Slx5-Slx8 ubiquitin ligase targets active pools of the Yen1 nuclease to limit crossover formation

Abstract

The repair of double-stranded DNA breaks (DSBs) by homologous recombination involves the formation of branched intermediates that can lead to crossovers following nucleolytic resolution. The nucleases Mus81-Mms4 and Yen1 are tightly controlled during the cell cycle to limit the extent of crossover formation and preserve genome integrity. Here we show that Yen1 is further regulated by sumoylation and ubiquitination. In vivo, Yen1 becomes sumoylated under conditions of DNA damage by the redundant activities of Siz1 and Siz2 SUMO ligases. Yen1 is also a substrate of the Slx5-Slx8 ubiquitin ligase. Loss of Slx5-Slx8 stabilizes the sumoylated fraction, attenuates Yen1 degradation at the G1/S transition, and results in persistent localization of Yen1 in nuclear foci. Slx5-Slx8-dependent ubiquitination of Yen1 occurs mainly at K714 and mutation of this lysine increases crossover formation during DSB repair and suppresses chromosome segregation defects in a mus81∆ background.

Fig. 1

Yen1 is sumoylated in vivo and in vitro. a A wild-type chromosomally tagged YEN1-HA strain was synchronized with alpha factor and released into fresh medium to observe phosphorylation of Yen1 by immunoblot (upper) and progression through the cell cycle by FACS (lower). b Wild-type strains expressing Yen1-HA, with (+) or without (−) pCUP-6xHIS-Smt3, were subjected to MMS challenge followed by denaturing Ni-NTA pull-down and immunoblot analysis. Yen1 was detected by anti-HA (top and middle) and a prominent sumoylated doublet is indicated (black rhombus). Membranes were also probed with anti-Smt3 (bottom). Note that un-sumoylated Yen1 binds to Ni due to a histine-rich region. c Yen1-HA was overexpressed in wild-type asynchronous cells, immunoprecipitated with anti-HA, eluted by HA peptide competition and mixed with Aos1-Uba2, Ubc9, and Smt3-3KR in the presence or absence of ATP. After immunoblotting with anti-HA sumoylated forms were detected in the presence of ATP that migrate at similar sizes to those detected in the PD experiments shown in b (far right duplicate for comparison). A control reaction was made with HA-immunoprecipitation of a yen1∆ strain eluted with the same amount of HA peptide. d yen1∆ cells expressing Yen1-HA from a Gal-inducible plasmid or harboring a control plasmid were subjected to MMS treatment (0.03%), and extracts were immunoprecipitated using anti-HA prior to immunoblotting with anti-Smt3 (left) or anti-HA (right). e Indicated strains (WT, siz1∆, siz2∆, or siz1∆ siz2∆) with (+) or without (−) pCUP-6xHIS-Smt3 were subjected to pull-down analysis of Smt3 as in b in conditions of MMS damage and eluates were analysed by immunoblot. The input used for PD was immunoblotted to allow normalization and comparison between strains (bottom). f Purified recombinant 6His-HA-Yen1 protein was incubated under sumoylation conditions with the indicated concentrations of Siz2 followed by immunoblotting with anti-HA. Asterisk indicates breakdown products of Yen1 carried from purification. g Sumoylation reactions were performed as in f but with increasing amounts of synthetic Holliday junction DNA, 458 nM Yen1 and in the absence of Siz2. All experiments were independently replicated at least three times and images are representative of the reproducible results obtained

Fig. 2

Yen1 interacts with Slx5–Slx8 in the nucleus. a Diploid strains carrying one allele of galactose inducible VC-Slx5 and VN-Yen1 with wild-type copies of YEN1 and SLX5 in the homologous chromosomes were observed by live microscopy. BiFC (white arrows) signal denotes an interaction between the two BiFC (Venus) epitopes. Control diploids lacking one or both of the epitope-tagged proteins (ϕ) were used to substract background signal. A plasmid carrying Nup49-mCherry was transformed on the diploid strain harboring VC-Slx5/VN-Yen1 to visualize the nuclear perimeter. BiFC interactions were only detected in the nuclear compartment. b Cells carrying either an empty vector or a pYES2 plasmid expressing GST-Slx5 under galactose control were grown in selective media and induced with galactose for 3 h. Lysates were then applied to a glutathione-sepharose column. After washing, the bound proteins were eluted and immunoblotted with α-HA (upper) or α-GST (lower). c Two-hybrid assays were performed with strains carrying the indicated activating domain (AD) or DNA binding domain (BD) fusions. Strains were grown in selective media lacking leucine (L) and tryptophan (W) prior to spotting on media lacking histidine (H) to detect a positive interaction. Experiments were independently replicated three times and images are representative of the reproducible results obtained

Fig. 3

Yen1 is a direct substrate of the Slx5–Slx8 ubiquitin ligase. a H6-HA-Yen1 (916 nM) was ubiquitinated in vitro in the presence of the indicated concentrations of either Slx5/Slx8 or the RING mutant Slx5–6/Slx8 and 0.2 µM DNA. Control lanes with H6-HA-Yen1 in the absence of E3 and Slx5/8 in the absence of Yen1 are shown. Breakdown products of Yen1 are marked with an asterisk. b The ubiquitination reaction was performed as above but with 50 nM Slx5–Slx8 and increasing amounts of DNA. c Strains expressing 6xHis-Ub were subjected to different growth conditions and lysed to pull-down ubiquitinated proteins under denaturing conditions, input Yen1-HA levels were controlled to allow comparisons. d Smt3 denaturing pull-downs were performed in wild type, slx5∆, or slx8∆ cells (all in a pdr5∆ background) after growth in the presence of MMS. The fold increase in the sumoylated fraction indicated at the bottom of the gel is an average of three trials. Inputs were controlled in each trial to allow comparison of the eluted sumoylated proteins

Fig. 4

Deletion of SLX8 alters the nuclear distribution and turnover of a fraction of Yen1. a Wild type and slx8∆ cells were synchronized in G1 and released to observe the phosphorylation of Yen1 as a function of cell-cycle progression. b Serial dilutions of the indicated strains were spotted onto YPAD media containing different genotoxins. c Cells with an endogenous HTA1-mCherry carrying plasmids expressing wild-type GFP-Yen1 were observed microscopically after a short induction of the fusion protein. Shown are cells presenting normal nuclear localization (lower) or presenting foci (white arrows, upper). d G1 and G2/M cells of the indicated genetic backgrounds were microscopically examined as in c and classified according to the number of foci they displayed. The graphs show the percentage of cells in each category. The total number of cells individually scored from three video recordings are indicated as (n). Categories were subjected to the Fischer’s exact test, asterisks denote significant levels at P < 0.001(***) or P < 0.005(**). e The duration of foci in the indicated genetic backgrounds was measured by video-microscopy analysis. The mean ± s.d. of the duration time and n are indicated, error bars denote s.d. Asterisks refer to significance at the P < 0.05 (**) and P < 0.001 (***) levels in unpaired two-tailed Student’s t-test. f Cells expressing GFP-Yen1 and Hta1-mCherry were observed by video-microscopy in 2′ time-lapse frames. GFP total intensity of the whole cell, the nucleus and the cytoplasm was determined for 5 z-planes and used to calculate the total GFP intensity in each compartment. The graph displays the time course of GFP intensity in a single cell g The indicated pdr5∆ strains, that are permeable to MG132, were synchronized in G1 and subjected to cycloheximide (CHX) treatment during their release from G1 arrest. Where indicated, cells were pre-treated with MG132 for 30 min before, and during release in the presence of CHX. PGK1 was used to normalize the amount of Yen1. h Quantitation of the fraction of Yen1, compared to G1, remaining at the indicated times after release into CHX. The mean ± s.d. of triplicate assays is shown; statistically significant difference in unpaired two-tailed Student’s t-test is indicated (**P < 0.05). i FACS analysis of cells at the beginning and at the end of the CHX treatment

Fig. 5

Yen1 foci are dynamic and localize preferentially to nucleolar sites in the absence of DNA damage. a slx8∆ cells carrying a SIK1-mCherry endogenous marker and an inducible GFP-Yen1 expressing plasmid were observed after short induction of the fusion protein. The white arrow denotes co-localizing signal of GFP-Yen1 with Sik1-mCherry. b Wild-type cells carrying a TetO-TetR array tag on chromosome XII and an inducible GFP-Yen1 expressing plasmid were observed after short induction of the fusion protein. c Cells were subjected to acute challenge with Zeocin (0.01 mg/ml) and observed during their recovery as in Fig. 4. Cells displaying the designated categories of GFP-Yen1 foci were scored at the indicated time points. The total number of cells analysed (n) from two independent recordings were as follows: WT 0 h (n_G1 = 81, n_G2/M = 267), WT 1.5 h (n_G1 = 78, n_G2/M = 124), WT 3.5 h (n_G1 = 110, n_G2/M = 118), slx8∆ 0 h (n_G1 = 107, n_G2/M = 79), slx8∆ 1.5 h (n_G1 = 40, n_G2/M = 73), slx8∆ 3.5 h (n_G1 = 52, n_G2/M = 62). d Cells were subjected to an acute challenge with 0.1% MMS and for foci were observed as in c. The total number of cells analysed (n) from two independent recordings were as follows: WT 0 h (n_G1 = 81, n_G2/M = 267), WT 1.5 h (n_G1 = 95, n_G2/M = 98), WT 3.5 h (n_G1 = 139, n_G2/M = 137), slx8∆ 0 h (n_G1 = 107, n_G2/M = 79), slx8∆ 1.5 h (n_G1 = 53, n_G2/M = 55), slx8∆ 3.5 h (n_G1 = 40, n_G2/M = 67). e slx8∆ cells carrying SIK1-mCherry were observed after Zeocin challenge to determine GFP-Yen1 co-localization. White arrows indicate GFP-Yen1 foci. f slx8∆ cells were observed as in e, but were subjected to MMS treatment. White arrows indicate GFP-Yen1 foci. Images are representative of the reproducible results obtained after three independent trials. Statistical significance at P < 0.0001 in Fischer’s exact test at 3.5 h recovery points is indicated by asterisks in c and d

Fig. 6

K714 is ubiquitinated by Slx5–Slx8. a Recombinant 6xHIS-HA-Yen1 and the mutant 6xHIS-HA-Yen1-K714R (916 nM) were subjected to in vitro ubiquitination as in Fig. 3a. b 6xHis-Ubiquitin pull-downs were performed on cells expressing 6His-Ub and carrying endogenous Yen1-HA or its variant Yen1-K714R-HA following treatment with the indicated genotoxics. c Strains carrying endogenous Yen1-HA or its K714R variant were synchronized with alpha factor in G1 and proteins extracted at indicated time points after G1 release and immunoblotted with anti-HA. At time points where Yen1 is modified by CDK1, extracts were subjected to phosphatase treatment (CIP+) and also subjected to phos-tag gel separation. d Strains carrying HTA1-mCherry and the indicated GFP-Yen1 expression plasmids were examined microscopically as in Fig. 4c to assess the presence of the proteins. Foci were quantified as a function of cell-cycle phase, which was determined by cell morphology. The total number of analysed cells (n) and independent video recordings (VR) were as follows: WT (n_G1 = 105, n_G2/M = 153, VR = 3), yen1-K714R (n_G1 = 106, n_G2/M = 64, VR = 3), mus81∆ (n_G1 = 294, n_G2/M = 337, VR = 3), mus81∆ yen1-K714R (n_G1 = 203, n_G2/M = 306, VR = 3), slx8∆ (n_G1 = 166, n_G2/M = 172, VR = 3). Statistical differences were estimated by the Fischer’s exact test and significance is indicated by asterisks P < 0.05(*), P < 0.005(**), P < 0.001(***). e Sensitivity to the indicated genotoxics was determined by spotting serial dilutions of different strains on the indicated media. f Cycloheximide chase experiment showing persistence of Yen1 after a G1 release in the presence of CHX. g Immunoprecipitated Yen1-K714R-HA was eluted and the protein was sumoylated with Aos1-Uba2, Ubc9, and Smt3-3KR in the presence or absence of ATP. Samples were de-phosphorylated with CIP before loading. h 6xHIS-Smt3 pull-downs were performed on cells expressing 6HIS-Smt3 and carrying YEN1-HA or its variant yen1-K714R-HA under conditions of MMS treatment as indicated. The average fold enrichment is indicated at the bottom of the blot

Fig. 7

Yen1-K714R increases COs and suppresses spontaneous chromosome segregation defects in mus81∆ cells. a Diagram explaining the Chr. XV DSB-induced crossover reporters. Recombination outcomes were scored in white/red sectored colonies of the indicated strains and normalized to its plating efficiency (PE). The number of independent experiment trials (T) and the total number of recombination events scored (n) were as follows: WT (T = 6, n = 158), yen1∆ (T = 9, n = 538), mus81∆ (T = 5, n = 168), mus81∆ yen1∆ (T = 9, n = 160), yen1-K714R (T = 5, n = 230), and mus81∆ yen1-K714R (T = 5, n = 230). b A strain harboring a lacO/GFP-LacI array tag on chromosome IV was followed by video-microscopy to discriminate chromosome segregation in timely manner from aberrant segregation. Images display a typical normal and aberrant segregation and its respective kymograph. c GFP foci of the indicated strains were observed by video-microscopy and chromosomal segregation was scored as to whether it displayed a proper phenotype (normal) or one of three types of defective phenotypes (non-disjunction, delay, and aberrant chromosome number (acn)). The total number of cells analysed (n) and independent video-recordings (VR) were as follows: WT (n = 606, VR = 3), yen1∆ (n = 133, VR = 3), yen1-K714R (n = 131, VR = 3), mus81∆ (n = 259, VR = 5), yen1-K714R mus81∆ (n = 448, VR = 3) and yen1∆ mus81∆ (n = 289, VR = 3). d Segregation events scored as in c were determined for mus81∆ and mus81∆ yen1∆ strains containing a pYES2 plasmid expressing Yen1-HA under Galactose-inducible control or an empty pYES2 and subjected to acute over-expression of Yen1-HA or the equivalent mock induction prior to the recording of the video-microscopy. The total number of cells analysed from three VRs were as follows: mus81∆ (+pYES2) n = 245, mus81∆ (+pYES2-Yen1) n = 391, mus81∆ yen1∆ (+pYES2) n = 117, mus81∆ yen1∆ (+pYES2-Yen1) n = 218. Statistically significant differences in a between CO and other outcomes and in c and d between normal and abnormal categories were determined by the Fischer’s exact test, asterisks refer to significance at the P < 0.001(***), P < 0.005(**) or P < 0.05(*)