Protein Folding
Anfinsen's Classic Experiment
The "Protein Folding Problem" asks a very simple question: "How does the primary structure of a protein determine its 2° and 3° structure?". We have known for many decades that proteins fold into their correct 3-D structures inside the cell. But correct folding during synthesis on the ribosome or later with assistance from unknown cellular factors could explain the in vivo results.
In the 1960's, C. B. Anfinsen and his coworkers at the NIH performed a series of seminal experiments in vitro that answered a key part of the problem. The original work led Anfinsen to propose his "Thermodynamic Hypothesis", which states that the native conformation of a protein is adopted spontaneously. In other words, there is sufficient information contained in the protein sequence to guarantee correct folding from any of a large number of unfolded states. A schematic diagram of Anfinsen's experiment is shown below in two parts:

1. The Observation
Ribonuclease A (RNaseA) is an extracellular enzyme of 124 residues with four disulfide bonds. In the first phase of the experiment, the S-S bonds were reduced to eight -SH groups (using mercaptoethanol, HS-CH2-CH2-OH); the protein was then denatured with 8 M urea. Under these conditions, the enzyme is inactive and becomes a flexible random polymer. In the second phase, the urea was slowly removed (dialysis); then the the -SH groups were oxidized back to S-S bonds. If the protein was able to regain its native structure spontaneously after removal of the urea, we expect that it would also regain its activity. In fact, the activity was >90% of the untreated enzyme. Moreover, sequence analysis showed that nearly all of the correct S-S bonds had been formed.
(cf. Fig. 4.20 in Campbell for this half of the experiment.)
2. The Control
A reasonable objection can be raised to the above result by suggesting that perhaps RNaseA was not completely unfolded in 8 M urea. To address this class of objections, RNAseA was first reduced and denatured as above. But in the second phase, the enzyme was first oxidized to form S-S bonds, and then the urea was removed, i.e. the order of steps in the second phase of the experiment was reversed. The resulting activity was only about 1-2% of the untreated enzyme. Sequence analysis showed a random assortment of S-S bonds ("Scrambled" in the diagram).
[Sidelight: Can we account for the 1-2% recovery of activity in the "Scrambled" sample?]