The Folded Question: A Narrative Review of Protein Folding, from Levinthal's Paradox to AlphaFold

Abstract

Protein folding is the chemistry of the sequence's decision: how a chain of amino acids—with astronomically many possible conformations—finds its native state in milliseconds, the paradox Cyrus Levinthal posed in 1968 and Christian Anfinsen's thermodynamic hypothesis answered: the sequence itself encodes the fold. This article presents a narrative review of the primary literature that built the field, from Sela, White, and Anfinsen's 1957 ribonuclease refolding and Anfinsen's 1973 principles, through Levinthal's 1968 paradox, Karplus and Weaver's 1976 diffusion-collision, Dill's 1985 hydrophobic collapse, Hemmingsen and colleagues' 1988 chaperonins, Ellis and van der Vies's 1991 chaperone synthesis, Wolynes, Onuchic, and Thirumalai's 1995 folding funnels, Wright and Dyson's 1999 intrinsically disordered proteins, Dobson's 2003 misfolding and disease, Dill and MacCallum's 2012 fifty-year assessment, and Jumper and colleagues' 2021 AlphaFold, whose neural prediction made the sequence's structure computable. The synthesis is organized around three themes: the thermodynamic settlement, in which the native state's stability and the paradox's resolution were established; the assisted and disordered revisions, in which chaperones and intrinsically disordered proteins extended the folding paradigm; and the computational settlement, in which funnels, misfolding, and AlphaFold closed the fifty-year question. It is concluded that protein folding's history is the conversion of a paradox into a science—and its latest chapter, the prediction of structure from sequence, into chemistry's most consequential computation.