Research

Biocatalysis

Biological systems are becoming increasing popular in synthetic organic chemistry.  These systems, particularly enzymes, have the potential to catalyze reactions of specific reactants with high enantio- and stereospecificity.  Enzymes are also advantageous because their reaction are typically mild, i.e. buffered saline at room temperature.  Our research is currently looking at a specific class of enzymes, the 2-keto-3-deoxy-6-phosphogluconate (KDPG) aldolases, for use in preparative scale synthetic organic chemistry.

    The KDPG aldolases catalyze in vivo the reversible cleavage of KDPG into glyceraldehyde-6-phosphate and pyruvate.  Under in vitro conditions, KDPG aldolases will accept nonnative substrates to give a single aldol reaction product.  For example, the KDPG aldolases for E. coli with catalyze the aldol reaction of 2-pyridine carboxaldehyde and pyruvate to give a single, enantiomerically pure product on a preparative scale.  Using another enzymatic step and without using any protecting groups, the aldol reaction product can be taken to produce the amino acid portion of the Nikkomycin K series of antifungal agents.
 
    One problem with enzymatic systems is their high substrate specificity.  Not all aldehydes and substituted pyruvates are accepted as substrates for the KDPG aldolases currently under investigation.  Our research centers on discovering aldolases that will accept substrates which are valuable synthetically.  This research is occuring by two separate methods.  The first method involves investigating aldolases from different organisms, such as P. putida and H. pylori.  The second method involves changing the substrate specificity of an aldolase.  We are currently attempting to change the substrate specificity of the KDPG aldolase from E. coli through the use of directed evolution.  Directed evolution involves making random mutations within the enzyme through the use of low-fidelity, error-prone PCR or with DNA shuffling.  These mutants are then screened for increased catalytic activity toward a particular substrate.  By selecting the appropriate mutants, we should be able to increase the utility of the KDPG aldolases in synthetic organic chemistry.

Protein-Carbohydrate Interactions

Many recognition events involve the interactions between an extracellular carbohydrate epitope and a receptor protein.  These processes have been observed in such diverse biological processes as fertilization, viral and bacterial infection, inflammatory responses, and the maintenance of lung integrity.  Although protein-carbohydrate interactions appear to be important in many biological events, they appear relatively weak when examined experimentally:  dissociation constants range from millimolar to micromolar in most cases.  Considering the importance of these interactions, it is not understood why the monomeric interactions are so weak.  While many different approaches have been used to better understand important factors in protein-carbohydrate interactions, our research efforts have focussed on multivalency and protein engineering.  Detailed below are the three subcategories of the protein-carbohydrate interactions project:  multivalent ligands in the model system concanavalin A (con A), multivalent inhibitors of Shiga-like toxin 1B (SLT1B), and directed evolution of carbohydrate-binding protein 35 (CBP-35).

Concanavalin A:

 For the purposes of our ligand-based studies, the protein used is concanavalin A (con A) which is a legume lectin, which is readily obtained from the Jack bean (Canavalia ensiformis).  Con A, which can exist as a dimer or tetramer depending on solution pH, binds specifically to mannose.  While it is not a medically relevant system, con A-mannoside interactions have been studied and characterized using many different techniques.  The binding site of con A has been shown to be a shallow cleft formed by the loop regions of the protein, similar to other legume lectins.  There are crystal structures available for con A both free and bound to mannosides, showing the binding sites and minimal epitopes that are bound.   In addition to a wealth of structural information, numerous thermodynamic studies have been carried out on this particular lectin.  A ligand-modifying strategy for creating glycodendrimers has been utilized to determine the effect of carbohydrate valency on association.  The synthesized glycodendrimers have defined valencies and known structures.  Information gathered from the different series (propyl, diglycine, and tetraethylene glycol based) will be pooled with previous observations from other structurally differentiated series to obtain a better understanding of valency effects with regard to binding phenomena.  Shown to the left is the x-ray crystallography structure of con A complexed with alpha-methyl-mannose as determined by James Naismith and co-workers (Acta Crystallogr.,Sect.D 50 pp. 847 (1994)). 
 
 

Shiga-like Toxin 1:

The Shiga-like toxin 1 (SLT1), a member of the two-component bacterial toxins, adheres to the surfaces of human kidney cells through binding interactions between the protein's pentameric 'B' binding subunit and the glycolipid globotriaosyl ceramide, expressed on cell surfaces.  The adherence of the toxin to the cell surface is followed by endocytosis of the toxic warhead 'A' subunit; the 'A' subunit disrupts protein synthesis on the 28 S rRNA subunit by acting as an N-glycosidase.  The resulting damage to the kidneys manifests itself as hemorrhagic colitis and in severe cases hemolytic uremic syndrome, for which there is currently no commercially available treatment.  Therapeutics based on inhibiting the toxin binding are an attractive yet elusive goal.  However, before such therapeutics can be developed a greater understanding of the driving forces of behind high affinity interactions must be gained.  Toward that end several members of the lab have been synthesizing multivalent ligands of various types in order to determine important aspects.  Shown to the right is the x-ray crystallography structure of SLT1B complexed with with a synthetic Pk-trisaccharide ligand as determined by Randy Read and co-workers (Biochemistry 37 pp. 1777 (1998)). 

Carbohydrate-Binding Protein 35:

A combination of error-prone PCR and DNA shuffling will be used to generate a library of CBP-35 mutants which can be displayed on filamentous phage and subsequently selected for affinity or selectivity through the use of a variety of affinity resins.  Mutants that show particularly interesting characteristics will be isolated and the protein fully analyzed by calorimetric binding studies to determine dH, KB, and dCp.  Thermodynamic solvent isotope effect studies, ELISA and crystallographic experiments may also be carried out in order to better characterize the underlying basis for binding differences.  Shown to the left is the x-ray crystallography structure of galectin-3 (the human homologue of CBP-35) complexed with N-acetyllactosamine as determined by J. Rini and co-workers (J.Biol.Chem. 273 pp. 13047 (1998)). 
 
All structures shown on this page are from the Protein Data Bank and were originally generated using Molscript and Raster 3D.