A new crystal form of the DNA-free full-length XPD helicase from Thermoplasma acidophilum

We present a 2.13 Å resolution structure of DNA-free XPD from T. acidophilum in a novel orthorhombic lattice obtained under high ionic strength conditions. Compared with prior hexagonal forms, this structure features fewer crystal contacts and uniquely resolves loop 510–514. Interestingly, Tyr425 adopts an inward-facing conformation that would clash with DNA, suggesting a possible autoinhibitory role. These findings highlight how crystallization conditions influence lattice stability and reveal functionally relevant conformations.

Abstract

The XPD helicase plays a critical role in DNA repair and serves as a model for structural studies of superfamily 2 (SF2) helicases. We report a novel ortho­rhombic crystal form of DNA-free Thermoplasma acidophilum XPD (TaXPD) obtained under high ionic strength conditions generated by sodium potassium tartrate and NaCl-based vapor-diffusion conditions, in contrast to earlier previously reported conditions that used polyols (PEG) or diols (MPD). The crystals belonged to space group P2₁2₁2₁ (a = 59.53, b = 96.00, c = 159.09 Å) and diffracted to 2.13 Å resolution, yielding the highest resolution TaXPD structure to date. Structural analysis showed that this crystal form contains fewer intermolecular interfaces than the previously reported hexagonal lattice, as supported by Protein Interfaces, Surfaces and Assemblies (PISA) analysis. This supports the determination of the 510–514 loop in the long-form DNA-free TaXPD, which was previously disordered in other structures. This work highlights how crystallization conditions influence lattice organization, structural completeness and diffraction quality. In this structure, Tyr425 adopts a conformation that may regulate DNA access in the DNA-free state.

1. Introduction

Nucleotide-excision repair (NER) is a highly conserved DNA-repair pathway that preserves genome integrity by recognizing and removing a wide variety of bulky DNA lesions, including those induced by ultraviolet radiation and environmental mutagens (D’Souza et al., 2022 ▸). A key player in this pathway is the superfamily 2 (SF2) helicase XPD, which unwinds DNA around damage sites in an ATP-dependent manner. In eukaryotic cells, XPD is a core component of the multiprotein TFIIH complex. While its ATP-dependent helicase activity is essential for NER, XPD functions solely as a structural scaffold during transcription initiation (Kuper et al., 2024 ▸). Genetic alterations in the human XPD helicase give rise to three clinically distinct hereditary diseases, xeroderma pigmentosum (XP), Cockayne syndrome (CS) and trichothiodystrophy (TTD), each illustrating the critical role of XPD in coordinating NER and maintaining genomic stability (Koch et al., 2016 ▸). Beyond its established roles in NER and transcription, TFIIH, and particularly its XPD subunit, has been implicated in additional cellular functions such as chromosome segregation and mitochondrial genome maintenance, further highlighting its importance in genome stability beyond canonical repair pathways (Zachayus et al., 2025 ▸).

Archaeal XPD homologues, including TaXPD from Thermoplasma acidophilum, have proven to be invaluable for elucidating helicase structure and mechanism due to their inherent thermostability and conservation of key functional domains (Fan et al., 2008 ▸; Liu et al., 2008 ▸; Wolski et al., 2008 ▸). Early high-resolution crystal structures from multiple archaeal species revealed that XPD contains two RecA-like motor domains and two auxiliary modules, the iron–sulfur (4Fe–4S) cluster and the Arch domain, that together coordinate DNA binding, unwinding and damage verification (Fan, 2013 ▸). These models postulated how single-stranded DNA threads through a tunnel formed by the Arch and FeS domains and established a structural framework for interpreting helicase action. Biochemical studies on archaea as well as eukaryotes support XPD helicase as a damage-verification enzyme by its ability to stall at DNA lesions (Buechner et al., 2014 ▸; Mathieu et al., 2010 ▸). Building on these foundations, cryo-EM reconstructions of human and yeast TFIIH complexes have resolved XPD in its native multi-protein context, highlighting how it coordinates with XPB and other TFIIH subunits to scan for DNA damage during NER (Greber et al., 2019 ▸; Kokic et al., 2019 ▸; Kim et al., 2023 ▸; van Eeuwen et al., 2021 ▸). More recently, cryo-EM structures captured eukaryotic XPD directly stalled on a cross-linked DNA lesion, providing direct evidence for how DNA damage obstructs helicase translocation (Kuper et al., 2024 ▸). Integrative structural modeling that combines cryo-EM, cross-linking mass spectrometry and AlphaFold2 predictions has further resolved the architecture of the human pre-incision complex, leading to a prediction for how XPD collaborates with TFIIH, XPG and XPF/ERCC1 to orchestrate precise dual incision around DNA lesions (Yu et al., 2024 ▸). Unlike its eukaryotic counterpart, archaeal XPD operates without TFIIH regulatory subunits such as p44, p62 and MAT1, offering a minimal system to examine how the Fe–S cluster and ARCH domain coordinate DNA unwinding and may facilitate redox-based lesion sensing (He et al., 2023 ▸).

Previous TaXPD structures (Table 1 ▸) crystallized in the hexagonal space group P6₅ using precipitants containing polyols such as PEG or diols such as MPD. Here, we report a DNA-free TaXPD structure crystallized under tartrate/NaCl-supported conditions, providing a new crystal form with improved diffraction quality. We further analyzed the crystal contacts using PISA (Krissinel & Henrick, 2007 ▸) to explore the relationship between crystallization conditions, lattice organization and diffraction outcomes. Additionally, there are two versions of TaXPD currently deposited in the PDB (Table 1 ▸): a short one of 602 amino acids in length (UniProt sequence Q9HM14) and a long one of 620 amino acids (RefSeq WP_010900485). The TaXPD structure presented in this work represents the long-form DNA-free version. Our structure now defines the two gaps that existed in the TaXPD–ssDNA complex (Kuper et al., 2012 ▸), where four nucleotides were modeled.

Table 1. Current and newly reported structures of T. acidophilum XPD in the PDB.

PDB code	Space group	a, b, c (Å)	dmin (Å)	Crystallization solution	Reference	TaXPD version†
2vsf, DNA-free	P65	78.9, 78.9, 174.0	2.9	0.1 M HEPES pH 5, 5–10% PEG 400, 0.2 M MgCl2	Wolski et al. (2008 ▸)	Short
4a15, DNA bound	P65	79.0, 79.0, 175.7	2.2	0.05 M MES pH 6.5, 0.1 M MgSO4, 15% MPD	Kuper et al. (2012 ▸)	Long
5h8w, DNA covalently bound‡	P65	78.6, 78.6, 177.8	2.2	0.2 M KCl, 0.1 M magnesium acetate·4H2O, 0.05 M sodium cacodylate·3H2O pH 6.5, 10% PEG 8K	Constantinescu-Aruxandei et al. (2016 ▸)	Short
This work, DNA-free	P212121	59.5, 96.0, 159.2	2.1	0.1 M bis-Tris propane, 1 M sodium potassium tartrate, 0.2 M NaCl pH 7.0	This work	Long

^†‘Long’ refers to the 620-residue version of TaXPD containing 19 additional N-terminal residues (RefSeq WP_010900485.1). ‘Short’ refers to the 601-residue version (UniProt Q9HM14).

^‡DNA is covalently cross-linked via a disulfide bond between an engineered cysteine (E542C) in TaXPD helicase domain 2 and a 5′-alkanethiol-modified DNA oligonucleotide.

2. Materials and methods

2.1. Macromolecule production

The gene encoding long-form TaXPD (RefSeq WP_010900485.1) was synthesized by PCR from T. acido­philum genomic DNA. The synthetic gene was cloned into pET-28a and transformed into Escherichia coli BL21(DE3) Rosetta cells (Novagen) for expression (Table 2 ▸). A 5 ml starter culture in lysogeny broth (LB) with 50 µg ml⁻¹ kanamycin and 34 µg ml⁻¹ chloramphenicol was inoculated from a glycerol stock and grown overnight. The next day, 6 l LB medium was inoculated with 2 ml starter, grown at 310 K to an OD₆₀₀ of 0.5, cooled to 295 K and induced with 0.1 mM isopropyl β-d-1-thiogalactopyranoside overnight.

Table 2. Recombinant expression and purification details for TaXPD.

Source organism	T. acidophilum
DNA source	Gene synthesis and cloning
Cloning vector	pET-28a
Expression host	E. coli BL21(DE3) Rosetta
Complete amino-acid sequence of the construct produced†	HHHHHHSSGLEVLFQGPHRGSEFELMYENRQYQVEAIDFLRSSLQKSYGVALESPTGSGKTIMALKSALQYSSERKLKVLYLVRTNSQEEQVIKELRSLSSTMKIRAIPMQGRVNMCILYRMVDDLHEINAESLAKFCNMKKREVMAGNEAACPYFNFKIRSDETKRFLFDELPTAEEFYDYGERNNVCPYESMKAALPDADIVIAPYAYFLNRSVAEKFLSHWGVSRNQIVIILDEAHNLPDIGRSIGSFRISVESLNRADREAQAYGDPELSQKIHVSDLIEMIRSALQSMVSERCGKGDVRIRFQEFMEYMRIMNKRSEREIRSLLNYLYLFGEYVENEKEKVGKVPFSYCSSVASRIIAFSDQDEEKYAAILSPEDGGYMQAACLDPSGILEVLKESKTIHMSGTLDPFDFYSDITGFEIPFKKIGEIFPPENRYIAYYDGVSSKYDTLDEKELDRMATVIEDIILKVKKNTIVYFPSYSLMDRVENRVSFEHMKEYRGIDQKELYSMLKKFRRDHGTIFAVSGGRLSEGINFPGNELEMIILAGLPFPRPDAINRSLFDYYERKYGKGWEYSVVYPTAIKIRQEIGRLIRSAEDTGACVILDKRAGQFRKFIPDMKKTSDPASDIYNFFISAQAREKYGA

^†The bold amino-acid sequence highlights the 19 amino acids present only in the long TaXPD construct. Underlined residues indicate the extra residues from the vector for N-terminal His-tag attachment.

Cells were harvested at 4000 rev min⁻¹ for 15 min at 277 K, resuspended in lysis buffer (50 mM Tris–HCl pH 7.0, 500 mM NaCl, 10% glycerol, 1 mM MgCl₂) and lysed by sonication using the following settings: 2 s on/3 s off at 100%, 80% and 60% amplitude, 2 min each. The lysate was clarified at 22 500 rev min⁻¹ for 45 min at 277 K.

The lysate was loaded onto a 5 ml Ni–NTA column equilibrated with lysis buffer plus 5 mM imidazole, washed with ten column volumes of 5 and 10 mM imidazole and eluted with 300 mM imidazole. Fractions were analyzed by SDS–PAGE. Elution fractions were pooled, applied onto a HiPrep S-200 HR size-exclusion chromatography column in 25 mM Tris–HCl pH 7.5, 200 mM NaCl, 5% glycerol (Fig. 1 ▸a), 1 mM MgCl₂, eluted at 0.3 ml min⁻¹, concentrated using a 50 kDa Amicon filter, quantified using a NanoDrop, flash-frozen and stored at 193 K. Crystals suitable for structure determination were obtained despite the presence of minor impurities (Fig. 1 ▸).

Figure 1.

Size-exclusion chromatography (SEC), SDS–PAGE analysis and crystallization of TaXPD. (a) SEC chromatogram of purified TaXPD. Inset: 4–20% SDS–PAGE gel showing the final TaXPD preparation (lane 2) and molecular-weight standards (lane 1; 250–15 kDa). The TaXPD band migrates at ∼74 kDa. (b) Representative image of the rod-shaped crystals obtained prior to maximal size and optimization. The red scale bar (∼200 µm) is approximate and shown for reference only; the central tick mark denotes 100 µm.

2.2. Crystallization

Crystallization was performed at the UCR Macromolecular X-ray Core Facility by sitting-drop vapor diffusion using Intelli-Plate 96-3 plates and a Phoenix robot. Protein solution (3 mg ml⁻¹) was mixed 1:1 with 200 nl reservoir solution. Condition 85 from MCSG kit 3 (0.1 M imidazole pH 8.0, 0.2 M NaCl, 1 M sodium potassium tartrate) yielded the best crystals and was subsequently optimized by replacing imidazole with 0.1 M bis-Tris propane pH 7.0 (Table 3 ▸ and Fig. 1 ▸b). A variety of DNA substrates (single-stranded, forked and bubble- and lesion-containing oligonucleotides) were prepared in 10 mM Tris–HCl pH 8.0 at a 1 mM concentration and were incubated with the protein at a 1.5-fold molar excess during crystallization screening and refinement trials. A single crystal was soaked in reservoir solution supplemented with 25% glycerol as a cryoprotectant, mounted in a nylon loop, flash-cooled in liquid nitrogen at 100 K and stored for data collection.

Table 3. Crystallization parameters for recombinant TaXPD.

Method	Vapor diffusion
Plate type	VDX 24-well (hanging drop)
Temperature (K)	293
Protein concentration (mg ml−1)	4.17
Buffer composition of protein solution	25 mM Tris–HCl pH 7.5, 200 mM NaCl, 5% glycerol, 1 mM MgCl2
Composition of reservoir solution	1 M sodium potassium tartrate, 200 mM NaCl, 0.1 M bis-Tris propane pH 7.0
Volume and ratio of drop	3 µl (2 µl protein:1 µl reservoir)
Volume of reservoir (µl)	500

2.3. Data collection and processing

A total of 33 crystals were screened during data collection. Complete data sets were obtained from six crystals, with diffraction quality ranging from 3.5 Å to better than 3.0 Å resolution, and in one case extending to the 2.13 Å resolution reported for the final structure. Phenix.xtriage (Adams et al., 2011 ▸) analysis showed a single mild anomaly at 2.67 Å without any loss of completeness, indicating no significant ice diffraction; visually, no ice rings were observed in the images. The crystal reported here yielded the highest resolution data set (Table 4 ▸).

Table 4. Data collection and processing.

Values in parentheses are for the outer shell.

Diffraction source	Beamline 8.3.1, Advanced Light Source
Wavelength (Å)	1.11583
Temperature (K)	100
Detector	PILATUS3 S 6M, Dectris
Crystal-to-detector distance (mm)	380
Rotation range per image (°)	0.15
Total rotation range (°)	360
Exposure time per image (s)	0.2
Space group	P212121
a, b, c (Å)	59.53, 96.00, 159.09
α, β, γ (°)	90, 90, 90
Mosaicity (°)	0.127
Resolution range (Å)	50.00–2.11 (2.18–2.11)
Total No. of reflections	672888 (37682)
No. of unique reflections	52613 (3828)
Completeness (%)	100 (99.9)
Multiplicity	12.8 (9.8)
〈I/σ(I)〉†	11.2 (0.3)
R meas	0.174 (5.694)
Overall B factor from Wilson plot (Å2)	50.2

^†The outer shell 〈I/σ(I)〉 falls below 2.0 in the 2.18–2.11 Å resolution bin, where it is 0.3.

2.4. Structure solution and refinement

X-ray diffraction data were processed using XDS (Kabsch, 2010 ▸). The processed data set was imported into CCP4 Cloud for structure determination (Krissinel et al., 2022 ▸). Molecular replacement was performed with Phaser (McCoy et al., 2007 ▸) using the TaXPD structure (PDB entry 2vsf; Wolski et al., 2008 ▸), which was pruned to only contain the protein and iron–sulfur center. Manual model building was performed in Coot (Emsley & Cowtan, 2004 ▸), followed by iterative refinement using REFMAC5 (Murshudov et al., 2011 ▸) (Table 5 ▸). Structural visualization was performed in PyMOL (DeLano, 2002 ▸). Although DNA substrates were included during crystallization screening and optimization, the resulting electron-density maps did not reveal any interpretable DNA density in the crystals, confirming that the final structure corresponds to the DNA-free form of the protein.

Table 5. Structure solution and refinement.

Values in parentheses are for the outer shell.

Resolution range (Å)	12.822–2.127 (2.181–2.127)
Completeness (%)	99.4
No. of reflections, working set	49231 (3488)
No. of reflections, test set	2530 (190)
Final Rcryst	0.202 (0.46)
Final Rfree	0.245 (0.47)
No. of non-H atoms
Total	5148
Protein	5040
Ion	12
Ligand	8
Water	88
R.m.s. deviations
Bond lengths (Å)	0.0117
Angles (°)	2.10
Average B factors (Å2)
Protein	70.3
Ion	100.0
Ligand	57.4
Solvent	68.6
Ramachandran plot
Most favored (%)	96.8
Allowed (%)	2.6

3. Results and discussion

3.1. Crystal packing and analysis

TaXPD crystallized in an orthorhombic form using sodium potassium tartrate and NaCl as the primary precipitants via vapor diffusion, differing from previously reported hexagonal crystals grown using polyols (PEG) or diols (MPD) as precipitants. The refined model includes an N-terminal expression tag of 25 residues, with electron density visible up to Ser−4. Residues −4 to 617 were modeled, with the final three residues disordered. Domain boundaries follow prior assignments, helicase domain 1 (HD1; −4–87, 178–225 and 366–407), FeS (88–177), Arch (226–365) and helicase domain 2 (HD2; 408–617), consistent with earlier work (Wolski et al., 2008 ▸).

To examine how crystallization conditions influence lattice organization and diffraction behavior, Protein Interfaces, Surfaces and Assemblies (PISA; Krissinel & Henrick, 2007 ▸) analysis was performed on the orthorhombic TaXPD structure and the previously reported hexagonal DNA-free form (PDB entry 2vsf; Wolski et al., 2008 ▸) for comparison. PISA evaluates crystal contacts by calculating interface area, solvation free-energy change upon interface formation (ΔⁱG) and statistical significance via p-values, where lower p-values indicate more specific, nonrandom interactions.

The orthorhombic crystals contain three principal interfaces (Fig. 2 ▸a and Table 6 ▸), whereas the hexagonal form exhibits six (Fig. 2 ▸b and Table 6 ▸), each characterized by distinct interface areas and thermodynamic parameters (Table 6 ▸). In the orthorhombic structure, two interfaces have negative ΔⁱG values (−6.8 and −1.0 kcal mol⁻¹), while the corresponding major interfaces in the hexagonal form show positive or near-neutral values (+4.0 and +0.9 kcal mol⁻¹). Interfaces 4–6 of PDB entry 2vsf display smaller buried areas (19.6–42.4 Ų) and are not discussed further here because they constitute the smallest interfaces observed and do not contain any salt bridges or hydrogen bonds; their parameters are listed in Table 6 ▸.

Figure 2.

Crystal-packing interfaces of TaXPD identified by PISA analysis. (a) Surface representation of TaXPD from this work, colored by domain: HD1 (pale green), FeS (light pink), Arch (yellow) and HD2 (salmon red). The three crystallographic interfaces identified by PISA are shown: interface 1 (blue), interface 2 (magenta) and interface 3 (teal green). Symmetry-related molecules are displayed as ribbons colored according to their respective symmetry operations. The right-hand view is rotated 180° about the vertical axis relative to the left-hand view. Complete hydrogen-bond and salt-bridge data for these interfaces are provided in Supplementary Table S1. (b) Equivalent surface representation of the previously reported DNA-free TaXPD structure (PDB entry 2vsf; Wolski et al., 2008 ▸) shown using the same domain and interface color scheme. Of the six interfaces identified by PISA for this model, only the three largest are displayed here. Detailed interaction data, including hydrogen bonds, salt bridges and symmetry operations, are listed in Supplementary Table S2.

Table 6. Comparison of PISA-reported crystal contacts.

This work (PDB entry 9p55) versus DNA-free TaXPD (PDB entry 2vsf).

PISA results, TaXPD	Crystal-contact interface	Symmetry operation	Interface area (Å2)	ΔiG (kcal mol−1)	ΔiG, p-value	N HB †	N SB ‡
This work	1	−x, y − 1/2, −z + 1/2	1001.5	−1	0.553	13	15
	2	−x − 1/2, −y + 1/2, −z	935.8	−6.8	0.213	7	3
	3	x, y, z	149.5	3	0.887	1	3
2vsf	1	x − y + 1, x + 1, z − 1/6	1103.3	4	0.769	10	8
	2	x − y, x, z − 1/6	744.9	0.9	0.588	9	0
	3	−y, x − y + 1, z − 1/3	348.4	0.3	0.543	5	2
	4	x, y − 1, z	42.4	−0.5	0.322	0	0
	5	−y + 1, x, y + 1/3	19.6	0.5	0.729	0	0
	6	x − y + 1, x + 1, z − 1/6	18.9	0.3	0.649	0	0

^†N_HB, number of hydrogen bonds.

^‡N_SB, number of salt bridges.

Both crystal forms involve comparable domain contacts but differ in the number and organization of lattice interfaces. Variation in ΔⁱG and p-values reflects differences in packing and space-group symmetry. PISA provides thermodynamic estimates of interfacial interactions, but does not account for mosaicity, molecular motion or lattice disorder. These analyses describe packing differences among DNA-free TaXPD structures obtained under distinct crystallization conditions, with the orthorhombic form showing slightly improved diffraction.

3.2. Loop 510–514 is resolved in long-form DNA-free TaXPD

Among all available TaXPD structures, only our DNA-free long-form model and the short-form DNA-bound structure (PDB entry 5h8w; Constantinescu-Aruxandei et al., 2016 ▸) are fully resolved without gaps. The long-form DNA-bound structure (light blue; PDB entry 4a15; Kuper et al., 2012 ▸) is missing residues 422–426 and 507–514 (Fig. 3 ▸). Notably, the 421–429 loop, adjacent to the DNA-binding site, remains disordered in PDB entry 4a15 despite DNA being bound. This is particularly significant because Tyr425, within this loop, engages in stacking interactions with DNA in PDB entry 5h8w (pink). The absence of this loop in PDB entry 4a15 suggests local flexibility or incomplete engagement of DNA under the crystallization conditions used.

Figure 3.

Superimposed view of TaXPD structures showing loop 510–514 and Tyr425 side-chain conformations. The DNA-bound long-form structure (light blue ribbons with a light green loop and green DNA; PDB entry 4a15) lacks residues 422–426 and 507–514. The short-form DNA-free structure (gray ribbons with a yellow loop; PDB entry 2vsf) also lacks loop 510–514. The short-form structure with covalently bound ssDNA is shown as pink ribbons with an orange loop and purple DNA (PDB entry 5h8w). This structure in this work is shown as salmon ribbons with a blue loop (PDB entry 9p55).

The short-form DNA-free structure (gray; PDB entry 2vsf; Wolski et al., 2008 ▸) also lacks residues 510–514, even though they are encoded in the construct. As shown in Fig. 3 ▸, our model (salmon) uniquely resolves this loop in the absence of DNA and within the context of the full-length protein, providing a new view of the conformation of HD2 under DNA-free conditions.

While Tyr425 is modeled in PDB entry 2vsf, it does not obstruct the DNA-binding groove to the extent observed in our structure. Interestingly, despite the structure being DNA-free, the authors of PDB entry 2vsf observed strong residual electron density in the central DNA passage, which they speculated may originate from oligonucleotide fragments or phosphate ions that are retained during purification.

Our orthorhombic structure, which crystallized in a distinct P2₁2₁2₁ lattice under conditions distinct from PEG/MPD formulations, presents a new conformation of Tyr425. In our model, this residue is oriented inwards, projecting towards the wedge region: a nearby α-helical feature proposed to aid in duplex strand separation by prying apart the DNA strands at the fork (Wolski et al., 2008 ▸; Constantinescu-Aruxandei et al., 2016 ▸). In this DNA-free specific conformation, Tyr425 would sterically clash with modeled DNA from either PDB entry 4a15 or 5h8w (Fig. 3 ▸, right), suggesting that the wedge and the adjacent loop act as a gate or autoinhibitory element in the absence of DNA. This arrangement may help to prevent nonspecific engagement or stabilize the helicase in a resting state prior to substrate loading.

4. Conclusions

We report the highest resolution DNA-free TaXPD structure to date, crystallized using sodium potassium tartrate with NaCl as the precipitant system in a novel orthorhombic crystal. Analysis of crystal contacts suggests that the improved lattice stability arose from fewer, but stronger and more specific interfaces compared with previous structures. These findings underscore how crystallization environment and symmetry influence crystal packing and diffraction outcomes.

In addition to the improved resolution and lattice order, our structure uniquely resolves loop 510–514 in the long-form, DNA-free TaXPD model. This flexible region was previously modeled only in the DNA-bound short-form structure (PDB entry 5h8w; Constantinescu-Aruxandei et al., 2016 ▸), but is missing from both the long-form DNA-bound structure (PDB entry 4a15; Kuper et al., 2012 ▸) and the short-form DNA-free structure (PDB entry 2vsf; Wolski et al., 2008 ▸). Its resolution in our model under DNA-free conditions provides new insight into the conformation of HD2 and highlights the structural completeness achieved under the optimized crystallization conditions used in this work.

Supplementary Material

PDB reference: DNA-free full-length XPD, 9p55

Conflict of interest

The authors declare no conflicts of interest

Data availability

The data supporting these findings are available within the article and supplementary material; coordinate and structure factor files have been deposited in the Protein Data Bank under accession code 9p55.