Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - ComPreValRther (Computational Prediction and Validation of RNA thermometer at transcriptome-wide scale in living cell)

Teaser

All life forms on earth are affected by temperature. Temperature is crucial to plant growth and development. Cold and heat stresses may drastically inhibit plant growth and cause yield losses in crops. Previous studies have focused on temperature-regulated transcriptional...

Summary

All life forms on earth are affected by temperature. Temperature is crucial to plant growth and development. Cold and heat stresses may drastically inhibit plant growth and cause yield losses in crops. Previous studies have focused on temperature-regulated transcriptional network; however, the transcriptional regulation of mRNA levels only partially correlates with translation. It is therefore important to elucidate the mechanisms underpinning post-transcription regulation of gene expression. RNA molecules fold into secondary and tertiary structures, which can play key roles in transcription, splicing, translation, localization and degradation. Recently, studies on RNA structures have drawn much attention.
So far RNA thermometers (RNATs) have mostly been identified in bacteria. RNATs respond to temperature shifts by changing their secondary structure. RNATs usually fold into a complex structure typically positioned in the 5’ UTR of mRNAs, to modulate translation by affecting ribosomal association in response to temperature shifts. Traditional RNA structure probing experiments are mainly based on in vitro treatment of a synthesised RNA sequence with various chemical or enzymatic probes to distinguish single and paired nucleotides. Thus, most RNATs in bacteria have been identified via in silico or in vitro conditions. However, the in silico / in vitro structure of RNA species will be different from those characterised in vivo as a living cell is usually out of equilibrium and the actual structure of RNA species will be heavily dependent on temperature, interacting proteins and nucleotides. In vivo RNA structure determination has always been a challenging but active interdisciplinary research field involving nucleic acids chemistry, biophysics, biochemistry and molecular biology. I took advantage of this novel method in vivo RNA structure profiling developed in the Ding lab to fundamentally understand the functional role of RNA structure in translation and how RNA structure alters in response to temperature in living cells. My project aimed to develop pipelines to quantitatively measure in vivo RNA structure features and identify their changes in response to temperature through modelling and data mining.

Work performed

I firstly developed new and reliable analysis pipelines for both RNA structurome and ribosome profiling datasets, generated by my wet-bench collaborator Dr. Yang. By evaluating both maximum-likelihood-estimation and log ratio methods, I found the maximum-likelihood-estimation method with optimized parameters provided accurate estimation of structure profiles where the false positive rate was low. I used this method to compare SHAPE modification probabilities of each mRNA across different temperatures. Globally, I found that SHAPE modification probabilities were generally higher (indicating more single-strandedness) at 37°C compared to 22°C. In contrast, SHAPE modification probabilities were generally lower (indicating more double-strandedness) at 4°C compared to 22°C. Having generated these SHAPE modification probability data, I performed RNA structure folding to each mRNA across different species and different temperatures. An average 21% base-pairs in each mRNA structure were disrupted at 37°C compared to mRNA structures at 22°C. Most of these disrupted base-pairs were long-distance base-pairing. In contrast, an average of 12.3% base-pairs in each mRNA structure were formed at 4°C compared to 22°C. These new base pairing formations included both local and long-distance base-pairing.
Secondly, in contrast to our original hypothesis, our results strongly indicated that rather than being affected by RNA structure, ribosomes reshape the RNA structure landscape instead. Faced with a mixed picture from both in vivo RNA structures and ribosome-remodeled RNA structures, I adopted a two-pronged approach to dissect this problem. One approach was to create a new mathematic model for dissecting the relationship between mRNA structure and translation across the whole translation process. This new model reflected a kinetics mode of action that the ribosome utilizes in both passive and active mechanisms for unwinding mRNA during translation. By incorporating the Totally Asymmetric Exclusion Process (TASEP) model, I was able to integrate the ribosome density information for translation initiation, elongation and termination and to associate RNA structure information across the whole translation process. This model is the first mathematical model which combines both in vivo RNA structure information and ribosome density datasets. In my second approach I dissected the ribosome data to find those mRNAs which were not remodeled by ribosome. I took advantage of those mRNA undergoing co-localization during translation. In this model, the translation of mRNAs encoding secretory or membrane proteins is initiated in the cytosol, and the mRNA–ribosome complexes are selectively recruited to the ER membrane. By dissecting these datasets, I found a 5-nt single-stranded region downstream of the transmembrane domain (TMD) region. This downstream single-stranded structure signal significantly correlated with ribosome density after TMD regions. This analysis may provide a direct causality relationship between mRNA structure and ribosome density whereby the mRNA structure feature may serve as a signal to facilitate co-localization translation recognition. By exploring structure features in the in vivo RNA structurome at 37°C, it has been possible to isolate those mRNAs containing 5-nt single-stranded region downstream of TMD region which were significantly altered at 37°C compared to normal yeast temperature at 28°C. This work will shed light on the functional understanding of mRNA co-translational targeting and the role of mRNA structure in regulating translation along with temperature regulation of RNA functions.
Results were presented at regular lab meetings, JIC departmental meetings, the 14th Nucleic Acid Forum, Computational RNA Biology Conference, Translation UK 2017, EMBO workshop “RNA structure meets function”, and ISCB (International Society for Computational Biology) conference 2018.

Final results

The new mathematic model is the first to combine both in vivo RNA structure information and ribosome density datasets. Once the model simulation and validation are complete, the corresponding manuscript will be written.
The in vivo RNA structure features I identified for co-localization translation will shed light on our functional understanding of mRNA co-translational targeting and the role of mRNA structure in regulating translation along with temperature regulation of RNA functions. This will also lead to a high impact publication. My project outputs have also led to new and reliable RNA structurome analysis pipelines which will be applied to other lab members’ publications.
The importance of RNA structure in regulating gene expression was publicly discussed in several public engagement events such as the Norwich Science Festival and the JIC 50 years science open day. I have discussed my work with local school children as well as UEA college students and helped them to understand basic RNA structure biology.

Website & more info

More info: https://www.jic.ac.uk/people/yilliang-ding/.