Search

The spatial and temporal organization of interactions between proteins underlie the regulation of most cellular processes. The requirement for such interactions to be specific predisposes a view that protein–protein interactions are relatively static and are formed through the stable complementarity of the interacting partners. A growing body of reports indicate, however, that many interactions lead to fuzzy complexes with an ensemble of conformations in dynamic exchange accounting for the observed binding. Here, we discuss how NMR has facilitated the characterization of these discrete, dynamic complexes and how such characterization has aided the understanding of dynamic, condensed phases of phase-separating proteins with exchanging multivalent interactions.

The power of structural information for informing biological mechanisms is clear for stable folded macromolecules, but similar structure-function insight is more difficult to obtain for highly dynamic systems such as intrinsically disordered proteins (IDPs) which must be described as structural ensembles. Here, we present IDPConformerGenerator, a flexible, modular open-source software platform for generating large and diverse ensembles of disordered protein states that builds conformers that obey geometric, steric, and other physical restraints on the input sequence. IDPConformerGenerator samples backbone phi (φ), psi (ψ), and omega (ω) torsion angles of relevant sequence fragments from loops and secondary structure elements extracted from folded protein structures in the RCSB Protein Data Bank and builds side chains from robust Monte Carlo algorithms using expanded rotamer libraries. IDPConformerGenerator has many user-defined options enabling variable fractional sampling of secondary structures, supports Bayesian models for assessing the agreement of IDP ensembles for consistency with experimental data, and introduces a machine learning approach to transform between internal and Cartesian coordinates with reduced error. IDPConformerGenerator will facilitate the characterization of disordered proteins to ultimately provide structural insights into these states that have key biological functions.

Tubulinopathies and neurodevelopmental ciliopathies are two groups of genetic disorders characterized by abnormal brain development resulting in structural brain malformations. Tubulinopathies are caused by dominant missense mutations in genes encoding for tubulins, the building blocks of microtubules. Neurodevelopmental ciliopathies are mostly recessive disorders caused by defects in the function of the primary cilium, a sensory organelle that modulates signaling pathways important for brain development. Though more than 40 genes have been associated with neurodevelopmental ciliopathies, many patients still do not have an identified genetic etiology. Herein, we present a novelde novoheterozygous missense variant in Tubulin Beta Class I (TUBB) identified through whole-genome sequencing analysis in a patient with both ciliopathy and tubulinopathy brain features. While microtubules are fundamental to primary cilia formation and function, no association between mutations in tubulin genes and neurodevelopmental ciliopathies has been found to date. Using patient-derived cells and gene-edited isogenic cell lines, we show that the identified variant impairs the early stages of cilia formation by altering microtubule dynamics and structure. Furthermore, we demonstrate that the disease mechanism is not haploinsufficiency and that other patient mutations inTUBBaffect cilia formationin vitro, putting forward defective ciliogenesis as a contributing pathogenic factor in a subset of tubulinopathy patients.

Algorithms for computational protein design usually begin with a 3D structure and design an amino acid sequence that will fold into that structure. Since intrinsically disordered protein regions lack stable 3D structures, strategies for their design have been limited. Here we describe and validate a general computational strategy for design of intrinsically disordered regions (IDRs). Our algorithm designs synthetic IDRs by minimizing the distance of an initially random amino acid sequence to a known IDR in a high-dimensional space of molecular properties, such as repeat content, charge and hydrophobicity. We tested a handful of IDRs designed to target proteins to mitochondria and heat-induced condensates, and several showed the expected patterns in cells.One sentence summaryWe design synthetic mitochondrial targeting signals and condensate forming intrinsically disordered regions

Supplementary Methods from Synuclein-γ Targeting Peptide Inhibitor that Enhances Sensitivity of Breast Cancer Cells to Antimicrotubule Drugs

Supplementary Figures 1-4 from Synuclein-γ Targeting Peptide Inhibitor that Enhances Sensitivity of Breast Cancer Cells to Antimicrotubule Drugs

Electrostatic interactions and charge balance are important for the formation of biomolecular condensates involving proteins and nucleic acids. However, a detailed, atomistic picture of the charge distribution around proteins during the phase-separation process is lacking. Here, we use solution NMR spectroscopy to measure residue-specific near-surface electrostatic potentials ( ϕ ENS ) of the positively charged carboxyl-terminal intrinsically disordered 103 residues of CAPRIN1, an RNA-binding protein localized to membraneless organelles playing an important role in messenger RNA (mRNA) storage and translation. Measured ϕ ENS values have been mapped along the adenosine triphosphate (ATP)–induced phase-separation trajectory. In the absence of ATP, ϕ ENS values for the mixed state of CAPRIN1 are positive and large and progressively decrease as ATP is added. This is coupled to increasing interchain interactions, particularly between aromatic-rich and arginine-rich regions of the protein. Upon phase separation, CAPRIN1 molecules in the condensed phase are neutral ( ϕ ENS ≈ 0 mV), with ∼five molecules of ATP associated with each CAPRIN1 chain. Increasing the ATP concentration further inverts the CAPRIN1 electrostatic potential, so that molecules become negatively charged, especially in aromatic-rich regions, leading to re-entrance into a mixed phase. Our results collectively show that a subtle balance between electrostatic repulsion and interchain attractive interactions regulates CAPRIN1 phase separation and provides insight into how nucleotides, such as ATP, can induce formation of and subsequently dissolve protein condensates.

We consider a generic representation problem of internal coordinates (bond lengths, valence angles, and dihedral angles) and their transformation to 3-dimensional Cartesian coordinates of a biomolecule. We show that the internal-to-Cartesian process relies on correctly predicting chemically subtle correlations among the internal coordinates themselves, and learning these correlations increases the fidelity of the Cartesian representation. This general problem has been solved with machine learning for proteins, but with appropriately formulated data is extensible to any type of chain biomolecule including RNA, DNA, and lipids. We show that the internal-to-Cartesian process relies on correctly predicting chemically subtle correlations among the internal coordinates themselves, and learning these correlations increases the fidelity of the Cartesian representation. We developed a machine learning algorithm, Int2Cart, to predict bond lengths and bond angles from backbone torsion angles and residue types of a protein, and allows reconstruction of protein structures better than using fixed bond lengths and bond angles, or a static library method that relies on backbone torsion angles and residue types on a single residue. The Int2Cart algorithm has been implemented as an individual python package at https://github.com/THGLab/int2cart.

The precise regulation of gene expression is fundamental to neurodevelopment, plasticity and cognitive function. Although several studies have profiled transcription in the developing human brain, there is a gap in understanding of accompanying translational regulation. In this study, we performed ribosome profiling on 73 human prenatal and adult cortex samples. We characterized the translational regulation of annotated open reading frames (ORFs) and identified thousands of previously unknown translation events, including small ORFs that give rise to human-specific and/or brain-specific microproteins, many of which we independently verified using proteomics. Ribosome profiling in stem-cell-derived human neuronal cultures corroborated these findings and revealed that several neuronal activity-induced non-coding RNAs encode previously undescribed microproteins. Physicochemical analysis of brain microproteins identified a class of proteins that contain arginine-glycine-glycine (RGG) repeats and, thus, may be regulators of RNA metabolism. This resource expands the known translational landscape of the human brain and illuminates previously unknown brain-specific protein products.

O-GlcNAc transferase (OGT) is an essential glycosylating enzyme that catalyzes the addition of N-acetylglucosamine to serine or threonine residues of nuclear and cytoplasmic proteins. The enzyme glycosylates a broad range of peptide sequences and prediction of glycosylation sites has proven challenging. The lack of an experimentally verified set of polypeptide sequences that are not glycosylated by OGT has made prediction of legitimate glycosylation sites more difficult. Here, we tested a number of intrinsically disordered protein regions as substrates of OGT to establish a set of sequences that are not glycosylated by OGT. The negative data set suggests an amino acid compositional bias for OGT targets. This compositional bias was validated by modifying the amino acid composition of the protein Fused in sarcoma (FUS) to enhance glycosylation. NMR experiments demonstrate that the tetratricopeptide repeat (TPR) region of OGT can bind FUS and that glycosylation-promoting mutations enhance binding. These results provide evidence that the TPR recognizes disordered segments of substrates with particular compositions to promote glycosylation, providing insight into the broad specificity of OGT.

The structural characterization of proteins with a disorder requires a computational approach backed by experiments to model their diverse and dynamic structural ensembles. The selection of conformational ensembles consistent with solution experiments of disordered proteins highly depends on the initial pool of conformers, with currently available tools limited by conformational sampling. We have developed a Generative Recurrent Neural Network (GRNN) that uses supervised learning to bias the probability distributions of torsions to take advantage of experimental data types such as nuclear magnetic resonance J-couplings, nuclear Overhauser effects, and paramagnetic resonance enhancements. We show that updating the generative model parameters according to the reward feedback on the basis of the agreement between experimental data and probabilistic selection of torsions from learned distributions provides an alternative to existing approaches that simply reweight conformers of a static structural pool for disordered proteins. Instead, the biased GRNN, DynamICE, learns to physically change the conformations of the underlying pool of the disordered protein to those that better agree with experiments.

The power of structural information for informing biological mechanism is clear for stable folded macromolecules, but similar structure-function insight is more difficult to obtain for highly dynamic systems such as intrinsically disordered proteins (IDPs) which must be described as structural ensembles. Here we present IDPConformerGenerator, a flexible, modular open source software platform for generating large and diverse ensembles of disordered protein states that builds conformers that obey geometric, steric and other physical restraints on the input sequence. IDPConformerGenerator samples backbone phi (φ), psi (ψ), and omega (ω) torsion angles of relevant sequence fragments from loops and secondary structure elements extracted from folded protein structures in the RCSB Protein Data Bank, and builds side chains from robust Monte Carlo algorithms using expanded rotamer libraries. IDPConformerGenerator has many user-defined options enabling variable fractional sampling of secondary structures, supports Bayesian models for assessing agreement of IDP ensembles for consistency with experimental data, and introduces a machine learning approach to transform between internal to Cartesian coordinates with reduced error. IDPConformerGenerator will facilitate the characterization of disordered proteins to ultimately provide structural insights into these states that have key biological functions.

Intrinsically disordered proteins and unfolded proteins have fluctuating conformational ensembles that are fundamental to their biological function and impact protein folding, stability and misfolding. Despite the importance of protein dynamics and conformational sampling, time-dependent data types are not fully exploited when defining and refining disordered protein ensembles. Here we introduce a computational framework using an elastic network model and normal mode displacements to generate a dynamic disordered ensemble consistent with NMR-derived dynamics parameters, including transverse [Formula: see text] relaxation rates and Lipari-Szabo order parameters ([Formula: see text] values). We illustrate our approach using the unfolded state of the drkN SH3 domain to show that the dynamical ensembles give better agreement than a static ensemble for a wide range of experimental validation data including NMR chemical shifts, J-couplings, nuclear Overhauser effects, paramagnetic relaxation enhancements, residual dipolar couplings, hydrodynamic radii, single-molecule fluorescence Förster resonance energy transfer, and small-angle X-ray scattering.

The AlphaFold Protein Structure Database contains predicted structures for millions of proteins. For the majority of human proteins that contain intrinsically disordered regions (IDRs), which do not adopt a stable structure, it is generally assumed these regions have low AlphaFold2 confidence scores that reflect low-confidence structural predictions. Here, we show that AlphaFold2 assigns confident structures to nearly 15% of human IDRs. By comparison to experimental NMR data for a subset of IDRs that are known to conditionally fold (i.e., upon binding or under other specific conditions), we find that AlphaFold2 often predicts the structure of the conditionally folded state. Based on databases of IDRs that are known to conditionally fold, we estimate that AlphaFold2 can identify conditionally folding IDRs at a precision as high as 88% at a 10% false positive rate, which is remarkable considering that conditionally folded IDR structures were minimally represented in its training data. We find that human disease mutations are nearly 5-fold enriched in conditionally folded IDRs over IDRs in general, and that up to 80% of IDRs in prokaryotes are predicted to conditionally fold, compared to less than 20% of eukaryotic IDRs. These results indicate that a large majority of IDRs in the proteomes of human and other eukaryotes function in the absence of conditional folding, but the regions that do acquire folds are more sensitive to mutations. We emphasize that the AlphaFold2 predictions do not reveal functionally relevant structural plasticity within IDRs and cannot offer realistic ensemble representations of conditionally folded IDRs.Significance StatementAlphaFold2 and other machine learning-based methods can accurately predict the structures of most proteins. However, nearly two-thirds of human proteins contain segments that are highly flexible and do not autonomously fold, otherwise known as intrinsically disordered regions (IDRs). In general, IDRs interconvert rapidly between a large number of different conformations, posing a significant problem for protein structure prediction methods that define one or a small number of stable conformations. Here, we found that AlphaFold2 can readily identify structures for a subset of IDRs that fold under certain conditions (conditional folding). We leverage AlphaFold2’s predictions of conditionally folded IDRs to quantify the extent of conditional folding across the tree of life, and to rationalize disease-causing mutations in IDRs.Classifications: Biological Sciences; Biophysics and Computational Biology

Intrinsically disordered proteins (IDPs) play critical roles in regulatory protein interactions, but detailed structural/dynamics characterization of their ensembles remain challenging, both in isolation and they form dynamic ‘fuzzy’ complexes. Such is the case for mRNA cap-dependent translation initiation, which is regulated by the interaction of the predominantly folded eukaryotic initiation factor 4E (eIF4E) with the intrinsically disordered eIF4E binding proteins (4E-BPs) in a phosphorylation-dependent manner. Single-molecule Förster resonance energy transfer showed that the conformational changes of 4E-BP2 induced by binding to eIF4E are non-uniform along the sequence; while a central region containing both motifs that bind to eIF4E expands and becomes stiffer, the C-terminal region is less affected. Fluorescence anisotropy decay revealed a nonuniform segmental flexibility around six different labelling sites along the chain. Dynamic quenching of these fluorescent probes by intrinsic aromatic residues measured via fluorescence correlation spectroscopy report on transient intra- and inter-molecular contacts on ns-μs timescales. Upon hyperphosphorylation, which induces folding of ~40 residues in 4E-BP2, the quenching rates decreased at most labelling sites. The chain dynamics around sites in the C-terminal region far away from the two binding motifs significantly increased upon binding to eIF4E, suggesting that this region is also involved in the highly dynamic 4E-BP2:eIF4E complex. Our time-resolved fluorescence data paint a sequence-level rigidity map of three states of 4E-BP2 differing in phosphorylation or binding status and distinguish regions that form contacts with eIF4E. This study adds complementary structural and dynamics information to recent studies of 4E-BP2, and it constitutes an important step towards a mechanistic understanding of this important IDP via integrative modelling.

Biomolecular condensates concentrate proteins, nucleic acids, and small molecules and play an essential role in many biological processes. Their formation is tuned by a balance between energetically favorable and unfavorable contacts, with charge-charge interactions playing a central role in some systems. The positively charged intrinsically disordered carboxy-terminal region of the RNA-binding protein CAPRIN1 is one such example, phase separating upon addition of negatively charged ATP or high concentrations of sodium chloride (NaCl). Using solution NMR spectroscopy, we measured residue-specific near-surface electrostatic potentials (ϕ