By analysing the intensities of many cross peaks in the spectrum, and paying particular attention to those at the site of the gap in the backbone, we have determined that the backbone is relatively undistorted, as compared to a standard “B DNA” even at the site opposite the gap. There does appear to be a slight over winding of the DNA at this site, giving a slightly greater than average local helix twist angle. We have also used the temperature dependence of the NMR spectrum to examine the melting in these sequences. We find that, as expected, melting occurs first as a separation of the dimer into monomers, followed at higher temperature by opening of the hairpin loop. For sequences with 4 GC base pairs in the overlap region, and six in the stem of the hairpin loop these melting events are separated by about 15°C, with the first transition having slower exchange kinetics than the latter. We have previously carried out NMR studies of the binding of the antibiotic distamycin-A to a specific DNA dodecamer. This drug shows a strong preference for binding at AT rich sequences, AATT in the cases which we have studied in detail. From the NMR data it was possible to position the drug relative to the DNA, and to qualitatively evaluate the degree of distortion of the DNA by the drug. We have now supplemented these measurements by carrying out energy minimization calculations with the AMBER program from UCSF. A general problem with such’ calculations is that the minimization algorithm cannot find a global minimum of energy in a reasonable time. We find that starting with the drug relatively far from the position determined from NMR,nft hydroponic convergence is obtained, but to a drug binding position which is far from that experimentally determined. However, when the calculation is started near the experimentally determined structure the energy obtained upon convergence is lower.
The coordinates obtained through such experimentally guided modelling are better than what can be determined from the NMR data alone. Presently we are trying to extend the measurements to other DNA sequences, and to analyse the differences in binding constant and kinetics of drug binding in light of the detailed structural information we now have about the bound state.First, we expect to be able to use cystines at specific· residues in proteins to predictably fold the remainder of the sequence. The expectations are based upon the observed folding in several naturally occuring peptides, which have common cystine positions. While we have previously assigned the resonances in two of these peptides, and qualitatively modeled their secondary structure, it is important to carry out a more quanititative analysis. To do this we have collected 2D NOE spectra and integrated cross peak intensities to obtain interproton distances for apamin, a small neurotoxic peptide from honey bee venom. With a relatively large number of such distances we have carried out distance geometry calculations. This approach takes the distance estimates,. including estimates for the experimental precision, and computes from these coordinates for structures which are consistant with all of the input data. From the range of structures obtained from repeated calculations we can analyse the precision with which the structure is determined, e.g. its effective “resolution”. In a similar fashion we have begun analysis. of the relationship between structure, function and immuno genicity of protein toxins isolated from sea anemones. These again are related through having common cystine positions, and several other conserved residues. However different toxins from this family are active against different receptors, and do not all show antibody cross reactivity.
We have now fully assigned resonances of the Radianthus paumotensis toxin II, and have established that its only regular secondary structure is beta sheet, with strands connected with a variety of loops and turns. We will now begin structure calculations for this protein, and at the same ti~e have begun assigning a related protein Rp III. We have just obtained the sequence of Rp III through a collaboration with Prof. Ken Walsh at the University of Washington, and have already established that there is a high degree of homology in the structures. From the refined structures for proteins in a related family, such as these, we should be. in a good position to look for the common structural features which give rise to the similar folding of the peptide chain, and yet be able to see the differences which lead to different activities, and immuno genicity. Phosphorus-31 NMR. spectroscopy is evolving into an important means for determining the in vivo concentrations of phosphorylated metabolites and has definite clinical implications and applications. Our previous contributions to this field demonstrated the feasibility of employing implanted radio frequency coils around organs of laboratory animals to permit elliciting the NMR. spectra over long periods to establish base line spectra. Using these devices and techniques we have determined phosphorus exchange reactions in rat hearts and kidneys, in situ, and. have demonstrated that there are pools of metabolic intermediates that are not directly visible -in the NMR. spectra. Comparison of the results from NMR spectroscopy with those obtained from radio labeling studies on Chick Embryo Fibroblasts also showed that there are significant pools of phosphorus not visible in the conventional P-31 NMR. spectrum. Both sets of studies suggest that compartmentation occurs. The invisibility of these pools is assumed to arise because of immobilization of the molecules by cellular macromolecules or organelles. The methods of solid state NMR. spectroscopy are being applied to render visible these solid like species. In particular we use the technique of magic angle sample spinning, together with cross polarization for signal enhancement.
Application of these methods to a large number of biological phosphorylated molecules, for which crystal structure data are avai1~ble, has permitted us to correlate the values of the chemical shielding tensor elements with details of chemical bonding within the phosphate moieties. Vpon application to lyophylized tissue, we observe phosphorus signals attributable to phospholipid head groups. The proton spectra of lyophylized tissue, ellicited with these techniques, are suprisingly rich and exhibit narrow features reminiscent of solution spectra. These narrow features are assigned to hydrocarbon chains of the membrane phospholipids of the tissue. Further support for such an interpretation is provided by the C-13 spectrum of these samples whose features are completely compatible with those of lipid chains .. We interpret these suprising findings to result from the fact that at the temperature of observation, ca room temperature, the membrane phospholipids of the tissue are in the liquid crystal state characteristic of their molecular composition. Normal functioning of the cellular membrane, as exemplified by the fluid-mosaic model,nft system is assumed to require a high degree of dynamic mobility. That we observe such high resolution proton spectra in lyophylized tissue is indeed dramatic support for such a model. The Chemical Bio-dynamics Division of LBL has inaugurated Tritium NMR. spectroscopy in conjunction with the establishment of the National Tritium Labeling Facility. The potential applications of TMR to problems in structural biology and biophysics are very great. They promise to extend the molecular weight range of molecules that can be profitably studied with NMR by several fold, will permit the study of interactions between en~ymes and bound substrates, between receptors and effectors, and between proteins and nucleit acids. This potential derives from the facts that the intrinsic sensitivity of· the triton is some 7% greater than that of the proton, that there will be zero interferring background signals and because the tritium spectrum will be sparse, arising only from those tritons at the sites specifically labeled. Importantly, the abundant protons can be decoupled from the tritons thus reducing their contribution to resonance broadening. We are able to work at the millicurie level of activity. In a first application of TMR to a biological problem, we have observed the conversion of glucose, tritiated at the C-l position, to lactate upon incubation with human erythrocytes. During this metabolism, additional resonances appear transiently indicative of metabolitic intermediates.
The most remarkable aspect of these spectra is the ability to observe the tritiated hydroxyl species at micromolar levels in the presence of 55 molar water without interferring background. During June of 1986 a new 300 MHz NMR spectrometer, specifically configured for optimum utility with TMR was installed in the laboratory.Our objective is to develop a molecular model for chemical mutagenesis from in vitro and in vivo studies of replication and transcription of chemically modified DNA templates. Many carcinogenic as well as chemotherapeutic agents cause covalent linkages between complementary strands of DNA. Cross linked DNA is a block to DNA replication which, if unrepaired, constitutes a lethal lesion to the cell. While the subject of DNA cross link repair has been an area of intensive study, the molecular events of this process have not been well characterized. Genetic studies of E. coli have demonstrated that ABC excision nuclease, coded for by the three unlinked genes uvr A, uvr B, and uvr C, plays a crucial role in DNA crosslink repair. To study the molecular events of ABC excision nuclease mediated DNA crosslink repair we have engineered a DNA fragment with a psoralen-DNA interstrand crosslink at a defined position, digested this substrate with the pure enzyme, and analyzed the reaction products on DNA sequencing gels. We find that the excision nuclease a) cuts only one of the two strands involved in the crosslink, cuts the crosslink by hydrolyzing the ninth phosphodiester bond 5′ and the third phosphodiester bond 3′ to the cross linked furan-side-thymine, and c) does not produce double strand breaks at any significant level. We have constructed a substrate for the ABC excision from 6 oligomers which were then ligated together to form a 40 bp DNA fragment containing the central 8-mer, TCGT*AGCT, in which the internal thymine is modified on the 3′ side by a proralen derivative 4′-hydroxymethyl- 4,5′ ,8-trimethylpsoralen . Both pyrone and furan monoadducts have been isolated, the latter of which reacts with the thymine in the opposite DNA strand to form a crosslink. The cross linked 40 bp DNA fragment was then purified on a denaturing polyacrylamide gel. The cross linked DNA was terminally labelled at either the 5′, 3′, of both termini, digested with ABC excision nuclease and the reaction products where analyzed on a DNA sequencing gel.The genome of human cells contains approximately to 9 nucleotide pairs organized into a particular sequence. The faithful replication of this amount of information into each daughter cells is obviously a formidable task. A clue as to how the· genome sequence is normally maintained in such a highly organized state during DNA replication is to learn something about the factors that destabilize the replication of the genome. Our research program centers its activity on the hypothesis that much of the control of genome replication takes place at the level of initiation of DNA replication within sections of the genome. We are interested in understanding what cellular factors regulate this initiation, how this regulation breaks down in various disease states, and how external environmental stresses can lead to aberrant initiation of DNA replication resulting in increased gene copy number. Because the human genome is so large and the mechanisms regulating its replication likely to be so complex, we have attempted to develop model systems which can help us understand these processes. One approach we have taken is to study the control of oncogenes thought to be involved with regulating the commitment of cells to DNA synthesis. The second study we have been carrying out is to study the aberrant initiation of DNA synthesis that results during gene amplification, and finally we have been investigating the effect of chemical carcinogens on DNA replication and mutation. We are investigating the involvement of protooncogene sequences in the regulation of initiation of DNA synthesis in cells growing in culture. Our hypothesis is that these sequences code for components the cells need to traverse the cell cycle and initiate DNA synthesis. We have focussed our attention on a member of the myc family of oncogenes – N-myc. The Nmyc oncogene is amplified and/or expressed at a high level in many cell lines derived from neuronal tumors; non-neuronal cells apparently do not express this gene.