Correspondences between low‐energy modes in enzymes: Dynamics‐based alignment of enzymatic functional families

Abstract

Proteins that show similarity in their equilibrium dynamics can be aligned by identifying regions that undergo similar concerted movements. These movements are computed from protein native structures using coarse-grained elastic network models. We show the existence of common large-scale movements in enzymes selected from the main functional and structural classes. Alignment via dynamics does not require prior detection of sequence or structural correspondence. Indeed, a third of the statistically significant dynamics-based alignments involve enzymes that lack substantial global or local structural similarities. The analysis of specific residue-residue correspondences of these structurally dissimilar enzymes in some cases suggests a functional relationship of the detected common dynamic features. Including dynamics-based criteria in protein alignment thus provides a promising avenue for relating and grouping enzymes in terms of dynamic aspects that often, though not always, assist or accompany biological function.Keywords: structure and function of enzymes; large-scale movements; protein alignment; dynamics-based alignment Supplemental material: see www.proteinscience.org Available alignment tools detect similarities among proteins in the first step of the logical cascade: sequence ! structure ! function. We introduce and apply a general and systematic algorithm to address functional relationships of enzymes by including dynamics in the alignment procedure. We focus on a common although not universal feature of enzymatic function, internal large-scale concerted movements. A large body of evidence links these movements to the structural changes that often accompany protein functions. For example, the displacements involved in allosteric changes in many proteins occur along the collective coordinates corresponding to the low-energy modes of the two stable structures (Delarue and Sanejouand 2002;Falke 2002;Rod et al. 2003;Alexandrov et al. 2005;Ming and Wall 2005;Smith et al. 2005;Zheng et al. 2007).The collective and large-scale character of these fluctuations has justified their characterization by simplified approaches, typically elastic network models (ENM) (Bahar et al. 1997;Hinsen 1998;Atilgan et al. 2001;Delarue and Sanejouand 2002;Micheletti et al. 2004;Sulkowska et al. 2007). These models rely on a simplified Reprint requests to: Cristian Micheletti, International School for Advanced Studies and CNR-INFM Democritos, Via Beirut 2-4, 34014 Trieste, Italy; e-mail: michelet@sissa.it; fax: 39-040-3787528.Abbreviations: ENM, elastic network model; MD, molecular dynamics; RMSD, root mean square distance; RMSIP, root mean square inner product.Article published online ahead of print. Article and publication date are at

Biophysics of protein evolution and evolutionary protein biophysics

“…Conformational flexibility, dynamics of protein folded states and allosteric transitions often can be deduced to a reasonable degree from the structure(s) of the protein in question using elastic network models for folded-state dynamics [191][192][193][194] or native-centric Gō-like potentials [195] with multiple folding basins ( [196]; reviewed in [177]). Similar to the aforementioned case for the probable existence of geometric/ topological constraints on the evolution of folding stability and cooperativity ( §2.5), the success of structure-based native-centric modelling in rationalizing conformational dynamics and allosteric transitions suggests that there are significiant structural constraints on the evolution of functional folded-state dynamics.…”

“…Similar to the aforementioned case for the probable existence of geometric/ topological constraints on the evolution of folding stability and cooperativity ( §2.5), the success of structure-based native-centric modelling in rationalizing conformational dynamics and allosteric transitions suggests that there are significiant structural constraints on the evolution of functional folded-state dynamics. The computational efficiency of elastic network models also allows enzymes that are dissimilar in sequence and structure yet probably perform similar functions to be detected by their similar dynamic properties [194,197], making it possible for relationships between evolutionary conservation and conformational dynamics to be explored [198].…”

Comparing the intrinsic dynamics of multiple protein structures using elastic network models

“…These show the interest of the community and the potential of ENMs for the characterisation of intrinsic dynamics in a way that can complement existing structural classifications systems. In addition, there have been efforts in using dynamic information as a means of aligning different proteins, and their developments have provided insight into comparing dynamics in general [69,[107][108][109].…”

DISCO: A New Algorithm for Detecting 3D Protein Structure Similarity

“…The result of a pairwise alignment or comparison procedure concerns two particular proteins, while the multiple comparison procedure gives similarity results between a certain protein and a list of other known proteins. Another term called 'dynamics-based-alignment' has also been recently introduced ( [11]) and is intended to compare the dynamic motions of different proteins. Furthermore, methods that perform structural alignment between a model protein and the true known structure of the protein have also been proposed ( [12]) and they are known as model comparison methodologies, but these last two subjects are out of the scope of the current work.…”