Site-directed mutagenesis of the bacterial metalloamidase UDP-(3-O-acyl)-N-acetylglucosamine deacetylase (LpxC). Identification of the zinc binding site.

UDP-3-O-(acyl)-N-acetylglucosamine deacetylase (LpxC) catalyzes the second step in the biosynthesis of lipid A in Gram-negative bacteria. Compounds targeting this enzyme are proposed to chelate the single, essential zinc ion bound to LpxC and have been demonstrated to stop the growth of Escherichia coli. A comparison of LpxC sequences from diverse bacteria identified 10 conserved His, Asp, and Glu residues that might play catalytic roles. Each amino acid was altered in both E. coli and Aquifex aeolicus LpxC and the catalytic activities of the variants were determined. Three His and one Asp residues (H79, H238, D246, and H265) are essential for catalysis based on the low activities (<0.1% of wild-type LpxC) of mutants with alanine substitutions at these positions. H79 and H238 likely coordinate zinc; the Zn(2+) content of the purified variant proteins is low and the specific activity is enhanced by the addition of Zn(2+). The third side chain to coordinate zinc is likely either H265 or D246 and a fourth ligand is likely a water molecule, as indicated by the hydroxamate inhibition, suggesting a His(3)H(2)O or His(2)AspH(2)O Zn(2+)-polyhedron in LpxC. The decreased zinc inhibition of LpxC mutants at E78 suggests that this side chain may coordinate a second, inhibitory Zn(2+) ion. Given the absence of any known Zn(2+) binding motifs, the active site of LpxC may have evolved differently than other well-studied zinc metalloamidases, a feature that should aid in the design of safe antibiotics.

Antibacterial agents that target lipid A biosynthesis in gram-negative bacteria. Inhibition of diverse UDP-3-O-(r-3-hydroxymyristoyl)-n-acetylglucosamine deacetylases by substrate analogs containing zinc binding motifs.

UDP-3-O-(R-3-hydroxymyristoyl)-N-acetylglucosamine deacetylase (LpxC) catalyzes the second step in the biosynthesis of lipid A, a unique amphiphilic molecule found in the outer membranes of virtually all Gram-negative bacteria. Since lipid A biosynthesis is required for bacterial growth, inhibitors of LpxC have potential utility as antibiotics. The enzymes of lipid A biosynthesis, including LpxC, are encoded by single copy genes in all sequenced Gram-negative genomes. We have now cloned, overexpressed, and purified LpxC from the hyperthermophile Aquifex aeolicus. This heat-stable LpxC variant (the most divergent of all known LpxCs) displays 32% identity and 51% similarity over 277 amino acid residues out of the 305 in Escherichia coli LpxC. Although A. aeolicus LpxC deacetylates the substrate UDP-3-O-(R-3-hydroxymyristoyl)-N-acetylglucosamine at a rate comparable with E. coli LpxC, a phenyloxazoline-based hydroxamate that inhibits E. coli LpxC with K(i) of approximately 50 nM (Onishi, H. R., Pelak, B. A., Gerckens, L. S., Silver, L. L., Kahan, F. M., Chen, M. H., Patchett, A. A., Galloway, S. M., Hyland, S. A., Anderson, M. S., and Raetz, C. R. H. (1996) Science 274, 980-982) does not inhibit A. aeolicus LpxC. To determine whether or not broad-spectrum deacetylase inhibitors can be found, we have designed a new class of hydroxamate-containing inhibitors of LpxC, starting with the structure of the physiological substrate. Several of these compounds inhibit both E. coli and A. aeolicus LpxC at similar concentrations. We have also identified a phosphinate-containing substrate analog that inhibits both E. coli and A. aeolicus LpxC, suggesting that the LpxC reaction proceeds by a mechanism similar to that described for other zinc metalloamidases, like carboxypeptidase A and thermolysin. The differences between the phenyloxazoline and the substrate-based LpxC inhibitors might be exploited for developing novel antibiotics targeted either against some or all Gram-negative strains. We suggest that LpxC inhibitors with antibacterial activity be termed "deacetylins."

UDP-3-O-(R-3-hydroxymyristoyl)-N-acetylglucosamine deacetylase of Escherichia coli is a zinc metalloenzyme.

The enzyme UDP-3-O-(R-3-hydroxymyristoyl)-GlcNAc deacetylase (LpxC) catalyzes the committed step in the biosynthesis of lipid A and is therefore a potential antibiotic target. Inhibition of this enzyme by hydroxamate compounds [Onishi, H. R.; Pelak, B. A.; Gerckens, L. S.; Silver, L. L.; Kahan, F. M.; Chen, M. H.; Patchett, A. A.; Stachula, S. S.; Anderson, M. S.; Raetz, C. R. H. (1996) Science 274, 980-982] suggested the presence of a metal ion cofactor. We have investigated the substrate specificity and metal dependence of the deacetylase using spectroscopic and kinetic analyses. Comparison of the steady-state kinetic parameters for the physiological substrate UDP-3-O-(R-3-hydroxymyristoyl)-GlcNAc and an alternative substrate, UDP-GlcNAc, demonstrates that the ester-linked R-3-hydroxymyristoyl chain increases kcat/KM (5 x 10(6))-fold. Metal-chelating reagents, such as dipicolinic acid (DPA) and ethylenediaminetetraacetic acid, completely inhibit LpxC activity, implicating an essential metal ion. Plasma emission spectroscopy and colorimetric assays directly demonstrate that purified LpxC contains bound Zn2+. This Zn2+ can be removed by incubation with DPA, causing a decrease in the LpxC activity that can be restored by subsequent addition of Zn2+. However, high concentrations of Zn2+ also inhibit LpxC. Addition of Co2+, Ni2+, or Mn2+ to apo-LpxC also activates the enzyme to varying degrees while no additional activity is observed upon the addition of Cd2+, Ca2+, Mg2+, or Cu2+. This is consistent with the profile of metals that substitute for catalytic zinc ions in metalloproteinases. Co2+ ions stimulate LpxC activity maximally at a stoichiometry of 1:1. These data demonstrate that E. coli LpxC is a metalloenzyme that requires bound Zn2+ for optimal activity.

Balanced biosynthesis of major membrane components through regulated degradation of the committed enzyme of lipid A biosynthesis by the AAA protease FtsH (HflB) in Escherichia coli.

The suppressor mutation, named sfhC21, that allows Escherichia coli ftsH null mutant cells to survive was found to be an allele of fabZ encoding R-3-hydroxyacyl-ACP dehydrase, involved in a key step of fatty acid biosynthesis, and appears to upregulate the dehydrase. The ftsH1(Ts) mutation increased the amount of lipopolysaccharide at 42 degrees C. This was accompanied by a dramatic increase in the amount of UDP-3-O-(R-3-hydroxymyristoyl)-N-acetylglucosamine deacetylase [the IpxC (envA) gene product] involved in the committed step of lipid A biosynthesis. Pulse-chase experiments and in vitro assays with purified components showed that FtsH, the AAA-type membrane-bound metalloprotease, degrades the deacetylase. Genetic evidence also indicated that the FtsH protease activity for the deacetylase might be affected when acyl-ACP pools were altered. The biosynthesis of phospholipids and the lipid A moiety of lipopolysaccharide, both of which derive their fatty acyl chains from the same R-3-hydroxyacyl-ACP pool, is regulated by FtsH.

Antibacterial and anti-inflammatory agents that target endotoxin.

Antibiotic-resistant bacterial infections are a major clinical problem. Lipid A, the active part of lipopolysaccharide endotoxins in Gram-negative bacteria, is an intriguing target for new antibacterial and anti-inflammatory agents. Inhibition of lipid A biosynthesis kills most Gram-negative bacteria, increases bacterial permeability to antibiotics and decreases endotoxin production.

Disruption of the active site solvent network in carbonic anhydrase II decreases the efficiency of proton transfer.

The importance of maintaining the active site water network for efficient proton transfer was investigated by substituting amino acids of varying size at position 65 in carbonic anhydrase II (including four amino acids found in other CA isozymes, F, L, S, and T, and two amino acids that do not occur naturally at position 65, G and H) and measuring the rate constants for the proton transfer reactions in the variant carbonic anhydrases. Intramolecular proton transfer between zinc-bound water and H64 is significantly inhibited by the introduction of bulky residues at position 65; kcat for CO2 hydration decreases up to 26-fold, comparable to the observed decrease in intramolecular proton transfer caused by removal of H64 [Tu, C., Silverman, D. N., Forsman, C., Jonsson, B.-H., & Lindskog, S. (1989) Biochemistry 28, 7913-7918]. Intermolecular proton transfer between protonated H64 and external buffer is also inhibited, although to a lesser degree. Furthermore, an alternative proton transfer pathway, consisting of an active site solvent-mediated proton transfer from zinc-water to imidazole buffer, is inhibited in the A65F, A65L, and A65H CAII variants. Therefore, the active solvent bridge between zinc-bound water and H64 is disrupted by substitutions at position 65. The inhibition of proton transfer reactions correlates with the disruption of the crystallographically observed solvent network in the CA active site and rotation of the proton acceptor, H64 [Scolnick, L. R., & Christianson, D. W. (1996) Biochemistry 35, 16429-16434], suggesting that this solvent network, including water molecules 292, 264, and 369, or a structurally related network, forms the proton transfer pathway in CAII for both intramolecular proton transfer and stimulation of proton transfer in imidazole buffers.