A defining feature of presynaptic nerve terminals is the presence of tight clusters of synaptic vesicles anchored to active zones of the presynaptic plasma membrane, and we have proposed that principles of liquid-liquid phase separation (LLPS) may explain the formation of these vesicle clusters. Recently, we have shown that ectopic expression in fibroblasts of synapsin 1, a peripheral synaptic vesicle-associated protein, and just an integral synaptic vesicle membrane protein, synaptophysin, is sufficient to generate clusters of small vesicles (a separate cytosolic phase or condensate) highly reminiscent of synaptic vesicle clusters at presynapses and with liquid like properties. We have now followed-up this study by investigating other synaptic vesicle proteins. In fibroblasts, unlike synaptophysin, other major integral synaptic vesicle membrane proteins such as synaptobrevin2 and synaptotagmin1, SCAMP5, vGlut and vGAT fail to form condensates with synapsin but can coassmeble into the vesicle clusters formed by synaptophysin and synapsin. Another vesicle protein present in nerve temrinal, ATG9A (the only transmembrane protein of the core autophagy machimeries), was recently shown to undergo activity-dependent exo-endocytosis at synapses in parallel with synaptic vesicle proteins, raising questions about the overlap of the traffic of this protein with the traffic of synaptic vesicles. We have found that ATG9A does not coassemble into the synaptophysin/synapsin condensates, but assembles with synapsin into a distinc phase. Our findings suggest that ATG9A undergoes differential sorting relative to synaptic vesicle proteins and also point to a dual role of synapsin in controlling the clustering at synapses of SVs and ATG9A vesicles.