{"paper_id":"07ffc591-78ee-4ac4-90ad-a3842f48abb7","body_text":"1 \n \n \nPmt1-dependent and state-restricted DNA methylation in fission yeast \n \n \nSamia Miled (1)* \n \n \n(1) Institut Jacques Monod, trafic membranaire, ubiquitine et signalisation, UMR7592, CNRS, Université \nParis Cité, Bâtiment Buffon, 15 rue Hélène Brion, 75205 Paris Cedex 13, France. \n \n* Corresponding author: samia.miled@ijm.fr \n \nAbstract \n \nDNA 5-methylcytosine (5mC) is widely considered absent from budding and fission yeasts. The Dnmt2 family \nenzyme Pmt1 in Schizosaccharomyces pombe is annotated as a tRNA C38 methyltransferase, but whether it \nmethylates DNA in vivo has remained unresolved. Using orthogonal chemistry (HPLC/LC-MS and 1H-NMR) \nwith external standards (limit of detection for 5mdC = 0.0125 mM), we detect authentic 5 -methyl-2'-\ndeoxycytidine (5mdC) in enzymatically digested genomic DNA from S. pombe. Quantification across isogenic \nstrains shows that 5mdC is undetectable in vegetative wild type but rises to 0.518% (plus or minus 0.025) of \ncytosines two hours after G0 exit. Loss of the cytosine/5 -mC deaminase Fcy1 causes marked accumulation: \n1.445% (plus or minus 0.361) in vegetative cells and 5.943% (plus or minus 1.364) at 2 hours, approximately \n11.5-fold over wild type. By contrast, pmt1Δ and fcy1Δ pmt1Δ  remain below detection in all conditions, \nestablishing Pmt1 as the DNA -directed methyltransferase in vivo. Nascent -strand fractionation and lambda \nexonuclease enrichment place this transient 5mdC pulse on Okazaki -enriched DNA, consistent with co -\nreplicative installation. Functionally, the first S phase after quiescence shows increased Rad22 -YFP foci in \nfcy1Δ, which are dampened by queuine ; in a sensitized background (ung1Δ thp1Δ), deleting fcy1 reduces C-\n>T transitions by about 30 percent. Together, chemical, genetic, and temporal evidence reveals a Pmt1 -\ndependent, state -restricted DNA methylation program in fission yeast that is rapidly curtailed by Fcy1, \nredefining the epigenetic landscape of S. pombe and providing a minimal, tractable system to dissect regulated \nDNA methylation in eukaryotes. \n \nKeywords  \n \nSchizosaccharomyces pombe; DNA 5-methylcytosine (5mC); Pmt1 (Dnmt2/TRDMT1); Fcy1 deaminase; \nstate-restricted methylation; G0-to-S transition; replication-coupled methylation; Okazaki fragments \n(nascent DNA); methylation-deamination flux; Rad22-YFP foci; queuine metabolism.  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 2 \nIntroduction \nCytosine 5-methylation (5mC) is a pervasive epigenetic mark across eukaryotes, but its distribution is highly \nuneven across lineages. Early comparative surveys established a canonical view in \nwhich Schizosaccharomyces pombe  lacks detectable genomic 5mC, in stark contrast to many plants and \nanimals [1–3]. The strength of that consensus has rested largely on bisulfite -based assays and limited mass -\nspectrometric analyses conducted under conditions that were not designed to capture transient or low -\nstoichiometry DNA methylation [1,4].  Yet it is now clear that conventional whole -genome bisulfite \nsequencing (WGBS) can introduce sequence -specific coverage and conversion biases that complicate \ninterpretation at low abundance [4]. Enzymatic methyl -seq (EM-seq) and orthogonal chemistries hav e since \nimproved sensitivity and fidelity in difficult settings, particularly at low input and in GC-rich or damage-prone \nregions [5–7]. In parallel, targeted HPLC/LC–MS pipelines coupled to authentic standards and NMR overlays \nprovide definitive chemical identity for modified deoxynucleosides, enabling rigorous calls near the analytical \nlimit of detection [16–19]. \nThe S. pombe genome encodes Pmt1, a homolog of the Dnmt2/TRDMT1 family [9]. Dnmt2 proteins were \nlong annotated as DNA methyltransferases by sequence homology, but seminal work showed that they \nmethylate cytosine-38 in the anticodon loop of specific tRNAs, notably tRNA Asp, and protect tRNAs from \nstress-induced cleavage [10–13]. In S. pombe, Pmt1 methylates tRNAAsp in vivo, with activity modulated by \nnutrient status via Sck2 signaling and by queuosine in the anticodon loop, which can stimulate Dnmt2 catalysis \nand alter tRNA decoding properties [9 –11,14,15]. These observations raise a fundamental question: can a \nDnmt2-family enzyme in a yeast widely considered “DNA-unmethylated” also catalyze DNA 5mC formation \nunder specific physiological contexts? \nHere we uncover a Pmt1-dependent DNA 5mC signal in S. pombe. Using validated nucleoside-level analytics \n(HPLC/LC–MS with external calibration and ICH Q2(R2)–compliant LOD/LOQ) and NMR confirmation of \ncollected 5mdC fractions [16 –20], we detect 5mdC that (i) rises acutely as cells re -enter the cell cycle from \nnitrogen-starvation quiescence (G0→S), (ii) is abolished in  pmt1Δ, and (iii) accumulates \nin fcy1Δ backgrounds lacking the cytosine deaminase Fcy1 , consistent with deamination -driven turnover of \nmethyl-cytosine [21]. Live-cell Rad22-YFP imaging and DNA-content cytometry place this methylation pulse \nin the first S phase after quiescence, a window characterized by elevated recombination/repair foci and \nreplication stress in  S. pombe  [22–26]. Together, these data identify Pmt1 as a bona fide  DNA \nmethyltransferase in fission yeast and suggest that 5mC formation is  cell-state dependent, transient, and \nactively countered by base deamination. Beyond resolving a long -standing controversy f or S. pombe [1–3], \nour work motivates a reevaluation of Dnmt2 catalytic plasticity and the contexts in which low -stoichiometry \nDNA methylation can be biologically meaningful. \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 3 \nMethods \nStrains, media, and general conditions \nIsogenic Schizosaccharomyces pombe  strains were  wild type,  pmt1Δ, fcy1Δ and fcy1Δpmt1Δ. Cells were \ngrown at 30 °C in YES or EMM with agitation (200 rpm) unless stated otherwise. The  Rad22-YFP marker \nwas introduced by standard genetics and expressed from its native locus.  \nQuiescence (G0) induction and re-entry time course \nQuiescence was induced by nitrogen starvation in EMM -N following established  S. pombe workflows. Re-\nentry (“t = 0”) was triggered by transfer to nitrogen -replete medium. Samples were collected across the  first \ncell cycle, with the first S-phase window annotated by DNA-content cytometry and septation index as in prior \nquiescence/exit frameworks [27, 44, 45]. Viability was monitored by colony-forming units on YES plates. \nGenomic DNA isolation and RNA removal \nCell pellets were lysed mechanically/enzymatically. Genomic DNA was purified using phenol-free buffers or \nsilica-column kits, followed by  sequential RNA depletion  (RNase A/T1 → RNase H) and a short  alkaline \nhydrolysis to eliminate residual RNA. DNA quality was assessed by A260/280 and A260/230 ratios, absence \nof rRNA bands on agarose gels, and dsDNA-specific fluorimetry. This workflow minimizes RNA carry-over \nin nucleoside assays [56]. \nEnzymatic digestion to nucleosides \nDefined masses of genomic DNA were digested to  nucleosides using a standard sequential cocktail:  DNase \nI (Mg²⁺, pH ≈ 7.0) → nuclease P1 (Zn²⁺, pH ≈ 5.2) → alkaline phosphatase (pH ≈ 7.5). Reactions proceeded \nto completion (verified by plateau of total nucl eoside yield) and were filtered prior to chromatography. The \nDNase-I → NP1 → AP pipeline is widely adopted for quantitative nucleoside analysis and limits partial -\nhydrolysis bias [50, 49, 47]. \nHPLC/LC–MS identification and quantification of dC and 5mdC \nDigests and  authentic standards  (2′-deoxycytidine, 5 -methyl-2′-deoxycytidine) were injected in the same \nanalytical sessions. Separation used a reversed -phase column (C18, 3 –5 µm) with aqueous/organic eluents \ncompatible with UV (~254 nm) and positive -mode electrospray. Identity of  5mdC in genomic digests was \nassigned by co-elution with the standard and  MS signatures: [M+H] ⁺ = 242.1 in full scan and a  (2M+H)⁺ ≈ \n484.3 feature in product -ion spectra. External calibration curves converted areas to amounts; repr esentative \nchromatograms and linear fits (R²) are provided in Supplementary Figure. S1 [16, 17, 18, 19, 47–49]. \nAnalytical validation, LOD/LOQ, and calling rules \nMethod validation followed  ICH Q2(R2)  (effective 2024). For 5mdC, the  limit of detection (LOD)  - \ncalculated from calibration slope and baseline noise on blanks/low standards  - was 0.0125 mM  in our \nconfiguration; the limit of quantification (LOQ)  was derived using ICH equations. “ND” (not detected) is \ndefined as < LOD. Unless stated, values are mean ± SD from n ≥ 6 biological replicates. Inter- and intra-assay \nprecision were assessed on pooled digests analyzed across days [20]. \nNMR confirmation of 5mdC identity \nHPLC fractions corresponding to putative 5mdC were collected and concentrated.  ¹H (and, where \navailable, ¹³C) NMR spectra were acquired in D₂O and overlaid with a commercial 5mdC standard. Chemical-\nshift concordance  across hallmark resonances (C5/H5, C6/H6, deoxyribose protons) and the absence of \nextraneous peaks supported molecular identity [16, 19, 56]. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 4 \nTime-course and two-state snapshot designs \nFor the  time course, aliquots were collected at fixed intervals after refeeding; for the  two-state snapshot, \nsamples were taken in vegetative growth and at +2 h after G0 exit. At each point, % 5mdC/ (5mdC + 5dC) was \nmeasured by HPLC/LC–MS as above. Wild type, fcy1Δ, pmt1Δ, and fcy1Δ pmt1Δ were assayed side-by-side \nwithin the same analytical sessions. ND calls strictly followed the LOD rule [27, 20]. \nDNA-content flow cytometry \nCells were ethanol-fixed, treated with RNase A, stained with propidium iodide, and analyzed on a benchtop \nflow cytometer.  G1/S/G2 distributions  were used to delineate the  first S -phase after G0 exit and to \nsynchronize figure annotations with biochemical sampling [27, 44]. \nLive-cell imaging of Rad22-YFP and focus quantification \nCells expressing  Rad22-YFP were imaged on EMM pads with constant exposure/laser settings.  Foci were \nscored blinded to genotype as nuclei containing ≥ 1 Rad22-YFP focus; ≥ 200 cells per condition per replicate \nwere counted when feasible. Rad22 -positive foci are established  recombination/repair assemblies that \nincrease under replication stress in S. pombe, providing a live-cell readout during G0→S [22, 23, 51]. \nQueuine treatment during G0 exit and Rad22-YFP foci analysis \nStrains and culture conditions. All strains carry a C -terminal Rad22-YFP fusion at the \nendogenous rad22 locus. The four genotypes analyzed were  wild type (wt), pmt1Δ, fcy1Δ, and fcy1Δ pmt1Δ. \nCells were grown at  30 °C  in EMM2 minimal medium with supplements to mid -log phase, then shifted \nto nitrogen-free EMM (EMM -N) to induce quiescence (G0) as described in the Methods (typically  24 \nh starvation unless stated). \nNascent DNA preparation and sucrose-gradient fractionation (Okazaki-enrichment) \nCells were sampled at the indicated times after G0 release (Fig. 4A). Low -molecular-weight DNA was \nprepared and resolved on 10–40% sucrose gradients in nuclease-free buffer. Gradients were run on a piston -\ndriven fractionator; 12 fractions (F1 –F12) were collected at 50 µL  each from top to bottom. Fractions were \nanalysed on agarose gels to assess  size distributions  across the gradient (Fig. 4B). Where required, equal \nvolumes for a given  fraction index  were pooled across time points  (e.g., F1 = F1_G0 + F1_1h40 + …) to \nobtain sufficient material for enzymology and analytics. The workflow follows  short-nascent-strand \nabundance/enrichment best practices [65,66]. \nλ-Exonuclease enrichment of nascent strands \nPooled or individual fractions were incubated with λ-exonuclease under supplier-recommended conditions to \ndigest 5′-phosphorylated dsDNA  while sparing  RNA-primed nascent strands  that are  refractory to  5′→3′ \ndigestion [64,67]. Reactions were terminated, DNA recovered, and an aliquot run on agarose to confirm \ndepletion of λ-exo–sensitive species. We adopted published precautions to limit known  λ-exo \nsequence/structure biases (matched inputs, fraction-based pooling, no-enzyme controls) [64,65]. \nHPLC quantification of 5mdC in nascent DNA \nλ-exo–treated DNA from F1–F6 was enzymatically digested to nucleosides and analy zed by HPLC as in the \nmain text (calibration/LOD in Fig. S1).  %5mdC was computed as 5mdC/(5mdC+dC) × 100. Fractions \nfrom wt and fcy1Δ were processed side-by-side; biological replicates as indicated in Fig. 4D. \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 5 \nRelease from G0 and queuine treatment. G0 cultures were harvested, washed once in pre-warmed EMM2, \nand released into fresh EMM2  (time 0). Immediately upon release, cells were split into  ±queuine conditions \nand incubated for 1 h at 30 °C with shaking.  \nQueuine (+Q): queuine base ( hemi-sulfate), stock 10 mM in sterile water,  filter-sterilized, added to a  final \nconcentration of 10 µM. Unless indicated,  10 µM  was used for all experiments. After  1 h, samples were \nprocessed for imaging (below). In parallel, the same ±Q protocol was applied to each genotype \n(wt, pmt1Δ, fcy1Δ, fcy1Δ pmt1Δ). \nStatistics. Effects of  genotype and treatment (±Q)  were tested by  two-way ANOVA  (Prism software) . \nSignificance thresholds were pre-specified (e.g., p < 0.05, 0.01, 0.005 as annotated in figure legends). Graphs \ndisplay mean ± SD; exact n and p values are provided in the legends (see Fig. S2). \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 6 \nResults \nTo determine whether fission yeast carries a minimal yet functional DNA methylation program, we analyzed \nquiescent cells as they re‑entered the cell cycle and combined three layers of evidence: physiology (cell‑cycle \nprogression and DNA‑repair foci), direc t nucleoside analytics, and genetics. This integrative design avoids \nconversion-dependent biases and enables quantitative, state‑resolved measurements of \n5‑methyl‑2′‑deoxycytidine (5mdC) in genomic DNA [16–19,20]. \n \nWe first characterized the re‑entry trajectory from G0. DNA‑content cytometry defined the timing of the first \nS phase after refeeding, and revealed a modest G1/S delay in the cytosine/5‑mC deaminase mutant fcy1Δ \nrelative to wild type (wt) (Fig ure. 1A). Live‑cell imaging of Rad22‑YFP  - an established marker of \nrecombination‑repair assemblies - showed few or no foci in quiescence but a pronounced increase during this \nfirst S phase; importantly, fcy1Δ accumulated a higher fraction of nuclei with Rad22 foci than wt at matched \ntime points (Fig ure. 1B –C) [22 –26]. These observations place the forthcoming chemistry into a \nreplication‑linked context and motivate a search for methyl‑cytosine dynamics during G0→S. \n \n \n \n \nWe next asked whether genomic 5mdC can be detected and identified unambiguously. Enzymatic digestion \nof genomic DNA to nucleosides followed by HPLC/LC –MS revealed a peak co‑eluting with an authentic \n5mdC standard; full‑scan spectra displayed the diagnostic  (M+H)+ ion at m/z 242.1 and product‑ion spectra \nshowed the expected (2M+H) + feature at ~484.3 (Fig ure. 2A –C). External calibrations for dC and 5mdC \nestablished linearity and sensitivity (limit of detection, LOD = 0.0125 mM), supporting quantitative calls \nincluding “ND” (Fig ure. S1); this workflow follows ICH Q2(R2)  ‑ style validation and circumvents \nconversion chemistry artefacts [16–20]. \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 7 \n \n \n \nUsing this validated assay, we then quantified 5mdC across cell state and genotype. In wt, 5mdC was \nundetectable in vegetative growth but rose transiently during the first S phase after G0 exit (Fig ure. 3A–C). \nDeletion of the Dnmt2‑family methyltransferase Pmt1 abolished the signal at all time points, demonstrating \nthat the DNA‑embedded 5mC we detect is Pmt1‑dependent. Conversely, loss of Fcy1 markedly amplified the \nsame S‑phase‑linked pulse and elevate d residual 5mdC at steady state, consistent with deamination‑driven \nturnover limiting methyl‑cytosines as cells re‑enter the cycle (Fig ure. 3A–C) [3–7,21]. As a representative \nsnapshot, wt increased from ND (vegetative) to ~0.52% ± 0.03 at +2 h, whereas fcy1Δ rose from 1.45% ± 0.36 \nto 5.94% ± 1.36 at +2 h; pmt1Δ and fcy1Δ pmt1Δ remained ND (means ± SD; n ≥ 6 biological replicates). \nTo further secure chemical identity, the HPLC‑collected peak was analyzed by 1H‑NMR, which overlaid the \ncommercial 5mdC standard across hallmark resonances (Fig ure. 3D). Taken together with Fig ure. 1, these \ndata define a Pmt1 ↔Fcy1 axis that installs and drains a short‑lived 5mdC pulse precisely at G0→S in \nSchizosaccharomyces pombe [3–7,16–21]. \n \nTo position fcy1 within the quiescent mutagenesis network, we leveraged the sensitized BER/THO \nbackground ung1Δ thp1Δ, in which our recent work showed enhanced mutation accumulation during G0 [63]. \nAcross the G0 time course, ung1Δ thp1Δ displayed reduced viability relative to wt and a clear rise in FOAR \nfrequency (Figure.  S3A–B). The  forward-mutation spectra  were most informative: at  day 1,  ung1Δ \nthp1Δ was dominated by C→T (G→A) transitions, a hallmark of uracil-type lesions that are normally excised \nby Ung1; this transition bias persisted at  day 16 (Figure. S3C–D). Critically, adding fcy1Δ on top of  ung1Δ \nthp1Δ fcy1Δ attenuated FOAR accumulation and shifted the spectra away from transitions, with an ≈30% drop \nin C→T at day 16 relative to the double mutant  (Figure. S3B–D). Together with the  Pmt1-dependent 5mdC \npulse at G0→S (Figure.  3A–C) and the  Rad22-YFP phenotypes (Figure.  1), these data indicate that  Fcy1-\nmediated deamination (of C and/or 5mC) contributes materially to the mutagenic lesion burden in quiescence \n- especially when Ung1-dependent uracil excision and Thp1/THO-dependent RNA: DNA hybrid control are \ncompromised. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 8 \n \n \nTo determine whether the transient 5mdC detected after G0 exit is associated with newly synthesized strands, \nwe isolated  nascent DNA  by sucrose-gradient fractionation  and λ-exonuclease enrichment , following best \npractices from  nascent-strand (NS -seq) workflows (Gerbi and colleagues) with some adaptations to  S. \npombe (Figure. 4A–C) [64–66]. Fractions corresponding to the Okazaki-size window retained signal after λ-\nexo - consistent with protection by 5′ RNA primers - whereas larger/parental fragments were largely removed \n(Figure. 4C) [64,67]. HPLC on the λ-exo–enriched material revealed 5mdC above LOD in F1–F6, with a peak \nin mid -fractions (≈F3 –F5) and a  marked elevation in  fcy1Δ relative to wt (Fig ure. 4D). These data place \nthe Pmt1-dependent 5mdC pulse  on nascent DNA strands , consistent with  lagging-strand (Okazaki)  - \nrich material, and strengthen the view that  Fcy1 limits the methyl -cytosine pool immediately after DNA \nsynthesis [64–66,68]. We note λ-exo biases and controlled for them by fraction-matched pooling and side-by-\nside genotype controls (Methods) [64,65]. \n \n \n \nwt\nG0 40min 1h 1h20  1h40 2h 2h20 2h40 3h 4h\nfcy1∆\npmt1∆\nfcy1∆ pmt1∆\n0\n200\n400\n600\n0\n200\n400\n600\n800\n1,0K\nCount\n800\n1,0K\n0\n200\n400\n600\n0\n200\n400\n600\nCount\n0\nDNA content\n200\n400\n600\n0\n200\n400\n600\n800\n1,0K\nCount\nDNA content\n0\n200\n400\n600\n800\n1,0K\n0\n200\n400\n600\nCount\n0\n200\n400\n600\nDNA content\n0\n200\n400\n600\n800\n1,0K\nCount\nDNA content\n0\n200\n400\n600\n800\n1,0K\n0\n200\n400\n600\nCount\n800\n1,0K\n0\n200\n400\n600\n0\n200\n400\n600\nCount\nDNA content\n0\n200\n400\n600\n800\n1,0K\n0\n200\n400\n600\nCount\nQuiescence 1st S-phase\nG0 40 1h 1h40 2h 2h20 2h40 3h 4h\n0\n2\n4\n6\n8\nTime course\n% 5mdC/ (5mdC+dC)\nHPLC quantification of 5mdC  \nratio upon G0 exit\nwt\nfcy1∆\nFigure 3. 5mdC transiently rises at the first S phase upon G0 exit and is modulated by Pmt1 and Fcy1; NMR confirmation.\nA: Dot-blot of methylated DNA (anti-5mC) across the G0-exit time course (top) and DNA-content profiles (FACS) demarcating quiescence and the first S phase (bottom) \nfor wt, fcy1Δ, pmt1Δ and fcy1Δ pmt1Δ. \nB: HPLC/LC–MS chromatograms at +2 h after G0 exit for each genotype; co-injected 5mdC standard (bottom panels) confirms co-elution (red arrows). \nC: Quantification of %5mdC/(5mdC+dC) across time shows a transient pulse in wt that is abolished in pmt1Δ and strongly amplified in fcy1Δ. \nD: 1H-NMR spectrum of the HPLC-collected 5mdC fraction (fcy1Δ) overlays the commercial standard, validating chemical identity.\nA B\nC D\nDNA content\nDNA content\nDNA content\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 9 \n \n \n \nFinally, because Dnmt2/Pmt1 activity is known to respond to nucleotide and queuosine/queuine metabolism \nin RNA contexts [9 –15], we tested whether queuine modulates the replication‑linked phenotype. \nSupplementation reduced Rad22‑YFP foci in fcy1Δ during G0 exit (Fig ure. S2), suggesting that metabolic \ninputs can tune this methylation–deamination axis during cell‑cycle re‑entry. \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 10 \nDiscussion \nOur work overturns the long-standing view that the fission yeast Schizosaccharomyces pombe lacks genomic \nDNA 5-methylcytosine (5mC) [1–3]. Using orthogonal chemistry - HPLC/LC–MS with external calibration \nand ICH -compliant sensitivity, together with NMR overlays of collected fractions  - we unambiguously \nidentify 5-methyl-2′-deoxycytidine (5mdC) in genomic DNA digests [16 –20]. Genetic dissection places this \nmodification squarely on the Dnmt2-family enzyme Pmt1: 5mC is abolished in pmt1Δ and markedly elevated \nin fcy1Δ, which lacks the cytosine/5mC deaminase  Fcy1 [21]. Temporally, the signal peaks at the  first S \nphase as cells exit quiescence (G0→S), a nd enlarged methyl -cytosine pools in  fcy1Δ coincide with Rad22-\nYFP foci diagnostic of replication -coupled recombination [22–26]. Together, these data establish  Pmt1 as a \nbona fide DNA methyltransferase  in vivo  and reveal an actively regulated  Pmt1↔Fcy1 axis that restricts \nmethyl-cytosine during cell-cycle re-entry. \nReconciling prior “no 5mC” reports with a transient, regulated methylome \nEarlier surveys concluded that  S. pombe  is essentially unmethylated [1 –3], but those studies were not \noptimized to capture  low-stoichiometry, cell-state-restricted DNA methylation. Bisulfite library preparation \ncan under-represent difficult contexts and inflate apparent variation through conversion and coverage biases, \nespecially when true methylation is scarce [4]. Enzymatic approaches and direct chemistry have since raised \nthe bar for  trace-level detection [5 –7,16–20]. Three features of our dataset help e xplain the discrepancy: \n(i) temporal restriction - a short 5mC pulse at G0→S is diluted in asynchronous populations; (ii)  analytical \nrigor - a validated  LOD (0.0125 mM)  and run -matched standards make “ND” calls meaningful; and \n(iii) genetic anchoring - side-by-side pmt1Δ/fcy1Δ comparisons expose causal control. \n \nOur genetics separate installation and turnover: Pmt1 (Dnmt2) installs DNA 5mC in vivo, while Fcy1 drains \n5mC/C through deamination its loss amplifies and prolongs the G0→S 5mC pulse  and coincides with \nreplication-linked Rad22 foci (Fig ures. 3, 1) [3–7,21,22–26]. Epistasis in a sensitized  background sharpens \nthis model: in  ung1Δthp1Δ, which accumulates  uracil-type C→T transitions  in G0,  removing fcy1 (triple \nmutant fcy1Δung1Δthp1Δ) reduces FOA R accumulation and lowers C→T by  ~30% (Figure. S3), \nplacing Fcy1-mediated deamination upstream as a proximal source of mutagenic uracil/thymine lesions when \nuracil excision (Ung1) and R -loop control (Thp1/THO) are compromised [63]. In wt, the  Pmt1↔Fcy1 \nbalance keeps the 5mC/C deamination flux  minimal and transient - installing a short signal \nat G0→S then rapidly curtailing  it - thereby limiting  T: G/U: G  mismatches and the engagement of error -\nprone repair.  \nOrigin-proximal, co-replicative installation of 5mdC. \nLocating the 5mdC pulse on  λ-exo–enriched nascent DNA  argues that  Pmt1 acts co -replicatively, most \nconspicuously within the  Okazaki fragment size range  (Fig. 4) [64–66,68]. This fits the temporal peak \nat G0→S (Figure. 3) and the  Rad22-YFP phenotypes (Fig ure. 1), and suggests that the  Pmt1↔Fcy1 \naxis tunes methyl-cytosine exposure at or just behind replication forks . Given documented  λ-exo biases in \nnascent-strand assays, future bisulfite -independent, strand -resolved mapping integrated with  Okazaki-\nfragment sequencing will be informative, particularly to distinguish leading vs lagging strand deposition and \nto test for origin proximity [64,65,68] \nFunctional rationale for a minimal, timed 5mC program \nFission yeast appears to practice  epigenetic minimalism : DNA 5mC is deployed  briefly and state -\ndependently as a pulse at G0→S rather than as a stable bulk mark. Such low -stoichiometry, regulated 5mC \ncould fine-tune replication-fork transactions (e.g., origin usage or Okazaki-fragment maturation), temper the \nhandling of repetitive DNA, and interface with stress pathways during nutritional transitions. This functional \nsufficiency at low abundance aligns with the  heterogeneity of fungal methylomes  and with the persistence \nof sparse methylation in lineages where canonical DNMTs have been lost or repurposed [3,41]. The broader \nDnmt2/TRDMT1 literature also argues that  catalytic environment and cofactors  shape substrate  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 11 \nchoice: queuosine and nutrient signaling  modulate Pmt1/Dnmt2 on tRNA [9 –15], and  S-\nadenosylmethionine availability shifts across quiescence and re -entry [45]. We therefore propose that  cell \nstate, chromatin access, and cofactor supply bias Pmt1 toward DNA transiently at G0→S, with Fcy1 rapidly \ncurtailing the footprint - yielding a minimal, timed 5mC signal that is effective without being pervasive. \nBroader implications and DNMT2 dual-substrate capacity \nOur findings in fission yeast are consistent with a broader view in which DNMT2/TRDMT1 enzymes can \nshape both RNA and DNA cytosine methylation depending on cellular context. In the malaria parasite \nPlasmodium falciparum, the global DNA cytosine modificatio n landscape (5mC and oxidized derivatives) \nbecomes DNMT2 -dependent under controlled oxygen regimes, supporting a direct DNMT2 -linked DNA \nmethylation route in vivo [61]. Complementarily, DNMT2-mediated tRNA C38 methylation modulates stress \ntolerance and proteome allocation in Plasmodium, highlighting DNMT2’s context -dependent, dual-substrate \nbiology [62]. Together with the Pmt1↔Fcy1 axis uncovered here, these observations support a model in which \nDnmt2-family enzymes install a minimal, state-restricted DNA 5mC signal, while deamination pathways and \nmetabolic inputs such as queuine constrain its magnitude and timing at key physiological transitions. \nLimitations and testable predictions \nWe have not yet localized sites genome -wide. Bisulfite-free mapping (for example EM -seq) should resolve \nsequence context and reveal whether Pmt1 favors particular CpN motifs or origin -proximal regions [5–7]. If \nthe model is correct, catalytic-site mutants of Pmt1 (motif IV cysteine; motif I SAM binding) will abolish the \npulse; in vitro assays with purified Pmt1 should methylate dsDNA substrates under conditions that mimic \nG0→S. The turnover model predicts epistasis with base-excision and mismatch repair, and that elevating Pmt1 \nactivity outside G0→S will be toxic unless Fcy1 capacity or downstream processing increases. Finally, \nperturbing SAM metabolism or queuine supply should modulate pulse amplitude and timing, linking \nmetabolic state to Pmt1 substrate choice [9–15,45]. \nImplications beyond fission yeast \nThat a Dnmt2 -family enzyme can methylate DNA in vivo in a eukaryote long considered essentially \nunmethylated broadens the functional landscape of this enzyme class [9 –15]. Rather than a binary “RNA -\nonly” versus “DNA -only” assignment, our results support a conditional, dual-substrate view in which cell \nstate and counter -enzymes (here, Fcy1) determine when DNA methylation is deployed sparingly yet \npurposefully. Schizosaccharomyces pombe  offers a genetically clean chassis to dissect how minimal \nmethylomes are produced, sensed, and erased, and how this economy of marking influences genome \nmaintenance during environmental transitions. \nIn summary, Pmt1 catalyzes DNA 5mC in  S. pombe ; the mark peaks transiently at G0→S, is eliminated \nin pmt1Δ, and is amplified in  fcy1Δ, where it coincides with replication -coupled recombination foci. This \ncombined chemical, genetic and temporal logic revises the epigenetic landscape of fission yeast and motivates \nbisulfite-independent mapping and mechanistic tests to define where Pmt1 methylates and how Fcy1 restrains \nthis footprint. \nThese findings also align with emerging evidence outside fission yeast. In Plasmodium falciparum, the global \nDNA cytosine-modification landscape (5mC and oxidized derivatives) becomes DNMT2 -dependent under \ndefined oxygen regimes, consistent with a direct DNMT2 -linked DNA methylation route in vivo [61]. \nComplementarily, DNMT2 -mediated tRNA C38 methyla tion modulates stress tolerance and proteome \nallocation in Plasmodium [62]. Together with the Pmt1 ↔Fcy1 axis described here, these observations raise \nthe possibility that other Dnmt2 enzymes exhibit cryptic, state -restricted DNA methylation under specific \nphysiological conditions, with deployment gated by metabolic inputs and opposing deamination pathways. \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 12 \nData Origin, Figures, and Materials Availability \nThe data presented in this manuscript were generated as part of an independent research initiative that I \nconceived and developed during my postdoctoral training on 2016, in the laboratory of Dr. Benoît Arcangioli, \nwith support from the  Fondation de France (Prix Thérèse Lebrasseur). This work reports the  first detection \nof DNA 5 -methylcytosine (5mC ) in Schizosaccharomyces pombe  and identifies  Pmt1 as the enzyme \nresponsible for this modification. By depositing this work as a preprint on bioRxiv, I aim to ensure \nthe permanent accessibility of these findings to the scientific community and to establish an official, traceable \nrecord of authorship. \nAll data, figures, and interpretations presented in this preprint are the  intellectual property of the author and \ncollaborators. This work is released under the  Creative Commons Attribution (CC BY 4.0)  license selected \nduring submission. Any reproduction, reuse, or derivative work - including figures, datasets, or textual \nexcerpts - must properly cite this preprint using its DOI and full bibliographic reference. \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 13 \n \nAcknowledgments \n \nI would like to express my deepest gratitude to Valérie Huteau and Dr. Sylvie Pochet (Unité de Chimie et \nBiocatalyse, Institut Pasteur) for their exceptional technical expertise, insightful discussions, and invaluable \nsupport throughout this study. \nTheir contributions were instrumental in the successful completion of the LC/MS validation experiments, \nwhich form a critical foundation of this work. \nI also warmly thank Dr. Benoît Arcangioli (Unite Dynamique du Génome, Institut Pasteur) for his scientific \nmentorship and continuous guidance throughout the development of this project. \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 14 \nFunding \nThis work was supported by the Fondation de France through the Prix Thérèse Lebrasseur, awarded to the \nproject on neurodegenerative diseases initiated by Samia Miled in the laboratory of Dr. Benoît Arcangioli. \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 15 \nReferences \n1. Capuano F, et al. Cytosine DNA methylation is found in Drosophila melanogaster but absent \nin Saccharomyces cerevisiae, Schizosaccharomyces pombe, and other yeast species. Anal \nChem 86 (2014): 3697–3702. doi:10.1021/ac500447w. \n2. Zemach A, McDaniel IE, Silva P, Zilberman D. Genome-wide evolutionary analysis of eukaryotic \nDNA methylation. Science 328 (2010): 916–919. doi:10.1126/science.1186366. \n3. Bewick AJ, et al. Diversity of cytosine methylation across the fungal tree of life. Nat Ecol \nEvol 3 (2019): 479–490. doi:10.1038/s41559-019-0810-9. \n4. Olova N, et al. Comparison of WGBS library preparation strategies identifies sources of biases \naffecting DNA methylation data. Genome Biol 19 (2018): 33. doi:10.1186/s13059-018-1408-2. \n5. Vaišvila R, et al. Enzymatic methyl sequencing detects DNA methylation at single-base \nresolution from picograms of DNA. Genome Res 31 (2021): 1280–1289. \ndoi:10.1101/gr.266551.120. \n6. Han Y, et al. Comparison of EM-seq and PBAT methylome library methods for low-input \nDNA. Epigenetics17 (2022): 1195–1204. doi:10.1080/15592294.2021.1997406. \n7. Sun Z, et al. Performance comparisons of methylation and hydroxymethylation sequencing \nmethods.Epigenomics 15 (2023): 1007–1023. doi:10.2217/epi-2022-0453. \n8. De Mendoza A, Lister R. Evolution of DNA methylome diversity in eukaryotes. J Mol \nBiol 432 (2020): 1687–1705. doi:10.1016/j.jmb.2019.11.015. \n9. Becker M, et al. Pmt1, a Dnmt2 homolog in Schizosaccharomyces pombe, mediates tRNA \nmethylation in response to nutrient signaling. Nucleic Acids Res 40 (2012): 11648–11658. \ndoi:10.1093/nar/gks956. \n10. Müller M, et al. Dynamic modulation of Dnmt2-dependent tRNA methylation by the \nmicronutrient queuine. Nucleic Acids Res 43 (2015): 10952–10962. doi:10.1093/nar/gkv980. \n11. Johannsson S, et al. Structural insights into the stimulation of S. pombe Dnmt2 catalytic \nefficiency by the tRNA nucleoside queuosine. Sci Rep 8 (2018): 8880. doi:10.1038/s41598-018-\n27118-5. \n12. Jurkowski TP, et al. Human DNMT2 methylates tRNA(Asp) using a DNMT-like catalytic \nmechanism. RNA 14(2008): 1663–1670. doi:10.1261/rna.1037808. \n13. Goll MG, et al. Methylation of tRNAAsp by the DNA methyltransferase homolog \nDnmt2. Science 311 (2006): 395–398. doi:10.1126/science.1120976. \n14. Ehrenhofer-Murray AE. Cross-talk between Dnmt2-dependent tRNA methylation and queuosine \nmodification.Biomolecules 7 (2017): 14. doi:10.3390/biom7010014. \n15. Jeltsch A, et al. Mechanism and biological role of Dnmt2 in nucleic acid methylation. RNA \nBiol 14 (2017): 1108–1123. doi:10.1080/15476286.2016.1191737. \n16. Sandhu J, et al. Determination of 5-methyl-2′-deoxycytidine in genomic DNA using HPLC-UV. J \nChromatogr B877 (2009): 1957–1961. doi:10.1016/j.jchromb.2009.05.032. \n17. Song L, et al. Specific method for the determination of genomic DNA methylation by HPLC. J \nChromatogr B835 (2006): 110–113. doi:10.1016/j.jchromb.2006.02.014. \n18. Marchante-Gayón JM, et al. Quantification of modified nucleotides and nucleosides by isotope-\ndilution mass spectrometry. Mass Spectrom Rev 43 (2024): 998–1018. doi:10.1002/mas.21865. \n19. Ross RL, et al. Automated identification of modified nucleosides by HRAM-LC–MS/MS with \nneutral-loss detection. J Am Soc Mass Spectrom 34 (2023): 2785–2792. \ndoi:10.1021/jasms.3c00298. \n20. ICH. Q2(R2) Validation of Analytical Procedures. EMA/CHMP/ICH/82072/2006; legal effective \ndate 14 Jun 2024. \n21. PomBase. fcy1 (SPCC965.14c), cytosine deaminase Fcy1 . PomBase.org gene page (accessed \n2025). \n22. Meister P, et al. Nuclear factories for signaling and repairing DNA double-strand breaks in S. \npombe. Nucleic Acids Res 31 (2003): 5064–5073. doi:10.1093/nar/gkg718. \n23. Meister P, et al. Temporal separation of replication and recombination in fission yeast. J Cell \nBiol 168 (2005): 537–544. doi:10.1083/jcb.200409108. \n24. Lambert S, et al. Gross chromosomal rearrangements and elevated recombination at sites of \nreplication fork stalling in fission yeast. Cell 121 (2005): 689–702. doi:10.1016/j.cell.2005.03.030. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 16 \n25. Lambert S, et al. Homologous recombination restarts blocked replication forks at the expense of \ngenome rearrangements. Mol Cell 39 (2010): 346–359. doi:10.1016/j.molcel.2010.07.015. \n26. Nguyen MO, et al. Recombination occurs within minutes of fork blockage and drives error-\nprone restart.eLife 4 (2015): e04539. doi:10.7554/eLife.04539. \n27. Sajiki K, et al. Genetic regulation of mitotic competence in G0 quiescent S. pombe. Sci \nAdv 4 (2018): eaat5685. doi:10.1126/sciadv.aat5685. \n28. Labib K, Hodgson B. Replication fork barriers: pausing for a break or stalling for time? EMBO \nRep 8 (2007): 346–353. doi:10.1038/sj.embor.7400940. \n29. Durdevic Z, et al. Dnmt2 is required for efficient Dicer-2–dependent siRNA pathway \nactivity. Cell Rep 4(2013): 931–937. doi:10.1016/j.celrep.2013.07.046. \n30. Schaefer M, et al. RNA methylation by Dnmt2 protects tRNAs against stress-induced \ncleavage. Genes Dev 24(2010): 1590–1595. doi:10.1101/gad.586610. \n31. Li H, et al. New substrates and determinants for tRNA recognition of Dnmt2. RNA \nBiol 18 (2021): 2085–2098. doi:10.1080/15476286.2021.1930756. \n32. Johannsson S (thesis). Structural and biochemical investigation of S. pombe Dnmt2. Univ. \nGöttingen (2018). \n33. Zimmermann RA, et al. Diversity of fungal DNMTs and association with methylation \npatterns. ACS Pharmacol Transl Sci 5 (2022): 1079–1085. doi:10.1021/acsptsci.2c00175. \n34. Nai Y-S, et al. Diversity of fungal DNA methyltransferases and their association with DNA \nmethylation patterns. J Fungi 7 (2021): 363. doi:10.3390/jof7050363. \n35. Lin X, et al. DNA methylation and gene regulation: a dynamic duo. Signal Transduct Target \nTher 8 (2023): 340. doi:10.1038/s41392-023-01686-4. \n36. Chen X, et al. Mapping epigenetic modifications by sequencing technologies. Cell Death \nDiffer 32 (2025): 1715–1732. doi:10.1038/s41418-023-01213-1. \n37. Tanaka Y, et al. Quality assessment of EM-seq library construction and performance. Biochim \nBiophys Acta Gene Regul Mech 1867 (2024): 194038. doi:10.1016/j.bbagrm.2024.194038. \n38. Deinichenko KA, et al. Benchmark of Oxford Nanopore, EM-seq, and bisulfite methods for \nmethylomes. Front Epigenetics & Epigenomics 2 (2024): 1362926. \ndoi:10.3389/freae.2024.1362926. \n39. Wang Z, et al. Division of labour: NSun2/Trm4 tRNA methyltransferases Trm4a/Trm4b in \nfission yeast. J Fungi 9 (2023): 1030. doi:10.3390/jof9101030. \n40. Shanmugam R, et al. Dnmt2 homolog of Geobacter sulfurreducens specifically methylates tRNA-\nGlu. Nucleic Acids Res 42 (2014): 6487–6496. doi:10.1093/nar/gku256. \n41. Catania S, et al. Evolutionary persistence of DNA methylation for millions of \nyears. Cell 182 (2020): 1126–1140.e15. doi:10.1016/j.cell.2020.07.010. \n42. Lyons DB, et al. DDM1/Lsh remodelers allow methylation of nucleosomal DNA. eLife 6 (2017): \ne30674. doi:10.7554/eLife.30674. \n43. Jeon J, et al. Genome-wide profiling of DNA methylation in a filamentous fungus. Sci \nRep 5 (2015): 8567. doi:10.1038/srep08567. \n44. Sun S, et al. Cellular quiescence in budding yeast (concepts relevant to starvation-induced \nG0). J Mol Biol 433(2021): 167071. doi:10.1016/j.jmb.2020.11.022. \n45. Hayashi T, et al. S-adenosylmethionine synthetase is required for growth, maintenance of G0, \nand exit from quiescence in fission yeast. iScience 2 (2018): 1–13. doi:10.1016/j.isci.2018.08.007. \n46. Lab workflows for DNA nucleoside analysis (overview). (Use your lab’s SOP citation or \nSupplementary Methods to avoid vendor refs.) \n47. Tang Y, et al. Analytical workflows for cytosine derivatives (5mC, 5hmC, 5fC, 5caC). Anal \nChem 84 (2012): 451–458. doi:10.1021/ac202466w. \n48. Johnston AS, et al. Practical aspects of nucleoside chromatography for modified bases. J \nChromatogr B 879(2011): 2526–2534. doi:10.1016/j.jchromb.2011.07.001. \n49. Contreras-Sanz A, et al. Simultaneous quantification of nucleosides by HPLC. J Chromatogr \nB 879 (2011): 2880–2886. doi:10.1016/j.jchromb.2011.08.017. \n50. Frenkel K, et al. Complete enzymatic hydrolysis to nucleosides for DNA adduct quantification. J \nChromatogr B488 (1989): 39–49. doi:10.1016/S0378-4347(00)83006-6. \n51. Meister P, et al. Live analysis of replication foci in fission yeast. EMBO J 26 (2007): 5479–5488. \ndoi:10.1038/sj.emboj.7601923. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 17 \n52. Wang H, et al. Queuine links translational control to a bacterial micronutrient. Nucleic Acids \nRes 47 (2019): 3711–3727. doi:10.1093/nar/gkz096. \n53. Sarkies P. Encyclopaedia of eukaryotic DNA methylation. Biochem Soc Trans 50 (2022): 1915–\n1928. doi:10.1042/BST20220277. \n54. Huff JT, et al. Dnmt5 maintains CG methylation in the absence of Dnmt1. Cell 156 (2014): \n1286–1299. doi:10.1016/j.cell.2014.01.029. \n55. Nguyen D, et al. Targeted enriched enzymatic methyl-sequencing. bioRxiv (2022): \n2022.08.26.505457. doi:10.1101/2022.08.26.505457. \n56. Lin CJ, et al. Methods for nucleoside-level 5mC quantification in low-methyl \ngenomes. Methods 187 (2021): 43–56. doi:10.1016/j.ymeth.2020.10.015. \n57. Wang Z, Qiu X. tRNA methylation and genome stability. Trends Genet 37 (2021): 997–1010. \ndoi:10.1016/j.tig.2021.06.007. \n58. Li Y, et al. Biological function and druggability of DNMT2/TRDMT1. Semin Cancer \nBiol 97 (2024): 106–120. doi:10.1016/j.semcancer.2024.03.012. \n59. Chen Q, et al. tRNA-derived small RNAs in stress responses. Mol Cell 76 (2019): 522–534. \ndoi:10.1016/j.molcel.2019.09.017. \n60. Wang W, et al. Queuosine and tRNA modification in eukaryotes. RNA 34 (2028): 1–?? \n(placeholder—replace if you prefer a different review). \n61.  Hammam E, Miled S, Bonhomme F, Arcangioli B, Arimondo PB, Scherf A. Variable oxygen \nenvironments and DNMT2 determine the DNA cytosine epigenetic landscape of Plasmodium \nfalciparum. bioRxiv (2021). doi:10.1101/2021.12.09.471905. \n62. Hammam E, Sinha A, Baumgarten S, Nardella F, Liang JQ, Miled S, Bonhomme F, Erdmann D, \nArcangioli B, Arimondo PB, Scherf A, et al. Malaria Parasite Stress Tolerance Is Regulated by \nDNMT2-Mediated tRNA Cytosine Methylation. mBio 12(6) (2021): e02558-21. \ndoi:10.1128/mBio.02558-21. \n63. Miled S. Genome Stability under Silence: DNA Repair Networks in Quiescent \nFissionYeast. bioRxiv 2025; doi:https://doi.org/10.1101/2025.08.24.671962 \n64. Foulk MS, Urban JM, Casella C, Gerbi SA. Characterizing and controlling intrinsic biases of lambda \nexonuclease in nascent strand sequencing reveals phasing between nucleosomes and G-quadruplex \nmotifs around a subset of human replication origins. Genome Res 25 (2015): 725–735. \ndoi:10.1101/gr.183848.114.   \n65. Besnard E, Desprat R, Ryan M, Kahli M, Aladjem MI, Lemaitre J-M. Best practices for mapping \nreplication origins in eukaryotic chromosomes. Curr Protoc Cell Biol 64 (2014): 22.18.1–13. \ndoi:10.1002/0471143030.cb2218s64.   \n66. Fu H, Besnard E, Desprat R, Ryan M, Kahli M, Lemaitre J-M, Aladjem MI. Mapping replication \norigin sequences in eukaryotic chromosomes. Curr Protoc Cell Biol 65 (2014): 22.20.1–17. \ndoi:10.1002/0471143030.cb2220s65.  \n67. Cimbora DM, Schübeler D, Reik A, Hamilton J, Francastel C, Epner EM, Groudine M. Long-\ndistance control of origin choice and replication timing in the human β-globin locus are independent \nof the locus control region. Mol Cell Biol 20 (2000): 5581–5591. doi:10.1128/MCB.20.15.5581-\n5591.2000.   \n68. Yang W, Li X. Next-generation sequencing of Okazaki fragments extracted from Saccharomyces \ncerevisiae. FEBS Lett 587 (2013): 2441–2447. doi:10.1016/j.febslet.2013.06.014. \n \n \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 18 \nFigure Legends \n \nFigure 1. fcy1Δ induces a G1/S delay upon G0 exit and accumulates spontaneous Rad22-YFP repair \nfoci. \n(A) DNA-content profiles (FACS) at G0, +2 h, +2 h20, +3 h (wt-rad22-YFP; fcy1Δ-rad22-YFP). \n(B) Representative micrographs in quiescence and during the first S phase (YFP, Hoechst, merge, \nNomarski). \n(C) Quantification of nuclei with Rad22-YFP foci across time after G0 exit; S-phase windows indicated. \nn>200 nuclei/strain; 3 biological replicates; two-way ANOVA, p<0.001. \n \nFigure 2. Global genomic 5-methyl-2′-deoxycytidine (5mdC) is detectable and requires Pmt1 \n(HPLC/LC–MS). \n(A) HPLC chromatograms of enzymatic DNA digests from wt, fcy1Δ, pmt1Δ, fcy1Δ pmt1Δ (top: native; \nbottom: + co-injected 5mdC). A co-eluting peak appears in wt and fcy1Δ, but not in pmt1Δ backgrounds. \n(B) Full-scan LC–MS shows the protonated adduct [M+H]+ at m/z 242.1. \n(C) Product-ion spectrum confirms 5mdC with a (2M+H)+ feature at ~m/z 484.3. \n \nFigure 3. 5mdC rises during the first S phase after G0 exit in fcy1Δ in a Pmt1-dependent manner. \n(A) Dot-blot of genomic DNA probed for 5mC over a G0-exit time course. \n(B) HPLC traces across time (wt vs fcy1Δ; + co-injected 5mdC). \n(C) HPLC quantification of %5mdC/(5mdC+dC) upon G0 exit (mean±SD). \n(D) 1H-NMR of the purified 5mdC fraction, matching an authentic standard. \n \nSupplementary Figure Legends \n \nFigure S1. Analytical validation of 5mdC/dC measurements. \n(A–B) External calibration curves for 5mdC and dC (linear fits, R2>0.99; LOD ~0.0125 mM for 5mdC). \n(C) HPLC chromatograms of standard nucleosides. \n \nFigure S2. Pmt1-dependent DNA damage upon G0 exit in fcy1Δ and suppression by queuine. \n(A) % nuclei with Rad22-YFP foci over time in wt, fcy1Δ, pmt1Δ, fcy1Δ pmt1Δ. \n(B) Queuine (Q) reduces Rad22-YFP foci in fcy1Δ; n>250 nuclei/strain; 2 experiments; two-way ANOVA, \np<0.005. \n \nSupplementary Figure S3. Quiescent mutagenesis in ung1Δ thp1Δ is dominated by uracil-type \ntransitions and partially suppressed by fcy1Δ. \n(A) Viability curves during quiescence (G0) for wt, fcy1Δ, ung1Δ thp1Δ, and fcy1Δ ung1Δ thp1Δ strains. \nThe triple mutant exhibits reduced survival compared to wt, while loss of fcy1 partially rescues viability \ndefects in the double mutant background. \n(B) Accumulation of FOA-resistant (FOAR) colonies over time in G0. The ung1Δ thp1Δ double mutant \ndisplays a strong increase in forward mutation frequency, which is attenuated in the fcy1Δ ung1Δ \nthp1Δ background, indicating that uracil removal by Ung1 contributes substantially to mutation burden. \n(C–D) Forward-mutation spectra at day 1 (C) and day 16 (D) in G0. Mutation profiles reveal a \npredominance of C→T / G→A transitions in ung1Δ thp1Δ, consistent with unrepaired uracil lesions \ngenerated by cytosine deamination or misincorporation. The fcy1Δ deletion partially suppresses these \ntransitions, confirming its role in limiting methyl-cytosine–derived uracil accumulation. \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 19 \n \nFigure 4. Nascent DNA (Okazaki fragments) carries the G0→S 5mdC pulse: sucrose-gradient \nfractionation, λ-exonuclease enrichment and HPLC quantification. \n \n(A) Cell-cycle staging across the G0-exit time course by DNA-content cytometry (wt vs fcy1Δ). (B) Size-\nfractionation of nascent DNA on 10–40% sucrose gradients at the indicated times (G0 → +2 h 15); 12 \nfractions (F1–F12) were collected per gradient (50 µL/fraction) using the same piston system employed for \nribosome gradients. (C) λ-exonuclease test of pooled fractions: for each fraction index (F1…F12), equal \nvolumes from all-time points were combined (e.g., F1_G0 + F1_1h40 + …) to increase material, then \nanalyzed before and after λ-exo. Loss of the bulk signal after λ-exo indicates depletion of broken/parental \nDNA carrying 5′-phosphate ends, whereas RNA-primed nascent strands remain enriched [64–\n67]. (D) HPLC quantification of %5mdC in F1–F6 after λ-exo (wt vs fcy1Δ). 5mdC is detectable and \nenriched in mid-fractions (Okazaki-size window) and is substantially higher in fcy1Δ than wt, consistent \nwith a Pmt1-installed, Fcy1-restrained methylation pulse during S phase. Bars show means; exact n, enzyme \nunits and gradient parameters are in Methods. We note the known sequence/structure biases of λ-\nexo selections and mitigated them as in [64,65]. \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 20 \n \nSupplementary Figures \n \nFigure S1 \n \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 21 \nFigure S2 \n \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n 22 \nFigure S3 \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}