XPB and XPD helicases in TFIIH orchestrate DNA duplex opening and damage verification to coordinate repair with transcription and cell cycle via CAK kinase

Abstract

Helicases must unwind DNA at the right place and time to maintain genomic integrity or gene expression. Biologically critical XPB and XPD helicases are key members of the human TFIIH complex; they anchor CAK kinase (cyclinH, MAT1, CDK7) to TFIIH and open DNA for transcription and for repair of duplex distorting damage by nucleotide excision repair (NER). NER is initiated by arrested RNA polymerase or damage recognition by XPC-RAD23B with or without DDB1/DDB2. XP helicases, named for their role in the extreme sun-mediated skin cancer predisposition xeroderma pigmentosum (XP), are then recruited to asymmetrically unwind dsDNA flanking the damage. XPB and XPD genetic defects can also cause premature aging with profound neurological defects without increased cancers: Cockayne syndrome (CS) and trichothiodystrophy (TTD). XP helicase patient phenotypes cannot be predicted from the mutation position along the linear gene sequence and adjacent mutations can cause different diseases. Here we consider the structural biology of DNA damage recognition by XPC-RAD23B, DDB1/DDB2, RNAPII, and ATL, and of helix unwinding by the XPB and XPD helicases plus the bacterial repair helicases UvrB and UvrD in complex with DNA. We then propose unified models for TFIIH assembly and roles in NER. Collective crystal structures with NMR and electron microscopy results reveal functional motifs, domains, and architectural elements that contribute to biological activities: damaged DNA binding, translocation, unwinding, and ATP driven changes plus TFIIH assembly and signaling. Coupled with mapping of patient mutations, these combined structural analyses provide a framework for integrating and unifying the rich biochemical and cellular information that has accumulated over forty years of study. This integration resolves puzzles regarding XP helicase functions and suggests that XP helicase positions and activities within TFIIH detect and verify damage, select the damaged strand for incision, and coordinate repair with transcription and cell cycle through CAK signaling.

Figure 1. Structural basis for early NER steps in humans (right) and bacteria (left)

During human global genome repair (GGR), DNA damage is recognized by the XPC/RAD23B complex (Rad4/Rad23 pdb 2QSG). CPD (dotted line) was not visible in the crystal [60]. For some damage, DDB1/DDB2 (pdb 3EI1 bound to 6-4PP [63]) complex aids recognition. During transcription, the RNA polymerase (RNAPII pdb 2JA7 with CPD [68]) stalls at a lesion, recruiting CSB and CSA. Lesions that cannot be repaired by base excision repair (BER) can be shuttled to NER by the ATL protein (pdb 3GYH shown bound to cigarette smoke derived lesion O(6)-4-(3-pyridyl)-4-oxobutylguanine [47]). For clarity, cartoons next to the crystal structure molecular surfaces show which strand is bound. Regardless of recognition method, XPA, XPG, RPA, and TFIIH bind such that the XPD (pdb 3CRV [79]) and XPB (pdb 2FWR [78]) helicases can open dsDNA into a 27-nt bubble suitable for excision by XPF/ERCC1 and XPG nucleases. The gap is then filled by DNA polymerase delta and kappa, or epsilon bound to PCNA, which is loaded onto DNA by RFC and ligated. In bacteria, the UvrB helicase (pdb 2FDC with DNA [86]) together with a dimer of UvrA (pdb 2R6F [177]) recognize the lesion. UvrB opens the helix and the recruited nuclease UvrC (pdb 2NRT C-terminal domain [178]) cuts on both sides of the lesion. The UvrD (pdb 2IS6 [84]) helicase removes the excised product and the gap is filled by DNA pol I.

Figure 2. NER helicase crystal structures and schematic sequence regions

AfXPB (pdb 2FWR [78]), SaXPD (pdb 3CRV [79]), BcUvrB (pdb 2FDC [86]), and EcUvrD (pdb 2R6F [84]) folds shown as ribbons with either DNA models (cartoons for AfXPB and SaXPD) or DNA co-structures (ribbons with bases for BcUvrB and EcUvrD). AfXPB DNA was docked manually using DNA from Hel308 (pdb 2P6R [110]) and SaXPD DNA was computationally docked as described [79]. Linear schematics are shown below each structure. Rad51/RecA domains are colored in cyan (helicase domain 1) or green (helicase domain 2). Accessory domains that are insertions in HD1 are colored in purple or light purple except for XPD FeS domain which is orange and insertions in HD2 are colored blue. The AfXPB DRD domain N-terminal to HD1 is colored magenta. Grey extensions at the N- and C-termini of the XPB schematic or the C-terminus of the XPD schematic represent extensions present in the human proteins. The seven helicase motifs (I, Ia, II-VI) shared among these helicases are shown in red.

Figure 3. XP helicase domains and functional conformational flexibility

Top, proposed 170° domain rotation of XPB upon DNA binding and ATP hydrolysis [78]. This domain rotation is key to bring HD1 and HD2 together in an active conformation. Bottom, proposed XPD conformational flexibility. Loading of XPD onto DNA would require an opening of the non-covalent interactions between the Arch and FeS domain, while flexibility between HD1 and HD2 is necessary for translocation along DNA during unwinding.

Figure 4. Molecular surface pockets and important residues in ATP binding, hydrolysis, and human disease

A-B. UvrB and UvrD crystal structures showing the binding pocket (transparent surface) and important coordinating residues (sticks) of Mg²⁺ and either ATP or ADPNP. C. Crystal structure (ribbons) of SaXPD (pdb 3CRV [79]) showing location of XP (red spheres), XP/CS (gold spheres), and TTD (purple spheres) human patient mutations. The location of an S. cerevisiae mutant that converts UV damage into replication-dependent DSBs is shown [167] (blue sphere T507). Computational DNA model (black ribbon) and position of ATP binding “ATP” are shown. D. Model of SaXPD ATP binding pocket with ATP analog, AMP-PCP or adenosine-5’;-[beta, gamma-methylene]triphosphate, from the Hjm helicase (pdb 2ZJA [179]) showing position of Gly34 Walker A residue mutated in XP/CS patients (G47R) and Asp180 Walker B residue mutated in XP patients (D234N).

Figure 5. TFIIH architecture with subunit structures, proposed assembly, and functional implications

Eight of ten TFIIH subunits with structural information are shown (ribbon folds and transparent surfaces) along proteins and domains without known structures (shapes). The positions of SaXPD (pdb 3CRV [79]) and AfXPB (pdb 2FWR and 2FZL [78]) in the TFIIH EM envelope [145] were determined computationally [79]. Immunolabeling experiments positioned Cdk7 (pdb 1UA2 [138]) at the TFIIH ring protrusion [145]. Cyclin H (pdb 1KXU [140]) was positioned based on other Cdk/cyclin structures. The Mat1 N-terminal domain (pdb 1G25 [141]) is connected to the C-terminal domain (brown oval) that stimulates Cdk7 by a central coiled-coiled domain (brown coils). The Mat1 coiled-coiled domain interacts with both XPD and XPB [26], so is positioned to monitor the conformation of both helicases. P44 (blue triangle) interacts with the C-terminal extension of XPD (not shown) and stimulates its activity [180]. A crystal structure of the yeast homolog of p8 (Tfb5) was solved with the C-terminal domain of yeast p52 (Tfb2) (pdb 3DGP) [150]. P52 stimulates XPB ATPase activity and contacts XPB as P52 N- and C-terminal domains interact with XPB [118]. Although not required for TFIIH assembly, the N-terminal domain of p62 (pdb 1PFJ [154]) interacts with the NER nuclease, XPG [154]. P34 (purple circle) is a member of core TFIIH [181]. Open area (white space) within TFIIH ring between XPB and XPD is predicted to contain the human protein domain extensions absent in the archaeal enzymes.

Figure 6. Unified testable model for XPB and XPD functions within TFIIH

After initial damage recognition and DNA binding by XPC-RAD23B (or other proteins not shown), TFIIH is recruited (with XPA, XPG, and RPA) and XPB binds opposite the damage. DNA binding induces a conformational change in XPB. ATP hydrolysis stabilizes this interaction and may aid nuclease recruitment. Working against the grip of XPC-RAD23B binding, XPB pries open ~ 5nts. XPD binds 5’ to the lesion so the distance from XPB to XPD in the TFIIH ring defines the 27nt size of the excision bubble. Stimulated by XPA, CAK disengages from TFIIH, which then stimulates XPD to unwind in a 5’ to 3’ direction to the lesion. XPD binding anchors TFIIH at the damaged site, recruits XPG, and marks the damaged strand for incision. During transcription (Insert, top left), only XPB helicase is engaged allowing promoter opening and CAK phosphorylation of the RNAPII C-terminal domain to occur. XP/CS mutations that lock conformation and/or affect signaling within TFIIH not only disrupt repair but block downstream events leading to cell death and tissue degeneration. TTD mutations destabilize TFIIH, affecting all TFIIH activities thus leading to cell death (center, right). XP mutations in XPD and XPB that cause defects in helicase or ATPase activities disrupt repair and lead to cancer (bottom, right).