Intrinsic Disorder and Protein Function

2 The dominant view of protein structure-function is that an amino acid sequence specifies a (mostly) fixed three-dimensional (3-D) structure that is a prerequisite to protein function. In contrast to the dominant view, many proteins display functions requiring the disordered state. Our purpose here is to provide a catalogue of disorder-function relationships. The very important molecular details in each example can be obtained from the references provided or from several excellent reviews and commentaries (1-9). For ordered protein, the ensemble members all have the same time-averaged canonical set of Ramachandran angles along their backbones. For intrinsically disordered protein, the ensemble members have different (and typically dynamic) Ramachandran angles. Such disorder has been characterized by a variety of methods including x-ray crystallography, NMR spectroscopy, CD spectroscopy, and protease sensitivity to name several. Each of these methods has advantages and limitations that are discussed in more detail elsewhere (10). Although a few disordered proteins and regions have been characterized by several methods as noted below, it would be useful to have more examples with multiple methods of characterization. In attempts to discover generalities from the known disorder examples, we recently used bioinformatics coupled with data mining (11-15). The results suggested that thousands of natively disordered proteins exist, representing a very substantial fraction of the proteins in the commonly used sequence databases (13, 16). From these and related database predictions and from a set of functionally important disordered proteins, Wright and Dyson (17) called for a re-assessment of the view that 3-D structure is always a prerequisite to protein function. In this article, we discuss the following topics: 1. how common is intrinsic disorder?; 2. intrinsic disorder in vivo; 3. functional annotations for 90 proteins having physically characterized regions of disorder; 4. disordered regions without known function 5. a structure-function proposal called " the protein trinity " ; 6. the functional repertoires of ordered and disordered protein, and 7. the need for a Disordered Protein Database (DisProt) to complement the Protein Data Bank (PDB). How Common is Intrinsic Disorder? A series of predictors of natural disordered regions (PONDRs) have been developed using amino acid sequence as inputs and giving intrinsic order or disorder tendencies as outputs (11, 14, 15, 18, 19). The various PONDRs are distinguished by different training sets, by different data representations for their inputs, and by different machine learning models for their development. For PONDR VL-XT 1 , currently the …