Content Detail

• Content META

  Title	Chemical reactivity of RNA and its modifications with hydrazine
  ID	6961770
  External ID	7552117794
  Author	Unknown
  Published Date	Feb. 14, 2025, 3:07 p.m.
  Summary	Proof of principle: RNA integrity and uridine conversion Starting from the reports that hydrazine converts uridine to urea-derivatives, we were wondering how the loss of the aromatic nucleobase impac
  Tags	None
  URL	https://www.nature.com/articles/s42004-025-01444-y
  Sentiment	Negative: 0.04970    Neutral: 0.90799    Positive: 0.04230
  Relevance	0.00844

• Content Details

  Content ID	4345009
  Key	ae75ae5b-31b7-47a9-843f-099198352f5a
  Content	Proof of principle: RNA integrity and uridine conversion Starting from the reports that hydrazine converts uridine to urea-derivatives, we were wondering how the loss of the aromatic nucleobase impacts the stability of the phosphodiester bond in the absence of aniline. We incubated a synthetic 20-nucleotide (nts) long RNA with 50% hydrazine at 4 °C11 for varying periods of time (1–24 h). Cleavage of RNA, which we define as RNA degradation, can be visualized by classical polyacrylamide gel electrophoresis (PAGE). PAGE of the aniline-free, hydrazine-treated 20-mer showed the full-length 20-mer and no smaller fragments were detected even after 24 h of incubation, indicating that prolonged exposure to hydrazine does not compromise the integrity of the RNA (Fig. S1). In further experiments, an unmodified transcript of tRNA isoleucine (tRNAIle) was exposed to varying concentrations of hydrazine for 1 h on ice. This 76 nts long RNA did not fragment under hydrazine treatment (Fig. 1b). After precipitation for removal of excess hydrazine, the RNA was subjected to isotope dilution mass spectrometry (LC–MS/MS)23 and the number of remaining canonical nucleosides was assessed. Figure 1c shows that neither C, A nor G are lost—not even at 50% hydrazine. U is also not converted in the presence of 1%–10% hydrazine. Starting at 20% hydrazine 1 U (out of 16 U per tRNAIle molecule) is lost and with rising hydrazine concentrations even more U loss is observed. At 50% hydrazine a maximum of 4 out of 16 Us are lost from tRNAIle. Thus, in our hands, we find a 25% conversion rate for the conditions commonly used in HydraPsiSeq11. Another construct, which was in vitro pseudouridylated by TruB at position 55 (tRNAIle ΔD, AC-arm), was used as a first modified RNA model. Quantification of canonical nucleosides and pseudouridine following hydrazine treatment confirmed that Us are lost while pseudouridine (and other nucleosides) were not compromised by the treatment (Fig. S2). Again, the RNA stayed intact and no shorter RNA fragments were observed. In conclusion, our findings confirm that aniline-free hydrazine selectively degrades uridine without affecting other nucleosides or compromising RNA integrity. Optimization of reaction conditions for complete uridine conversion Using the HydraPsiSeq conditions, we could show that on average 4 out of 16 Us in in-vitro transcribed tRNAIle undergo conversion while pseudouridine stays intact. A higher conversion of U would enhance the U-Ψ discrimination potential and reduce false-positive Ψ assignments. To achieve a full conversion of every U, we systematically tested variations of the reaction conditions. We examined the effects of incubation time and temperature using 30% hydrazine as a starting point. Figure 2a illustrates that extending the incubation time from 1 to 6 h on ice enhances uridine loss as determined by LC–MS/MS analysis while other canonical nucleosides do not react (Fig. S3A). RNA integrity was maintained, as confirmed by gel electrophoresis in Fig. 2b. It is noteworthy that we see a faster migrating band with longer incubation times which correlates with the substantial loss of Us from the RNA. This shift is most likely caused by the loss of pyrazolone from U which reduces the molecular weight of the Us in RNA by 12%. As we see no sign of shorter fragments or a smear on the gel, this shift is not considered RNA degradation. Next, we tested the impact of temperature on the reaction. Figure 2c, d demonstrates that higher incubation temperatures, e.g. 50 °C, lead to nearly complete loss of all Us (15/16) after both 1 and 6 h of incubation. However, at 50 °C cytidine is lost as well and in addition the quantification becomes obscure as suddenly more G can be detected (Fig. S3B). The explanation for the obscure quantification results can be found in the respective gels in Fig. 2c, d. Here it becomes clear that the RNA is fully degraded at high incubation temperatures, as shown by the smear of faster migrating RNA fragments on the gel. From these experiments, we can see that RNA remains stable at 0 and 10 °C at 30% hydrazine concentration after both 1 and 6 h of reaction. Incubation with 30% hydrazine for 1 h at 20 °C does not lead to pronounced degradation while 30 °C does. After 6 h of incubation at 20 °C the RNA is clearly fragmenting at this temperature as well. These results indicate that an improvement of U conversion at 0 and 10 °C upon longer incubation times improves U conversion at 0 and 10 °C, while higher temperatures lead to RNA degradation, highlighting the need to balance the conditions for optimal results. Fig. 2: Impact of incubation time and temperature on uridine conversion and RNA integrity. a Abundance of uridine in tRNAIle after 30% hydrazine incubation for 15 min up to 6 h. b 20% TBE–urea polyacrylamide gel, stained with GelRed, of tRNAIle after 30% hydrazine incubation for 15 min up to 6 h. c, d Temperature dependence: abundance of uridine and cytidine in tRNAIle after 30% (0* = 50%) hydrazine incubation for 1 h (c) or 6 h (d). 20% TBE–urea polyacrylamide gel, stained with GelRed, of tRNAIle after 30% (0* = 50%) hydrazine incubation for 1 h (top) or 6 h (bottom). e Summary of the optimization experiments indicating the remaining abundance of U on the y-axes in the tRNA after 1 h (top) and 6 h (bottom). Full size image In a final step, we combined the results from the optimization experiments which results in Fig. 2e. Here, we plotted the remaining abundance of U in tRNAIle under the various conditions. Approximately 40% uridine conversion can be achieved within 1 h at either 50% hydrazine/0 °C or 30% hydrazine/20 °C, with the latter condition yielding even higher conversion rates of more than 50% without causing RNA degradation. By extending the incubation time to 6 h, more than 70% conversion occurs at 30% hydrazine/10 °C, while even more than 90% conversion is achieved at 50% hydrazine/0 °C without degrading the RNA. Aniline-free hydrazine treatment allows analysis of RNA by MALDI-MS Next, we wanted to assess whether aniline-free hydrazine treatment is suitable for oligonucleotide mass spectrometry (ONTS-MS). For this purpose, we incubated a synthetic 20-nts long RNA with 30% hydrazine at 0 °C for 1 h. The RNA was precipitated and analyzed by MALDI-MS analysis. The control showed the expected singly charged oligonucleotide at m/z 6550 and to a minor degree the doubly charged species at m/z 3276 (Fig. S4). After hydrazine treatment, three additional peaks arise which differ in mass by 52 Da (Fig. S5). Additional peaks that would indicate unwanted RNA fragmentation are not observed in either MALDI spectrum. The mass difference of 52 indicates the conversion of uridine into urea–ribose (Fig. 1a). In total, we observe the loss of 1, 2 and 3 Us, with the peak corresponding to the loss of 1 U being the main product. To address the question how RNA structure might protect certain Us from conversion, we incubated an unmodified tRNA with hydrazine, performed an RNase T1 digestion and compared the MS results to the corresponding non-hydrazine treated but T1 digested tRNA. In Fig. 3a, the peaks of two U-containing ONTS are shown before and after hydrazine treatment. The first one is the ONT from U61-G69 and contains 2 Us. After hydrazine treatment two new peaks arise which are 52 or 104 Da lighter due to U conversion compared to the untreated tRNA. The second ONT ranges from C28-G35 and contains 1 U. Here, only one new peak with a Δm/z of −52 compared to the untreated tRNA is observed which again, indicates U conversion. In total, we could identify six ONTS from the tRNA hydrolysate covering all parts of the tRNA (Fig. S6). This indicates that RNA structure is not a factor in the conversion of U, as we observed products in all U-containing ONTS. Yet, unmodified tRNA has a less stable 3D structure compared to native, modified tRNA. Therefore, we incubated commercially available S. cerevisiae tRNAPhe with hydrazine and performed RNase T1 treatment. Here, we can identify two ONTS, namely A[Cm]U[Gm]AA[Y]AΨ[m5C]UG and [m1A]UCCACAG which is in accordance with the literature15. The first ONT contains 2 Us and 1 Ψ. As seen in Fig. 3b, we see a conversion of 2 Us (Δ−52 and Δ−104) and no third conversion which indicates that U39 is fully pseudouridylated. Interestingly, we also see that Y, the hypermodified nucleoside wybutosine, shows no reactivity toward hydrazine. The second ONT contains 1 U and indeed we see conversion of U by the occurrence of a peak with Δ−52. In conclusion, we observe U conversion in both tRNAs in loop regions but also stem regions which indicates that U conversion can also occur in highly structured RNAs. Fig. 3: MALDI-MS analysis of hydrazine-treated tRNA. a Cloverleaf display of unmodified tRNAIle and MALDI-MS spectra of untreated (−) and hydrazine-treated (+) RNase-T1 hydrolyzed tRNA. Full spectra with annotation of all peaks can be found in Fig. S6. b Cloverleaf display of native S.c. tRNAPhe and MALDI-MS spectra of untreated (−) and hydrazine treated (+) RNase-T1 hydrolyzed tRNA. Hydrazine reaction conditions were 30% hydrazine, 0 °C, 1 h. Full size image While MALDI-MS is a highly useful technology to analyze purified RNAs quickly, it does not allow for automated spectrum interpretation like NASE24. To test the compatibility of aniline-free hydrazine treatment with NASE, we treated a synthetic 20-mer with 20% hydrazine for 1 h at 0 °C. The product was separated on a nano-flow reverse phase chromatographic system before injection for high-resolution mass spectrometry (HRMS). The resulting HRMS peaks are shown in Fig. S7. The unreacted RNA showed three clear peaks in the MS which correspond to the expected precursor ion of the full-length RNA in a +4 charge state (charging with either four protons, m/z 1614.23, or three protons +1 of either Na+ or K+). After hydrazine treatment, the unreacted 20-mer is still clearly detectable at m/z 1614.23. Additional peaks are now present in the spectrum at m/z 1601.22 and at lower abundance at m/z 1588.22. The difference among the peaks at the +4 charge state is 13.01 and 26.01, respectively. This corresponds again to a mass loss of 52 Da and urea–ribose formation. We selected the precursor ion of m/z 1588.22 for higher-energy collisional dissociation (HCD) fragmentation. The fragment spectra were successfully analyzed using NASE24 and the locations of the converted Us were assigned, e.g. to positions 5 and 8 (Fig. S7). In total, MS/MS analysis hinted toward a random distribution of U conversion. Impact of hydrazine treatment on RNA modifications in native RNA In Fig. 3b, we observed that modified nucleosides, such as Y, 2′-O-methylcytidine (Cm), 2′-O-methylguanosine (Gm), m5C, 1-methyladenosine (m1A) and Ψ do not react with hydrazine under aniline-free conditions. We next assessed how hydrazine treatment impacts up to 25 native modifications by incubating native total tRNA from either E. coli or HEK cells with hydrazine. As we expect a higher reactivity of modified nucleosides with hydrazine, we chose a low concentration of hydrazine (30%) and incubated for only 1 h at 0 °C. We used our established LC–MS/MS workflow23 and quantified the abundance of 20 modified nucleosides from E. coli tRNA and 23 modified nucleosides from human tRNA. For normalization of the amount of injected RNA, we used the signal of guanosine. As expected, pseudouridine is unaltered by the incubation with hydrazine (Fig. 4a, b) as is 5-methyluridine (m5U). Ribose-methylated uridine (Um) reacts partly with hydrazine. In addition, thiolated uridine modifications such as 4-thiouridine (s4U) in E. coli and 5-methoxycarbonylmethyl-2-thiouridine (mcm5s2U) are lost after hydrazine treatment. In contrast, 2-thiocytidine (s2C) from E. coli does not react under the 30% hydrazine conditions. In both, E. coli and HEK, dihydrouridine (D) is decreased at least tenfold and N4-acetylcytidine (ac4C) is completely lost. Similarly, 3-methylcytidine (m3C) is no longer detectable in human tRNA after hydrazine treatment. With regard to m7G and 6-threonylcarbamoyladenosine (t6A), we find a differential behavior in dependence of the organism. While m7G decreases fourfold in human tRNA (p value < 0.0001), no statistically significant loss is observed in tRNA from E. coli. For t6A, we find a loss in E. coli tRNA, but not in human tRNA. Chemically, there is no difference between E. coli m7G and human m7G, or t6A, respectively, and thus the result does not reflect differential hydrazine selectivity but potentially other factors such as accessibility of the nucleosides20,25 or the presence of catalysts that enhance the reaction. Fig. 4: Absolute quantification of modified nucleosides in total tRNA. a Absolute abundance of modified nucleosides from human embryonic kidney cells (HEK) tRNA (left) and tRNA integrity on a 10% PAGE (right). b Absolute abundance of modified nucleosides from E. coli tRNA (left) and tRNA integrity on a 10% PAGE. a, b Average of n = 3 biological replicates, error bars reflect standard deviation. L = ladder of 76 and 123 nts long RNAs. c Abundance of nucleosides after hydrazine treatment (1 h, 0 °C) using rising hydrazine concentrations. Data taken from E. coli total tRNA unless marked with *, which indicates data from human total tRNA. From n = 1 replicate. Full size image Intrigued by the observation that some modifications are fully lost, while others are only partially lost, we repeated the experiment with E. coli and HEK total tRNA with different hydrazine concentrations at 0 °C for 1 h. In accordance with Fig. 2e, we see a 50% conversion of U under the highest hydrazine concentration and no U conversion at concentrations below 10% (Fig. S8). Here, only one replicate was analyzed due to the sigmoidal decrease of modifications with rising hydrazine concentrations. The hydrazine-sensitive modifications are more reactive than U as their abundance decreases even at low hydrazine concentrations (Fig. 4c). For example, at 10% hydrazine, s4U, m3C and mcm5s2U are nearly fully lost, while only 5% of U is converted. These modifications appear to have the highest reactivity with hydrazine. At 30% hydrazine, ac4C and D are fully lost, m7G is reduced to ~25% of its original abundance, and acp3U decreases by ~40%. At even higher concentrations, s2C is also lost from the tRNA. Figure 5 summarizes the differential chemical reactivity of the hydrazine-sensitive nucleosides. Regarding m7G and t6A which appeared to be differentially reactive in regard to the tRNA’s organismal origin, we see a comparable reactivity of m7G in E. coli and HEK tRNA at all concentrations, except at 30% hydrazine (Fig. S8). It is not yet fully understood why m7G exhibits this reproducible but unexpected behavior at this particular hydrazine concentration in E. coli tRNA, while human tRNA shows the expected sigmoidal concentration dependence (Fig. 4 and Fig. S8). Similarly, t6A is still behaving unexpectedly. Again, we see loss of t6A from E. coli tRNA, even at the lowest hydrazine concentrations, while t6A does not change in abundance in HEK-derived tRNA. t6A is uniquely found at position 37 in both organisms’ tRNAs and therefore the modification is perfectly accessible to hydrazine. While assuming that the chemical structures of t6A are identical in bacterial and human tRNA, we conclude that the bacterial tRNA preparation might contain a cofactor, that catalyzes the hydrazine-dependent decomposition of t6A. Fig. 5: Reaction products and yields for modified nucleosides treated with hydrazine. Each box depicts the nucleoside and its product after hydrazine conversion. The concentration of hydrazine and incubation time for near-complete conversion is indicated next to the arrow. The conversion yield (c.y.) is given below the product in green font. The reaction product structures of Uridine4,7, m3C and m7G12 were previously reported. Full size image In summary, we see that aniline-free hydrazine treatment enables the reaction of modified nucleosides even at hydrazine concentrations that mainly prevent U conversion. Thus, hydrazine might be a valuable reagent for sequencing-based detection of modified nucleosides. Reaction products of modified nucleosides with hydrazine For the use of hydrazine in MS- or sequencing-based RNA modification analysis, it is important to know the yield (Fig. 4) of the reaction but also the reaction products. For this purpose, we took commercially available modified nucleoside lyophilized reference standards, dissolved them in water and performed the aniline-free hydrazine reaction (0 °C, 1 h, 30%). After the removal of hydrazine by freeze-drying, the nucleosides were analyzed by high-accuracy TOF-MS (time-of-flight). Note that mcm5s2U and t6A could not be used due to their current unavailability as dry powder. Starting with ac4C, here we see C as the main product which indicates a hydrazine-induced deacetylation. Similar to U, the C6 appears to react as an electrophile in the case of acp3U, m3C, ac4C and s2C which lose the pyrazolone moiety after hydrazine reaction, potentially following the same reaction mechanism as U (Fig. 1a). The reaction products are shown in Fig. 5, HRMS spectra in Table S1, MS/MS spectra in Table S2 and theoretical mass and mass error in Table S3. The reaction mechanisms of s4U, D and m7G are unique and shown in Fig. 6a–c. From the MS data, we propose that C4 acts as the electrophile for both D and s4U due to the lower electrophilicity of C6 of these modifications21. Subsequently, D and s4U add hydrazine to C4 of the nucleobase. In the case of s4U, the sulfur is subsequently eliminated and thus the aromaticity of the base is maintained (Fig. 6a). The nucleoside loses the nucleobase in MS2 and further fragmentation of the nucleobase (MS3) reveals loss of ammonia (−17) and hydrazine (−32) which further confirms the structure. As the accuracy is fairly low for this product (15.6 ppm) but the fragmentation spectra appear to fit (4.3 ppm), we are confident in the product’s structure. In alkaline environments D is reported to hydrolyze with subsequent ring opening26. We assume this is the first step, before hydrazine adds to the C4 which is supported by the accuracy of 2.9 ppm in the HRMS (Fig. 6b). In MS2, we observe again the neutral loss of ribose and the remaining construct undergoes loss of ammonia and hydrazine in MS3 which substantiates our proposed structure. Furthermore, the classical HNCO loss (−43) is observed from classical uracil nucleobase fragmentation27. Finally, m7G does not react with the hydrazine molecule but instead with hydroxyl-ions (or hydroxyl-radicals if bivalent cations are present in the solution28). Thus, the well-known reaction product N7-methyl-FAPy-G (Fig. 6c) is formed which is confirmed by our HRMS (MS1 7.5 ppm and MS2 3.8 ppm) and MS3 spectra29. This product is also formed under alkaline treatment as recently reported for AlkAnilinSeq12.
  Content Reference URL	None

• Publisher

  Publisher Name	Nature (#2153)
  Publisher URL	https://www.nature.com/
  Description	None
  Trust Flow Score	61
  Citation Flow Score	58

• Data Source

  Data Source Name	Quantexa News (#5)
  Data Source Type	news
  Data Source URL	https://app.aylien.com/dashboard
  Description	Welcome to Quantexa News API — the most powerful way to aggregate, filter, and integrate global news data into your apps and models.
  Weight	10