SUMO-targeted ubiquitin ligase activity can either suppress or promote genome instability, depending on the nature of the DNA lesion

Abstract

The posttranslational modifiers SUMO and ubiquitin critically regulate the DNA damage response (DDR). Important crosstalk between these modifiers at DNA lesions is mediated by the SUMO-targeted ubiquitin ligase (STUbL), which ubiquitinates SUMO chains to generate SUMO-ubiquitin hybrids. These SUMO-ubiquitin hybrids attract DDR proteins able to bind both modifiers, and/or are degraded at the proteasome. Despite these insights, specific roles for SUMO chains and STUbL in the DDR remain poorly defined. Notably, fission yeast defective in SUMO chain formation exhibit near wild-type resistance to genotoxins and moreover, have a greatly reduced dependency on STUbL activity for DNA repair. Based on these and other data, we propose that a critical role of STUbL is to antagonize DDR-inhibitory SUMO chain formation at DNA lesions. In this regard, we identify a SUMO-binding Swi2/Snf2 translocase called Rrp2 (ScUls1) as a mediator of the DDR defects in STUbL mutant cells. Therefore, in support of our proposal, SUMO chains attract activities that can antagonize STUbL and other DNA repair factors. Finally, we find that Taz1TRF1/TRF2-deficiency triggers extensive telomeric poly-SUMOylation. In this setting STUbL, together with its cofactor Cdc48p97, actually promotes genomic instability caused by the aberrant processing of taz1Δ telomeres by DNA repair factors. In summary, depending on the nature of the initiating DNA lesion, STUbL activity can either be beneficial or harmful.

Fig 1. Forced nucleation of SUMO on chromatin engages STUbL activity.

A, Wild-type or slx8-29 cells with 256-repeats of the LacO sequence integrated near cen1 that express LacI-GFP from the P_nmt1 promoter and LacI-mCherry-SUMO from the attenuated P_nmt81 promoter. Cells were grown overnight at 30°C in filtered EMM-LUAH with 5x Thiamine (B1) to allow leaky protein expression from the nmt promoters, and imaged live. B, Percentage of cells containing co-localized mCherry and GFP foci over total cells containing GFP foci was plotted against each genotype. Over 200 cells were counted for each genotype.

Fig 2. STUbL mutant phenotypes are mitigated by mutations that counteract SUMO chain accumulation.

A to D, five fold serial dilutions of the indicated strains were spotted onto YES plates, with or without the indicated genotoxic challenge, and grown at indicated temperature. Cells in D were grown at 25°C. E, anti-SUMO and tubulin Western blots of total proteins from the indicated strains grown at 25°C to log phase.

Fig 3. STUbL mutant phenotypes are mitigated by Rrp2 deletion.

A, five fold serial dilutions of the indicated strains were spotted onto YES plates, with or without hydroxyurea (HU) or camptothecin (CPT), and grown at the indicated temperatures. B & C, anti-SUMO and tubulin Western blots of total proteins of indicated strains grown in YES (B) or minimal selection (UAH) medium (C). To induce Rrp2 overexpression (Rrp2 O.E.), cells containing pREP41-mCherry-FLAG-Rrp2 were grown in the absence of Thiamine (B1) for 24 h. D, five fold serial dilutions of the indicated strains were spotted onto YES plates, with or without camptothecin (CPT), and grown at 25°C.

Fig 4. Rescue of taz1Δ cold sensitivity by mutations that reduce SUMO chain formation.

A to D, five fold serial dilutions of the indicated strains were spotted onto YES plates, the cells were incubated at 25 or 19°C for three to six days. B, dilution series of the indicated strains were spotted onto minimum medium (EMM-UAH) without thiamine (B1) to induce the expression of Pli1 and incubated at 19°C or 30°C. Cells in D were grown at 30°C.

Fig 5. Rescue of taz1Δ cold sensitivity and checkpoint activation by STUbL and Cdc48/p97 mutants.

A, five fold serial dilutions of the indicated strains were spotted onto YES plates, and incubated at 25 or 19°C for three to six days. B, dilution series of the indicated strains were spotted onto minimum medium (EMM-UAH) with or without thiamine (B1) and incubated at 19°C for six days. C & D, differential interference contrast (DIC) light microscopy of the indicated strains grown for two days in liquid culture at 19°C. Bar, 5 μM.

Fig 6. The effect of SUMO pathway mutations on telomere length.

A, genomic DNA samples of the indicated genotypes grown for two days at 19°C were analyzed for telomere length by Southern blot hybridization with a telomere DNA probe. B, DNA samples of the indicated genotypes grown at 25°C were analyzed for telomere length by Southern blot hybridization with a telomere DNA probe.

Fig 7. SUMO chain accumulation at taz1Δ telomeres.

A, Our genetic analysis supports a model wherein Pli1-dependent SUMO chains persistently hyper-accumulate at telomeres in taz1Δ cells. Reducing SUMO chain formation by mutating Pli1, Nup132, or SUMO itself curbs SUMO chain accumulation at telomeres. In turn, this prevents the engagement of STUbL and Cdc48(p97) activities that drive telomere entanglement. B, C, D. ChIP analysis of SUMO at telomere of the indicated genotypes. Telomeric association of SUMO was measured by quantitative real-time PCR, and normalized against background (see Materials and Methods for details on calculation). Plots show mean values ± standard error for two or more independent experiments. Cells in C & D were grown at 19°C.

Fig 8. Model integrating known STUbL functions with data herein.

A, STUbL and Cdc48(p97) activities are known to act at DNA lesions to make them permissive for downstream DDR events [48]. However, the nature of the DDR-inhibitory signal(s) upon STUbL and Cdc48(p97) inactivation remains unclear. We propose that SUMO chains are DDR-inhibitory, and their formation by the locally hyperactive SUMO conjugation pathway must be actively antagonized by STUbL and Cdc48(p97). STUbL could antagonize SUMO chains by (i) ubiquitinating SUMO's amino-terminal lysine residues and "capping" SUMO chain growth [35] (ii) driving the local degradation/extraction (via Cdc48/p97) of the major SUMO ligases Pli1 and SIZ1 [29, 72] or (iii) degrading/extracting other SUMOylated proteins that are "licensed" for SUMOylation e.g. MDC1 at DNA lesions, and therefore constantly nucleate the SUMOylation machinery [48]. B, in the absence of STUbL, SUMO chains accumulate and inhibit DNA repair including homologous recombination (HR) and non-homologous end joining (NHEJ) [48]. SUMO chains may exert intrinsic/steric DDR-inhibitory effects, but could also attract proteins such as the Swi2/Snf2 DNA translocase Rrp2 (and orthologs) that has known DNA repair modulatory functions. This model is consistent with the broadly compromised DDR in STUbL mutant cells, and suppression of this phenotype by reducing SUMO chain formation.