OB-fold Up-down bundle SH3-like roll After the remarkable early successes in X-ray diffraction studies of easily crystalliz able proteins, the idea became fixed that every protein probably had well-defined secondary and tertiary structures. As more and more proteins of diverse functions were investigated, however, it became clear that the situation was more complicated. Researchers began to find regions in polypeptide chains that simply did not show up in the crystallographic maps. These regions differed from those portions of the chain that do not have any standard type of secondary structure but are still well defined in the map, such as loops and turns. The sections that are not found on crystallographic maps instead correspond to chain segments that have such mobility that they differ in conformation in different molecules in the crystal or at different times. These are referred to as intrinsically disordered proteins or regions. Evidence for such structural features of proteins in solution also comes from other techniques, notably high-resolution NMR . It became clear that many parts of many proteins have intrinsic flexibility. Indeed, we now know some cases in which an entire protein molecule behaves very much like a random coil, devoid of regular secondary and tertiary structure. These are often referred to as intrinsically disordered proteins or ID proteins. It is important to remember that intrinsic disorder is just as much a consequence of amino acid sequence as is order. ID proteins are often characterized by a combination of low hydrophobicity and high charge, both of which favor their opening up to solvents.
It is believed that the high flexibility of ID proteins facilitates their interaction with multiple, other proteins; their conformation can fit to different partners. Figure 1 presents one well-studied example, the p53 tumor suppressor protein. Mutations in p53 are associated with a number of human malignancies. The p53 protein participates in a number of biological pathways, interacting with a different partner in each case. A key to this lies in the disordered N- and C-terminal portions of p53, which can adopt many conformations. The p53 example illustrates another point: most post translational protein modifications occur in these disordered regions, presumably to tailor the protein to specific partners. Figure 2 presents an example of a protein known to recognize and react with methylated lysines and with other groups. Again, flexible peptides can adapt to the geometry of the binding site, promoting strong interactions.
Fig1. Tumor suppressor protein p53: A protein that contains both a structured region and two intrinsically disordered domains. (A) p53 domain structure and protein partners that bind to the different domains. More than 70% of the protein partners interact with the nonfolded regions of p53, which comprise only 29% of the length of the polypeptide. Some proteins, as well as DNA, interact with the structured DNA-binding domain. The exact portions of the p53 polypeptide chain that interact with the respective partners are depicted by boxes of different colors. Note that some partners interact with more than one region in p53. The boxed, numbered proteins have been co-crystallized with p53; the crystal structures with cyclin A (4), sirtuin (5), CBP (8), and S100ββ (10) are presented in part B. Post-translational modification sites are shown by colored vertical bars. Note that the majority of these sites are located in disordered regions. (B) Comparison of experimentally determined X-ray structures of the same C-terminal disordered region, in shades of red or in green, which adopts different structures upon interacting with different partners. [A, adapted from Dunker AK, Silman I, Uversky VN et al. (2008) Curr Opin Struct Biol 18: 756–764. With permission from Elsevier. B, adapted from Oldfield CJ, Meng J, Yang JY et al. (2008) BMC Genomics 9(Suppl 1): S1. With permission from BioMed Central.
Fig2. Protein segments with different disordered sequences use their flexibility to adapt to a common binding site. Crystal structures of five intrinsically disordered peptides bound to the highly structured binding pocket of protein 14-3-3ζ. The protein, shown in light blue, binds to various protein partners by interacting with their disordered segments: peptides from serotonin N-acetyltransferase, shown in green, and histone H3, shown in purple, as well as three other distinct peptides, shown in orange, red, and cyan. Each peptide has a distinct sequence and modifications and follows a different path through the 14-3-3ζ peptide binding pocket. [From Oldfield CJ, Meng J, Yang JY et al. (2008) BMC Genomics 9(Suppl 1): S1. With permission from BioMed Central.]
Functional uses for disordered protein regions have also been found in enzymes, where such regions may promote promiscuity, in the sense that a given enzyme may accommodate a variety of similar substrates. This may have facilitated enzyme evolution. The disordered regions correspond more frequently to regulatory sites rather than to the catalytic site. This is believed to be because catalysis requires a fairly rigid molecular structure, complementary to the transition state in the reaction, whereas regulation may occur through binding of a number of different effectors, which a disordered region can accomplish.
Perhaps the most striking utilization of unstructured protein regions is in signaling proteins, which must interact with a variety of other proteins as part of a signaling pathway. To date, no signaling protein–protein interactions that involve structured domains have been identified. Thus, there is a vast class of proteins with functions in signaling and regulation that utilize a lack of defined protein structure.
