The sugar-binding site of the L-arabinose binding protein, an essential component of the high-affinity L-arabinose uptake system in Escherichia coli, is located deep in a cleft formed by the two domains of the protein. The site was unambiguously identified with the electron-rich substrate analog 6-bromo-6-deoxy-D-galactose in a difference Fourier analysis. The observation that the original structure might have been solved with bound L-arabinose necessitated the synthesis of the heavy-atom analog, its structure consistent with the sugar-binding specificity of the protein. Difference Fourier maps showed a peak which was partially coincident with the "extraneous" density found in the native map. This "extraneous" density was previously attributed to a bound L-arabinose molecule, and its presence accounts for early failures of difference Fourier analyses of crystals soaked in or co-crystallized with L-arabinose to locate the binding site. The location of a C6 substituent by difference Fourier analysis of crystals soaked in solutions of the C6 analog D-galactose and a fit of an L-arabinose molecule to the "extraneous" electron density allowed for a model for L-arabinose binding to the L-arabinose binding protein to be proposed. The L-arabinose molecule binds in a chair conformation with the OH(1) axial ((beta)-anomer). The anomeric hydroxyl is able to hydrogen bond to the main chain oxygen of Asp-89 and to Lys-10, the latter amino acid also forms a hydrogen bond with the hydroxyl at C2. Glu-14 and Asn-205 hydrogen bond with OH(3), and OH(4) hydrogen bonds to Asn-205 and Asn-232. Small angle X-ray scattering experiments indicate that the L-arabinose-binding protein undergoes a substantial conformational change upon the binding of substrate: the radius of gyration of the protein molecule decreases by (TURN)1 A with the addition of L-arabinose. It is proposed that this change in the radius of gyration is due to movement of one domain with respect to the other about a "hinge" deep in the base of the cleft separating the two domains of the bilobal protein.

A variety of 14(alpha)-methyl, 14(alpha)-hydroxymethyl, and 14(alpha)-hydroxy sterols were prepared by chemical synthesis for use in investigations of the mechanism of the enzymatic removal of the 14(alpha)-methyl group (carbon atom 32) of lanosterol during the biosynthesis of cholesterol. A comparison of the in vitro metabolism of {2,4-('3)H}14(alpha)-hydroxymethyl-5(alpha)-cholest-7-en-3(beta)-ol (I) and {2,4-('3)H}14(alpha)-hydroxymethyl-5(alpha)-cholest-7-ene-3(beta),15(alpha)-diol (II) was made using 10,000 x g supernatant fractions of liver homogenates from male rats. Compound I was found to be efficiently converted to tritium-labeled cholesterol upon aerobic incubation with these enzyme preparations supplemented with NAD and NADP; in the absence of a cofactor supplement, a variety of other labeled C(,27) monohydroxy sterols, in addition to cholesterol, were isolated and characterized, including 5(alpha)-cholest-8(14)-en-3(beta)-ol, 5(alpha)-cholest-7-en-3-ol, 5(alpha)-cholesta-8,14-dien-3(beta)-ol, and cholesta-5,7-dien-3(beta)-ol. Compound II, which differed from I only by the addition of a 15(alpha)-hydroxyl group, was not significantly metabolized to cholesterol in either the presence or the absence of a cofactor supplement. These results do not favor the intermediary role of 15(alpha)-hydroxy sterols during the enzymatic removal of carbon atom 32 of 14(alpha)-methyl sterols. Incubation of compound I with a washed microsomal preparation (derived from the livers of male rats) supplemented with NAD and an NADPH generator under aerobic conditions yielded cholesterol and 5(alpha)-cholest-7-en-3(beta)-ol as the principal labeled C(,27) monohydroxy sterol products. However, incubation of I with washed microsomes under aerobic conditions in the absence of added cofactors or in the presence or absence of supplemental NAD and NADP under anaerobic conditions yielded 5(alpha)-cholest-8(14)-en-3(beta)-ol as the major ((GREATERTHEQ) 90%) labeled C(,27) monohydroxy sterol product. The extent of conversion of I to C(,27) monohydroxy sterols was roughly the same in all of the microsomal incubations, regardless of the presence or absence of oxygen and/or a cofactor supplement. These results indicate that oxygen is not required for the enzymatic removal of carbon atom 32 of 14(alpha)-hydroxymethyl-5(alpha)-cholest-7-en-3(beta)-ol. It is further suggested by these results that oxygen is required only for the initial hydroxylation of the 14(alpha)-methyl group during the removal of carbon atom 32 of lanosterol, and that the immediate product of this demethylation process is a sterol containing a (DELTA)('8(14)) double bond. Washed microsomal preparations were also found to catalyze, under anaerobic conditions, the isomerization of 14(alpha)-methyl-5(alpha)-cholest-7-en-3(beta)-ol and 14(alpha)-hydroxymethyl-5(alpha)-cholest-7-en-3(beta)-ol to 14(alpha)-methyl-5(alpha)-cholest-8-en-3(beta)-ol and 14(alpha)-hydroxymethyl-5(alpha)-cholest-8-en-3(beta)-ol, respectively. Two sterols with the "unnatural" cis-C-D ring junction were prepared in labeled form: {2,4-('3)H}5(alpha),14(beta)-cholest-7-en-3(beta),15(beta)-diol (III) and {2,4-('3)H}5(alpha),14(beta)-cholest-7-en-3(beta),15(beta)-diol (IV). Compound III, but not its 15(beta)-hydroxy epimer (IV), was found to be efficiently converted to cholesterol in 10,000 x g supernatant fractions of liver homogenates from male and female rats. After the incubation of III, a number of other labeled products, in addition to cholesterol, were isolated and characterized, including 5(alpha)-cholest-8-en-3(beta),ol, 5(alpha)-cholest-7-en-3(beta)-ol, 5(alpha)-cholesta-8,14-dien-3(beta)-ol, and cholesta-5,7-dien-3(beta)-ol.

The reactions of thirteen isonitriles with hemoglobin, and its isolated subunits were characterized in order to examine the effects of ligand size and stereochemistry on hemoglobin function. All of the isonitriles bind to hemoglobin in a cooperative manner. The exact amount of cooperativity decreases with ligand length but increases with increasing substitution on the alkyl side chain. The affinities of hemoglobin for these compounds exhibit a more complicated dependence on stereochemistry, which has been interpreted in terms of competing hydrophobic and steric interactions. Subtraction of hydrophobic effcts from the observed free energy changes of ligand binding allowed the construction of a rough map of steric hindrances at the heme site. The association and dissociation rates of the high affinity forms of hemoglobin (dimers and isolated chains) have also been examined in terms of bound free energy potentials as above and in terms of kinetic barrier potentials. For the n-series, the observed association rates exhibited a V-shaped dependence on ligand length with methyl and n-hexyl isocyanides binding at the fastest rates and n-propyl the slowest. The large association rates for the longer isonitriles argue against the concept of a small solvent channel limiting the entrance of ligands into the heme site. Rather, it appears that the ligand molecules rapidly dissolve within the hydrophobic interior of the protein and then diffuse to the heme group. Access to the iron atom appears to be attenuated by its coordination geometry and by local steric interactions. Kinetic studies using native hemoglobin showed that the isonitrile ligands can be divided into two mechanistic classes. (1) The larger ligands ((GREATERTHEQ) butyl) exhibit biphasic time courses at all ligand concentrations; this is the result of the (beta) subunits within the low affinity form of hemoglobin reacting with ligand 5-75 times faster than the (alpha) subunits. (2) The smaller isocyanides, exhibit biphasic time courses only at low ligand concentrations where the large differences between the dissociation rate constants for the high and low affinity conformations are expressed; at higher concentrations monophasic or accelerating behavior is observed. A complete description of the binding of each of the isonitriles to human deoxyhemoglobin was obtained by fitting simultaneously an equilibrium curve and a set of kinetic time courses to an expanded 2-state allosteric model. The rate and affinity constants of the R (high affinity) and T (low affinity) states were used to calculate potentials for the ligands at the kinetic barrier and bound to the heme iron atom within the individual (alpha) and (beta) subunits of each protein conformation. The R state free energies were virtually identical to those determined for the isolated (alpha) and (beta) chains, whereas the T state values were 1 - 4 Kcal/mole higher. For (beta) subunits, the barrier and bound potential differences between the R and T states were independent of ligand length and stereochemistry. Thus steric interactions at the binding site exert little or no effect on the expression of cooperativity by this subunit. The differences between the R and T potentials for the (alpha) subunits are larger so that a greater part of the overall observed cooperativity is due to these subunits. The magnitude of these differences decreases with increasing ligand length, which suggests that steric effects are important for the expression of cooperativity by (alpha) subunits.

Mammalian cytochrome oxidase was studied by spectroscopic and kinetic methods to examine the function of cytochromes a and a(,3) in the catalytic mechanism of electron transfer from cytochrome c to molecular oxygen. Various electronic states of cytochrome oxidase were characterized by Reso

The L-arabinose-binding protein, an essential component of the high affinity L-arabinose transport system of Escherichia coli, forms a covalent adduct with the chromophoric probe 2-chloromercuri-4-nitrophenol (MNP) at a single site, cysteine residue 64. The addition of L-arabinose to MNP-modified L-arabinose-binding protein causes an increase in the nitrophenol pK to approximately 8.0 and a blue shift in the wavelength of maximum absorbance of the nitrophenolate species. The rate of reaction of MNP or 5,5'-dithiobis (2-nitrobenzoic acid) (DTNB) with native L-arabinose-binding protein in the presence L-arabinose was found to be inversely related to the concentration of sugar. In addition, L-arabinose-binding activity is dramatically reduced by thiol reagents; an SCN derivative, prepared by reaction of KCN with thionitrobenzoate-modified binding protein was less inhibited than binding protein modified with either MNP or DTNB. The residual binding activity and the apparent dependence upon the size of the group attached to the cysteine suggests that whereas the cysteine is not "essential" it is near the binding site. This conclusion has been confirmed by X-ray crystallographic data. The reactions of specific ligands with the L-arabinose-, D-galactose-,D-maltose- and L-histidine-binding proteins were examined by stopped flow rapid mixing techniques. The change in tryptophan fluorescence that occurs in these reactions is adequately described in all cases by a rapid second order process, no additional phases of fluorescence charge were observed. The resultant time courses were fitted to a simple equilibrium model of binding with association rate constants on the order of 10('7) to 10('8) M('-1)s('-1) and dissociation rate constants in the range of 1 to 100 s('-1). These findings indicate that the ligand-induced conformational changes proposed for these proteins must proceed at a rapid rate. It was also established that purified L-arabinose('-), D-galactose('-), and L-histidine-binding proteins contain approximately stoichiometric amounts of specifically bound substrates which can be readily removed by treatment with concentrated solutions of guanidine-HCl. The presence of glucose bound to the D-galactose-binding protein may account for the heterogeneous galactose equilibrium curves which have been reported in the past by other investigations.

An improved purification scheme for rat kidney (gamma)-glutamylcysteinesynthetase ((gamma)GCSase) was developed. Such a preparation yields only onemajor band and one minor band on polyacrylamide gel electrophoresis.Further, the preparation is devoid of contaminating adenylate kinaseactivity. The K(,eq) for the following reaction was estimated to be 400 atpH 8.2: Glu + (alpha)AB ('M('2+)) (gamma)Glu(alpha)AB + ADP + P(,i) The reaction mechanism of (gamma)GCSase was studied using isotope exchange at equilibrium techniques and inhibition studies. Of the four isotope exchanges measured ((alpha)AB()(gamma)Glu(alpha)AB, Glu()(gamma)Glu(alpha)AB, ATP()ADP and P(,i)()ATP) only the ATP()ADP exchange did not show a saturable effect when substrate concentrations were elevated to three times their Km values. Inhibition studies using L-methionine-S-sulfoximine (MSOX) yielded results which showed MSOX was competitive with respect to Glu, noncompetitive with respect to (alpha)AB and uncompetitive with respect to ATP. Results from both experiments are consistent with a random BC-random RQ reaction mechanism. Only ATP and ADP bind oligatorally to free enzyme. The immune function of mice was studied using two inbred strains of mice maintained on one of three diets: rodent chow, nucleotide-free (NF) diet or nucleotide-free + RNA (NFRNA) diet. Immune response towards allogeneic challenges was measured in vivo using either free allografting of caridao tissue or allogeneic spleen cell injections (popliteal lymph node weight gain assay). In both allogeneic model systems, mice maintained on the NF diet exhibited diminished immune response to the challenges relative to the responses shown by the two control groups (chow and NFRNA). A syngeneic model was also tested. A lymphoid leukemia was injected subcutaneously into the flanks of mice, and tumor growth and recipient mouse mortality were monitored. Mice maintained on a NF diet survived longer than mice in the two control groups. Tumor growth in mice fed a NF diet was slower than tumor growth in mice fed either control diet.

This project has focused on the effects of incorporation of a synthetic analog of creatine-P on the energy metabolism of brain under both in vivo and in vitro conditions. Specifically, it has been shown that the derivative of the creatine analog, cyclocreatine (1-carboxymethyl-2-iminoimidazoli

The lactose repressor protein binds specifically to the operator region of Escherichia coli DNA, physically blocking the transcription of the genes coding for the lactose catabolic enzymes. Initial work with tryptic digestion of repressor to produce a tetrameric core protein (inducer binding) and four NH(,2)-terminal peptides (DNA binding) suggested contributions of the protein regions to ligand interactions. Chemical studies, however, suggested that an alteration of this view was required. At low molar ratios of N-bromosuccinimide (NBS) to repressor, cysteine oxidation caused loss of operator DNA binding activity with simultaneous retention of inducer and nonspecific DNA binding activities. The region surrounding cysteine 107 was found to be influential in maintaining intact operator DNA binding function in repressor. Matthews demonstrated specific, inducible DNA binding affinity in core fragment by direct binding studies (Matthews (1979) J. Biol. Chem. 254, 3348). In light of this activity, application of methylation protection techniques using core protein fragment were of interest. The modification pattern with core protein contained components of the uninduced repressor-operator methylation pattern reported previously. These similarities in the pattern were restricted to the central, asymmetric region of the operator sequence. The effects of inducer and anti-inducer binding to both core and repressor were reflected by changes in the methylation patterns (the anti-inducer-repressor pattern has not been published previously). Exonuclease III digestion was used to probe the 3'-boundary of DNA-protein complexes, and DNase I treatment to probe the susceptibility to digestion of the operator region in the intact complex. Exposure to exonuclease III in the presence of the tryptic core and intact repressor yielded different 3'-boundaries for the complexes formed. The tryptic core protein protected the operator region from DNase I digestion, though to a lesser degree than repressor and in a more restricted region of the operator DNA. The effects of sugar binding on the availability of DNA to enzymatic action for both complexes were consistent with the expected alterations in affinity. A model for core and repressor binding to operator DNA is proposed.

Part I. Previously the {bis-2,6{1-(2-imidazol-4-ylethylimino)-^ethyl}pyridine}copper(I) cation, {Cu('I)(imidH)(,2)DAP}('+), was found to reversibly bind dioxygen, whereas the {bis-2,6-{1-(2-pyridin-2-ylethylimino)ethyl}pyridine}copper(I) cation, {Cu('I)(py)(,2)DAP}('+), was unreactive toward O(,2). Magnetochemical and resonance Raman experiments were conducted to characterize the {Cu('I)(imidH)(,2)DAP}('+)-dioxygen interaction. {Cu('I)(imidR)(py)DAP}('+), {Cu('I)(imidH)(py)DAP}('+), {Cu('I)(imidR)(,2)DAP}('+), and {Cu('I)(imidR)(imidH)DAP}('+) where R = -CH(,2)(C(,6)H(,4))CH(,3), were synthesized to examine the role of the acidic imidazole proton of {Cu('I)(imidH)(,2)DAP}('+) in the reversible oxygenation process. While {Cu('I)(imidR)(py)DAP}('+) was dioxygen-inactive, {Cu('I)(imidH)(py)DAP}('+) reacted irreversibly with dioxygen. {Cu('I)(imidR)(,2)DAP}('+) and {Cu('I)(imidR)(imidH)DAP}('+) reacted reversibly with dioxygen (1 mole O(,2):2 moles Cu), paralleling the dioxygen reactivity for {Cu('I)(imidH)(,2)DAP}('+). Correlation between the redox potential of the Cu(II) (DBLHARR) Cu(I) couple and the reaction of the Cu(I) center toward O(,2) was proposed. Analogous Cu(II) and Zn(II) complexes of all the Cu(I) species were characterized. The relevance of these copper(I) compounds as the first synthetic copper(I) dioxygen carriers to model the active site function of hemocyanin was discussed. Part II. (mu)-Imidazolato mixed-metal complexes were synthesized^to model the proposed imidazolate-bridged {Cyt.a(,3)('3+)(imid)Cu('2+)}^active site structure of cytochrome oxidase where -J(,(Fe)III(,-Cu)II(,)(, ))>(, )200 cm('-1). The compounds were derived from {Fe('III)(TPP)X} (TPP('2-) = tetraphenylporphyrinato and X = Cl('-) or OSO(,2)CF(,3)('-)) and {M('II)(imidH)(,2)DAP}('2+) or {M('II)(imidH)(py)DAP}('2+) (M = Zn or Cu) to yield species containing the following (mu)-imidazolato cationic cores: {ClFe('III)(imid)Zn('II)}('+) {compound (1)}, {ClFe('III)(imid)Cu('II)}('+) {compound (2)}, {(OSO(,2)CF(,3))Fe('III)(imid)Zn('II)}('+) {compound (3)}, {(OSO(,2)CF(,3))Fe('III)(imid)Cu('II)}('+) {compound (4)}, {ClFe('III)(imid)Zn('II)(py)}('+) {compound (5)}, and {ClFe('III)(imid)Cu('II)(py)}('+) {compound (6)}. The variable-temperature magnetochemical and EPR data obtained on the solid state samples indicated that the iron(III) and copper(II) centers of the (2) and (6) complexes are electronically isolated with -J(,(Fe)III(,-Cu)II(,)) (, )200 cm('-1). Compound (4) is the first synthetic (mu)-imidazolato iron(II)-copper(II) metalloporphyrin complex where strong antiferromagnetic exchange interaction between metal centers has been observed. Reinvestigation of the magnetic properties of the "2Co('II)2Cu('II)" superoxide dismutase derivative confirmed that strong antiferromagnetic coupling {-J(,(Co)II(,-CU)II(,)(, ))>(, )300 cm('-1)}, mediated by a bridging imidazolate moiety, exists between the Co(II) and Cu(II) centers. These results provided evidence supporting the possibility of an imidazolate bridge at the active site of cytochrome oxidase.

The characteristics of the oxidation reaction and of the thermodynamics of electron transfer in xanthine oxidase were analyzed by stopped flow spectrophotometry and room temperature potentiometric titrations. Product formation during the oxidation of xanthine oxidase was examined directly by using cytochrome c peroxidase and cytochrome c as trapping agents for H(,2)O(,2) and O(')(,2) respectively. When fully reduced enzyme is mixed with high concentrations of oxygen, 2 molecules of H(,2)O(,2)/flavin are produced rapidly, while 1 molecule of O(')(,2)/flavin is produced rapidly and another much more slowly. Time courses for superoxide formation and those for the absorbance changes due to the enzyme oxidation were fitted successfully to the mechanism proposed earlier (Olson, J. S., Ballou, D. P., Palmer G., and Massey, V., (1974), J. Biol. Chem. 249, 4363-4382). In this scheme, fully reduced enzyme is reoxidized by a series of steps in which sets of 2, 2, 1 and 1 electrons are removed in succession to form 2 equivalents of H(,2)O(,2) and 2 equivalents of O(')(,2). The observed pH effects on the reaction rates can be attributed to proton binding to the FAD-H(,2)O(,2) adduct. A series of potentiometric titration of xanthine oxidase was carried out at room temperature in the pH range 6.2-9.9. Reduction of the two Fe/S centers was monitored by CD, and that of the FAD and Mo center by EPR. The Fe/S centers behave as centers having a protonable group whose pK(,a) changes with reduction state (E = -344 mV, pK(,o) = 6.4, pK(,r) = 8.1 for Fe/S I; and E = -249 mV, pK(,o) = 6.4, pK(,r) = 8.0 for Fe/S II). The flavin and the two types of molybdenum centers show varying behavior but in all the cases electron addition is accompanied by protonation. The sequence for FAD is: reduction, protonation, reduction, protonation with E(,1) = -398 mV, E(,2) = -240 mV, pK(,1) = 9.5, pK(,2) = 7.4. For "rapid" molybdenum the sequence is: protonation, reduction, protonation, reduction with E(,1) = -369 mV, E(,2) = -301 mV, pK(,1) = 7.9, pK(,2) = 8.4; and for "slow" molybdenum: protonation, reduction, reduction, protonation with E(,1) = -320 mV, E(,2) = -477 mV, pK(,1) = 7.5, pK(,2) = 9.5. Comparison to data obtained previously at cryogenic temperatures (Cammack, R., Barber, M. J., and Bray, R. C., (1976), Biochem. J. 157, 469-478 and Barber, M. J., and Siegel, L. M., (1981) in "Flavins and Flavoproteins", (Massey, V., and Williams, C. H., eds.) in press) showed the centers to have significant temperature dependence. . . . (Author's abstract exceeds stipulated maximum length. Discontinued here with permission of school.) UMI