Precise mapping of single-stranded DNA breaks by using an engineered, error-prone DNA polymerase for sequence-templated erroneous end-labelling

doi:10.21203/rs.3.rs-5071189/v1

Download PDF

Article

Precise mapping of single-stranded DNA breaks by using an engineered, error-prone DNA polymerase for sequence-templated erroneous end-labelling

https://doi.org/10.21203/rs.3.rs-5071189/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The ability to analyze whether DNA includes lesions is important in identifying mitogenic substances. Until now, the detection of single-stranded DNA breaks (SSBs) has lacked precise methods. To overcome this limitation, we have engineered a chimeric DNA polymerase, Sloppymerase, that is able to replicate DNA in the absence of one nucleotide. In addition to polymerase activity, Sloppymerase demonstrates 5´-3´exonuclease activity. We characterized the activity of Sloppymerase and utilized the enzyme to develop a method for sequence-templated erroneous end-labelling sequencing (STEEL-seq) that is relevant to the mapping of SSBs. Following the omission of a specific nucleotide, e.g., dATP, from the reaction mixture, Sloppymerase introduces mismatches directly downstream of SSBs at positions that should contain deoxyadenosine. The ability to retain sequence information after end-labelling ensures that hits are bona fide SSBs. STEEL-seq works with a variety of sequencing technologies, shown by our successful experiments using Sanger, Illumina, PacBio and Nanopore systems.

Biological sciences/Molecular biology/DNA damage and repair/Single-strand DNA breaks

Biological sciences/Biological techniques/Molecular engineering/Protein design

Biological sciences/Biochemistry/DNA

Biological sciences/Biotechnology/Sequencing/DNA sequencing

Cellular DNA is constantly subjected to threats, both from the environment via exposure to radiation or chemicals and as a consequence of natural cellular processes, e.g., the production of endogenous reactive oxygen species (ROS) or DNA replication and transcription¹. In the worst case, these threats can cause DNA lesions, which can either be chemical modifications of nucleobases or strand breaks. Single-stranded DNA breaks (SSBs) are the most common type of damage to DNA, with a frequency of one SSB per cell every 1–10 s¹. Although this type of damage is not as threatening as double-stranded DNA breaks (DSBs), SSBs are transformed into DSBs when independent SSBs on opposite DNA strands are in close proximity. The repair of SSBs shows high fidelity, as the opposite DNA strand serves as the template for the missing nucleotide. In contrast, the repair of DSBs causes losses in genetic material and may result in chromosomal translocation^{2, 3}. In the case of damaged nucleobases, specific DNA excision repair pathways remove the damaged nucleotide and convert the lesion into an SSB ⁴. Although the repair of SSBs is highly efficient, this process involves a risk of introducing mutations. This has been exploited by B cells, which introduce somatic hypermutations during the affinity maturation of antibodies. Furthermore, translesion DNA polymerase η fills in the gaps in immunoglobulin genes, generated by the concerted action of the enzymes AID, UDG and APEI, in an error-prone manner⁵.

Methods that efficiently detect DNA damage, such as SSBs, DSBs, or damaged nucleobases are necessary for accurately determining the consequences of a damaging agent, as well as the subsequent extent of repair. Examples of current methods that provide a gross estimate of DNA fragmentation are the COMET assay⁶ and visualization of DNA damage in cells via the incorporation of fluorophore-labeled nucleotides⁷. We recently developed a method that combines different DNA polymerases to selectively label SSBs or DSBs⁸. Recent years have seen the development of sequencing-based assays that are based on end-labeling⁹, either via methylated nucleotides (RADAR-seq)¹⁰ or synthetic noncanonical deoxyribonucleotides (DENT-seq)¹¹. Moreover, several of the approaches that exist for the detection of DSBs, including BLESS or BLISS, are based on end-labeling¹². However, most of the current methods used to detect SSBs require extensive manipulation of the extracted DNA, including sonication and heat denaturation, which create secondary DNA breaks that do not represent what occurs in the cellular environment. Methods such as SSiNGLe¹³, Nick-seq¹⁴, and GLOE-seq¹⁵ label the 3´-OH end of an SSB by adaptor ligation or via terminal deoxynucleotidyl transferase (TdT). However, as free 3´-OH ends are present in both SSBs and DSBs, these approaches cannot exclusively identify SSBs¹⁶. Furthermore, no methodology can precisely determine whether multiple sequencing reads originate from different cells or represent multiple PCR amplificons of one molecule. Thus, there is currently a pressing need for methods that provide precise descriptions of SSBs, i.e., with a precision of a few nucleotides, to identify where these lesions are in the genome and whether the identified breaks are shared between cells (an indicator of vulnerable parts in the genome).

To address this need, we set out to develop a robust method for the detection of SSBs. We envisioned using a highly error-prone DNA polymerase to initiate DNA synthesis from a SSB when only three types of nucleotides are present, e.g., omission of dATP from the nucleotide mixture. Hence, this DNA polymerase would introduce mismatches directly downstream from any detected SSBs (Fig. 1). The inclusion of biotinylated nucleotides would enable researchers to extract any modified DNA regions. The DNA could subsequently be cut using any restriction enzyme, which would cut the DNA outside of the modified regions due to mismatches, to enable ligation of sequencing adapters. This method would facilitate the sequence-templated erroneous end-labelling sequencing (STEEL-seq) of SSBs and confer each individual molecule with a unique stretch of modified nucleotides. The sequence reads will contain sequence information downstream from the SSBs that will validate the presence of bona fide SSBs. The DNA polymerase required for this STEEL-seq method, i.e., the detection of SSBs, needs several features, namely, it must be highly error-prone, have the ability to extend from a mismatched base, tolerate modified nucleotides (e.g., those that are biotinylated), lack 3´-5´exonuclease activity (i.e., no proof-reading), yet demonstrate 5´-3´ exonuclease activity to remove the strand in front of the nick while synthesizing a new strand.

DNA polymerases in eukaryotes and prokaryotes can be divided into seven different families, based on sequence homology: A; B; C; D; X; Y; and RT¹⁷. Numerous specialized DNA polymerases are responsible for DNA replication and maintenance¹⁸. Despite these specialized roles, most DNA polymerases share a common structure and mechanism of action. For instance, these enzymes fold into a conformation that resembles the shape of a right-hand which comprises fingers, palm, and thumb subdomains^{19, 20}. Binding to DNA via the thumb subdomain elicits a structural change, upon which the thumb and fingers form a complete circle around the DNA strand. The palm subdomain of the polymerase contains the active site, which interacts with the template strand along the minor grove, whereas the fingers subdomain is essential for nucleotide recognition and binding¹⁹. DNA polymerases also share other common traits, such as the ability to only synthesize DNA in the 5’-3’ direction. Moreover, they require all four dNTPs, a DNA template, and a DNA primer from which to initiate synthesis, as well as divalent magnesium (Mg²⁺) as a cofactor²¹. There are, however, some exceptions to these requirements, such as terminal deoxynucleotidyl transferase (TdT), which does not require a template for DNA synthesis²². This special feature makes TdT a valuable tool in many biotechnological applications. Polymerases belonging to the A, B, C, D, and RT families are – with some exceptions – mostly specialized in replication and repair²⁰. Replicative polymerases are highly selective during dNTP incorporation and even demonstrate 3’-5’ exonuclease activity (proofreading), which checks the newly synthesized strand and excises any mismatches that may have occurred during replication. The high fidelity of DNA synthesis, when considered together with proofreading functions, means that most polymerases demonstrate very low error rates²³. A consequence of this elevated selectivity is decreased tolerance to DNA lesions, which stop replicative polymerases from bypassing the lesion, instead resulting in replication fork stalling and collapse^{23, 24}. Other polymerases possess a more open active site, which enables the replication of damaged DNA templates without stalling. This process, referred to as translesion synthesis (TLS), is facilitated by polymerases belonging to the Y family (which includes DNA polymerase η)^{20, 25}. In the case of Y-family polymerases, the DNA-binding surface is extended by a polymerase-associated domain²⁶, a little finger domain, while the selection of incoming dNTPs is guided by hydrogen-bond complementarity rather than shape ²⁷. X-family polymerases are specialized in small gap-filling repair synthesis during Base excision repair (BER) and non-homologous end joining (NHEJ)²⁸.

Several mutated DNA polymerases with reduced fidelity have been reported^{29, 30}, but we opted to try to develop an engineered chimeric DNA polymerase for even lower fidelity. We herein describe how we designed the DNA polymerase – Sloppymerase – and assessed the performance when one or more nucleotides were omitted from the dNTP mixture. We then used Sloppymerase to develop the STEEL-seq method and tested the performance using different sequencing technologies.

Design of Sloppymerase

To ensure the robust detection of SSBs, the DNA polymerase applied to the STEEL-seq method must have several features, namely, highly error-prone functioning, the ability to extend from a mismatched base, high tolerance to modified nucleotides (e.g., biotinylated nucleotides), a lack of 3´-5´exonuclease activity (i.e., no proof-reading), and 5´-3´exonuclease activity to remove the strand in front of the nick while synthesizing a new strand. This combination of features has not yet been found in any DNA polymerase that has evolved in the natural world; this is unsurprising, as such a DNA polymerase would not provide the faithful replication of organismal DNA. Hence, we set out to engineer a DNA polymerase that would demonstrate all of the aforementioned specifications.

DNA polymerases have evolved to demonstrate high fidelity so that the error-rate during replication is minimized. The disadvantage of this high fidelity is that DNA polymerases are unable to bypass damaged bases, e.g., cis-syn thymine dimers, abasic sites, and 7,8-dihydro-8-oxoguanine (8-oxoG), when replicating DNA. However, both pro- and eukaryotes have certain translesion DNA polymerases that can synthesize DNA in the opposite direction from such damaged bases. As a consequence of the relaxed requirement for the identification of nucleotides, translesion DNA polymerases show higher frequencies of base mis-incorporation when applied to undamaged DNA. For instance, the translesion DNA polymerase η has an base substitution error rate of 10^− 2 to 10^− 3. However, polymerase η demonstrates low processivity, incorporating only a few nucleotides before falling off the DNA strand; moreover, this polymerase does not have 5´-3´exonuclease activity³¹. Hence, to be applicable to our research, DNA polymerase η would need domains that possess 5´-3´exonuclease activity; E. coli DNA polymerase I demonstrates this type of activity, which is necessary for removing the RNA primer of Okazaki fragments during replication and filling gaps in the lagging strand. This DNA polymerase I contains three domains: the first is responsible for DNA polymerase activity; the second shows 3´-5´exonuclease activity for proof-reading; while the third has 5´-3´exonuclease activity for the removal of the RNA or DNA strand ahead of polymerization. DNA polymerase η from S. cerevisiae was codon-optimized for expression in E. coli and connected, via a 4xTGS spacer, at the 5´end to the 5´-3´exonuclease domain (nucleotides 1-969) of E. coli DNA polymerase I. This yielded our engineered chimeric DNA polymerase, called Sloppymerase, with the full sequence presented in Supplementary Fig. 1. Sloppymerase was equipped with a HIS-tag at the 5´end for subsequent purification and was expressed in E. coli under the control of an L-arabinose-inducible promotor. The purified Sloppymerase had a predicted Mw of 109 kDa and the identity was validated by mass spectrometry (Supplementary Fig. 2).

Characterization of Sloppymerase

To test the performance of Sloppymerase, we utilized a nicked DNA hairpin because it provides an easy approach to determine both 5´-3´exonuclease and polymerase activity via denaturing polyacrylamide gel electrophoresis (PAGE). A nicked DNA hairpin, when subjected to denaturing PAGE, will separate into two fragments: one small fragment consisting of the 3´end of the nicked hairpin and one large fragment consisting of the 5´end of the nicked hairpin. 5´-3´exonuclease activity can then be monitored based on the degradation of the small fragment, while polymerase activity can be monitored based on the increasing size of the large fragment. We first determined whether Sloppymerase can extend the hairpin by using all four dNTPs, after which we omitted dCTP from the dNTP mixture. The data confirm that Sloppymerase possesses both 5´-3´ exonuclease and polymerase activities, and is also able to polymerize DNA when dCTP is missing from the nucleotide mixture (Fig. 2A). The omission of dCTP does, however, noticeably decrease Sloppymerase speed. Next, we assayed how the removal of each of the four dNTPs affects the polymerase activity of Sloppymerase. The omission of dATP only slightly reduced the efficiency of Sloppymerase relative to what was observed when all four dNTPs were present. The removal of dCTP from the dNTP mixture further slowed down the speed of Sloppymerase, and the omission of either dTTP or dGTP resulted in the slowest Sloppymerase speed. Interestingly, Sloppymerase even works when only two types of nucleotides (dGTP and dTTP) are available, although the efficiency drops noticeably (Fig. 2B). The data clearly show that Sloppymerase can replicate DNA when one or more of the four required nucleotides are omitted. An illustration of the predicted folding of Sloppymerase, generated with AlphaFold ^{32, 33}, is presented in Fig. 2C.

We next investigated whether Sloppymerase can incorporate modified nucleotides when synthesizing DNA. The ability to incorporate biotinylated nucleotides is essential to pulling down Sloppymerase-modified DNA. To test this, the hairpins were extended, as described above, with the presence of biotin-dUTP in the dNTP mixture. The extended hairpins were then separated using PAGE and stained using Streptavidin-IRDye 800CW. The results showed that Sloppymerase can utilize biotinylated nucleotides when synthesizing DNA, as shown by the Streptavidin-IRDye 800CW results (Supplementary Fig. 3). Further, we determined that Sloppymerase can be used to label SSBs prior to DNA extraction, a characteristic which ensures that any detected SSBs were not generated as a consequence of DNA preparation. We used a dNTP mixture containing AF555-dUTP to stain fixed/permeabilized cells treated with the nickase Nt.BsmAI to generate SSBs at GTCTCN*N. In line with previous findings from DNA Polymerase I, Sloppymerase incorporated AF555-dUTP in fixed and permeabilized cells⁸ treated with Nt.BsmAI (Fig. 2D).

Sloppymerase generates unique molecular signatures

Having shown that Sloppymerase is able to modify the sequence downstream of SSBs, we next investigated whether the engineering polymerase demonstrates any bias in base substitutions. We therefore ligated an adaptor to the extended hairpin and performed PCR using primers specific for the loop of the hairpin and the adaptor. The PCR product was cloned into TOPO vectors and transformed into E. coli bacteria. Single colonies were collected and sequenced using Sanger sequencing. The extended hairpins that were synthesized with all four dNTPs present showed few alterations in the predicted sequence (Fig. 3A, top). In contrast, the extended hairpins that were synthesized when dATP was omitted from the dNTP mixture (Fig. 3A, left) shared that deoxyadenosine was mainly replaced with deoxyguanosine. In addition, several insertions and deletions were observed. We also performed Sanger sequencing of extended hairpins that had been synthesized when dCTP was omitted from the dNTP mixture (Fig. 3A, right). In this case, deoxycytidine was mainly replaced with deoxythymidine while several insertions and deletions were also observed. Hence, the data show that Sloppymerase introduces other nucleotides when the correct one is missing, with a clear bias to incorporate dTTP when dCTP is missing, along with the incorporation of dGTP when dATP is missing. In other words, purines are replaced with purines, while pyrimidines are replaced with pyrimidines, i.e., transitions rather than transversions. Sloppymerase also introduces deletions and insertions. As a high degree of variation in the substitutions, as well as the length of Sloppymerase-modified extended hairpins, was observed, we needed to determine whether our strategy produces unique sequence modifications downstream of SSBs. Hence, we performed Illumina (MiSeq Nano) sequencing to generate a large representation of the possible sequences generated. Sequence analysis of a 54-nt-long DNA molecule that had been modified by Sloppymerase in the absence of dATP generated 33 000 different sequences, with the vast majority (30 000) of these sequences only observed once. The sequences differ in both the complete length of the modified sequence, i.e., Sloppymerase has not progressed through the entire hairpin, base substitutions, and presence of insertions and deletions. Thus, the results provide strong evidence that Sloppymerase yields unique molecular signatures that can be used to determine whether products represent the same cell or not. (Fig. 3B).

Sequence-templated erroneous end-labelling sequencing (STEEL-seq)

Having confirmed that Sloppymerase can modify the DNA downstream of SSBs, we decided to sequence Sloppymerase-treated human DNA using PacBio (Sequel, HiFi reads), Nanopore (PromethION), and Illumina (NovaSeq 6000) sequencing to show that the Sequence-templated erroneous end-labelling sequencing (STEEL-seq) strategy (Fig. 1) is applicable to the major sequencing platforms. DNA collected from TK6 (for Illumina and PacBio) or HaCaT cells (for Nanopore) was treated with the Nickase Nt.BsmAI to generate SSBs at GTCTCN*N positions. The DNA was then subjected to Sloppymerase that utilized a dNTP mixture lacking dATP. For the samples that would undergo PacBio sequencing, the dNTP mixture was also spiked with biotinylated dUTP to allow the capture of Sloppymerase-modified DNA. Sequencing adaptors were attached to the Illumina sample by tagmentation³⁴. The PacBio sample was digested with a mixture of restriction enzymes (EcoRI, BamHI, NcoI and HindIII), followed by the ligation of PCR adaptors and pull down with magnetic streptavidin beads. After a PCR step, which would increase the amount of sample and overwrite deoxyuridine, sequencing adaptors were ligated. The sample that would be subjected to Nanopore sequencing was digested with PmlI and PmeI, followed by the ligation of sequencing adaptors. To find SSBs sites, we used a custom break-detection script that finds a minimum number of consecutive replaced deoxyadenosines as signature of Sloppymerase activity. We have herein used a treshold of five consecutive deoxyadenosine replacements to identify an SSB from PacBio reads and Nanopore reads, while three consecutive deoxyadenosine replacements were used to identify an SSB from Illumina reads. The SSB is, hence, located between the last remaining deoxyadenosine and the 5´-end of the first replaced deoxyadenosine. Examples of reads produced by the different sequencing technologies are presented in Fig. 4. The data show that unique Sloppymerase signatures are introduced downstream of Nt.BsmAI nicking sites. In Table 1, we compare results of detecting SSBs in Nickase-treated versus untreated WGS libraries to show that STEEL-seq works reliably on a larger scale. For both Nanopore and Illumina, the rate of predicted SSBs per million reads is roughly an order of magnitude higher in the two Nickase-treated libraries. Also, up to 73% of predicted SSBs coincide with nickase target sites. To determine how much overlap would be expected by chance, we generated an annotation track with the same number of nickase sites, but at random locations. This track results in only 5–7% of sites intersecting detected SSBs (not shown in table). We conclude that STEEL-seq can reliably detect SSBs with high specificity using any of the three tested sequencing technologies.

Table 1

Overview of the analyzed Sloppymerase-treated WGS libraries from all three sequencing technologies. Alignments are considered usable if they are primary (i.e., split alignments of long reads are exluded) and have at most 10% errors. SSBs were detected using the script and parameters as described in Methods. A detected break is considered to coincide with a Nt.BsmAI nickase site if its distance to the nickase motif is at most 10 bp (using bedtools -w 10).
Accession	SSBs induced by	Technology	Reads	Usable alignments	Predicted unique SSBs	SSBs per 1M reads	SSBs coinciding w/nickase site
ERR13611926	-	Nanopore	3,808,852	3,686,404	14,925	3918.5	7.6%
ERR13611927	nickase	Nanopore	4,033,986	3,700,481	117,950	29239.1	59.5%
ERR13611928	-	Nanopore	2,969,154	2,884,053	13,721	4621.2	8.4%
ERR13611929	nickase	Nanopore	3,544,808	3,266,454	104,262	29412.6	62.2%
ERR13611931	nickase	PacBio HiFi	388,822	216,324	8,131	20911.9	73.0%
ERR13611932	-	Illumina	72,745,911	4,455,285	21,413	294.4	5.9%
ERR13611933	-	Illumina	68,231,584	3,624,677	17,536	257.0	6.2%
ERR13611934	nickase	Illumina	61,827,843	3,761,922	40,097	648.5	49.7%
ERR13611935	nickase	Illumina	66,612,438	4,129,381	51,883	778.9	52.4%
ERR13611936	-	Illumina	65,807,144	4,416,310	22,074	335.4	6.6%
ERR13611937	-	Illumina	61,293,833	5,620,433	37,141	606.0	6.4%
ERR13611938	irradiation	Illumina	64,172,242	5,327,813	31,779	495.2	6.1%
ERR13611939	irradiation	Illumina	72,118,308	5,653,558	33,193	460.3	6.5%

We then investigated the presence of spontaneous SSBs in genomic DNA by using HaCaT cells labelled through STEEL-seq. The Nanopore sequencing results demonstrate an enrichment of SSBs in regions proximal to transcription start sites (TSS), a finding which agrees with what has been reported in previous publications using SSiNGLe to identify SSBs³⁵. The STEEL-seq data from Nanopore sequencing show that 11.8% of all detected SSBs were located at a promotor region within 1kb from TSS; these regions constitute 3.88% of the genome (Fig. 5). To confirm the result, we also performed STEEL-seq with Illumina sequencing. DNA was extracted from TK6 cells which had either received 10 Gy of irradiation or no irradiation. The results again revealed an enrichment of SSBs at promotor regions in non-irradiated cells, with these SSBs representing 11.76% of all SSBs; in contrast, the irradiated cells had a reduced fraction of SSBs at promotor regions (9.95% of all SSBs) (Fig. 5). The Illumina sequencing, when replicated, showed that 12.76% of all SSBs occurred at promotor regions in non-irradiated cells, while 10.31% of all SSBs occurred at promotor regions in irradiated cells (Supplementary Fig. 4). It is important to note that the irradiated cells contain both spontaneous SSBs, e.g., the large share observed in promotor regions, and additional irradiation-induced SSBs. The finding that irradiated cells demonstrate a lower share of total SSBs at promoter regions in comparison to non-irradiated cells could indicate that irradiation-induced SSBs are less targeted to promoter regions.

We do not yet have a complete understanding of exactly when and where SSBs form in the genome, which means that our understanding of how SSBs contribute to mutations and cancer is lacking. Studies of SSBs have – for several years – been limited by the lack of specific methods. However, recent years have been characterized by significant developments in several new methods for the precise detection of SSBs. Methods such as SSB-seq, in which the region downstream of a SSB is modified by the incorporation of biotinylated nucleotides, can pull down modified regions. However, this technique does not allow researchers to pinpoint exactly where in the DNA the break occurred. Methods such as SSiNGLe, Nick-seq, and GLOE-seq provide information on the positions of free 3´-OH ends, but cannot distinguish if these were the result of an SSB or DSB. The method presented in this paper, STEEL-seq, combines the advantages of the aforementioned methods as well as provides information, within a few bases, of where SSBs are located by labelling sequences that contains SSBs. By labeling SSBs via a template-specific DNA polymerase, instead of TdT, STEEL-seq provides sequence information both upstream and downstream of SSBs to enable the identification bona fide SSBs. This required the engineering of a highly error-prone DNA polymerase, which we have described – in detail – in this paper. Another advantage of the described approach is that the random insertion of mismatches will label each molecule with a unique molecular signature, which will allow researchers to determine whether multiple sequencing reads originate from different cells. The STEEL-seq approach also supports the labeling of SSBs in fixed, permeabilized cells prior to DNA extraction, which ensures that detected SSBs are not a consequence of DNA extraction or downstream handling. Although STEEL-seq cannot be used to exactly map where the SSB is located, it still provides very high resolution, i.e., within a few nucleotides of the SSB. We have validated that STEEL-seq is compatible with various sequencing technologies and can be used with different types of sample handling, including PCR, tagmentation, enzymatic digestion, ligation of adapters, and the pull down of biotinylated DNA fragments. We have herein used a threshold for PacBio and Nanopore of five consecutively replaced deoxyadenosines, while for Illumina used three. Increasing this threshold will also improve the confidence that the a detected SSB represents a bona fide SSB; nevertheless, it is possible to miss certain events where Sloppymerase has only replicated a shorter stretch. To ameliorate this shortcoming, increasing the time for Sloppymerase treatment or the concentration of Sloppymerase will increase the length of the modified stretch. The sequencing method can also easily be modified to determine the number of damaged bases (e.g., cis-syn thymine dimers or 8-oxoG). By utilizing enzymes that remove these structures, such as T4 pyrimidine DNA glycosylase (T4 PDG) or Formamidopyrimidine DNA Glycosylase (FPG), ssDNA breaks will be created. These ssDNA breaks can then be labeled with Sloppymerase. By omitting another nucleotide, e.g., dCTP, different labels will be produced that may allow researchers to simultaneously monitor multiple types of DNA damage.

We have herein shown that our engineered chimeric DNA polymerase, Sloppymerase, demonstrates an ability to replicate DNA when one, or even two, nucleotides are missing. As Sloppymerase is equipped with a domain that possesses 5´-3´ exonuclease activity, it will degrade the DNA strand in front of the DNA polymerase, which entails an ability to exchange nucleotides over a longer stretch. We have also shown that Sloppymerase can be used to selectively label SSBs. However, we envision several other applications for Sloppymerase that we will explore in the coming years. The purpose of translesion DNA polymerase η is to bypass damaged bases. The ability of Sloppymerase to also degrade the strand in front via 5´-3´exonuclease activity could be exploited to replace damaged bases in, for example, ancient DNA. Such applications would require all four nucleotides to be present so that Sloppymerase can exchange damaged nucleobases on the synthesized strand as well as when the template strand includes a damaged nucleobase. Another possible application of Sloppymerase would be the intentional introduction of mutations, as the error rates associated with Sloppymerase and translesion DNA polymerase η are orders of magnitude higher than conventional DNA polymerases. In this way, Sloppymerase could be used in a way that mimics the generation of somatic hypermutations observed in B cells via translesion DNA polymerase η activity. This ability could be utilized in both the in vitro and in vivo modification of DNA. As Sloppymerase modifies regions downstream of SSBs with a higher substitution rate that what would naturally occur, mutations will be directed to regions directly downstream of SSBs. It should be noted that SSBs naturally occur during replication, especially in the lagging strand, as well as during transcription. Furthermore, SSBs also play crucial roles in recombinant DNA and genome editing technologies; an example is CRISPR Cas9, where the RNA-guided DNA endonuclease Cas9 and the associated nickase variants can create site-specific nicks and DSBs to enable the removal or addition of genes^{11, 36}. As such, it may be advantageous to co-express Sloppymerase together with CRISPR Cas9 nickase in cells in order to direct mutagenesis to defined regions of the genome. These possible applications are only speculative at this point, but something that we will address in the coming years.

We have described the development of the highly error-prone DNA polymerase Sloppymerase and demonstrated how it can be utilized to selectively map SSBs. The presented STEEL-seq method offers valuable opportunities for the characterization of SSBs in a genome. The aim of the project was, effectively, to develop the worst DNA polymerase that has ever existed to enable the robust development of STEEL-seq. The properties of Sloppymerase may, in addition to STEEL-seq, be exploited in several types of biotechnological applications.

Acknowledgement

This study has been funded by grants from the Swedish Cancer Foundation (22 2306 Pj, OSö).

Illumina Sequencing was performed by the SNP&SEQ Technology Platform in Uppsala. The facility is part of the National Genomics Infrastructure (NGI) Sweden and Science for Life Laboratory. The SNP&SEQ Platform is also supported by the Swedish Research Council and the Knut and Alice Wallenberg Foundation. For PacBio and Nanopore sequencing the authors would like to acknowledge support of the National Genomics Infrastructure (NGI) /Uppsala Genome Center and UPPMAX for providing assistance in massive parallel sequencing and computational infrastructure. Work performed at NGI / Uppsala Genome Center has been funded by RFI / VR and Science for Life Laboratory, Sweden. Sequence analysis was supported by the SciLifeLab & Wallenberg Data Driven Life Science Program, Knut and Alice Wallenberg Foundation (grants: KAW 2020.0239 and KAW 2017.0003), and by the National Bioinformatics Infrastructure Sweden (NBIS) at SciLifeLab.

Conflict of interest

LW, JH & OSö are inventors on the patent application (WO Patent 2022/093091 A1) covering the design and use of Sloppymerase.

Author contribution

LW, JH, FAS, EJ, AD & BS performed experiments, the vast majority by LW. MM, BS & WS performed sequence analysis. YE, XC, OSp & OSö supervised experiments and/or sequence analysis. LW & OSö drafted the first version of the manuscript; all coauthors contributed to the final version of the manuscript. OSö conceived the project, designed Sloppymerase and coordinated all the work on the project.

Data and code availability

The raw sequencing reads used in this study are available from the European Nucleotide Archive (ENA) using accession PRJEB79373.

The break-detection software and a pipeline for replicating results from this paper are available at https://github.com/NBISweden/Sloppymerase/.

Production of Sloppymerase

To purify Sloppymerase, an overnight culture was created from BL21 E. coli (Thermo Fisher Scientific) that have been transformed by heat-shock with the Sloppymerase vector (VectorBuilder). The bacteria were inoculated with 4 ml lysogeny broth (LB) containing 100 μg/ml ampicillin (Merck) at 37°C with 225 rpm. The next day the overnight culture was transferred to a 1 L flask with 200 ml LB containing 100 μg/ml ampicillin. The culture was grown at 37°C with 225 rpm to an OD600 of 0.5. The expression of Sloppymerase was induced with 0.2% L-arabinose (Merck) and allowed to grow at 16°C overnight with 225 rpm. The next day the cells were harvested in a Sorwall centrifuge for 15 mins at 6000 x g at 4°C in a GSA rotor. After centrifugation the cells were lysed with 20 ml binding buffer (50 mM sodium phosphate, 500 mM NaCl, pH 7.4, 1 mM MgCl₂, 0.25 % Triton-X, 1x c0mplete protease inhibitor EDTA free (Merck), 0.2 mg/ml lysozyme (Thermo Fisher Scientific)) and incubated for 30 mins at 4°C. To clear the lysate, the samples were centrifuged for 15 mins at 13000 x g at 4°C and passed through a 0.45 mm Filtropur S syringe filter (Sarstedt). The His GraviTrap ™ Talon® (Cytivia) were equilibrated according to the producer’s manual before loading the samples. The columns were washed with 3 x 10 ml washing buffer (50 mM sodium phosphate, 500 mM NaCl, 5 mM imidazole, pH 7.4) and thereafter the samples were eluted with elution buffer (50 mM sodium phosphate, 500 mM NaCl, 50 mM imidazole pH 7.4). To exchange the buffer to the storage buffer (25mM Tris-HCl, 1 mM DL-Dithiothreitol (DTT), 0.2 mM EDTA (Merck), pH 7.4), the enzyme was concentrated down with Amicon® 10kDa Ultra Centrifugal filters (Merck) and reconstituted to the 2x sample volume with 2x storage buffer. This was repeated twice before the concentration was measured with nanodrop and glycerol (Merck) was added to a final concentration of 50%. To check the purity of the enzyme, samples from the purification were run on a NuPAGE Novex 4-12% Bis-Tris gel and thereafter stained with Coomassie Brilliant Blue G-250 Dye (Thermo Fisher Scientific) for 30 min. After destaining with water, the gel was scanned at 700 nm with an Odyssey® Fc imaging system (Li-Cor).

LC-MS Analysis

To confirm the amino acid sequence of the generated enzyme, a sample was prepared for LC-MS analysis. For filter-aided sample preparation, 20 µg of protein lysate were placed on the filter unit (Microcon-30 kDa; Merck, Darmstadt, Germany) and washed with a buffer containing 8 M urea and 100 mM Tris (pH 8.5). First 8 mM DTT were added followed by an incubation at 56 °C for 15 mins and then 50 mM of IAA. After 20 mins of incubating at room temperature, excess IAA was removed with 8 mM DTT (incubation at 56 °C for 15 min). After each incubation the sample was washed twice with Tris buffer. Finally, washing with NH₄HCO₃ was performed, trypsin was added [enzyme–protein ratio 1:50 (w/w)] and the samples were placed in a wet chamber at 37 °C. After incubation overnight, the resulting peptides were washed from the filter by adding 50 mM NH₄HCO₃ and centrifuging at 14,000g for 10 mins twice. Trifluoroacetic acid [final concentration of 1% (v/v)] was added, the samples were dried and reconstituted in a solution containing 3% acetonitrile and 0.1% formic acid in water to a final concentration of 150 ng protein/ µL.

For tryptic peptide analysis, a nanoAcquity UPLC system equipped with a C18, 5 μm, 180 μm × 20 mm trap column and a HSS-T3 C18 1.8 μm, 75 μm × 100 mm analytical column (Waters Corporation, Manchester, UK) was coupled to a Synapt G2 Si HDMS mass spectrometer with an electrospray ionization source (Waters Corporation, Manchester, UK). Mobile phase A contained 0.1% formic acid and 3% dimethyl sulfoxide in water and mobile phase B 0.1% formic acid and 3% dimethyl sulfoxide in acetonitrile. 300 ng of protein was injected in trapping mode. The peptides were separated at 40 °C with a gradient run from 3 to 40% (v/v) mobile phase B at a flow rate of 0.3 μl/mins over 120 min. Via the reference channel, a lock mass solution composed of [Glu1]-fibrinopeptide B (0.1 μM) and leu-enkephalin (1 μM) was introduced every 60 s. Peptide analysis was performed in positive ionization mode using the ultra-definition MSE (UDMSE) approach. The reproducibility and stability of the method were controlled with a commercially available HeLa digest (Thermo Scientific, Waltham, MA).

ProteinLynx Global Server (PLGS) (version 3.0.3, Waters Corporation, Milford, MA) was used for data processing. The samples were searched with a false discovery rate (FDR) of 0.01 against a randomized UniProt human database (Uni-ProtKB version 14/01/2020) with the addition of the sequence for the engineered enzyme. Search parameters were carbamidomethyl cysteine set as a fixed modification; acetyl lysine, C-terminal amidation, asparagine deamidation, glutamine deamidation, and methionine oxidation as variable modification; and trypsin as the digest reagent. One missed cleavage was allowed. Minimum peptide matches per protein were 2, and minimum ion matches per peptide and protein were 1 and 3, respectively.

Sloppymerase activity

To test the activity of the purified enzyme, an incomplete hairpin was used. Two oligonucleotides (hairpin A1 and A2, (Integrated DNA Technologies)) were ligated together by mixing 20 μM of each oligonucleotide and ligate the oligos for 48 hrs at 4°C end-over-end in 1x T4 DNA ligation buffer (50 mM Tris-HCl (pH 7.6), 10 mM MgCl2, 1 mM ATP, 1 mM DTT, 5% (w/v) polyethylene glycol-8000) and 0.1 U/μl T4 DNA ligase (Thermo Fisher Scientific). The reaction was heat inactivated at 65°C for 20 min. The incomplete hairpin was hybridized to a complementary oligonucleotide (hairpin A3) to create a nicked hairpin. 0.02 μM of the hairpin was then treated with Sloppymerase with either all four dNTPs: (0.1 mM dATP, 0.1 mM dCTP, 0.1 mM dGTP, 0.05 mM dTTP (Thermo Fisher Scientific) and 0.05 mM Biotin-11-dUTP (Jena Bioscience)) or omitting either dATP or dCTP in 1x Neb2.1 buffer (New England Biolabs) with 0.1 mM MnCl₂ (Merck) and 0.035 μg/μl Sloppymerase. The samples were incubated at 37°C for 60 minutes unless stated otherwise and then heat inactivated at 75°C for 20 min. The samples were mixed to a final concentration of 1x with Novex™ TBE-Urea Sample Buffer (2X) (Thermo Fisher Scientific) and heated up to 95°C for 5 mins before running the samples on Novex™ TBE-Urea gels, 10% (Thermo Fisher Scientific) at denaturing conditions to evaluate DNA polymerase and 5’-3’ exonuclease activity. The gel was stained with SYBR™ Gold nucleic acid gel stain (Thermo Fisher Scientific) and scanned at 600 nm with an Odyssey® Fc imaging system (Li-Cor). For evaluating if Sloppymerase can incorporate biotin-dUTP, the gel was additionally stained with IRDye® 800CW Streptavidin (926-32230. LI-COR) in a final concentration of 0.2 μg/ml before scanning the gel at 800 nm.

DNA oligonucleotides:

Hairpin A1	CCCAAACCCAATTAATGTACTGCAGAATTCAGCTCGAAGCTTGG CCGGATCCGTGAGCTGTCGTC
Hairpin A2	/5Phos/TCAGATCGGATACGGCGACCACCGAGATCTA CACCCTGCGGGACACTCTTTCCCTACACGACGCTCT TCCGATCTGAGACGACAGCTCAC
Hairpin A3	CCGGCCAAGCTTCGAGCTGAATTCTGCAGTACATTAATTGGGTTTGGG

SSB labelling in fixed cells

The SSB labelling was conducted as described in Bivehed et al. ¹. Briefly, HaCat cells were seeded in chamber slides, cultivated for 24 hours and thereafter fixed with ice-cold ethanol (70%) for 30 mins, followed by ethanol washes (96%-99.5%). To induce SSBs, the slides were subjected to 125 mU/μl Nt.BsmAI in 1x CutSmart buffer for 1 hour at 37°C followed by 2x5 mins washes with TBS. For SSB detection, the slides were incubated with 0,315 μg/μl Sloppymerase diluted in 1x NEB 2.1 buffer, 0.1mM MnCl₂, 0.1 mM dATP/dGTP/dCTP, 0.08 mM dTTP, 0.02 mM Aminoallyl-dUTP-XX-AF555 for 60 mins at 37 °C in a moisture chamber. Thereafter, the slides were subjected to a series of washes: 3 times quickly rinsed with H₂O, 2x5 mins with TBS-tween, 5 mins with 10 mg/ml Hoechst 33342 diluted in TBS-tween, and finally 2x5 mins with TBS. The slides were mounted with ProLong Glass antifade (Thermofisher Scientific) and cured overnight before being imaged. All experiments were conducted three independent times and at least 3 images were acquired per replicate.

Images were acquired with a Zeiss imager M2 microscope equipped with a Plan-Apochromat 63X/1.4 Oil objective, HXP 120 V light source, Hamamatsu C11440 camera and Zen 2 software (blue edition). Cube filter sets 43HE and 49 were used (all from Zeiss). Signal strength was equally enhanced for visualization purposes.

Sanger sequencing of Sloppymerase-treated Hairpin

For the sanger sequencing, 20 μM of hairpin S1, S2 and S3 were mixed together to create a nicked hairpin. The oligonucleotides were ligated overnight as described above. 0.02 μM of the hairpin was then treated with Sloppymerase with either all four dNTPs: (0.1 mM dATP, 0.1 mM dCTP, 0.1 mM dGTP and 0.1 mM dTTP or omitting either dATP or dCTP in 1x Neb2.1 buffer with 0.1 mM MnCl₂ and 0.035 μg/μl Sloppymerase. The samples were incubated at 37°C for 60 mins and thereafter heat inactivated for 20 mins at 75°C. To prepare the sequences for PCR, adapter S1 was ligated over night at 4°C with 1 mM ATP, 0.1 U/μl T4 DNA Ligase and 0.3 U/μl T4 Polynucleotide Kinase (PNK) (New England Biolabs).

For the PCR 0.04 μM of treated hairpin was added to 1x Phusion Green HF buffer (Thermo Fisher Scientific), 0.2 mM dNTPs (Thermo Fisher Scientific), 0.5 μM of forward and reverse primers S1 with 0.02 U/μl of Phusion U Hot Start DNA Polymerase (Thermo Fisher Scientific). The hairpin was amplified with a thermocycler for 7 seconds at 98°C, 20 seconds at 60°C and 20 seconds at 72°C for 20 cycles. The PCR product was cloned with Zero Blunt™ TOPO™ PCR Cloning Kit (Thermo Fisher Scientific) for sequencing according to the manufacturer’s manual. Another PCR was performed with the cloning products by dipping a pipette tip in a single colony and then in a tube with master mix consisting of 1x Platinum II HS buffer, 0.2 mM dNTPs, 0.04 U/μl Platinum II HS polymerase (Thermo Fisher Scientific) and 0.2 μM forward and reverse primers S2. The reaction was amplified with a thermocycler for 15 seconds at 94°C, 15 seconds at 60°C and 15 seconds at 68°C for 30 cycles.

The PCR products were cleaned up with 0.8 U/μl exonuclease I and 0.03 U/μl shrimp alkaline phosphatase (New England Biolabs) for 30 mins at 37°C and heat inactivated at 80°C for 20 min. The samples were then prepared for sequencing according to the platform’s instructions for TubeSeq Service (Eurofins Genomics). The samples were sequenced on ABI 3730XL sequencing machines using cycle sequencing technology (dideoxy chain termination/ cycle sequencing).

DNA oligonucleotides:

Hairpin S1	CCCAAACCCAATTAATGTACTGCAGAATT CAGCTCGAAGCTTGGCCGGATCCAGCGTGGGACTGAGTC
Hairpin S2	/5Phos/GTCTCGTGTCTGTAAAAACGTACGTAGATGCCATT TCTAAAAAAACAGACACGAGACGACTCAGTCCCACGCT
Hairpin S3	CCGGCCAAGCTTCGAGCTGAATTC TGCAGTACATTAATTGGGTTTGGG
Adapter S1	/5Phos/CGCACTGAGACTGATATGTGAAAAATTAGATT GGATAACTGCGCAGAAAAACACATATCAGTCTCAGTGCG
Fwd primer S1	CTGCGCAGTTATCCAATCTAA
Rev primer S1	CGTACGTAGATGCCATTTCTA
Fwd primer S2	GTAAAACGACGGCCAG
Rev primer S2	CAGGAAACAGCTATGAC

Illumina sequencing of Sloppymerase-treated Hairpin

For the experiment with Illumina sequencing, the hairpin was generated by ligating hairpin I1, I2 and I3.

The ligation and Sloppymerase treatment of the hairpin were performed as described above. For the PCR a different set of adapters was used as each adapter contained a unique barcode for each sample. For the omitting dATP sample adapter I1 was ligated to the Sloppymerase treated hairpin, and for the sample with all four dNTPs adapter I2 was used.

To open up the hairpins, the samples were treated with 1 unit of Uracil-DNA Glycosylase (UNG) and 2 units of endonuclease IV (Endo IV) (Thermo Fisher Scientific).

The PCR was again performed as before, but with only 10 cycles followed by a cleanup and buffer exchange to Tris-EDTA (TE) buffer pH 8.0 (Thermo Fisher Scientific) with Amicon Ultra-0.5 Centrifugal Filter Unit (Merck) according to the manufacturer’s protocol. The samples were then diluted and prepared according to the sequencing platform’s requirements and were sequenced with Illumina. SNP&SEQ Technology Platform performed 75 cycles paired-end sequencing in a MiSeq Nano v2 flowcell.

DNA oligonucleotides:

Hairpin I1	CCCAAACCCAATTAATGTACTGCAGAATTCAG CTCGAAGCTTGGCCGGATCCGTGAGCTGTCGTC
Hairpin I2	/5Phos/TCAGAUCGGAATGATACGGCGACCACC GAGATCTACACCCTGCGGGACACTCTTTCCCTA CACGACGCTCTTCCGATCTGAGACGACAGCTCAC
Hairpin I3	CCGGCCAAGCTTCGAGCTGAATTCTGCA GTACATTAATTGGGTTTGGG
Adapter I1	AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC AGACCAGCATCTCGTATGCCGTCTTCTGCTTG
Adapter I2	AGATCGGAAGAGCACACGTCTGAACTCCAGTC ACAAATACAGATCTCGTATGCCGTCTTCTGCTTG

STEEL-seq using Illumina sequencing

For Illumina sequencing, DNA has been extracted from TK-6 lymphoblast cells that have either been left untreated (WT) or irradiated (IR) to introduce DNA damage. Irradiation was performed by incubating 22.510⁶ TK6 cells, in a 15 ml Falcon tube on ice, with 10 Gy of 225 kV X-rays (X-RAD iR225, Precision X-Ray Inc., North Branford, CT, USA) at a dose-rate of 1.5 Gy/mins using an inherent Ba filter (0.8 mm) and an external Cu filter (0.3 mm). The DNA was extracted with DNeasy Blood & Tissue Kit (Qiagen) according to the manufacturer’s manual. 100 ng of the extracted WT DNA was additionally incubated with 20 U of Nt.BsmAI (New England Biolabs) and 1x Cutsmart buffer (New England Biolabs) to introduce nicks. After a buffer exchange to ddH₂O with Zeba™ Spin Desalting Columns, 7K MWCO, 0.5 ml (Thermo Fisher Scientific), the samples were incubated with 1x Neb2.1 buffer, a mixture of dNTPs (0.1 mM dCTP, 0.2 mM dGTP, 0.1 mM dTTP), 0.1 mM MnCl₂(Merck) and 0.035 μg/μl Sloppymerase for 120 mins at 37°C. Thereafter, the samples were heat inactivated for 20 mins at 75°C. For tagmentation,10 ng of purified DNA/sample was added to a transposase mixture with 12.5 µL 2x TD buffer, 1.25 µL Tn5 transposase (2 µM), produced and assembled according to Picelli et al.² and ddH₂O to bring the reaction volume up to 50 µL. The Tn5 adapters used were:I3, I4 and I5.

The transposase mixture was incubated at 55 ℃ for 7 minutes and purified using the QIAquick PCR cleanup kit (QIAGEN), following manufacturer’s protocol and eluted in 20 µL ddH₂O. The purified DNA was added to 25 µL NEBNext High-Fidelity 2x PCR Master Mix (New England Biolabs), 2.5 µL of forward primer: I1 (25 nM) and 2.5 µL of reverse primer: I1, I2 or I3 (25 nM). It was amplified with a thermocycler following 72 ℃ for 5 minutes, 98 ℃ for 30 seconds and thermocycling at 98℃ for 10 seconds, 63 ℃ for 30 seconds and 72 ℃ for 1 minute for 9 cycles. After amplification, the samples were purified using SPRIselect beads (Beckman) at a 1:1 ratio, following the manufacturer’s protocol for the Left workflow. The purified samples were analyzed with the 2200 TapeStation System (Agilent) and kept at -20 ℃ for sequencing. The samples were then sequenced with Illumina NovaSeq 6000. SNP&SEQ Technology Platform performed 150 cycles paired-end sequencing in one lane of a SP flowcell.

DNA oligonucleotides:

Adapter I3	TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG
Adapter I4	GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG
Adapter I5	CTGTCTCTTATACACATCT
Fwd primer I1	AATGATACGGCGACCACCGAGATCTACACCCTGCGGG TCGTCGGCAGCGTCAGATGTGTAT
Rev primer I1	CAAGCAGAAGACGGCATACGAGATC TGTATTTGTCTCGTGGGCTCGGAGATGTG
Rev primer I2	CAAGCAGAAGACGGCATACGAGATGAC CCAAGGTCTCGTGGGCTCGGAGATGTG
Rev primer I3	CAAGCAGAAGACGGCATACGAGATT TTAACGCGTCTCGTGGGCTCGGAGATGTG

STEEL-seq using PacBio sequencing

To test the functionality of the STEEL method with PacBio sequencing, extracted and purified DNA from TK6 lymphoblast cells was treated with Sloppymerase and the DNA was sequenced. The DNA was extracted with DNeasy Blood & Tissue Kit (Qiagen) according to the manufacturer’s manual. Before treatment 8,7 ug of purified DNA from untreated wild type TK-6 cells was incubated with 25 units of the nickase Nt.BsmAI for 45 mins at 37°C and then heat inactivated for 20 mins at 65°C to introduce nicks. The reaction was prepared in 1,5 ml eppendorf tubes with 1x Neb2.1 buffer (New England Biolabs), a mixture of dNTPs (0.1 mM dCTP, 0.2 mM dGTP, 0.05 mM dTTP) and 0.05 mM Biotin-11-dUTP, 0.1 mM MnCl₂and 0.035 μg/μl Sloppymerase. The DNA was added to the reaction at a concentration of 130 ng/μl. The sample was then incubated at 37°C for 90 mins followed by heat inactivation for 20 mins at 75°C. To fill any potential gaps, the sample was treated with 10 U Klenow fragment (3’-5’ exo-) (New England Biolabs) after adding 0.2 mM dATP, 0.1 mM dCTP and 0.15 mM dTTP to account for different concentrations of dNTPs. The sample was incubated for 30 mins at 37°C, followed by heat inactivation for 20 mins at 75°C

The sample was digested with a combination of EcoRI-HF, BamHI-HF, NcoI-HF and HindIII-HF (New England Biolabs) (10 U/enzyme) for 60 mins at 37°C followed by a buffer exchange to ligation buffer with Amicon Ultra-0.5 Centrifugal Filter Unit according to the manufacturer’s protocol. In the next step, a mixture of adapters (P1, P2, P3, P4, P5 and P6) with unique overhangs matching the different restriction sides and 30 U of T4 DNA ligase was added. The sample was then incubated overnight end over end at 4°C.

To extract the manipulated DNA fragments containing biotinylated dUTPs, a Dynabeads™ kilobaseBINDER™ Kit (Thermo Fisher Scientific) was used with a 3x higher concentration of beads than the manufacturer’s recommendation to ensure a high yield, otherwise the beads were prepared and washed according to the original protocol. The solution with biotinylated DNA was incubated with the beads end-over-end overnight at 4°C and an additional 2 hrs at RT the next morning. The bead solution was washed according to the manufacturer’s protocol before they were eluted in ddH₂O at 75°C for 5 min. The concentration of the recovered DNA was measured with nanodrop for PCR amplifciation. The PCR reaction was mixed according to the manufacturer’s protocol with 175 ng of DNA, 0.2 mM dNTPs, 0.5μM forward primer P1 and reverse primer P2, 1,5 mM MgCl₂ and 1,25 U Taq polymerase.

The DNA was amplified in a thermocycler for 20 cycles at 94°C for 45 seconds, at 51°C for 30 seconds and 2.5 mins at 72°C. The samples were run on a Novex™ TBE Gels, 10% (Thermo Fisher Scientific) for quality control followed by PCR clean up with MinElute PCR Purification Kit (Qiagen) according to the manufacturer’s protocol and eluted in nuclease free water. After purification, the samples were run on an agarose gel and the concentration was measured with nanodrop before they were sent to PacBio sequencing. The sequencing was performed by SNP & Seq technology platform using PacBio Revio system to generate 1 million HiFi reads.

DNA oligonucleotides:

Adapter P1	/5Phos/AATTGTTCCCTACACGGACTGAATACTCTGGCCGTCGTTTTAC
Adapter P2	/5Phos/GATCGTTCCCTACACGGACTGAATACTCTGGCCGTCGTTTTAC
Adapter P3	/5Phos/CATGCTTCCCTACACGGACTGAATACTCTGGCCGTCGTTTTAC
Adapter P4	/5Phos/AGCTGTTCCCTACACGGACTGAATACTCTGGCCGTCGTTTTAC
Adapter P5	/5Phos/CTTCCCTACACGGACTGAATACTCTGGCCGTCGTTTTAC
Adapter P6	CAGGAAACAGCTATGACAGTATTCAGTCCGTGTAGGGAAC
Fwd primer P1	GTAAAACGACGGCCAG
Rev primer P1	CAGGAAACAGCTATGAC

STEEL-seq using Nanopore sequencing

DNA extracted from HaCat keratinocyte cells was either left untreated, or nicked with 10 U of Nt.BsmAI for 60 mins at 37°C and thereafter heat inactivated for 20 mins at 65°C. The DNA was added to 1x Neb2.1 buffer, a mixture of dNTPs (0.2 mM dCTP, 0.2 mM dGTP, 0.15 mM dTTP) (Thermo Fisher Scientific) and 0.05 mM Biotin-11-dUTP (Jena Bioscience), 0.1 mM MnCl₂(Merck) and 0.035 μg/μl Sloppymerase for 60 mins at 37°C and heat inactivated for 20 mins at 75°C. To fill any remaining gaps, the samples were then treated with 10 U Klenow Fragment (3'→5' exo-), 60 U T4 DNA ligase and 0.1 mM dATP for 60 mins at 37°C, followed by heat inactivation for 20 mins at 75°C. The DNA was fragmentized with 20 U of PmeI and 40 U of PmlI (New England Biolabs) for 60 mins at 37°C with 1 volume of 1x cutsmart buffer and heat inactivated for 20 mins at 65°C. The samples were then sequenced with Oxford Nanopore. The sequencing was performed by SNP&SEQ technology platform using a PromethION system and two flow cells.

Data analysis

Sequence Analysis of Sloppymerase treated Hairpin

Comparison of sample sequences to the reference was performed using a custom Python script. Reads were filtered by matching to a primer sequence and the number of errors from the template. Sequence alignment was performed through the Needleman-Wunch global fit algorithm ³.

Alignment of STEEL-seq reads

GRCh38 was used as reference sequence. Illumina reads were aligned with strobealign ⁴ using default settings. Nanopore and PacBio HiFi reads were aligned with minimap2⁵ using -ax map-ont and -ax map-hifi, respectively.

Break detection

We developed a Python script that finds potential SSB sites in aligned reads from STEEL-seq libraries prepared without dATP in the reaction mix. The main signature of Sloppymerase activity in this setting is substitution of adenine with a different nucleotide – typically guanine, but we do not impose this as a restriction. The script thus starts by searching each aligned read for regions in which all adenines appear mutated (substituted or deleted). As also the reverse strand may have been sequenced, it also searches for regions where all thymines appear mutated. We call these regions events below. Supplementary alignments are ignored in the current version of the script. Used software libraries include pysam (https://github.com/pysam-developers/pysam) and pyfaidx (https://dx.doi.org/10.7287/peerj.preprints.970v1).

Filtering

Since sequencing errors can easily conspire to look like Sloppymerase activity, we apply the following filters to ensure we remove most false positives.

1) Aligned reads with an overall mutation rate (counting substitutions, insertions and deletions) above 10% are removed. Since both sequencing errors and mutations caused by Sloppymerase are counted, this threshold is set to a value well above the typical sequencing error rate.

2) An event is only kept if the number of mutated adenines (or thymines when on the reverse strand) is at least five for PacBio HiFi and Nanopore reads, three for Illumina reads.

3) The average base quality of substituted bases must be at least 10.

4) For Nanopore reads, at least three of the mutations in an event must be substitutions (i.e., not deletions).

Choice of which filters to implement and the values for filtering thresholds were informed by comparing the script output to a list of manually annotated ground truth events on a 1 Mbp region of a nickase-treated sample.

Output

The script computes the regions within which the SSB site must be located. These are between the last unmutated adenine upstream of the mutated region and its first mutated nucleotide for the case of mutated adenines, and between the last mutated nucleotide and the first unmutated thymine downstream of the region for the case of mutated thymines. The output is a BED track with extra annotations in a format thatIGV ⁶ understands and displays (including read name, mapping quality, base qualities). In a postprocessing step, the BED file is sorted using BEDTools ⁷ and duplicate entries are removed. Additional output is a BAM file with the subset of the input reads on which events were detected.

Single Strand Breaks Annotations

SSB hits from the Illumina sequencing were deduplicated based on the position. PacBio and NanoPore data where not deduplicated as they correspond to unique break events.
Non-canonical chromosomes were filtered out before annotation. For Illumina data, non-irradiated and non-treated cells data were used as references and, Sloppymerase treated samples were filtered using the references. The ChIPseeker, a R Bioconductor package was used to annotate genomic regions (promoter, 5’UTR, 3’UTR, exon, intergenic regions) and visualize forward and reverse strand breaks ⁸. Notebooks for data exploration, preprocessing and annotation are available on github (https://github.com/barslmn/sloppymerase-annotations).

References (Method section)

1. Bivehed, E. et al. Visualizing DNA single- and double-strand breaks in the Flash comet assay by DNA polymerase-assisted end-labelling. Nucleic Acids Res 52, e22 (2024).

2. Picelli, S. et al. Tn5 transposase and tagmentation procedures for massively scaled sequencing projects. Genome Res 24, 2033-2040 (2014).

3. Needleman, S.B. & Wunsch, C.D. A general method applicable to the search for similarities in the amino acid sequence of two proteins. J Mol Biol 48, 443-453 (1970).

4. Sahlin, K. Strobealign: flexible seed size enables ultra-fast and accurate read alignment. Genome Biol 23, 260 (2022).

5. Li, H. Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics 34, 3094-3100 (2018).

6. Robinson, J.T. et al. Integrative genomics viewer. Nat Biotechnol 29, 24-26 (2011).

7. Quinlan, A.R. & Hall, I.M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841-842 (2010).

8. Yu, G., Wang, L.G. & He, Q.Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics 31, 2382-2383 (2015).

Caldecott, K.W. DNA single-strand break repair and human genetic disease. Trends Cell Biol 32, 733-745 (2022).
Jackson, S.P. & Bartek, J. The DNA-damage response in human biology and disease. Nature 461, 1071-1078 (2009).
Blasiak, J. Single-strand annealing in cancer. Int J Mol Sci 22 (2021).
Caldecott, K.W. Single-strand break repair and genetic disease. Nat Rev Genet 9, 619-631 (2008).
Zeng, X. et al. DNA polymerase eta is an A-T mutator in somatic hypermutation of immunoglobulin variable genes. Nat Immunol 2, 537-541 (2001).
Ostling, O. & Johanson, K.J. Microelectrophoretic study of radiation-induced DNA damages in individual mammalian cells. Biochem Biophys Res Commun 123, 291-298 (1984).
Holton, N.W., Ebenstein, Y. & Gassman, N.R. Broad spectrum detection of DNA damage by Repair Assisted Damage Detection (RADD). DNA Repair 66-67, 42-49 (2018).
Bivehed, E. et al. Visualizing DNA single- and double-strand breaks in the Flash comet assay by DNA polymerase-assisted end-labelling. Nucleic Acids Res 52, e22 (2024).
Hu, J., Lieb, J.D., Sancar, A. & Adar, S. Cisplatin DNA damage and repair maps of the human genome at single-nucleotide resolution. Proc Natl Acad Sci U S A 113, 11507-11512 (2016).
Zatopek, K.M. et al. RADAR-seq: A RAre DAmage and Repair sequencing method for detecting DNA damage on a genome-wide scale. DNA Repair 80, 36-44 (2019).
Elacqua, J.J., Ranu, N., DiIorio, S.E. & Blainey, P.C. DENT-seq for genome-wide strand-specific identification of DNA single-strand break sites with single-nucleotide resolution. Genome Res 31, 75-87 (2021).
Mirzazadeh, R., Kallas, T., Bienko, M. & Crosetto, N. Genome-wide profiling of DNA double-strand breaks by the BLESS and BLISS methods. Methods Mol Biol 1672, 167-194 (2018).
Cao, H. et al. Novel approach reveals genomic landscapes of single-strand DNA breaks with nucleotide resolution in human cells. Nat Commun 10, 5799 (2019).
Cao, B. et al. Nick-seq for single-nucleotide resolution genomic maps of DNA modifications and damage. Nucleic Acids Res 48, 6715-6725 (2020).
Sriramachandran, A.M. et al. Genome-wide nucleotide-resolution mapping of DNA replication patterns, single-strand breaks, and lesions by GLOE-Seq. Mol Cell 78, 975-985 e977 (2020).
Zilio, N. & Ulrich, H.D. Exploring the SSBreakome: genome-wide mapping of DNA single-strand breaks by next-generation sequencing. FEBS J 288, 3948-3961 (2021).
Jain, R., Aggarwal, A.K. & Rechkoblit, O. Eukaryotic DNA polymerases. Curr Opin Struct Biol 53, 77-87 (2018).
Tang, K.H. & Tsai, M.D. Structure and function of 2:1 DNA polymerase complexes. J Cell Physiol 216, 315-320 (2008).
Rothwell, P.J. & Waksman, G. Structure and mechanism of DNA polymerases. Adv Protein Chem 71, 401-440 (2005).
Yang, W. & Gao, Y. Translesion and repair DNA polymerases: diverse structure and mechanism. Annu Rev Biochem 87, 239-261 (2018).
Hübscher, U., Spadari, S., Villani, G. & Maga, G. DNA Polymerases Discovery, Characterization and Functions in Cellular DNA Transactions. (World Scientific Publishing Co., 2010).
Motea, E.A. & Berdis, A.J. Terminal deoxynucleotidyl transferase: the story of a misguided DNA polymerase. Biochim Biophys Acta 1804, 1151-1166 (2010).
McCulloch, S.D. & Kunkel, T.A. The fidelity of DNA synthesis by eukaryotic replicative and translesion synthesis polymerases. Cell Res 18, 148-161 (2008).
Cruet-Hennequart, S. et al. DNA polymerase eta, a key protein in translesion synthesis in human cells. Subcell Biochem 50, 189-209 (2010).
Sutton, M.D. Coordinating DNA polymerase traffic during high and low fidelity synthesis. Biochim Biophys Acta 1804, 1167-1179 (2010).
Trincao, J. et al. Structure of the catalytic core of S. cerevisiae DNA polymerase eta: implications for translesion DNA synthesis. Mol Cell 8, 417-426 (2001).
Alt, A. et al. Bypass of DNA lesions generated during anticancer treatment with cisplatin by DNA polymerase eta. Science 318, 967-970 (2007).
Moon, A.F. et al. The X family portrait: Structural insights into biological functions of X family polymerases. DNA Repair 6, 1709-1725 (2007).
Minnick, D.T. et al. Side chains that influence fidelity at the polymerase active site of Escherichia coli DNA polymerase I (Klenow fragment). J Biol Chem 274, 3067-3075 (1999).
Hohlbein, J. et al. Conformational landscapes of DNA polymerase I and mutator derivatives establish fidelity checkpoints for nucleotide insertion. Nat Commun 4, 2131 (2013).
Washington, M.T., Johnson, R.E., Prakash, S. & Prakash, L. Fidelity and processivity of Saccharomyces cerevisiae DNA polymerase eta. J Biol Chem 274, 36835-36838 (1999).
Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583-589 (2021).
Varadi, M. et al. AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Res 50, D439-D444 (2022).
Picelli, S. et al. Tn5 transposase and tagmentation procedures for massively scaled sequencing projects. Genome Res 24, 2033-2040 (2014).
Cao, H. et al. Hotspots of single-strand DNA "breakome" are enriched at transcriptional start sites of genes. Front Mol Biosci 9, 895795 (2022).
Jinek, M. et al. A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science 337, 816-821 (2012).

Yes there is potential Competing Interest. LW, JH & OSö are inventors on the patent application (WO Patent 2022/093091 A1) covering the design and use of Sloppymerase.

Supplementary.pdf

Download PDF

Version 1

posted

You are reading this latest preprint version

Precise mapping of single-stranded DNA breaks by using an engineered, error-prone DNA polymerase for sequence-templated erroneous end-labelling

Status:

Version 1

Abstract

Figures

Introduction

Results

Design of Sloppymerase

Characterization of Sloppymerase

Sloppymerase generates unique molecular signatures

Sequence-templated erroneous end-labelling sequencing (STEEL-seq)

Discussion

Declarations

Methods

Production of Sloppymerase

LC-MS Analysis

Sloppymerase activity

References

Additional Declarations

Supplementary Files

Status:

Version 1