Protein Structure and Function. Application of Bioinformatics Methods - John Rigden 2014

Bioinformatics methods for studying the structure and function of disordered proteins
Functional classification of IDPs
Structural elements of IDPs associated with functioning

IDP transitional structural elements are involved in Molecular recognition. This functional feature is directly related to sequence-based function prediction. The presence of such elements, which are often distinguishable at both the sequence and structural levels, can be leveraged for function prediction. During the evolution of ordered Proteins, A large number of domains associated with specialized recognition Functions emerged (Pawson and Nash 2003; Seet et al. 2006), whereas their related partners—such as those from the SH3 domain (Hiroaki et al. 2001; Ferreon and Hilser 2004), the 14-3-3 domain (Busto and Iglesias 2006), or the PTB domain (Obenauer et al. 2003)—more commonly appear as short motifs within flexible protein regions. Several distinct yet interrelated concepts explain the existence of such short motifs, focusing on either Structure or sequence when discussing their functions.

Class="center">5.4.3.1. Predetermined Structural Elements

METABOLISM/2.html">THE CONCEPT OF predetermined structural elements originated from the analysis of IDP structures in complexes with their molecular partners.

A key question addressed during this analysis was whether the local structure of an IDP in complex with its molecular partner can be predicted using Secondary structure prediction algorithms (Fuxreiter et al. 2004). It was found that the prediction accuracy for these IDP secondary structure elements is higher than the analogous prediction performance for their ordered protein partners. This indicates that IDPs exhibit a pronounced predisposition toward the Conformations they adopt in bound states: presumably, recognition is mediated by elements whose existence is (partially) predetermined already in solution. This correlation is most pronounced for helices and least pronounced for coils. It has been confirmed by NMR studies of various free IDPs whose structures in isolation closely resemble their bound-state conformations. Such IDPs include the KID domain of CREB (Radhakrishnan et al. 1998), the KID domain of the Cdk inhibitor p27Kip2 (Kriwacki et al. 1996; Lacy et al. 2004), and the transactivation domain of the tumor suppressor p53 (Lee et al. 2000).

5.4.3.2. Molecular Recognition Elements and Features

The aforementioned findings are directly related to how successfully the presence of recognition elements can be predicted in IDP molecules, as demonstrated within the framework of molecular recognition features/elements (MoRFs/MoREs) (MoREs/MoRFs, standing for “molecular recognition elements/features”). A series of studies has focused on analyzing protein-Protein Complexes from the PDB database in which one molecule had fewer than 30 amino acid residues and the other had more than 30 (Oldfield et al. 2005b), or where one molecule comprised between 10 and 70 amino acid residues and the other was a globular protein (Mohan et al. 2006; Vacic et al. 2007). These studies identified 372 molecular recognition features (MoRFs), which are also referred to as molecular recognition elements (MoREs). Based on the predominant secondary structure elements, four categories of molecular recognition features are conventionally distinguished: α-MoRFs, β-MoRFs, ι-MoRFs, and mixed MoRFs. Overall, within MoRFs, 27% of residues reside in an α-helical conformation, 12% in a β-strand conformation, and about 48% in a disordered conformation; the remaining 13% of residues lack atomic coordinates. The close connection between the MoRF concept and predetermined structural elements and disorder is supported by the fact that the local structural predispositions of MoRFs are readily predictable—surpassing Globular proteins in prediction simplicity—and that the presence of MoRFs correlates with disorder in the unbound state (Mohan et al. 2006).

This analysis of MoRFs led to the idea that MoRFs can be identified by their characteristic disorder profiles. Typically, a local dip in disorder scores, particularly using PONDR VL-XT (Iakoucheva et al. 2002), serves as a strong indicator of a functionally significant recognition element. Combined with Sequence Motifs and the functional Analysis of proteins containing MoRFs, the analysis of these elements allows for well-founded hypotheses regarding the functions of IDPs/IDRs. It is worth noting that a large number of MoRF-containing proteins have been discovered among molecules performing signaling functions (Mohan et al. 2006; Vacic et al. 2007).

5.4.3.3. Short Linear Recognition Motifs

Sequence analysis of Protein-Protein Interactions has shown that in certain proteins, the recognition element is a short, highly conserved motif, often represented as a consensus sequence. Such sequences are involved, for example, in kinase substrate modification or binding to SH3 domains (Neduva and Russell 2005). These motifs are evolutionarily labile and typically consist of several conserved Specificity-determining residues interspersed among highly variable residues, with an overall motif length ranging from 5 to 25 residues. They are frequently referred to as linear motifs, eukaryotic linear motifs, or short linear motifs. An analysis of linear motifs compiled in the Eukaryotic Linear Motif database (Puntervoll et al. 2003) using disorder predictors demonstrated that linear motifs and their flanking segments of about 20 residues generally reside in locally disordered regions (Fuxreiter et al. 2007). Sequence-based prediction of linear motifs, combined with Disorder Prediction and additional functional information, can be extremely useful in predicting the functions of IDPs/IDRs.