Genetic Basis of Dynamic Light Response Signalling and Photoprotection in Plants
This research will reveal the genetic signalling network controlling photosynthetic and developmental responses to spectrally dynamic light conditions in Arabidopsis and Brachypodium. Genome wide association studies, novel mutant screens, gene expression and landscape genomic studies, will be used to identify new adaptive alleles, genes and conserved signalling pathways that optimize light capture and use. The integration of these different data sets will reveal genetic control points in the signaling network to rationalise further molecular and physiological characterization. This work will determine how naturalized plants do, and crops plants can, survive and thrive in environments which differ in the amount, quality, and timing of light.
Light intensity and quality change by orders of magnitude during the day due to sun angle and shading, markedly impacting photosynthesis, development and yield. Our research will identify genes, adaptive variants, and conserved pathways that provide protection and developmental acclimation enabling improved yield under fluctuating conditions.
The aims of this proposal are to determine the genetic architecture and conserved signalling pathways controlling dynamic light response in two plant model species. Dynamic growth chamber and field experiments will reveal key candidate genes underlying light adaptation defining natural control points of photosynthesis, photoprotection, and shade signalling. Together, the alleles, genes, and pathways identified will provide targets for crop breeding and suggest approaches for genetic conservation and restoration of key populations of foundation plant species.
AIM 1: Discovery of dynamic light response alleles and genes using genome wide association studies and novel mutant screens
AIM 2: Field and genetic studies to identify and confirm adaptive alleles
AIM 3: Define molecular function of new alleles and genes regulating light acclimation
What is the problem and how will we achieve a breakthrough?
Plant performance and crop yield are impacted by the photosynthetic apparatus having to cope with variation in ambient light. In a given hour, day, week, and season the light intensity and spectra fluctuates by orders of magnitude due to sun angle, clouds, and variable canopy shading. Light and the status of the chloroplast regulate photomorphogenesis and development controlling leaf morphology, plant architecture, and flowering time. Decades of research has identified: many of the sensors of light for development (phytochromes, cryptochromes, phototropins, etc.); the structural components of the photosynthetic apparatus; and photoprotective mechanisms that adjust light harvesting on a short, medium, and long term time scales.
However, except for key studies, few have considered the complexity of dynamic changes in light that occur in natural environments, nor have they identified natural genetic diversity as a source of novel insight into light response mechanisms. There are literally thousands of papers on photoprotection in plants (ISI web of Science), yet just a handful of studies, including from our groups, have shown that there is natural variation (Borevitz et al 2002, Botto and Smith 2002) and an ability to acclimate in response to fluctuating light (Alter 2012, Caliandro et al, 2013, Gordon et al, 2012, Jung and Niyogi, 2009). Similarly, few studies, again mostly by our groups, have shown that absence of one or more photoprotective mechanisms has little influence under continuous high or low light, but can be lethal or severely impair seed yield under fluctuating light (Tikkanen et al 2010, Suorsa et al 2012, Kulhiem 2002). This is because such studies are typically limited by the breadth of researchers’ backgrounds, molecular and computational skills, germplasm, and until recently LED lighting technology.
Our team has a uniquely diverse skill set (high light and shade signalling, adaptation/conservation and mechanism), germplasm (natural populations and mutants), technology (dynamic multispectral LED lights, plant growth and photosynthetic imaging, high-throughput genomics), and experimental data to ensure a systematic investigation. We will perform association studies and expression analysis controlling dynamic light responses of photosynthesis, photoprotection, and key developmental events in natural populations and mutants impaired in fluctuating light. We will then determine the adaptive importance of the underlying genetic variation in the field. This knowledge will be used to validate new alleles and genes and prioritize characterization of light response and photoprotective processes.
What is our strategy and what will be the breakthrough findings?
The challenge, because of the central role of light in photosynthesis and development, is to design experiments that identify novel regulators and control points, not just re-cloning known structural photosynthetic genes. Thus, a central strategy is to determine natural genetic variation in two species under contrasting ‘field like’ light regimes. In parallel, we will screen for suppressors in a fluctuating light-sensitive mutant backgrounds. This combinatorial strategy will identify new alleles, genes, and conserved pathways, specifying dynamic light response signalling. It will also test their role in adaptation in the field.
Dynamic diurnal light regimes will simulate field like conditions in growth chambers to mimic sun, shade, and fluctuating light environments.
Photosynthesis, photoprotection, and photomorphogenic growth traits (Phenomics) will be measured in high-throughput on natural and mutagenized populations.
Global and Australian populations will be used for genome wide association studies on Arabidopsis and Brachypodium.
Landscape genomic field studies on current and new collections will identify adaptive loci via genotype associations with light environmental variables.
Mechanistic insight into the regulation of photoprotection will be obtained from above strategies, gene expression analysis, and a suppressor screen of the fluctuating light sensitive mutants.
The results will yield quantitative trait loci (candidate genes and their signalling pathways), their effects (developmental and environmental sensitivities), their pleiotropy (trait correlations), and their geographic distribution. Expression analysis and the suppressor mutant screens under the same dynamic conditions will also identify new players in the acclimation pathways, which can serve as candidate genes underlying each QTL.
Why two species for the natural variation studies?
Together, studying Arabidopsis and Brachypodium allows for deep evolutionary (>100MYA) genomic comparisons separating monocots dicots, including crops such as wheat, barley and canola. Yet they are both ‘agricultural weeds’ and have similarities ecologically – namely, life history strategies of overwintering and/or rapid cycling (Fig. 1). Identifying the genetic basis of dynamic light response signalling and photoprotection allows comparisons of the types of mutations, genes, and pathways utilized by two deeply diverged model plant species.
Why combine GWAS and forward genetics?
Integrating the, at first glance, seemingly disparate strategies in parallel makes data collection more streamlined and enables each strategy to inform the other, thereby minimising effort in trait or mutant validation – the respective bottlenecks. Furthermore, traditional forward genetics has dissected the structural components for the major photoprotective processes and some components of retrograde signalling pathways (reviewed in Chan et al 2013, Murchie and Niyogi, 2011, Tikkanen et al 2012), but have provided little insight into regulation of photosynthetic efficiency and photoprotection with respect to fitness or yield. Integrating GWAS and forward genetics bypasses these limitations, as will new multispectral LED lights better simulate dynamic conditions in a physiologically and ecologically relevant way. Fine mapping and cloning of novel loci will be aided by next generation sequencing, but it is the cross comparison of candidate QTL genes from Aim 1 and eQTL from Aim 3 together with the suppressor screens from Aim 1 and 3 that will enable rapid and accurate selection of the key control genes with adaptive value.
Why low and high light, shade and sun?
It is the dynamic changes among high and low light where physiological mechanisms are necessary to cope with rapid light variation as revealed by Aro’s group (Suorsa et al 2012; Tikkanen et al 2010). Thus, in nature signalling from both shade and sun-regulated pathways would occur concurrently and are in fact two sides of the same coin. Indeed, Pogson and Botto groups have already initiated a joint project on a high light signalling pathway (SAL1-PAP pathway, Estavillo et al 2011) that also functions in shade. Signalling from these pathways is integrated into developmental changes affecting reproduction and yield.
The research projects proposed here aims to determine the adaptive genetic basis of light response variation in photosynthesis, development, and reproduction among contrasting light regimes in two key plant model organisms, Arabidopsis and Brachypodium. We expect to provide new insight into previous limitations with spectral control in LED chambers that allow simulation of dynamic light under competition. Our studies aim to identify these alleles, genes, and pathways available to natural or artificial selection.
Agriculture and Adaptation
Light intensity and quality fluctuate during the day and throughout the year perturbing photosynthesis. Rapid to long term protective mechanisms minimise the negative effects. A loss or reduction in photoprotection can be lethal or result in major seed yield decreases (Kulheim et al 2002). We seek to identify novel genetic regulators of light signalling and photoprotection and determine variants in known photoprotective and light acclimation genes that confer tolerance to fluctuating light. The ‘green revolution’ was brought about by an architectural transformation of plants, allowed much higher planting densities, partially shading each other and thereby changing exposure to light intensity and spectrum through the crop canopy. This highlights an example where artificial selection for community or plot performance is opposite to that of natural selection on the individual. Complicating these studies, is the fact that density dependent phenotypes of photoprotection and shade avoidance, measured on individuals in the lab, often do not transfer to plots in the field. However, plots in the field are highly variable making it difficult to measure or reproduce individual genetic effects. In contrast, local adaptation is important for both natural selection and artificial selection in agriculture. Adapting crop varieties to local environments generally increases sustainable yield (Stokes and Howden, 2010).
Local Adaptation to Light Environment
Light environmental signals can forecast the period of the year with suitable temperatures and available moisture suitable for growing. Growing seasons vary qualitatively, such as overwintering or summer annual. Within season variation is often quantified by photo-thermal hours, the cumulative sum of the product of light and temperature (Wilczek et al, 2009). Plants are exquisitely tuned to these seasonal environmental signals and their development integrates variable weather (light, temp, moisture) into discrete phenological stages that are changing dramatically with climate change (Rosenzweig et al, 2008). As temperatures change and become more variable, growing season signals are being uncoupled from dependable light signals (day length, light quantity and quality), driving adaptation, migration, or causing extinction. Not only are light signals perceived by phytochromes and cryptochromes, the chloroplast is also a sensor of the environment that, in turns, signals development and acclimatory changes.
With respect to the influence of genetics on plant adaptation certain alleles in certain environments confer greater fitness leading to enhanced survivorship in the next generation and an adaptive advantage. This local adaptation leaves a signature of environmental filtering on the landscape, however, it may be unclear exactly what traits are advantageous in a given environment. Powerful new landscape genomic approaches allow us to look directly at DNA sequence variation in situ for the signs of selection. When allelic distribution uniquely coincides with environmental gradients this indicates the range of selection. The alternative experimental approach in common environments, is genome wide association studies (GWAS), which directly identifies alleles underlying potentially adaptive traits. In GWAS the phenotypes are accurately measured, however the experimental conditions may not be the same as those in the field. The research strategy proposed here, unites landscape genomics and GWAS with each informing the other.
Genetic Basis and Geographic Scale of Local Adaptation in Arabidopsis
Prerequisite work in the Borevitzlab resolved genetic diversity and geographic patterns among nearly 6000 wild Arabidopsis accessions (Platt et al, 2010) to develop a genetically balanced global set for fine mapping alleles linked to light (Li et al, 2010) and salt tolerance adaptive traits (Baxter et al 2010). A parallel approach to field based, common garden experiments was taken. We assayed local adaptation in the laboratory by simulating dynamic light regimes and temperatures mimicking climates in growth chambers (Li et al 2006, 2010, 2013). Our key study found the adaptive genetic basis and environmental sensitivity of flowering time and yield. We identified novel and known genes where selection was detected across the latitudinal range of Europe (Li et al, 2010). Subsequently, landscape genomic studies of local adaptation made use of our Arabidopsis line set planted in field conditions. Hancock et al (2011) scanned the genome for SNPs that are correlated with local climate variables while accounting for geographic and genomic correlations and the beneficial alleles that were identified could explain fitness in a local ‘common garden’. In parallel, Fournier-Level et al (2011) performed four common garden experiments across the European range. Alleles that confer a fitness disadvantage in a particular location were found to be more sparse in that region suggesting that they were excluded by natural selection. These studies cover a broad continental range and provide preliminary evidence for selection acting on standing variation, a promising mechanism of rapid adaptation.
Figure 1. Life history and local habitat variation. Arabidopsis, a) overwintering, b, c) annual rapid cycling in spring and/or autumn where lack of dormancy can result in two generations per year. d) Typical annual grain agriculture where weeds display multiple life histories of rapid cycling in open fields and overwintering in shaded field margins. Brachypodium and Arabidopsis have qualitatively similar life history variation.
High light (HL) causes damage to DNA, proteins and lipids, including components of the photosynthetic apparatus (reviewed in Pogson et al 2008, Chan et al 2013, Murchie and Niyogi, 2011, Tikkanen et al 2012). Exposure to prolonged periods of HL increases the generation of reactive oxygen species (ROS) and alters the redox state of photosynthetic components such as the electron carrier, plastoquinone. These components provide important retrograde signals that communicate the chloroplast status to the nucleus driving transcriptional activation of defence systems (Pogson et al, 2008; Estavillo et al, 2011). HL induces (1) pathways that allow for the dissipation of excess energy; (2) systems that detoxify the harmful by-products of HL; and (3) mechanisms that reduce the amount of light absorbed by the plant. Plants have also evolved different mechanisms that facilitate the dissipation of accumulated excess energy absorbed under HL conditions, including state transitions, cyclic electron flow (CEF), photorespiration and non-photochemical quenching (NPQ).
However, rather than the LL to HL shift laboratory experiment, dawn till sunset plants are subjected to varying light intensities due to the angle of the sun and transient shade from clouds, wind-induced movement of leaves, and shading from neighbouring plants. Living in such an environment creates ‘hot spots’ of solar energy, such hotspots can trigger rapid acclimation (Gordon et al, 2012). A key question is what is the extent to which transient changes in light intensity result in a “memory” of the event leading to acclimation responses. Indeed, we observed that treating leaves with fluctuating light resulted in a substantive increase in nonphotochemical quenching capacity (Gordon et al, 2012). But, what are the regulatory pathways that induce acclimation? Using the LED chambers we will impose four sets of light regimes (sun, shade, and two fluctuating conditions of minutes and hours) superimposed with the diurnal shift in light and temperature of a typical ‘day’ on natural populations and key photoprotective mutants (see below).
A new ARC infrastructure award to Borevitz, Pogson, et al ‘Spectral climate chamber facilities for phenomic studies of plant light response adaptation’ (LE130100081 $500k ARC + $340k internal) underpins this proposal. It enables the fine control of light intensity and spectrum (simulating sun, shade, and fluctuating transitions), temperature, and moisture, to simulate local and regional field like conditions at particular locations and seasons (Fig. 2A). This major equipment facility is equipped with high spatial and temporal resolution (NIR-RGB) image based phenotyping of 300 plants in each of 4 chambers. Recent work validated the platform for assaying real-time growth under simulated seasons (Zhang et al 2012). Here, images were taken every 10 minutes and processed to calculate rosette area on hundreds of plants (Fig. 2BC). GWAS was then be performed at each time point to uncover four and five loci separately controlling juvenile and adult growth.
Figure 2. Multispectral lights and quantitative growth analysis. A) 7 wavelengths dynamically adjust to simulate mixed sun and shade with diurnal and seasonal variation (Brachypodium is growing below), B) RGB and Fluorescence time lapse of Arabidopsis growth, C) quantification of plant area (green pixels) every 10 minutes over 5 weeks showing daily rhythms. Each chamber tracks 300 plants.
AIM 1: Discovery of dynamic light response alleles and genes
1.1 GWAS for Arabidopsis – photosynthesis and development under dynamic light conditions
1.2 GWAS for Brachypodium – light response traits under Australian dynamic light conditions
1.3 Identifying regulators of light acclimation by novel suppressor screens
AIM 2: Determine adaptive alleles via landscape genomics and field genetic studies
2.1 Landscape Genomics for light adaptation using public climate data and current populations
2.2 New diverse Brachypodium collections across light environments in Australia
2.3 Evaluate fitness of accessions and mutants in Canberra ‘sandbox’ under sun and partial shade
AIM 3: Define molecular function of new alleles and genes in known light acclimation pathways
3.1 Transcriptional changes that underlie variation in light acclimation (eQTL)
3.2 Candidate gene confirmation of light sensitive QTL
3.3 Characterization of key genes and alleles that modulate the dynamic light response
AIM 1 Discovery of dynamic light response alleles and genes
This proposal integrates three complementary sets of experiments that would traditionally be run separately and sequentially. Here we take advantage of our team’s unique expertise, germplasm, and genomics and phenomics technology to run these experiments in parallel. This has the advantage of genetic cross validation across the three aims, deeper insight from integration, critical mass of researchers and economies of scale, reducing time making an ambitious proposal feasible.
AIM 1.1 Genome wide association studies (Arabidopsis)
The experiments proposed here focus on model plants grown in special chambers under dynamic lighting conditions. They will be used to identify the genetic basis for light sensitivity in photosynthesis and photoprotection, development (photomorphogenesis), and reproduction (flowering time). Genome wide association studies (GWAS) will determine the genetic architecture (number of loci and their environmental sensitivities) and identify candidate genes and pathways underlying light adaptation.
GWAS was established in Arabidopsis thaliana (Atwell et al, 2010, Li et al, 2010, Horton et al, 2012) with 250,000 SNPs. It is now being improved ~5M common SNPs using data from 1001genomes.org and imputation based on extensive linkage disequilibrium (LD). The proposed studies will utilize the balanced global HapMap set of 300 lines (Li et al, 2010) to identify common variation typically available for adaptation into new environments. This block of lines will be phenotyped in each of four chambers running contrasting light regimes. A second round of experiments in year 2 will validate QTL in a reduced genetic diversity set using 144 Cvi X Ler RILs (Borevitz et al, 2002, Botto and Coluccio 2007). These plants will be also used for RNAseq analysis (AIM 3.1).
Simulating growing conditions with dynamic and transient light variation
Most prior studies have mapped QTL under static lights conditions that contrast light quality (Borevitz et al, 2002, Botto and Coluccio 2007, Jiménez-Gómez et al, 2010, Filleau and Maloof 2012). Our dynamic light environments contrast fluctuating light quantity and quality, integrated with diurnal and seasonal signals to identify potentially adaptive and novel loci. Light conditions to be simulated:
open sun (500uE/m2/s, Red/FarRed 1.1)
permanent partial shade (50uE/m2/s, R/FR 0.5)
sun/shade transitions every 5 minutes
sun/shade transitions every 2 hours
Rapid transitions are relevant to light filtering through dense stands or a canopy, and slow transitions mimic diurnal shade in wider woodland openings (Alter et al, 2012). Diurnal and seasonal changes in temperature and day length, typical of the spring growing season will be included as background under each light regime to capture other important field like parameters regulating photosynthesis and development. These conditions mimic environmental variation typical both within local fields sites and among regional locations. We will simulate seasonal variation with a planting date of September 1st in southern Australia which is similar to March 1st in southern Europe. The goal is that by phenotyping in controlled, but ‘field-like’ conditions, we can identify adaptive QTL which will show patterns of allelic variation on the landscape correlating with the light environment variation where the samples were collected (AIM 2.1 and 2.2).
Plasticity among light environments
To compare the cumulative effect of changing light environments through the simulated growing season, a photothermal model incorporating radiation and temperature will be used. Light response traits such as flowering time will be quantified in these cumulative units across environments to account for the variation in amount of light at different times of the day and season (Wilczek et al 2009). Photothermal units normalize growing seasons by down-weighting shorter, shaded, and cooler days relative to longer, bright, and warmer days so the growth potential across various growing seasons can be more directly compared. This allows seasonal plasticity across environments to be calculated in comparable units for individual accessions.
Phenomics: High throughput phenotyping of photosynthesis, photoprotection and development
In addition to measuring the date of flowering, real time quantification of growth and development will be performed in high throughput with digital imaging technology. Our SpectralPhenoClimatron chambers contain RGB+IR cameras to record plant growth and development. Leaf emergence and petiole elongation rate will be assayed continuously in large populations under controlled conditions (Zhang et al, 2012). ANU also has one of the first TrayScan systems to quantify photosynthetic parameters and leaf temperature parameters at the leaf pixel level for 300 plants per hour. This special camera setup can image chlorophyll fluorescence and quantify photosynthetic efficiency, photoinhibition, and photoprotection, via Non-Photochemical Quenching (NPQ). Photosynthetic efficiency has been dissected through traditional QTL mapping (Jung and Niyogi 2009, and our new preliminary data) but not in the context of GWAS or dynamic environments as proposed here. Each block of accessions from the 4 environments will be processed by TrayScan, every second day, from 3 weeks until leaves out grow the pots ~10 weeks. TrayScan will also be used for characterization of fluctuating light suppressor mutants (Aims 1.3 and 3.3).
GWAS method: Phenotype ~ Genotype x Environment
The statistical method part of this proposal is technically straightforward and is now routine in Human Genetics. QTL are identified by a non-random association between the given phenotype (e.g. NPQ) and SNP genotype, while controlling for background relatedness estimated across all SNPs (Kang et al 2010). Once identified, major QTL are subsequently included in the model using a forward stepwise procedure as implemented in QTLRel (Cheng et al 2011). The author of QTLRel, Dr Riyan Cheng, is a statistician in the Borevitzlab and is budgeted 50% to this project. Our association mapping experimental design includes light environmental interactions as fixed effects.
AIM 1.2 Genome Wide Association Studies (Brachypodium)
Brachypodium, is closely related to wheat and barley but with a high quality genome sequence (International Brachypodium Initiative. Nature. 2010). It has a short stature, performs well in our growth chambers, and is maintained as self-pollinating inbred accessions that are widely shared in the scientific community. Collections have been made from hundreds of sites across the Northern Mediterranean range (Fig. 4C) and 51 accessions have recently been deeply sequenced by the Joint Genome Institute. To enable GWAS in Brachy, we are working with international collaborators (Pilar Catalan, Sam Hazen, Todd Mockler, and John Vogel) on a 1002genomes.org project to characterize species wide genetic variation. Reduced representation Genotyping by Sequencing (GBS), as we and others have published (Morris et al, 2011, Lu et al 2013) was used to type ~400 lines resolving isolation and admixture among western Spanish and eastern Turkish collections. This is the first step to building regional association mapping lines (RegMap, Brachi et al, 2011, Horton et al, 2012).
Once a balanced diversity mapping set is selected, the next step is to capture common haplotype diversity by sequencing 300 lines at a genomic depth required by LD (~1-2X for 100-200kb). Shotgun based, low coverage and low cost (~$40/sample) sequencing of accessions is possible by imputation based on linkage disequilibrium and the 51 reference genomes (Huang et al, 2011). GWAS for photosynthetic and developmental light response traits, replicated across same four dynamic light conditions, will be performed using this core Brachypodium hapmap set as described above for Arabidopsis AIM1.1.
Brachypodium has also naturalized in Australia (Fig. 3D) and the Borevitzlab has recently made new collections from more than 40 locations along a transect from Canberra to Adelaide. New collections and sequencing are proposed in AIM 2.2 to define a second balanced Australian hapmap set. In year 2, GWAS will be performed on this population under similar, but locally modified, dynamic light conditions to identify and confirm adaptive light response QTL.
AIM 1.3 Identifying regulators of light acclimation by novel suppressor screens
Strategy: A suppressor screen of light sensitive mutants, stn7, pgr5, and TLP18.3 under fluctuating light will identify regulators that increase the capacity of plants to survive fluctuating light stress
A key challenge in photoprotection has been identifying regulators. To date nearly all mutants have been a loss of function and were in structural components of the different photoprotective mechanisms. The novelty of this strategy arises from the application of fluctuating light to exemplars of each process (NPQ – npq4, cyclic electron flow – pgr5 and state transitions – stn7 and PSII repair cycle – TPL18.3) (Tikkanen et al 2010, Suorsa et al, 2012, Sirpiö et al,, 2007). Under fluctuating conditions stn7 and TPL18.3 growth is impaired and pgr5 is lethal, but all exemplar mutants grow normally under static light intensities. Thus, a revertant screen for suppressors that enables growth under fluctuating light selects readily for gain-of-function alleles and genes of the key regulators for each mechanism. We expect mutants to include restoration of the impaired mechanism (as for szl1 supressing npq1, see Li et al 2009) or a compensatory increase in different photoprotective mechanism(s).
The lethal fluctuating light sensitized genetic background provides a unique opportunity to explore the pathways required for dynamic light sensitivity. Aro’s group (led by Tikkanen and Suorsa) have developed mutagenized populations of stn7. Additional mutagenized populations of pgr5 and TPL18.3 are be prepared during 2013 and 2014. ANU groups will screen for revertants and undertake the gene identification using next generation sequencing (Schneeberger et al 2011). Confirmation and analysis will be performed in Aim 3.3). Indeed, Pogon’s group has sequenced 14 revertants on a different project in the past 2 months (see progress report for FS100100085).
Figure 3. Arabidopsis mutants and their phenotypes under dynamic light. All mutants have a wild-type appearance and growth rate under static low or high light for germination, growth and seed collection. Excitingly, growth slows or stops (stn7) or the plants die (pgr5) under shifts to fluctuating light (Tikkanen et al 2010, Soursa et al 2012). This makes for a rapid, ideal suppressor (revertant) screen.
AIM 2 Determine adaptive alleles via landscape genomics and field studies
Landmark papers for natural variation in Arabidopsis surveyed hypocotyl length among hundreds of accessions grown under different wavelengths (Maloof et al, 2001, Borevitz et al, 2002, Botto and Smith, 2002). Our studies unexpectedly found new change of function protein substitutions in the Phytochrome A photoreceptor. They also found a clear relationship of white and red light response, with latitude of origin, implicating natural selection as shaping components of this important trait. Now, our recent studies have resolved genetic basis for phenotype by environment associations. Here, environmental associations at QTL alleles stand out among other loci (Baxter et al, 2010, Li et al 2010). This is being formalized by new landscape genomic approaches that utilize population sequence information, and broad environmental data, to directly scan for associations that result from local adaptation, while controlling for background population structure (Coop et al, 2010). The rationale for this is that SNP alleles tightly linked to locally adaptive traits will provide a selective advantage. Over time, the geographic distribution of adaptive alleles should mirror environmental gradients, because selection on adaptive traits has filtered the alleles controlling them. Thus, landscape genomics is complementary to GWAS that directly identifies trait loci, without knowing if they are adaptive. QTL showing an environmental signature are adaptive.
AIM 2.1 Landscape Genomics for light adaptation using public climate data and current populations
In Arabidopsis, >4,400 accessions, spanning considerable geographic range, habitats, and light environments, have been genotyped at high density (Fig. 4A, Horton et al 2012). We will apply landscape genomic methods (BayEnv, Coop et al 2010), and other emerging ‘multitrait’ methods, to test SNP alleles for an association with latitude, light intensity, proportion of cloudy days (Fig. 4B), along with other available climate data (Hancock et al, 2011, Fourier-Level et al 2011), to identify candidate adaptive alleles and genes. Vegetation cover (NDVI) will also be investigated as a covariate at sites where this information is current and perhaps more reliable. We will then look for linked candidate genes and enrichment of associations in known light response pathways. If landscape genomics identifies the same loci that we identity through GWAS (AIM 1.1), then this will be strong evidence for the identification of a QTL conferring local adaptation.
Figure 4. Population samples cover a wide range of light environments. A) European subset of >4400 genotyped Arabidopsis accessions (Horton et al, 2012). B) Average annual cloudy days (Banta et al, 2012). C) Mediterranean Brachypodium core accessions (www.brachypodium.org) D) Available Australian Brachypodium populations red dots (Borevitzlab) and solar radiation variation in the driest quarter (ala.org.au).
Light response landscape genomics will be applied to current populations Brachypodium. This includes the Mediterranean set (Fig. 4C) and preliminary data from Australian collections (Fig. 4D) genotyped and phenotyped under AIM 1.2. These lines encompass a wide range of light environments and growing seasons across the world providing the opportunity for repeated environmental filtering on standing genetic variation and/or independent selection on alternative alleles and genetic loci. If QTL identified by GWAS overlap with loci associated with local light environment, this will again be strong evidence for adaptation. In addition, critical conserved pathways involved in light response adaptation may be jointly identified in Arabidopsis and Brachypodium such as photoreceptors and early signalling genes.
AIM 2.2 New diverse Brachypodium collections across light environments in Australia
The Australian Brachypodium collections give us a unique local opportunity to resample key populations spanning light environments (Fig. 4D) that are observed to be genetically diverse and segregating light response QTL. In 2015 PhD student Jared Streich will make new collections of ~100 accessions each from ~6 populations. Individuals within populations will be selected based on local light variation (e.g. open field vs. shaded field margin Fig. 1D). Low pass sequencing for 600 lines will be performed to determine whole genome haplotypes and resolve recombinations events. Subsequently landscape genomics analysis will scan for loci associated with local and regional light environmental variation as in AIM2.1. This local sampling will provide high power to detect and confirm adaptive loci filtered by local light environmental variation.
AIM 2.3 Sandbox survival of Arabidopsis and Brachypodium
A final test of the adaptive value of light response will include competitive growth trials in an outdoor location. Sandbox experiments on university campus have been successfully used to measure fitness in Arabidopsis (Hancock et al 2011). Xiang Go, a Borevitzlab masters student, proposed this experiment in 2012. He identified 3 suitable pairs of sunny and shady sites and received approval from campus facilities to install and remotely monitor them. Jared Streich (Borevitzlab PhD candidate) has vigorous outdoor Australian Brachypodium plants starting to flower at one site. The experimental design will mimic the chamber studies with 300 lines of each Arabidopsis and Brachypodium planted into each of 6 sandboxes with 5x5cm spacing. Hourly growth, rosette diameter, bolting status, and number of siliques/heads, produced will be observed by high resolution time lapse (gigavision.org), similar to our growth chambers. Positional coordinates define the genotypes in the same grid pattern as in the chambers. Remote environmental monitoring will record soil and air moisture and temperature and light intensity and quality every 5 minutes throughout the growing season. Although these outdoor conditions are more variable than the controlled light environments of the chambers we should be able to make predictions about the relative success of certain genotypes in sun vs actual dynamic shade environments. This has been shown previously but for continental scale climate variation (Hancock et al, 2011). It will be interesting to see whether adaptive loci identified from landscape genomics across a wider range, or QTL identified in controlled but dynamic conditions, are better predictors of fitness outdoors. It is also possible that entirely new QTL are found in our sandboxes.
AIM 3 Molecular characterization of new alleles and genes
In this proposal, the integration of traditionally separate strategies, provides the opportunity for a broad comparative analysis to distil conserved light response signalling pathways. Indeed, it is the power of this integration that is the breakthrough of this grant. Namely: GWAS results across populations and species is the first filter to identify ‘natural control points’. Landscape genomics is the second filter on the GWAS results to determine adaptive alleles and the real environments where they act. Gene expression analysis and expression QTL mapping (eQTL) sit at the interface between the GWAS and suppressor screen by highlighting the key genes and shared expression cascades. This is possible as the expression analysis includes natural populations, photoprotection mutants and suppressors (from Aim 3.3). Thus, eQTL, GWAS, and landscape genomics, combined with genes identified by the suppressor screens, enable us to prioritize the most novel and biologically relevant loci controlling plant response to dynamic light environments.
AIM 3.1 Dynamic light response gene induction and eQTL
Expression analysis under dynamic light conditions in mutant and natural mapping populations will reveal the shared expression cascades and conserved response pathways.
Transcriptional profiling is a powerful method to resolve the underlying signalling cascade of complex traits and environmental interactions. Here we propose high throughput RNAseq experiments on accessions, mutants, and recombinant inbred lines to reveal differential expression light response signalling pathways and identify causal candidate genes. With deep sequence coverage now available (>200M reads/lane) the critical issue is sample breadth to provide the biological replication needed to identify significant changes. Recently former colleagues developed a 96 sample/lane RNAseq library construction method (Kumar et al 2012) coupled with established RNAseq bioinformatic methods (EdgeR trimmed mean of M values, Robinson and Oshlack, 2010) that makes this possible. The first experiment will profile 8 genotypes (accessions and mutants), across 4 light environments, and 3 time points. These 96 samples will be replicated in 3 experimental blocks. After normalization, we will fit a model to identify genes differentially expressed by genotype, environment, and time, and as well as key interactions. This will yield critical response pathways, molecular control points, and when light induced expression level polymorphisms map to QTL, candidate genes. The second experiment will profile a well-studied mapping population (Cvi X Ler 144 RILs, Borevitz et al 2002, Botto and Coluccio 2007) before and after dynamic light treatment. This will allow us to genetically map gene expression variation and identify master regulatory loci controlling dynamic light response, aka expression QTL (eQTL) mapping as we have previously performed (Zhang et al 2011). Together these studies will identify candidate genes and the transcriptional pathways under QTL control, revealing light adaptive signalling cascades.
AIM 3.2 QTL candidate gene confirmation
Beginning in the final year and continuing after this grant, QTL identified by GWAS will be characterized and confirmed in both Arabidopsis and Brachypodium. This will be done by examination the list from gene expression studies AIM 3.1 and further genetic mapping. This will be performed through an evaluation of select hybrids and F2s that differ at candidate QTL using 5-10 F2 families and 24-48 plants each. They would then be phenotyped and genotyped and subjected to nested association mapping (Buckler et al, 2009). Once QTL are confirmed, often at the genomic resolution of 5-20kb, candidate genes will be selected. Knock out lines will be ordered and used for quantitative complementation and as a null background for the introduction of transgenic alleles. Knock out lines are available for most genes in Arabidopsis and the knock out collection in Brachypodium is expanding rapidly. Finally, the function of causal quantitative trait nucleotides (QTN) will be verified in transgenic plants by substituting specific allelic variants.
AIM 3.3 Molecular, biochemical and physiological characterization of key genes and alleles that modulate the dynamic light response in natural and mutant populations
We expected to identify 20-30 revertants and candidate genes in AIM 1.3. The ‘traditional’ strategy would be to subject these to genetic complementation, perform next generation sequencing on segregating pools (Schneeberger and Weigel, 2010) to identify the causative mutation in perhaps 10 genes. This would then be sequentially followed by detailed biochemical and physiological characterization of each mutant. Rather, we will utilize data sets from Aims 1, 2 and 3.1 and 3.2 to enable prioritization of the 20-30 suppressors which are biologically relevant under dynamic light. Key regulators will then be subject to detailed genetic, molecular, biochemical and physiological analysis. Furthermore, the mutant screen will improve the power of GWAS, that is, loci that only are marginally significant would be normally discounted, but if a marginal GWAS locus co-locates with a suppressor locus and there is an underlying change in expression in the eQTL analysis then this locus will be analysed. Aim 3.3 will commence during the funding period, but depending upon the level of funding and rate of progress will extend beyond the current scope of this proposal.
1. Genetic and Molecular. Standard complementation experiments, expression analysis, subcellular targeting will be performed. All are routinely undertaken by the Pogson, Botto and Aro groups (Gangappa et al 2013, Soursa et al, 2012, Estavillo 2011, Millar et al 2009, Botto and Coluccio 2007). This will include mutant and ecotype complementation experiments by different alleles and mutants of the key genes. ChIPseq and Proteomics experiments will follow up targets identified by RNAseq. This will be led by collaborating lab of Eva-Mari Aro who is establishing these methods now and they will be ready when the grants starts in 2014. The Pogson group also has expertise in histone modifying enzymes and ChIPseq (Cazzonelli et al 2009).
2. Biochemical and Physiological Analyses. The exact experiment would depend upon the gene identified and its subcellular location. If for example it is a chloroplast targeted kinase then its function in photoprotection and photosystem assembly and performance under different light regimes would be analysed. Again our groups have expertise in protein complexes, protein modifications, enzymatic activity (Tikkanen et al 2010, Soursa et al 2012, Estavillo et al 2011, Albrecht et al 2010, Tanz et al, 2012).
3. Photosynthesis and Photoprotection. A significant aspect of this proposal is to target genes that increase photoprotective capacity under dynamic light and ideally without a yield penalty under static light. Key analyses will include: electron transport rates, max photosynthetic efficiency, NPQ (qI, qE etc), state transitions, and cyclic electron transport using the PSI trayscan for high-throughput, plus LICOR 6400s for gas exchange and PAMs, dual PAMS and imaging PAMs for detailed fluorescence analyses. Aro’s group are world leaders in such analyses and Pogson’s group have published significantly in this area as well (for example Tikkanen et al 2010, Soursa et al 2012, Estavillo et al 2011, Albrecht et al 2010, Tanz et al, 2012 and Gordon et al 2012).