Sequence-specific interactions of UvrABC endonuclease with psoralen interstrand cross-links

Modeling the interactions of the nucleotide excision repair UvrA(2) dimer with DNA

The UvrA protein initiates the DNA damage recognition process by the bacterial nucleotide excision repair (NER) system. Recently, crystallographic structures of holo-UvrA(2) dimers from two different microorganisms have been released (Protein Data Bank entries 2r6f , 2vf7 , and 2vf8 ). However, the details of the DNA binding by UvrA(2) and other peculiarities involved in the damage recognition process remain unknown. We have undertaken a molecular modeling approach to appraise the possible modes of DNA-UvrA(2) interaction using molecular docking and short-scale guided molecular dynamics [continuum field, constrained, and/or unrestricted simulated annealing (SA)], taking into account the three-dimensional location of a series of mutation-identified UvrA residues implicated in DNA binding. The molecular docking was based on the assumptions that the UvrA(2) dimer is preformed prior to DNA binding and that no major protein conformational rearrangements, except moderate domain reorientations, are required for binding of undamaged DNA. As a first approximation, DNA was treated as a rigid ligand. From the electrostatic relief of the ventral surface of UvrA(2), we initially identified three, noncollinear DNA binding paths. Each of the three resulting nucleoprotein complexes (C1, C2, and C3) was analyzed separately, including calculation of binding energies, the number and type of interaction residues (including mutated ones), and the predominant mode of translational and rotational motion of specific protein domains after SA to ensure improved DNA binding. The UvrA(2) dimer can accommodate DNA in all three orientations, albeit with different binding strengths. One of the UvrA(2)-DNA complexes (C1) fulfilled most of the requirements (high interaction energy, proximity of DNA to mutated residues, etc.) expected for a natural, high-affinity DNA binding site. This nucleoprotein presents a structural organization that is designed to clamp and bend double-stranded DNA. We examined the binding site in more detail by docking DNAs of significantly different (AT- vs CG-enriched) sequences and by submitting the complexes to DNA-unrestricted SA. It was found that in a manner independent of the DNA sequence and applied MD protocols, UvrA(2) favors binding of a bent and unwound undamaged DNA, with a kink positioned in the proximity of the Zn3 hairpins, anticollinearly aligned at the bottom of the ventral protein surface. It is further hypothesized that the Zn3 modules play an essential role in the damage recognition process and that the apparent existence of a family of DNA binding sites might be biologically relevant. Our data should prove to be useful in rational (structure-based) mutation studies.

Repair of mitomycin C mono- and interstrand cross-linked DNA adducts by UvrABC: a new model

Mitomycin C induces both MC-mono-dG and cross-linked dG-adducts in vivo. Interstrand cross-linked (ICL) dG-MC-dG-DNA adducts can prevent strand separation. In Escherichia coli cells, UvrABC repairs ICL lesions that cause DNA bending. The mechanisms and consequences of NER of ICL dG-MC-dG lesions that do not induce DNA bending remain unclear. Using DNA fragments containing a MC-mono-dG or an ICL dG-MC-dG adduct, we found (i) UvrABC incises only at the strand containing MC-mono-dG adducts; (ii) UvrABC makes three types of incisions on an ICL dG-MC-dG adduct: type 1, a single 5′ incision on 1 strand and a 3′ incision on the other; type 2, dual incisions on 1 strand and a single incision on the other; and type 3, dual incisions on both strands; and (iii) the cutting kinetics of type 3 is significantly faster than type 1 and type 2, and all of 3 types of cutting result in producing DSB. We found that UvrA, UvrA + UvrB and UvrA + UvrB + UvrC bind to MC-modified DNA specifically, and we did not detect any UvrB- and UvrB + UvrC–DNA complexes. Our findings challenge the current UvrABC incision model. We propose that DSBs resulted from NER of ICL dG-MC-dG adducts contribute to MC antitumor activity and mutations.

Mammalian nucleotide excision repair proteins and interstrand crosslink repair

Although various schemes for interstrand crosslink (ICL) repair incorporate DNA recombination, replication, and double-strand break intermediate steps, action of the nucleotide excision repair (NER) system or some variation of it is a common feature of most models. In the bacterium Escherichia coli, the NER enzyme UvrABC can incise on either side of an ICL to unhook the crosslink, and can proceed via a subsequent recombination step. The relevance of NER to ICL repair in mammalian cells has been challenged. Of all NER mutants, it is clear that ERCC1 and XPF-defective cells show the most pronounced sensitivities to ICL-inducing agents, and defects in ICL repair. However, there is good evidence that cells defective in NER proteins including XPA and XPG are also more sensitive than normal to ICL-inducing agents. These results are summarized here, together with evidence for defective crosslink removal in NER-defective cells. Studies of incision at sites of ICL by cell extracts and purified proteins have been done, but these studies are not all consistent with one another and further research is required.

Differential survival of Escherichia coli uvrA, uvrB, and uvrC mutants to psoralen plus UV-A (PUVA): Evidence for uncoupled action of nucleotide excision repair to process DNA adducts

The nucleotide excision repair mechanism (NER) of Escherichia coli is responsible for the recognition and elimination of more than twenty different DNA lesions. Herein, we evaluated the in vivo role of NER in the repair of DNA adducts generated by psoralens (mono- or bi-functional) and UV-A light (PUVA) in E. coli. Cultures of wild-type E. coli K12 and mutants for uvrA, uvrB, uvrC or uvrAC genes were treated with PUVA and cell survival was determined. In parallel, kinetics of DNA repair was also evaluated by the comparison of DNA sedimentation profiles in all the strains after PUVA treatment. The uvrB mutant was more sensitive to PUVA treatment than all the other uvr mutant strains. Wild-type strain, and uvrA and uvrC mutants were able to repair PUVA-induced lesions, as seen by DNA sedimentation profiles, while the uvrB mutant was unable to repair the lesions. In addition, a quadruple fpg nth xth nfo mutant was unable to nick PUVA-treated DNA when the crude cell-free extract was used to perform plasmid nicking. These data suggest that DNA repair of PUVA-induced lesions may require base excision repair functions, despite proficient UvrABC activity. These results point to a specific role for UvrB protein in the repair of psoralen adducts, which appear to be independent of UvrA or UvrC proteins, as described for the classical UvrABC endonuclease mechanism.

Efficient processing of TFO-directed psoralen DNA interstrand crosslinks by the UvrABC nuclease

Photoreactive psoralens can form interstrand crosslinks (ICLs) in double-stranded DNA. In eubacteria, the endonuclease UvrABC plays a key role in processing psoralen ICLs. Psoralen-modified triplex-forming oligonucleotides (TFOs) can be used to direct ICLs to specific genomic sites. Previous studies of pyrimidine-rich methoxypsoralen–modified TFOs indicated that the TFO inhibits cleavage by UvrABC. Because different chemistries may alter the processing of TFO-directed ICLs, we investigated the effect of another type of triplex formed by purine-rich TFOs on the processing of 4′-(hydroxymethyl)-4,5′,8-trimethylpsoralen (HMT) ICLs by the UvrABC nuclease. Using an HMT-modified TFO to direct ICLs to a specific site, we found that UvrABC made incisions on the purine-rich strand of the duplex ∼3 bases from the 3′-side and ∼9 bases from the 5′-side of the ICL, within the TFO-binding region. In contrast to previous reports, the UvrABC nuclease cleaved the TFO-directed psoralen ICL with a greater efficiency than that of the psoralen ICL alone. Furthermore, the TFO was dissociated from its duplex binding site by UvrA and UvrB. As mutagenesis by TFO-directed ICLs requires nucleotide excision repair, the efficient processing of these lesions supports the use of triplex technology to direct DNA damage for genome modification.

'Close-fitting sleeves': DNA damage recognition by the UvrABC nuclease system

DNA damage recognition represents a long-standing problem in the field of protein-DNA interactions. This article reviews our current knowledge of how damage recognition is achieved in bacterial nucleotide excision repair through the concerted action of the UvrA, UvrB, and UvrC proteins.

New insights on how nucleotide excision repair could remove DNA adducts induced by chemotherapeutic agents and psoralens plus UV-A (PUVA) in Escherichia coli cells

Chemotherapeutic agents such as mitomycin C or nitrogen mustards induce DNA inter-strand cross-links (ICL) and are highly toxic, thus constituting an useful tool to treat some human degenerative diseases, such as cancer. Additionally, psoralens plus UV-A (PUVA), which also induce ICL, find use in treatment of patients afflicted with psoriasis and vitiligo. The repair of DNA ICL generated by different molecules involves a number of multi-step DNA repair pathways. In bacteria, as in eukaryotic cells, if DNA ICL are not tolerated or repaired via nucleotide excision repair (NER), homologous recombination or translesion synthesis pathways, these DNA lesions may lead to mutations and cell death. Herein, we bring new insights to the role of Escherichia coli nucleotide excision repair genes uvrA, uvrB and uvrC in the repair of DNA damage induced by some chemotherapeutic agents and psoralen derivatives plus UV-A. These new observations point to a novel role for the UvrB protein, independent of its previously described role in the Uvr(A)BC complex, which could be specific for repair of monoadducts, intra-strand biadducts and/or ICL.

Reduced sulfhydryls maintain specific incision of BPDE-DNA adducts by recombinant thermoresistant Bacillus caldotenax UvrABC endonuclease

Prokaryotic DNA repair nucleases are useful reagents for detecting DNA lesions. Escherichia coli UvrABC endonuclease can incise DNA containing UV photoproducts and bulky chemical adducts. The limited stability of the E. coli UvrABC subunits leads to difficulty in estimating incision efficiency and quantitative adduct detection. To develop a more stable enzyme with greater utility for the detection of DNA adducts, thermoresistant UvrABC endonuclease was cloned from the eubacterium Bacillus caldotenax (Bca) and individual recombinant protein subunits were overexpressed in and purified from E. coli. Here, we show that Bca UvrC that had lost activity or specificity could be restored by dialysis against buffer containing 500 mM KCl and 20mM dithiothreitol. Our data indicate that UvrC solubility depended on high salt concentrations and UvrC nuclease activity and the specificity of incisions depended on the presence of reduced sulfhydryls. Optimal conditions for BCA UvrABC-specific cleavage of plasmid DNAs treated with [3H](+)-7R,8S-dihydroxy-9S,10R-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE) (1-5 lesions/plasmid) were developed. Preincubation of substrates with UvrA and UvrB enhanced incision efficiency on damaged substrates and decreased non-specific nuclease activity on undamaged substrates. Under optimal conditions for damaged plasmid incision, approximately 70% of adducts were incised in 1 nM plasmid DNA (2 BPDE adducts/5.4 kbp plasmid) with UvrA at 2.5 nM, UvrB at 62.5 nM, and UvrC at 25 nM. These results demonstrate the potential usefulness of the Bca UvrABC for monitoring the distribution of chemical carcinogen-induced lesions in DNA.

Differential incision of bulky carcinogen-DNA adducts by the UvrABC nuclease: comparison of incision rates and the interactions of Uvr subunits with lesions of different structures

The UvrABC nuclease system from Escherichia coli removes DNA damages induced by a wide range of chemical carcinogens with variable efficiencies. The interactions with UvrABC proteins of the following three lesions site-specifically positioned in DNA, and of known conformations, were investigated: (i) adducts derived from the binding of the (-)-(7S,8R,9R,10S) enantiomer of 7,8-dihydroxy-9, 10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene [(-)-anti-BPDE] by cis-covalent addition to N(2)-2'-deoxyguanosine [(-)-cis-anti-BP-N(2)-dG], (ii) an adduct derived from the binding of the (+)-(1R,2S,3S,4R) enantiomer of 1,2-dihydroxy-3,4-epoxy-1,2,3, 4-tetrahydro-5-methylchrysene [(+)-anti-5-MeCDE] by trans addition to N(2)-2'-deoxyguanosine [(+)-trans-anti-MC-N(2)-dG], and (iii) a C8-2'-deoxyguanosine adduct (C8-AP-dG) formed by reductively activated 1-nitropyrene (1-NP). The influence of these three different adducts on UvrA binding affinities, formation of UvrB-DNA complexes by quantitative gel mobility shift analyses, and the rates of UvrABC incision were investigated. The binding affinities of UvrA varied among the three adducts. UvrA bound to the DNA adduct (+)-trans-anti-MC-N(2)-dG with the highest affinity (K(d) = 17 +/- 2 nM) and to the DNA containing C8-AP-dG with the least affinity (K(d) = 28 +/- 1 nM). The extent of complex formation with UvrB was also the lowest with the C8-AP-dG adduct. 5' Incisions occurred at the eighth phosphate from the modified guanine. The major 3' incision site corresponded to the fifth phosphodiester bond for all three adducts. However, additional 3' incisions were observed at the fourth and sixth phosphates in the case of the C8-AP-dG adduct, whereas in the case of the (-)-cis-anti-BP-N(2)-dG and (+)-trans-anti-MC-N(2)-dG lesions additional 3' cleavage occurred at the sixth and seventh phosphodiester bonds. Both the initial rate and the extent of 5' and 3' incisions revealed that C8-AP-dG was repaired less efficiently in comparison to the (-)-cis-anti-BP-N(2)-dG and (+)-trans-anti-MC-N(2)-dG containing DNA adducts. Our study showed that UvrA recognizes conformational changes induced by structurally different lesions and that in certain cases the binding affinities of UvrA and UvrB can be correlated with the incision rates. The size of the bubble formed around the damaged site with mismatched bases also appears to influence the incision rates. A particularly noteworthy finding in this study is that UvrABC repair of a substrate with no base opposite C8-AP-dG was quite inefficient as compared to the same adduct with a C opposite it. These findings are discussed in terms of the available NMR solution structures.

Triple helix-directed psoralen crosslinks are recognized by Uvr(A)BC excinuclease

Pyrimidine oligonucleotides bind to the major groove of an oligopyrimidine-oligopurine DNA sequence by triple helix formation. A 14-mer oligopyrimidine 3'-psoralen-conjugate (P) and a doubly modified 5'-acridine/3'-psoralen-oligonucleotide (PA) were photo-crosslinked to their target site. The crosslinked complexes were tested regarding their sensitivity to Uvr(A)BC excinuclease/DNA complex formation and excision, and compared to free psoralen crosslinked to the same site (M). An electrophoretic mobility-shift assay showed that the crosslinked triple-helix did not hamper formation of the (A)2B complex under conditions where the third strand was bound to its target. In vitro excision experiments performed on damaged DNA fragments containing crosslinked 5-methoxypsoralen (M-target) confirmed that the psoralen photoadduct was recognized by Uvr(A)BC and that excision occurred at the crosslinked site. The major cleavage reaction took place on the 5'-side of oligopurine strand. The excision was less efficient on the 5'-side of the pyrimidine strand. The 3'-side incision either on the purine or pyrimidine strand was even weaker. With optimal Uvr(A) concentrations, it was observed that the incision reaction on (P)- and (PA)-modified targets was clearly inhibited compared to the (M)-modified target, reflecting an effect of the oligonucleotide on the recognition/excision process. These results demonstrate that a triple helix is efficient in promoting inhibition of Uvr(A)BC excision nuclease activity. These results could account for divergent findings concerning the effects of triple helix-forming oligonucleotides on repair systems and open new perspectives to study DNA repair processes through the use of bi-substituted triple helix-forming oligonucleotides.

Dynamics in psoralen-damaged DNA by 1H-detected natural abundance 13C NMR spectroscopy

The dynamics of the DNA oligomer d(GCGTACGC)2 and the 4'-(hydroxymethyl)-4,5',8-trimethylpsoralen-DNA furanside monoadduct (MAf) of this oligomer have been determined from NMR relaxation parameters. Longitudinal and transverse 13C relaxation rates and heteronuclear NOE relaxation data have been measured at natural abundance and have been analyzed in the context of the Lipari and Szabo model-free formalism. The generalized order parameters for methine carbons in the octamer sequence d(GCGTACGC)2 (UM) are relatively and uniformly high for the entire molecule. The generalized order parameters for methine carbons in the MAf are significantly lower for the deoxyribose bearing the damaged thymidine base and for the bases flanking the lesion on the undamaged strand, indicating additional conformational flexibility due to the lesion. The order parameters for the bases on the damaged strand flanking the lesion remain high. Analysis of the relaxation data indicates substantial chemical exchange for the adenosine residues in the UM TpA site, and this chemical exchange is quenched upon MAf formation. These data are discussed in terms of a model for DNA damage recognition by the nucleotide excision repair system.

Rate of incision of N-acetyl-2-aminofluorene and N-2-aminofluorene adducts by UvrABC nuclease is adduct- and sequence-specific: comparison of the rates of UvrABC nuclease incision and protein-DNA complex formation

The UvrABC nuclease, the nucleotide excision repair complex from Escherichia coli, is able to incise a variety of types of DNA damage and the repair efficiency of this enzyme complex appears to be influenced by the structure of the damage and the sequence context within which the damage is positioned. In order to better establish these relationships, we have constructed two DNA sequences each containing a site-specifically positioned N-2-aminofluorene (AF) or N-acetyl-2-aminofluorene (AAF) adduct and have determined both the kinetics of UvrABC nuclease incision and the kinetics of UvrABC nuclease-substrate complex formation. It is well established that these two adducts induce very different structures in the DNA and that these structures also depend on the sequence context. We have found that the rate of incision of both AAF- and AF-DNA adducts is significantly faster when they are positioned in the mutation hotspot NarI sequence (5-GGCG*CC-3') than when located in a normal or non-NarI sequence (5'-GATG*ATA-3') and that the rate of incision for AAF-DNA adducts is faster that for AF adducts in both sequences. Most siginificantly, we find that the rate of UvrB and UvrBC-substrate complex formation correlates with the rate of UvrABC nuclease incision.

The Escherichia coli DNA repair protein UvrA can re-associate with the UvrB: aflatoxin B1-DNA complex in vitro

The UvrA and UvrB proteins form part of the UvrABc endonuclease, which is responsible for nucleotide excision repair in Escherichia coli. Using a mobility shift gel assay we have studied the binding of UvrA dimer, UvrB monomer and UvA(2)B trimer complexes with 40, 50 and 136 bp (32)P-end-labelled DNA fragments adducted with aflatoxin B(1). UvrA was shown to re-associate with adduct specific UvrB: DNA complexes, a phenomenon which could be reversed by the addition of 500 mM potassium chloride or anti-UvrA anti-sera. Re-association was shown to be UvrA concentration dependent. Re-association of UvrA(2)B to the UvrB:DNA complex was not seen. We have also shown that the UvrB:DNA complex, in the case of aflatoxin B(1), is extremely stable with a half-life excess of 400 min and that fragment termini are not a specific substrate for UvrA binding.

Sequence specificity of DNA-psoralen photoproduct formation in supercoiled plasmid DNA (pUC19)

Supercoiled pUC19 DNA, photoreacted with psoralen derivatives (xanthotoxin (8-methoxypsoralen, 8-MOP), 4,5',8-trimethylpsoralen (TMP) and angelicin), influences the enzymatic activity of restriction enzymes in a different manner, although all the enzymes employed contain, within their recognition sites, suitable nucleic acid bases for photoproduct formation. The activity of the enzymes is strongly influenced by the photomodification of thymine residues within their recognition sites. 5'-TpA sequences favour intercalation as an essential prerequisite of the photoreaction, while 5'ApT sequences do not. This, in turn, influences photoproduct formation and the inhibition of the action of the restriction enzymes KpnI, SspI, DraI and RsaI, but not EcoRI and BamHI. The inhibitory effect is independent of the number of cleavage sites and also of whether monoaddition products or crosslinks are formed. Psoralen intercalation alone does not affect the activity of the restriction enzymes used.

The reactivity of 4,5',8-trimethylpsoralen with oligonucleotides containing AT sites

Pyrimidine bases of duplex DNA, of appropriate sequence context, are photoreactive toward 4,5',8-trimethylpsoralen in the presence of long-wavelength UV light. It is generally believed that a 5'-AT site is less photoreactive with psoralen than a 5'-TA site. We have compared the reactivities of these two sites using oligonucleotide duplexes of different sequence context and found that 5'-TA and 5'-AT sites are equally reactive in certain sequences. The presence of alternating pyrimidine and purine (5'-PyATPu-3') bases in oligonucleotide duplexes optimizes the reactivity of 4,5',8-trimethylpsoralen in the 5'-AT sites.