Alpha-synuclein (SNCA), an abundantly expressed presynaptic protein, is implicated in Parkinson's disease (PD). Since over-expression of human SNCA (hSNCA) leads to death of dopaminergic (DA) neurons in human, rodent and fly brain, hSNCA gene silencing may reduce levels of toxic forms of SNCA and ameliorate degeneration of DA neurons in PD. To begin to develop a gene therapy for PD based on hSNCA gene silencing, two AAV gene silencing vectors were designed, and tested for efficiency and specificity of silencing, as well as toxicity in vitro. The same hSNCA silencing sequence (shRNA) was used in both vectors, but in one vector, the shRNA was embedded in a microRNA backbone and driven by a pol II promoter, and in the other the shRNA was not embedded in a microRNA and was driven by a pol III promoter. Both vectors silenced hSNCA to the same extent in 293T cells transfected with hSNCA. In DA PC12 cells, neither vector decreased expression of rat SNCA, tyrosine hydroxylase (TH), dopamine transporter (DAT) or the vesicular monoamine transporter (VMAT). However, the mir30 embedded vector was significantly less toxic to both PC12 and SH-SY5Y cells. Our in vitro data suggest that this miRNA-embedded silencing vector may be ideal for chronic in vivo SNCA gene silencing in DA neurons.