Structure and Function of the Burkholderia cepacia 2,2-dialkylglycine decarboxylase. This remarkable enzyme is a pyridoxal 5'-phosphate dependent aminotransferase that catalyzes sequential decarboxylation of simple 2-methyl-substituted amino acids, followed by amino group transfer to pyruvate or α-ketobutyrate. Currently we are investigating several aspects of this enzyme including (1) structure determination and kinetics of inactivation by optically active 2-trifluoromethylalanine mechanism based inhibitors, and (2) studies of the role of quaternary structure in regulation and catalysis.

Application of Functional Genomics to Soil Amino Acid Metabolism. We are currently cloning dgd genes (dgdA, which encodes the dialkylglycine decarboxylase, and dgdR, which encodes a specific DNA binding regulatory protein) from various bacteria that utilize aminoisobutyrate (Aib) as a sole nitrogen source. Some of these are unknown soil isolates, and others are well known species containing orthologues of the B. cepacia dgd genes. Over 20 different species that utilize Aib have been identified by16S rRNA gene sequencing. Do all these organisms contain dgd-type genes? If so, how similar are the encoded proteins to known proteins, and can we use their sequence variations to help build 3D models of these proteins?

Computational Chemistry. We have undertaken several computational studies based on student projects in undergraduate lecture and lab classes. These take advantage of UAF's excellent facilities for computational chemistry, including a 2312-processor Sun supercomputer at the Arctic Region Supercomputing Center. One ongoing project concerns the properties of a carboxylic acid "bimolecule", a complex of formic acid and trifluoroacetic acid whose gas-phase vibrational spectrum has been obtained in our lab. This complex contains doubly-H-bonded molecules similar to those found in a DNA base pair. One question we are pursuing is, Do H's in complexes like this exchange between acid molecules? And if so, how fast? In another project related to the possible occurrence of low-barrier hydrogen bonds in enzyme active sites, we are carrying out density functional theory studies on various 1-methylimidazole-carboxylic acid hydrogen bonded complexes. These are thought to be related to histidine-aspartic acid complexes in enzymes, however the detailed structure and vibrational properties of the complexes are not known.

1. J. W. Keller, "Lewis Acid Catalyzed Diels-Alder Reaction of Carvone with Isoprene. Using Two-Dimensional NMR and Molecular Modeling to Solve a Stereo- and Regiochemical Puzzle", The Chemical Educator, 11, 262-266 (2006).
2. E. J. Fogle, See-Tarn Woon, J. W. Keller, and M. D. Toney, "Role of Q52 in Catalysis of Decarboxylation and Transamination in Dialkylglycine Decarboxylase", Biochemistry, 44, 16392-16404 (2005).
3. J. W. Keller, "The Formic Acid-Trifluoroacetic Acid Bimolecule. Gas-Phase Infrared Spectrum and Computational Studies", Journal of Physical Chemistry A, 108, 4610-4618 (2004).

Dr. Keller, shown with a space-filling model of DNA, is a leader in the use of molecular modeling and computational chemistry in undergraduate education.

• B.S. 1968, The Ohio State University (Chemistry)
• Ph.D. 1976, University of Wisconsin-Madison (Chemistry)

phone: 907.474.6042

fax: 907.474.5640

email: jwkeller@alaska.edu

address: John W. Keller Department of Chemistry & Biochemistry University of Alaska Fairbanks Fairbanks, AK 99775-6160

JK at the col between Blackcomb and Horstman glaciers, Whistler-Blackcomb ski area, Whistler, B.C. 3-11-2009