Skip to main content

Exploring natural variation of Shade Avoidance Syndrome in Arabidopsis using high-throughput phenotyping and Genome-wide association studies

Final Report Summary - SAS-ARABIDOPSIS-GWA (Exploring natural variation of Shade Avoidance Syndrome in Arabidopsis using high-throughput phenotyping and Genome-wide association studies)

Project context and objectives

Higher plants adapt to a changing environmental condition, in particular to their light environment, by altering their morphology accordingly. Understanding the molecular basis for this plasticity is an important aspect in future plant breeding to maintain high-yielding varieties for changing environments. The aim of the project was to study growth responses of a large range of Arabidopsis genotypes (mutants and accessions) in response to light competition, which is also denoted as shade avoidance syndrome (SAS). Our understanding of the molecular mechanisms of SAS has mainly come from studies on the hypocotyl (embryonic stem) of Arabidopsis plants. However, SAS responses of older plants, in particular of leaves, have been investigated with much less attention. Though this aspect is of great importance towards an understanding of yield formation in dense canopies. In this we wanted to quantitatively measure SAS traits on rosette-stage Arabidopsis plants using a custom-built high-throughput phenotyping device (Scanalyzer HTS). The aim was to compare our molecular knowledge of SAS responses of hypocotyls with leaves and to revise the known molecular signaling network. We further wanted to study natural genetic variation in field-collected Arabidopsis accessions and to potentially discover novel components in the SAS signaling network of leaves.

Description of project work

The first goal of the project was to establish the phenotyping pipeline from ‘seed to data’, which would meet the throughput demands of the project. Therefore a growth protocol was established to grow plants in a highly standardized environment, which was essential to minimize plant-to-plant variability and to compare data between experiments. Next a transformation matrix was derived to convert the laser-scanner images into 3D data, which in turn was used for a quantitative analysis (Fig. 1). Software tools in Matlab were developed to process 3D point clouds measured for individual rosette-stage plants. Two traits were computed for each plant related to mean leaf angle and plant size. This project needed the screening of hundreds of genotypes, which required proper automation of data flow from the originally measured laser scanner images to measured traits in a dedicated database. To do this we established collaboration with the Swiss Institute of Bioinformatics. Together we have developed Matlab and Pearl scripts to automatically transfer and process the data, which is ultimately stored in a MySQL database. This database is online (http://plantgrowth.vital-it.ch) and all published data will be made publically available. The whole method including the imaging pipeline, scripts and transformation matrices has been published in the journal Functional Plant Biology (http://dx.doi.org/10.1071/FP12018). Once the analysis pipeline was set-up we used it to phenotype 145 field-collected accessions and their Shade avoidance response.

For the Genome Wide Association (GWA) to link the observed phenotype to specific genomic loci we established a collaboration with Zoltan Kutalik who was post-doctoral researcher in the Computational Biology Group of Sven Bergmann http://www2.unil.ch/cbg at that time. Meanwhile Z. Kutalik established his own Statistical Genetics group as Assistant professor at the University of Lausanne http://www3.unil.ch/wpmu/sgg/. Work on this has already been initiated and data analysis is still on going. In parallel to the GWA analysis we developed a new image-processing algorithm to analyse growth and movements of individual leaves (see below).

Additional project work

During the development of the image-processing algorithm, it became apparent that the 3D data was well suited to monitor growth and movements of individual leaves. Phenotyping techniques to non-destructively measure leaf growth and movements with enough throughput for large-scale experiments such as genetic screen have not yet been developed. Yet it was very promising that our approach could meet these demands and open up new ways of studying leaf development. Work on the leaf-specific image processing started in January 2012. In Mai 2012, Olivier Michaud joined the group as Master Student under the supervision of the Marie-Curie-Fellow Tino Dornbusch. The master thesis was specifically dedicated to characterize growth and movements of individual leaves. It contained computational as well as experimental work. Both researchers have worked closely together and developed a semi-automated algorithm to measure growth characteristics of individual leaves. The work has led to a highly successful master thesis of Olivier Michaud defended in March 2013. Based on this excellent work, O. Michaud has been offered a follow-up PhD project in the group of C. Fankhauser to continue the initiated work on Arabidopsis leaf development. Results of this joint work have been presented at several international conferences and are currently prepared for publication in a top-tier scientific journal and expected to be published soon.

Results and their potential impact and use

Shade-avoidance of rosette-stage plants

The key hypothesis of the project was that the genes PIF4/PIF5 and SAV3 act in different SAS pathways leading to different growth in response to a shade treatment. Note that PIF4/PIF5 are transcription factors involved in regulating growth and SAV3 is an enzyme involved in auxin biosynthesis in leaves. This hypothesis was derived from data on hypocotyl length as phenotypic marker. However, our analysis showed that role of these genes for the shade-avoidance response of leaves differs from their role in hypocotyls. Showing that the gene regulatory network controlling shade-induced hypocotyl versus leaf growth differ is very interesting and justifies our choice of studying this important adaptive response in leaves. Work on shade avoidance in leaves is currently being followed up by Olivier Michaud and Dr Mieke de Wit and expected to be published soon. Understanding the mechanisms underlying shade avoidance in leaves is of potential impact on plant breeding in the future. To obtain high yields, plants need to be grown at high density and in these conditions shade avoidance leads to yield losses due to competition and lodging.

Natural variation

Our knowledge on understanding SAS or plant competition is increasingly becoming better, but the whole underlying dynamic genetic network is still not yet described in sufficient detail for the model plant Arabidopsis. In addition to testing the function of genes selected based on their function in the control of hypocotyl growth by shade (see above), we performed a screen of 145 field-collected accessions from worldwide collection sites. We found considerable variation between accession regarding both traits: leaf angle and rosette size (Fig. 2). Preliminary GWA analysis (collaboration with Statistical Genetics Group) has identified interesting loci, which were correlated to the observed phenotype with high statistical probability. Interestingly all hits (genetic loci) were in the coding region of specific genes. We are currently following up these genes by aligning and comparing their sequences to look for nucleotide changes that may have functional consequences. The most promising hits will then be functionally studied.

Phenotyping of growth and movements of individual leaves

The developed image-processing algorithm based on laser-scanning data allowed simultaneously measuring leaf growth as well as up- and downward leaf movements. Both, growth and movements followed a rhythmically oscillating diurnal pattern. Growth was minimal during the night around zeitgeber time 20 (ZT20) and peaked in the morning at ZT2-4. Leaf angle was minimal shortly after dawn around ZT2 and reached a maximum in the evening at ZT13-14. Supported by further experiments we could show that growth and movements are inversely correlated. During phases of increasing growth rate we observed a decrease in leaf angle and vice-versa. The precision of our laser- scanning data also allowed to monitor leaf circumnutations, which are helical movement of growing leaves around the growth axis with a period length of 1-2h and amplitudes of 5-10°. At the beginning of this project we did not foresee that the precision of our measurements would allow us to monitor nutation. These plant movements have been described since Darwin but their function and the mechanisms underlying nutations remain unknown. Hence our laser scanning approach on individual leaves now opens up the unique possibility to address many future question on leaf development for which proper phenotyping techniques have been missing up to now.