Quantitative and contextual modelling of transcriptional responses to auxin

Auxin is a plant hormone that plays an important role in many, very different aspects of plant growth and development. For example, on one hand auxin regulates the pattern in which leaves emerge from the growing shoot tip, while on the other it mediates environmental responses such as the bending growth observed in shoots growing towards a light source. Because auxin is integral to so much of plant development, a knowledge of how auxin works is absolutely fundamental to our ability both to understand how plants grow and to improve important crop traits. Although it is unclear how auxin, a small and simple molecule, is able to regulate such a diversity of processes, it is known that changes in gene expression are important and that several hundred genes are either turned on or off in response to auxin. Work from several labs has established a network of signalling proteins whose complex interactions translate increases in auxin levels within a cell into gene expression changes. However, this model of auxin signalling is generic and contains several gaps which mean that in most cases it offers only a theoretical basis for understanding how auxin operates throughout the plant. We can identify the new information required to remedy this situation and with this work we intend to use newly available techniques to obtain these data and correct the deficiencies of the current model. The complexity of the interactions in this auxin response system mean that the data that describe it are similarly complex, and therefore we need to adopt a new approach for analysing the information that we will gather in which we build mathematical models that describe the functioning of the system. These models can then be used to test our predictions of how the auxin response system works, generate new hypotheses about its functioning, and ultimately tell us if we have understood its essential details. This higher level of understanding of auxin action will be useful to a number of groups, including plant biologists, crop scientists, and scientists studying similarly complex signalling mechanisms in other organisms.

We are taking a context-specific and quantitative modelling approach to transcriptional auxin response. We will generate quantitative data sets that will parameterise mathematical models which will then be used to test hypotheses about the functioning of the auxin signalling mechanism. The main methodologies that will be used to achieve the specific objectives of the work are outlined below: Objective 1. Data describing contextual parameters To obtain cell specific information for auxin levels and transcriptional auxin response in the root we will use the combination of protoplasting and fluorescence activated cell sorting (FACS) of cell-specific GFP-marker lines. We will sample epidermal cell types both treated and untreated with the auxin, IAA and perform microarray and quantitative RT-PCR analysis of RNA extracted from these sorted cells. As well as performing the protoplasting/FACS, the Ljung lab will also quantitate IAA levels by GC-MS from samples of the same cells. Objective 2. Data describing interaction parameters We will quantify Aux/IAA-ARF and ARF-DNA interactions using surface plasmon resonance (SPR). His- and GST-tagged Aux/IAAs and ARFs will be expressed by in vitro transcription/translation in wheatgerm extract and the cis regulatory sequences of auxin-regulated genes will be amplified by PCR using a biotinylated primer to facilitate tethering the DNA to the SPR chip. Objective 3. Data describing abundance parameters Auxin-induced Aux/IAA induction will be measured by qRT-PCR and the auxin-induced destabilisation characteristics of the Aux/IAAs will be quantified by analysing Aux/IAA-luciferase fusion proteins. These data will enable us to model Aux/IAA protein abundance. Objective 4. Mathematical modelling Initially, models will be constructed using ordinary differential equations (ODEs) but partial differential equation (PDE)-based approaches will be adopted should?stochastic and/or spatial effects prove significant.