Protein folding takes place by interactions between reactive surfaces. Typically these surfaces consist of exposed hydrophobic side chains. Their interactions form a hydrophobic core. The intrinsic reactivity of these surfaces means that incorrect interactions may occur unless the process is controlled. Figure 8.7 illustrates what would happen. As a newly synthesized protein emerges from the ribosome, any hydrophobic patch in the sequence is likely to aggregate with another hydrophobic patch. Such associations are likely to occur at random and therefore will probably not represent the desired conformation of the protein. (For an account of the molecular interactions involved in folding see Fersht and Daggett, 2002).
Chaperones are proteins that mediate correct assembly by causing a target protein to acquire one possible conformation instead of others (for review see Ellis and van der Vies, 1991; Hartl and Hayer-Hartl, 2002). This is accomplished by binding to reactive surfaces in the target protein that are exposed during the assembly process, and preventing those surfaces from interacting with other regions of the protein to form an incorrect conformation. Chaperones function by preventing formation of incorrect structures rather than by promoting formation of correct structures. Figure 8.8 shows an example in which a chaperone in effect sequesters a hydrophobic patch, allowing interactions to occur that would not have been possible in its presence, as can be seen by comparing the result with Figure 8.7.
An incorrect structure may be formed either by misfolding of a single protein or by interactions with another protein. The density of proteins in the cytosol is high, and "macromolecular crowding" can increase the efficiencies of many reactions compared to the rates observed in vitro. Crowding can cause folding proteins to aggregate, but chaperones can counteract this effect (van den Berg, Ellis, and Dobson, 1999). So one role of chaperones may be to protect a protein so that it can fold without being adversely affected by the crowded conditions in the cytosol.
We do not know what proportion of proteins can self-assemble as opposed to those that require assistance from a chaperone. (It is not axiomatic that a protein capable of self-assembly in vitro actually self-assembles in vivo, because there may be rate differences in the two conditions, and chaperones still could be involved in vivo. However, there is a distinction to be drawn between proteins that can in principle self-assemble and those that in principle must have a chaperone to assist acquisition of the correct structure.)