Local adaptation in maize and teosintes
Strictly defined, a genotype can be considered locally adapted if it has a higher fitness at its native site than any other non-native genotypes (Kawecki and Ebert 2004). Locally adapted alleles can be either neutral or deleterious in other environments. Two models depict those situations, namely conditional neutrality and antagonistic pleiotropy (Anderson et al. 2013).
Despite numerous studies, the genetic processes underlying local adaptation in natural populations are still poorly understood. Studies showed that highlands maize landraces outperform lowland maize populations in their native environment but perform worse than any other population at lower elevation sites (Mercer, Martínez-Vásquez, and Perales 2008), suggesting strong adaptation for high altitude. Interestingly, an ancient DNA study shows that, by 4000 years ago, maize was already largely cultivated in the lowlands of southwestern United States but the adaptation to the highland of Colorado Plateau took an extra 2000 years. This delay is probably the result of a long time to adapt to local conditions (Swarts et al. 2017).
Natural selection acts on phenotypic traits, changing the frequency of underlying alleles and shifting population phenotypes toward local optima. Since these optima rely on local conditions, genes ecologically important usually differ between sub-populations in heterogeneous environments, which results in divergence in allele frequencies over time. This characteristic has been utilized in genome scans to mine correlations between allele frequencies and environmental variables (Fig. 1A). Such studies have revealed that, in teosintes, loci associated with environmental variables impact flowering time and adaptation to soil composition (Aguirre-Liguori et al. 2017; Fustier et al. 2017; Pyhäjärvi et al. 2013). Flowering time was also a key component of maize’s local adaptation to higher latitudes during post-domestication. Maize evolved a reduced sensitivity to photoperiod, in part due to a CACTA-like TE insertion in the promoter region of the ZmCCT gene that drives photoperiod response in early flowering maizes (Hung et al. 2012; Yang et al. 2013). An example of adaptation driven by soil interactions is the tolerance of maize and teosinte to aluminum in highly acidic soils. In these lines, the adaptation is linked to tandem duplications of the MATE1 gene involved in the extrusion of toxic compounds (Maron et al. 2013).
Numerous other biotic and abiotic factors are likely involved in adaptation in maize and teosinte, including predation, parasitism, moisture and herbicide (Linhart and Grant 1996; Valverde 2007). For example, a study on parviglumis has shown that in response to herbivory, immunity genes involved in the inhibition of insect’ digestive proteases experienced a recent selective sweep in a region of Mexico, probably reflecting their local adaptation (Moeller and Tiffin 2008). These measures are however more difficult to gather making their study less common.
Interestingly, three large inversion polymorphisms seem to play an important role in local adaptation. Among them, a 50Mb inversion on chromosome 1 is found at high frequency in parviglumis (20-90%), low frequency in mexicana (10%) and is absent in maize. This inversion is highly correlated with altitude and significantly associated with temperature and precipitation (Fang et al. 2012; Pyhäjärvi et al. 2013). Inversions on chromosomes 4 and 9 also displayed environmental association in teosintes (Pyhäjärvi et al. 2013). Local adaptation to different habitats or niches is a gradual process that can promote divergence and, in the long run, ecological speciation (Schluter 2009). Genotyping of a broad sample of 49 populations covering the entire geographic range of teosintes has recently provided some evidence of this. Aguirre-Liguori et al. 2017 showed that both within parviglumis and mexicana, populations distributed at the edge of the ecological niche, but not the range edge, experience stronger local adaptation. This suggests that local adaptation may have contributed to divergence between these two subspecies.
What is the role of phenotypic plasticity?
Phenotypic plasticity is defined as the capacity of a genotype to produce a range of expressed phenotypes in distinct environments. This is achieved through differential developmental pathways in response to changing conditions (Beldade, Mateus, and Keller 2011; Gilbert and Epel 2009). Studies have shown that plasticity is an important process for the evolution of novel traits during adaptation. Indeed, populations with flexible phenotypes are predicted to better cope with environmental changes, to colonize broader niches, and to display a greater potential for expansion (Wennersten and Forsman 2012). This process is particularly important for plants as they are fixed in a specific location and not sheltered from the environment (Des Marais, Hernandez, and Juenger 2013b).
When the environment changes, the phenotypic optimum of a population is likely altered as well. As a result, individuals that show a plastic response in the direction of the new optimum, will have a fitness advantage. In contrast, individuals that exhibit no plasticity or that produce phenotypes too far from this optimum, will be selected against.
However, plasticity has some limits and may entail a fitness cost. For instance, compared to developmentally fixed phenotypes, plastic individuals in constant environments may display lower fitness or produce a less adapted phenotype. Possible reasons include sensory mechanisms that have a high energetic cost, the epistatic effects of regulatory genes involved in the plastic response, lag time between the perception and the phenotypic response and genetic correlations among traits (Auld, Agrawal, and Relyea 2010; DeWitt, Sih, and Wilson 1998; Nicotra et al. 2010).
Phenotypic plasticity is difficult to study as it arises from genetic and environmental interactions which are often hard to disentangle. Moreover, phenotypic plasticity is fundamentally intertwined with genetic adaptation, furthering the difficulty of determining the causality of a phenotype. The difference between genetic adaptation and phenotypic plasticity is that for the former the phenotype is genetically determined, whereas plastic phenotypes plasticity may be heritable, the phenotype is largely determined by the environment (Kawecki and Ebert 2004; van Kleunen and Fischer 2005). In addition, plasticity can be lost, and the phenotype constitutively expressed by the fixation of genetic variation after a number of generations of constant selection, a process called genetic assimilation (Diggle and Miller 2013; Kuzawa and Bragg 2012; Standen, Du, and Larsson 2014). Hence an initially plastic phenotype may become a genetic adaptation after genetic assimilation. Some examples of plastic responses are well documented in plants, for example, the response to vernalization in Arabidopsis regulating flowering time in some ecotypes (Nicotra et al. 2010). Another example is the change in seed dormancy in response to the environment which prevents germination when conditions are unlikely to lead to the survival of the plant (Nicotra et al. 2010).
Mechanisms of genetic adaptation in maize and teosintes
Populations of teosinte have long evolved under natural selection. In contrast, maize populations have been under artificial human selection that moved phenotypes towards optimal traits tailored to agriculture during a shorter time frame of ~9,000 years (Piperno and Flannery 2001; Matsuoka et al. 2002; Fukunaga et al. 2005). These time scales have left distinct genetic signatures. In theory, traits fixed by domestication should involve genes with larger effect sizes, and standing variation should be a major contributor to domestication (Wallace, Larsson, and Buckler 2014). This is supported by crosses between maize and teosinte that led to the discovery of six main QTLs responsible for major phenotypic differences between them, notably vegetative architecture and inflorescence sexuality (Beadle 1972; Briggs et al. 2007). Among these QTLs, genes with major phenotypic effects have been discovered such as tb1 and tga1 (teosinte glume architecture1). In addition to these major genes, a collection of targets (2 to 4% of the genome according to Wright et al. 2005 and Hufford, Xu, et al. 2012) have likely contributed to the domesticated phenotype. In contrast, Genome Wide Association (GWA) studies on traits selected over much longer time scale such as drought tolerance or flowering time have highlighted only minor effect loci that rarely contribute to more than 5% of the phenotypic variation (Buckler et al. 2009; Cook et al. 2012; Wallace, Larsson, and Buckler 2014).
In addition to the time frame over which adaptation occurs, another important factor for evolution is the nature of variation for selection to act on. Maize and teosintes are genetically very diverse, with as much nucleotide diversity in coding regions between two maize lines as there are between humans and chimpanzees (Tian, Stevens, and Buckler 2009). This diversity is even higher in intergenic regions (Tenaillon et al. 2001; Buckler, Gaut, and McMullen 2006). Some adaptive mutations are found in coding sequences. Examples include non-synonymous changes in the tga1 gene responsible for the “naked kernel” maize phenotype, and in the diacylglycerol acyltransferase (DGAT1-2) gene resulting in elevated kernel oil content in maize lines (Wang et al. 2005; Zheng et al. 2008). But most observations support adaptation from regulatory non-coding sequences. Indeed, in comparison with Arabidopsis, where adaptive variants are enriched in coding sequences (Hancock et al. 2011), in maize and teosinte these are predominantly found in non-coding region: estimates in Zea show that non-coding variation may explain as much of the phenotypes as the coding regions (Chia et al. 2012; Rodgers-Melnick et al. 2016). Selection on regulatory sequences drive important expression changes; hence, genes displaying footprints of selection in maize are usually more expressed than in teosintes (Hufford, Xu, et al. 2012), and are associated with modified co-expression networks (Swanson-Wagner et al. 2012).
Introgression from wild maize in highland populations
Adaptive introgression from the wild teosinte taxon Zea mays ssp. mexicana (hereafter, mexicana) has previously been observed in maize in the highlands of Mexico (Hufford et al. 2013). Our broad sampling allowed us to investigate whether introgressed mexicana haplotypes have spread to highland maize populations outside of Mexico, potentially playing a role in adaptation in other regions. In order to test this hypothesis, we calculated Patterson’s D statistic (Durand et al. 2011) across all maize populations. All individuals from both the Mexican and Guatemalan highlands exhibited highly significant evidence for shared ancestry with mexicana (Additional file 1: Figure S4). Maize from the southwestern USA also showed more limited evidence of introgression, consistent with findings from ancient DNA suggesting this region was originally colonized by admixed maize from the highlands of Mexico (Da Fonseca et al. 2015). In contrast, the distribution of z-scores for South American populations overlapped zero, providing no evidence for substantial spread of mexicana haplotypes to this region.
We localized introgression to chromosomal regions through genome-wide calculation of the fd^ statistic (Martin, Davey, and Jiggins 2015). Megabase-scale regions of introgression were identified in both Mexican and Guatemalan highland populations that correspond to those reported by (Hufford et al. 2013) on chromosomes 4 and 6 (Fig. 2; Additional file 1: Figure S5). On chromosome 3 (at around 75−90 Mb), a large, previously unidentified region of introgression can be found in the Mexican and southwestern US highlands (Fig. 2; Additional file 1: Figure S5).
This region overlaps a putative chromosomal inversion associated with flowering time in maize landraces (Romero Navarro et al. 2017) and in the maize nested association mapping population (Buckler et al. 2009) and may be an example of mexicana contribution to modern maize lines.
The influence of demography on accumulation of deleterious alleles
Population-specific changes in historical N e should influence the efficiency of purifying selection and alter genome-wide patterns of deleterious variants (Fu et al. 2014). Introgression from a species with substantially different N e may also influence the abundance and distribution of deleterious alleles in the genome (Harris and Nielsen 2016; Juric, Aeschbacher, and Coop 2016). Below we evaluate the effects of major demographic events during the pre-Columbian history of maize on patterns of deleterious alleles.
Domestication and deleterious alleles
We first compared counts of deleterious alleles in Mexican lowland maize individuals to four parviglumis individuals from a single population in the Balsas River Valley. Maize from the Mexican lowlands has not experienced substantial introgression from wild relatives and is near the center of maize origin (van Heerwaarden et al. 2011), and thus best reflects the effects of domestication alone. After identifying putatively deleterious mutations using Genomic Evolutionary Rate Profiling (GERP) (Cooper et al. 2005), we calculated the number of derived deleterious alleles per genome under both an additive and a recessive model across four levels of mutation severity (see Methods for details). Maize showed significantly more deleterious alleles than teosinte under both additive (<1 0% more; p=0.0079, Wilcoxon test; Additional file 1: Figure S6) and recessive (< 20−30% more; p=0.0079; Fig. 3) models across all categories (Additional file 1: Figure S7). Additionally, maize contained more than twice as many fixed deleterious alleles than teosinte (57,881 versus 26,947) and 10% fewer segregating deleterious alleles (429,837 versus 478,594), effects expected under a domestication bottleneck (Fig. 3c; (Simons et al. 2014)). GERP load (GERP score × frequency of deleterious alleles), a more direct proxy of mutation load quantified at the population level, revealed a similar trend (additive model: maize median =23.635, teosinte median =22.791, p=0.008, Wilcoxon test; recessive model: maize median =14.922, teosinte median =12.231, p=0.008). Maize, like other domesticates (Marsden et al. 2016; Liu et al. 2017; Renaut and Rieseberg 2015; Günther and Schmid 2010), thus appears to have a higher mutation load compared to its wild progenitor parviglumis.
The effect of the Andean founder event on deleterious alleles
The extreme founder event observed in the Andes could potentially alter genome-wide patterns of deleterious variants beyond the effects of domestication alone. Under a recessive model, maize from the Andes contains significantly more deleterious alleles than any other population (Fig. 3b; Additional file 1: Figure S7; all p values <0.02, Wilcoxon test), and this difference becomes more extreme when considering more severe (i.e., higher GERP score) mutations (Additional file 1: Figure S7). In contrast, we observe no significant difference under an additive model (Additional file 1: Figure S6; Additional file 1: Figure S7). The Andean founder event therefore appears to have resulted in higher mutation load than seen in other maize populations. This result is further supported by a higher proportion of fixed deleterious alleles within the Andes and fewer segregating deleterious alleles (Additional file 1: Figure S10; Fig. 3d), a result comparable to the differences observed between maize and parviglumis.
Introgression decreases the prevalence of deleterious alleles
Highly variable rates of mexicana introgression were detected across our landrace populations (Fig. 2; Additional file 1: Figure S4; Additional file 1: Figure S5). To explore the potential effects of introgression on the genomic distribution of deleterious alleles, we fit a linear model in which the number of deleterious sites is predicted by introgression (represented by fd^) and gene density (exonic base pairs per centimorgan) in 10-kb non-overlapping windows in the Mexican highland population where we found the strongest evidence for mexicana introgression. Gene density was included as a predictor in the regression to control for the positive correlation observed between gene density and both introgression (p=3.48e−08) and deleterious alleles (p≈0).
When identifying deleterious alleles under both additive and recessive models, we found a strong negative correlation with introgression (i.e., fewer deleterious alleles in introgressed regions; p≈0 under both models). These findings likely reflect the larger ancestral N e and more efficient purifying selection in mexicana.
Table of contents :
Synthèse en français
II. GENOMICS OF LONG- AND SHORT- TERM ADAPTATION IN MAIZE AND TEOSINTES
II.2 How to explore adaptation?
II.3 Local adaptation in maize and teosintes
II.4 What is the role of phenotypic plasticity?
II.5 How convergent is adaptation?
II.6 Mechanisms of genetic adaptation in maize and teosintes
II.7 What constraints adaptation?
III. THE INTERPLAY OF DEMOGRAPHY AND SELECTION DURING MAIZE DOMESTICATION AND EXPANSION
Maize population size change during domestication and expansion
Introgression from wild maize in highland populations
The influence of demography on accumulation of deleterious alleles
Historical changes in maize population size
The prevalence of gene flow during maize diffusion
Impacts of demography on accumulation of deleterious variants
Population size and deleterious variants
Introgression and deleterious variants
Samples, whole genome resequencing, and read mapping
Demography of maize domestication and diffusion
Population structure, genetic diversity, and inbreeding coefficients
Runs of homozygosity
Detection of introgression
Estimating burden of deleterious mutations
III.7 Acknowledgements and funding
III.9 Supplementary Material
IV. THE POTENTIAL ROLE OF GENETIC ASSIMILATION DURING MAIZE DOMESTICATION
IV.3 Material and methods
Growth chamber experiment
Additional data sets
IV.9 Supporting information
V. BOTH HARD AND SOFT SWEEPS CONTRIBUTE TO LOCAL ADAPTATION OF NATURAL POPULATIONS OF TEOSINTES
V.2 Results and discussion
V.4 Material and methods
Samples and whole genome re-sequencing
Read mapping and SNP calling
Population genetics parameters
Site frequency spectrums
Hard sweeps and soft sweeps
Percentage of genic regions in hard sweeps
Modeling heritabilities explained by selective sweeps
V.6 Supplementary information
VI. DISCUSSION AND PERSPECTIVES
VI.1 Main results
VI.3 Future directions