Targets for repair: detecting and quantifying DNA damage with fluorescence-based methodologies

Abstract

Detection and characterization of DNA damage is essential for evaluating genotoxicity, monitoring DNA repair, developing biomarkers for exposures, and evaluating the efficacy of chemotherapies. These diverse applications for DNA damage measurements have spurred the continual development and refinement of methodologies for detecting, characterizing, and quantifying DNA damage from isolated DNA and in cells and tissues. Current damage detection methods cover a wide range of techniques from radiolabeling to mass spectrometry, and use of these techniques varies widely based on expense, expertise, and knowledge of adduct formation. More generalizable, easy-to-use methods for detecting and quantifying DNA damage are needed, and there has been an emergence of fluorescence-based methodologies to address this need. Developments in these fluorescence-based strategies are reviewed here.

Graphical Abstract.

DNA adducts are induced endogenously via the production of reactive metabolic byproducts and during cellular propagation, and exogenously through environmental exposures to DNA damaging chemicals and radiation. The methods that require DNA purification are labeled in vitro, and methods capable of detecting adducts within a cellular context are identified as in situ.

Introduction

DNA damage is a natural consequence of daily life. Endogenous metabolism and exposures to environmental, dietary, and therapeutic chemicals, create modified DNA bases and strand breaks that if left unrepaired can produce mutations, chromosomal aberrations, and induce cell death. Detection, characterization, and quantification of these modifications, or adducts, and strand breaks provides important information about genotoxic exposures, their mechanistic actions, and cellular responses. These data are critical for risk assessment and for gauging the efficiency of DNA repair pathways, the effectiveness of chemopreventives, or the efficacy of chemotherapies [1–3].

Given these diverse applications, a large number of methods for the detection and quantification of DNA adducts have been developed. These methods range from the highly specific detection and quantification of DNA adducts using liquid chromatography and mass spectrometry (reviewed in [1,2,4]) to more broad spectrum detection of total DNA damage through polymerase chain reactions (PCR) [5,6].

An emerging tool in DNA damage detection are fluorescence-based strategies for detecting, characterizing, and quantifying DNA damage. Fluorescence methods offer sensitive and specific detection of damage and avoid the use of hazardous radiolabeled compounds. They are compatible with a large number of read-outs from single molecule imaging, which allows the heterogeneity of a system to be analyzed, to high content imaging with flow cytometry or microplate readers, which allows population level differences to be measured. These methods also allow spectrally distinct labeling strategies that may be multiplexed, increasing the potential throughput of experiments, and they are capable of measuring damage on isolated DNA and within cells and tissues.

The advantages of fluorescence-based methods are not new to DNA damage detection [1,3]. Antibody-based detection of damage in cells and tissues has been conducted for decades, typically using fluorophore-conjugated primary or secondary antibodies [1,7]. DNA intercalating dyes have been used to detect DNA strand breaks in single cell gel electrophoresis (SGCE) [8–10] or chromosomal instability by monitoring supercoiled DNA (the “halo” assay) [11] or formation of micronuclei [12]. However, new fluorescence-based DNA damage detection methodologies for use in vitro and in situ are emerging that provide sensitive and specific detection of damage, are compatible with single molecule and whole cell imaging, and can be coupled with genomic sequence analysis to better understand the damage context. Here, we will review these techniques and recent advances in these methodologies.

Molecular markers for in vivo detection of DNA damage

Antibodies against DNA adducts, like thymine dimers or 8-hydroxy-2’-deoxyguanosine (8-oxodG), or DNA repair proteins like Ku, XRCC1, or phosphorylated H2AX (γH2AX) have all been used to characterize DNA damage within isolated DNA and in cells [13–16]. However, all of these techniques rely on the generation of highly specific, high affinity antibodies for targets of interest. Given the natural affinity and specificity of DNA repair proteins for damage sites, a more direct detection strategy would be to use specific DNA repair or DNA binding proteins to directly characterize induced damage (Fig. 1).

Figure 1.

Overview of fluorescent reporters for DNA damage in vivo. Cells are transfected with a fluorescent reporter bound to a double strand break (DSB) binding protein, Gam or 53BP-1. Endogenous DSBs or exogenously induced DSBs can be detected via foci formation at the DSB site.

For example, the Gam protein from bacteriophage Mu recognizes double strand breaks (DBSs) generated during integration and protects the linear DNA from degradation [17,18]. This specific recognition of DNA ends led Shee et al. to fuse Gam to green fluorescent protein (Gam-GFP) and monitor the induction of double strand breaks in E. coli and mammalian systems [19]. In E. coli, specific detection of DSBs generated during replication, by I-SceI cleavage, and from genotoxic exposure was achieved with an estimated detection efficiency of 71–82% [19]. In mammalian systems, expressed Gam-GFP has been shown to detect laser and radiation-induced DSBs as well as failed mitotic events and shown to co-localized with phosphorylated H2AX (γH2AX) and 53BP-1 [19,20]. However, the detection efficiency in the mammalian systems was not as high. Competition with endogenous Ku proteins, to which Gam is related, likely contributes to this loss in detection efficiency [17,19]. Despite some limitations for mammalian systems, Gam-GFP provides specific and sensitive detection of in vivo DSBs and has been used to characterize DNA repair dynamics in E. coli [21,22]. Additionally, the competition with Ku in mammalian systems may provide another DNA repair linked application for this damage reporter, where Ku’s recruitment and role in nonhomologous end-joining (NHEJ) is dissected.

Similarly, another DSB reporter has also been developed for in vivo single-cell imaging based on truncating 53BP-1 (amino acids 1220–1711) and fusing it to the mApple fluorescent protein [23]. 53BP-1 recognizes DSBs and promotes NHEJ-mediated repair [24]. The truncated 53BP-1 was shown to accumulate at DSBs in a number of cell lines including breast, ovarian and neuronal cells, allowing study of the dynamics of strand break formation and repair after genotoxic exposures [23,25,26].

While these methodologies are still developing, engineered proteins may provide new reagents for the detection of strand breaks or DNA adducts in vivo. More work is needed to generate reagents compatible with single strand breaks (SSBs) and other DNA adducts, but in vivo imaging of induced damage would provide new insight into the genotoxicity of environmental and therapeutic chemicals and allow dynamics of DNA repair to be visualized.

Enzyme-mediated damage detection assays

Engineered protein reporters exploit the natural activities of DNA repair proteins, similarly a number of techniques have emerged utilizing DNA repair enzymes to excise and specifically label DNA damage sites [27–31]. In 2013, Lee et al. demonstrated that molecular combing of λ DNA molecules could be used to detect strand breaks induced by ultraviolet light (UV) [28]. Stretched DNA molecules were irradiated and visualized by fluorescent staining with the DNA intercalator, YOYO-1. Gaps in the YOYO-1 staining revealed DSBs in the DNA. Further damage characterization was carried out by coupling enzymatic processing by pyrimidine dimer glycosylase (PDG) with the incorporation of fluorescently-labeled dNTPs by DNA polymerase I to characterize the presence of single strand breaks (SSBs) and UV-induced base lesions [28].

Zirkin et al. refined and extended this technique to detect abasic sites, uracils, oxidative, and UV-induced DNA lesions from DNA isolated from treated cells or human tissues. Isolated DNA was treated prior to combing with an enzymatic cocktail, PreCR®, which contains bacterial DNA glycosylases, EndoIV, EndoVIII, Fpg, T4 PDG, UDG, and Bst DNA polymerase [32]. DNA damage is recognized and excised by these enzymes and damage sites are detected by the incorporation of fluorescently-labeled nucleotides by the DNA polymerase. Modified versions of this assay, with different enzyme cocktails, have been utilized to detect oxidative DNA damage in viral genomes [27] and alcohol-induced DNA damage in the E. coli genome [29].

Recently, a further modification of this assay, Repair Assisted Damage Detection (RADD) was developed with a refined enzyme cocktail for the detection of abasic sites, uracils, strand breaks, and oxidative and UV-induced DNA lesions within cells (Fig. 2) [31].

Figure 2.

Overview of enzyme-mediated DNA damage detection strategies that can be performed on purified DNA or on fixed cells. First, DNA repair enzymes process the DNA damage and create the DNA end chemistry needed for DNA polymerase to fill the gap with a fluorescently tagged nucleotide. Damage sites are imaged and quantified by counting the number of lesions per length of DNA or through measurement of the fluorescence intensity of the nuclear area.

While the RADD in cell assay allows relative quantification of DNA damage [31], the DNA combing strategy can detect very low levels of endogenous DNA damage 0–10 labels per Mbp and significant increases ~ 1–1.5 fold in damage after genotoxic exposures [27–30]. Additionally, since the DNA context is preserved and damage sites are labeled, these techniques are compatible with optical sequence mapping [27–29] or next generation sequencing of isolated damage fragments (similar to [33,34]), allowing new insight into genome damage susceptibility.

These DNA repair enzyme-mediated damage detection strategies offer significant improvements over other enzyme detection strategies such as terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) [35–37]. TUNEL has been employed to detect DNA strand breaks, typically in the context of apoptosis, but also in some cases DNA adducts across a variety of biological samples [38–40]. However, TUNEL is highly specific for DNA strand breaks with 3’-OH ends, limiting the utility for detecting different types of DNA damage. While several TUNEL modifications have emerged extending its ability to detect other DNA ends (i.e., 3’-PO₄) or improving DNA damage detection by incorporating FPG to excise oxidative DNA adducts [41], the repair enzyme-mediated strategies outlined here offer more robust strategies for detecting specific classes of DNA damage or a broad spectrum of DNA damage [30,31]. The only limitation to these techniques are the specificity of known DNA repair enzymes, and engineering of enzymes for specific lesions would open new opportunities to study damage and its sequence context.

Fluorescent modification of DNA adducts

The intrinsic fluorescence of some DNA adducts and fluorescent conjugates DNA adducts by HPLC techniques [1,2]. While these fluorescent strategies are limited in their applications, they highlight that covalent modification of DNA adducts with fluorescent molecules could enhance detection and quantification. To this end, new synthetic strategies have been developed for chemical modifiers that specifically and covalently bind some DNA adducts. These modifiers may be fluorophores, but they could also be functional groups, such as azides, alkynes, amines, or other functional groups compatible with fluorescent labeling.

Fluorophore conjugation directly to DNA adducts is less common, though a near-infrared (NIR) fluorescent probe has been developed for in vitro and in situ imaging of abasic sites [42] (Fig. 3). This probe binds abasic sites with higher affinity than the two current detection reagents for abasic sites, aldehyde-reactive probe (ARP, alkoxyamine) or methoxyamine (MX). Sensitivity of the probe was demonstrated by quantifying chemotherapy-induced abasic sites in DNA isolated from colorectal cancer cells [42]. While this probe has potential in vivo applications, the cell permeability was not demonstrated and inhibition of AP-endonuclease 1 (APE1) above ARP and MX levels was observed [42]. With resolution of these issues, this reagent may be a new analytical tool for assessing abasic sites in vivo and in vitro.

Figure 3.

Two different approaches for fluorescent modification of DNA adducts. Here, an absaic site is modified directly with the NIR probe (top, [42]) or with AA3, a modified alkoxyamine (bottom, [44]). Labeling of the AA3-conjugated molecule is achieved by the addition of an azide-conjugated fluorophore or biotin (bottom, [44]).

Abasic site detection is more commonly conducted through reactions of the abasic sites with ARP modified with biotin, which reacts with the open form of deoxyribose sugar forming an oxime. The biotinylated abasic site can then be readily detected. However, ARP has numerous drawbacks, including competition with endogenous biotin, low reactivity, and production of side products [43]. An improved abasic detection reagent, AA3, has been synthesized, which addresses these shortcomings by functionalizing the abasic site with an alkyne group, which can be subsequently labeled via Click chemistry (Fig. 3) [44]. This functionalization allows users to dictate the fluorescent molecules or functional groups attached to the abasic site. AA3 also inhibits the activity of APE1, but the inhibition is similar to MX [44].

A reagent for interstrand crosslink detection has also been generated by attaching an alkyne group to psoralen, 8-propargyloxypsoralen (8-POP). Psoralen generates interstrand crosslinks when activated by UV light, the 8-POP derivative generates interstrand crosslinks and labels them with an alkyne group, offering fluorophore compatible labeling strategies for the detection of DNA damage in vivo and in vitro

Direct labeling of DNA damage sites with fluorophores or chemical modifiers that facilitate fluorescent labeling offer new strategies for the detection and quantification of DNA damage, particularly of abasic sites and interstrand crosslinks, which are difficult to detect. While the efficiency of these reactions have not been adequately measured, these tools offer new labeling opportunities that overcome antibody affinity and dissociation issues due to their covalent nature. However, users need to be aware of potential genotoxic side effects of these agents and their competition with endogenous repair proteins.

Conclusions

With continual improvements in optics and microscopy and the affordability of laser emitting diodes, fluorescence-based methodologies have significantly expanded. Further, single molecule techniques and fluorescence reading devices have become cheaper, smaller, and more readily available. These advancements have improved sequencing technologies and will likely significantly improve the detection and quantification of DNA damage. Here, we reviewed emerging fluorescence-based techniques for the detection and quantification of DNA damage. While these techniques are not as well-known as mass spectrometry, HPLC, or radiolabeling for adduct detection, they offer new tools for damage characterization that are compatible with other characterization methods, such as live cell imaging and microirradiation [19,31], genomic sequencing [27–29], and are capable of being multiplexed with the analysis of other DNA modifications, such as epigenetic marks [30]. While the sensitivity and detection efficiency of these fluorescence-based techniques still require improvement compared to well accepted adduct detection methods (Table 1), overall these offer new opportunities to characterize genotoxicity and examine DNA repair.

Table 1:

Comparison of established DNA adduct detection methods with reviewed emerging fluorescence-based methods.

Established methods	Maximum Sensitivity	Total DNA required (μg)	Citations
HPLC-ECD/fluorescence	1 in 108	1–100	[1,2]
Immunoassay	1 in 108	1–200	[1,2]
Radio labeled compounds	1 in 108	1–100	[1,2]
32P-post-labelling	1 in 1010	1–10	[1,2]
LC-MS and GC-MS	1 in 108	1–100	[2,7]
AMS	1 in 1012	1–2000	[1,2]
Emerging methods	Maximum Sensitivity	Total DNA required (μg)	Citations
Strand break reporter	1 in 105*	NA	[19,23]
In vitro RADD	1 in 105*	0.1–1	[27–30]
In vivo RADD	ND	NA	[31]
Fluorescent modifiers	ND	0.1–1	[42,44]

^*estimated from Mbp measurements

Highlights.

• DNA adducts are important biological markers of exposure and DNA repair

• Fluorescence-based methods offer new opportunities for assessing DNA adducts

• Fluorescent-tagging of DNA repair proteins allow in vivo damage measurements

• Labeled dNTPs inserted at damage sites by repair enzymes also allow damage detection

• Specific DNA adducts are labeled using covalent binding to chemical modifiers

• Fluorescent methods are improving the detection and quantification of DNA damage