Initially, the co-existence of fast-acting neurotransmitters in a given terminal was an occasional observation but presently it is known to occur in various animal species and neuronal systems, from invertebrates to humans.

Recently, two research groups have shown that coexistence of VGLUT1 and VGAT may occur at cortical and hippocampal glutamatergic synapses (Fattorini et al., 2009; Zandler et al., 2010). However, a clear functional proof of GABA co-release at central glutamatergic synapses is still lacking and its functional significance is greatly debated.

Using patch-clamp recordings, we demonstrate that GABA is effectively co-released at hippocampal excitatory autapses. The amount of co-released GABA is not sufficient to activate mismatched GABA_A receptors expressed at the postsynaptic sites of glutamatergic autapses. On the contrary, due to its ten-fold higher affinity, the co-released GABA is able to activate presynaptic GABA_B receptors that are highly expressed on glutamatergic terminals. Presynaptic GABA_B receptors are constitutively activated by co-released GABA, maintaining a "brake" on the probability of release and on the readily releasable pool of glutamatergic autapses.

These results indicate that the co-release of GABA at excitatory nerve terminals, by operating an autocrine GABA_B-mediated down regulation of excitatory transmission, might play a homeostatic role to restrain neuronal excitability under physiopathological conditions.