We utilized a two-virus system, in combination with Cre-expressing mouse lines (Gong et al., 2007), to target genetically specified projection neuron subtypes in the striatum and specifically label their monosynaptic inputs (Haubensak et al., 2010 and Wall et al., 2010). The first

virus is a Cre-dependent adenoassociated virus (AAV) that expresses TVA and rabies glycoprotein; these proteins are necessary for infection and monosynaptic spread of a modified rabies virus, respectively. The second virus is a monosynaptic rabies virus that has been modified in two ways: first, the native rabies glycoprotein in the viral membrane has been replaced with an avian sarcoma leucosis virus envelope protein (EnvA), preventing infection of AZD6738 mammalian neurons in the absence of its binding partner, TVA. Second, the glycoprotein gene from the rabies virus genome has been deleted, preventing new particles from spreading retrogradely in the absence of another source of glycoprotein. Once TVA from the AAV is expressed in Cre+ neurons, the rabies virus

specifically infects these cells. Since the Cre-dependent AAV provides Cre+ cells with a source of rabies glycoprotein, newly formed rabies virus particles can spread retrogradely from these Cre+ cells to their directly connected inputs. These input cells do not contain Cre (and thus

do not express Selleckchem Rapamycin TVA or rabies glycoprotein), preventing the rabies virus from spreading beyond this step. This technique L-NAME HCl effectively restricts rabies virus infection to only Cre+ cells and their direct, monosynaptic inputs. We injected either D1R-Cre mice, D2R-Cre mice, or wild-type C57 control mice with 180 nl of helper virus (Figure 1A), followed 3 weeks later with 180 nl of modified rabies virus injected at the same location, but along a different injection tract (Figure 1B), to avoid potential double-labeling of dopamine receptor-expressing cells along the injection tract. We then waited one week for the rabies virus to replicate and spread monosynaptically before tissue processing and analysis (Figure 1C). We mounted every second section and stained against dsRed to amplify mCherry expression from the rabies virus, and counterstained with a fluorescent Nissl marker (Neurotrace 500/525). We then scanned each slide on a semiautomatic fluorescence slide scanner and counted labeled somata to determine the numbers of retrogradely labeled cells in each brain region. Mice with fewer than 50 input cells originating outside of striatum were excluded from analysis to prevent small number bias, yielding a final data set comprising inputs from 9 D1R-Cre mice and 10 D2R-Cre mice.