Biology subjects

Lehner, B.

Publications and source records attributed to Lehner, B..

At least 37 records · Page 2Linked to original sources

bioRxiv2024

Energetic portrait of the amyloid beta nucleation transition state

Amyloid protein aggregates are pathological hallmarks of more than fifty human diseases but how soluble proteins nucleate to form amyloids is poorly understood. Here we use combinatorial mutagenesis, a kinetic selection assay, and machine learning to massively perturb the energetics of the nucleation reaction of amyloid beta (A{beta}42), the protein that aggregates in Alzheimers disease. In total, we quantify the nucleation rates of >140,000 variants of A{beta}42. This allows us to accurately quantify the changes in reaction activation energy for all possible amino acid substitutions in a protein for the first time and, in addition, to quantify >600 energetic interactions between mutations. The data reveal the simple and interpretable genetic architecture of an amyloid nucleation reaction. Strikingly, strong energetic couplings are rare and identify a subset of structural contacts in mature fibrils. Together with the activation energy changes, this strongly suggests that the A{beta}42 nucleation reaction transition state is structured in a short C-terminal region, providing a structural model for the reaction that may initiate Alzheimers disease. We believe this approach can be widely applied to probe the energetics and transition state structures of protein reactions.

biophysics↗

bioRxiv2024

Interpretably deep learning amyloid nucleation by massive experimental quantification of random sequences

Protein aggregation is a pathological hallmark of more than fifty human diseases and a major problem for biotechnology. Methods have been proposed to predict aggregation from sequence, but these have been trained and evaluated on small and biased experimental datasets. Here we directly address this data shortage by experimentally quantifying the amyloid nucleation of >100,000 protein sequences. This unprecedented dataset reveals the limited performance of existing computational methods and allows us to train CANYA, a convolution-attention hybrid neural network that accurately predicts amyloid nucleation from sequence. We adapt genomic neural network interpretability analyses to reveal CANYAs decision-making process and learned grammar. Our results illustrate the power of massive experimental analysis of random sequence-spaces and provide an interpretable and robust neural network model to predict amyloid nucleation.

genomics↗

bioRxiv2024

Genetics, energetics and allostery during a billion years of hydrophobic protein core evolution

Protein folding is driven by the burial of hydrophobic amino acids in a tightly-packed core that excludes water. The genetics, biophysics and evolution of hydrophobic cores are not well understood, in part because of a lack of systematic experimental data on sequence combinations that do - and do not - constitute stable and functional cores. Here we randomize protein hydrophobic cores and evaluate their stability and function at scale. The data show that vast numbers of amino acid combinations can constitute stable protein cores but that these alternative cores frequently disrupt protein function because of allosteric effects. These strong allosteric effects are not due to complicated, highly epistatic fitness landscapes but rather, to the pervasive nature of allostery, with many individually small energy changes combining to disrupt function. Indeed both protein stability and ligand binding can be accurately predicted over very large evolutionary distances using additive energy models with a small contribution from pairwise energetic couplings. As a result, energy models trained on one protein can accurately predict core stability across hundreds of millions of years of protein evolution, with only rare energetic couplings that we experimentally identify limiting the transplantation of cores between highly diverged proteins. Our results reveal the simple energetic architecture of protein hydrophobic cores and suggest that allostery is a major constraint on sequence evolution.

biophysics↗

bioRxiv2024

The allosteric landscape of the Src kinase

Enzymes catalyze the reactions of life and are the targets of most small molecule drugs. Most drugs target conserved enzyme active sites, often causing problems of specificity and toxicity. Targeting allosteric sites can increase specificity, overcome resistance mutations, and allow fine-tuning of activity. However, most enzymes have no known allosteric sites and methods do not exist to comprehensively identify them. Here we present a general and fast approach to chart allosteric communication in enzymes and apply it to the Src kinase to produce the first comprehensive map of negative and positive allosteric control of an enzymatic activity. Allostery in the Src kinase domain is pervasive, anisotropic, partially predictable, and modulated by regulatory domains. Multiple surface pockets of Src are allosterically active and so genetically-prioritized for the development of inhibitory and activating drugs. Using this approach it should be possible to chart global allosteric maps of many kinases and other enzymes important for medicine and biotechnology. HighlightsO_LIFirst comprehensive map of negative and positive allosteric control of an enzymatic activity, the Src kinase. C_LIO_LIAllosteric communication is pervasive, distance dependent, and anisotropic. C_LIO_LIAllostery is conserved and modulated in the presence of the Src regulatory domains. C_LIO_LIGenetic prioritization of druggable surface pockets for Src inhibition and activation. C_LIO_LIAllosteric maps can now be constructed for many medically and industrially important kinases and enzymes. C_LI

biophysics↗

bioRxiv2024

Site saturation mutagenesis of 500 human protein domains reveals the contribution of protein destabilization to genetic disease

Missense variants that change the amino acid sequences of proteins cause one third of human genetic diseases1. Tens of millions of missense variants exist in the current human population, with the vast majority having unknown functional consequences. Here we present the first large-scale experimental analysis of human missense variants across many different proteins. Using DNA synthesis and cellular selection experiments we quantify the impact of >500,000 variants on the abundance of >500 human protein domains. This dataset, Human Domainome 1, reveals that >60% of pathogenic missense variants reduce protein stability. The contribution of stability to protein fitness varies across proteins and diseases, and is particularly important in recessive disorders. Combining stability measurements with protein language models annotates functional sites across proteins. Mutational effects on stability are largely conserved in homologous domains, allowing accurate stability prediction across entire protein families using energy models. Domainome 1 demonstrates the feasibility of assaying human protein variants at scale and provides a large consistent reference dataset for clinical variant interpretation and the training and benchmarking of computational methods.

genomics↗

bioRxiv2024

A complete map of specificity encoding for a partially fuzzy protein interaction

The human genome encodes the affinities and specificities of >half a million molecular interactions. Changes in specificity drive evolution, disease and therapeutic efficacy, but how specificity is encoded and reprogrammed is not well understood. Here we present a complete map of the encoding of specificity in a protein sequence and how mutations quantitatively interact to reprogram binding. By measuring >200,000 energetic interactions we identify 17 major sites that directly and allosterically encode the specificity for six sites in a ligand. Combining mutations allows specificity for all six sites to be simultaneously re-programmed. However, binding reprogramming is not fully independent, with energetic couplings within both the protein and ligand required to accurately predict binding. This approach of measuring the specificities of thousands of proteins in a single experiment will allow many different types of molecular interaction to be understood and interpretably reprogrammed. HighlightsO_LIMassively parallel quantification of >200,000 energetic couplings between mutations in a protein and its ligand. C_LIO_LISpecificity is directly and allosterically encoded in 17 sites for six positions in the ligand. C_LIO_LISpecificity for all six sites can be simultaneously re-programmed with accurate binding scores for >150,000 intermediately reprogrammed variants. C_LIO_LIAccurate binding prediction requires energetic interactions between the protein and ligand and within both the protein and ligand. C_LI

biophysics↗

bioRxiv2024

Deep indel mutagenesis reveals the regulatory and modulatory architecture of alternative exon splicing

Altered splicing is a frequent mechanism by which genetic variants cause disease and antisense oligonucleotides (AONs) that target pre-mRNA splicing have been approved as therapeutics for multiple pathologies including patient-customized treatments for rare diseases. However, the regulatory architecture of human exons remains poorly understood and AON discovery is currently slow and expensive, limiting the wider adoption of the approach. Here we show that that systematic deletion scans -which can be made experimentally at very low cost - provide an efficient strategy to chart the regulatory landscape of human exons and to rapidly identify effective splicing-modulating oligonucleotides in a fully quantitative manner. Our results suggest a mechanism for the evolutionary origins of unusually short microexons and the repression of transmembrane domain-encoding exons, and reveal a checkerboard architecture of sequential enhancers and silencers in a model alternative exon. Accurate prediction of the effects of deletions using deep learning provides a resource, DANGO, that maps the splicing regulatory landscape of all human exons and predicts effective splicing-altering antisense oligonucleotides genome-wide.

genetics↗

bioRxiv2024

MoCHI: neural networks to fit interpretable models and quantify energies, energetic couplings, epistasis and allostery from deep mutational scanning data

The massively parallel nature of deep mutational scanning (DMS) allows the quantification of the phenotypic effects of thousands of perturbations in a single experiment. We have developed MoCHI, a software tool that allows the parameterisation of arbitrarily complex models using DMS data. MoCHI simplifies the task of building custom models from measurements of mutant effects on any number of phenotypes. It allows the inference of free energy changes, as well as pairwise and higher-order interaction terms (energetic couplings) for specified biophysical models. When a suitable user-specified mechanistic model is not available, global nonlinearities (epistasis) can be estimated directly from the data. MoCHI also builds upon and leverages theory on ensemble (or background-averaged) epistasis to learn sparse predictive models that can incorporate higher-order epistatic terms and are informative of the genetic architecture of the underlying biological system. The combination of DMS and MoCHI allows biophysical measurements to be performed at scale, including the construction of complete allosteric maps of proteins. MoCHI is freely available (https://github.com/lehner-lab/MoCHI) and implemented as an easy-to-use python package relying on the PyTorch machine learning framework.

systems biology↗

bioRxiv2023

The genetic architecture of protein stability

There are more ways to synthesize a 100 amino acid protein (20100) than atoms in the universe. Only a miniscule fraction of such a vast sequence space can ever be experimentally or computationally surveyed. Deep neural networks are increasingly being used to navigate high-dimensional sequence spaces. However, these models are extremely complicated and provide little insight into the fundamental genetic architecture of proteins. Here, by experimentally exploring sequence spaces >1010, we show that the genetic architecture of at least some proteins is remarkably simple, allowing accurate genetic prediction in high-dimensional sequence spaces with fully interpretable biophysical models. These models capture the non-linear relationships between free energies and phenotypes but otherwise consist of additive free energy changes with a small contribution from pairwise energetic couplings. These energetic couplings are sparse and caused by structural contacts and backbone propagations. Our results suggest that artificial intelligence models may be vastly more complicated than the proteins that they are modeling and that protein genetics is actually both simple and intelligible.

biophysics↗

bioRxiv2023

The genetic architecture of protein interaction affinity and specificity

Proteins function in crowded cellular environments in which they must bind to specific target proteins but also avoid binding to many other off-target proteins. In large protein families this task is particularly challenging because many off-target proteins have very similar structures. How this specificity of physical protein-protein interactions in cellular networks is encoded and evolves is not very well understood. Here we address the question of specificity-encoding by comprehensively quantifying the effects of all mutations in one protein, JUN, on its binding to all other members of a protein family, the 54 human basic leucine zipper transcription factors. Fitting a global thermodynamic model to the data reveals that most affinity changing mutations equally affect JUNs propensity to bind to all its interaction partners. Mutations that alter the specificity of binding are much rarer. These specificity-altering mutations are, however, distributed throughout the JUN interaction interface. JUNs interaction specificity is encoded by both positive determinants that promote on-target interactions and negative determinants that prevent off-target interactions. Indeed, about half of the specificity-defining residues in JUN have dual functions and both promote on-target binding and prevent off-target binding. Whereas nearly all mutations that alter specificity are pleiotropic and also alter the affinity of binding to all interaction partners, the converse is not true with mutations outside of the interface able to tune affinity without affecting specificity. Our results provide the first global view of how mutations in a protein affect binding to all its potential interaction partners and reveal the distributed encoding of specificity and affinity in an interaction interface. They also show how the modular architecture of coiled-coils provides an elegant solution to the challenge of optimising specificity and affinity in a large protein family.

systems biology↗

bioRxiv2023

Deep indel mutagenesis reveals the impact of insertions and deletions on protein stability and function

Amino acid insertions and deletions (indels) are an abundant class of genetic variants. However, compared to substitutions, the effects of indels on protein stability are not well understood and are poorly predicted. To better understand indels here we analyze new and existing large-scale deep indel mutagenesis (DIM) of structurally diverse proteins. The effects of indels on protein stability vary extensively among and within proteins and are not well predicted by existing computational methods. To address this shortcoming we present INDELi, a series of models that combine experimental or predicted substitution effects and secondary structure information to provide good prediction of the effects of indels on both protein stability and pathogenicity. Moreover, quantifying the effects of indels on protein-protein interactions suggests that insertions can be an important class of gain-of-function variants. Our results provide an overview of the impact of indels on proteins and a method to predict their effects genome-wide.

genomics↗

bioRxiv2023

Genome-scale quantification and prediction of pathogenic stop codon readthrough by small molecules

Premature termination codons (PTCs) cause [~]10-20% of Mendelian diseases and are the major mechanism of tumor suppressor gene inactivation in cancer. A general strategy to alleviate the effects of PTCs would be to promote translational readthrough. Nonsense suppression by small molecules has proven effective in diverse disease models, but translation into the clinic is hampered by ineffective readthrough of many PTCs. Here we directly tackle the challenge of defining drug efficacy by quantifying readthrough of [~]5,800 human pathogenic stop codons by 8 drugs. We find that different drugs promote readthrough of complementary subsets of PTCs defined by local sequence context. This allows us to build interpretable models that accurately predict drug-induced readthrough genome-wide. Accurate readthrough quantification and prediction will empower clinical trial design and the development of personalized nonsense suppression therapies.

genomics↗

bioRxiv2023

An extension of the Walsh-Hadamard transform to calculate and model epistasis in genetic landscapes of arbitrary shape and complexity

Accurate models describing the relationship between genotype and phenotype are necessary in order to understand and predict how mutations to biological sequences affect the fitness and evolution of living organisms. The apparent abundance of epistasis (genetic interactions), both between and within genes, complicates this task and how to build mechanistic models that incorporate epistatic coefficients (genetic interaction terms) is an open question. The Walsh-Hadamard transform represents a rigorous computational framework for calculating and modeling epistatic interactions at the level of individual genotypic values (known as genetical, biological or physiological epistasis), and can therefore be used to address fundamental questions related to sequence-to-function encodings. However, one of its main limitations is that it can only accommodate two alleles (amino acid or nucleotide states) per sequence position. In this paper we provide an extension of the Walsh-Hadamard transform that allows the calculation and modeling of background-averaged epistasis (also known as ensemble epistasis) in genetic landscapes with an arbitrary number of states per position (20 for amino acids, 4 for nucleotides, etc.). We also provide a recursive formula for the inverse matrix and then derive formulae to directly extract any element of either matrix without having to rely on the computationally intensive task of constructing or inverting large matrices. Finally, we demonstrate the utility of our theory by using it to model epistasis within both simulated and empirical multiallelic fitness landscapes, revealing that both pairwise and higher-order genetic interactions are enriched between physically interacting positions. Author SummaryAn important question in genetics is how the effects of mutations combine to alter phenotypes. Genetic interactions (epistasis) describe non-additive effects of pairs of mutations, but can also involve higher-order (three- and four-way etc.) combinations. Quantifying higher-order interactions is experimentally very challenging requiring a large number of measurements. Techniques based on deep mutational scanning (DMS) represent valuable sources of data to study epistasis. However, the best way to extract the relevant pairwise and higher-order epistatic coefficients (genetic interaction terms) from this data for the task of phenotypic prediction remains an unresolved problem. The Walsh-Hadamard transform represents a rigorous computational framework for calculating and modeling epistatic interactions at the level of individual genotypic values. Critically, this formalism currently only allows for two alleles (amino acid or nucleotide states) per sequence position, hampering applications in more biologically realistic scenarios. Here we present an extension of the Walsh-Hadamard transform that overcomes this limitation and demonstrate the utility of our theory by using it to model epistasis within both simulated and empirical multiallelic genetic landscapes.

systems biology↗

bioRxiv2023

SUNi mutagenesis: scalable and uniform nicking for efficient generation of variant libraries

Multiplexed assays of variant effects (MAVEs) have made possible the functional assessment of all possible mutations to genes and regulatory sequences. A core pillar of the approach is generation of variant libraries, but current methods are either difficult to scale or not uniform enough to enable MAVEs at the scale of gene families or beyond. We present an improved method called Scalable and Uniform Nicking (SUNi) mutagenesis that combines massive scalability with high uniformity to enable cost-effective MAVEs of gene families and eventually genomes.

genomics↗

bioRxiv2022

The energetic and allosteric landscape for KRAS inhibition

Thousands of proteins have now been genetically-validated as therapeutic targets in hundreds of human diseases. However, very few have actually been successfully targeted and many are considered undruggable. This is particularly true for proteins that function via protein-protein interactions: direct inhibition of binding interfaces is difficult, requiring the identification of allosteric sites. However, most proteins have no known allosteric sites and a comprehensive allosteric map does not exist for any protein. Here we address this shortcoming by charting multiple global atlases of inhibitory allosteric communication in KRAS, a protein mutated in 1 in 10 human cancers. We quantified the impact of >26,000 mutations on the folding of KRAS and its binding to six interaction partners. Genetic interactions in double mutants allowed us to perform biophysical measurements at scale, inferring >22,000 causal free energy changes, a similar number of measurements as the total made for proteins to date. These energy landscapes quantify how mutations tune the binding specificity of a signalling protein and map the inhibitory allosteric sites for an important therapeutic target. Allosteric propagation is particularly effective across the central beta sheet of KRAS and multiple surface pockets are genetically-validated as allosterically active, including a distal pocket in the C-terminal lobe of the protein. Allosteric mutations typically inhibit binding to all tested effectors but they can also change the binding specificity, revealing the regulatory, evolutionary and therapeutic potential to tune pathway activation. Using the approach described here it should be possible to rapidly and comprehensively identify allosteric target sites in many important proteins.

biophysics↗

bioRxiv2022

Dominance vs. epistasis: the biophysical origins and plasticity of genetic interactions within and between alleles

A central challenge in genetics, evolutionary biology and biotechnology is to understand and predict how mutations combine to alter phenotypes, including molecular activities, fitness and disease. In diploid organisms, two mutations in the same gene can either combine on the same chromosome or on different chromosomes, with interactions between the mutations quantified as epistasis and dominance, respectively. However, a direct comparison of the extent, sign and stability of interactions within and between alleles is lacking. Here we show that, even in the simplest biophysical systems, interactions between mutations are frequent, context-dependent and different when variants are combined within and between alleles. Whereas protein folding alone generates epistasis, the addition of a single molecular interaction is sufficient to cause dominance. Epistasis and dominance interactions change quantitatively, qualitatively and differently as a system becomes more complicated or the conditions change. Altering the concentration of a ligand can, for example, switch an allele from dominant to recessive. Our results show that epistasis and dominance should be widely expected in even the simplest biological systems but also reinforce the view that they are plastic system properties and so a formidable challenge to predict. Accurate prediction of epistasis and dominance will require either detailed mechanistic understanding and experimental parameterization or brute-force measurement and learning.

genetics↗

bioRxiv2022

An atlas of amyloid aggregation: the impact of substitutions, insertions, deletions and truncations on amyloid beta fibril nucleation

Multiplexed assays of variant effects (MAVEs) guide clinical variant interpretation and reveal disease mechanisms. To date, MAVEs have focussed on a single mutation type - amino acid (AA) substitutions - despite the diversity of coding variants that cause disease. Here we use Deep Indel Mutagenesis (DIM) to generate the first comprehensive atlas of diverse variant effects for a disease protein, the amyloid beta (A{beta}) peptide that aggregates in Alzheimers disease (AD) and is mutated in familial AD (fAD). The atlas identifies known fAD mutations and reveals many variants beyond substitutions accelerate A{beta} aggregation and are likely to be pathogenic. Truncations, substitutions, insertions, single- and internal multi-AA deletions differ in their propensity to enhance or impair aggregation, but likely pathogenic variants from all classes are highly enriched in the polar N-terminus of A{beta}. This first comparative atlas highlights the importance of including diverse mutation types in MAVEs and provides important mechanistic insights into amyloid nucleation.

genomics↗

bioRxiv2021

The impact of rare germline variants on human somatic mutation processes

Somatic mutations are an inevitable component of ageing and the most important cause of cancer. The rates and types of somatic mutation vary across individuals, but relatively few inherited influences on mutation processes are known. We performed a comprehensive gene-based rare variant association study with diverse mutational processes, using human cancer genomes from over 11,000 individuals of European ancestry. By combining burden and variance tests, we identify 207 associations involving 15 somatic mutational phenotypes and 42 genes that replicated in an independent data set at a FDR of 1%. We associated rare inherited deleterious variants in novel genes such as MSH3, EXO1, SETD2, and MTOR with two different forms of DNA mismatch repair deficiency, and variants in genes such as EXO1, PAXIP1, and WRN with deficiency in homologous recombination repair. In addition, we identified associations with other mutational processes, such as APEX1 with APOBEC-signature mutagenesis. Many of the novel genes interact with each other and with known mutator genes within cellular sub-networks. Considered collectively, damaging variants in the newly-identified genes are prevalent in the population. We suggest that rare germline variation in diverse genes commonly impacts mutational processes in somatic cells.

genomics↗