Prions are the proteinaceous infectious agents responsible for Transmissible Spongiform Encephalopathies. Compelling evidence supports the hypothesis that prions are composed exclusively of a misfolded version of the prion protein (PrP(Sc)) that replicates in the body in the absence of nucleic acids by inducing the misfolding of the cellular prion protein (PrP(C)). The most common form of human prion disease is sporadic, which appears to have its origin in a low frequency event of spontaneous misfolding to generate the first PrP(Sc) particle that then propagates as in the infectious form of the disease. The main goal of this study was to mimic an early event in the etiology of sporadic disease by attempting de novo generation of infectious PrP(Sc)in vitro. For this purpose we analyzed in detail the possibility of spontaneous generation of PrP(Sc) by the protein misfolding cyclic amplification (PMCA) procedure. Under standard PMCA conditions, and taking precautions to avoid cross-contamination, de novo generation of PrP(Sc) was never observed, supporting the use of the technology for diagnostic applications. However, we report that PMCA can be modified to generate PrP(Sc) in the absence of pre-existing PrP(Sc) in different animal species at a low and variable rate. De novo generated PrP(Sc) was infectious when inoculated into wild type hamsters, producing a new disease phenotype with unique clinical, neuropathological and biochemical features. Our results represent additional evidence in support of the prion hypothesis and provide a simple model to study the mechanism of sporadic prion disease. The findings also suggest that prion diversity is not restricted to those currently known, and that likely new forms of infectious protein foldings may be produced, resulting in novel disease phenotypes.