A novel DDB2 mutation causes defective recognition of UV-induced DNA damages and prevalent equine squamous cell carcinoma

Abstract

Squamous cell carcinoma (SCC) occurs frequently in the human Xeroderma Pigmentosum (XP) syndrome and is characterized by deficient UV-damage repair. SCC is the most common equine ocular cancer and the only associated genetic risk factor is a UV-damage repair protein. Specifically, a missense mutation in horse DDB2 (T338M) was strongly associated with both limbal SCC and third eyelid SCC in three breeds of horses (Halflinger, Belgian, and Rocky Mountain Horses) and was hypothesized to impair binding to UV-damaged DNA. Here, we investigate DDB2-T338M mutant’s capacity to recognize UV lesions in vitro and in vivo, together with human XP mutants DDB2-R273H and -K244E. We show that the recombinant DDB2-T338M assembles with DDB1, but fails to show any detectable binding to DNA substrates with or without UV lesions, due to a potential structural disruption of the rigid DNA recognition β-loop. Consistently, we demonstrate that the cellular DDB2-T338M is defective in its recruitment to focally radiated DNA damages, and in its access to chromatin. Thus, we provide direct functional evidence indicating the DDB2-T338M recapitulates molecular defects of human XP mutants, and is the causal loss-of-function allele that gives rise to equine ocular SCCs. Our findings shed new light on the mechanism of DNA recognition by UV-DDB and on the initiation of ocular malignancy.

Introduction

Ultraviolet (UV) irradiation induces covalent crosslinks between neighboring pyrimidines. If left unrepaired, error-prone replication of this damaged DNA leads to increased rate of mutagenesis and genome instability [1,2]. Specialized damage-surveillance factor UV-DDB recognizes diverse photo-lesions, crucial for the dynamic process of nucleotide excision repair (NER) [3]. Loss of function mutations in the gene encoding the DDB2 subunit of UV-DDB are associated with a rare autosomal recessive syndrome Xeroderma Pigmentosum (XP), which is characterized by sun sensitivity and severe risk for skin cancers [4].

Cellular UV-DDB activity requires two protein subunits, DDB1 and DDB2. DDB2′s WD40 domain provides the sole recognition interface for diverse photo-damage [5], including those buried in nucleosomes [6], whereas DDB1 is needed to facilitate the folding and stability of DDB2 [7]. DDB1 also serves as a substrate adaptor for Cul4 E3 ubiquitination ligase complexes [8,9]. DDB2′s role in UV-DDB seem to reflect a more recent evolutionary adaption to repair UV-lesions: while DDB1 is conserved from fission yeast to human, DDB2′s homologs have only been previously reported in vertebrates [4]. Inactivating DDB2 mutants R273H and K244E have been found in XP complementation group E (XP-E) human patients, whereas no DDB1 mutation has been reported. In addition, UV-DDB is limited by the transcription level of DDB2 mRNA, which is activated by the transcription factor P53 in human cells, but not in rodents, where very low level of UV-DDB can be detected [4].

Squamous cell carcinoma (SCC) is frequently detected among XP patients [10], and is also the most common ocular cancer in horses [11]. We previously studied the Haflinger horse breed with high SCC occurrence of the limbus, and found an autosomal recessive mode of inheritance [12]. A genome-wide association study (GWAS) followed by candidate gene sequencing identified a major genetic risk locus, where a missense mutation T338M of DDB2 gene was found as a candidate SCC allele [13]. Later, the same variant was found to be a risk factor for SCC initiated from the third eye lid, also known as the nictitating membrane [14], and from additional horse breeds [15,16]. This work led to the speculation that this was a loss-of-function variant and a causal risk factor for ocular SCC as computational modeling predicted that this mutation altered conformation of the β loop involved in photolesion recognition [13].

Here, we provide direct functional evidence indicating that the T338M mutant indeed is a loss-of-function allele that abolishes UV-DDB recognition of damaged DNA. We present a structural rationale for T338 to anchor the critical β-hairpin loop of DDB2 for proper DNA-engagement. Our biochemical and cellular characterizations of the T338M and human XP-E mutants prompted us to propose a 2-step damage-scanning model that critically depends on the β-hairpin loop anchored by the T338 residue.