Regulation of alternative splicing by G-quadruplexes: molecular mechanisms and tools to manipulate gene expression

When RNA polymerase starts to transcribe a gene into mRNA, the sequence and thus the activity of the protein encoded by the RNA depend on the pattern in which large portions of the RNA are spliced out. The processes by which the sites of splicing are selected are very complex, and they are still understood poorly. Understanding them is hugely important both because splicing is an essential process that, more than anything else, enables highly complex organisms such as ourselves to have developed despite having only the same number of genes as much simpler organisms and because, by controlling splicing, we could shift the expression of a gene from one type of protein to another for therapeutic purposes. Indeed, the first drugs targeting splicing in muscle and the CNS have been approved recently by the US FDA and others are in trials.

There has been great excitement recently over the discovery that quadruplexes (G4s) might regulate splicing. G4s are small, four-stranded structures that can form in the RNA from four sequences of GG or GGG in close proximity. They could open up new ways of understanding and manipulating splicing. However, it has been very difficult to prove that they form in long RNA molecules in functional splicing conditions, and nothing is known of how they might affect splicing. We have recently published a new method, called FOLDeR, that enables us to map the regions of a pre-mRNA that form G4s in splicing conditions. We have applied this to Bcl-X, a gene expressing two isoforms of protein: one promotes cell survival and the other promotes apoptosis. The difference results from the choice between two 5' splice sites. We have shown that there are two G4-forming sequences in Bcl-X, one close to each splice site.

Many small molecules are known to bind to and stabilize G4s. We have tested a range of 33 of these on Bcl-X. Both in nuclear extracts and in cells, one reagent shifts splicing so much that the usually minor pro-apoptotic isoform becomes predominant, and we have shown that it affects the structures of the two G4-forming regions in the RNA, probably by enhancing G4 formation. Moreover, it switches the splicing of another gene crucial for some cancers, Mcl-1, to express only the pro-apoptotic isoform. Accordingly, it promotes apoptosis. The same reagent has no effects on some other alternative splicing events, and others that affect different alternative splicing events have no effects on Bcl-X. Most of the other 32 compounds show mild or no effects. Importantly, the effects of each one on Bcl-X splicing are similar in nuclear extracts with purified pre-mRNA and in cells, showing that the molecules affect splicing directly. This suggests that G4 stabilizers might each target a defined set of genes. Are these genes in sets with common biological functions? If not, could we investigate how the small molecules and their cognate G4s work so that we can prevent unwanted effects and develop useful and selective drugs? Could we predict the sites of action of such molecules? Could we use their target sequences to develop ligand dependent splicing switches, enabling a gene to switch from one function to another?

We propose to address these exciting possibilities using four approaches. (i) The first is use high-throughput sequencing to identify all the changes in expression and splicing brought about by several G4-binding molecules in cells, which will inform us about the range of effects and the common features associated with the targets of each molecule. (ii) We will identify and test the exact nucleotides and contacts associated with G4 formation and small molecule binding by methods including NMR and X-ray crystallography. (iii) We will synthesize and test a range of new analogues to help in defining interactions and, using structural information, to improve selectivity. (iv) We will determine how G4s affect splicing and use this to test whether G4s can be inserted as switches into new positions.

G-quadruplexes (G4) play crucial roles in gene expression but their function in RNA processing and alternative splicing in particular remain poorly understood and controversial. We have shown that the pre-mRNA of the Bcl-x gene contains G4s that influence alternative splice site choices and identified one chemical compound (GQC-05) that promotes the formation of the minor pro-apoptotic isoform of Bcl-x both in vitro and in cells. This is of importance because this molecule could be further developed to induce cancer cell deaths.
To better understand the function and specificity of GQC-05 in alternative splicing, we propose to:

1. Determine the effects of GQC-05 and other G4 ligands on transcription and alternative splicing transcriptome-wide using RNA-seq experiments from cells incubated with various G4 ligands. We will gain information on their specificity of action, correlate the structural features of the ligands with their effects in cells and identify correlations with the distributions of potential G4-forming sequences.

2. Provide a molecular description of G4 formation and binding of GQC-05 in Bcl-x pre-mRNA. We will identify the G nucleotide involved in G4 formation in Bcl-x by a novel method based on phosphorothioate nucleotide analogue interference (NAIM) and confirm the contribution of these nucleotides in alternative splicing by specific chemical modifications. We will determine the structural basis of the specificity of GQC-05 to Bcl-x pre-mRNA G4 sequences.

3. Synthesize GQC-5 analogues with novel properties. We will define a structure-activity profile of GQC-05 analogues to improve their affinity to RNA and specificity in alternative splicing regulation.

4. Decipher the mechanisms by which G4s ligands affect splice site selection by investigating whether they either perturb the binding of splicing factors with critical roles in splice site selection or promote their recruitment using biochemical methods and TIRF single-molecule imaging.

Underpinning this research is our new technology for identifying functional G-quadruplexes (G4s) in pre-mRNA and our evidence that (a) G4s do affect splicing in Bcl-X, (b) the splicing patterns of Bcl-X and Mcl-1 respond strongly in vitro and in cells to only one of a wide range of G4 ligands, and (c) we have shown that this ligand binds and affects the structure of the RNA around the known G4-forming regions.

Both Bcl-X and Mcl-1 are important determinants of cancer cell survival and the best ligand shifts splicing strongly to the pro-apoptotic (death) patterns. We will investigate this process in detail (the first mechanistic analysis of the actions of G4s on splicing) and use this knowledge to design, synthesize and test a suite of biologically active compounds and specific tools in collaboration with GlaxoSmithKline (GSK).

1. Potential Economic Impact of the Research
Therapeutics & tool compounds for biotechnology and pharmaceuticals: new opportunities for the development of a new class of splice-switching small molecules and probes of splicing; new methods of controlling gene expression by creating RNA motifs that provide ligand-inducible splicing switches; new routes for developing cancer drugs with higher selectivity targeting known sets of genes.
Outputs: Patent protection of intellectual property; exploration of licensees, develop new chemical biological methods for industrial applications with GSK.

Mechanisms of delivering Economic Impact: we will apply to the many funds, some administered by UoL and UoS, available for translational research (see PTI for details) and seek alliances with GSK via an EPSRC industrial CASE studentship and with CRT.

2. People and Skills Development

PDRA1 will learn RNA biochemistry (very rare in the UK) and transcriptomics. PDRA2 will learn a wide range of skills in structural biology of RNA, under-represented in the UK, and single molecule methods available only in ICE's laboratory. The synthetic PDRA3 will develop skills in reaction optimization and photo-crosslinking using dedicated high-throughput facilities in Dr Jacob Bush's laboratory. All three PDRAs will be mentored, appraised annually and disseminate their findings at major conferences.

Outputs: Enhanced industrial-academic collaboration and upskilling of post-doctoral researchers; International exposure to academic excellence by presenting at conferences.

3. Potential Societal Impact of the Research

The aim of this research is to understand how G-quadruplexes (G4s) and small molecules that bind to them (G4 stabilizers) influence gene expression by altering the outcome of RNA alternative splicing. Public engagement will enable broader appreciation of the importance of chemical biology approaches. Cancer - see above.

Outputs: Public discussion of the role of Chemical Biology in bioscience (Glasgow Science Centre; Glasgow Science Festival). Communication of our research in specific publications aimed at the general public as well as stakeholders and policy makers such as Horizon2020 portal.
Mechanisms of delivering Societal Impact: Delivery of presentations at the Glasgow Science Festival and Glasgow Science Centre events; dissemination of results through the mainstream press on UoS and UoL websites; industrial-facing workshops and symposia via GSK.

4. Academic impact

PDRAs will present their findings at national and international conferences, and SET for Britain in order to promote our findings to the scientific community and to the broader public.

Summary of Resources for the Delivery of Impact-Related Activities

i. one-day workshop and two-day symposium, attendance at conferences,
ii. travel costs associated with collaborative meetings,
iii. industrial engagement (one-day symposium)
iv. costs associated with 3 month secondment of PDRA3 to GSK.