RNF138 interacts with RAD51D and is required for DNA interstrand crosslink repair and maintaining chromosome integrity

Abstract

The RAD51 family is integral for homologous recombination (HR) mediated DNA repair and maintaining chromosome integrity. RAD51D, the fourth member of the family, is a known ovarian cancer susceptibility gene and required for the repair of interstrand crosslink DNA damage and preserving chromosomal stability. In this report, we describe the RNF138 E3 ubiquitin ligase that interacts with and ubiquitinates the RAD51D HR protein. RNF138 is a member of an E3 ligase family that contains an amino-terminal RING finger domain and a putative carboxyl-terminal ubiquitin interaction motif. In mammalian cells, depletion of RNF138 increased the stability of the RAD51D protein, suggesting that RNF138 governs ubiquitin-proteasome-mediated degradation of RAD51D. However, RNF138 depletion conferred sensitivity to DNA damaging agents, reduced RAD51 focus formation, and increased chromosomal instability. Site-specific mutagenesis of the RNF138 RING finger domain demonstrated that it was necessary for RAD51D ubiquitination. Presence of RNF138 also enhanced the interaction between RAD51D and a known interacting RAD51 family member XRCC2 in a yeast three-hybrid assay. Therefore, RNF138 is a newly identified regulatory component of the HR mediated DNA repair pathway that has implications toward understanding how ubiquitination modifies the functions of the RAD51 paralog protein complex.

Keywords: Chromosome integrity; DNA double-stranded breaks; DNA interstrand crosslinks; DNA repair; E3 ligase; Homologous recombination; RAD51D; RNF138; Ubiquitination.

Fig. 1

RNF138 interacts with RAD51D. (A) RNF138 clones isolated from yeast two-hybrid screening with RAD51D. (B) Yeast two-hybrid analysis of interactions between RNF138 and RAD51 paralogs. AH109 haploids were co-transformed with the indicated GAL4 DBD and AD fusion constructs and serially diluted on to non-selective and selective growth medium. (C) Protein interaction was quantified by measuring β-galactosidase activity. Data represent mean +/− SEM from three independent experiments conducted in triplicate. (D) Lysates from cells transfected with the indicated expression constructs were subject to anti-HA immunoprecipitation followed by immunoblot analysis with anti-Myc (RNF138, top panel) and anti-HA (RAD51D, bottom panel) antibodies. (E) A reciprocal immunoprecipitation experiment was performed for HA-RNF138 and Myc-RAD51D. (F) Cells were transfected with vectors for either wild-type (WT) or RING finger domain mutant and Myc-RNF138 for anti-HA immunoprecipitation. Abbreviations: DBD Fusion- GAL4 DNA binding domain fusion, AD Fusion- GAL4 activation domain fusion, IP- immunoprecipitate, WC- whole cell lysate, FT- immunoprecipitation flow through, W- immunoprecipitation wash, E- first eluate from immunoprecipitation.

Fig. 2

RNF138 modifies interaction with RAD51C and XRCC2. (A) ONPG analysis of Y187 haploids transformed with the indicated RNF138 and RAD51D expression constructs. (B) Yeast three-hybrid analysis of Y190 haploids containing the indicated two-hybrid plasmids with either pVT100u–RNF138 or empty vector. (C) Yeast two-hybrid analysis of AH109 haploids transformed with the RNF138 splice variants and full-length RAD51D as indicated. Data represent mean +/− SEM from three independent experiments performed in triplicate, and * indicates P < 0.05, ** indicates P < 0.01.

Fig. 3

RT-PCR analysis of Mus musculus Rnf138 alternative transcripts. (A) RT-PCR for Rnf138 was performed from eight tissues. A schematic of the Rnf138 gene is shown (top) with numbered boxes representing exons. (B and C) Quantitative Real-time PCR was performed for Rnf138 from the same eight tissues. Black arrows in Rnf138 exons indicate the locations of forward and reverse PCR primers: Rnf138 exons 2 and 3 (B) or exons 6 and 7 (C). Note that Rnf138 alternative transcripts lacking exon 7 predominant in testis are not amplified in (C). Rnf138 expression was normalized to Gapdh mRNA levels. Error bars indicate standard deviation from two independent experiments performed in duplicate.

Fig. 4

RNF138 deficiency confers increased sensitivity to Mitomycin C. (A) Rnf138 expression in Rad51d^+/+ (Rad51d^+/+, Trp53^−/−) and Rad51d^−/− (Rad51d^−/−, Trp53^−/−) MEFs was analyzed by qPCR 24 h following siRNA transfection and normalized to both β-actin and Gapdh. Data represent mean +/− STD from a representative experiment performed in triplicate, and *** indicates P < 0.001. (B) Cell viability relative to scrambled siRNA control was measured by colony forming assay (CFA). Error bars represent mean +/− SEM from three independent experiments performed in duplicate. (C and D) MEFs were treated with the indicated amounts of Mitomycin C and cell viability assessed by (C) MTT or (D) CFA. Error bars represent +/− SEM from four independent experiments performed in triplicate for MTT assays and three independent experiments performed in duplicates for CFA using siRNA1. Note that Trp53 encodes the p53 protein and that RNF138 KD represents the RNF138 siRNA knockdown.

Fig. 5

RNF138 is necessary for increased RAD51 foci formation in response to DNA damage. (A) Immunofluorescence visualization of RAD51 foci in Rad51d^+/+ (Rad51d^+/+, Trp53^−/−) MEFs. (B) Quantification of RAD51 foci formation. Data represent mean +/− SEM from three independent experiments using RNF138 siRNA2 and * indicates P < 0.05. (C) Western analysis of RAD51 protein levels. β-actin was used as a loading control.

Fig. 6

RNF138 maintains chromosome integrity. (A) Metaphase spreads were prepared from Rad51d^+/+ (Rad51d^+/+, Trp53^−/−) and Rad51d^−/− (Rad51d^−/−, Trp53^−/−) MEFs following mock transfection or transfection with 30 nM Rnf138 siRNA1. Cells were either untreated (left panel) or treated with 50 ng/mL MMC (right panel). Spreads were scored for chromatid breaks/gaps, chromosome breaks/gaps, and chromosome radials, represented by arrows. The frequency of each aberration per chromosome is displayed above the bar (10⁻³). Note the change in scale between untreated and treated MEFs. (B) Giemsa stained cells were scored for the presence of anaphase bridges. Error bars represent mean +/− STD from two independent experiments and * indicates P < 0.05.

Fig. 7

RNF138 mediates degradation of RAD51D through the ubiquitin-proteasome pathway. (A) Analysis of RNF138 facilitated RAD51D ubiquitination in vivo. (B) Rad51d^−/− Trp53^−/− HARad51d MEFs were transfected with Myc-RNF138 (upper panels) or 30nM Rnf138 siRNA1 (lower panels). HA-RAD51D protein levels were assessed 2, 4, and 6 hours following initiation of CHX block. HA-RAD51D band intensity was normalized to β-actin and plotted as percent protein remaining for each time point. Data represent mean +/− STD from two representative experiments. (C) MEFs were treated with CHX (100µg/mL) or CHX in combination with MG132 (20µg/mL). Cell lysates were prepared at the 4 h time point, and HA-RAD51D band intensity was normalized to β-actin.