46

Biology

ANNUAL REPORT OF OSAKA UNIVERSITY—Academic Achievement—2004-2005

UV-Induced Ubiquitylation of XPC Protein Mediated by UV-DDB-Ubiquitin Ligase ComplexPaper in journals: this is the first page of a paper published in Cell.[Cell] 121, 387-400 (2005)

� Reprinted from Cell, 121, Sugasawa, K. et al., UV-Induced Ubiquitylation of XPC Protein Mediated by UV-DDB-Ubiquitin Ligase Complex, 387-400, Copyright 2005, with permission from Elsevier.

47

Osaka University 100 Papers : 10 Selected Papers

ANNUAL REPORT OF OSAKA UNIVERSITY—Academic Achievement—2004-2005

The following is a comment on the published paper shown on the preceding page.

UV-Induced Ubiquitylation of XPC Protein Mediated by UV-DDB-Ubiquitin Ligase ComplexHANAOKA Fumio(Graduate School of Frontier Biosciences)

Introduction

Nucleotide excision repair (NER) is a versatile DNA repair path-way, which eliminates a wide variety of helix-distorting base

lesions, including ultraviolet light (UV)-induced cyclobutane pyrim-idine dimers (CPDs) and pyrimidine-pyrimidone (6-4) photo-products (6-4PPs), as well as bulky adducts induced by numerouschemical compounds. Impaired NER activity has been associatedwith several rare autosomal recessive disorders in humans, suchas xeroderma pigmentosum (XP) and Cockayne syndrome (CS)[1, 2]. XP patients are clinically characterized by cutaneous hyper-sensitivity to sunlight exposure and predisposition to skin cancer.Seven NER-deficient genetic complementation groups of XP (XP-A through G) and two groups of CS (CS-A and B) have been iden-tified, and all of the corresponding genes (XPA~XPG, CSA, andCSB) have now been cloned.

Accumulating evidence indicates that two protein complexesplay essential roles in NER-specific damage recognition. One isthe XP group C (XPC) protein complex existing in vivo as a het-erotrimeric complex with one of the two mammalian homologs ofS. cerevisiae Rad23p (HR23A or HR23B) and centrin 2 [3, 4].This complex binds to various NER-type lesions in vitro, includ-ing UV-induced 6-4PP [5]. The other is UV-damaged DNA bind-ing protein (UV-DDB) consisting of DDB1 and DDB2 (the XP-E responsible gene product), and has a much higher binding affin-ity and specificity for damaged DNA than XPC. Interestingly, UV-DDB is contained in a large complex with cullin 4A, Roc1, andCOP9 signalosome, which are components of ubiquitin ligase (E3)[6]. Although UV-DDB-associated E3 appears to be activated uponUV irradiation of cells, the physiological substrates of E3 activi-ty have not yet been determined. Here we present evidence thatthe UV-DDB-E3 complex ubiquitylates XPC in response to UVirradiation of cells.

The XPC protein is ubiquitylated in response to UV irradiationTo examine the in vivo response of the XPC complex to UV

irradiation, the SV40-transformed normal human fibroblast cellline was UV irradiated and then incubated for various time peri-ods in the presence of cycloheximide to inhibit de novo proteinsynthesis after UV irradiation. After cell lysis, the resulting solu-ble fractions were subjected to immunoblotting with anti-XPCantibodies. As shown in Fig. 1A, slowly-migrating heterogeneousbands appeared as early as 5 min after irradiation, peaked around60 min, and declined thereafter. The observed mobility shift ofXPC upon UV irradiation was large and heterogeneous, raisingthe possibility that ubiquitylation might be involved in this band

shift. To demonstrate the ubiquitylation of XPC directly, we firstestablished a stable transformant of the XPC-deficient cell line thatexpresses FLAG-tagged XPC at physiological levels. Then hemag-glutinin (HA)-tagged ubiquitin was transiently overexpressed inthis transformant and FLAG-XPC was immunoprecipitated. Whenthe precipitated samples were subjected to immunoblotting withanti-HA antibody, ubiquitylated FLAG-XPC was detected, par-ticularly when the transfected cells were UV irradiated before theextract preparation (Fig. 1B, lane 8). From these data, we concludethat UV irradiation induces the ubiquitylation of XPC. We foundthat even when de novo protein synthesis was inhibited, the totalamount of XPC was not significantly reduced after UV irradiation(Fig. 1A). Therefore, the ubiquitylation of XPC appears to bereversible and does not serve as a signal for degradation by theproteasome.

Figure 1. XPC is ubiquitylated in response to UV irradiation(A) Immunoblot analyses of the soluble lysates from normal human fibroblast cells irradiatedwith UV and incubated for various times. (B) Immunoblot analyses of the soluble lysates fromXP-V cells expressing FLAG-XPC and HA-ubiquitin. The arrows indicate the authentic unmod-ified XPC bands, and the brackets indicate the modified forms of XPC.

48 ANNUAL REPORT OF OSAKA UNIVERSITY—Academic Achievement—2004-2005

UV-DDB is required for the modification of XPC and physicallyinteracts with XPC

To explore possible relationships between XPC modificationand the NER process, we examined the UV-induced shift in mol-ecular weight of XPC in human cell lines belonging to differentgenetic complementation groups of XP and CS. Among cell linesexamined, XP-E cell lines lacking UV-DDB did not induce theband shifting of XPC upon UV irradiation. To further investigatethe involvement of UV-DDB in XPC modification, we used V79Chinese hamster cell line that lacks UV-DDB activity. Whenhuman DDB2 (untagged or FLAG-tagged) was expressed in V79cells, the XPC modification upon UV irradiation was observed(Fig. 2A). The above findings indicating some functional interac-tion between XPC and UV-DDB prompted us to examine whetherthey interacted physically as well. In fact, we found with purifiedproteins that XPC physically interacts with UV-DDB dependingon DDB2 (Fig. 2B).

The UV-DDB-E3 complex catalyzes the ubiquitylation of XPC invitro

To test whether the UV-DDB-E3 complex is responsible for theubiquitylation of XPC, cell-free ubiquitylation assays were carriedout. When XPC-HR23B-His was incubated with UV-DDB-E3complex in the presence of E1, E2 (UbcH5a), ubiquitin, and ATP,a shift in the molecular weight of XPC was detected (Fig. 3). Theshift appeared to depend on each of the protein components, andthe use of GST-tagged ubiquitin instead of normal ubiquitin result-ed in an altered pattern of the shifted XPC bands, strongly sug-gesting that the observed band shift is due to ubiquitin conjugation.

Furthermore, when a lysine-less ubiquitin was employed in the reac-tion, the band shift was significantly reduced, suggesting that theslow mobility species shown in the complete reaction were the resultof polyubiquitin chain formation. In the same reactions, DDB2 andcullin 4A were found to be ubiquitylated extensively as well, regard-less of the presence or absence of XPC-HR23B-His (Fig. 3).

Ubiquitylation Alters the DNA-binding Properties of XPC and UV-DDB

To obtain insights into roles the ubiquitylation plays in NER,we examined the DNA-binding properties of ubiquitylated XPCand UV-DDB. For this purpose, synthetic DNA containing UVlesion (CPD or 6-4PP) or the undamaged sequence as a controlwas immobilized on paramagnetic beads. Cell-free ubiquitylationreactions were carried out in the presence of one of these DNAbeads. After unbound fractions were saved and the beads were washed,the proteins retained by the DNA were subjected to immunoblot-ting for detection of XPC as well as subunits of the UV-DDB-E3complex (Fig. 4). Under these conditions, both DDB1 and DDB2were bound to the 6-4PP beads in a nearly quantitative manner,while only a part of XPC was detected in the bound fractions. Intrigu-ingly, when all factors required for ubiquitylation were present,only ubiquitylated XPC was detected in the DNA-bound fractions,although a significant part of XPC still remained unmodified inthe unbound fraction. In these reactions, DDB2 was extensivelyubiquitylated and very little DDB1 or DDB2 was detected in thebound fractions with any of the DNA beads. Thus, the ubiquityla-tion of UV-DDB abolished its DNA-binding activity almost com-

Figure 2. Functional and physical interactions of XPC with UV-DDB(A) UV-induced XPC modification depends on the presence of functional UV-DDB. (B) XPCphysically interacts with UV-DDB in purified protein system.

Figure 3. Cell-free ubiquitylation of XPC mediated by the UV-DDB-E3 complexCell-free ubiquitylation reactions were performed using the indicated sets of purified proteincomponents. The arrows indicate the unmodified form of each protein.

49

Osaka University 100 Papers : 10 Selected Papers

ANNUAL REPORT OF OSAKA UNIVERSITY—Academic Achievement—2004-2005

Figure 5. Model for UV-induced UV-DDB-dependent ubiquitylation of XPCUpon UV irradiation, UV-DDB translocates onto the damaged chromatin by binding to lesions.Its dissociation from the CSN and the neddylation of cullin 4A activates E3, and XPC, DDB2,and cullin 4A are ubiquitylated by the activated UV-DDB-E3 at the lesion site. Polyubiquity-lated UV-DDB loses its damaged DNA binding activity, whereas the DNA binding of XPC ispotentiated by its ubiquitylation. This results in the displacement of UV-DDB by XPC on thelesion.

References

1. Friedberg, E. C., How nucleotide excision repair protects against cancer., NatRev Cancer 1, 22-33 (2001).

2. Hoeijmakers, J. H. J., Genome maintenance mechanisms for preventing can-cer., Nature 411, 366-374 (2001).

3. Masutani, C., Sugasawa, K., Yanagisawa, J., Sonoyama, T., Ui, M., Enomo-to, T., Takio, K., Tanaka, K., van der Spek, P.J., Bootsma, D., Hoeijmakers,J.H.J., & Hanaoka, F., Purification and cloning of a nucleotide excision repaircomplex involving the xeroderma pigmentosum group C protein and a humanhomolog of yeast RAD23., EMBO J 13, 1831-1843 (1994).

4. Araki, M., Masutani, C., Takemura, M., Uchida, A., Sugasawa, K., Kondoh,J., Ohkuma, Y., & Hanaoka, F., Centrosome protein centrin 2/caltractin 1 ispart of the xeroderma pigmentosum group C complex that initiates globalgenome nucleotide excision repair., J Biol Chem 276, 18665-18672 (2001).

5. Sugasawa, K., Okamoto, T., Shimizu, Y., Masutani, C., Iwai, S., & Hanaoka,F., A multistep damage recognition mechanism for global genomic nucleotideexcision repair., Genes Dev 15, 507-521 (2001).

6. Groisman, R., Polanowska, J., Kuraoka, I., Sawada, J., Saijo, M., Drapkin, R.,Kisselev, A.F., Tanaka, K., & Nakatani, Y., The ubiquitin ligase activity inthe DDB2 and CSA complexes is differentially regulated by the COP9 sig-nalosome in response to DNA damage., Cell 113, 357-367 (2003).

pletely. In contrast, UV-DDB still retained its DNA-binding activ-ity as well as specificity for the UV-lesions when methylated ubiq-uitin was included instead of normal ubiquitin, while DDB2 stillshowed significant band shifts. Reductive methylation of lysineresidues in ubiquitin blocks the elongation of polyubiquitin chains.These data indicate that the formation of polyubiquitin chains abovea certain length abrogate the damage-binding activity of UV-DDB,whereas polyubiquitylated XPC retains its ability to bind DNA.

Concluding remarks Based on our present results, we proposed that UV-DDB-depen-

dent polyubiquitylation is involved in the displacement of UV-DDB by XPC from 6-4PP (Fig. 5). One of the roles UV-inducedXPC ubiquitylation may play is that it reinforces the DNA bind-ing of XPC, which helps displace UV-DDB from the lesion (eventhough UV-DDB initially binds to the UV lesion more stronglythan XPC). These observations together appear to add quite novelinsights into the damage recognition mechanisms involved in DNArepair, as well as to the functions of ubiquitylation.

Figure 4. Ubiquitylation alters the DNA binding properties of XPC and UV-DDBXPC-HR23B was incubated with paramagnetic beads bearing undamaged DNA (N) or DNAcontaining CPD (C) or 6-4PP (6). UV-DDB-E3 and other factors required for ubiquitylationwere also included as indicated. The proteins that did not bind to DNA (unbound) and thoseretained by DNA (bound) were subjected to immunoblot analyses.