Structural and functional analysis of cystatin E reveals enzymologically relevant dimer and amyloid fibril states.

Protein activity is often regulated by altering the oligomerization state. One mechanism of multimerization involves domain swapping, wherein proteins exchange parts of their structures and thereby form long-lived dimers or multimers. Domain swapping has been specifically observed in amyloidogenic proteins, for example the cystatin superfamily of cysteine protease inhibitors. Cystatins are twin-headed inhibitors, simultaneously targeting the lysosomal cathepsins and legumain, with important roles in cancer progression and Alzheimer's disease. Although cystatin E is the most potent legumain inhibitor identified so far, nothing is known about its propensity to oligomerize. In this study, we show that conformational destabilization of cystatin E leads to the formation of a domain-swapped dimer with increased conformational stability. This dimer was active as a legumain inhibitor by forming a trimeric complex. By contrast, the binding sites toward papain-like proteases were buried within the cystatin E dimer. We also showed that the dimers could further convert to amyloid fibrils. Unexpectedly, cystatin E amyloid fibrils contained functional protein, which inhibited both legumain and papain-like enzymes. Fibril formation was further regulated by glycosylation. We speculate that cystatin amyloid fibrils might serve as a binding platform to stabilize the pH-sensitive legumain and cathepsins in the extracellular environment, contributing to their physiological and pathological functions.

Insights in progressive myoclonus epilepsy: HSP70 promotes cystatin B polymerization.

Cystatin B (CSTB) is an anti-protease frequently mutated in progressive myoclonus epilepsy (EPM1), a devastating degenerative disease. This work shows that rat CSTB is an unstable protein that undergoes structural changes following the interaction with a chaperone, either prokaryotic or eukaryotic. Both the prokaryotic DnaK and eukaryotic HSP70 promote CSTB polymerization. Denaturated CSTB is polymerized by the chaperone alone. Native CSTB monomers are more stable than denatured monomers and require Cu(2+) for chaperone-dependent polymerization. Cu(2+) interacts with at least two conserved histidines, at positions 72 and 95 modifying the structure of native monomeric CSTB. Subsequently, CSTB becomes unstable and readily responds to the addition of DnaK or HSP70, generating polymers. This reaction depends strictly on the presence of this divalent metal ion and on the presence of one cysteine in the protein chain. The cysteine deletion mutant does not polymerize. We propose that Cu(2+) modifies the redox environment of the protein, allowing the oxidation of the cysteine residue of CSTB that triggers polymerization. These polymers are sensitive to reducing agents while polymers obtained from denatured CSTB monomers are DTT resistant. We propose that the Cu(2+)/HSP70 dependent polymers are physiological and functional in eukaryotic cells. Furthermore, while monomeric CSTB has anti-protease function, it seems likely that polymeric CSTB fulfils different function(s).