Identification of Low-Complexity Domains by Compositional Signatures Reveals Class-Specific Frequencies and functions across the domains of life.

Low-complexity domains (LCDs) in proteins are typically enriched in one or two predominant amino acids. As a result, LCDs often exhibit unusual structural/biophysical tendencies and can occupy functional niches. However, for each organism, protein sequences must be compatible with intracellular biomolecules and physicochemical environment, both of which vary from organism to organism. This raises the possibility that LCDs may occupy sequence spaces in select organisms that are otherwise prohibited in most organisms. Here, we report a comprehensive survey and functional analysis of LCDs in all known reference proteomes (>21k organisms), with added focus on rare and unusual types of LCDs. LCDs were classified according to both the primary amino acid and secondary amino acid in each LCD sequence, facilitating detailed comparisons of LCD class frequencies across organisms. Examination of LCD classes at different depths (i.e., domain of life, organism, protein, and per-residue levels) reveals unique facets of LCD frequencies and functions. To our surprise, all 400 LCD classes occur in nature, although some are exceptionally rare. A number of rare classes can be defined for each domain of life, with many LCD classes appearing to be eukaryote-specific. Certain LCD classes were consistently associated with identical functions across many organisms, particularly in eukaryotes. Our analysis methods enable simultaneous, direct comparison of all LCD classes between individual organisms, resulting in a proteome-scale view of differences in LCD frequencies and functions. Together, these results highlight the remarkable diversity and functional specificity of LCDs across all known life forms.

The LCD-Composer webserver: high-specificity identification and functional analysis of low-complexity domains in proteins.

SUMMARY: Low-complexity domains (LCDs) in proteins are regions enriched in a small subset of amino acids. LCDs exist in all domains of life, often have unusual biophysical behavior, and function in both normal and pathological processes. We recently developed an algorithm to identify LCDs based predominantly on amino acid composition thresholds. Here, we have integrated this algorithm with a webserver and augmented it with additional analysis options. Specifically, users can (i) search for LCDs in whole proteomes by setting minimum composition thresholds for individual or grouped amino acids, (ii) submit a known LCD sequence to search for similar LCDs, (iii) search for and plot LCDs within a single protein, (iv) statistically test for enrichment of LCDs within a user-provided protein set and (v) specifically identify proteins with multiple types of LCDs. AVAILABILITY AND IMPLEMENTATION: The LCD-Composer server can be accessed at http://lcd-composer.bmb.colostate.edu. The corresponding command-line scripts can be accessed at https://github.com/RossLabCSU/LCD-Composer/tree/master/WebserverScripts.

Expansion and functional analysis of the SR-related protein family across the domains of life.

Serine/arginine-rich (SR) proteins comprise a family of proteins that is predominantly found in eukaryotes and plays a prominent role in RNA splicing. A characteristic feature of SR proteins is the presence of an S/R-rich low-complexity domain (RS domain), often in conjunction with spatially distinct RNA recognition motifs (RRMs). To date, 52 human proteins have been classified as SR or SR-related proteins. Here, using an unbiased series of composition criteria together with enrichment for known RNA binding activity, we identified >100 putative SR-related proteins in the human proteome. This method recovers known SR and SR-related proteins with high sensitivity (∼94%), yet identifies a number of additional proteins with many of the hallmark features of true SR-related proteins. Newly identified SR-related proteins display slightly different amino acid compositions yet similar levels of post-translational modification, suggesting that these new SR-related candidates are regulated in vivo and functionally important. Furthermore, candidate SR-related proteins with known RNA-binding activity (but not currently recognized as SR-related proteins) are nevertheless strongly associated with a variety of functions related to mRNA splicing and nuclear speckles. Finally, we applied our SR search method to all available reference proteomes, and provide maps of RS domains and Pfam annotations for all putative SR-related proteins as a resource. Together, these results expand the set of SR-related proteins in humans, and identify the most common functions associated with SR-related proteins across all domains of life.

Phase separation by the SARS-CoV-2 nucleocapsid protein: Consensus and open questions.

In response to the recent SARS-CoV-2 pandemic, a number of labs across the world have reallocated their time and resources to better our understanding of the virus. For some viruses, including SARS-CoV-2, viral proteins can undergo phase separation: a biophysical process often related to the partitioning of protein and RNA into membraneless organelles in vivo. In this review, we discuss emerging observations of phase separation by the SARS-CoV-2 nucleocapsid (N) protein-an essential viral protein required for viral replication-and the possible in vivo functions that have been proposed for N-protein phase separation, including viral replication, viral genomic RNA packaging, and modulation of host-cell response to infection. Additionally, since a relatively large number of studies examining SARS-CoV-2 N-protein phase separation have been published in a short span of time, we take advantage of this situation to compare results from similar experiments across studies. Our evaluation highlights potential strengths and pitfalls of drawing conclusions from a single set of experiments, as well as the value of publishing overlapping scientific observations performed simultaneously by multiple labs.

Generalizable Compositional Features Influencing the Proteostatic Fates of Polar Low-Complexity Domains.

Protein aggregation is associated with a growing list of human diseases. A substantial fraction of proteins in eukaryotic proteomes constitutes a proteostasis network-a collection of proteins that work together to maintain properly folded proteins. One of the overarching functions of the proteostasis network is the prevention or reversal of protein aggregation. How proteins aggregate in spite of the anti-aggregation activity of the proteostasis machinery is incompletely understood. Exposed hydrophobic patches can trigger degradation by the ubiquitin-proteasome system, a key branch of the proteostasis network. However, in a recent study, we found that model glycine (G)-rich or glutamine/asparagine (Q/N)-rich prion-like domains differ in their susceptibility to detection and degradation by this system. Here, we expand upon this work by examining whether the features controlling the degradation of our model prion-like domains generalize broadly to G-rich and Q/N-rich domains. Experimentally, native yeast G-rich domains in isolation are sensitive to the degradation-promoting effects of hydrophobic residues, whereas native Q/N-rich domains completely resist these effects and tend to aggregate instead. Bioinformatic analyses indicate that native G-rich domains from yeast and humans tend to avoid degradation-promoting features, suggesting that the proteostasis network may act as a form of selection at the molecular level that constrains the sequence space accessible to G-rich domains. However, the sensitivity or resistance of G-rich and Q/N-rich domains, respectively, was not always preserved in their native protein contexts, highlighting that proteins can evolve other sequence features to overcome the intrinsic sensitivity of some LCDs to degradation.

LCD-Composer: an intuitive, composition-centric method enabling the identification and detailed functional mapping of low-complexity domains.

Low complexity domains (LCDs) in proteins are regions predominantly composed of a small subset of the possible amino acids. LCDs are involved in a variety of normal and pathological processes across all domains of life. Existing methods define LCDs using information-theoretical complexity thresholds, sequence alignment with repetitive regions, or statistical overrepresentation of amino acids relative to whole-proteome frequencies. While these methods have proven valuable, they are all indirectly quantifying amino acid composition, which is the fundamental and biologically-relevant feature related to protein sequence complexity. Here, we present a new computational tool, LCD-Composer, that directly identifies LCDs based on amino acid composition and linear amino acid dispersion. Using LCD-Composer's default parameters, we identified simple LCDs across all organisms available through UniProt and provide the resulting data in an accessible form as a resource. Furthermore, we describe large-scale differences between organisms from different domains of life and explore organisms with extreme LCD content for different LCD classes. Finally, we illustrate the versatility and specificity achievable with LCD-Composer by identifying diverse classes of LCDs using both simple and multifaceted composition criteria. We demonstrate that the ability to dissect LCDs based on these multifaceted criteria enhances the functional mapping and classification of LCDs.

From Prions to Stress Granules: Defining the Compositional Features of Prion-Like Domains That Promote Different Types of Assemblies.

Stress granules are ribonucleoprotein assemblies that form in response to cellular stress. Many of the RNA-binding proteins found in stress granule proteomes contain prion-like domains (PrLDs), which are low-complexity sequences that compositionally resemble yeast prion domains. Mutations in some of these PrLDs have been implicated in neurodegenerative diseases, including amyotrophic lateral sclerosis and frontotemporal dementia, and are associated with persistent stress granule accumulation. While both stress granules and prions are macromolecular assemblies, they differ in both their physical properties and complexity. Prion aggregates are highly stable homopolymeric solids, while stress granules are complex dynamic biomolecular condensates driven by multivalent homotypic and heterotypic interactions. Here, we use stress granules and yeast prions as a paradigm to examine how distinct sequence and compositional features of PrLDs contribute to different types of PrLD-containing assemblies.

A proposed role for the SARS-CoV-2 nucleocapsid protein in the formation and regulation of biomolecular condensates.

To date, the recently discovered SARS-CoV-2 virus has afflicted >6.9 million people worldwide and disrupted the global economy. Development of effective vaccines or treatments for SARS-CoV-2 infection will be aided by a molecular-level understanding of SARS-CoV-2 proteins and their interactions with host cell proteins. The SARS-CoV-2 nucleocapsid (N) protein is highly homologous to the N protein of SARS-CoV, which is essential for viral RNA replication and packaging into new virions. Emerging models indicate that nucleocapsid proteins of other viruses can form biomolecular condensates to spatiotemporally regulate N protein localization and function. Our bioinformatic analyses, in combination with pre-existing experimental evidence, suggest that the SARS-CoV-2 N protein is capable of forming or regulating biomolecular condensates in vivo by interaction with RNA and key host cell proteins. We discuss multiple models, whereby the N protein of SARS-CoV-2 may harness this activity to regulate viral life cycle and host cell response to viral infection.

Composition-based prediction and rational manipulation of prion-like domain recruitment to stress granules.

Mutations in a number of stress granule-associated proteins have been linked to various neurodegenerative diseases. Several of these mutations are found in aggregation-prone prion-like domains (PrLDs) within these proteins. In this work, we examine the sequence features governing PrLD localization to stress granules upon stress. We demonstrate that many yeast PrLDs are sufficient for stress-induced assembly into microscopically visible foci that colocalize with stress granule markers. Additionally, compositional biases exist among PrLDs that assemble upon stress, and these biases are consistent across different stressors. Using these biases, we have developed a composition-based prediction method that accurately predicts PrLD assembly into foci upon heat shock. We show that compositional changes alter PrLD assembly behavior in a predictable manner, while scrambling primary sequence has little effect on PrLD assembly and recruitment to stress granules. Furthermore, we were able to design synthetic PrLDs that were efficiently recruited to stress granules, and found that aromatic amino acids, which have previously been linked to PrLD phase separation, were dispensable for this recruitment. These results highlight the flexible sequence requirements for stress granule recruitment and suggest that PrLD localization to stress granules is driven primarily by amino acid composition, rather than primary sequence.

Atypical structural tendencies among low-complexity domains in the Protein Data Bank proteome.

A variety of studies have suggested that low-complexity domains (LCDs) tend to be intrinsically disordered and are relatively rare within structured proteins in the Protein Data Bank (PDB). Although LCDs are often treated as a single class, we previously found that LCDs enriched in different amino acids can exhibit substantial differences in protein metabolism and function. Therefore, we wondered whether the structural conformations of LCDs are likewise dependent on which specific amino acids are enriched within each LCD. Here, we directly examined relationships between enrichment of individual amino acids and secondary structure tendencies across the entire PDB proteome. Secondary structure tendencies varied as a function of the identity of the amino acid enriched and its degree of enrichment. Furthermore, divergence in secondary structure profiles often occurred for LCDs enriched in physicochemically similar amino acids (e.g. valine vs. leucine), indicating that LCDs composed of related amino acids can have distinct secondary structure tendencies. Comparison of LCD secondary structure tendencies with numerous pre-existing secondary structure propensity scales resulted in relatively poor correlations for certain types of LCDs, indicating that these scales may not capture secondary structure tendencies as sequence complexity decreases. Collectively, these observations provide a highly resolved view of structural tendencies among LCDs parsed by the nature and magnitude of single amino acid enrichment.

Natural and pathogenic protein sequence variation affecting prion-like domains within and across human proteomes.

BACKGROUND: Impaired proteostatic regulation of proteins with prion-like domains (PrLDs) is associated with a variety of human diseases including neurodegenerative disorders, myopathies, and certain forms of cancer. For many of these disorders, current models suggest a prion-like molecular mechanism of disease, whereby proteins aggregate and spread to neighboring cells in an infectious manner. The development of prion prediction algorithms has facilitated the large-scale identification of PrLDs among "reference" proteomes for various organisms. However, the degree to which intraspecies protein sequence diversity influences predicted prion propensity has not been systematically examined. RESULTS: Here, we explore protein sequence variation introduced at genetic, post-transcriptional, and post-translational levels, and its influence on predicted aggregation propensity for human PrLDs. We find that sequence variation is relatively common among PrLDs and in some cases can result in relatively large differences in predicted prion propensity. Sequence variation introduced at the post-transcriptional level (via alternative splicing) also commonly affects predicted aggregation propensity, often by direct inclusion or exclusion of a PrLD. Finally, analysis of a database of sequence variants associated with human disease reveals a number of mutations within PrLDs that are predicted to increase prion propensity. CONCLUSIONS: Our analyses expand the list of candidate human PrLDs, quantitatively estimate the effects of sequence variation on the aggregation propensity of PrLDs, and suggest the involvement of prion-like mechanisms in additional human diseases.

Sky1: at the intersection of prion-like proteins and stress granule regulation.

Serine-arginine (SR) protein kinases regulate diverse cellular activities, including various steps in RNA maturation and transport. The yeast Saccharomyces cerevisiae expresses a single SR kinase, Sky1. Sky1 has a bipartite kinase domain, separated by an aggregation-prone prion-like domain (PrLD). The assembly of PrLDs is involved in the formation of various membraneless organelles, including stress granules; stress granules are reversible ribonucleoprotein assemblies that form in response to a variety of stresses. Here, we review a recent study suggesting that Sky1's PrLD promotes Sky1 recruitment to stress granules, and that Sky1 regulates stress granule dissolution by phosphorylating the RNA-shuttling protein Npl3.

The prion-like protein kinase Sky1 is required for efficient stress granule disassembly.

Stress granules are membraneless protein- and mRNA-rich organelles that form in response to perturbations in environmental conditions. Stress granule formation is reversible, and persistent stress granules have been implicated in a variety of neurodegenerative disorders, including amyotrophic lateral sclerosis. However, characterization of the factors involved in dissolving stress granules is incomplete. Many stress granule proteins contain prion-like domains (PrLDs), some of which have been linked to stress granule formation. Here, we demonstrate that the PrLD-containing yeast protein kinase Sky1 is a stress granule component. Sky1 is recruited to stress granules in part via its PrLD, and Sky1's kinase activity regulates timely stress granule disassembly during stress recovery. This effect is mediated by phosphorylation of the stress granule component Npl3. Sky1 can compensate for defects in chaperone-mediated stress granule disassembly and vice-versa, demonstrating that cells have multiple overlapping mechanisms for re-solubilizing stress granule components.

Aggregation and degradation scales for prion-like domains: sequence features and context weigh in.

Protein aggregation in vivo is generally combated by extensive proteostatic defenses. Many proteostasis factors specifically recognize aggregation-prone features and re-fold or degrade the targeted protein. However, protein aggregation is not uncommon, suggesting that some proteins employ evasive strategies to aggregate in spite of the proteostasis machinery. Therefore, in addition to understanding the inherent aggregation propensity of protein sequences, it is important to understand how these sequences affect proteostatic recognition and regulation in vivo. In a recent study, we used a genetic mutagenesis and screening approach to explore the aggregation or degradation promoting effects of the canonical amino acids in the context of G-rich and Q/N-rich prion-like domains (PrLDs). Our results indicate that aggregation propensity scales are strongly influenced by the interplay between specific PrLD features and proteostatic recognition. Here, we briefly review these results and expand upon their potential implications. In addition, a preliminary exploration of the yeast proteome suggests that these proteostatic regulation heuristics may influence the compositional features of native G-rich and Q/N-rich domains in yeast. These results improve our understanding of the features affecting the aggregation and proteostatic regulation of prion-like domains in a cellular context, and suggest that the sequence space for native prion-like domains may be shaped by proteostatic constraints.

Proteome-scale relationships between local amino acid composition and protein fates and functions.

Proteins with low-complexity domains continue to emerge as key players in both normal and pathological cellular processes. Although low-complexity domains are often grouped into a single class, individual low-complexity domains can differ substantially with respect to amino acid composition. These differences may strongly influence the physical properties, cellular regulation, and molecular functions of low-complexity domains. Therefore, we developed a bioinformatic approach to explore relationships between amino acid composition, protein metabolism, and protein function. We find that local compositional enrichment within protein sequences is associated with differences in translation efficiency, abundance, half-life, protein-protein interaction promiscuity, subcellular localization, and molecular functions of proteins on a proteome-wide scale. However, local enrichment of related amino acids is sometimes associated with opposite effects on protein regulation and function, highlighting the importance of distinguishing between different types of low-complexity domains. Furthermore, many of these effects are discernible at amino acid compositions below those required for classification as low-complexity or statistically-biased by traditional methods and in the absence of homopolymeric amino acid repeats, indicating that thresholds employed by classical methods may not reflect biologically relevant criteria. Application of our analyses to composition-driven processes, such as the formation of membraneless organelles, reveals distinct composition profiles even for closely related organelles. Collectively, these results provide a unique perspective and detailed insights into relationships between amino acid composition, protein metabolism, and protein functions.

Sequence features governing aggregation or degradation of prion-like proteins.

Enhanced protein aggregation and/or impaired clearance of aggregates can lead to neurodegenerative disorders such as Alzheimer's Disease, Huntington's Disease, and prion diseases. Therefore, many protein quality control factors specialize in recognizing and degrading aggregation-prone proteins. Prions, which generally result from self-propagating protein aggregates, must therefore evade or outcompete these quality control systems in order to form and propagate in a cellular context. We developed a genetic screen in yeast that allowed us to explore the sequence features that promote degradation versus aggregation of a model glutamine/asparagine (Q/N)-rich prion domain from the yeast prion protein, Sup35, and two model glycine (G)-rich prion-like domains from the human proteins hnRNPA1 and hnRNPA2. Unexpectedly, we found that aggregation propensity and degradation propensity could be uncoupled in multiple ways. First, only a subset of classically aggregation-promoting amino acids elicited a strong degradation response in the G-rich prion-like domains. Specifically, large aliphatic residues enhanced degradation of the prion-like domains, whereas aromatic residues promoted prion aggregation without enhancing degradation. Second, the degradation-promoting effect of aliphatic residues was suppressed in the context of the Q/N-rich prion domain, and instead led to a dose-dependent increase in the frequency of spontaneous prion formation. Degradation suppression correlated with Q/N content of the surrounding prion domain, potentially indicating an underappreciated activity for these residues in yeast prion domains. Collectively, these results provide key insights into how certain aggregation-prone proteins may evade protein quality control degradation systems.

Manipulating the aggregation activity of human prion-like proteins.

Considerable advances in understanding the protein features favoring prion formation in yeast have facilitated the development of effective yeast prion prediction algorithms. Here we discuss a recent study in which we systematically explored the utility of the yeast prion prediction algorithm PAPA for designing mutations to modulate the aggregation activity of the human prion-like protein hnRNPA2B1. Mutations in hnRNPA2B1 cause multisystem proteinopathy in humans, and accelerate aggregation of the protein in vitro. Additionally, mutant hnRNPA2B1 forms cytoplasmic inclusions when expressed in Drosophila, and the mutant prion-like domain can substitute for a portion of a yeast prion domain in supporting prion activity in yeast. PAPA was quite successful at predicting the effects of PrLD mutations on prion activity in yeast and on in vitro aggregation propensity. Additionally, PAPA successfully predicted the effects of most, but not all, mutations in the PrLD of the hnRNPA2B1 protein when expressed in Drosophila. These results suggest that PAPA is quite effective at predicting the effects of mutations on intrinsic aggregation propensity, but that intracellular factors can influence aggregation and prion-like activity in vivo. A more complete understanding of these intracellular factors may inform the next generation of prion prediction algorithms.

The effects of glutamine/asparagine content on aggregation and heterologous prion induction by yeast prion-like domains.

Prion-like domains are low complexity, intrinsically disordered domains that compositionally resemble yeast prion domains. Many prion-like domains are involved in the formation of either functional or pathogenic protein aggregates. These aggregates range from highly dynamic liquid droplets to highly ordered detergent-insoluble amyloid-like aggregates. To better understand the amino acid sequence features that promote conversion to stable, detergent-insoluble aggregates, we used the prediction algorithm PAPA to identify predicted aggregation-prone prion-like domains with a range of compositions. While almost all of the predicted aggregation-prone domains formed foci when expressed in cells, the ability to form the detergent-insoluble aggregates was highly correlated with glutamine/asparagine (Q/N) content, suggesting that high Q/N content may specifically promote conversion to the amyloid state in vivo. We then used this data set to examine cross-seeding between prion-like proteins. The prion protein Sup35 requires the presence of a second prion, [PIN+], to efficiently form prions, but this requirement can be circumvented by the expression of various Q/N-rich protein fragments. Interestingly, almost all of the Q/N-rich domains that formed SDS-insoluble aggregates were able to promote prion formation by Sup35, highlighting the highly promiscuous nature of these interactions.

Amino acid composition predicts prion activity.

Many prion-forming proteins contain glutamine/asparagine (Q/N) rich domains, and there are conflicting opinions as to the role of primary sequence in their conversion to the prion form: is this phenomenon driven primarily by amino acid composition, or, as a recent computational analysis suggested, dependent on the presence of short sequence elements with high amyloid-forming potential. The argument for the importance of short sequence elements hinged on the relatively-high accuracy obtained using a method that utilizes a collection of length-six sequence elements with known amyloid-forming potential. We weigh in on this question and demonstrate that when those sequence elements are permuted, even higher accuracy is obtained; we also propose a novel multiple-instance machine learning method that uses sequence composition alone, and achieves better accuracy than all existing prion prediction approaches. While we expect there to be elements of primary sequence that affect the process, our experiments suggest that sequence composition alone is sufficient for predicting protein sequences that are likely to form prions. A web-server for the proposed method is available at http://faculty.pieas.edu.pk/fayyaz/prank.html, and the code for reproducing our experiments is available at http://doi.org/10.5281/zenodo.167136.

Effects of Mutations on the Aggregation Propensity of the Human Prion-Like Protein hnRNPA2B1.

Hundreds of human proteins contain prion-like domains, which are a subset of low-complexity domains with high amino acid compositional similarity to yeast prion domains. A recently characterized mutation in the prion-like domain of the human heterogeneous nuclear ribonucleoprotein hnRNPA2B1 increases the aggregation propensity of the protein and causes multisystem proteinopathy. The mutant protein forms cytoplasmic inclusions when expressed in Drosophila, the mutation accelerates aggregation in vitro, and the mutant prion-like domain can substitute for a portion of a yeast prion domain in supporting prion activity. To examine the relationship between amino acid sequence and aggregation propensity, we made a diverse set of point mutations in the hnRNPA2B1 prion-like domain. We found that the effects on prion formation in Saccharomyces cerevisiae and aggregation in vitro could be predicted entirely based on amino acid composition. However, composition was an imperfect predictor of inclusion formation in Drosophila; while most mutations showed similar behaviors in yeast, in vitro, and in Drosophila, a few showed anomalous behavior. Collectively, these results demonstrate the significant progress that has been made in predicting the effects of mutations on intrinsic aggregation propensity while also highlighting the challenges of predicting the effects of mutations in more complex organisms.