- Population genomics
- Evolution of meiosis
- Gene network evolution
- Mechanisms of habitat adaptation and morphological evolution
Mechanisms of habitat adaptation
in Arabidopsis arenosa
Flowering time differences associated with habitat:
Arabidopsis arenosa is unusual compared to its relatives in its ability to survive in a range of distinct habitats; it has been reported from limestone outcrops and scree, from river cobbles, from limestone, from serpentine, from lead, zinc and cadmium contaminated soils, from beaches, and from acidic peat bogs. A striking contrast is between populations found on railway versus shaded outcrop sites, and the tetraploid A. arenosa is very abundant in both.
Plants growing in railway sites are exposed to greater drought risk, elevated temperature, higher UV exposure, poor soil quality, and annual herbicide applications. How did these plants manage this habitat shift? Are they just very plastic, or did they adapt genetically? Our initial common garden experiments suggest that there was at least some genetic adaptation; the plants have heritable differences in life history and developmental traits. We are using genomic analyses and genetic mapping to ask which genes might have been involved in adaptation to these distinct habitats, with an initial focus on differences in reproductive timing.
Examples of two typical A. arenosa habitats. Left, a shaded limestone rock outcrop, and right, a railway bed (Photos: left, K Bomblies; right, Sang-Tae Kim).
It was previously described that A. arenosa from lowland railways are rapid-cycling annuals and populations from mountain or hill outcrops are late-flowering perennials. We have confirmed in the lab that this is largely true, replicable in common gardens, and that this difference is at least in part genetic: in our controlled conditions, progeny of mountain plants flower much later than those of flatland plants. When plants are treated with extended cold (vernalization), the flowering of mountain plants, but not flatland plants, is greatly accelerated. Thus flatland plants are early flowering in the lab because they have the lost the vernalization response. We are currently using genetic mapping and transcriptome sequencing to explore the causal loci for flowering time variation in tetraploid A. arenosa.
Flowering time of non-vernalized (left) and vernalized(right) plants. Populations are connected by lines. Outcrop populations (red) are late flowering, but have a much stronger response to vernalization than railway (yellow) populations. A river cobble population shows an intermediate response.
In addition to phenotyping we have done RNAseq analyses (primarily this is work of postdoc Ben Hunter and PhD student Pierre Baduel) that shows that while most late flowering plants have expected expression patterns from studies in A. thaliana, it has also become clear that not all early-flowering plants are early-flowering for the same reason. One railway population from a mountain region is early flowering like flatland railway populations, but shows gene expression levels of flowering genes more similar to that seen in mountain accessions.
Since a major gene controlling flowering time and the loss of vernalization in A. thaliana is FLC, we have examined this node in particular. We found that flatland railway populations have little or no FLC expression, fitting with their early flowering. They also show high levels of expression of an FLC-suppressor. However, the early flowering mountain railway population has high levels of FLC expression. This finding suggests that these railway populations are early flowering for different reasons and likely independently evolved early flowering.
Together with David Salt (University of Aberdeen)we inititated a study of substrate adaptation in A. arenosa. The Salt lab analyzed over 100 A. arenosa isolates from our collection using their Ionomics platform to measure nutrient uptake. Several exciting trends emerged, for example one population is extremely efficient at phosphorous uptake and accumulation. Another, collected from serpentine, excludes magnesium. There is extensive variation for sodium uptake. We plan to test candidate genes for these differences. Our genomic analyses have also identified a number of candidate genes involved in substrate adaptation to limestone (calcareous) versus silicaceous soils. These include numerous ion channels and heavy metal scavenging proteins, as well as genes implicated in redox regulation. Over time we hope to understand better the molecular mechanisms these plants used for substrate adaptation, how easily substrate shifts are achieved, whether these edaphic adaptations restrict gene flow among populations, and whether populations found on similar substrate, but in geographically discontiguous regions, migrated there or “re-adapted” from neighboring populations. In cases where independent adaptations to similar substrates occurred, we can ask whether the same genes are involved.
Limestone outcrop in the Košice region in Slovakia (photo B Arnold).
Non-calcareous rock outcrop site in a forest in the Steiermark region of central Austria (photo B Arnold and J Hollister).
Brian Arnold and Jesse Hollister also collected A. arenosa plants from a serpentine site in Austria, where beautiful green rocks abound. Serpentine is a geochemically challenging substrate for plants, with a high level of nickel and low Calcium to magnesium ratios (photo B Arnold and J Hollister).
Mechanisms of morphological evolution
in Arabidopsis arenosa
In addition to the flowering time and substrate differences mentioned above, A. arenosa populations vary extensively in morphology. We see particularly striking morphological variation in root system architecture and gravitropism, inflorescence branch outgrowth, and leaf shape.
Root architecture: Subspecies arenosa grows in flat ruderal sits, especially railways, while subspecies borbasii, grows on steeply sloped rock outcrops and scree slopes. Interestingly, subspecies arenosa has gravitropic roots (like A. thaliana and A. lyrata) but subspecies borbasii has "wandering" roots (see picture below). Root architecture differences of this sort have been implicated in other species in adaptation to growing on slopes, nutrient and moisture acquisition.
Inflorescence architecture: Subspecies arenosa also has more extensive branch outgrowththan subspecies borbasii. Rosette branches are greater in number and growth rate. Rate of growth and height at flowering of the main inflorescence is also higher in subspecies arenosa. There is, however, no difference in the lateral branches that grow out from the main inflorescence.
Leaf shape: There is extensive variation among A. arenosa populations in leaf shape. Though not a hard-and-fast rule, subspecies arenosa tends to have simpler leaves (see bottom image below for leaf scans from wild plants).
We plan to use genetic mapping approaches to identify the causal genes underlying variation for these traits.
Roots of A. arenosa, A. lyrata and A. thaliana plants grown on vertical plates at pH5 or pH8. Notice that A. arenosa roots grow better at pH 8 than at pH 5, which is not true of A. thaliana. The strains used here originate from limestone outcrops that naturally have high pH (~7.5-8.5). A. lyrata also shows a preference for elevated pH, but the difference is less dramatic than for A. arenosa. The roots of A. arenosa also show a tendency to “wander” from vertical and show horizontally growing rather than downward curving lateral roots. This feature is unique to our outcrop samples. Plants from railway populations have roots that grow vertically.
Arabidopsis arenosa plants grown in laboratory conditions showing genetically controlled branching differences among strains.
Leaves from wild A. arenosa plants showing extensive variation. Left of the line are leaves from plants from limestone outcrops in the Swabian Alb region of southwestern Germany, while right of the line are leaves from sandy forest sites in the Rhine valley area.