1. Physics of Living Systems
Download icon

Single molecule microscopy reveals key physical features of repair foci in living cells

  1. Judith Miné-Hattab  Is a corresponding author
  2. Mathias Heltberg
  3. Marie Villemeur
  4. Chloé Guedj
  5. Thierry Mora
  6. Aleksandra M Walczak
  7. Maxime Dahan
  8. Angela Taddei  Is a corresponding author
  1. Institut Curie, PSL University, Sorbonne Université, CNRS, Nuclear Dynamics, France
  2. Laboratoire de Physique de l’Ecole Normale Supérieure, PSL University, CNRS, Sorbonne Université , Université de Paris, France
  3. Institut Curie, PSL University, Sorbonne Université, CNRS, Physico Chimie Curie, France
  4. Cogitamus Laboratory, France
Research Article
  • Cited 0
  • Views 752
  • Annotations
Cite this article as: eLife 2021;10:e60577 doi: 10.7554/eLife.60577

Abstract

In response to double strand breaks (DSB), repair proteins accumulate at damaged sites, forming membrane-less sub-compartments or foci. Here we explored the physical nature of these foci, using single molecule microscopy in living cells. Rad52, the functional homolog of BRCA2 in yeast, accumulates at DSB sites and diffuses ~6 times faster within repair foci than the focus itself, exhibiting confined motion. The Rad52 confinement radius coincides with the focus size: foci resulting from 2 DSBs are twice larger in volume that the ones induced by a unique DSB and the Rad52 confinement radius scales accordingly. In contrast, molecules of the single strand binding protein Rfa1 follow anomalous diffusion similar to the focus itself or damaged chromatin. We conclude that while most Rfa1 molecules are bound to the ssDNA, Rad52 molecules are free to explore the entire focus reflecting the existence of a liquid droplet around damaged DNA.

Introduction

The cell nucleus contains membrane-less sub-compartments inside which specific proteins are more concentrated than elsewhere in the nucleus (Meldi and Brickner, 2011; Miné-Hattab and Taddei, 2019; Rippe, 2007). Such regions of high local protein concentration, called ‘foci’, are hypothesized to help proteins coordinate and collectively perform their function (Banani et al., 2017; Cabianca et al., 2019; Chubb and Bickmore, 2003; Hübner and Spector, 2010; Miné-Hattab and Taddei, 2019). The formation of these foci at the right place in the nucleus and within a well-defined time window is essential for the functioning of the cell. Here, we focus on repair sub-compartments formed in response to double strand breaks (DBS) (Lisby et al., 2004). DNA repair is an essential process for preserving genome integrity. Among the different kinds of DNA damages, DSBs are the most genotoxic (Iliakis et al., 2004). Failure to repair such lesions leads to genomic instability or cell death, and mutations in DNA repair genes lead to diseases such as Werner, Bloom and other cancer predisposition syndromes (Rassool, 2003; Thorslund and West, 2007). Eukaryotic organisms use mainly two major mechanisms to repair DSBs: non-homologous end-joining (NHEJ) and homologous recombination (HR) (Wyman and Kanaar, 2006). HR occurs primarily in S/G2 phase cells and uses an undamaged homologous DNA sequence as a template for copying the missing information (Lisby and Rothstein, 2015). The biochemistry and the genetics of DSB repair by HR have been extensively investigated in vitro (Kaniecki et al., 2018; Miné et al., 2007) and in vivo (Lisby and Rothstein, 2009; Sánchez et al., 2017). HR is orchestrated by mega-Dalton multi-protein complexes of 500 to 2000 proteins that colocalize with the DSB (Lisby et al., 2001). In living cells, these protein centers can be visualized as fluorescent foci using tagged HR proteins. Among the proteins occupying these centers are the enzymes of the highly conserved Rad52 epistasis group, including Rad51, Rad52 and Rad54 (Symington et al., 2014). When a DSB forms, the 5’ ends of the DNA break are resected by nucleases to yield 3’ single-stranded DNA (ssDNA) tails that are rapidly coated by the RPA complex. The Rad52 protein, the functional analog of human BRCA2 in yeast, then stimulates the removal of RPA and recruits the recombinase Rad51 to the ssDNA tail on which it polymerizes (Symington et al., 2014). The Rad51-ssDNA complex, called a nucleo-filament, has the capacity to search and identify a region of homology and to promote strand invasion of the homologous duplex DNA (Bordelet and Dubrana, 2019). Once homology is found, the invading strand primes DNA synthesis of the homologous template, ultimately restoring genetic information disrupted at the DSB. How repair foci and more generally membrane-less sub-compartments are formed, maintained and disassembled at time scales relevant for their biological function remains unknown. Several models are intensively debated in the literature to understand the nature of membrane-less sub-compartments (Erdel and Rippe, 2018; Hyman et al., 2014; McSwiggen et al., 2019a; Miné-Hattab and Taddei, 2019; Strom et al., 2017). In particular, recent works proposed that repair proteins exhibit some properties of liquid-like droplets such as Rad52 in Saccharomyces cerevisiae yeast (Oshidari et al., 2020) and 53BP1 in human cells (Altmeyer et al., 2015; Kilic et al., 2019; Pessina et al., 2019).

Here, we present new insights into the dynamics and the nature of repair foci using Single Particle Tracking (SPT) and Photo Activable Localization Microscopy (PALM) in Saccharomyces cerevisiae cells. Watching how proteins move and interact within a living cell is crucial for better understanding their biological mechanisms. SPT is a powerful technique that makes these observations possible by taking ‘live’ recordings of individual molecules in a cell at high temporal and spatial resolution (50 Hz, 30 nm) (Dolgin, 2019; Manley et al., 2008; Oswald et al., 2014). Based on the way individual molecules move in vivo, SPT allows for (i) sorting proteins into subpopulations characterized by their apparent diffusion coefficients, (ii) quantifying their motion, (iii) estimating residence times in specific regions of the nucleus, (iv) and testing the existence of a potential attracting or repelling molecules within distances smaller than the diffraction limit. To complement this approach, PALM allows for measuring the position of molecules at 30 nm resolution to establish density maps of the molecules of interest.

Using SPT and PALM in the presence or absence of DSB, we have assessed the dynamics, for the first time at the single molecule level, of 2 repair proteins: Rad52 and the ssDNA-binding protein Rfa1, a subunit of the RPA complex. We find that inside repair foci, Rad52 molecules are surprisingly mobile: they diffuse an order of magnitude faster than Rfa1, the whole focus itself, and damaged DNA, indicating that most of the Rad52 molecules are not bound to damaged DNA. Instead, Rad52 explores the volume of the focus through confined motion and exhibits a sharp change of diffusion coefficient when entering or escaping foci. In response to multiple DSBs, Rad52 form larger foci than the ones formed upon a single DSB. When foci size is larger, the confinement radius of individual Rad52 inside the focus scales with focus size. Our statistical analysis of single molecule trajectories reveals the existence of an attractive potential maintaining Rad52 molecules within a focus. Altogether, our results using single molecule microscopy indicate that while Rfa1 diffuses similarly to damaged DNA, Rad52 motion exhibits physical properties consistent with diffusion within a Liquid-Liquid Phase Separated droplet.

Results

Rad52 exhibits two diffusive behaviors in the absence of DSB

We first investigated the mobility of individual Rad52 molecules in vivo in the absence of DNA damage by SPT. To image Rad52 without altering its endogenous expression level, we generated haploid cells expressing the endogenous Rad52 fused to Halo (Figure 1A, Figure 1—figure supplement 1 and Materials and methods). Prior to visualization on a PALM microscope (see Materials and methods), exponentially growing cells were incubated with fluorescent and fluorogenic JF646, a dye emitting light only once bound to Halo (Grimm et al., 2015). We used a low concentration of JF646 allowing for the observation of individual molecules (Ranjan et al., 2020Figure 1—figure supplement 2). Rad52-Halo bound to JF646 (Rad52-Halo/JF646) were visualized at 20 ms time intervals (50 Hz) in 2-dimensions during 1000 frames until no signal was visible. A typical individual cell is shown in Figure 1B. After detection and tracking (see Materials and methods), we calculated density and displacement maps of Rad52 molecules (Figure 1B). Since Rad52 tracking is performed in 2-dimensions, molecules are observable as long as they stay within the focal plan representing a z-section of about 400 nm (Hansen et al., 2018Figure 1C). After examining 23 cells in S/G2 phase of the cell cycle as assessed by the presence and the size of a bud on the transmission image, we obtained an exponentially decreasing distribution of trace lengths with an average trajectory length of 7.5 frames, the longest trace reaching 65 frames (1.3 s) (Figure 1D). Since the half-life time of JF646 is 2.1 s (Figure 1—figure supplement 3), the short trace lengths observed here are due to molecules moving out of the observable z-section and not the photo bleaching of the JF646 dyes.

Figure 1 with 5 supplements see all
Rad52 diffusion observed at the single molecule level, in the absence of DNA damage.

(A) Design of the experiment. Haploid yeast strain harboring Rad52 endogenously fused to Halo and incubated with fluorogenic JF646 dyes are visualized by Single Particle Tracking (SPT). Individual Rad52-Halo/JF646 are tracked at 20 ms time intervals (50 Hz), movies are 1000 frames long. (B) Typical S/G2 haploid cells harboring Rad52-Halo coupled with JF646 (Rad52-Halo/JF646). From left to right: transmission image; time-projection of a typical SPT acquisition (see Materials and methods); Rad52 detections: each spot represents a single detection of Rad52-Halo/JF646, the color map indicates the number of Rad52 neighbors inside a 50 nm radius disk; Rad52 traces: each line represents the trajectory of a detection, the color map indicates the distance covered in 20 ms. The bar scale represents 1 μm. This typical nucleus exhibits 682 detections and 129 traces. (C) Schematic of SPT experiment: individual molecules are tracked in 2-dimensions. The blue rectangle represents the observable z-section. (D) Distribution of tracks length of Rad52-Halo/JF646 in the absence of DNA damage. The histogram combines 23 S/G2 phase cells representing 964 traces (mean length of 7.5 fames), and 6252 displacements of 20 ms. S/G2 cells were selected based on the presence and the size of a bud on the transmission image. (E) Probability Density Function (PDF) of Rad52-Halo/JF646 molecules of haploid S/G2-phase cells in the absence of DNA damage. The time interval is 20 ms. Red: 2-population fit; dashed purple: slow population; dashed orange: fast population; Black: 1-population fit. (F) Cumulative Density Function (CDF) of Rad52-Halo/JF646 molecules in haploid S/G2-phase cells in the absence of DNA damage. Dashed green line: data; Red: 2-population fit; Black: 1-population fit. The p-values are indicated in parenthesis (see Materials and methods). (G) Probability Density Function (PDF) of Rad52-Halo/JF646 molecules in haploid S/G2-phase cells in the absence of DNA damage, for time-points spaced by 20, 40, 60, 80, 100 and 120 ms. The lines show the 2-population fit performed on each PDF. (H) Diffusion coefficient D (μm2/s) obtained from the fits shown in G for each Δt. The fits (dotted lines) represent the expected fraction calculated from simulations (see Materials and methods). (I) Typical cell harboring Rad52-Halo/JF646. Left: transmission image; right: time-projection of a typical SPT acquisition (1000 frames). (J) Displacement histogram of S/G2 cells (blue) versus G1 cells (red). (K) Probability Density Function (PDF) of Rad52-Halo/JF646 molecules in haploid G1 cells in the absence of DNA damage. Black: 1-population fit; Red: 2-population fit; dashed purple: slow molecule; dashed orange: fast molecules.

To estimate the apparent diffusion coefficient of Rad52 in the absence of DSB, we calculated their displacement histogram (Hansen et al., 2017; Klein et al., 2019; Stracy and Kapanidis, 2017). This method allows us to test whether Rad52 molecules exhibit a single diffusive regime across the nucleus or if they exist in several subpopulations characterized by distinct diffusion coefficients (see Materials and methods). By fitting the displacement histogram with 1- and 2-population fits, we observed that Rad52 exhibits two distinct diffusive behaviors (Figure 1E–F, p=0.0001 and p=0.67, two-sided Kolmogorov-Smirnoff (KS) test for the 1- and 2-population fits respectively, see Materials and methods). We obtained D1/S-G2 = 1.16 ± 0.08 μm2/s, D2/S-G2 = 0.28 ± 0.02 μm2/s as the best-fit diffusion coefficients (see Materials and methods), with the fraction of slow molecules being 0.62 ± 0.04. To accurately estimate the apparent diffusion coefficient of Rad52, we then calculated the displacement histograms for several frame rates (20, 40, 60, 80 and 100 ms time intervals) and fitted these experimental histograms with a t2-population fit (Figure 1G). When mobility is observed at larger time intervals, mean square displacement per unit time decreases, which stems from the confinement of molecules within the nuclear space (we checked that confined diffusion explained the trend better than anomalous diffusion, see Figure 1—figure supplement 4). Fitting these points with a straight line, we extracted the diffusion coefficients to be D1/S-G2 = 1.16 ± 0.06 μm2/s, D2/S-G2 = 0.29 ± 0.02 μm2/s, which agree with the measurements obtained from the displacement histogram with Δt = 20 ms (Figure 1G–H).

To investigate the origin of the two Rad52 populations observed in the nucleus in the absence of DBS, we first asked whether the existence of a slower population of Rad52 is due to its transient association with RPA present at replication forks during S phase. To test this hypothesis, we measured Rad52 mobility in G1 cells. We obtained similar displacement histograms and diffusion coefficients in G1 as in S/G2 cells (Figure 1I, J and K, D1,G1 = 1.08 ± 0.07 μm2/s, D2,G1 = 0.27 ± 0.03 μm2/s) with similar proportions. Thus, Rad52 diffusion does not change significantly during the cell cycle and the two observed Rad52 populations cannot be simply explained by free versus Rad52 molecules transiently associated with the replication fork. We then examined if the two observed populations could correspond to the monomeric and multimeric forms of Rad52 observed in vitro (Saotome et al., 2018; Shinohara et al., 1998). We calculated the predicted diffusion coefficient of Rad52 monomers and multimers using the Stokes-Einstein equation (see Materials and methods). For that, we first estimated the dynamic viscosity inside yeast nuclei by measuring the mobility of free NLS-Halo/JF646 (Figure 1—figure supplement 5). We found a diffusion coefficient of DfreeHalo-NLS = 1.90 ± 0.06 μm2/s, consistent with previous studies by FRAP (Larson et al., 2011). From this measurement, we extracted a dynamic viscosity in the yeast nucleus of 0.122 ± 0.004 Pa.s. Using this measured viscosity, the predicted diffusion coefficients of Rad52 monomers and multimers are Dmonomer = 1.20 ± 0.04 μm2/s and Dmultimer = 0.28 ± 0.01 μm2/s (see Materials and methods), consistent with the two populations D1 = 1.16 ± 0.08 and D2 = 0.28 ± 0.02 obtained experimentally. Thus, our measurements of Rad52 diffusion by SPT are consistent with the existence of a mixed population of Rad52 inside nuclei,~2/3 being multimers.

Inside foci, Rad52 exhibits a slower diffusion coefficient and a confined motion

We next investigated how Rad52 diffusion is affected in response to DSB. A single DSB was induced at an I-SceI target site inserted at the LYS2 locus in a strain harboring Rad52-Halo (Figure 2A and Materials and methods). Of note, in this setting one can follow only the first step of HR, as no donor sequence is available in the genome. Cells were observed 2 hr after inducing the expression of the endonuclease I-SceI driven by the inducible GAL promoter. In these conditions, the I-SceI site is cleaved in 95% of the cells (Batté et al., 2017) leading to the formation of Rad52 foci (Miné-Hattab and Rothstein, 2012) but D-loop extension is not started yet (Barzel and Kupiec, 2008; Piazza et al., 2019). During the second hour of DSB induction, fluorogenic JF646 were added similarly to the previous experiment without DSB and cells were imaged with the same illumination conditions as in the absence of DSB. Inside foci, we observe at most o1detection of Rad52-Halo/JF646 per frame during the movie, a necessary condition to avoid mislinking when tracking single molecules (Figure 1—figure supplement 2). Unlike the experiment in the absence of DSB, the Rad52 density map exhibits a strong accumulation of molecules corresponding to the repair focus in G2/S cells. S/G2 cells harboring a Rad52 focus were selected for further analysis (Figure 2B). As shown by the displacement map (Figure 2B, right panel), molecules localized in Rad52 foci cover shorter distances in a 20 ms step; since the tracking is performed in two dimensions, we can observe molecule for a longer time before they move out of the focal plane. Thus, following DSB, the histogram of trace lengths presents a longer tail than before damage (Figure 2C and Figure 2—figure supplement 1).

Figure 2 with 5 supplements see all
Rad52 diffusion at the single molecule level in the presence of a single I-SceI induced DSB.

(A) Design of the experiment. We induced a single DSB in haploid cells harboring Rad52 endogenously fused to Halo as well as an I-SceI cut site inducible under galactose promoter. The DSB is induced for 2 hours and fluorgenic JF646 dyes are added to the medium during the last hour of incubation prior visualization by SPT.for 2 hr and fluorogenic JF646 dyes are added to the medium during the last hour of incubation prior visualization by SPT. Individual Rad52-Halo/JF646 are tracked at 20 ms time intervals (50 Hz), during 1000 frames. Only cells harboring a Rad52 focus are analyzed. (B) Typical S/G2-phase haploid cells harboring Rad52-Halo/JF646. From left to right: transmission image; time-projection of a typical SPT acquisition; Rad52 detections: each spot represents a single detection of Rad52-Halo/JF646, the color map indicates the number of Rad52 neighbors inside a 50 nm radius disk; Rad52 traces: each line represents the trajectory of a detection, the color map indicates the distance in μm covered in 20 ms. The bar scale represents 1 μm. This particular nucleus exhibits 682 detections and 129 traces. (C) Distribution of tracks length of Rad52-Halo/JF646 in the presence and the absence of a single DSB. The histogram combines 27 S/G2 phase cells, all of them harboring a Rad52 focus, representing 1061 traces (mean length of 8 fames), and 8495 displacements of 20 ms time-intervals. (D) Probability Density Function (PDF) of Rad52-Halo/JF646 molecules in haploid S/G2-phase cells in the presence of a single DSB. The time interval is 20 ms. Green: Rad52 data (27 cells, 1061 trajectories); Black: 1-population fit; Red: 2-population fit; Yellow: 3-population fit. (E) Cumulative Density Function (CDF) of Rad52-Halo/JF646 molecules in haploid S/G2-phase cells in the presence of single DSB. Dashed green line: data; Black: 1-population fit; Red: 2-population fit; Yellow: 3-poulations fit. The p-values are indicated in parenthesis (see Materials and methods). (F) Left: density map of a typical nucleus following the induction of a single DSB. The nucleus is divided in two zones based on a density threshold: the Rad52 focus (highlighted in red) and the rest of the nucleus (highlighted in blue). Right: displacements histograms of trajectories contained inside foci (red) versus outside foci (green). If a trajectory crosses the focus boundary, it is cut into two parts: the part inside and the part outside of the focus. The blue histogram represents the displacement of Rad52 molecules in the absence of DSB (shown in Figure 1E). (G) Mean Square Displacement (MSD) curve of Rad52/Halo/JF646 molecules inside foci. The dotted line shows a fit of the MSD with a confined model (see Materials and methods). (H) Effect of SUMOylation: we induced a single DSB in haploid cells expressing Rad52 endogenously fused to SUMO and Halo, as well as an I-SceI cut site at LYS2 locus inducible under galactose promoter. The DSB is induced for 2 hr and fluorogenic JF646 dyes are added to the medium during the last hour of incubation prior to visualization by SPT. Individual Rad52-Halo/JF646 are tracked at 20 ms time intervals (50 Hz), during 1000 frames. (I) Displacement histogram of traces inside foci of Rad52 wild type cells versus cells expressing Rad52-SUMO in S/G2 phase cells. A single I-SceI DSB is induced for 2 hr, and only cells harboring a Rad52 focus are analyzed. Red: Rad52 wild type; Blue: Rad52-SUMO. (J) MSD curves of individual Rad52 molecules inside foci (red, same data as Figure 3E) versus individual Rad52-SUMO molecules inside foci (pink). MSD curves are fitted with a confined model (see Materials and methods).

To estimate the apparent diffusion coefficient of Rad52 in the presence of a single DSB, we tested a 1-, 2- and 3-population fit of the Rad52 displacement histograms (Figure 2D). For the three populations, we obtained the best fitted values of D1 = 1.17 ± 0.06, D2 = 0.24 ± 0.04 and finally D3 = 0.054 ± 0.007 μm2/s (Figure 2E, p=0.0001, p=0.27 and p=0.67, two-sided Kolmogorov-Smirnoff (KS) test for the 1-, 2- and 3-population fits respectively). Only the 3-population fit is not rejected. In addition, it allows for recovering the same two populations previously observed in the absence of DSB, but supplemented by a third slow population (with diffusivity D3). The value of D3 obtained in this way is overestimated because of detection noise (for diffusion coefficients D < 0.1 μm2/s, we use mean square displacement analysis allowing us to subtract the noise, see Methods and below). A quick inspection of the spatial distribution of step sizes in the nuclear space indicates that the slowest population is found in the zone where Rad52 molecules are dense, that is within the repair focus (Figure 2B, displacement map). Using a density threshold to separate the trajectories located within the focus from the rest of the nucleus (Figure 2F, left), we then calculated the displacement histograms of the corresponding trajectories (Figure 2F, right). We confirm that Rad52 molecules outside foci exhibit similar displacements as in the absence of damage, whereas Rad52 molecules inside the focus are less mobile (Figure 2F, right panel). We also confirmed that the diffusion of Rad52 molecules inside foci is well described by a 1-population fit (see Methods and Figure 2—figure supplement 2). To accurately estimate the diffusion coefficient of Rad52 inside the focus accounting for detection noise, and also to measure the nature of Rad52 motion, we used mean squared displacement (MSD) analysis (see Methods). The MSD quantifies the amount of space a molecule explores, and its shape versus time reveals the nature of its movement. This analysis revealed that Rad52 molecules inside foci exhibit a confined motion with a confinement radius Rc = 124 ± 3 nm and a diffusion coefficient DRad52,inside = 0.032 ± 0.006 μm2/s (see Methods and Figure 2G).

Constitutive Rad52 SUMOylation regulates its dynamics within foci

The post-translational modification by the small ubiquitin-like modifier (SUMO) protein has been proposed as a key factor regulating the composition and formation of membrane-less compartments (Banani et al., 2017). Furthermore, several studies suggest that this modification plays an important role in mitotic and meiotic recombination (Altmannova et al., 2010; Esta et al., 2013Torres-Rosell et al., 2007). We thus tested the observables accessible by SPT in a strain expressing Rad52 fused to SUMO. Despite being mildly sensitive to MMS and γ-Rays (Esta et al., 2013), Rad52 SUMOylation has several effects in vitro and in vivo. In particular, SUMOylated Rad52 exhibits lower affinity towards ssDNA and dsDNA, has reduced single-strand annealing activity (Altmannova et al., 2010), and partially destabilizes Rad51 filaments compensating the effect of Srs2 deletion (Esta et al., 2013). Using a strain expressing Rad52 fused to SUMO and Halo, as well as a single I-SceI cut-site at LYS2 (Figure 2H), we measured the mobility of Rad52-SUMO after 2 hr of DSB induction in cells harboring a focus. We found that Rad52 mobility is significantly slower in the SUMO version of Rad52 than in the wild type (p=1, t2-sided KS test, Figure 2I, Figure 2—figure supplement 3). As reported in Silva et al., 2016, in vivo, foci intensity remains similar in SUMOylated Rad52 (Figure 2—figure supplement 3). Inside foci, SUMOylated Rad52 also exhibits confined diffusion although with slower diffusion coefficient (DRad52-SUMO, inside = 0.009 ± 0.002 μm2/s, Figure 2J and Supplementary file 1) and a smaller confinement radius than wild type Rad52 (Rc Rad52-SUMO = 83 ± 12 nm versus Rc Rad52 = 124 ± 3 nm). Overall, SUMOylated foci are smaller, denser than in wild type cells, and molecules inside foci are less mobile.

Inside foci, individual Rad52 molecules diffuse faster than the repair focus and the ssDNA-binding protein Rfa1

We next compared the mobility of individual Rad52 molecules inside foci with the mobility of the focus itself (Figure 3A). After inducing a single DSB for 2 hr in cells harboring Rad52-mMaple, we used high photo-activation illumination to simultaneously activate all Rad52-mMaple and image the repair focus as a single entity (Figure 3B). Using both displacement histograms and MSD analysis (Figure 3C–E and Figure 3—figure supplement 1), we found that Rad52 molecules diffuse 6.4 times faster than the whole focus (DRad52 inside = 0.032 ± 0.006 μm2/s versus Dwhole focus = 0.005 ± 0.002 μm2/s, assuming normal diffusion with a time difference of 20 ms for the whole focus). The diffusion coefficient we found is consistent with the one previously reported in response to a single DSB (0.002 ± 0.002 μm2/s in haploid imaged at 1 s time interval Dion et al., 2012). Furthermore, the shape of the MSD fits better with an anomalous diffusion of anomalous exponent 0.51 ± 0.05, as previously reported for damaged chromatin at that time scale (Miné-Hattab et al., 2017). Thus, our results show that individual Rad52 molecules are highly mobile within the repair focus.

Figure 3 with 1 supplement see all
Individual Rad52 molecules are more mobile than the whole focus.

(A) Schematic of the experiment: to shade in light the internal dynamics of Rad52, we compare the mobility of the entire Rad52 foci, shown in light red, with the mobility of individual Rad52 molecules located inside foci, in red. The mobility of the entire Rad52 focus is measured using a strain harboring Rad52-mMaple and a single I-SceI DSB at LYS2 (see Materials and methods). We used high photo-activation illumination to simultaneously activate all Rad52-mMaple and image the entire foci as a single large spot. Rad52 foci were tracked at 20 ms time-intervals in two dimensions. (B) Typical image of a haploid cell harboring a Rad52-mMaple focus after a 2h-induction of a single DSB. Left: transmission image; Middle: typical frame of a movie in which we see the whole focus; Right trajectory of the whole focus after analysis. The bar scale represents 1 μm. (C) Displacements histogram of the entire Rad52 foci (grey), compared with individual Rad52-Halo located inside foci (red, same data shown in Figure 2F in red). Both displacements are measured after 2 hr of galactose induction at 20 ms time intervals in 2-dimensions. F14S/G2 cells harboring a Rad52 focus are analyzed, representing 131 traces (mean length 24 frames), and 3015 translocations of 20 ms time-intervals. (D) PDF of the entire Rad52 foci calculated for 20, 40 and 60 ms time intervals. Plain lines represent a 1-population fit of the PDF. (E) MSD of the entire Rad52 foci (green) versus individual Rad52 molecules located inside foci (red, same data as Figure 2G). The MSD of individual Rad52 molecules inside foci is fitted with a confined model (dotted line) while the MSD of the entire Rad52 focus is fitted using an anomalous model (see Methods).

We then compared the mobility of individual Rad52 molecules with another component of repair foci, the Rfa1 subunit of the single-strand binding factor RPA (Symington et al., 2014; Figure 4A). Using a strain expressing Rfa1-Halo and experiencing a single I-SceI DSB, we imaged individual Rfa1-Halo/JF646 2 hr after DSB induction with the same illumination conditions as Rad52-Halo (Figure 4B). Comparing the displacement histograms of individual Rfa1 and Rad52 molecules, we found that Rfa1 is less mobile than individual Rad52 molecules inside foci (Figure 4C) and exhibits a mobility similar to whole Rad52 foci (Figure 4D). Using both Rfa1 displacement histograms and MSD analysis inside foci, Rfa1 exhibits an apparent normal diffusion coefficient DRfa1 = 0.006 ± 0.001 μm2/s (Figure 4E, assuming normal diffusion with a time difference of 20 ms). Moreover, the MSD of Rfa1 inside foci is similar to the entire focus, with an anomalous exponent of 0.56 ± 0.05 (Figure 4E and Figure 3—figure supplement 1).

Figure 4 with 1 supplement see all
Rad52 and Rfa1 behavior strongly differs inside foci (A) Simplified view of a DSB repaired by HR where RPA is shown.

(B) Typical S/G2-phase haploid cells harboring Rfa1-Halo/JF646 after 2 hr if galactose induction. From left to right: transmission image; Time-projection of the whole movie; Rfa1 density map: each spot represents a single detection of Rfa1-Halo/JF646, the color map indicates the number of Rfa1 neighbors inside a 50 nm radius disk; Rfa1 displacement map: each line represents the trajectory of a detection, the color map indicates the distance in mm covered in 20 ms. The bar scale represents 1 μm. (C) Displacement histograms of individual Rfa1 molecules (yellow) and individual Rad52 molecules (red) inside foci. The time interval is 20 ms. 29 S/G2 cells were analyzed, representing 621 traces (mean length of 12.4 frames) and 7095 displacement of 20 ms time intervals. (D) Displacement histograms of individual Rfa1 molecules inside foci (yellow) and whole Rad52 foci (grey). The time interval is 20 ms. (E) MSD of individual Rfa1 molecules inside foci (yellow), individual Rad52 molecules inside foci (red, same data as Figure 3E) and the whole Rad52 focus (green, same data as Figure 3E). MSDs of Rfa1 molecules and the whole Rad52 focus are fitted with an anomalous model while the MSD of individual Rad52 is fitted with a confined model. (F) Typical haploid cells expressing Rad52-RFP and an I-SceI inducible DSB at LYS2, observed by wide field microscopy. First panel: a DSB is induced for 2 hr; second panel: a DSB is induced for 2 hr including 30 min with 10% hexanediol and 10 μg/ml of digitonin; third panel: a DSB is induced for 2 hr and cells are fixed with fixation with 4% paraformaldehyde or 10 min; fourth panel: quantification of the images presenting the percentage of S/G2 phase cells with a Rad52 focus. 200 cells are analyzed in each condition. The bar scale represents 1 μm. (G) Typical haploid cells expressing Rfa1-RFP and an I-SceI inducible DSB at LYS2, observed by wide field microscopy. Cells are images in the same conditions as in figure (F). Since Rfa1 form several foci per nucleus, the quantification shows the number of foci per nucleus (0bar 3 foci). The bar scale represents 1 μm.

Overall, we found that individual Rad52 diffuses approximately six times faster than the whole focus or individual Rfa1 molecules, and exhibits confined diffusion while Rfa1 and the repair focus follow anomalous diffusion.

Rad52 high internal mobility and type of motion inside foci is conserved in diploid cells

We then verified if our observations held in diploid cells in which a repair template is available. We investigated the mobility of individual Rad52-Halo/JF646 molecules in living diploid cells harboring an I-SceI cut site at one of the LYS2 loci (see Supplementary file 2 and Figure 2—figure supplement 5). Using the same illumination settings as in haploid cells, we measured the mobility of Rad52-Halo/JF646 after 2 hr of DSB induction in cells harboring a Rad52 focus. Consistent with the results obtained in haploid cells, Rad52 exhibits three distinct diffusive behaviors, the slowest population being localized in the repair focus (Figure 2—figure supplement 5, D1 = 1.06 ± 0.05, D2 = 0.22 ± 0.02 and D3 = 0.033 ± 0.002 μm2/s, p=<10−7 0, p=10−7 and p=0.34, two-sided KS test for the 1-, 2- and 3-population fits respectively). The MSD analysis revealed a confined motion inside foci with a confinement radius similar to the one obtained in haploid cells (Rc Haploid, inside foci = 124 ± 3 nm and Rc Diploid inside foci = 125 ± 3 nm, Figure 2—figure supplement 5). Also similarly to the haploid cells, the diffusion coefficient of molecules inside foci is eight fold faster that the focus itself (Ddiploid, inside = 0.016 ± 0.001 μm2/s, Figure 2—figure supplement 5, versus Dwhole focus = 0.002 ± 0.001 μm2/s from Miné-Hattab et al., 2017, assuming normal diffusion with a time difference of 20 ms).

We thus conclude that the high internal dynamics of Rad52 inside foci, and the capacity of Rad52 to explore the whole focus conserved in diploid cells where DNA damage can be repaired.

Rad52 but not Rfa1 foci are partially dissolved under 1.6-hexanediol treatment

The aliphatic alcohol hexanediol has been proposed as a tool to differentiate between liquid-like and solid-like assemblies in living cells (Kroschwald et al., 2017). We thus tested the effect of 1.6-hexanediol treatment both on Rad52 and Rfa1 foci formed in response to a single I-SceI induced DSB. Aliphatic alcohol 1,6-hexanediol perturbs weak hydrophobic interactions. Several in vivo studies in yeast and mammalian cells showed that 1.6-hexanediol treatment dissolves dynamic liquid-like assemblies such as P bodies, whereas solid-like assemblies, such as protein aggregates and cytoskeletal assemblies, are largely resistant to hexanediol (Kroschwald et al., 2017; Strom et al., 2017). We first verified the efficiency of 1.6-hexanediol to dissolve liquid-like assemblies using a control strain expressing fluorescent P-bodies (Edc3-GFP) as described in Kroschwald et al., 2017 (See Figure 4—figure supplement 1 and Methods). After 30 min of treatment, the number of Edc3 foci per cell significantly decreases (Figure 4—figure supplement 1 and Methods). We thus applied the same treatment on strains expressing Rad52-FRP or Rfa1-RFP and harboring a single I-SceI cut-site, and we imaged these cells on a regular wide field microscope (Figure 4F and G respectively). Three conditions were analyzed: (i) cells after 2 hr of DSB induction (left panel), (ii) cells after 2 hr of DSB induction including 30 min of 1.6-hexanediol treatment (middle panel) and (iii) cells after 2 hr of DSB induction followed by 10 min of fixation with paraformaldehyde (see Methods). We found that the percentage of S-phase cells containing a Rad52 focus significantly decreases after 1.6-hexanediol treatment (22 ± 2% to 8 ± 3% respectively), consistent with recent work (Oshidari et al., 2020). In contrast, the percentage of Rfa1 foci remains similar with or without hexanediol treatment (Figure 4G, left and middle panels). We also noticed that after fixation with several classical fixation methods (Methods), Rad52 foci are not visible and Edc3 foci are significantly less abundant (Figure 4F and Figure 4—figure supplement 1 respectively), while Rfa1 foci are not altered by fixation (Figure 4G).

In conclusion, Rad52 but not Raf1 foci show a behavior similar to P-bodies in response to 1.6-hexanediol or paraformaldehyde treatments indicating different modes of foci formation for these two proteins, in agreement with their difference in mobility.

Rad52 molecules change diffusion coefficient when entering or escaping repair foci

The way molecules diffuse around the boundary of membrane-less sub-compartments is crucial for defining their physical properties. We analyzed Rad52 diffusion following DSB in more detail (Figure 5A). First, we estimated the residence time of Rad52 molecules inside foci by calculating the survival probability of Rad52 molecules in a focus. Importantly, by comparing this curve with the bleaching time of the JF646, we checked that JF646 bleaching is not a limiting factor (Figure 5B). We found that the mean residence time of Rad52 inside foci, defined as the integral of the survival probability curve, is ~240 ms while Rfa1 molecules stay longer (560 ms) (Figure 5B).

Figure 5 with 1 supplement see all
Rad52 diffusion coefficient changes when molecules enter and escape repair foci.

(A) Illustration of the three categories of Rad52 traces observed in response to a single DSB in the nucleus: (i) traces staying inside repair foci (plain red), (ii) traces crossing foci boundaries (dotted lines), with red-dotted lines for the part inside foci and blue-dotted lines for the part outside; (iii) traces staying outside of the foci (plain blue). (B) Survival probability curve of Rad52 molecules inside foci (red), Rfa1 (yellow) and renormalized bleaching curve of the JF646 (black) (see Figure 1—figure supplement 3). (C) Light red: displacement histogram of traces represented as dotted red lines in Figure 4A (travelers, part inside foci). Dark red: displacement histogram of traces represented in plain red in Figure 4A (traces inside foci). (D) Light blue: displacement histogram of traces represented as dotted blue lines in Figure 4A (travelers, part outside foci). Dark blue: displacement histogram of traces represented in plain blue in Figure 4A (traces outside foci). (E) Left: Step size for traces inside the focus (green), crossing (blue), and outside the focus (red). The x-axis is squeezed so that all traces take the same space, in order to visually compare the step sizes. Above: traces starting inside and ending outside. Below: traces starting outside and ending inside the focus. Right: Bar plot showing the estimated diffusion coefficient calculated from all the traces.

So far, we compared Rad52 diffusion inside versus outside repair foci; however, some trajectories cross the focus boundary, and those were separated in two parts. To understand how Rad52 diffuses around foci boundary, we first quantified the different categories of traces: (i) traces staying inside repair foci during the time of the acquisition, (ii) traces crossing focus boundaries and (iii) traces staying outside foci (Figure 5A). Among the first 2 categories of traces, 70% cross a focus boundary at least once during the 20 s movie, while 30% stay inside foci, indicating that Rad52 foci are in constant exchange with the rest of the nucleus. As a comparison, only 33% of Rfa1 molecules cross a focus boundary. Of course, these exchange rates are specific to our acquisition settings.

We then focused only on Rad52 trajectories crossing repair focus boundaries, which we call ‘travelers’ (dotted lines in Figure 5A). To test whether Rad52 molecules change their diffusive behavior when entering or escaping repair foci, we calculated the displacement histograms of these traces, distinguishing portions inside and outside foci (Figure 5A, dotted red and dotted blue traces respectively). We observed that Rad52 molecules change diffusive behavior sharply when crossing the focus boundary. Molecules crossing into foci (category 2) diffuse similarly to molecules staying inside foci, as long as they are in the focus (Figure 5C); when crossing the boundary outward, they start diffusing like molecules outside of foci (Figure 5D). To have a visual representation of this change in diffusion, we displayed the norm of the displacement vectors over time of all travelers, centering all the traces when they cross the boundary (Figure 5E). In other words, Rad52 molecules clearly change diffusion coefficient when entering or escaping repair foci, adopting the diffusive behavior of the surrounding Rad52 molecules in their environment.

Following multiple DSBs, Rad52 confinement scales with focus size

Following the induction of multiple DSBs, it has been shown in yeast,Drosophila and human cells that repair sites move and collapse into larger units, or ‘clusters’ (Aten et al., 2004; Aymard et al., 2017; Chiolo et al., 2011; Krawczyk et al., 2012; Lisby et al., 2003; Neumaier et al., 2012; Oshidari et al., 2020; Schrank et al., 2018; Waterman et al., 2019), presumably to facilitate DSB repair progression by increasing the local concentration of repair proteins. However, the structure of collapsed DSBs remains unknown. We thus used PALM to investigate the structure of Rad52 foci formed in response to multiple DSBs at the single molecule level. We compared haploid strains harboring a single I-SceI site (Figure 2A) with haploid strains harboring 2 I-SceI sites on different chromosomes (Figure 6A and Methods). In both cases, DSB(s) were induced for 2 hr and we observed a single focus per nucleus in most cells, as previously described (Lisby et al., 2003). First, we measured the mobility of the entire Rad52 focus induced by a single DSB versus t2DSBs, at 20 ms time intervals. Using strains expressing Rad52-mMaple with high photo-activation, we found that both foci exhibit similar mobility (Figure 6B). This is consistent with previous studies reporting no change in chromatin diffusion coefficients when inducing more DSBs despite a dramatic increase in the chromatin confinement radius (Miné-Hattab and Rothstein, 2012).

Rad52 foci in the presence of multiple DSBs.

(A) Design of the experiment. We induced 2 I-SceI DSBs (at LYS2 and HIS3) in haploid cells harboring Rad52-Halo (see Methods). The DSB is induced for 2 hr and fluorogenic JF646 dyes are added to the medium during the last hour of incubation prior visualization by SPT. Individual Rad52-Halo/JF646 are tracked at 20 ms time intervals (50 Hz), during 1000 frames. Only cells harboring a Rad52 focus are analyzed. (B) Displacement histogram whole Rad52 foci: Blue: after the induction of a single DSB (same data as 3C); Red: after the induction of 2 DSBs. Rad52 foci were tracked at 20 ms time intervals after 2 hr of galactose induction. For the experiments following the induction of 2 DSBs, we examined 14 cells, representing 702 traces (mean length of 9.8 frames) and 6220 displacements of 20 ms time intervals. (C) Top left: Rad52 Focus size measured for 1 DSB-induced foci versus 2 DSBs-induced foci. We performed live PALM on cells harboring Rad52-mMaple strains (see Methods). Bottom left: Estimation of the number of Rad52 molecules inside 1 versus 2 DSBs-induced foci, measured by live PALM. Right: Rad52 Foci density of 1 versus 2 DSBs-induced foci. (D) Illustration of the ‘Cluster’ and the ‘Fusion’ scenarii in the case of multiple DSBs. (E) MSD of 1 versus 2 DSBs-induced foci (red and blue curves respectively). Dotted lines represented a fit of the experimental MSDs, using a model of confined diffusion. (F) Typical example of a Rad52 trajectory represented in blue. The whole focus is shown in the background using a Gaussian blur of each Rad52 detections contained in this focus. Left: 1 DSB-induced focus; right: 2 DSBs-induced focus.

We then compared the size, number of Rad52 molecules and density of foci formed in response to one versus tDSBs (Figure 6C). Since the focus size is under the diffraction limit, we used PALM in haploid cells expressing Rad52-mMaple. Since we observed that Rad52 foci are not visible after fixation with classical fixation methods, we performed PALM in living cells (Izeddin et al., 2011) (refereed as live PALM, see Methods). We found that foci induced by 2 DSBs are larger than those induced by a single DSB, (170 ± 20 nm versus 116 ± 20 nm respectively, p=1×10−4, Wilcoxon-Mann-Whitney test). In addition, the number of Rad52 molecules forming foci varies from 880 ± 600 for 1 DSB induced foci to 2300 ± 811 for 2-DSBs foci (p=8×10−4, Wilcoxon-Mann-Whitney test). Thus, focus density is indistinguishable between foci induced by versus 2 DSBs (Figure 6C, right panel, p=0.73, Wilcoxon-Mann-Whitney test).

We proposeviews of DSBs clustering illustrated in Figure 6D. In the first view, multiple DSBs cluster and stay close without merging (top figure). In the second, multiple DSBs fuse together and molecules are shared between the different DSBs (bottom figure). In the latter case, different foci merge together giving rise to a larger focus where Rad52 molecules explore the entire space; thus the radius of confinement should scale with the focus size. In the former case, Rad52 molecule should have the same confinement radius as in the situation of single double strand break if the clustered foci remain sufficiently distant. If the two foci are sufficiently close for Rad52 molecules to ‘jump’ from a focus to another one, we expect to observe bimodal trajectories, exhibiting confined diffusion in the first focus, followed by one or several larger steps and confined diffusion in a second focus. Since the average residence time inside a focus induced by 1 DSB is 240 ms (12 frames) and traces are often longer than that inside foci (Figure 5—figure supplement 1), such traces should be easily observable if present. Here, we found that upon 1 DSB, Rad52 diffuse inside foci with smaller confinement radius (124 ± 3 nm) compared to foci induced by 2 DSBs (156 ± 3 nm), supporting the second view where foci fuse (Figure 6E). Importantly, this confinement radius found by SPT is very close to the mean size of Rad52 foci previously obtained by live PALM. The correlation between focus size and confinement radius indicates that Rad52 is able to explore the entire focus space. To have a visual representation of Rad52 exploration inside foci, we overlaid a reconstruction of a focus with one of the longest Rad52 trajectories contained inside that focus. The reconstruction of the whole focus is obtained by taking all the detections inside it, except the ones belonging to the overlaid trajectory. These detections are represented by a blurred spot using a Gaussian smoothing kernel around each detection. Two typical foci are shown in this representation: a focus formed in response to a single DSB (Figure 6F, left panel) versus a focus induced by 2 DSBs (Figure 6F, right panel). This representation clearly shows that Rad52 explores the focus and that its confinement radius scales with the focus size. In addition, using this representation, we never observe bimodal trajectories, allowing us to rule out clustered but non-merged foci.

Rad52 shows attractive motion around the focus

Next, we wanted to estimate whether there was a region of attraction for the molecules of Rad52 inside the focus. For each cell, we considered a set of traces, of which a subset was defined to be inside the focus (Figure 7A). We quantified the radial movement of molecules relative to the center of the focus in terms of the change in the distance to that center between two subsequent time steps, which we denote Δr (Figure 7B). To obtain this measure we needed to estimate the position of the center of the focus. Since we observed that the focus itself was diffusing ~6 times more slowly than the Rad52 molecules, we approximated the focus as the average particle position in each analyzed trace (see Methods). Concentrating on molecules close to the boundary, we identified a region of attraction with an average movement of molecules towards the center of the focus (Figure 7C).

Rad52 shows attractive motion around the focus.

(A) Representative cell, with traces outside the focus (black) and inside the focus (blue). Cell nucleus is estimated from the traces. (B) Schematic showing the definition of the radial displacement Δr, defined by two subsequent data-points. (C) Sign of the average radial displacement as a function of particle position relative to the focus, averaged over all cells. (D) Radial displacement Δr versus r (distance to the center of the focus prior to displacement) over all cells with 1 DSB. Cyan line indicates the mean value. Black line corresponds to the expected movement for free diffusion with diffusion coefficient equal the one inside the focus (assuming the center of focus is exactly known). Red line corresponds to a free diffusion, simulated with same trace length and the same diffusion coefficient as the data traces and the same method for estimating the center of the focus. (E) Same as (D) but for cells with 2 DSBs.

To get a more detailed map of displacements, we plotted the radial movements of all particles as a function of their original position relative to the center for both 1 and 2 DSB’s (Figure 7E and F). The mean radial movement, outside the focus but close to its boundary, has negative values, which is consistent with an attractive force towards the center of the focus. This does not mean that this potential attracts molecules to the center of the focus, but it hinders the escape from the focus region. Inspired by the observation that molecules are attracted to the center of the focus, we computed the difference in the free energy. Using the estimated fraction of molecules belonging to the slow diffusion population (Figure 2), we extracted a change in the Gibbs free energy difference U0 between the inside of the focus and the surroundings. This was done by estimating the increased density of molecules inside the focus using Boltzmann’s law for the fraction of slowly diffusing molecules, p=e-U0/kBTVFV0-VF+e-U0/kBTVF, where VF is the volume of the focus, as calculated by the MSD, V0 is the volume of the observable frame. Inverting this relation for U0 and injecting the measured values of p, V0, and VF yields U0kBT-5.5 (Figure 7G). However, detection noise (as estimated in Figure 2) and low statistics far away from the focus make it difficult distinguish between a continuous potential and a surface potential.

Discussion

The nucleus contains membrane-less sub-compartments that form and disassemble according to the needs of the cell at times scales relevant for their biological function. Repair foci provide a powerful example for studying such sub-compartments because it is possible to induce their formation at will in vivo and compare the behavior of repair proteins before and after DSB induction. Following DSB, HR proteins relocalize from a diffuse nuclear distribution to a sub-nuclear focus at the DNA damaged site (Lisby et al., 2004). To understand this process, here we observed for the first time how Rad52 molecules diffuse inside living nuclei in the presence and in the absence of DSB using single molecule microscopy.

Dynamic behavior of individual Rad52 in living cells

Tracking single molecules of the repair factor Rad52 at 50 Hz, we observed that in the absence of DSB, Rad52 exhibits two distinct diffusive behaviors characterized by apparent diffusion coefficients D1 = 1.15 μm2/s, and D2 = 0.27 μm2/s with a ratio2/1 = 2/3 (see Supplementary file 1). These two populations were observed both in G1 and S phase suggesting that the slower population does not correspond to transient binding of Rad52 at replication forks. Assuming Brownian diffusion, a diffusion coefficient of 1.15 μm2/s implies that Rad52 can go across a yeast nucleus in as little as 700 ms. Moreover, tracking free Halo-NLS, we estimate the dynamic viscosity inside yeast nuclei to 0.1 Pa.s, compared to 10−3 Pa.s for water at 20°C, 0.013 P.s in Hela cells nuclei (Liang et al., 2009) and 6 Pa.s for honey. Using the Stokes-Einstein equation, we showed that the two diffusion coefficients measured for Rad52 agree with the ones predicted for the monomeric and multimeric forms of Rad52 observed in vitro (Saotome et al., 2018; Shinohara et al., 1998). Thus, single molecule tracking of Rad52 suggests that both forms co-exist in living nuclei, two thirds of them being multimers.

In response to a DSB, Rad52 accumulates at the damaged site forming highly concentrated foci (Lisby et al., n.d.). The presence of a DSB does not influence the mobility of Rad52 outside of the focus, consistent with (Essers et al., 2002), but inside foci, Rad52 exhibits an intermediate diffusive behavior, slower than free Rad52 but faster than damaged DNA (DRad52,inside = 0.032 ± 0.006 μm2/s). Taking the measured diffusion coefficient and confinement radius, Rad52 molecules inside foci start feeling the effect of its boundary at tc = Rc2/(6D)=  80 ± 3 ms, representing four time-steps of 20 ms (see Materials and methods and [Klein et al., 2019]). Assuming a multimeric form inside foci, the Rad52 diffusion coefficient measured here would be consistent with a viscosity inside the focus of 1 ± 0.1 Pa.s, and it would reach 2.4 ± 0.1 Pa.s assuming a monomeric form. These viscosities are respectively 8 to 20 times higher than the nucleus viscosity estimated before damage (νyeast nucleus = 0.122 ± 0.004 Pa.s). In human cells (U2OS), a similar viscosity inside repair foci has been found (2.5 Pa.s) by measuring the diffusion of 53BP1 molecules within foci (Pessina et al., 2019). Although human nuclei are ~25 times less viscous than yeast, the viscosity inside repair foci appears similar in yeast and humans.

Different dynamics for Rad52 and Rfa1 within repair foci

Comparing the dynamics of Rad52 with the single strand binding protein Rfa1, we observed a very different behavior in response to a single DSB. Inside foci, individual Rfa1 molecules move ~6 times slower than Rad52 molecules. Furthermore, individual molecules of Rfa1 and Rad52 exhibit a different type of motion (Figure 4 and Figure 3—figure supplement 1). Rfa1 molecules inside foci follow anomalous diffusion with an anomalous exponent (αRfa1 = 0.56 ± 0.05) consistent with the Rouse model, a behavior similar to the focus itself (αfocus0.51 ± 0.05) or to damaged chromatin when imaged at the same frequency (50 Hz) (Miné-Hattab and Rothstein, 2012). In contrast, individual Rad52 molecules are highly mobile and exhibit confined motion inside the focus (Figure 4 and Figure 3—figure supplement 1). Such differences in diffusion coefficient and type of motion between Rad52 and Rfa1 indicate that most of Rad52 molecules are not bound to the Rfa1-coated ssDNA inside foci, in agreement with (Essers et al., 2002). In addition to high internal dynamics, Rad52 exhibits a higher rate of exchange than Rfa1with the surrounding nucleoplasm. The rapid exchange rate observed here for individual Rad52 molecules is in good agreement with FLIP (fluorescence loss in photobleaching) experiments in yeast cells (Oshidari et al., 2020). Fluorescence Recovery After Photobleaching (FRAP) experiments in mammalian cells also revealed different exchange rates and residence times depending on the repair proteins: Rad54 and Rad52 inside foci rapidly exchange with the rest of the nucleus while Rad51 stays immobilized at the site of DNA damage (Essers et al., 2002).

The intermediate mobility of Rad52 molecules inside foci (faster than chromatin or Rfa1 molecules but slower than free molecules) could in principle be explained by rapid equilibrium binding of Rad52 to its target, which would effectively decrease its mobility. However, in that case we should be able to capture with SPT some intervals of time where Rad52 is bound. During these bound intervals, Rad52 should diffuse slowly with the same diffusion coefficient as Rfa1 and such events would lead to a strongly multi-modal distribution of displacements inside foci. We do not observe such multi-modal distribution inside foci in our data at 20 ms resolution. To be consistent with the smoothness of our displacement histogram, the binding specificity would have to be very weak, Kd = 1/(4 D a τ) ≫ 2 μM (Berg and Purcell, 1977), where a ≈ 10 nm is the linear size of the binding site, D ≈ 1 μm2/s, and τ is the mean bound time ≪ 20 ms, conservatively assuming diffusion-limited binding. This dissociation constant is 20 to 400 times larger than the values observed in vitro (Saotome et al., 2018). Assuming a realistic range of binding dissociation constants Kd ~1 nM to 1 μM, the mean bound time of τ = 1/(Kd D a) would range between 40 ms to 40 s. These values are all larger than our resolution and should be easily observable if we had a binding and free diffusion inside foci. Altogether, our data and analysis rather suggest that most Rad52 freely diffuse inside foci and only a very small fraction of Rad52 molecules are bound to the Rfa1-coated ssDNA. Since with single molecule experiments, we can follow only a fraction of the molecules, it is very unlikely to catch the few bound molecules in our assay.

Rad52, but not Rfa1, behavior inside foci is consistent with LLPS model

Different models have been proposed to account for the formation of membrane-less compartments. The simplest model is the binding model where molecules bind and unbind their target sites (i.e. ssDNA for Rfa1). In this case, the local concentration of proteins reflects the number of binding sites, the amount of proteins available and their affinity for their binding sites, without phase separation. If the binding sites are present on chromatin or DNA, these binding sites can be close to each other increasing their local density. In a second scenario, some binding proteins can form bridges between different chromatin loci by creating loops or by stabilizing interactions between distant loci along the chromatin fiber. These interactions can be driven by interactions between chromatin binding proteins or chromatin components. In this model, referred to as ‘Bridging’ Model, or Polymer Polymer Phase Separation (PPPS) model (Erdel and Rippe, 2018; Miné-Hattab and Taddei, 2019), the existence of sub-compartments relies on both the binding and bridging properties of these proteins to chromatin. A third scenario is the Liquid-Liquid Phase Separation (LLPS) Model, also referred to as ‘Droplet Model’ (Hyman et al., 2014). Unlike the binding and the PPPS models, in LLPS, proteins self-organize into liquid-like droplets that grow around a nucleation site allowing certain molecules to become concentrated while excluding others. In this framework, one should distinguish proteins able to initiate a liquid phase separation on their own (scaffold proteins) from proteins concentrating and diffusing freely within this droplet without being responsible for its formation (client proteins) (Banani et al., 2017).

Despite a large literature in the field, many in vivo tests commonly used to probe the nature of sub-compartments are extremely phenomenological and insufficient to rule out other possible mechanisms (McSwiggen et al., 2019b; Miné-Hattab and Taddei, 2019). In addition, most of the optical methods used to discriminate models are at the limit of the diffraction and suffer from severe artefacts (McSwiggen et al., 2019a; Miné-Hattab and Taddei, 2019). Overall, there is a lack of solid experimental criteria to confirm one model versus another in living cells, in particular at the microscopic level.

Our results reveal that Rad52, but not Rfa1, motion shares several properties of LLPS models which are conserved from haploid to diploid cells.

  • McSwiggen et al., 2019a). In the LLPS model, even in the presence of a nucleation core, the large majority of proteins are not bound to chromatin (Erdel and Rippe, 2018; Miné-Hattab and Taddei, 2019). Consistent with the LLPS model, Rad52 molecules inside foci exhibit high mobility compared to the focus itself, and the statistics of confined diffusion with a confinement radius of the same size as the whole focus. In contrast with Rad52, the motion of Rfa1 molecules, which coat ssDNA, have similar diffusion coefficient and type of motion as the focus.

  • Second, LLPS are characterized by a change of motion upon entering or escaping the sub-compartments (Banani et al., 2017; McSwiggen et al., 2019a). Here, we observed a sharp change in the diffusion coefficient of Rad52 molecules crossing foci boundary.

  • Third, in the LLPS model, it is energetically more favorable for molecules to stay inside the sub-compartment than to leave because of the presence of an energetic potential maintaining the sub-compartment. Here, we observed the existence of an attractive potential maintaining Rad52 molecules inside the focus even if molecules are not physically bound to a polymer substrate. However, given the noise and the variable size of foci, we cannot currently describe the shape of the potential.

  • Fourth, an important hallmark of LLPS is their ability to fuse, two droplets for radius R leading to a bigger droplet of doubled volume. Fusion of repair foci has been previously observed in yeast (Oshidari et al., 2020; Waterman et al., 2019) and in human cells (Kilic et al., 2019; Pessina et al., 2019; Schrank et al., 2018; Sollazzo et al., 2018) by tracking repair foci at larger time scales (several minutes). Here, using live PALM, we further show that foci resulting from 2 DSBs are 1.99 ± 0.2 larger in volume than those formed by a single DSB (156 ± 3 nm versus 124 ± 3 nm respectively), and present similar Rad52 densities. Moreover, focus sizes measured by PALM agree with the confinement radius of individual Rad52 molecules inside foci, showing the ability of Rad52 to explore the whole sub-compartment. Thus, our results indicate that upon 2 DSBs, Rad52 foci do not merely cluster, but form a focus of size consistent with the fusion of two foci inside which molecules explore the entire larger sub-compartment.

  • Fifth, at the macroscopic level, aliphatic alcohol hexanediol, a component proposed as a tool to differentiate liquid-like from solid-like assemblies, partially dissolves Rad52 foci and P-bodies while Rfa1 foci remains as abundant. Of note, Rad52 foci as well as P-bodies (Edc3 foci) are not visible after fixation (see Materials and methods), whereas Rfa1 foci resist to the same fixation treatment, suggesting that paraformaldehyde might not be able to fix LLPS.

Other predictions of the LLPS versus PPPS properties have been proposed in the literature, in particular when over-expressing a protein able to form a nuclear sub-compartment. In a simple LLPS model where a sub-compartment is formed by a single species, increasing protein amounts should increase the focus size while concentrations remain the same inside foci and in the nucleoplasm (Erdel and Rippe, 2018; Miné-Hattab and Taddei, 2019). When a LLPS is formed but the observed molecule is not the main species driving the LLPS (referred as a ‘client’ of a LLPS), the concentration inside foci and in the background increases linearly with over-expression of the observed molecule (Hyman et al., 2014). Here, we found that upon different levels of Rad52 over-expression, the background concentration increases as predicted for a ‘client’ (see Appendix 1 and Appendix 1- figure 1) suggesting that Rad52 might not be the main driving molecule responsible for the LLPS formed at the damaged site. This is in agreement with the observation that several DSBs coalesce and dissociate independently of Rad52 (Waterman et al., 2019). However, a recent study pointed out that LLPS composed by multiple proteins with heterotypic multicomponent interactions do not exhibit a fixed saturation concentration (Riback et al., 2020). Since sub-compartments are formed by a complex mixture of proteins, it is possible that the formation of a LLPS is not driven by a single species defined as a driver, but rather by the combination of several interacting proteins.

Recent studies across different model systems proposed that repair foci are LLPS (Altmeyer et al., 2015; Kilic et al., 2019; Oshidari et al., 2020). In human cells, it has been suggested that DSB-induced transcriptional promoters drive RNA synthesis and stimulate phase separation of repair proteins (Pessina et al., 2019). As mentioned in McSwiggen et al., 2019b, some of the criteria used to define a LLPS are at the limit of the optical resolution or are not sufficient to distinguish between a LLPS and a PPPS. For example, sub-nuclear compartments can phase separate by PPPS and collapse of such foci is sometimes wrongly interpreted as a fusion of liquid droplets (Erdel et al., 2020; McSwiggen et al., 2019a). Here, by quantifying the physical properties of Rad52 foci at the single molecule level, our study confirms that Rad52 diffusion is consistent with a LLPS at damaged sites, although Rad52 might not be the main driver of droplet formation. Here we discuss two possible models, LLPS versus PPPS, but there is a gradient of models that can be proposed and might better describe nuclear sub-compartments.

Finally, we found that SUMOylation modulates the physical nature of Rad52 foci. First, the mobility of SUMOylated Rad52 molecules inside foci remains confined but becomes significantly slower than wild type Rad52 (DRad52-Sumo, inside = 0.009 ± 0.002 μm2/s compared to DRad52 inside = 0.032 ± 0.006 μm2/s). Second, SUMOylated Rad52 foci are 3.4 times smaller in volume but exhibit similar intensity than wild type foci, indicating that they are denser. We propose that when all Rad52 molecules are SUMOylated, Rad52 foci transition to a gel-like sub-compartment where most of the molecules are still free to explore the whole focus but much slower than in wild type cells. Since SUMOylated Rad52 foci are smaller, it is possible that the formation of LLPS around damage sites is a mechanism to increase the pull of molecules at the DSB.

Working model for Rad52 foci

Summing up all our observations, we propose the following model. In response to a DSB, Rad52 diffusion is not altered in the nucleus, thus Rad52 molecules likely do not exhibit a strong collective or directive motion toward the damaged DNA site. They rather arrive by simple diffusion at the DSB where they bind ssDNA coated by the RPA complex. The focus is then formed by a small seed of Rad52 molecules bound RPA, around which a large cloud of Rad52 rapidly diffuse exploring the whole focus. Similar to LLPS, Rad52 foci are dense but surprisingly dynamic: Rad52 molecules inside foci are ~6 times more mobile than ssDNA-binding proteins Rfa1 with permanent exchange with the rest of the nucleus. When escaping foci, Rad52 molecules change their diffusion behavior outside foci, suggesting that their diffusive behavior is defined by its environment. Rad52 molecules are retained inside the focus by a potential. The existence of such attractive potential supports a model where DNA is dispensable to maintain the phase separation once a given saturating concentration of multivalent binder molecules is reached. Overall, our results at the single molecule level reveal that Rad52 foci exhibit many features of LLPS at the microscopic level, consistent with recent studies using macroscopic criteria both in yeast and in human cells (Altmeyer et al., 2015; Kilic et al., 2019; Oshidari et al., 2020; Pessina et al., 2019). Our results provide a quantitative measurement of Rad52 behavior once the focus is formed and future studies will be necessary to investigate Rad52 foci formation.

In the future, single molecule microscopy combined with genetics will be a powerful method to shed light on the role of mutants affecting diffusion, residence time, focus size and the type of motion. In particular, further studies i different genetic backgrounds might reveal which exact mechanisms drive the formation and the physical nature of Rad52 foci. More generally, applying single molecule microscopy to other biological processes may reveal that nuclear sub-compartments with very similar aspects and sharing some macroscopic features of LLPS are in fact formed by other alternative mechanisms. For example, silencing sub-compartments resemble Rad52 foci: they fuse and disassemble in specific metabolic conditions. However, single molecule microscopy reveals that Sir3 proteins diffuse very differently than Rad52 (our unpublished results) and do not behave as expected in a LLPS (Miné-Hattab and Taddei, 2019). In human cells, Herpes Simplex Virus appears to share many properties commonly attributed to LLPS, but SPT measurements have recently shown that they are formed through a distinct compartmentalization mechanism (McSwiggen et al., 2019b). Similarly, although several studies proposed that heterochromatin sub-compartments are LLPS, recent results question a simple LLPS mechanism and are consistent with a collapsed chromatin-globule model (Erdel et al., 2020). As the list of proteins that can undergo phase separation is growing in the literature (Mészáros et al., 2019; You et al., 2020), it will be essential to confirm the nature of sub-compartments using in vivo single molecule approaches. As we obtain higher-resolution biophysical insight, what appear to be droplets at first glance might in fact follow more complex models. Therefore it is becoming necessary to define solid criteria and observables at the microscopic level to distinguish between different models and to develop alternative models of membrane-less sub-compartments.

Materials and methods

All strains used in this work are isogenic to RAD5+ W303 derivatives (Supplementary file 2).

Strain constructions

Request a detailed protocol

Strain harboring Rad52-mMaple (McEvoy et al., 2012) or Rad52-Halo are constructed using crisper-Cas9 technic: the fusion is made at the Rad52 N-terminus with a 48 bases linker (CGTACGCTGCAGGTCGACGGAGCAGGTGCTGGTGCTGGTGCTGGAGCA) or a 15 bases linker (AGTGGAGGCGGAGGT), respectively. For strains harboring Rad52-mMaple, we first built plasmid pAT475 containing the cas9 sequence and a gRNA targeting Rad52. The donor sequence containing the mMaple sequence was produced from plasmid pAT446 (pUM003, kindly provided by Jonas Ries laboratory, EMBL), using primers am1941 and am1942 (see Supplementary file 2). Cells were transformed simultaneously with the gRNA-RAD52 plasmid (pAT475) and the donor DNA on –LEU plates. For strains harboring Rad52-Halo, we amplified the HaloTag sequence from the plasmid pAT496 (pBS-SK-Halo-KanMX), kindly provided by the Carl Wu laboratory using primers am2182 and am2402. Cells were co-transformed with the gRNA-RAD52 plasmid (pAT475) and the donor DNA on –LEU plates. After clones validation by PCR (am384, am1944 and am1947), cells are re-streaked several times on YPD until the loss of the gRNA-RAD52 plasmid.

In Rad52-Halo strains, the pleiotropic drug resistance PDR5 gene was replaced by the N-acetyltransferase (NAT) gene using deletion plasmid collection from the toolbox (Janke et al., 2004) and S1/S2 primers: pAT197, am2165 and am2166. The replacement of PDR5 gene with the 3-isopropylmalate dehydrogenase (LEU2) gene was performed by PCR-based gene targeting using pAT297; primers am2609 and am2610. The PDR5 gene deletion was checked by PCR using following primers: am431, am2167, am2168 and am2421.

The functionality of Rad52 in these strains was verified using a dilution assay on plates containing MMS at different concentrations (Figure 1—figure supplement 1).

Induction of a DSB

Before microscopy, strains harboring an I-SceI cut site under galactose promoter were pre-grown in SC medium and diluted at OD600nm = 0.01 in SC +3% raffinose medium overnight. In the morning, cells were diluted at OD600nm = 0.2 in SC +3% raffinose medium until they reach OD600nm = 0.4. To induce the DSB, 2% galactose were then added directly into the tube for 2 hr, and fluorogenic JF646 (Grimm et al., 2015) were added directly into the tube during the last hour of galactose induction.

For the control experiments performed in the absence of DSB (referred as ‘no DSB’ in the figures), we used a strain without an I-SceI cut site, grown in the same conditions as the strain used to induce a DSB to ensure a fair comparison between the two conditions.

Sample preparation for the microscopy

Request a detailed protocol

A key aspect of fluorescent experiment is the choice of a suitable fluorophore. Several parameters have to be taken into account, depending on the experiments: specificity of the dyes, density of labeling, brightness and photo-stability. For SPT, in which we want to follow individual molecules as long as possible, we choose fluorogenic HaloTag JF646 dyes (Grimm et al., 2015) for their brightness and specificity. Indeed, fluorogenic JF646 dyes have the ability to emit light only when they are bound to a Halo molecule and no unspecific signal was observed. Cells are prepared in dark tubes and incubated with 5 nM JF646 during 1 hr.

For live PALM, in which we want to observe each Rad52 molecules only once, we used a strain harboring endogenously fused Rad52-mMaple (McEvoy et al., 2012). Unlike the SPT experiments using external dyes, here, all the Rad52 molecules are fused to mMaple, allowing the observation of all Rad52 molecules for quantification of focus size, number of Rad52 molecules and density. Due to the photo-conversion properties of mMaple, it is estimated that 70% of mMaple can be photo-converted (McEvoy et al., 2012), which is taken into account in our analysis.

In vivo tests of LLPS dissolution

To test the physical nature of Rad52 and Rfa1 foci, we used in vivo treatment known to disrupt LLPS but not solid aggregates. As described in Kroschwald et al., 2017, cells were treated with 1.6 hexanediol (Aldrich reference 240117) at 10% and digitonine (Merk Millipore reference 300410) at 0.001% for 30 min. Note that as reported in several studies (Chang et al., 2018; Kroschwald et al., 2017; McSwiggen et al., 2019a), extended exposure of cells to 1.6-hexanediol leads to loss of membrane integrity but a 30 min treatment does not alter membrane integrity (Kroschwald et al., 2017).

To compare the effect of fixation on Edc3, Rad52 and Rfa1 foci, fixation was performed during 10 min with 4% paraformaldehyde. We also tested several other fixation buffers (Kaplan and Ewers, 2015; Schnell et al., 2012, see fixation protocols below) but we found that Rad52 foci were not visible for these protocols.

Detailed protocols fixation

Request a detailed protocol
  • Protocol 1: 4% paraformaldehyde for 10 min, three rinsing steps in 2x SC media for 5 min each.

  • Protocol 2: 4% paraformaldehyde for 10 min followed by 1% Glutaraldeide for 10 min, three rinsing steps in 2x SC media for 5 min each.

  • Protocol 3: Methanol (à −20°C) for 10 min.

  • Protocol 4: Ethanol (à −20°C) for 10 min.

All paraformaldehyde solutions were dissolved in the following fixation buffer: PFA 4%, 1X PBS, 2% sucrose.

Single Particle Tracking

Request a detailed protocol

We acquired single molecule images on a custom setup based on a Nikon iSPT-PALM inverted microscope. For SPT experiments, we used fluorogenic JF646 dyes excited with a 647 nm laser with a power of 1.9 kW/cm2 at the sample. The fluorescent signal is captured with an EM-CCD camera (Ixon Ultra 897 Andor) using a 100x/1.45NA (Nikon) objective. Using this objective, the image pixel size was 160 nm. We controlled the microscope with NIS software (Nikon). Experiments were performed at 30°C using a Tokai temperature control (STXG-TIZWX-SET) placed at the objective. We detected and connected the spots obtained in the movies with a custom algorithm derived from Sergé et al., 2008 and we used home-made programs to visualize density and displacements detections. We kept all traces equal or longer than two points for further analysis.

Estimation of diffusion coefficients

Displacement histogram

Request a detailed protocol

Probability Density Functions (PDF) are obtained after normalization of the displacement histograms and fitted with a 1-, 2-, or 3-population model (Klein et al., 2019; Stracy and Kapanidis, 2017):

1-population model: PDF(r,t)=2r4Dtexp(-r24Dt), where D is the diffusion coefficient and r is the position at time t.

2-population model: PDFr,t=f2r4D1texp-r24D1t+(1-f)2r4D2texp(-r24D2t), where D1 and D2 are the diffusion coefficients of sub-population 1 and 2 respectively, and f is the fraction of sub-population 1.

3-population model: PDFr,t=f12r4D1texp-r24D1t+f22r4D2texp-r24D2t+1-f1-f22r4D3texp-r24D3t, where D1, D2 and D3 are the diffusion coefficients of sub-population 1, 2, and, 3 respectively, and f1 and f2 are the fractions of sub-population 1 and 2.

In order to estimate the diffusion coefficients, we started out with the null hypothesis that traces were described by one diffusion coefficient. We simulated trajectories using the maximum likelihood estimate for the diffusion coefficient and compared the simulated and experimental CDF using a two-sided Kolmogorov-Smirnoff (KS) test to test the one diffusion coefficient hypothesis.

If one diffusion coefficient could not describe the data, we used maximum likelihood to find the two diffusion coefficients and the population fraction diffusing with each coefficient. We used a Metropolis Hastings algorithm to sample the two diffusion coefficients from the posterior distribution. Based on the distribution of the diffusion coefficients we estimated the uncertainty in the extracted parameters.

Criteria to select traces inside repair foci

Request a detailed protocol

We use two methods to select traces inside foci:

  1. Since foci are much denser than the rest of the nucleus, we selected traces based on a density threshold. We first detect all the Rad52 molecules in our movies, and calculate a density map showing the number of Rad52 neighbours inside a sphere of 50 nm in radius (Figures 1B and 2B, third panel). At this stage of the analysis, the tracking is not performed yet: we simply localize molecules and we have no information yet on Rad52 mobility. The focus is then determined by a density threshold as it corresponds to the region in the nucleus where repair proteins are highly concentrated. The program calculated the coordinates of a polygon, which defines the focus boundaries. Then, we perform the tracking step to connect Rad52 detections and draw a displacement map (Figure 2B, fourth panel): the traces inside a focus are obtained by overlying the polygon on the displacement map. For these traces, we tested a 2-population fit (Dinside-1 = 0.26 ± 0.02 μm2/s, and Dinside-2 = 0.066 ± 0.004 μm2/s, p=0.56, two-sided KS test), with slow molecules being the large majority (0.76 ± 0.04). However, the diffusion of fast molecules is similar to D2 (0.24 ± 0.04), and by removing boundary points, we obtain a very good fit using a 1-population fit (p=0.99, two-sided KS test) inside the focus (data not shown). This analysis is referred as ‘cropped data’ in Supplementary file 1.

  2. Since no trace longer than 70 time-points were found in the absence of DSB, we selected traces longer than 70 time-points and assumed that they would be majorly populated by a slow diffusion coefficient. However, since molecules can travel back and forth between the focus and rest of the nucleus, we separated the traces belonging to the slow diffusion coefficient by assaying each displacement a probability based on the slow diffusion coefficient D3 = 0.054 and removed points with probability less than 1/10 000. This analysis is refereed as ‘Traces longer than 70 time-points’ in Supplementary file 1.

After selecting traces inside foci with each method, we used MSD analysis to calculate the diffusion coefficient of molecules and decipher the nature of Rad52 motion inside foci (see below). Both methods lead to similar results (see Supplementary file 1).

Mean square displacement

Request a detailed protocol

In the case of slow diffusion coefficients (D < 0.1 μm2/s), we needed to extract the diffusion coefficient from the experimental noise. For that, we calculated the MSD:

MSD(nΔt)=1Nni=1Nn[(xi+nxi)2+(yi+nyi)2],

where N is the number of points in the trajectory, (x, y) the coordinates of the locus in two -dimensions and Δt the time interval used during the acquisition. To obtain a precise estimation of the confinement radius and the diffusion coefficient, we calculated time-ensemble-averaged MSD over several trajectories, which are simply referred to as ‘MSD’ in the Figures. To check possible artefacts due to the variability between molecules, we also computed individual MSD that is MSD calculated from a single trajectory (data not shown).

When the motion appeared confined, we fit the MSD with MSD(t)=R2(1e2dDt/R2)+2, to extract the experimental noise level σ2, the radius of confinement Rc=R2+d2, where d is the dimension of the motion, and the diffusion coefficient D (Klein et al., 2019). The characteristic equilibration time after which the effect of boundaries appears is defined by the time at which the MSD curve starts bending and is given by tc=Rc2/(6D), where Rc is the confinement radius and D the diffusion coefficient inside foci (Klein et al., 2019).

When the motion is not a simple confinement but is modulated in time and space with scaling properties, it is called anomalous sub-diffusion. In this case, sub-diffusive loci are constrained, but, unlike confined loci, they can diffuse without boundaries and thus reach further targets if given enough time. For sub-diffusive motion, the MSD exhibits a power law and is fitted with:

MSD(t)=Atα+ε, where α, the anomalous exponent, is smaller than 1, A is the anomalous diffusion coefficient, and ε is the noise due to the experimental measurements (Miné-Hattab et al., 2017, p.). For the wild type data, we fitted both functions to the data, and found a better agreement with the confinement fit.

Estimation of the dynamic viscosity and the diffusion coefficient of Rad52 monomers versus multimers

Request a detailed protocol

To estimate the dynamic viscosity of the nucleus in the absence of DSB, we use Stokes-Einstein equation:

D=kbT6πηr

where D is the diffusion coefficient, r is the radius of a spherical molecule, η is the dynamic viscosity, kB the Boltzmann constant and T the absolute temperature.

We used 1.5 nm and 9 nm ±1 for the radius of a Rad52 monomer and the radius of a Rad52 multimer (Saotome et al., 2018; Shinohara et al., 1998).

Errors on D and η are calculated using error propagations:Δη=|Dη|ΔDandΔD=|ηD|Δη.

Simulations

Request a detailed protocol

In order to extract as much information as possible from the traces, we performed simulations to correctly quantify the effects of: (1) the reduction in the number of observed fast molecules, due to their larger rate to exiting the z-frame undetected than slow molecules, and (2) the reduced diffusion coefficient, due to the collisions with the nucleus boundary.

To understand (1), we simulated 10000 particles for 10 s, inside a sphere of 1 μm radius, where traces were only measured if the particle within the window of |z| < 0.15 μm. We counted the number of detected data points, as a function of the size of the time window using time windows of ts = 5–100 ms. We found the number of detected data points decayed exponentially for all diffusion coefficients, but the magnitude of the decay rate grew as a function of the real diffusion coefficient. Therefore we took all of the computed decay rates, and fitted them to a power law, giving us an expression for the rate at which the number of observed molecules decays as a function of the time window and the true diffusion coefficient. We do not know the total number of particles in each population and we cannot determine this relation directly. We use this empirical scaling of the relation between the two populations.

In the same manner, we estimated (2) how the diffusion coefficients were affected. We measured the MSD and found that the diffusion coefficient obtained from this measure decayed as a function of the true diffusion coefficient. We directly calculated the diffusion coefficient from the maximum likelihood, and measured the gradient in the increase of the diffusion coefficient, assuming the considered time window is constant. By doing this we could fit the gradient and determine the expected diffusion coefficient.

Estimation of radial motion

We used data cropped to movement inside the focus to estimate radial motion. We used all traces longer than 30 points and estimated the center of the focus by calculating their mean value for each trace in the x-direction and y-direction. We calculated the distance to the center for all points, and defined the change in radius as Δr. We grouped data points from all traces and compared the statistics with a random walk with no attraction but with a similar diffusion coefficient (dashed lines Figure 7D, E).

Live PALM

Request a detailed protocol

For live PALM experiments, we used strains harboring Rad52-mMaple as well as 1 or 2 I-SceI cut-sites inducible under galactose. Samples were prepared with living cells and mMaple was excited with a 561 nm laser at power of 6 kW/cm2 at the sample (max power) as well as a pulsed 405 nm laser to photo-activate mMaple. Movies were performed at 20 ms time interavals, during 2000 frames (no more spot were detectable at the end of the acquisition). Spots were detected using a custom algorithm derived from Sergé et al., 2008 and home-made programs to visualize spots density and quantify foci.

Appendix 1

Consequences of increasing the concentration of Rad52

To evaluate the consequences of Rad52 concentration, we measured the concentration of Rad52 in the nucleoplasm for different levels of Rad52 over-expression. We built two haploid strains expressing Rad52 under the ADH or TEF1 promoters (5 and 12 times over-expression respectively) (Hocher et al., 2018). In addition, a single I-SceI cut site allows the induction of a single DSB under the galactose promoter. Cells were grown at 25°C due to a slight growth defect of cells overexpressing Rad52. A single DSB was induced for 2 hr; during the last hour of induction, cells were incubated with JF646 dyes at 50 nM, a concentration 10 times higher than the one used for single molecule tracking. Such concentration allows the observation of the entire Rad52 focus as a single spot. Cells were observed on an inverted wide field microscope and Rad52 signal was quantified using a home-made software Q-foci (Guidi et al., 2015). Rad52 concentration in the nucleoplasm is calculated for each nucleus (nucleoplasm intensity divided by nucleoplasm volume).

We found that Rad52 intensity in the nucleoplasm is significantly different between wild type cells and cells expressing over-expressed level of Rad52 (see Appendix 1—figure 1, pWT-ADH = 6.5.10−8, p ADH-TEF1 = 7.10−32 and p WT-TEF1 = 3.5.10-8, Wilcoxon-Mann-Whitney test).

Appendix 1—figure 1
Consequences of increasing the concentration of Rad52 molecules in the cell.

(A) Typical images of Rad52 after the induction a single DSB in wild type cells, and for two levels of Rad52 over-expression. The bar scale represents 1 μm. (B) Effect of Rad52 over-expression on the nucleoplasm concentration. 11, 94 and 141 cells were analyzed for wild type, ADH and TEF1 promoters respectively.

Data availability

All data generated or analysed during this study are included in the manuscript and supporting files. Source data files are available on zenodo using the following link: https://zenodo.org/record/4495116.

The following data sets were generated
    1. Miné-Hattab J
    2. Heltberg M
    3. Villemeur M
    4. Guedj C
    5. Mora T
    6. Walczak AM
    7. Dahan M
    8. Taddei A
    (2021) Zenodo
    Single molecule microscopy reveals key physical features of repair foci in living cells.
    https://doi.org/10.5281/zenodo.4495116

References

Decision letter

  1. Irene E Chiolo
    Reviewing Editor; University of Southern California, United States
  2. Jessica K Tyler
    Senior Editor; Weill Cornell Medicine, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

The revised manuscript shows elegant and compelling evidence that the dynamics of Rad52 molecules inside repair foci are consistent with the existence of a liquid droplet surrounding repair sites. Using single particle tracking (SPT) and Photo-activatable Localization Microscopy (PALM), the study also reveal that the ssDNA-binding protein RPA display a different behavior from Rad52, being mostly chromatin associated. Together, this analysis and the related conclusions significantly advance the field of nuclear dynamics in the context of DNA repair.

Decision letter after peer review:

Thank you for submitting your article "Single molecule microscopy reveals key physical features of repair foci in living cells" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Jessica Tyler as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

As the editors have judged that your manuscript is of interest, but as described below that additional experiments are required before it is published, we would like to draw your attention to changes in our revision policy that we have made in response to COVID-19 (https://elifesciences.org/articles/57162). First, because many researchers have temporarily lost access to the labs, we will give authors as much time as they need to submit revised manuscripts. We are also offering, if you choose, to post the manuscript to bioRxiv (if it is not already there) along with this decision letter and a formal designation that the manuscript is "in revision at eLife". Please let us know if you would like to pursue this option.

Summary:

In this manuscript by Miné-Hattab and colleagues, the authors use single-molecule imaging approaches to investigate local dynamics of Rad52 foci at DSBs in budding yeast, which is an important area of investigation. They show that the dynamics of Rad52 molecules inside foci are consistent with protein movement within LLPS domains, while Rfa1 dynamics are not. Their data also provide supporting evidence to previous observations that repair sites cluster within the nuclei, and suggest that clustered foci behave as larger phase separated structures. While the idea that Rad52 and other repair proteins form phase separated domains is not novel, this study presents higher resolution data in support of this model. The reviewers generally agree that the study is interesting and well conducted, but the conceptual advancement is limited and a significant revision is needed for publication in eLife. Specifically, more convincing experiments demonstrating that the observed Rad52 dynamics reflect LLPS are required. Evidence that the dynamics are relevant for DNA repair and genome stability should also be provided. Additionally, the study should be better integrated with previous studies, statistical analyses need to be more rigorous/better presented, and a revisions of the text should include a clearer separation between observations and speculations.

The following is a summary of the main points brought up by the reviewers, including after consulting with each other.

Experimental revisions:

1) Additional data need to be provided to draw conclusions about whether or not the authors' observations are reflective of phase separation. Specifically, additional mobility studies in conditions that disrupt LLPS (chemical treatments and mutations) are needed, both for the individual protein and for the foci.

2) Regarding the possible categories of traces evaluated, one category is not included in the study. The surface tension that defines LLPS-dependent bodies is known to both help maintain focus integrity and partly counter LLPS body fusions. So if the foci represent true phase-separated bodies, have the authors then observed traces where Rad52 molecules interact with yet fail to enter the larger Rad52 foci?

3) How is it possible to distinguish a cluster of binding sites from liquid-liquid phase separation? In the absence of breaks, there are two Rad52 diffusion populations (D=1.2 and 0.3 um2/s), which the authors attribute to monomers and multimers. They don't verify these multimers by alternative approaches (say number and brightness analysis), but it seems like a reasonable possibility. After a break, a third component – slower than the previous two – becomes evident. This slow population coincides with the break. In the vicinity of the break, there is now only 1 component diffusion (D=0.03 um2/s). Also, the motion is now more confined, but not absolutely so. Also, Rad52 diffuses faster than Rfa1, which is bound to ssDNA. At this point, there is no data to distinguish between two possibilities: slow diffusion *or* diffusion + binding. Except, if it were diffusion + binding, one might perhaps expect to still see the free diffusion component.

The authors then turn to diffusion at the boundary (Figure 5), which I agree can be a more informative measure. Here, they see changes in the diffusion estimator for trajectories which cross the boundary, using displacement which they argue is more robust for slow diffusion. The problem is that the “boundary” is determined by the very thing they are trying to measure, not some independent marker of the compartment. In other words, Rad52 defines the compartment, unless I missed something fundamental in the experimental design. Ideally, the way such an experiment would be done to test the hypothesis that Rad52 is forming a LLPS compartment is to look at the diffusion of an inert tracer as it comes in and out of the compartment. As designed, I frankly do not see how the observation of different diffusivities in and out of the compartment distinguishes between a cluster of binding sites and an LLPS. If you accept that DNA-binding is in no way biasing the kinetics, then the authors' interpretation seems like the most sensible one. But the fact that Rad52 is involved in DNA repair makes that a hard assumption to swallow.

Furthermore, I'm not sure I entirely grasp the significance of Figure 6. Since Rad52 can easily escape one focus and enter another, regardless of whether it is a cluster of binding sites or a phase, I don't see how the radius of confinement measurement distinguishes between these two alternatives. The observation that the foci are 2x larger in diploids but at similar density is compelling, although recent data from the Brangwynne lab point out that conserved density need not be the case (PMID: 32405004).

4) In the syntax of this paper, Rad52 is a client in the LLPS, leaving the question of the scaffold unaddressed. After all, the Rad52 focus ultimately disappears, meaning that something caused this phase to be dispersed. So is RPA the scaffold? It might be possible to address both this point and point #3 knowing what is responsible for forming the LLPS in the first place.

Along the same lines, how do the authors reconcile previous findings indicating that recombinant DNA repair proteins phase separate in vitro with their claim that "Rad52 acts as a client of the LLPS but does not drive its formation"?

5) What is the evidence that the biophysical properties observed are of direct relevance to DNA repair? For example, is the mobility of Rad52 within the repair focus important for repair? Is the difference in diffusion kinetics within and outside of the repair focus important for genome stability? What could the authors do to alter that diffusion profile and what would be the consequence on repair? Also, addressing this point implies the need to use a more physiologically relevant system with repairable DSBs, and not the irreparable DSB system used here.

One should easily compare wild-type Rad52 to known Rad52 mutants (that partly or fully abrogate function) to see if there is any correlation between intra-focus mobility profiles and repair. In other words, experiments along these lines may indicate whether a particular intra-focus mobility profile consistently correlates with repair, while this profile is absent in non-functional mutants, or a particular type of mutant. These experiments should be very feasible.

6) Can the authors visualize the fusion of the Rad52 foci/DSBs in live cells within their experimental systems?

7) The statistical significance of most presented data is either lacking or unclear. This needs to be carefully addressed. Providing additional data files may also help the authors strengthen their findings.

Text revisions:

– Several statements made are not supported by the data and without clearly stating that the statements represent speculations. E.g. longer tail is due to Rad52 molecules diffusing slowly inside the focus; observing the 2 populations also in G1 does not necessarily mean that the 2 populations in S/G2 do not reflect replication forks at all. The authors need to carefully revise their claims/statements and consider alternative explanations. Also, the writing is often unclear or confusing and the authors should consider substantially revising it to clarify their claims, clearly indicate speculations that are not supported by the data, and make the text as accessible as possible to non-specialists.

– How was the cell cycle stage determined? This should be better explained.

– Figure 1—figure supplement 1 data appear to show the existence of a partial loss of Rad52 function in the Rad52-Halo cells. This should be clearly stated in the Results and consequent limitations/caveats discussed. Also, please clarify whether Figure 1—figure supplement 1 shows the viability of Rad52-Halo cells in the presence or absence of JF646.

– The authors present no direct evidence for an "attractive potential" that drives molecules towards the centre of the focus. For example, what if the “attractive potential” is simply the focus' boundary surface tension creating a barrier against which some of the molecules inside the focus bounce back towards the centre of the focus?

– The authors state that "Here, we found that upon different levels of Rad52 over-expression, the background concentration increases (Appendix 1—figure 1) suggesting that Rad52 might not be the driving molecule responsible for the LLPS formed at the damaged site." The logical transition here is unclear.

– It is difficult to judge the novelty of this work, as key papers that showed similar conclusions or datasets are not cited. Without direct comparison with other data sets, it is difficult to see exactly where this paper goes beyond published studies.

Here are a few key examples:

a) In the last year the Haber lab published a very similar study in Plos Genetics (Waterman et al., 2019). Although they tracked Ddc2 and Rad51, they also looked at the behavior of separate foci and this paper is not even cited. The data should be compared at the very least.

b) The characteristics of 53BP1 foci have been extensively studied by many labs including those of Altmeyer, Scherthan, DeLange and others, with very similar findings as Miné-Hattab reports for Rad52 (for example, Kilic et al., 2019; Sollazzo et al., 2018, as well as the single molecule work of the lab of Eric Greene). Moreover both rad52 and PCNA foci were studied by Essers et al. (Kanaar and Vermeulen) MCB 2005. 25(21): 9350-9359 and EMBO J. 2002 Apr 15. Comparisons with these studies needs to be made.

c) A number of earlier studies followed Rad52 foci in budding yeast on induced double strand breaks (even using the I-Sce1-cut system used here) that are not taken into consideration. The diffusion coefficients presented here have to be compared with these earlier studies and differences should be resolved by comparing techniques and conditions of imaging. For instance, Dion et al., 2012).

– It is unclear if the “absence of DNA damage” condition discussed in the first section of the Results is the non-induced version of the system described in the second section of the Results. Also regarding these sections, it seems that the “absence of DNA damage” control conditions were not conducted as part of the same experiments with the I-SceI DSB.

– Figure 2C is a little underwhelming. I would like to see one data panel in the main text with the cumulative distribution +/- DSB.

– "in the case of diffusion coefficients D<0.1 μm2/s, we use mean square displacement analysis allowing us to substrate the noise." Subtract the noise.

– One technical point. The fluorescence lifetime of a freely diffusing fluorophore (reported half-life of JF646 is 2.1 sec) is not the same as if it is bound to a protein; in the latter context the fluorophore would diffuse more slowly in the illumination path and be subjected to more rapid photobleaching, making 2.1 sec an overestimation of the half-life/underestimation of the decay rate. At a frame interval of 20ms, photobleaching may be a competitive rate for the disappearance of signal and I am not sure the lack of applying photobleaching correction or saying that the short tracks are not affected by that rate is properly justified.

https://doi.org/10.7554/eLife.60577.sa1

Author response

Experimental revisions:

1) Additional data need to be provided to draw conclusions about whether or not the authors' observations are reflective of phase separation. Specifically, additional mobility studies in conditions that disrupt LLPS (chemical treatments and mutations) are needed, both for the individual protein and for the foci.

We agree and thank the reviewer for this suggestion. In the new version of the manuscript, we now provide additional data in conditions that disrupt LLPS, for both Rad52 and Rfa1 foci. We used aliphatic alcohol hexanediol (1.6-hexanediol), a component proposed in the literature as a tool to differentiate between liquid-like and solid-like assemblies in living cells (Kroschwald et al., 2017, Strom et al., 2017, McSwiggen et al., 2019, Oshidari, 2020). In budding yeast, it has been shown that 1.6-hexanediol dissolves dynamic, liquid-like assemblies, such as P bodies, whereas solid-like assemblies, such as protein aggregates and cytoskeletal assemblies, are largely resistant to hexanediol (Kroschwald et al., 2017). We observed a clear difference between Rad52 and Rfa1 foci: following the induction of a single DSB, Rad52 foci were significantly less abundant in cells treated with hexanediol (8 % against 22 %) consistently with a recent report from the Meckhail laboratory (Oshidari et al., 2020), while Rfa1 foci remains stable. These results are presented and discussed in the revised manuscript (see Figure 4F and G and Figure 4—figure supplement 1).

In addition, we also tested the effect of 1.6-hexanediol at the single molecule level: we measured the mobility of individual Rad52 molecules (Rad52-Halo/JF646) following DSB induction and hexanediol treatment using the conditions in which the number of Rad52 foci is significantly decreased. However, comparing the histogram of Rad52 displacements in the nucleus with and without 30 minutes of hexanediol treatment, we did not observe any significant change in Rad52 mobility neither in the nucleoplasm nor in the few cells still harboring a Rad52 focus (2-sided Kolmogorov-Smirnoff test, p=0.995).

Finally, we also found that Rad52 foci are destabilized after fixation (10 minutes with 4% paraformaldehyde), as well as P bodies (Edc3 foci) described in the literature for their LLPS properties. However, Rfa1 foci resist to the same fixation treatment, suggesting that paraformaldehyde might not be able to fix LLPS (see Figure 4F and G and Figure 4—figure supplement 1).

2) Regarding the possible categories of traces evaluated, one category is not included in the study. The surface tension that defines LLPS-dependent bodies is known to both help maintain focus integrity and partly counter LLPS body fusions. So if the foci represent true phase-separated bodies, have the authors then observed traces where Rad52 molecules interact with yet fail to enter the larger Rad52 foci?

This is an excellent point and something we have investigated heavily, but this is something that is not possible for two reasons. First, the focus is diffusing as well, and therefore in order to find a specific trace at the boundary of the focus, one would need to know the exact position of the center of the focus, which due to movements can be shifted during the time steps considered. Secondly, and most importantly, the noise level in the experimental observations are of the order 25 nm, and since the estimated radius of the focus is around 100 nm, it is impossible to show if one particular trace has been repelled at the surface in the time between two successive observations. However statistically we have grouped observations based on their relative change in radial distance to the centre of the focus (this is shown in Figure 7). Here we have tried to measure exactly the presence of these events, but data-points of traces just outside the focus, are quite scarce and not something we have been able to draw any conclusions about.

3) How is it possible to distinguish a cluster of binding sites from liquid-liquid phase separation? In the absence of breaks, there are two Rad52 diffusion populations (D=1.2 and 0.3 um2/s), which the authors attribute to monomers and multimers. They don't verify these multimers by alternative approaches (say number and brightness analysis), but it seems like a reasonable possibility.

The principle of single molecule microscopy is to observe one molecule at a time, this molecule being bound to a single dye. In this study, we used flurogenic JF646 dyes (dye emitting light only once bound to Halo) that are not photo-activable. To reach the single molecule regime (as shown in Figure 1—figure supplement 2), we use an extremely low concentration of JF646 allowing us to obtain 0, 1 or maximum 2 detections per nucleus in the same frame. Thus, only a small fraction of Rad52-Halo molecules are bound to a JF646 dye and in case of multimer, the JF dye binds only 1 molecule of a multimer.

Thus, we cannot observe a difference in brightness: all detections have the same intensity regardless of whether they are monomeric or part of bigger multimeric complex.

After a break, a third component – slower than the previous two – becomes evident. This slow population coincides with the break. In the vicinity of the break, there is now only 1 component diffusion (D=0.03 um2/s). Also, the motion is now more confined, but not absolutely so. Also, Rad52 diffuses faster than Rfa1, which is bound to ssDNA. At this point, there is no data to distinguish between two possibilities: slow diffusion *or* diffusion + binding. Except, if it were diffusion + binding, one might perhaps expect to still see the free diffusion component.

We agree with the reviewer that in principle, our observations could be consistent with both hypotheses of “slow diffusion” or “diffusion + binding”. However, in the second case, we would expect to capture with SPT some intervals of time where Rad52 is bound. During these bound intervals, Rad52 should be immobile or diffuse slowly with the same coefficient as Rfa1. This would lead to a strongly multi-modal distribution of displacements inside foci, corresponding to 3 populations of freely-diffusing, bound, and mixed states (intervals in which the particle may bind or unbind). We do not observe such multi-modal behavior at our resolution of 20 ms intervals. This does not rule out “diffusion + binding”, but it puts hard constraints on the rates of binding and unbinding, which must be much faster than 20ms to explain the smoothness of the displacement histogram.

We can try to quantify this by making the conservative assumption that binding is diffusion limited, with diffusivity D = 1 μm2/s, and a linear binding size a = 10 nm. In that case the binding rate of individual molecules to a given target is kon = 4Da (Berg and Purcell, 1977). Assuming a realistic range of binding dissociation constants Kd ~ 1nM to 1μM and using the relation Kd = koff / kon, this implies a mean bound time of 1/koff = 1/Kd/kon ~ 40 ms to 40 s, values which are all larger than our resolution. In fact, to explain the smoothness of our displacement histogram, the binding specificity would have to be very weak, Kd >> 2 μM, which is far above values observed in vivo.

In the revised version of the manuscript, we now discuss this point.

The authors then turn to diffusion at the boundary (Figure 5), which I agree can be a more informative measure. Here, they see changes in the diffusion estimator for trajectories which cross the boundary, using displacement which they argue is more robust for slow diffusion. The problem is that the “boundary” is determined by the very thing they are trying to measure, not some independent marker of the compartment. In other words, Rad52 defines the compartment, unless I missed something fundamental in the experimental design. Ideally, the way such an experiment would be done to test the hypothesis that Rad52 is forming a LLPS compartment is to look at the diffusion of an inert tracer as it comes in and out of the compartment. As designed, I frankly do not see how the observation of different diffusivities in and out of the compartment distinguishes between a cluster of binding sites and an LLPS. If you accept that DNA-binding is in no way biasing the kinetics, then the authors' interpretation seems like the most sensible one. But the fact that Rad52 is involved in DNA repair makes that a hard assumption to swallow.

We thank the reviewer for this remark because it is an important point to clarify. A repair focus is defined by a high concentration of repair proteins at a site of DNA damage. In our experiments, we first detect all the Rad52 molecules in our videos, and calculate a density map showing the number of Rad52 neighbours inside a sphere of 50 nm in radius (Figure 1B and 2B, third panels). At this stage of the analysis, the tracking is not performed yet: we simply localize molecules and we have no information yet on Rad52 mobility. The focus is then determined by a density threshold as it corresponds to the region in the nucleus where repair proteins are highly concentrated. The program calculated the coordinates of a polygon, which defines the focus boundaries. Then, we realized the tracking step to connect Rad52 detections and draw a displacement map (Figure 2B, forth panel): the traces inside a focus are obtained by overlying the polygon on the displacement map. Thus, in this method, the determination of the focus is independent of Rad52 mobility. We then found that inside foci, Rad52 mobility is slower, and that traces crossing the focus boundary sharply change mobility as they come in and out the sub-compartment. We now clarify the method (Figure 2—figure supplement 2).

To estimate the mobility of Rad52 molecules within foci, we also used a second method to identify the traces inside foci: since no trace longer than 70 time-points were found in the absence of DSB, we selected traces longer than 70 time-points and define these ones as the ones belonging to the focus. Both methods give similar distribution of step sizes and fitting the probability density function leads to the similar diffusion coefficient, as shown in Figure 2—figure supplement 2.

We agree that looking at the diffusion of an inert tracer as it comes in and out of the compartment would be interesting. However, we have not found yet an inert tracer that penetrates into repair foci and that we can follow at the single molecule level. Further assays would need to be developed to go in this direction.

Furthermore, I'm not sure I entirely grasp the significance of Figure 6. Since Rad52 can easily escape one focus and enter another, regardless of whether it is a cluster of binding sites or a phase, I don't see how the radius of confinement measurement distinguishes between these two alternatives.

Thank you for this comment. We agree that this point needs clarifying. In a simplified view where multiple DSBs cluster and stay close without merging, we expect to observe the same confinement radius than following a single DSB. In another extreme view, multiple DSBs fuse, giving rise to a large single droplet, and Rad52 confinement radius increases.

Since exchange is important, we now discuss a third intermediate view in which DSBs cluster without merging, but we include a high rate of exchange making possible for a Rad52 molecule to “jump” between a focus to another one. In this view, we should observe bimodal trajectories, exhibiting confined diffusion in the first focus, followed by a confined diffusion in the second focus. Such traces should be observable since the mean residence inside a focus is 240 ms (12 frames) and traces are often longer than this (we added the histogram of Rad52 traces length inside foci, see new Figure 5—figure supplement 1).

After examining carefully individual trajectories inside a focus induced by 2 DSBs, we never observe bimodal trajectories. Instead, in typical trajectories (Figure 6F, right panel), Rad52 is homogenously distributed in time within a larger focus. In the revised version of the manuscript, we now comment on this point. In the limit case where the two foci are so close that we can’t visualize jumps the second and third views are effectively the same.

The observation that the foci are 2x larger in diploids but at similar density is compelling, although recent data from the Brangwynne lab point out that conserved density need not be the case (PMID: 32405004).

First, we would like to clarify this point: in Figure 6, we observed that foci induced by 2 DSBs are 2x larger than those induced by a single DSB. We have not performed experiments in diploid cells in the first version of the manuscript.

Then, we explain that 2 droplets of radius R would lead to a bigger droplet of doubled volume when they fuse, with no significant difference in density (see p 13). We thank the reviewer for bringing our attention to the recent study by the Brangwynne lab. In the experiment following 2 DSBs, there is no change of Rad52 concentration in the nucleus, so we believe the recent study by the Brangwynne lab is of no direct relevance for these results. However, this study led us to revisit our results following Rad52 over-expression. We now comment this point.

4) In the syntax of this paper, Rad52 is a client in the LLPS, leaving the question of the scaffold unaddressed. After all, the Rad52 focus ultimately disappears, meaning that something caused this phase to be dispersed. So is RPA the scaffold? It might be possible to address both this point and point #3 knowing what is responsible for forming the LLPS in the first place.

We do not know what is the scaffold of repair foci but it could not be Rfa1 because one expects the scaffold to form a LLPS, unlike Rfa1.

Along the same lines, how do the authors reconcile previous findings indicating that recombinant DNA repair proteins phase separate in vitro with their claim that "Rad52 acts as a client of the LLPS but does not drive its formation"?

We thank the reviewer for this interesting remark. Recent studies have shown that the physical nature of proteins complexes can differ from in vitro to in vivo assays. In a recent publication on the nature of heterochromatin (Erdel et al., 2020), the authors confirm that the HP1α can form droplets in vitro as also observed by Larson et al., Nature 2018. However, HP1α droplets formation in vitro requires a high protein concentration (> 40 μM): it is not certain that the in vitro behavior of highly concentrated and phosphorylated HP1 accurately recapitulates the in vivo biology with physiological HP1 level (~ 1 μM). Indeed, the results in vivo presented by Erdel et al., 2020, clash with a simple LLPS mechanism for heterochromatin subcompartments; instead, they are more consistent with a collapsed “chromatin-globule model” (or polymer-polymer phase separation model, PPPS).

As underlined in the recent literature (Gitler et al., Mol Cell 2020 Previews, McSwiggen et al., 2019), with the advances in super resolution microscopy and biophysical approaches, what appear to be droplets at first glance might in fact follow more complex models.

Taking into consideration the work from the Brangwynne lab showing that LLPS composed by multiple proteins with heterotypic multicomponent interactions do not exhibit a fixed saturation concentration, we propose that the formation of repair foci could be driven by a combination of several interacting proteins including Rad52. We now comment on this point.

5) What is the evidence that the biophysical properties observed are of direct relevance to DNA repair? For example, is the mobility of Rad52 within the repair focus important for repair? Is the difference in diffusion kinetics within and outside of the repair focus important for genome stability? What could the authors do to alter that diffusion profile and what would be the consequence on repair? Also, addressing this point implies the need to use a more physiologically relevant system with repairable DSBs, and not the irreparable DSB system used here.

Thank you for the remark and we agree with the reviewer’s comment. To verify if our observations are conserved in a system with repairable DSBs, we measured Rad52 mobility in diploids cells, after the induction of a single DSB at the same locus as in haploids (LYS2). We obtained results very similar to the one obtained in haploid cells: with Rad52 molecules exhibiting: (i) 3 populations, (ii) confined motion inside foci, (iii) a difference in diffusion coefficient of 8 between internal Rad52 mobility inside foci and the whole focus mobility. It is important to note that all our experiments in haploid cells where performed 2 hours after induction which correspond to the step of search for an ectopic homologous sequence (Piazza et al., 2019). This could explain why Rad52 has the same behavior in the presence and in the absence of a donor sequence.

We now present the results in Figure 2—figure supplement 5, and we discuss this point in the text.

One should easily compare wild-type Rad52 to known Rad52 mutants (that partly or fully abrogate function) to see if there is any correlation between intra-focus mobility profiles and repair. In other words, experiments along these lines may indicate whether a particular intra-focus mobility profile consistently correlates with repair, while this profile is absent in non-functional mutants, or a particular type of mutant. These experiments should be very feasible.

It will of course be interesting to combine SPT and genetics to test to which extent repair proteins mobility, residence time inside foci internal dynamics and type of motion are modified when DNA repair is deficient. In the revised manuscript, we show that Rad52-SOMUylation affect some of the observables accessible by SPT (Figure 2, Figure 2—figure supplement 3,). We agree that testing different Rad52 mutant will be very interesting but this will be the subject of another manuscript.

6) Can the authors visualize the fusion of the Rad52 foci/DSBs in live cells within their experimental systems?

The fusion of repair foci has been already observed both in yeast (Oshidari et al., 2020 and Waterman et al., 2019) and in human cells (Altmeyer et al., 2015) by tracking repair foci at larger time scales (several minutes). We now comment this point in the Discussion. Here we focused on foci size and internal dynamics inside foci in response to 1 versus 2 DSBs, which brings new information that are only accessible by single molecule microscopy.

7) The statistical significance of most presented data is either lacking or unclear. This needs to be carefully addressed. Providing additional data files may also help the authors strengthen their findings.

Thank you for this remark. We have added statistical tests throughout the revised manuscript. To be more specific, we added:

– goodness of fit for all the MSD fit, indicating the Pearson's chi-squared (see revised Table 1)

– the p-values of the Wilcoxon-Mann-Whitney test for the data with over-expressed levels of Rad52 (Figure 6—figure supplement 1)

Text revisions:

– Several statements made are not supported by the data and without clearly stating that the statements represent speculations. E.g. longer tail is due to Rad52 molecules diffusing slowly inside the focus; observing the 2 populations also in G1 does not necessarily mean that the 2 populations in S/G2 do not reflect replication forks at all.

We now modulate our interpretation of the 2 populations in S/G2.

The authors need to carefully revise their claims/statements and consider alternative explanations. Also, the writing is often unclear or confusing and the authors should consider substantially revising it to clarify their claims, clearly indicate speculations that are not supported by the data, and make the text as accessible as possible to non-specialists.

In the revised version, we made the text more accessible, for example,

– we moved in the method or figure legend some technical details explaining how the diffusion coefficient is estimated inside repair foci (see Results and legend of Figure 2—figure supplement 2).

– We explained the distribution of traces length in a more accessible manner.

– We removed some redundancies in the Discussion and discuss the models of sub-compartments in a more accessible manner.

– How was the cell cycle stage determined? This should be better explained.

Cell cycle stage was assessed by the presence and the size of a bud on the transmission image. We now clarified this point in the text (figure legend of Figure 1).

– Figure 1—figure supplement 1 data appear to show the existence of a partial loss of Rad52 function in the Rad52-Halo cells. This should be clearly stated in the Results and consequent limitations/caveats discussed. Also, please clarify whether Figure 1—figure supplement 1 shows the viability of Rad52-Halo cells in the presence or absence of JF646.

We have commented on this point in the legend of Figure 1—figure supplement 1.

– The authors present no direct evidence for an "attractive potential" that drives molecules towards the centre of the focus. For example, what if the “attractive potential” is simply the focus' boundary surface tension creating a barrier against which some of the molecules inside the focus bounce back towards the centre of the focus?

We thank the reviewer for this comment that definitely needs clarification. When we refer to an attractive potential, we do not mean a potential that attracts the molecules towards the center of the focus, but a potential that, exactly like it would be the case for surface tension, hinders the passage out of the focus for the molecules. To clarify this important passage, we have updated the text, in the last paragraph of the Results part, following text for Figure 7.

– The authors state that "Here, we found that upon different levels of Rad52 over-expression, the background concentration increases (Appendix 1—figure 1) suggesting that Rad52 might not be the driving molecule responsible for the LLPS formed at the damaged site." The logical transition here is unclear.

We have clarified this point and added a reference.

– It is difficult to judge the novelty of this work, as key papers that showed similar conclusions or datasets are not cited. Without direct comparison with other data sets, it is difficult to see exactly where this paper goes beyond published studies.

Here are a few key examples:

a) In the last year the Haber lab published a very similar study in Plos Genetics (Waterman et al., 2019). Although they tracked Ddc2 and Rad51, they also looked at the behavior of separate foci and this paper is not even cited. The data should be compared at the very least.

The work by Waterman et al., 2019 is actually different from our present work although complementary. Waterman et al. did not monitor the dynamics of single molecules but the dynamics of repair foci decorated either by Rad51 or Ddc2 in a strain experiencing 3 DSB. They monitored the aggregation and mobility of DSBs into a single focus and show that these events were independent of Rad52.

This is paper is now cited in the revised version.

b) The characteristics of 53BP1 foci have been extensively studied by many labs including those of Altmeyer, Scherthan, DeLange and others, with very similar findings as Miné-Hattab reports for Rad52 (for example, Kilic et al., 2019; Sollazzo et al., 2018, as well as the single molecule work of the lab of Eric Greene). Moreover both rad52 and PCNA foci were studied by Essers et al. (Kanaar and Vermeulen) MCB 2005. 25(21): 9350-9359 and EMBO J. 2002 Apr 15. Comparisons with these studies needs to be made.

Thank you, we added these references and discuss them in the revised manuscript.

c) A number of earlier studies followed Rad52 foci in budding yeast on induced double strand breaks (even using the I-Sce1-cut system used here) that are not taken into consideration. The diffusion coefficients presented here have to be compared with these earlier studies and differences should be resolved by comparing techniques and conditions of imaging. For instance, Dion et al., 2012).

We now compare the diffusion coefficient of the whole focus measured here with the one obtained in Dion et al., 2012.

– It is unclear if the “absence of DNA damage” condition discussed in the first section of the Results is the non-induced version of the system described in the second section of the Results.

The “absence of DNA damage” condition is indeed the same experiment described in the second section of the Results and referred as “no DSB” in the figures. For this condition, we used a strain without a I-SceI cut site and grown in the same conditions as the strain used to induce a DSB to ensure a fair comparison between the two conditions (with or without DSB). We now clarify this point in the Materials and methods section.

Also regarding these sections, it seems that the “absence of DNA damage” control conditions were not conducted as part of the same experiments with the I-SceI DSB.

As mentioned above, experiments performed in the absence or in the presence of DNA damage uses a different strain and were done separately. We have repeated these experiments 3 to 6 times, and one set of data includes the experiments with and without DSB done the same day. We also use the “absence of DNA damage” condition as a control experiment for the stability of our set-up: this experiment has been systematically performed and analysed once a year and similar diffusion coefficients were obtained.

– Figure 2C is a little underwhelming. I would like to see one data panel in the main text with the cumulative distribution +/- DSB.

We now present the cumulative distribution of traces length for Rad52-Halo/JF646, in the absence and in the presence of a DSB (see new Figure 2—figure supplement 1).

– "in the case of diffusion coefficients D<0.1 μm2/s, we use mean square displacement analysis allowing us to substrate the noise." Subtract the noise.

We made the correction.

– One technical point. The fluorescence lifetime of a freely diffusing fluorophore (reported half-life of JF646 is 2.1 sec) is not the same as if it is bound to a protein; in the latter context the fluorophore would diffuse more slowly in the illumination path and be subjected to more rapid photobleaching, making 2.1 sec an overestimation of the half-life/underestimation of the decay rate. At a frame interval of 20ms, photobleaching may be a competitive rate for the disappearance of signal and I am not sure the lack of applying photobleaching correction or saying that the short tracks are not affected by that rate is properly justified.

We thank the reviewer for this remark. The JF646 used in this study are flurogenic, meaning that they emit light only when they are bound to a Halo molecule. Thus, there is no signal when free JF646 dyes diffuse inside a cell that does not express Halo. To measure the half life time of JF646, we used the strain harbouring Rad52-Halo, in the absence of DSB and we incubated the JF646 with the cells for 1h for 50 nM (a concentration 10 times higher than the one used to be in the single molecule regime). This point is explained in Figure 1—figure supplement 3.

https://doi.org/10.7554/eLife.60577.sa2

Article and author information

Author details

  1. Judith Miné-Hattab

    Institut Curie, PSL University, Sorbonne Université, CNRS, Nuclear Dynamics, Paris, France
    Contribution
    Conceptualization, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Writing - original draft, Writing - review and editing
    For correspondence
    judith.Mine@curie.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9986-4092
  2. Mathias Heltberg

    1. Institut Curie, PSL University, Sorbonne Université, CNRS, Nuclear Dynamics, Paris, France
    2. Laboratoire de Physique de l’Ecole Normale Supérieure, PSL University, CNRS, Sorbonne Université , Université de Paris, Paris, France
    Contribution
    Software, Formal analysis, Validation, Investigation, Methodology, Writing - original draft
    Competing interests
    No competing interests declared
  3. Marie Villemeur

    Institut Curie, PSL University, Sorbonne Université, CNRS, Nuclear Dynamics, Paris, France
    Contribution
    Methodology, Design and construction of the strains
    Competing interests
    No competing interests declared
  4. Chloé Guedj

    Institut Curie, PSL University, Sorbonne Université, CNRS, Nuclear Dynamics, Paris, France
    Contribution
    Conceptualization, Methodology, Design and construction of the strains
    Competing interests
    No competing interests declared
  5. Thierry Mora

    Laboratoire de Physique de l’Ecole Normale Supérieure, PSL University, CNRS, Sorbonne Université , Université de Paris, Paris, France
    Contribution
    Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-5456-9361
  6. Aleksandra M Walczak

    Laboratoire de Physique de l’Ecole Normale Supérieure, PSL University, CNRS, Sorbonne Université , Université de Paris, Paris, France
    Contribution
    Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing - review and editing
    Competing interests
    Reviewing editor, eLife
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-2686-5702
  7. Maxime Dahan

    Institut Curie, PSL University, Sorbonne Université, CNRS, Physico Chimie Curie, Paris, France
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Validation, Investigation, Methodology
    Competing interests
    No competing interests declared
  8. Angela Taddei

    1. Institut Curie, PSL University, Sorbonne Université, CNRS, Nuclear Dynamics, Paris, France
    2. Cogitamus Laboratory, Paris, France
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Validation, Investigation, Methodology, Project administration, Writing - review and editing
    For correspondence
    angela.taddei@curie.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3217-0739

Funding

Agence Nationale de la Recherche (ANR-11-LABEX-0044 DEEP)

  • Angela Taddei

Agence Nationale de la Recherche (ANR-10-IDEX-0001-02 PSL)

  • Angela Taddei

Agence Nationale de la Recherche (ANR-15-CE12-0007)

  • Angela Taddei

Agence Nationale de la Recherche (ANR-12-PDOC0035–01)

  • Judith Miné-Hattab

Fondation pour la Recherche Médicale (DEP20151234398)

  • Thierry Mora
  • Aleksandra M Walczak
  • Angela Taddei
  • Judith Miné-Hattab

Conseil National de la Recherche Scientifique (ANR-17-CONV-0005)

  • Angela Taddei

France Bioimaging National Infrastructure (ANR-10-INBS-04)

  • Angela Taddei
  • Judith Miné-Hattab

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

This work is dedicated to the memory of our friend and colleague Maxime Dahan. The authors thank Vu Nguyen and Carl Wu for fruitful exchanges on the use of JF dyes and for sharing plasmids. We thank Eric Coic for sharing with us a plasmid harboring Rad52-SUMO and to Siyu Liu for help in the revisions. We thank Antoine Coulon, Maxime Woringer, for their comments on the manuscript, as well as Bassam Hajj for his help during the construction of the PALM set-up. We also thank Mickael Garnier for his help on image analysis and Gaizka Le Goff for his help on preliminary experiments performed during his L3 internship. This work is at the origin of the Muse-IC project, a collaborative project between musicians and composers aiming to create musical pieces inspired by recent scientific discoveries (Dolgin, 2019).

AT team was financially supported by funding from the Labex DEEP (ANR-11-LABEX-0044 DEEP and ANR-10-IDEX-0001–02 PSL), from the ANR DNA-Life (ANR-15-CE12-0007), ANR-12-PDOC-0035–01, Fondation pour la Recherche Médicale (DEP20151234398), the AT team, TM and AMW. were supported by CNRS grant 80prime PhONeS and Q-life (ANR-17-CONV-0005), the authors greatly acknowledge the PICT-IBiSA@Pasteur Imaging Facility of the Institut Curie, member of the France Bioimaging National Infrastructure (ANR-10-INBS-04).

Senior Editor

  1. Jessica K Tyler, Weill Cornell Medicine, United States

Reviewing Editor

  1. Irene E Chiolo, University of Southern California, United States

Publication history

  1. Received: June 30, 2020
  2. Accepted: January 26, 2021
  3. Accepted Manuscript published: February 5, 2021 (version 1)
  4. Version of Record published: March 1, 2021 (version 2)
  5. Version of Record updated: March 5, 2021 (version 3)

Copyright

© 2021, Miné-Hattab et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 752
    Page views
  • 179
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Physics of Living Systems
    Farha Naaz et al.
    Research Article

    It is well known that flagellated bacteria, such as Escherichia coli, sense chemicals in their environment by a chemoreceptor and relay signals via a well-characterised signalling pathway to the flagellar motor. It is widely accepted that the signals change the rotation bias of the motor without influencing the motor speed. Here, we present results to the contrary and show that the bacteria is also capable of modulating motor speed on merely sensing a ligand. Step changes in concentration of non-metabolisable ligand cause temporary recruitment of stator units leading to a momentary increase in motor speeds. For metabolisable ligand, the combined effect of sensing and metabolism leads to higher motor speeds for longer durations. Experiments performed with mutant strains delineate the role of metabolism and sensing in the modulation of motor speed and show how speed changes along with changes in bias can significantly enhance response to changes in its environment.

    1. Computational and Systems Biology
    2. Physics of Living Systems
    Lakshya Chauhan et al.
    Research Article Updated

    Phenotypic (non-genetic) heterogeneity has significant implications for the development and evolution of organs, organisms, and populations. Recent observations in multiple cancers have unraveled the role of phenotypic heterogeneity in driving metastasis and therapy recalcitrance. However, the origins of such phenotypic heterogeneity are poorly understood in most cancers. Here, we investigate a regulatory network underlying phenotypic heterogeneity in small cell lung cancer, a devastating disease with no molecular targeted therapy. Discrete and continuous dynamical simulations of this network reveal its multistable behavior that can explain co-existence of four experimentally observed phenotypes. Analysis of the network topology uncovers that multistability emerges from two teams of players that mutually inhibit each other, but members of a team activate one another, forming a ‘toggle switch’ between the two teams. Deciphering these topological signatures in cancer-related regulatory networks can unravel their ‘latent’ design principles and offer a rational approach to characterize phenotypic heterogeneity in a tumor.