Pmt1-dependent and state-restricted DNA methylation in fission yeast

DNA 5-methylcytosine (5mC) is widely considered absent from budding and fission yeasts. The Dnmt2 family enzyme Pmt1 in Schizosaccharomyces pombe is annotated as a tRNA C38 methyltransferase, but whether it methylates DNA in vivo has remained unresolved. Using orthogonal chemistry (HPLC/LC-MS and 1H-NMR) with external standards (limit of detection for 5mdC = 0.0125 mM), we detect authentic 5 -methyl-2'- deoxycytidine (5mdC) in enzymatically digested genomic DNA from S. pombe. Quantification across isogenic strains shows that 5mdC is undetectable in vegetative wild type but rises to 0.518% (plus or minus 0.025) of cytosines two hours after G0 exit. Loss of the cytosine/5 -mC deaminase Fcy1 causes marked accumulation: 1.445% (plus or minus 0.361) in vegetative cells and 5.943% (plus or minus 1.364) at 2 hours, approximately 11.5-fold over wild type. By contrast, pmt1Δ and fcy1Δ pmt1Δ remain below detection in all conditions, establishing Pmt1 as the DNA -directed methyltransferase in vivo. Nascent -strand fractionation and lambda exonuclease enrichment place this transient 5mdC pulse on Okazaki -enriched DNA, consistent with co - replicative installation. Functionally, the first S phase after quiescence shows increased Rad22 -YFP foci in fcy1Δ, which are dampened by queuine ; in a sensitized background (ung1Δ thp1Δ), deleting fcy1 reduces C- >T transitions by about 30 percent. Together, chemical, genetic, and temporal evidence reveals a Pmt1 - dependent, state -restricted DNA methylation program in fission yeast that is rapidly curtailed by Fcy1, redefining the epigenetic landscape of S. pombe and providing a minimal, tractable system to dissect regulated DNA methylation in eukaryotes.

Schizosaccharomyces pombe; DNA 5-methylcytosine (5mC); Pmt1 (Dnmt2/TRDMT1); Fcy1 deaminase; state-restricted methylation; G0-to-S transition; replication-coupled methylation; Okazaki fragments (nascent DNA); methylation-deamination flux; Rad22-YFP foci; queuine metabolism. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 2

Cytosine 5-methylation (5mC) is a pervasive epigenetic mark across eukaryotes, but its distribution is highly uneven across lineages. Early comparative surveys established a canonical view in which Schizosaccharomyces pombe lacks detectable genomic 5mC, in stark contrast to many plants and animals [1–3]. The strength of that consensus has rested largely on bisulfite -based assays and limited mass - spectrometric analyses conducted under conditions that were not designed to capture transient or low - stoichiometry DNA methylation [1,4]. Yet it is now clear that conventional whole -genome bisulfite sequencing (WGBS) can introduce sequence -specific coverage and conversion biases that complicate interpretation at low abundance [4]. Enzymatic methyl -seq (EM-seq) and orthogonal chemistries hav e since improved sensitivity and fidelity in difficult settings, particularly at low input and in GC-rich or damage-prone regions [5–7]. In parallel, targeted HPLC/LC–MS pipelines coupled to authentic standards and NMR overlays provide definitive chemical identity for modified deoxynucleosides, enabling rigorous calls near the analytical limit of detection [16–19]. The S. pombe genome encodes Pmt1, a homolog of the Dnmt2/TRDMT1 family [9]. Dnmt2 proteins were long annotated as DNA methyltransferases by sequence homology, but seminal work showed that they methylate cytosine-38 in the anticodon loop of specific tRNAs, notably tRNA Asp, and protect tRNAs from stress-induced cleavage [10–13]. In S. pombe, Pmt1 methylates tRNAAsp in vivo, with activity modulated by nutrient status via Sck2 signaling and by queuosine in the anticodon loop, which can stimulate Dnmt2 catalysis and alter tRNA decoding properties [9 –11,14,15]. These observations raise a fundamental question: can a Dnmt2-family enzyme in a yeast widely considered “DNA-unmethylated” also catalyze DNA 5mC formation under specific physiological contexts? Here we uncover a Pmt1-dependent DNA 5mC signal in S. pombe. Using validated nucleoside-level analytics (HPLC/LC–MS with external calibration and ICH Q2(R2)–compliant LOD/LOQ) and NMR confirmation of collected 5mdC fractions [16 –20], we detect 5mdC that (i) rises acutely as cells re -enter the cell cycle from nitrogen-starvation quiescence (G0→S), (ii) is abolished in pmt1Δ, and (iii) accumulates in fcy1Δ backgrounds lacking the cytosine deaminase Fcy1 , consistent with deamination -driven turnover of methyl-cytosine [21]. Live-cell Rad22-YFP imaging and DNA-content cytometry place this methylation pulse in the first S phase after quiescence, a window characterized by elevated recombination/repair foci and replication stress in S. pombe [22–26]. Together, these data identify Pmt1 as a bona fide DNA methyltransferase in fission yeast and suggest that 5mC formation is cell-state dependent, transient, and actively countered by base deamination. Beyond resolving a long -standing controversy f or S. pombe [1–3], our work motivates a reevaluation of Dnmt2 catalytic plasticity and the contexts in which low -stoichiometry DNA methylation can be biologically meaningful. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 3

Strains, media, and general conditions Isogenic Schizosaccharomyces pombe strains were wild type, pmt1Δ, fcy1Δ and fcy1Δpmt1Δ. Cells were grown at 30 °C in YES or EMM with agitation (200 rpm) unless stated otherwise. The Rad22-YFP marker was introduced by standard genetics and expressed from its native locus. Quiescence (G0) induction and re-entry time course Quiescence was induced by nitrogen starvation in EMM -N following established S. pombe workflows. Re- entry (“t = 0”) was triggered by transfer to nitrogen -replete medium. Samples were collected across the first cell cycle, with the first S-phase window annotated by DNA-content cytometry and septation index as in prior quiescence/exit frameworks [27, 44, 45]. Viability was monitored by colony-forming units on YES plates. Genomic DNA isolation and RNA removal Cell pellets were lysed mechanically/enzymatically. Genomic DNA was purified using phenol-free buffers or silica-column kits, followed by sequential RNA depletion (RNase A/T1 → RNase H) and a short alkaline hydrolysis to eliminate residual RNA. DNA quality was assessed by A260/280 and A260/230 ratios, absence of rRNA bands on agarose gels, and dsDNA-specific fluorimetry. This workflow minimizes RNA carry-over in nucleoside assays [56]. Enzymatic digestion to nucleosides Defined masses of genomic DNA were digested to nucleosides using a standard sequential cocktail: DNase I (Mg²⁺, pH ≈ 7.0) → nuclease P1 (Zn²⁺, pH ≈ 5.2) → alkaline phosphatase (pH ≈ 7.5). Reactions proceeded to completion (verified by plateau of total nucl eoside yield) and were filtered prior to chromatography. The DNase-I → NP1 → AP pipeline is widely adopted for quantitative nucleoside analysis and limits partial - hydrolysis bias [50, 49, 47]. HPLC/LC–MS identification and quantification of dC and 5mdC Digests and authentic standards (2′-deoxycytidine, 5 -methyl-2′-deoxycytidine) were injected in the same analytical sessions. Separation used a reversed -phase column (C18, 3 –5 µm) with aqueous/organic eluents compatible with UV (~254 nm) and positive -mode electrospray. Identity of 5mdC in genomic digests was assigned by co-elution with the standard and MS signatures: [M+H] ⁺ = 242.1 in full scan and a (2M+H)⁺ ≈ 484.3 feature in product -ion spectra. External calibration curves converted areas to amounts; repr esentative chromatograms and linear fits (R²) are provided in Supplementary Figure. S1 [16, 17, 18, 19, 47–49]. Analytical validation, LOD/LOQ, and calling rules

validation followed ICH Q2(R2) (effective 2024). For 5mdC, the limit of detection (LOD) - calculated from calibration slope and baseline noise on blanks/low standards - was 0.0125 mM in our configuration; the limit of quantification (LOQ) was derived using ICH equations. “ND” (not detected) is defined as < LOD. Unless stated, values are mean ± SD from n ≥ 6 biological replicates. Inter- and intra-assay precision were assessed on pooled digests analyzed across days [20]. NMR confirmation of 5mdC identity HPLC fractions corresponding to putative 5mdC were collected and concentrated. ¹H (and, where available, ¹³C) NMR spectra were acquired in D₂O and overlaid with a commercial 5mdC standard. Chemical- shift concordance across hallmark resonances (C5/H5, C6/H6, deoxyribose protons) and the absence of extraneous peaks supported molecular identity [16, 19, 56]. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 4 Time-course and two-state snapshot designs For the time course, aliquots were collected at fixed intervals after refeeding; for the two-state snapshot, samples were taken in vegetative growth and at +2 h after G0 exit. At each point, % 5mdC/ (5mdC + 5dC) was measured by HPLC/LC–MS as above. Wild type, fcy1Δ, pmt1Δ, and fcy1Δ pmt1Δ were assayed side-by-side within the same analytical sessions. ND calls strictly followed the LOD rule [27, 20]. DNA-content flow cytometry Cells were ethanol-fixed, treated with RNase A, stained with propidium iodide, and analyzed on a benchtop flow cytometer. G1/S/G2 distributions were used to delineate the first S -phase after G0 exit and to synchronize figure annotations with biochemical sampling [27, 44]. Live-cell imaging of Rad22-YFP and focus quantification Cells expressing Rad22-YFP were imaged on EMM pads with constant exposure/laser settings. Foci were scored blinded to genotype as nuclei containing ≥ 1 Rad22-YFP focus; ≥ 200 cells per condition per replicate were counted when feasible. Rad22 -positive foci are established recombination/repair assemblies that increase under replication stress in S. pombe, providing a live-cell readout during G0→S [22, 23, 51]. Queuine treatment during G0 exit and Rad22-YFP foci analysis Strains and culture conditions. All strains carry a C -terminal Rad22-YFP fusion at the endogenous rad22 locus. The four genotypes analyzed were wild type (wt), pmt1Δ, fcy1Δ, and fcy1Δ pmt1Δ. Cells were grown at 30 °C in EMM2 minimal medium with supplements to mid -log phase, then shifted to nitrogen-free EMM (EMM -N) to induce quiescence (G0) as described in the Methods (typically 24 h starvation unless stated). Nascent DNA preparation and sucrose-gradient fractionation (Okazaki-enrichment) Cells were sampled at the indicated times after G0 release (Fig. 4A). Low -molecular-weight DNA was prepared and resolved on 10–40% sucrose gradients in nuclease-free buffer. Gradients were run on a piston - driven fractionator; 12 fractions (F1 –F12) were collected at 50 µL each from top to bottom. Fractions were analysed on agarose gels to assess size distributions across the gradient (Fig. 4B). Where required, equal volumes for a given fraction index were pooled across time points (e.g., F1 = F1_G0 + F1_1h40 + …) to obtain sufficient material for enzymology and analytics. The workflow follows short-nascent-strand abundance/enrichment best practices [65,66]. λ-Exonuclease enrichment of nascent strands Pooled or individual fractions were incubated with λ-exonuclease under supplier-recommended conditions to digest 5′-phosphorylated dsDNA while sparing RNA-primed nascent strands that are refractory to 5′→3′ digestion [64,67]. Reactions were terminated, DNA recovered, and an aliquot run on agarose to confirm depletion of λ-exo–sensitive species. We adopted published precautions to limit known λ-exo sequence/structure biases (matched inputs, fraction-based pooling, no-enzyme controls) [64,65]. HPLC quantification of 5mdC in nascent DNA λ-exo–treated DNA from F1–F6 was enzymatically digested to nucleosides and analy zed by HPLC as in the main text (calibration/LOD in Fig. S1). %5mdC was computed as 5mdC/(5mdC+dC) × 100. Fractions from wt and fcy1Δ were processed side-by-side; biological replicates as indicated in Fig. 4D. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 5 Release from G0 and queuine treatment. G0 cultures were harvested, washed once in pre-warmed EMM2, and released into fresh EMM2 (time 0). Immediately upon release, cells were split into ±queuine conditions and incubated for 1 h at 30 °C with shaking. Queuine (+Q): queuine base ( hemi-sulfate), stock 10 mM in sterile water, filter-sterilized, added to a final concentration of 10 µM. Unless indicated, 10 µM was used for all experiments. After 1 h, samples were processed for imaging (below). In parallel, the same ±Q protocol was applied to each genotype (wt, pmt1Δ, fcy1Δ, fcy1Δ pmt1Δ). Statistics. Effects of genotype and treatment (±Q) were tested by two-way ANOVA (Prism software) . Significance thresholds were pre-specified (e.g., p < 0.05, 0.01, 0.005 as annotated in figure legends). Graphs display mean ± SD; exact n and p values are provided in the legends (see Fig. S2). .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 6

To determine whether fission yeast carries a minimal yet functional DNA methylation program, we analyzed quiescent cells as they re‑entered the cell cycle and combined three layers of evidence: physiology (cell‑cycle progression and DNA‑repair foci), direc t nucleoside analytics, and genetics. This integrative design avoids conversion-dependent biases and enables quantitative, state‑resolved measurements of 5‑methyl‑2′‑deoxycytidine (5mdC) in genomic DNA [16–19,20]. We first characterized the re‑entry trajectory from G0. DNA‑content cytometry defined the timing of the first S phase after refeeding, and revealed a modest G1/S delay in the cytosine/5‑mC deaminase mutant fcy1Δ relative to wild type (wt) (Fig ure. 1A). Live‑cell imaging of Rad22‑YFP - an established marker of recombination‑repair assemblies - showed few or no foci in quiescence but a pronounced increase during this first S phase; importantly, fcy1Δ accumulated a higher fraction of nuclei with Rad22 foci than wt at matched time points (Fig ure. 1B –C) [22 –26]. These observations place the forthcoming chemistry into a replication‑linked context and motivate a search for methyl‑cytosine dynamics during G0→S. We next asked whether genomic 5mdC can be detected and identified unambiguously. Enzymatic digestion of genomic DNA to nucleosides followed by HPLC/LC –MS revealed a peak co‑eluting with an authentic 5mdC standard; full‑scan spectra displayed the diagnostic (M+H)+ ion at m/z 242.1 and product‑ion spectra showed the expected (2M+H) + feature at ~484.3 (Fig ure. 2A –C). External calibrations for dC and 5mdC established linearity and sensitivity (limit of detection, LOD = 0.0125 mM), supporting quantitative calls including “ND” (Fig ure. S1); this workflow follows ICH Q2(R2) ‑ style validation and circumvents conversion chemistry artefacts [16–20]. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 7 Using this validated assay, we then quantified 5mdC across cell state and genotype. In wt, 5mdC was undetectable in vegetative growth but rose transiently during the first S phase after G0 exit (Fig ure. 3A–C). Deletion of the Dnmt2‑family methyltransferase Pmt1 abolished the signal at all time points, demonstrating that the DNA‑embedded 5mC we detect is Pmt1‑dependent. Conversely, loss of Fcy1 markedly amplified the same S‑phase‑linked pulse and elevate d residual 5mdC at steady state, consistent with deamination‑driven turnover limiting methyl‑cytosines as cells re‑enter the cycle (Fig ure. 3A–C) [3–7,21]. As a representative snapshot, wt increased from ND (vegetative) to ~0.52% ± 0.03 at +2 h, whereas fcy1Δ rose from 1.45% ± 0.36 to 5.94% ± 1.36 at +2 h; pmt1Δ and fcy1Δ pmt1Δ remained ND (means ± SD; n ≥ 6 biological replicates). To further secure chemical identity, the HPLC‑collected peak was analyzed by 1H‑NMR, which overlaid the commercial 5mdC standard across hallmark resonances (Fig ure. 3D). Taken together with Fig ure. 1, these data define a Pmt1 ↔Fcy1 axis that installs and drains a short‑lived 5mdC pulse precisely at G0→S in Schizosaccharomyces pombe [3–7,16–21]. To position fcy1 within the quiescent mutagenesis network, we leveraged the sensitized BER/THO

ung1Δ thp1Δ, in which our recent work showed enhanced mutation accumulation during G0 [63]. Across the G0 time course, ung1Δ thp1Δ displayed reduced viability relative to wt and a clear rise in FOAR frequency (Figure. S3A–B). The forward-mutation spectra were most informative: at day 1, ung1Δ thp1Δ was dominated by C→T (G→A) transitions, a hallmark of uracil-type lesions that are normally excised by Ung1; this transition bias persisted at day 16 (Figure. S3C–D). Critically, adding fcy1Δ on top of ung1Δ thp1Δ fcy1Δ attenuated FOAR accumulation and shifted the spectra away from transitions, with an ≈30% drop in C→T at day 16 relative to the double mutant (Figure. S3B–D). Together with the Pmt1-dependent 5mdC pulse at G0→S (Figure. 3A–C) and the Rad22-YFP phenotypes (Figure. 1), these data indicate that Fcy1- mediated deamination (of C and/or 5mC) contributes materially to the mutagenic lesion burden in quiescence - especially when Ung1-dependent uracil excision and Thp1/THO-dependent RNA: DNA hybrid control are compromised. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 8 To determine whether the transient 5mdC detected after G0 exit is associated with newly synthesized strands, we isolated nascent DNA by sucrose-gradient fractionation and λ-exonuclease enrichment , following best practices from nascent-strand (NS -seq) workflows (Gerbi and colleagues) with some adaptations to S. pombe (Figure. 4A–C) [64–66]. Fractions corresponding to the Okazaki-size window retained signal after λ- exo - consistent with protection by 5′ RNA primers - whereas larger/parental fragments were largely removed (Figure. 4C) [64,67]. HPLC on the λ-exo–enriched material revealed 5mdC above LOD in F1–F6, with a peak in mid -fractions (≈F3 –F5) and a marked elevation in fcy1Δ relative to wt (Fig ure. 4D). These data place the Pmt1-dependent 5mdC pulse on nascent DNA strands , consistent with lagging-strand (Okazaki) - rich material, and strengthen the view that Fcy1 limits the methyl -cytosine pool immediately after DNA synthesis [64–66,68]. We note λ-exo biases and controlled for them by fraction-matched pooling and side-by- side genotype controls (Methods) [64,65]. wt G0 40min 1h 1h20 1h40 2h 2h20 2h40 3h 4h fcy1∆ pmt1∆ fcy1∆ pmt1∆ 0 200 400 600 0 200 400 600 800 1,0K Count 800 1,0K 0 200 400 600 0 200 400 600 Count 0 DNA content 200 400 600 0 200 400 600 800 1,0K Count DNA content 0 200 400 600 800 1,0K 0 200 400 600 Count 0 200 400 600 DNA content 0 200 400 600 800 1,0K Count DNA content 0 200 400 600 800 1,0K 0 200 400 600 Count 800 1,0K 0 200 400 600 0 200 400 600 Count DNA content 0 200 400 600 800 1,0K 0 200 400 600 Count Quiescence 1st S-phase G0 40 1h 1h40 2h 2h20 2h40 3h 4h 0 2 4 6 8 Time course % 5mdC/ (5mdC+dC) HPLC quantification of 5mdC ratio upon G0 exit wt fcy1∆ Figure 3. 5mdC transiently rises at the first S phase upon G0 exit and is modulated by Pmt1 and Fcy1; NMR confirmation. A: 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) for wt, fcy1Δ, pmt1Δ and fcy1Δ pmt1Δ. B: HPLC/LC–MS chromatograms at +2 h after G0 exit for each genotype; co-injected 5mdC standard (bottom panels) confirms co-elution (red arrows). C: Quantification of %5mdC/(5mdC+dC) across time shows a transient pulse in wt that is abolished in pmt1Δ and strongly amplified in fcy1Δ. D: 1H-NMR spectrum of the HPLC-collected 5mdC fraction (fcy1Δ) overlays the commercial standard, validating chemical identity. A B C D DNA content DNA content DNA content .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 9 Finally, because Dnmt2/Pmt1 activity is known to respond to nucleotide and queuosine/queuine metabolism in RNA contexts [9 –15], we tested whether queuine modulates the replication‑linked phenotype. Supplementation reduced Rad22‑YFP foci in fcy1Δ during G0 exit (Fig ure. S2), suggesting that metabolic inputs can tune this methylation–deamination axis during cell‑cycle re‑entry. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 10

Our work overturns the long-standing view that the fission yeast Schizosaccharomyces pombe lacks genomic DNA 5-methylcytosine (5mC) [1–3]. Using orthogonal chemistry - HPLC/LC–MS with external calibration and ICH -compliant sensitivity, together with NMR overlays of collected fractions - we unambiguously identify 5-methyl-2′-deoxycytidine (5mdC) in genomic DNA digests [16 –20]. Genetic dissection places this modification squarely on the Dnmt2-family enzyme Pmt1: 5mC is abolished in pmt1Δ and markedly elevated in fcy1Δ, which lacks the cytosine/5mC deaminase Fcy1 [21]. Temporally, the signal peaks at the first S phase as cells exit quiescence (G0→S), a nd enlarged methyl -cytosine pools in fcy1Δ coincide with Rad22- YFP foci diagnostic of replication -coupled recombination [22–26]. Together, these data establish Pmt1 as a bona fide DNA methyltransferase in vivo and reveal an actively regulated Pmt1↔Fcy1 axis that restricts methyl-cytosine during cell-cycle re-entry. Reconciling prior “no 5mC” reports with a transient, regulated methylome Earlier surveys concluded that S. pombe is essentially unmethylated [1 –3], but those studies were not optimized to capture low-stoichiometry, cell-state-restricted DNA methylation. Bisulfite library preparation can under-represent difficult contexts and inflate apparent variation through conversion and coverage biases, especially when true methylation is scarce [4]. Enzymatic approaches and direct chemistry have since raised the bar for trace-level detection [5 –7,16–20]. Three features of our dataset help e xplain the discrepancy: (i) temporal restriction - a short 5mC pulse at G0→S is diluted in asynchronous populations; (ii) analytical rigor - a validated LOD (0.0125 mM) and run -matched standards make “ND” calls meaningful; and (iii) genetic anchoring - side-by-side pmt1Δ/fcy1Δ comparisons expose causal control. Our genetics separate installation and turnover: Pmt1 (Dnmt2) installs DNA 5mC in vivo, while Fcy1 drains 5mC/C through deamination its loss amplifies and prolongs the G0→S 5mC pulse and coincides with replication-linked Rad22 foci (Fig ures. 3, 1) [3–7,21,22–26]. Epistasis in a sensitized background sharpens this model: in ung1Δthp1Δ, which accumulates uracil-type C→T transitions in G0, removing fcy1 (triple mutant fcy1Δung1Δthp1Δ) reduces FOA R accumulation and lowers C→T by ~30% (Figure. S3), placing Fcy1-mediated deamination upstream as a proximal source of mutagenic uracil/thymine lesions when uracil excision (Ung1) and R -loop control (Thp1/THO) are compromised [63]. In wt, the Pmt1↔Fcy1 balance keeps the 5mC/C deamination flux minimal and transient - installing a short signal at G0→S then rapidly curtailing it - thereby limiting T: G/U: G mismatches and the engagement of error - prone repair. Origin-proximal, co-replicative installation of 5mdC. Locating the 5mdC pulse on λ-exo–enriched nascent DNA argues that Pmt1 acts co -replicatively, most conspicuously within the Okazaki fragment size range (Fig. 4) [64–66,68]. This fits the temporal peak at G0→S (Figure. 3) and the Rad22-YFP phenotypes (Fig ure. 1), and suggests that the Pmt1↔Fcy1 axis tunes methyl-cytosine exposure at or just behind replication forks . Given documented λ-exo biases in nascent-strand assays, future bisulfite -independent, strand -resolved mapping integrated with Okazaki- fragment sequencing will be informative, particularly to distinguish leading vs lagging strand deposition and to test for origin proximity [64,65,68] Functional rationale for a minimal, timed 5mC program Fission yeast appears to practice epigenetic minimalism : DNA 5mC is deployed briefly and state - dependently as a pulse at G0→S rather than as a stable bulk mark. Such low -stoichiometry, regulated 5mC could fine-tune replication-fork transactions (e.g., origin usage or Okazaki-fragment maturation), temper the handling of repetitive DNA, and interface with stress pathways during nutritional transitions. This functional sufficiency at low abundance aligns with the heterogeneity of fungal methylomes and with the persistence of sparse methylation in lineages where canonical DNMTs have been lost or repurposed [3,41]. The broader Dnmt2/TRDMT1 literature also argues that catalytic environment and cofactors shape substrate .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 11 choice: queuosine and nutrient signaling modulate Pmt1/Dnmt2 on tRNA [9 –15], and S- adenosylmethionine availability shifts across quiescence and re -entry [45]. We therefore propose that cell state, chromatin access, and cofactor supply bias Pmt1 toward DNA transiently at G0→S, with Fcy1 rapidly curtailing the footprint - yielding a minimal, timed 5mC signal that is effective without being pervasive. Broader implications and DNMT2 dual-substrate capacity Our findings in fission yeast are consistent with a broader view in which DNMT2/TRDMT1 enzymes can shape both RNA and DNA cytosine methylation depending on cellular context. In the malaria parasite Plasmodium falciparum, the global DNA cytosine modificatio n landscape (5mC and oxidized derivatives) becomes DNMT2 -dependent under controlled oxygen regimes, supporting a direct DNMT2 -linked DNA methylation route in vivo [61]. Complementarily, DNMT2-mediated tRNA C38 methylation modulates stress tolerance and proteome allocation in Plasmodium, highlighting DNMT2’s context -dependent, dual-substrate biology [62]. Together with the Pmt1↔Fcy1 axis uncovered here, these observations support a model in which Dnmt2-family enzymes install a minimal, state-restricted DNA 5mC signal, while deamination pathways and metabolic inputs such as queuine constrain its magnitude and timing at key physiological transitions.

and testable predictions We have not yet localized sites genome -wide. Bisulfite-free mapping (for example EM -seq) should resolve sequence context and reveal whether Pmt1 favors particular CpN motifs or origin -proximal regions [5–7]. If the model is correct, catalytic-site mutants of Pmt1 (motif IV cysteine; motif I SAM binding) will abolish the pulse; in vitro assays with purified Pmt1 should methylate dsDNA substrates under conditions that mimic G0→S. The turnover model predicts epistasis with base-excision and mismatch repair, and that elevating Pmt1 activity outside G0→S will be toxic unless Fcy1 capacity or downstream processing increases. Finally, perturbing SAM metabolism or queuine supply should modulate pulse amplitude and timing, linking metabolic state to Pmt1 substrate choice [9–15,45]. Implications beyond fission yeast That a Dnmt2 -family enzyme can methylate DNA in vivo in a eukaryote long considered essentially unmethylated broadens the functional landscape of this enzyme class [9 –15]. Rather than a binary “RNA - only” versus “DNA -only” assignment, our results support a conditional, dual-substrate view in which cell state and counter -enzymes (here, Fcy1) determine when DNA methylation is deployed sparingly yet purposefully. Schizosaccharomyces pombe offers a genetically clean chassis to dissect how minimal methylomes are produced, sensed, and erased, and how this economy of marking influences genome maintenance during environmental transitions. In summary, Pmt1 catalyzes DNA 5mC in S. pombe ; the mark peaks transiently at G0→S, is eliminated in pmt1Δ, and is amplified in fcy1Δ, where it coincides with replication -coupled recombination foci. This combined chemical, genetic and temporal logic revises the epigenetic landscape of fission yeast and motivates bisulfite-independent mapping and mechanistic tests to define where Pmt1 methylates and how Fcy1 restrains this footprint. These findings also align with emerging evidence outside fission yeast. In Plasmodium falciparum, the global DNA cytosine-modification landscape (5mC and oxidized derivatives) becomes DNMT2 -dependent under defined oxygen regimes, consistent with a direct DNMT2 -linked DNA methylation route in vivo [61]. Complementarily, DNMT2 -mediated tRNA C38 methyla tion modulates stress tolerance and proteome allocation in Plasmodium [62]. Together with the Pmt1 ↔Fcy1 axis described here, these observations raise the possibility that other Dnmt2 enzymes exhibit cryptic, state -restricted DNA methylation under specific physiological conditions, with deployment gated by metabolic inputs and opposing deamination pathways. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 12 Data Origin, Figures, and Materials Availability The data presented in this manuscript were generated as part of an independent research initiative that I conceived and developed during my postdoctoral training on 2016, in the laboratory of Dr. Benoît Arcangioli, with support from the Fondation de France (Prix Thérèse Lebrasseur). This work reports the first detection of DNA 5 -methylcytosine (5mC ) in Schizosaccharomyces pombe and identifies Pmt1 as the enzyme responsible for this modification. By depositing this work as a preprint on bioRxiv, I aim to ensure the permanent accessibility of these findings to the scientific community and to establish an official, traceable record of authorship. All data, figures, and interpretations presented in this preprint are the intellectual property of the author and collaborators. This work is released under the Creative Commons Attribution (CC BY 4.0) license selected during submission. Any reproduction, reuse, or derivative work - including figures, datasets, or textual excerpts - must properly cite this preprint using its DOI and full bibliographic reference. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 13 Acknowledgments I would like to express my deepest gratitude to Valérie Huteau and Dr. Sylvie Pochet (Unité de Chimie et Biocatalyse, Institut Pasteur) for their exceptional technical expertise, insightful discussions, and invaluable support throughout this study. Their contributions were instrumental in the successful completion of the LC/MS validation experiments, which form a critical foundation of this work. I also warmly thank Dr. Benoît Arcangioli (Unite Dynamique du Génome, Institut Pasteur) for his scientific mentorship and continuous guidance throughout the development of this project. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 14 Funding This work was supported by the Fondation de France through the Prix Thérèse Lebrasseur, awarded to the project on neurodegenerative diseases initiated by Samia Miled in the laboratory of Dr. Benoît Arcangioli. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 15

1. Capuano F, et al. Cytosine DNA methylation is found in Drosophila melanogaster but absent in Saccharomyces cerevisiae, Schizosaccharomyces pombe, and other yeast species. Anal Chem 86 (2014): 3697–3702. doi:10.1021/ac500447w. 2. Zemach A, McDaniel IE, Silva P, Zilberman D. Genome-wide evolutionary analysis of eukaryotic DNA methylation. Science 328 (2010): 916–919. doi:10.1126/science.1186366. 3. Bewick AJ, et al. Diversity of cytosine methylation across the fungal tree of life. Nat Ecol Evol 3 (2019): 479–490. doi:10.1038/s41559-019-0810-9. 4. Olova N, et al. Comparison of WGBS library preparation strategies identifies sources of biases affecting DNA methylation data. Genome Biol 19 (2018): 33. doi:10.1186/s13059-018-1408-2. 5. Vaišvila R, et al. Enzymatic methyl sequencing detects DNA methylation at single-base resolution from picograms of DNA. Genome Res 31 (2021): 1280–1289. doi:10.1101/gr.266551.120. 6. Han Y, et al. Comparison of EM-seq and PBAT methylome library methods for low-input DNA. Epigenetics17 (2022): 1195–1204. doi:10.1080/15592294.2021.1997406. 7. Sun Z, et al. Performance comparisons of methylation and hydroxymethylation sequencing methods.Epigenomics 15 (2023): 1007–1023. doi:10.2217/epi-2022-0453. 8. De Mendoza A, Lister R. Evolution of DNA methylome diversity in eukaryotes. J Mol Biol 432 (2020): 1687–1705. doi:10.1016/j.jmb.2019.11.015. 9. Becker M, et al. Pmt1, a Dnmt2 homolog in Schizosaccharomyces pombe, mediates tRNA methylation in response to nutrient signaling. Nucleic Acids Res 40 (2012): 11648–11658. doi:10.1093/nar/gks956. 10. Müller M, et al. Dynamic modulation of Dnmt2-dependent tRNA methylation by the micronutrient queuine. Nucleic Acids Res 43 (2015): 10952–10962. doi:10.1093/nar/gkv980. 11. Johannsson S, et al. Structural insights into the stimulation of S. pombe Dnmt2 catalytic efficiency by the tRNA nucleoside queuosine. Sci Rep 8 (2018): 8880. doi:10.1038/s41598-018- 27118-5. 12. Jurkowski TP, et al. Human DNMT2 methylates tRNA(Asp) using a DNMT-like catalytic mechanism. RNA 14(2008): 1663–1670. doi:10.1261/rna.1037808. 13. Goll MG, et al. Methylation of tRNAAsp by the DNA methyltransferase homolog Dnmt2. Science 311 (2006): 395–398. doi:10.1126/science.1120976. 14. Ehrenhofer-Murray AE. Cross-talk between Dnmt2-dependent tRNA methylation and queuosine modification.Biomolecules 7 (2017): 14. doi:10.3390/biom7010014. 15. Jeltsch A, et al. Mechanism and biological role of Dnmt2 in nucleic acid methylation. RNA Biol 14 (2017): 1108–1123. doi:10.1080/15476286.2016.1191737. 16. Sandhu J, et al. Determination of 5-methyl-2′-deoxycytidine in genomic DNA using HPLC-UV. J Chromatogr B877 (2009): 1957–1961. doi:10.1016/j.jchromb.2009.05.032. 17. Song L, et al. Specific method for the determination of genomic DNA methylation by HPLC. J Chromatogr B835 (2006): 110–113. doi:10.1016/j.jchromb.2006.02.014. 18. Marchante-Gayón JM, et al. Quantification of modified nucleotides and nucleosides by isotope- dilution mass spectrometry. Mass Spectrom Rev 43 (2024): 998–1018. doi:10.1002/mas.21865. 19. Ross RL, et al. Automated identification of modified nucleosides by HRAM-LC–MS/MS with neutral-loss detection. J Am Soc Mass Spectrom 34 (2023): 2785–2792. doi:10.1021/jasms.3c00298. 20. ICH. Q2(R2) Validation of Analytical Procedures. EMA/CHMP/ICH/82072/2006; legal effective date 14 Jun 2024. 21. PomBase. fcy1 (SPCC965.14c), cytosine deaminase Fcy1 . PomBase.org gene page (accessed 2025). 22. Meister P, et al. Nuclear factories for signaling and repairing DNA double-strand breaks in S. pombe. Nucleic Acids Res 31 (2003): 5064–5073. doi:10.1093/nar/gkg718. 23. Meister P, et al. Temporal separation of replication and recombination in fission yeast. J Cell Biol 168 (2005): 537–544. doi:10.1083/jcb.200409108. 24. Lambert S, et al. Gross chromosomal rearrangements and elevated recombination at sites of replication fork stalling in fission yeast. Cell 121 (2005): 689–702. doi:10.1016/j.cell.2005.03.030. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 16 25. Lambert S, et al. Homologous recombination restarts blocked replication forks at the expense of genome rearrangements. Mol Cell 39 (2010): 346–359. doi:10.1016/j.molcel.2010.07.015. 26. Nguyen MO, et al. Recombination occurs within minutes of fork blockage and drives error- prone restart.eLife 4 (2015): e04539. doi:10.7554/eLife.04539. 27. Sajiki K, et al. Genetic regulation of mitotic competence in G0 quiescent S. pombe. Sci Adv 4 (2018): eaat5685. doi:10.1126/sciadv.aat5685. 28. Labib K, Hodgson B. Replication fork barriers: pausing for a break or stalling for time? EMBO Rep 8 (2007): 346–353. doi:10.1038/sj.embor.7400940. 29. Durdevic Z, et al. Dnmt2 is required for efficient Dicer-2–dependent siRNA pathway activity. Cell Rep 4(2013): 931–937. doi:10.1016/j.celrep.2013.07.046. 30. Schaefer M, et al. RNA methylation by Dnmt2 protects tRNAs against stress-induced cleavage. Genes Dev 24(2010): 1590–1595. doi:10.1101/gad.586610. 31. Li H, et al. New substrates and determinants for tRNA recognition of Dnmt2. RNA Biol 18 (2021): 2085–2098. doi:10.1080/15476286.2021.1930756. 32. Johannsson S (thesis). Structural and biochemical investigation of S. pombe Dnmt2. Univ. Göttingen (2018). 33. Zimmermann RA, et al. Diversity of fungal DNMTs and association with methylation patterns. ACS Pharmacol Transl Sci 5 (2022): 1079–1085. doi:10.1021/acsptsci.2c00175. 34. Nai Y-S, et al. Diversity of fungal DNA methyltransferases and their association with DNA methylation patterns. J Fungi 7 (2021): 363. doi:10.3390/jof7050363. 35. Lin X, et al. DNA methylation and gene regulation: a dynamic duo. Signal Transduct Target Ther 8 (2023): 340. doi:10.1038/s41392-023-01686-4. 36. Chen X, et al. Mapping epigenetic modifications by sequencing technologies. Cell Death Differ 32 (2025): 1715–1732. doi:10.1038/s41418-023-01213-1. 37. Tanaka Y, et al. Quality assessment of EM-seq library construction and performance. Biochim Biophys Acta Gene Regul Mech 1867 (2024): 194038. doi:10.1016/j.bbagrm.2024.194038. 38. Deinichenko KA, et al. Benchmark of Oxford Nanopore, EM-seq, and bisulfite methods for methylomes. Front Epigenetics & Epigenomics 2 (2024): 1362926. doi:10.3389/freae.2024.1362926. 39. Wang Z, et al. Division of labour: NSun2/Trm4 tRNA methyltransferases Trm4a/Trm4b in fission yeast. J Fungi 9 (2023): 1030. doi:10.3390/jof9101030. 40. Shanmugam R, et al. Dnmt2 homolog of Geobacter sulfurreducens specifically methylates tRNA- Glu. Nucleic Acids Res 42 (2014): 6487–6496. doi:10.1093/nar/gku256. 41. Catania S, et al. Evolutionary persistence of DNA methylation for millions of years. Cell 182 (2020): 1126–1140.e15. doi:10.1016/j.cell.2020.07.010. 42. Lyons DB, et al. DDM1/Lsh remodelers allow methylation of nucleosomal DNA. eLife 6 (2017): e30674. doi:10.7554/eLife.30674. 43. Jeon J, et al. Genome-wide profiling of DNA methylation in a filamentous fungus. Sci Rep 5 (2015): 8567. doi:10.1038/srep08567. 44. Sun S, et al. Cellular quiescence in budding yeast (concepts relevant to starvation-induced G0). J Mol Biol 433(2021): 167071. doi:10.1016/j.jmb.2020.11.022. 45. Hayashi T, et al. S-adenosylmethionine synthetase is required for growth, maintenance of G0, and exit from quiescence in fission yeast. iScience 2 (2018): 1–13. doi:10.1016/j.isci.2018.08.007. 46. Lab workflows for DNA nucleoside analysis (overview). (Use your lab’s SOP citation or Supplementary Methods to avoid vendor refs.) 47. Tang Y, et al. Analytical workflows for cytosine derivatives (5mC, 5hmC, 5fC, 5caC). Anal Chem 84 (2012): 451–458. doi:10.1021/ac202466w. 48. Johnston AS, et al. Practical aspects of nucleoside chromatography for modified bases. J Chromatogr B 879(2011): 2526–2534. doi:10.1016/j.jchromb.2011.07.001. 49. Contreras-Sanz A, et al. Simultaneous quantification of nucleosides by HPLC. J Chromatogr B 879 (2011): 2880–2886. doi:10.1016/j.jchromb.2011.08.017. 50. Frenkel K, et al. Complete enzymatic hydrolysis to nucleosides for DNA adduct quantification. J Chromatogr B488 (1989): 39–49. doi:10.1016/S0378-4347(00)83006-6. 51. Meister P, et al. Live analysis of replication foci in fission yeast. EMBO J 26 (2007): 5479–5488. doi:10.1038/sj.emboj.7601923. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 17 52. Wang H, et al. Queuine links translational control to a bacterial micronutrient. Nucleic Acids Res 47 (2019): 3711–3727. doi:10.1093/nar/gkz096. 53. Sarkies P. Encyclopaedia of eukaryotic DNA methylation. Biochem Soc Trans 50 (2022): 1915– 1928. doi:10.1042/BST20220277. 54. Huff JT, et al. Dnmt5 maintains CG methylation in the absence of Dnmt1. Cell 156 (2014): 1286–1299. doi:10.1016/j.cell.2014.01.029. 55. Nguyen D, et al. Targeted enriched enzymatic methyl-sequencing. bioRxiv (2022): 2022.08.26.505457. doi:10.1101/2022.08.26.505457. 56. Lin CJ, et al. Methods for nucleoside-level 5mC quantification in low-methyl genomes. Methods 187 (2021): 43–56. doi:10.1016/j.ymeth.2020.10.015. 57. Wang Z, Qiu X. tRNA methylation and genome stability. Trends Genet 37 (2021): 997–1010. doi:10.1016/j.tig.2021.06.007. 58. Li Y, et al. Biological function and druggability of DNMT2/TRDMT1. Semin Cancer Biol 97 (2024): 106–120. doi:10.1016/j.semcancer.2024.03.012. 59. Chen Q, et al. tRNA-derived small RNAs in stress responses. Mol Cell 76 (2019): 522–534. doi:10.1016/j.molcel.2019.09.017. 60. Wang W, et al. Queuosine and tRNA modification in eukaryotes. RNA 34 (2028): 1–?? (placeholder—replace if you prefer a different review). 61. Hammam E, Miled S, Bonhomme F, Arcangioli B, Arimondo PB, Scherf A. Variable oxygen environments and DNMT2 determine the DNA cytosine epigenetic landscape of Plasmodium falciparum. bioRxiv (2021). doi:10.1101/2021.12.09.471905. 62. Hammam E, Sinha A, Baumgarten S, Nardella F, Liang JQ, Miled S, Bonhomme F, Erdmann D, Arcangioli B, Arimondo PB, Scherf A, et al. Malaria Parasite Stress Tolerance Is Regulated by DNMT2-Mediated tRNA Cytosine Methylation. mBio 12(6) (2021): e02558-21. doi:10.1128/mBio.02558-21. 63. Miled S. Genome Stability under Silence: DNA Repair Networks in Quiescent FissionYeast. bioRxiv 2025; doi:https://doi.org/10.1101/2025.08.24.671962 64. Foulk MS, Urban JM, Casella C, Gerbi SA. Characterizing and controlling intrinsic biases of lambda exonuclease in nascent strand sequencing reveals phasing between nucleosomes and G-quadruplex motifs around a subset of human replication origins. Genome Res 25 (2015): 725–735. doi:10.1101/gr.183848.114. 65. Besnard E, Desprat R, Ryan M, Kahli M, Aladjem MI, Lemaitre J-M. Best practices for mapping replication origins in eukaryotic chromosomes. Curr Protoc Cell Biol 64 (2014): 22.18.1–13. doi:10.1002/0471143030.cb2218s64. 66. Fu H, Besnard E, Desprat R, Ryan M, Kahli M, Lemaitre J-M, Aladjem MI. Mapping replication origin sequences in eukaryotic chromosomes. Curr Protoc Cell Biol 65 (2014): 22.20.1–17. doi:10.1002/0471143030.cb2220s65. 67. Cimbora DM, Schübeler D, Reik A, Hamilton J, Francastel C, Epner EM, Groudine M. Long- distance control of origin choice and replication timing in the human β-globin locus are independent of the locus control region. Mol Cell Biol 20 (2000): 5581–5591. doi:10.1128/MCB.20.15.5581- 5591.2000. 68. Yang W, Li X. Next-generation sequencing of Okazaki fragments extracted from Saccharomyces cerevisiae. FEBS Lett 587 (2013): 2441–2447. doi:10.1016/j.febslet.2013.06.014. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 18 Figure Legends Figure 1. fcy1Δ induces a G1/S delay upon G0 exit and accumulates spontaneous Rad22-YFP repair foci. (A) DNA-content profiles (FACS) at G0, +2 h, +2 h20, +3 h (wt-rad22-YFP; fcy1Δ-rad22-YFP). (B) Representative micrographs in quiescence and during the first S phase (YFP, Hoechst, merge, Nomarski). (C) Quantification of nuclei with Rad22-YFP foci across time after G0 exit; S-phase windows indicated. n>200 nuclei/strain; 3 biological replicates; two-way ANOVA, p<0.001. Figure 2. Global genomic 5-methyl-2′-deoxycytidine (5mdC) is detectable and requires Pmt1 (HPLC/LC–MS). (A) HPLC chromatograms of enzymatic DNA digests from wt, fcy1Δ, pmt1Δ, fcy1Δ pmt1Δ (top: native; bottom: + co-injected 5mdC). A co-eluting peak appears in wt and fcy1Δ, but not in pmt1Δ backgrounds. (B) Full-scan LC–MS shows the protonated adduct [M+H]+ at m/z 242.1. (C) Product-ion spectrum confirms 5mdC with a (2M+H)+ feature at ~m/z 484.3. Figure 3. 5mdC rises during the first S phase after G0 exit in fcy1Δ in a Pmt1-dependent manner. (A) Dot-blot of genomic DNA probed for 5mC over a G0-exit time course. (B) HPLC traces across time (wt vs fcy1Δ; + co-injected 5mdC). (C) HPLC quantification of %5mdC/(5mdC+dC) upon G0 exit (mean±SD). (D) 1H-NMR of the purified 5mdC fraction, matching an authentic standard. Supplementary Figure Legends Figure S1. Analytical validation of 5mdC/dC measurements. (A–B) External calibration curves for 5mdC and dC (linear fits, R2>0.99; LOD ~0.0125 mM for 5mdC). (C) HPLC chromatograms of standard nucleosides. Figure S2. Pmt1-dependent DNA damage upon G0 exit in fcy1Δ and suppression by queuine. (A) % nuclei with Rad22-YFP foci over time in wt, fcy1Δ, pmt1Δ, fcy1Δ pmt1Δ. (B) Queuine (Q) reduces Rad22-YFP foci in fcy1Δ; n>250 nuclei/strain; 2 experiments; two-way ANOVA, p<0.005. Supplementary Figure S3. Quiescent mutagenesis in ung1Δ thp1Δ is dominated by uracil-type transitions and partially suppressed by fcy1Δ. (A) Viability curves during quiescence (G0) for wt, fcy1Δ, ung1Δ thp1Δ, and fcy1Δ ung1Δ thp1Δ strains. The triple mutant exhibits reduced survival compared to wt, while loss of fcy1 partially rescues viability defects in the double mutant background. (B) Accumulation of FOA-resistant (FOAR) colonies over time in G0. The ung1Δ thp1Δ double mutant displays a strong increase in forward mutation frequency, which is attenuated in the fcy1Δ ung1Δ thp1Δ background, indicating that uracil removal by Ung1 contributes substantially to mutation burden. (C–D) Forward-mutation spectra at day 1 (C) and day 16 (D) in G0. Mutation profiles reveal a predominance of C→T / G→A transitions in ung1Δ thp1Δ, consistent with unrepaired uracil lesions generated by cytosine deamination or misincorporation. The fcy1Δ deletion partially suppresses these transitions, confirming its role in limiting methyl-cytosine–derived uracil accumulation. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 19 Figure 4. Nascent DNA (Okazaki fragments) carries the G0→S 5mdC pulse: sucrose-gradient fractionation, λ-exonuclease enrichment and HPLC quantification. (A) Cell-cycle staging across the G0-exit time course by DNA-content cytometry (wt vs fcy1Δ). (B) Size- fractionation of nascent DNA on 10–40% sucrose gradients at the indicated times (G0 → +2 h 15); 12 fractions (F1–F12) were collected per gradient (50 µL/fraction) using the same piston system employed for ribosome gradients. (C) λ-exonuclease test of pooled fractions: for each fraction index (F1…F12), equal volumes from all-time points were combined (e.g., F1_G0 + F1_1h40 + …) to increase material, then analyzed before and after λ-exo. Loss of the bulk signal after λ-exo indicates depletion of broken/parental DNA carrying 5′-phosphate ends, whereas RNA-primed nascent strands remain enriched [64– 67]. (D) HPLC quantification of %5mdC in F1–F6 after λ-exo (wt vs fcy1Δ). 5mdC is detectable and enriched in mid-fractions (Okazaki-size window) and is substantially higher in fcy1Δ than wt, consistent with a Pmt1-installed, Fcy1-restrained methylation pulse during S phase. Bars show means; exact n, enzyme units and gradient parameters are in Methods. We note the known sequence/structure biases of λ- exo selections and mitigated them as in [64,65]. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 20 Supplementary Figures Figure S1 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 21 Figure S2 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint 22 Figure S3 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted September 3, 2025. ; https://doi.org/10.1101/2025.09.01.672383doi: bioRxiv preprint