Alternative splicing is a prominent and sophisticated molecular machinery in gene regulation. It is well-known that temperature, as a crucial factor, can influence alternative splicing to further modulate gene expression.
In this thesis, we first focused on the RNA binding motif protein 3 (RBM3), a well-known cold-induced protein, to decipher the molecular mechanism how cold induces RBM3 expression. A “poison” exon, exon 3a, in the RBM3 mRNA isoforms was successfully identified to be responsible for the degradation of RBM3 transcripts at normal and high temperature. Briefly, the “poison” exon 3a containing premature termination codons (PTCs) can elicit non-sense mediated decay (NMD) of the transcripts. At normal and high temperature, the “poison” exon is substantially included, which induces the NMD of RBM3 transcripts, and RBM3 expression is downregulated; at low temperature, exon 3a is majorly exclusive, resulting that RBM3 transcripts escape from the NMD, and RBM3 expression increases. Next, two strong exonic splicing cis-regulatory enhancer (M2 and M4) were determined via mutagenesis screening to potently promote exon 3a inclusion at normal and high temperature. Antisense oligonucleotides (ASOs) (2'-O-MOE and morpholino chemistry modification), targeting M2, M4 motif and 5’ splice site (SS) in the exon 3a RNA, efficiently increase endogenous RBM3 expression at both mRNA and protein levels in both HEK293T and neuroblastoma N2a cells. Given that the evolutionary conservation of “poison” exon 3a of RBM3 between humans and mice, it is conceivable that the effective ASOs can be potentially and therapeutically used in clinic, such as cardiac arrest, moderate or severe neonatal encephalopathy or even some neurodegenerative diseases, as RBM3 has been well demonstrated to protect neurons in vivo through mediating structural plasticity in mice before.
Next, for the first time we report that in mammalian cells RNA itself can directly sense temperature alterations though “spliceswitch”, G-quadruplex (G4) structures to modulate alternative splicing and further regulate gene expression. Utilizing RNA sequencing data of human cells at different temperatures, we found that RNA G4 (rG4) motifs are significantly enriched around splice sites of cassette exons repressed upon cold shock. We further validated that formation/stabilization of rG4 indeed effectively mask splice sites, which decreased exon inclusion in several cold-repressed exons of several different genes, showing that G4 stabilizer ligand (PDS) could decrease the inclusion of cold-repressed exons at four temperatures. Taking RBM3 as an example, we found that stabilizing G4 through PDS can significantly decrease the inclusion of RBM3 exon 3a in cells. Meanwhile, the putative G4 motifs (G1) of RBM3 is manifested to form real rG4 in vitro by biophysical assays. Under low KCl condition, G4 signal is temperature sensitive, showing higher signal at low temperature; while under high KCl condition, its temperature sensitivity is not significantly obvious. Intriguingly, the stability of G4s across different temperatures is reversible, presenting that rG4 signal gradually decreases from 33°C to 40°C, and then if the temperature goes back to 33°C from 40°C it returns to the signal level at 33°C. Furthermore, stabilizing rG4s in RBM3 exon 3a is also capable of increasing endogenous RBM3 expression at both mRNA and protein levels at only high temperature, but not at low temperature. In conclusion, rG4s are stable enough at low temperature to mask splice sites, thereby promoting “poison” exon 3a exclusion, so stabilizers are less effective for the stabilization of rG4s in this scenario; while at high temperature rG4s are predominantly unstable to make the G4 stabilizers present their effects more easily. Our results unveil an unexpected mechanism how temperature perturbations are directly sensed by RNAs and integrated into gene expression programs via spliceswitch, which might provide a novel avenue to treat disease caused by splicing defects.
Finally, considering that both cis-regulatory enhancer elements M2 and M4 and repressive G4 motifs can modulate the alternative splicing of RBM3 “poison” exon 3a, we asked how these elements cooperate each other, and further whether there exist other cis-regulatory element networks to be involved in the regulation of alternative splicing. Systematic CRISPR-dCasRx perturbation sliding on the “poison” exon 3a and its flanking introns clearly show that several gRNA clusters targeting M2, M4, 3’SS (G1), G3 and 5’SS, increase endogenous RBM3 expression with higher fold-changes at high temperature than that at low temperature. This indicates that M2, M4 enhancer and G4 structure (G1 and G3 motif) may be the strongest cis-regulatory elements modulating exon 3a inclusion. By virtue of the relatively low resolution of CRISPR-dCasRx screening, to further deeply dissect the cis-regulatory network regulating exon 3a inclusion, we sequentially mutated the sequence surrounding alternative splice site in the exon 3a and its flanking intronic region. An exonic cis-regulatory element in the proximity to the downstream of alternative splice site was successfully identified to inhibit the usage of alternative splice site. Meanwhile, several intronic cis-regulatory elements in vicinity to each other in distance were determined to be involved in enhancing or repressing exon 3a inclusion. To further understand the coordinated network of cis-inhibitory and -enhancing regulatory elements in achieving the regulation of temperature sensitive RBM3 exon 3a inclusion and simplify the model as well, we took enhancer M4 and repressive motif G1 and G3 as representative examples to build a working model which manipulates RBM3 exon 3a inclusion. At low temperature (33°C) both G1 and G3 repressive activity are dominant, while the enhancer activity of M4 is more powerful than the repressive activity of G3 at high temperature (40°C). At 37°C, repressive activity of G1 and G3 is almost identical to the enhancer activity of M4. Finally, we showed that both phosphatase and CLK inhibitor influenced the inclusion of “poison” exon 3a just at high temperature, but not at low temperature, suggesting that G4 activity is indispensable of phosphorylation, and M2 and M4 function dependently on the phosphorylation status of trans-acting factors.
In summary, three independent and connected stories with high novelty in the dissertation were successfully conducted: 1) discovered the mystery how RBM3 expression is cold-induced in a post-transcription level and developed ASOs to boost RBM3 expression, which may potentially protect neurons at normal temperature in clinic; 2) proposed and demonstrated RNA serves as physiologically direct thermo-sensors via spliceswitch formed by G-quadruplex in mammalian cells. Importantly, the “poison” exon 3a of RBM3 is one of the cold-repressed exons (CRE) candidates with G4 near splice sites. This illuminates that stable RNA G4s at low temperature in the splice sites of exon3a, which influence the utilization of splice sites, are one of the reasons that cold represses “poison” exon 3a inclusion; 3) fine-tuned mapping cis-regulatory elements , that modulate RBM3 exon 3a inclusion and deciphered the coordinated cis-regulatory network of alternative splicing of exon 3a.
