Exploiting computational modelling to study comparative leaf development

A key challenge in biology is to understand how different organisms come to have different forms. In plants this variation in form is obvious in the many different leaf shapes we see when eating a salad or walking in a park. For example, spinach has simple leaves whereas parsley has complex, subdivided leaves. Leaves are also interesting to study because they play a key role in the food chain being the main photosynthetic organs of land plants and thus responsible for CO2 fixation in terrestrial ecosystems. For these reasons, understanding how diversity in leaf form is generated is of considerable interest to scientists. To study this problem we work with the hairy bittercress (Cardamine hirsuta), which is a plant that has complex leaves subdivided into leaflets. The presence of leaflets makes this plant very different to its close relative the thale cress (Arabidopsis thaliana), which has simple, undivided leaves. We already know a lot about how a simple leaf shape is produced in thale cress because it is easy to do experiments with. Hairy bittercress is also very easy to work with in the lab, so we use it to understand how leaflets are produced and ultimately why this plant makes leaflets whereas its relative the thale cress does not. One important problem with such comparisons is that while we can identify individual proteins that might be responsible for generating the different leaf shapes of these two plants it is very difficult to understand by what 'rules' the 'sum' of all the possible processes that regulate shape is 'created' in plants and how this 'sum' influences the timing, position and direction of cellular growth to direct formation of different leaf shapes. To resolve this problem we will collaborate with computer scientists who by considering when and where particular proteins that influence bitter cress development are expressed, produce models that can help clarify what might be these fundamental 'rules' that govern leaflet formation and determine final shape. Additionally, we will improve our knowledge of how exactly hairy bittercress leaflets grow and this will involve two methods. Firstly, we will use a laser based visualization methodology that will allow us to capture images of leaves while they are growing without destroying them, and produce time lapse movies of their growth, not unlike those seen in David Attenborough movies except the structures we will be observing will be tiny. Secondly to directly observe the hairy bitter cress cells that divide to produce leaflets we will use a method that renders the tissue 'see-through' and will hence allow us to obtain three-dimensional images of the developing leaflets when they are still very small and inaccessible to dissection. This method, which is similar in its logic to medical tomography will allow us to directly visualize dividing cells and hopefully pinpoint the locations of cell division at successive stages of development. Information obtained from these two methods will be used to producer more accurate models of how leaves of different species end up having different shapes.

We aim to understand how species-specific differences in action of key developmental genes are translated into natural variation in leaf form. To this end we developed C.hirsuta, a dissected-leafed plant related to the simple-leafed model organism A.thaliana into a powerful genetic system, and we wish to understand why C.hirsuta unlike A.thaliana produces leaflets. Understanding how development of these two closely related species differs is challenging because initiating leaf primordia appear identical, but the two species attain very different final leaf shapes due to poorly understood differences in the location, timing and direction of cellular growth after leaf initiation. To address this problem we will develop an interdisciplinary research direction where computational modelling and leading-edge imaging technologies will be deployed in C.hirsuta to enrich developmental genetics approaches, thus allowing us to define the mechanisms underlying leaflet formation and hence species-specific elaboration of leaf geometry. Our recent data suggest firstly that only marginal cellular lineages contribute to leaflet formation and secondly that the small indolic hormone auxin and KNOTTED1 homedomain transcription factors promote leaflet formation by acting in a negative feedback loop. Together with P. Prusinkiewicz we propose to examine whether formalized computational models based on these data can produce leaflet patterns such as those observed in C.hirsuta. We also propose to exploit Optical Projection Tomography and live Confocal Microscopy methodologies to directly study leaf growth pattern and gene expression dynamics and thus further enrich models that conceptualize the fundamental principles that distinguish leaf growth of the two species.