Elucidating mechansims and roles of alternative polyadenylation

Our genes are made of DNA, but when they are switched on, copies are made in a related molecule called RNA and this RNA goes on to code for the protein products of our genes. As the gene is copied into RNA, the RNA is cut and a string of Adenine molecules (A for short) are added at the end. This so-called 'poly A tail' functions to protect the RNA from being degraded, and helps to transport the RNA around the cell and stimulates the formation of protein from the RNA. The site at which the poly A tail is added is not always the same, even for the same gene. For example, half of all human genes have RNAs with more than one site for adding a poly A tail. Controlling the site at which the poly A tail is added is very important because it ultimately affects how genes function. However, this is a process we know surprisingly little about. It's not just human RNAs that have different poly A tails, other animals and plants do too. We have been studying how plants control the time at which they flower, a process where genes are very precisely controlled. In the course of this work, we have discovered that three factors called FCA, FY and, most recently, FPA, function to control poly A site selection of some RNAs. Such basic aspects of gene expression are very similar in plants and animals and it turns out that there are human proteins highly related to FY and FPA. It is possible therefore, that these proteins control poly A site selection in humans too, but very little is known about them. As we have found that FCA and FPA don't need each other to control poly A site choice, we think they must be doing this in different ways. This gives us a chance to understand how poly A site choice can be controlled. In this proposal we plan to build on what we know about FCA and FPA in plants, but this knowledge should be of much more general interest. We want to know two things: (1) How do FCA and FPA control the site at which a poly A tail is added (2) What genes do FCA and FPA regulate by controlling alternative poly A site choice? We will work out how FCA and FPA control poly A sites by identifying the features of the RNA required. This should be quite straightforward. We will make test genes containing different parts of the target gene and see how they affect poly A site selection when placed back in plants. In order to find the other genes whose normal poly A tail depends on FCA and FPA, we will look at where RNAs are polyadenylated in normal plants and in mutant plants that lack FCA or FPA. It is now possible for us to look at nearly all the RNAs in a cell thanks to Next Generation Sequencing, a technology that is revolutionizing modern biology by giving us huge amounts of sequence data, very quickly and at a fraction of the cost to before. This technology has been developed to look at RNA by sequencing a short part of every RNA, sufficient to identify it, called a 'tag'. To find the tag, scientists use the poly A tail and sequence what is next to it. This is a happy coincidence for us, because it means that in addition to tagging a particular RNA, this method also tells us where a poly A tail has been added to RNA. To analyse the large amounts of data and make comparisons, we will need to develop specialized computational tools. Because we already know genes where FCA and FPA control poly A site selection, we should be able to find changes in these 'tags' if our tools are working well. Once we are sure they are, we can look for other shifts in 'tags' to identify other genes controlled by FCA and FPA. As lots of other scientists are also using this sequencing technology, but for completely different reasons, we can use our analysis tools to look at changes in polyadenylation in their data too. In this way we will be able to identify cell-types and situations where alternative polyadenylation is an important part of gene regulation.

Alternative polyadenylation (pA) is a commonplace, but surprisingly poorly characterised aspect of eukaryotic gene regulation. We have discovered that the Arabidopsis RNA binding proteins, FCA and FPA (regulators of flowering and RNA silencing) function genetically independently to control the site of RNA 3' end formation. Our work is unique because no other trans-acting regulators of RNA 3' end formation, that are not components of the splicing or polyadenylation machinery, have been identified. Our discovery therefore provides a genetically tractable system to dissect alternative pA. We will first define the level at which FCA and FPA control pA site selection through cis-element analysis, which we will couple with PolII ChIP and transcription run-on assays. A combination of RNA immunoprecipitation (which we have successfully developed) and RNA specificity-swap assays will be used to determine how directly these proteins regulate 3' end formation. Next, a genome-wide identification of pA sites regulated by FCA and FPA will be made by utilising recent developments in next generation sequencing: Fortunately for us, Digital Transcriptomics (DT) involves sequencing short 'tags' of RNA adjacent to pA tails, thereby providing positional information on pA sites. We will develop a bioinformatics pipeline to analyse DT data to quantify changes in pA site selection, working first with different FCA and FPA genetic backgrounds that we know have contrasting patterns of pA. With these tools in place, we will be able analyse other DT data releases and associate alternative pA with diverse backgrounds, cell types or treatments. Our work will have widespread impacts in understanding gene regulation because it will define mechanisms by which alternative pA can be controlled, clarify the connections between 3' end formation and RNA silencing and establish generic bioinformatics tools to identify alternative pA in next generation sequencing data.