Replication Stress and eccDNA: Hydroxyurea, Cell Cycle Effects, and Study Design Considerations

Small extrachromosomal circular DNA (eccDNA) is an adaptable byproduct—and sometimes a driver—of genome plasticity. Replication stress amplifies the conditions under which DNA fragments can be excised and ligated into circles. Hydroxyurea (HU), a well-characterized ribonucleotide reductase inhibitor, provides a controllable model to induce replication stress in cell culture. This advanced tutorial explains how HU-induced stalled forks contribute to eccDNA formation, what to expect across the cell cycle, and how to design experiments that yield interpretable, reproducible measurements—strictly for research use only (RUO).

We focus on the core mechanism: S‑phase fork stalling and collapse under HU depleting dNTP pools, which can drive loop‑out excision and microhomology‑mediated end joining (MMEJ/alt‑EJ) to form circles. We then translate this mechanistic insight into practical steps for dose–time selection, sampling windows, enrichment and library construction for stressed, low‑input samples, artifact controls, and bioinformatics reporting standards. Throughout, we highlight where rigorous QC prevents confounding by apoptosis and mitochondrial DNA.

Why replication stress is a catalyst for genome plasticity

Replication stress refers to conditions that slow, stall, or collapse DNA replication forks—commonly through nucleotide pool depletion, polymerase impediments, or helicase–polymerase uncoupling. HU primarily inhibits ribonucleotide reductase (RNR), limiting conversion of NDPs to dNDPs and reducing dNTP pools required for polymerase progression. Stalled forks accumulate single‑stranded DNA (ssDNA), recruit RPA, and activate ATR–Chk1 checkpoints; origin firing is suppressed, and cells attempt to stabilize and restart forks. Under sustained stress, some forks collapse, producing one‑ended double‑strand breaks that must be processed and repaired.

In this landscape, eccDNA can emerge. Loop‑out excision of small genomic segments followed by ligation into circles is favored when resection exposes microhomologies and alternative end‑joining (MMEJ/TMEJ) operates robustly—conditions prevalent in S and G2. In essence, "replication stress eccdna" captures how fork perturbations alter break processing and pathway choice, biasing toward circularization in certain sequence contexts (e.g., repeats) and repair configurations.

Mechanistically, HU has two interlocking effects: dNTP depletion via RNR inhibition and, at some doses, reactive oxygen species (ROS) that oxidize Fe–S clusters in replicative polymerases, promoting polymerase dissociation even when nucleotide pools are partially adequate. Together, these push forks into unstable states primed for collapse and error‑prone repair.

Hydroxyurea as a replication‑stress model: mechanism and dosing guardrails

Hydroxyurea is widely used to create controlled replication stress. It stalls forks by depleting dNTPs and can generate ROS that affect polymerase function. The dual mechanism increases the likelihood of helicase–polymerase uncoupling, ssDNA exposure, and checkpoint activation.

To use HU effectively, distinguish "stress" from "toxicity." You want robust checkpoint signatures (p‑CHK1, γH2AX foci; reduced EdU incorporation) with minimal apoptosis (low Annexin V/PI and cleaved PARP) at harvest. Pilot titrations are indispensable because sensitivity differs across cell lines and primary cells.

Practical guardrails for cell models (non‑prescriptive; adjust per line after pilots) are best deployed as a narrative plan rather than a checklist. Start by defining an acute pulse paradigm in HeLa or U2OS: expose cells to 0.5–1 mM HU for 2–4 hours to induce clear stalling without widespread death, then release into fresh medium and sample the late‑S/G2 windows. In parallel, consider a chronic, low‑dose paradigm (around 0.2 mM HU for several days) if you are probing sequence‑context effects such as microsatellite‑associated circles; this regime perturbs polymerase progression and origin firing subtly, with distinct impacts on circle structure.

As you escalate dose or extend time, monitor the balance between checkpoint engagement and apoptosis. A moderate pulse at 1 mM for four hours will produce stronger ATR–Chk1 activation and more γH2AX foci than a 0.5 mM pulse, but viability should remain high in checkpoint‑competent lines. By contrast, extending 1–2 mM HU for 16–24 hours can tilt the response toward apoptosis in some genetic backgrounds, confounding eccDNA readouts with apoptotic fragments. Primary fibroblasts often demand gentler conditions—explore 0.2–0.5 mM for 2–6 hours, confirm stress markers, and only then lengthen exposures.

Checkpoint and viability readouts anchor these decisions. Run EdU pulses and measure incorporation drops to confirm slowed DNA synthesis; quantify γH2AX foci as a proxy for fork collapse and DSBs; blot for p‑CHK1 to confirm checkpoint activation. In the same samples, gate out conditions with high Annexin V/PI or cleaved PARP. Finally, profile DNA content by flow cytometry (PI or DAPI) to document %G1/S/G2 at harvest and during recovery.

Figure 1: Mechanism of HU‑stalled forks and loop‑out circularization

Caption: Schematic showing HU (RNR↓) inducing fork stalling and collapse, resection exposing microhomologies, and MMEJ/TMEJ‑mediated loop‑out excision and ligation into eccDNA. Alternative fates (reintegration, micronuclei) are noted. This figure illustrates the core "replication stress eccdna" concept.

Cell cycle context: formation in S‑phase and persistence into G2

Do circles form in S? Yes—S‑phase provides the substrate (active replication, emerging ssDNA, resection) and the enzymatic environment where alternative end‑joining pathways operate alongside homologous recombination. As forks stall or collapse, small segments can be excised and circularized. Do they persist in G2? Often, yes. Circles may remain detectable through G2 and segregate at mitosis, with daughter cells inheriting eccDNA while corresponding chromosomal deletions persist or are repaired.

The practical implication is that synchronization and sampling windows determine what you measure. If you collect at the end of HU treatment, you enrich for late‑S circles associated with acute stalling. If you sample 6–24 hours after release, you capture the persistence phase in G2/M and potential decline toward G1. Double‑thymidine block helps enrich a cohort entering S uniformly, while HU synchronization (strong, temporary arrest in early S) followed by release lets you follow a stress‑conditioned wave through S and into G2. Always verify fractions by flow cytometry and overlay EdU pulse data to confirm active replication during your sampling windows.

Understanding biogenesis mechanisms helps interpret this—see the detailed mechanistic context in Linking eccDNA to Genome Instability: alt‑EJ, Replication Stress, and Retrotransposons: Mechanisms linking replication stress and eccDNA.

Image 2: Flow cytometry profiles under HU

Caption: Conceptual flow cytometry histogram comparing DNA content in control vs HU‑treated cells. HU increases the fraction in early S and suppresses progression; recovery windows show re‑entry into late S/G2. Use this readout to align eccDNA sampling with cell‑cycle status.

Replication stress eccdna: from mechanism to measurement

Placing the exact phrase front and center, "replication stress eccdna" describes a measurable outcome of fork perturbation. It is not just about more circles; it is about which circles form, where junctions occur, and how repair pathway choice and sequence context shape the spectrum you detect. In repeat‑dense regions, microhomology can scaffold loop‑out excision. Under chronic low stress, circle structures can diversify without necessarily producing a steep increase in total counts. Under stronger acute pulses, you may observe a transient rise in junctions consistent with fork breakage and templated end‑joining.

Designing your assay to capture this nuance means choosing dose–time conditions that elicit checkpoint‑verified stress while preserving viability, sampling across S and G2 recovery, and reporting junction‑level evidence rather than only global counts.

Study design: a narrative plan for interpretable eccDNA measurements

A robust study does not hinge on a single checklist; it integrates model selection, dosing, synchronization, sampling, controls, enrichment, verification, and analysis. Here is how that typically unfolds.

Begin with a hypothesis anchored in sequence context or pathway dependence. If you are testing whether microsatellites produce circles preferentially under stress, favor a chronic low‑dose regimen (around 0.2 mM HU for several days) and include an aphidicolin arm, because a polymerase‑impediment stressor can serve as an orthogonal confirmation. If you are mapping acute fork collapse and repair trajectories, design 0.5–1 mM HU pulses for 2–4 hours and plan a recovery series to capture late‑S formation and G2 persistence.

Synchronize your cohorts. Double‑thymidine block produces a relatively uniform entry into S; HU synchronization arrests early S and creates a wave on release. Choose the approach that best matches your objectives and that your cell line tolerates well. Verify synchronization and progression by flow cytometry and EdU pulses.

Controls keep your interpretation honest. Untreated baselines provide a reference for circle counts and junction motifs in unstressed cells. An aphidicolin arm allows you to confirm that replication‑stress‑specific phenomena recur when polymerases are slowed by a different agent. If HU's ROS component is suspected at your doses, introduce antioxidant pre‑treatments cautiously to test whether circle signatures shift when oxidative contributions are mitigated. Where permitted, pathway probes—POLQ (MMEJ/TMEJ) inhibition versus DNA‑PKcs (c‑NHEJ) inhibition—can help confirm the reliance on microhomology‑mediated joining.

Plan enrichment and library prep for low input. Stressed cells yield less DNA and more fragile circles. Exonuclease digestion (ATP‑dependent nuclease) removes linear DNA; verify digestion with qPCR assays targeting linear loci. Rolling circle amplification (Phi29) increases sensitivity to small circles but introduces bias; balance RCA with PCR‑free steps for larger circles where feasible. Hybrid capture panels targeting repeats or loci of interest can increase signal in low‑input contexts, and long‑read confirmation for selected samples disambiguates repeat‑rich junctions.

Ensure your library prep accommodates low input—see the detailed guidance in Experimental Workflow for eccDNA Sequencing: Enrichment, library prep, and common pitfalls.

Verification and QC are the backbone of reproducibility. Confirm replication stress (EdU, γH2AX, checkpoint blots) and exclude apoptosis (Annexin V/PI, cleaved PARP). Validate eccDNA structures with inverse PCR across junctions and divergent primer PCR; when available, corroborate junctions using long‑read data. Track mitochondrial reads separately and filter robustly—mtDNA circles are a common confounder in enriched libraries. Quality control is critical when handling stressed samples—see: Quality Metrics for eccDNA Sequencing.

Add logistical rigor. Align your sampling and submission workflows with documented inputs and QC thresholds. If you will outsource sequencing or analysis, review the Sample Submission Guidelines to ensure buffer, volume, concentration, and shipping conditions protect circular integrity: Sample Submission Guidelines.

Bioinformatics and reporting: concrete parameters and examples

Analysis begins with clean data and transparent thresholds. For Illumina libraries enriched for circles, adapter trimming (e.g., Trim Galore or fastp) should be followed by mapping with split‑read awareness. Bowtie2 in end‑to‑end mode or BWA‑MEM can be configured to capture soft‑clipped alignments; many eccDNA tools wrap these mappers and extract junction candidates.

An example Illumina short‑read pipeline might look like this:

• Quality control: run FastQC and MultiQC; ensure per‑base quality > Q30 for the majority of reads; document duplication and GC content.
• Trimming: fastp with adapter auto‑detection; trim to remove low‑quality trailing bases (—cut_right) while avoiding over‑trimming that disrupts junction signals.
• Mapping: BWA‑MEM with default seeding, reporting secondary alignments; allow soft clips to surface junction sequences.
• Junction calling: a tool such as ecc_finder or Circle‑Map aggregates split reads and discordant pairs; set minimum support to ≥4 split reads plus ≥4 discordant pairs per junction, and require ≥99% coverage of the circle sequence in enriched libraries.
• Repeat annotation: intersect junctions and circles with RepeatMasker/Dfam tracks; flag repeat‑dense calls and require stronger evidence or orthogonal validation.
• Mitochondrial filtering: exclude or separately report mtDNA‑derived circles; if mtDNA dominates, revisit enrichment and RCA settings.
• Size distribution and genomic context: report median and interquartile ranges for circle sizes; annotate proximity to genes, enhancers, or repeat families.
• Depth and normalization: state per‑sample read counts; normalize counts to cell number or input nuclear DNA; include spike‑in recovery estimates if used.

For long‑read validation of enriched libraries, the emphasis shifts to continuity and junction confirmation. Map with minimap2 using parameters tuned for circular templates (—secondary=no to reduce multi‑mapping noise in repeats, with careful consideration for true duplicates). Require reads traversing the junction with high identity and confirm consensus junction sequences. Report per‑circle read counts and coverage, and, when feasible, share representative junction sequences.

Transparent reporting strengthens interpretation. Include the mapper and version, parameters, minimum junction support, repeat filtering strategy, mitochondrial handling, per‑sample depths, and normalization choices. For downstream analysis support, consult CD Genomics' Bioinformatics Services. To plan coverage realistically, frame targets with the explainer: Sequencing Depth and Coverage.

Data availability (RUO)

Public examples that can be used to reproduce circle‑calling and analysis pipelines include GEO accession GSE261856 (Circle‑seq of human bone‑marrow mesenchymal stem cells; short‑read Circle‑seq; see GEO landing page) and GSE165919 (rolling‑circle amplified full‑length eccDNA profiled by Oxford Nanopore; see GEO landing page). As a methods reference, see ATAC‑eccDNA data from Kumar et al., Sci Adv (2020) DOI:10.1126/sciadv.aba2489. Minimal analysis parameters for reproducibility: mapper (BWA‑MEM or minimap2), minimum junction support ≥4 split reads plus ≥4 discordant pairs, and explicit mitochondrial filtering. If available, include accession MD5s or file versions for exact replication.

Troubleshooting: artifacts and misreads, handled in prose

Apoptotic fragments are the most common confounder. If Annexin V/PI positivity rises, your "eccDNA" counts may inflate due to fragmented linear DNA that survived digestion or ligated into circles post‑lysis. Tighten the viability gate, shorten exposure, and collect earlier in recovery. Verify digestion efficiency by qPCR against linear loci and consider a second digestion step with fresh ATP cofactor.

Residual linear DNA can survive exonuclease treatment when cofactors are depleted or buffers are off. Re‑optimize nuclease conditions, verify reagent freshness, and include restriction enzyme pre‑clears that target known linear contaminants (e.g., plasmids) before exonuclease.

Mitochondrial dominance in enriched libraries is a widely reported issue. Minimize RCA bias toward small circles by adjusting reaction times and temperatures, and filter mitochondrial reads rigorously during analysis. In some contexts, pre‑clearing mitochondria before extraction helps, but weigh the risk to nuclear circles.

Repeat‑rich false positives arise when reads map ambiguously across tandem elements. Increase junction support thresholds, annotate repeats comprehensively, and require orthogonal confirmation (iPCR or long‑read junctions) for high‑impact calls. Think of repeats as acoustic mirrors: they bounce reads around; your pipeline needs enough signal to separate echo from source.

Low library yield in stressed samples is more logistics than theory. Pool biological replicates, use low‑input adapters and kits validated for nanogram‑level inputs, and keep lysis gentle to preserve circular integrity. If RCA becomes too biased, pivot to hybrid capture for loci of interest and layer in long‑read confirmation for junctions that remain ambiguous.

Image 3: Conceptual eccDNA fold‑change across dose and recovery

Caption: Schematic, not empirical data. Moderate HU doses and early recovery windows (late S–G2) often yield the highest measurable eccDNA counts, with declines as cells return to G1.

Enrichment and library considerations for stressed cells

Choosing the right enrichment strategy is essential for reproducibility. Exonuclease digestion removes linear DNA; verify with qPCR and avoid over‑digestion that may degrade fragile circles. Rolling circle amplification (Phi29) is sensitive to small circles but can bias representation; combine RCA with PCR‑free steps for larger circles where possible. Targeted capture can increase signal for low‑input contexts, and long‑read validation disambiguates repeat‑rich junctions.

Practical example workflow (RUO)

To ground the steps above, here is a neutral, example workflow that a PI might adopt in a HeLa or U2OS model.

Objective: Measure eccDNA formation under HU‑induced replication stress, focusing on S‑phase formation and G2 persistence. Design HU arms at 0.5 mM for two hours (acute), 1 mM for four hours (stronger acute), and 0.2 mM for four days (chronic). Include an aphidicolin arm (0.2 µM, two days) as an orthogonal stressor. Sample at the end of treatment and at 6, 24, and 48 hours post‑release.

Verification should show EdU incorporation drop, γH2AX foci, and Annexin V/PI negativity at harvest. Enrichment proceeds with exonuclease digestion to remove linear DNA and RCA for small circles; junctions are validated by inverse PCR. Sequencing on Illumina targets 30–50 million reads per sample for enriched libraries, with long‑read validation for a subset to confirm junction continuity. Analysis uses split‑read and discordant‑pair evidence to call circles, reports junction support and size distributions, annotates repeat content, and filters mtDNA robustly.

For sequencing support and downstream analysis, CD Genomics' Next Generation Sequencing and Bioinformatics Services can be used in RUO studies to process eccDNA libraries and generate standard reports. Disclosure: CD Genomics is our product. If you plan to submit samples, review the Sample Submission Guidelines to align volumes, buffers, and QC with low‑input requirements.

Conclusion and next steps

Stress models like HU reveal rules of eccDNA formation: S‑phase forks are primed for loop‑out excision and MMEJ‑mediated ligation, circles can persist into G2, and measured abundance depends strongly on dose–time choices and sampling windows. The most interpretable studies separate replication stress from toxicity, synchronize and verify cell‑cycle states, implement low‑input‑aware enrichment and library prep, and apply artifact‑resistant bioinformatics with transparent reporting.

To capture transient events effectively, choose the right enrichment methods: Exonuclease digestion, RCA, capture, and controls. And remember—Quality metrics determine confidence when handling stressed samples.

Finally, one phrase you may encounter in searches is "eccdna and hydroyurea." While it is a misspelling of hydroxyurea, the intent is the same: exploring eccDNA under HU‑induced replication stress.

Author

Yang H. — Senior Scientist, CD Genomics; University of Florida.

Yang is a genomics researcher with over 10 years of research experience in genetics, molecular and cellular biology, sequencing workflows, and bioinformatic analysis. Skilled in both laboratory techniques and data interpretation, Yang supports RUO study design and NGS-based projects.

References:

1. Musiałek MW. Hydroxyurea—The Good, the Bad and the Ugly. Genes (Basel) (2021). doi:10.3390/genes12071096. https://pubmed.ncbi.nlm.nih.gov/34356112/ (PMCID: PMC8304116).
2. Shaw AE et al. Revised mechanism of hydroxyurea‑induced cell cycle arrest. Proc Natl Acad Sci U S A (2024). doi:10.1073/pnas.2404470121. https://pubmed.ncbi.nlm.nih.gov/38496404/
3. Conti BA, Dewar JM. Mechanisms of direct replication restart at stressed replisomes. Nat Rev Mol Cell Biol (2020). doi:10.1038/s41580-020-00322-9. https://pubmed.ncbi.nlm.nih.gov/32853827/
4. Gadgil RY et al. Microsatellite break‑induced replication generates highly mutagenized extrachromosomal circular DNAs. NAR Cancer (2024). doi:10.1093/nar/zcae027. https://academic.oup.com/narcancer/article/6/2/zcae027/7689927
5. Hull RM, King M, Pizza G, Krueger F, Vergara X, Houseley J. Transcription‑induced formation of extrachromosomal DNA during yeast ageing. PLoS Biol (2019). doi:10.1371/journal.pbio.3000471. PubMed PMID: 31794573. https://pmc.ncbi.nlm.nih.gov/articles/PMC6890164/
6. Zuo S., Yi Y., Wang C., et al. "Extrachromosomal Circular DNA (eccDNA): From Chaos to Function." Frontiers in Cell and Developmental Biology (2021). doi:10.3389/fcell.2021.792555. Full text: Frontiers article (Zuo et al., 2021).
7. Zuo S, Yi Y, Wang C, et al. Extrachromosomal Circular DNA (eccDNA): From Chaos to Function. Front Cell Dev Biol. 2021;9:792555. doi:10.3389/fcell.2021.792555. https://pubmed.ncbi.nlm.nih.gov/35083218/
8. Paulsen T, Malapati P, Shibata Y, et al. MicroDNA levels are dependent on MMEJ, repressed by c‑NHEJ pathway, and stimulated by DNA damage. Nucleic Acids Res. 2021;49(20):11787–11799. doi:10.1093/nar/gkab984. (Full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC8599734/)
9. Kumar P, Kiran S, Saha S, et al. ATAC‑seq identifies thousands of extrachromosomal circular DNA in cancer and cell lines. Sci Adv. 2020;6(20):eaba2489. doi:10.1126/sciadv.aba2489. (Full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC7228742/)
10. Zhang P, Peng H, Llauro C, Mirouze M, et al. ecc_finder: a robust and accurate tool for detecting extrachromosomal circular DNA from sequencing data. Front Plant Sci. 2021;12:743742. doi:10.3389/fpls.2021.743742. (Full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC8672306/)
11. Wanchai V., Jenjaroenpun P., Leangapichat T., et al. CReSIL: Accurate Identification of Extrachromosomal Circular DNA from Long‑read Sequences. Briefings in Bioinformatics. 2022;23(6):bbac422. doi:10.1093/bib/bbac422. PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC10144670/
12. dos Santos CR, Bøllehuus Hansen L, Rojas‑Triana M, Johansen AZ, Perez‑Moreno M, Regenberg B. Variation of extrachromosomal circular DNA in cancer cell lines. Computational and Structural Biotechnology Journal. 2023;21:4207–4214. doi:10.1016/j.csbj.2023.08.027. PubMed: https://pubmed.ncbi.nlm.nih.gov/37705597/ .
13. Apraiz A., de Menezes RX, Forcales SV, et al. Studying Cell Cycle‑regulated Gene Expression by Two Complementary Cell Synchronization Protocols. Journal of Visualized Experiments (JoVE). 2017; doi:10.3791/55543. PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC5608252/
14. Chen G., Bradford WD., Sefton MT., et al. Cell Synchronization by Double Thymidine Block. Bio‑Protocol. 2018; doi:10.21769/BioProtoc.3038. PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC6156087/