DNA repair pathways present in all cells serve to preserve genome stability, but in cancer cells they also act reduce the efficacy of chemotherapy. The endonuclease ERCC1-XPF has an important role in the repair of DNA damage caused by a variety of chemotherapeutic agents and there has been intense interest in the use of ERCC1 as a predictive marker of therapeutic response in non-small cell lung carcinoma, squamous cell carcinoma and ovarian cancer. We have previously validated ERCC1 as a therapeutic target in melanoma, but all small molecule ERCC1-XPF inhibitors reported to date have lacked sufficient potency and specificity for clinical use. In an alternative approach to prevent the repair activity of ERCC1-XPF, we investigated the mechanism of ERCC1 ubiquitination and found that the key region was the C-terminal (HhH) 2 domain which heterodimerizes with XPF. This ERCC1 region was modified by non-conventional lysine-independent, but proteasome-dependent polyubiquitination, involving Lys33 of ubiquitin and a linear ubiquitin chain. XPF was not polyubiquitinated and its expression was dependent on presence of ERCC1, but not vice versa. To our surprise we found that ERCC1 can also homodimerize through its C-terminal (HhH) 2 domain. We exploited the ability of a peptide containing this C-terminal domain to destabilise both endogenous ERCC1 and XPF in human melanoma cells and fibroblasts, resulting in reductions of up to 85% in nucleotide excision repair and near two-fold increased sensitivity to DNA damaging agents. We suggest that the ERCC1 (HhH) 2 domain could be used in an alternative strategy to treat cancer.