP, 2 mM Cl⁻ and then either 14 mM Cl⁻ (A) or 4 mM glutamate (B). The traces in black indicate vesicles without any further addition. At the arrow indicated baf/val, 250 nM bafilomycin (red) or 50 nM valinomycin (gray) or both (light red) were added. The potassium ionophore valinomycin accelerates the dissipation of ΔpH by bafilomycin to a much greater extent in vesicles acidified with glutamate than with Cl⁻. (C) Data from the first 60 seconds after baf/val addition were fit to either a straight line (no treatment or val) or single exponential (baf or val+baf) to derive initial rates of alkalinization (Vo ΔF/t). Within every condition tested, Cl⁻-acidified vesicles alkalinized significantly faster than those acidified by glutamate (p < 0.05 by two-tailed t-test). The addition of val also accelerated alkalinization under each condition, with the exception of baf-treated vesicles acidified using Cl⁻ (p < 0.01 by two-tailed t-test). Data presented as means ± SEM.

(A,B) Tissue monoamine content was measured by HPLC coupled to an electrochemical detector. Punches from the dorsal striatum (A) showed no difference across genotype in the levels of dopamine (DA), dopamine metabolites dihydroxyphenylacetic acid (DOPAC) and homovanillic acid (HVA) and serotonin (5-HT). (n=10/group) However, DA, DOPAC and HVA were significantly reduced in the ventral striatum (B) of cKO mice relative to control littermates, with no change in 5-HT levels (n=10/group; *p < 0.05 by 2-tailed t-test). (C,D) The dopamine release evoked by single action potentials was measured in striatal slices using fast-scan cyclic voltammetry (FSCV). (C) The traces show representative examples of evoked dopamine release and reuptake in the ventral striatum. (D) The ventral striatum (NAc shell) of cKO mice released less dopamine than controls, but the two genotypes did not differ in the dorsal striatum. (n=9 mice and 29 paired slices per genotype for dorsal striatum, 10 mice and 34 paired slices per genotype for ventral striatum). **p < 0.01 by 2-tailed, paired t-test. Data indicate mean ± SEM. See also Figure S2.

(A) Postnatal midbrain cultures from control and cKO mice were grown on a glial monolayer and immunostained for YFP (green) and TH (red). The almost complete co-localization indicates that CreR has mediated recombination both efficiently and specifically in DA neurons of both cKO and control mice. (B) Double staining of the cultures for YFP and VGLUT2 demonstrates that CreR has eliminated the expression of VGLUT2 by YFP+ dopamine neurons of cKO but not control mice. Size bars in A and B, 200 µm. (C) Higher magnification images of double stained cultures show colocalization of VGLUT2 with YFP in the synaptic boutons of dopamine neurons from control but not cKO mice (arrowheads). (D) A single YFP⁺ dopamine neuron grown in microculture was recorded in whole-cell patch-clamp mode. From a holding potential of −60mV, brief depolarization to +20 mV induced an unclamped action potential followed by a fast EPSC in the control (ctrl) but not cKO cell. (E) The mean excitatory autaptic current differs between control and cKO dopamine neurons. n=8−11, ***p < 0.001 by 2-tailed t-test. Data indicate means ± SEM.

(A) Spontaneous locomotor activity over a 72-hour period did not differ between cKO mice and control littermates (n=8/genotype; 2-tailed t-test). (B) Assessed on the rotarod, motor performance and learning showed no difference between genotypes by repeated measures ANOVA (control n=13, cKO n=14). (C) Naïve cKO and control mice were injected with cocaine (20 mg/kg, i.p.) and locomotor activity monitored over the following hour. cKO mice exhibited substantially less locomotor activity than controls (control n=15, cKO n=16; **p < 0.01 by Mann-Whitney rank sum test). (D) To assess locomotor sensitization, naïve mice were injected daily with cocaine (20 mg/kg i.p.) over 5 consecutive days and again 72 h after the fifth injection. cKO mice exhibited less locomotor activity than controls for the first few cocaine injections. By the fifth injection, however, their response did not differ significantly from control mice. (control n=7, cKO n=8) repeated measures ANOVA revealed a significant effect of genotype (p ≤ 0.01) and a significant effect of treatment (p < 0.001). *p < 0.05, ** p< 0.01 by Tukey post hoc comparison across genotype. Data presented as mean ± SEM. See also Figure S1.

(A,B) Glutamatergic responses evoked by stimulation in the VTA were recorded from medium spiny neurons in the medial shell of the NAc using 9–11 day (A) and 21–23 day (B) control and cKO mice. Cumulative probability plots of EPSC amplitudes show that at both ages, a larger proportion of cells from cKO mice show no or smaller responses than controls. Inset traces show representative examples of No Response (1), Minimal Response (2) and Maximal Response (3), with their location on the cumulative probability plot indicated with numbered open circles. Inset bar graphs of mean amplitudes from all recorded cells (including those without a response) show a major reduction in both younger and older cKO mice. The values in parentheses indicate the number of animals (above) and cells (below). *p < 0.05, *** p < 0.001 in unpaired t-test. (C–E) Glutamate input to YFP+ VTA dopamine neurons does not differ between P21-24 cKO mice and controls in terms of mean evoked EPSC amplitude (C), spontaneous EPSC frequency (D) or amplitude (E).

The quenching of acridine orange fluorescence was used to assess synaptic vesicle ΔpH in response to the addition of ATP and other anions. (A) The traces show representative responses to the sequential addition of 1 mM ATP, 2 mM Cl⁻ (to provide allosteric activation of the VGLUTs) and either 4 mM Cl⁻ (Cl⁻, blue), glutamate (glut, red) or aspartate (asp, black). Bafilomycin (baf, 250 nM) inhibits the vacuolar H⁺-ATPase and allows dissipation of ΔpH. Under these conditions, glutamate produced greater acidification than equimolar Cl⁻, and aspartate had no effect. In addition, vesicles acidified with glutamate lost ΔpH more slowly in response to bafilomycin than those acidified with Cl⁻. (B) Anion-induced acidification was fit to a single exponential and the maximum change in fluorescence (ΔF) extrapolated. Although glutamate produced greater acidification at low concentrations, Cl⁻ acidified to a greater extent at high concentrations (>10 mM). (C) Fitting to a single exponential was also used to derive Initial rates of acidification (Vo ΔF/t). Cl⁻ acidifies more rapidly than glutamate, particularly at anion concentrations >4 mM. Data presented as means ± SEM. The sequential (D) or simultaneous (E) addition of 4 mM glutamate produced greater acidification than 20 mM Cl⁻ alone. (F) Synaptic vesicles acidified to approximately the same extent with either 4 mM glutamate or 14 mM Cl⁻, but the vesicles acidified with glutamate lost ΔpH much more slowly in response to bafilomycin than those acidified with Cl⁻.

(A,B) Synaptic vesicles (SV) prepared from rat ventral striatum were immunoprecipitated (IP) with antibodies to synaptophysin (syp), VGLUT2, VMAT2, Golgi matrix protein GM130, or no primary antibody, and the isolated vesicles immunoblotted for VMAT2 (A) and VGLUT2 (B). Relative to the other samples, one tenth of the indicated samples was immunoblotted to avoid saturation of the signal. Antibodies to VGLUT2 isolated SVs containing VMAT2 (A) and antibodies to VMAT2 isolated vesicles containing VGLUT2 (B), with minimal or no isolation in the absence of antibodies or with antibodies to GM130. (C) The uptake of ³H-5-HT was measured over 5 minutes in membranes from HEK293 cells transiently expressing VMAT2 with or without VGLUT2. (³H-5-HT was used for these experiments because VMAT2 recognizes all the monoamine transmitters and 5-HT is more stable than the catecholamines). Relative to aspartate, which is not recognized by the VGLUTs and has no effect on monoamine uptake (data not shown), equimolar glutamate (10 mM) stimulated transport in the presence, but not the absence, of VGLUT2, and reserpine (10 µM) largely eliminated the activity. (D) Monoamine uptake was measured over 10 minutes using synaptic vesicles from rat ventral striatum. Glutamate (10 mM) stimulated monoamine uptake relative to aspartate in the presence of 2 mM Cl⁻ (left). Increasing Cl⁻ concentration from 2 to 20 mM also stimulated monoamine uptake in the absence of glutamate (middle). Despite the sensitivity of VMAT activity to Cl⁻, glutamate stimulates monoamine uptake even at higher, physiological concentrations of Cl⁻ (20 mM) (right). **, p < 0.01; ***, p < 0.001 by two-tailed t-test. Data presented as means ± SEM.