Protein-Protein Interactions in the core Nucleotide Excision Repair pathway

Abstract

Nucleotide excision repair (NER) clears genomes of DNA adducts formed by UV light, environmental agents, and antitumor drugs. Gene mutations that lead to defects in the core NER reaction cause the skin cancer-prone disease xeroderma pigmentosum. In NER, DNA lesions are excised within an oligonucleotide of 25–30 residues via a complex, multi-step reaction that is regulated by protein-protein interactions. These interactions were first characterized in the 1990s using pull-down, co-IP and yeast two-hybrid assays. More recently, high-resolution structures and detailed functional studies have started to yield detailed pictures of the progression along the NER reaction coordinate. In this review, we highlight how the study of interactions among proteins by structural and/or functional studies have provided insights into the mechanisms by which the NER machinery recognizes and excises DNA lesions. Furthermore, we identify reported, but poorly characterized or unsubstantiated interactions in need of further validation.

Introduction

The nucleotide excision repair (NER) pathway is an essential mechanism for the removal of bulky DNA adducts [1]. NER operates through the sequential action of more than 30 proteins that assemble as a dynamic series of multi-protein complexes at DNA lesions and excise them as part of an oligonucleotide [2–4]. NER is versatile in its ability to remove a range of structurally diverse lesions such as UV induced cyclobutane–pyrimidine dimers (CPDs) and (6–4) pyrimidine–pyrimidone photoproducts ((6–4) PPs) or intrastrand crosslinks caused by anticancer drugs such as cisplatin [1, 5]. Hereditary mutations in NER genes are associated with several disorders, including xeroderma pigmentosum (XP), a prototypical DNA repair disorder characterized by an extreme hypersensitivity to sunlight and more than 2000-fold increase in skin cancer. NER gene mutations are additionally associated with Cockayne syndrome, trichothiodystrophy, and UV sensitive syndrome, which display more complex pathologies including developmental and neurological abnormalities as well as premature aging, due to effects on factors outside of the core NER pathway [6, 7]. NER can be initiated through two damage recognition sub-pathways: global-genome NER (GG-NER) and transcription-coupled NER (TC-NER). In GG-NER, lesions are detected across the entire genome, while TC-NER is responsible for the accelerated detection of lesions in the transcribed strands of active genes (Figure 1). GG-NER is initiated by the recognition of bulky lesions by XPC-RAD23B-CETN2 [8–10], but for non-distorting lesions and in the context of chromatin, UV-DDB (DDB1-DDB2) and its associated ubiquitin ligase are needed for damage detection [11]. Additionally, numerous factors are involved in facilitating NER in the context of chromatin (reviewed in [12]). XPC-RAD23B-CETN2 recruits the 10-subunit TFIIH to the damage, which then uses its helicase subunits, XPB and XPD, to open the DNA duplex and verify the presence of the lesion [13–15]. TC-NER is initiated when an RNA polymerase II (RNAPII) stalls at the lesion site. Several proteins, including CSB, CSA, and UVSSA then assemble at the stalled RNAPII and together with ELOF1, which is part of the RNAPII complex, facilitate the engagement of TFIIH with the lesion site [16, 17]. In both GG-NER and TC-NER, the opening of the DNA by TFIIH generates a bubble with which XPA and RPA scaffold proteins engage to guide incision and removal of the damaged strand by the two endonucleases [18–20]; the first incision 5’ to the lesion is made by ERCC1-XPF, followed by the 3’ incision by XPG [21]. Filling of the resulting gap and sealing of the nick complete the NER process [22, 23].

Figure 1.

Illustration of the steps of global genome (GG-) and transcription coupled (TC-) nucleotide excision repair (NER). GG-NER and TC-NER differ at the damage recognition stage but converge upon formation of the pre-incision complex and subsequent dual incision.

Many excellent reviews on NER are available (including in the “cutting edge perspectives in genome maintenance series” in DNA Repair, [24–26]). This review is concerned with a central driving force of the NER pathway: protein-protein interactions among the six core NER factors - XPC-RAD23B-CETN2, TFIIH, XPA, RPA, XPG, and ERCC1-XPF. We discuss recent advances, formulate open questions, and importantly, provide a guide for the reader as to which of the many reported interactions among NER proteins we consider to be important, and which ones have not been held up to the scrutiny of vigorous experimentation after their initial reports. To keep our discussion focused, we refer the reader to recent reviews that discuss interactions specific to the TC-NER initiation complex [16, 17] or recognition of lesions in chromatin by the UV-DDB-CRL4 complex [12, 27].

Stage 1: Damage recognition by XPC-RAD23B-CETN2 and verification by TFIIH and XPA

GG-NER is initiated by XPC, which senses lesions through the thermodynamic destabilization induced in DNA [28]. High-resolution structures, first of the yeast ortholog Rad4 [29] and subsequently of the human XPC protein [30], show that the transglutaminase domain (TGD) and the first β-hairpin domain (BHD1) bind DNA independently of damage and sequence, while two additional β-hairpin domains BHD2 and BHD3 domains interact with the damaged site, encircling the DNA opposite the lesion (see Figure 2A, B for domain map and schematic of the complex). XPC forms a constitutive heterotrimer with RAD23B [31, 32] and Centrin 2 (CETN2) [10, 33] (Table 1, Figure 2B, C). A ~55 amino acid helical domain in RAD23B (amino acids 275–332) binds yeast Rad4/XPC via its N-terminal helix in the TGD domain [29, 33, 34] (Table 1). RAD23B has been shown to stabilize XPC and promote its binding to damaged DNA in biochemical and cellular experiments [10, 32, 35]. The C-terminal domain of XPC interacts with CETN2, a small EF-hand calcium-binding protein homologous to calmodulin, through residues 847 to 866, which are part of the long C-terminal helix of XPC that links it to XPB [30] (Table 1, Figure 2C, D). Cellular and biochemical studies using XPC mutants that cannot physically interact with CETN2 showed that the interaction between CETN2 and XPC had a stimulatory effect on the NER incision reaction, and contributed to resistance of cells to UV irradiation, although the interaction is not required for NER activity [10]. Interactions with RAD23B and CETN2 increase the cellular stability of XPC and its binding to DNA. They do not directly bind DNA or as far as is currently known, other NER proteins. It is worth noting that the interaction domains among XPC, RAD23B, and CETN2 determined in early biochemical studies were confirmed in structures of the larger NER recognition complexes (Table 1) [30]. Cellular studies have suggested that RAD23B dissociates from XPC following binding of XPC to UV lesions, raising the possibility of the dissociation of XPC-RAD23B-CETN2 prior to assembly of the NER pre-incision complex [35]. The cryo-EM structures of XPC and TFIIH bound to damaged DNA show the presence of both RAD23B and CETN2 suggesting that XPC-RAD23B-CETN2 might disassemble later, likely upon its displacement from the lesion and the NER complex [30].

Figure 2.

Interactions between XPC and TFIIH coordinate damage recognition in NER. (A) Domain map of XPC and TFIIH with their interacting regions highlighted. Overlapping domains are shaded as white stripes. HEL denotes the helicase domain, FeS the iron-sulfur cluster domain, and Zn indicates a Zn finger domain. (B) A schematic model for the interaction of XPC and TFIIH complexes in the presence of a DNA substrate. (C) Structure of the CETN2-XPC-TFIIH/XPB complex showing the interaction required for TFIIH recruitment (PDB: 8EBS). (D) CENT2-XPC interaction with residues L848/L851/L855 of XPC in the CETN2 binding pocket highlighted in red. (E) Solution structure of the AS motif of XPC interacting with the PH domain in p62 of TFIIH (PDB: 2RVB). Residue W133 and V136 of XPC in the binding pocket are highlighted in red.

Table 1.

List of NER protein-protein interactions verified by structural and functional studies.

Protein	Interacting NER factors	Interaction domain (residues)	Methods	PDB ID	Mutations	UV sensitivity	UV repair	NER activity	Reference
XPC	RAD23B	200–220	EM	8EBS, 8EBV	N/D				[29, 30]
	Centrin2	847–866	X-ray, ITC, Mut, EM	2OBH, 8EBS, 8EBV, 8EBT	W848A/L851A/L855A	+	++	+	[10, 30, 80]
	p62/TFIIH	124–137, 165–175	NMR, ITC, Mut	2RVB, 8EBS	W133A	++	+++	N/D	[30, 42]
	XPB/TFIIH	870–940, 849–910	PD, EM	8EBS, 8EBV, 8EBT	N/D				[30, 37]
TFIIH	XPC	p62: 1–106, 176–186	NMR, ITC	2RVB, 8EBS	N/D				[30, 42]
		XPB: 680–720	EM	8EBS, 8EBV, 8EBT	N/D				[30]
	XPA	p8: 1–15	EM	8EBY	Deletion of 1–15 of p8	N/D	N/D	+++	[30, 46]
		p52: 413–459	EM	8EBY	N/D				[30]
		XPB: 705–722	EM	8EBY	N/D				[30]
XPA	TFIIH (XPB, p8, p52)	XPB: 229–241	EM	8EBY	N/D				[30, 45]
		p52: 247–270	EM	8EBY	N/D				[30, 45]
		p8: 241–273	EM	8EBY	H244R	++	++	++	[30, 45, 49]
	RPA	RPA32: 29–46	NMR, X-ray, ITC, SAXS	1DPU, PDBDEV : 124, 128	R30E/K31E/R32E/R34E/R39E/R42E	++	++	+++	[50, 57]
		RPA70: 98–129	ITC, SAXS, Mut	PDBDEV : 039, 040, 124, 128	D101N/E106K/K110E/E111K/F112A/D114N	++	++	+++	[51, 57]
	ERCC1	67–90	X-ray, Mut	2A1I	F75A, Deletion of G73, G74	N/D	N/D	+++	[70, 81]
RPA	XPA	RPA32: 218–270	NMR, X-ray, ITC, SAXS	1DPU, PDBDEV: 124, 128	N/D				[50, 57]
		RPA70: 192–389	Mut, NMR, SAXS	PDBDEV : 039, 040, 124, 128	N/D				[51, 57]
XPG	Active site	1–77, 791–812	X-ray	6TUW, 6TUX	E791A D812A	+++	N/D	+++	[60, 61]
XPF	ERCC1	848–905	EM	6SXA,6SXB	N/D				[73, 75]
	Active site	656–813	EM	6SXB	D687A D731A	N/D	N/D	+++	[76]
ERCC1	XPF	220–297	EM, X-ray	6SXA, 6SXB,2A1J	N/D				[73, 75]
	XPA	110–152	Mut, NMR	2JNW	N110A, Y145A	++	++	++	[69, 71]

EM: Cryogenic Electron Microscopy; FA: Fluorescence Anisotropy; ITC: Isothermal Calorimetry; Mut: Site-directed/deletion mutagenesis; NMR: Nuclear Magnetic Resonance Spectroscopy; PD: Pull-down assay; SAXS: Small Angle X-ray Scattering; X-ray: X-ray crystallography.

After initial damage recognition, XPC-RAD23B-CETN2 recruits TFIIH to the DNA around the lesion. TFIIH contains 10 subunits that are organized into the core (XPB, XPD, p62, p52, p44, p34, p8) and the CAK complex (CDK7, Cyclin H, and MAT1). The CAK complex dissociates from TFIIH once it is engaged in NER; the core alone is fully competent in NER [36]. The first interaction between XPC and TFIIH was shown to be with XPB through the C-terminal domain of XPC (Table 1) [37, 38]. Although this interaction was also seen in cryo-EM structures of NER complexes (Figure 2C) [30, 39], its functional importance for NER has not been thoroughly examined. XPC also interacts with TFIIH through its N-terminus, specifically through its “acidic string (AS)” motif (amino acids 124–137). This motif is rich in acidic residues interspersed with the hydrophobic residues W133 and V136 and interacts with the pleckstrin homology (PH) domain of p62 [40] (Figure 2E, Table 1). This interaction was also observed, albeit not at high resolution, in the cryo-EM structure of the yeast TFIIH/Rad4-Rad23-Rad33 complex on damaged DNA [39]. The PH domain of p62 interacts with many proteins, including XPC, XPG (see below), UVSSA, p53, and VP16 to regulate the function of TFIIH in various pathways [40–42]. With respect to XPC, a W133A mutation in the AS motif of XPC sensitizes the cells to UV and greatly reduces (6–4) PP repair, showing the importance of the AS motif in the assembly of the NER complex (Table 1) [40]. Interestingly, the PH domain of p62 is not visible in the cryo-EM structure of the human NER damage recognition complex [30], suggesting that it is flexible within TFIIH to aid its role in interacting with diverse proteins.

Once TFIIH engages with the XPC-lesion-containing DNA complex, its role is to open the helix, verify the lesion and provide a platform for the assembly of additional factors required to complete the NER incision complex. Central to this role is the XPB translocase, which is believed to open the DNA helix by binding 5’ to the lesion and induce negative supercoiling. The second ATPase of TFIIH, XPD, is a 5’->3’ helicase that engages with the open DNA and tracks along the DNA until it stalls at the bulky lesion [43]. To engage in this process, TFIIH undergoes a significant configurational change, brought about by the departure of the CAK complex. The key to this transition is the loss of the long α-helix of MAT1 that keeps XPB and XPD at a fixed distance. Upon the departure of MAT1 and association with the XPC-lesion-containing DNA complex, core TFIIH opens to assume a horseshoe shape [44] with XPB bound to duplex DNA and XPD placed away from the DNA but ready to engage and exert its activity in lesion scanning. Upon binding of XPA, core TFIIH rearranges to a closed σ-shape [45]. In this state, XPA is bound to DNA and TFIIH, forming an elongated arch over the DNA that bridges XPB and XPD (Figure 3A, 3B) [45]. Once engaged, the XPD helicase scans the damaged strand in the 5’ to 3’ direction and verifies the lesion. The disordered C-terminal domain of XPA engages with three different subunits of TFIIH (p8, p52, and XPB) [30] (Table 1, Figure 3C, 3D). The N-terminus (residues 1–15) of p8, the smallest subunit of TFIIH, has been shown to be essential for the association of XPA with NER complexes [46] (Table 1, Figure 2A). This interaction was clearly visualized in cryo-EM structures as a β-strand that interacts with the p8/p52 dimer [30, 45]. Conversely, a motif containing two conserved His (242, 244) and two conserved Cys (261, 264) residues in the C-terminal domain of XPA has long been known to interact with TFIIH (Table 1) [47, 48]. The H244R allele of XPA, which is found in patients with a mild XP phenotype, but neurological abnormalities, is in this interaction interface and has been shown to reduce NER activity and inhibit the interaction of XPA with TFIIH after UV irradiation [49] (Table 1, Figure 2F). Intriguingly, the H244R mutation exhibits a more pronounced defect in TC-NER than GG-NER, pointing to a possible difference in the interaction of XPA and TFIIH in the two sub-pathways.

Figure 3.

Interactions between TFIIH and XPA coordinate damage verification in NER. (A) Domain map of TFIIH and XPA with their interacting regions highlighted. (B) A schematic model of TFIIH as it undergoes major configurational rearrangement to interact with XPA and DNA in order to perform the lesion verification step. (C) Structure of the interaction of XPA with XPB-p8-p52 subunits of TFIIH (PDB: 8EBY). (D) Deletion of residues 1–15 of p8 (highlighted in magenta) disrupts interaction with XPA.

Stage 2: Formation of pre-incision complex by XPA-RPA mediated scaffolding of NER factors

Following damage verification, the pre-incision complex (PIC) is formed and stabilized by the addition of XPA, RPA, and XPG. The cryo-EM structures of TFIIH-XPA-DNA suggest that XPA is likely recruited first by TFIIH. Here, XPA is positioned at the ss-ds DNA junction of the NER bubble, to set the stage for proper opening of the DNA bubble and allow for damage verification by the XPD helicase. Subsequently, XPA serves important scaffolding functions in assembling the PIC and excising the damaged strand. Despite its small size of only 273 residues, XPA interacts with multiple NER proteins in the NER pathway: TFIIH, RPA, ERCC1-XPF, DDB2, and PCNA (Table 1) [48, 50–54]. In its function as NER scaffolding, XPA works in coordination with the ubiquitous ssDNA binding protein RPA, which occupies the undamaged strand. XPA and RPA together guide the accurate positioning of other NER proteins, especially the two endonucleases XPG and ERCC1-XPF [55].

XPA has a central globular DNA binding domain (DBD) spanning residues 98 to 239 with a zinc binding motif and disordered, flexible N-terminal (1–98) and C-terminal (240–273) domains. RPA is a heterotrimer consisting of the RPA70, RPA32, and RPA14 subunits. Its ssDNA binding apparatus consists of 4 OB-fold domains (RPA70A, RPA70B, RPA70C, and RPA32D) that can bind up to 30 nucleotides of ssDNA with a defined 5’ to 3’ polarity [56] (Table 1, Figure 4A). XPA and RPA interact through two distinct binding motifs: a primary site that involves the XPA N-terminal domain (residues 29–46) and the RPA32C winged-helix recruitment domain [49], and a secondary interaction involving the DNA binding domains (DBD) RPA70AB and XPA DBD [51] (Table 1, Figure 4B, 4C). The XPA DBD–RPA70AB interaction is crucial as it helps the positioning of XPA and RPA on the NER bubble [51]. A series of XPA variants were characterized that disrupted the interaction with the RPA32 and RPA70 subunits and showed that the two interaction surfaces synergistically contribute to NER (Table 1) [57]. Mutations in the RPA32 interacting domain abolished XPA localization to DNA damage, whereas mutations in the RPA70 interacting domain led to proper localization, but delayed departure of XPA from UV damaged sites. These results suggest that the stronger contact with RPA32C is responsible for XPA recruitment and localization to damage, whereas the weaker contact with RPA70AB is indispensable for the completion of NER. Based on SAXS data supported by computational modeling, it was observed that XPA and RPA can engage the two interactions simultaneously and that the XPA-RPA70 interaction localizes the two ss-dsDNA junctions in close proximity to facilitate the dual incision (Figure 4B) [57].

Figure 4.

Interactions in the NER pre-incision complex. (A) Domain map of XPA, RPA, XPG, XPF-ERCC1 with their interacting regions highlighted. Overlapping domains are identified with white stripes. NUC denotes nuclease domains. (B) A schematic model of the NER preincision complex with interactions among the component proteins indicated. (C) Structures of the dual interaction of XPA-RPA through two distinct binding motifs: XPA_29–46/RPA32C (PDB: 1DPU) and XPA DBD/RPA70AB (PDB: PDBDEV_00000039) show the essential scaffolding role in NER. Mutation of residues in XPA that disrupt the RPA32 interaction (R30, K31, R32, R34, R39, R42) and the RPA70 interaction (D101, E106, K110, E111, F112, D114) are highlighted in red. (D) Structure of the acidic region of Rad2/XPG interacting with the pH domain of TFIIH-Tfb1/p62 (PDB: 2LOX), which is the same interaction interface as XPC/TFIIH-p62. This interaction interface has not been characterized using functional studies. (E) Structure showing how ERCC1-XPF is recruited to the pre-incision complex through an interaction between XPA and ERCC1 (PDB: 2JNW). Mutation of residues that disrupt the interaction - G72-G74, F75 of XPA are highlighted in magenta and N110, Y145 of ERCC1 are highlighted in red.

With TFIIH, XPA, and RPA bound at the lesion, the stage is set for the two endonucleases to engage. Although cellular experiments have shown that the two endonucleases can localize to the site of damage independently [9], it is likely that XPG engages with the complex first due to its tight interaction with TFIIH (Table 2) [58, 59], and observations showing that XPG needs to be present (although not catalytically active) for ERCC1-XPF to make the first incision [60, 61].

Table 2.

List of NER protein-protein interactions that are missing or need further validation.

Protein	Interacting NER factors	Interaction domain (residues)	Methods	Reference
XPC	XPA *	154–334	PD	[33]
TFIIH	XPG	XPB, XPD, p62, and p44	PD	[82]
	XPF	XPB RPA32: 4–29;	Mut	[77]
XPA	RPA *	RPA70: 153–176 aa	PD, Mut	[52, 83, 84]
RPA	XPG	N/D	co-IP	[85]
	XPF	RPA70	YH	[86]
XPG	TFIIH	N/D	PD	[82]
	RPA	668–747	co-IP	[85]
XPF	TFIIH	N/D	Mut	[77]
	RPA	N/D	YH	[86]

CoIP: Co-Immunoprecipitation; Mut: Site directed/deletion mutagenesis, PD: pull-down assay, YH: Yeast two-hybrid assay.

^*Interaction interfaces not supported by current experimental structures and functional studies

XPG is a member of the FEN-1 (flap endonuclease) family of nucleases with a split active site consisting of the highly conserved N-terminal (N)- and internal (I)-nuclease regions separated by a large spacer region of 680 residues [62]. Its extended C-terminus is unique to XPG compared to other proteins in the FEN-1 family [63]. The spacer region in XPG is highly acidic and does not contain any known structural or functional motifs. The spacer region is known to interact with TFIIH (Table 2) [64, 65], although the molecular basis of the interactions between XPG and the subunits of TFIIH remains to be elucidated. In yeast, two acidic regions of Rad2/XPG (residues 359–383 and 642–690) were found to interact with the PH domain of Tfb1/p62 [65, 66] (Table 2, Figure 4D). This binding mode is similar to the interaction of Rad4/XPC with Tfb1/p62 and NMR titrations show that Rad2/XPG can outcompete Rad4/XPC to bind to the same interaction interface in Tfb1/p62. Since experimental evidence suggests that XPC and XPG do not co-exist in the NER complexes [19], it is tempting to speculate that the common interaction of XPC and XPG with p62 are key to a handover mechanism for the replacement of XPC with XPG. Cross-linking mass spectrometry provided additional information on the interaction interface of XPG with TFIIH (Table 2) [45]. The N-terminal region of XPG was localized near the FeS cluster and Arch domains of XPD and the C-terminal region of XPG is positioned near XPB and p52 subunits. It has been shown that XPG can stimulate the helicase and translocase activities of XPD and conversely that XPD can influence XPG cleavage activity [45, 68]. Further studies will be needed to fully define the architecture and functionality of XPG-TFIIH interactions.

Stage 3: Dual incision by ERCC1-XPF and XPG endonucleases

The active incision complex requires the second endonuclease ERCC1-XPF. The TFIIH-XPA-RPA-XPG complex provides the necessary preorganization to enable recruitment of ERCC1-XPF, which occurs through a direct interaction of XPA with ERCC1 (Table 1) [69–71]. ERCC1 engages a short 14 residue motif in the XPA N-terminal domain (residues 67–80) (Table 1). The XPA motif undergoes a disorder-to-order transition that is formed as a loop containing G72-G74 and F75 is inserted into an XPA binding pocket in ERCC1 (Table 1, Figure 4E). Deletion of the conserved glycines or mutation of F75 to A inhibits NER activity in vitro (Table 1) [71]. Mutation of two conserved residues (N110A/Y145A) in the XPA binding pocket of ERCC1 dramatically reduced NER activity in vitro and in vivo as ERCC1-XPF could not be recruited due to the weakened interaction with XPA (Table 1) [69].

ERCC1 and XPF form an obligate heterodimer through two evolutionarily related regions [72]. The two helix-hairpin-helix (HhH) domains at the C-terminus of XPF and ERCC1 form a tight heterodimer and bind the DNA at the junction [73–75]. Similarly, the central regions of ERCC1 and XPF dimerize, with this region in XPF making up the active site of the nuclease [76] and providing the XPA interaction domain in ERCC1 (Table 1) [71]. Less is known about the function of the helicase-like domain (HLD) of XPF in NER, although this HLD is likely to contribute to junction binding and/or protein-protein interactions. High resolution cryo-EM structures were determined for both DNA-free and DNA-bound ERCC1-XPF and show that the XPF HLD regulates XPF’s catalytic activity by forming an extensive interface with the ERCC1 HhH domain and blocking the catalytic site [75]. This autoinhibitory conformation of XPF needs to be released for XPF to bind to the DNA junction and perform 5’ incision. One mode of XPF activation could be the interaction of XPF HLD with the XPB subunit of TFIIH. S751 in XPB is a crucial residue in the C terminal region of the protein and its phosphorylation inhibits NER in vivo and regulates 5′ incision by the ERCC1- XPF (Table 2) [77]. The interaction between XPB and XPF seems ripe for structural and functional studies.

With the engagement of ERCC1-XPF, the stage is set for dual incision to occur with the first incision believed to be by ERCC1-XPF 5’ to the lesion [21]. Evidence for this order was provided by experiments showing that incision by ERCC1-XPF and initiation of repair synthesis requires the presence but not the catalytic function of XPG. Based on existing structural and biochemical evidence and in support of distinct structural and catalytic roles of XPG, the protein is likely to undergo a structural rearrangement from inactive to an active configuration to perform incision [78]. It has furthermore been shown that the incision activity of XPG can be stimulated by XPD if once it is stalled at the lesion and its ATPase activity is blocked [68]. Incision by ERCC1-XPF produces a free 3’-OH that can be used by the replication machinery to initiate repair synthesis [21]. After initiation of DNA repair synthesis, XPG incises 3’ to the lesion and marks the transition from the incision to the repair synthesis steps. While we focused on interactions that regulate formation of the incision complex, specific protein-protein interactions between NER and replication factors are certain to be involved in the later steps in NER. These are needed to provide a tight coordination of the two phases of NER and thereby avoid the formation of long-lived and toxic gaps upon removal of the oligonucleotide containing the damage [79].

Conclusion

Characterization of interactions among the constituent proteins using a broad range of biochemical, structural, and biophysical techniques is important to completely understand the inner workings of the complex NER machinery. Conversely, it is imperative to test the functional importance of the models derived from these results in biochemical and cellular NER assays. Through the years many interactions among NER proteins were reported based on yeast two-hybrid, co-IP, and pull-down studies (Table 2). Some of these reports lacked proper structural and functional characterization and were not validated in subsequent studies. In the main body of this review, we discussed the interactions that we consider to be properly validated by functional and/or structural studies. These interactions were critical to gain insight into the complex NER mechanisms. For completion, in Table 2 we include the list of interactions that were reported between NER proteins that we consider to be lacking proper validation using structural or functional studies or have been proven to be artefactual in subsequent studies. Figure 4 shows a summary of the interactions between the NER core factors discussed in this review (solid arrows) and partially characterized or expected interactions (dotted arrow) in the pre-incision complex. The latter await functional and structural characterization and once available, will provide further insight into how multiple proteins shape the pre-incision complex to license the incision reaction (Table 2). With advances in high-resolution cryo-EM structural studies and the development of deep-learning algorithms to predict structures of multi-protein complexes, we are now at a point in time where we can expect to gain a more complete understanding of the dynamic protein-protein interactions in NER that seemed out of reach just a few years ago.

Figure 5.

Summary of protein-protein interactions between core NER factors. Interactions validated using structural and functional studies are depicted as straight solid lines. Less well characterized and potential interactions are depicted as curved dotted lines.

Highlights.

• NER is dynamic multistep pathway for the repair of bulky DNA lesions

• Protein-protein interaction regulate the steps in the NER reaction

• Structural and functional studies have elucidated the roles of protein-protein interactions in NER

• This review provides a guide to the protein-protein interactions in NER