What makes prions adopt toxic, self-perpetuating conformations? The question has baffled researchers for years. In the May 9 Nature online, Peter Tessier and Susan Lindquist from Boston’s Whitehead Institute of Biomedical Research suggest that a few short sections of amino acids may be the answer. They report that small recognition elements facilitate conversion of normal to toxic conformations, and that they also confer strain and species specificity on prions. The study implies that toxic prions fold in a fundamentally different way from how proteins with globular domains typically do it. Folding of the latter is driven by large numbers of intramolecular interactions that act together to fold up the whole domain, whereas this study suggests two separate prions fold into an amyloid structure when very small bits of their sequence interact. The discovery is based on protein behavior on peptide arrays. If confirmed, it could change the way scientists view and study protein aggregation.
Prions may play beneficial, if poorly understood, roles when in their normal conformation. In the brain, they may even facilitate learning and memory (see ARF related news story). But once they adopt a toxic conformation, they begin converting other normal prions to the poisonous form. Like the amyloid-β peptide, toxic prions have a high affinity for each other and form large amyloid aggregates.
To study the intricacies of the prion conversion process, Tessier and Lindquist turned to surface chemistry. They used peptide arrays to investigate what sequences within the baker’s yeast prion, Sup35, are capable of promoting aggregation. The researchers spotted overlapping 20-amino-acid-long Sup35 peptides onto glass slides and then incubated them with NM prion, a fragment of Sup35 that consists of the combined N-terminal and middle region and retains prion properties.
Only five of the 20-mer peptides bound strongly to NM. They spanned a region (amino acids 9-39) that mutation and cross-linking analysis had previously linked to prion aggregation, suggesting that these peptides may represent prion-interaction motifs. In support of this idea, the authors found that NM continued to accumulate on the spotted peptides for up to five days. The accumulation resisted detergent, as do prion amyloids, and when the researchers scraped the aggregates from the slides and examined them by electron microscopy, they found prion-like fibers. The data “establish that with an extraordinary degree of specificity, small elements of the Sup35 protein sequence are able to recognize molten conformers of full-length NM and convert them via this recognition to a self-templating state,” write the authors. The authors found that peptides from Candida albicans Sup35 can similarly seed growth of C. albicans NM fibers.
Prions are generally strain- and species-specific. They only convert their own kind and rarely jump the species barrier—though in the case of mad cow disease, or bovine spongiform encephalopathy (BSE), ingested bovine prions have occasionally caused human disease. True to form, C. albicans and Saccharomyces cerevisiae NM, which share little sequence identity, cannot cross-seed. But there is a genetically engineered NM chimera that can seed production of Sup35 aggregates in both yeasts. Tessier and Lindquist found that the same peptides that seed S. cerevisiae and C. albicans NM aggregates can also seed formation of chimeric fibers. Furthermore, seeding by the peptides followed previously established rules of temperature and sequence specificity. At colder temperatures (4 degrees centigrade), C. albicans peptides failed to seed formation of chimera fibers, while at warmer temperatures (37 degrees) the S. cerevisiae peptides were poor seeds. Similarly, mutations that render the chimera incapable of seeding aggregation of C. albicans Sup35 also made this prion resistant to C. albicans peptides. According to the authors, the results suggest that small, highly localized elements within prions govern self-recognition and conversion of normal prion to toxic conformations.
This result came as a surprise to the authors. Previous data had suggested that the N-terminal domain of Sup35 can be jumbled up and still seed formation of aggregates, pointing to amino acid composition rather than primary sequence as crucial for the conversion from wild-type to toxic. Tessier and Lindquist suggest that scrambled NM proteins may form prions at low efficiency, or that new prion elements are formed during the scrambling process. Whether the high concentration of peptides in the array spots (~300-fold that in solution) may also partly explain their ability to nucleate aggregation is unclear. Primary structure may not be the sole determinant of prion activity, either. Independent work from several groups suggests that secondary structure can influence whether a prion can jump the species barrier (see ARF related news story).
In more general terms, the protein arrays used here may turn out to be a boon for studying protein aggregation, and may be useful to AD researchers studying Aβ aggregates. “From a methodological perspective, it is empowering that peptide arrays, which are typically used to study the specificities of folded proteins, can also be used to identify sequences that drive structural transitions in natively unfolded proteins,” the authors write.—Tom Fagan
- Tessier PM, Lindquist S. Prion recognition elements govern nucleation, strain specificity and species barriers. Nature. 2007 May 31;447(7144):556-61. PubMed.