Pseudouridine (Y), a simple glycosidic N1 to C5 isomer of uridine, is one of the most prevalent of the modified ribonucleotides. It is found in a variety of non-coding RNAs - ribosomal RNA (rRNA), transfer RNA (tRNA), and small nuclear RNA (snRNA) associated with spliceosomes. However, recent papers suggest it may also function in coding messenger RNA (mRNA).
A seminal paper to suggest this expanded role for pseudouridine in coding mRNA was by Karijolich and Yu who found that replacing U with Y in the stop codon UAA allowed read-through by the ribosome and suppression of the translation termination in vitro using a synthetic mRNA transcript.1 To further their argument, they then used the cell’s own pseudouridylation machinery, the H/ACA RNAs, to convert in vivo a stop codon introduced into the CUP1 gene. The gene product of CUP1 infers tolerance to copper sulfate in the culture media by the chelation of Cu2+. By plating the cells in media containing CuSO4, they had a convenient reporter system to monitor suppression of translation termination.
By mutating a naturally occurring H/ACA RNA guide sequence, they were able to specifically target the U of the stop codon UAA, which they introduced in the CUP1 gene, and convert it to YAA. They found that the transformed S. cerevisiae only survived in the high Cu2+ media when the guide RNA specifically targeted the UAA codon for pseudouridylation.
The authors went on to look at the suppression of termination of the stop codons UAA, UAG and UGA in plasmids containing the TRM4 gene. Again, site-specific pseudouridylation, using guide RNA strands targeting the stop codons introduced in the TRM4 gene, led to observable TRM4 protein production. When the resulting TRM4 protein was purified and analyzed by mass spectrometry, it was found that the translation of the YAA and YAG codons led to the incorporation of serine and threonine in roughly equal frequency for YAA and predominantly serine for YAG. Whereas YGA led to the incorporation of predominantly tyrosine (with some phenylalanine observed).
The lab of Venki Ramakrishnan went on to crystalize the 30S ribosomal subunit docked to the tRNA(ser), which presents the anticodon AGI bound to the YAG codon. Intriguingly, the structure showed that the N1 imino proton of pseudouridine does not form any hydrogen bonds and the canonical Y-A Watson-Crick base pair was observed. However, the pseudouridine seems to exert a subtle but profound effect upon the anticodon stem loop structure, leading to a purine-purine A-G Hoogsteen base pair with the adenosine of the YAG codon in an unusual syn conformation.2
However tantalizing the results of Karijolich and Yu, there was still no demonstration of naturally occurring pseudouridylation of mRNA in cells. This soon changed with a second seminal paper on pseudouridine that was recently published by the Gilbert lab at MIT, who used next-generation sequencing to demonstrate that pseudouridylation of mRNAs occurred naturally in both human and yeast cells.3
The Gilbert lab utilized techniques developed by Bakin and Ofengand4, who had found that pseudouridine would specifically and irreversibly react with N-Cyclohexyl-N'-(ß-[N-methylmorpholino]ethyl)carbodiimide p-toluenesulfonate (CMC). After reacting the RNA with CMC and treating with sodium carbonate to remove non-specific reactions with G and U bases, the resulting carbodiimide-Y adduct blocked reverse transcription. By comparing the +CMC transcripts with -CMC controls to correct for Y-independent transcription stops, the locations of pseudouridine could be determined in RNA (a sequencing procedure termed Pseudo-seq).
After confirming the correct Y-calling in non-coding RNA, they then used poly-dT cellulose beads to pull-down poly-A+ transcripts to look for pseudouridylation in mRNA. Remarkably, they found hundreds - conservatively, 260 pseudouridylated sites - in over 238 mRNA coding transcripts. Not only that, they also found that the pseudouridylation was regulated depending upon environmental conditions. As the yeast growth went from exponential (log phase) to stationary (as the nutrients were depleted in the media), pseudouridylation was found to be upregulated for some mRNA and down-regulated for others, indicating pseudouridylation was responsive to environmental stress. Not unexpectedly, pseudouridylation sites of rRNA changed little.
The authors then looked at the sites of pseudouridylation in mRNA to see if they corresponded to H/ACA RNA guide sequences and found that most were not. Therefore, they turned their attention to the family of pseudouridine synthases in yeast (PUS1-9) that do not require guide RNA sequences. When PUS deletion strains were grown and pseudouridylation of mRNAs was analyzed for the different ΔPUS strains, PUS1 was found to be responsible for most of the mRNA pseudouridylation. However, PUS2-4, PUS6-7 and PUS9 all had unique mRNA targets.
Knowing that pseudouridine synthases are conserved in all eukaryotes, HeLa cells were analyzed by Pseudo-seq during normal growth and serum-starvation conditions. Their conservative estimate suggested that 89 human mRNAs were pseudouridylated, some in response to growth-state conditions. As a number of diseases are associated with mutations in PUS genes, the authors raised the possibility that misregulation of pseudouridylation in mRNAs may contribute to the diseased state.
These results taken together hint at a second layer of functions and activities of pseudouridine in not only non-coding RNA but coding RNA as well.
We envision continuing interest in PseudoUridine CE Phosphoramidite as researchers continue to investigate termination suppression in premature termination codons, potential alternative codon coding, and the function of conditional pseudouridylation of mRNA.