The rank ordering of genotypic fitness values predicts genetic constraint on natural selection on landscapes lacking sign epistasis

Sewall Wright's genotypic fitness landscape makes explicit one mechanism by which epistasis for fitness can constrain evolution by natural selection. Wright distinguished between landscapes possessing multiple fitness peaks and those with only a single peak and emphasized that the former class imposes substantially greater constraint on natural selection. Here I present novel formalism that more finely partitions the universe of possible fitness landscapes on the basis of the rank ordering of their genotypic fitness values. In this report I focus on fitness landscapes lacking sign epistasis (i.e., landscapes that lack mutations the sign of whose fitness effect varies epistatically), which constitute a subset of Wright's single peaked landscapes. More than one fitness rank ordering lacking sign epistasis exists for L > 2 (where L is the number of interacting loci), and I find that a highly statistically significant effect exists between landscape membership in fitness rank-ordering partition and two different proxies for genetic constraint, even within this subset of landscapes. This statistical association is robust to population size, permitting general inferences about some of the characteristics of fitness rank orderings responsible for genetic constraint on natural selection.     

Evidence for heterogeneous selective pressures in the evolution of the env gene in different human immunodeficiency virus type 1 subtypes

Recent studies have demonstrated the emergence of human immunodeficiency virus type 1 (HIV-1) subtypes with various levels of fitness. Using heterogeneous maximum-likelihood models of adaptive evolution implemented in the PAML software package, with env sequences representing each HIV-1 group M subtype, we examined the various intersubtype selective pressures operating across the env gene. We found heterogeneity of evolutionary mechanisms between the different subtypes with a category of amino acid sites observed that had undergone positive selection for subtypes C, F1, and G, while these sites had undergone purifying selection in all other subtypes. Also, amino acid sites within subtypes A and K that had undergone purifying selection were observed, while these sites had undergone positive selection in all other subtypes. The presence of such sites indicates heterogeneity of selective pressures within HIV-1 group M subtype evolution that may account for the various levels of fitness of the subtypes.     

Pervasive Pairwise Intragenic Epistasis among Sequential Mutations in TEM-1 β-Lactamase

Interactions between mutations play a central role in shaping the fitness landscape, but a clear picture of intragenic epistasis has yet to emerge. To further reveal the prevalence and patterns of intragenic epistasis, we present a survey of epistatic interactions between sequential mutations in TEM-1 β-lactamase. We measured the fitness effect of ~ 12,000 pairs of consecutive amino acid substitutions and used our previous study of the fitness effects of single amino acid substitutions to calculate epistasis for over 8000 mutation pairs. Since sequential mutations are prone to physically interact, we postulated that our study would be surveying specific epistasis instead of nonspecific epistasis. We found widespread negative epistasis, especially in beta-strands, and a high frequency of negative sign epistasis among individually beneficial mutations. Negative epistasis (52%) occurred 7.6 times as frequently as positive epistasis (6.8%). Buried residues experienced more negative epistasis that surface-exposed residues. However, TEM-1 exhibited a couple of hotspots for positive epistasis, most notably L221/ R222 at which many combinations of mutations positively interacted. This study is the first to systematically examine pairwise epistasis throughout an entire protein performing its native function in its native host.     

Evolution in alternating environments with tunable interlandscape correlations

Natural populations are often exposed to temporally varying environments. Evolutionary dynamics in varying environments have been extensively studied, although understanding the effects of varying selection pressures remains challenging. Here, we investigate how cycling between a pair of statistically related fitness landscapes affects the evolved fitness of an asexually reproducing population. We construct pairs of fitness landscapes that share global fitness features but are correlated with one another in a tunable way, resulting in landscape pairs with specific correlations. We find that switching between these landscape pairs, depending on the ruggedness of the landscape and the interlandscape correlation, can either increase or decrease steady‐state fitness relative to evolution in single environments. In addition, we show that switching between rugged landscapes often selects for increased fitness in both landscapes, even in situations where the landscapes themselves are anticorrelated. We demonstrate that positively correlated landscapes often possess a shared maximum in both landscapes that allows the population to step through sub‐optimal local fitness maxima that often trap single landscape evolution trajectories. Finally, we demonstrate that switching between anticorrelated paired landscapes leads to ergodic‐like dynamics where each genotype is populated with nonzero probability, dramatically lowering the steady‐state fitness in comparison to single landscape evolution.     

Perspective: Sign epistasis and genetic constraint on evolutionary trajectories

Epistasis for fitness means that the selective effect of a mutation is conditional on the genetic background in which it appears. Although epistasis is widely observed in nature, our understanding of its consequences for evolution by natural selection remains incomplete. In particular, much attention focuses only on its influence on the instantaneous rate of changes in frequency of selected alleles via epistatic contribution to the additive genetic variance for fitness. Thus, in this framework epistasis only has evolutionary importance if the interacting loci are simultaneously segregating in the population. However, the selective accessibility of mutational trajectories to high fitness genotypes may depend on the genetic background in which novel mutations appear, and this effect is independent of population polymorphism at other loci. Here we explore this second influence of epistasis on evolution by natural selection. We show that it is the consequence of a particular form of epistasis, which we designate sign epistasis. Sign epistasis means that the sign of the fitness effect of a mutation is under epistatic control; thus, such a mutation is beneficial on some genetic backgrounds and deleterious on others. Recent experimental innovations in microbial systems now permit assessment of the fitness effects of individual mutations on multiple genetic backgrounds. We review this literature and identify many examples of sign epistasis, and we suggest that the implications of these results may generalize to other organisms. These theoretical and empirical considerations imply that strong genetic constraint on the selective accessibility of trajectories to high fitness genotypes may exist and suggest specific areas of investigation for future research.     

Detecting epistasis from an ensemble of adapting populations

The role that epistasis plays during adaptation remains an outstanding problem, which has received considerable attention in recent years. Most of the recent empirical studies are based on ensembles of replicate populations that adapt in a fixed, laboratory controlled condition. Researchers often seek to infer the presence and form of epistasis in the fitness landscape from the time evolution of various statistics averaged across the ensemble of populations. Here, we provide a rigorous analysis of what quantities, drawn from time series of such ensembles, can be used to infer epistasis for populations evolving under weak mutation on finite‐site fitness landscapes. First, we analyze the mean fitness trajectory—that is, the time course of the ensemble average fitness. We show that for any epistatic fitness landscape and starting genotype, there always exists a non‐epistatic fitness landscape that produces the exact same mean fitness trajectory. Thus, the presence of epistasis is not identifiable from the mean fitness trajectory. By contrast, we show that two other ensemble statistics—the time evolution of the fitness variance across populations, and the time evolution of the mean number of substitutions—can detect certain forms of epistasis in the underlying fitness landscape.     

Extracting characteristic properties of fitness landscape from in vitro molecular evolution: a case study on infectivity of fd phage to E.coli

We have developed a methodology for extracting characteristic properties of a fitness landscape of interest by analyzing fitness data on an in vitro molecular evolution. The in vitro evolution is required to be conducted as the following "adaptive walk": a single parent sequence generates N mutant sequences as its offsprings, and the fittest individual among the N offsprings will become a new parent in the next generation. N is the library size of mutants to be screened in a single generation. Our theory of the adaptive walk on the "NK landscape" suggests the following: the adaptive walker starting from a random sequence climbs the landscape easily in an early stage, and then reaches a stationary phase in which the mutation-selection-random drift balance sets in. The stationary fitness value is nearly proportional to square root of ln N. Our analysis is performed from the following points: (1) stationary fitness values, (2) time series of fitness in the transitional state, (3) mutant's fitness distribution, and (4) the strength of selection pressure. Applying our methodology, we analyzed experimental data on the in vitro evolution of a random polypeptide (139 amino acids) toward acquiring infectivity (= ability to infect) of fd phage. As a result, we estimated that k is about 27 in this system, indicating that an arbitrary residue in a sequence is affected from other 23% residues. In this article, we demonstrated that the experimental data is consistent with our theoretical equations quantitatively, and that our methodology for extracting characteristic properties of a fitness landscape may be effective.     

Adaptive evolution of butterfly wing shape: from morphology to behaviour

Butterflies display extreme variation in wing shape associated with tremendous ecological diversity. Disentangling the role of neutral versus adaptive processes in wing shape diversification remains a challenge for evolutionary biologists. Ascertaining how natural selection influences wing shape evolution requires both functional studies linking morphology to flight performance, and ecological investigations linking performance in the wild with fitness. However, direct links between morphological variation and fitness have rarely been established. The functional morphology of butterfly flight has been investigated but selective forces acting on flight behaviour and associated wing shape have received less attention. Here, we attempt to estimate the ecological relevance of morpho‐functional links established through biomechanical studies in order to understand the evolution of butterfly wing morphology. We survey the evidence for natural and sexual selection driving wing shape evolution in butterflies, and discuss how our functional knowledge may allow identification of the selective forces involved, at both the macro‐ and micro‐evolutionary scales. Our review shows that although correlations between wing shape variation and ecological factors have been established at the macro‐evolutionary level, the underlying selective pressures often remain unclear. We identify the need to investigate flight behaviour in relevant ecological contexts to detect variation in fitness‐related traits. Identifying the selective regime then should guide experimental studies towards the relevant estimates of flight performance. Habitat, predators and sex‐specific behaviours are likely to be major selective forces acting on wing shape evolution in butterflies. Some striking cases of morphological divergence driven by contrasting ecology involve both wing and body morphology, indicating that their interactions should be included in future studies investigating co‐evolution between morphology and flight behaviour.     

The impact of epistatic selection on the genomic traces of selection

The rapid accumulation of genomic data has led to an explosion of studies searching for signals of past selection left within DNA sequences. Yet the majority of theoretical studies investigating the traces of selection have assumed a simple form of selection, without interactions among selectively fixed sites. Fitness interactions—‘epistasis’—are commonplace, however, and take on a myriad of forms ( Whitlock et al. 1995 ; Segrèet al. 2005 ; Phillips 2008 ). It is thus important to determine how such epistasis would influence selective sweeps. On p. 5018 of this issue, Takahasi (2009) explores the effect of epistasis on genetic variation neighbouring two sites that interact in determining fitness, finding that such epistasis has a dramatic impact on the genetic variability in regions surrounding the interacting sites.     

Amino Acid Covariation in a Functionally Important Human Immunodeficiency Virus Type 1 Protein Region Is Associated With Population Subdivision

The frequently reported amino acid covariation of the highly polymorphic human immunodeficiency virus type 1 (HIV-1) exterior envelope glycoprotein V3 region has been assumed to reflect fitness epistasis between residues. However, nonrandom association of amino acids, or linkage disequilibrium, has many possible causes, including population subdivision. If the amino acids at a set of sequence sites differ in frequencies between subpopulations, then analysis of the whole population may reveal linkage disequilibrium even if it does not exist in any subpopulation. HIV-1 has a complex population structure, and the effects of this structure on linkage disequilibrium were investigated by estimating within- and among-subpopulation components of variance in linkage disequilibrium. The amino acid covariation previously reported is explained by differences in amino acid frequencies among virus subpopulations in different patients and by nonsystematic disequilibrium among patients. Disequilibrium within patients appears to be entirely due to differences in amino acid frequencies among sampling time points and among chemokine coreceptor usage phenotypes of virus particles, but not source tissues. Positive selection explains differences in allele frequencies among time points and phenotypes, indicating that these differences are adaptive rather than due to genetic drift. However, the absence of a correlation between linkage disequilibrium and phenotype suggests that fitness epistasis is an unlikely cause of disequilibrium. Indeed, when population structure is removed by analyzing sequences from a single time point and phenotype, no disequilibrium is detectable within patients. These results caution against interpreting amino acid covariation and coevolution as evidence for fitness epistasis.     

WHY EPISTASIS IS IMPORTANT FOR SELECTION AND ADAPTATION

Organisms are built from thousands of genes that interact in complex ways. Still, the mathematical theory of evolution is dominated by a gene‐by‐gene perspective in which genes are assumed to have the same effects regardless of genetic background. Gene interaction, or epistasis, plays a role in some theoretical developments such as the evolution of recombination, reproductive isolation, and canalization, but is strikingly missing from our standard accounts of phenotypic adaptation. This absence is most puzzling within the field of quantitative genetics, which, despite its polygenic perspective and elaborate statistical representation of epistasis, has not found a single important role for gene interaction in evolution. To the contrary, there is a widespread consensus that epistasis is evolutionary inert, and that all we need to know to predict evolutionary dynamics is the additive component of the genetic variance. This view may have roots in convenience, but also in theoretical results showing that the response to selection derived from epistatic variance components is not permanent and will decay when selection is relaxed. I show that these results are tied to a conceptual confusion, and are misleading as general statements about the significance of epistasis for the selection response and adaptation.     

MULTIDIMENSIONAL ADAPTIVE EVOLUTION OF A FEED‐FORWARD NETWORK AND THE ILLUSION OF COMPENSATION

When multiple substitutions affect a trait in opposing ways, they are often assumed to be compensatory, not only with respect to the trait, but also with respect to fitness. This type of compensatory evolution has been suggested to underlie the evolution of protein structures and interactions, RNA secondary structures, and gene regulatory modules and networks. The possibility for compensatory evolution results from epistasis. Yet if epistasis is widespread, then it is also possible that the opposing substitutions are individually adaptive. I term this possibility an adaptive reversal. Although possible for arbitrary phenotype‐fitness mappings, it has not yet been investigated whether such epistasis is prevalent in a biologically realistic setting. I investigate a particular regulatory circuit, the type I coherent feed‐forward loop, which is ubiquitous in natural systems and is accurately described by a simple mathematical model. I show that such reversals are common during adaptive evolution, can result solely from the topology of the fitness landscape, and can occur even when adaptation follows a modest environmental change and the network was well adapted to the original environment. The possibility of adaptive reversals warrants a systems perspective when interpreting substitution patterns in gene regulatory networks.     

“Cancer Reviews 2021” series: Cell competition in intratumoral and tumor microenvironment interactions

Tumors are complex cellular and acellular environments within which cancer clones are under continuous selection pressures. Cancer cells are in a permanent mode of interaction and competition with each other as well as with the immediate microenvironment. In the course of these competitive interactions, cells share information regarding their general state of fitness, with less‐fit cells being typically eliminated via apoptosis at the hands of those cells with greater cellular fitness. Competitive interactions involving exchange of cell fitness information have implications for tumor growth, metastasis, and therapy outcomes. Recent research has highlighted sophisticated pathways such as Flower, Hippo, Myc, and p53 signaling, which are employed by cancer cells and the surrounding microenvironment cells to achieve their evolutionary goals by means of cell competition mechanisms. In this review, we discuss these recent findings and explain their importance and role in evolution, growth, and treatment of cancer. We further consider potential physiological conditions, such as hypoxia and chemotherapy, that can function as selective pressures under which cell competition mechanisms may evolve differently or synergistically to confer oncogenic advantages to cancer.     

A population genetics-phylogenetics approach to inferring natural selection in coding sequences

Through an analysis of polymorphism within and divergence between species, we can hope to learn about the distribution of selective effects of mutations in the genome, changes in the fitness landscape that occur over time, and the location of sites involved in key adaptations that distinguish modern-day species. We introduce a novel method for the analysis of variation in selection pressures within and between species, spatially along the genome and temporally between lineages. We model codon evolution explicitly using a joint population genetics-phylogenetics approach that we developed for the construction of multiallelic models with mutation, selection, and drift. Our approach has the advantage of performing direct inference on coding sequences, inferring ancestral states probabilistically, utilizing allele frequency information, and generalizing to multiple species. We use a Bayesian sliding window model for intragenic variation in selection coefficients that efficiently combines information across sites and captures spatial clustering within the genome. To demonstrate the utility of the method, we infer selective pressures acting in Drosophila melanogaster and D. simulans from polymorphism and divergence data for 100 X-linked coding regions.     

Strong Epistatic Selection on the RNA Secondary Structure of HIV

A key question in evolutionary genomics is how populations navigate the adaptive landscape in the presence of epistasis, or interactions among loci. This problem can be directly addressed by studying the evolution of RNA secondary structures, for which there is constraint to maintain pairing between Watson-Crick (WC) sites. Replacement of a nucleotide at one site of a WC pair reduces fitness by disrupting binding, which can be restored via a compensatory replacement at the interacting site. Here, I present the first genome-scale analysis of epistasis on the RNA secondary structure of human immunodeficiency virus type 1 (HIV-1). Comparison of polymorphism frequencies at ancestrally conserved sites reveals that selection against replacements is ∼2.7 times stronger at WC than at non-WC sites, such that nearly 50% of constraint can be attributed to epistasis. However, almost all epistatic constraint is due to selection against conversions of WC pairs to unpaired (UP) nucleotides, whereas conversions to GU wobbles are only slightly deleterious. This disparity is also evident in pairs with second-site compensatory replacements; conversions from UP nucleotides to WC pairs increase median fitness by ∼4.2%, whereas conversions from GU wobbles to WC pairs only increase median fitness by ∼0.3%. Moreover, second-site replacements that convert UP nucleotides to GU wobbles also increase median fitness by ∼4%, indicating that such replacements are nearly as compensatory as those that restore WC pairing. Thus, WC peaks of the HIV-1 epistatic adaptive landscape are connected by high GU ridges, enabling the viral population to rapidly explore distant peaks without traversing deep UP valleys.     

Fitness landscapes emerging from pharmacodynamic functions in the evolution of multidrug resistance

Adaptive evolution often involves beneficial mutations at more than one locus. In this case, the trajectory and rate of adaptation is determined by the underlying fitness landscape, that is, the fitness values and mutational connectivity of all genotypes under consideration. Drug resistance, especially resistance to multiple drugs simultaneously, is also often conferred by mutations at several loci so that the concept of fitness landscapes becomes important. However, fitness landscapes underlying drug resistance are not static but dependent on drug concentrations, which means they are influenced by the pharmacodynamics of the drugs administered. Here, I present a mathematical framework for fitness landscapes of multidrug resistance based on Hill functions describing how drug concentrations affect fitness. I demonstrate that these ‘pharmacodynamic fitness landscapes’ are characterized by pervasive epistasis that arises through (i) fitness costs of resistance (even when these costs are additive), (ii) nonspecificity of resistance mutations to drugs, in particular cross‐resistance, and (iii) drug interactions (both synergistic and antagonistic). In the latter case, reciprocal drug suppression may even lead to reciprocal sign epistasis, so that the doubly resistant genotype occupies a local fitness peak that may be difficult to access by evolution. Simulations exploring the evolutionary dynamics on some pharmacodynamic fitness landscapes with both constant and changing drug concentrations confirm the crucial role of epistasis in determining the rate of multidrug resistance evolution.     

Epistasis and intramolecular networks in protein evolution

Proteins are molecular machines composed of complex, highly connected amino acid networks. Their functional optimization requires the reorganization of these intramolecular networks by evolution. In this review, we discuss the mechanisms by which epistasis, that is, the dependence of the effect of a mutation on the genetic background, rewires intramolecular interactions to alter protein function. Deciphering the biophysical basis of epistasis is crucial to our understanding of evolutionary dynamics and the elucidation of sequence-structure-function relationships. We featured recent studies that provide insights into the molecular mechanisms giving rise to epistasis, particularly at the structural level. These studies illustrate the convoluted and fascinating nature of the intramolecular networks co-opted by epistasis during the evolution of protein function.     

Coordinated Evolution of Influenza A Surface Proteins

The surface proteins hemagglutinin (HA) and neuraminidase (NA) of human influenza A virus evolve under selection pressures to escape adaptive immune responses and antiviral drug treatments. In addition to these external selection pressures, some mutations in HA are known to affect the adaptive landscape of NA, and vice versa, because these two proteins are physiologically interlinked. However, the extent to which evolution of one protein affects the evolution of the other one is unknown. Here we develop a novel phylogenetic method for detecting the signatures of such genetic interactions between mutations in different genes – that is, inter-gene epistasis. Using this method, we show that influenza surface proteins evolve in a coordinated way, with mutations in HA affecting subsequent spread of mutations in NA and vice versa, at many sites. Of particular interest is our finding that the oseltamivir-resistance mutations in NA in subtype H1N1 were likely facilitated by prior mutations in HA. Our results illustrate that the adaptive landscape of a viral protein is remarkably sensitive to its genomic context and, more generally, that the evolution of any single protein must be understood within the context of the entire evolving genome.     

Visualizing fitness landscapes

Fitness landscapes are a classical concept for thinking about the relationship between genotype and fitness. However, because the space of genotypes is typically high-dimensional, the structure of fitness landscapes can be difficult to understand and the heuristic approach of thinking about fitness landscapes as low-dimensional, continuous surfaces may be misleading. Here, I present a rigorous method for creating low-dimensional representations of fitness landscapes. The basic idea is to plot the genotypes in a manner that reflects the ease or difficulty of evolving from one genotype to another. Such a layout can be constructed using the eigenvectors of the transition matrix describing the evolution of a population on the fitness landscape when mutation is weak. In addition, the eigendecomposition of this transition matrix provides a new, high-level view of evolution on a fitness landscape. I demonstrate these techniques by visualizing the fitness landscape for selection for the amino acid serine and by visualizing a neutral network derived from the RNA secondary structure genotype-phenotype map.