Similarly, filaments with a mean diameter of 9 nm and 10 nm were found in cytoplasmic and neuritic inclusions, respectively, of FTLD-U brains, whereas fibrils with a mean diameter of 18 nm were present in intranuclear inclusions. Inclusions present in the brain of these patients showed an inability to bind amyloid-specific dyes, a finding confirmed in the spinal cords of FTLD-U patients, where no ThS positive inclusions were found. However, in the same study showing a large ThS positivity in ALS spinal cords, a remarkable and diffuse ThS staining of TDP-43 inclusions in FTLD-U Chlorhexidine hydrochloride brains was also reported. Likewise, studies in vitro describing aggregation of purified fulllength or fragmented TDP-43 have reported conflicting reports as to the structure of the resulting TDP-43 aggregates. In a first report, full-length TDP-43 was shown to form filaments unable to bind CR and ThT. By contrast, other reports have shown that peptides encompassing the most highly aggregating region of C-terminal TDP-43 are able to form, 10 nm fibrils with b-sheet secondary structure and dye binding. In this work we have Chloroquine Phosphate addressed this point by overexpressing full-length and C-terminal TDP-43 in E. coli, purifying the resulting TDP-43 containing inclusion bodies and subjecting them to a number of biophysical analyses to assess their structure and morphology. Indeed, it is increasingly accepted that bacterial IBs mainly consist of amyloid-like aggregates. Therefore, the detection of amyloid structure in the TDP-43 aggregates present in FL TDP-43 IBs and Ct TDP-43 IBs could be informative on the intrinsic propensity of this protein to form this type of protein aggregate. By contrast, the absence of amyloid-like structure in bacterial IBs would suggest a propensity to form another type of deposit. In addition to performing a biophysical investigation of the TDP-43 IBs, we have transfected eukaryotic cell cultures with FL TDP-43 IBs to evaluate their inherent toxicity. We will show that the TDP-43 aggregates contained in the bacterial IBs do not bind ThT and CR, possess random coil and b-turn secondary structure, and are highly susceptible to proteinase K digestion, thus possessing none of the amyloid distinctive hallmarks. In such unstructured/amorphous form, however, the FL TDP-43 aggregates appear to be highly toxic when transfected to the cytosol of the cultured cells, revealing their inherent ability to cause cell dysfunction. The difficulty of purifying TDP-43 has limited the study of TDP-43 aggregation in vitro and is likely to be a limiting factor in future research on TDP-43. One report exists, however, on the characterization of the aggregates formed in vitro from TDP-43 after its purification. This study has confirmed a filamentous morphology of TDP-43 aggregates, in the absence of ThT and CR binding. By contrast, four other reports have emphasized an amyloid-like structure of TDP-43 aggregates, but all of them have referred to short peptides of 13 to 50 residues, all obtained from the sequence of TDP-43. It is well known that protein fragments have generally aggregation properties different from those of the full-length protein and it was recently reported that amyloid-like aggregation, as opposed to structurally undefined aggregation, is favored by small peptides and proteins. Hence, our data and analysis suggest that FL and Ct TDP-43 form amorphous aggregates rather than amyloid-like. The finding that 
most, if not all, TDP-43 aggregates are amorphous in the spinal cord and brain of ALS and FTLD-U patients, in bacterial IBs and finally in aggregates formed in vitro from purified TDP-43.