Live Cell Single-Molecule Imaging to Study DNA Repair in Human Cells

Abstract

The genetic material in human cells is continuously exposed to a wide variety of insults that can induce different DNA lesions. To maintain genomic stability and prevent potentially deleterious genetic changes caused by DNA damage, mammalian cells have evolved a number of pathways that repair specific types of DNA damage. These DNA repair pathways vary in their accuracy, some providing high-fidelity repair while others are error-prone and are only activated as a last resort. Adding additional complexity to cellular mechanisms of DNA repair is the DNA damage response which is a sophisticated a signaling network that coordinates repair outcomes, cell-cycle checkpoint activation, and cell fate decisions. As a result of the sheer complexity of the various DNA repair pathways and the DNA damage response there are large gaps in our understanding of the molecular mechanisms underlying DNA damage repair in human cells. A key unaddressed question is how the dynamic recruitment of DNA repair factors controls the kinetics and repair pathway choice in human cells. Methodological advances in live cell single-molecule imaging over the last decade now allow researchers to directly observe and analyze the dynamics of DNA repair proteins in living cells with high spatiotemporal resolution. Live cell single-molecule imaging combined with single-particle tracking can provide direct insight into the biochemical reactions that control DNA repair and has the power to identify previously unobservable processes in living cells. This review summarizes the main considerations for experimental design and execution for live cell single molecule imaging experiments and describes how they can be used to define the molecular mechanisms of DNA damage repair in mammalian cells.

1. Introduction

All organisms constantly acquire damage to their genomes as a result of both natural physiological processes and exposure to endogenous and exogenous mutagens [1,2]. If left unrepaired or repaired improperly, DNA damage can lead to mutations and loss of genetic information [1,3,4]. For this reason, eukaryotic cells have evolved a complex set of specialized and dynamically controlled DNA repair pathways, that work in concert with one another in time and space to safeguard genomic stability. These pathways are controlled by a conserved signaling cascade known as the DNA damage response (DDR) that regulates and coordinates DNA repair at DNA lesions [3,5,6]. Pointing to the importance of these pathways for organismal physiology, loss-of-function mutations in a variety of DNA repair proteins leads to the development of a wide variety of diseases including immunodeficiencies, neurological disorders, and cancer [3,4,7–10]. Programmed base damage and DNA breaks play essential roles in a number of cellular processes such as V(D)J recombination, somatic hypermutation, class-switch recombination, meiotic recombination, replication, and transcription [9,11–14]. Additionally, a myriad of endogenous and exogenous mutagens such as reactive oxygen species, metabolites, metals, ultraviolet light, and ionizing radiation also induce many distinct types of DNA damage which differ in chemistry and complexity including base damage, protein-DNA crosslinks, inter- and intra-strand crosslinks, single-strand breaks (SSBs), and double-strand breaks (DSBs) [1,15]. Therefore, functionally diverse DNA repair pathways are critical to mitigate the potentially deleterious consequences of these lesions for genomic stability by ensuring their repair in a timely manner and with as much accuracy as possible [4,15].

Mammalian DNA repair pathways differ dramatically in the cadre of proteins involved, kinetics of repair, and regulation and can be broadly categorized based by the type of damage each resolves. Major DSB repair pathways in mammalian cells include homologous recombination (HR), non-homologous end joining (NHEJ), microhomology-mediated end joining (MMEJ), and single-strand annealing (SSA) [16]. Conversely, base and nucleotide excision repair, as well as mismatch repair are dedicated to resolving single-stranded DNA nicks, chemically modified or mismatched bases, as well as bulky single-stranded DNA lesions, such as thymine dimers and 6–4 photoproducts [17–19]. It is also important to consider the complex nature of the DDR by recognizing that these pathways are not monolithic but rather are dynamically interconnected with many factors overlapping with one another in certain contexts [1,2,5,15]. Additionally, most DDR factors do not function as single functional units, but rather have diverse or multiple functions within the same or different pathways depending on cell cycle stage, genomic context of the DNA damage, and incorporation into different protein complexes [15,20]. In the context of this review, we will largely focus on DSB repair but many of the concepts can be generalized and applied to the pathways that process other types of DNA lesions.

A key difference between the distinct DSB repair pathways is their differing outcomes, some lead to high-fidelity repair while others are highly error prone. HR is a high-fidelity DSB repair pathway that is largely restricted to S and G2 phase due to a requirement of the sister-chromatid to serve as a repair template [15,16]. In contrast, NHEJ, which is the predominant DSB repair pathway in mammalian cells, as well as MMEJ and SSA are all prone to the introduction of insertions or deletions (indels) as a consequence of the repair process [16,21,22]. While many of the key proteins involved in these repair pathways have been identified, the molecular mechanisms by which individual breaks are repaired by a specific pathway is subject to multiple levels of regulation including cellular abundance of repair factors, extent of DNA end resection, cell cycle phase, and genomic context of the break (e.g., heterochromatin vs. euchromatin and sites actively undergoing DNA replication or transcription) [1,2,5,15]. Importantly, many unanswered questions remain regarding the spatiotemporal dynamics and regulation of DSB repair pathway choice. In addition to traditional biochemical, cell biological, genetic techniques and imaging approaches, the past two decades have seen the development of exciting new methodologies enabling increasingly detailed studies of DNA repair [23]. For example, new sequencing approaches have provided insights into the characteristics and regulation of end-resection at DNA breaks and revealed that specific genomic loci are predisposed to a particular repair pathway. A variety of genomics-based approaches such as ChIP-seq, ATAC-seq, END-Seq, 3C- and 4C-seq have also fundamentally transformed the DNA repair field by providing detailed analysis of DDR signaling and repair factor recruitment to damaged chromatin, as well as revealing how chromatin modifications, accessibility, and 3D-genome organization influences damage induction and repair [24–29]. Finally, advances in cryogenic-electron microscopy (cryoEM) has led the determination of detailed structures of nucleosomes, repair proteins, and repair protein complexes contributing to an increased understanding of the molecular mechanisms underlying DNA repair [30,31]. While these technologies have had a significant impact on the DNA repair field, our knowledge of the subcellular dynamics of DNA repair proteins and the kinetics of DNA repair in living cells is still limited. Live-cell imaging has been extensively used to analyze the recruitment of DNA repair factors in DNA repair foci and laser-microirradiation induced DNA lesions [32]. However, many DNA repair proteins do not form visible foci because their high protein abundance leads to a significant background signal that makes foci undetectable or because only a small number of these proteins is recruited to DNA breaks (e.g., core NHEJ factors) [31]. To define how the dynamic recruitment of DNA repair factors controls DNA repair pathway choice, the kinetics of DNA repair, and ultimately genome integrity, we therefore require new methods that provide the high sensitivity, spatiotemporal resolution, and are capable of analyzing the repair of wide variety of DNA lesions in living cells.

Despite the advances of our understanding of DNA repair in human cells, many significant mechanistic questions remain unanswered. For instance, it is unclear how the repair kinetics and outcomes of DNA repair are controlled by genomic context, the chemical nature of the DNA lesions, cell-cycle stage, and the relative abundance of repair factors. The development of live-cell single-molecule imaging methods provides unprecedented sensitivity and spatiotemporal resolution to begin answering these critical questions. This powerful tool enables real-time analysis of DNA repair at the single molecule level in living cells and facilitates the precise determination of biochemical and biophysical properties of DNA repair proteins in their endogenous context. In particular, when combined with genome editing to introduce fluorescent tags at the endogenous locus of DNA repair proteins, live cell single-molecule imaging will undoubtably lead to significant advancement of understanding of DNA repair in human cells. Importantly, the quantitative information generated by live cell single-molecule imaging studies can be used to test predictions made by mathematical models of DNA repair or serve as input parameters for such models [33–35]. In this review, we will discuss live-cell single-molecule imaging and its many potential applications to the study of DNA repair. We will describe methods for performing these experiments including expressing fluorescent fusion proteins, various strategies to induce DNA-damage, imaging approaches to detect single-molecules in living cells and highlight how the quantitative data generate by single-particle tracking can be used to answer critical unanswered questions in the DNA repair field.

2. General considerations for using live-cell single-molecule imaging to study DNA repair

For macromolecules to carry out their function in human cells, they transition through a number of distinct molecular (e.g. complex assembly, target binding) and cellular (e.g. sub-cellular trafficking) states. These states often co-exist within the cell and molecules can rapidly exchange between them. A simple example would be a DNA repair factor that is imported into the nucleus, binds to a co-factor, and then rapidly binds and dissociates from non-cognate chromatin sites (chromatin scanning) on its path to tightly associating with a DNA lesion (Fig. 1). To precisely define how DNA repair factors carry out their function in living cells requires a methodology that can identify distinct molecular states and measure the transition of molecules between them with high spatiotemporal resolution. Live-cell single-molecule imaging facilitates the identification molecular states that result in a change of the diffusion dynamics or sub-cellular localization of the molecule analyzed (Fig. 1). The rate at which molecules transition between states reports on the biochemical properties of the underlying interactions. Together, these unique capabilities make live-cell single-molecule imaging a powerful tool to discover otherwise unobservable molecular states and essentially allows biochemical measurements in living cells. A key requirement to study the diffusion and chromatin-binding activities of DNA repair proteins at the single-molecule level in living cells is that only a small number of proteins are visualized at a given time (at most ~50 per nucleus). For low abundance DNA repair factors such as the components of the Shieldin complex (~5,000 molecules per cell) a labeling density of ~100 particles per nucleus corresponds to 2% of the cellular protein being labeled [36]. In contrast, for highly abundant proteins like DNA-PKcs (~100,000 molecules per cell) this corresponds to only ~0.1% of total DNA-PKcs being fluorescently labeled [36]. To use this approach for monitoring the recruitment of sparsely labeled, abundant DNA repair factors to DNA lesions, a significant fraction of the total protein pool must be recruited to sites of DNA damage. Importantly, this can be controlled by the number of lesions that are induced in single-molecule imaging experiments by titrating DNA damaging agents. Therefore, to successfully perform live-cell single-molecule imaging experiments to study DNA repair, the interplay between protein abundance, the number of DNA damage sites, and protein or drug kinetics must be taken into account to produce biologically relevant and reliable measurements of biochemical or biophysical properties of the DNA repair factors studied.

Figure 1.

Distinct molecular states of DNA repair factors can be distinguished by the rate of their diffusion through the cell. Single-molecule imaging can identify these different states by determining individual particle trajectories with high spatial and temporal resolution.

3. Expression of fluorescently tagged DNA repair proteins

Fluorescence microscopy is a powerful technique that can be used to visualize biological processes in living cells and organisms. The detection of fluorescently labeled biomolecules such as a protein or RNA requires their direct fusion to a fluorescent protein (FP) or RNA-aptamer and expression of these molecules in the cells of interest. In the past, expression of fluorescently tagged biomolecules mainly relied upon transient or stable expression using plasmid transfection or viral transduction, respectively [37–43]. However, these approaches can lead to overexpression, which can perturb physiological protein localization, function, and/or kinetics. In addition, stable integration of virally delivered expression constructs can lead to unwanted alterations of the host genome. Finally, these expression strategies can make it challenging to assess the functionality of the tagged molecule if the endogenous untagged protein is present.

Advances in genome editing technologies now enable stable insertion of fluorescent tags directly into the genomic loci of the molecules of interest [44,45]. In particular, the CRISPR-Cas9 system in combination with homology directed repair can be used to guide site-specific insertion of fluorescent tag at the N-or C-terminus of essentially any gene (Fig. 2A). Without prior knowledge it is hard to predict whether fusion of a fluorescent tag to either terminus of a protein will impact its function, but in many cases fluorescent protein tagged DNA repair factors have been transiently expressed in prior studies and can guide the decision-making process. A flexible linker that includes a protease cleavage site has been successfully used in the past to minimize the functional impact of fusing the HaloTag to DNA repair factors [36]. Importantly, direct insertion of the fluorescent tag into the endogenous loci of the biomolecules of interest retains the regulatory mechanisms conferred by the endogenous promoter, splicing patterns, and homozygous insertion facilitates exclusive expression of the tagged molecule, making functional validation of the tagged protein much more straightforward. To interpret data from live-cell single-molecule imaging experiments, it is critical to confirm that tagging does not interfere with protein localization or functionality. In previous work the functionality of homozygously tagged DNA repair proteins was assessed in a variety of ways including subcellular localization, DNA repair foci formation, and recruitment to laser microirradiation induced DNA damage sites [36]. However, while these experiments demonstrate if the tagged protein can be recruited to DNA damage sites, they provide little information on whether the tagged protein can support proper DNA repair. Therefore, it is critical to ensure full functionality of the tagged DNA repair factor using drug sensitivity assays (e.g., clonogenic survival assays) to compare the responses of parental and genome-edited cells exclusively expressing tagged protein, ideally using a knock-out of the tagged factor as a complete loss of function control [36].

Figure 2.

(A) Different strategies to insert the HaloTag at the endogenous loci of protein coding genes in human cells. The HaloTag can either be directly inserted or include a selectable marker (SV40-PuroR), that can be subsequently removed using Cre-recombinase. (B) Chemical basis of covalent modification of the self-labeling HaloTag using the HaloTag ligand conjugated to JF646.

The mostly widely used tags for live-cell imaging of proteins are FPs and self-labeling protein tags (e.g., HaloTag and SNAP-tag) [46]. In recent years, many modifications to FPs have been made that improve their protein folding, brightness, and photostability and FPs with a wide range of spectral properties are available for live-cell imaging. In addition, photo-activatable and photo-convertible FPs have been developed that can be used to mark specific sub-populations of tagged proteins by exposure to a specific wavelength of light which can be spatially and temporally controlled [46–51]. In comparison to FPs, self-labeling tags are more versatile allowing multi-color, low-, or high-density protein labeling with fluorescent small-molecule ligands. In addition to their compatibility with traditional live-cell imaging approaches, these self-labeling protein tags can also be used for live-cell single-molecule imaging [52]. Self-labeling proteins tags typically catalyze the formation of a covalent bond between the tag and a small molecule ligand which can be modified with a wide range of membrane-permeable, bright, and photostable small-molecule fluorophores. The three most widely used self-labeling technologies are the HaloTag (34 kDa), the SNAP-tag (14 kDa), and to a lesser extent the CLIP-tag (20 kDa) [53,54]. These proteins are engineered variants of natural enzymes that covalently bind to ligands with chemically distinct reactive groups (Fig. 2B). Because the HaloTag and the SNAP-tag react with distinct small molecule ligands they can also be used simultaneously. It is important to note that the self-labeling kinetics of the HaloTag are approximately 10-fold faster compared to the SNAP-tag, which reduces the concentration of fluorescent ligand and time required to label it leading to substantially lower background signal [55]. For this reason, the HaloTag is most commonly used for live-cell single-molecule imaging experiments.

Key advantages of self-labeling tags are the availability of extremely bright and photo-stable fluorescent ligands and the ability to label only a subset of the molecules with small-molecule HaloTag and SNAP-tag ligands, which are both critical for single-molecule imaging (Fig. 2B). Labeling a well-controlled fraction of the protein of interest is critical for the sparse-labeling required to detect single-molecules in living cells. There is an extensive toolbox of small-molecule probes for self-labeling protein tags that can be used to answer specific scientific questions [49]. The development of the Janelia Fluor dyes by Jonathan Grimm and Luke Lavis and others has dramatically increased the versatility of self-labeling proteins tags through the generation of membrane-permeable, bright, and photostable HaloTag and SNAP-tag ligands with diverse spectral properties [47,48,52,56]. Many of these ligands are freely available as part of the HHMI/Janelia Research Campus Open Science Initiative (https://janeliamaterials.azurewebsites.net/). In most cases it is preferable to covalently attach the small-molecule fluorophores to self-labeling tags, but it is worth noting that HaloTag ligands were recently developed that bind non-covalently allowing for fluorophore exchange [57]. These ligands can be desirable for long-term imaging or applications that require transient fluorescence detection like stochastic optical reconstruction microscopy (STORM). For live-cell single-molecule imaging the membrane permeable Janelia Fluor dyes JF549, JF646 and the deuterated JFX554 and JFX650 perform exceedingly well for imaging HaloTagged proteins. In particular, JFX650 stands out due to its high brightness, photostability, and because its spectral properties avoid interference from cellular auto-fluorescence and common dyes used in tissue culture media [47,48,52,56]. In addition, JF646 and JFX650 are fluorogenic dyes whose fluorescence increases dramatically (>20-fold) upon conjugation to the HaloTag, significantly reducing non-specific background fluorescence of unbound dye [47]. In previous live-cell single-molecule imaging studies of HaloTagged DNA repair proteins labeled with either Halo-JF646 or Halo-JFX650, low-density protein labeling was achieved with very short labeling time and incubation with low concentrations of the ligands (30 seconds – 1 min using 0.1 – 10 nM of fluorescent HaloTag ligand depending on the expression level of the protein of interest) [36]. Therefore, we recommend using the HaloTag in combination with the Halo-JFX650 ligand for studying DNA repair factors at the single-molecule level in human cells.

4. Methods to induce DNA damage

To analyze the recruitment and dissociation of DNA repair factors to and from DNA lesions in living cells at the single-molecule level the well-controlled induction of DNA damage is essential. Three important variables to consider when choosing the method of DNA damage induction are temporal control (how fast do lesions occur), genomic sites that are damaged (the number of damages sites, specific vs. nonspecific damage induction), and the chemical nature of the lesions (DSBs, DNA adducts, etc.). The three techniques frequently used in live-cell imaging experiments to introduce DNA damage are nucleases, chemical agents, and laser micro-irradiation [23]. In the following section, we will discuss the utility of these methodologies for live-cell single-molecule imaging experiments.

Chemical agents can be used to directly induce specific types of DNA lesions such as base damage, bulky single-stranded DNA lesions, SSBs, inter- and intra-strand crosslinks, as well as DSBs with unique end chemistries [23] (TABLE 1). In addition, small molecule drugs that inhibit topoisomerases or poly(ADP) ribose polymerase (PARP) result in the formation of DSBs that is greatly enhanced during DNA replication in S-phase of the cell-cycle [58]. A major limitation of inducing DNA damage with chemical agents is that in most cases they induce DNA damage non-specifically throughout the entire genome which cannot be spatially controlled [23]. Additionally, there is limited information on the kinetics with which DNA lesions are induced making it important to optimize treatment time as well as drug concentration to ensure enough lesions are being formed to observe a change in protein binding to DNA damage sites. For example, PARPi requires DNA replication to occur before most DSBs are induced, which significantly delays the occurrence of DNA lesions relative to the exposure to the drug and can lead to dramatically different effects between individual cells unless cell cycle stage is accounted for [59]. In addition, to add the drug to the cells the cell culture medium has to be changed. Therefore, reliable imaging of the same cells before and after drug addition requires a flow chamber to seamless change the culture media without physically moving the imaging dish.

Table 1.

List of commonly used DNA-damage inducing agents

Agents		Examples		DNA damage		Major mammalian DNA repair pathway involved		Mechanism of action
Antitumor Antibiotic		Bleomycin		DSB, SSB, oxidized bases		NHEJ/HR		Generation of free radicals/inhibition of DNA synthesis
		Zeocin		DSB		NHEJ/HR		Generation of free radicals
		Calicheamicin		DSB		NHEJ/HR		Generation of free radicals (highly reactive species known as diradicals)
		Doxorubucin		DSBs with top- DNA adducts		NHEJ/HR		Inhibition of topoisomerase II /can intercalate into DNA leading to introduce torsional stress/generate free radicals
		Etoposide		DSBs with top- DNA adducts		NHEJ/HR		Inhibition of topoisomerase II/inhibition of DNA re-ligation
		Teniposide		DSBs with top- DNA adducts		NHEJ/HR		Inhibition of topoisomerase II/inhibition of DNA re-ligation
		Daunorubicin		DSBs with top- DNA adducts		NHEJ/HR		Inhibition of topoisomerase II/inhibition of DNA re-ligation
		Zorubicin		DSBs with top- DNA adducts		NHEJ/HR		Inhibition of topoisomerase II/inhibition of DNA re-ligation
		Camptothecin		DSBs with top- DNA adducts		NHEJ/HR		Inhibition of topoisomerase I
		Topotecan		DSBs with top- DNA adducts		NHEJ/HR		Inhibition of topoisomerase I
		Irinotecan		DSBs with top- DNA adducts		NHEJ/HR		Inhibition of topoisomerase I
Antimetabolites		5-Fluorouracil (5-FU)		SSB, DSB		BER/HR/MMR		Inhibition of thymidylate synthase/inhibition of DNA synthesis
		Cytarabine (Ara-C)		SSB, DSB		HR		Inhibition of DNA synthesis
		Gemcitabine		SSB, DSB		HR/NER		Nucleoside analog/ antimetabolite/inhibition of DNA synthesis/blocks the progression of cells through the G1/S-phase boundary
		Methotrexate		SSB, DSB		HR		Inhibition of nucleotide synthesis (including dihydrofolate reductase, thymidylate synthase, aminoimidazole caboxamide ribonucleotide transformylase
		Hydroxyurea		SSB, DSB		HR		Inhibition of DNA synthesis/inhibiting the activity of the enzyme ribonucleotide reductase
Platinum analogs		Cisplatin		intra- and interstrand cross-links		NER/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases/preventing DNA synthesis and RNA transcription
		Carboplatin		intra- and interstrand cross-links		NER/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases/preventing DNA synthesis and RNA transcription
		Oxaliplatin		intra- and interstrand cross-links		NER/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases/preventing DNA synthesis and RNA transcription
Alkylating agents		Bendamustine		intra- and interstrand cross-links,SSB, DSB		NER/FA/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases
		Mitomycin C		Interstrand cross-links (ICLs)		NER/FA/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases
		Chlorambucil		intra- and interstrand cross-links,SSB, DSB		NER/FA/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases
		Cyclophosphamide		intra- and interstrand cross-links,SSB, DSB		NER/FA/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases
		Melphalan		intra- and interstrand cross-links		NER/FA/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases
		Ifosfamide		intra- and interstrand cross-links		NER/FA/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases
		Procarbazine		intra- and interstrand cross-links,SSB, DSB		NER/FA/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases
		Busulfan		intra- and interstrand cross-links,SSB, DSB		NER/FA/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases
		Carmustine		intra- and interstrand cross-links,SSB, DSB		NER/FA/HR		Binds to the DNA molecule, causing cross-links to form between adjacent bases
PARPi		Olaparib		SSB, DSB		BER/HR		Selective competitive inhibitor of NAD+ at the catalytic site of PARP enzymatic activity and increased formation of PARP- DNA complexes
		Rucaparib		SSB, DSB		BER/HR		Selective competitive inhibitor of NAD+ at the catalytic site of PARP enzymatic activity and increased formation of PARP- DNA complexes
		Niraparib		SSB, DSB		BER/HR		Selective competitive inhibitor of NAD+ at the catalytic site of PARP enzymatic activity and increased formation of PARP- DNA complexes
		Talazoparib		SSB, DSB		BER/HR		Selective competitive inhibitor of NAD+ at the catalytic site of PARP enzymatic activity and increased formation of PARP- DNA complexes
		Veliparib		SSB, DSB		BER/HR		Selective competitive inhibitor of NAD+ at the catalytic site of PARP enzymatic activity and increased formation of PARP- DNA complexes

In contrast to chemical agents, nucleases like I-SceI, FokI, Zinc-finger nucleases, and Cas9 facilitate the introduction of DSBs at specific genomic loci. These tools can provide excellent specificity DNA damage induction; however, the number of breaks that can be induced at once is low making single-molecule imaging difficult. Cas9 is an exception because it can be programmed to target a repeated element that occurs thousands of times in the human genome. A key disadvantage of using nucleases is that it is challenging to temporally control the induction of the DNA break. In addition, once expressed, nucleases are in the cell for an extended period of time and can repeatedly cleave genomic DNA. To some degree temporal control could be provided by inducible expression of the nuclease or controlled nuclear import by fusing the nuclease to the estrogen receptor, which translocates to the nucleus after ligand binding. In contrast to the examples described so far, the recently developed very fast CRISPR (vfCRISPR) on demand approach allows for the precise spatial and temporal induction of a DSB by the Cas9 endonuclease [60]. vfCRISPR on demand uses a photocaged sgRNA that enables localization of the RNP to DNA but does not allow cleavage until the removal of the caged groups from the sgRNA using blue light [60]. In combination with the specificity provided by the sgRNA, vfCRISPR on demand offers unprecedented temporal control of DSB induction and is therefore an ideal tool to combine with live cell single-molecule imaging. Moreover, the combination of vfCRISPR on demand with single-particle tracking is an extremely promising approach to define the molecular mechanism by which Cas9-induced DNA breaks are repaired, which will be critical for future development of therapeutic genome-editing approaches. Drawbacks of vfCRISPR on demand include the requirement for nucleofection of Cas9 loaded with the caged sgRNA and the very slow dissociation rate of Cas9 from cleaved DNA (t_½ > 43 hours in vitro), which could bias repair pathway choice due to the steric hinderance imposed by the Cas9 bound to the DNA break [60,61].

Finally, the most widely used method to induce DDR in live-cell imaging is laser microirradiation (LMI). This powerful tool uses a laser to induce DNA damage at a specific location in the nucleus. LMI has been extensively used to study the recruitment of DNA repair factors to DNA lesions and the dynamics of DNA repair protein binding to damaged DNA using fluorescence recovery after photobleaching (FRAP) [62]. Furthermore, LMI has been used to define the sequence of recruitment of repair proteins to the site of DNA damage by comparing the recruitment kinetics of studied proteins, shed light on differences between repair in various nuclear compartments (e.g., nucleoli, euchromatin, heterochromatin, etc.), provide structural details of DNA damage foci, and investigate the role of liquid-liquid phase separation (LLPS) in the DNA repair [63–66]. Key drawbacks of using LMI are that they are challenging to reproduce between laboratories because of key differences in experimental parameters (e.g., wavelength, pulse length, pulse frequency and energy), and the induction of complex DNA lesions (a mixture of base damage, abasic sites, SSBs and DSBs) [32]. A particular challenging aspect of using LMI for single-molecule live cell imaging is the photo-bleaching associated with the high intensity laser pulse required to induce damage. Because samples have to be sparsely labeled, LMI will likely bleach most if not all of the fluorophores in the cell. In future studies is might be possible to optimize the labeling density to yield an optimal fluorophore density after LMI. Alternatively, LMI could be combined with photo-activatable fluorescent dyes, which would lead to an increased fluorophore density after laser exposure.

In summary, each of these approaches represents a feasible strategy to induce DNA damage in live cell single-molecule imaging experiments. Each has certain benefits related to the spatial, temporal, and numerical induction of DNA breaks as well as specific limitations that should be considered when designing single-molecule imaging experiments to study DNA repair mechanisms in living cells.

5. Live-cell single-molecule imaging of DNA repair

Live-cell single-molecule imaging in combination with single-particle tracking is a powerful tool that extracts biochemical information from molecule trajectories and sub-cellular localization and has many potential applications for studying DNA repair. Because this method is performed in living cells it avoids many potential confounding variables introduced by many in vitro biochemical and single-molecule approaches (e.g. missing co-factors, post-translational modifications). On the other hand, because the experiments are carried out in living cells and not in a controlled system with a specified sub-set of components the biochemical interactions that lead to a change in molecular behavior are not immediately apparent. It is therefore critical to include genetic or chemical perturbations to identify the molecular interactions responsible for the observed molecular properties. The extraction of biologically relevant information from single-particles in live cells involves several key steps that include low density protein labeling, implementation of a sensitive microscopy approach, single-particle tracking, and analysis of molecule trajectories to determine important biochemical properties. In this section we will discuss key principles of live-cell single-molecule imaging and single-particle tracking, experimental details and considerations, applications for studying mechanisms of DNA repair as well as the benefits of this approach versus other types of biochemical and imaging approaches.

5.1. Principles of live-cell single-molecule imaging

To detect single-molecules it is critical to sparsely label the molecule of interest with a bright, photostable fluorophore (as described above) and to image the cells using a highly sensitive microscopy approach (Fig. 3A–B). It is important to note that in most cases single-molecule live cell imaging experiments are carried out in a single-focal plane, which means that molecules can diffuse out of the focal plane. The most commonly used microscopy methods for single-molecule live cell imaging are highly inclined laminated optical sheet microscopy (HILO) (Fig. 3A), total internal reflection fluorescence microscopy (TIRFM), and lattice light sheet microscopy (LLSM) [67]. By illuminating a subsection of the sample, these methods reduce out-of-focus fluorescence background. TIRFM is used to illuminate a 100–200 nm thin section of the sample at the interface between the glass coverslip and the cell, which is ideal for studying plasma membrane associated molecules, but has limited utility for studying nuclear processes such as DNA repair. In LLSM a thin sheet of light is used to precisely excite fluorophores in the focal plane of the detection objective using an illumination objective oriented at a 90-degree angle relative to the detection objective [68]. LLSM is ideal for limiting photo-bleaching of the sample and to facilitate imaging of thicker specimen (e.g. organoids) but it requires water dipping objectives, which limits the sensitivity of LLSM. In addition, commercial solutions for lattice light sheet microscopes are thus far limited and can require complex sample mounting procedures. HILO uses a TIRF illuminator to expose the cell to a sheet of light that can be angled into the sample to excite fluorophores in the nucleus (Fig. 3A) [69]. Importantly, HILO can be implemented on most TIRF microscopes and uses high numerical aperture TIRF objectives to maximize sensitivity. All of the microscopy methods described here can take advantage of recent advances in camera technology. Highly sensitive scientific Complementary Metal-Oxide Semiconductor (sCMOS) cameras (e.g. Hamamatsu ORCA Quest, Hamamatsu ORCA Fusion BT) can reliably detect single-molecules in living cells at high imaging rates (>100 Hz) and over large fields of view. Because of its straightforward implementation using widely available commercial TIRF microscopes, we believe HILO microscopy is the ideal approach for the broader scientific community to carry our live cell single-molecule imaging experiments to study DNA repair processes.

Figure 3.

(A) Principle of HILO microscopy for single-molecule live cell imaging. Only labeled proteins in the path of the excitation laser will be detected. In combination with sparse labeling this approach provides single-molecule sensitivity. (B) Workflow for live cell single molecule imaging. Timing of induction of DNA damage depends on the specific scientific question being asked and mechanism to induce DNA damage. After allowing cells to attach to the plate, proteins are sparsely labeled with a fluorescent HaloTag or SNAP-tag ligand and fluorescently labeled proteins are imaged using HILO microscopy. Next, live-cell single-molecule imaging movies are processed using a multiple-target tracing algorithm which goes through each frame and detects particles and links them into trajectories. Finally, quantitative information from protein trajectories is extracted using model-based data analysis tools such as Spot-On and ExTrack.

The reliable analysis of protein behavior by live-cell single-molecule imaging of tagged proteins requires several considerations related to sample preparation. In these experiments cells are seeded onto high precision coverslips or multi-well plates. To reduce fluorescence background due to adherence of the fluorescent small-molecule ligands used for the self-labeling protein tags the glass surface can be washed with ethanol and potassium hydroxide solution. However, in our experience, single-molecule imaging of nuclear proteins in U2OS cells can be efficiently carried out using commercially available glass-bottom plates [36]. U2OS cells have the advantage of being relatively flat, while the Z-dimensions of some other adherent or non-adherent cell lines differs dramatically, therefore a careful optimization of which type of glass bottom plate yields the highest quality imaging data will need to be determined for each individual cell line. For adherent cell lines, it is important to seed cells allowing sufficient time for cells to fully attach to the plate and “flatten out” prior to imaging. While it is more challenging, live cell single-molecule imaging has also been successfully carried out using cells that grow in suspension like lymphocytes [70].

The true power of live-cell single-molecule imaging is the ability to monitor events or protein behavior with high spatiotemporal resolution. In addition to the importance of optimizing protein labeling for reliable detection of single molecules (see above), the imaging rate is very important to maximize temporal resolution. In previous work a variety of DNA repair proteins were imaged at the single-molecule level using frame rates of up to 200 Hz, making the time resolution of the experiments (time between imaging frames) ~ 5 ms [36,71]. It is important to note that increasing the imaging rate reduces the number of photons that can be emitted and detected during a single imaging frame. To allow single-molecule sensitivity at high imaging rates the laser-power often has to be increased, which accelerates photobleaching. It is therefore critical to carefully optimize the imaging rate and the laser-power depending on the kinetics of the biological process that is being studied. Extremely fast processes require fast imaging rates and high laser-power, while long-lasting binding events are best studied using lower laser-power, longer exposure times, and slower overall imaging rates.

Once high-quality live cell single-molecule movies have been acquired, single-particle tracking involves three main steps: 1. Detection of particles in each imaging frame, 2. Linking detected particles into trajectories, and 3. Inference of molecular properties from trajectories (Fig. 3B). The multiple-target tracing (MTT) algorithm developed by Serge et al. and implemented for MATLAB by Hansen et al. is commonly used to carry out single-particle detection and tracking, but other tools like TrackIt are also available [72–74]. These algorithms generate a list of particle trajectories that typically include particle coordinates, signal intensity, and in some cases information on localization precision. The particle trajectories contain a tremendous amount of information regarding the biochemical properties of the molecule analyzed. In the following sections we will discuss the many potential applications of this approach and provide examples of how the trajectories of DNA repair proteins can be analyzed to define the molecular mechanism of DNA repair pathways (Figure 4).

Figure 4.

Applications with single-molecule sensitivity include visualization of lowly abundant proteins and nuclear analysis of protein binding and diffusion. Applications combining multiple fluorescent proteins includes protein dynamics in subnuclear regions, inside and outside of DNA repair foci, throughout the cell cycle as well as protein trafficking between the nucleus and cytoplasm. Lastly, the principle of proximity assisted photoactivation (PAPA) demonstrates another potential application of live-cell single-molecule imaging to detect interactions between two different fluorescently labeled proteins.

6. Applications of live-cell single-molecule imaging to the study of DNA repair

6.1. Protein diffusion and chromatin binding:

There is a variety of different ways that biologically relevant information can be extracted from particle trajectories. A core assumption is that all biomolecules can exist in different states that have distinct movement patterns. In the context of DNA repair proteins, the simplest case would be a factor than can either randomly diffuse through the nucleus moving with large step sizes or can be bound to a DNA lesion leading to small step sizes because the underlying chromatin can be considered static over the time course of a single-molecule imaging movie (Figure 4). To determine what fraction of molecules is chromatin bound the overall step size distribution can be analyzed with tools like SPOT-ON [74]. SPOT-ON explicitly takes into account that mobile molecules will be underrepresented because they can diffuse out of the focal plane of the experiment, while static molecules will be overrepresented because they remain in the focal plane for an extended period of time. In addition to reporting the fraction of molecules that are static (in this example considered chromatin bound) SPOT-ON also determines the diffusion coefficient (a direct measure of the magnitude of molecule movement over time) of the freely diffusing and static particles [74]. This approach can be used to great effect to analyze chromatin recruitment of DNA repair factors in response to DNA damage [36,71]. Additionally, it has led to new insights into the mechanisms by which other nuclear factors search for specific genomic sites and associate with chromatin [75,76]. For example, live-cell single-molecule imaging of HaloTagged CTCF and Rad21 revealed the dynamic properties of CTCF and cohesin which bind the same genomic sites [76]. Live-cell single-molecule imaging of HaloTagged PRC2 also revealed the chromatin binding properties of this methyltransferase as well as the importance of several accessory proteins for the PRC2-chromatin interaction [75]. In addition, global analysis of step size distributions using SPOT-ON can be expanded to include a third molecular state with an intermediate diffusion coefficient (e.g. molecules scanning chromatin, or contained in a phase separated droplet). Importantly, some prior knowledge of the number of potential molecular states and their associated diffusion coefficient is required to run SPOT-ON. It is worth noting that tools exist to predict the number of distinct states based on the step size distribution of single-particle trajectory data sets. Importantly, proteins can be associated with DNA or other nuclear structures in many different ways such as being incorporated into nucleosomes, directly bound to chromatin or to a modified chromatin mark, indirectly associated with chromatin through another protein, or slowly diffusing within a phase-separated focus. In principle, live-cell single-molecule imaging allows for the exploration of any of these types of interactions given the proper chemical, genetic, and imaging tools. In previous work CRISPR-Cas9 was used to knock a 3xFLAG-HaloTag into the endogenous loci of 12 DNA repair factors, and the diffusion and chromatin binding characteristics of each protein was systematically analyzed in both unperturbed conditions and after induction of DNA DSBs with zeocin (see videos 11–22) [36]. As important controls for these experiments transiently expressed the 3xFLAG-HaloTag fused to a nuclear localization sequence (Halo-NLS) and HaloTagged H2B (Halo-H2B) were imaged [36]. These controls were important to define the expected dynamics of the fastest diffusion of unbound particles (Halo-NLS) and slowest diffusion of bound particles (Halo-H2B). Additionally, these controls also represented the lower and upper bounds of the expected fraction of bound and unbound particles. All these proteins were analyzed using a 2-state model using SpotOn in MATLAB and a wide range in diffusion coefficients were observed of both freely diffusing and bound particles consistent with differences in protein size, interactions, and modes of chromatin binding leading to detectable changes in diffusion rates [36]. In addition, a wide range in the fraction of each protein bound to chromatin was detected, including the unexpected observations that MDC1 and RIF1 are constitutive chromatin-binding proteins [36]. While it was possible to demonstrate that proteins like MDC1 and RIF1 are tightly associated with chromatin it is not immediately known which molecular interactions are responsible for this observation. It is therefore critical to perturb the system using genetic or chemical means to identify the biochemical basis underlying the observed binding. In the case of MDC1, individual deletion of two known chromatin interactions domains in MDC1, the Proline-Serine-Threonine (PST) domain and the BRCA1 C-teminal (BRCT) domain, lead to the observation that the PST domain was responsible for the constitutive chromatin binding of MDC1 while the BRCT domain is required for its recruitment to DNA breaks [36]. It was also demonstrated that analysis of the chromatin bound fraction of a DNA repair factor can be used to study its recruitment to DNA breaks. The fraction of static DNA-PKcs particles was significantly increased after treatment with zeocin, consistent with its recruitment to DNA breaks [36]. This was an important advance because only two molecules of DNA-PKcs are expected to be recruited to a DSB highlighting the sensitivity of the single-molecule imaging approach. The ability to analyze the recruitment of other NHEJ factors (Ku70, XRCC4, XLF) to chromatin using single-molecule imaging was further highlighted in a recent pre-print (see movies S7-S9) [71]. In addition, this work demonstrated that single-molecule imaging can be used to analyze the overall kinetics of DSB repair by NHEJ after treatment with the drug Calicheamicin in human cancer cells (see movie S10) [71]. Mahadevan et al. recently used live-cell single-molecule imaging to analyze the effect of four FDA-approved PARP inhibitors (PARPi) (olaparib, rucaparib, niraparib, and talazoparib) on dynamics states of PARP1/2 [77]. They demonstrated that HaloTagged PARP1/2 proteins exist in fast diffusing, slow diffusing, transiently and stably chromatin bound states and that the PARP inhibitors had different effects on PARP1/2 immobilization, which is thought to be the mechanism of action of these drugs [77]. These studies serve as a proof-of-principle that single-molecule imaging is a valuable tool to analyze the dynamic properties of DNA repair proteins and has the potential to advance our understanding of the molecular mechanisms of DNA repair in mammalian cells. While our previous study provided evidence of the feasibility to monitor DNA repair protein dynamics, this approach could also be combined with many other tools to answer a range of biological questions (Figure 4). For example, DNA damage could be induced using different chemicals that result in distinct types of DNA damage or CRISPR-Cas9 and other nucleases to induce site-specific and/or temporally controlled DNA damage. Additionally, simultaneous imaging of live-cell sensors of cell cycle phase, such as the fluorescent ubiquitin-based cell cycle indicator (FUCCI) could also be useful to monitor DNA repair protein activities in specific cell cycle stages (Figure 4) [78]. Finally, this methodological approach can also enable visualization of endogenous proteins that have low abundance that have been previously undetectable, which was recently demonstrated by HaloTagging and imaging polymerase theta (Polθ) (Figure 4) [79].

6.2. Residence time analysis:

The time frame of a chromatin binding event (residence time), i.e. the length of a static particle trajectory can be used to determine the dissociation rate constant of the underlying biochemical interaction. The dissociation rate constant is an extremely valuable piece of information that can for example be used to estimate the affinity of a DNA repair factor for a DNA lesion in living cells. Importantly, determination of residence time has previously been used to define the kinetic properties of transcription factors and chromatin remodelers and their dynamic associations with chromatin [80,81]. Accurate measurement of the residence time requires continuous tracking of a bound molecule. Imaging under low signal to noise conditions makes tracking challenging because even using the most lenient tracking settings, long binding events are often represented by multiple short trajectories because the tracking algorithm is not 100% accurate. It is critical to assure that the tracking algorithm continuously tracks the binding events, rather than breaking up a single binding event into multiple shorter trajectories, which will lead to an overestimation of the dissociation rate constant. There are several experimental and analytical approaches that allow reliable analysis of the residence time of static molecules. On the experimental end, one can increase the exposure time to 100–1000 ms (see movies S6-S9) [82]. This long integration time will amplify the signal of static molecules and “blur out” mobile molecules. It is important to choose an exposure time that is substantially shorter than the expected residence time. In addition to minimizing the signal contributed by mobile molecules, this approach also requires lower laser power which in turn minimizes the contribution of photo-bleaching. For long binding events it is critical to consider the contribution of photo-bleaching, because static molecules can either disappear due to dissociation or as a result of photo-bleaching. A simple approach to rule out photo-bleaching as a contributing factor to the observed kinetics is to use two different laser powers for the same experiment. If the observed kinetics are independent of the laser power, photo-bleaching is not a confounding variable. An analytical approach to analyze the residence time of chromatin bound DNA repair factors is to constrain the expected diffusion coefficient in the tracking algorithm [83]. This method will exclusively track static molecules because freely diffusing molecules move too far to be connected into trajectories. To increase the reliability of tracking static molecules, the fluorescence signal can be averaged over multiple frames minimizing fluctuations in the fluorescence signal caused by photo-blinking (see video 25) [36]. The track length distribution can then be fit with one or multiple exponential decay functions to determine the dissociation rate constant of the underlying interactions. It is important to note, that the HaloTag-NLS alone displays a small fraction of static molecules, indicating that it can associate with chromatin or other static nuclear structures. One should therefore always expect at least two kinetic components when analyzing the residence time distributions of HaloTagged DNA repair factors.

6.3. Kinetic analysis of state transitions:

The residence time analysis described above is limited to analyzing the transition from a static state to a freely diffusing state. It is also possible to determine the transition rates between other states with distinct diffusion patterns (e.g. chromatin scanning, complex formation, chromatin binding). To determine the rates at which molecules transition between different states, long particle tracks are required in which the tracked molecule occupies multiple distinct states over the course of the trajectory. Transition rates inform on the kinetics of underlying biochemical process. For example, the transition from a freely diffusing to a bound state reports on the association rate constant. Precise determination of transition rates is complicated by diffusion of molecules out of the focal plane of the microscope or photobleaching of the imaged fluorophore because they lead to track termination. In single-molecule imaging experiments of yeast and bacterial cells transition rates can be reliably determined because the dimensions of the cells are sufficiently small to eliminate the possibility of molecules diffusing out of the focal plane [84–87]. Importantly, the recently developed Ex-Track tool could potentially fill this gap for the analysis of single-particle trajectories in human cells [88].

6.4. Subnuclear protein dynamics—applying masks:

In addition to studying recruitment of proteins to DNA damage sites, the use of other fluorescent tools can dramatically expand the possible applications of live-cell single-molecule imaging in the study of DNA repair. The nucleus is composed of many sub-compartments that possess unique functions and have different biochemical characteristics including the nucleoplasm, nucleolus, Cajal bodies, heterochromatin and euchromatin, and PML bodies [89,90]. In addition to performing nucleus-wide analyses of protein dynamics and chromatin binding, this approach can also be adapted to analyze proteins within specific nuclear sub-compartments (Figure 4). Examining the behavior of proteins in one or more specific nuclear sub-compartments requires the expression of a marker protein tagged with a spectrally distinct fluorescent protein, which can be used to generate a mask of the sub-nuclear compartment. The mask can used to filter out molecule trajectories in the region of interest that can then be analyzed in any of the ways described above. This approach was recently used to analyze the subnuclear dynamics of endogenously HaloTagged Telomerase reverse transcriptase (TERT) inside and outside of nucleoli by marking nucleoli with GFP-NPM1, which revealed new insights into the mechanism of telomerase assembly in human cells (see video S6-S10) [91]. While many HaloTagged DNA repair proteins are exclusively or predominantly localized in the nucleus, many others are also localized to the cytoplasm (e.g., HaloTagged Rev7, Shld1, Shld2, Shld3, XLF, and Polθ) [36,71,79]. To specifically analyze the nuclear function of these proteins in DNA repair, a nuclear mask was used to filter out and exclude cytoplasmic protein trajectories from analysis. This can be accomplished in several possible ways. First, a brightfield image of the cell acquired at the time of single-molecule imaging can be used to manually create a nuclear mask. Second, a mask can be generated using a fluorescent image of the nucleus marked either by incubating cells prior to imaging with a membrane-permeable DNA stain such as Hoechst 33342 or by transiently expressing a fluorescently tagged histone protein such as H2B. In our previous work, these approaches were used to isolate and analyze nuclear trajectories of HaloTagged Rev7, SHLD1, SHLD2, SHLD3, XLF, and Polθ which have substantial nuclear and cytoplasmic fractions [36,71,79].

6.5. Analysis of protein dynamics inside and outside of repair foci:

Next, it may be of interest to many research groups to compare the behavior of proteins inside and outside of DNA repair foci (Figure 4). This may also be particularly interesting for the study of proteins incorporated within biomolecular condensates or that undergo liquid-liquid phase separation (LLPS) at DNA damage sites such as 53BP1, RPA, Rad52, and TopBP1 [63,92–94]. It is possible to analyze a protein-dynamics of a HaloTagged DNA repair factor at the single molecule and simultaneously use the same protein as a marker for DNA repair foci. Initial low-density labeling marks single particles which can be followed by high-density labeling to mark the remaining cellular protein pool with a spectrally distinct HaloTag ligand. Using the masking approach described above, trajectories can then be separated into molecules within our outside of DNA repair foci. As a proof-of-principle for this type of analysis, low and high-density labeling of Halo-53BP1 was combined after induction of DSBs with zeocin (see video 24) [36]. Static 53BP1-molecules were identified which were associated with DNA repair foci and interestingly also outside of the foci, suggesting that 53BP1 chromatin binding is not exclusively restricted to DNA repair foci. Therefore, simultaneous low and high-density labeling of a tagged DNA repair factor with different HaloTag ligands is a powerful approach to analyze protein dynamics in distinct sub-nuclear compartments and for the in vivo analysis of LLPS and its role in DNA repair foci formation at the single-molecule level.

6.6. Proximity activation to analyze complex formation:

Finally, it is quite challenging to assess interactions or co-diffusion of two different DNA repair factors fused with distinct tags (e.g. HaloTag and SNAP-tag) at the single-molecule level due to low-density labeling. For example, if two proteins constitutively interact in cells and only 1% of both proteins are labeled the chance of observing co-diffusion of two labeled proteins is 1:10,000. However, the recent development of proximity assisted photoactivation (PAPA) could potentially overcome this limitation (Figure 4) [95]. PAPA relies on the transfer of energy from a low wavelength fluorophore to reactivate fluorescence a longer wavelength fluorophore from a dark state in close proximity in a manner similar to Förster resonance energy transfer (FRET) but with a greater working distance. In contrast to FRET, PAPA does not result in the emission of fluorescence from the acceptor dye, but rather reactivates it for subsequent imaging at the single-molecule level (see video 1) [95]. To analyze complex formation by two DNA repair factors protein 1 would be fused to HaloTag and Protein 2 would be fused to SNAP-tag. Next, the SNAP-tagged protein is densely labeled with a lower wavelength ligand such as SNAP-JF549 while the other protein is labeled at a single-molecule density with a far-red fluorescent ligand (JF646). The far-red fluorophore is then converted into a dark state by intensely illuminating the sample with far-red light. Excitation of the JF549 labeled protein with a 561 nm wavelength then leads to re-activation of the JF646 labeled protein, but only if the two proteins are in close proximity. The reactivated protein can then by analyzed at the single-molecule level. Essentially this approach allows specific analysis of a subset of protein molecules that are in complex with another factor, which can be leveraged to analyze protein-protein interaction networks and to define the biophysical properties of DNA repair factor complexes.

7. Concluding remarks

Live-cell single-molecule imaging joins a growing field of imaging approaches used to study DNA repair including wide-field fluorescence and super-resolution microscopy (e.g. Stimulated Emission Depletion (STED) Microscopy, Structured Illumination Microscopy (SIM), and Stochastic Optical Reconstruction Microscopy (STORM)). While these approaches can be powerful tools to study DNA repair in cells, they are limited in their ability to provide information about the kinetics of protein behavior with single-molecule resolution. Considering the limitations of these approaches, live-cell single-molecule imaging represents a complementary method that enables new investigations of DNA repair by providing high spatial and temporal resolution of individual protein dynamics in living cells. This imaging technique will undoubtedly open up entirely new areas of research in the DNA repair field by allowing for the direct observation and monitoring of proteins and events that have been previously unobservable.