The processes of transcription and replication require the separation of complementary DNA strands to allow the corresponding machineries to access a single-strand template. This leads to the over- or under-winding of strands and the accumulation of supercoils that can impede or impact subsequent nucleic acid transactions if not appropriately resolved. Topoisomerases relax supercoiled DNA through a strand cleavage-ligation mechanism that entails a transient covalent DNA-enzyme intermediate [1]. Type I topoisomerases nick one DNA strand to allow controlled rotation around the intact strand, while type II enzymes nick both strands to create a double-strand break (DSB) through which an intact duplex can pass. Although covalent cleavage intermediates are rapidly and efficiently reversed to restore DNA integrity, local DNA lesions and topoisomerase poison drugs can impede the religation step and thereby lead to toxic DSBs and genetic instability [2]. The current study focuses on the eukaryal type IB topoisomerases (Top1), which use an active site tyrosine to attack the phosphodiester backbone, giving rise to a nick flanked by 3’-linked Top1 (3’-phosphotyrosyl bond) and a 5’-OH.

The mutagenic potential of Top1 was discovered in Saccharomyces cerevisiae when examining the consequences of high levels of transcription on genome stability [3], [4]. In a forward mutation assay, yeast Top1 (yTop1) produces a distinctive deletion signature at hotspots that coincide with short tandem repeats. The incorporation of ribonucleotides into DNA and their persistence generates a similar signature [5], which is dependent on Top1 [6]. Vaccinia is a prototype poxvirus and early biochemical studies with its Top1 (vTop1) demonstrated that incision at a DNA-embedded ribonucleotide leads to vTop1-catalyzed attack of the vicinal ribose-O2’ on the DNA 3’-phosphotyrosyl bond to generate a 2’,3’-cyclic phosphate end [7]. These biochemical data, along with genetic analyses, led to a sequential cleavage model for ribonucleotide- and Top1-dependent mutations (Fig. 1A; [8]). Following Top1 cleavage at a ribonucleotide and its subsequent release, Top1 cleaves the same strand upstream of the nick left by the first cleavage. Release of the resulting short oligonucleotide traps the Top1 cleavage complex adjacent to a small gap, with the 5’-OH created by the first incision being used for subsequent ligation and release of Top1. Although Top1-mediated ligation occurs across gaps in vitro [9], [10], [11], the location of deletions at short repeats in vivo suggested a model in which repeat-mediated realignment of complementary strands is required to convert the gap to a more efficiently ligated nick. This sequential reaction has been confirmed in vitro using yTop1 as well as the human enzyme [12], [13]. Recent studies have shown that a similar mutation signature in human cancers is ribonucleotide driven and reflects a Top1-dependent mechanism [14]. Although most Top1-dependent deletions in yeast involve incision at an embedded ribonucleotide, we previously identified a deletion hotspot that appears to be ribonucleotide independent [8].

The position of yTop1 or human Top1 (hTop1) cleavage is difficult to predict, with the most conserved feature being incision at a thymidine (Tp↓N) [15]. We thus have relied on deletion hotspots identified in yeast forward-mutation assays to infer positions of yTop1 cleavage. To facilitate genetic analyses of the deletion process, individual hotspots have been incorporated into LYS2-based frameshift-reversion assays [3]. We previously used a ribonucleotide-dependent, 5-bp deletion hotspot [(CCCTT)2] identified in the URA3 forward mutation assay [5] to characterize yTop1-dependent mutagenesis in parallel in vitro and in vivo analyses. Manipulation of the hotspot provided strong support for the sequential-cleavage model of ribonucleotide-initiated deletion by demonstrating that the distance between yTop1-generated nicks in vitro corresponds to the deletion size in vivo [16].

In contrast to the unpredictable cleavage sites of the yeast and human proteins, the consensus sequence for vTop1 cleavage is the pentameric motif (T/C)CCTTp↓ [17], [18]. In the current study, substrates that mimic a ribonucleotide-generated intermediate were used to demonstrate robust vTop1-generated deletions at CCCTT tandem repeats in vitro. vTop1 expression in a top1Δ yeast background similarly elevated 5-bp deletions at CCCTT repeats but, in stark contrast to yTop1-dependent events, vTop1-initiated events were not affected by the level of ribonucleotides in DNA. In addition to yTop1-generated deletions, we previously showed that genetic or chemical stabilization of the cleavage intermediate gives rise to large deletions that are dependent on the nonhomologous end-joining (NHEJ) pathway of DSB repair [19]. We suggested that the NHEJ-dependent deletions reflect the joining of two yTop1-initiated DSBs. To model this event using the unique sequence specificity of vTop1, we created a reporter with CCCTT cleavage sites separated by 42 bp. Although the predicted 47-bp deletion was observed, it was NHEJ independent. Additional experiments support a model in which vTop1 can directly mediate the joining of independent DSBs initiated by the enzyme.