Direct coupling analysis of ankyrin-repeat protein

ESPADA, R; MORA, T; WALCZAK, AM; FERREIRO, DU

San Pablo

Simposio; Symposium on Current Topics in Molecular Biophysics; 2014

USP

The amino acids sequence of natural proteins contains information about the folding and motion of biomolecules, defining the energy landscape. Due to conflicting signals between folding (physical) and functional (biological) constraints and the inherent evolutionary noise, it is not trivial to decompose how these factors contribute to shaping the amino acid sequences. Here we analyze the ankyrin-repeat protein family, which have copies of a ~33 amino-acid motif and typically fold into elongated objects of stacked alpha-helices an beta-turns. The apparent simple architecture of repeat proteins suggests that the folding properties of a complete domain (the stability and the cooperativity of the array of repeats) can be derived from the microscopic description of the balance of the energetic terms of each element and its interactions with its near neighbors. By analyzing the correlation structure of alignments of ankyrin repeats, we quantify the Mutual and the Direct Information between the positions and distinguish positions with varying information content. We detect sporius highly correlated positions due to the repetitive nature of this family of proteins. We propose a correction to deal with this bias and improve the detection of positions with high information generated by correlated mutations. We find that most amino acid pair-interactions can be well predicted by the corrected direct information ranking, even when no structural information was used as input.