The role of lysine 55 in determining the specificity of the purine repressor for its operators through minor groove interactions.

The interaction of the dimeric Escherichia coli purine repressor (PurR) with its cognate sequences leads to a 45 degrees to 50 degrees kink at a central CpG base step towards the major groove, as dyad-related leucine side-chains interdigitate between these bases from the minor groove. The resulting broadening of the minor groove increases the accessibility of the six central base-pairs towards minor groove interactions with residues from PurR. It has been shown that lysine 55 of PurR makes a direct contact with the adenine base (Ade8) directly 5' to the central CpG base-pair step in the high-affinity purF operator sequence. We have investigated the importance of this interaction in the specificity and affinity of wild-type PurR (WT) for its operators and we have studied a mutant of PurR in which Lys55 is replaced with alanine (K55A). Complexes of WT and K55A with duplex DNA containing pur operator sequences varied at position 8 were investigated crystallographically, and binding studies were performed using fluorescence anisotropy. The structures of the protein-DNA complexes reveal a relatively unperturbed global conformation regardless of the identity of the base-pair at position 8 or residue 55. In all structures the combination of higher resolution and a palindromic purF operator site allowed several new PurR.DNA interactions to be observed, including contacts by Thr15, Thr16 and His20. The side-chain of Lys55 makes productive, though varying, interactions with the adenine, thymine or cytosine base at position 8 that result in equilibrium dissociation constants of 2.6 nM, 10 nM and 35 nM, respectively. However, the bulk of the lysine side-chain apparently blocks high-affinity binding of operators with guanine at position 8 (Kd620 nM). Also, the high-affinity binding conformation appears blocked, as crystals of WT bound to DNA with guanine at position 8 could not be grown. In complexes containing K55A, the alanine side-chain is too far removed to engage in van der Waals interactions with the operator, and, with the loss of the general electrostatic interaction between the phosphate backbone and the ammonium group of lysine, K55A binds each operator weakly. However, the mutation leads to a swap of specificity of PurR for the base at position 8, with K55A exhibiting a twofold preference for guanine over adenine. In addition to defining the role of Lys55 in PurR minor groove binding, these studies provide structural insight into the minor groove binding specificities of other LacI/GalR family members that have either alanine (e.g. LacI, GalR, CcpA) or a basic residue (e.g. RafR, ScrR, RbtR) at the comparable position.

The stereospecificity of the ferrous-ion-dependent alcohol dehydrogenase from Zymomonas mobilis.

Alcohol dehydrogenase from Zymomonas mobilis has been found to transfer the pro-R hydrogen of NADH to acetaldehyde. This is the first report of the stereospecificity of a dehydrogenase in the mechanistic and structural class of Fe2+-dependent alcohol dehydrogenases and offers an opportunity to expand mechanistic hypotheses relating stereospecificity, reaction mechanism and reaction thermodynamics in dehydrogenases.

The stereospecificities of seven dehydrogenases from Acholeplasma laidlawii. The simplest historical model that explains dehydrogenase stereospecificity.

Stereospecificities are reported for seven dehydrogenases from Acholeplasma laidlawii, an organism from an evolutionarily distinct branch of life which has not previously been studied from a stereochemical point of view. Three of the activities examined (alcohol dehydrogenase, lactate dehydrogenase, and alanine dehydrogenase) catalyze the transfer of the pro-R (A) hydrogen from NADH. Four other activities (3-hydroxy-3-methylglutaryl-CoA reductase, glyceraldehyde-3-phosphate dehydrogenase, glucose-6-phosphate dehydrogenase, and NADH oxidase) catalyze the transfer of the pro-S (B) hydrogen from NAD(P)H. The stereospecificity of hydroxymethylglutaryl-CoA reductase is notable because it is the opposite of that of hydroxymethylglutaryl-CoA reductases from yeast and rat. These data are used to derive the simplest historical model capable of explaining available experimental facts.

Structural determinants of stereospecificity in yeast alcohol dehydrogenase.

Replacing Leu-182 by Ala in yeast alcohol dehydrogenase (YADH; alcohol:NAD+ oxidoreductase, EC 1.1.1.1) yields a mutant that retains 34% of its kcat value and makes one stereochemical "mistake" every 850,000 turnovers (instead of approximately 1 error every 7,000,000,000 turnovers in native YADH) in its selection of the 4-Re hydrogen of NADH. Half of the decrease in stereochemical fidelity comes from an increase in the rate of transfer of the 4-Si hydrogen of NADH. The mutant also accepts 5-methylnicotinamide adenine dinucleotide, a cofactor analog not accepted by native YADH. The stereospecificity of the mutant is lower still with analogs of NADH where the carboxamide group of the nicotinamide ring is replaced by groups with weaker hydrogen bonding potential. For example, with thio-NADH, the mutant enzyme makes 1 stereochemical "mistake" every 450 turnovers. Finally, the double mutant T157S/L182A, in which Thr-157 is replaced by Ser and Leu-182 is replaced by Ala, also shows decreased stereochemical fidelity. These results suggest that Si transfer in the mutant enzymes arises from NADH bound in a syn conformation in the active site and that this binding is not obstructed in native YADH by side chains essential for catalysis.

Crystallographic studies of the mechanism of xylose isomerase.

The mechanism of xylose isomerase (EC 5.3.1.5) has been studied with X-ray crystallography. Four refined crystal structures are reported at 3-A resolution: native enzyme, enzyme + glucose, enzyme + glucose + Mg2+, and enzyme + glucose + Mn2+. One of these structures (E.G.Mg) was determined in a crystal mounted in a flow cell. The other structures were equilibrium experiments carried out by soaking crystals in substrate containing solution. These structures and other studies suggest that, contrary to expectation, xylose isomerase may not use the generally expected base-catalyzed enolization mechanism. A mechanism involving a hydride shift is consistent with the structures presented here and warrants further investigation. Additional evidence in support of a hydride shift comes from comparing xylose isomerase with triosephosphate isomerase which is known to catalyze an analogous reaction via an enediol intermediate. Evidence is presented that suggests that aldose-ketose isomerases can be divided into two groups. Phospho sugar isomerases generally do not require a metal ion for activity and show exchange of substrate protons with solvent. In contrast, simple sugar isomerases all require a metal ion and show very low solvent exchange. These observations are rationalized on the basis of the need for stereospecific sugar binding.

The stereospecificity of hydrogen transfer to NAD(P)+ catalyzed by lactol dehydrogenases.

The stereochemistry of hydrogen transfer to NAD(P)+ has been determined for five lactol dehydrogenases. It was found that D-glucose dehydrogenases from Bacillus megaterium and Cryptococcus uniguttulatus and L-rhamnose dehydrogenase from Aureobasidium pullulans are pro-S (B) specific, while D-glucose dehydrogenase from Thermoplasma acidophilum and D-xylose dehydrogenase from procine liver are pro-R (A) specific. The latter two enzymes are the first examples of A-specific dehydrogenases oxidizing aldoses at the anomeric carbon. These findings are discussed in terms of functional and historical models that seek to make predictive generalizations regarding dehydrogenase stereospecificity.

The X-ray structure of the PurR-guanine-purF operator complex reveals the contributions of complementary electrostatic surfaces and a water-mediated hydrogen bond to corepressor specificity and binding affinity.

The purine repressor, PurR, is the master regulatory protein of de novo purine nucleotide biosynthesis in Escherichia coli. This dimeric transcription factor is activated to bind to cognate DNA operator sites by initially binding either of its physiologically relevant, high affinity corepressors, hypoxanthine (Kd = 9.3 microM) or guanine (Kd = 1.5 microM). Here, we report the 2.5-A crystal structure of the PurR-guanine-purF operator ternary complex and complete the atomic description of 6-oxopurine-induced repression by PurR. As anticipated, the structure of the PurR-guanine-purF operator complex is isomorphous to the PurR-hypoxanthine-purF operator complex, and their protein-DNA and protein-corepressor interactions are nearly identical. The former finding confirms the use of an identical allosteric DNA-binding mechanism whereby corepressor binding 40 A from the DNA-binding domain juxtaposes the hinge regions of each monomer, thus favoring the formation and insertion of the critical minor groove-binding hinge helices. Strikingly, the higher binding affinity of guanine for PurR and the ability of PurR to discriminate against 2-oxopurines do not result from direct protein-ligand interactions, but rather from a water-mediated contact with the exocyclic N-2 of guanine, which dictates the presence of a donor group on the corepressor, and the better electrostatic complementarity of the guanine base and the corepressor-binding pocket.

Characterization of crystals of xylose isomerase from Streptomyces violaceoniger.

Crystals of the tetrameric xylose isomerase from Streptomyces violaceoniger have been examined by x-ray analysis. Octahedral crystals with a maximum dimension of 0.7 mm were grown from ammonium sulfate solution. They possess the symmetry of P4(1)2(1)2 or P4(3)2(1)2 space groups, which are crystallographically indistinguishable. The unit cell dimensions are a = b = 140 A and c = 134 A. There is one tetramer of molecular weight 160,000 per asymmetric unit. The crystals diffract to 2.2 A.

Isotopic exchange plus substrate and inhibition kinetics of D-xylose isomerase do not support a proton-transfer mechanism.

The D-xylose isomerase of Streptomyces olivochromogenes is a Mg2+- or Mn(2+)-dependent enzyme that catalyzes the aldose-ketose isomerization of xylose to xylulose or of glucose to fructose. Proton exchange into water during enzyme-catalyzed isomerization of C-2 tritiated glucose at 15, 25 and 55 degrees C shows < 0.6% exchange (the loss of one proton in every billion turnovers). High concentrations of guanidine hydrochloride and extremes of pH had no effect on the amount of exchange detected. Such a low percentage of exchange is inconsistent with a proton-transfer mechanism as the main kinetic pathway for isomerization. 19F NMR experiments showed no release of fluoride after incubation of the enzyme for 4 weeks with 800 mM 3-deoxy-3-fluoroglucose or 3-deoxy-3-fluoroallose (both are competitive inhibitors with Ki values of 600 mM). This result is also inconsistent with a proton-transfer mechanism. A hydride-shift mechanism following ring opening has been proposed for the isomerization. Enzyme-catalyzed ring opening was directly measured by demonstrating H2S release upon reaction of xylose isomerase with 1-thioglucose. D-Xylose isomerase-catalyzed interconversion of glucose to fructose exhibited linear Arrhenius behavior with an activation energy of 14 kcal/mol from 0 to 50 degrees C. No change in rate-determining step occurs over this temperature range. 13C NMR experiments with glucose show that enzyme-bound magnesium or manganese does not interact specifically with any one site on the sugar. These results are consistent with nonproductive binding modes for the substrate glucose in addition to productive binding.

Role of the divalent metal ion in sugar binding, ring opening, and isomerization by D-xylose isomerase: replacement of a catalytic metal by an amino acid.

The distinct roles of the two magnesium ions essential to the activity of D-xylose isomerase from Streptomyces olivochromogenes were examined. The enzyme-magnesium complex was isolated, and the stoichiometry of cation binding determined by neutron activation analysis to be 2 mol of magnesium per mole of enzyme. A plot of Mg2+ added versus Mg2+ bound to enzyme is consistent with apparent KD values of < or = 0.5-1.0 mM for one Mg2+ and < or = 2-5 mM for the second. A site-directed mutant of D-xylose isomerase was designed to remove the tighter, tetracoordinated magnesium binding site (site 1, Mg-1); Glu180 was replaced with Lys180. The stoichiometry of metal binding to this mutant, E180K, is 1 mol of magnesium per mole of enzyme. Ring-opening assays with 1-thioglucose (H2S released upon ring opening) show E180K catalyzes the opening of the sugar ring at 20% the rate of the wild-type, but E180K does not catalyze isomerization of glucose to fructose. Thus, the magnesium bound to Glu180 is essential for isomerization but not essential for ring opening. The X-ray crystallographic structures of E180K in the absence of magnesium and in the presence and absence of 250 mM glucose were obtained to 1.8-A resolution and refined to R factors of 17.7% and 19.7%, respectively. The wild-type and both E180K structures show no significant structural differences, except the epsilon-amino group of Lys180, which occupies the position usually occupied by the Mg-1.(ABSTRACT TRUNCATED AT 250 WORDS)

Natural selection, protein engineering, and the last riboorganism: rational model building in biochemistry.

A detailed study of the chemical behavior of modern catalysts (here, exemplified by dehydrogenases dependent on NAD+) allows us to construct models that distinguish between selected and drifting behaviors in biological macromolecules. These models enable us to manipulate rationally the properties of enzymes, here to design an "acetaldehyde reductase" dependent on NAD+ that is faster than any given us by nature. When applied to the origin of protein catalysis, models that explain the structures of ribo-cofactors (e.g., NAD+) must postulate a metabolically complex breakthrough organism. This means that: (1) The view from the present day back to the truly primeval organism is obscured; it is futile to try to deduce the detailed structure of the first life by examining the behaviors of modern organisms. (2) Riboorganisms dominated life on earth for a long time before translation evolved; indeed, fossils of riboorganisms might already be known. (3) Using organic synthesis, we have expanded the number of bases available for making RNA and making accessible RNA molecules that are likely to be intrinsically better catalysts.