Resonant fluctuations of selection pressure exponentially accelerate fitness valley crossing

Two functional protein sequences can sometimes be separated by a fitness valley - a series of low or non-functional intermediate mutations that must be traversed to reach a more optimal or refined function. Time-varying selection pressure modulates evolutionary sampling of such valleys. Yet, how the amplitude and frequency of fluctuating selection influence the rate of protein evolution is poorly understood. Here, we derive a simple equation for the time-dependent probability of crossing a fitness valley as a function of evolutionary parameters: valley width, protein size, mutation rate, and selection pressure. The equation predicts that, under low selection pressure, the valley crossing rate is magnified by a factor that depends exponentially on valley width. However, after a characteristic time set by the evolutionary parameters, the rate rapidly decays. Thus, there is an optimal frequency of selection-pressure fluctuations that maximizes the rate of protein optimization. This result is reminiscent of the resonance frequency in mechanical systems. The equation unites empirical and theoretical results that were previously disconnected, and is consistent with time-dependent in vitro and clinical data. More generally, these results suggest that seasonal and climate oscillations could synchronously drive protein evolution at the resonant frequency across a range of organism hosts and timescales. This theory could also be applied to optimize de novo protein evolution in laboratory directed evolution using time-varying protocols.

Modeling the trajectory of SARS-CoV-2 spike protein evolution in continuous latent space using a neural network and Gaussian process

Viral vaccines can lose their efficacy as the genomes of targeted viruses rapidly evolve, resulting in new variants that may evade vaccine-induced immunity. This process is apparent in the emergence of new SARS-CoV-2 variants which have the potential to undermine vaccination efforts and cause further outbreaks. Predictive vaccinology points to a future of pandemic preparedness in which vaccines can be developed preemptively based in part on predictive models of viral evolution. Thus, modeling the trajectory of SARS-CoV-2 spike protein evolution could have value for mRNA vaccine development. Traditionally, in silico sequence evolution has been modeled discretely, while there has been limited investigation into continuous models. Here we present the Viral Predictor for mRNA Evolution (VPRE), an open-source software tool which learns from mutational patterns in viral proteins and models their most statistically likely evolutionary trajectories. We trained a variational autoencoder with real-time and simulated SARS-CoV-2 genome data from Australia to encode discrete spike protein sequences into continuous numerical variables. To simulate evolution along a phylogenetic path, we trained a Gaussian process model with the numerical variables to project spike protein evolution up to five months in advance. Our predictions mapped primarily to a sequence that differed by a single amino acid from the most reported spike protein in Australia within the prediction timeframe, indicating the utility of deep learning and continuous latent spaces for modeling viral protein evolution. VPRE can be readily adapted to investigate and predict the evolution of viruses other than SARS-CoV-2 in temporal, geographic, and lineage-specific pathways.

Unconventional conservation reveals structure-function relationships in the synaptonemal complex

Functional requirements constrain protein evolution, commonly manifesting in a conserved amino acid sequence. Here, we extend this idea to secondary structural features by tracking their conservation in essential meiotic proteins with highly diverged sequences. The synaptonemal complex (SC) is a ~100-nm-wide ladder-like meiotic structure present in all eukaryotic clades, where it aligns parental chromosomes and regulates exchanges between them. Despite the conserved ultrastructure and functions of the SC, SC proteins are highly divergent within Caenorhabditis. However, SC proteins have highly conserved length and coiled-coil domain structure. We found the same unconventional conservation signature in Drosophila and mammals, and used it to identify a novel SC protein in Pristionchus pacificus, Ppa-SYP-1. Our work suggests that coiled-coils play wide-ranging roles in the structure and function of the SC, and more broadly, that expanding sequence analysis beyond measures of per-site similarity can enhance our understanding of protein evolution and function.

Structure and stability of the designer protein WRAP-T and its permutants

Abstract$$\beta $$ β -Propeller proteins are common natural disc-like pseudo-symmetric proteins that contain multiple repeats (‘blades’) each consisting of a 4-stranded anti-parallel $$\beta $$ β -sheet. So far, 4- to 12-bladed $$\beta $$ β -propellers have been discovered in nature showing large functional and sequential variation. Using computational design approaches, we created perfectly symmetric $$\beta $$ β -propellers out of natural pseudo-symmetric templates. These proteins are useful tools to study protein evolution of this very diverse fold. While the 7-bladed architecture is the most common, no symmetric 7-bladed monomer has been created and characterized so far. Here we describe such a engineered protein, based on a highly symmetric natural template, and test the effects of circular permutation on its stability. Geometrical analysis of this protein and other artificial symmetrical proteins reveals no systematic constraint that could be used to help in engineering of this fold, and suggests sequence constraints unique to each $$\beta $$ β -propeller sub-family.

Frequent lineage-specific substitution rate changes support an episodic model for protein evolution

Since the inception of the molecular clock model for sequence evolution, the investigation of protein divergence has revolved around the question of a more or less constant change of amino acid sequences, with specific overall rates for each family. Although anomalies in clock-like divergence are well known, the assumption of a constant decay rate for a given protein family is usually taken as the null model for protein evolution. However, systematic tests of this null model at a genome-wide scale have lagged behind, despite the databases’ enormous growth. We focus here on divergence rate comparisons between very closely related lineages since this allows clear orthology assignments by synteny and reliable alignments, which are crucial for determining substitution rate changes. We generated a high-confidence dataset of syntenic orthologs from four ape species, including humans. We find that despite the appearance of an overall clock-like substitution pattern, several hundred protein families show lineage-specific acceleration and deceleration in divergence rates, or combinations of both in different lineages. Hence, our analysis uncovers a rather dynamic history of substitution rate changes, even between these closely related lineages, implying that one should expect that a large fraction of proteins will have had a history of episodic rate changes in deeper phylogenies. Furthermore, each of the lineages has a separate set of particularly fast diverging proteins. The genes with the highest percentage of branch-specific substitutions are ADCYAP1 in the human lineage (9.7%), CALU in chimpanzees (7.1%), SLC39A14 in the internal branch leading to humans and chimpanzees (4.1%), RNF128 in gorillas (9%), and S100Z in gibbons (15.2%). The mutational pattern in ADCYAP1 suggests a biased mutation process, possibly through asymmetric gene conversion effects. We conclude that a null model of constant change can be problematic for predicting the evolutionary trajectories of individual proteins.

Sequencing refractory regions in bird genomes are hotspots for accelerated protein evolution

Background Approximately 1000 protein encoding genes common for vertebrates are still unannotated in avian genomes. Are these genes evolutionary lost or are they not yet found for technical reasons? Using genome landscapes as a tool to visualize large-scale regional effects of genome evolution, we reexamined this question. Results On basis of gene annotation in non-avian vertebrate genomes, we established a list of 15,135 common vertebrate genes. Of these, 1026 were not found in any of eight examined bird genomes. Visualizing regional genome effects by our sliding window approach showed that the majority of these "missing" genes can be clustered to 14 regions of the human reference genome. In these clusters, an additional 1517 genes (often gene fragments) were underrepresented in bird genomes. The clusters of “missing” genes coincided with regions of very high GC content, particularly in avian genomes, making them “hidden” because of incomplete sequencing. Moreover, proteins encoded by genes in these sequencing refractory regions showed signs of accelerated protein evolution. As a proof of principle for this idea we experimentally characterized the mRNA and protein products of four "hidden" bird genes that are crucial for energy homeostasis in skeletal muscle: ALDOA, ENO3, PYGM and SLC2A4. Conclusions A least part of the “missing” genes in bird genomes can be attributed to an artifact caused by the difficulty to sequence regions with extreme GC% (“hidden” genes). Biologically, these “hidden” genes are of interest as they encode proteins that evolve more rapidly than the genome wide average. Finally we show that four of these “hidden” genes encode key proteins for energy metabolism in flight muscle.

Convergent patterns of gene expression and protein evolution associated with adaptation to desert environments in rodents

Desert specialization has arisen multiple times across rodents and is often associated with a suite of convergent phenotypes, including modification of the kidneys to mitigate water loss. However, the extent to which phenotypic convergence in desert rodents is mirrored at the molecular level is unknown. Here, we sequenced kidney mRNA and assembled transcriptomes for three pairs of rodent species to search for convergence in gene expression and amino acid sequence associated with adaptation to deserts. We conducted phylogenetically-independent comparisons between a desert specialist and a non-desert relative in three families representing ∼70 million years of evolution. Overall, patterns of gene expression faithfully recapitulated the phylogeny of these six taxa. However, we found that 8.6% of all genes showed convergent patterns of expression evolution between desert and non-desert taxa, a proportion that is much higher than expected by chance. In addition to these convergent changes, we observed many species-pair specific changes in gene expression indicating that different instances of adaptation to deserts include a combination of unique and shared changes. Patterns of protein evolution revealed a small number of genes showing evidence of positive selection, the majority of which did not show convergent changes in gene expression. Overall, our results suggest convergent changes in gene regulation play a primary role in the complex trait of desert adaptation in rodents.

Ancestral protein topologies draw the rooted bacterial tree of life

Many experimental and predicted observations remain unanswered in the current proposed trees of life (ToL). Also, the current trend in reporting phylogenetic data is based in mixing together the information of dozens of genomes or entire conserved proteins. In this work, we consider the modularity of protein evolution and, using only two domains with duplicated ancestral topologies from a single, universal primordial protein corresponding to the RNA binding regions of contemporary bacterial glycyl tRNA synthetase (bacGlyRS), archaeal CCA adding enzyme (arch-CCAadd) and eukaryotic rRNA processing enzyme (euk-rRNA), we propose a rooted bacterial ToL that agrees with several previous observations unaccounted by the available trees.