Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Mar 11;13(1):1271.
doi: 10.1038/s41467-022-28817-4.

Secondary structure prediction for RNA sequences including N6-methyladenosine

Elzbieta Kierzek  1 , Xiaoju Zhang  2 , Richard M Watson  2 , Scott D Kennedy  2 , Marta Szabat  3 , Ryszard Kierzek  3 , David H Mathews  4
Affiliations

Secondary structure prediction for RNA sequences including N6-methyladenosine

Elzbieta Kierzek et al. Nat Commun. .

Abstract

There is increasing interest in the roles of covalently modified nucleotides in RNA. There has been, however, an inability to account for modifications in secondary structure prediction because of a lack of software and thermodynamic parameters. We report the solution for these issues for N6-methyladenosine (m6A), allowing secondary structure prediction for an alphabet of A, C, G, U, and m6A. The RNAstructure software now works with user-defined nucleotide alphabets of any size. We also report a set of nearest neighbor parameters for helices and loops containing m6A, using experiments. Interestingly, N6-methylation decreases folding stability for adenosines in the middle of a helix, has little effect on folding stability for adenosines at the ends of helices, and increases folding stability for unpaired adenosines stacked on a helix. We demonstrate predictions for an N6-methylation-activated protein recognition site from MALAT1 and human transcriptome-wide effects of N6-methylation on the probability of adenosine being buried in a helix.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Overview.
In this study, we advanced the RNAstructure software package (at center) to be capable of predicting secondary structures for sequences with the m6A nucleotide. RNA secondary structure prediction by RNAstructure relies on nearest neighbor parameters for estimating folding stability and dynamic programming algorithms for estimating structures and base pair probabilities. Here we fitted nearest neighbor parameters for m6A to optical melting data and revised software to be capable of considering any user-specified sequence alphabet.
Fig. 2
Fig. 2. A comparison of base pairing stability for m6A to A.
A The nearest neighbor parameters for helix stacks. The position of the m6A is indicated by 6. The stacking parameters, determined by linear regression, are compared for methylated (blue; i.e., m6A-U base pairs) and unmethylated (red; i.e., A-U base pairs) sequences for analogous nearest neighbors. The unmethylated stacks (i.e., A-U base pairs) are those of Xia et al. for adjacent Watson–Crick pairs and those of Chen et al. for adjacent G-U pairs. Stacks with m6A-U pairs are generally less stabilizing than analogous stacks with A-U pairs. Uncertainty estimates are the standard errors of the regression. B Terminal m6A-U pairs are not destabilizing. Plotted are the duplex stabilities as folding free energy change from the linear fit to the TM−1 vs. ln(CT/a) plots of the optical melting data. The top two sequences (Watson–Crick paired with a complementary strand) have the same nearest neighbor stacks, but the second helix has two terminal A-U pairs. This costs 0.7 kcal/mol of stability. The bottom two sequences also have the same nearest neighbor stacks, but the second has two terminal m6A-U pairs. Here the stability cost is 0.18 kcal/mol and not outside of the uncertainty estimate, which is approximated as 4% of the total free energy change. On average, terminal A-U pairs cost 0.45 kcal/mol of stability, but terminal m6A-U pairs are not destabilizing.
Fig. 3
Fig. 3. M6A stacking on a helix end stabilizes secondary structure as compared to A stacking.
The ΔΔG°37 (kcal/mol) for dangling ends and terminal mismatches as a result of N6-methylation (Supplementary Table S4) is shown, where negative values mean greater folding stability for m6A than A. The motifs shown here have a terminal base pair (left side of motif), and either a dangling end or terminal mismatch right (right side of motif). On average, the methylated motifs are more stabilizing than the unmethylated motifs, although the extent of the stabilization is sequence dependent. Uncertainty estimates are propagated from the uncertainty from the individual optical melting experiments (see “Error propagation” in the “Methods”).
Fig. 4
Fig. 4. Tests of the m6A nearest neighbor parameters and RNAstructure software.
A The structure of MALAT1 RNA. The predicted secondary structure for the HNRNPC binding site is the closed conformation both with and without N6-methylation at A22 (green arrow). This is also supported by the NMR NOESY walk (Supplementary Fig. S4) and the similar chemical shifts for imino proton resonances with and without methylation (Supplementary Fig. S6). We model the binding of HNRNPC protein as the conformation that exposes the recognition sequence (marked in red nucleotides). When A22 is methylated to m6A, we estimate the cost of opening the binding site is reduced by 0.6 kcal/mol as compared to the unmethylated sequence. B The average probability that A or m6A are buried in a helix at the position of high-confidence m6A sites in the human transcriptome. The mean probability that an A or m6A is base paired and stacked between two adjacent pairs for 18,026 sites of N6-methylation, as estimated by RNAstructure. Position 0 is the site of methylation. N6-methylation is estimated to further open the structure at the methylation site. C The average PARS scores for accessibility for the 18,026 sites of N6-methylation in the human transcriptome. Lower PARS scores indicate higher counts of nuclease S1 cleavage relative to nuclease V1 cleavage and therefore a higher likelihood of being unpaired. The RNAstructure predictions and the PARS data both show considerable single-stranded character at the site of N6-methylation.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Phizicky EM, Hopper AK. tRNA biology charges to the front. Genes Dev. 2010;24:1832–1860. - PMC - PubMed
    1. Li X, Xiong X, Yi C. Epitranscriptome sequencing technologies: decoding RNA modifications. Nat. Methods. 2016;14:23–31. - PubMed
    1. Gilbert WV, Bell TA, Schaening C. Messenger RNA modifications: form, distribution, and function. Science. 2016;352:1408–1412. - PMC - PubMed
    1. Sakurai M, et al. A biochemical landscape of A-to-I RNA editing in the human brain transcriptome. Genome Res. 2014;24:522–534. - PMC - PubMed
    1. Carlile TM, et al. Pseudouridine profiling reveals regulated mRNA pseudouridylation in yeast and human cells. Nature. 2014;515:143–146. - PMC - PubMed

Publication types