Versatile cell-based assay for measuring DNA alkylation damage and its repair

DNA alkylation damage induced by environmental carcinogens, chemotherapy drugs, or endogenous metabolites plays a central role in mutagenesis, carcinogenesis, and cancer therapy. Base excision repair (BER) is a conserved, front line DNA repair pathway that removes alkylation damage from DNA. The capacity of BER to repair DNA alkylation varies markedly between different cell types and tissues, which correlates with cancer risk and cellular responses to alkylation chemotherapy. The ability to measure cellular rates of alkylation damage repair by the BER pathway is critically important for better understanding of the fundamental processes involved in carcinogenesis, and also to advance development of new therapeutic strategies. Methods for assessing the rates of alkylation damage and repair, especially in human cells, are limited, prone to significant variability due to the unstable nature of some of the alkyl adducts, and often rely on indirect measurements of BER activity. Here, we report a highly reproducible and quantitative, cell-based assay, named alk-BER (alkylation Base Excision Repair) for measuring rates of BER following alkylation DNA damage. The alk-BER assay involves specific detection of methyl DNA adducts (7-methyl guanine and 3-methyl adenine) directly in genomic DNA. The assay has been developed and adapted to measure the activity of BER in fungal model systems and human cell lines. Considering the specificity and conserved nature of BER enzymes, the assay can be adapted to virtually any type of cultured cells. Alk-BER offers a cost efficient and reliable method that can effectively complement existing approaches to advance integrative research on mechanisms of alkylation DNA damage and repair.


Scientific Reports
| (2021) 11:18393 | https://doi.org/10.1038/s41598-021-97523-w www.nature.com/scientificreports/ new, conserved regulators of the BER pathway by using complementary model eukaryotic systems, including fungal model organisms and human cells. Application of the alk-BER assay could also facilitate identification of BER-deficient cancer sub-types, which might represent suitable targets for therapy with alkylating agents and/ or PARP inhibitors.

DNA damage and time course of repair in yeast cells. Yeast (S. cerevisiae) liquid cell cultures were
inoculated from single colonies and grown in 10 ml of YPD (Yeast extract -Peptone-Dextrose) medium for ~ 16 h at 30 °C in an orbital shaker. The following day, cells were sub-cultured in fresh media and grown until the cultures reached the logarithmic stage of growth, as determined by measuring the optical density of the cell culture (e.g., OD 600 ~ 0.6). Next, MMS was added to the liquid cultures at a final concentration of 20 mM and cells were incubated for 10 min at 30 °C in an orbital shaker. Alternatively, cells were treated with 3.5 mM MMS for 1-3 h, followed by removal of media containing MMS and repair time course in fresh media for 1-6 h. Cells were then harvested by centrifugation, and the supernatant fractions containing MMS were removed and disposed following DEHS guidelines. Cell pellets were washed with ice-cold phosphate buffered saline (1X PBS), re-suspended in a pre-warmed YPD media, and allowed to repair for a total of 3 h. Extended repair time points (longer than 4-5 h) were avoided to ensure that restoration of the genome integrity was due to the activity of BER, and not due to lesion bypass and DNA replication. Cultures were incubated with continuous shaking and cell aliquots were collected at different repair time points 49,50 .
Yeast genomic DNA isolation. Yeast genomic DNA was extracted with the glass bead method following previously established protocols 51  Alkaline agarose gel electrophoresis. Alkaline agarose gel electrophoresis was performed following previously published protocols with modifications 48,53 . The large gel electrophoresis box and casting tray were used. The agarose solution was prepared by adding 4.3 g of agarose to 360 ml of H 2 0 in a 1 L Erlenmeyer flask and then heating in a microwave oven until the agarose was dissolved. The solution was cooled to 55 °C and then followed by addition of a 0.1 volume (40 ml) of 10X alkaline agarose gel electrophoresis buffer (500 mM NaOH, 10 mM EDTA, pH 8.0). Addition of 10X alkaline buffer to a hot agarose solution should be avoided because NaOH in the buffer may cause hydrolysis of the agar. The agarose solution was poured into a large gel casting tray. After the gel was completely solidified, it was mounted in the electrophoresis tank. Then the tank was filled with 1X alkaline electrophoresis buffer until the gel was covered with the buffer at a depth of 3-5 mm above the gel. DNA samples after AAG and APE1 digestion were collected and 6X alkaline gel-loading buffer (300 mM NaOH, 6 49,51 . Briefly, using the ImageQuant software functions, each data point on the gel image is marked with a box encompassing the entire length of the lane to give the total area of each lane. The data point corresponding to 1/2 the total area, designated as X med , is then determined. The X med value indicates the median migration distance of the DNA fragments. X med is converted to the median length L med of DNA molecules by using a standard curve generated from the migration of DNA-size markers. The number average molecular length, L n , is calculated from L med by using the equation L n = 0.6 L med 54 , assuming a Poisson distribution of DNA fragments. Numbers of SSBs/kb is calculated using the following equation; SSBs/kb = 1/L n (+ AAG + APE1) − 1/L n (-AAG-APE1). Calculated numbers of SSBs per kb DNA fragment indicate numbers of MDAs.
DNA damage and time course of repair in multicellular fungus Neurospora crassa. Liquid cultures of N. crassa were grown for 11 h in VMM (Vogel's minimal medium) media in an orbital shaker at 30 °C. MMS was added to liquid cultures at a final concentration of 3.5 mM. Cells were incubated in the presence of MMS for 3 h. Next, the cells were collected by using a Buchnner funnel, and washed with 500 ml of VMM media to remove MMS. Washed cells were transferred to pre-warmed VMM media and then allowed to grow at 30 °C for 4 h to enable repair of damaged DNA. Aliquots of cells were harvested and immediately frozen in liquid nitrogen. Genomic DNA was isolated and 300 ng of genomic DNA was digested with APE1 endonuclease (cat # M0282S, New England Biolabs), AAG glycosylase (cat # M0313S, New England Biolabs), or both enzymes in MOPS reaction buffer (70 mM MOPS, pH 7.5,1 mM dithiothreitol DTT, 1 mM EDTA, 5% glycerol) for 1 h and 15 min at 37 °C. Reactions were stopped by adding alkaline DNA loading buffer (50 mM NaOH, 1 mM EDTA, 3% Ficoll). Samples were resolved on a 1.2% alkaline agarose gel. Agarose gels were run in a cold room at 25 V for 17 h and then incubated in neutralization buffer (1.5 M NaCl, 1 M Tris-Cl pH 7.6) for 45 min before being stained with SYBR Gold (cat # S-11494, Life Technologies) for 40 min and de-stained for 30 min before imaging 55 .

DNA damage and time course of repair in human cells. Human adrenal carcinoma cells (SW13)
were a gift from Dr. Feng Gong laboratory at the University of Miami. Human fibroblast CHON-002 were a gift from Dr. Rafael Contreras laboratory at the Hormel Institute, University of Minnesota. Human haploid HAP1 cells were purchased from Horizon Discovery Biosciences. SW13 cells were cultured in DMEM (Sigma, cat# D5796) supplemented with 10% fetal bovine serum FBS (Gibco, cat# 26140-079). Human fibroblast CHON-002, leukemia HAP1, and lymphoblastoid cells GM12878 were cultured according the ATCC guidelines. Cell cultures were routinely tested for mycoplasma by using a mycoplasma detection kit (ATCC, cat# 30-1012 K). Cell doubling time was determined following ATCC guidelines and viability was routinely monitored with trypan blue. Cells were seeded in T25 flasks (~ 500,000 cells per dish) and grown for 16-24 h until cells reached 60-70% confluence. One T25 flask was set for each data point to be collected. In DNA damage dose response experiments SW13 cells were treated with (0-20 mM) MMS in 1XPBS for 10 min at RT (room temperature). In DNA damage and repair experiments, SW13 cells were treated with 10 mM (0.1%) MMS for 10 min in 1xPBS at RT, or alternatively ice-cold, serum-free media can be also used for MMS treatments. Ice-cold treatment is used to inhibit endogenous background BER during the MMS treatment. MMS was removed and cells were washed with 1X PBS. Next, fresh pre-warmed media were added and cells were allowed to repair for 0, 3, 8, or 22 h. GM 12878 cells were treated with 5 mM (0.05%) MMS for 5 min. Total genomic DNA was purified from each time point using the PureLink genomic DNA mini kit (K 182001, Thermo Fisher Scientific). DNA was processed with AAG and APE1 enzymes as described above. DNA was resolved on a 1.2% alkaline agarose gel, run at 30 V for 22 h, stained with SYBR gold, and quantified as described above.

Results
Alk-BER assay workflow. The conserved nature of the BER pathway enables easy adaptation of the assay to various cell types and model organisms from lower eukaryotes to human cells. We initially developed the alk-BER assay to measure BER in yeast cells (S. cerevisiae) and successfully adapted the assay to other fungal model organism (N. crassa) and human cells. The alk-BER assay is based on the enzymatic conversion of the MMS-induced damaged DNA bases to single strand breaks (SSBs), which are subsequently resolved on alkaline agarose gels and quantified. The rate of BER can be analyzed by performing a DNA damage and time course for repair and assessing the rate of removal of damaged bases from the total genomic DNA. The assay is performed in 5 simple steps (Fig. 1) and can be completed in 3 days. The first step involves exposing the cells to sub-lethal doses of MMS to induce formation of methyl DNA adducts (mainly 7meG and 3meA). In the second step, cell aliquots corresponding with DNA damage and repair time points are collected and total genomic DNA is isolated. The third step involves conversion of MDAs to SSBs with BER enzymes (AAG glycosylase and APE1 endonuclease) that specifically bind and cleave the DNA at sites of MDAs. The fourth step involves running the www.nature.com/scientificreports/ samples on alkaline agarose gels to separate DNA fragments containing SSBs from the bulk genomic DNA that does not contain damage. The final step involves staining of the separated DNA fragments in the gel, acquiring an image of the gel and performing quantitation of MDAs. The alk-BER method directly quantifies the numbers of MDAs in purified genomic DNA and analyzes the kinetics of DNA repair at the whole-genome level. In order to assess the capacity of BER using the alk-BER assay, cells should be exposed to sub-lethal doses of MMS, that induce detectable levels of MDAs, but do not induce substantial cell death. To determine sub-lethal doses for specific cell line it is recommended to perform MMS dose response experiment encompassing range of MMS doses from low to high. Cell death is induced when unrepaired lesions persist due to inability of BER to efficiently repair abnormally high levels of DNA damage or when BER activity is compromised (e.g., BER gene mutants). BER capacity should also be analyzed within a specific, experimentally determined window of time following DNA damage, typically 0-3 h for yeast cells, or 0-24 h for human cells, to avoid interference from lesion bypass and DNA replication.

Alk-BER in fungal cells.
A DNA damage dose response assay was performed by exposing yeast cells (strain BY4741) to increasing doses of MMS (5, 10, 20 mM) for 10 min at 30 °C. Total genomic DNA was isolated and treated with AAG and APE1 enzymes. DNA samples were resolved by alkaline agarose gel electrophoresis and numbers of MDAs were quantified as described above. As the gel is run at alkaline pH, hydrogen bonding between the two DNA strands is broken to facilitate separation of strands containing breaks from non-damaged genomic DNA. The denatured DNA is maintained in a single-stranded state and migrates through the alkaline gel as a function of its size, forming a distinct smear. Increased formation of MMS-induced DNA lesions in response to increased doses of MMS is shown as increases in lower molecular weight DNA molecules, also visible as a smear (Fig. 2A). The frequency of MDAs (7meG and 3meA) was calculated and plotted as the number of methyl A, G per kb fragment as a function of increasing doses of MMS (Fig. 2B). The proportional relationship between increasing MMS doses and numbers of MDAs indicates a high sensitivity of the alk-BER method, ~ 1.0 MDAs per 10,000 bases, that is induced by 5 mM MMS dose, and ~ 4.0 MDAs per 10,000 bases induced by 20 mM MMS in yeast cells (Fig. 2B). www.nature.com/scientificreports/ DNA damage and time course of repair has been performed to evaluate the overall rate of BER to repair alkylation DNA damage in the yeast wild type strain BY4741. DNA molecules containing MMS-induced MDAs are converted to SSBs. Proficient DNA repair and restoration of genome integrity can be visually monitored as progressive shortening of the DNA smear in migration and restoration of the genome integrity by formation of high molecular weight DNA. DNA was processed with AAG and APE1 and resolved on alkaline agarose gels as described previously. A representative gel image demonstrating DNA damage and repair in the BY4741 yeast strain is shown (Fig. 2C) and corresponding quantitative analysis of the gel is also shown (Fig. 2D). BER in the WT BY4741 yeast strain is proficient, and over 80% of the total MDAs in the genome are repaired after 3 h repair time at the dose used (Fig. 2D). The specificity of the alk-BER method was validated using yeast and Neurospora mutant cells deficient in BER. The yeast mag1∆ mutant has no Mag1 glycosylase (orthologue of human AAG) and is deficient in cleaving MMS-induced 7meG and 3meA from the DNA. Cells from the WT and mag1∆ strains were exposed to MMS for a total of 3 h followed by MMS removal and a 6 h-long repair time course to allow cells to repair damaged DNA. During the 3 h-long MMS exposure, mutant cells displayed higher levels of MDA formation as compared to WT because endogenous BER is inactive in the mutant cells, which results in accumulation of MDAs under the conditions of continuous MMS exposure. After removal of MMS, WT cells with proficient BER were able to clear most of the lesions during the repair time course, whereas BER-deficient mag1∆ cells contained high levels of unrepaired MDAs (Fig. 3A-C). The alk-BER assay was successfully adapted to multicellular fungal model Neurospora crassa. DNA damage and repair time course was performed with Neurospora WT and mag1∆ cells 55 . Neurospora cells of the wild type laboratory strain were exposed to 3.5 mM MMS continuously for 3 h, followed by repair time course in media without MMS for 5 h. Genomic DNA was isolated and processed with combinations of AAG and APE1 enzymes and resolved on alkaline agarose gel. BER We found that the cocktail of both enzymes AAG and APE1 works most efficient in converting MDAs to SSBs (Fig. 4C). The BER capacity in SW13 cells was analyzed by performing DNA damage and a time course of repair as described in the methods section. Cells were exposed to 10 mM MMS for 10 min in 1X PBS at RT, followed by removal of MMS and repair for 22 h in fresh media. The MMS dose used was a sublethal dose, that did not trigger significant cell death, as demonstrated by the cell viability data (Fig. 4F). After 22 h post MMS exposure nearly 70% of the genome was restored in SW13 cells. Representative data showing the image of the alkaline agarose gel and data quantitation are presented (Fig. 4D,E). Other human cell lines were subject to alk-BER assay, including untransformed fibroblast cells CHON-002, and leukemia cancer cells HAP1. The BER capacity to remove MMS-induced adducts was quantitated over distinct repair time points. The removal of MDAs appears very slow (~ 0-30% repair) during the first 0-8 h post MMS exposure and is consistently observed in many different human cell lines we tested. Interestingly, the repair rates vary significantly between different cell lines, and unlike in fungal cells, the rates do not appear to correlate well with the levels of AAG enzyme in the panel of cell lines we tested (Fig. 5A,B). Additionally, alk-BER assay was performed with human lymphoblastoid cell line GM12878. Cells were exposed to 5 mM MMS for 5 min, and repair rate was analyzed at 2.5 and 5 h in the repair time in media without MMS. Clearly, very slow repair was detected during the initial 5 h repair (Fig. S3). Specificity of alk-BER was analyzed by exposing SW13 cells to increasing doses of oxidative agent hydrogen peroxide (H2O2), and temozolomide (TMZ), clinically used SN1 alkylator, known to induce 7meG, 3meA and O6meG adducts 57 . Exposure to H2O2 does not result in formation of a dose-dependent DNA smear, indicating that alk-BER assay is specific to methyl adducts, and it cannot detect oxidative DNA damage (Fig. S1A). As expected, alk-BER can specifically detect TMZ-induced DNA methyl adducts in a dose dependent manner (Fig. S1B,C), although the TMZ potency to induce MDAs appears lower than the potency of MMS.

Discussion
The efficiency of BER in repairing alkylation DNA damage varies substantially between different cells, tissues, and individuals, and has important implications in cancer development and treatment 1,2,58,59 . BER efficiency is a result of a complex interplay between genetic and epigenetic factors influencing the abundance and activity of   www.nature.com/scientificreports/ the BER enzymes, and the individual steps in the BER pathway. Ability to measure the formation and rates of repair of alkyl DNA adducts in genomic DNA provides a direct assessment of BER efficiency in a given cell type.
Here, we report a quantitative cell-based assay, alk-BER (alkylation Base Excision Repair), developed for measuring efficiency and rates of BER following alkylation DNA damage in fungal model organisms (S. cerevisiae, N. crassa) and human cell lines (SW13, CHON-002, HAP1 and GM12878). Alk-BER offers a simple, time-and cost-efficient, cell-based method for quantitative analysis of alkylation DNA damage and repair in the genomic DNA. The alk-BER assay can be used to determine BER efficiency in various cell types, by assessment of the rates of methyl DNA adducts removal from the DNA over time.
Alk-BER has been developed as an adaptation and extension of the previous methods pioneered by Sutherland et al., to quantify UV-induced DNA damage and repair 48 . The original methods were based on quantification of DNA migration in ethidium bromide-stained alkaline agarose gels and scanning the image on film negatives with a densitometer. Alk-BER method reported here relies on the similar principles of quantifying DNA damage from alkaline agarose gels, but alk-BER assay has been developed and optimized specifically for the analysis of DNA alkylation repair. In addition, alk-BER involves several technological advances that improve method sensitivity and overcome many of the limitations of quantifying DNA from ethidium bromide fluorescence, including the linear range of detection. For example, we use SYBR Gold fluorescent dye to stain DNA in the gel. SYBR gold is about tenfold more sensitive than ethidium bromide 60 . SYBR gold also has increased signal to background ratio, as demonstrated by more than 1000-fold signal enhancement when bound to nucleic acids, which is approximately 50 times greater than that of ethidium bromide 60 . Furthermore, in our method gel images are acquired with modern, highly sensitive imaging systems (e.g., ChemiDoc MP, Bio-Rad) that allow for significant improvements of the linear dynamic range of detection. The improved linear range of detection permits accurate quantification of the DNA from a gel, where there is large variability in DNA band intensity (e.g. low intensity when bands disappear right after damage, and high intensity when bands reappear during repair) 51 . In summary, we adapted the principles of quantifying DNA damage from alkaline gel electrophoresis to analyze repair of DNA alkylation, and extended the method to analyze repair by BER at the genome level in various model organisms, including fungal cells (S. cerevisiae, N. crassa) and human cancer cell lines.
Yeast cells have been extensively used to study DNA repair processes in eukaryotic cells 61 . Yeast have a robust and conserved BER pathway to repair alkylation DNA damage. Using alk-BER assay, we found that in wild type strains of fungal model organisms (S. cerevisiae and N. crassa), BER proceeds quite rapidly, following removal of the MMS from the growth media. We showed that BER in the BY4741 strain of S. cerevisiae is nearly completed within 3-4 h post exposure since nearly 80-90% of the lesions were removed from the genome (Fig. 2C,D). Similarly, repair of alkylation DNA damage has been nearly completed within 2-4 h in the wild type laboratory strain of N. crassa 55 (Fig. S2). We also showed that BER-deficient mag1∆ yeast mutant cells were deficient to clear MMS-induced lesions and accumulated MDAs over time, which validates specificity of the alk-BER assay ( Fig. 3A-C). The human genome is much larger therefore repair is expected to take longer as compared to lower eukaryotes having smaller genomes. We found that the majority of the genome (~ 70% lesions removed) was restored following 22 h post MMS exposure in SW13 cells (Figs. 4D,E, 5A), which is consistent with previous studies reporting that majority of DNA alkylation repair in mammalian cells can be completed within 24 h post exposure, as revealed by results generated with various quantification methods 39,62-66 . We found that the rates of BER to remove MMS-induced DNA adducts vary significantly between SW13 and CHON-002, and HAP1 cell lines and do not correlate well with the endogenous levels of AAG enzyme in these cell lines (Fig. 5A,B). These rates can be influenced by the endogenous levels and activities of various BER enzymes, including additional glycosylases, and other regulators of DNA repair. Future studies are needed to further investigate the mechanisms and regulators of DNA alkylation repair in human cells.
We found that in lower eukaryotes (S. cerevisiae, N. crassa), the rate of repair of MMS-induced methyl DNA adducts is strongly dependent on the functional MAG1 glycosylase, where mag1∆ mutants demonstrate abolished ability to repair MDAs. Interestingly, repair of MDAs in certain mammalian cells does not appear to be exclusively dependent on AAG glycosylase. It has been reported that the alkylated bases 3meA and 7meG, both AAG substrates generated from MMS treatment, are removed from the genome of AAG-deficient embryonic stem cells, with slower kinetics for 3meA but comparable kinetics for 7meG 67 . Other study revealed that similar levels of 7meG were detected in livers of AAG + / + and AAG − / − mice 24 h after exposure to MNU 68 . These studies suggest that in mammalian cells methyl DNA adducts can be excised and repaired in the absence of AAG enzyme, perhaps by involvement of other glycosylases, or spontaneous depurination. Future application of alk-BER could facilitate further understanding of the role of AAG and other factors in regulation of human BER.
The rates of repair of DNA alkylation can vary significantly between different organisms and cell types, and can be influenced by the levels and activities of the BER enzymes, genome size, chromosome landscape, DNA replication and translesion DNA synthesis (TSL). Therefore, it is necessary to experimentally determine a time range for accurate analysis of BER activity for each cell type under study. This is accomplished by carefully monitoring DNA damage induction and repair over a repair time course. The time range for BER to repair alkylation lesions is also dependent on the cell doubling time, which is much shorter (~ 90 min) in yeast cells, and longer (~ 16-24 h) in human cancer cell lines. It should be noted that exposure to DNA damage triggers transient cell cycle arrest as part of DNA damage response, that allows the time for DNA repair to take place before the next round of cell division 69 . The activity of BER should be analyzed immediately after induction of DNA damage, and within the 3-4 h for S. cerevisiae and Neurospora, and 24-48 h in human cell lines, when majority of the lesions are expected to be repaired by BER. After majority of lesions have been repaired, the DNA replication and cell cycle progression will resume. A small proportion of DNA base lesions, those which are left unrepaired, can be bypassed by translesion DNA synthesis (TSL) that occurs during DNA replication. TSL involves specialized replicative polymerases, which can perform error-free DNA synthesis over DNA base lesions 70  www.nature.com/scientificreports/ order to avoid possible interference from DNA replication and lesion bypass, BER activity should be analyzed at early repair time points and within experimentally determined cell-specific time rage. The alk-BER could serve as useful framework for number of approaches to study repair of DNA alkylation. For example, alk-BER assay could be used to distinguish, in a quantitative way, between the levels of MDAs and levels of downstream repair intermediates, such as AP sites. Highly specific and sensitive detection of AP sites could also be performed by processing of the sample with the AAG enzyme only (converts MDA to AP sites) and subsequent detection of AP sites using a highly sensitive AP site detection kit (e.g., Abcam, ab 211154). Alk-BER assay can also serve as a framework for quantification of gene-specific repair when coupled with Southern blot and hybridization of gene-specific probes. Alk-BER could also be useful in detection and quantification of MMS-induced methyl DNA adducts in preparation and optimization of samples for the approaches involving next generation sequencing, such as NMP-seq.
In summary, the alk-BER assay offers a versatile, reliable and affordable approach for quantitative analysis of DNA damage formation and repair following exposure to DNA methylating alkylating agents. The alk-BER assay can be easily optimized to be used in any type of cultured cells, and integrated with the existing approaches to study mechanisms regulating BER balance and capacity. The assay has the ability to detect imbalances in the activity of the BER process, that is highly relevant to cancer development and treatment. Quantitative analyses of DNA alkylation damage and repair using fungal genetic model organisms and human cell lines offer unique opportunities to identify novel, conserved regulators of BER.