and has been transfected with Lipofectamine (Invitrogen)/transferrin (Sigma- Aldrich), as previously described ( Marron et al., 2005; Simeoni et al., 2000 ). In the experiments involving steroid hormone treat- ments, the fetal bovine serum (FBS) was replaced with charcoal stripped-FBS (CS-FBS), to eliminate endogenous steroids. Transient transfections for immuno ?uorescence or microscopy analysis were performed on NSC34 plated in 12-well multiwell plates containing coverslips at 70,000 cells/mL hydralazine density and using with 0.7  g of plasmids coding for AR.Q23 and AR.Q46 or the ?uorescent variants GFP-AR.Q22 or GFP-AR.Q48, 3  l of transferrin solution and 2  l of Lipofectamine. Co-transfections were performed adding 0.05  g of plasmid coding for GFPu or 0.2  g of plasmid coding for mRFP-LC3 and ds-RED monomer. For Western blot analysis and

lter retardation assay NSC34 were plated in 12-wells multiwell at 80,000 cell/mL density and using 0.7  g of plasmids coding for AR.Q23 or AR.Q46, wt or G93A-SOD1, FL TDP-43 or  C TDP-43, 3  l of transferrin solution and 2  l of Lipofectamine for each sample; whereas samples for western blot analysis of proteasome functions were obtained by co-transfecting 0.7  g of AR.Q23 or AR.Q46 plasmids, 0.05  g of YFPu plasmid for each hydralazine 304-20-1 sample. To analyze the effect of 17-AAG after LC3 silencing, NSC34 were co-transfected with 0.7  g of AR.Q46 and 0.7  g shRNA against LC3 or shRNA scrambled control, 3  l of transferrin solution and 2  l of Lipofectamine for each sample. Samples for cyto ?uorimetric analysis were obtained from NSC34 cells plated in 12-well multiwell plates at 90,000 cells/ml density

co-transfected with (a) 0.7  g of GFP-ARQ.48 with 0.2  g of DsRed Monomer, (b) 0.7  g of AR.Q46 plasmid, 0.1  g of GFPu, 3  l of transferrin solution and 2  l of Lipofectamine for each sample. For real-time PCR, NSC34 were plated at 80,000 cells/mL in 6-well buy hydralazine multiwell plates. Transfections were performed with (a) 1  g of AR. Q46 plasmid, (b) 1  g of shRNA against LC3 or shRNA scrambled control, 4  l of transferrin solution and 3  l of Lipofectamine for each sample. For proteasome activity, NSC34 were plated at 80,000 cells/mL in 6-well multiwell plates. Transfections were performed with (a) 1  g of AR.Q46 plasmid, 4  l of transferrin solution and 3  l of Lipofectamine for each sample. Fluorescence, immuno ?uorescence and microscopy on NSC34 cells Cells were allowed to grow for 48 h and then ?xed and processed as previously described ( Sau et al., 2007 ). AR.Qn was detected using the rabbit polyclonal AR(N-20) (Santa Cruz Biotech, SantaCruz, CA, USA) 1:50 in milk as primary antibody, and Alexa 594 anti-rabbit (Invitrogen) at the dilution of 1:1000 in milk as secondary antibody. Cells were stained with DAPI to visualize the nuclei.

The 17-AAG effects on testosterone induced ARpolyQ total level and on cytoplasmic aggregates were analyzed by microscopy analysis 48 h after treatment. GFP-AR.Q48 bearing aggregates were estimated using a PL 10X/20 eyepiece with graticules (100 mm  10 mm in a 100 grid divisions). Transfected cells were evaluated by their staining pattern as diffusely labelled, or as containing cytoplasmic aggregates. The percentage of GFP-AR.Q48 positive cells was obtained dividing the number of cells of sausage expressing GFP-AR.Q48 by the total number of DsRed monomer transfected cells. The percentage of cells presenting GFP-AR.Q48 cytoplasmic aggregates was divided by the total number of the transfected DsRed monomer positive cells. At least 150 cells/ ?eld were counted, and three ?eld for each coverslips were analyzed. The experiments were performed in triplicate. To routinely analyze transfection ef ?ciency in living immortalized motorneuronal cells, an Axiovert 200 microscope (Zeiss Instr., Germany) equipped with FITC/TRITC/DAPI was used. Fluorescence images were captured with a Photometric CoolSnap CCD camera (Ropper Scienti ?c, NJ, USA). Images were processed using Metamorph software (Universal Imag