Implementation of a Bayesian Segmentation Algorithm to the analysis of receptor conformational changes in multidimensional single-molecule data

Intercellular communication is achieved via secretion into the blood of chemical messengers, an important class of which are the growth factors. The main function of growth factors is the promotion of cell multiplication. Without growth factors life is not viable, as they are required to orchestrate cell growth in a manner that maintains the integrity and function of each organ. To detect growth signals cells insert into their surface receptor proteins whose function is to bind the growth factors in the blood and then transduce, probably via a change in conformation and/or interaction with other receptors, the signal across the plasma membrane to the cell interior. When intracellular proteins recognise in the receptor the onset of signalling, they initiate the reactions that lead to the amplification of this signal, an ultimately to DNA replication and cell division. Recent years have seen an explosion of discoveries in growth factor receptors, including direct clinical applications such as Herceptin, as these receptors are linked to the development of a number of human tumours. This link is hardly surprising, as a malignant tumour is just the product of uncontrolled cell proliferation. Understanding the molecular mechanisms behind the delicate balance between growth-promoting and growth-inhibiting events giving rise to uncontrolled cell growth and malignancy is currently one of the biggest endeavours in science. Such events need to be understood at the molecular level and in the context of the living cell, as receptor signalling is by nature an unsynchronised event whose outcome for the cell depends on the particular characteristics of the cell environment and on which receptor partners are allowed or not in the vicinity. For this purpose we have developed during previous grants unique imaging instrumentation that allows the visualisation through a fluorescence microscope of the signalling behaviour of individual receptor molecules in a living cell. In this application we aim to implement a programme of work by which we will be able to analyse our single molecule receptor data using a state-of-the art image analysis algorithm originally developed to extract quantitative information from extremely faint, noisy and blurred astronomical data. In single molecule research, the combination of weak signals from single receptors with the unavoidable presence of the fluorescence background from living cells implies that essential information in the data can be blurred and/or hidden by the background noise. In the imaging analysis method we seek to implement, pre-existing information on the receptor data can be explicitly included in the image analysis algorithms, creating a much more powerful algorithm with demonstrated trustworthy results. In combination with these novel image analysis methods the instrument we have built will become an accurate spectroscopic ruler and protractor at the molecular scale to report on receptor conformational dynamics in the natural cell environment. It is hoped that the quantitative elucidation of the mechanisms of behaviour of normal and oncogenic growth factor receptors will ultimately help in the designing of new inhibitor molecules that antagonise the action of mutant or overexpressed receptor in cancer.

Growth factor receptor signalling is an unsynchronised molecular event mediated by interactions between receptors, structural reorganisations, the local cell environment, and the nature of the intracellular signalling partners. High-resolution observations in live cells are therefore required to understand receptor-mediated signal transduction. The multidimensional single molecule microscopy (MSMM) technique developed at Daresbury is beginning to allow the determination of crucial aspects of receptor signalling, such as the number and nature of interacting receptor partners, and/or the conformational changes in receptor complexes during signalling in the living cell. The proposed programme of work aims to apply a state-of-the art Bayesian segmentation image analysis method, originally developed to extract quantitative information from extremely faint, noisy, and blurred astronomical data, to the study of the molecular mechanisms by which growth signals are transduced across the plasma membrane by ErbB receptors. In MSMM, the combination of weak signals from single fluorophores, the need to partition signals to measure receptor distances and orientation, and the unavoidable presence of some fluorescence background from living cells implies that essential information in the data can be blurred and/or hidden by the background noise. We seek to implement an imaging analysis method in which we will integrate the wealth of a-priori knowledge of ErbB multidimensional single molecule data collected during previous work into the existing Bayesian segmentation algorithms. This will create a powerful algorithm expected to allow us to achieve in live cells a signal-to-noise at least as good as what it is currently accomplished in cell-free systems, and likely much better. This will open the field to very accurate measurements of angle, and inter- and intra-molecular distances, so far not accomplished in the living cell.