Characterising new roles for APOBEC4 and ADAR deaminases

Abstract

Deamination or the hydrolytic removal of one hydroxyl group from a base in DNA or RNA can lead to changes in the transcript and protein produced. Examples of this are the deamination of cytosine residues in DNA by activation induced deaminase (AID) during antibody diversification, or deamination of adenosine at the Q/R site in the GluR-B transcript by adenosine deaminase acting on RNA 2 (ADAR2), which regulates calcium permeability in neurons. The initial focus of my thesis was to characterise a putative novel deaminase APOBEC4. APOBEC4 was identified in a bioinformatic search for proteins containing the core catalytic residues common to the whole family of Cytidine Deaminase enzymes. The aim of the project was to express and purify recombinant APOBEC4 for in vitro characterization, however despite using different expression systems and purification conditions the majority of the recombinant protein was inherently insoluble and I could not isolate sufficient amounts of protein for further studies. Recombinant protein with a GST-tag was used to generate polyclonal antibodies which recognised recombinant protein but were unable to detect endogenous APOBEC4. The focus of my thesis then changed to the process of adenosine to inosine editing in RNA, which is a post-transcriptional mechanism for generating protein diversity. The enzyme family responsible for catalysing this reaction is known as ADAR, and Drosophila melanogaster has only one Adar gene. Flies lacking the Adar gene show locomotion defects and age-dependent neurodegeneration, however little is known about the molecular mechanism underlying these defects. To investigate this phenotype I performed microarray analysis on RNA isolated from heads of 5 day old flies lacking the Adar gene to characterize gene expression changes in the fly heads before neurodegeneration caused secondary effects. Analysis was also performed on Adar-null flies expressing either an active Adar gene or a catalytically-inactive Adar gene in cholinergic neurons to determine which transcripts could be directly regulated by Adar. I confirmed the microarray results by real-time PCR, and demonstrated that the changes in transcript level could be reversed by expression of either active or catalytically-inactive Adar. Expression of edited transcripts did not change dramatically. Filter-binding analysis and electrophoretic mobility shift assay revealed that recombinant ADAR could bind to all RNA transcripts analysed with similar affinity; both known substrates and potential new substrates for Adar, as well as transcripts that were chosen as negative controls due to their expression not altering in the expression microarray. Recombinant ADAR bound to dsRNA with a very high affinity; other transcripts investigated bound with considerably lower affinity, yet all transcripts investigated were bound by ADAR. Further analysis of transcript changes in Adar-null flies was investigated by performing microarray analysis with a custom-made splicing-sensitive microarray. Analysis revealed that a subset of transcripts were differentially spliced in Adar-null flies; however this group of transcripts was distinct from the group identified as being altered on the expression microarray, indicating that the splicing changes are independent of changes in expression. Analysis of exon-specific probes on the splicing array confirmed the transcript changes identified in the expression array. Real-time PCR confirmed the changes in splicing, and these transcripts were further examined by sequence analysis. This revealed several transcripts identified as altered by the AS array showed use of alternative polyadenylation sites indicating ADAR may have a role in detemining polyadenylation site selection.