Analysis of Large-Scale Mutagenesis Data To Assess the Impact of Single Amino Acid Substitutions

Rational design of monomeric IL37 variants guided by stability and dynamical analyses of IL37 dimers

IL37 plays important roles in the regulation of innate immunity and its oligomeric status is critical to these roles. In its monomeric state, IL37 can effectively inhibit the inflammatory response of IL18 by binding to IL18Rα, a capacity lost in its dimeric form, underlining the pivotal role of the oligomeric status of IL37 in its anti-inflammatory action. Until now, two IL37 dimer structures have been deposited in PDB, reflecting a substantial difference in their dimer interfaces. Given this discrepancy, we analyzed the PDB structures of the IL37 dimer (PDB IDs: 6ncu, 5hn1) along with a AF2-multimer prediction by molecular dynamics (MD) simulations. Results showed that the 5hn1 and AF2-predicted dimers have the same interface and stably maintained their conformations throughout simulations, while the recent IL37 dimer (PDB ID: 6ncu) with a different interface did not, proposing a possible issue with the recent IL37 dimer structure (6ncu). Next, focusing on the stable dimer structures, we have identified five critical positions of V71/Y85/I86/E89/S114, three new positions compared to the literature, that would reduce dimer stability without affecting the monomer structure. Two quintuple mutants were tested by MD simulations and showed partial or complete dissociation of the dimer. Overall, the insights gained from this study reinforce the validity of the 5hn1 and AF2 multimer structures, while also advancing our understanding of the IL37 dimer interface through the generation of monomer-locked IL37 variants.

RudS: bacterial desulfidase responsible for tRNA 4-thiouridine de-modification

In this study, we present an extensive analysis of a widespread group of bacterial tRNA de-modifying enzymes, dubbed RudS, which consist of a TudS desulfidase fused to a Domain of Unknown Function 1722 (DUF1722). RudS enzymes exhibit specific de-modification activity towards the 4-thiouridine modification (s4U) in tRNA molecules, as indicated by our experimental findings. The heterologous overexpression of RudS genes in Escherichia coli significantly reduces the tRNA 4-thiouridine content and diminishes UVA-induced growth delay, indicating the enzyme's role in regulating photosensitive tRNA s4U modification. Through a combination of protein modeling, docking studies, and molecular dynamics simulations, we have identified amino acid residues involved in catalysis and tRNA binding. Experimental validation through targeted mutagenesis confirms the TudS domain as the catalytic core of RudS, with the DUF1722 domain facilitating tRNA binding in the anticodon region. Our results suggest that RudS tRNA modification eraser proteins may play a role in regulating tRNA during prokaryotic stress responses.

An invariant C-terminal tryptophan in McdB mediates its interaction and positioning function with carboxysomes

Bacterial microcompartments (BMCs) are widespread, protein-based organelles that regulate metabolism. The model for studying BMCs is the carboxysome, which facilitates carbon fixation in several autotrophic bacteria. Carboxysomes can be distinguished as type α or β, which are structurally and phyletically distinct. We recently characterized the maintenance of carboxysome distribution (Mcd) systems responsible for spatially regulating α- and β-carboxysomes, consisting of the proteins McdA and McdB. McdA is an ATPase that drives carboxysome positioning, and McdB is the adaptor protein that directly interacts with carboxysomes to provide cargo specificity. The molecular features of McdB proteins that specify their interactions with carboxysomes, and whether these are similar between α- and β-carboxysomes, remain unknown. Here, we identify C-terminal motifs containing an invariant tryptophan necessary for α- and β-McdBs to associate with α- and β-carboxysomes, respectively. Substituting this tryptophan with other aromatic residues reveals corresponding gradients in the efficiency of carboxysome colocalization and positioning by McdB in vivo. Intriguingly, these gradients also correlate with the ability of McdB to form condensates in vitro. The results reveal a shared mechanism underlying McdB adaptor protein binding to carboxysomes, and potentially other BMCs. Our findings also implicate condensate formation as playing a key role in this association.

Flattening the curve-How to get better results with small deep-mutational-scanning datasets

Proteins are used in various biotechnological applications, often requiring the optimization of protein properties by introducing specific amino-acid exchanges. Deep mutational scanning (DMS) is an effective high-throughput method for evaluating the effects of these exchanges on protein function. DMS data can then inform the training of a neural network to predict the impact of mutations. Most approaches use some representation of the protein sequence for training and prediction. As proteins are characterized by complex structures and intricate residue interaction networks, directly providing structural information as input reduces the need to learn these features from the data. We introduce a method for encoding protein structures as stacked 2D contact maps, which capture residue interactions, their evolutionary conservation, and mutation-induced interaction changes. Furthermore, we explored techniques to augment neural network training performance on smaller DMS datasets. To validate our approach, we trained three neural network architectures originally used for image analysis on three DMS datasets, and we compared their performances with networks trained solely on protein sequences. The results confirm the effectiveness of the protein structure encoding in machine learning efforts on DMS data. Using structural representations as direct input to the networks, along with data augmentation and pretraining, significantly reduced demands on training data size and improved prediction performance, especially on smaller datasets, while performance on large datasets was on par with state-of-the-art sequence convolutional neural networks. The methods presented here have the potential to provide the same workflow as DMS without the experimental and financial burden of testing thousands of mutants. Additionally, we present an open-source, user-friendly software tool to make these data analysis techniques accessible, particularly to biotechnology and protein engineering researchers who wish to apply them to their mutagenesis data.

Understanding activity-stability tradeoffs in biocatalysts by enzyme proximity sequencing

Understanding the complex relationships between enzyme sequence, folding stability and catalytic activity is crucial for applications in industry and biomedicine. However, current enzyme assay technologies are limited by an inability to simultaneously resolve both stability and activity phenotypes and to couple these to gene sequences at large scale. Here we present the development of enzyme proximity sequencing, a deep mutational scanning method that leverages peroxidase-mediated radical labeling with single cell fidelity to dissect the effects of thousands of mutations on stability and catalytic activity of oxidoreductase enzymes in a single experiment. We use enzyme proximity sequencing to analyze how 6399 missense mutations influence folding stability and catalytic activity in a D-amino acid oxidase from Rhodotorula gracilis. The resulting datasets demonstrate activity-based constraints that limit folding stability during natural evolution, and identify hotspots distant from the active site as candidates for mutations that improve catalytic activity without sacrificing stability. Enzyme proximity sequencing can be extended to other enzyme classes and provides valuable insights into biophysical principles governing enzyme structure and function.

Scaling up Functional Analyses of the G Protein-Coupled Receptor Rhodopsin

Eukaryotic cells use G protein-coupled receptors (GPCRs) to convert external stimuli into internal signals to elicit cellular responses. However, how mutations in GPCR-coding genes affect GPCR activation and downstream signaling pathways remain poorly understood. Approaches such as deep mutational scanning show promise in investigations of GPCRs, but a high-throughput method to measure rhodopsin activation has yet to be achieved. Here, we scale up a fluorescent reporter assay in budding yeast that we engineered to study rhodopsin's light-activated signal transduction. Using this approach, we measured the mutational effects of over 1200 individual human rhodopsin mutants, generated by low-frequency random mutagenesis of the GPCR rhodopsin (RHO) gene. Analysis of the data in the context of rhodopsin's three-dimensional structure reveals that transmembrane helices are generally less tolerant to mutations compared to flanking helices that face the lipid bilayer, which suggest that mutational tolerance is contingent on both the local environment surrounding specific residues and the specific position of these residues in the protein structure. Comparison of functional scores from our screen to clinically identified rhodopsin disease variants found many pathogenic mutants to be loss of function. Lastly, functional scores from our assay were consistent with a complex counterion mechanism involved in ligand-binding and rhodopsin activation. Our results demonstrate that deep mutational scanning is possible for rhodopsin activation and can be an effective method for revealing properties of mutational tolerance that may be generalizable to other transmembrane proteins.

Dominant epitopes of cross-reactive anti-domain III human antibody response change from early to late convalescence of infection with dengue virus

Cross-neutralizing activity of human antibody response against Dengue virus complex (DENV) changes importantly over time. Domain III (DIII) of the envelope protein of DENV elicits a potently neutralizing and mostly type-specific IgG response. We used sera from 24 individuals from early- or late convalescence of DENV1 infection to investigate the evolution of anti-DIII human IgG with the time lapse since the infection. We evaluated the correlation between the serotype-specific reactivity against recombinant DIII proteins and the neutralization capacity against the four serotypes, and examined its behavior with the time of convalescence. Also, we use a library of 71 alanine mutants of surface-exposed amino acid residues to investigate the dominant epitopes. In early convalescence anti-DIII titers and potency of virus neutralization were positively associated with correlation coefficients from 0.82 to 1.0 for the four serotypes. For late convalescence, a positive correlation (r = 0.69) was found only for DENV1. The dominant epitope of the type-specific response is centered in the FG-loop (G383, E384, and K385) and includes most of the lateral ridge. The dominant epitope of the anti-DIII cross-reactive IgG in secondary infections shifts from the A-strand during early convalescence to a site centered in residues E314-H317 of the AB-loop and I352-E368 of the DI/DIII interface, in late convalescence. An immunoassay based on the detection of IgG anti-DIII response can be implemented for detection of infecting serotype in diagnosis of DENV infection, either primary or secondary. Human dominant epitopes of the cross-reactive circulating antibodies change with time of convalescence.

Affinity maturation of antibody fragments: A review encompassing the development from random approaches to computational rational optimization

Routinely screened antibody fragments usually require further in vitro maturation to achieve the desired biophysical properties. Blind in vitro strategies can produce improved ligands by introducing random mutations into the original sequences and selecting the resulting clones under more and more stringent conditions. Rational approaches exploit an alternative perspective that aims first at identifying the specific residues potentially involved in the control of biophysical mechanisms, such as affinity or stability, and then to evaluate what mutations could improve those characteristics. The understanding of the antigen-antibody interactions is instrumental to develop this process the reliability of which, consequently, strongly depends on the quality and completeness of the structural information. Recently, methods based on deep learning approaches critically improved the speed and accuracy of model building and are promising tools for accelerating the docking step. Here, we review the features of the available bioinformatic instruments and analyze the reports illustrating the result obtained with their application to optimize antibody fragments, and nanobodies in particular. Finally, the emerging trends and open questions are summarized.

Discovering functionally important sites in proteins

Proteins play important roles in biology, biotechnology and pharmacology, and missense variants are a common cause of disease. Discovering functionally important sites in proteins is a central but difficult problem because of the lack of large, systematic data sets. Sequence conservation can highlight residues that are functionally important but is often convoluted with a signal for preserving structural stability. We here present a machine learning method to predict functional sites by combining statistical models for protein sequences with biophysical models of stability. We train the model using multiplexed experimental data on variant effects and validate it broadly. We show how the model can be used to discover active sites, as well as regulatory and binding sites. We illustrate the utility of the model by prospective prediction and subsequent experimental validation on the functional consequences of missense variants in HPRT1 which may cause Lesch-Nyhan syndrome, and pinpoint the molecular mechanisms by which they cause disease.

Germline-encoded amino acid-binding motifs drive immunodominant public antibody responses

Despite the vast diversity of the antibody repertoire, infected individuals often mount antibody responses to precisely the same epitopes within antigens. The immunological mechanisms underpinning this phenomenon remain unknown. By mapping 376 immunodominant "public epitopes" at high resolution and characterizing several of their cognate antibodies, we concluded that germline-encoded sequences in antibodies drive recurrent recognition. Systematic analysis of antibody-antigen structures uncovered 18 human and 21 partially overlapping mouse germline-encoded amino acid-binding (GRAB) motifs within heavy and light V gene segments that in case studies proved critical for public epitope recognition. GRAB motifs represent a fundamental component of the immune system's architecture that promotes recognition of pathogens and leads to species-specific public antibody responses that can exert selective pressure on pathogens.

Explaining Black Box Drug Target Prediction Through Model Agnostic Counterfactual Samples

Many high-performance DTA deep learning models have been proposed, but they are mostly black-box and thus lack human interpretability. Explainable AI (XAI) can make DTA models more trustworthy, and allows to distill biological knowledge from the models. Counterfactual explanation is one popular approach to explaining the behaviour of a deep neural network, which works by systematically answering the question "How would the model output change if the inputs were changed in this way?". We propose a multi-agent reinforcement learning framework, Multi-Agent Counterfactual Drug-target binding Affinity (MACDA), to generate counterfactual explanations for the drug-protein complex. Our proposed framework provides human-interpretable counterfactual instances while optimizing both the input drug and target for counterfactual generation at the same time. We benchmark the proposed MACDA framework using the Davis and PDBBind dataset and find that our framework produces more parsimonious explanations with no loss in explanation validity, as measured by encoding similarity. We then present a case study involving ABL1 and Nilotinib to demonstrate how MACDA can explain the behaviour of a DTA model in the underlying substructure interaction between inputs in its prediction, revealing mechanisms that align with prior domain knowledge.

Evolution-inspired engineering of anthracycline methyltransferases

Streptomyces soil bacteria produce hundreds of anthracycline anticancer agents with a relatively conserved set of genes. This diversity depends on the rapid evolution of biosynthetic enzymes to acquire novel functionalities. Previous work has identified S-adenosyl-l-methionine-dependent methyltransferase-like proteins that catalyze 4-O-methylation, 10-decarboxylation, or 10-hydroxylation, with additional differences in substrate specificities. Here we focused on four protein regions to generate chimeric enzymes using sequences from four distinct subfamilies to elucidate their influence in catalysis. Combined with structural studies we managed to depict factors that influence gain-of-hydroxylation, loss-of-methylation, and substrate selection. The engineering expanded the catalytic repertoire to include novel 9,10-elimination activity, and 4-O-methylation and 10-decarboxylation of unnatural substrates. The work provides an instructive account on how the rise of diversity of microbial natural products may occur through subtle changes in biosynthetic enzymes.

Analysis of BlaEC family class C beta-lactamase

Recent years have witnessed an increased prevalence of intrinsic and acquired beta-lactamase-producing bacteria, severely limiting human and veterinary medicine therapeutic options. The present study aimed to design specific oligonucleotides for rapid PCR detection of the cephalosporinase-encoding gene blaEC (BlaEC family class C beta-lactamase). A total of three primers were designed to detect 2281 variants of the blaEC gene and two sets of primer pairs were also tested against DNA from 11 strains. The study indicates that the proposed primers should be able to detect 100% of all described blaEC genes in different bacterial strains and monitor their spread. After comparing the amino acid sequences, a phylogenetic tree was created based on the presence of conserved amino acids and homologous motifs. More than 24 760 mutations in BlaEC enzymes have been identified. The mutations involving 371 amino acid positions and these hotspots can change the structure and activity of the monitored enzymes. We predicted several BlaEC enzymes with a broadened substrate activity against higher-generation cephalosporins.

Joint estimation and imputation of variant functional effects using high throughput assay data

Deep mutational scanning assays enable the functional assessment of variants in high throughput. Phenotypic measurements from these assays are broadly concordant with clinical outcomes but are prone to noise at the individual variant level. We develop a framework to exploit related measurements within and across experimental assays to jointly estimate variant impact. Drawing from a large corpus of deep mutational scanning data, we collectively estimate the mean functional effect per AA residue position within each gene, normalize observed functional effects by substitution type, and make estimates for individual allelic variants with a pipeline called FUSE (Functional Substitution Estimation). FUSE improves the correlation of functional screening datasets covering the same variants, better separates estimated functional impacts for known pathogenic and benign variants (ClinVar BRCA1, p=2.24×10-51), and increases the number of variants for which predictions can be made (2,741 to 10,347) by inferring additional variant effects for substitutions not experimentally screened. For UK Biobank patients who carry a rare variant in TP53, FUSE significantly improves the separation of patients who develop cancer syndromes from those without cancer (p=1.77×10-6). These approaches promise to improve estimates of variant impact and broaden the utility of screening data generated from functional assays.

Integrating deep mutational scanning and low-throughput mutagenesis data to predict the impact of amino acid variants

BACKGROUND

Evaluating the impact of amino acid variants has been a critical challenge for studying protein function and interpreting genomic data. High-throughput experimental methods like deep mutational scanning (DMS) can measure the effect of large numbers of variants in a target protein, but because DMS studies have not been performed on all proteins, researchers also model DMS data computationally to estimate variant impacts by predictors.

RESULTS

In this study, we extended a linear regression-based predictor to explore whether incorporating data from alanine scanning (AS), a widely used low-throughput mutagenesis method, would improve prediction results. To evaluate our model, we collected 146 AS datasets, mapping to 54 DMS datasets across 22 distinct proteins.

CONCLUSIONS

We show that improved model performance depends on the compatibility of the DMS and AS assays, and the scale of improvement is closely related to the correlation between DMS and AS results.

Intragenic compensation through the lens of deep mutational scanning

A significant fraction of mutations in proteins are deleterious and result in adverse consequences for protein function, stability, or interaction with other molecules. Intragenic compensation is a specific case of positive epistasis when a neutral missense mutation cancels effect of a deleterious mutation in the same protein. Permissive compensatory mutations facilitate protein evolution, since without them all sequences would be extremely conserved. Understanding compensatory mechanisms is an important scientific challenge at the intersection of protein biophysics and evolution. In human genetics, intragenic compensatory interactions are important since they may result in variable penetrance of pathogenic mutations or fixation of pathogenic human alleles in orthologous proteins from related species. The latter phenomenon complicates computational and clinical inference of an allele's pathogenicity. Deep mutational scanning is a relatively new technique that enables experimental studies of functional effects of thousands of mutations in proteins. We review the important aspects of the field and discuss existing limitations of current datasets. We reviewed ten published DMS datasets with quantified functional effects of single and double mutations and described rates and patterns of intragenic compensation in eight of them.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12551-022-01005-w.

Entropies Derived from the Packing Geometries within a Single Protein Structure

A fast, simple, yet robust method to calculate protein entropy from a single protein structure is presented here. The focus is on the atomic packing details, which are calculated by combining Voronoi diagrams and Delaunay tessellations. Even though the method is simple, the entropies computed exhibit an extremely high correlation with the entropies previously derived by other methods based on quasi-harmonic motions, quantum mechanics, and molecular dynamics simulations. These packing-based entropies account directly for the local freedom and provide entropy for any individual protein structure that could be used to compute free energies directly during simulations for the generation of more reliable trajectories and also for better evaluations of modeled protein structures. Physico-chemical properties of amino acids are compared with these packing entropies to uncover the relationships with the entropies of different residue types. A public packing entropy web server is provided at packing-entropy.bb.iastate.edu, and the application programing interface is available within the PACKMAN (https://github.com/Pranavkhade/PACKMAN) package.

In Silico Identification of Potential Sites for a Plastic-Degrading Enzyme by a Reverse Screening through the Protein Sequence Space and Molecular Dynamics Simulations

The accumulation of polyethylene terephthalate (PET) seriously harms the environment because of its high resistance to degradation. The recent discovery of the bacteria-secreted biodegradation enzyme, PETase, sheds light on PET recycling; however, the degradation efficiency is far from practical use. Here, in silico alanine scanning mutagenesis (ASM) and site-saturation mutagenesis (SSM) were employed to construct the protein sequence space from binding energy of the PETase–PET interaction to identify the number and position of mutation sites and their appropriate side-chain properties that could improve the PETase–PET interaction. The binding mechanisms of the potential PETase variant were investigated through atomistic molecular dynamics simulations. The results show that up to two mutation sites of PETase are preferable for use in protein engineering to enhance the PETase activity, and the proper side chain property depends on the mutation sites. The predicted variants agree well with prior experimental studies. Particularly, the PETase variants with S238C or Q119F could be a potential candidate for improving PETase. Our combination of in silico ASM and SSM could serve as an alternative protocol for protein engineering because of its simplicity and reliability. In addition, our findings could lead to PETase improvement, offering an important contribution towards a sustainable future.

Optimization of the antimicrobial peptide Bac7 by deep mutational scanning

BACKGROUND

Intracellularly active antimicrobial peptides are promising candidates for the development of antibiotics for human applications. However, drug development using peptides is challenging as, owing to their large size, an enormous sequence space is spanned. We built a high-throughput platform that incorporates rapid investigation of the sequence-activity relationship of peptides and enables rational optimization of their antimicrobial activity. The platform is based on deep mutational scanning of DNA-encoded peptides and employs highly parallelized bacterial self-screening coupled to next-generation sequencing as a readout for their antimicrobial activity. As a target, we used Bac71-23, a 23 amino acid residues long variant of bactenecin-7, a potent translational inhibitor and one of the best researched proline-rich antimicrobial peptides.

RESULTS

Using the platform, we simultaneously determined the antimicrobial activity of >600,000 Bac71-23 variants and explored their sequence-activity relationship. This dataset guided the design of a focused library of ~160,000 variants and the identification of a lead candidate Bac7PS. Bac7PS showed high activity against multidrug-resistant clinical isolates of E. coli, and its activity was less dependent on SbmA, a transporter commonly used by proline-rich antimicrobial peptides to reach the cytosol and then inhibit translation. Furthermore, Bac7PS displayed strong ribosomal inhibition and low toxicity against eukaryotic cells and demonstrated good efficacy in a murine septicemia model induced by E. coli.

CONCLUSION

We demonstrated that the presented platform can be used to establish the sequence-activity relationship of antimicrobial peptides, and showed its usefulness for hit-to-lead identification and optimization of antimicrobial drug candidates.

Turning Failures into Applications: The Problem of Protein ΔΔG Prediction

After nearly two decades of research in the field of computational methods based on machine learning and knowledge-based potentials for ΔG and ΔΔG prediction upon variations, we now realize that all the approaches are poorly performing when tested on specific cases and that there is large space for improvement. Why this is so? Is it wrong the underlying assumption that experimental protein thermodynamics in solution reflects the thermodynamics of a single protein? Both machine learning and knowledge-based computational methods are rigorous and we know the solid theory behind. We are now in a critical situation, which suggests that predictions of protein instability upon variation should be considered with care. In the following, we will show how to cope with the problem of understanding which protein positions may be of interest for biotechnological and biomedical purposes. By applying a consensus procedure, we indicate possible strategies for the result interpretation.

Functional and structural segregation of overlapping helices in HIV-1

Overlapping coding regions balance selective forces between multiple genes. One possible division of nucleotide sequence is that the predominant selective force on a particular nucleotide can be attributed to just one gene. While this arrangement has been observed in regions in which one gene is structured and the other is disordered, we sought to explore how overlapping genes balance constraints when both protein products are structured over the same sequence. We use a combination of sequence analysis, functional assays, and selection experiments to examine an overlapped region in HIV-1 that encodes helical regions in both Env and Rev. We find that functional segregation occurs even in this overlap, with each protein spacing its functional residues in a manner that allows a mutable non-binding face of one helix to encode important functional residues on a charged face in the other helix. Additionally, our experiments reveal novel and critical functional residues in Env and have implications for the therapeutic targeting of HIV-1.

Genome interpretation using in silico predictors of variant impact

Estimating the effects of variants found in disease driver genes opens the door to personalized therapeutic opportunities. Clinical associations and laboratory experiments can only characterize a tiny fraction of all the available variants, leaving the majority as variants of unknown significance (VUS). In silico methods bridge this gap by providing instant estimates on a large scale, most often based on the numerous genetic differences between species. Despite concerns that these methods may lack reliability in individual subjects, their numerous practical applications over cohorts suggest they are already helpful and have a role to play in genome interpretation when used at the proper scale and context. In this review, we aim to gain insights into the training and validation of these variant effect predicting methods and illustrate representative types of experimental and clinical applications. Objective performance assessments using various datasets that are not yet published indicate the strengths and limitations of each method. These show that cautious use of in silico variant impact predictors is essential for addressing genome interpretation challenges.

Toward Compilation of Balanced Protein Stability Data Sets: Flattening the ΔΔG Curve through Systematic Enrichment

Often studies analyzing stability data sets and/or predictors ignore neutral mutations and use a binary classification scheme labeling only destabilizing and stabilizing mutations. Recognizing that highly concentrated neutral mutations interfere with data set quality, we have explored three protein stability data sets: S2648, PON-tstab, and the symmetric S^sym that differ in size and quality. A characteristic leptokurtic shape in the ΔΔG distributions of all three data sets including the curated and symmetric ones was reported due to concentrated neutral mutations. To further investigate the impact of neutral mutations on ΔΔG predictions, we have comprehensively assessed the performance of 11 predictors on the PON-tstab data set. Correlation and error analyses showed that all of the predictors performed the best on the neutral mutations, while their performance became gradually worse as the ΔΔG of the mutations departed further from the neutral zone regardless of the direction, implying a bias toward dense mutations. To this end, after unraveling the role of concentrated neutral mutations in biases of stability data sets, we described a systematic enrichment approach to balance the ΔΔG distributions. Before enrichment, mutations were clustered based on their biochemical and/or structural features, and then three mutations were selected from every 2 kcal/mol of each cluster. Upon implementation of this approach by distinct clustering schemes, we generated five subsets varying in size and ΔΔG distributions. All subsets showed improved ΔΔG and frequency distributions. We ultimately reported that the errors toward enriched subsets were higher than those toward the parent data sets, confirming the enrichment of difficult-to-predict mutations in the subsets. In summary, we elaborated the prediction bias toward a concentrated neutral zone and also implemented a rational strategy to tackle this and other forms of biases. Ultimately, this study equipping us with an extended view of shortcomings of stability data sets is a step taken toward development of an unbiased predictor.

Droplet-based screening of phosphate transfer catalysis reveals how epistasis shapes MAP kinase interactions with substrates

The combination of ultrahigh-throughput screening and sequencing informs on function and intragenic epistasis within combinatorial protein mutant libraries. Establishing a droplet-based, in vitro compartmentalised approach for robust expression and screening of protein kinase cascades (>10⁷ variants/day) allowed us to dissect the intrinsic molecular features of the MKK-ERK signalling pathway, without interference from endogenous cellular components. In a six-residue combinatorial library of the MKK1 docking domain, we identified 29,563 sequence permutations that allow MKK1 to efficiently phosphorylate and activate its downstream target kinase ERK2. A flexibly placed hydrophobic sequence motif emerges which is defined by higher order epistatic interactions between six residues, suggesting synergy that enables high connectivity in the sequence landscape. Through positive epistasis, MKK1 maintains function during mutagenesis, establishing the importance of co-dependent residues in mammalian protein kinase-substrate interactions, and creating a scenario for the evolution of diverse human signalling networks.

Here, the authors use a droplet-based screen for phosphate transfer catalysis, testing variants of the human protein kinase MKK1 for its ability to activate its downstream target ERK2. Data reveal a flexible motif in the MKK1 docking domain that promotes efficient activation of ERK2, and suggest epistasis between the residues within that sequence.

Directed mutational scanning reveals a balance between acidic and hydrophobic residues in strong human activation domains

Acidic activation domains are intrinsically disordered regions of the transcription factors that bind coactivators. The intrinsic disorder and low evolutionary conservation of activation domains have made it difficult to identify the sequence features that control activity. To address this problem, we designed thousands of variants in seven acidic activation domains and measured their activities with a high-throughput assay in human cell culture. We found that strong activation domain activity requires a balance between the number of acidic residues and aromatic and leucine residues. These findings motivated a predictor of acidic activation domains that scans the human proteome for clusters of aromatic and leucine residues embedded in regions of high acidity. This predictor identifies known activation domains and accurately predicts previously unidentified ones. Our results support a flexible acidic exposure model of activation domains in which the acidic residues solubilize hydrophobic motifs so that they can interact with coactivators. A record of this paper’s transparent peer review process is included in the supplemental information.

Transcriptional activation domains are poorly conserved, intrinsically disordered regions of the transcription factors that remain difficult to predict from protein sequences. A high-throughput method reveals how strong activation domains require a balance between acidic and hydrophobic residues. This balance powers an accurate predictor of activation domains on human transcription factors.

Escherichia coli alanyl-tRNA synthetase maintains proofreading activity and translational accuracy under oxidative stress

Aminoacyl-tRNA synthetases (aaRSs) are enzymes that synthesize aminoacyl-tRNAs to facilitate translation of the genetic code. Quality control by aaRS proofreading and other mechanisms maintains translational accuracy, which promotes cellular viability. Systematic disruption of proofreading, as recently demonstrated for alanyl-tRNA synthetase (AlaRS), leads to dysregulation of the proteome and reduced viability. Recent studies showed that environmental challenges such as exposure to reactive oxygen species can also alter aaRS synthetic and proofreading functions, prompting us to investigate if oxidation might positively or negatively affect AlaRS activity. We found that while oxidation leads to modification of several residues in Escherichia coli AlaRS, unlike in other aaRSs, this does not affect proofreading activity against the noncognate substrates serine and glycine and only results in a 1.6-fold decrease in efficiency of cognate Ala-tRNA^Ala formation. Mass spectrometry analysis of oxidized AlaRS revealed that the critical proofreading residue in the editing site, Cys666, and three methionine residues (M217 in the active site, M658 in the editing site, and M785 in the C-Ala domain) were modified to cysteine sulfenic acid and methionine sulfoxide, respectively. Alanine scanning mutagenesis showed that none of the identified residues were solely responsible for the change in cognate tRNA^Ala aminoacylation observed under oxidative stress, suggesting that these residues may act as reactive oxygen species “sinks” to protect catalytically critical sites from oxidative damage. Combined, our results indicate that E. coli AlaRS proofreading is resistant to oxidative damage, providing an important mechanism of stress resistance that helps to maintain proteome integrity and cellular viability.

Predicting and interpreting large-scale mutagenesis data using analyses of protein stability and conservation

Understanding and predicting the functional consequences of single amino acid changes is central in many areas of protein science. Here, we collect and analyze experimental measurements of effects of >150,000 variants in 29 proteins. We use biophysical calculations to predict changes in stability for each variant and assess them in light of sequence conservation. We find that the sequence analyses give more accurate prediction of variant effects than predictions of stability and that about half of the variants that show loss of function do so due to stability effects. We construct a machine learning model to predict variant effects from protein structure and sequence alignments and show how the two sources of information support one another and enable mechanistic interpretations. Together, our results show how one can leverage large-scale experimental assessments of variant effects to gain deeper and general insights into the mechanisms that cause loss of function.

Pathogenic variants in MRPL44 cause infantile cardiomyopathy due to a mitochondrial translation defect

Cardiac dysfunction is a common phenotypic manifestation of primary mitochondrial disease with multiple nuclear and mitochondrial DNA pathogenic variants as a cause, including disorders of mitochondrial translation. To date, five patients have been described with pathogenic variants in MRPL44, encoding the ml44 protein which is part of the large subunit of the mitochondrial ribosome (mitoribosome). Three presented as infants with hypertrophic cardiomyopathy, mild lactic acidosis, and easy fatigue and muscle weakness, whereas two presented in adolescence with myopathy and neurological symptoms. We describe two infants who presented with cardiomyopathy from the neonatal period, failure to thrive, hypoglycemia and in one infant lactic acidosis. A decompensation of the cardiac function in the first year resulted in demise. Exome sequencing identified compound heterozygous variants in the MRPL44 gene including the known pathogenic variant c.467 T > G and two novel pathogenic variants. We document a combined respiratory chain enzyme deficiency with emphasis on complex I and IV, affecting heart muscle tissue more than skeletal muscle or fibroblasts. We show this to be caused by reduced mitochondrial DNA encoded protein synthesis affecting all subunits, and resulting in dysfunction of complex I and IV assembly. The degree of oxidative phosphorylation dysfunction correlated with the impairment of mitochondrial protein synthesis due to different pathogenic variants. These functional studies allow for improved understanding of the pathogenesis of MRPL44-associated mitochondrial disorder.

Gene Erosion Can Lead to Gain-of-Function Alleles That Contribute to Bacterial Fitness

Despite our extensive knowledge of the genetic regulation of heat shock proteins (HSPs), the evolutionary routes that allow bacteria to adaptively tune their HSP levels and corresponding proteostatic robustness have been explored less. In this report, directed evolution experiments using the Escherichia coli model system unexpectedly revealed that seemingly random single mutations in its tnaA gene can confer significant heat resistance. Closer examination, however, indicated that these mutations create folding-deficient and aggregation-prone TnaA variants that in turn can endogenously and preemptively trigger HSP expression to cause heat resistance. These findings, importantly, demonstrate that even erosive mutations with disruptive effects on protein structure and functionality can still yield true gain-of-function alleles with a selective advantage in adaptive evolution.

Dissection of FixK2 protein-DNA interaction unveils new insights into Bradyrhizobium diazoefficiens lifestyles control

The FixK₂ protein plays a pivotal role in a complex regulatory network, which controls genes for microoxic, denitrifying, and symbiotic nitrogen-fixing lifestyles in Bradyrhizobium diazoefficiens. Among the microoxic-responsive FixK₂ -activated genes are the fixNOQP operon, indispensable for respiration in symbiosis, and the nnrR regulatory gene needed for the nitric-oxide dependent induction of the norCBQD genes encoding the denitrifying nitric oxide reductase. FixK₂ is a CRP/FNR-type transcription factor, which recognizes a 14 bp-palindrome (FixK₂ box) at the regulated promoters through three residues (L195, E196, and R200) within a C-terminal helix-turn-helix motif. Here, we mapped the determinants for discriminatory FixK₂ -mediated regulation. While R200 was essential for DNA binding and activity of FixK₂ , L195 was involved in protein-DNA complex stability. Mutation at positions 1, 3, or 11 in the genuine FixK₂ box at the fixNOQP promoter impaired transcription activation by FixK₂ , which was residual when a second mutation affecting the box palindromy was introduced. The substitution of nucleotide 11 within the NnrR box at the norCBQD promoter allowed FixK₂ -mediated activation in response to microoxia. Thus, position 11 within the FixK₂ /NnrR boxes constitutes a key element that changes FixK₂ targets specificity, and consequently, it might modulate B. diazoefficiens lifestyle as nitrogen fixer or as denitrifier.

Exploring amino acid functions in a deep mutational landscape

Amino acids fulfil a diverse range of roles in proteins, each utilising its chemical properties in different ways in different contexts to create required functions. For example, cysteines form disulphide or hydrogen bonds in different circumstances and charged amino acids do not always make use of their charge. The repertoire of amino acid functions and the frequency at which they occur in proteins remains understudied. Measuring large numbers of mutational consequences, which can elucidate the role an amino acid plays, was prohibitively time-consuming until recent developments in deep mutational scanning. In this study, we gathered data from 28 deep mutational scanning studies, covering 6,291 positions in 30 proteins, and used the consequences of mutation at each position to define a mutational landscape. We demonstrated rich relationships between this landscape and biophysical or evolutionary properties. Finally, we identified 100 functional amino acid subtypes with a data-driven clustering analysis and studied their features, including their frequencies and chemical properties such as tolerating polarity, hydrophobicity or being intolerant of charge or specific amino acids. The mutational landscape and amino acid subtypes provide a foundational catalogue of amino acid functional diversity, which will be refined as the number of studied protein positions increases.

Deep Mutational Scanning of Viral Glycoproteins and Their Host Receptors

Deep mutational scanning or deep mutagenesis is a powerful tool for understanding the sequence diversity available to viruses for adaptation in a laboratory setting. It generally involves tracking an in vitro selection of protein sequence variants with deep sequencing to map mutational effects based on changes in sequence abundance. Coupled with any of a number of selection strategies, deep mutagenesis can explore the mutational diversity available to viral glycoproteins, which mediate critical roles in cell entry and are exposed to the humoral arm of the host immune response. Mutational landscapes of viral glycoproteins for host cell attachment and membrane fusion reveal extensive epistasis and potential escape mutations to neutralizing antibodies or other therapeutics, as well as aiding in the design of optimized immunogens for eliciting broadly protective immunity. While less explored, deep mutational scans of host receptors further assist in understanding virus-host protein interactions. Critical residues on the host receptors for engaging with viral spikes are readily identified and may help with structural modeling. Furthermore, mutations may be found for engineering soluble decoy receptors as neutralizing agents that specifically bind viral targets with tight affinity and limited potential for viral escape. By untangling the complexities of how sequence contributes to viral glycoprotein and host receptor interactions, deep mutational scanning is impacting ideas and strategies at multiple levels for combatting circulating and emergent virus strains.

Understanding the Origins of Loss of Protein Function by Analyzing the Effects of Thousands of Variants on Activity and Abundance

Understanding and predicting how amino acid substitutions affect proteins are keys to our basic understanding of protein function and evolution. Amino acid changes may affect protein function in a number of ways including direct perturbations of activity or indirect effects on protein folding and stability. We have analyzed 6,749 experimentally determined variant effects from multiplexed assays on abundance and activity in two proteins (NUDT15 and PTEN) to quantify these effects and find that a third of the variants cause loss of function, and about half of loss-of-function variants also have low cellular abundance. We analyze the structural and mechanistic origins of loss of function and use the experimental data to find residues important for enzymatic activity. We performed computational analyses of protein stability and evolutionary conservation and show how we may predict positions where variants cause loss of activity or abundance. In this way, our results link thermodynamic stability and evolutionary conservation to experimental studies of different properties of protein fitness landscapes.

Disentangling the Contribution of Each Descriptive Characteristic of Every Single Mutation to Its Functional Effects

Mutational effects predictions continue to improve in accuracy as advanced artificial intelligence (AI) algorithms are trained on exhaustive experimental data. The next natural questions to ask are if it is possible to gain insights into which attribute of the mutation contributes how much to the mutational effects and if one can develop universal rules for mapping the descriptors to mutational effects. In this work, we mainly address the former aspect using a framework of interpretable AI. Relations between the physicochemical descriptors and their contributions to the mutational effects are extracted by analyzing the data on 29,832 variants from eight systematic deep mutational scan studies. An opposite trend in the dependence of fitness and solubility on the distance of the amino acid from the catalytic sites could be extracted and quantified. The dependence of the mutational effect contributions on the position-specific scoring matrix (PSSM) score for the amino acid after mutation or the BLOSUM score of the substitution showed universal trends. Our attempts in the present work to explain the quantitative differences in the dependence on conservation and SASA across proteins were not successful. The work nevertheless brings transparency into the predictions and development of rules, and will hopefully lead to empirically uncovering the universalities among these rules.

Prediction of Function Determining and Buried Residues Through Analysis of Saturation Mutagenesis Datasets

Mutational scanning can be used to probe effects of large numbers of point mutations on protein function. Positions affected by mutation are primarily at either buried or at exposed residues directly involved in function, hereafter designated as active-site residues. In the absence of prior structural information, it has not been easy to distinguish between these two categories of residues. We curated and analyzed a set of twelve published deep mutational scanning datasets. The analysis revealed differential patterns of mutational sensitivity and substitution preferences at buried and exposed positions. Prediction of buried-sites solely from the mutational sensitivity data was facilitated by incorporating predicted sequence-based accessibility values. For active-site residues we observed mean sensitivity, specificity and accuracy of 61, 90 and 88% respectively. For buried residues the corresponding figures were 59, 90 and 84% while for exposed non active-site residues these were 98, 44 and 82% respectively. We also identified positions which did not follow these general trends and might require further experimental re-validation. This analysis highlights the ability of deep mutational scans to provide important structural and functional insights, even in the absence of three-dimensional structures determined using conventional structure determination techniques, and also discuss some limitations of the methodology.

Composition-dependent multivalency of peptide-peptide interactions revealed by tryptophan-scanning mutagenesis

We have examined in this contribution the composition dependence of binding characteristics in peptide-peptide interactions between an oligopeptide octa-glycine and a series of tryptophan-containing octapeptides. The binding energy associated with tryptophan-glycine interactions manifests pronounced stepwise binding characteristics as the number of tryptophan increases from 0 to 8 in the octapeptides consisting only of glycine and can be attributed to mono-, di-, and tri-valent peptide-peptide interactions. At the same time, only weak fluctuations in binding energy were observed as the number of tryptophan increases from 2 to 7. Such distinctive nonlinearity of composition-dependent tryptophan-glycine binding energy characteristics due to continuously varying tryptophan compositions in the octapeptides could be considered as a reflection of combinatorial contributions due to the hydrogen bonds originated from the indole moieties of tryptophan with the main chains of octapeptide of glycine containing N-H and C=O moieties and the van der Waals interactions (including π-π and π-CH interactions) between peptides.

A large-scale survey of pairwise epistasis reveals a mechanism for evolutionary expansion and specialization of PDZ domains

Deep mutational scanning (DMS) facilitates data-driven models of protein structure and function. Here, we adapted Saturated Programmable Insertion Engineering (SPINE) as a programmable DMS technique. We validate SPINE with a reference single mutant dataset in the PSD95 PDZ3 domain and then characterize most pairwise double mutants to study epistasis. We observe wide-spread proximal negative epistasis, which we attribute to mutations affecting thermodynamic stability, and strong long-range positive epistasis, which is enriched in an evolutionarily conserved and function-defining network of "sector" and clade-specifying residues. Conditional neutrality of mutations in clade-specifying residues compensates for deleterious mutations in sector positions. This suggests that epistatic interactions between these position pairs facilitated the evolutionary expansion and specialization of PDZ domains. We propose that SPINE provides easy experimental access to reveal epistasis signatures in proteins that will improve our understanding of the structural basis for protein function and adaptation.

Genomic analyses of glycine decarboxylase neurogenic mutations yield a large-scale prediction model for prenatal disease

Hundreds of mutations in a single gene result in rare diseases, but why mutations induce severe or attenuated states remains poorly understood. Defect in glycine decarboxylase (GLDC) causes Non-ketotic Hyperglycinemia (NKH), a neurological disease associated with elevation of plasma glycine. We unified a human multiparametric NKH mutation scale that separates severe from attenuated neurological disease with new in silico tools for murine and human genome level-analyses, gathered in vivo evidence from mice engineered with top-ranking attenuated and a highly pathogenic mutation, and integrated the data in a model of pre- and post-natal disease outcomes, relevant for over a hundred major and minor neurogenic mutations. Our findings suggest that highly severe neurogenic mutations predict fatal, prenatal disease that can be remedied by metabolic supplementation of dams, without amelioration of persistent plasma glycine. The work also provides a systems approach to identify functional consequences of mutations across hundreds of genetic diseases. Our studies provide a new framework for a large scale understanding of mutation functions and the prediction that severity of a neurogenic mutation is a direct measure of pre-natal disease in neurometabolic NKH mouse models. This framework can be extended to analyses of hundreds of monogenetic rare disorders where the underlying genes are known but understanding of the vast majority of mutations and why and how they cause disease, has yet to be realized.

Building models of human genetic disease, both computational and animal, is an essential part of understanding the disease, designing treatments, and testing therapies. Here, we have developed new in silico tools to build models for the rare neurological disorder non-ketotic hyperglycinemia (NKH), which is caused by mutations in glycine decarboxylase (GLDC), a protein that degrades glycine. We first applied a mutation scoring tool to GLDC in both the human and mouse genomes, and then used this data to develop a computational model for predicting which mutations would be well-modeled in mice, and how severe their disease would be. We then validated this computational model by genetically-engineering a mutation predicted to cause mild disease and another predicted to cause severe disease. Our predictions were correct and we used them to develop a model relevant for over a hundred major and minor neurogenic mutations that suggests that the more severe the mutation, the greater chance it will cause disease that starts before birth and is likely to be fatal unless rescued by modifying diet. This study also demonstrates the power of in silico analyses for guiding the development of genetic disease models and incorporating them into scalable models that can be applied to understand hundreds of mutations that cause disease.

Directed evolution of cyclic peptides for inhibition of autophagy

In recent decades it has become increasingly clear that induction of autophagy plays an important role in the development of treatment resistance and dormancy in many cancer types. Unfortunately, chloroquine (CQ) and hydroxychloroquine (HCQ), two autophagy inhibitors in clinical trials, suffer from poor pharmacokinetics and high toxicity at therapeutic dosages. This has prompted intense interest in the development of targeted autophagy inhibitors to re-sensitize disease to treatment with minimal impact on normal tissue. We utilized Scanning Unnatural Protease Resistant (SUPR) mRNA display to develop macrocyclic peptides targeting the autophagy protein LC3. The resulting peptides bound LC3A and LC3B-two essential components of the autophagosome maturation machinery-with mid-nanomolar affinities and disrupted protein-protein interactions (PPIs) between LC3 and its binding partners in vitro. The most promising LC3-binding SUPR peptide accessed the cytosol at low micromolar concentrations as measured by chloroalkane penetration assay (CAPA) and inhibited starvation-mediated GFP-LC3 puncta formation in a concentration-dependent manner. LC3-binding SUPR peptides re-sensitized platinum-resistant ovarian cancer cells to cisplatin treatment and triggered accumulation of the adapter protein p62 suggesting decreased autophagic flux through successful disruption of LC3 PPIs in cell culture. In mouse models of metastatic ovarian cancer, treatment with LC3-binding SUPR peptides and carboplatin resulted in almost complete inhibition of tumor growth after four weeks of treatment. These results indicate that SUPR peptide mRNA display can be used to develop cell-penetrating macrocyclic peptides that target and disrupt the autophagic machinery in vitro and in vivo.

Massively parallel functional testing of MSH2 missense variants conferring Lynch syndrome risk

The lack of functional evidence for the majority of missense variants limits their clinical interpretability and poses a key barrier to the broad utility of carrier screening. In Lynch syndrome (LS), one of the most highly prevalent cancer syndromes, nearly 90% of clinically observed missense variants are deemed “variants of uncertain significance” (VUS). To systematically resolve their functional status, we performed a massively parallel screen in human cells to identify loss-of-function missense variants in the key DNA mismatch repair factor MSH2. The resulting functional effect map is substantially complete, covering 94% of the 17,746 possible variants, and is highly concordant (96%) with existing functional data and expert clinicians’ interpretations. The large majority (89%) of missense variants were functionally neutral, perhaps unexpectedly in light of its evolutionary conservation. These data provide ready-to-use functional evidence to resolve the ∼1,300 extant missense VUSs in MSH2 and may facilitate the prospective classification of newly discovered variants in the clinic.

Toward Developing Intuitive Rules for Protein Variant Effect Prediction Using Deep Mutational Scanning Data

Protein structure and function can be severely altered by even a single amino acid mutation. Predictions of mutational effects using extensive artificial intelligence (AI)-based models, although accurate, remain as enigmatic as the experimental observations in terms of improving intuitions about the contributions of various factors. Inspired by Lipinski's rules for drug-likeness, we devise simple thresholding criteria on five different descriptors such as conservation, which have so far been limited to qualitative interpretations such as high conservation implies high mutational effect. We analyze systematic deep mutational scanning data of all possible single amino acid substitutions on seven proteins (25153 mutations) to first define these thresholds and then to evaluate the scope and limits of the predictions. At this stage, the approach allows us to comment easily and with a low error rate on the subset of mutations classified as neutral or deleterious by all of the descriptors. We hope that complementary to the accurate AI predictions, these thresholding rules or their subsequent modifications will serve the purpose of codifying the knowledge about the effects of mutations.

The N-terminal length and side-chain composition of CXCL13 affect crystallization, structure and functional activity

CXCL13 is the cognate chemokine agonist of CXCR5, a class A G-protein-coupled receptor (GPCR) that is essential for proper humoral immune responses. Using a `methionine scanning' mutagenesis method on the N-terminus of CXCL13, which is the chemokine signaling region, it was shown that minor length alterations and side-chain substitutions still result in CXCR5 activation. This observation indicates that the orthosteric pocket of CXCR5 can tolerate these changes without severely affecting the activity. The introduction of bulk on the ligand was well tolerated by the receptor, whereas a loss of contacts was less tolerated. Furthermore, two crystal structures of CXCL13 mutants were solved, both of which represent the first uncomplexed structures of the human protein. These structures were stabilized by unique interactions formed by the N-termini of the ligands, indicating that CXCL13 exhibits substantial N-terminal flexibility while the chemokine core domain remains largely unchanged. Additionally, it was observed that CXCL13 harbors a large degree of flexibility in the C-terminal extension of the ligand. Comparisons with other published structures of human and murine CXCL13 validate the relative rigidity of the core domain as well as the N- and C-terminal mobilities. Collectively, these mutants and their structures provide the field with additional insights into how CXCL13 interacts with CXCR5.

Multiplexed measurement of variant abundance and activity reveals VKOR topology, active site and human variant impact

Vitamin K epoxide reductase (VKOR) drives the vitamin K cycle, activating vitamin K-dependent blood clotting factors. VKOR is also the target of the widely used anticoagulant drug, warfarin. Despite VKOR's pivotal role in coagulation, its structure and active site remain poorly understood. In addition, VKOR variants can cause vitamin K-dependent clotting factor deficiency or alter warfarin response. Here, we used multiplexed, sequencing-based assays to measure the effects of 2,695 VKOR missense variants on abundance and 697 variants on activity in cultured human cells. The large-scale functional data, along with an evolutionary coupling analysis, supports a four transmembrane domain topology, with variants in transmembrane domains exhibiting strongly deleterious effects on abundance and activity. Functionally constrained regions of the protein define the active site, and we find that, of four conserved cysteines putatively critical for function, only three are absolutely required. Finally, 25% of human VKOR missense variants show reduced abundance or activity, possibly conferring warfarin sensitivity or causing disease.

Computational analysis of androgen receptor (AR) variants to decipher the relationship between protein stability and related-diseases

Although more than 1,000 androgen receptor (AR) mutations have been identified and these mutants are pathologically important, few theoretical studies have investigated the role of AR protein folding stability in disease and its relationship with the phenotype of the patients. Here, we extracted AR variant data from four databases: ARDB, HGMD, Cosmic, and 1,000 genome. 905 androgen insensitivity syndrome (AIS)-associated loss-of-function mutants and 168 prostate cancer-associated gain-of-function mutants in AR were found. We analyzed the effect of single-residue variation on the folding stability of AR by FoldX and guanidine hydrochloride denaturation experiment, and found that genetic disease-associated mutations tend to have a significantly greater effect on protein stability than gene polymorphisms. Moreover, AR mutants in complete androgen insensitivity syndrome (CAIS) tend to have a greater effect on protein stability than in partial androgen insensitive syndrome (PAIS). This study, by linking disease phenotypes to changes in AR stability, demonstrates the importance of protein stability in the pathogenesis of hereditary disease.

Identification of biochemically neutral positions in liver pyruvate kinase

Understanding how each residue position contributes to protein function has been a long-standing goal in protein science. Substitution studies have historically focused on conserved protein positions. However, substitutions of nonconserved positions can also modify function. Indeed, we recently identified nonconserved positions that have large substitution effects in human liver pyruvate kinase (hLPYK), including altered allosteric coupling. To facilitate a comparison of which characteristics determine when a nonconserved position does vs does not contribute to function, the goal of the current work was to identify neutral positions in hLPYK. However, existing hLPYK data showed that three features commonly associated with neutral positions-high sequence entropy, high surface exposure, and alanine scanning-lacked the sensitivity needed to guide experimental studies. We used multiple evolutionary patterns identified in a sequence alignment of the PYK family to identify which positions were least patterned, reasoning that these were most likely to be neutral. Nine positions were tested with a total of 117 amino acid substitutions. Although exploring all potential functions is not feasible for any protein, five parameters associated with substrate/effector affinities and allosteric coupling were measured for hLPYK variants. For each position, the aggregate functional outcomes of all variants were used to quantify a "neutrality" score. Three positions showed perfect neutral scores for all five parameters. Furthermore, the nine positions showed larger neutral scores than 17 positions located near allosteric binding sites. Thus, our strategy successfully enriched the dataset for positions with neutral and modest substitutions.

Position-coded multivalent peptide-peptide interactions revealed by tryptophan-scanning mutagenesis

We demonstrate in this contribution the evidence that significant cooperative binding effect can be identified for the amino acid sites that are determinant to the binding characteristics in peptide-peptide interactions. The analysis of tryptophan-scanning mutagenesis of the 14-mer peptide consisting only of glycine provides a mapping of position-dependent contributions to the binding energy. The pronounced tryptophan-associated peptide-peptide interactions are originated from the indole moieties with the main chains of 14-mer glycines containing N-H and CO moieties. Specifically, with the presence of two tryptophans as determinant amino acids, cooperative binding can be observed, which are dependent on relative positions of the two tryptophans with a "volcano"-like characteristics. An optimal separation of 6-10 amino acids between two adjacent binding sites can be identified to achieve maximal binding interactions.

Peptide Folding and Binding Probed by Systematic Non-canonical Mutagenesis

Many proteins and peptides fold upon binding another protein. Mutagenesis has proved an essential tool in the study of these multi-step molecular recognition processes. By comparing the biophysical behavior of carefully selected mutants, the concert of interactions and conformational changes that occur during folding and binding can be separated and assessed. Recently, this mutagenesis approach has been radically expanded by deep mutational scanning methods, which allow for many thousands of mutations to be examined in parallel. Furthermore, these high-throughput mutagenesis methods have been expanded to include mutations to non-canonical amino acids, returning peptide structure-activity relationships with unprecedented depth and detail. These developments are timely, as the insights they provide can guide the optimization of de novo cyclic peptides, a promising new modality for chemical probes and therapeutic agents.

Comprehensive exploration of the translocation, stability and substrate recognition requirements in VIM-2 lactamase

Metallo-β-lactamases (MBLs) degrade a broad spectrum of β-lactam antibiotics, and are a major disseminating source for multidrug resistant bacteria. Despite many biochemical studies in diverse MBLs, molecular understanding of the roles of residues in the enzyme's stability and function, and especially substrate specificity, is lacking. Here, we employ deep mutational scanning (DMS) to generate comprehensive single amino acid variant data on a major clinical MBL, VIM-2, by measuring the effect of thousands of VIM-2 mutants on the degradation of three representative classes of β-lactams (ampicillin, cefotaxime, and meropenem) and at two different temperatures (25°C and 37°C). We revealed residues responsible for expression and translocation, and mutations that increase resistance and/or alter substrate specificity. The distribution of specificity-altering mutations unveiled distinct molecular recognition of the three substrates. Moreover, these function-altering mutations are frequently observed among naturally occurring variants, suggesting that the enzymes have continuously evolved to become more potent resistance genes.

TLmutation: Predicting the Effects of Mutations Using Transfer Learning

A reccurring challenge in bioinformatics is predicting the phenotypic consequence of amino acid variation in proteins. With the recent advancements in sequencing techniques, sufficient genomic data has become available to train models that predict the evolutionary statistical energies, but there is still inadequate experimental data to directly predict functional effects. One approach to overcome this data scarcity is to apply transfer learning and train more models with available data sets. In this study, we propose a set of transfer learning algorithms we call TLmutation, which implements a supervised transfer learning algorithm that transfers knowledge from survival data of a protein to a particular function of that protein. This is followed by an unsupervised transfer learning algorithm that extends the knowledge to a homologous protein. We explore the application of our algorithms in three cases. First, we test the supervised transfer on 17 previously published deep mutagenesis data sets to complete and refine missing data points. We further investigate these data sets to identify which mutations build better predictors of variant functions. In the second case, we apply the algorithm to predict higher-order mutations solely from single point mutagenesis data. Finally, we perform the unsupervised transfer learning algorithm to predict mutational effects of homologous proteins from experimental data sets. These algorithms are generalized to transfer knowledge between Markov random field models. We show the benefit of our transfer learning algorithms to utilize informative deep mutational data and provide new insights into protein variant functions. As these algorithms are generalized to transfer knowledge between Markov random field models, we expect these algorithms to be applicable to other disciplines.

Thermodynamic and Evolutionary Coupling between the Native and Amyloid State of Globular Proteins

The amyloid-like aggregation propensity present in most globular proteins is generally considered to be a secondary side effect resulting from the requirements of protein stability. Here, we demonstrate, however, that mutations in the globular and amyloid state are thermodynamically correlated rather than simply associated. In addition, we show that the standard genetic code couples this structural correlation into a tight evolutionary relationship. We illustrate the extent of this evolutionary entanglement of amyloid propensity and globular protein stability. Suppressing a 600-Ma-conserved amyloidogenic segment in the p53 core domain fold is structurally feasible but requires 7-bp substitutions to concomitantly introduce two aggregation-suppressing and three stabilizing amino acid mutations. We speculate that, rather than being a corollary of protein evolution, it is equally plausible that positive selection for amyloid structure could have been a driver for the emergence of globular protein structure.