Protein interaction energy landscapes are shaped by functional and also non-functional partners

Abstract

38In the crowded cell, a strong selective pressure operates on the proteome to limit the 39 competition between functional and non-functional protein-protein interactions. We 40 developed an original theoretical framework in order to interrogate how this competition 41 constrains the behavior of proteins with respect to their partners or random encounters. Our 42 theoretical framework relies on a two-dimensional (2D) representation of interaction energy 43 landscapes with 2D energy maps that reflect in a synthetic way the propensity of a protein to 44 interact with another protein. We investigated the propensity of protein surfaces to interact 45 with functional and arbitrary partners and asked whether their interaction propensity is 46 conserved during the evolution. Therefore, we performed several thousands of cross-docking 47 behavior of a protein in interaction with hundreds of selected partners and opens the way for 56 further developments to study the behavior of proteins in a specific environment. 57 58 Biomolecular interactions are central for many physiological processes and are of utmost 59 importance for the functioning of the cell. Particularly protein-protein interactions have 60 attracted a wealth of studies these last decades [1-5]. The concentration of proteins in a cell 61 has been estimated to be approximately 2-4 million proteins per cubic micron [6]. In such a 62 highly crowded environment, proteins constantly encounter each other and numerous non-63 specific interactions are likely to occur [7-10]. For example, in the cytosol of S. cerevisiae a 64 protein can encounter up to 2000 different proteins [11]. In this complex jigsaw puzzle, each 65protein has evolved to bind the right piece(s) in the right way (positive design) and to prevent 66 misassembly and non-functional interactions (negative design) [12][13][14][15][16]). 67Consequently, positive design constrains the physico-chemical properties and the evolution of 68 protein-protein interfaces. Indeed, a strong selection pressure operates on binding sites to 69 maintain the functional assembly including the functional partner and the functional binding 70 mode. For example, homologs sharing at least 30% sequence identity almost invariably 71 interact in the same way [17]. Conversely, negative design prevents proteins to be trapped in 72 the numerous competing non-functional interactions inherent to the crowded environment of 73 the cell. Many studies were reported on the relationship between the propensity of proteins for 74 promiscuous interactions and their abundances or surface properties [18][19][20][21]. Particularly, it 75 has been shown that the misinteraction avoidance shapes the evolution and physico-chemical 76properties of abundant proteins, resulting in a slower evolution and less sticky surfaces than 77 what is observed for less abundant ones [18,[22][23][24][25][26]. The whole surface of abundant proteins 78 is thus constrained, preventing them to engage deleterious non-specific interactions that could 79 be of dramatic impac...

Using collections of structural models to predict changes of binding affinity caused by mutations in protein–protein interactions

“…We obtain several poses or conformations of a PPI. In line with Lopes and coworkers, 32 some of them may share a similar interface, or overlap with similar interfaces of the same interaction, while others may reproduce different interacting regions. Consequently, we follow two different approaches to analyze the models: (a) using all conformations to estimate the binding; or (b) classifying and clustering similar interfaces and analyzing them separately.…”

“…Therefore, by using a collection of poses we ensure the inclusion of many suitable conformations producing the interaction. Interestingly, not all poses of an interaction have the same or similar interface, 32 which affects the relevance of mutations in the change of affinity.…”

“…Conformational variability in PPIs has also been invoked in the model of “conformational selection” 30,31 which suggests that binding entails a choice of the correct conformer for binding out of a rapidly interconverting ensemble 29 . Recent works have also introduced the relevant effect that regions outside the interface can play on the affinity of an interaction, which may imply the contribution of different conformations 32 . Methods to characterize conformational variability in PPIs can be classified into those that generate an ensemble of discrete conformations and those that generate an ensemble of continuous conformations 33 .…”

Moving pictures: Reassessing docking experiments with a dynamic view of protein interfaces

“…Structural data regarding macromolecular assemblies is available in the Protein Data Bank (PDB) [10], in particular thanks to recent advances in cryo-EM techniques, which enabled to reach atomic resolution for very large complexes [11,12]. On the other hand, in silico methods represent a complementary approach for the identification of protein partners [13][14][15] and determining the 3-dimensional structure of protein assemblies [16][17][18].…”

Protein Abundance Biases the Amino Acid Composition of Disordered Regions to Minimize Non-functional Interactions

“…Importantly, however, properties of disordered regions that promote functional recognition are also expected to promote dysfunctional recognition [43], [44], [45]. Indeed, disordered regions naturally exhibit high solvent exposure, and we know that exposed protein surfaces show a natural tendency to bind one another promiscuously [46], [47], suggesting that a delicate balance exists in cells, between functional (selected) and non-functional (non-selected) binding [44], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], [60]. Such a delicate balance is reflected in the dosage sensitivity associated with over-expression of proteins that contain disordered regions or that are highly promiscuous [43], [44], [61], [62].…”