Recently, electrophysiological evidence was given for inhibitory postsynaptic responses at dopaminergic striatal synapses. These responses were independent of the vesicular GABA transporter, VGAT, but dependent on the vesicular dopamine transporter VMAT2. The identity and the exact source of the released molecule, as well as the presence of the putative inhibitory transmitter in VMAT2 containing synaptic vesicles remain to be shown. To clarify this, in particular to determine whether GABA is responsible for the inhibitory response at dopaminergic synapses, we used the electron microscopic immunogold method to label in vivo perfusion fixed striatal tissue with antibodies recognising GABA, VGAT, VMAT2 and tyrosine hydroxylase. We show that about 13 % of tyrosine hydroxylase positive and 11 % of VMAT2 axonal terminals in the caudo-putamen contain significant labelling for GABA. Immunogold signals for tyrosine hydroxylase and VGAT was totally segregated into different pools of nerve terminals. Quantitative analyses of the distance between gold particles signalling GABA and synaptic vesicles showed that GABA was as closely associated with synaptic vesicles in tyrosine hydroxylase positive as in tyrosine hydroxylase negative nerve terminals. Likewise, in dopaminergic terminals GABA and VMAT2 immunogold particles showed a close spatial localization, strongly suggesting the presence of GABA in VMAT2 positive synaptic vesicles. Our results suggest that GABA is exocytosed together with dopamine from dopaminergic nerve terminals in the caudo-putamen through VGAT negative and VMAT2 positive synaptic vesicles.