Structural and biochemical characterisation of the nucleotide excision repair helicase XPD as a potential drug target

DNA damage is an ever-present threat to genomic integrity. With constant incidents of damage, living organisms have evolved intricate networks to recognise, and repair DNA lesions. Nucleotide excision repair (NER) is such a system. NER is responsible for dealing with a broad variety of damages such as bulky lesions and damages originating from sources like ultraviolet light (UV). Interruptions in the NER cascade can have serious consequences for genomic maintenance and cell survival. The general transcription factor II H (TFIIH) is a multi-protein complex that is deeply involved in NER and RNA polymerase II-based transcription. Holo-TFIIH contains 10 subunits which are subdivided into core-TFIIH and the cyclin dependent kinase activating complex (CAK). The former is consisting of XPD, XPB, p62, p52, p44, p34, and p8, and the latter CDK7, MAT1 and CyclinH. Congenital mutations in TFIIH genes can lead to genetic disorders like xeroderma pigmentosum (XP), a Cockayne syndrome (CS) /XP combined phenotype and trichothiodystrophy (TTD). All of these diseases are associated with higher UV sensitivity as phenotypic hallmark. In XP patients UV induced damage is one of the main causes of early onset skin cancer. Also, melanoma the deadliest form of skin cancer has highly elevated occurrence in these patients. XPD is a superfamily 2 (SF2) DNA helicase that is involved in both transcription and NER as part of the TFIIH. Unlike other TFIIH components XPDs enzymatic activity is exclusive to NER. In NER it serves as a DNA helicase that opens a DNA bubble, scans and verifies the lesion site and in transcription initiation only serves as a scaffold for the CAK complex aiding the CAK activity to phosphorylate RNA polymerase II’s C-terminal domain. In recent years there have been reports of bladder cancer patients with somatic XPD mutations being sensitized to platinum-based chemotherapeutics and XP patients seem to be more susceptible to immune checkpoint therapy. Thus, XPD as an NER exclusive target presents itself for new options in cancer therapy. A previous high throughput screening (HTS) campaign in our group using FMP libraries CBB₁-6 and the LOPAC library revealed over 170 small molecules that could inhibit XPD helicase activity by more than 50%. A promising candidate named N3 was identified, and several structurally similar and commercially available chemical derivatives were selected for further screening. In this study, we used biochemical and biophysical techniques to evaluate the inhibitory capacity of different N3 derivatives. Using helicase assays, we established inhibition parameters for multiple N3 derivatives and identified several with improved helicase inhibition profiles. DNA binding and ATPase studies were conducted to assess the effects of the inhibitors on XPD’s affinity for DNA and ATPase activity. We also determined which small molecules bind directly to XPD using biolayer interferometry (BLI). By comparing helicase activity results with BLI data, we confirmed a relationship between inhibition and direct interaction with XPD indicating on target effects of the most potent compounds in vitro. Specificity of the N3 derivatives was additionally supported by the fact that small functional group changes to N3 led to significant differences in inhibitory efficacy. We also aimed to resolve the binding positions of our top candidate using cryo-electron microscopy (Cryo-EM) to elucidate an inhibitor bound structure of XPD. These would allow us to do structure-based drug design, improving the parameters of inhibition. Although our cryo- EM studies did not resolve inhibitor binding, they provided structures of multiple states of XPD ATP processivity, offering insights into the mechanism by which XPD unwinds double-stranded DNA. Using a novel DNA interstrand cross-link substrate we were able to solve a structure of XPD stalled upon approaching a DNA interstrand cross-link a non-canonical NER lesion representing the first structure of XPD stalled on a lesion for verification. In addition, this non canonical lesion gave us vital insights into how XPD could approach a canonical NER damage. Eventually, we succeeded to test our inhibitors in newly established cell-based assays systems. Using CRISPR/Cas9 we generated melanoma cell models mimicking XPD helicase insufficiency (R683W and K48R) to allow us to investigate and validate XPD inhibition as a promising target in cancer.