The TFIIH subunits p44/p62 act as a damage sensor during nucleotide excision repair

Abstract

Nucleotide excision repair (NER) in eukaryotes is orchestrated by the core form of the general transcription factor TFIIH, containing the helicases XPB, XPD and five 'structural' subunits, p62, p44, p34, p52 and p8. Recent cryo-EM structures show that p62 makes extensive contacts with p44 and in part occupies XPD's DNA binding site. While p44 is known to regulate the helicase activity of XPD during NER, p62 is thought to be purely structural. Here, using helicase and adenosine triphosphatase assays we show that a complex containing p44 and p62 enhances XPD's affinity for dsDNA 3-fold over p44 alone. Remarkably, the relative affinity is further increased to 60-fold by dsDNA damage. Direct binding studies show this preference derives from p44/p62's high affinity (20 nM) for damaged ssDNA. Single molecule imaging of p44/p62 complexes without XPD reveals they bind to and randomly diffuse on DNA, however, in the presence of UV-induced DNA lesions these complexes stall. Combined with the analysis of a recent cryo-EM structure, we suggest that p44/p62 acts as a novel DNA-binding entity that enhances damage recognition in TFIIH. This revises our understanding of TFIIH and prompts investigation into the core subunits for an active role during DNA repair and/or transcription.

Figure 1.

SEC analysis of XPD, the dimeric p44/p62 and the ternary XPD/p44/p62 complexes. SEC chromatograms of XPD (92 kDa), p44/p62 (132 kDa) and the XPD/p44/p62 (224 kDa) complex were performed in 20 mM HEPES pH 7.5, 250 mM NaCl and 1 mM TCEP on a Superdex200 Increase 3.2/300. XPD was mixed with p44/p62 at an equal molar ratio (final concentration of 40 μM) prior to SEC. The inset shows the SDS-PAGE analysis of the center fractions from the XPD/p44/p62 elution, the markers on the x-axis correspond to the fractions in the gel.

Figure 2.

Steady-state ATPase and helicase activity of XPD complexes in the presence of various DNA substrates and core TFIIH proteins. (A) XPD’s helicase activity is stimulated by N-p44 (hashed) and p44/p62 (white) on an open fork substrate with lengths indicated in nucleotides. XPD alone displays no helicase activity (21). Errors are shown as S.E.M from nine repeats. Statistical significance determined using a student's t-test where * = P < 0.05, n.s = not statistically significant. (B) Summary table of the k_cat values and K_M values obtained from DNA substrate titrations. Errors are S.E.M fit values. (C) Summary of the effect of DNA substrates on XPD-p44/p62 (white) and XPD-N-p44 (hashed) affinities. XPD-p44/p62 shows a 60-fold enhancement in affinity versus XPD-N-p44. (D) XPD-p44/p62’s ATPase is stimulated by the presence of DNA substrates. Damaged dsDNA (squares) binds much more tightly than undamaged DNA (triangles), and approximately equivalent to ssDNA (circles). ATPase rates are normalized to the protein concentration used so that the asymptote reports the k_cat. Errors are shown as S.E.M from three or more repeats. In the absence of DNA, XPD possesses a slow ATPase that is unaffected by N-p44, p44/p62 or damaged DNA (Supplementary Table S1) indicating that p44/p62 displays no ATPase activity by itself.

Figure 3.

p44/p62 binding affinities for different DNA substrates. Fluorescence polarization was used to determine the binding of various substrates to increasing concentrations of p44/p62. p44/p62 displayed an intermediate affinity for dsDNA irrespective of the presence of a fluorescein DNA damage, but a higher affinity for open fork and ssDNA. The tightest binding was observed with ssDNA containing a fluorescein DNA damage. Data were plotted and fitted in GraphPad, presented here as normalized values with error bars representing the SD from at least three repeats.

Figure 4.

Imaging of p44/p62’s motility on DNA tightropes. (A) Schematic drawing of the DNA tightrope assay: Qdot-labeled proteins are imaged interacting with single molecules of DNA suspended between 5 μm diameter beads. (B) Diffusion constant versus exponent (alpha) at high (circles) and low salt (crosses). The average diffusion constant values are given in the main text. Average alpha exponent values were 0.89 ± 0.04 and 0.91 ± 0.02 in high and low salt, respectively. (C) Single molecule fluorescence imaging reveals that the number of diffusing p44/p62 complexes on dsDNA tightropes decreases in the presence of UV-induced DNA damage. Data are average percentages from 8 (No UV) or 6 (UV) experimental repeats with the S.E.M as errors bars, Statistical significance determined using a student's t-test where * = P < 0.05. (D) Example kymograph of a molecule showing constrained diffusion. The scale-bar represents 10 s and 1 μm.

Figure 5.

The position of p62 in apo- and DNA bound TFIIH structures. (A) Representation of the Kokic et al. DNA-bound TFIIH structure (PDB ID: 6RO4) shown in cartoon mode with the EM-envelope displayed as gray mesh. (B) The unmodeled cryo-EM density around p44/p34 was fit using p62 from the Greber et al. apo-TFIIH structure (PDB ID: 6NMI) and the three helical bundle (residues 450–547) superimposed with an rmsd of 0.5 Å after rebuilding. Residues 397–450 differ in orientation due the concerted movement of XPD-p44 in 6RO4. The additional part of p62 interacting with XPD was modeled according to the cryo-EM density with no defined amino acid sequence. (C) Close-up view of XPD and p62 based on the apo-TFIIH structure, with the iron sulphur cluster in XPD shown as spheres. (D) Equivalent view of XPD from the DNA-bound TFIIH structure with the additional cryo-EM density modeled as p62. The modeled structure has been deposited in the worldwide protein databank, accession code: 7AD8.