Crystallographic studies on alcohol dehydrogenase from Drosophila
Gordon, Elspeth Jane
MetadataShow full item record
Alcohol dehydrogenases are enzymes that catalyse the oxidation of alcohols to aldehydes and ketones. Alcohol dehydrogenase from the fruit-fly Drosophila, is a particularly efficient enzyme for this reaction and it is a member of the short chain dehydrogenase protein family. Little structural information has been determined for this family, although extensive biochemical and enzymological studies have been carried out. The aim of this project was to determine the crystal structure of the alcohol dehydrogenase enzyme from Drosophila. Chapter one contains an introduction to dehydrogenases, concentrating mainly on the short chain dehydrogenase family and in particular on the Drosophila alcohol dehydrogenase. The short chain family contains more then twenty proteins from a diversity of natural sources, including prokaryotes and humans. The widespread nature of this family indicates its general importance as a protein family, and makes it essential to produce a structural profile for it. The structure of alcohol dehydrogenase from Drosophila together with biochemical data, may be used to validate the reaction mechanisms that have already been proposed. The structure of alcohol dehydrogenase from Drosophila will also provide an unique opportunity for comparisons to be made between a short chain and a medium chain alcohol dehydrogenase. Such a comparison will give insight into the evolutionary origin of the different alcohol dehydrogenase protein types, and may suggest why more than one protein family has evolved to carry out a single detoxification function. Alcohol dehydrogenase from Drosophtla has been crystallized and two crystal forms have been observed. Most crystals were plate-like (form A) and only 0.05 mm in their shortest dimension. Form A crystals diffract X-rays weakly and initial crystal characterization was carried out using the Synchrotron Radiation Source, Daresbury, U.K. Twinning was a severe problem with this crystal form. The second crystal form (form B) was grown in the absence of NAD+ and with DTT added to all crystallization buffers. This form is more suitable for X-ray diffraction studies since the crystals diffract X-rays to better than 2 A resolution. Form B crystals are monoclinic with unit cell dimensions, a = 81.24(6), b = 55.75(4), c = 109.60(7)A and /? = 94.26(9)° and they have two dimers in the asymmetric unit. However, it appears that a smaller rotated cell is also valid at low resolution, with unit cell dimensions, a = 70.60, b = 55.75, c = 65.74 A and f3 = 106.95° and with one dimer in the asymmetric unit. Native data and derivative data have been collected on both crystal forms. Most data were collected on form B crystals and analysis and statistics for this data are included. Several data collection systems were used in the course of this study and the relative advantages of each system are discussed. Attempts to phase the crystallographic data have included both isomorphous replacement and molecular replacement techniques. The molecular replacement study was possible as a result of the recent structural determination of another short chain dehydrogenase, 3a, 20/3-hydroxysteroid dehydrogenase. However, only a partially refined polyalanine model of this dehydrogenase was available and this made phase determination by molecular replacement a challenging problem. A preliminary solution has been refined. This molecular replacement solution gives an R-factor = 38.9% for 9237 reflection. The solution was refined using simulated annealing with data (|F| > 2a) between 3-15A. Two isomorphous derivative data sets have been used to calculate an MIR map. These derivatives are weakly substituted and the strongest site is common to both derivatives. Eight heavy atom positions were refined using maximum likelihood phase refinement. They gave an mean figure of merit 0.43 for 4684 acentric reflections. Both a molecular replacement map and a MIR map have been calculated (to 3.5A resolution). The maps were of poor quality and at this stage an unambiguous chain trace is impossible. However, the quality of the maps has been improved by using density modification techniques. A preliminary interpretation of these maps has been made and is discussed.