An evolving approach to understanding plant adaptation
(Author for correspondence: tel +1 541 737 8488; fax +1 541 737 1393; email email@example.com)
The advent of genomics and other ‘omics’ has stimulated movement towards linking gene function and molecular mechanisms to whole-plant physiology, ecology and evolution. The impetus for this feature issue was the 12th New Phytologist Symposium, ‘Functional genomics of environmental adaptation in Populus’ (DiFazio, 2005); hence, most papers in this issue focus on poplars or other forest trees. Although studies on the molecular basis of adaptation are progressing in Arabidopsis (Stenøien et al., 2002; McKay et al., 2003), and in other plants that are long-standing models of adaptive evolution (e.g. Mimulus; Bradshaw & Schemske, 2003), studies on forest trees will have an important niche. Trees are good models for studying the genetics of adaptation because of their extensive natural populations with little confounding substructure, random mating, sufficient nucleotide diversity, rapid decay of linkage disequilibrium, economic importance and long history of ecological genetic research (Morgenstern, 1996; Neale & Savolainen, 2004). Although difficulties in genetic transformation will constrain research in most forest trees, transformation is relatively easy in Populus, allowing hypotheses to be tested using controlled experimentation with homologous or heterologous gene insertion (Busov et al., pp. 9–18; Man et al., pp. 31–39; both in this issue). Furthermore, the availability of the complete genome sequence in Populus (http://genome.jgi-psf.org/Poptr1/Poptr1.home.html) will unlock possibilities for ecological genetic research that were unforeseen just a few years ago. Despite this progress, a complete understanding of the molecular genetic basis of adaptation will require cross-disciplinary collaboration on a scale that has been rarely seen, including collaborations among experts in molecular biology, physiology, ecology, population biology, statistics and bioinformatics (Wright & Gaut, 2004; DiFazio, 2005). Fortunately, this area is beginning to form the dendritic connections between diverse disciplines that are needed to answer some of the most fundamental questions in biology.
‘In the broad sense, adaptations are phenotypic traits that have been favored by natural selection, and can be identified by being variable, heritable and responsible for variation in Darwinian fitness.’
How do plants adapt?
The phenotypic changes we see over evolutionary time, across diverse environments and among taxa often reflect adaptive evolution (Orr, 1998; Kawecki & Ebert, 2004). In the broad sense, adaptations are phenotypic traits that have been favored by natural selection, and can be identified by being variable, heritable and responsible for variation in Darwinian fitness. Individual plants also adjust (acclimate) their physiology and development in response to changes in their environment. For example, the morphological and physiological changes that occur in temperate-zone trees from summer to winter are dramatic examples of adaptive phenotypic plasticity. The genetics, ecology and evolution of phenotypic plasticity was the focus of a recent issue of New Phytologist (Vol. 166, No. 1, April 2005) and new information on phenotypic plasticity is presented in this issue.
Local adaptation may occur in species exposed to diverse abiotic or biotic environments – and may contribute to speciation (Schluter, 2001; Kawecki & Ebert, 2004). Furthermore, an understanding of local adaptation is important for designing effective breeding programs and conserving genes (Aitken, 2000). In contrast to locally adapted phenotypes, characters that are unconditionally adaptive tend to become fixed within species (Kawecki & Ebert, 2004), and may distinguish closely related species. The best measure for judging the relative importance of an adaptive trait is its effect on fitness itself, or well-documented components of fitness such as seed production, germination, survival and vegetative biomass. Eventually, however, we would like to understand the multifaceted biochemical, physiological and morphological bases of these adaptations, and their genetic control. For example, which component traits contribute to survival and vegetative biomass – and which genes are responsible for genetic variation in these traits? Studies of the biochemical and physiological bases of fitness components have a long history, which continues in this issue, including analyses of Populus productivity and architecture (Cooke et al., pp. 41–52; Monclus et al., pp. 53–62), survival and biomass of Arabidopsis on serpentine-like nutrient solutions (Bradshaw, pp. 81–88), in situ respiration of tree roots (Cheng et al., pp. 297–307), and carbon transport and partitioning in Populus (Babst et al., pp. 63–72). Furthermore, this issue addresses both adaptation and acclimation to a variety of environmental challenges, including edaphic factors such as soil moisture and nutrients (Bradshaw; Cooke et al.; Di Baccio et al., pp. 73–80; Man et al.; Monclus et al.), competition (e.g. productivity, Monclus et al.; and tension wood formation; Paux et al., pp. 89–100), and pathogens (Jorge et al., pp. 113–127).
Other ways to judge adaptive traits include looking for evidence that they are under strong natural selection, including nonrandom interactions between genotypes and environments (Kawecki & Ebert, 2004) and nonrandom distributions of genotypes relative to important environmental variables (Howe et al., 2003). Bradshaw, for example, chose to study tolerance to serpentine soils because reciprocal transplant studies demonstrate that plant adaptation to serpentine soils has a genetic basis. Furthermore, Bradshaw mimicked these transplant studies by growing wild-type and mutant Arabidopsis genotypes on normal and ‘faux serpentine’ nutrient solutions, and demonstrated that the cax1 mutants have most of the phenotypes associated with tolerance to serpentine soils. In their review article, Cooke & Weih (pp. 19–30) noted the genecological characteristics of seasonal nitrogen cycling, and how its timing is closely tied to the environmental cues that control the annual growth cycle in poplar trees.
Articles in this issue also shed light on both short-term and long-term acclimation to changing environmental conditions, including changes in poplar biomass accumulation, plant architecture and gene expression in response to nitrate levels (Cooke et al.), changes in carbon transport and partitioning in response to jasmonic acid (a signaling molecule that mimics the induction of plant resistance by herbivores; Babst et al.), and changes in gene expression in response to tree bending (Paux et al.) and elevated levels of CO2 and O3 (Gupta et al., pp. 129–142; Taylor et al., pp. 143–154). Although the extent and adaptive significance of phenotypic plasticity remains to be fully understood, it is clear that large-scale modifications in gene expression occur rapidly in response to many environmental changes. Gene expression profiling using microarrays not only provides a remarkably sensitive tool for studying phenotypic plasticity, but also provides an entry into uncovering the actual genes responsible for natural variation in adaptive traits. Moreover, by revealing molecular genetic networks, microarrays also provide inroads into understanding how a gene's position within a pathway affects its potential to contribute to adaptation (Cork & Purugganan, 2004).
Genetics of adaptation
The prospects for understanding physiological, ecological and evolutionary genetics have changed dramatically in recent years – we now have the tools to identify the specific genes responsible for adaptive genetic variation in natural populations. Despite a rich literature describing empirical studies of quantitative genetic variation and expectations from theoretical treatments of adaptation and speciation (Orr, 1998, 2003; Mauricio, 2001), we still know little about the genetic basis of adaptation. Which classes of genes vary among ecotypes, ecoclines, varieties and species? Are the same classes of genes involved? Does adaptation and speciation involve few genes with large effects or many genes with small effects? In the near future, there will be a wealth of new information on each of these questions.
Although we have the tools in-hand, understanding the genetic basis of adaptation is still a daunting task. We only have a rudimentary understanding of which traits are most important, and the numbers of genes controlling each of these traits is large. Therefore, the most thorough (and objective) approach would be to associate genes with phenotypes by scanning the entire genome, presumably with single nucleotide polymorphism (SNP) markers. The practical difficulties of this approach, however, are enormous. Therefore, most researchers are focusing on a small to modest number of candidate genes – genes believed to be associated with particular phenotypes based on indirect or circumstantial evidence (Pot et al., pp. 101–112). How are these candidates chosen? The most obvious approach is to choose genes based on their known function – either in the plant of interest or in other species. Of course, it is best if the candidates have been directly tested in the species of interest. The review by Busov et al. reveals that transgenic approaches have been widely used to demonstrate that individual genes can influence specific adaptive traits. Man et al., for example, shows that a pine glutamine synthetase gene can enhance nitrogen assimilation efficiency in transgenic poplar, leading to increased leaf area, plant height and leaf dry weight. Traditional mutational approaches and activation-tagging may be even more valuable because they can be used to find ‘new’ genes that are associated with adaptive phenotypes (Busov et al.). Using activation-tagging in Arabidopsis, Bradshaw found that mutations in a gene that encodes a tonoplast calcium-proton antiporter (CAX1) results in plants that have most of the phenotypes that are associated with plants adapted to serpentine soils. Bradshaw's data and physiological model suggest that mutations in this gene might be involved in the adaptation of a wide variety of natural plant populations to serpentine soils. Although physiological and biochemical studies can be used to test the functional relationships between specific genes and phenotypes (Bradshaw; Man et al.), these tests do not indicate whether these genes have any role in explaining adaptive genetic variation in nature. This requires a comparative analysis of naturally occurring alleles within a single species, or orthologous loci in sister taxa, perhaps using transgenic approaches (Bradshaw).
Candidate genes can also be identified based on their positions on quantitative trait locus (QTL) maps or patterns of gene expression. Jorge et al. mapped QTL for qualitative and quantitative resistance to leaf rust (Melampsora) in poplar hybrids and noted that analogs of NBS-LRR resistance genes have been found in the vicinity of Melampsora resistance loci in other hybrid poplar pedigrees (Zhang et al., 2001). Paux et al. identified candidate genes for tension wood formation based on the differential and distinctive patterns of gene expression in bent vs unbent Eucalyptus trees. ‘Reaction wood’, which forms in leaning or bent trees, helps maintain upright growth of the main stem and adds support to large branches.
The approaches described above generally use ‘forward genetics’ to proceed from adaptive phenotype to adaptive genes. An alternative path toward uncovering the genetic basis of adaptation uses approaches more akin to ‘reverse genetics’. In its purest form, the first step is to simply identify genes that have strong ‘signatures’ of natural selection (e.g. based on DNA sequence analysis and population differentiation), and then determine their phenotypes and specific roles in adaptation. Based on an exploratory study of nucleotide diversity in eight wood-related genes in pines, Pot et al. identified two genes with unusually high levels of population differentiation in Pinus pinaster, suggesting that artificial or natural selection was operating. Pp1 encodes a glycine-rich protein thought to provide elasticity and tensile strength, whereas KORRIGAN is involved in cellulose–hemicellulose assembly. In contrast to approaches that incorporate careful physiological analyses, sequence-based approaches may use evidence of strong population differentiation or the distinctive characteristics of DNA sequences themselves as evidence for the important adaptive roles of these genes (Wright & Gaut, 2004). The effects of underlying population structure should be incorporated into these analyses to obtain valid conclusions about the role of natural selection on individual genes (Wright & Gaut, 2004). One way to do this is to characterize variation in neutral genetic markers. The microsatellite markers and sampling strategies described by Cole (pp. 155–164, this issue) may be suitable for these kinds of analyses.
‘Because of its status as a model plant species, poplar research could provide a more direct means to assess the influence of generation interval, longevity, organism size and effective population size on genome evolution and adaptation.’
Recent studies of plant genome sequences and large expressed sequence tag (EST) databases show that Arabidopsis, rice and several crop plants are paleopolyploids (Bowers et al., 2003; Blanc & Wolfe, 2004a; Wang et al., 2005). Although all angiosperms and most land plants may be polyploid, the number, antiquity and types of genome duplications and subsequent genome reorganizations vary among taxa (Soltis, 2005). The mechanisms that direct the evolutionary fate of duplicated genes are still unclear. For example, the duplication itself (or successive rearrangements) might make the duplicated gene immediately different from its parental copy by placing it under the influence of new cis-regulatory elements (Lynch & Katju, 2004). In a study of duplicated genes originating from the most recent genome duplication in Arabidopsis, Blanc & Wolfe (2004b) found that certain classes of regulatory genes (such as kinases) were preferentially retained, whereas others (such as those involved in DNA repair) were preferentially lost. Is this a common pattern in plants? Are there lineage-specific patterns of gene retention and loss that are indicative of specialized functions or adaptations? With the rice and poplar genome sequences in hand, and other plant genome sequences on the horizon, we can begin to answer these questions. By analyzing ESTs from several Populus species, Sterck et al. (pp. 165–170) showed that a genome duplication occurred early in the evolution of the genus, before the species diverged. Based on synonymous nucleotide substitution rates, the authors estimated that the duplication occurred 8–13 Myr ago, which conflicts with fossil evidence that suggests that the species diverged 18–58 Myr ago (Eckenwalder, 1996). While misinterpretation of the fossil record is one possible explanation, Sterck et al. also note the potential impact of generation interval on synonymous substitution rates, which have been mostly estimated from plants that flower and reproduce at very young ages compared with poplar trees. Although an individual poplar tree typically lives for around 100 years, poplar's propensity for clonal propagation allows a genotype to survive and sexually reproduce for millennia. Because of its status as a model plant species (Busov et al.), poplar research could provide a more direct means to assess the influence of generation interval, longevity, organism size and effective population size on genome evolution and adaptation (Lynch & Conery, 2003).
In addition to genome duplication and restructuring, a recent study revealed a previously unknown process for genome-wide sequence changes. Lolle et al. (2005) found that Arabidopsis plants carrying mutations in the HOTHEAD gene can inherit allele-specific DNA sequences at multiple loci that were not present in the genomes of their parents, but were present in an earlier ancestor. The authors propose that a cache of stable RNA serves as the template for this extra-genomic mechanism of DNA sequence reversion, and may provide a way for self-fertilizing species to circumvent the negative consequences of inbreeding or recover from the effects of a genotype that is maladapted to its present environment. As described by Chen et al. (pp. 171–180), flax exhibits phenotypic and genomic changes associated with environmental factors. When an inbred variety of flax (P1) is grown in an inducing environment, it gives rise to progeny called genotrophs that exhibit stable changes in size, branching, seed hairs, isozymes, hormone levels, nuclear DNA contents, number of ribosomal genes and number of other repetitive sequences. Although the parents remain phenotypically plastic when grown in different environments, the altered phenotypes of the genotrophs are stable. Building on previous work that showed that the flax genome undergoes highly specific DNA changes at multiple loci from parents to progeny, they identified a site-specific insertion sequence (LS-1) in the genotrophs that is also found in natural populations of flax. Reminiscent of the changes associated with the HOTHEAD mutants, an intact LS-1 is not present in the genome of the progenitor, P1. These studies suggest the possibility that an extra-chromosomal, RNA-based mechanism of inheritance might be operating in flax as well.
This feature issue illustrates that we are at an exciting and challenging crossroads in the study of plant adaptation. A powerful suite of tools is now available for Arabidopsis, and other taxa that are highly divergent with respect to evolutionary distance (rice) and perennial growth habit (poplar). Moreover, genomic resources are becoming increasingly available for other tree species and plants that are models of adaptive evolution. Because EST databases and high-throughput techniques for assessing both molecular and phenotypic variation are becoming widely available, progress will not be limited to model plants. Perhaps the most difficult challenge is not the development of tools and resources, but the complete merging of genomic, ecological and evolutionary perspectives.