C. elegans ring finger protein RNF-113 is involved in interstrand DNA crosslink repair and interacts with a RAD51C homolog

Abstract

The Fanconi anemia (FA) pathway recognizes interstrand DNA crosslinks (ICLs) and contributes to their conversion into double-strand DNA breaks, which can be repaired by homologous recombination. Seven orthologs of the 15 proteins associated with Fanconi anemia are functionally conserved in the model organism C. elegans. Here we report that RNF-113, a ubiquitin ligase, is required for RAD-51 focus formation after inducing ICLs in C. elegans. However, the formation of foci of RPA-1 or FCD-2/FANCD2 in the FA pathway was not affected by depletion of RNF-113. Nevertheless, the RPA-1 foci formed did not disappear with time in the depleted worms, implying serious defects in ICL repair. As a result, RNF-113 depletion increased embryonic lethality after ICL treatment in wild-type worms, but it did not increase the ICL-induced lethality of rfs-1/rad51C mutants. In addition, the persistence of RPA-1 foci was suppressed in doubly-deficient rnf-113;rfs-1 worms, suggesting that there is an epistatic interaction between the two genes. These results lead us to suggest that RNF-113 and RFS-1 interact to promote the displacement of RPA-1 by RAD-51 on single-stranded DNA derived from ICLs.

Figure 1. Effects of RNF-113 depletion on C. elegans survival after interstrand DNA crosslinking (ICL) and the intracellular localization of RNF-113.

(A) Comparison of embryonic hatching rates after knocking down RNF-113 expression with and without ICL treatment. rnf-113 RNAi was performed from the P0 young adult stage by feeding wild-type C. elegans worms E. coli cells expressing double-strand RNA for rnf-113. F1 worms at the L4 stage were treated with TMP (25 µg/ml) plus UVA, and eggs were collected over 24 h and their hatching scored 20 h later. This set of experiments in the wild-type background, was also performed in fcd-2(tm298) and rfs-1(ok1372) mutants. The error bars are SEM (standard errors of the mean). (B) Intracellular localization of RNF-113 in the germ cells of the mitotically proliferating region of C. elegans gonads was detected using polyclonal antibodies against RNF-113. Immuno-staining was performed at 3, 9, 16, and 24 h after ICL treatment in the wild type. The scale bar is 10 µm.

Figure 2. Focus formation by FCD-2 after ICL treatment is not affected by RNF-113 depletion.

The mitotically proliferating regions of wild-type and rnf-113(RNAi) gonads were immuno-stained using FCD-2 polyclonal antibody at 18 h after ICL (TMP/UVA) treatment. The scale bar is 10 µm.

Figure 3. RNF-113 depletion attenuates RAD-51 focus formation at ICLs in wild-type worms, but not in rfs-1 mutants.

(A) The mitotically proliferating regions of gonads from wild-type, rnf-113(RNAi), rfs-1(ok1372), and rfs-1(ok1372);rnf-113(RNAi) worms are shown after staining for RAD-51 at 16 h following TMP/UVA treatment. (B) The immuno-staining in (A) was repeated at 3 h after IR (ionizing radiation, 75 Gy) treatment instead of ICL treatment. The scale bars are 10 µm. (C) A focal plane having a maximum number of RAD-51 foci was chosen for each nucleus of germ cells (n = 150 for wild-type and rnf-113(RNAi); n = 30 for rfs-1(ok1372) and rfs-1(ok1372);rnf-113(RNAi)) 16 h after ICL treatment as in (A), and the average numbers of RAD-51 foci per nuclear focal plane are plotted. The error bars are SEM.

Figure 4. RNF-113 depletion does not affect the formation of RPA-1 foci after ICL treatment, but greatly retards the dissipation of RPA-1 foci.

The mitotically proliferating regions of gonads from wild-type, rnf-113(RNAi), rfs-1(ok1372), and rfs-1(ok1372);rnf-113(RNAi) worms are shown after immuno-staining for RPA-1 at 4 h, 12 h, and 24 h following TMP/UVA treatment. The scale bar is 10 µm.

Figure 5. RNF-113 has E3 ubiquitin ligase activity in vitro.

Recombinant 6×HIS::RNF-113 was purified from E. coli lysates using Ni-NTA agarose and incubated with E1, HA-ubiquitin, and UbcH5c as E2 enzyme. The reaction products were electrophoresed in two separate 6–18% SDS-polyacrylamide gels and probed with (A) HA and (B) HIS antibodies, respectively.

Figure 6. Proposed model on the roles of RNF-113 and RFS-1 in loading RAD-51 to replication forks stalled at ICLs.

(A) In the wild-type background, a replication fork stalls spontaneously at an ICL and is cleaved. The resulting one-ended DSB is resected to produce ssDNA (single-strand DNA), to which RPA-1, RFS-1 and an unknown protein ‘X’ bind. RNF-113 is proposed to ubiquitinate X, and the ubiquitinated X together with RFS-1 promotes replacement of RPA-1 on ssDNA by RAD-51. (B) In the absence of RNF-113, X is not ubiquitinated so that RPA-1 cannot be displaced from ssDNA. The presence of RPA-1, together with RFS-1 and unmodified X on ssDNA prevents the loading of RAD-51, and the DNA intermediate is at a dead end. (C) In the absence of RFS-1, RAD-51 cannot be recruited to ssRNA, even though RNF-113 ubiquitinates X on ssDNA. Nevertheless, the DNA intermediate with RPA-1 and ubiquitinated X bound is shunted to an error-prone repair pathway involving nucleotide excision repair (NER) and translesion DNA synthesis (TLS). In the absence of both RNF-113 and RFS-1, the DNA intermediate with RPA-1 and unmodified X is also shunted to the error-prone repair pathway.