Linking eccDNA to Genome Instability: alt-EJ, Replication Stress, and Retrotransposons

Abstract and scope

This advanced guide synthesizes how extrachromosomal circular DNA (eccDNA) emerges from genome instability and how it may persist, reintegrate, or modulate genome function under replication stress. We adopt a cautious balance regarding the idea that retrotransposons hijack alt-EJ for DNA replication and eccDNA biogenesis: we lay out supportive evidence, counter-examples, and testable predictions, with Drosophila embryogenesis used as the primary experimental anchor. The content is for research use only.

Why eccDNA sits at the center of genome instability

eccDNA is both a symptom and a driver of genomic instability. Catastrophic events like chromothripsis, error-prone repair through microhomology-mediated end joining (MMEJ, also called alt-EJ), and replication stress can all generate circular fragments ranging from sub-kilobase microDNAs to multi-megabase ecDNAs. Once formed, these circles can alter gene dosage, titrate transcription factors, or reinsert into chromosomes to remodel the genome. For a broader discussion of downstream pathological effects in oncology models, see the overview of eccDNA in cancer in the companion resource, eccDNA in Cancer: Gene Amplification, Oncogene Regulation, and Research Applications.

Across systems, a recurring observation is the presence of microhomology at circular junctions, suggestive of alt-EJ/MMEJ activity, while stress on replication forks correlates with increased eccDNA yield and altered fate dynamics. But how strong is the causal chain, and where do retrotransposons fit in? That is the focus of this guide.

Biogenesis models that leave readable scars in the genome

Multiple mechanisms can generate eccDNA, and each tends to leave characteristic sequence signatures that careful sequencing and analysis can recover.

We recognize at least four routes that are recurrent in the literature. First, chromothripsis produces many fragments after one or a few rounds of catastrophic breakage; some fragments circularize rather than reintegrate. Second, breakage–fusion–bridge cycles repeatedly break and rejoin dicentric chromosomes, creating amplifications and complex rearrangements that may spin off circular by-products. Third, fork stalling and template switching creates junctions with templated insertions and microhomology when replication forks encounter barriers. Fourth, alt-EJ/MMEJ exposes short microhomologies by resection and can resolve looped-out segments as eccDNA while sealing a microhomology scar at the chromosomal locus.

What should you expect to see in the data? In short-read libraries, true circular junctions appear as outward-facing read pairs and split reads bridging the circle junction. In long reads, the junction resolves as a head-to-tail concatemer or single-contig circle with a sharp breakpoint, often showing 2–10 bp microhomology, small templated insertions, or short indels. These signatures collectively point toward error-prone rejoining pathways over classical NHEJ.

alt-EJ/MMEJ and the target hypothesis: retrotransposons hijack alt-EJ for DNA replication and eccDNA biogenesis

MMEJ, frequently discussed interchangeably with alt-EJ, relies on limited resection around a double-strand break to reveal short tracts of microhomology. Polymerase theta (Polθ) can stabilize annealing and extend from the aligned microhomology, with Ligase III and XRCC1 sealing the final junction. If the break architecture contains a looped segment or repetitive element that forms a flap during alignment, the flap can be excised and circularized as eccDNA.

Figure 1. Polθ‑mediated MMEJ/alt‑EJ pathway producing eccDNA. End resection exposes short 3′ overhangs that are aligned via 2–10 bp microhomologies by Polθ (helicase and polymerase activities); non‑annealed flaps or looped segments are excised, gaps are filled and sealed by Polθ polymerase and Ligase III/XRCC1, and the excised loop can circularize as eccDNA carrying a characteristic microhomology‑flanked junction scar.

Evidence snapshot with inline sources

Several lines of evidence connect microhomology usage to eccDNA formation. A 2025 synthesis in Nucleic Acids Research argued that MMEJ is a frequent route for circle formation by exposing microhomology tracts during resection and ligating looped segments as circles; see the detailed mechanism review in Molecular mechanisms of extrachromosomal circular DNA formation — Nucleic Acids Research (2025) (DOI 10.1093/nar/gkaf122). According to the NAR Cancer discussion of 2024 on microsatellite-driven eccDNA junctions, replication-associated template switching often co-occurs with microhomology features at circle junctions (NAR Cancer, 2024; DOI 10.1093/narcancer/zcae027). At replication-challenged microsatellites, Gadgil et al. (2024) used inverse PCR and sequencing to validate eccDNAs with templated insertions and 2–10 bp microhomology, consistent with break-induced replication and MMEJ-like scars (DOI: https://doi.org/10.1093/nar/gkae417). In an engineered setting, CRISPR-induced circle formation proceeds via distinct repair pathways with different inhibitor sensitivities, implying contributions from both NHEJ and MMEJ depending on context, as shown in Cancer Discovery (2025; DOI: https://doi.org/10.1158/2159-8290.CD-23-1117).

Interpretation

The weight of evidence supports alt-EJ/MMEJ involvement in eccDNA junction formation, especially under replication stress, and particularly near repeats where microhomology is abundant. However, universal Polθ dependence is not established across all systems. Junction microhomology is a strong indicator, but not exclusive proof, of MMEJ—some NHEJ or templated repair events can mimic parts of this signature. Robust conclusions require orthogonal validation and pathway perturbations.

Retrotransposons, circular intermediates, and what the evidence really shows

The hypothesis of interest is often phrased dramatically: retrotransposons hijack alt-EJ for DNA replication and eccDNA biogenesis. What should we make of it?

LTR elements generate cDNA via reverse transcription and rely on integrase for insertion; non‑LTR elements typically use target‑primed reverse transcription and generate linear cDNA intermediates. Some viral systems demonstrate circular LTR products (e.g., HIV 2‑LTR circles), but endogenous retrotransposons are a different matter. During Drosophila aging, eccDNA containing transposable element sequences accumulates, especially when RNA silencing is perturbed. This demonstrates that TE sequences appear in circular DNA fractions after exonuclease enrichment and outward‑facing PCR validation, as documented by Yang et al., PLOS Genetics (2022) (DOI 10.1371/journal.pgen.1010024). In contrast, DIRS‑1 non‑LTR retrotransposons in Dictyostelium amplify via linear, single‑stranded cDNA intermediates rather than circles, arguing against a universal circular step for endogenous elements, as shown in Godiska et al., Nucleic Acids Research (2020) (DOI 10.1093/nar/gkaa289).

Figure 2. Retrotransposon lifecycle and disputed circular intermediates — LTR retrotransposons can produce one‑ and two‑LTR circular by‑products during reverse transcription and intrachromatid recombination, while evidence for stable circular intermediates from non‑LTR elements is weaker.

Cautious synthesis

Supportive patterns include the detectability of TE sequences in eccDNA pools in flies, the frequent microhomology at eccDNA junctions, and strong links between replication stress and circle formation near repeats. Limits and gaps persist: direct, Drosophila-specific molecular tracing of endogenous LTR circle intermediates remains sparse; non-LTR elements often favor linear intermediates. Therefore, the phrase retrotransposons hijack alt-EJ for DNA replication and eccDNA biogenesis should be read as a testable, context-dependent model, not a universal law.

For computational handling of TE-rich circles—including repeat-aware mapping, artifact suppression, and junction reporting—see the companion bioinformatics guide focused on repeat filtering and standards.

• Computational filtering for repetitive elements is covered in Bioinformatics for eccDNA.

Replication stress, chromothripsis, and circle fate

Replication stress is a reliable upstream driver of genome instability, and it correlates with increased eccDNA generation in multiple cell systems. Hydroxyurea and aphidicolin perturb replication dynamics and produce distinctive landscapes of copy‑number alterations, micronuclei, and rearrangements in human cells, as reported in Nature Communications (2022) — replication stress and copy‑number landscapes (DOI 10.1038/s41467-022-33485-4). In parallel, work at microsatellites demonstrates that replication‑associated mechanisms can yield eccDNAs with microhomology scars, as shown by Gadgil et al., Nucleic Acids Research (2024) — microsatellite break‑induced eccDNA (DOI 10.1093/nar/gkae417).

What happens to circles once they form? Several fates are possible. Some circles persist as autonomous-like elements that segregate unevenly across divisions. Others reintegrate into chromosomes, potentially forming homogeneously staining regions. Many probably degrade or dilute if not replicated or selected.

Figure 3. Chromothripsis model: catastrophic chromosome shattering generates multiple fragments that can be re‑ligated, reintegrated into chromosomes, or circularized to form eccDNA/ecDNA; these circles may persist, be lost, or later reintegrate, producing focal amplifications or genomic remodeling.

Chromothripsis adds another layer. Catastrophic shattering creates numerous fragments that can be re‑ligated in complex patterns; some fragments circularize into eccDNA or ecDNA, while others reintegrate. Long‑read sequencing of chromothripsis cases has visualized these rearrangement landscapes and supports the plausibility of circle generation in such contexts (see the long‑read chromothripsis study inLong‑read sequencing reveals chromothripsis in a molecularly unsolved case of Cornelia de Lange syndrome — Frontiers in Genetics (2024) (DOI: 10.3389/fgene.2024.1358334); DOI 10.3389/fgene.2024.1358334). For a deeper discussion of whether circles can maintain copy number independently versus mainly reintegrating to drive amplifications, see the companion resource on Autonomous Replication in eccDNA.

Cautious synthesis

There is strong evidence that replication stress elevates eccDNA yield and that both reintegration and autonomous-like maintenance can occur, depending on cell state and selective pressures. Which fate dominates is context dependent. Disentangling them experimentally requires orthogonal readouts: long-read assemblies, cytogenetic imaging, and circle-specific enrichment plus junction validation.

Drosophila embryogenesis as a testbed for mechanism and measurement

Early Drosophila embryogenesis features extraordinarily fast S phases with minimal gap phases, making replication stress inherently likely at repetitive loci. The genome carries diverse LTR and non-LTR elements, microsatellites, and tandem repeats—fertile ground for template switching and microhomology-mediated repair. The system is genetically tractable, allowing pathway perturbations in Polθ/TMEJ components and small-molecule stressors. While direct reports of HU or aphidicolin inducing eccDNA in embryos are limited, evidence from other systems and from adult/aging flies justifies using embryos as a rigorous, testable model.

A reproducible experiment can be organized in stages. First, collect and stage 0–2 hour embryos, dechorionate, and focus on pre-cellularization cycles where S phases are rapid. Maintain tight timing windows to limit developmental variability. Second, titrate hydroxyurea or aphidicolin exposure in permeabilized embryos or cultured embryonic cells to introduce fork perturbations without causing lethality; parallel vehicle controls are mandatory. Third, extract DNA and enrich for circles using exonuclease digestion, spiking circular standards to quantify enrichment efficiency. If input is limiting, employ rolling circle amplification or Circle‑Seq-style library preparation. Fourth, prepare both Illumina short-read and nanopore long-read libraries; enrich for outward-facing reads where possible, and for long reads target N50 sufficient to span expected circles. Finally, validate junctions via inverse PCR and Sanger sequencing and, if desired, junction‑specific qPCR for quantification across timepoints or genotypes.

Controls should include mock-treated embryos, exonuclease-minus libraries to estimate background, and heat‑inactivated enzyme controls. Positive controls may include adult tissues known to accumulate TE‑eccDNA with age and plasmid spike‑ins to verify enrichment and library recovery. Genetic perturbations such as Polθ loss-of-function or Ligase III/XRCC1 depletion help test MMEJ contributions, while DNA‑PKcs or ATM perturbations probe NHEJ/HR involvement. Expected signatures include an enrichment of circle junctions with 2–10 bp microhomology, small templated insertions, increased TE‑mapping circles under stress, shifts in circle size distributions, and gene‑proximal circles near replication‑challenged repeats.

Disclosure and neutral example of outsourcing support

Disclosure: CD Genomics is our product. In practice, teams without in-house library prep or long-read capacity sometimes outsource circle enrichment and sequencing to a vendor to accelerate iteration. For example, researchers can use the eccDNA (circular DNA) sequencing workflow to combine exonuclease-based enrichment with short- and long-read platforms, then validate selected junctions with PCR and Sanger. See the non-clinical service overview at the CD Genomics eccDNA sequencing service page for a description of typical inputs, QC metrics, and deliverables.

• Example workflow overview at eccDNA Sequencing Service.

Bioinformatics and reporting standards for repeat-rich circles

Repeat-aware mapping and artifact control are the backbone of credible eccDNA calls, especially when transposable elements or microsatellites are involved. A practical two‑path pipeline works well. On the short‑read side, employ circle‑aware callers such as Circle‑Map, AA, or integrated frameworks like eccDNA‑pipe, requiring split reads and outward‑facing pairs that span the junction and carefully de‑duplicating to manage PCR artifacts. On the long‑read side, map with minimap2 using parameters tuned for repeats, assemble suspect circles when coverage permits, validate head‑to‑tail structure, and compute per‑circle coverage relative to flanking regions. For each junction, report microhomology length and sequence, any templated insertions, and indel sizes; record leftmost coordinates and offsets to represent ambiguous alignments in repetitive contexts. Artifact filters should remove junctions explained by chimeric PCR products, ligation artifacts, or mapping bleed‑through across highly similar repeats; spike‑ins and simulated reads help benchmark false positives.

A minimal reporting template should include sample and condition metadata (stress regimen and stage), enrichment metrics (fraction linear DNA removed; recovery of circular spike‑ins), a junction table (coordinates, microhomology, read support, TE annotation, size, validation status), and reproducibility metrics (inter‑replicate overlap and abundance correlations; sensitivity relative to spike‑ins). For a methodological review and consolidated tooling, see eccDNA‑pipe, Briefings in Bioinformatics (2024).

Open questions and testable predictions in flies

Given current evidence, several experiments could sharpen understanding. Does Polθ loss reduce TE‑containing eccDNA specifically under replication stress in embryos? A drop in microhomology‑rich junctions alongside a shift toward NHEJ‑like junctions would support MMEJ contributions. Are LTR retrotransposon circles directly observable as discrete head‑to‑tail long reads in embryos, or do TE‑eccDNAs largely reflect by‑products of repair at TE‑flanking repeats? Enrichment of LTR–LTR junctions would support true circles; enrichment of TE–host repeat junctions would favor repair by‑products. Does replication stress increase the ratio of eccDNA persistence versus reintegration in embryonic lineages? Time‑course long‑read sequencing and imaging could reveal uneven segregation (autonomous‑like persistence) versus homogeneously staining regions (reintegration). Finally, can template switching and FoSTeS‑like signatures be cleanly separated from MMEJ scars at embryonic microsatellites? The presence of templated insertions spanning non‑adjacent templates would support FoSTeS/MMBIR contributions.

These experiments can be executed in Drosophila with careful staging, pathway perturbations, and modern circle‑aware sequencing, providing a rigorous test of when and how retrotransposons appear to engage alt‑EJ/MMEJ machinery.

Putting the hypothesis in perspective

The phrase retrotransposons hijack alt-EJ for DNA replication and eccDNA biogenesis captures an appealing unifying model. Evidence in favor includes TE‑eccDNA accumulation in flies, microhomology usage at many circle junctions, and strong links between replication stress and circle formation near repeats. Counterpoints include established linear intermediates for certain non‑LTR elements and the possibility that many TE‑eccDNAs are repair by‑products rather than functional circular intermediates in a retrotransposon lifecycle. On balance, the claim is best treated as a context‑dependent, testable hypothesis that should be evaluated locus by locus and system by system.

A closing, RUO‑friendly option

If you need additional capacity for circle‑enriched sequencing or independent junction validation as you evaluate these mechanisms, a research‑use provider can help with standardized QC metrics and deliverables so you can focus on the experimental variables. See the earlier example link for an outline of typical sequencing inputs and outputs.

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. Gadgil RY, Zhou B, Dannenberg JH, et al. Microsatellite break‑induced replication generates highly mutagenized extrachromosomal circular DNAs. Nucleic Acids Research. 2024. DOI: 10.1093/nar/gkae417. https://doi.org/10.1093/nar/gkae417
2. Eugen‑Olsen RAB, Madsen BS, Pallesen S, et al. Molecular mechanisms of extrachromosomal circular DNA formation. Nucleic Acids Research. 2025. DOI: 10.1093/nar/gkaf122. https://doi.org/10.1093/nar/gkaf122
3. Hu J, Liu Y, Chen X, et al. Microhomology‑mediated circular DNA formation from nucleosome‑protected DNA fragments in germ cells. eLife. 2023. DOI: 10.7554/eLife.87115. https://doi.org/10.7554/eLife.87115
4. NAR Cancer editorial/article on microsatellite‑driven eccDNA junctions with microhomology features. NAR Cancer. 2024.DOI: 10.1093/nar/zcae027. https://academic.oup.com/narcancer/article/6/2/zcae027/7689927
5. Sun L, Liu Y, Zhang H, et al. Disparate Pathways for Extrachromosomal DNA Biogenesis and Genomic DNA Repair. Cancer Discovery. 2025. DOI: 10.1158/2159-8290.CD-23-1117. https://doi.org/10.1158/2159-8290.CD-23-1117
6. Yang N, Srivastav SP, Rahman R, et al. Transposable element landscapes in aging Drosophila. PLOS Genetics. 2022. DOI: 10.1371/journal.pgen.1010024. https://doi.org/10.1371/journal.pgen.1010024
7. Nelson JO, Watase GJ, Warsinger‑Pepe N, et al. The retrotransposon R2 maintains Drosophila ribosomal DNA repeats. Proceedings of the National Academy of Sciences. 2023. DOI: 10.1073/pnas.2221613120. https://doi.org/10.1073/pnas.2221613120
8. Godiska R, Yang S, Durand‑Dubief M, et al. DIRS retrotransposons amplify via linear, single‑stranded cDNA intermediates. Nucleic Acids Research. 2020. DOI: 10.1093/nar/gkaa289. https://doi.org/10.1093/nar/gkaa289
9. Cortés‑Ciriano I, Lee JJ‑K, Xi R, et al. Comprehensive analysis of chromothripsis in 2,658 human cancers using whole‑genome sequencing. Nature Genetics. 2020. DOI: 10.1038/s41588‑019‑0576‑7. https://doi.org/10.1038/s41588-019-0576-7
10. Bestetti I, Perrone MD, Cesare F, et al. Long‑read sequencing reveals chromothripsis in a constitutional case. Frontiers in Genetics. 2024. DOI: 10.3389/fgene.2024.1358334. https://doi.org/10.3389/fgene.2024.1358334
11. Zhao Y, Xie J, Zhang L, et al. Extrachromosomal circular DNA: Current status and future prospects. eLife. 2022. DOI: 10.7554/eLife.81412. https://doi.org/10.7554/eLife.81412
12. eccDNA‑pipe consortium. eccDNA‑pipe integrates detection and analysis of extrachromosomal circular DNA. Briefings in Bioinformatics. 2024. DOI: 10.1093/bib/bbae034. https://doi.org/10.1093/bib/bbae034
13. Hanscom T, Woodward N, Batorsky R, Brown AJ, Roberts SA, McVey M. Characterization of sequence contexts that favor alternative end joining at Cas9‑induced double‑strand breaks. Nucleic Acids Research. 2022;50(13):7465–7478. DOI: 10.1093/nar/gkac575. https://academic.oup.com/nar/article/50/13/7465/6640225
14. Collins AS, Johnson SC, et al. Replication stress generates distinctive landscapes of DNA copy number alterations in human cells. Genome Biology. 2022;23:207. DOI: 10.1186/s13059-022-02781-0. https://doi.org/10.1186/s13059-022-02781-0