Proteins can aggregate into disordered aggregates or ordered assemblies such as amyloid fibrils. These two distinct phases serve differing roles in function and disease. How protein sequence determines the preferred phase is unknown. Here we establish a statistical mechanical disorder-to-order transition condition for compact polymer aggregates, including proteins. The theory produces a simple universal equation determining the favored phase as a function of temperature, polymer length, and interaction energy variance. We show that the sequence-dependent energy variance is efficiently calculated using atomistic simulations, so that the theory has no adjustable parameters. The equation accurately predicts experimental length-dependent crystallization temperatures of synthetic polymers. The equation also predicts that all protein sequences that aggregate will also favor ordering. Consequently, energy must be expended to maintain the steady-state disordered phase if it is not kinetically metastable on physiological timescales. More broadly, the theory suggests that aggregates of organic polymers will generally tend to order on habitable planets.

Most biomolecules become functional and bioactive by forming protein complexes through interaction with ligands that are diverse in size, shape, and physicochemical properties. In the complex biological milieu, the interaction is ligand-specific, driven by molecular sensing, and involves the recognition of a binding interface localized within a protein structure. Mapping interfaces of protein complexes is a highly sought area of research as it delivers fundamental insights into proteomes and pathology and hence strategies for therapeutics. While X-ray crystallography and electron microscopy remain the gold standard for structural elucidation of protein complexes, their artificial and static analytic nature often produces a non-native interface that otherwise might be negligible or non-existent in a biological environment. Recently, the mass spectrometry-coupled approaches, chemical crosslinking (CLMS) and hydrogen-deuterium exchange (HDMS) have become valuable analytic complements to the traditional techniques. These methods explicitly identify hot residues and motifs embedded in binding interfaces, especially when the interaction is predominantly dynamic, transient, and/or caused by an intrinsically disordered domain. Here, we review the principal role of CLMS and HDMS in protein structural biology with a particular emphasis on the contribution of recent examples to exploring biological interfaces. Additionally, we describe recent studies that utilized these methods to expand our understanding of protein complex formation and the related biological processes, to increase the probability of structure-based drug design.

Understanding how protein-protein binding affinity is determined from molecular interactions at the interface is essential in developing protein therapeutics such as antibodies, but this has not yet been fully achieved. Among the major difficulties are the facts that it is generally difficult to decompose thermodynamic quantities into contributions from individual molecular interactions and that the solvent effect-dehydration penalty-must also be taken into consideration for every contact formation at the binding interface. Here, we present an atomic-level thermodynamics analysis that overcomes these difficulties and illustrate its utility through application to SARS-CoV-2 neutralizing antibodies. Our analysis is based on the direct interaction energy computed from simulated antibody-protein complex structures and on the decomposition of solvation free energy change upon complex formation. We find that the formation of a single contact such as a hydrogen bond at the interface barely contributes to binding free energy due to the dehydration penalty. On the other hand, the simultaneous formation of multiple contacts between two interface residues favorably contributes to binding affinity. This is because the dehydration penalty is significantly alleviated: the total penalty for multiple contacts is smaller than a sum of what would be expected for individual dehydrations of those contacts. Our results thus provide a new perspective for designing protein therapeutics of improved binding affinity.

Many steps in programmed cell death are evolutionarily conserved across different species. The Caenorhabditis elegans proteins CED-9, CED-4 and EGL-1 involved in apoptosis are respectively homologous to anti-apoptotic Bcl-2 proteins, Apaf-1 and the "BH3-only" pro-apototic proteins in mammals. In the linear apoptotic pathway of C. elegans, EGL-1 binding to CED-9 leads to the release of CED-4 from CED-9/CED-4 complex. The molecular events leading to this process are not clearly elucidated. While the structures of CED-9 apo, CED-9/EGL-1 and CED-9/CED-4 complexes are known, the CED-9/CED-4/EGL-1 ternary complex structure is not yet determined. In this work, we modeled this ternary complex and performed molecular dynamics simulations of six different systems involving CED-9. CED-9 displays differential dynamics depending upon whether it is bound to CED-4 and/or EGL-1. CED-4 exists as an asymmetric dimer (CED4a and CED4b) in CED-9/CED-4 complex. CED-4a exhibits higher conformational flexibility when simulated without CED-4b. Principal Component Analysis revealed that the direction of CED-4a's winged-helix domain motion differs in the ternary complex. Upon EGL-1 binding, majority of non-covalent interactions involving CARD domain in the CED-4a-CED-9 interface have weakened and only half of the contacts found in the crystal structure between α/β domain of CED4a and CED-9 are found to be stable. Additional stable contacts in the ternary complex and differential dynamics indicate that winged-helix domain may play a key role in CED-4a's dissociation from CED-9. This study has provided a molecular level understanding of potential intermediate states that are likely to occur when CED-4a is released from CED-9.

Many proteins must interact with molecular chaperones to achieve their native state in the cell. Yet, how chaperone binding-site characteristics affect the folding process is poorly understood. The ubiquitous Hsp70 chaperone system prevents client-protein aggregation by holding unfolded conformations and by unfolding misfolded states. Hsp70 binding sites of client proteins comprise a nonpolar core surrounded by positively charged residues. However, a detailed analysis of Hsp70 binding sites on a proteome-wide scale is still lacking. Further, it is not known whether proteins undergo some degree of folding while chaperone bound. Here, we begin to address the above questions by identifying Hsp70 binding sites in 2258 Escherichia coli (E. coli) proteins. We find that most proteins bear at least one Hsp70 binding site and that the number of Hsp70 binding sites is directly proportional to protein size. Aggregation propensity upon release from the ribosome correlates with number of Hsp70 binding sites only in the case of large proteins. Interestingly, Hsp70 binding sites are more solvent-exposed than other nonpolar sites, in protein native states. Our findings show that the majority of E. coli proteins are systematically enabled to interact with Hsp70 even if this interaction only takes place during a fraction of the protein lifetime. In addition, our data suggest that some conformational sampling may take place within Hsp70-bound states, due to the solvent exposure of some chaperone binding sites in native proteins. In all, we propose that Hsp70-chaperone-binding traits have evolved to favor Hsp70-assisted protein folding devoid of aggregation.

TDP-43 protein is associated with many neurodegenerative diseases and has been shown to adopt various oligomeric and fibrillar states. However, a detailed kinetic understanding of the structural transformation of the native form of the protein to the fibrillar state is missing. In this study, we delineate the temporal sequence of structural events during the amyloid-like assembly of the functional nucleic acid-binding domain of TDP-43. We kinetically mapped the aggregation process using multiple probes such as tryptophan and thioflavin T (ThT) fluorescence, circular dichroism (CD), and dynamic light scattering (DLS) targeting different structural events. Our data reveal that aggregation occurs in four distinct steps-very fast, fast, slow, and very slow. The "very fast" change results in partially unfolded forms that undergo conformational conversion, oligomerization and bind to ThT in the "fast step" to form higher order intermediates (HOI). The temporal sequence of the formation of ThT binding sites and conformational conversion depends upon the protein concentration. The HOI further undergoes structural rearrangement to form protofibrils in the "slow" step, which, consequently, assembles in the "very slow" step to form an amyloid-like assembly. The spectroscopic properties of the amyloid-like assembly across the protein concentration remain similar. Additionally, we observe no lag phase across protein concentration for all the probes studied, suggesting that the aggregation process follows a linear polymerization reaction. Overall, our study demonstrates that the amyloid-like assembly forms in multiple steps, which is also supported by the temperature dependence of the kinetics.

Native topology is known to determine the folding kinetics and the energy landscape of proteins. Furthermore, the circular permutation (CP) of proteins alters the order of the secondary structure connectivity while retaining the three-dimensional structure, making it an elegant and powerful approach to altering native topology. Previous studies elucidated the influence of CP in proteins with different folds such as Greek key β-barrel, β-sandwich, β-α-β, and all α-Greek key. CP mainly affects the protein stability and unfolding kinetics, while folding kinetics remains mostly unaltered. However, the effect of CP on metalloproteins is yet to be elaborately studied. The active site of metalloproteins poses an additional complexity in studying protein folding. Here, we investigate a CP variant (cpN42) of azurin-in both metal-free and metal-bound (holo) forms. As observed earlier in other proteins, apo-forms of wild-type (WT) and cpN42 fold with similar rates. In contrast, zinc-binding accelerates the folding of WT but decelerates the folding of cpN42. On zinc-binding, the spontaneous folding rate of WT increases by >250 times that of cpN42, which is unprecedented and the highest for any CP to date. On the other hand, zinc-binding reduces the spontaneous unfolding rate of cpN42 by ~100 times, making the WT and CP azurins unfold at similar rates. Our study demonstrates metal binding as a novel way to modulate the unfolding and folding rates of CPs compared to their WT counterparts. We hope our study increases the understanding of the effect of CP on the folding mechanism and energy landscape of metalloproteins.

Riboflavin is an essential water-soluble vitamin that needs to be provided through the diet because of the conversion into flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN), important cofactors in hundreds of flavoenzymes. The adsorption and distribution of riboflavin is mediated by transmembrane transporters of the SLC52 family, namely RFVT1-3, whose mutations are mainly associated with two diseases, MADD and the Brown-Vialetto-Van Laere syndrome. Interest in RFVTs as pharmacological targets has increased in the last few years due to their overexpression in several cancer cells, which can be exploited both by blocking the uptake of riboflavin into the cancerous cells, and by performing cancer targeted delivery of drugs with a high affinity for RFVTs. In this work, we propose three-dimensional structural models for all three human riboflavin transporters obtained by state-of-the-art artificial intelligence-based methods, which were then further refined with molecular dynamics simulations. Furthermore, two of the most notable mutations concerning RFVT2 and RFVT3 (W31S and N21S, respectively) were investigated studying the interactions between the wild-type and mutated transporters with riboflavin.

The glutathione degrading enzyme ChaC1 is highly upregulated in several cancers and viral infections making it a potential pharmacological target for cancer therapy. As an enzyme, however, ChaC1 has a relatively high Km (~2 mM) towards its natural substrate, and therefore finding its inhibitors becomes very difficult. Given this limitation, a careful mapping of the active site has become necessary. In the current study, the enzyme-substrate complex was generated by docking glutathione with the modeled hChaC1 structure. Using a combination of in silico and wet lab approaches, the active site residues forming direct interactions with the substrate glutathione were identified and validated. Furthermore, the role of residues exclusively conserved in the ChaC family and forming the surface of the active site were also explored for their putative role in active site stabilization. Mutants of these residues have been analysed for their structural stability and interaction with the substrate through MD simulations and MMGBSA binding energy calculations. These findings were experimentally validated by assessment of their function through in vivo assays in yeast. The experimental evidences along with the molecular modeling suggest that residues 38'YGSL'41, D68, R72, E115, and Y143 are responsible for high affinity binding of hChaC1 with the substrate/inhibitor, whereas the residues exclusive to the ChaC family are required for the structural stability of the enzyme and its active site. Such a characterization of essential active site and conserved residues is significant as a key step toward rational design of novel inhibitors of the ChaC1 enzyme.