The INO80 chromatin-remodeling complex is a multi-subunit regulator of DNA-templated processes, yet the mechanisms that control remodeler function in vivo are not completely known. In this study, we report that the Ies6 subunit of the INO80 complex encodes a prion-like domain (PLD) within a larger region of predicted disorder. After transient inducible overexpression, both full-length Ies6 and the PLD domain alone aggregate in a manner that evades proteasome-mediated degradation, suggestive of prion-like behavior. Cytosolic aggregates are also visible with fluorescent microscopy following expression of the PLD alone. Loss of the protein chaperone Hsp70 increases Ies6 aggregation. In addition, deletion of ARP5, another INO80 subunit and binding partner of Ies6, also results in elevated protein aggregation. Finally, transcriptome analysis indicates that loss of PLD-mediated aggregation alters the expression of telomere-proximal and metabolic stress-responsive genes, including those involved in glucose starvation and respiratory programs, even under glucose-replete conditions. Together, these findings support a model in which Ies6 couples the INO80 complex to metabolic gene regulation via prion-like behavior, providing a potential new mechanism for tuning chromatin-based metabolic adaptation.

Most proteins act through interactions with other molecules, yet predicting how single mutations perturb these interactions-defined as 'protein codes'-remains a central challenge in computational biology. Here we introduce eSIG-Net, the edgetic mutation sequence-based interaction grammar network, a language model that integrates protein sequence embeddings with syntax-aware and evolution-aware mutation encoding and contrastive learning to predict mutation-driven interaction changes. eSIG-Net outperforms state-of-the-art sequence-based and structure-based methods, nominates causal variants and provides mechanistic insights. Together, eSIG-Net is a mutation-centric interaction language model that accurately predicts interaction-specific network rewiring from sequence information alone and generalizes across biological contexts.

Mitochondria, which evolved from symbiotic bacteria, possess their own genomes (mtDNA) and support independent transcription and translation within the organelle. Given the essential role of mtDNA in energy production, metabolism, as well as cellular homeostasis, and the high density of confirmed pathogenic mutations that map to mtDNA, there is a pressing need for versatile methods to study and manipulate this genome. Although CRISPR technology has revolutionized the editing of nuclear genomes, it has not been successfully extended to mtDNA, primarily due to the challenge of delivering single guide RNAs (sgRNAs) across both outer and inner mitochondrial membranes. Here we develop a survival-based reporter in Saccharomyces cerevisiae to screen for potential RNA import motifs. We identify a 40-nucleotide aptamer (IM83) that facilitates sgRNA entry into the mitochondrial matrix, enabling CRISPR editing by a mitochondrially-localized adenine base editor. We show that mitochondrial import of IM83 is ATP-dependent and enhanced by the tRNA synthetase Msk1. Further investigations identify barriers to efficient CRISPR editing of mtDNA, including loss of membrane potential associated with mitochondrial targeting of the base editor. These insights lay the groundwork for future improvements in CRISPR-based editing of mtDNA in eukaryotes.

Most proteins exert their functions in complex with other interactors. Single mutations can exhibit a profound impact on perturbing protein interactions, leading to human disease. However, predicting the effect of single mutations on protein interactions remains a major computational challenge. Deep learning, particularly protein language models or transformers, has become an effective tool in bioinformatics for protein structure prediction. However, the functional divergence of mutations makes it difficult to predict their interaction perturbation profiles. To address this fundamental challenge, we present eSIG-Net (edgetic mutation Sequence-based Interaction Grammar Network), a novel sequence-based "Interaction Language Model" for predicting protein interaction alterations caused by single mutations. eSIG-Net combines various protein sequence embeddings, introduces a mutation-encoding module with syntax and evolutionary insights, and employs contrastive learning to evaluate mutation-induced interaction changes. eSIG-Net significantly outperforms current state-of-the-art sequence-based and structure-based prediction methods at predicting mutational impact on protein interactions. We highlight examples where eSIG-Net nominates causal variants with high confidence and elucidates their functional role under relevant biological contexts. Together, eSIG-Net is a first-in-kind "interaction language model" that can accurately predict interaction-specific rewiring by single mutations with only sequence information, and exhibits generalizability across biological contexts.

Mutations supply the raw material for evolution, yet because most are neutral or deleterious, elevated mutation rates are typically transient. Nonetheless, modeling predicts that a mechanism for heritable but reversible "mutagenesis switches" would be advantageous in some selective contexts. Here, we report that frequent prion-based switching of DNA repair and recombination proteins alters mutagenesis in Saccharomyces cerevisiae populations from diverse ecological niches, including the laboratory and clinic, providing adaptive benefits in short-term evolution under strong selective pressure. Self-templating protein assembly alters the activities and interactions of multiple DNA-fidelity factors, reshaping adaptive outcomes while maintaining resilience to genotoxic stress. In the WHO priority pathogen Candida albicans, which diverged from S. cerevisiae ∼300 million years ago, a key regulator of prion inheritance accelerates the rapid emergence of fluconazole resistance. These findings suggest that protein self-assembly can generate epigenetic memory that tunes genome diversification over multiple generations, enabling rapid adaptation in challenging environments.

Polyglutamine (polyQ) expansion is associated with pathogenic protein aggregation in neurodegenerative disorders. However, long polyQ tracts are also found in many transcription factors (TFs), such as FOXP2, a TF implicated in human speech. Here, we explore how FOXP2 and other glutamine-rich TFs avoid unscheduled assembly. Throughout interphase, DNA binding, irrespective of sequence specificity, has a solubilizing effect. During mitosis, multiple phosphorylation events promote FOXP2's eviction from chromatin and supplant the solubilizing function of DNA. Further, human-specific amino acid substitutions linked to the evolution of speech map to a mitotic phospho-patch, the "EVO patch," and reduce the propensity of the human FOXP2 to assemble. Fusing the pathogenic form of Huntingtin to either a DNA-binding domain, a phosphomimetic variant of this EVO patch, or a negatively charged peptide is sufficient to diminish assembly formation, suggesting that hijacking mechanisms governing solubility of glutamine-rich TFs may offer new strategies for treatment of polyQ expansion diseases.

Drug resistance is the chief cause of treatment failure for therapies targeting chronic and infectious diseases. Whether the emergence of resistance is accelerated by environmental exposure to low levels of therapeutics remains controversial. Here, we report a non-genetic mechanism of stress adaptation that promotes heritable resistance to the widely used antifungal drug fluconazole. In the human fungal pathogen Candida albicans, transient exposure to subtherapeutic fluconazole doses induces a protective response that we term para-resistance. Like conventional resistance mechanisms, para-resistance is heritable. However, it does not arise from genetic mutations and can revert spontaneously. Systematic analyses of para-resistant isolates suggest that its key regulators include the stress-activated MAP kinase Hog1, the histone deacetylase subunit Snt1, the chromatin regulator Rap1, and the Sko1 transcriptional factor. Notably, molecules that disrupt biomolecular condensation and prion propagation - crucial for the inheritance of protein assemblies - block the induction of para-resistance, whereas inhibiting histone deacetylases facilitates its induction. We find that para-resistance is common in clinical isolates and, remarkably, passage through the mammalian gut triggers its acquisition, compromising fluconazole's therapeutic efficacy. Our work defines a pervasive, prion-like epigenetic mechanism of stress adaptation and highlights potential strategies to mitigate the rapid emergence of drug resistance.

Understanding how genetic variation translates into complex phenotypes remains a fundamental challenge. In this work, we address this by mapping genome-to-proteome relationships in 800 progeny of a cross between two yeast strains adapted to distinct environments. Despite the modest genetic distance between the parents, we observed notable proteomic diversity and mapped more than 6400 genotype-protein associations, with more than 1600 linked to individual genetic variants. Proteomic adaptation emerged from a conserved network of cis- and trans-regulatory variants, often originating from proteins not traditionally linked to gene regulation. This atlas allowed us to forecast organismal fitness effects across diverse conditions. By connecting genomic and proteomic landscapes at unprecedented resolution, our study provides a framework for predicting the phenotypic outcomes of natural genetic variation.

Mutations are not always phenotypically active or show different effects in different individuals. While the mechanisms underlying this variable relationship between mutations and phenotypes are largely elusive, some specific genes may influence the phenotypic effects of cryptic variation. We employ the toolbox of Saccharomyces cerevisiae to perform a genome-wide screen aimed at identifying these so-called genetic buffer genes. Measuring the fitness of 1.8 million mutated strains identified a small set of evolutionary conserved buffer genes involved in protein folding and chromatin organization, including GIM3, SSA2, HOG1 and FKH2. Deletion of these genes increases the fitness effect of de novo mutations as well as standing genetic variation, with some mutations becoming adaptive. Moreover, losing a buffer gene results in a decline of standing genetic variation. Together, these results reveal a set of conserved genes that influence the phenotypic outcome of mutations and play a role in complex genetics and incomplete penetrance. ### Competing Interest Statement The authors have declared no competing interest.

Despite abundant genomic and phenotypic data across individuals and environments, the functional impact of most mutations on phenotype remains unclear. Here, we bridge this gap by linking genome to proteome in 800 meiotic progeny from an intercross between two closely related Saccharomyces cerevisiae isolates adapted to distinct niches. Modest genetic distance between the parents generated remarkable proteomic diversity that was amplified in the progeny and captured by 6,476 genotype-protein associations, over 1,600 of which we resolved to single variants. Proteomic adaptation emerged through the combined action of numerous cis- and trans-regulatory mutations, a regulatory architecture that was conserved across the species. Notably, trans-regulatory variants often arose in proteins not traditionally associated with gene regulation, such as enzymes. Moreover, the proteomic consequences of mutations predicted fitness under various stresses. Our study demonstrates that the collective action of natural genetic variants drives dramatic proteome diversification, with molecular consequences that forecast phenotypic outcomes.

Global modifier genes influence the mapping of genotypes onto phenotypes and fitness through their epistatic interactions with genetic variants on a massive scale. The first such factor to be identified, Hsp90, is a highly conserved molecular chaperone that plays a central role in protein homeostasis. Hsp90 is a “hub of hubs” that chaperones proteins engaged in many key cellular and developmental regulatory networks. These clients, which are enriched in kinases, transcription factors, and E3 ubiquitin ligases, drive diverse cellular functions and are themselves highly connected. By contrast to many other hub proteins, the abundance and activity of Hsp90 changes substantially in response to shifting environmental conditions. As a result, Hsp90 modifies the functional impact of many genetic variants simultaneously in a manner that depends on environmental stress. Studies in diverse organisms suggest that this coupling between Hsp90 function and challenging environments exerts a substantial impact on what parts of the genome are visible to natural selection, expanding adaptive opportunities when most needed. In this Perspective, we explore the multifaceted role of Hsp90 as global modifier of the genotype-phenotype-fitness map as well as its implications for evolution in nature and the clinic.

Protein aggregation is a hallmark of age-related neurodegeneration. Yet, aggregation during normal aging and in tissues other than the brain is poorly understood. Here, we leverage the African turquoise killifish to systematically profile protein aggregates in seven tissues of an aging vertebrate. Age-dependent aggregation is strikingly tissue specific and not simply driven by protein expression differences. Experimental interrogation in killifish and yeast, combined with machine learning, indicates that this specificity is linked to protein-autonomous biophysical features and tissue-selective alterations in protein quality control. Co-aggregation of protein quality control machinery during aging may further reduce proteostasis capacity, exacerbating aggregate burden. A segmental progeria model with accelerated aging in specific tissues exhibits selectively increased aggregation in these same tissues. Intriguingly, many age-related protein aggregates arise in wild-type proteins that, when mutated, drive human diseases. Our data chart a comprehensive landscape of protein aggregation during vertebrate aging and identify strong, tissue-specific associations with dysfunction and disease.

Protein aggregation, which can sometimes spread in a prion-like manner, is a hallmark of neurodegenerative diseases. However, whether prion-like aggregates form during normal brain aging remains unknown. Here, we use quantitative proteomics in the African turquoise killifish to identify protein aggregates that accumulate in old vertebrate brains. These aggregates are enriched for prion-like RNA-binding proteins, notably the ATP-dependent RNA helicase DDX5. We validate that DDX5 forms aggregate-like puncta in the brains of old killifish and mice. Interestingly, DDX5's prion-like domain allows these aggregates to propagate across many generations in yeast. In vitro, DDX5 phase separates into condensates. Mutations that abolish DDX5 prion propagation also impair the protein's ability to phase separate. DDX5 condensates exhibit enhanced enzymatic activity, but they can mature into inactive, solid aggregates. Our findings suggest that protein aggregates with prion-like properties form during normal brain aging, which could have implications for the age-dependency of cognitive decline.

Cell size is tightly controlled in healthy tissues and single-celled organisms, but it remains unclear how cell size influences physiology. Increasing cell size was recently shown to remodel the proteomes of cultured human cells, demonstrating that large and small cells of the same type can be compositionally different. In the present study, we utilize the natural heterogeneity of hepatocyte ploidy and yeast genetics to establish that the ploidy-to-cell size ratio is a highly conserved determinant of proteome composition. In both mammalian and yeast cells, genome dilution by cell growth elicits a starvation-like phenotype, suggesting that growth in large cells is restricted by genome concentration in a manner that mimics a limiting nutrient. Moreover, genome dilution explains some proteomic changes ascribed to yeast aging. Overall, our data indicate that genome concentration drives changes in cell composition independently of external environmental cues. Here the authors establish that ploidy-to-cell size ratio is a highly conserved determinant of proteome composition. In both mammals and yeast, they find that growth in large cells is restricted by genome concentration as if it were a limiting nutrient.

The essential stress-responsive chaperone Hsp90 impacts development and adaptation from microbes to humans. Yet despite evidence of its role in evolution, pathogenesis, and oncogenic transformation, the molecular mechanisms by which Hsp90 alters the consequences of mutations remain vigorously debated. Here we exploit the power of nucleotide-resolution genetic mapping in Saccharomyces cerevisiae to uncover more than 1,000 natural variant-to-phenotype associations governed by this molecular chaperone. Strikingly, Hsp90 more frequently modified the phenotypic effects of cis-regulatory variation than variants that altered protein sequence. Moreover, these interactions made the largest contribution to Hsp90-dependent heredity. Nearly all interacting variants-both regulatory and protein-coding-fell within clients of Hsp90 or targets of its direct binding partners. Hsp90 activity affected mutations in evolutionarily young genes, segmental deletions, and heterozygotes, highlighting its influence on variation central to evolutionary novelty. Reconciling the diverse epistatic effects of this chaperone, synthetic transcriptional regulation and reconstructions of natural alleles by genome editing revealed a central role for Hsp90 in regulating the fundamental relationship between activity and phenotype. Our findings establish that non-coding variation is a core driver of Hsp90's influence on heredity, offering a mechanistic explanation for the chaperone's strong effects on evolution and development across species.

Fusion oncoproteins (FOs) arise from chromosomal translocations in ~17% of cancers and are often oncogenic drivers. Although some FOs can promote oncogenesis by undergoing liquid-liquid phase separation (LLPS) to form aberrant biomolecular condensates, the generality of this phenomenon is unknown. We explored this question by testing 166 FOs in HeLa cells and found that 58% formed condensates. The condensate-forming FOs displayed physicochemical features distinct from those of condensate-negative FOs and segregated into distinct feature-based groups that aligned with their sub-cellular localization and biological function. Using Machine Learning, we developed a predictor of FO condensation behavior, and discovered that 67% of ~3000 additional FOs likely form condensates, with 35% of those predicted to function by altering gene expression. 47% of the predicted condensate-negative FOs were associated with cell signaling functions, suggesting a functional dichotomy between condensate-positive and -negative FOs. Our Datasets and reagents are rich resources to interrogate FO condensation in the future.

Oxygen deprivation and excess are both toxic. Thus, the body’s ability to adapt to varying oxygen tensions is critical for survival. While the hypoxia transcriptional response has been well studied, the post-translational effects of oxygen have been underexplored. In this study, we systematically investigate protein turnover rates in mouse heart, lung, and brain under different inhaled oxygen tensions. We find that the lung proteome is the most responsive to varying oxygen tensions. In particular, several extracellular matrix (ECM) proteins are stabilized in the lung under both hypoxia and hyperoxia. Furthermore, we show that complex 1 of the electron transport chain is destabilized in hyperoxia, in accordance with the exacerbation of associated disease models by hyperoxia and rescue by hypoxia. Moreover, we nominate MYBBP1A as a hyperoxia transcriptional regulator, particularly in the context of rRNA homeostasis. Overall, our study highlights the importance of varying oxygen tensions on protein turnover rates and identifies tissue-specific mediators of oxygen-dependent responses.

Rapid mutation fuels the evolution of many cancers and pathogens. Much of the ensuing genetic variation is detrimental, but cells can survive by limiting the cost of accumulating mutation burden. We investigated this behavior by propagating hypermutating yeast lineages to create independent populations harboring thousands of distinct genetic variants. Mutation rate and spectrum remained unchanged throughout the experiment, yet lesions that arose early were more deleterious than those that arose later. Although the lineages shared no mutations in common, each mounted a similar transcriptional response to mutation burden. The proteins involved in this response formed a highly connected network that has not previously been identified. Inhibiting this response increased the cost of accumulated mutations, selectively killing highly mutated cells. A similar gene expression program exists in hypermutating human cancers and is linked to survival. Our data thus define a conserved stress response that buffers the cost of accumulating genetic lesions and further suggest that this network could be targeted therapeutically.