Construction of the RAD52-GFP Fusion Gene.

A red-shifted enhanced version (10C) of the GFP gene (28) was a generous gift from R. Tsien (University of California, San Diego, La Jolla, CA). This version of GFP was fused by PCR to either the 5′- or 3′-two-thirds of the Kluyveromyces lactis URA3 gene, using adaptamer technology (29, 30). Next, the resulting fragments were ligated into the SacII site of pRS423 (31) to produce plasmids pWJ1164 and pWJ1165. By using the principle in our previously described allele replacement method (32), these plasmids serve as basic tools to target GFP specifically to any site in the genome. In brief, two GFP-URA3 PCR products were amplified from pWJ1164 and pWJ1165 by using appropriate primers (GFPstart-F: 5′-ATGAGTAAAGGAGAAGAAC-3′, URA3int-R: 5′-GAGCAATGAACCCAATAACGAAATC-3′ and GFPend-R: 5′-TTTGTATAGTTCATCCATGC-3′, URA3int-F: 5′-CTTGACGTTCGTTCGACTGATGAGC-3′, respectively; Fig. ​Fig.11A). Next, two pairs of PCR primers were designed to amplify ≈400 base pairs upstream (UF: 5′-CCTTTGTTACAGCTAAGGC-3′, URtag1: 5′-gttcttctcctttactcatCCCAGTAGGCTTGCGTGCATG-3′) and downstream (DFtag2: 5′-ggatgaactatacaaataaCCCGCTTCCTGGCCGAAAC-3′, DR: 5′-AATGAACCTAAGGATTCCGC-3′) of the GFP insertion site at the 3′ end of RAD52 by using genomic DNA (or plasmid DNA) as a template (Fig. ​(Fig.11A). The upstream and downstream fragments each contain an additional 19 base pairs of DNA sequence (tag1 and tag2 shown in small letters) complementary to parts of the two GFP-URA3 PCR products amplified from pWJ1164 and pWJ1165. The reverse and complementing sequences at the ends (the tags) facilitate the fusion of the two pairs of PCR products to generate the two fragments shown in Fig. ​Fig.11B. After cotransformation of these two PCR products into W2109-14C, in vivo homologous recombination generates a GFP direct repeat flanking an intact K. lactis URA3 gene downstream of RAD52 (Fig. ​(Fig.11C). Loss of the URA3 gene resulting from direct repeat recombination is selected on 5-FOA medium to generate a “clean” genomic RAD52-GFP fusion. Western blot analysis of Rad52 was performed essentially as described (16) except that the Rad52 antibody was affinity purified according to standard procedures by employing a Rad52-Sepharose 4B resin.

Open in a separate window

Figure 1

GFP-tagging strategy for construction of a RAD52-GFP strain. (A) In the first PCR reactions, two pairs of PCR primers were used to amplify sequences upstream and downstream of the GFP insertion site at the 3′ end of RAD52 (see Materials and Methods). (B) Adaptamer-mediated fusion of PCR products. The resulting upstream and downstream fragments each contained an additional 19 base pairs complementary to GFP sequence to allow the fusion of these two PCR products to DNA sequences consisting of GFP fused to either the 5′- or 3′-two-thirds of the K. lactis URA3 gene. (C) Integration by in vivo recombination. The two DNA fragments produced in the second set of PCR reactions were transformed into S. cerevisiae to integrate a direct repeat of GFP immediately upstream of the endogenous RAD52 stop codon. Subsequently, pop-out recombinants were selected on 5-FOA to remove the URA3 marker and the extra copy of GFP.

Microscopy.

Unless otherwise noted, all experiments were performed at 23°C to allow the GFP chromophore to form efficiently (32). Before epifluorescence microscopy, cells were washed twice in either SC (mitotic experiments) or SPO (meiotic experiments) medium. Next, cells were immobilized on a glass slide by mixing them with a 37°C solution of 1.2% (wt/vol) low melting agarose (NuSieve 3:1 from FMC) containing the appropriate medium. To prevent evaporation, the cover glass was sealed with a mixture of equal volumes of molten bees' wax, petroleum jelly (VWR Scientific), and lanolin (Sigma). The filters used to visualize Rad52-GFP (excitation 480 nm; emission 535 nm) and 4′,6-diamidino-2-phenylindole (DAPI; excitation 365 nm, emission 450 nm) were from Omega Optical (Brattleboro, VT). Images were acquired by using a cooled CCD camera (Star-1 from Photometrics, Tucson, AZ) mounted on a Zeiss Axioplan II microscope with a Plan-Apochromat 100×, 1.4 numerical aperture (NA) objective lens. The illumination source was a 100-W mercury arc lamp. Images were obtained at 0.2-μm intervals along the z axis by using a piezo-controlled stepper and Rad52-GFP foci were counted by inspecting all of the focal planes intersecting each cell. Integration time for acquisition of the fluorescent images was 200 ms. Images were acquired in the IP LAB software (Scanalytics, Billerica, MA) and prepared for publication in SCION V1.62 (National Institutes of Health) and ADOBE PHOTOSHOP (Adobe Systems, Mountain View, CA).

DNA in living cells was stained for visualization by adding 10 μg/ml DAPI to the culture 30 min before imaging. Selected strains were made rho⁰ (mitochondrial DNA negative) before staining to eliminate any signal from mitochondrial DNA (27). Localization of Rad52-GFP was not affected by the absence of mitochondria nor by DAPI staining (data not shown).

γ-Irradiation.

The γ-ray sensitivities of strains were determined by growing cultures in YPD to mid-log phase at 23°C. An appropriate number of cells were plated on YPD plates and exposed to different doses of γ-rays by using a Gammacell-220 ⁶⁰Co irradiator (Atomic Energy, Ottawa). Cells analyzed by microscopy were pregrown in YPD at 23°C until OD₆₀₀ reached 0.2. At this point, the liquid cultures were exposed to defined doses of irradiation and aliquots of the cultures were processed immediately for imaging.

To estimate the number of rad52-repairable lesions induced by γ-irradiation, we made the following two assumptions: (i) γ-ray-induced DNA damage occurs randomly in a population of cells and (ii) one rad52-repairable lesion is sufficient to kill a cell (33). The number (n) of γ-ray-induced DNA lesions per cell that are lethal in the absence of RAD52 can be described by a Poisson distribution [P_n(x) = e^−μμⁿ/n!, −μ = kx, n = 0, 1, 2, . . . , where x is the γ-ray dose and μ the mean number of lethal lesions per cell at that dose]. Accordingly, the probability of a cell surviving a given γ-ray dose (x) is P₀(x) = e^kxμ⁰/0! = e^kx ln(P₀(x)) = kx. By fitting the γ-ray survival curve of rad52Δ strains (Fig. ​(Fig.22B) to this equation, k is estimated to be −0.5. Thus, the γ-ray dose required to generate an average of one rad52-repairable lesion per cell (μ = 1) is −μ/k = −1/−0.5 = 2 krad.

Open in a separate window

Figure 2

Characterization of the RAD52-GFP strain. (A) Western blot analysis of Rad52 and Rad52-GFP. Protein extracts obtained from RAD52 (W2297-12D) and RAD52-GFP (W2297-4A) strains were fractionated by SDS/PAGE. Rad52 and Rad52-GFP were visualized by using anti-Rad52 antibody. (B) γ-Ray survival curves for RAD52 (W2297-12D), RAD52-GFP (W2297-4A), and rad52 null (W2366-6B) strains. Each data point represents the mean of three independent trials.

Induction of DSBs by the HO Endonuclease.

Strains containing no, one, or two accessible HO-cut sites were transformed with pJH132 (kindly provided by Jim Haber, Brandeis University, Waltham, MA) containing a galactose-inducible HO gene and a TRP1 selectable marker (34). Transformants were grown at 23°C in SC-Trp (W2334–1D and W2322–5B) or SC-Trp-Ura (W2613–22A and W2322–8A) medium containing 2% raffinose. At OD₆₀₀ ≈ 0.2, HO was induced by addition of galactose to a final concentration of 2% and incubation continued for 1 h before processing of cells for microscopy. The efficiency of mating-type switching was calculated as 1 − Na_after/(Na_after + Nα_after − (Nα_before (Na_after + Nα_after)/(Na_before + Nα_before))), where N is the number of cells with the indicated mating-type before and after 1.5 h of HO induction, respectively.

Rad52-GFP Forms Nuclear DNA Repair Foci During S Phase.

Rad52-GFP was visualized in unsynchronized mitotic haploid cells by epifluorescence microscopy. Based on cellular morphology, cells were grouped into G₁ phase (unbudded), S (small-budded), and G₂/M (large-budded) (39). Simultaneous staining of the DNA with DAPI shows that in most cells Rad52-GFP displays a diffuse nuclear localization throughout the mitotic cell cycle (Fig. ​(Fig.33A). Interestingly, Rad52-GFP relocalizes and forms foci in 22% of S phase cells suggesting a specific role of Rad52 during DNA replication (Fig. ​(Fig.33A and Table ​Table3).3). We previously provided (40) a strong link between replication and recombination by showing that Holliday junction intermediates are formed in rDNA specifically during S phase. The amount of these structures dramatically increases in a temperature-sensitive pol12-100 strain that encodes a mutated subunit of the polymerase α complex. Because formation of these recombination intermediates depends on RAD52 (40), we examined whether the high level of Holliday junction formation in pol12-100 strains affects Rad52-GFP distribution. Indeed, even at the permissive temperature we found a striking increase in the frequency of Rad52-GFP foci in S phase (Table ​(Table3).3). Furthermore, transfer to the nonpermissive temperature results in accumulation of cells in G₂/M that contain Rad52-GFP foci. Thus, replication defects caused by an impaired polymerase α complex trigger Rad52-GFP focus formation.

Open in a separate window

Figure 3

Induction of nuclear Rad52-GFP foci by γ-irradiation. (A and B) Microscopy of cells (J861) expressing the Rad52-GFP fusion protein. Pictures shown are pseudocolored monochrome images: Rad52-GFP in green and DAPI-stained DNA in blue. (A) Diffuse nuclear localization of Rad52-GFP in nonirradiated cells. Note that one cell contains a spontaneous Rad52-GFP focus. (B) Relocalization of Rad52-GFP to distinct foci in cells after exposure to 20 krad of γ-irradiation. At this dose, most cells contain Rad52-GFP foci. (C–F) Dose dependence of Rad52 focus formation in G₁ (open diamonds), S (open triangles), and G₂/M (solid squares) phase haploid (W2297–4A) and diploid (W2682) cells. All focal planes were inspected for Rad52 foci. The number of foci observed after γ-irradiation was constant in the time frame from 15 to 90 min after irradiation and all numbers were obtained within this time frame. (C) Percentage of haploid cells with ≥1 focus. (D) The average number of Rad52 foci that are observed in haploid cells with ≥1 focus. (E) Percentage of diploid cells with ≥1 focus. (F) The average number of Rad52 foci that are observed in diploid cells with ≥1 focus.

Table 3

Frequencies of Rad52-GFP focus formation in G₁, S, and G₂/M cells

Strain	G1 (unbudded)	S (small budded)	G2/M (large budded)
Wild type,* 23°C	0/316(0%)	24/109(22%)	1/82(1%)
pol12-100,† 23°C	0/100(0%)	21/43(49%)	5/50(13%)
pol12-100,† 37°C	2/67(3%)	5/43(12%)	97/157(62%)
Wild type,* 37°C	0/284(0%)	10/87(11%)	11/67(16%)
mec1 sml1,‡ 23°C	15/186(8%)	29/56(52%)	19/34(54%)
sml1,§ 23°C	2/202(1%)	19/77(25%)	4/81(5%)

Open in a separate window

^*Strain W2297-4A. 

^† Strain UM10-25D. 

^‡ Strain W2381-10B. The essential function of MEC1 is suppressed by mutation of sml1, an inhibitor of ribonucleotide reductase, but importantly, mec1 sml1 strains still fail to arrest at the G₁ and G₂/M checkpoints during genotoxic stress (68). Hence, we employed the mec1 sml1 strain background to test the effect of a DNA damage checkpoint defect on Rad52-GFP focus formation. 

^§ Strain W2382-2A. 

To investigate the possibility that Rad52-GFP foci represent sites of DNA damage repair, we determined their formation during RAD52-dependent DNA repair and recombination processes. First, DNA damage was induced in the RAD52-GFP strains by exposing them to 20 krad of γ-irradiation. A dramatic increase in the proportion of cells containing Rad52 foci was observed, suggesting that formation of these aggregates correlates with recombinational repair (Fig. ​(Fig.33B). Surprisingly, cells on average contain only one Rad52-GFP focus despite the fact that a dose of 20 krad generates ≈10 DNA lesions per cell that require Rad52-dependent repair in order for the cell to survive (see Materials and Methods). This observation suggests that one focus may represent the repair of multiple lesions. To investigate this possibility, the γ-ray dose dependence of focus formation was analyzed in a dose range (2.5–160 krad) expected to generate 1.25–80 lesions per cell. If each focus represents repair of a single DNA lesion, then the number of foci is expected to increase proportionally to the γ-ray dose. First, we noticed that the process of focus formation is more easily triggered in S phase and G₂/M (Fig. ​(Fig.33C). Thus, after a dose that produces 1.25 lesions per cell (2.5 krad), more than 50% of the cells that are either in S phase or in G₂/M form at least one Rad52-GFP focus. In contrast, cells in G₁ require a 16-fold higher dose (40 krad) to reach this level. Among the cells in S or G₂/M phase that contain at least one focus, we found that a 64-fold increase in the γ-ray dose (160 krad or 80 lesions per cell) results in only a 2-fold increase in the average number of Rad52-GFP foci per cell (Fig. ​(Fig.33D). Furthermore, cells in G₁ rarely display more than one focus irrespective of γ-ray dose. Thus, the number of Rad52-GFP foci and γ-ray dose are not directly proportional, indicating that each focus likely represents the repair of multiple DNA lesions.

The biased distribution of Rad52-GFP foci in the cell cycle may reflect that haploid cells in G₁ lack a homologous chromosome that can be used as a template for recombinational repair. To evaluate this possibility, induction of Rad52 foci by γ-irradiation was studied in diploid RAD52-GFP/RAD52-GFP cells, where the homologous chromosome can provide a DNA substrate for recombinational repair throughout the entire cell cycle. Accordingly, the number of Rad52 foci formed in diploid cells after exposure to doses of γ-irradiation ranging from 1.25 to 80 krad (equivalent to 1.25–80 lesions per cell) was determined. Interestingly, the number of Rad52 foci induced by γ-irradiation in diploid cells is approximately twice the number observed in haploid cells, but is still not proportional to the number of DNA lesions induced (compare Figs. ​Figs.33D and ​and22F). However, like haploid G₁ cells, diploid G₁ cells require a 16-fold higher dose than S phase cells to reach the level where 50% of the cells contain at least one focus (Fig. ​(Fig.33E). This observation demonstrates that DNA DSBs induced in the presence of a homologous DNA substrate alone is not sufficient to elicit Rad52 focus formation. Indeed, the cell also needs to be in the proper phase of the cell cycle to form this structure.

Numerous Rad52 Foci Are Formed During Meiotic Recombination.

Because γ-irradiation produces a variety of lesions (41) that are likely to have different repair requirements, we tested Rad52-GFP focus formation specifically during DNA DSB repair. Multiple DSBs are generated by Spo11 during the onset of meiotic recombination at early prophase of meiosis I (42, 43). Efficient repair of these breaks requires Rad52 (44). After shifting RAD52-GFP/RAD52-GFP diploid strains to SPO medium, we found that Rad52-GFP relocalizes to form ≈15 foci per nucleus (Fig. ​(Fig.44A). The number of meiotic Rad52-GFP foci observed in vivo is similar to that seen after immunostaining of Rad52 in nuclear spreads (45). We next examined meiotically induced spo11/spo11 mutants, which are blocked for DSB formation (46). In contrast to SPO11/SPO11 diploids, multiple Rad52-GFP foci were not observed and the occasional single focus observed in these mutants is likely due to repair of DNA damage that arose during premeiotic S phase (data not shown).

Open in a separate window

Figure 4

Formation of Rad52-GFP foci during sporulation. Microscopy of sporulating cells (W2682) expressing the Rad52-GFP fusion. (A) A representative sporulating cell before the first meiotic division after 11 h in SPO medium. Four optical sections separated by 0.4 μm are displayed (see schematic). (B) A four-spored tetrad after 15 h in SPO medium is displayed showing that Rad52 foci have disassembled after the second meiotic division.

Rad52 Foci Form After Induction of Specific DSBs by the HO Endonuclease.

Well defined DSBs were also induced by the HO endonuclease (47) in a RAD52-GFP strain. Two different HO-cut sites were investigated. One is the HO-cut site that exists naturally at the mating-type locus (MAT-HO). When this site is cleaved by the HO endonuclease, the resulting DSB is repaired by gene conversion (48). The other HO-cut site is situated in a LEU2 direct repeat at the LEU2 locus (leu2-HO). In this case, the HO endonuclease induces a DSB that is repaired by single-strand annealing (35). Strains were constructed that contain no accessible HO-cut sites, a single HO-cut site (either MAT-HO or leu2-HO), or two HO-cut sites (both MAT-HO and leu2-HO). After induction of the HO endonuclease in strains that contained one HO-cut site at either MAT or leu2, Rad52 foci were observed in 50–60% of S phase cells and 20–30% of G₂/M phase cells (Fig. ​(Fig.55A). In contrast, foci are formed in only 15% of S and 10% of G₂/M cells in the strain that contains no accessible HO-cut site. Similar to what was observed after γ-irradiation (Fig. ​(Fig.33C), very few G₁ cells form foci upon HO induction. Thus, HO-induced breaks, like γ-ray-induced DNA damage, preferentially generate Rad52 foci in S phase. Interestingly, when the strains containing two HO-cut sites were analyzed, the average number of foci observed per cell did not double when compared with strains that contain only one HO-cut site (Fig. ​(Fig.55B). Hence, the number of Rad52-GFP foci formed in vivo is largely independent of the number of DNA DSBs, regardless of whether they are induced by γ-irradiation or by HO. This result suggests that one focus may represent the repair of multiple DNA lesions within a limited number of centers in the nucleus.

Open in a separate window

Figure 5

Rad52-GFP foci induced by the HO endonuclease. (A) The percentages of cells in G₁, S, and G₂/M containing foci after induction of HO in strains with no (W2334–1D), one (W2322–5B and W2613–22A), or two (W2322–8A) HO-cut sites, are shown. (B) The average number of Rad52 foci that are observed in G₁ (open diamonds), S (open triangles), and G₂/M (solid squares) phase cells with ≥1 focus.

We thank members of the Rothstein laboratory and Aleksei Aleksenko for helpful discussions concerning this work, as well as Liza Pon and members of her laboratory for skillful assistance during microscopy. We also thank Roger Tsien for providing the enhanced GFP clone. This research was supported by grants from the Dagmar Marshall Foundation, the Leo Research Foundation, the Neye Foundation, the Danish Cancer Society, and the Alfred Benzon Foundation (to M.L.); grants from the Tonnessen Foundation, the Danish Medical Research Council and the Danish Technical Research Council (to U.H.M.); Grants HG01620 and GM50237 from the National Institutes of Health (to R.R.); and National Institutes of Health Shared Instrument Grant S10 RR10506 and National Institutes of Health Grant P30 CA13696 to the Columbia Cancer Center.