A quantitative model predicts how m6A reshapes the kinetic landscape of nucleic acid hybridization and conformational transitions

N6-methyladenosine (m6A) is a post-transcriptional modification that controls gene expression by recruiting proteins to RNA sites. The modification also slows biochemical processes through mechanisms that are not understood. Using temperature-dependent (20°C–65°C) NMR relaxation dispersion, we show that m6A pairs with uridine with the methylamino group in the anti conformation to form a Watson-Crick base pair that transiently exchanges on the millisecond timescale with a singly hydrogen-bonded low-populated (1%) mismatch-like conformation in which the methylamino group is syn. This ability to rapidly interchange between Watson-Crick or mismatch-like forms, combined with different syn:anti isomer preferences when paired (~1:100) versus unpaired (~10:1), explains how m6A robustly slows duplex annealing without affecting melting at elevated temperatures via two pathways in which isomerization occurs before or after duplex annealing. Our model quantitatively predicts how m6A reshapes the kinetic landscape of nucleic acid hybridization and conformational transitions, and provides an explanation for why the modification robustly slows diverse cellular processes.

This paper describes a detailed relaxation dispersion NMR investigation of the role of m6A methylation in regulating strand formation and melting. Using sophisticated NMR dynamics methods, a series of nucleic acid constructs, different temperatures and labeling that stabilizes conformational preferences, a model is described consisting of separate induced fit and conformational selection pathways that explains the kinetics of duplex formation/melting. Notably, the intermediate formed in the induced fit pathway is one where the syn conformation is retained in the duplex state -the first time that such a conformation has been observed. However, the authors provide convincing evidence that indeed this is in fact the case. This has potential implications for recognition of nucleic acids by cognate proteins and further enhances our knowledge of the different structures that these molecules can form.
The work represents an important contribution to the literature and a stunning example of how to dissect what appears to be a complex 4 state pathway. I have only 1 simple question. Could the authors provide an intuitive explanation for why at 65oC the syn double strand conformation is much lower than at 55oC, that is, what influences the DH of the process so significantly?
Reviewer #2 (Remarks to the Author): Liu et al. present a study of a quantitative model to predict kinetic behaviors of nucleic acid consisting of m6A methylation in hybridization and conformational transitions. The m6A methylation is an important posttranscriptional modification in the regulation of gene expression. This study shows that m6A modulates the annealing/melting behavior of nucleic acids through syn/anti-conversion and conformational selection pathways. The study is highly significant; the approach and data analysis appear to be rigorous. After addressing the following concerns, the manuscript is appropriate for the publication of this journal. 1. One of the fundamental concerns is the whole process of the annealing appears to be driven by a single modification at one of ~nine bps, where most of them are GC bps. One would expect the annealing process should largely be driven and dominated by GC bp interactions, and thus intuitively would consider an alternative 3-state kinetic model to count for the RD data: ssRNA-synm6A + ssRNA-complementary <-> dsRNA-with-syn "bulge"m6A <-> dsRNA I'd suggest re-analyze the RD and CEST data under this alternative 3-state model. I suspect this 3state model could also at least be possible, especially for long helices where it's both thermodynamically and kinetically favorable to form, regardless of what happens in one of the positions. 2. Explain why the presence of Mg2+ has an effect on RD of m6AMP. 3. Include data for C8, C10 of m6AMP at all temperatures in Suppl. Table 1. Do RD data measured at C2, C10, and C8 agree with each other at various temperatures? From the data presented in the ms, it's not clear. 4. The main conclusion appears to be made based on the 2-state fit/constrained 3-state simulation of the data acquired at T=65. It would be helpful if the authors discuss what would be at 37C, a temperature more relevant to the physiological condition? 5. Reconcile/rewording seemly contradictory tandem statements "…in which the methylamino group rotates into the energetically favored syn isomer. Although such a conformation is predicted to be highly energetically disfavored…" 6. Given its relatively small size by Mw of the RNAs in this study, S/N would possibly permit detection of the two HB present in the m6Aanti:U basepair. Did the authors attempt to detect the presence of the two hydrogen bonds in dsRNAanti using HNN-cosy for the (A)N1---H-N3(U) H-bond and NOE for the another? 7. I suggest plotting the CS difference between m6A and m62A and m6A vs. residue number at three temperatures in Ext. fig.7. The plots would better illustrate the chemical shift differences in adjacent residues. 8. The proposed 4-state model is an underdetermined system with multiple unknowns and assumptions. It's always possible to fit kinetic models, given enough number of "floating" parameters to fit based on my experience and literature. I'd suggest the authors discuss the possibility. 9. Fig 5 appears to suggest even distributions of 50% of IF and CS at T=55C in the 4-state model. More convincing experimental data would be direct detection of species, which might be feasible in this case given the equal population distribution using a high-field spectrometer.
Reviewer #3 (Remarks to the Author): Review of Nature Communications Liu et al. "A quantitative model predicts how m6A reshapes the kinetic landscape of nucleic acid hybridization and conformational transitions" The manuscript of Liu et al. describes detailed conformational analysis of the m6A methylation modification in RNA. Posttranscriptional m6A methylation and demethylation are critical features of regulation in eukaryotic organisms. While the biological importance of these modifications has been clearly outlined, their structural and biophysical effects on RNA structure have remained unclear; m6A modified As can pair with U similar to unmodified As, but only in the anti-form of the methyl amino group. Multiple groups have shown that m6A destabilizes the ability of RNA-RNA interactions to form, but the origins of this kinetic effect were unknown. Liu et al apply elegant NMR exchange methods to identify the syn-anti conformational exchange of the methyl group as the key modulator of RNA base pairing kinetics. The preferred syn form blocks pairing with U, slowing duplex formation while transition to anti in the final structure leads to no effect on the dissociation rates. This effect occurs by pathways that depend on whether the isomerization of the m6A group occurs before or after base pairing. The authors determine the rates of these processes and create detailed kinetic models that explain the m6A effects on stability and exchange. The results of this rigorous and important study are of broad interest to the readers of Nature Communications and beyond, and finally place the effects of m6A on a firm physical foundation. As such the manuscript absolutely deserves publication in Nature Communications. Below I make some suggestions and possible additional data that will improve the final published manuscript.
1. The presentation of the relaxation dispersion measurements and fitting is not clear in the main text. The methods and mathematics, albeit complex, are presented in the methods and supp material. This reader, who understands a bit of NMR, always needs to be reminded of the exchange theory used to extract the excited states. I assume non-NMR readers may struggle further. A nice figure panel in either Fig 1 or better still Fig 2 that outlines the exchange measurements and makes better sense of the field dependent curves fits in subsequent panels would go miles in helping a reader through the data 2. The duplex association and dissociation rates were extracted from NMR exchange data at high temperature. I would like to see these rates confirmed by an alternative method, such as T-jump or fluorescence or another method independent of NMR. If this is too much work, a temperature dependence of the rates that give the appropriate thermodynamic parameters. In short, it would be nice to "trust" the rates as measured by NMR as true association and dissociation rates. These parameters are critical to the models presented here, and methods are referred to in (21). 3. Relatedly, the authors jump around to make measurements at different temperatures. I think I extracted the logic of moving from RT to 55 or 65C, in order to populate excited states and accelerate rates to make the exchange measurements. However to the novice reader I think this point might be lost. A clearer outline of the experimental logic would help greatly. 4. It would have been interesting to see the effect of m6A on base triple formation, since now the methyl group, even in anti, would inhibit triple formation. 5. All these minor critiques are meant to improve the impact of the paper to the non-NMR reader. These results are exciting and explain for example our own results on tRNA and release factor binding in the A site. Publish away! Jody Puglisi Reviewer #1 (Remarks to the Author): This paper describes a detailed relaxation dispersion NMR investigation of the role of m6A methylation in regulating strand formation and melting. Using sophisticated NMR dynamics methods, a series of nucleic acid constructs, different temperatures and labeling that stabilizes conformational preferences, a model is described consisting of separate induced fit and conformational selection pathways that explains the kinetics of duplex formation/melting. Notably, the intermediate formed in the induced fit pathway is one where the syn conformation is retained in the duplex state -the first time that such a conformation has been observed. However, the authors provide convincing evidence that indeed this is in fact the case. This has potential implications for recognition of nucleic acids by cognate proteins and further enhances our knowledge of the different structures that these molecules can form.
We thank the reviewer for his/her positive comments. Based on RD NMR measurements on the hairpin construct, in which the exchange contribution from the ssRNA state is essentially eliminated (see Fig. 4a-c and Extended Data Fig. 6a), the dsRNA syn population is similar at 55°C (~1.1%) and 65°C (~1.6%). The dsRNA syn species was not directly observable at 65°C in our RD NMR experiments on the duplex construct because its exchange contribution was probably masked by the much larger contribution due to the ssRNA state with population ~22%. By reducing the population of the ssRNA to 5%, we were able to unmask and observe the dsRNA syn intermediate at 55°C in the duplex construct.
The reviewer may be curious about why the flux along the IF pathway is lower at 65°C (10%) versus 55°C (50%). Since the hybridization step is rate-limiting for both CS and IF pathways, the flux is primarily determined by the population of the ssRNA species and annealing rate constants. The smaller flux along the IF pathway at 65°C versus 55°C can be attributed to a slower annealing rate along the IF pathway at 65°C due to a 2-fold reduction in population of the ssRNA syn relative to ssRNA anti and comparatively 2.5-fold slower annealing rate constants for ssRNA syn along the IF pathway relative to ssRNA anti along the CS pathway.
To clarify these points, we added the following two paragraphs to the main text on pages 12 and 20, respectively: "Although we did not observe any evidence for the IF dsRNA syn intermediate, simulations indicate that its RD contribution was probably masked by the larger RD contribution from the ssRNA with population ~22%. We therefore repeated the CEST measurements at a slightly lower temperature T = 55°C. This reduced the ssRNA population to ~5%, but it remained large enough to permit accurate measurements of hybridization kinetics using NMR RD.
Repeating the measurements at a different temperature also allowed us to test the robustness of the CS model." "The smaller flux along the IF pathway at 65°C versus 55°C can be attributed to a slower annealing rate along the IF pathway at 65°C due to a 2-fold reduction in population of the ssRNA syn relative to ssRNA anti and comparatively 2.5-fold slower annealing rate constant of ssRNA syn along the IF pathway relative to ssRNA anti along the CS pathway." Reviewer #2 (Remarks to the Author):

Liu et al. present a study of a quantitative model to predict kinetic behaviors of nucleic acid consisting of m6A methylation in hybridization and conformational transitions. The m6A methylation is an important posttranscriptional modification in the regulation of gene
expression. This study shows that m6A modulates the annealing/melting behavior of nucleic acids through syn/anti-conversion and conformational selection pathways. The study is highly significant; the approach and data analysis appear to be rigorous. After addressing the following concerns, the manuscript is appropriate for the publication of this journal.
We thank the reviewer for his/her positive comments.
1. One of the fundamental concerns is the whole process of the annealing appears to be driven by a single modification at one of ~nine bps, where most of them are GC bps. One would expect the annealing process should largely be driven and dominated by GC bp interactions, and thus intuitively would consider an alternative 3-state kinetic model to count for the RD data: We also added a method section entitled "3-state IF simulations and constrained fits for the dsGGACU m6A RD data measured at T = 65°C and 55°C" on page 45, describing the approach we used to perform IF pathway constrained fitting.
In addition, to avoid confusion, we changed the introduction on page 4 and updated Fig. 1c to acknowledge early on the possibility for the induced fit pathway in addition to conformational selection: "Kinetic mechanisms involving binding and conformational change can occur via pathways wherein the conformational change occurs prior or post binding 1 . We therefore hypothesized that m 6 A could slow hybridization via at least two pathways in which isomerization of the methylamino group occurs either before or following duplex formation (Fig. 1c). In the conformational selection (CS) pathway, hybridization proceeds via an unpaired intermediate (ssRNA anti ) with m 6 A in the energetically disfavored anti conformation (Fig. 1c). In the induced fit (IF) pathway, the more populated ssRNA syn species with m 6 A in the syn conformation initially hybridizes to form a double-stranded intermediate (dsRNA syn ) that entails the loss of at least one Watson-Crick H-bond between m 6 A and the partner uridine (Fig 1a). This is then followed by isomerization to form the Watson-Crick bp (dsRNA anti ) with m 6 A in the anti conformation (Fig. 1c) We thank the reviewer for this question. In the case of m 6 AMP, we only measured RD for C2 and C8. The RD profiles for C8 were all flat as was the case for ssGGACU m6A (Extended Data Fig. 2a) and the hpGGACU m6A (Extended Data Fig. 6a). This is most likely because the C8 chemical shift is insensitive to isomerization of the methylamino group. We did not have C10 labelled m 6 AMP and did not measure C10 RD at natural abundance because the signal was too weak. We did however measure RD for C2, C8 and C10 in ssGGACU m6A at 25°C and 37°C and the data is in good agreement, the population and k ex is close to within error (Supplementary Table 1). Similarly, the C2 and C10 CEST data measured at 55°C and 65°C for hpGGACU m6A could be combined in a global fit (Supplementary Table 1). Note that C8 is sensitive to duplex melting, which is why we observe C8 RD at 55°C and 65°C in dsRNA constructs (Extended Data Fig. 4a, 5) but not at 37°C (Extended Data Fig. 5) due to the low population (< 0.1%) of the ssRNA species.
In the revision, we updated Supplementary We thank the reviewer for raising this important point. In the original submission, we briefly addressed this point, but this was buried in the Methods section on page 51 "Predict m 6 Ainduced slowdown of DNA hybridization in the mouse genome". In the revision, we moved this discussion from the methods to discussion on page 27: "Our NMR measurements had to be performed under high temperature conditions so that hybridization falls within the detection limits of RD. However, we were able to observe isomerization of the methylamino group in both ssRNA and dsRNA at T = 37°C (Extended Data Fig. 2b and 6a) We thank the reviewer for pointing this out. We meant to say the syn conformation is energetically favored in unpaired m 6 A while it is predicted to be highly disfavored in dsRNA. We revised this sentence on page 14 as follows: "Although never observed previously, one possibility is that the new intermediate is a dsRNA conformation in which the methylamino group rotates into the syn conformation. Such a conformation is predicted to be highly energetically disfavored in dsRNA, given the loss of at least one Watson-Crick H-bond. However, this loss in energetic stability would be partly compensated for by a gain in stability of ~-1.5 kcal/mol from restoring the energetically favored syn isomer."

Given its relatively small size by Mw of the RNAs in this study, S/N would possibly permit detection of the two HB present in the m6Aanti:U basepair. Did the authors attempt to detect the presence of the two hydrogen bonds in dsRNAanti using HNN-cosy for the (A)N1---H-N3(U) H-bond and NOE for the another?
The conformation of the m 6 A anti :U base pair in the dsGGACU m6A has been extensively characterized previously based on NOEs in a prior study 3 . Additionally, prior structural studies using NMR 6 and crystallography 7 have shown that m 6 A anti :U forms the two Watson-Crick hydrogen bonds. We made changes to page 4 to clarify this point: (Fig. 1a)."

"Rather, when paired with uridine, the methylamino group rotates into the energetically disfavored anti isomer and forms a canonical m 6 A-U Watson-Crick bp that retains both (A)N1···H-N3(U) and (A)N6···H-O4(U) H-bonds
7. I suggest plotting the CS difference between m6A and m62A and m6A vs. residue number at three temperatures in Ext. fig.7. The plots would better illustrate the chemical shift differences in adjacent residues.
We thank the reviewer for this suggestion. In the revision, we added plots showing the chemical shift differences in Extended Data Fig. 7c. We observe upfield shifted C8, C2 and C1ʹ in the dimethylated residue in both RNA duplexes, consistent with DFT calculations (Extended Data Fig. 7c). In addition, the chemical shifts of the dimethylated A partner and neighboring residues also show perturbations, indicating that dimethylation only locally affects RNA structure.
New Extended Data 7c. Chemical shift perturbations comparing dimethylated and m 6 A modified dsGGACU and dsA6 RNA constructs (left). Shown on the right are the chemical shift perturbations for C2, C8 and C1ʹ measured for dsGGACU and dsA6 and calculated using DFT.

The proposed 4-state model is an underdetermined system with multiple unknowns and assumptions. It's always possible to fit kinetic models, given enough number of "floating" parameters to fit based on my experience and literature. I'd suggest the authors discuss the possibility.
We thank the reviewer for this thoughtful suggestion. We agree that with a 4-state model there is a risk of over-fitting the data. It is for this reason that we initially assessed the CS+IF 4-state model using explicit simulations in which no parameter was allowed to float and in which approximate rate constants for each step were measured independently using appropriate constructs or conditions tailored to isolate as much as possible a particular step. These values and the 4-state model were then used to simulate the CEST profiles without any adjustable parameter (see Extended Data Fig. 9a). We observed good agreement between simulated and experimentally measured CEST data. Afterwards, we performed a constrained 4-state fit allowing the parameters to float around these average values by an amount determined by experimental uncertainty (one standard deviation). To clarify these points, we added the following paragraph to the methods section on page 43:   While the reviewer is correct that the flux through the CS and IF pathways is 50%, the two intermediates have low populations and/or are in fast exchange on the NMR timescale, as a consequence they cannot be observed directly as separate resonances. This underscores a general point, which is that many reaction intermediates are low-populated and short-lived, and difficult to resolve using many conventional biophysical approaches. The low-populated species would be difficult to directly observe in NMR spectra even with higher field due to fast exchange. We do observe separate resonances for the starting duplex reactant and single stranded product, because here the exchange is slow on the NMR timescale.

Review of Nature Communications Liu et al. "A quantitative model predicts how m6A reshapes the kinetic landscape of nucleic acid hybridization and conformational transitions"
The We thank Jody for his positive comments. These are great suggestions. We have added the full CS+IF pathway schematic in Fig. 1c, laying out all the definitions of the various species and kinetic rate constants.

The presentation of the relaxation dispersion measurements and fitting is not clear in
Revised Fig. 1. c. Schematic of the CS+IF model with kinetic rate constants defined as follows: and are the forward and backward rate constants for methylamino isomerization in ssRNA, respectively. and are the forward and backward rate constants for methylamino isomerization in dsRNA, respectively.
, and , are the annealing and melting rate constants, respectively when m 6 A adopts anti conformation in both ssRNA and dsRNA.
, and , are the annealing and melting rate constants, respectively when m 6 A adopts syn conformation in both ssRNA and dsRNA.
We also added a brief introduction to the R 1ρ and CEST experiment on page 7: "NMR RD experiments can be used to characterize conformational exchange between a dominant ground state (GS) and short-lived low-populated "excited-state" (ES). The R 1ρ experiment measures the line-broadening contribution (R ex ) to the transverse relaxation rate (R 2 ) during a relaxation period in which a continuous radiofrequency (RF) field is applied with variable power (ω SL ) and frequency (ω RF ). The RF field reduces the R ex contribution in a manner dependent on ω SL and ω RF and the exchange parameters of interest (see below). The RD profiles are typically displayed by plotting the measured R 2 + R ex as a function of ω SL and ω RF . For detectable exchange, a peak is observed centered at the difference between the chemical shift of the GS and ES (-∆ω, assuming ω GS = 0 and ω ES = ∆ω). The CEST experiment measures the impact of conformational exchange on longitudinal GS magnetization during a relaxation period following application of a continuous RF field with variable power (ω SL ) and frequency (ω RF ). When applied on resonance with the ES, the RF field saturates the ES magnetization and this saturation can be transferred via conformational exchange to the GS. This typically results in a reduced signal intensity for the GS and a minor dip centered at ω ES = ∆ω when the RF is on resonance with ES. A major dip is also observed centered at ω GS = 0 when the RF field is on resonance with the GS. The dependencies of R 2 + R ex (R 1ρ ) or the GS signal intensity (CEST) on ω SL and ω RF can be fit to the Bloch-McConnell equations describing n-site exchange to determine exchange parameters of interest (see below)." We also provide a description of the R 1ρ and CEST data when they are first introduced on page 8: "Shown in Fig. 2c on the left is the CEST profile recorded for the m 6 A-C10 methyl carbon in ssGGACU m6A as a function of RF. As is typical for CEST profiles, a major dip is observed when the RF field is on-resonance with the GS chemical shift at ∆ω = 0. In addition, a minor dip was observed indicative of conformational exchange with a sparsely populated ES. The dip was observed at a chemical shift ∆ω C10 = ⍵ ES -⍵ GS = 3 ppm, which was in good agreement with the value predicted for the anti isomer (∆ω C10 = 3-5 ppm) using density functional theory (DFT) calculations 8 Fig. 2c on the right is the R 1ρ profile measured for m 6 A-C2 in ssGGACU m6A as a function of RF field. A peak was observed at -∆ω C2 = 0.6 ppm indicative of conformational exchange. A similar C2 RD was observed in methylated but not unmethylated AMP, as expected if the RD is reporting on isomerization (Extended Data Fig. 2a)." 2. The duplex association and dissociation rates were extracted from NMR exchange data at high temperature. I would like to see these rates confirmed by an alternative method, such as T-jump or fluorescence or another method independent of NMR. If this is too much work, a temperature dependence of the rates that give the appropriate thermodynamic parameters. In short, it would be nice to "trust" the rates as measured by NMR as true association and dissociation rates. These parameters are critical to the models presented here, and methods are referred to in (21).

(Methods). Shown in
We thank Jody for these suggestions. We did perform something analogous to the requested comparison when we first introduced the NMR RD methodology in the prior (ref 21) paper 9 . In this prior ref 21, we benchmarked the NMR RD method using two 12-mer DNA duplexes. The kinetic rate constants for duplex hybridization measured by NMR RD fell within the expected range of values reported previously using fluorescence spectroscopy for duplexes of similar length. For example, for 12-mer DNA duplexes, k on ranged between 2x10 5 -4x10 6 M -1 s -1 compared to our value of ~1.2x10 6 M -1 s -1 while k off ranged between 0.4-200 s -1 , compared to our value of 5-60 s -1 . In addition, the kinetic rate constants for duplex hybridization measured by NMR RD exhibited the well-established strong dependence of k off on sequence while k on showed virtually no sequence dependence, in agreement with prior studies 10 . In addition, the chemical shifts deduced by NMR RD for both DNA and RNA duplexes were shown to be in excellent agreement with those measured for the isolated single-stranded species, and this internal consistency further supports that the NMR RD data is reporting on the duplex hybridization kinetics.
To further test the NMR RD approach, we followed Jody's suggestion and performed temperature dependent 13 C CEST measurements on dsGGACU m6A . These new data are included in the revision in Extended Data Fig. 11. k off was strongly dependent on temperature as reported in prior studies of duplex hybridization kinetics employing fluorescent spectroscopic techniques 11,12 and could be fit to a standard van't Hoff equation (R 2 = 0.97). On the other hand, k on showed a much weaker dependence on temperature and exhibited deviations from the Arrhenius behavior (R 2 = 0.68), in very good agreement with prior studies 11,12 , which also report a weak temperature dependence and non-Arrhenius behavior for k on . For comparison, both the forward and backward rate constants for methylamino isomerization showed a strong temperature dependence and both could be fit to a standard van't Hoff equation (R 2 > 0.96, see Extended Data Fig. 2c, 6d).
Following Jody's suggestion, we used the NMR RD data to determine thermodynamic parameters ∆ ° and ∆ ° for duplex annealing based on the temperature dependence of p B (ssRNA population) which was used to deduce the value of ∆ ° at each temperature. We then compared these thermodynamic parameters to counterparts measured previously for dsGGACU m6A using UV melting experiments 3 . Indeed, we find good agreement between the two measurements particularly for ∆G°; the difference is < 0.2 kcal/mol (Extended Data Fig. 11b). Although the ∆ ° and ∆ ° values are not within error, small deviations are to be expected given that these parameters are generally not as well determined as for ∆ ° and that different techniques were used employing very different duplex concentrations (NMR ~1 mM versus UV ~3 µM).
In addition, we also observed good agreement between ∆ ° measured using CEST and UV melting experiments for 9 additional DNA/RNA duplexes at temperatures varying between 45°C to 65°C (Extended Data Fig. 11c). We suscept that the small systematic differences probably arise in part due to small differences in the temperature calibration on the two instruments, given that the ∆ ° values are reported near T m making them particularly sensitive to temperature. In addition, duplex-duplex interactions at the higher duplex concentration used in the NMR experiments could explain the slightly higher duplex stability observed by NMR relative to UV. Extended Data Fig. 11. Independent tests of NMR RD measurements of m 6 A RNA hybridization. a, Temperature dependence of the melting (k off ) and annealing (k on ) rate constants of dsGGACU m6A . b, Temperature dependence of annealing free energy ∆ °, which is derived from population of ssGGACU m6A measured by CEST. c, Comparison of thermodynamic parameters measured using optical melting experiments (UV) 3 and temperature dependent CEST experiments (CEST) for dsGGACU m6A . d, Comparison of ∆ ° measured 9 for 10 DNA/RNA duplexes at temperatures ranging from 45°C to 65°C.
In the revision, we make sure to point the reader to the prior paper describing the original RD NMR approach on page 3: "Recently, we developed and validated an NMR relaxation-dispersion (RD) [13][14][15] based method to measure the hybridization kinetics in DNA and RNA duplexes 9 . Using this approach, we showed that m 6 A preferentially slows the apparent rate of RNA duplex annealing by ~5-10-fold while having little effect on the apparent rate of duplex melting 9 (Fig.  1b)." We also included a description of the new temperature dependent hybridization kinetics in the methods section on page 37 along with the new Extended Data Fig. 11: "Validation of NMR RD measurements on m 6 A RNA hybridization. We have previously 9 shown that hybridization kinetics measured from NMR RD on unmodified DNA and RNA duplexes are consistent with those measured using other techniques employing fluorescence spectroscopy. As an additional test, we performed temperature dependent RD measurements for dsGGACU m6A (Extended Data Fig. 11a). The annealing rate constant k on did not have a strong temperature dependence, consistent with prior studies reporting non-Arrhenius behavior for k on in unmodified duplexes 11,12 . On the other hand, the melting rate constant k off showed a strong temperature dependence, also consistent with prior studies 11,12 . The extrapolated annealing thermodynamic parameters including ∆ °, ∆ ° and ∆ ° measured from NMR experiments are in good agreement with those measured from UV melting experiments 3 Fig. 11b-c). We also observed a good agreement between the annealing free energy (∆ °) measured using CEST and UV melting experiments for 9 additional DNA/RNA duplexes at temperatures ranging from 45°C to 65°C 9 (Extended Data Fig. 11d)." 3. Relatedly, the authors jump around to make measurements at different temperatures. I think I extracted the logic of moving from RT to 55 or 65C, in order to populate excited states and accelerate rates to make the exchange measurements. However to the novice reader I think this point might be lost. A clearer outline of the experimental logic would help greatly.

(Extended Data
We agree. The reason we started at high temperatures is because the hybridization kinetics can be better characterized using NMR RD owing to larger ssDNA population and faster exchange kinetics. To clarify this point, we added the following sentence on page 10: "The CEST experiments were performed at high temperature because at 37°C, the ssRNA is too lowly populated (<0.1%) and the hybridization is too slow (<50 s -1 ) to be effectively characterized by RD." 4. It would have been interesting to see the effect of m6A on base triple formation, since now the methyl group, even in anti, would inhibit triple formation.
We thank Jody for sharing this interesting idea! We agree that m 6 A should in principle disrupt the base triple structure. We do plan to survey for base triple structures and examine the effect of m 6 A.
5. All these minor critiques are meant to improve the impact of the paper to the non-NMR reader. These results are exciting and explain for example our own results on tRNA and release factor binding in the A site. Publish away! Jody Puglisi Thank you Jody!