AlphaFold2 changed structural biology by providing high-quality structure predictions for all possible proteins. Since its inception, a plethora of applications were built on AlphaFold2, expediting discoveries in virtually all areas related to protein science. In many cases, however, optimism seems to have made scientists forget about data leakage, a serious issue that needs to be addressed when evaluating machine learning methods. Here we provide a rigorous benchmark set that can be used in a broad range of applications built around AlphaFold2/3.

Motivation: Intrinsic disorder (ID), i.e. the lack of a unique folded conformation at physiological conditions, is a common feature for many proteins, which requires specialized biochemical experiments that are not high-throughput. Missing X-ray residues from the PDB have been widely used as a proxy for ID when developing computational methods. This may lead to a systematic bias, where predictors deviate from biologically relevant ID. Large benchmarking sets on experimentally validated ID are scarce. Recently, the DisProt database has been renewed and expanded to include manually curated ID annotations for several hundred new proteins. This provides a large benchmark set which has not yet been used for training ID predictors. Results: Here, we describe the first systematic benchmarking of ID predictors on the new DisProt dataset. In contrast to previous assessments based on missing X-ray data, this dataset contains mostly long ID regions and a significant amount of fully ID proteins. The benchmarking shows that ID predictors work quite well on the new dataset, especially for long ID segments. However, a large fraction of ID still goes virtually undetected and the ranking of methods is different than for PDB data. In particular, many predictors appear to confound ID and regions outside X-ray structures. This suggests that the ID prediction methods capture different flavors of disorder and can benefit from highly accurate curated examples. Availability and implementation: The raw data used for the evaluation are available from URL: http://www.disprot.org/assessment/. Contact: [email protected]. Supplementary information: Supplementary data are available at Bioinformatics online.

BACKGROUND: Analyzing the amino acid sequence of an intrinsically disordered protein (IDP) in an evolutionary context can yield novel insights on the functional role of disordered regions and sequence element(s). However, in the case of many IDPs, the lack of evolutionary conservation of the primary sequence can hamper the study of functionality, because the conservation of their disorder profile and ensuing function(s) may not appear in a traditional analysis of the evolutionary history of the protein. RESULTS: Here we present DisCons (Disorder Conservation), a novel pipelined tool that combines the quantification of sequence- and disorder conservation to classify disordered residue positions. According to this scheme, the most interesting categories (for functional purposes) are constrained disordered residues and flexible disordered residues. The former residues show conservation of both the sequence and the property of disorder and are associated mainly with specific binding functionalities (e.g., short, linear motifs, SLiMs), whereas the latter class correspond to segments where disorder as a feature is important for function as opposed to the identity of the underlying sequence (e.g., entropic chains and linkers). DisCons therefore helps with elucidating the function(s) arising from the disordered state by analyzing individual proteins as well as large-scale proteomics datasets. CONCLUSIONS: DisCons is an openly accessible sequence analysis tool that identifies and highlights structurally disordered segments of proteins where the conformational flexibility is conserved across homologs, and therefore potentially functional. The tool is freely available both as a web application and as stand-alone source code hosted at http://pedb.vib.be/discons .

Synonymous constraint elements (SCEs) are protein-coding genomic regions with very low synonymous mutation rates believed to carry additional, overlapping functions. Thousands of such potentially multi-functional elements were recently discovered by analyzing the levels and patterns of evolutionary conservation in human coding exons. These elements provide a good opportunity to improve our understanding of how the redundant nature of the genetic code is exploited in the cell. Our premise is that the protein segments encoded by such elements might better comply with the increased functional demands if they are structurally less constrained (i.e. intrinsically disordered). To test this idea, we investigated the protein segments encoded by SCEs with computational tools to describe the underlying structural properties. In addition to SCEs, we examined the level of disorder, secondary structure, and sequence complexity of protein regions overlapping with experimentally validated splice regulatory sites. We show that multi-functional gene regions translate into protein segments that are significantly enriched in structural disorder and compositional bias, while they are depleted in secondary structure and domain annotations compared to reference segments of similar lengths. This tendency suggests that relaxed protein structural constraints provide an advantage when accommodating multiple overlapping functions in coding regions.

Vesicle trafficking systems play essential roles in the communication between the organelles of eukaryotic cells and also between cells and their environment. Endocytosis and the late secretory route are mediated by clathrin-coated vesicles, while the COat Protein I and II (COPI and COPII) routes stand for the bidirectional traffic between the ER and the Golgi apparatus. Despite similar fundamental organizations, the molecular machinery, functions, and evolutionary characteristics of the three systems are very different. In this work, we compiled the basic functional protein groups of the three main routes for human and yeast and analyzed them from the structural disorder perspective. We found similar overall disorder content in yeast and human proteins, confirming the well-conserved nature of these systems. Most functional groups contain highly disordered proteins, supporting the general importance of structural disorder in these routes, although some of them seem to heavily rely on disorder, while others do not. Interestingly, the clathrin system is significantly more disordered (~23%) than the other two, COPI (~9%) and COPII (~8%). We show that this structural phenomenon enhances the inherent plasticity and increased evolutionary adaptability of the clathrin system, which distinguishes it from the other two routes. Since multi-functionality (moonlighting) is indicative of both plasticity and adaptability, we studied its prevalence in vesicle trafficking proteins and correlated it with structural disorder. Clathrin adaptors have the highest capability for moonlighting while also comprising the most highly disordered members. The ability to acquire tissue specific functions was also used to approach adaptability: clathrin route genes have the most tissue specific exons encoding for protein segments enriched in structural disorder and interaction sites. Overall, our results confirm the general importance of structural disorder in vesicle trafficking and suggest major roles for this structural property in shaping the differences of evolutionary adaptability in the three routes.

Molecular recognition by intrinsically disordered proteins (IDPs) commonly involves specific localized contacts and target-induced disorder to order transitions. However, some IDPs remain disordered in the bound state, a phenomenon coined "fuzziness", often characterized by IDP polyvalency, sequence-insensitivity and a dynamic ensemble of disordered bound-state conformations. Besides the above general features, specific biophysical models for fuzzy interactions are mostly lacking. The transcriptional activation domain of the Ewing's Sarcoma oncoprotein family (EAD) is an IDP that exhibits many features of fuzziness, with multiple EAD aromatic side chains driving molecular recognition. Considering the prevalent role of cation-π interactions at various protein-protein interfaces, we hypothesized that EAD-target binding involves polycation- π contacts between a disordered EAD and basic residues on the target. Herein we evaluated the polycation-π hypothesis via functional and theoretical interrogation of EAD variants. The experimental effects of a range of EAD sequence variations, including aromatic number, aromatic density and charge perturbations, all support the cation-π model. Moreover, the activity trends observed are well captured by a coarse-grained EAD chain model and a corresponding analytical model based on interaction between EAD aromatics and surface cations of a generic globular target. EAD-target binding, in the context of pathological Ewing's Sarcoma oncoproteins, is thus seen to be driven by a balance between EAD conformational entropy and favorable EAD-target cation-π contacts. Such a highly versatile mode of molecular recognition offers a general conceptual framework for promiscuous target recognition by polyvalent IDPs.

Chromosomal translocations, which often generate chimeric proteins by fusing segments of two distinct genes, represent the single major genetic aberration leading to cancer. We suggest that the unifying theme of these events is a high level of intrinsic structural disorder, enabling fusion proteins to evade cellular surveillance mechanisms that eliminate misfolded proteins. Predictions in 406 translocation-related human proteins show that they are significantly enriched in disorder (43.3% vs. 20.7% in all human proteins), they have fewer Pfam domains, and their translocation breakpoints tend to avoid domain splitting. The vicinity of the breakpoint is significantly more disordered than the rest of these already highly disordered fusion proteins. In the unlikely event of domain splitting in fusion it usually spares much of the domain or splits at locations where the newly exposed hydrophobic surface area approximates that of an intact domain. The mechanisms of action of fusion proteins suggest that in most cases their structural disorder is also essential to the acquired oncogenic function, enabling the long-range structural communication of remote binding and/or catalytic elements. In this respect, there are three major mechanisms that contribute to generating an oncogenic signal: (i) a phosphorylation site and a tyrosine-kinase domain are fused, and structural disorder of the intervening region enables intramolecular phosphorylation (e.g., BCR-ABL); (ii) a dimerisation domain fuses with a tyrosine kinase domain and disorder enables the two subunits within the homodimer to engage in permanent intermolecular phosphorylations (e.g., TFG-ALK); (iii) the fusion of a DNA-binding element to a transactivator domain results in an aberrant transcription factor that causes severe misregulation of transcription (e.g. EWS-ATF). Our findings also suggest novel strategies of intervention against the ensuing neoplastic transformations.

UNLABELLED: The TOPDOM database is a collection of domains and sequence motifs located consistently on the same side of the membrane in alpha-helical transmembrane proteins. The database was created by scanning well-annotated transmembrane protein sequences in the UniProt database by specific domain or motif detecting algorithms. The identified domains or motifs were added to the database if they were uniformly annotated on the same side of the membrane of the various proteins in the UniProt database. The information about the location of the collected domains and motifs can be incorporated into constrained topology prediction algorithms, like HMMTOP, increasing the prediction accuracy. AVAILABILITY: The TOPDOM database and the constrained HMMTOP prediction server are available on the page http://topdom.enzim.hu CONTACT: [email protected]; [email protected].

Intrinsically disordered/unstructured proteins (IDPs) are extremely sensitive to proteolysis in vitro, but show no enhanced degradation rates in vivo. Their existence and functioning may be explained if IDPs are preferentially associated with chaperones in the cell, which may offer protection against degradation by proteases. To test this inference, we took pairwise interaction data from high-throughput interaction studies and analyzed to see if predicted disorder correlates with the tendency of chaperone binding by proteins. Our major finding is that disorder predicted by the IUPred algorithm actually shows negative correlation with chaperone binding in E. coli, S. cerevisiae, and metazoa species. Since predicted disorder positively correlates with the tendency of partner binding in the interactome, the difference between the disorder of chaperone-binding and non-binding proteins is even more pronounced if normalized to their overall tendency to be involved in pairwise protein-protein interactions. We argue that chaperone binding is primarily required for folding of globular proteins, as reflected in an increased preference for chaperones of proteins in which at least one Pfam domain exists. In terms of the functional consequences of chaperone binding of mostly disordered proteins, we suggest that its primary reason is not the assistance of folding, but promotion of assembly with partners. In support of this conclusion, we show that IDPs that bind chaperones also tend to bind other proteins.

MOTIVATION: The dynamic nature of protein interaction networks requires fast and transient molecular switches. The underlying recognition motifs (linear motifs, LMs) are usually short and evolutionarily variable segments, which in several cases, such as phosphorylation sites or SH3-binding regions, fall into locally disordered regions. We probed the generality of this phenomenon by predicting the intrinsic disorder of all LM-containing proteins enlisted in the Eukaryotic Linear Motif (ELM) database. RESULTS: We demonstrated that LMs in average are embedded in locally unstructured regions, while their amino acid composition and charge/hydropathy properties exhibit a mixture characteristic of folded and disordered proteins. Overall, LMs are constructed by grafting a few specificity-determining residues favoring structural order on a highly flexible carrier region. These results establish a connection between LMs and molecular recognition elements of intrinsically unstructured proteins (IUPs), which realize a non-conventional mode of partner binding mostly in regulatory functions.

Summary: Intrinsically unstructured/disordered proteins and domains (IUPs) lack a well-defined three-dimensional structure under native conditions. The IUPred server presents a novel algorithm for predicting such regions from amino acid sequences by estimating their total pairwise interresidue interaction energy, based on the assumption that IUP sequences do not fold due to their inability to form sufficient stabilizing interresidue interactions. Optional to the prediction are built-in parameter sets optimized for predicting short or long disordered regions and structured domains. Availability: The IUPred server is available for academic users at http://iupred.enzim.hu Contact: [email protected]