The formation of protein structures is still a puzzling question. The first theories were developed 55 years ago, when Anfinsen et al. demonstrated that proteins can fold by themselves [1]. Additionally, Levinthal and coworkers suggested that an amino acid chain can adapt a high number of possible folds and random searching through the many possible conformations could never happen in a reasonable folding time [2]. Proteins are indeed able to fold on a time scale of milliseconds, which led Levinthal and coworkers to the conclusion that proteins follow a programmed structure formation pathway.

The current theory of the folding funnel describes folding based on an energy landscape model. An unfolded protein chain folds to the native state (global energy minimum) by passing local minima on the folding pathway. These local minima are called intermediates. High-resolution optical tweezers is a well-suited method to shed light on the folding processes of proteins. Using this method, the molecule of interest can be unfolded and refolded several times by successively stretching and relaxing the protein. Unfolding of a protein reveals on the one hand the stability of the protein determined by the unfolding force; on the other hand, unfolding intermediates which show stable intermediate structures on the unfolding pathway can be detected. In the refolding process of a completely stretched amino acid chain, refolding intermediates can be populated and be further investigated in the terms of kinetic and structural characterization. This information gives great insights into the folding pathway in terms of folding seeds and misfolded structures, which can accelerate or delay protein folding, respectively.