3D Domain swapping is a mechanism of protein aggregation, in which a structural element of a protein fold is replaced by an identical element from another subunit. Some proteins, for instance RNase A, can assume many domain- swapped forms, thus undermining the dogma,'one sequence ? one structure' in a particularly spectacular way. Completed in a mutual fashion, it is a mechanism of protein oligomerization. In an open-ended fashion, 3D domain swapping could be a mechanism of amyloid fibril formation. In another mechanism, possibly operating together with domain swapping, a specific sequence, such as a glutamine expansion, could form a ?-spine of the fibril in a motif called steric zipper. The first connection between 3D domain swapping and amyloidogenicity was established in human cystatin C (HCC), the second - in the prion protein (PrP). In both cases, a disulfide bridge (natural in PrP, engineered in HCC) can be used for redox control of 3D domain swapping and to demonstrate that it is indeed involved in amyloid fibril formation. HCC has a naturally occurring L68Q mutant with drastically increased propensity for aggregation. The L68Q mutation occurs at the closed interface, or protein core. Mutations in other areas, such as the flexible hinge (especially deletions and insertions) can also be used to control 3D domain swapping and aggregation. Paradoxically, 3D domain swapping can also be used by Nature for prevention of nucleation processes that lead to high-order aggregates or crystals. Such a situation exists in the eye lens, where despite astronomical concentration of crystallins, the solution remains clear. One of the Nature's tricks to achieve polydispersity is to use a palindromic sequence for the swapped domain, thereby frustrating the growth of aggregates by constantly changing the interaction topology.