1. Selective protein-protein interactions between neurotransmitter transporters and their synaptic targets play important roles in regulating chemical neurotransmission. We screened a yeast two-hybrid library with bait containing the C-terminal amino acids of VGLUT1 and obtained clones that encode endophilin 1 and endophilin 3, proteins considered to play an integral role in glutamatergic vesicle formation. 2. Using a modified yeast plasmid vector to enable more cost-effective screens, we analyzed the selectivity and specificity of this interaction. Endophilins 1 and 3 selectively recognize only VGLUT1 as the C-terminus of VGLUT2 and VGLUT3 do not interact with either endophilin isoform. We mutagenized four conserved stretches of primary sequence in VGLUT1 that includes two polyproline motifs (Pro1, PPAPPP, and Pro2, PPRPPPP), found only in VGLUT1, and two conserved stretches (SEEK, SYGAT), found also in VGLUT2 and VGLUT3. The absence of the VGLUT conserved regions does not affect VGLUT1-endophilin association. Of the two polyproline stretches, only one (Pro2) is required for binding specificity to both endophilin 1 and endophilin 3. 3. We also show that endophilin 1 and endophilin 3 co-localize with VGLUT1 in synaptic terminals of differentiated rat neocortical neurons in primary culture. These results indicate that VGLUT1 and both endophilins are enriched in a class of excitatory synaptic terminals in cortical neurons and there, may interact to play an important role affecting the vesicular sequestration and synaptic release of glutamate.

VGLUT2 is one of three vesicular glutamate transporters that play crucial roles in glutamatergic excitatory neurotransmission. We explored the functional properties of the rat VGLUT2 by heterologous expression of VGLUT2 in Xenopus oocytes. Immunocytochemical analysis indicated that most VGLUT2 protein was expressed in intracellular compartments but that some expression occurred also on the plasma membrane. Functional analysis revealed VGLUT2 to be active in two independent modes, namely, uptake into intracellular organelles and efflux at the plasma membrane. VGLUT-specific transport was identified based on the strong preference for glutamate over aspartate--in contrast to plasma-membrane or mitochondrial glutamate transporters--and sensitivity to known VGLUT blockers. VGLUT2 expression in oocytes (1) stimulated the influx of L-[(3)H]glutamate, but not D-[(3)H]aspartate, into digitonin-permeabilized oocytes and (2) stimulated efflux of L-glutamate, but not L-aspartate, from intact oocytes preinjected with (3)H-labeled amino acids. In the latter assay, cellular efflux of glutamate (which was blocked by rose bengal and trypan blue) may be analogous to vesicular packaging of glutamate. Our data are consistent with VGLUT2-mediated H(+)/L-glutamate antiport, but not antiport with chloride. Expression of mammalian VGLUT1 and VGLUT3 also stimulated L-[(3)H]glutamate efflux from Xenopus oocytes, suggesting that this phenomenon is a general feature of vesicular glutamate transporters. Our findings support the idea that vesicular glutamate transporters, when transiently expressed on the neuronal plasma membrane, may mediate Ca(2+)-independent glutamate leakage in addition to their traditional role of packaging glutamate into synaptic vesicles for Ca(2+)-dependent exocytosis.