Structural analysis of mammalian protein phosphorylation at a proteome level

Abstract

Highlights d Phosphorylation sites are structurally stratified into four distinct groups d Dynamic core phospho-sites are more functional than static ones d Dynamic core and interface phospho-sites depend on cell cycle more than static ones d Conformational Q9 change during phosphorylation is associated with functionality

The structural context of posttranslational modifications at a proteome-wide scale

“…To assess functional relevance of PTMs on a more global scale, Beltrao and colleagues recently presented a machine learning model that uses information from different features indicative of proteomic, structural, regulatory, or evolutionary relevance to derive a functional score for a large catalog of phosphosites [ 6 ]. Another recent study directly evaluated the functional relevance of phosphorylations purely based on available structural information [ 7 ]. Based on these and many previous studies, we know that phosphorylations are predominantly observed on spatially accessible amino acids and particularly in intrinsically disordered regions (IDRs) [ 7 – 9 ].…”

The structural context of PTMs at a proteome wide scale

“…Although mass-spectrometry based proteomics routinely enables the identification and quantification of thousands of PTMs, systematic assessment of their functional relevance remains a persisting challenge. While previous work already demonstrated the merits of structural information for PTM analyses (Beltrao et al, 2012; Iakoucheva et al, 2004; Kamacioglu et al, 2021; Ochoa et al, 2019; Tyanova et al, 2013), the recent revolution in computational protein structure prediction (Baek et al, 2021; Jumper et al, 2021) only now enables the proteome-wide integration of structural information with PTM data. In this study, we provide a first overview of how the comprehensive structural context of all detected PTMs can provide global and protein-specific insights into biological mechanisms, to filter in vitro datasets for physiological relevance and to identify promising candidates for biochemical follow-up studies.…”

“…Based on these and many previous studies, we know that phosphorylations are predominantly observed on spatially accessible amino acids and particularly in intrinsically disordered regions (IDRs) (Iakoucheva et al, 2004;Kamacioglu et al, 2021;Tyanova et al, 2013). Furthermore, it stands to reason that phosphorylations in flexible regions within folded domains and on binding interfaces are more likely to be functional compared to those that are buried or less accessible in rigidly folded regions (Kamacioglu et al, 2021). Although these studies impressively highlight the value of integrating structural information into the analysis of PTMs, they have been limited to phosphorylation and to the set of available experimentally derived structures deposited in PDB, which furthermore inherently favor stable regions of proteins (Burley et al, 2021).…”

“…To assess functional relevance of PTMs on a more global scale, Beltrao and co-workers recently presented a machine learning model that uses information from different features indicative of proteomic, structural, regulatory or evolutionary relevance to derive a functional score for a large catalog of phosphosites (Ochoa et al, 2019). Another recent study directly evaluated the functional relevance of phosphorylations purely based on available structural information (Kamacioglu et al, 2021). Based on these and many previous studies, we know that phosphorylations are predominantly observed on spatially accessible amino acids and particularly in intrinsically disordered regions (IDRs) (Iakoucheva et al, 2004; Kamacioglu et al, 2021; Tyanova et al, 2013).…”

“…Another recent study directly evaluated the functional relevance of phosphorylations purely based on available structural information (Kamacioglu et al, 2021). Based on these and many previous studies, we know that phosphorylations are predominantly observed on spatially accessible amino acids and particularly in intrinsically disordered regions (IDRs) (Iakoucheva et al, 2004; Kamacioglu et al, 2021; Tyanova et al, 2013). Furthermore, it stands to reason that phosphorylations in flexible regions within folded domains and on binding interfaces are more likely to be functional compared to those that are buried or less accessible in rigidly folded regions (Kamacioglu et al, 2021).…”