The genetics of adaptation to novel environments: selection on germination timing in Arabidopsis thaliana



Brook T. Moyers, Fax: (604) 822-6089; E-mail:


When studying selection during adaptation to novel environments, researchers have often paid little attention to an organism’s earliest developmental stages. Despite this lack of attention, early life history traits may be under strong selection during colonization, as the expression of adaptive phenotypes at later points is contingent upon early survival. Moreover, the timing of early developmental transitions can constrain the timing of later transitions, with potentially large effects on fitness. In this issue, Huang et al. (2010) underscore the importance of early life history traits in the adaptation of Arabidopsis thaliana to old-field sites in North America. Using a new population of mapped recombinant inbred lines, the authors examined germination timing and total lifetime fitness of A. thaliana while varying site latitude, dispersal season, and maternal photoperiod. Huang et al. (2010) discovered several Quantitative Trait Loci (QTL) with large effects on fitness that colocalized with QTL for field germination timing and seed dormancy—demonstrating that fitness is genetically associated with these early life history traits, and that these loci are likely under strong selection during adaptation to novel environments. In the epistatic interactions of some loci, recombinant genotypes outperformed parental genotypes, supporting the potentially adaptive role of recombination. This study provides elegant evidence that traits expressed early in an organism’s development can play an important role during adaptive evolution.

Arabidopsis thaliana is an attractive system for answering ecological genetic questions because the wealth of genetic data and genotyped lines available makes it possible to quantify fitness effects of particular genetic differences in natural environments (e.g. Wilczek et al. 2009). Moreover, its predominantly selfing breeding system, which makes it a great laboratory organism, also makes it well suited for doing quantitative genetics in the field, as inbred lines are less likely to be affected by inbreeding depression than in outcrossing species (Mitchell-Olds 2001). A previously estimated outcrossing rate of less than 1% (Abbott & Gomes 1989) means that recombinant inbred lines are actually quite likely to occur in nature; the experiments of Huang et al. (2010) therefore have recreated the types of genetic variation that might occur naturally in cases of recent A. thaliana introduction. Moreover, Huang et al. (2010) used recombinant inbred lines (RILs) derived from a cross between wild accessions rather than strains adapted for laboratory experiments, again increasing the ecological relevance of this study.

Huang et al. (2010) present results from a field experiment which examined phenotypic plasticity and gene by environment effects for a broad suite of traits. They used RILs derived from a relatively unstudied cross between wild-collected lines from Tacoma, Washington, USA (`Tac’) and Calver, England (`Cal’). The authors dispersed seeds from the same RIL population, matured under two different maternal photoperiods (14-h and 10-h daylengths), at two latitudes (in Rhode Island and Kentucky), and during two seasons (June in KY and RI, November in RI). Environmental conditions experienced during seed maturation, especially photoperiod, have been shown to affect seed dormancy and germination timing (e.g. Munir et al. 2001). All of the manipulated factors reflect real variation in environmental conditions experienced by natural A. thaliana populations, although not all combinations (e.g. a short maternal photoperiod for seeds dispersed in June) are natural. The field experiments tracked total lifetime fitness and phenology, from seed germination (Fig. 1) to reproduction (Fig. 2), and in doing so were able to detect allele frequencies changing in a single generation. Huang et al. (2010) additionally supplemented these field experiments with laboratory studies of primary and secondary seed dormancy and adult life history traits in the same RIL population—the authors used these studies to look for genetic associations between these traits and traits expressed in the field.

Figure 1.

 Young Arabidopsis seedlings germinated in the field. Photo credit: Amity Wilczek.

Figure 2.

 Flowering at the wrong time can result in fitness costs. Photo credit: Engseng Ho.

Several striking results demonstrate the importance of including early life history traits in studies of adaptive evolution. Crucially, the authors identified a single QTL on chromosome 3 with large effects on both fitness (explaining up to 20% of the variance in fitness) and germination timing (explaining up to 45% of the variance in germination timing) in the field. This QTL also colocalizes with a QTL for primary seed dormancy induction in the lab, and includes a well-studied seed dormancy candidate gene. Chromosome 5 contained another QTL with significant effects on fitness and germination phenology in the Kentucky field location and overlaps with a previously characterized QTL affecting germination in the laboratory (Clerkx et al. 2004). This colocalization suggests that germination timing explains the QTL’s observed strong fitness effects in June-dispersed seeds. Because the strong selection on early life history traits eliminated much of the genetic variation, later life history traits such as reproductive timing and rosette leaf number could only be mapped in the lab. Unlike germination timing, QTL for adult life history traits did not colocate with the field-measured fitness QTL, again indicating the primary importance of early life history traits in this experiment. Interestingly, fitness QTL had significant epistatic effects, with Cal alleles at both loci showing higher fitness than any other combination. A full genome scan for epistasis revealed that Cal–Cal combinations of alleles were commonly associated with higher fitness, although in one-third of cases where significant epistasis existed, the recombinant Cal-Tac genotypes were superior.

Selection on early life history traits has the potential to be quite strong during adaptation to a novel environment, as Huang et al. (2010) demonstrate. In a single generation, the allele frequency of the large-effect fitness QTL on chromosome 3 reached nearly 90% in some treatments. This occurred despite some initial segregation distortion of the RILs in the opposite direction. This rapid genetic change is likely to be related to the cascading effect of germination timing on later developmental transitions, especially on the timing of the transition from vegetative growth to reproduction. In A. thaliana, and likely other annual plants, germinating at an inappropriate time can have large effects on survival and lifetime fitness (Wilczek et al. 2009). As such, control of germination timing through the induction of or release from primary and secondary dormancy is a form of habitat selection. Habitat selection allows an organism to target an environment to which it is already adapted, and may enable colonizing populations to become established.

Huang et al. (2010) present the most complete experimental evidence for the importance of early life history traits in adaptation to novel environments. Their work has important implications for studies of range expansion, invasiveness, and adaptation to novel climates, and provides support for the adaptive potential of recombination between divergent lineages. More specifically, this study points to factors which may have enabled A. thaliana to become broadly cosmopolitan and demonstrates the usefulness of this species as a tool for ecological genetic studies. Two previous papers from this group (Donohue et al. 2005a,b) examined the role of phenotypic plasticity and maternal effects in this population—the current study goes further by explicitly linking genotype and phenotype with lifetime fitness. The result gives us an important message: in order to fully understand how organisms adapt to novel environments, we must understand selection at the earliest life history stages.