HOW ENZYMES WORK

An understanding of enzyme-catalyzed reaction mechanisms is fundamental to many aspects of medical science, including rational drug design and toxicology. The focus of this laboratory is to decipher the chemical mechanisms of single substrate enzymes using a variety of biochemical and biophysical techniques (e.g., NMR, EPR, MS, kinetics, etc.). Our approach takes advantage of normal phylogenetic variations, as well as site-directed mutagenesis and recently incorporated synthetic gene technology. We are particularly interested in deciphering the roles of metal ions in enzyme catalysis and in understanding the biological chemistry of lead poisoning.

The tetrapyrroles are the pigments of life-making blood red and foliage green. During the past year, our efforts have focused on porphobilinogen synthase (PBGS), a metalloenzyme that catalyzes the formation of porphobilinogen, the monopyrrole precursor of all biological tetrapyrroles (e.g., porphyrin, chlorophyll, vitamin B₁₂). The importance of PBGS to human health is that in trace amounts the environmental pollutant lead, acts as a powerful inhibitor of the enzyme. An increasing body of literature correlates two common alleles for human PBGS with a differential susceptibility to lead poisoning. We have established that metal ion usage in PBGS is not phylogenetically conserved. Thus, although tetrapyrrole biosynthesis is essential to bacteria, plants and other higher organisms, lead inhibition of PBGS appears limited to the animal kingdom. Current studies focus on determining the uniquely related roles of zinc, magnesium, and lead in PBGS from animals, plants, and bacteria.

PBGS is a homo-octameric protein that catalyzes the asymmetric condensation of two molecules of 5-aminolevulinic acid. The only characterized intermediate in the PBGS-catalyzed reaction is a Schiff base that forms between one of the two substrate molecules and a conserved lysine. Mammalian PBGS is established to have four functional active sites. The crystal structure of yeast PBGS shows eight spatially distinct and structurally equivalent subunits and, thus, eight active site lysines. Remarkably, data for E. coli PBGS have been interpreted to support both four and eight active sites.

A unifying hypothesis is that Schiff base formation triggers a conformational change that results in the evolution of asymmetry. In E. coli PBGS, Lys246 is the Schiff base-forming residue. In this study, the catalytically inactive E. coli PBGS mutants K246H, K246M, K246W, K246N, and K246G were prepared by site-directed mutagenesis. Product binding studies with wild type and K246G demonstrate that both forms of PBGS bind porphobilinogen at four per octamer although the latter cannot form the Schiff base from substrate (see Figure 1).

Thus, Schiff base formation does not solely dictate half-sites reactivity. ¹³C NMR was used to view [3,5-¹³C]porphobilinogen bound at the active site of these mutants. Only minor chemical shift differences relative to wild type were observed. Schiff base trapping studies strongly support a stoichiometry of four functional active sites per homo-octamer. Data from other laboratories supports this view, even though some of this data has previously been interpreted in the context of eight active sites.

FIGURE 1. Porphobilinogen binding to E. coli PBGS as determined by equilibrium dialysis; (·) represents wild-type protein and a filled square represents the mutant K246G.

STRUCTURAL AND FUNCTIONAL ANALYSIS OF THE Mg²⁺-DEPENDENT PORPHOBILINOGEN SYNTHASE FROM PSEUDOMONAS AERUGINOSA. JAFFE, in collaboration with FRANKENBERG,^b HEINZ,^b JAHN^b

Pseudomonas aeruginosa PBGS was overexpressed in E. coli and purified as a fusion protein with glutathione-S-transferase (GST). After removal of GST, a molecular mass of 280,000 ± 10,000 with a Stokes radius of 57 Å was determined for native PBGS, indicating a homo-octameric structure of the enzyme. Fusion of GST to the N-terminus and mutation of PBGS N-terminal amino acids changed the oligomeric state and reduced the activity of the enzyme, thus, revealing the importance of this region for oligomerization and activity. EDTA treatment severely inhibited enzymatic activity, which could be completely restored by the addition of Mg²⁺ or Mn²⁺. A pH-dependent stimulatory effect of monovalent cations like K⁺ was also detected. A K_m of 0.33 mM for ALA and a V_max of 60 mmole/h/mg at pH optimum of 8.8 in the presence of Mg²⁺ and K⁺ was found. These properties are similar to those we have revealed through studies of Bradyrhizhobium japonicum PBGS, a Type IV PBGS that uses both a catalytic and an allosteric Mg²⁺ and exhibits a similar K⁺ effect. The technique of pH-dependent kinetics, in combination with protein modification studies, was employed to determine ionizing groups of the active site involved in catalysis. pK_a values of approximately 7.9 (pK₁) and 9.3 (pK₂) were determined which are postulated as ionization of an active site lysine residue and of the substrate during catalysis, respectively.

Human PBGS is a primary target in lead poisoning. Two common alleles encoding human PBGS have been correlated with differential susceptibility of humans to lead poisoning. Because PBGS exhibits a phylogenetic variation in metal ion usage, non-mammalian forms are poor models for lead poisoning and human PBGS is required. Previous expression systems for human PBGS do not yield the quantities of protein needed for thorough physical/chemical and mechanistic studies of the two human PBGS isozymes. We have used synthetic gene technology to rectify this problem. A synthetic gene encoding human PBGS was designed to resemble the highly expressed E. coli hemB gene, which encodes E. coli PBGS, and to remove rare codons that can confound expression of heterologous proteins in E. coli. Recursive PCR was used to construct the synthetic gene (designer gene) with 5' and 3' regions natural to the E. coli hemB gene. The 1.1 kilobase pair PCR product, EJhum, was designed for high level constitutive expression of the designer gene. However, constitutive expression of human PBGS in E. coli proved toxic. EJhum was shortened to contain only the designer gene, which was then placed in a pET plasmid for controlled expression in a T7 expression system. The construct, BLR(DE3)(pMVhum), is a superb expression system for human PBGS.

MODELING THE STRUCTURE OF HOMO-OCTAMERIC HUMAN PORPHOBILINOGEN SYNTHASE. JAFFE, in collaboration with DUNBRACK^§

The determination of the crystal structure of the asymmetric PBGS enzymes (with four functional active sites per homo-octamer) has been problematic. However, during this year, the coordinates were released for yeast PBGS (law5.pdb), which does not show the asymmetry expected for such half-sites reactivity. The yeast structure has been used to prepare a model for homo-octameric human PBGS (Figure 2). One dimer of the octamer is shown in Figure 2A. The TIM-like ab-barrel structure is easily seen for the lower subunit. The two zinc ions of each monomer are shown as balls in the center of the ab barrel. The active site lysine is seen as a ball and stick model jutting into the center of the barrel. Figure 2B shows the octamer, with each subunit shaded differently. The K59N variation that defines the two common alleles for human PBGS is shown on the surface of the octamer, as a space-filled model of the asparagine residue.

Mammalian PBGS has two distinct Zn(II) binding sites (designated ZnA and ZnB) in the active site. The precise roles of these Zn(II) in catalysis remain unclear. Evidence exists that both sites can bind Pb(II), but the site of inhibition of biological activity by Pb(II) is not established. At full occupancy, the homo-octamer binds a total of eight zinc ions with two zincs per active site (presumably in only four of the eight monomers). The enzyme is fully active at half occupancy with zinc supposedly occupying only four of the eight possible ZnA sites. Previous, extended X-ray absorption fine structure (EXAFS) studies on bovine PBGS indicated that the ZnA environment is pentacoordinate with mixed nitrogen/sulfur content, the ZnB site is tetracoordinate with four sulfurs, and product binding does not dramatically change the EXAFS spectra. Recent crystallographic data of yeast PBGS at 2.3 Å resolution show a Zn(II) analogous to ZnB with only three Cys ligands and a second Zn(II), analogous to ZnA, at low occupancy with one His and one Cys ligand. These ligands are as we had predicted. The poor resolution around the ZnA environment is partially due to the existence of a disordered loop region adjacent to the ZnA site. It has been suggested that this loop becomes ordered upon substrate binding and substrate has also been suggested as a bidentate ligand to ZnA. We have carried out EXAFS analysis on recombinant human PBGS and on a mutant protein where the cysteine ligands to ZnB are replaced by alanine. These studies confirm the presence of two distinct Zn(II) environments and will be extended to study the ZnA site more closely.

Paper in press at time of previous report:

SHIMONI, L., CARRELL, H.L., WAGNER, B., KATZ, A., AFSHAR, C., VOLIN, M., JAFFE, E.K., GLUSKER, J.P. Crystallization and preliminary X-ray diffraction studies of E. coli porphobilinogen synthase and its heavy atom derivatives. Acta Crystallogr. D. 54:438-4401, 1998.

^§   Fox Chase researcher

^a   L.W. Mitchell: Chemistry Department, St. Joseph's University, Philadelphia, PA 19131

^b   N. Frankenberg, D. Jahn, D. Heinz: Albert-Ludwigs-Universität, Institut für Organische Chemie und Biochemie, Albertstr. 21, 79104 Freiburg im Breisgau, Germany

^c   M. Sazinsky, J. Ehrenfeld, R.C. Scarrow: Department of Chemistry, Haverford College, Haverford, PA 19041

Illustrations or unpublished data in these reports should not be used without permission of the author.

Fox Chase Cancer Center

Scientific Report 1998