Molecular mechanisms of signal transduction and cellular control
Signal transduction and the corresponding molecular pathways play fundamental role in cell biology and defects in these pathways are underlying causes of many pathological conditions. One of the long-term goals of my research is the elucidation of the mechanisms by which mitogen-activated protein kinases (MAPKs), one of the central players in eukaryotic signal transduction, regulate cellular processes such as differentiation, cell proliferation, polarization and cytoskeletal dynamics. A particular focus of my research is protein-protein interaction and protein phosphorylation.
In eukaryotic cells MAPKs engage in a vast and as yet unmapped network of molecular interactions which facilitate compartmentalization and association with protein substrates. In collaborations with groups from US and UK we are conducting ongoing studies to identify novel MAPK interaction partners and phosphorylation targets and to elucidate the corresponding molecular pathways. We use models based on the yeast pheromone pathway and cultured mammalian cells. We employ post-genomic approaches and methods based on affinity techniques and mass spectrometry as well as imaging, and molecular approaches such as the yeast two-hybrid screen and phage display.
We use phosphoaffinity techniques, mass spectrometry and bioinformatics to identify novel targets of MAPK-dependent phosphorylation and then use classical biochemical and imaging techniques to study the mechanisms in question. We also rely increasingly on quantitative mass spectrometry to study the dynamics of MAPK-dependent phosphorylation of the proteins that have been identified by system-wide screens.
We believe that our basic research on cell signalling will eventually result in applied biomedical developments as well – through the identification of novel protein targets and phosphorylation events and protein-protein interactions that might be deregulated in disease. It is for this reason that many researchers expect that phospho- and interaction proteomics will become one of the preferred approaches for discovery of novel drug targets and disease biomarkers in the near future.
Translational and clinical proteomics
Over the last 5 years my group has developed a collaborative program for translational research that aims to identify novel cancer drug targets and protein biomarkers by mass spectrometry based studies of tumor specimens. In collaboration with oncologists and pathologists from UK, Australia and Germany we apply quantitative methods to define the proteomes of different types of breast and colorectal tumors.
To achieve this we have created a proteomics pipeline for large-scale analysis of tumor samples, which uses shot-gun as well as top-down approaches. So far we have accumulated several millions of peptide MS/MS spectra from different types of tumors and from tissue peripheral to the tumors. The dataset allows the identification of more than 10,000 proteins and facilitates the comparative analysis of their abundance in tumors, surrounding tissue and non-cancerous control samples. The dataset is constantly growing with new samples coming from our clinical collaborators. The analytical pipeline is based on the interactive CPAS software developed at Fred Hutchinson Cancer Centre which runs on a dedicated server in my lab and is accessible through a web portal. It uses a 40-node cluster to run the protein identification search engine.
Concomitant with the proteomics data acquisition and analysis we perform bioinformatics analyses that use public domain gene expression data and systems biology tools to further characterize the proteins identified as promising drug target or biomarker candidates. For this we collaborate with statisticians and bioinformaticians from the Department of Mathematics at University of Essex.
One of the long-term objectives of our translational research is to create a quantitative atlas of protein expression in tumor tissue, which not only lists the identities of the present proteins, but also defines their relative abundance in different types of tumors. Such a quantitative tool can be very useful for drug target and biomarker discovery and validation, and for in silico studies of tumorigenesis, metastasis, and responses to therapy. It can act as a necessary complement of the available and coming tumor DNA array data, and whole-genome sequences obtained from individual tumour samples. It is now recognized that gene expression data alone cannot provide the necessary information for comprehensive analysis of complex biological systems. The abundance of the individual proteins in the different types of cells found inside the tumor and in the tumor microenvironment, and the post-translational modification state of these proteins are dimensions of complexity, which only proteomics can grasp.
In addition, we are working actively to extend the analysis to the phosphoproteomes and protein interaction networks of the tumors and to develop and implement higher resolution sample preparation strategies utilizing microdissection and cell sorting.
The field of mass spectrometry based proteomics has now matured. In the last few years a new breed of hybrid high-resolution and high-accuracy instruments have entered the labs and proteomics core facilities which now allow studies at an analytical depth of more than 1,000 proteins per run and a throughput of tens of runs per day. We firmly believe that the combination of this analytical power with new and improved data analysis algorithms will eventually allow proteomics to fulfil its promise in the field of drug target discovery and disease diagnosis and monitoring. We believe that we will be one of the groups that will contribute substantially to this process.