Fig 1.
POLQ-deficient animals are hypersensitive to EMS and UV/TMP.
A. Sensitivity to EMS exposure. B. Sensitivity to UV/TMP treatment. L4 animals of the indicated genotype were exposed to DNA damaging treatments and survival was quantified by counting dead embryos versus living progeny in the next generation. C-D. The total brood (eggs + larvae) was determined for P0 animals of the indicated genotype that were mock treated or treated with EMS (C) or UV/TMP (D). Lines represent the median for each dataset. E-F. The total brood was determined for F1 animals that originated from P0 animals that were either mock treated or treated with EMS (E) or UV/TMP (F). Lines represent the median for each dataset.
Fig 2.
EMS and UV/TMP-induced deletion alleles are dependent on POLQ.
A-B. Size distribution for all confirmed deletion events found in EMS (A) or UV/TMP (B) mutagenized libraries. Red bars represent the median deletion size. C. Fraction of populations that contained unc-93(e1500) revertant animals. At least 250 populations were assayed per experimental condition. D. Distribution of unc-93 reversion-footprints for the indicated genotype and experimental condition. The class of 50-1000bp was found to be statistically different between treated unc-93 and unc-93 polq-1 animals. The category ‘other’ includes wild-type sized PCR products, which based on previous experiments mostly reflect base substitutions. (p<0.001, Fisher’s exact test, indicated by ***) E. Size distribution of EMS- and UV/TMP-induced deletions generated by the C. elegans community. Only the deletions 50–1,000 bp (918 and 6,063 for EMS and UV/TMP-induced deletions, respectively) were used in subsequent analyses. F. Graphic representation of the two different types of deletions. The upper panel illustrates a simple deletion, in which only sequence is lost; the bottom panel reflects a delins, in which loss of sequence is accompanied with the insertion of de novo sequence. G-H. Pie chart representation of the fraction of deletions and delins that were isolated from EMS (G) and UV/TMP (H) mutagenized libraries. Deletions + SNV represent cases where a SNV is found in close proximity to a deletion.
Fig 3.
Replication approaches to one nucleotide from the damage.
A. Schematic illustration of the concept that one junction of DNA-damage-induced deletions is defined by the nascent strand blocked at sites of DNA damage. In this hypothesis, the replication-blocking lesion may dictate position -1, being the outermost nucleotide of the lost sequence. B-C. The base composition of all breakpoints, normalized to the relative AT/CG content around the breakpoints (from +100 to -100). Position +100 to +1 reflects the sequence that is retained in the deletion alleles; position -1 to -100 reflects the sequence that is lost. Dashed lines represent three times the SD. Data points outside these boundaries are marked with a dot.
Fig 4.
POLQ-mediated repair is characterized by single nucleotide homology.
A. Schematic illustration of a replication fork blocked at an UV/TMP-induced crosslink that subsequently leads to a DSB, which is repaired by POLQ leading to a deletion of the intervening sequence. One reactive end of the DSB is determined by the nascent strand blocked by an UV/TMP-induced crosslink that predominantly links thymines in opposite strands when in a 5’TA configuration. B. Deletion alleles that contain a 5’TA at the (+1, -1) position of one of their breakpoints are analysed (n = 1,248) for the base composition at the opposite breakpoint. Dashed lines represent three times the SD, which is determined by the base composition of the region between -100 and +100. C. Schematic illustration of how microhomology between breakpoints is determined in an unbiased manner. For each allele a table is constructed that allows for the scoring of homology between both breakpoints that give rise to a deletion. Each position of the upstream breakpoint (purple) is compared to each position of the downstream breakpoint (black). Identical nucleotides score 1, non-identical score 0. Subsequently, a heat map is constructed by summing all scores for all events at each position divided by the number of events. For reference purposes, a heat map was constructed for 7,000 deletions randomly created in silico throughout the genome. Of note, all alleles are annotated in keeping with maximal 5’ conservation, which here dictates that the base at the -1 position at the 5’ side is never identical to the +1 position at the 3’ side: in such a case, that base will shift to the +1 position at the 5’ side. As a consequence of this rule, the position marked by a cross will have no microhomology score, while the +1,-1 position is slightly elevated. The extent of this methodological skewing can be noticed in the analysis of the random set of deletions. D. Heat maps for UV/TMP- and EMS-induced deletions. Heat map contains 16 bases overlapping each breakpoint; 8 bases immediately flanking the deletion (light grey) and 8 bases immediately inside the deletion (dark grey).
Fig 5.
Hallmarks and genesis of delins.
A. Schematic illustration of the concept that templated insertions are generated by POLQ-mediated extension of one reactive 3’ end (e.g. the nascent strand blocked at sites of base damage) using the other end as a template: single nucleotide priming and disrupted extension can lead to delins formation. B. Size distribution of insertions found in EMS- and UV/TMP-derived delins. For 47–50% of delins the insert size is too small (<5 bp) to uniquely identify their origin. 37–44% of delins can be mapped to within 20 bp flanking the breakpoint. Another 2–3% of delins are copied from inter- or intrachromosomal (>1000 bp away from deletion) locations. For 6–10% of delins no apparent source could be identified. C. Schematic illustration for how microhomology is determined between the sequence that was used as a template for the generation of an insertion (the template) and the opposite breakpoint (the primer). A typical delins is portrayed at the sequence level as an example in which both the insertion (in blue) as its identified origin (in striped blue) is indicated. Underneath is another representation of the same delins, now containing the deleted sequence. This configuration is used in the subsequent analysis, where for each delins a table is constructed in which the bases overlapping with the 5’ side of the insertion origin (black) are compared to the bases that are overlapping the opposite breakpoint (purple). Identical nucleotides score 1, non-identical score 0. Subsequently, a heat map is constructed by summing all scores for all events divided by the number of events at each position. For reference purposes, a heat map was constructed for ~6,000 delins with perfect templated flank insertion randomly created in silico throughout the genome. Of note, at one position such a comparison cannot be done because the start and end nucleotide of an insertion is never identical to the deleted part of a delins and are thus always 0 (crossed out). As a result some other positions become slightly overrepresented as can be appreciated from the in silico generated delins. D. Heat map for UV/TMP- and EMS-induced delins for which the origin of the inserts are mapped. E. Visual representation of the origins of flank insertions for UV/TMP- and EMS-induced delins. A single line represents one mapped flank insertion and is drawn relative to its cognate breakpoint with ‘-’ for deleted and ‘+’ for retained sequences.
Fig 6.
Primer-template switching results in delins formation.
A. Schematic illustration of how primer template switching followed by POLQ-mediated extension and resolution results in a templated insertion. The requirement of single-nucleotide homology in POLQ-mediated end joining predicts that the nucleotide directly 3’ of the templated insertion (blue line) is typically identical to the outermost nucleotide of the ‘acceptor’ breakpoint. This prediction is highlighted by the red box. B. As in Fig 5C, but here for the end of the origin of templated insert and the adjacent deletion junction. As an example a typical delins is portrayed at the sequence level in which both the insertion (in blue) as its identified origin (in striped blue) is indicated. Underneath is another representation of the same delins, now containing the deleted sequence. This configuration is used in the subsequent analysis. Of note, at one position such a comparison cannot be done because the start and end nucleotide of an insertion is never identical to the deleted part of a delins and are thus always 0 (crossed out). As a result some other positions become slightly overrepresented as can be appreciated from the in silico generated delins. C. Heat map for UV/TMP and EMS-induced delins where the insertion origin could be faithfully traced back to the immediate flank.
Fig 7.
A. The fraction of templated flank insertions derived from a single origin is greatly increased when we allow a SNV or a slippage-event in a microsatellite (≥4 bp). B. The relative position of mismatches in delins is plotted for each mutagen relative to the insertion. C. Fraction of incorrect incorporated nucleotides in EMS and UV/TMP deletions, grouped by nucleotide misincorporation.
Fig 8.
Mutagen-induced deletions are the product of DSB repair.
A. Schematic illustration of a replication-blocking lesion that is converted to a DSB and finally results in a templated flank insertion. The 5’TA causing the deletion defines one end of the break, while the composition of the other end is unknown. By using the 5’TA together with the side of origin of templated insertions we can determine the reactivity of both 3’ break ends: if the 5’TA is on the opposite side of the templated insertion origin, repair initiated from the damaged side. On the other hand if both are on the same side then repair is initiated from the non-damaged side. B. Examples of two delins, portrayed at the sequence level, where either the 5’ side (left drawing) or the 3’ side (right drawing) potentially served as a primer to initiate repair. C. Analysis that probes the (+1,-1) junction of the side opposite to the flank containing the insertion origin. Dashes lines represent 3.5 times the SD. Only the largest and smallest variations for individual dinucleotides are shown. Only dinucleotide sets containing at least one position (marked by dots) that is >3.5 times the SD are shown in color. D. As in C, but in this case the (+1,-1) junction of the side that contains the insertion origin is analysed.