RepA, the replication initiator protein of Pseudomonas pPS10 plasmid, is made of two winged-helix (WH) domains. RepA dimers undergo a structural transformation upon binding to origin DNA sequences (iterons), resulting in monomerization and ␣-helix into -strand conversion. This affects the N-terminal domain (WH1) and generates a metastable intermediate. Here it is shown that the interaction of short dsDNA oligonucleotides, including iteron or operator RepA targets, with the isolated WH1 domain promotes the assembly of different nanostructures. These range from irregular aggregates to amyloid spheroids and fibers. Their intrinsic order inversely correlates with the extent of the transformation induced by each DNA sequence on RepA. However, DNA is not a constituent of the assembled fibers, in agreement with the protein-only principle for amyloid structure. Thus, the RepA-WH1 domain on DNA binding mimics the behavior of the mammalian prion protein. The stretch of amino acids responsible for WH1 aggregation has been identified, leading to the design of mutants with enhanced or reduced amyloidogenicity and the synthesis of a peptide that assembles into a cross- structure. RepA amyloid assemblies could have a role in the negative regulation of plasmid replication. This article underlines the potential of specific nucleic acid sequences in promoting protein amyloidogenesis at nearly physiological conditions. amyloidogenesis ͉ conformational changes ͉ DNA binding ͉ fiber assembly ͉ RepA protein T he formation of protein assemblies with amyloid properties is a subject of intense research because of their involvement in severe human diseases (1, 2). An unsolved issue is the nature of the signal(s) responsible for amyloidogenesis, which includes proteolytic processing of a functional precursor and mutations, and also hyperphosphorylation (3) or the binding to elements of the cell envelope (4) or nucleic acids (5-8). The major (if not the only) component of any mature amyloid assembly is protein (9, 10). The studies of amyloidogenesis have been extended to model systems based in proteins not involved in any disease (1).Our current understanding of protein aggregation into amyloids suggests that some partially unfolded intermediates that populate the funnel-shaped folding energy landscape of virtually any protein aggregate into a stable, low-energy conformation distinct from the protein native state (1, 2). Supramolecular assembly is a common functional resource for proteins by using an ample repertoire of protein-protein interfaces. However, the family of structures responsible for the unique staining (11) and diffraction (12) properties of any protein amyloid is known as cross-: a -sheet of indefinite length is formed by the stacking, orthogonal to its axis, of -strands from each protein monomer (13). Amyloidogenesis often includes the refolding as -strands of protein regions that, although compatible with such secondary structure, are found as ␣-helices or coils in the native structure (14-18). In other cases, nearly...