Jan Kerschgens
The mis-regulation of gene transcription in the living cell can be the cause of complex human diseases like cancer. One challenge of the post-genomic era of biomedical research is the development of in vitro techniques for the rapid and cost-effective analysis of regulatory transcriptional patterns of disease. Two crucial points for this endeavour have to be considered: (1) Sensitivity, since biologically active transcription factors are expressed at low amounts in the nucleus of the living cell and (2) the as close as feasible reconstitution of the nuclear microenvironment in order to guarantee as close to natural conditions for the performance of experiments as possible. The first part of this thesis reports on the application of protein-binding microarrays (PBM) bearing long-stranded DNA molecules for the analysis of a human disease-related protein, the human tumor suppressor activating enhancer-binding protein 2 alpha (AP2α). The second part of this thesis describes how chromatin-immunoprecipitation combined with high-throughput sequencing (ChIP-seq) and subsequent data analysis led to the discovery of a novel AP2α-binding motif in a cancer cell line model. We first developed a setup and protocol to perform on-chip protein-DNA molecule interaction analyses. Following this, recombinant AP2α protein was produced and purified using an E.coli host system. AP2α was subsequently assessed for its binding specificity to an array of ∼6000 human promoter and intergenic sequences. These experiments confirmed previously established AP2α DNA binding sequences and allowed the identification of novel ones. High and low binder sequences were selected from the PBM data and analyzed for specific AP2α binding in vivo utilizing reporter transfection and chromatin-immunoprecipitation (ChIP) assays. In contrast to the lowest AP2α-bound sequences, the highest PBM-bound sequences had significant effects on reporter gene expression. ChIP-qPCR revealed these sequences were bound by AP2α in vivo. From these results Kallikrein 5, a member of the kallikrein family of extracellular proteases that includes the prostate-specific antigen (PSA), which is currently emerging as one of the most prominent biomarkers of tumor progression for various types of cancers, and the growth-arrest specific 2 (GAS2) protein, which modulates cell susceptibility to p53-dependent apoptosis, were identified as novel AP2α targets. Finally, the binding patterns of AP2α from nuclear extracts of breast cancer tissue biopsies were assessed on the 6k human promoter arrays and compared to the binding patterns of 'healthy' tissue extracts. Overall these experiments confirmed known targets of AP2α. Interestingly a discrepancy of 26% was revealed when assessing the correlation of binding data from healthy and tumor tissue biopsies. This finding suggested that depending on the source of the protein, that is of either healthy or tumor tissue origin, the DNA-binding specificity of the protein factor under study might be altered and most importantly this change in binding specificity can be monitored by PBM. The second part of this thesis describes the analysis of the binding specificity of AP2α by combining chromatin-immunoprecipitation with high-throughput sequencing (ChIP-seq). The binding specificity of the human tumor suppressor protein AP2α in the living cell was assessed in a colon cancer cell model. Analysis of the ChIP-seq enriched sequences resulted in the discovery of a novel recognition motif for the in vivo binding of AP2α. Interestingly, this sequence motif resembles the motifs discovered from data sets with recombinant protein alone. Further analysis of a previously described in vitro and the new ChIP-seq discovered binding motifs were performed by electrophoretic mobility shift assays, which demonstrated that the newly discovered motif shows the same effects with colon cancer and breast cancer nuclear extracts, thereby suggesting a general relevance of the motif independent of the cell type. In these DNA-binding competition experiments (EMSA), the previously computed, in vitro derived DNA recognition sequences does not abolish native AP2α protein (obtained from crude nuclear cell extracts) interaction with the ChIP-seq discovered motif. This observation indicates for a complex situation in vivo that might require protein co-factors in order to allow specific AP2α-DNA interaction. Therefore additional experiments will be required in order to assess the novel motif further. In conclusion, this thesis work demonstrates that protein-binding microarrays (PBM) can be used for the analysis of binding patterns of a human disease-relevant protein. It could be shown that the PBM setup developed was capable of identifying novel cancer-linked target genes for AP2α. It was also shown to be sensitive enough to directly probe nuclear extracts from tissue biopsies for specific DNA-binding. Finally, a novel DNA-recognition motif for AP2α was discovered from cellular assays. Additional experiments are required to assess whether this motif can be utilized for the reliable assignment of novel probes that may subsequently be integrated onto protein-binding microarrays. Together with the analyses for additional protein markers of disease, this avenue might serve as a platform for developing PBM technology further towards biotechnological applications such as diagnostics of regulatory transcriptional markers and patterns of cancer.
EPFL2009
