NMR-derived solution structure of a 17mer hydroxymethyluracil-containing DNA.

Incorporation of 5-(hydroxymethyl)-2'-deoxyuridine into DNA in place of thymine by SPO1, a Bacillus subtilis bacteriophage, allows the viral DNA to bind selectively to transcription factor 1. We have synthesized a TF1-binding site: d(5'-ACCHACHCHHHGHAGGT-3')-d(5'-ACCHACAAAGAGHAGGT-3') and studied this molecule using NMR spectroscopy. The chemical shifts of exchangeable and non-exchangeable protons were sequentially assigned. Absence of corresponding NOEs in the imino-imino region suggested that the end base pairs did not form Watson-Crick hydrogen bond. Restrained molecular dynamics calculation yielded a family of B-DNA structures whose r.m.s.d. was 0.66 A (all atoms) for the internal 15 bp. The helical twist was 38.5 degrees per step. The base pairs were situated directly on the helix axis (X-displacement = -0.2 A). All sugars exhibited C2'-endo puckering with P = 167.3 degrees and upsilon(max)= 38.2 degrees. The OH groups of all hmU bases resided on the 3' side of the base plane and may affect the base orientation relative to the sugar plane as the average chi value for all hmU was 4 degrees more positive than that of other nucleosides (258 degrees versus 254 degrees ). Positive roll angles (rho) and small flanking twists (omega) at hmU suggested that the two hmU-A base pair steps open toward the minor grooves.

Manganese(II) as a paramagnetic probe of the tertiary structure of transfer RNA.

The effect of manganese on both the low field (10--15 ppm) and the high field (o--3 ppm) NMR spectra of unfractionated tRNA and yeast tRNAPhe has been investigated. Trace amounts of Mn2+ cause selective broadening of resonances which are assigned to specific tertiary interactions. The order in which resonances broaden is the same as the order in which they are stabilized by the addition of magnesium, namely s4U8 - A14, U33 and A58 - T54. From this we conclude that three of the strong binding sites probably are the same for both Mn2+ and Mg2+, and that these sites are located close to the tertiary interactions which are stabilized by the strongly bound metals. The broadening data, taken in conjunction with published X-ray data on yeast tRNAPhe, permit us to suggest some plausible locations for the strong binding sites.

Effect of magnesium and polyamines on the structure of yeast tRNAPhe.

The effect of magnesium and polyamines (spermine, spermidine, putrescine and cadaverine) on the structure of yeast tRNAPhe has been investigated. It is found that magnesium induces structural changes and stabilizes hydrogen bonds in the temperature range 22--44 degrees C in 0.17 M sodium. The number of Mg2+ which affect tRNA structure increases from 1 +/- 1 at 22 degrees C to 4 +/- 1 at 44 degrees C and the number of additional base pairs formed in the presence of magnesium increases from 1 +/- 1 at 22 degrees C to 4 +/- 1 at 44 degrees C. The spectral changes are more-or-less sequential. The polyamine spermine stabilizes some, but not all, of the structural features stabilized by magnesium at 44 degrees C, and the combination of magnesium and spermine, at low levels, is more effective than either cation alone in stabilizing tRNA structure. Comparison of the effects of spermine, spermidine, putrescine and cadaverine indicates that it is the asymmetric triamine unit which is important in the stabilization. Some spectral changes induced by magnesium can be assigned to stabilization of specific tertiary structure interactions and to alteration of stacking adjacent to U8-A14.

Specific conversion of s4 U to U in Escherichia coli tRNA by iodate oxidation.

The minor nucleoside 4-thiouridine in Escherichia coli tRNA is transformed selectively to uridine by iodate oxidation at acidic pH. The four major nucleotides were found to be inert under these conditions. The iodate oxidation appears to be more specific than the previous conversion methods reported, and has the advantage that it does not affect the chargeability of most tRNA.

Tertiary structure in E. coli tRNA Arg and tRNA Val.

300 MHz proton NMR has been used to demonstrate that both Escherichia coli tRNA Arg and tRNA(1) Val have a tertiary structure base pair between 4-thiouridine (s-4U) at position 8 and A at position 14. Formation of this s-4U(8)-A(14) base pair leads to a very low field (14.8 ppm) resonance in the spectra of both molecules. When s-4U is converted to U the 14.8 resonance is replaced by a new resonance which appears at 14.3 ppm. The presence of the tertiary structure s-4U(8)-A(14) base pair greatly constrains the folding of these tRNAs in solution.

NMR investigation of the effect of selective modifications in the anticodon loop on the conformation of yeast transfer RNA-Phe.

Changes in the 300 MHz proton NMR spectrum of yeast tRNA-Phe induced by the removal of the Y-base from the anticodon loop are reversed when the dye proflavine is incorporated in its place. These observations correlate with our earlier interpretation of the NMR data that removal of the Y-base causes a change in the conformation of the anticodon stem. Such changes in stem conformation may in part be responsible for the differences in the biochemical properties of tRNA-Phe, tRNA-PhePF and tRNA-Phe-Y.

Mechanisms for the enhanced thermal stability of a mutant of transcription factor 1 as explained by (1)H, (15)N and (13)C NMR chemical shifts and secondary structure analysis.

A variant of the bacteriophage SPO1-encoded transcription factor 1 (TF1) with two site-specific mutations (E15G and T32I) was shown to be more thermally stable and bind DNA more tightly compared to the wild-type protein. In order to understand the biochemical mechanisms underlying these properties, we are engaged in determining the solution structures of this mutant alone and in complex with DNA using nuclear magnetic resonance (NMR) spectroscopy. The first phase of this project is reported here, as we have completed most of the backbone and sidechain sequential NMR assignments of the mutant protein, TF1-G15/I32. Insights derived from the (1)H, (15)N and (13)C chemical shifts and from the secondary structure analysis provide us with an explanation for the noted increase in thermal stability of TF1-G15/I32. Compared to the structure of the wild-type protein, the beta-sheet and the C-terminal helix remain largely unaffected whereas the mutations cause great changes in the first two helices and their enclosed loop. Specifically, we have found that the second helix is extended by one residue at its N-terminus and rotated in a way that allows Ala-37 to interact with Tyr-94 of the C-terminal helix. The loop has been found to become more rigid as a result of hydrophobic interactions between the flanking second and first helices and also between the second helix and the loop itself. Furthermore, the T32I mutation allows tighter packing between the second helix and the beta-sheet. Collectively, these changes contribute to a more tightly associated dimer and hence, to a greater thermal stability.

Effect of environment, conformation, sequence and base substituents on the imino proton exchange rates in guanine and inosine-containing DNA, RNA, and DNA-RNA duplexes.

High-resolution 1H nuclear magnetic resonance in H2O has been used to study the effect of sequence, conformation, environmental factors and base substituents on the exchange behavior of the hydrogen-bonded imino protons of guainine X cytosine and inosine X cytosine base-pairs in DNA, RNA, and DNA-RNA duplexes. The exchange rates were determined by measurement of the spin-lattice relaxation rates of the imino protons as a function of temperature. The exchange was not altered by the presence of high concentrations of salt, and the inability of phosphate to catalyze the exchange indicates that the exchange is limited by formation of a solvent-accessible "open" state. The exchange behavior depends on the duplex conformation and sequence. Exchange from the Z form polymers was orders of magnitude slower than the corresponding duplexes in the B conformation, and the A form RNA duplexes exchanged more slowly than the B form DNA polymers with the same sequence. The exchange behavior of the DNA-RNA hybrids was dependent on whether the purine or the pyrimidine strand contained the deoxyribose sugar. For both the guanine and inosine-containing duplexes, the homopolymer duplexes exchange more slowly than the more stable alternating copolymers. For the alternating duplexes, substitution of cytosine with 5-bromo- or 5-methylcytosine slowed the exchange and increased the activation energy for exchange. The inosine-containing duplexes exchanged more rapidly than the guanosine-containing duplexes, but both showed similar changes in exchange behavior in response to changes in sequence and base substituents. The activation energies for base-pair opening in B form DNA are correlated with the van der Waals contribution to the base-base interaction energy, suggesting that the purine base is partially unstacked in the open state. Using the relaxation measurements to set an upper limit on the exchange rate in poly(dG-dC) and the tritium exchange behavior at low temperature, we find that even though Z-DNA exchanges very slowly, the activation energy is similar to that observed in the A and B form duplexes, suggesting that exchange occurs from a similar open state.

Static disorder and librational motions of the purine bases in films of oriented Li-DNA.

Solid-state 2H nuclear magnetic resonance line shapes have been obtained from folded films of oriented Li-DNA molecules with the purine bases selectively labeled with deuterium at the 8-position. From line shape simulations, the static base tilts as well as the anisotropic motional amplitudes were determined as a function of hydration level and temperature. It was found that the average tilt angle of the bases is close to 0 degrees and at a hydration of ten water molecules per nucleotide the distribution width of tilt angles about this average cannot be larger than 9 degrees (standard deviation). A slightly increased distribution width is observed at low hydration levels. The motional amplitudes are hydration dependent, with the tilting motion ranging from 4 degrees for the driest, up to 15 degrees for the wettest sample, and slightly larger amplitudes are observed for the twisting motion. The amplitude of the twisting motion is unaffected by a temperature decrease down to -60 degrees C, in contrast to the tilting motion that is suppressed at low temperatures.

Solution structure of a mutant of transcription factor 1: implications for enhanced DNA binding.

An NMR solution structure of a mutant of the homodimer protein transcription factor 1, TF1-G15/I32 (22 kDa), has been solved to atomic resolution, with 23 final structures that converge to an r.m. s.d. of 0.78 A. The overall shape of TF1-G15/I32 remains similar to that of the wild-type protein and other type II DNA-binding proteins. Each monomer has two N-terminal alpha-helices separated by a short loop, followed by a three-stranded beta-sheet, whose extension between the second and third beta-strands forms an antiparallel beta-ribbon arm, leading to a C-terminal third alpha-helix that is severely kinked in the middle. Close examination of the structure of TF1-G15/I32 reveals why it is more stable and binds DNA more tightly than does its wild-type counterpart. The dimeric core, consisting of the N-terminal helices and the beta-sheets, is more tightly packed, and this might be responsible for its increased thermal stability. The DNA-binding domain, composed of the top face of the beta-sheet, the beta-ribbon arms and the C-terminal helices, is little changed from wild-type TF1. Rather, the enhancement in DNA affinity must be due almost exclusively to the creation of an additional DNA-binding site at the side of the dimer by changes affecting helices 1 and 2: helix 2 of TF1-G15/I32 is one residue longer than helix 2 of the wild-type protein, bends inward, and is both translationally and rotationally displaced relative to helix 1. This rearrangement creates a longer, narrower fissure between the V-shaped N-terminal helices and exposes additional positively charged surface at each side of the dimer.

1H nuclear magnetic resonance study of the dynamic properties of the B and Z-forms of poly[d(A-br5C).d(G-T)].

Poly[d(A-br5C).d(G-T)], a synthetic polynucleotide with a 50% A-T base composition, undergoes a reversible, highly co-operative transition between the right-handed B and left-handed Z conformations. The latter is stabilized at both elevated temperature and ionic strength. The B and Z-forms of poly[d(A-br5C).d(G-T)] coexist in 4.6 M-NaCl at 45 degrees C. Due to slow exchange, two sets of Tim and Gim resonances are observed and can be assigned to the B and Z conformations (the chemical shifts are, respectively, Tim = 13.4, 14.1 p.p.m. (parts/million); and Gim = 11.9, 12.4 p.p.m.). Measurements of the 1H spin-lattice (R1) and spin-spin (R2) relaxation rates of the exchangeable thymine (Tim) and guanine (Gim) imino protons have been used to probe the internal dynamics of the B and Z-forms of poly[d(A-br5C).d(G-T)] and the mechanism of the B-Z transition. The proton exchange behavior in the B and Z conformations is quite different. At elevated temperature, R1 for both Tim and Gim in the B conformation is dominated by exchange with the solvent, with Tim exchanging more rapidly than Gim. This demonstrates that exchange involves the opening of single base-pairs and that neighboring A-T and G-br5C base-pairs exchange independently of each other. B-form poly[d(A-br5C).d(G-T)] is unusual in that there is an acceleration of the Tim exchange rate with increasing NaCl concentration. Conversion to the Z-form by addition of 4.5 M-NaCl dramatically reduces both the Tim and Gim exchange rates (estimated to be less than 2 s-1 at 70 degrees C). Thus, the G-br5C base-pair and, in particular, the A-T base-pair are stabilized in the Z conformation. By measuring relaxation rates at 45 to 50 degrees C where the B and Z-forms are in equilibrium, we find that the B-Z interconversion rates are less than two per second. In the B conformation at 25 degrees C, the dipolar contributions to the imino proton relaxation rates are about one-third of those expected on the basis of a rigid rod model for 65 base-pair fragments, a difference we assign to large amplitude (30 degrees high frequency (less than 100 ns) out-of-plane motions of the bases. Conversion to the Z conformation has little effect on the dipolar contributions to relaxation, i.e. on the internal motions.(ABSTRACT TRUNCATED AT 400 WORDS)

Structure of the Bacillus subtilis phage SPO1-encoded type II DNA-binding protein TF1 in solution.

The solution structure of a type II DNA-binding protein, the bacteriophage SPO1-encoded transcription factor 1 (TF1), was determined using NMR spectroscopy. Selective 2H-labeling, 13C-labeling and isotopic heterodimers were used to distinguish contacts between and within monomers of the dimeric protein. A total of 1914 distance and dihedral angle constraints derived from NMR experiments were used in structure calculations using restrained molecular dynamics and simulated annealing protocols. The ensemble of 30 calculated structures has a root-mean-square deviation (r.m.s.d.) of 0.9 A, about the average structure for the backbone atoms, and 1.2 A for all heavy-atoms of the dimeric core (helices 1 and 2) and the beta-sheets. A severe helix distortion at residues 92-93 in the middle of helix 3 is associated with r.m.s.d. of approximately 1.5 A for the helix 3 backbone. Deviations of approximately 5 A or larger are noted for the very flexible beta-ribbon arms that constitute part of a proposed DNA-binding region. A structural model of TF1 has been calculated based on the previously reported crystal structure of the homologous HU protein and this model was used as the starting structure for calculations. A comparison between the calculated average solution structure of TF1 and a solution structure of HU indicates a similarity in the dimeric core (excluding the nine amino acid residue tail) with pairwise deviations of 2 to 3 A. The largest deviations between the average structure and the HU solution structure were found in the beta-ribbon arms, as expected. A 4 A deviation is found at residue 15 of TF1 which is in a loop connecting two helical segments; it has been reported that substitution of Glu15 by Gly increases the thermostability of TF1. The homology between TF1 and other proteins of this family leads us to anticipate similar tertiary structures.

Interactions of antitumor drugs with natural DNA: 1H NMR study of binding mode and kinetics.

1H NMR has been used to study the interactions of over 70 clinical and experimental antitumor drugs with DNA. Spectra of the low-field (H-bonded imino proton) resonances of DNA were studied as a function of drug per base pair ratio. From the spectral changes observed, it was possible to distinguish three modes of drug binding (intercalation, groove binding, and nonspecific outside binding), to determine the kinetics of drug binding (approximate lifetime of the bound drug), and, in favorable cases, to determine the specificity of the drugs for A X T or G X C base pairs. This method is a useful assay for general drug-binding characteristics. For the intercalating compounds there appears to be a correlation between drug-binding kinetics and useful antitumor activity.

Substituent effects on the binding of ethidium and its derivatives to natural DNA.

The binding of eight ethidium derivatives to short (approximately 35 base-pair), random sequence DNA has been investigated using 1H-NMR. At 35 degrees C, all drugs cause upfield shifts of the DNA imino proton resonances characteristic of intercalative binding to DNA, but the line shapes vary significantly with the nature of the drug. The results confirm our previous proposal that removal of the amino group at position-3, but not at position-8, on the parent ethidium shortens the lifetime of the intercalative state (less than 1-2 ms at 35 degrees C). These results suggest that hydrogen-bonding interactions with the 3-NH2 group are involved in stabilization of the drug-DNA complex or that changes in charge distribution that accompany removal of the 3-NH2 group reduce the complex stability. The magnitude of the shift of the drug-DNA spectra indicates a slight preference for binding of the drugs adjacent to G X C base-pairs.

A fluorescence study of the binding of Hoechst 33258 and DAPI to halogenated DNAs.

We have studied the time-resolved and the steady-state fluorescence of the DNA groove binders 4',6-diamidino-2-phenylindole (DAPI) and Hoechst 33258 with the double stranded DNAs poly(dA-dU) and poly(dI-dC) and their halogenated analogs, poly(dA-I5dU) and poly(dI-Br5dC). These studies were prompted by earlier observations that steady-state fluorescence of Hoechst 33258 is quenched on binding to halogenated DNAs (presumably due to an intermolecular heavy atom effect involving the halogen atom in the major groove), and recent studies which clearly point to a binding-site in the minor groove of DNA. Measurements of the time resolved fluorescence decay demonstrate that the fluorescence of Hoechst 33258 is quenched on binding to the halogenated DNAs, in agreement with previous observations. However, quenching studies carried out using the free halogenated bases IdUrd and BrdCyd in solution yielded bimolecular rate constants more than one order of magnitude larger than those expected for an intermolecular heavy atom effect. Moreover, the quenching of the Hoechst 33258 fluorescence was accompanied by an accelerated photochemical destruction of Hoechst 33258. We therefore conclude that the fluorescence quenching observed with halogenated DNAs is probably due to a photochemical reaction involving Hoechst 33258, rather than direct contact of Hoechst 33258 with the halogen substituents in the major groove of the DNA. The fluorescence decay measurements however, do provide clear evidence for at least two different modes of binding. Taking into account the alternating sequences used in this study and the possibility of two different conformations for bound dye, at least four different modes of binding are plausible. Our present data do not allow us to distinguish between these alternatives. The time-resolved fluorescence decays and fluorescence quantum yields of DAPI are not affected by the presence of the heavy atom substituents in the DNA major groove. Based on this observation and earlier reports that DAPI binds in one of the DNA grooves, we conclude that the high affinity sites for DAPI on DNA are located in the minor groove.

Multidimensional NMR spectroscopy of DNA-binding proteins: structure and function of a transcription factor.

The solution structure of a type II DNA-binding protein (DBPII), transcription factor 1 (TF1), has been determined using NMR spectroscopy. A multidimensional, heteronuclear strategy was employed to overcome assignment ambiguities due to resonance overlap and broadened crosspeaks. This approach involved the use of selectively deuteriated, 13C- and 15N-labeled samples and 'isotopic heterodimers' to distinguish between intra- and intermonomeric NOEs. A comparison with the crystal structure and NMR analysis of the E. coli HU protein suggests that other homologous proteins in this family will possess similar tertiary structures. This NMR strategy is applicable to the study of other proteins and their biomolecular complexes.

Secondary structure of T4 gene 33 protein. Fourier transform infrared and circular dichroic spectroscopic studies.

The secondary structure of bacteriophage T4 gene 33 protein (gp33) has been quantitatively examined by using Fourier transform infrared (FT-IR) and circular dichroism (CD) spectroscopy. Resolution enhancement techniques, including Fourier deconvolution and derivative spectroscopy were used to quantitate the spectral information from the amide I bands. The relative areas of these component bands indicate 21% alpha-helix, 25% beta-sheet, 34% turn, 12% random coil and 8% other undefined structures in gp33. An analysis of the CD spectrum of gp33 at the same pH and temperature revealed 19% alpha-helix, 25% beta-sheet, 13% turn and 43% random coil structures. The possible reasons for the discrepancies in estimates of the contributions to the secondary structure from turns and random coils are discussed.