Finding Order in Randomness: Single-Molecule Studies Reveal Stochastic RNA Processing

Observing single RNA in living cells reveals the kinetics of RNA synthesis and processing.

Observing single RNA in living cells reveals the kinetics of RNA synthesis and processing.

Producing a functional eukaryotic messenger RNA (mRNA) requires the coordinated activity of several large protein complexes to initiate transcription, elongate nascent transcripts, splice together exons, and cleave and polyadenylate the 3’ end. Kinetic competition between these various processes has been proposed to regulate mRNA maturation, but this model could lead to multiple, randomly determined, or stochastic, pathways or outcomes. Regulatory checkpoints have been suggested as a means of ensuring quality control. However, current methods have been unable to tease apart the contributions of these processes at a single gene or on a time scale that could provide mechanistic insight. To begin to investigate the kinetic relationship between transcription and splicing, Daniel Larson, Ph.D., of CCR’s Laboratory of Receptor Biology and Gene Expression, and his colleagues employed a single-molecule RNA imaging approach to monitor production and processing of a human β-globin reporter gene in living cells.

This reporter gene incorporated 24 PP7 RNA hairpin structures in the second intron as well as 24 MS2 hairpins in the 3’ untranslated region. By expressing red-tagged PP7 coat proteins and green-tagged MS2 coat proteins, which bind their respective RNA hairpins, the researchers could use time-lapse fluorescence microscopy to track the splicing of single transcripts in three dimensions in osteosarcoma cells. Unspliced RNA appeared in the red and green channels, while the spliced product was in green only. Imaging revealed a diffraction-limited spot with fluctuating fluorescence in both colors over time, suggesting that this spot is the transcription site. Observing the introns and exons at the same time allowed the investigators to determine when a transcript was spliced. They found that individual transcripts could be spliced before or after release of the mRNA from the transcription site and that diffusion away from the site is rapid, resulting in steep drops in the fluorescence signal.

The scientists then developed an analysis method to extract kinetic information from the fluctuations in fluorescence and generated functions from approximately 2,000 individual transcripts that describe how a change in one channel is correlated with a change in the same channel or in the other channel after a specific time delay. Based on these correlation functions, they found that splicing occurs before release from the transcription site about 15 percent of the time, indicating that splicing and release do not have to occur in a particular order. The functions also revealed other aspects of the transcription cycle, such as the elongation rate and the time the transcript remains at the transcription site. The researchers verified their assignments of these features by treating the cells with drugs known to alter specific aspects of RNA production, specifically splicing or elongation.

To gain mechanistic insights, the investigators evaluated five mathematical models relating the shape of the correlation functions with the timing of the underlying molecular processes. The model that best fit the correlation curves was one in which splicing and elongation/release undergo kinetic competition. In this model splicing occurs a specific amount of time after the 3’ splice site is transcribed, while transcript release takes place after reaching the poly(A) site and a stochastic delay. The timing of reporter gene splicing was consistent with previous studies, suggesting the PP7 hairpins did not affect splicing kinetics. Likewise, the fact that the drug treatments only affected a single process supports the assertion that splicing and elongation are kinetically independent. Because the time to release the transcript can be shorter or longer than the intron removal time, splicing may occur before or after release.

Since most splicing takes place after a transcript is released, the scientists asked where and when this splicing occurs. They observed mobile unspliced transcripts enriched near the transcription site and spliced transcripts throughout the nucleus, suggesting that splicing occurs faster than diffusion. In fact, the researchers found that splicing is 10-fold faster on diffusing transcripts than those associated with chromatin.

Finally, the research team decided to investigate whether proteins could regulate splicing by altering kinetic competition. They focused on U2 auxiliary factor 1 (U2AF1), an essential 3’ splice site recognition factor with a recurrent serine to phenylalanine point mutation observed in myelodysplastic syndrome, leukemias, and breast and lung cancers. Using their time-lapse imaging method, the scientists found that expression of a tagged U2AF1 point mutant completely blocked pre-release splicing and that post-release splicing occurred more slowly, potentially encouraging alternative splicing. However, the mutant had no effect on overall splicing efficiency. The investigators confirmed their results with the β-globin reporter gene by using the endogenous FXR1 mRNA, which previous studies showed is alternatively spliced in the presence of mutant U2AF1.

The results of these studies reveal the existence of multiple pathways for intron removal at single molecule resolution in living cells. They argue against the presence of a checkpoint for this process and indicate that splicing and release can occur in either order, depending on kinetic competition. Changes in this kinetic balance can lead to changes in splicing, which may have an important role in the development of diseases, such as cancer.

Summary Posted: 11/2014


Coulon A, Ferguson M, de Turris V, Palangat M, Chow CC, Larson DR. Kinetic competition during the transcription cycle results in stochastic RNA processing. eLife. October 1, 2014 PubMed Link