Thermostability of double-stranded deoxyribonucleic acids: effects of covalent additions of a psoralen

We have carried out a thermodynamic study on the effects of covalent additions of the psoralen derivative HMT, 4'-(hydroxymethyl)-4,5',8-trimethylpsoralen, on the stability of double-stranded deoxyoligonucleotides. This was done with two systems. The first was a double-stranded DNA formed by two non-self-complementary oligonucleotides, 5'-GAAGCTACGAGC-3' and 5'-GCTCGTAGCTTC-3', where we site specifically placed an HMT molecule on the thymidine residue in oligonucleotide 5'-GAAGCTACGAGC-3' as either a furan-side monoadduct or a pyrone-side monoadduct. The second was a double-stranded DNA formed by a self-complementary oligonucleotide, 5'-GGGTACCC-3', where we placed an HMT molecule on the thymidine residue of each strand as a furan-side monoadduct or cross-linked the two strands with an HMT molecule linked to the two thymidines. We found that HMT cross-linking of the two strands stabilizes the double helix formed by 5'-GGGTACCC-3', as one might expect. Less predictable results were that the monoaddition of a psoralen stabilizes the double helix formed by the two non-self-complementary oligonucleotides by as much as 1.3 kcal/mol as a furan-side monoadduct and 0.7 kcal/mol as a pyrone-side monoadduct at 25 degrees C in 50 mM NaCl. In contrast, the monoaddition of a psoralen on each of the two thymidines in the double helix formed by 5'-GGGTACCC-3' destabilizes the helix by 1.8 kcal/mol at 25 degrees C in 1 M NaCl. This destabilization arises from an unfavorable enthalpy change (8.6 kcal/mol) and a favorable entropy change (23 cal/K X mol) due to the two HMT molecules at the centers of each strand.(ABSTRACT TRUNCATED AT 250 WORDS)     

In vitro repair of psoralen-DNA cross-links by RecA, UvrABC, and the 5'-exonuclease of DNA polymerase I

Psoralens produce DNA interstrand cross-links which are thought to be repaired via a sequential excision and recombination mechanism in Escherichia coli. The first round of incision by UvrABC has been characterized: it results in 11-base oligonucleotide cross-linked to an intact DNA strand (Van Houten, B., Gamper, B., Holbrook, S.R., Hearst, J.E., and Sancar, A. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 8077-8081). In the present work, DNA substrates containing 4'-hydroxymethyl-4,5',8-trimethylpsoralen (HMT) cross-links in defined positions are constructed and used to analyze the other steps in repair. It is shown that RecA protein mediates strand transfer past an oligonucleotide cross-linked to a single-stranded DNA circle and that the resulting heteroduplex is a substrate for the UvrABC complex: it excises a double-stranded oligonucleotide which contains the HMT cross-link. It is also found that the first round of UvrABC incision does not lead directly to strand exchange but that an intervening step is needed. That step is carried out in vitro by the 5'-exonuclease activity of DNA polymerase I (pol I) which creates a single-stranded DNA region (a gap) at an incised cross-link such that RecA can initiate strand exchange. Studies using cross-linked oligonucleotides showed that the gap produced by pol I results from the inability of the polymerase to add nucleotides to a 3'-OH end two to three nucleotides away from the furan side of an HMT cross-link. Pol I can, however, extend a 3'-OH end next to the pyrone side of the cross-link. Since UvrABC incises predominantly the furan side of psoralen cross-links in duplex DNA, this discrepancy has important consequences for repair.     

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.     

Cooperativity in A-tract structure and bending properties of composite TnAn blocks

The existence of intrinsically curved DNA molecules incorporating short runs of adenines is undisputed, but none of the current models can explain the entire experimental data set. Recently, Burkhoff and Tullius [Burkhoff, A. M., & Tullius, T. D. (1988) Nature 331, 455-457] offered an explanation for Hagerman's observations on A4T4N2 vs T4A4N2 polymers [Hagerman, P. J. (1986) Nature 321, 449-450], which showed that A4T4N2 multimers migrate anomalously slowly on polyacrylamide gels and T4A4N2 multimers migrate normally. In A4T4N2 multimers Burkhoff and Tullius observe a hydroxy-radical cutting pattern associated with bent DNA and a B-like cutting pattern in T4A4N2. They attribute this difference in cutting pattern to a clash in the TA step of T4A4N2 and suggest that TA4N5 might already adopt an unbent B-DNA conformation [Tullius T. D., & Burkhoff, A. M. (1988) in Structure and Expression. Vol. 3: DNA Bending and Curvature (Olson, W. K., Sarma, M. H., Sarma, R. H., & Sundaralingam, M., Eds.) pp 77-85, Adenine Press, Guilderland, NY]. We show that the conformation adopted by TnAn blocks is similar to that of AnTn blocks. Two A-tract structures of opposite polarity coexist in both blocks. Moreover, we demonstrate a cooperative buildup of a T-tract structure adjacent to an A-tract structure that cannot be predicted by any of the current models. We conclude that AA steps do not assume the same conformation in long tracts of A's as in isolated AA steps. Therefore, the assumption of nearest-neighbor models, that global curvature is an additive phenomenon of local effects, is invalid.     

Identification of sites of 4'-(hydroxymethyl)-4,5',8-trimethylpsoralen cross-linking in Escherichia coli 23S ribosomal ribonucleic acid

The reagent 4'-(hydroxymethyl)-4,5',8-trimethylpsoralen (HMT) was used to cross-link 23S rRNA from Escherichia coli under 50S ribosomal subunit reconstitution conditions. Following partial digestion of the RNA with ribonuclease T1, two-dimensional diagonal electrophoresis in denaturing polyacrylamide gels was used to isolate fragments derived from the cross-linked sites. These fragments were analyzed by digestion with ribonucleases T1 and A and their positions in the 23S RNA sequence identified. Fragment a1 (positions 1325-1426) is cross-linked to a2 (positions 1574-1623); fragment b1 (positions 1700-1731) is cross-linked to b2 (positions 1732-1753); and a cross-link is formed within fragment c (or c') (positions 863-916). In the latter case, the cross-link was located precisely, linking residues C867 and U913. All three HMT-mediated cross-links are consistent with a proposed secondary structure model for 23S RNA [Noller, H. F., Kop, J., Wheaton, V., Brosius, J., Gutell, R. R., Kopylov, A. M., Dohme, F., Herr, W., Stahl, D. A., Gupta, R., & Woese, C. R. (1981) Nucleic Acids Res. 9, 6167-6189].     

Nonhistone proteins HMG1 and HMG2 unwind DNA double helix.

In a previous communication we have shown that both HMG1 and HMG2 nonhistone proteins change the DNA helical structure and the binding of HMG1 and HMG2 to DNA induces a net unwinding equivalent of DNA double helix (Javaherian, K., Liu, L. F. and Wang, J. C. (1978) Science, 199, 1345-1346). Employing melting absorption technique, we now show that in the presence of salt HMG1 and HMG2 destabilize DNA whereas in the absence of salt, they both stabilize DNA molecules. Consequently the folded structure of HMG must play an important role in melting DNA. Furthermore, by measuring topological winding number using competition unwinding experiments, we conclude that HMG1 has a higher affinity for a single-stranded DNA relative to double-stranded DNA. These results together suggest that HMG1 and HMG2 unwind DNA double helix by local denaturation of the DNA base pairs. The net unwinding angles have been measured to be 22 degrees and 26 degrees per molecule of HMG1 and HMG2 respectively.     

DNA-supercoiling is affected in vitro by the peptide antibiotics tyrocidine and gramicidin

Tyrocidine, a peptide antibiotic produced by Bacillus brevis (ATCC 8185), relaxes superhelical DNA in a biphasic manner and induces 'packaging' of the DNA at higher concentrations. This was concluded from studies using the sensitive 4,5',8-trimethylpsoralen photobinding technique [Sinden, R. R., Carlson, J. O. & Pettijohn, D.-E. (1980) Cell 21, 773-783]. Relaxed DNA is not affected by tyrocidine whereas linearized molecules become packaged. The linear gramicidin synthesized by the same strain reverses the tyrocidine-induced relaxation as well as the packaging, an observation which might be of biological relevance.     

Lambda phage cro repressor interaction with its operator DNA: 2'-deoxy-5-fluorouracil OR3 analogues

The experiments here show that chemically synthesized DNA containing fluorine at selected sites can be used to test specific predictions of a model for cro repressor--operator interaction. This is done by observation of the perturbation to the fluorine-19 NMR spectra of analogues of OR3 synthesized with 2'-deoxy-5-fluorouracil at specific positions in the DNA helix. Although the three-dimensional structure of the cro repressor from phage lambda has been determined by Matthews and co-workers [Anderson, W., Ohlendorf, D., Takeda, Y., & Matthews, B. (1981) Nature (London) 290, 754-758], direct structural observations on the complex of the protein with its specific DNA recognition sequence, OR3, are limited. From that structure of the protein, alone, a model of its complex to DNA was built by fitting B-form DNA, with some distortion [Ohlendorf, D., Anderson, W., Fisher, R., Takeda, Y., & Matthews, B. (1982) Nature (London) 298, 718-723]. That model proposes that the cro repressor contacts only one side of this DNA double helix and a number of specific protein--DNA contacts. To test the model, 2'-deoxy-5-fluorouracil was used to place the fluorine-19 nuclear spin-label on the side of the DNA contacting the cro repressor and on the opposite side facing away from the cro repressor. The results presented here are consistent with the prediction that lambda phage cro repressor contacts only one side of the DNA double helix.     

DNA bending by the bulge defect

Comparative gel electrophoresis measurements were used to characterize DNA bending in molecules containing an extra adenosine on one strand, the so-called bulge defect. We used oligomers containing A6 tracts separated from the bulged base by varying numbers of nucleotides to determine the direction and magnitude of the bulge bend. Helix unwinding by the bulge was determined from the electrophoretic anomaly as a function of the size of the repeated monomers. We conclude that the bulge bend is 21 degrees +/- 3 degrees, primarily in the direction of tilt away from the bulged base. The total helical advance of the DNA at the bulge site is smaller than would be the case if the complementary T were present, corresponding to an unwinding by 25 degrees +/- 6 degrees. These values are in good agreement with the results of NMR and energy minimization studies of the bulged base in double-helical deoxyoligonucleotides [Woodson, S. A., & Crothers, D.M. (1988) Biochemistry 27, 3130-3141]     

On the question of DNA bending: two-dimensional NMR studies on d(GTTTTAAAAC)2 in solution

It is very well documented that the presence of an An.Tn tract causes intrinsic DNA bending. Hagerman demonstrated that the sequence in which the An.Tn tracts are joined plays a very crucial role in determining DNA bending. For example, Hagerman showed that the polymer with a repeat of d(GA4T4C)n greater than or equal to 10 is bent but the polymer with a repeat of d(GT4A4C)n greater than or equal to 10 is not bent [Hagerman, P. J. (1986) Nature (London) 326, 720-722]. Earlier we have shown that the decamer repeat d(GA4T4C)2 is itself bent with a finite structural discontinuity at the A----T sequence [Sarma, M. H., Gupta, G., & Sarma, R. H. (1988) Biochemistry 27, 3423-3432]. In the present article, we summarize our studies on the decamer repeat d(GT4A4C)2 structure in solution. By employment of 1D and 2D 1H NMR studies at 500 MHz a complete sequential assignment has been made for the exchangeable and nonexchangeable protons belonging to the ten nucleotides. NOESY data were collected for d(GT4A4C)2 at 17 degrees C in D2O for three mixing times, 150, 100, and 50 ms. A quantitative NOESY simulation technique was employed to arrive at a structural model of d(GT4A4C)2 in solution. Our detailed analyses revealed the following structural features: (i) The duplex adopts the gross morphology of a B-DNA. (ii) All the A.T pairs are propeller twisted (less than or equal to -15 degrees). (iii) Although both A and T nucleotides belong to the C2'-endo,anticonformational domain, there is a mild variation in the actual conformation of the A and T residues. (iv) Even though there is a subtle conformational difference in the A and T nucleotides, two structural frames of T4.A4 segments are joined at the T----A sequence in such a way that there is no finite discontinuity at the junction; i.e., two neighboring frames exactly coincide at the T----A junction. Thus, our studies on d(GA4T4C)2 (Sarma et al., 1988) and on d(GT4A4C)2 (this article) reveal the structural peculiarity of the An.Tn tract and the effect of A----T/T----A sequence in causing DNA bending.     

Interactions of norharman and harman with DNA.

The interactions of norharman (9H-pyrido [3,4-b] indole) and harman (1-methyl-9H-pyrido [3,4-b] indole) with DNA were studied. DNA caused remarkable fluorescence quenching and change in the absorption spectra of the dyes. Scatchard plots obtained by optical titration gave Kd values of 2.2 X 10(-5)M and 7.7 X 10(-6)M, and apparent numbers of binding sites of 0.13/base and 0.12/base for norharman and harman, respectively. Agarose gel electrophoresis of circular DNA, closed in the presence or absence of norharman revealed that the dye intercalates DNA, thereby causing 17 +/- 3 degrees unwinding of the double helix.     

N(4)C-alkyl-N(4)C cross-linked DNA: bending deformations in duplexes that contain a -CNG- interstrand cross-link

Short DNA duplexes containing a 1,3-N(4)C-alkyl-N(4)C interstrand cross-link that joins the two C residues of a -CNG- sequence were prepared using either a phosphoramidite or convertible nucleoside approach. The alkyl cross-link consists of 2, 4, or 7 methylene groups. The duplexes, which contain a seven-base-pair core and A(3)/T(3) complementary 3'-overhanging ends, were characterized by enzymatic digestion and MALDI-TOF mass spectrometry. Ultraviolet thermal denaturation studies showed that the duplexes denature in a cooperative manner and that the length of the cross-link affects the thermal stability. Thus, the transition temperature of the ethyl cross-linked duplex, 42 degrees C, is 16 degrees C higher than the melting temperature of the corresponding non-cross-linked control, whereas the transition temperatures of the butyl and heptyl cross-linked duplexes, 73 and 72 degrees C, respectively, are 46-47 degrees C higher. The reduced molecularity of denaturation of the cross-linked duplexes versus melting of the non-cross-linked duplex most likely accounts for these differences. Examination of molecular models suggests that the ethyl cross-link is too short to span the distance between the two C residues at the site of the cross-link in B-form DNA without causing distortion of the helix, whereas less and no distortion would be expected for the butyl and heptyl cross-links, respectively. The circular dichroism spectra, which show greatest deviation in the ethyl cross-linked duplex from B-form DNA, are consistent with this expectation. Anomalous mobilities on native polyacrylamide gels of multimers produced by self-ligation of each of the cross-linked duplexes suggest that the ethyl and butyl cross-linked duplexes undergo bending deformations, whereas multimers derived from the heptyl cross-linked duplex migrated normally. The bending angle was estimated to be 20 degrees, 13 degrees, and 0 degrees for the ethyl, butyl, and heptyl cross-linked duplexes, respectively. Thus, it appears that the degree of bending in these N(4)C-alkyl-N(4)C cross-linked duplexes is controlled by the length of the cross-link.     

Tyrocidine-induced modulation of the DNA conformation in Bacillus brevis

Using the [3H]trimethylpsoralen photobinding method [Sinden, R.R., Carlson, J.O. & Pettijohn, D.E. (1980) Cell 21, 773-783], a decrease in unrestrained torsional tension of DNA was detected in Bacillus brevis cells when they had entered the sporulation phase. This decrease in superhelicity was found in cells which synthesized the peptide antibiotic tyrocidine and which were stimulated to sporulate. Fluctuations in superhelicity probably reflect a highly complicated picture of tension-relaxing and tension-inducing activities. Addition of tyrocidine to vegetative cells reduced by one-half the torsional tension from DNA, whereas ethidium bromide relaxes DNA completely. Cross-links between DNA and tyrocidine were introduced with ultraviolet light in vitro and in vivo indicating that the modulation of the DNA conformation in the cell may in fact be due to a DNA-tyrocidine interaction. In a growing B. brevis culture exogenous [3H]tyrocidine could only be photobound to DNA after the cells had entered the sporulation phase. Our results could mean that the peptide antibiotic tyrocidine is active in B. brevis on the DNA level as one regulatory factor controlling DNA functions.     

Molecular mechanics and dynamics calculations on (dA)10.(dT)10 incorporating distance constraints derived from NMR relaxation measurements

Structural constraints derived from proton NMR relaxation measurements on poly(dA).poly(dT) in the form of interproton separations and orientation have been combined with molecular mechanics and annealed molecular dynamics calculations to derive a model for the solution-state structure of this molecule. Three different possible starting configurations, including the standard A and B forms of Arnott and Hukins [Arnott, S., & Hukins, D. W. L. (1972) Biochem. Biophys. Res. Commun. 47, 1506-1509] and the heteronomous (H) structure [Arnott, S., Chandrasekaran, R., Hall, I. H., & Puigjaner, L. C. (1983) Nucleic Acids Res. 11, 4141-4155], were examined. Both the B- and H-DNA structures converged to the same B-like structure (approximately C2'-endo conformation on both the A and T sugars, glycosidic bond torsional angle of 63-73 degrees) with the same energies and average helical parameters that gave good fits of the NMR relaxation rates. This model also accounts for the experimental observation [Behling, R. W., & Kearns, D. R. (1986) Biochemistry 25, 3335-3346] that the AH2 proton interacts more strongly with the H1' sugar proton on the T strand than on the A strand. Although the helix repeat angle (39 degrees) is larger than that for standard B-DNA (36 degrees), this does not result in a significantly smaller minor groove, as monitored by the interstrand P-P separation. Calculations starting with the A-DNA structure lead to a very high energy structure that gave a poorer fit of the NMR data.     

Structural model for an oligonucleotide containing a bulged guanosine by NMR and energy minimization

We present three-dimensional structural models for a DNA oligomer containing a bulged guanosine based on proton NMR data and energy minimization computations. The nonexchangeable proton resonances of the duplex 5'd(GATGGGCAG).d(CTGCGCCATC) are assigned by nuclear Overhauser effect spectroscopy (NOESY) and correlated spectroscopy connectivities, and the NMR spectrum is compared with that of a regular 8-mer of similar sequence, 5'd(GATGGCAG).d(CTGCCATC). Experimental proton-proton distances are obtained from NOESY spectra acquired with mixing times of 100, 150, and 200 ms. A refined three-dimensional structure for the bulge-containing duplex is calculated from regular B DNA starting coordinates by using the AMBER molecular mechanics program [Weiner, S. J., Kollman, P. A., Case, D. A., Singh, U. C., Ghio, C., Alagona, G., Profeta, S., & Weiner, P. (1984) J. Am. Chem. Soc. 106, 765-784]. We compare structures obtained by building the helix in three and four base pair increments with structures obtained by direct minimization of the entire nine base sequence, with and without experimental distance constraints. The general features of all the calculated structures are very similar. The helix is of the B family, with the extra guanine stacked into the helix, and the helix axis is bent by 18-23 degrees, in agreement with gel mobility data for bulge-containing sequences [Rice, J. A. (1987) Ph.D. Thesis, Yale University].     

Potency and property of DNA conformational damage determine ATM activation: two new DNA topological probes

We have shown previously the intercalation geometry of a series of acenaphtho [1,2-b] pyrrole derivatives with DNA double helix in vitro. In this report we chose a couple of intercalating analogues and a Chinese traditional medicine Tanshinone IIA as probes to investigate the response of DNA damage sensor ataxia-telangiectasia mutated (ATM) protein toward the DNA topological change in vivo. The two analogues (1)a (3-(4-Methyl-piperazin)-8-oxo-8H-acenaphtho [1,2-b]pyrrole -9-carbonitrile) and (3)a (3-(3-Dimethylamino-propylamino)-8-oxo-8H-acenaphtho[1,2-b]pyrrole-9- carbonitrile) could unwind double helix to different extents, whereas Tanshinone IIA could wind the double helix. Using a combination of circular dichroism (CD) studies and immunoflurescence assays, we found for the first time that the ATM protein kinase can respond to the unwinding chromatin conformational damage caused by (1)a and (3)a, while it could not be activated by the winding effects caused by Tanshinone IIA. Moreover, the amount of ATM protein phosphorylation is consistent with the degree of unwinding conformational damage. The average number of ATM foci in an MCF-7 cell is 32 +/- 1.5 at 6 microM (1)a, which is significantly higher than the 8 microM (3)a exposure (15 +/- 0.5, p < 0.5). A new couple of DNA topological probes, (1)a and (3)a have been found for the future semi-quantitative investigation of factors involved in the DNA damage pathway.     

Wavelength dependence for the photoreversal of a psoralen-DNA cross-link

We report an action spectrum for the photoreversal of a psoralen cross-link joining two self-complementary DNA oligonucleotides. The cross-link was formed between two thymines (T) on opposite strands of the DNA oligomers and 4'-(hydroxymethyl)-4,5',8-trimethylpsoralen (HMT). For comparison, we also present an action spectrum for the photoreversal of the isolated diadduct T-HMT-T. The wavelength dependence for the diadduct photoreversal parallels its absorption spectrum. Both the diadduct and the cross-linked DNA can be photoreversed by exposure to light with wavelengths between 240 and 313 nm. We did not observe photoreversal at 334 nm or above. At least two distinct absorption bands appear to contribute to photoreversal. We measured a quantum yield of 0.16 for photoreversal of the isolated diadduct at wavelengths between 240 and 266 nm. For wavelengths above 280 nm, the quantum yield is 0.30. We also observed a preferential photoreversal at the furan end of the psoralen in the T-HMT-T diadduct. In contrast, the cross-linked DNA oligonucleotides preferentially photoreversed at the pyrone end of the psoralen adduct. The rate constant for photoreversal of the cross-linked DNA is larger than that for the isolated diadduct at wavelengths below 300 nm.     

Unique cross-link and monoadduct repair characteristics of a xeroderma pigmentosum revertant cell line

Monoadducts and cross-links formed in DNA of human cells by a psoralen derivative, 4'-hydroxy-methyl-4,5',8-trimethylpsoralen (HMT), have been measured by a new, simple method, based on S1 nuclease digestion of 3H-labeled adducts in DNA, that provides rapid information on the repair of both classes of lesions. Normal human fibroblasts and cells from patients with dyskeratosis congenita and xeroderma pigmentosum (XP) group C were capable of removing both monoadducts and cross-links, whereas XP groups A and D failed to remove either. An XP revertant, isolated from a group A cell line on the basis of an acquired mutagen-induced resistance to ultraviolet light, has the unique property of being capable of removing cross-links but not monoadducts. Consistent with this property, the XP revertant was found to be resistant to cell killing by the cross-linking psoralen derivative, HMT, but as sensitive as its parental cell line to a monofunctional psoralen derivative, 5-methylisopsoralen.     

Induction of DNA replication-mediated double strand breaks by psoralen DNA interstrand cross-links

The effect of DNA interstrand cross-links (cross-links) on DNA replication was examined with a cell-free SV40 origin-dependent DNA replication system. A defined template DNA with a single psoralen cross-link and the SV40 origin of replication was replicated by HeLa cell-free extract in the presence of SV40 large T antigen. The psoralen cross-link inhibited DNA replication by terminating chain elongation at 1-50 nucleotides before the cross-linked sites. The termination of DNA replication by the cross-links mediated the generation of double strand breaks near the cross-linked sites. These results are the first biochemical evidence of the generation of double strand breaks by DNA replication.