DihydrouridinE RNA modification marks: written and erased by the same enzyme – DERASE

New Analytics for Dihydrouridine mapping in RNA :
We have so far developed two related, but still distinct, approaches for the detection of the modified nucleotide Dihydrouridine (D) in RNA sequences. Both methods give clear signals for D-residues, but need further development and thorough validation for reliable and quantitative D detection in RNAs. Both methods are based on hydrolytic opening of the D ring under alkaline conditions, leading to an abasic site, which can be used in RNAseq (AlkAnilineSeq protocol). Here, our preliminary data obtained for E.coli tRNAs clearly indicate that inclusion of NaBH4 reduction significantly enhances the observed D signals. Alternatively, combination with NaBH4 reduction allows Schiff base formation with e.g. a hydrazine or primary amine bearing a fluorescent and/or a biotin label. This, in turn, will be developed towards quantification and affinity enrichment protocols. After optimization of both methods we aim to combine biotin-affinity enrichment and AlkAnilineSeq to perform global D mapping in tRNAs/rRNAs, as well as in other coding and non-coding RNAs from model organisms (E.coli and yeast).
Dus properties are characterized in more detail by fast kinetic studies using a stopped-flow.Another major objective is to provide a molecular basis for the D-dynamic by X-ray crystallography, with particular attention to the enzyme active site. A quite challenging task will be to obtain structures of oxidized Dus in complex with tRNAs containing D residues at the targeted sites.
The demonstration of reversibility in living cells apply various means of manipulating D-levels, in combination with LC-MS and other previously developed analytics as readout. In particular, manipulation of the cells’ redox status is conducted via treatment with paraquat (and related agents) and in mutants lacking e.g. key enzymes of the pentose pathway, a major source of reductive equivalents in the form of NADPH.
As accompanying investigations, we monitor general features of tRNA stability and turnover at different redox states using e.g. stable isotope pulse labeling. Individual tRNA levels are monitored using microscale thermophoresis (MST), and their turnover/stability by SLAM-seq.

Development of AlkAniline Seq for D detection and quantification :
AlkAnilineSeq protocol was initially designed and optimized for m7G detection, but inspection of tRNA profiles revealed that all conventional D positions also give moderate signals. This low level of AlkAnilineSeq signals might be related to inefficiency of cleavage and only partial formation of abasic site, subsequentily cleaved by aniline. Alternatively, low signals might result from substoichiometric modification level. In order to adapt AlkAnilineSeq to D detection we tweaked the scoring of AlkAnilineSeq signals in order to detect such sub-stoichiometric D residues in the presence of very strong signals of stoichiometric m7G546. Stop ratio (ratio of starting and passing reads for position) and NormCount (signal normalized to local background in 11 nt window). This scoring system allows efficient detection of low D levels formed in vivo and in vitro conditions.

Fluorescent and affinity labeling of D-sites for quantification and biotin based enrichment :
We wanted to develop a simple and rapid techniques for detection of in vivo D levels in different RNA molecules. To set-up a strictly specific D-labelling, we wanted to take advantage of the presence of -COOH group of hydrolyzed D-cycle. Adjustments to the method are still required, especially concerning the quantitative nature of the technique. However, this specific D-labeling has been used to analyze the D-content of ribosomal RNA extracted from several E. coli strains. This revealed that some strains were deprived of D at position 2449 in the ribosomal 23S RNA and allowed us to identify the protein responsible for ribosomal D deposition. This discovery constitutes a breakthrough in DUS identification as the newly identified protein has a DUS activity, but is structurally and biochemically completely different from all known DUS.

In vitro activity assays probing D-reversible formation in tRNAs:
We chose to prepare fully modified tRNA from cells grown in the presence of an isotopic source of carbon allowing every cellular molecule to only contain 13C. On the other end, D-less tRNA was extracted from an E. coli triple mutant grown with a regular 12C source. Mixing both populations of tRNA with yeast DUS2 enzyme without NADP+, nor NADPH, LC-MS analysis showed a decrease in D-content of 13C tRNA coupled to an increase in D-content of 12C tRNA. Further analysis in deuterated water are now required to prove that hydrogen transfer occurs directly between D-containing and D-less tRNA.

Participation of oxidized vs reduced tRNA in translation:
The tRNA pool as analyzed in 3D was compared with that present on actively translating polysomes. The most visible difference between both populations was the relative scarcity of initiator tRNA on polysomes, which is a quite plausible observation.

In view of the data presented we obtained, we consider it highly plausible that the in vivo level of D-residues is (i) dynamic, (ii) governed by reversible reduction of uridine and effected by Dus enzymes and (iii), probably tied to the physiological reductive state in the cytosol. Furthermore, when the above (i)-(iii) will be explored, very strong and far reaching implications would concern (iv) a coupling of the cell’s redox state to translation, and (v) massive impact on cancer biology, given experimental findings on Dus overexpression and reports on aberrant redox levels in cancer cells. Along the lines of the latter, one might indeed hypothesize, that the observed increase in D under physiopathological conditions could be the results of an increase of NADPH production, potentially generated by the pentose-phosphate anabolism. Up to now, we focused the study of D-dynamics in two model organisms namely E. coli for prokaryotes and S. cerevisiae for eukaryotes, then this approach could be extended later on to pathogenic organisms and mammals/human with long-term perspectives for public health issues.

Lombard, M., Reed, C.J., Pecqueur, L., Faivre, B., Toubdji, S., Sudol, C., Brégeon, D., de Crécy-Lagard, V., and Hamdane, D. (2022). Evolutionary Diversity of Dus2 Enzymes Reveals Novel Structural and Functional Features among Members of the RNA Dihydrouridine Synthases Family. Biomolecules 12, 1760. 10.3390/biom12121760.

Brégeon, D., Pecqueur, L., Toubdji, S., Sudol, C., Lombard, M., Fontecave, M., de Crécy-Lagard, V., Motorin, Y., Helm, M., and Hamdane, D. (2022). Dihydrouridine in the Transcriptome: New Life for This Ancient RNA Chemical Modification. ACS Chem. Biol. 17, 1638–1657. 10.1021/acschembio.2c00307.

Marchand, V., Bourguignon-Igel, V., Helm, M., and Motorin, Y. (2021). Mapping of 7-methylguanosine (m7G), 3-methylcytidine (m3C), dihydrouridine (D) and 5-hydroxycytidine (ho5C) RNA modifications by AlkAniline-Seq. Methods Enzymol. 658, 25–47. 10.1016/bs.mie.2021.06.001.

Dihydrouridine (D) is a post-transcriptional RNA modification that is frequent in tRNA as the namesake of the D-loop, whose structure becomes more flexible upon reduction of certain uridine residues by dihydrouridine synthases (Dus). However, beyond a growth phenotype, presumably a consequence of a generic effect on translation, the biology of D has remained surprisingly ill understood. In this project, the applicants display and pursue exciting evidence for a timely feature of modern RNA modification research, namely dynamics of intracellular D-levels coupled to reversibility of the enzymatic modification reaction. A consortium of four reputed RNA modification labs has produced two key findings, which, are being supplied here as preliminary data, and which form the basis for an intriguing working hypothesis. There is, on one hand, the biochemical demonstration of enzymatic reversion of D residues in a tRNA scaffold to the unmodified uridines, which irrefutably show that D as an epitranscriptomic mark can be removed by the same enzyme that is responsible for its biogenesis. On the other hand, in vivo data show a reduction of the D-levels in cells treated with a paraquat, an agent known to reduce intracellular levels of NADPH, the Dus cofactor for D-synthesis. From these flows the working hypothesis, that the intracellular ratio of NADPH/NADP+ level should govern the level of D in tRNAs, and potentially affect translation. The present application features a balanced work programme for a profound elucidation of the working hypothesis based on high-end analytics, mechanistic and structural investigations, as well as an in vivo investigations of tRNA modification patterns and their participation in translation. Successfully proving the working hypothesis will be equivalent to identifying the first epitranscriptomic writer/eraser pair in the same enzyme. Its far-reaching consequences in the life sciences include in particular a new principle of molecular regulation, acting between intracellular redox status and RNA components of the translation machinery.