The cyclization of farnesyl diphosphate and nerolidyl diphosphate by a purified recombinant delta-cadinene synthase.

The first step in the conversion of the isoprenoid intermediate, farnesyl diphosphate (FDP), to sesquiterpene phytoalexins in cotton (Gossypium barbadense) plants is catalyzed by delta-cadinene (CDN) synthase. CDN is the precursor of desoxyhemigossypol and hemigossypol defense sesquiterpenes. In this paper we have studied the mechanism for the cyclization of FDP and the putative intermediate, nerolidyl diphosphate, to CDN. A purified recombinant CDN synthase (CDN1-C1) expressed in Escherichia coli from CDN1-C1 cDNA isolated from Gossypium arboreum cyclizes (1RS)-[1-2H](E, E)-FDP to >98% [5-2H]and [11-2H]CDN. Enzyme reaction mixtures cyclize (3RS)-[4,4,13,13,13-2H5]-nerolidyl diphosphate to 62.1% [8,8,15,15,15-2H5]-CDN, 15.8% [6,6,15,15,15-2H5]-alpha-bisabolol, 8.1% [6,6,15,15,15-2H5]-(beta)-bisabolene, 9.8% [4,4,13,13-2H4]-(E)-beta-farnesene, and 4.2% unknowns. Competitive studies show that (3R)-nerolidyl diphosphate is the active enantiomer of (3RS)-nerolidyl diphosphate that cyclized to CDN. The kcat/Km values demonstrate that the synthase uses (E,E)-FDP as effectively as (3R)-nerolidyl diphosphate in the formation of CDN. Cyclization studies with (3R)-nerolidyl diphosphate show that the formation of CDN, (E)-beta-farnesene, and beta-bisabolene are enzyme dependent, but the formation of alpha-bisabolol in the reaction mixtures was a Mg2+-dependent solvolysis of nerolidyl diphosphate. Enzyme mechanisms are proposed for the formation of CDN from (E,E)-FDP and for the formation of CDN, (E)-beta-farnesene, and beta-bisabolene from (3RS)-nerolidyl diphosphate. The primary structures of cotton CDN synthase and tobacco epi-aristolochene synthase show 48% identity, suggesting similar three-dimensional structures. We used the SWISS-MODEL to test this. The two enzymes have the same overall structure consisting of two alpha-helical domains and epi-aristolochene synthase is a good model for the structure of CDN synthase. Several amino acids in the primary structures of both synthases superimpose. The amino acids having catalytic roles in epi-aristochene synthase are substituted in the CDN synthase and may be related to differences in catalytic properties.

Crystal structures of Escherichia coli glycerol kinase variant S58-->W in complex with nonhydrolyzable ATP analogues reveal a putative active conformation of the enzyme as a result of domain motion.

Escherichia coli glycerol kinase (GK) displays "half-of-the-sites" reactivity toward ATP and allosteric regulation by fructose 1, 6-bisphosphate (FBP), which has been shown to promote dimer-tetramer assembly and to inhibit only tetramers. To probe the role of tetramer assembly, a mutation (Ser58-->Trp) was designed to sterically block formation of the dimer-dimer interface near the FBP binding site [Ormo, M., Bystrom, C., and Remington, S. J. (1998) Biochemistry 37, 16565-16572]. The substitution did not substantially change the Michaelis constants or alter allosteric regulation of GK by a second effector, the phosphocarrier protein IIAGlc; however, it eliminated FBP inhibition. Crystal structures of GK in complex with different nontransferable ATP analogues and glycerol revealed an asymmetric dimer with one subunit adopting an open conformation and the other adopting the closed conformation found in previously determined structures. The conformational difference is produced by a approximately 6.0 degrees rigid-body rotation of the N-terminal domain with respect to the C-terminal domain, similar to that observed for hexokinase and actin, members of the same ATPase superfamily. Two of the ATP analogues bound in nonproductive conformations in both subunits. However, beta, gamma-difluoromethyleneadenosine 5'-triphosphate (AMP-PCF2P), a potent inhibitor of GK, bound nonproductively in the closed subunit and in a putative productive conformation in the open subunit, with the gamma-phosphate placed for in-line transfer to glycerol. This asymmetry is consistent with "half-of-the-sites" reactivity and suggests that the inhibition of GK by FBP is due to restriction of domain motion.

Glycerol kinase from Escherichia coli and an Ala65-->Thr mutant: the crystal structures reveal conformational changes with implications for allosteric regulation.

BACKGROUND: Glycerol kinase (GK) from Escherichia coli is a velocity-modulated (V system) enzyme that has three allosteric effectors with independent mechanisms: fructose-1,6-bisphosphate (FBP); the phosphocarrier protein IIAGlc; and adenosine nucleotides. The enzyme exists in solution as functional dimers that associate reversibly to form tetramers. GK is a member of a superfamily of ATPases that share a common ATPase domain and are thought to undergo a large conformational change as an intrinsic step in their catalytic cycle. Members of this family include actin, hexokinase and the heat shock protein hsc70. RESULTS: We report here the crystal structures of GK and a mutant of GK (Ala65-->Thr) in complex with glycerol and ADP. Crystals of both enzymes contain the same 222 symmetric tetramer. The functional dimer is identical to that described previously for the IIAGlc-GK complex structure. The tetramer interface is significantly different, however, with a relative 22.3 degrees rotation and 6.34 A translation of one functional dimer. The overall monomer structure is unchanged except for two regions: the IIAGlc-binding site undergoes a structural rearrangement and residues 230-236 become ordered and bind orthophosphate at the tetramer interface. We also report the structure of a second mutant of GK (IIe474-->Asp) in complex with IIAGlc; this complex crystallized isomorphously to the wild type IIAGlc-GK complex. Site-directed mutants of GK with substitutions at the IIAGlc-binding site show significantly altered kinetic and regulatory properties, suggesting that the conformation of the binding site is linked to the regulation of activity. CONCLUSIONS: We conclude that the new tetramer structure presented here is an inactive form of the physiologically relevant tetramer. The structure and location of the orthophosphate-binding site is consistent with it being part of the FBP-binding site. Mutational analysis and the structure of the IIAGlc-GK(IIe474-->Asp) complex suggest the conformational transition of the IIAGlc-binding site to be an essential aspect of IIAGlc regulation.

Escherichia coli glycerol kinase: role of a tetramer interface in regulation by fructose 1,6-bisphosphate and phosphotransferase system regulatory protein IIIglc.

Escherichia coli glycerol kinase (EC 2.7.1.30; ATP:glycerol 3-phosphotransferase) is a key element in a signal transduction pathway that couples expression of genes required for glycerol metabolism to the relative availability of glycerol and glucose. Its catalytic activity is inhibited by protein-protein interactions with IIIglc, a phosphotransferase system protein, and by fructose 1,6-bisphosphate (FBP); each of these allosteric effectors constitutes a positive signal that glucose is available. Loss of glucose inhibition of glycerol metabolism was used to screen for regulatory mutants of glycerol kinase after hydroxylamine mutagenesis of the cloned glpK gene. Two mutant enzymes were identified and shown by DNA sequencing to contain the mutations alanine 65 to threonine (A65T) and aspartate 72 to asparagine (D72N). Initial velocity studies show the mutations do not significantly affect the catalytic properties, hence active-site structures, of the enzymes. Both mutations decrease inhibition by FBP; A65T eliminates the inhibition while D72N appears to decrease the affinity for FBP and the extent of the inhibition. However, neither mutation significantly affects inhibition by IIIglc. Gel-permeation chromatography studies show that both of the mutations alter the dimer-tetramer assembly reaction of the enzyme and the effect of FBP in increasing the molecular weight. The effects of the mutations on the assembly reaction are consistent with the locations of these two amino acid residues in the X-ray structure, which shows them to be associated with an alpha-helix that constitutes one of the two subunit-subunit interfaces within the tetramer.(ABSTRACT TRUNCATED AT 250 WORDS)

Cation-promoted association of a regulatory and target protein is controlled by protein phosphorylation.

A central question in molecular biology concerns the means by which a regulatory protein recognizes different targets. IIIGlc, the glucose-specific phosphocarrier protein of the bacterial phosphotransferase system, is also the central regulatory element of the PTS. Binding of unphosphorylated IIIGlc inhibits several non-PTS proteins, but there is little or no sequence similarity between IIIGlc binding sites on different target proteins. The crystal structure of Escherichia coli IIIGlc bound to one of its regulatory targets, glycerol kinase, has been refined at 2.6-A resolution in the presence of products, adenosine diphosphate and glycerol 3-phosphate. Structural and kinetic analyses show that the complex of IIIGlc with glycerol kinase creates an intermolecular Zn(II) binding site with ligation identical to that of the zinc peptidase thermolysin. The zinc is coordinated by the two active-site histidines of IIIGlc, a glutamate of glycerol kinase, and a water molecule. Zn(II) at 0.01 and 0.1 mM decreases the Ki of IIIGlc for glycerol kinase by factors of about 15 and 60, respectively. The phosphorylation of one of the histidines of IIIGlc, in its alternative role as phosphocarrier, provides an elegant means of controlling the cation-enhanced protein-protein regulatory interaction. The need for the target protein to supply only one metal ligand may account for the lack of sequence similarity among the regulatory targets of IIIGlc.

Structure of the regulatory complex of Escherichia coli IIIGlc with glycerol kinase.

The phosphocarrier protein IIIGlc is an integral component of the bacterial phosphotransferase (PTS) system. Unphosphorylated IIIGlc inhibits non-PTS carbohydrate transport systems by binding to diverse target proteins. The crystal structure at 2.6 A resolution of one of the targets, glycerol kinase (GK), in complex with unphosphorylated IIIGlc, glycerol, and adenosine diphosphate was determined. GK contains a region that is topologically identical to the adenosine triphosphate binding domains of hexokinase, the 70-kD heat shock cognate, and actin. IIIGlc binds far from the catalytic site of GK, indicating that long-range conformational changes mediate the inhibition of GK by IIIGlc. GK and IIIGlc are bound by hydrophobic and electrostatic interactions, with only one hydrogen bond involving an uncharged group. The phosphorylation site of IIIGlc, His90, is buried in a hydrophobic environment formed by the active site region of IIIGlc and a 3(10) helix of GK, suggesting that phosphorylation prevents IIIGlc binding to GK by directly disrupting protein-protein interactions.

Nucleotide regulation of Escherichia coli glycerol kinase: initial-velocity and substrate binding studies.

Substrate binding to Escherichia coli glycerol kinase (EC 2.7.1.30; ATP-glycerol 3-phosphotransferase) was investigated by using both kinetics and binding methods. Initial-velocity studies in both reaction directions show a sequential kinetic mechanism with apparent substrate activation by ATP and substrate inhibition by ADP. In addition, the Michaelis constants differ greatly from the substrate dissociation constants. Results of product inhibition studies and dead-end inhibition studies using 5'-adenylyl imidodiphosphate show the enzyme has a random kinetic mechanism, which is consistent with the observed formation of binary complexes with all the substrates and the glycerol-independent MgATPase activity of the enzyme. Dissociation constants for substrate binding determined by using ligand protection from inactivation by N-ethylmaleimide agree with those estimated from the initial-velocity studies. Determinations of substrate binding stoichiometry by equilibrium dialysis show half-of-the-sites binding for ATP, ADP, and glycerol. Thus, the regulation by nucleotides does not appear to reflect binding at a separate regulatory site. The random kinetic mechanism obviates the need to postulate such a site to explain the formation of binary complexes with the nucleotides. The observed stoichiometry is consistent with a model for the nucleotide regulatory behavior in which the dimer is the enzyme form present in the assay and its subunits display different substrate binding affinities. Several properties of the enzyme are consistent with negative cooperativity as the basis for the difference in affinities. The possible physiological importance of the regulatory behavior with respect to ATP is considered.

Inactivation of Escherichia coli glycerol kinase by 5'-[p-(fluorosulfonyl)benzoyl]adenosine: protection by the hydrolyzed reagent.

Incubation of Escherichia coli glycerol kinase (EC 2.7.1.30; ATP:glycerol 3-phosphotransferase) with 5'-[p-(fluorosulfonyl)benzoyl]adenosine (FSO2BzAdo) at pH 8.0 and 25 degrees C results in the loss of enzyme activity, which is not restored by the addition of beta-mercaptoethanol or dithiothreitol. The FSO2BzAdo concentration dependence of the inactivation kinetics is described by a mechanism that includes the equilibrium binding of the reagent to the enzyme prior to a first-order inactivation reaction in addition to effects of reagent hydrolysis. The hydrolysis of the reagent has two effects on the observed kinetics. The first effect is deviation from pseudo-first-order kinetic behavior due to depletion of the reagent. The second effect is the novel protection of the enzyme from inactivation due to binding of the sulfonate hydrolysis product. The rate constant for the hydrolysis reaction, determined independently from the kinetics of F- release, is 0.021 min-1 under these conditions. Determinations of the reaction stoichiometry with 3H-labeled FSO2BzAdo show that the inactivation is associated with the covalent incorporation of 1.08 mol of reagent/mol of enzyme subunit. Ligand protection experiments show that ATP, AMP, dAMP, NADH, 5'-adenylyl imidodiphosphate, and the sulfonate hydrolysis product of FSO2BzAdo provide protection from inactivation. The protection obtained with ATP is not dependent on Mg2+. Less protection is obtained with glycerol, GMP, etheno-AMP, and cAMP. No protection is obtained with CMP, UMP, TMP, etheno-CMP, GTP, or fructose 1,6-bisphosphate. The results are consistent with modification by FSO2BzAdo of a single adenine nucleotide binding site per enzyme subunit.

Inactivation of Escherichia coli glycerol kinase by 5,5'-dithiobis(2-nitrobenzoic acid) and N-ethylmaleimide: evidence for nucleotide regulatory binding sites.

Glycerol kinase (EC 2.7.1.30, ATP:glycerol 3-phosphotransferase) from Escherichia coli is inactivated by 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB) and by N-ethylmaleimide (NEM) in 0.1 M triethanolamine at pH 7 and 25 degrees C. The inactivation by DTNB is reversed by dithiothreitol. In the cases of both reagents, the kinetics of activity loss are pseudo first order. The dependencies of the rate constants on reagent concentration show that while the inactivation by NEM obeys second-order kinetics (k2app = 0.3 M-1 s-1), DTNB binds to the enzyme prior to the inactivation reaction; i.e., the pseudo-first-order rate constant shows a hyperbolic dependence on DTNB concentration. Complete inactivation by each reagent apparently involves the modification of two sulfhydryl groups per enzyme subunit. However, analysis of the kinetics of DTNB modification, as measured by the release of 2-nitro-5-thiobenzoate, shows that the inactivation is due to the modification of one sulfhydryl group per subunit, while two other groups are modified 6 and 15 times more slowly. The enzyme is protected from inactivation by the ligands glycerol, propane-1,2-diol, ATP, ADP, AMP, and cAMP but not by Mg2+, fructose 1,6-bisphosphate, or propane-1,3-diol. The protection afforded by ATP or AMP is not dependent on Mg2+. The kinetics of DTNB modification are different in the presence of glycerol or ATP, despite the observation that the degree of protection afforded by both of these ligands is the same.(ABSTRACT TRUNCATED AT 250 WORDS)

Binding of regulatory ligands to rabbit muscle phosphofructokinase. A model for nucleotide binding as a function of temperature and pH.

The binding of nucleoside triphosphates to rabbit muscle phosphofructokinase has been determined in 0.05 M phosphate buffers by changes in intrinsic protein fluorescence and by direct binding measurements. These experiments have been performed over a wide range of pH, temperature, and effector concentration. Quenching of protein fluorescence is shown to measure binding of nucleotides to a site which is not the active site but rather a site responsible for inhibition of the kinetic activity. This site is relatively specific for either ATP or MgATP with free ATP binding about 10-fold more tightly than MgATP. A model to describe binding to this site as a function of pH and temperature is proposed. This model assumes that the apparent affinity for ATP is determined by protonation of two ionizable groups (per subunit) and that ATP binds exclusively to protonated enzyme forms. Several ligands which affect the apparent affinity for nucleotide binding at the inhibitory site act by shifting the apparent pK of the ionizable groups. NH4+ and citrate do not influence nucleotide binding to the inhibitory site. At pH 6.9 in 0.05 M phosphate, low concentrations of MgATP or MgGTP enhance the protein fluorescence due to binding at the active site. The fluorescence studies and direct binding studies show that there is one active site and one inhibitory site per subunit. As described elsewhere (Pettigrew, D. W., and Frieden, C. (1978) J. Biol. Chem. 253, 3623-3627), there is a third nucleotide binding site on each subunit which is specific for cAMP, AMP, and ADP.

Rabbit muscle phosphofructokinase. Modification of molecular and regulatory kinetic properties with the affinity label 5'-p-(fluorosulfonyl)benzoyl adenosine.

The affinity label 5'-p-(fluorosulfonyl)benzoyl adenosine modifies rabbit muscle phosphofructokinase to the extent of one group/subunit. Modification appears to occur at a binding site specific for AMP, cyclic AMP, and ADP, i.e. those adenine nucleotides which are activators under conditions where regulatory kinetic behavior is obtained. The consequences of the modification are consistent with the model proposed previously for correlation between the pK of specific ionizable groups, regulatory kinetic behavior, ligand binding, and the reversible cold inactivation of the enzyme (Frieden, C., Gilbert. H. R., and Bock, P. E. (1976) J. Biol. Chem. 251, 5644-5647). Thus, the modification shifts the apparent pK of the essential ionizable groups from 6.9 to 6.4 at 25 degrees C, with the result that regulatory kinetic behavior at pH 6.9 and 25 degrees C is lost. Furthermore, the apparent affinity of a site (other than the active site) for ATP, as measured by ATP-dependent quenching of intrinsic protein fluorescence at pH 6.9 and 25 degrees C, is decreased by the modification. Regulatory kinetic behavior for both substrates is obtained with the modified enzyme at a lower pH, consistent with the downward shift in the pK of the ionizable groups, but sensitivity to cAMP activation is abolished by the modification. The loss of regulatory kinetic behavior upon modification of sulfhydryl groups does not appear to be the same as that due to modification by the affinity label.