An RNA modification enzyme directly senses reactive oxygen species for translational regulation in Enterococcus faecalis

Lee, Wei Lin; Sinha, Ameya; Lam, Ling Ning; Loo, Hooi Linn; Liang, Jiaqi; Ho, Peiying; Cui, Liang; Chan, Cheryl Siew Choo; Begley, Thomas; Kline, Kimberly Ann; Dedon, Peter

doi:10.1038/s41467-023-39790-x

Download PDF

Article
Open access
Published: 11 July 2023

An RNA modification enzyme directly senses reactive oxygen species for translational regulation in Enterococcus faecalis

Nature Communications volume 14, Article number: 4093 (2023) Cite this article

6714 Accesses
1 Citations
105 Altmetric
Metrics details

Subjects

Abstract

Bacteria possess elaborate systems to manage reactive oxygen and nitrogen species (ROS) arising from exposure to the mammalian immune system and environmental stresses. Here we report the discovery of an ROS-sensing RNA-modifying enzyme that regulates translation of stress-response proteins in the gut commensal and opportunistic pathogen Enterococcus faecalis. We analyze the tRNA epitranscriptome of E. faecalis in response to reactive oxygen species (ROS) or sublethal doses of ROS-inducing antibiotics and identify large decreases in N²-methyladenosine (m²A) in both 23 S ribosomal RNA and transfer RNA. This we determine to be due to ROS-mediated inactivation of the Fe-S cluster-containing methyltransferase, RlmN. Genetic knockout of RlmN gives rise to a proteome that mimics the oxidative stress response, with an increase in levels of superoxide dismutase and decrease in virulence proteins. While tRNA modifications were established to be dynamic for fine-tuning translation, here we report the discovery of a dynamically regulated, environmentally responsive rRNA modification. These studies lead to a model in which RlmN serves as a redox-sensitive molecular switch, directly relaying oxidative stress to modulating translation through the rRNA and the tRNA epitranscriptome, adding a different paradigm in which RNA modifications can directly regulate the proteome.

Nuclear mRNA decay: regulatory networks that control gene expression

Article 18 April 2024

DNA-targeting short Argonautes complex with effector proteins for collateral nuclease activity and bacterial population immunity

Article 15 April 2024

Phage defence system CBASS is regulated by a prokaryotic E2 enzyme that imitates the ubiquitin pathway

Article 22 April 2024

Introduction

Reactive oxygen species (ROS) such as superoxide (O₂⁻) and hydrogen peroxide (H₂O₂) play fundamental roles in shaping bacterial evolution¹. Bacteria are exposed to ROS from a variety of sources, endogenously as byproducts of aerobic respiration and exogenously from redox-active natural products and the respiratory/oxidative bursts of activated mammalian immune cells². If not neutralized, ROS damage essential cellular components including DNA, lipids, carbohydrates, and proteins¹. Bacteria have thus evolved ROS defense systems such as superoxide dismutases (SOD), catalases, glutathione, thioredoxin systems, peroxidases, and nitrate/nitrite reductases³. These defenses are often regulated transcriptionally by ROS-sensing transcription factors such as OxyR, PerR, OhrR, and SoxR^4,5,6.

Recent evidence points to mechanisms of translational regulation of bacterial stress response systems. For example, the hypoxic stress response in mycobacteria involves reprogramming of dozens of modified ribonucleosides on tRNAs—the tRNA epitranscriptome—to cause selective translation of codon-biased mRNAs from hypoxia response genes including the DosR transcription factor and its regulon⁷. Modifications on other forms of RNA also participate in diverse cellular processes by altering RNA stability, structure, localization, and protein-RNA interactions⁸.

Here we report the discovery of an ROS-sensing RNA-modifying enzyme that regulates translation of stress-response proteins in the gut commensal and opportunistic pathogen Enterococcus faecalis. Following exposure to the superoxide generator, menadione, or sublethal doses of ROS-inducing erythromycin and chloramphenicol, analysis of 24 modified ribonucleosides of the epitranscriptome revealed large decreases in N²-methyladenosine (m²A) in both 23 S ribosomal RNA and transfer RNA possibly caused by ROS-mediated inactivation of the Fe-S cluster-containing methyltransferase, RlmN. Loss of RlmN altered protein expression in a way that mimicked menadione exposure, such as increased superoxide dismutase and decreased virulence proteins. These studies suggest that RlmN acts as a redox-sensitive molecular switch that links environmental and antibiotic-induced ROS exposure to epitranscriptome dynamics in ribosomal and transfer RNA to effect translation of stress response proteins.

Results

While transcriptional regulation in response to stress is well established in bacteria, translational regulation is less well understood. Our demonstration of hypoxia-induced epitranscriptome reprogramming and codon-biased translation in mycobacteria⁷ led us to hypothesize that a similar mechanism might hold true for the response of Enterococcus faecalis to the stress of antibiotic exposure. Here we quantified 24 modifications in the rRNA and tRNA in two E. faecalis strains: OG1RF, a strain derived from the human commensal oral isolate OG1⁹, and V583¹⁰, a multidrug resistant clinical isolate. V583 possesses an erythromycin resistance methyltransferase (ErmB) that methylates the N⁶-position of adenosine (m⁶A, m^6,6A) in 23 S rRNA at position 2058 (Escherichia coli numbering)¹¹ and prevents binding of macrolides (e.g., erythromycin), lincosamides, and streptogramin B (MLS)¹¹, but only confers partial resistance to erythromycin¹². OG1RF lacks ErmB and is thus about 100-fold more sensitive to erythromycin than V583.

We first explored antibiotic effects on V583’s epitranscriptome by growing cells in the presence of sub-inhibitory concentrations of erythromycin (10–200 μg/mL; Fig. 1a). Following purification of 23 S and 16 S rRNAs and tRNA and hydrolysis to ribonucleosides, 24 modified ribonucleosides were quantified by chromatography-coupled mass spectrometry (LC-MS) in each type of RNA (Fig. 1b–d, Supplementary Table 1)⁷. Neither m⁶A nor m^6,6A increased in 23 S rRNA with erythromycin treatment, confirming that Erm expression in V583 is constitutive¹². While most of the monitored modifications remained relatively unchanged with treatment, a striking reduction in 2-methyladenosine (m²A) in both 23 S rRNA and tRNA was observed for erythromycin exposure (Fig. 1b–d), with dose-dependency (Fig. 1e, f).

**Fig. 1: Epitranscriptomic profiling of 23 S and 16 S rRNA and tRNA of *Enterococcus faecalis* V583 grown in the presence of growth-permissive concentrations of erythromycin.**

To assess if the m²A reduction was unique to V583, we repeated the study with E. faecalis OG1RF at growth-permissive erythromycin concentrations (0.1–0.3 μg/mL; Supplementary Table 2). We again observed a significant concentration-dependent decrease in m²A in both 23 S rRNA and tRNA (Fig. 2a, b), with insignificant changes in the other 23 modifications (Supplementary Fig. 3). We next asked if m²A reduction was unique to erythromycin or a general response to all antibiotics. At 10–25% of the MICs of the different antibiotics (Supplementary Table 2), m²A was reduced by macrolides erythromycin and spiramycin and the phenicol antibiotic chloramphenicol, all of which are bacteriostatic, but not by the bactericidal ciprofloxacin, ampicillin, or aminoglycosides kanamycin and gentamicin (Fig. 2a–c). Macrolides and chloramphenicol bind to the large (50 S) ribosomal subunit at the peptide exit tunnel and peptidyl transferase center, respectively, while aminoglycosides target the small (30 S) ribosomal subunit, in contrast to ciprofloxacin and ampicillin as gyrase inhibitor and cell wall disruptor, respectively¹³. The antibiotic-induced reduction of m²A is less in OG1RF compared to V583 (Fig. 2c), most likely due to the markedly lower concentrations of erythromycin and chloramphenicol used with OG1RF. So far, the data reveal that exposure of E. faecalis to sub-inhibitory concentrations of antibiotics that target the 50 S ribosomal subunit selectively reduce m²A in 23 S rRNA and tRNA, which raises the question of the mechanistic basis for this epitranscriptome behavior.

**Fig. 2: Reduction of m2A in V583 and OG1RF is specific to bacteriostatic macrolides and chloramphenicol.**

In Escherichia coli, m²A formation is catalyzed by RNA methyltransferase RlmN, which methylates A2503 in 23 S rRNA at the peptidyl transferase center in the 50 S ribosomal subunit and A37 in the subset of tRNAs bearing adenine at this position in the anticodon stem loop¹¹. Since RlmN in E. faecalis has not been previously characterized, we analyzed m²A levels in a ΔrlmN deletion mutant in OG1RF and found complete loss of the modification in 23 S rRNA and tRNA (Fig. 3a). We first asked if the reduction of m²A by macrolides and chloramphenicol involved transcriptional or translational regulation of RlmN. Neither the level of rlmN mRNA (qPCR) nor the level of RlmN protein (targeted proteomics) was affected significantly by erythromycin treatment (Fig. 3b, c). These data suggested that the activity of RlmN was regulated post-translationally by antibiotic exposure.

**Fig. 3: RlmN is regulated at the protein level by reactive oxygen species ROS.**

Among possible mechanisms for regulating RlmN activity, one involves antibiotic-induced oxidative stress. RlmN is a radical S-adenosylmethionine (SAM) enzyme containing a [4Fe-4S] cluster that is sensitive to disruption by reactive oxygen species (ROS)². To test this model, we grew OG1RF in the presence of the superoxide radical generator, menadione¹⁴, at sub-inhibitory concentrations and then measured m²A in 23 S rRNA and tRNA. Menadione treatment caused a dose-dependent decrease in m²A in both rRNA and tRNA (Fig. 3d, e). This result implies that the macrolides and chloramphenicol also induced ROS in OG1RF and V583, which we assessed using the fluorogenic superoxide-specific probe CellROX Green to quantify antibiotic-induced ROS levels in OG1RF¹⁵. As expected, both erythromycin and chloramphenicol increased CellROX Green fluorescence at sub-inhibitory and higher concentrations (Fig. 3f). Given the impact of antibiotics on bacteria cell shape and size that may lead to artifacts on the interpretation of flow cytometric analyses of ROS-detecting fluorescent dyes and the importance of setting FSC and SSC gates¹⁶, we confirmed that the fluorescence changes were due to ROS using an optimized gating and other flow cytometric parameters that minimize artifacts in dye-based ROS detection (Supplementary Figs. 5, 6)¹⁵. Indeed, two observations confirm the absence of size artifacts. First, erythromycin reduced mean FSC-H and SSC-H values at sub-inhibitory concentrations (Supplementary Fig. 6), suggesting smaller cell size, yet mean fluorescence intensity of CellROX Green Dye increased directly with erythromycin concentration (Fig. 3f ). On the other hand, ampicillin caused increases in the mean FSC-H and SSC-H (Supplementary Fig. 6) but did not cause an increase in mean fluorescence intensity of CellROX Green Dye (Fig. 3f ). These two observations convincingly show that antibiotic-induced changes in fluorescence in the CellROX dye studies were due to ROS activation of the dye and not flow cytometry size artifacts. Not all ribosome-targeting antibiotics generate ROS in OG1RF. Sub- and supra-inhibitory concentrations of tetracycline, which binds to the 30 S subunit, induced an increase in CellROX Green fluorescence whereas the aminoglycosides streptomycin and kanamycin did not (Fig. 3f). Further, the mechanistically distinct ampicillin and ciprofloxacin also did not increase CellROX Green fluorescence at both sub- or supra-inhibitory concentrations (Fig. 3f). These observations would seem to contradict Léger et al., who reported that the β-lactams amoxicillin and cefotaxime caused superoxide production by reduction of demethylmenaquinone (DMK) in E. faecalis¹⁷. However, they did not measure intracellular superoxide. It is well established that E. faecalis produces high levels of extracellular superoxide by way of externally-facing, membrane-bound DMK, but the superoxide is unable to diffuse through cell walls and be detected as intracellular superoxide by CellRox and is also rapidly dismutated to hydrogen peroxide outside the cell¹⁸; Léger et al. measured extracellular hydrogen peroxide as their surrogate for superoxide¹⁷. That RlmN activity observed here could be caused by antibiotic-induced reduction in SAM levels is not possible given the lack of change in methylation-based modifications other than m²A. Further, while we did not measure tRNA levels here, the constancy of all other RNA modification levels confirms that antibiotic treatment did not significantly change the levels of the rRNA and tRNA substrates of RlmN. Though we our studies do not provide proof of a direct reaction of ROS with RlmN, the dose-response for RlmN inhibition with menadione- (Fig. 3e), erythromycin- (Figs. 2c, 3f), and chloramphenicol-induced ROS (Figs. 2c, 3f) strongly suggests that RlmN is inactivated by intracellular superoxide resulting from exposure of E. faecalis to macrolides and chloramphenicol.

Given the controversial model that bactericidal antibiotics share a common mechanism of generating cytotoxic ROS^15,19,20, we tested antibiotics for their bactericidal and bacteriostatic activity in E. faecalis. Erythromycin and chloramphenicol have been classified as bacteriostatic²¹, with bactericidal activity at high concentrations against Streptococcus pneumoniae²¹, while ciprofloxacin, ampicillin, and streptomycin are classified as bactericidal²¹. However, our data show that erythromycin and chloramphenicol induce superoxide production at low concentrations in OG1RF, while ciprofloxacin, ampicillin, and streptomycin do not. Léger et al. observed ROS generation by the related β-lactam, amoxicillin, at supra-lethal concentrations, again by measuring hydrogen peroxide which was likely generated from superoxide produced extracellularly. To establish bactericidal activity, we performed killing assays with OG1RF in the presence of antibiotics at ten times their MIC and found that antibiotics that increase CellROX Green fluorescence (i.e., erythromycin, chloramphenicol, tetracycline) are bacteriostatic in OG1RF, while ciprofloxacin, ampicillin, and streptomycin, which do not induce ROS, are bactericidal (Fig. 3g). These data not only further disprove the link between bactericidal antibiotics, ROS, and cell death, but also raise the question of the role of RlmN in E. faecalis antibiotic sensitivity.

We next investigated the effect of RlmN activity on E. faecalis antibiotic sensitivity, using the ΔrlmN deletion mutant and a strain over-expressing RlmN. Here we created a constitutive over-expression mutant, OG1RFprlmN, which carried an RlmN-expressing plasmid under the constitutive Sortase A promoter in vector pGCP123²²; and a control strain, OG1RFpEmpty, which carried the same plasmid lacking the coding sequence for RlmN. Over-expression of RlmN did not affect the MIC for erythromycin, streptomycin, kanamycin, ampicillin, tetracycline, or ciprofloxacin (Supplementary Table 2), but did increase the sensitivity of OG1RFprlmN to killing by the bactericidal antibiotics ampicillin and ciprofloxacin by 10-fold (Fig. 4a, b). Loss of RlmN, however, had no effect on E. faecalis growth in the presence of up to 5-times the wild-type MIC for the 5 antibiotics (Supplementary Table 2, Fig. 4c, d), except for a 16-fold increase in MIC for chloramphenicol (Fig. 4e, Supplementary Table 3). Taken together, (1) over-expression of RlmN increases E. faecalis sensitivity to ampicillin and ciprofloxacin and (2) loss of RlmN activity confers resistance to chloramphenicol suggesting that RlmN could play a role in regulating antibiotic tolerance.

**Fig. 4: Phenotypic characterization of rlmN KO (ΔrlmN) and over-expressed RlmN (OG1RFprlmN).**

We next defined the effects of RlmN activity on the cell proteome. Our results establish that RlmN’s methyltransferase activity is attenuated by superoxide, raising the possibility that RlmN serves as a redox sensor that regulates cell stress response. To test this idea, we performed quantitative proteomics using multiplexed isobaric Tandem Mass Tags (TMT) to identify differentially expressed proteins between OG1RF, ΔrlmN, and OG1RF grown in menadione (Fig. 5a, b; Supplementary Table 3; Supplementary Data 1). Only a few proteins were upregulated in ΔrlmN as compared to OG1RF treated with menadione, suggesting precision of RlmN as a molecular switch. Indeed, we found a positive correlation of R = 0.900 between statistically significant (p < 0.05) protein changes in ΔrlmN and OG1RF treated with menadione (Fig. 5d). One of these was the antioxidant defense enzyme, superoxide dismutase, for which there is a single gene (sodA) in E. faecalis²³. Of note, another well-known antioxidant defense enzyme, catalase (KatA), was not consistently detected in the proteomic analyses (Supplementary Data 1). Interestingly, the common set of proteins that are downregulated in both datasets include proteins associated with virulence (Supplementary Table 4). Pilus subunit protein EbpA is a major virulence factor in E. faecalis, involved in biofilm formation, endocarditis, and catheter-associated urinary tract infection²⁴; and WxL domain-containing proteins^25,26, which, in Enterococus faecium, have been implicated in survival in bile salt and the pathogenesis of endocarditis²⁷.

**Fig. 5: Loss of rlmN and treatment with menadione cause similar changes in the OG1RF proteome.**

Discussion

Based on the results presented here, we propose a mechanism (Fig. 6) in which RlmN serves as a ROS-sensitive molecular switch that modulates physiological responses to oxidative stress and confers phenotypic resistance to environmental stresses and antimicrobial agents. This is perhaps not surprising given the importance of other Fe-S proteins, such as SoxR, Fnr, and aconitase, as ROS sensors linked to changes in gene expression and cell phenotype⁶. RlmN and m²A are absent in eukaryotes and, within prokaryotes, RlmN is the only enzyme that synthesizes m²A. The inertness of the C2 of adenosine to electrophilic attack and the low acidity of the C2 proton requires a free radical SAM intermediate. This mechanism, distinctly different from typical SAM-dependent methyltransferases, is unique to RlmN²⁸ and the chloramphenicol-florfenicol resistance methyltransferase (Cfr)²⁹. Here we showed that RlmN activity is not only strongly dependent upon intracellular superoxide levels but also regulates levels of SodA, which promotes antibiotic tolerance³⁰ and facilitates survival in macrophages²³ in E. faecalis and other bacteria.

**Fig. 6: Working model of RlmN as a sensor for oxidative stress.**

So how does exposure of E. faecalis to ribosome-binding antibiotics lead to elevated superoxide levels? While there is no universal mechanism by which antibiotic exposure causes increases in ROS in bacteria, in spite of earlier claims²⁰, there are numerous pathways for generating superoxide and other ROS and for environmental exposure of bacteria to ROS². E. faecalis generates large amounts of extracellular superoxide¹⁸, with Léger et al. showing that supra-lethal ampicillin doses increase these levels¹⁷. While we did not measure extracellular ROS, superoxide cannot diffuse through the bacterial cell wall and our results show that neither sub- nor supra-lethal amoxicillin or other bactericidal antibiotics cause intracellular superoxide formation in E. faecalis (Fig. 3f). How sublethal concentrations of erythromycin and chloramphenicol cause superoxide levels to increase could relate to the cell stress caused by inhibition of translation or by mistranslation, with the proteotoxic stress leading to increased reductive metabolism and thus increases in superoxide. The mechanism of erythromycin-induced superoxide production awaits further study.

How does RlmN-catalyzed m²A play a role in the phenotypic changes caused by loss of RlmN or exposure to superoxide? Two possibilities come to mind based on RlmN activity on both rRNA and tRNA. From the rRNA perspective, m²A might facilitate the ribosome stalling that leads to ErmBL nascent peptide activation of ErmB expression³¹. RlmN catalyzes m²A formation at A2503 of 23 S rRNA, a conserved nucleotide that resides in the peptidyl transferase center (PTC) of the ribosome near the entrance to the exit channel for the nascent polypeptide (Fig. 7) and is involved in fine-tuning ribosome–nascent peptide interactions, relaying the stalling signal to the PTC³². A2503 is very close spatially to the erythromycin binding site³², to the ErmBL nascent peptide that activates ErmB expression upon antibiotic binding³³, and to the A2058 that is modified with m⁶A by ErmB to prevent antibiotic binding (Fig. 7). Though speculative, one hypothesis is that loss of RlmN activity reduces m²A2503 and thus facilitates ErmBL-induced activation of ErmB synthesis.

**Fig. 7: Structure of the ribosome showing superimposed binding sites for erythromycin, chloramphenicol, and positions of modified nucleotides A2503 and A2058.**

From the tRNA perspective, RlmN is one of only two methyltransferases known to modify both rRNA and tRNA¹¹. In E. coli, m²A37 occurs in six tRNAs: tRNA^ArgICG, tRNA^AspQUC, tRNA^Glncmnm⁵sUUG, tRNA^GlnCUG, tRNA^Glumnm⁵s²UUC, and tRNA^HisQUG³⁴. Modifications at position 37 are important for maintaining the reading frame³⁵, while loss of RlmN increases stop codon readthrough³⁶. The latter is likely not relevant for synthesis of selenoproteins in E. faecalis since there are no apparent genes encoding selenoproteins in the E. faecalis genome³⁷. Further studies are underway to determine if E. faecalis uses tRNA reprogramming and codon-biased translation to regulate expression of stress response genes as observed in mycobacteria⁷.

Finally, our results raise the question of the uniqueness of RlmN’s sensitivity to inactivation by superoxide, compared to other Fe-S cluster proteins in E. faecalis and other organisms. E. faecalis is unusual among human commensal and pathogenic bacteria in lacking many Fe-S cluster proteins, such as fumarase, aconitase, isocitrate dehydrogenase, and succinate dehydrogenase in the tricarboxylic acid cycle³⁸. This lack of a tricarboxylate cycle is shared by Listeria monocytogenes, which also shares with E. faecalis a lack of ROS formation when exposed to bactericidal antibiotics³⁹. Other Fe-S cluster proteins absent in E. faecalis include the GrxD iron transport regulator, 1 of 3 systems for Fe-S cluster biogenesis (NIF, ISC, and SUF; only SUF is present in E. faecalis⁴⁰), and MiaB. The latter is corroborated by our inability to detect m²si⁶A in the presence of i⁶A (Fig. 1d). However, E. faecalis possesses several Fe-S cluster proteins, including QueE and QueG involved in queuosine (Q) biosynthesis⁴¹. Clearly more work is needed to determine the generality of ROS-sensitive Fe-S cluster RNA-modifying proteins in other organisms and homologous epitranscriptome-mediated signaling networks. It is certainly possible that RlmN is uniquely sensitive as a potential redox signaling node in E. faecalis. However, in addition to MiaB, there are other Fe-S cluster-containing RNA-modifying enzymes⁴² that, like RlmN, may serve as redox responsive regulators. A systematic analysis of ROS-sensitive RNA modifications in different microbes would advance our understanding of oxidative stress response networks in bacteria.

In conclusion, we showed that RlmN activity is not only strongly dependent upon superoxide levels but also regulates levels of SodA. In E. faecalis and other bacteria, SodA promotes antibiotic tolerance³⁰ and facilitates survival in macrophages²³. In all, RlmN, widely distributed across bacteria genera⁴³, may serve as a redox switch relaying redox sensing to both the rRNA and tRNA epitranscriptome for direct modulation of translation for protective oxidative stress response.

Methods

Bacteria strains, plasmids, and growth conditions

Enterococcus faecalis strains OG1RF and V583 are grown in tryptic soy broth (TSB) or plated on tryptic soy agar under aerobic conditions at 37 °C. All mutant strains are derivatives of OG1RF. The rlmN knockout in OG1RF (ΔrlmN) was generated by an in-frame deletion of rlmN from OG1RF by allelic replacement using vector pGCP213²². The 250 bp regions upstream and downstream of the coding sequence of rlmN was amplified from OG1RF genomic DNA and stitched together by overlapping PCR and inserted into the XhoI/KpnI sites of pGCP213 (Supplementary Tables 5 and 6). This plasmid was transformed into OG1RF and introduced into the specific sites on the chromosome of the parental strain by recombinase-mediated gene replacement.

RlmN over-expressor, OG1RFprlmN, was generated by the introduction of the gene coding for RlmN into the plasmid pGCP123 under the constitutive sortase promoter²². The coding sequence for RlmN was PCR amplified with a His6 tag at the C-terminus and inserted into the XhoI/NotI sites of pGCP123 under a Sortase A promoter (Supplementary Tables 5, 6). OG1RFpEmpty is OG1RF carrying the plasmid pGCP123, but without introduction of the gene encoding RlmN. All the plasmids used in this study are listed in Supplementary Table 5. All the primers for cloning and RT-qPCR for making the mutant strains are listed in Supplementary Tables 6 and 7. All plasmid constructions were verified by Sanger sequencing. MICs used for OG1RF are 1 µg/mL erythromycin, 1 µg/mL spiramycin, 4 µg/mL chloramphenicol, 64 µg/mL kanamycin, 128 µg/mL streptomycin, 1 µg/mL ampicillin, 0.5 µg/mL tetracycline and 1 µg/mL ciprofloxacin. MICs used for V583 are >1024 µg/mL erythromycin, 8 µg/mL chloramphenicol, 256 µg/mL gentamicin, 1 µg/mL ampicillin, and 1 µg/mL ciprofloxacin. OG1RFprlmN and OG1RFpEmpty are grown in 500 µg/mL kanamycin to maintain the pGCP123 plasmid.

RNA isolation and purification

An overnight culture was diluted 1:20 fold and then grown at 37 °C to reach mid-log phase. This mid-log phase culture was then diluted 1:20 into media containing sublethal concentrations of antibiotics and grown at 37 °C with shaking at 180 rpm. The cultures are harvested at an optical density (OD₆₀₀) of ~0.6–0.8 after 4–5 doublings by centrifugation at 4000 × g for 10 min at 4 °C. RNA was extracted and purified following Hia et al.⁴⁴. Briefly, 100 mL of bacteria culture was lysed in the presence of phenol:chloroform:isoamyl alcohol and 100 mM sodium acetate pH 5.0 by bead-beating with 0.1-mm zirconia-silica beads using Qiagen TissueLyser II for 12 min at 30 Hz. Large and small RNA species were differentially recovered using the PureLink miRNA Isolation Kit (Invitrogen) with 35% ethanol and 70% ethanol respectively. 23 S and 16 S rRNA are separately isolated to purity from the large RNA fraction following HPLC using the Bio SEC-5 column (Agilent; 7.8 mm, length: 300 mm, particle size: 5 μm, pore size: 1000 Å); and tRNA was isolated to purity from the small RNA fraction following HPLC on the Bio SEC-3 column (Agilent; 7.8 mm, length: 300 mm, particle size: 5 μm, pore size: 300 Å). All separations were run with 100 mM ammonium acetate, at 60 °C with a flow rate of 0.5 mL/min. The elution profiles for all samples are shown in Supplementary Fig. 1, which reveals highly purified samples of high-quality 16 S and 23 S rRNAs and tRNAs. HPLC fractions containing the target RNA populations were combined, concentrated, and subjected to buffer exchange into 10 mM ammonium acetate using 3000 Da cutoff size exclusion filters (Millipore). RNA was quantified by NanoDrop spectroscopy.

Quantification of ribonucleosides by chromatography-coupled tandem mass spectrometry

For quantification of ribonucleosides, RNA (5 µg) was enzymatically hydrolyzed for 4 h at 37 °C with benzonase (99% purity, Novagen 70664), bacterial phosphatase (ThermoFisher 18011015) and phosphodiesterase I (Sigma P3243), in the presence of magnesium chloride, antioxidants and deaminase inhibitors including desferroxamine (Sigma D9533), butylated hydroxyltoluene (Sigma W218405), pentostatin (Sigma SML0508), tetrahydrouridine (Calbiochem 584222), and internal standard [¹⁵N]₅-deoxyadenosine. Samples were cleaned up using a 10 kDa cutoff filter (Nanosep). The ribonucleoside mixtures were resolved on a Hypersil C18 analytical column (2.1 × 100 mm, 1.8 mm; Agilent) mounted on an Agilent 1290 HPLC system and linked to an Agilent 6490 triple-quadrupole mass spectrometer using multiple reaction monitoring in positive-ion mode. ESI parameters used were gas temperature 80 °C; gas flow 11 L/min; nebulizer pressure 20 psi; sheath gas temperature 300 °C; capillary voltage 1800 V; nozzle voltage 2000 V.

The ribonucleosides were identified using the retention times of standards and MS/MS mass transitions involving loss of either ribose (136 m/z) or 2’-O-methyl-ribose (146 m/z) (Supplementary Table 1). For relative quantitation of modifications among the same batch of samples, the signal intensity is normalized against the combined intensity of the four canonical ribonucleosides to correct for variation in RNA quantities. Spectral signals are also normalized against spiked internal standard ([¹⁵N]₅−2’-deoxyadenosine) to adjust for variations in instrument sensitivity. For absolute quantification of m²A and adenosine, a series of concentrations of nucleoside standards for m²A and adenosine were run with every batch of samples to obtain standard calibration curves. The concentrations of nucleosides were then obtained by fitting the signal intensities onto the calibration curves, and these were then used to obtain the molar ratio of m²A/A.

Absolute quantification of m²A levels in rRNA and tRNA using calibration curves for adenosine and m²A reveals that m²A is present in 23 S rRNA in a ratio of approximately 0.00025 m²A per adenosine and at a 2.5-times higher ratio of 0.001 in tRNA. The level of reduction in m²A is identical between rRNA and tRNA, which decreases by ten-fold between untreated E. faecalis and cultures grown in 100-200 µg/mL of erythromycin (Fig. 1c, d). We approximated the abundance of m²A in rRNA and tRNA based on 23 S rRNA at 2904 nucleotides long and tRNA typically 76-90 nucleotides long, which equates to ~1m²A in 3.6 23 S rRNAs and ~1 in every 60 tRNAs in untreated V583 grown aerobically.

Measurement of rlmN mRNA levels

Total RNA (2 µg) was subjected to DNA removal using the TURBO DNA-free kit (Ambion, Life Technologies) following manufacturer’s protocol. 600 ng in 15 µL was used for reverse transcription using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA, USA). The reverse transcription program was run as follows: 25 °C for 5 min, 42 °C for 30 min, and 85 °C for 5 min, followed by a cooling step at 4 °C. Two-step real-time quantitative polymerase chain reaction (qPCR) was then performed using the BlitzAmp qPCR mastermix (MiRxes, Singapore). Primer sequences can be found in Supplementary Table 8. The qPCR program was run as follows: 95 °C for 5 min followed by 40 cycles of denaturation at 95 °C for 10 s and annealing/extension at 60 °C for 30 s on a Bio-Rad CFX384 real-time PCR instrument and analyzed using the CFX manager 3.1. A melting curve analysis consisting of 0.5 °C increments from 65 to 95 °C was performed for all reactions to ascertain the specificity of the primers. RpoA served as an internal loading control.

Measurement of RlmN protein levels

Since we could not detect RlmN in the TMT quantitative proteomics analyses, we performed targeted proteomics to quantify RlmN. Bacteria pellets from 10 mL of log phase culture of OG1RF and V583 were resuspended in 250 µL of 50 mM Hepes pH 8, 8 M urea, 1 mM DTT, and homogenized by bead-beating followed by clarification by centrifugation at 16,000 × g for 30 min at 4 °C. Following quantification of total protein in the supernatant using the bicinchoninic acid protein assay (ThermoFisher Scientific), 50 µg of protein was mixed with SDS-PAGE loading dye and separated on a 14% SDS-PAGE gel. Following staining and destaining, a gel slice corresponding to 30–45 kDa, encompassing target protein RlmN 40.9 kDa and reference proteins RpoA 35.05 kda and Gap2 35.77 kDa, was excised and cut into 1–2 mm pieces (Supplementary Fig. 4). The gel pieces were destained, reduced and alkylated followed by overnight trypsin digestion and peptide extraction following manufacturer’s instructions (In-Gel Tryptic Digestion kit, ThermoFisher Scientific). Extracted peptides were vacuum dried and redissolved in 2% acetonitrile in 0.1% formic acid in water.

Targeted quantification of RlmN was then achieved by first using Skyline (http://proteome.gs.washington.edu/software/skyline) to identify precursor peptides and transitions for RlmN and reference proteins RpoA and Gap2 (Supplementary Table 8). We analyzed the protein sequences of OG1RF RlmN, RpoA, and Gap2 against a background of the OG1RF proteome, looking for peptides 8–14 amino acids long, with the following Skyline settings: (1) trypsin cleavage; (2) minimum length 7 amino acids, maximum length of 25 amino acids, excluding 25 N-terminal amino acids; (3) excluding peptides containing Cys, Met, His, NXT/NXS, RP/KP; (4) including potential structural modifications (e.g., carbamidomethyl); and (5) maximum variable modifications at 3 and maximum neutral losses at 1. For RlmN, Skyline identified 3 target peptides: (1) K.QVIVQEAQDGTVK.Y [67, 79], (2) K.YLFELPDK.N [80, 87], and (3) K.GLAIGAR.H [184, 190], with peptides #1 and #2 selected. For Gap2, Skyline identified 5 target peptides: (1) K.YDTTQGR.F [46, 52], (2) K.AIGLVIPELNGK.L [215, 226], (3) K.LDGAAQR.V [227, 233], (4) R.VPVATGSLTELVTVLDK.E [234, 250], and (5) R.TLEYFANL.- [325, 332], with peptide #2 selected. For RpoA, Skyline identified 2 target peptides: (1) R.EDVTQIILNIK.G [70, 80] and (2) K.LYAEEEK.T [86, 92], with peptide #1 selected. Selected peptides were synthesized at 90% purity and analyzed by LC-MS/MS on a Hypersil C18 analytical column (2.1 × 100 mm, 1.8 mm; Agilent) mounted on an Agilent 1290 infinity LC system coupled to an Agilent 6490 QQQ spectrometer in positive-ion mode. Agilent Automated MRM Method Optimizer for Peptides was used to optimize collision energies and fragmentation voltages for their MRM transitions and peptides were used at a concentration of 10 µg/mL for determination of retention times. Reversed-phase chromatography was performed with a fixed flow rate of 0.25 mL/min with a gradient of water and acetonitrile (solvent B) acidified with 0.1% (v/v) formic acid. Gradients used were as follow: 0–29% solvent B from 0–29 min, 29–90% from 29–30 min, 90% for 38 min, 90 to 0 % from 38 to 39 min, and 0% for 45 min. Source conditions: gas temperature 325 °C, gas flow 10 L/min, nebulizer 32 psi, sheath gas temperature 300 °C, sheath gas flow 11 L/min, capillary 2000 V, charging 500 V. Columns were incubated at 40 °C. The top two precursor peptides by peak area and number of transitions (minimum 2) were selected as qualification and quantification ions. Following the definition of the analytical parameters for the synthetic peptides, the peptides extracted from the gel slices were then analyzed on the same LC-MS/MS system. Signal intensities for each peptide were normalized to the total signal intensity for the LC-MS/MS run to account for peptide loading differences and the fold-change data were calculated by dividing normalized signal intensities for antibiotic-treated samples by those in control samples.

Flow cytometry assays

CellROX Green (ThermoFisher) is a proprietary oxidation-sensitive dye whose fluorescence at 500–550 nm after excitation at 488 nm increases substantially on oxidation in the presence of dsDNA. Cellrox green reacts to hydroxyl radical and superoxide but not hydrogen peroxide¹⁵. Log-phase cultures were diluted to OD₆₀₀ of 0.1 in 10% TSB in the presence of menadione or antibiotics and incubated for 30 min at 37 °C followed by the addition of CellROX green (final 0.5 µM) for a further 30 mins at 37 °C in the dark with shaking at 180 rpm. Samples were analyzed using a HTS autosampler system on the Attune Nxt v4.2.0 (ThermoFisher Scientific) and the flow rate was set to 25 µL/min with a 30 µL injection volume. Samples were detected with a 530/30 nm band-pass emission and recording 50,000 events in the bacterial gate. Gating was set using unstained samples for the bacterial population by forward-scatter (FSC; correlates with cell size) and side-scatter (SSC; correlates with cell internal granularity) of light to determine background fluorescence. Data were analyzed with the Attune Nxt software.

Proteomics

Fresh mid-log phase cultures (OD₆₀₀ of 0.6) were diluted 1:20 into TSB media with or without treatment and grown at 37 °C with shaking at 180 rpm. The cultures are harvested at an optical density (OD₆₀₀) of ~0.6–0.8 after 4–5 doublings by centrifugation at 4000 × g for 10 min at 4 °C. Bacterial pellets were resuspended in 250 µL of 50 mM Hepes pH 8, 8 M urea, 1 mM DTT, and homogenized by bead-beating followed by clarification by centrifugation at 16,000 × g for 30 min at 4 °C. Following protein quantification in the supernatant using the bicinchoninic acid protein assay (ThermoFisher Scientific), 200 µg of protein was digested with trypsin after being reduced with DTT and alkylated with iodoacetamide. After washing 3 times with 0.5 M TEAB followed by fractionation using the 10 kDa ultrafiltration system, ~100 μg of digested peptides from each group, including two biological replicates, was labeled using the six-plex TMT isobaric and isotopic mass-tagging kit (ThermoFisher Scientific), which was performed according to manufacturer’s instructions.

Peptides were separated by reverse phase HPLC (Thermo Easy nLC1000) using a precolumn (made in house, 6 cm of 10 µm C18) and a self-pack 5 µm tip analytical column (15 cm of 5 µm C18, New Objective) over a 150 min gradient before nanoelectrospray using a QExactive mass spectrometer (ThermoFisher). The mass spectrometer was operated in a data-dependent mode. The parameters for the full scan MS were as follow: resolution of 70,000 across 350–2000 m/z, AGC 3e6, and maximum IT 50 ms. The full MS scan was followed by MS/MS for the top 15 precursor ions in each cycle with a NCE of 28 and dynamic exclusion of 30 s. Raw mass spectral data files (.raw) were searched using Proteome Discoverer (ThermoFisher) and Sequest. Search parameters were as follows: 10 ppm mass tolerance for precursor ions; 0.8 Da for fragment ion mass tolerance; 2 missed cleavages of trypsin; fixed modification was carbamidomethylation of cysteine, and N-term and Lysine TMT-label; variable modifications were methionine oxidation and serine, threonine and tyrosine phosphorylation. Only peptides with a Scorer score ≥2, reporter channel intensity >500 and an isolation interference ≤30 were included in the data analysis. Proteomics data are presented in Supplementary Data 1.

Bactericidal activity analysis

A log phase culture of OG1RF was diluted to a final concentration of 10⁸ CFU/mL (OD₆₀₀ ~ 0.1) in Tryptic Soy Broth in the presence of antibiotics at 37 °C with shaking. Aliquots were drawn from each respective tube at various time points and serially diluted until 10⁻⁸-fold. An aliquot (2.5 µL) of each dilution was spotted onto TSB agar and incubated at 37 °C overnight, with colonies counted 24 h after spotting.

Determination of minimal inhibitory concentrations (MIC) of antibiotics

Two-fold serial dilutions of antibiotics in TSB were performed in separate rows of a polystyrene 96-well plate (Corning) with each plate containing an inoculum of respective bacteria. The inoculum was a 1:500 dilution from a culture at log phase (OD₆₀₀ = 0.5) grown at 37 °C. The plate was incubated with shaking at 37 °C and the optical density of each well was measured at a wavelength of 600 nm (BIOTEK, Synergy 4). The MIC values were taken as the lowest concentration for which no growth was discernible (<0.05 OD₆₀₀) after 24 h. All tests were performed three times independently with two samples in each test. MIC data are presented in Supplementary Table 2.

Antibiotic killing assays

The OG1RF strains were cultured aerobically in TSB 37 °C for approximately 16 h with shaking (180 rpm) followed by 1:20 dilution and cultured to mid-log growth. Strains OG1RFpEmpty and OG1RFprlmN harboring pGCP123 plasmids were grown in 500 µg/mL kanamycin. The mid-log culture was diluted to OD₆₀₀ ~ 0.06 (~5 × 10⁷ cfu/mL). An aliquot was plated to enumerate the colony-forming units (cfu) (Time 0) before the addition of antibiotics with final concentrations at 20 µg/mL ciprofloxacin and 20 µg/mL ampicillin. An aliquot was removed at the indicated time points and washed with sterile PBS. The cells were serially diluted and plated on tryptic soy agar to enumerate the survivors.

Statistical analysis

Statistical significance was assessed using appropriate tests using Prism 8 (GraphPad) software, detailed in their respective figure legends. Asterisks indicate the level of statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. P values < 0.05 were considered significant. Experiments were repeated at least three times.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The data supporting the findings of this study are available from the corresponding authors upon reasonable request. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the accession code PXD038178. The source data used to generate plots are provided as a Source Data file. Source data are provided with this paper.

References

Imlay, J. A. Cellular defenses against superoxide and hydrogen peroxide. Annu. Rev. Biochem. 77, 755–776 (2008).
Article CAS PubMed PubMed Central Google Scholar
Imlay, J. A. Where in the world do bacteria experience oxidative stress? Environ. Microbiol. 21, 521–530 (2019).
Article CAS PubMed Google Scholar
Staerck, C. et al. Microbial antioxidant defense enzymes. Microb. Pathog. 110, 56–65 (2017).
Article CAS PubMed Google Scholar
Chiang, S. M. & Schellhorn, H. E. Regulators of oxidative stress response genes in Escherichia coli and their functional conservation in bacteria. Arch. Biochem. Biophys. 525, 161–169 (2012).
Article CAS PubMed Google Scholar
Fassler, R., Zuily, L., Lahrach, N., Ilbert, M. & Reichmann, D. The central role of redox-regulated switch proteins in bacteria. Front. Mol. Biosci. 8, 706039 (2021).
Article CAS PubMed PubMed Central Google Scholar
Green, J. & Paget, M. S. Bacterial redox sensors. Nat. Rev. Microbiol. 2, 954–966 (2004).
Article CAS PubMed Google Scholar
Chionh, Y. H. et al. tRNA-mediated codon-biased translation in mycobacterial hypoxic persistence. Nat. Commun. 7, 13302 (2016).
Article CAS PubMed PubMed Central ADS Google Scholar
Hofer, K. & Jaschke, A. Epitranscriptomics: RNA modifications in bacteria and archaea. Microbiol. Spectr. 6, (2018).
Bourgogne, A. et al. Large scale variation in Enterococcus faecalis illustrated by the genome analysis of strain OG1RF. Genome Biol. 9, R110 (2008).
Article PubMed PubMed Central Google Scholar
Paulsen, I. T. et al. Role of mobile DNA in the evolution of vancomycin-resistant Enterococcus faecalis. Science 299, 2071–2074 (2003).
Article CAS PubMed ADS Google Scholar
Vester, B. & Long, K. Antibiotic resistance in bacteria caused by modified nucleosides in 23s ribosomal RNA. Madame Curie Bioscience Database [Internet] (2009). https://www.ncbi.nlm.nih.gov/books/NBK6514/.
Aakra, A. et al. Transcriptional response of Enterococcus faecalis V583 to erythromycin. Antimicrob Agents Chemother. 2005;49: 2246–2259.
Golan, D., Armstrong, E. & Armstrong, A. Principles of pharmacology. The pathophysiology basis of drug therapy. 7th edn, (Wolters Kluwer, 2017).
Iyanagi, T. On the mechanism of one-electron reduction of quinones by microsomal flavin enzymes: the kinetic analysis between cytochrome B5 and menadione. Free. Radic Res. Commun. 8, 259–268 (1990).
Article CAS PubMed Google Scholar
McBee, M. E. et al. Production of superoxide in bacteria is stress- and cell state-dependent: a gating-optimized flow cytometry method that minimizes ROS measurement artifacts with fluorescent dyes. Front. Microbiol. 8, 459 (2017).
Article PubMed PubMed Central Google Scholar
Imlay, J. A. Diagnosing oxidative stress in bacteria: not as easy as you might think. Curr. Opin. Microbiol. 24, 124–131 (2015).
Article CAS PubMed PubMed Central Google Scholar
Leger, L. et al. Beta-lactam exposure triggers reactive oxygen species formation in enterococcus faecalis via the respiratory chain component DMK. Cell Rep. 29, 2184–2191.e2183 (2019).
Article CAS PubMed Google Scholar
Huycke, M. M. et al. Extracellular superoxide production by Enterococcus faecalis requires demethylmenaquinone and is attenuated by functional terminal quinol oxidases. Mol. Microbiol. 42, 729–740 (2001).
Article CAS PubMed Google Scholar
Van Acker, H. & Coenye, T. The role of reactive oxygen species in antibiotic-mediated killing of bacteria. Trends Microbiol. 25, 456–466 (2017).
Article PubMed Google Scholar
Dwyer, D. J., Kohanski, M. A. & Collins, J. J. Role of reactive oxygen species in antibiotic action and resistance. Curr. Opin. Microbiol. 12, 482–489 (2009).
Article CAS PubMed PubMed Central Google Scholar
Pankey, G. A. & Sabath, L. D. Clinical relevance of bacteriostatic versus bactericidal mechanisms of action in the treatment of Gram-positive bacterial infections. Clin. Infect. Dis. 38, 864–870 (2004).
Article CAS PubMed Google Scholar
Nielsen, H. V. et al. The metal ion-dependent adhesion site motif of the Enterococcus faecalis EbpA pilin mediates pilus function in catheter-associated urinary tract infection. mBio 3, e00177–00112 (2012).
Article CAS PubMed PubMed Central Google Scholar
Peppoloni, S. et al. Role of the (Mn)superoxide dismutase of Enterococcus faecalis in the in vitro interaction with microglia. Microbiol. 157, 1816–1822 (2011).
Article CAS Google Scholar
Sillanpaa, J. et al. Contribution of individual Ebp Pilus subunits of Enterococcus faecalis OG1RF to pilus biogenesis, biofilm formation and urinary tract infection. PLoS One 8, e68813 (2013).
Article CAS PubMed PubMed Central ADS Google Scholar
Antunes, A. et al. The phosphocarrier protein HPr contributes to meningococcal survival during infection. PLoS One 11, e0162434 (2016).
Article PubMed PubMed Central Google Scholar
Herro, R. et al. How seryl-phosphorylated HPr inhibits PrfA, a transcription activator of Listeria monocytogenes virulence genes. J Mol. Microbiol. Biotechnol. 9, 224–234 (2005).
CAS PubMed Google Scholar
Galloway-Peña Jessica, R. et al. The identification and functional characterization of wxl proteins from enterococcus faecium reveal surface proteins involved in extracellular matrix interactions. J. Bacteriol. 197, 882–892 (2015).
Article PubMed PubMed Central Google Scholar
Grove, T. L. et al. A radically different mechanism for S-adenosylmethionine-dependent methyltransferases. Science 332, 604–607 (2011).
Article CAS PubMed ADS Google Scholar
Yan, F. & Fujimori, D. G. RNA methylation by radical SAM enzymes RlmN and Cfr proceeds via methylene transfer and hydride shift. Proc. Natl Acad. Sci. USA 108, 3930–3934 (2011).
Article CAS PubMed PubMed Central ADS Google Scholar
Bizzini, A., Zhao, C., Auffray, Y. & Hartke, A. The Enterococcus faecalis superoxide dismutase is essential for its tolerance to vancomycin and penicillin. J. Antimicrob. Chemother. 64, 1196–1202 (2009).
Article CAS PubMed Google Scholar
Weisblum, B. Erythromycin resistance by ribosome modification. Antimicrob. Agents Chemother. 39, 577–585 (1995).
Article CAS PubMed PubMed Central Google Scholar
Vázquez-Laslop, N., Ramu, H., Klepacki, D., Kannan, K. & Mankin, A. S. The key function of a conserved and modified rRNA residue in the ribosomal response to the nascent peptide. EMBO J. 29, 3108–3117 (2010).
Article PubMed PubMed Central Google Scholar
Wang, S. et al. Translational attenuation mechanism of ermb induction by erythromycin is dependent on two leader peptides. Front. Microbiol. 12, 690744 (2021).
Article PubMed PubMed Central Google Scholar
Stojković, V., Chu, T., Therizols, G., Weinberg, D. E. & Fujimori, D. G. miCLIP-MaPseq, a substrate identification approach for radical sAM RNA methylating enzymes. J. Am. Chem. Soc. 140, 7135–7143 (2018).
Article PubMed PubMed Central Google Scholar
Urbonavicius, J., Qian, Q., Durand, J. M., Hagervall, T. G. & Björk, G. R. Improvement of reading frame maintenance is a common function for several tRNA modifications. EMBO J. 20, 4863–4873 (2001).
Article CAS PubMed PubMed Central Google Scholar
Benítez-Páez, A., Villarroya, M. & Armengod, M. E. The Escherichia coli RlmN methyltransferase is a dual-specificity enzyme that modifies both rRNA and tRNA and controls translational accuracy. RNA 18, 1783–1795 (2012).
Article PubMed PubMed Central Google Scholar
Zhang, Y., Romero, H., Salinas, G. & Gladyshev, V. N. Dynamic evolution of selenocysteine utilization in bacteria: a balance between selenoprotein loss and evolution of selenocysteine from redox active cysteine residues. Genome Biol. 7, R94 (2006).
Article PubMed PubMed Central Google Scholar
Ramsey, M., Hartke, A. & Huycke, M. In: Enterococci: From Commensals to Leading Causes of Drug Resistant Infection (eds. M. S. Gilmore, D. B. Clewell, Y. Ike, & N. Shankar) (Massachusetts Eye and Ear Infirmary, 2014).
Feld, L., Knudsen, G. M. & Gram, L. Bactericidal antibiotics do not appear to cause oxidative stress in Listeria monocytogenes. Appl. Environ. Microbiol. 78, 4353–4357 (2012).
Article CAS PubMed PubMed Central ADS Google Scholar
Riboldi, G. P., Larson, T. J. & Frazzon, J. Enterococcus faecalis sufCDSUB complements Escherichia coli sufABCDSE. FEMS Microbiol. Lett. 320, 15–24 (2011).
Article CAS PubMed Google Scholar
Hutinet, G., Swarjo, M. A. & de Crecy-Lagard, V. Deazaguanine derivatives, examples of crosstalk between RNA and DNA modification pathways. RNA Biol. 14, 1175–1184 (2017).
Article PubMed Google Scholar
Kimura, S. & Suzuki, T. Iron-sulfur proteins responsible for RNA modifications. Biochim. Biophys. Acta 1853, 1272–1283 (2015).
Article CAS PubMed Google Scholar
Atkinson, G. C. et al. Distinction between the Cfr methyltransferase conferring antibiotic resistance and the housekeeping rlmn methyltransferase. Antimicrob. Agents Chemother. 57, 4019–4026 (2013).
Article CAS PubMed PubMed Central Google Scholar
Hia, F. et al. Mycobacterial RNA isolation optimized for non-coding RNA: high fidelity isolation of 5S rRNA from Mycobacterium bovis BCG reveals novel post-transcriptional processing and a complete spectrum of modified ribonucleosides. Nucleic Acids Res. 43, e32 (2015).
Article PubMed Google Scholar

Download references

Acknowledgements

The authors thank Dr. Michael DeMott for his assistance with the proteomics work. This research was supported by the National Research Foundation, Prime Minister’s Office, Singapore, under its Campus for Research Excellence and Technological Enterprise (CREATE) program, through the Singapore–MIT Alliance for Research and Technology Antimicrobial Resistance Interdisciplinary Research Group. We acknowledge funding support from NMRC OFIRG20nov-0079. Contributions of LNL were supported by the National Research Foundation and Ministry of Education Singapore under its Research Center of Excellence Program at SCELSE. A.S. and J.L. acknowledge support from the Singapore‐MIT Alliance (SMA) Graduate Fellowship and MOE Tier 2 Grant MOE2018-T2-2-13. Proteomics work was performed in part in the MIT Center for Environmental Health Sciences Bioanalytical Core, which is supported by Center grant P30‐ES002109 from the National Institute of Environmental Health Sciences with the aid of Dr. Michael Demott.

Author information

Ameya Sinha
Present address: Helmholtz-Zentrum für Infektionsforschung GmbH, Inhoffenstraße 7, 38124, Braunschweig, Germany
Ling Ning Lam
Present address: Department of Oral Biology, University of Florida College of Dentistry, Gainesville, FL, USA
Jiaqi Liang
Present address: School of Chemistry, Chemical Engineering and Biotechnology, College of Engineering, Nanyang Technological University, Singapore, Singapore
Cheryl Siew Choo Chan
Present address: Critical Analytics for Manufacturing Personalized-Medicine IRG, Singapore MIT Alliance for Research and Technology, Singapore, Singapore

Authors and Affiliations

Antimicrobial Resistance IRG, Singapore MIT Alliance for Research and Technology, Singapore, Singapore
Wei Lin Lee, Ameya Sinha, Hooi Linn Loo, Jiaqi Liang, Peiying Ho, Liang Cui, Cheryl Siew Choo Chan, Kimberly Ann Kline & Peter Dedon
School of Biological Sciences, Nanyang Technological University, Singapore, Singapore
Ameya Sinha, Ling Ning Lam & Kimberly Ann Kline
Singapore Centre for Environmental Life Sciences Engineering, Nanyang Technological University, Singapore, Singapore
Ling Ning Lam & Kimberly Ann Kline
Department of Biological Sciences and The RNA Institute, University at Albany, Albany, NY, USA
Thomas Begley
Department of Microbiology and Molecular Medicine, Faculty of Medicine, University of Geneva, Geneva, Switzerland
Kimberly Ann Kline
Dept. of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA
Peter Dedon

Authors

Wei Lin Lee
View author publications
You can also search for this author in PubMed Google Scholar
Ameya Sinha
View author publications
You can also search for this author in PubMed Google Scholar
Ling Ning Lam
View author publications
You can also search for this author in PubMed Google Scholar
Hooi Linn Loo
View author publications
You can also search for this author in PubMed Google Scholar
Jiaqi Liang
View author publications
You can also search for this author in PubMed Google Scholar
Peiying Ho
View author publications
You can also search for this author in PubMed Google Scholar
Liang Cui
View author publications
You can also search for this author in PubMed Google Scholar
Cheryl Siew Choo Chan
View author publications
You can also search for this author in PubMed Google Scholar
Thomas Begley
View author publications
You can also search for this author in PubMed Google Scholar
Kimberly Ann Kline
View author publications
You can also search for this author in PubMed Google Scholar
Peter Dedon
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

W.L.L., K.A.K., and P.D. designed research; W.L.L., A.S., L.N.L., H.L.L., J.L., P.H., and L.C. performed research; C.S.C.C. and T.B. contributed new reagents/analytic tools; W.L.L., A.S., T.B., K.A.K., and P.D. analyzed data, all authors participated in writing the manuscript.

Corresponding author

Correspondence to Peter Dedon.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Lee, W.L., Sinha, A., Lam, L.N. et al. An RNA modification enzyme directly senses reactive oxygen species for translational regulation in Enterococcus faecalis. Nat Commun 14, 4093 (2023). https://doi.org/10.1038/s41467-023-39790-x

Download citation

Received: 28 October 2022
Accepted: 27 June 2023
Published: 11 July 2023
DOI: https://doi.org/10.1038/s41467-023-39790-x

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.