CDK8/19 Mediator kinases potentiate induction of transcription by NFκB.

The nuclear factor-κB (NFκB) family of transcription factors has been implicated in inflammatory disorders, viral infections, and cancer. Most of the drugs that inhibit NFκB show significant side effects, possibly due to sustained NFκB suppression. Drugs affecting induced, but not basal, NFκB activity may have the potential to provide therapeutic benefit without associated toxicity. NFκB activation by stress-inducible cell cycle inhibitor p21 was shown to be mediated by a p21-stimulated transcription-regulating kinase CDK8. CDK8 and its paralog CDK19, associated with the transcriptional Mediator complex, act as coregulators of several transcription factors implicated in cancer; CDK8/19 inhibitors are entering clinical development. Here we show that CDK8/19 inhibition by different small-molecule kinase inhibitors or shRNAs suppresses the elongation of NFκB-induced transcription when such transcription is activated by p21-independent canonical inducers, such as TNFα. On NFκB activation, CDK8/19 are corecruited with NFκB to the promoters of the responsive genes. Inhibition of CDK8/19 kinase activity suppresses the RNA polymerase II C-terminal domain phosphorylation required for transcriptional elongation, in a gene-specific manner. Genes coregulated by CDK8/19 and NFκB include IL8, CXCL1, and CXCL2, which encode tumor-promoting proinflammatory cytokines. Although it suppressed newly induced NFκB-driven transcription, CDK8/19 inhibition in most cases had no effect on the basal expression of NFκB-regulated genes or promoters; the same selective regulation of newly induced transcription was observed with other transcription signals potentiated by CDK8/19. This selective role of CDK8/19 identifies these kinases as mediators of transcriptional reprogramming, a key aspect of development and differentiation as well as pathological processes.

Mammalian PER2 regulates AKT activation and DNA damage response.

PER2 is a key mammalian circadian clock protein. It also has a tumor suppressive function. Down regulation of PER2 in the cultured cancer cells accelerates cell proliferation, while overexpression of PER2 inhibits cell growth and induces apoptosis. The Per2 mutant mice have a cancer prone phenotype and an altered DNA damage response. Here we report that PER2 regulates AKT activity. Cells with down-regulated PER2 expression have prolonged high levels of AKT T308 phosphorylation after growth factor stimulation or DNA damage. PER2 down-regulation delays DNA damage induced Chk2 activation and overrides DNA damage induced apoptosis and cell cycle arrest.

Induction of intrachromosomal homologous recombination in human cells by raltitrexed, an inhibitor of thymidylate synthase.

Thymidylate deprivation brings about "thymineless death" in prokaryotes and eukaryotes. Although the precise mechanism for thymineless death has remained elusive, inhibition of the enzyme thymidylate synthase (TS), which catalyzes the de novo synthesis of TMP, has served for many years as a basis for chemotherapeutic strategies. Numerous studies have identified a variety of cellular responses to thymidylate deprivation, including disruption of DNA replication and induction of DNA breaks. Since stalled or collapsed replication forks and strand breaks are generally viewed as being recombinogenic, it is not surprising that a link has been demonstrated between recombination induction and thymidylate deprivation in bacteria and lower eukaryotes. A similar connection between recombination and TS inhibition has been suggested by studies done in mammalian cells, but the relationship between recombination and TS inhibition in mammalian cells had not been demonstrated rigorously. To gain insight into the mechanism of thymineless death in mammalian cells, in this work we undertook a direct investigation of recombination in human cells treated with raltitrexed (RTX), a folate analog that is a specific inhibitor of TS. Using a model system to study intrachromosomal homologous recombination in cultured fibroblasts, we provide definitive evidence that treatment with RTX can stimulate accurate recombination events in human cells. Gene conversions not associated with crossovers were specifically enhanced several-fold by RTX. Additional experiments demonstrated that recombination events provoked by a double-strand break (DSB) were not impacted by treatment with RTX, nor was error-prone DSB repair via nonhomologous end-joining. Our work provides evidence that thymineless death in human cells is not mediated by corruption of DSB repair processes and suggests that an increase in chromosomal recombination may be an important element of cellular responses leading to thymineless death.

Phosphorylation of nucleotide excision repair factor xeroderma pigmentosum group A by ataxia telangiectasia mutated and Rad3-related-dependent checkpoint pathway promotes cell survival in response to UV irradiation.

DNA damage triggers complex cellular responses in eukaryotic cells, including initiation of DNA repair and activation of cell cycle checkpoints. In addition to inducing cell cycle arrest, checkpoint also has been suggested to modulate a variety of other cellular processes in response to DNA damage. In this study, we present evidence showing that the cellular function of xeroderma pigmentosum group A (XPA), a major nucleotide excision repair (NER) factor, could be modulated by checkpoint kinase ataxia-telangiectasia mutated and Rad3-related (ATR) in response to UV irradiation. We observed the apparent interaction and colocalization of XPA with ATR in response to UV irradiation. We showed that XPA was a substrate for in vitro phosphorylation by phosphatidylinositol-3-kinase-related kinase family kinases whereas in cells XPA was phosphorylated in an ATR-dependent manner and stimulated by UV irradiation. The Ser196 of XPA was identified as a biologically significant residue to be phosphorylated in vivo. The XPA-deficient cells complemented with XPA-S196A mutant, in which Ser196 was substituted with an alanine, displayed significantly higher UV sensitivity compared with the XPA cells complemented with wild-type XPA. Moreover, substitution of Ser196 with aspartic acid for mimicking the phosphorylation of XPA increased the cell survival to UV irradiation. Taken together, our results revealed a potential physical and functional link between NER and the ATR-dependent checkpoint pathway in human cells and suggested that the ATR checkpoint pathway could modulate the cellular activity of NER through phosphorylation of XPA at Ser196 on UV irradiation.

Specific and efficient binding of xeroderma pigmentosum complementation group A to double-strand/single-strand DNA junctions with 3'- and/or 5'-ssDNA branches.

Human XPA is an important DNA damage recognition protein in nucleotide excision repair (NER). We previously observed that XPA binds to the DNA lesion as a homodimer [Liu, Y., Liu, Y., Yang, Z., Utzat, C., Wang, G., Basu, A. K., and Zou, Y. (2005) Biochemistry 44, 7361-7368]. Herein we report that XPA recognized undamaged DNA double-strand/single-strand (ds-ssDNA) junctions containing ssDNA branches with binding affinity (Kd = 49.1 +/- 5.1 nM) much higher than its ability to bind to DNA damage. The recognized DNA junction structures include the Y-shape junction (with both 3'- and 5'-ssDNA branches), 3'-overhang junction (with a 3'-ssDNA branch), and 5'-overhang junction (with a 5'-ssDNA branch). Using gel filtration chromatography and gel mobility shift assays, we showed that the highly efficient binding appeared to be carried out by the XPA monomer and that the binding was largely independent of RPA. Furthermore, XPA efficiently bound to six-nucleotide mismatched DNA bubble substrates with or without DNA adducts including C8 guanine adducts of AF, AAF, and AP and the T[6,4]T photoproducts. Using a set of defined DNA substrates with varying degrees of DNA bending, we also found that the XPC-HR23B complex recognized DNA bending, whereas neither XPA nor the XPA-RPA complex could bind to bent DNA. We propose that, besides DNA damage recognition, XPA may also play a novel role in stabilizing, via its high affinity to ds-ssDNA junctions, the DNA strand opening surrounding the lesion for stable formation of preincision NER intermediates. Our results provide a plausible mechanistic interpretation for the indispensable requirement of XPA for both global genome and transcription-coupled repairs. Since ds-ssDNA junctions are common intermediates in many DNA metabolic pathways, the additional potential role of XPA in cellular processes is discussed.

Preferential localization of hyperphosphorylated replication protein A to double-strand break repair and checkpoint complexes upon DNA damage.

RPA (replication protein A) is an essential factor for DNA DSB (double-strand break) repair and cell cycle checkpoint activation. The 32 kDa subunit of RPA undergoes hyperphosphorylation in response to cellular genotoxic insults. However, the potential involvement of hyperphosphorylated RPA in DSB repair and checkpoint activation remains unclear. Using co-immunoprecipitation assays, we showed that cellular interaction of RPA with two DSB repair factors, Rad51 and Rad52, was predominantly mediated by the hyperphosphorylated species of RPA in cells after UV and camptothecin treatment. Moreover, Rad51 and Rad52 displayed higher affinity for the hyperphosphorylated RPA than native RPA in an in vitro binding assay. Checkpoint kinase ATR (ataxia telangiectasia mutated and Rad3-related) also interacted more efficiently with the hyperphosphorylated RPA than with native RPA following DNA damage. Consistently, immunofluorescence microscopy demonstrated that the hyperphosphorylated RPA was able to co-localize with Rad52 and ATR to form significant nuclear foci in cells. Our results suggest that hyperphosphorylated RPA is preferentially localized to DSB repair and the DNA damage checkpoint complexes in response to DNA damage.

Interactions of human replication protein A with single-stranded DNA adducts.

Human RPA (replication protein A), a single-stranded DNA-binding protein, is required for many cellular pathways including DNA repair, recombination and replication. However, the role of RPA in nucleotide excision repair remains elusive. In the present study, we have systematically examined the binding of RPA to a battery of well-defined ssDNA (single-stranded DNA) substrates using fluorescence spectroscopy. These substrates contain adducts of (6-4) photoproducts, N-acetyl-2-aminofluorene-, 1-aminopyrene-, BPDE (benzo[a]pyrene diol epoxide)- and fluorescein that are different in many aspects such as molecular structure and size, DNA disruption mode (e.g. base stacking or non-stacking), as well as chemical properties. Our results showed that RPA has a lower binding affinity for damaged ssDNA than for non-damaged ssDNA and that the affinity of RPA for damaged ssDNA depends on the type of adduct. Interestingly, the bulkier lesions have a greater effect. With a fluorescent base-stacking bulky adduct, (+)-cis-anti-BPDE-dG, we demonstrated that, on binding of RPA, the fluorescence of BPDE-ssDNA was significantly enhanced by up to 8-9-fold. This indicated that the stacking between the BPDE adduct and its neighbouring ssDNA bases had been disrupted and there was a lack of substantial direct contacts between the protein residues and the lesion itself. For RPA interaction with short damaged ssDNA, we propose that, on RPA binding, the modified base of ssDNA is looped out from the surface of the protein, permitting proper contacts of RPA with the remaining unmodified bases.

Expression and regulation of RAD51 mediate cellular responses to chemotherapeutics.

There is evidence that RAD51 expression associates with resistance to commonly used chemotherapeutics. Our previous work demonstrated that inhibitors of thymidylate synthase (TS) induced RAD51-dependent homologous recombination (HR), and depleting the RAD51 recombinase sensitized cells to TS inhibitors. In this study, the consequences of RAD51 over-expression were studied. Over-expression of wild-type RAD51 (∼6-fold above endogenous RAD51) conferred resistance to TS inhibitors. In contrast, over-expression of a mutant RAD51 (T309A) that is incapable of being phosphorylated rendered cells more chemosensitive. Moreover, over-expression of the T309A mutant acted in a dominant negative manner over endogenous RAD51 by causing the reduced localization of RAD51 foci following treatment with TS inhibitors. To measure the effect of mutant RAD51 on the cellular response to other DNA damaging chemotherapeutics, the topoisomerase poison etoposide was utilized. Cells over-expressing wild-type RAD51 showed reduced DNA strand breaks, while cells over-expressing the mutant RAD51 showed more than twice as many strand breaks, suggesting that the mutant RAD51 was actively inhibiting strand break resolution. To directly demonstrate an effect on HR, wild-type RAD51 and T309A mutant RAD51 were transiently expressed in HeLa cells that contained an HR reporter construct. HR events provoked by DNA breaks induced by the I-SceI endonuclease increased in cells expressing wild-type RAD51 and decreased in cells expressing the T309A mutant. Collectively, the data suggest that interference with the activation of RAD51-mediated HR represents a potentially useful anticancer target for combination therapies.

Targeting subcellular localization through the polo-box domain: non-ATP competitive inhibitors recapitulate a PLK1 phenotype.

The polo-box domain (PBD) has critical roles in the mitotic functions of polo-like kinase 1 (PLK1). The replacement with partial ligand alternative through computational enrichment (REPLACE) strategy to develop inhibitors of protein-protein interactions has identified alternatives for the N-terminal tripeptide of a Cdc25C substrate. In addition, a peptide structure-activity relationship described key determinants and novel information useful for drug design. Fragment-ligated inhibitory peptides (FLIP) were generated with comparable affinity to peptide PBD inhibitors and possessed antiproliferative phenotypes in cells consistent with the observed decrease in PLK1 centrosomal localization. These FLIPs showed evidence of enhanced PLK1 inhibition in cells relative to peptides and induced monopolar and multipolar spindles, which stands in contrast to previously reported small-molecule PBD inhibitors that display phenotypes only partially representative of PLK1 knockdown. Progress obtained applying REPLACE validates this approach for identifying fragment alternatives for determinants of the Cdc25C-binding motif and extends its applicability of the strategy for discovering protein-protein interaction inhibitors. In addition, the described PBD inhibitors retain high specificity for PLK1 over PLK3 and therefore show promise as isotype selective, non-ATP competitive kinase inhibitors that provide new impetus for the development of PLK1-selective antitumor therapeutics.

DNA damage and homologous recombination signaling induced by thymidylate deprivation.

DNA damage is accepted as a consequence of thymidylate deprivation induced by chemotherapeutic inhibitors of thymidylate synthase (TS), but the types of damage and signaling responses remain incompletely understood. Thymidylate deprivation increases dUTP and uracil in DNA, which is removed by base excision repair (BER). Because BER requires a synthesis step, strand break intermediates presumably accumulate. Thymidylate deprivation also induces cell cycle arrest during replication. Homologous recombination (HR) is a means of repairing persistent BER intermediates and collapsed replication forks. There are also intimate links between HR and S-phase checkpoint pathways. In this study, the goals were to determine the involvement of HR-associated proteins and DNA damage signaling responses to thymidylate deprivation. When RAD51, which is a central component of HR, was depleted by siRNA cells were sensitized to raltitrexed (RTX), which specifically inhibits TS. To our knowledge, this is the first demonstration in mammalian cells that depletion of RAD51 causes sensitivity to thymidylate deprivation. Activation of DNA damage signaling responses was examined following treatment with RTX. Phosphorylation of replication protein A (RPA2 subunit) and formation of damage-induced foci were strikingly evident following IC(50) doses of RTX. Induction was much more striking following RTX treatment than with hydroxyurea, which is commonly used to inhibit replication. RTX treatment also induced foci of RAD51, gamma-H2AX, phospho-Chk1, and phospho-NBS1, although the extent of co-localization with RPA2 foci varied. Collectively, the results suggest that HR and S-phase checkpoint signaling processes are invoked by thymidylate deprivation and influence cellular resistance to thymidylate deprivation.

Structural characterization of human RPA sequential binding to single-stranded DNA using ssDNA as a molecular ruler.

Human replication protein A (RPA), a heterotrimer composed of RPA70, RPA32, and RPA14 subunits, contains four single-stranded DNA (ssDNA) binding domains (DBD): DBD-A, DBD-B, and DBD-C in RPA70 and DBD-D in RPA32. Although crystallographic or NMR structures of these DBDs and a trimerization core have been determined, the structure of the full length of RPA or the RPA-ssDNA complex remains unknown. In this article, we have examined the structural features of RPA interaction with ssDNA by fluorescence spectroscopy. Using a set of oligonucleotides (dT) with varying lengths as a molecular ruler and also as the substrate, we have determined at single-nucleotide resolution the relative positions of the ssDNA with interacting intrinsic tryptophans of RPA. Our results revealed that Trp528 in DBD-C and Trp107 in DBD-D contact ssDNA at the 16th and 24th nucleotides (nt) from the 5'-end of the substrate, respectively. Evaluation of the relative spatial arrangement of RPA domains in the RPA-ssDNA complex suggested that DBD-B and DBD-C are spaced by about 4 nt ( approximately 19 A) apart, whereas DBD-C and DBD-D are spaced by about 7 nt ( approximately 34 A). On the basis of these geometric constraints, a global structure model for the binding of the major RPA DBDs to ssDNA was proposed.

Recognition and incision of oxidative intrastrand cross-link lesions by UvrABC nuclease.

Nucleotide excision repair (NER) is a repair pathway that removes a variety of bulky DNA lesions in both prokaryotic and eukaryotic cells. The perturbation of DNA helix structure caused by the oxidative intrastrand lesions could render them good substrates for the NER pathway. Here we employed Escherichia coli NER enzymes, i.e., UvrA, UvrB, and UvrC, to examine the incision efficiency of duplex DNA carrying three different oxidative intrastrand cross-link lesions, that is, G[8-5]C, G[8-5m]mC, and G[8-5m]T, and two dithymine photoproducts, namely, the cis,syn-cyclobutane pyrimidine dimer (T[c,s]T) and the pyrimidine(6-4)pyrimidone product (T[6-4]T). Our results showed that T[6-4]T was the best substrate for UvrA binding, followed by G[8-5]C, G[8-5m]mC, and G[8-5m]T, and then by T[c,s]T. The efficiencies of the UvrABC incisions of these lesions were consistent with their UvrA binding affinities: the stronger the binding to UvrA, the higher the rate of incision. In addition, flanking DNA sequences appeared to have little effect on the binding affinity of UvrA for G[8-5]C as AG[8-5]CA was only slightly preferred over CG[8-5]CG. Consistently, these two sequences exhibited almost no difference in incision rates. Furthermore, we investigated the thermal stability of dodecameric duplexes containing G[8-5m]mC or G[8-5m]T, and our results revealed that these two lesions destabilized the duplex, due to an increase in the free energy for duplex formation at 37 degrees C, by approximately 5.4 and 3.6 kcal/mol, respectively. The destabilizations to the DNA helix caused by those lesions, for the most part, are correlated with the binding affinities of UvrA and incision rates of UvrABC. Taken together, the results from this study suggest that oxidative intrastrand lesions might be substrates for NER enzymes in vivo.

Recognition and incision of gamma-radiation-induced cross-linked guanine-thymine tandem lesion G[8,5-Me]T by UvrABC nuclease.

Nucleotide excision repair (NER) plays an important role in maintaining the integrity of DNA by removing various types of bulky or distorting DNA adducts in both prokaryotic and eukaryotic cells. In Escherichia coli, the excision repair proteins UvrA, UvrB, and UvrC recognize and incise the bulky DNA damages induced by UV light and chemical carcinogens. In this process, when a putative lesion in DNA is identified initially by UvrA, a subsequent strand opening is carried out by UvrB that not only ensures that the distortion is indeed due to a damaged nucleotide but also recognizes the chemical structure of the modified nucleotides with varying efficiencies. UvrB also recruits UvrC that catalyzes both the 3'- and the 5'-incisions. Herein, we examined the interaction of UvrABC with a DNA substrate containing a single G[8,5-Me]T cross-link and compared it with T[6,4]T (the 6-4 pyrimidine-pyrimidone photoproduct) and the C8 guanine adduct of N-acetyl-2-aminofluorene (AAF). The intrastrand vicinal cross-link G[8,5-Me]T containing a covalent bond between the C8 position of guanine and the 5-methyl carbon of the 3'-thymine is formed by X-radiation, while T[6,4]T is a vicinal cross-link induced by the UV light. We also selected the AAF adduct for comparison because it represents a highly distorting monoadduct containing a covalent linkage at the C8 position of guanine. The dissociation constants (K(d)) for UvrA protein binding to DNA substrates containing the G[8,5-Me]T, T[6,4]T, and AAF adducts, as determined by gel mobility shift assays, were 3.1 +/- 1.3, 2.8 +/- 0.9, and 8.2 +/- 1.9, respectively. Although UvrA had a considerably higher affinity for G[8,5-Me]T than for the AAF adduct, the G[8,5-Me]T intrastrand cross-link was incised by UvrABC much less efficiently than the T[6,4]T intrastrand cross-link and the AAF adduct. Similar incision results also were obtained with the DNA substrates containing the adducts in a six-nucleotide bubble, indicating that the inefficient incision of G[8,5-Me]T cross-link by UvrABC was probably due to the lack of efficient recognition of the adduct by UvrB at the second step of DNA damage recognition in the E. coli NER. Indeed, as compared to T[6,4]T and AAF substrates, which clearly showed UvrB-DNA complex formation, very little UvrB complex was detectable with the G[8,5-Me]T substrate. Our result suggests that G[8,5-Me]T intrastrand cross-link is more resistant to excision repair in comparison with the T[6,4]T and AAF adducts and thus will likely persist longer in E. coli cells.

Cooperative interaction of human XPA stabilizes and enhances specific binding of XPA to DNA damage.

Human xeroderma pigmentosum group A (XPA) is an essential protein for nucleotide excision repair (NER). We have previously reported that XPA forms a homodimer in the absence of DNA. However, what oligomeric forms of XPA are involved in DNA damage recognition and how the interaction occurs in terms of biochemical understanding remain unclear. Using the homogeneous XPA protein purified from baculovirus-infected Sf21 insect cells and the methods of gel mobility shift assays, gel filtration chromatography, and UV-cross-linking, we demonstrated that both monomeric and dimeric XPA bound to the DNA adduct of N-acetyl-2-aminofluorene (AAF), while showing little affinity for nondamaged DNA. The binding occurred in a sequential and protein concentration-dependent manner. At relatively low-protein concentrations, XPA formed a complex with DNA adduct as a monomer, while at the higher concentrations, an XPA dimer was involved in the specific binding. Results from fluorescence spectroscopic and competitive binding analyses indicated that the specific binding of XPA to the adduct was significantly facilitated and stabilized by the presence of the second XPA in a positive cooperative manner. This cooperative binding exhibited a Hill coefficient of 1.9 and the step binding constants of K(1) = 1.4 x 10(6) M(-)(1) and K(2) = 1.8 x 10(7) M(-)(1). When interaction of XPA and RPA with DNA was studied, even though binding of RPA-XPA complex to adducted DNA was observed, the presence of RPA had little effect on the overall binding efficiency. Our results suggest that the dominant form for XPA to efficiently bind to DNA damage is the XPA dimer. We hypothesized that the concentration-dependent formation of different types of XPA-damaged DNA complex may play a role in cellular regulation of XPA activity.

Dimerization of human XPA and formation of XPA2-RPA protein complex.

XPA plays an important role in the DNA damage recognition during human nucleotide excision repair. Here we report that the XPA is a homodimer either in the free state or as a complex with human RPA in solution under normal conditions. The human XPA protein purified from baculovirus-infected sf21 insect cells has a molecular mass of 36 317 Da, as determined by mass spectroscopy. However, the apparent molecular mass of XPA determined by the native gel filtration chromatography was about 71 kDa, suggesting that XPA is a dimer. This observation was supported by a native PFO-PAGE and fluorescence spectroscopy analysis. XPA formed a dimer (XPA2) in a broad range of XPA and NaCl concentrations, and the dimerization was not due to the disulfide bond formation. Furthermore, a titration analysis of the binding of XPA to the human RPA indicated that it was the XPA2 that formed the complex with RPA. Finally, the difference between the mass spectrometric and the calculated masses of XPA implies that the protein contains posttranslational modifications. Taken together, our data suggest that the dimerization of XPA may play an important role in the DNA damage recognition of nucleotide excision repair.

Effects of DNA adduct structure and sequence context on strand opening of repair intermediates and incision by UvrABC nuclease.

DNA damage recognition of nucleotide excision repair (NER) in Escherichia coli is achieved by at least two steps. In the first step, a helical distortion is recognized, which leads to a strand opening at the lesion site. The second step involves the recognition of the type of chemical modification in the single-stranded region of DNA during the processing of the lesions by UvrABC. In the current work, by comparing the efficiencies of UvrABC incision of several types of different DNA adducts, we show that the size and position of the strand opening are dependent on the type of DNA adducts. Optimal incision efficiency for the C8-guanine adducts of 2-aminofluorene (AF) and N-acetyl-2-aminofluorene (AAF) was observed in a bubble of three mismatched nucleotides, whereas the same for C8-guanine adduct of 1-nitropyrene and N(2)-guanine adducts of benzo[a]pyrene diol epoxide (BPDE) was noted in a bubble of six mismatched nucleotides. This suggests that the size of the aromatic ring system of the adduct might influence the extent and number of bases associated with the opened strand region catalyzed by UvrABC. We also showed that the incision efficiency of the AF or AAF adduct was affected by the neighboring DNA sequence context, which, in turn, was the result of differential binding of UvrA to the substrates. The sequence context effect on both incision and binding disappeared when a bubble structure of three bases was introduced at the adduct site. We therefore propose that these effects relate to the initial step of damage recognition of DNA structural distortion. The structure-function relationships in the recognition of the DNA lesions, based on our results, have been discussed.

DNA damage recognition of mutated forms of UvrB proteins in nucleotide excision repair.

The DNA repair protein UvrB plays an indispensable role in the stepwise and sequential damage recognition of nucleotide excision repair in Escherichia coli. Our previous studies suggested that UvrB is responsible for the chemical damage recognition only upon a strand opening mediated by UvrA. Difficulties were encountered in studying the direct interaction of UvrB with adducts due to the presence of UvrA. We report herein that a single point mutation of Y95W in which a tyrosine is replaced by a tryptophan results in an UvrB mutant that is capable of efficiently binding to structure-specific DNA adducts even in the absence of UvrA. This mutant is fully functional in the UvrABC incisions. The dissociation constant for the mutant-DNA adduct interaction was less than 100 nM at physiological temperatures as determined by fluorescence spectroscopy. In contrast, similar substitutions at other residues in the beta-hairpin with tryptophan or phenylalanine do not confer UvrB such binding ability. Homology modeling of the structure of E. coli UvrB shows that the aromatic ring of residue Y95 and only Y95 directly points into the DNA binding cleft. We have also examined UvrB recognition of both "normal" bulky BPDE-DNA and protein-cross-linked DNA (DPC) adducts and the roles of aromatic residues of the beta-hairpin in the recognition of these lesions. A mutation of Y92W resulted in an obvious decrease in the efficiency of UvrABC incisions of normal adducts, while the incision of the DPC adduct is dramatically increased. Our results suggest that Y92 may function differently with these two types of adducts, while the Y95 residue plays an unique role in stabilizing the interaction of UvrB with DNA damage, most likely by a hydrophobic stacking.