APOBEC-Induced Mutagenesis in Cancer.

The initiation, progression, and relapse of cancers often result from mutations occurring within somatic cells. Consequently, processes that elevate mutation rates accelerate carcinogenesis and hinder the development of long-lasting therapeutics. Recent sequencing of human cancer genomes has identified patterns of mutations, termed mutation signatures, many of which correspond to specific environmentally induced and endogenous mutation processes. Some of the most frequently observed mutation signatures are caused by dysregulated activity of APOBECs, which deaminate cytidines in single-stranded DNA at specific sequence motifs causing C-to-T and C-to-G substitutions. In humans, APOBEC-generated genetic heterogeneity in tumor cells contributes to carcinogenesis, metastasis, and resistance to therapeutics. Here, we review the current understanding of APOBECs' role in cancer mutagenesis and impact on disease and the biological processes that influence APOBEC mutagenic capacity.

Her2 amplification, Rel-A, and Bach1 can influence APOBEC3A expression in breast cancer cells.

APOBEC-induced mutations occur in 50% of sequenced human tumors, with APOBEC3A (A3A) being a major contributor to mutagenesis in breast cancer cells. The mechanisms that cause A3A activation and mutagenesis in breast cancers are still unknown. Here, we describe factors that influence basal A3A mRNA transcript levels in breast cancer cells. We found that basal A3A mRNA correlates with A3A protein levels and predicts the amount of APOBEC signature mutations in a panel of breast cancer cell lines, indicating that increased basal transcription may be one mechanism leading to breast cancer mutagenesis. We also show that alteration of ERBB2 expression can drive A3A mRNA levels, suggesting the enrichment of the APOBEC mutation signature in Her2-enriched breast cancer could in part result from elevated A3A transcription. Hierarchical clustering of transcripts in primary breast cancers determined that A3A mRNA was co-expressed with other genes functioning in viral restriction and interferon responses. However, reduction of STAT signaling via inhibitors or shRNA in breast cancer cell lines had only minor impact on A3A abundance. Analysis of single cell RNA-seq from primary tumors indicated that A3A mRNA was highest in infiltrating immune cells within the tumor, indicating that correlations of A3A with STAT signaling in primary tumors may be result from higher immune infiltrates and are not reflective of STAT signaling controlling A3A expression in breast cancer cells. Analysis of ATAC-seq data in multiple breast cancer cell lines identified two transcription factor sites in the APOBEC3A promoter region that could promote A3A transcription. We determined that Rel-A, and Bach1, which have binding sites in these peaks, elevated basal A3A expression. Our findings highlight a complex and variable set of transcriptional activators for A3A in breast cancer cells.

Genetic inhibitors of APOBEC3B-induced mutagenesis.

The cytidine deaminases APOBEC3A (A3A) and APOBEC3B (A3B) are prominent mutators of human cancer genomes. However, tumor-specific genetic modulators of APOBEC-induced mutagenesis are poorly defined. Here, we used a screen to identify 61 gene deletions that increase A3B-induced mutations in yeast. We also determined whether each deletion was epistatic with Ung1 loss, which indicated whether the encoded factors participate in the homologous recombination (HR)-dependent bypass of A3B/Ung1-dependent abasic sites or suppress A3B-catalyzed deamination by protecting against aberrant formation of single-stranded DNA (ssDNA). We found that the mutation spectra of A3B-induced mutations revealed genotype-specific patterns of strand-specific ssDNA formation and nucleotide incorporation across APOBEC-induced lesions. Combining these three metrics, we were able to establish a multifactorial signature of APOBEC-induced mutations specific to (1) failure to remove H3K56 acetylation, (2) defective CTF18-RFC complex function, and (3) defective HR-mediated bypass of APOBEC-induced lesions. We extended these results by analyzing mutation data for human tumors and found BRCA1/2-deficient breast cancers display three- to fourfold more APOBEC-induced mutations. Mirroring our results in yeast, Rev1-mediated C-to-G substitutions are mainly responsible for increased APOBEC-signature mutations in BRCA1/2-deficient tumors, and these mutations associate with lagging strand synthesis during replication. These results identify important factors that influence DNA replication dynamics and likely the abundance of APOBEC-induced mutation during tumor progression. They also highlight a novel role for BRCA1/2 during HR-dependent lesion bypass of APOBEC-induced lesions during cancer cell replication.

Genome-wide maps of rare and atypical UV photoproducts reveal distinct patterns of damage formation and mutagenesis in yeast chromatin.

Ultraviolet (UV) light induces different classes of mutagenic photoproducts in DNA, namely cyclobutane pyrimidine dimers (CPDs), 6-4 photoproducts (6-4PPs), and atypical thymine-adenine photoproducts (TA-PPs). CPD formation is modulated by nucleosomes and transcription factors (TFs), which has important ramifications for Ultraviolet (UV) mutagenesis. How chromatin affects the formation of 6-4PPs and TA-PPs is unclear. Here, we use UV damage endonuclease-sequencing (UVDE-seq) to map these UV photoproducts across the yeast genome. Our results indicate that nucleosomes, the fundamental building block of chromatin, have opposing effects on photoproduct formation. Nucleosomes induce CPDs and 6-4PPs at outward rotational settings in nucleosomal DNA but suppress TA-PPs at these settings. Our data also indicate that DNA binding by different classes of yeast TFs causes lesion-specific hotspots of 6-4PPs or TA-PPs. For example, DNA binding by the TF Rap1 generally suppresses CPD and 6-4PP formation but induces a TA-PP hotspot. Finally, we show that 6-4PP formation is strongly induced at the binding sites of TATA-binding protein (TBP), which is correlated with higher mutation rates in UV-exposed yeast. These results indicate that the formation of 6-4PPs and TA-PPs is modulated by chromatin differently than CPDs and that this may have important implications for UV mutagenesis.

CTCF binding modulates UV damage formation to promote mutation hot spots in melanoma.

Somatic mutations in DNA-binding sites for CCCTC-binding factor (CTCF) are significantly elevated in many cancers. Prior analysis has suggested that elevated mutation rates at CTCF-binding sites in skin cancers are a consequence of the CTCF-cohesin complex inhibiting repair of UV damage. Here, we show that CTCF binding modulates the formation of UV damage to induce mutation hot spots. Analysis of genome-wide CPD-seq data in UV-irradiated human cells indicates that formation of UV-induced cyclobutane pyrimidine dimers (CPDs) is primarily suppressed by CTCF binding but elevated at specific locations within the CTCF motif. Locations of CPD hot spots in the CTCF-binding motif coincide with mutation hot spots in melanoma. A similar pattern of damage formation is observed at CTCF-binding sites in vitro, indicating that UV damage modulation is a direct consequence of CTCF binding. We show that CTCF interacts with binding sites containing UV damage and inhibits repair by a model repair enzyme in vitro. Structural analysis and molecular dynamic simulations reveal the molecular mechanism for how CTCF binding modulates CPD formation.

Essential Roles for Polymerase θ-Mediated End Joining in the Repair of Chromosome Breaks.

DNA polymerase theta (Pol θ)-mediated end joining (TMEJ) has been implicated in the repair of chromosome breaks, but its cellular mechanism and role relative to canonical repair pathways are poorly understood. We show that it accounts for most repairs associated with microhomologies and is made efficient by coupling a microhomology search to removal of non-homologous tails and microhomology-primed synthesis across broken ends. In contrast to non-homologous end joining (NHEJ), TMEJ efficiently repairs end structures expected after aborted homology-directed repair (5' to 3' resected ends) or replication fork collapse. It typically does not compete with canonical repair pathways but, in NHEJ-deficient cells, is engaged more frequently and protects against translocation. Cell viability is also severely impaired upon combined deficiency in Pol θ and a factor that antagonizes end resection (Ku or 53BP1). TMEJ thus helps to sustain cell viability and genome stability by rescuing chromosome break repair when resection is misregulated or NHEJ is compromised.

Characterization of sequence contexts that favor alternative end joining at Cas9-induced double-strand breaks.

Alternative end joining (alt-EJ) mechanisms, such as polymerase theta-mediated end joining, are increasingly recognized as important contributors to inaccurate double-strand break repair. We previously proposed an alt-EJ model whereby short DNA repeats near a double-strand break anneal to form secondary structures that prime limited DNA synthesis. The nascent DNA then pairs with microhomologous sequences on the other break end. This synthesis-dependent microhomology-mediated end joining (SD-MMEJ) explains many of the alt-EJ repair products recovered following I-SceI nuclease cutting in Drosophila. However, sequence-specific factors that influence SD-MMEJ repair remain to be fully characterized. Here, we expand the utility of the SD-MMEJ model through computational analysis of repair products at Cas9-induced double-strand breaks for 1100 different sequence contexts. We find evidence at single nucleotide resolution for sequence characteristics that drive successful SD-MMEJ repair. These include optimal primer repeat length, distance of repeats from the break, flexibility of DNA sequence between primer repeats, and positioning of microhomology templates relative to preferred primer repeats. In addition, we show that DNA polymerase theta is necessary for most SD-MMEJ repair at Cas9 breaks. The analysis described here includes a computational pipeline that can be utilized to characterize preferred mechanisms of alt-EJ repair in any sequence context.

Asymmetric repair of UV damage in nucleosomes imposes a DNA strand polarity on somatic mutations in skin cancer.

Nucleosomes inhibit excision repair of DNA damage caused by ultraviolet (UV) light, and it has been generally assumed that repair inhibition is equivalent on both sides of the nucleosome dyad. Here, we use genome-wide repair data to show that repair of UV damage in nucleosomes is asymmetric. In yeast, nucleosomes inhibit nucleotide excision repair (NER) of the nontranscribed strand (NTS) of genes in an asymmetric manner, with faster repair of UV damage occurring on the 5' side of the nucleosomal DNA. Analysis of genomic repair data from UV-irradiated human cells indicates that NER activity along the NTS is also elevated on the 5' side of nucleosomes, consistent with the repair asymmetry observed in yeast nucleosomes. Among intergenic nucleosomes, repair activity is elevated on the 5' side of both DNA strands. The distribution of somatic mutations in nucleosomes shows the opposite asymmetry in NER-proficient skin cancers, but not in NER-deficient cancers, indicating that asymmetric repair of nucleosomal DNA imposes a strand polarity on UV mutagenesis. Somatic mutations are enriched on the relatively slow-repairing 3' side of the nucleosomal DNA, particularly at positions where the DNA minor groove faces away from the histone octamer. Asymmetric repair and mutagenesis are likely caused by differential accessibility of the nucleosomal DNA, a consequence of its left-handed wrapping around the histone octamer.

Mechanistic basis for microhomology identification and genome scarring by polymerase theta.

DNA polymerase theta mediates an end joining pathway (TMEJ) that repairs chromosome breaks. It requires resection of broken ends to generate long, 3' single-stranded DNA tails, annealing of complementary sequence segments (microhomologies) in these tails, followed by microhomology-primed synthesis sufficient to resolve broken ends. The means by which microhomologies are identified is thus a critical step in this pathway, but is not understood. Here we show microhomologies are identified by a scanning mechanism initiated from the 3' terminus and favoring bidirectional progression into flanking DNA, typically to a maximum of 15 nucleotides into each flank. Polymerase theta is frequently insufficiently processive to complete repair of breaks in microhomology-poor, AT-rich regions. Aborted synthesis leads to one or more additional rounds of microhomology search, annealing, and synthesis; this promotes complete repair in part because earlier rounds of synthesis generate microhomologies de novo that are sufficiently long that synthesis is more processive. Aborted rounds of synthesis are evident in characteristic genomic scars as insertions of 3 to 30 bp of sequence that is identical to flanking DNA ("templated" insertions). Templated insertions are present at higher levels in breast cancer genomes from patients with germline BRCA1/2 mutations, consistent with an addiction to TMEJ in these cancers. Our work thus describes the mechanism for microhomology identification and shows how it both mitigates limitations implicit in the microhomology requirement and generates distinctive genomic scars associated with pathogenic genome instability.

Polymerase δ promotes chromosomal rearrangements and imprecise double-strand break repair.

Recent studies have implicated DNA polymerases θ (Pol θ) and ß (Pol ß) as mediators of alternative nonhomologous end-joining (Alt-NHEJ) events, including chromosomal translocations. Here we identify subunits of the replicative DNA polymerase δ (Pol δ) as promoters of Alt-NHEJ that results in more extensive intrachromosomal mutations at a single double-strand break (DSB) and more frequent translocations between two DSBs. Depletion of the Pol δ accessory subunit POLD2 destabilizes the complex, resulting in degradation of both POLD1 and POLD3 in human cells. POLD2 depletion markedly reduces the frequency of translocations with sequence modifications but does not affect the frequency of translocations with exact joins. Using separation-of-function mutants, we show that both the DNA synthesis and exonuclease activities of the POLD1 subunit contribute to translocations. As described in yeast and unlike Pol θ, Pol δ also promotes homology-directed repair. Codepletion of POLD2 with 53BP1 nearly eliminates translocations. POLD1 and POLD2 each colocalize with phosphorylated H2AX at ionizing radiation-induced DSBs but not with 53BP1. Codepletion of POLD2 with either ligase 3 (LIG3) or ligase 4 (LIG4) does not further reduce translocation frequency compared to POLD2 depletion alone. Together, these data support a model in which Pol δ promotes Alt-NHEJ in human cells at DSBs, including translocations.

APOBEC3A is a prominent cytidine deaminase in breast cancer.

APOBEC cytidine deaminases are the second-most prominent source of mutagenesis in sequenced tumors. Previous studies have proposed that APOBEC3B (A3B) is the major source of mutagenesis in breast cancer (BRCA). We show that APOBEC3A (A3A) is the only APOBEC whose expression correlates with APOBEC-induced mutation load and that A3A expression is responsible for cytidine deamination in multiple BRCA cell lines. Comparative analysis of A3A and A3B expression by qRT-PCR, RSEM-normalized RNA-seq, and unambiguous RNA-seq validated the use of RNA-seq to measure APOBEC expression, which indicates that A3A is the primary correlate with APOBEC-mutation load in primary BRCA tumors. We also demonstrate that A3A has >100-fold more cytidine deamination activity than A3B in the presence of cellular RNA, likely explaining why higher levels of A3B expression contributes less to mutagenesis in BRCA. Our findings identify A3A as a major source of cytidine deaminase activity in breast cancer cells and possibly a prominent contributor to the APOBEC mutation signature.

Recurrent Noncoding Mutations in Skin Cancers: UV Damage Susceptibility or Repair Inhibition as Primary Driver?

Somatic mutations arising in human skin cancers are heterogeneously distributed across the genome, meaning that certain genomic regions (e.g., heterochromatin or transcription factor binding sites) have much higher mutation densities than others. Regional variations in mutation rates are typically not a consequence of selection, as the vast majority of somatic mutations in skin cancers are passenger mutations that do not promote cell growth or transformation. Instead, variations in DNA repair activity, due to chromatin organization and transcription factor binding, have been proposed to be a primary driver of mutational heterogeneity in melanoma. However, as discussed in this review here, recent studies indicate that chromatin organization and transcription factor binding also significantly modulate the rate at which UV lesions form in DNA. The authors propose that local variations in lesion susceptibility may be an important driver of mutational hotspots in melanoma and other skin cancers, particularly at binding sites for ETS transcription factors.

Repair of base damage within break-induced replication intermediates promotes kataegis associated with chromosome rearrangements.

Break induced replication (BIR) is a double strand break repair pathway that can promote genetic instabilities similar to those observed in cancer. Instead of a replication fork, BIR is driven by a migration bubble where asynchronous synthesis between leading and lagging strands leads to accumulation of single-stranded DNA (ssDNA) that promotes mutation. However, the details of the mechanism of mutagenesis, including the identity of the participating proteins, remain unknown. Using yeast as a model, we demonstrate that mutagenic ssDNA is formed at multiple positions along the BIR track and that Pol ζ is responsible for the majority of both spontaneous and damage-induced base substitutions during BIR. We also report that BIR creates a potent substrate for APOBEC3A (A3A) cytidine deaminase that can promote formation of mutation clusters along the entire track of BIR. Finally, we demonstrate that uracil glycosylase initiates the bypass of DNA damage induced by A3A in the context of BIR without formation of base substitutions, but instead this pathway frequently leads to chromosomal rearrangements. Together, the expression of A3A during BIR in yeast recapitulates the main features of APOBEC-induced kataegis in human cancers, suggesting that BIR might represent an important source of these hyper-mutagenic events.

Nucleosome positions establish an extended mutation signature in melanoma.

Ultraviolet (UV) light-induced mutations are unevenly distributed across skin cancer genomes, but the molecular mechanisms responsible for this heterogeneity are not fully understood. Here, we assessed how nucleosome structure impacts the positions of UV-induced mutations in human melanomas. Analysis of mutation positions from cutaneous melanomas within strongly positioned nucleosomes revealed a striking ~10 base pair (bp) oscillation in mutation density with peaks occurring at dinucleotides facing away from the histone octamer. Additionally, higher mutation density at the nucleosome dyad generated an overarching "translational curvature" across the 147 bp of DNA that constitutes the nucleosome core particle. This periodicity and curvature cannot be explained by sequence biases in nucleosomal DNA. Instead, our genome-wide map of UV-induced cyclobutane pyrimidine dimers (CPDs) indicates that CPD formation is elevated at outward facing dinucleotides, mirroring the oscillation of mutation density within nucleosome-bound DNA. Nucleotide excision repair (NER) activity, as measured by XR-seq, inversely correlated with the curvature of mutation density associated with the translational setting of the nucleosome. While the 10 bp periodicity of mutations is maintained across nucleosomes regardless of chromatin state, histone modifications, and transcription levels, overall mutation density and curvature across the core particle increased with lower transcription levels. Our observations suggest structural conformations of DNA promote CPD formation at specific sites within nucleosomes, and steric hindrance progressively limits lesion repair towards the nucleosome dyad. Both mechanisms create a unique extended mutation signature within strongly positioned nucleosomes across the human genome.

Parp3 promotes long-range end joining in murine cells.

Chromosomal rearrangements, including translocations, are early and essential events in the formation of many tumors. Previous studies that defined the genetic requirements for rearrangement formation have identified differences between murine and human cells, most notably in the role of classic and alternative nonhomologous end-joining (NHEJ) factors. We reported that poly(ADP)ribose polymerase 3 (PARP3) promotes chromosomal rearrangements induced by endonucleases in multiple human cell types. We show here that in contrast to classic (c-NHEJ) factors, Parp3 also promotes rearrangements in murine cells, including translocations in murine embryonic stem cells (mESCs), class-switch recombination in primary B cells, and inversions in tail fibroblasts that generate Eml4-Alk fusions. In mESCs, Parp3-deficient cells had shorter deletion lengths at translocation junctions. This was corroborated using next-generation sequencing of Eml4-Alk junctions in tail fibroblasts and is consistent with a role for Parp3 in promoting the processing of DNA double-strand breaks. We confirmed a previous report that Parp1 also promotes rearrangement formation. In contrast with Parp3, rearrangement junctions in the absence of Parp1 had longer deletion lengths, suggesting that Parp1 may suppress double-strand break processing. Together, these data indicate that Parp3 and Parp1 promote rearrangements with distinct phenotypes.

Genome-wide maps of alkylation damage, repair, and mutagenesis in yeast reveal mechanisms of mutational heterogeneity.

DNA base damage is an important contributor to genome instability, but how the formation and repair of these lesions is affected by the genomic landscape and contributes to mutagenesis is unknown. Here, we describe genome-wide maps of DNA base damage, repair, and mutagenesis at single nucleotide resolution in yeast treated with the alkylating agent methyl methanesulfonate (MMS). Analysis of these maps revealed that base excision repair (BER) of alkylation damage is significantly modulated by chromatin, with faster repair in nucleosome-depleted regions, and slower repair and higher mutation density within strongly positioned nucleosomes. Both the translational and rotational settings of lesions within nucleosomes significantly influence BER efficiency; moreover, this effect is asymmetric relative to the nucleosome dyad axis and is regulated by histone modifications. Our data also indicate that MMS-induced mutations at adenine nucleotides are significantly enriched on the nontranscribed strand (NTS) of yeast genes, particularly in BER-deficient strains, due to higher damage formation on the NTS and transcription-coupled repair of the transcribed strand (TS). These findings reveal the influence of chromatin on repair and mutagenesis of base lesions on a genome-wide scale and suggest a novel mechanism for transcription-associated mutation asymmetry, which is frequently observed in human cancers.

Clustered mutations in yeast and in human cancers can arise from damaged long single-strand DNA regions.

Mutations are typically perceived as random, independent events. We describe here nonrandom clustered mutations in yeast and in human cancers. Genome sequencing of yeast grown under chronic alkylation damage identified mutation clusters that extend up to 200 kb. A predominance of "strand-coordinated" changes of either cytosines or guanines in the same strand, mutation patterns, and genetic controls indicated that simultaneous mutations were generated by base alkylation in abnormally long single-strand DNA (ssDNA) formed at double-strand breaks (DSBs) and replication forks. Significantly, we found mutation clusters with analogous features in sequenced human cancers. Strand-coordinated clusters of mutated cytosines or guanines often resided near chromosome rearrangement breakpoints and were highly enriched with a motif targeted by APOBEC family cytosine-deaminases, which strongly prefer ssDNA. These data indicate that hypermutation via multiple simultaneous changes in randomly formed ssDNA is a general phenomenon that may be an important mechanism producing rapid genetic variation.

Mutational heterogeneity in cancer and the search for new cancer-associated genes.

Major international projects are underway that are aimed at creating a comprehensive catalogue of all the genes responsible for the initiation and progression of cancer. These studies involve the sequencing of matched tumour-normal samples followed by mathematical analysis to identify those genes in which mutations occur more frequently than expected by random chance. Here we describe a fundamental problem with cancer genome studies: as the sample size increases, the list of putatively significant genes produced by current analytical methods burgeons into the hundreds. The list includes many implausible genes (such as those encoding olfactory receptors and the muscle protein titin), suggesting extensive false-positive findings that overshadow true driver events. We show that this problem stems largely from mutational heterogeneity and provide a novel analytical methodology, MutSigCV, for resolving the problem. We apply MutSigCV to exome sequences from 3,083 tumour-normal pairs and discover extraordinary variation in mutation frequency and spectrum within cancer types, which sheds light on mutational processes and disease aetiology, and in mutation frequency across the genome, which is strongly correlated with DNA replication timing and also with transcriptional activity. By incorporating mutational heterogeneity into the analyses, MutSigCV is able to eliminate most of the apparent artefactual findings and enable the identification of genes truly associated with cancer.

Avoidance of APOBEC3B-induced mutation by error-free lesion bypass.

APOBEC cytidine deaminases mutate cancer genomes by converting cytidines into uridines within ssDNA during replication. Although uracil DNA glycosylases limit APOBEC-induced mutation, it is unknown if subsequent base excision repair (BER) steps function on replication-associated ssDNA. Hence, we measured APOBEC3B-induced CAN1 mutation frequencies in yeast deficient in BER endonucleases or DNA damage tolerance proteins. Strains lacking Apn1, Apn2, Ntg1, Ntg2 or Rev3 displayed wild-type frequencies of APOBEC3B-induced canavanine resistance (CanR). However, strains without error-free lesion bypass proteins Ubc13, Mms2 and Mph1 displayed respective 4.9-, 2.8- and 7.8-fold higher frequency of APOBEC3B-induced CanR. These results indicate that mutations resulting from APOBEC activity are avoided by deoxyuridine conversion to abasic sites ahead of nascent lagging strand DNA synthesis and subsequent bypass by error-free template switching. We found this mechanism also functions during telomere re-synthesis, but with a diminished requirement for Ubc13. Interestingly, reduction of G to C substitutions in Ubc13-deficient strains uncovered a previously unknown role of Ubc13 in controlling the activity of the translesion synthesis polymerase, Rev1. Our results highlight a novel mechanism for error-free bypass of deoxyuridines generated within ssDNA and suggest that the APOBEC mutation signature observed in cancer genomes may under-represent the genomic damage these enzymes induce.

Secondary structure forming sequences drive SD-MMEJ repair of DNA double-strand breaks.

Alternative end-joining (alt-EJ) repair of DNA double-strand breaks is associated with deletions, chromosome translocations, and genome instability. Alt-EJ frequently uses annealing of microhomologous sequences to tether broken ends. When accessible pre-existing microhomologies do not exist, we have postulated that new microhomologies can be created via limited DNA synthesis at secondary-structure forming sequences. This model, called synthesis-dependent microhomology-mediated end joining (SD-MMEJ), predicts that differences between DNA sequences near double-strand breaks should alter repair outcomes in predictable ways. To test this hypothesis, we injected plasmids with sequence variations flanking an I-SceI endonuclease recognition site into I-SceI expressing Drosophila embryos and used Illumina amplicon sequencing to compare repair junctions. As predicted by the model, we found that small changes in sequences near the I-SceI site had major impacts on the spectrum of repair junctions. Bioinformatic analyses suggest that these repair differences arise from transiently forming loops and hairpins within 30 nucleotides of the break. We also obtained evidence for 'trans SD-MMEJ,' involving at least two consecutive rounds of microhomology annealing and synthesis across the break site. These results highlight the importance of sequence context for alt-EJ repair and have important implications for genome editing and genome evolution.