Gregory K. Farber
The Structure and Function of Enzymes [Research at Pennsylvania State University 1990-2000]
Most of the work in the laboratory is aimed at trying to use the models derived from protein crystal structures to understand interesting biological problems. We have begun by examining the motions which occur as an enzyme catalyzes a reaction. Enzymatic reactions usually have several distinct intermediates. These enzyme bound intermediates can have half lives ranging from hours to microseconds. We originally planned to observe these intermediates using the ultrafast Laue method of data collection. However, we have discovered evidence that this method has several significant limitations. Instead of trying to collect data faster than an enzyme can react, we have chosen to turn a kinetic problem into a thermodynamic problem. We do this by adjusting the substrate and product concentrations around a crystal to a sufficiently high level so that all of the active sites in the crystal are occupied. In some cases, different intermediates which precede a rate determining step can be trapped by flowing substrate through a crystal (a steady state experiment) rather than by allowing the system to come to equilibrium. Our work over the past four years at Penn State has been to generalize the method to enzymes with more than one substrate and product.
Since it is not possible to vary the concentration of water in water, trapping enzyme bound intermediates in which water is a substrate or a product is impossible. Many enzymes fall into this category, so we have investigated the possiblity of moving these crystals into nonaqueous solvents. The first project that we have attempted is to trap the tetrahedral intermediate of reaction catalyzed by chymotrypsin. This intermediate has never been observed either spectroscopically or crystallographically using a true substrate. We developed techniques to transfer the crystals without crosslinking into a number of organic solvents. We have shown that chymotrypsin is catalytically active in hexane and that the mechanism has not changed in the organic solvent since it shows the expected burst phase kinetics with p-nitrophenylacetate. We have been able to trap the tetrahedral intermediate (in the peptide bond synthesis direction) in the crystal in hexane.
During the work on chymotrypsin in hexane, we discovered that adding a small amount of isopropanol to either water or to hexane resulted in a doubling of the number of observable waters in the crystal. We have also discovered that organic solvent molecules sometimes displace waters on the surface of the protein. These sites are quite interesting since they seem to be general small molecule binding sites. If that hypothesis is true, it could also be true that these are the sites at which protein unfolding begins in the presence of chemical denaturants. To test this idea, we have solved the structures of both wild type and mutant dihydrofolate reductases in water and in a 2M urea solution. As with the isopropanol, urea immobilizes a number of waters. Unexpectedly, it also decreases the mobility of the protein in the crystal. We have identified a number of urea binding sites, and we are now repeating the experiment with other denaturants. If there are common denaturant binding sites in the two structures, we believe that these sites mark the start of the unfolding reaction. We plan to experimentally test these predictions by making mutants at the denaturant binding sites.
Finally, we are now working on two new structures. These enzymes are involved in either amino acid degradation or synthesis. In both cases, we plan to trap all of the intermediates which occur along the reaction pathway. The two enzymes are aspartate ammonia-lyase and ketol-acid reductoisomerase (also known as acetohydroxy acid isomeroreductase). The structure of aspartate ammonia-lyase is well underway and we are now interpreting electron density maps. For ketol-acid reductoisomerase, we have just obtained crystals which are suitable for data collection.