Selection on growth rate and local adaptation drive genomic adaptation during experimental range expansions in the protist Tetrahymena thermophila

Abstract

Populations that expand their range can undergo rapid evolutionary adaptation of life-history traits, dispersal behaviour and adaptation to the local environment. Such adaptation may be aided or hindered by sexual reproduction, depending on the context. However, few empirical and experimental studies have investigated the genetic basis of adaptive evolution during range expansions. Even less attention has been given to the question how sexual reproduction may modulate such adaptive evolution during range expansions. We here studied genomic adaptation during experimental range expansions of the protist Tetrahymena thermophila in landscapes with a uniform environment or a pH gradient. Specifically, we investigated two aspects of genomic adaptation during range expansion. First, we investigated adaptive genetic change in terms of the underlying numbers of allele frequency changes from standing genetic variation and de novo variants. We focused on how sexual reproduction may alter this adaptive genetic change. Second, we identified genes subject to selection caused by the expanding range itself, and directional selection due to the presence or absence of the pH gradient. We focused this analysis on alleles with large frequency changes that occurred in parallel in more than one population to identify the most likely candidate targets of selection. We found that sexual reproduction altered adaptive genetic change both in terms of de novo variants and standing genetic variation. However, sexual reproduction affected allele frequency changes in standing genetic variation only in the absence of long-distance gene flow. Adaptation to the range expansion affected genes involved in cell divisions and DNA repair, whereas adaptation to the pH gradient additionally affected genes involved in ion balance and oxidoreductase reactions. These genetic changes may result from selection on growth and adaptation to low pH. In the absence of gene flow, sexual reproduction may have aided genetic adaptation. Gene flow may have swamped expanding populations with maladapted alleles, thus reducing the extent of evolutionary adaptation during range expansion. Sexual reproduction also altered the genetic basis of adaptation in our evolving populations via de novo variants, possibly by purging deleterious mutations or by revealing fitness benefits of rare genetic variants.

Keywords: Tetrahymena; gene ontology; life history; pH; range expansions; whole genome resequencing.

FIGURE 1

Two‐patch landscapes. i.1: Two‐patch landscapes consist of connected tubes. i.2: To initiate dispersal, the connection is opened for 1 hour, allowing cells to actively swim from the home patch to the target patch. i.3: If we found no cells in the target patch (unsuccessful dispersal), we transferred the contents of the home patch to a new two‐patch landscape. i.4: If we did detect cells in the target patch (successful dispersal), we transferred the content of the target patch to a new two‐patch landscape [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 2

Gene swamping affects fitness changes during range expansion. Fitness change calculated as the ratio of the logarithm (base 2) of the intrinsic population growth rate ($r_0$) of an evolved population and the intrinsic population growth rate ($r_0$) of the ancestral population. Circles represent data for individual populations. Black lines and shaded areas represent mean predictions and 95% confidence intervals of the best model. Colours represent mode of reproduction (yellow = asexual, blue = sexual). Left subplots (panels a and c) show data and model predictions for populations expanding into a uniform environment, right subplots (panels b and d) for populations expanding into a pH gradient. Upper subplots (panels a and b) show data and model predictions for populations where gene flow is absent, and lower subplots (panels c and d) for populations where gene flow is present. Adapted from Moerman, Fronhofer, et al. (2020b) [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 3

De novo variants increase to higher frequency in sexual populations. The figure shows the number of novel variants for the evolved populations expanding into a uniform environment (panels a and c) or into a pH gradient (panels b and d). The x‐axis shows the cut‐off value in minimum allele frequency. The y‐axis shows the number of de novo variants whose frequency is above a given cut‐off. Circles represent the data, with thin lines connecting data from the same replicate population. The thicker opaque lines and shaded areas show the predicted means and 95% confidence intervals for the best model. Colours represent the mode of reproduction [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 4

Reproduction alters genetic changes in standing genetic variation when gene flow is absent. Number of alleles that changed significantly in frequency during range expansion into a uniform environment (panels a and c) or into a pH gradient (panels b and d). The x‐axis shows the cut‐off value in the magnitude of allele frequency. The y‐axis shows the number of variants analysed at each cut‐off value. Circles represent allele frequency data, with thin lines connecting data from the same replicate population. Thick opaque lines show the model predictions for the best model. Shaded areas show 95%‐confidence intervals from the posterior distribution. Colours represent mode of reproduction (yellow: asexual, blue:sexual) [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 5

A positive correlation between fitness change and genetic change when sexual populations expand into a uniform environment. Statistical association between change in fitness and the number of allele frequency changes in standing genetic variation, for populations expanding into a uniform environment (left panels) and populations expanding into a pH gradient (right panels). The x‐axis shows the number of standing genetic variants that changed significantly in allele frequency. The y‐axis shows the change in growth rate (fitness) of evolved populations compared to the ancestral population. Horizontal subplots show the data for different cut‐off values used for the minimum change in allele frequency at the end of evolution for inclusion in the analysis. Each geometric symbol within a subplot represents a single population, with colour representing reproductive mode (yellow = asexual, blue = sexual), and shape representing gene flow (circle = gene flow absent, triangle = gene flow present). Text insets show the Pearson correlation coefficient r and significance p, based on a paired sample correlation test for sexual populations (blue text) and asexual populations (yellow text). Non‐significant correlations are shown in light colour (yellow or blue), whereas significant correlations are shown in dark colour. For the most stringent cut‐off value (allele frequency change$>0.8$), we did not calculate the correlation, because several populations did not harbour any variants that changed so dramatically in allele frequency [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 6

General and gradient‐specific adaptations during range expansion. Partitioning of all enriched gene ontology terms (i.e. terms describing biological or molecular function of genes, as well as cellular structure) among 13 major functional categories for genes involved in general adaptations (panel a) and gradient‐specific adaptations (panel b). Colours correspond to the 13 major GO categories we identified. Numbers in each rectangle represent the number and percentage of genes whose annotation fell in each category. Differentially enriched categories are displayed as filled rectangles and are boxed in the colour legend. The remaining categories are displayed as open rectangles with a coloured border [Colour figure can be viewed at wileyonlinelibrary.com]