Research - Summary of Recent Findings The use of crystalline DNA has made major advances possible. One reason for this is that the samples are exceptionally well defined (known to atomic resolution) and homogeneous. Another is that the sample structure is exceptionally reproducible, greatly reducing standard errors. Our progress reflects these advantages. Two milestones were reached over the past 5 years. First, free radical yields in DNA were measured with an accuracy that surpassed our expectations (and all previous efforts); using crystalline DNA, the variance for relative yields is less than +/- 10%. Second, we were able to measure the yields of end products on the same samples as were used for EPR measurement, and with comparable accuracy. These advances have made it possible to bridge between the wealth of information on free radical events obtained by EPR and the stable products terminating the free radical reactions. Five years ago the predominant view was that essentially all the holes formed by ionization of the sugar phosphate backbone transferred into the base stack, ending up on guanine. Our EPR studies and product analysis on crystalline DNA show that this is not the case. A large fraction (about 1/2) of the holes initially formed in the sugar-phosphate backbone become fixed on the sugar. This solves the long standing conundrum as to how direct ionization causes strand breaks under conditions where all the holes transfer to guanine and the guanine radical is not a precursor to strand scission. The answer is that all the holes do not transfer to guanine. Five years ago it was widely believed that hole/electron mobility is much larger between stacked bases than between strands and effectively zero between separate duplexes. We now know that hole transfer (and most likely electron transfer) is intrastrand, interstrand, and intermolecular. Through our findings and the work of numerous laboratories around the world, we have formulated a working model that moves us significantly closer to achieving our long term goal of being able to predict, a priori, the type, yield, and spatial distribution of radiation damage in DNA in vivo. The model makes predictions that can be, and need to be, tested. Our predictions begin with the chemical physics of track expansion, in which the electron and hole move by tunneling and hopping. These two mechanisms of transfer are different at a fundamental level. Tunneling rates are not governed by temperature and are strongly dependent on distance. Hopping rates are governed by an activation energy and the rates are temperature dependent. We propose that hopping rates depend on reversible proton transfer rates. This can be tested using DNA crystals, in which we take advantage of a precise knowledge of distances and the ability to vary temperature over a wide range. Our predictions extend to the point in time where the expanded track becomes fixed as stable products. For example, the prediction that reductive damage will occur at cytosine with equal or higher probability than thymine deviates from prior expectations. Unexpected is the proposition that the ratio of sugar-phosphate lesions to base lesions is about one-to-one. Our on going research projects test these predictions.