DNA sequence specificity of 4,5',8-trimethylpsoralen cross-linking. Effect of neighboring bases on cross-linking the 5'-TA dinucleotide

The DNA sequence specificity for 4,5',8-trimethylpsoralen cross-linking of DNA has been examined using chemically synthesized DNA fragments containing all possible pyrimidine and purine base pair combinations. We confirm our previous findings that the 5'-TA dinucleotide represents a preferred cross-link site. Other dinucleotides that form cross-links are 5'-AT much greater than 5'-TG greater than 5'-GT. Although 5'-TA represents a preferred cross-link site, the rate of cross-linking can vary 3-4-fold depending on the base composition flanking the cross-linkable 5'-TA dinucleotide. In some cases, changes in DNA sequence 3 base pairs (bp) away from 5'-TA resulted in significant changes in the rate of cross-linking. We also see evidence for a long-range sequence context effect on the rate of cross-linking. A 30-bp fragment cross-linked at a relative rate of 2.9 compared to the rate of cross-linking of a 20-bp fragment when cloned contiguously in plasmid DNA. When cross-linked as separate fragments, the 30-bp fragment cross-linked at a relative rate of 1.9 compared to the 20-bp fragment. The results suggest that the reactivity of DNA to psoralens, and perhaps other intercalating drugs, is dependent on the dinucleotide sequence, the bases flanking the dinucleotide, and the long-range sequence context of the DNA.     

Sequence specificity of psoralen photobinding to DNA: a quantitative approach.

The effects of different DNA sequences on the photoreaction of various furocoumarin derivatives was investigated from a quantitative point of view using a number of self-complementary oligonucleotides. These contained 5'-TA and 5'-AT residues, having various flanking sequences. The furocoumarins included classical bifunctional derivatives, such as 8-methoxy- and 5-methoxypsoralen, as well as monofunctional compounds, such as angelicin and benzopsoralen. Taking into an account the thermodynamic constant for noncovalent binding of each psoralen to each DNA sequence, the rate constants for the photobinding process to each fragment were evaluated. The extent of photoreaction is greatly affected by the DNA sequence examined. While sequences of the type 5'-(GTAC)n are quite reactive towards all furocoumarins, 5'-TATA exhibited a reduced rate of photobinding using monofunctional psoralens. In addition terminal 5'-TA groups were the least reactive with 5- and 8-methoxypsoralen, but not with angelicin or benzopsoralen. Also 5'-AT-containing fragments exhibited remarkably variable responses toward monofunctional or bifunctional psoralen derivatives. As a general trend the photoreactivity rate of the former is less sequence-sensitive, the ratio between maximum and minimum being less than 2 for the examined fragments. The same ratio is about 3.4 for 8-methoxypsoralen and 6.2 for 5-methoxypsoralen. This approach, in combination with footprinting studies, appears to be quite useful for a quantitative investigation of the process of covalent binding of psoralens to specific sites in DNA.     

Psoralen photofootprinting of protein-binding sites on DNA

Using a BAL31 exonuclease assay to determine the sites of 4,5',8-trimethylpsoralen photocrosslinking in DNA we have shown that 5'-TA sites which are accessible to psoralen DNA interstrand photocrosslinking in naked DNA become inaccessible when protein, in casu, lambda-repressor E. coli or RNA polymerase are bound at their recognition DNA sequences (OR1 operator or deo1 promoter, respectively). These results show that psoralens can be used as photofootprinting reagents to study specific protein-DNA interactions.     

Laser-induced photo-cross-linking of cisplatin-modified DNA to HMG-domain proteins.

Laser-induced photo-cross-linking was investigated for DNA, modified with cisplatin at specific sites, bound to structure-specific recognition domains of proteins in the high-mobility group (HMG) class. The efficiency of photo-cross-linking depends on the wavelength and power of the laser, the nature of the protein domain, and the oligodeoxyribonucleotide sequences flanking the platinated site. Introduction of 5-iodouridine at thymine sites of the oligodeoxyribonucleotide as an additional photoreactive group did not increase the photo-cross-linking yield. Formation of platinum-mediated DNA-DNA interstrand cross-linking observed previously upon irradiation with 302 nm light [Kane, S. A., and Lippard, S. J. (1996) Biochemistry 35, 2180-2188] was significantly reduced with laser irradiation. HMG1 domain B is superior to domain A for platinum-mediated photo-cross-linking, a result attributed to the different positioning of the proteins with respect to the platinum adduct and the greater ability of domain B to access photolabilized platinum in the major groove. Studies with proteins containing specifically mutated amino acids, and with DNA probes in which the sequences flanking the platinum cross-link site were varied, suggest that the most effective photo-cross-linking occurs for protein domains bound symmetrically and flexibly to cisplatin-modified DNA. The thermodynamic equilibrium between the protein-platinated DNA complex and its components, revealed in gel electrophoretic mobility shift assays (EMSAs), is significantly shifted to the right upon irreversible photo-cross-linking. Thus, only upon photo-cross-linking can the interaction of cisplatin-DNA 1,3-intrastrand d(GpTpG) or interstrand cross-links with HMG1 domain B protein be detected. Photo-cross-linking is thus an effective tool for investigating the interaction of cisplatin-modified DNA with damage-recognition proteins under heterogeneous conditions such those in cell extracts or living cells.     

The specific cleavage of DNA with ultrasound

Cleavages of double-stranded DNA fragments of known base pair sequence upon ultrasound irradiation at 22 and 44 kHz were studied by gel electrophoresis. The cleavage rate is found to be strongly dependent on the DNA fragment length, pH, temperature and ionic strength of the solution under study. The cleavage of double-stranded DNA occurs predominantly at sites containing alternating 5'-CpG-3' sequences. The breakage of phosphatediester bond takes place between C and G in such a way that phosphate group at the 5'-end of the guanine residue remain intact. The cleavage rate at a given DNA site is found to depend on base pair sequences at adjacent sites. Distinctly different cleavage patterns are observed when free DNA and DNA complexes with cys-diammine-Pt-bridged bis-netropsin were irradiated by ultrasound. The observed effect can be attributed to local DNA conformation changes induced upon complex formation between ligand and DNA.     

Chemical reaction of DNA with cross-linking agents: Influences of base permutation on chemical reactivity

Certain psoralens can photoreact with the pyrimidine bases of DNA and form interstrand cross-links. In this communication, theoretical mechanism of the photoreaction of DNA with psoralens in the presence of near-ultraviolet light (320–400 nm, u.v.-A) is diagrammatically and quantitatively presented.The number of the receptor binding sites, the theorized photoreaction sites, can be estimated from the base composition of DNA if the extent of the photoreactivity of cytosine and the possibly steric hindrance effects of DNA double helix in photoreactivity can be experimentally determined.The photoreactivity of DNA can be expressed by the number of psoralens photobound. It is predicted that the photoreactivity of DNA is less than, or at most equal to, the number of receptor binding sites depending on the number of cross-links formed during photoreaction. The number of cross-links formed is the consequence of the base permutation which determines the constitution of cross-linking sites, and of the manners of psoralens intercalated in the stacked DNA bases. In addition, the formation of the cross-linking sites is also influenced by the photoreactivity of cytosine which is presently not understood.It is likely that theoretical mechanism of the chemical reaction of DNA with other cross-linking agents can be similarly hypothesized.     

Site and sequence specificity of the daunomycin-DNA interaction

The site and sequence specificity of the daunomycin-DNA interaction was examined by equilibrium binding methods, by deoxyribonuclease I footprinting studies, and by examination of the effect of the antibiotic on the cleavage of linearized pBR322 DNA by restriction endonucleases PvuI and EcoRI. These three experimental approaches provide mutually consistent results showing that daunomycin indeed recognizes specific sites along the DNA lattice. The affinity of daunomycin toward natural DNA increases with increasing GC content. The quantitative results are most readily explained by binding models in which daunomycin interacts with sites containing two adjacent GC base pairs, possibly occurring as part of a triplet recognition sequence. Deoxyribonuclease I footprinting studies utilizing the 160 base pair (bp) tyrT DNA fragment and 61 and 53 bp restriction fragments isolated from pBR322 DNA further define the sequence specificity of daunomycin binding. Specific, reproducible protection patterns were obtained for each DNA fragment at 4 degrees C. Seven protected sequences, ranging in size from 4 to 14 bp, were identified within the tyrT fragment. Relative to the overall tyrT sequence, these protected sequences were GC rich and contained a more limited and distinct distribution of di- and trinucleotides. Within all of the protected sequences, a triplet containing adjacent GC base pairs flanked by an AT base pair could be found in one or more copies. Nowhere in the tyrT fragment did that triplet occur outside a protected sequence. The same triplet occurred within seven out of nine protected sequences observed in the fragments isolated from pBR322 DNA. In the two remaining cases, three contiguous GC base pairs were found. We conclude that the preferred daunomycin triplet binding site contains adjacent GC base pairs, of variable sequence, flanked by an AT base pair. This conclusion is consistent with the results of a recent theoretical study of daunomycin sequence specificity [Chen, K.-X., Gresh, N., & Pullman, B. (1985) J. Biomol. Struct. Dyn. 3, 445-466]. Adriamycin and the beta-anomer of adriamycin produce the same qualitative pattern of protection as daunomycin with the tyrT fragment. Daunomycin inhibits the rate of digestion of pBR322 DNA by PvuI (recognition sequence 5'-CGATCG-3') to a greater extent than it does EcoRI (recognition sequence 5'-GAATTC-3'), a finding consistent with the conclusions derived from our footprinting studies. Our results, as a whole, are the clearest indication to date that daunomycin recognizes a specific DNA sequence as a preferred binding site.     

DNA sequence determinants for binding of the Escherichia coli catabolite gene activator protein.

The consensus DNA site for binding of the Escherichia coli catabolite gene activator protein (CAP) is 22 base pairs in length and is 2-fold symmetric: 5'-AAATGTGATCTAGATCACATTT-3'. Positions 4 to 8 of each half of the consensus DNA half-site are the most strongly conserved. In this report, we analyze the effects of substitution of DNA base pairs at positions 4 to 8, the effects of substitution of thymine by uracil and by 5-methylcytosine at positions 4, 6, and 8, and the effect of dam methylation of the 5'-GATC-3' sequence at positions 7 to 10. All DNA sites having substitutions of DNA base pairs at positions 4 to 8 exhibit lower affinities for CAP than does the consensus DNA site, consistent with the proposal that the consensus DNA site is the ideal DNA site for CAP. Specificity for T:A at position 4 appears to be determined solely by the thymine 5-methyl group. Specificity for T:A at position 6 and specificity for A:T at position 8 appear to be determined in part, but not solely, by the thymine 5-methyl group. dam methylation has little effect on CAP.DNA complex formation. The thermodynamically defined consensus DNA site spans 28 base pairs. All, or nearly all, DNA determinants required for maximal affinity for CAP and for maximal thermodynamically defined CAP.DNA ion pair formation are contained within a 28-base pair DNA fragment that has the 22-base pair consensus DNA site at its center. The quantitative data in this report provide base-line thermodynamic data required for detailed investigations of amino acid-base pair and amino acid-phosphate contacts in this protein-DNA complex.     

Dissociation of the lactose repressor-operator DNA complex: effects of size and sequence context of operator-containing DNA

The dissociation kinetics for repressor-32P-labeled operator DNA have been examined by adding unlabeled operator DNA to trap released repressor or by adding a small volume of concentrated salt solution to shift the Kd of repressor-operator interaction. The dissociation rate constant for pLA 322-8, an operator-containing derivative of pBR 322, was 2.4 X 10(-3) s-1 in 0.15 M KCl. The dissociation rate constant at 0.15 M KCl for both lambda plac and pIQ, each of which contain two pseudooperator sequences, was approximately 6 X 10(-4) s-1. Elimination of flanking nonspecific DNA sequences by use of a 40 base pair operator-containing DNA fragment yielded a dissociation rate constant of 9.3 X 10(-3) s-1. The size and salt dependences of the rate constants suggest that dissociation occurs as a multistep process. The data for all the DNAs examined are consistent with a sliding mechanism of facilitated diffusion to/from the operator site. The ability to form a ternary complex of two operators per repressor, determined by stoichiometry measurements, and the diminished dissociation rates in the presence of intramolecular nonspecific and pseudooperator DNA sites suggest the formation of an intramolecular ternary complex. The salt dependence of the dissociation rate constant for pLA 322-8 at high salt concentrations converges with that for a 40 base pair operator. The similarity in dissociation rate constants for pLA 322-8 and a 40 base pair operator fragment under these conditions indicates a common dissociation mechanism from a primary operator site on the repressor.     

Interstrand psoralen cross-links do not introduce appreciable bends in DNA

Analysis of the X-ray crystallographic structure of an 8-methoxypsoralen-thymine monoadduct has led to the suggestion that psoralen cross-links would bend DNA by as much as 70 degrees [Peckler, S., Graves, B., Kanne, D., Rapoport, H., Hearst, J. E., & Kim, S.-H. (1982) J. Mol. Biol. 162, 157-172]. DNA can exist in a stably bent configuration in solution as recently demonstrated from analysis of polyacrylamide gel electrophoresis and differential decay of birefringence. Using these techniques, we have investigated the structure of DNA cross-linked with 8-methoxypsoralen and 4,5',8-trimethylpsoralen. The results are not consistent with cross-links introducing any appreciable stable bend in double-stranded DNA molecules. Results suggest that photobound 4,5',8-trimethylpsoralen molecules lengthen DNA by the equivalent of about one base pair per photobound adduct. We have also determined that 4,5',8-trimethylpsoralen cross-links are introduced preferentially into 5'-TA compared to 5'-AT DNA sequences.     

High-resolution analysis of a histone H1 binding site in a rat albumin gene

Interaction of rat liver histone H1 fraction with the 5'-end of the rat serum albumin gene was localized within a 346 base pair (bp) restriction fragment. Sequence analysis of the fragment showed the fragment was 72 mol % adenosine-thymidine, which is significantly greater than the mole percent adenosine-thymidine composition of the rat genome. Gel retardation assays of the histone H1-DNA interaction indicate the complex formed behaves as previously characterized H1-DNA and shows a high-affinity H1 binding site within the enriched albumin restriction site. Deoxyribonuclease I (DNase I) protection assays on the H1 binding site define three protected regions only on the template strand of the DNA fragment. The three sites lie 55 and 110 bp apart (165 bp between the first and third binding site) with a consensus binding sequence of 5'-GA-ATA-CTGGCTT-C-TT-CTA-G-3'. The sequences between the protected DNA regions are highly enriched in adenosine-thymidine bases (79.3 and 86 mol % adenosine-thymidine, respectively). The functional significance is not understood.     

Avian Retrovirus pp32 DNA-Binding Protein I. Recognition of Specific Sequences on Retrovirus DNA Terminal Repeats

The avian retrovirus pp32 protein possesses a DNA-nicking activity which prefers supercoiled DNA as substrate. We have investigated the binding of pp32 to avian retrovirus long terminal repeat (LTR) DNA present in both supercoiled and linear forms. The cloned viral DNA was derived from unintegrated Schmidt-Ruppin A (SRA) DNA. A subclone of the viral DNA in pBR322 (termed pPvuII-DG) contains some src sequences, tandem copies of LTR sequences, and partial gag sequences in the order src-U(3) U(5):U(3) U(5)-gag. Binding of pp32 to supercoiled pPvuII-DG DNA followed by digestion of this complex with a multicut restriction enzyme (28 fragments total) permitted pp32 to preferentially retain on nitrocellulose filters two viral DNA fragments containing only LTR DNA sequences. In addition, pp32 also preferentially retained four plasmid DNA fragments containing either potential promoters or Tn3 "left-end" inverted repeat sequences. Mapping of the pp32 binding sites on viral LTR DNA was accomplished by using the DNase I footprinting technique. The pp32 protein, but not the avian retrovirus alphabeta DNA polymerase, is able to form a unique protein-DNA complex with selected regions of either SRA or Prague A LTR DNAs. Partial DNase I digestion of a 275-base pair SRA DNA fragment complexed with pp32 gives upon electrophoresis in denaturing gels a unique ladder pattern, with regions of diminished DNase I susceptibility from 6 to 10 nucleotides in length, in comparison with control digests in the absence of protein. The binding of pp32 to this fragment also yields enhanced DNase I-susceptible sites that are spaced between the areas protected from DNase I digestion. The protected region of this unique complex was a stretch of 170 +/- 10 nucleotides that encompasses the presumed viral promoter site in U(3), which is adjacent to the src region, extends through U(5), and proceeds past the joint into U(3) for about 34 base pairs. No specific protection or DNase I enhancement by pp32 was observed in experiments with a 435-base pair SRA DNA fragment derived from a part of U(3) and the adjacent src region or a 55-base pair DNA fragment derived from another part of U(3). The DNA sequence of Prague A DNA at the fused LTRs differs from that of SRA DNA. The alteration in the sequence at the juncture of the LTRs prevented pp32 from forming a stable complex in this region of the LTR. Our results are relevant to two aspects of the interaction between pp32 and LTR DNA. First, the pp32 protein in the presence of selected viral DNA restriction fragments possibly forms a higher order oligomer analogous to Escherichia coli DNA gyrase-DNA complexes or eucaryotic nucleosome structures. Second, the specificity of the binding suggests a role for pp32 and the protected DNA sequences in the retrovirus life cycle. The preferred sequences to which pp32 binds include two adjacent 15-base pair inverted terminal repeats at the joint between U(5) and U(3) in SRA DNA. This region is involved in circularization of linear DNA and is perhaps the site that directs integration into cellular DNA.     

Cloning and characterization of a highly reiterated 5.8-kilobase pair nucleolar EcoRI DNA fragment found in Novikoff hepatoma ascites cells

The DNA of Novikoff hepatoma ascites cells was found to contain a 3.6-megadalton EcoRI restriction fragment, referred to as EcoRI fragment A (Parker et al., 1979). C0t analyses demonstrated an enrichment of fragment A sequences in Novikoff hepatoma genome relative to normal rat liver DNA. This fragment was cloned in lambda gtWES to determine its molecular structure and sequence organization. The DNA from a positive clone was labeled by nick translation and hybridized to a Southern blot of EcoRI digested Novikoff DNA. Distinct hybrids formed with the region corresponding to fragment A. The greater degree of hybridization to the nucleolar fraction suggested a nucleolar enrichment of fragment A. Fragment A has a PstI site approximately 300 base pairs from one terminus which was used to generate mono-5'-32P-labeled fragments. The larger PStI subfragment, 5500 base pairs, labeled at a single terminus, was used to evolve a restriction enzyme map. The 300 base pair fragment was partially sequenced, revealing the presence of a repetitive sequence "island", TT(GTCT)8(GAAT)5G-. C0t analysis, utilizing the purified clone as a probe, confirmed the enrichment of fragment A sequences in the tumor relative to the normal rat liver control.     

High resolution psoralen mapping reveals an altered DNA helical structure in the SV40 regulatory region.

Preferential psoralen photobinding sites have been mapped in vitro on restriction fragments spanning the SV40 origin region and surrounding sequences by a new fine structure analysis technique. Purified DNA fragments were photoreacted with 3H-5-methylisopsoralen (3H-5-MIP), a psoralen derivative which forms only monoadducts. Fragments were then end-labeled and digested with lambda exonuclease, a 5' processive enzyme which we have determined pauses at 5-MIP monoadducts. When photobinding sites were mapped on denaturing sequencing gels, it was observed that 5-MIP binds preferentially to 5'-TA sites, and to a lesser degree to 5'-AT sites. Utilizing this approach, we have identified a psoralen hypersensitive region in which the binding sites were much stronger than those in the surrounding sequences. This region extends from 150 base pairs (bp) to the late side of the enhancers to the early enhancer/promoter boundary. We suggest that this region contains a sequence directed structural alteration of the DNA helix which can be detected by the psoralen mapping approach described.     

Phe 771 of Escherichia coli DNA polymerase I (Klenow fragment) is the major site for the interaction with the template overhang and the stabilization of the pre-polymerase ternary complex

To identify the sites in the Klenow fragment of Escherichia coli DNA polymerase I that interact with the ssDNA overhang of the template strand in the pre-polymerase ternary complex, we carried out UV-mediated photo-cross-linking of the enzyme-DNA-dNTP ternary complex. The template strand contained a nine-nucleotide overhang and was radiolabeled at the 5'-end. Since the enzyme-TP-dNTP ternary complex but not the E-TP binary complex is stable at high ionic strengths, the cross-linking was carried out in the presence of 0.5 M NaCl. The cross-linked E-TP-dNTP complex was purified and subjected to trypsin digestion. The radiolabeled TP cross-linked peptide was further purified by DEAE-Sepharose and C18 column chromatography and subjected to amino acid sequencing. The release of radiolabeled DNA during each sequencing cycle was also monitored. The sequencing results as well as the radioactivity release pattern show that F771, contained in a peptide spanning amino acids 759-775 of pol I, is the unequivocal site of the template cross-linking. A qualitative assessment of the cross-linking efficiency of the template overhang containing a TT sequence at different positions in the ternary complex further suggests that the major cross-linking site within the template overhang is at the second and/or third nucleotide. An examination of the F771A mutant enzyme showed that it was able to form the E-TP binary as well as E-TP-dNTP ternary complex; however, it could not cross-link to the template-primer in the ternary complex. Furthermore, the ternary complex with F771A was qualitatively defective and exhibited some salt sensitivity. These results suggest that F771 participates in the stabilization of the pre-polymerase ternary complex.     

DNA sequence context affects UV-induced mutagenesis in Escherichia coli

We investigated the effect of altering the DNA sequence surrounding a mutable target site on the production of ultraviolet light (UV) induced mutations. Site-directed base substitutions were incorporated on both sides of a TAA sequence encoding a UAA nonsense defect in the tyrA14 allele of Escherichia coli. This allele is readily revertable by UV and a total of eight different base substitution mutations can be recovered. Five different strains harboring DNA sequences allowing the formation of 5'-TT, 5'-CT and 5'-TA* photoproducts were constructed and exposed to UV. DNA sequence analysis was used to determine the spectrum of the revertants that were recovered. The results showed that changes at the 3'-base of a TT site were predominantly T to C transitions and T to A transversions. However, unlike the TT site, a 5'-CT site produced a relatively high frequency of T to G transversions. In addition, T to A transversions that could not have been targeted by a cyclobutane-type or [6-4]-type pyrimidine dimer were produced; this result suggested that these mutations may be targeted by a TA* photoproduct. Also, a distinct strand bias was noted for two mechanistically identical base substitutions in a strain having a palindromic target sequence; this result may reflect an unequal damage distribution or processing of photoproducts as a consequence of asymmetric DNA replication. Finally, our results show that DNA sequences expected to allow the greatest density of UV-induced DNA damage produce the highest mutation frequencies. Overall, these findings provide new insights regarding the role of DNA photoproducts in UV mutagenesis.     

Structure of the DNA interstrand cross-link of 4,5',8-trimethylpsoralen.

4,5',8-Trimethylpsoralen (TMP) cross-links a 5' TpA or a 5' ApT site by photoreacting with one thymine moiety in each DNA strand. We are interested in whether psoralen interstrand cross-links all share one structure or whether there are significant differences. In this paper, we employed a rapid method for probing the structure of the cross-link by making a series of TMP cross-linked duplexes containing specific base-pair mismatches. The relative stability provided by a base pair can be correlated with neighboring base pairs by comparing the extents of gel retardation when base-pair mismatches happen in each position. From our studies, we infer that with respect to the furan-side strand, the 5'T.A base pair of the two T.A base pairs in the TpA site is not hydrogen bonded. Immediately on each side of the cross-linked TpA site is a highly stabilized base pair. Next, a region of decreased stability occurs in each arm of a cross-linked duplex and these base pairs of least stability are located farther away from the cross-linked thymines as the lengths of the arms of the cross-linked helix increase. Finally, even in 7 M urea at 49 degrees C the cross-linked helix is hydrogen bonded at both ends of a duplex of 22 base pairs. We propose that the structures of interstrand cross-links in DNA vary appreciably with the DNA sequence, the length of the DNA duplex, and the structures of the DNA cross-linking agents.