Substrate and metal ion binding to carbamate kinase: NMR and EPR studies

Metal ion and substrate binding to carbamate kinase from Streptococcus faecalis was studied by nuclear magnetic resonance (NMR) and electron paramagnetic resonance using Mn²⁺ as the paramagnetic probe. The enzyme binds Mn²⁺ weakly (K_D = 0.45 ± 0.05 mm) with a stoichiometry of one per two subunits. However, in the presence of nucleotides, tighter binding of Mn²⁺ was observed with K_D = 44 ± 4 μm in the presence of ADP and K_D = 23 ± 4 μm with ATP present. Proton relaxation rate enhancement studies were conducted on water molecules interacting with ternary enzyme-Mn²⁺-nucleotide and binary enzyme-Mn²⁺ complexes. Mn²⁺ bound to carbamate kinase enhances the proton relaxation rate of water giving a binary enhancement value of ?_b = 9.3 ± 0.4. When enzyme-Mn²⁺ was titrated with ADP or ATP, a bell-shaped titration curve was obtained typical of many other enzyme-Mn²⁺-nucleotide ternary complexes. Computer fits to the titration data gave ternary enhancement values of ?_t^ADP = 14 ± 1 and ?_t^ATP = 19 ± 1. The dissociation constants for Mn-ADP and Mn-ATP binding to carbamate kinase were also obtained from these data analyses and are K₁ = 2.5 ± 0.5 μm and K₁ = 50 ± 8 μm, respectively. Therefore, these data demonstrate the formation of a ternary enzyme-metal-nucleotide bridge complex at the nucleotide substrate site of carbamate kinase. Distance measurements were conducted by NMR techniques with ¹³C-enriched carbamate and demonstrate that carbamate is 4–8 Å from enzyme-bound Mn²⁺. Thus carbamate binds near the metal-nucleotide substrate site of carbamate kinase.     

Stereochemistry of binding of thiophosphate analogs of ATP and ADP to carbamate kinase, glutamine synthetase, and carbamoyl-phosphate synthetase

Thiophosphate analogs of adenine nucleotides were used to establish the absolute stereochemistry of nucleotide substrates in the reactions of carbamate kinase (Streptococcus faecalis), unadenylylated glutamine synthetase (Escherichia coli), and carbamoyl-phosphate synthetase (E. coli). ³¹P NMR was used to determine that carbamate kinase uses the B isomer of Ado-5′-(2-thioPPP) in the presence of Mg²⁺. The stereospecificity of the reaction with carbamate kinase was not reversed by Cd²⁺ suggesting that the metal ion does not bind to the β-phosphoryl group or that both Mg²⁺ and Cd²⁺ bind to the sulfur atom. Carbamate kinase uses both A and B isomers of Ado-5′-(1-thioPP) with Mg²⁺ and Cd²⁺. We have previously reported that carbamoyl-phosphate synthetase uses the A isomer of Ado-5′-(2-thioPPP) at both ATP sites with Mg²⁺ (Raushel et al., 1978J. Biol. Chem.253, 6627). Current experiments show that the stereospecificity is reversed by Cd^2? and that both A and B isomers are used when Zn²⁺ is present. With Ado-5′-(1-thioPPP), the B isomer is used with Mg²⁺, the A isomer with Cd²⁺, and both isomers with Zn²⁺. Neither carbamate kinase nor carbamoyl-phosphate synthetase utilized Co(III)(NH₃)₄ATP as a substrate and thus we can only speculate that the Δ chelate ring configuration is the chelate structure utilized by carbamoyl-phosphate synthetase (based on the analogy between thiophosphate-ATP analogs and Co³⁺-ATP analogs utilized by hexokinase (E. K. Jaffe, and M. Cohn, 1978Biochemistry17, 652). If the sulfur of the β-phosphoryl of Ado-5′-(2-thioPPP) binds to the metal ion with carbamate kinase, then the Δ chelate ring is also used in this enzyme that catalyzes one of the steps in the overall reaction catalyzed by carbamoyl-phosphate synthetase. Glutamine synthetase reacts with the B isomer of both Ado-5′-(2-thioPPP) and Ado-5′-(1-thioPPP) in the presence of Mg²⁺. When Co²⁺ is used with this enzyme the A and B isomers of both thio-ATP compounds are substrates. Co(III)(NH₃)₄ATP is not a substrate for glutamine synthetase. Glutamine synthetase is therefore different from the two previously mentioned enzymes in that it used the opposite A ring configuration for the metal-ATP chelate.     

Kinetic studies of 18O exchange from inorganic phosphate catalyzed by Mg2+-activated unadenylylated glutamine synthetase (E. coli w).

Oxygen-18 exchange out of [¹⁸O]Pi catalyzed by Mg²⁺-activated unadenylated glutamine synthetase from $\text{E.}$$\text{coli}$ was followed by ³¹P-NMR in the presence of the other substrates, ADP and L-glutamine. The pattern of the ${}^{16}\text{O}{}^{18}\text{O}$ in the species P¹⁸O₄, P¹⁸O₃¹⁶O₁, P¹⁸O₂¹⁶O₂, P¹⁸O₁¹⁶O₃, P¹⁶O₄ during the exchange followed a binomial distribution consistent with indiscriminate removal of any of the four oxygens of P_i. The rate constant for ${}^{16}\text{O}{}^{18}\text{O}$ exchange was 410±40 min^?1 while the rate constant for net reaction (ATP formation) was 62±4 min^?1. Thus exchange proceeds ～7 times faster than net reaction, a finding in accord with that of Stokes and Boyer ($\text{J.}$$\text{Biol.}$$\text{Chem.}$ (1976) $\text{251}$, 5558) for the Mn²⁺-activated adenylylated glutamine synthetase. A model for the overall catalytic events first derived from rapid kinetic fluorescence experiments (Rhee and Chock, Proc. Natl. Acad. Sci. USA, (1976) $\text{73}$, 476) was successfully used to fit the oxygen exchange data in this paper.     

Mn-Mn interaction in adenylylated and unadenylylated glutamine synthetase

The distance between the two catalytically important metal ions of glutamine synthetase was determined by electron paramagnetic resonance (EPR). Mn(II) binds more tightly to the n1 site of this enzyme in the presence of methionine sulfoximine and the influence of Mn(II) bound at the n2 site on the EPR spectrum of Mn(II) at n1 was studied. A monotonic increase in the EPR spectrum of Mn(II) was observed at Mn:E (subunit) ratios of 0 to 0.8. After this point as Mn(II) was added to about 1.8 Mn:E, a decrease in the EPR signal was observed. This phenomenon was found for both adenylylated and unadenylylated forms of glutamine synthetase. The data were analyzed using a theory for dipolar electron-electron relaxation and a distance of 10-12 A was computed for the Mn(II)-Mn(II) separation. These data demonstrate that both modified and unmodified forms of glutamine synthetase which have different catalytic activities have a similar spatial relationship between the two catalytic metal ion sites.     

Mitochondrial respiratory chain of Tetrahymena pyriformis The properties of submitochondrial particles and the soluble b and c type pigments

Submitochondrial particles isolated from Tetrahymena pyriformis contain essentially the same redox carriers as those present in parental mitochondria: at pH 7.2 and 22 °C there are two b-type pigments with half-reduction potentials of −0.04 and −0.17 V, a c-type cytochrome with a half reduction potential of 0.215 V, and a two-component cytochrome a₂ with E_m7.2 of 0.245 and 0.345 V.

EPR spectra of the aerobic submitochondrial particles in the absence of substrate show the presence of low spin ferric hemes with g values at 3.4 and 3.0, a high spin ferric heme with g = 6, and a g = 2.0 signal characteristic of oxidized copper. In the reduced submitochondrial particles signals of various iron-sulfur centers are observed.

Cytochrome c₅₅₃ is lost from mitochondria during preparation of the submitochondrial particles. The partially purified cytochrome c₅₅₃ is a negatively charged protein at neutral pH with an E_m7.2 of 0.25 V which binds to the cytochrome c-depleted Tetrahymena mitochondria in the amount of 0.5 nmol/mg protein with a K_D of 0.8 · 10⁻⁶ M. Reduced cytochrome c₅₅₃ serves as an efficient substrate in the reaction with its own oxidase. The EPR spectrum of the partially purified cytochrome c₅₅₃ shows the presence of a low spin ferric heme with the dominant resonance signal at g = 3.28.

A pigment with an absorption maximum at 560 nm can be solubilized from the Tetrahymena cells with butanol. This pigments has a molecular weight of approx. 18 000, and E_m7.2 of −0.17 V and exhibits a high spin ferric heme signal at g = 6.     

Regulation of concentrations of glycolytic enzymes and creatine-phosphate kinase in “fast-twitch” and “slow-twitch” skeletal muscles of the chicken

The distinctive contractile and metabolic characteristics of different skeletal muscle fiber types are associated with different protein populations in these cells. In the present work, we investigate the regulation of concentrations of three glycolytic enzymes (aldolase, enolase, glyceraldehyde-3-phosphate dehydrogenase) and creatine-phosphate kinase in “fast-twitch” (breast) and “slow-twitch” (lateral adductor) muscles of the chicken. Results of short-term amino acid incorporation experiments conducted both in vivo and with muscle explants in vitro showed that these enzymes turnover at different rates and that aldolase turns over 2 to 3 times faster than the other three enzymes. However, these differences in turnover rates were difficult to detect in long-term double-isotope incorporation experiments, presumably because extensive reutilization of labeled amino acids occurred during these long-term experiments. Mature muscle fibers synthesize these four cytosolic enzymes at very high rates. For example, 11 to 14% of the total labeled leucine incorporated into protein by breast muscle fibers was found in the enzyme aldolase. Results of short-term amino acid incorporation experiments also showed that the relative rates of synthesis of the three glycolytic enzymes were about fourfold higher in mature “fast-twitch” muscle fibers than in mature “slow-twitch” ones while the relative rates of synthesis of creatine-phosphate kinase were similar in the two fiber types. The relative rates of synthesis of these four enzymes and cytosolic proteins in general were found to be very similar in immature muscles of both types. More profound changes in the relative rates of synthesis of major cytosolic proteins, including the glycolytic enzymes, occurred during postembryonic maturation of fast-twitch fibers than occurred during maturation of slow-twitch fibers. Our work demonstrates that (1) the synthesis of creatine-phosphate is independently regulated with respect to the synthesis of the glycolytic enzymes in muscle fibers; and (2) the approximate fourfold higher steady-state concentrations of glycolytic enzymes in fast-twitch muscle fibers as compared with slow-twitch fibers are determined predominantly by regulatory mechanisms operating at the level of protein synthesis rather than protein degradation. Our demonstration that more profound changes in the relative rates of synthesis of major cytosolic proteins occur during maturation of fast-twitch fibers as compared with slow-twitch fibers is discussed in terms of the mode(s) of fiber-type differentiation proposed by others.     

Hydrophobic interaction chromatography of myosin fragments: Potential use in purification

Myosin fragments were fractionated on columns of the hydrophobic gel phenyl-Sepharose CL-4B. In the presence of high NaCl concentrations the fragments bound tightly to the columns; they could be eluted by decreasing the ionic strength, by increasing the pH, or by applying various concentrations of ethylene glycol. In myosin subfragment-1 (S-1), the light chains underwent partial dissociation from the heavy chain and bound separately to the column matrix. The order of strength of binding of the various species to the column was heavy chain > A1 light chain > A2 light chain > native S-1 > denatured heavy chain or S-1. Thus the hydrophobic gel appears to be able to differentiate between enzymatically active and inactive S-1. Under appropriate elution conditions it was possible to obtain S-1 preparations depleted from nicked heavy chains and with specific ATPase activities 34–130% higher than those of untreated S-1. When S-1(A2) was fractionated on phenyl-Sepharose a fivefold enrichment of the heavy chain with respect to the light chains was obtained, while the ATPase activity was equal or larger than that of the original S-1, implying that the light chains are not essential for ATPase activity. Thus, it seems that chromatography of S-1 on phenyl-Sepharose is a potentially useful method for obtaining a purified myosin heavy-chain fragment with a high ATPase specific activity.     

Synthesis of long chain acyl-enzyme thioesters by modified fatty acid synthetases and their hydrolysis by a mammary gland thioesterase.

The fatty acid synthetase multienzyme from lactating rat mammary gland was modified either by removal of the two thioesterase I domains with trypsin or by inhibiting the thioesterase I activity with phenylmethanesulfonyl fluoride. The modified multienzymes are able to convert acetyl-CoA, malonyl-CoA, and NADPH to long chain acyl moieties (C₁₆C₂₂), which are covalently bound to the enzyme through thioester linkage, but they are unable to release the acyl groups as free fatty acids. A single enzyme-bound, long chain acyl thioester is formed by each molecule of modified multienzyme. Kinetic studies showed that the modified multienzymes rapidly elongate the acetyl primer moiety to a C₁₆ thioester and that further elongation to C₁₈, C₂₀, and C₂₂ is progressively slower. Thioesterase II, a mammary gland enzyme which is not part of the fatty acid synthetase multienzyme, can release the acyl moiety from its thioester linkage to either modified multienzyme. Kinetic data are consistent with the formation of an enzyme—substrate complex between thioesterase II and the acylated modified multienzymes. The present study demonstrates that the ability of thioesterase II to modify the product specificity of normal fatty acid synthetase is most likely attributable to the capacity of thioesterase II for hydrolysis of acyl moieties from thioester linkage to the multienzyme.     

Liver protein phosphatases: studies of the presumptive native forms of phosphorylase phosphatase activity in liver extracts and their dissociation to a catalytic subunit of Mr 35,000.

Several aspects of the properties of phosphorylase phosphatase in crude rat liver extracts were investigated. Treatment of tissue extracts with either trypsin, ethanol, or urea was found to dissociate phosphorylase phosphatase activity to a form of M_r 35,000. The M_r 35,000 enzyme form was derived from three native enzyme forms. The major cytosolic form of phosphorylase phosphatase had a molecular weight of 260,000 as determined by gel filtration and was dissociated to a M_r 35,000 form by treatment with either ethanol or urea. Treatment of the M_r 260,000 form with trypsin led to its conversion to M_r 225,000 and a M_r 35,000 form. A minor cytosolic form of M_r 200,000 was also present. This minor activity was latent until activated by trypsin treatment and was converted to a M_r 35,000 form by such treatment. The third form was found to chromatograph as a form of molecular weight greater than 500,000 on gel filtration and, like the M_r 200,000 form, was only detected after activation with trypsin. Subsequent to this treatment, it too behaved as a M_r 35,000 enzyme. Although a single major enzyme form was present in the cytosol, multiple molecular weight forms could be generated in crude extracts simply by the use of vigorous mechanical homogenization procedures. This suggested that artifactual forms of enzyme may readily be produced, possibly by proteolytic cleavage of the native enzyme.     

Specific modification of fatty acid synthetase from lactating rat mammary gland by chymotrypsin and trypsin.

Fatty acid synthetase from lactating rat mammary gland after limited proteolysis with chymotrypsin or trypsin synthesizes longer chain fatty acids than those produced by the native enzyme. Of the seven partial reactions of the multienzyme complex, only the thioesterase activity was decreased. The results suggest that modification of the fatty acid synthetase product specificity by chymotrypsin and trypsin results from a specific action of these proteases on the thioesterase component. Trypsin, but not chymotrypsin, cleaved a catalytically active thioesterase from the complex; it thus appears that limited trypsinization will be a useful tool for the isolation of the thioesterase component of the multienzyme.     

Comparison of steady-state kinetics of thiophosphorylated versus unphosphorylated smooth muscle myosin

(i) The steady-state kinetic data obtained with purified gizzard and uterus smooth muscle myosins indicated the presence of a plateau region on the substrate-saturation curves. Hill plots of these data provided evidence for mixed positive and negative cooperative interactions. In contrast, when gizzard myosin was prepared according to the method of A. Sobieszek and R.D. Bremel (1975, Eur. J. Biochem.55, 49–60), the saturation curve in the presence of CaATP was hyperbolic and no cooperativity of the binding site(s) was discerned. However, in the presence of MgATP although the curve appeared hyperbolic the Hill plot of the data was biphasic with negative cooperativity at low MgATP concentration, (ii) When thiophosphorylated gizzard myosin was used for kinetic analysis, the plateau region in the presence of MnATP was eliminated from the saturation curve and this curve became hyperbolic. However, in the presence of MgATP, although the plateau was almost eliminated, the saturation curve was still biphasic with either no or greatly reduced negative cooperativity of binding sites at low MgATP concentrations but positive cooperativity of binding at high MgATP concentrations. In addition, the thiophosphorylation of myosin also increased the K_m and V of MgATP and MnATP, thus indicating weaker affinity for these substrates with thiophosphorylated myosin. (iii) Gizzard myosin also hydrolyzed other nucleotides (the order of rates being CTP = ITP > ATP = UTP > GTP), therefore saturation kinetics using different nucleotides as substrates was also carried out. The saturation curves with each nucleotide were different i.e., hyperbolic with CTP, sigmoid with GTP, hyperbolic with biphasic Hill plot with ITP, and possessing plateau with UTP. In addition, it was observed that the kinetic pattern with each nucleotide was very sensitive to temperature and pH.     

Further characterization and thiophosphorylation of smooth muscle myosin

(i) Myosin from chicken gizzards was purified by a modification of an earlier procedure (M. N. Malik, 1978,Biochemistry17, 27–32). When this myosin, as well as that prepared by the method of A. Sobieszek and R. D. Bremel (1975,Eur. J. Biochem.55, 49–60), was analyzed by gradient slab gel using the discontinuous buffer system of Neville (1971,J. Biol. Chem.246, 6328–6334), a closely spaced doublet in the heavy chain and four light chains were observed as opposed to one heavy chain and two light chains with the method of Weber and Osborn (1969, J. Biol. Chem.244, 4406–4412). These findings raise the possibility of the existence of myosin isoenzymes in smooth muscle. (ii) The purified gizzard myosin was found to be free of kinase and phosphatase. Phosphorylation or thiophosphorylation of myosin was observed only by exogenously adding kinase. A maximum of 1.2 mol of ³²P/mol of myosin and 2.3 mol of ³⁵S/mol of myosin were obtained. The actin-activated ATPase activity depended upon the extent of thiophosphorylation of myosin; a four- to fivefold increase in the activity was observed when myosin was fully thiophosphorylated. Thiophosphorylated myosin was found to be more stable than phosphorylated myosin.     

Expression of fetal thymidine kinase in human cobalamin or folate deficient lymphocytes.

In extracts of peripheral blood lymphocytes of cobalamin or folate deficient patients thymidine kinase activity is increased three fold and exhibits properties of the fetal isoenzyme. Appropriate vitamin therapy results in reduction of this activity to normal levels and change from fetal to adult isoenzyme. The occurrence in cobalamin or folate deficiency of fetal thymidine kinase activity in non proliferating human lymphocytes is unique and may reflect events in the deficient marrow lymphoid progenitor cells.     

Allosteric behavior of platelet myosin.

The allosteric properties of platelet actomyosin and myosin have been further studied. At pH 7.2, both exhibit sigmoid kinetics with at least two interacting ATP binding sites. At pH 8.9, the velocity versus substrate curve is shifted to the right and becomes more sigmoidal. In contrast, at pH 5.5, the enzyme appears to follow hyperbolic kinetics and the K_m is reduced. In the presence of 1.4 m urea, the sigmoidicity is lost and the enzyme obeys Michaelis-Menten kinetics. The effect of ADP on the ATPase activity was also investigated. ADP shows characteristics of a competitive inhibitor; it increases K_m (shifts sigmoid curve to the right) without affecting V. When the enzyme is desensitized by low pH (5.5) or urea (1.4 m), the allosteric interaction is abolished without impairing the catalytic activity and ADP is no longer inhibitory. These findings suggest that platelet myosin possesses two interacting sites and that ADP binds to the allosteric site which appears to be different from the catalytic site.     

The kinetics and mechanism of cobalamin-dependent methyl and ethyl transfer to mercuric ion

The chemical transfer of alkyl groups from alkylcobalamins to mercuric ion has been studied in detail by using ultraviolet-visible conventional and stopped-flow kinetics and, in the case of methyl group transfer, by 220 MHz NMR spectroscopy. These experiments show that heterolytic cleavage of the Co–C δ-bond occurs during electrophilic attack by mercuric ion to give alkylmercury and aquocobalamin as the reaction products. Equilibrium and kinetic experiments are consistent with the initial displacement of 5,6-dimethylbenzimidazole by mercuric ion which results in a deactivaion toward dealkylation by a second mercuric ion. Consequently the main dealkylation reaction at pH 5.0 occurs with uncomplexed alkylcobalamin with the overall rate k_obd being controlled by the above equilibrium. Both the displacement of 5,6-dimethylbenzimidazole (“fast reaction”) and dealkylation (“slow reaction”) are first order in the active mercuric species.     

The asymmetric orientation of cytochrome b561 in bovine chromaffin granule membranes

The topological arrangement of cytochrome b561 in the bovine adrenal medullary chromaffin granule membrane was investigated by radiolabeling and immunoprecipitation techniques using antibody raised against the purified cytochrome. The first labeling procedure involved a membrane-permeable amino group labeling reagent, ethyl acetimidate, and two membrane-nonpermeable amino group labeling reagents, isethionyl acetimidate and trinitrobenzenesulfonic acid. The second radiolabeling procedure involved lactoperoxidase-catalyzed iodination of the exposed tyrosines on the membrane-bound proteins. The labeled cytochrome b561 was isolated by immunoprecipitating detergent extracts of treated membranes, followed by electrophoresis of the precipitated cytochrome in polyacrylamide-dodecyl sulfate. From the analysis of both labeling techniques, cytochrome b561 appeared to be a transmembrane protein and a major portion of this protein was cytoplasmically exposed.     

Histidine as an essential residue in the active site of the copper enzyme galactose oxidase.

The pH dependence of the oxidation of β-methyl-d-galactopyranoside by galactose oxidase at 1.33 mm O₂ has been determined. The k_cat exhibits a bell-shaped dependence on the ionization of at least two groups in the enzyme-substrate complex, pK_b' = 6.3 and pK_a' = 7.1, respectively. The pH-independent value for k_cat at 1.33 mm O₂ (nonsaturating) and saturating glycoside is 1435 s^?; the pH optimum is 6.7. Galactose oxidase is inactivated rapidly by iodoacetamide. Although the reaction is much slower, iodoacetate also inactivates the enzyme. The inactivation by iodoacetamide obeys saturation kinetics; at pH 7.0 k₃ = 2.19 min^?1 and K_i = 5.1 mM; k₃ but not K_i exhibits a bell-shaped pH dependence, with pK_a values of 6.3 and 7.6, respectively. Labeling with [¹⁴C]iodoacetamide establishes that one carboxamidomethyl group is incorporated per enzyme molecule. This incorporation parallels the loss of enzymatic activity. Only N-3-carboxymethylhistidine is detected in chromatograms following hydrolysis of the labeled protein. The protein-bound copper is not lost as a consequence of alkylation. Apogalactose oxidase does not react with iodoacetamide. The alkylation is inhibited by the oxidation of an active center tryptophan residue (s) by N-bromosuccinimide. The fraction of residual enzyme activity remaining after tryptophan oxidation corresponds to the extent of labeling by [¹⁴C]iodoacetamide. Although alkylation causes little change in the spin Hamiltonian parameters of the Cu(II) atom, it nearly abolishes both the optical activity and optical absorbance of the metal. The native tryptophan fluorescence of the enzyme, which is a sensitive probe of its active site, is also markedly affected. Since binding of a substrate, β-methyl-d-galactopyranoside, reduces fluorescence as it does in the active enzyme and binding of CN^? at the Cu(II) site as detected by electron spin resonance appears unaffected by the alkylation, the effect of alkylation is on catalysis, per se. Both a catalytic and a subtle conformational role for the active site histidine are inferred from the results.     

Glutamine synthetase from Salmonella typhimurium: manganese(II), substrate, and inhibitor interaction with the unadenylylated enzyme.

Unadenylylated glutamine synthetase (EC 6.3.1.2) was isolated and purified to homogeneity from Salmonella typhimurium. The enzyme molecule is a symmetrical aggregate of 12 subunits arranged in two hexagonal layers, as is evident from electron micrographs. The subunit molecular weight of the enzyme was found to be approximately 50,000 by polyacrylamide gel electrophoresis in sodium dodecyl sulfate when compared to Escherichia coli glutamine synthetase and other protein standards. A long tube of glutamine synthetase was formed as a single-stranded coil resulting from incubation of the enzyme in a low ionic strength buffer. A study of Mn(II) binding to the unadenylylated enzyme at 25 °C was conducted as a function of pH. At pH 7.1 two classes of metal ion sites per subunit were found with K_D values of 3.7 × 10^?6 and 1.7 × 10^?4m, while at pH 6.8 these values were 1.1 × 10^?5 and 1.0 × 10^?4m, respectively. Only one set of binding sites was observed at pH 6.2 with a K_D value of 1.0 × 10^?4m. The metal ion binding sites were further investigated by monitoring proton relaxation rates (prr) and the epr spectrum of enzyme-bound Mn(II). The longitudinal prr of water protons at pH 7.1 indicate that protons interacting with enzyme-Mn(II) at the “tight” site (K_D = 3.7 × 10^?6) are de-enhanced (?_b1 = 0.42) and result from water protons beyond the inner coordination sphere. The second Mn(II) site has a value of ?_b2 = 35 for the binary enhancement, suggesting that this site probably has two to three rapidly exchanging water molecules in its coordination sphere. The epr spectrum of enzyme-bound Mn(II) at the “tight” site is isotropic and is dramatically sharpened by adding the substrate analog methionine sulfoximine. Subsequent addition of ATP or the ATP analog, AMP-PCP (adenylyl methylene diphosphate) produced anisotropic spectra that were similar, suggesting that both ATP and AMP-PCP bind similarly on the enzyme surface. However, a marked change in the Mn(II) environment from anisotropic to near cubic results from the addition of ADP to the quaternary enzyme-Mn(II)-sulfoximine- (AMP-PCP) complex, indicating that ADP displaces AMP-PCP. No change in the anisotropic spectrum due to the enzyme-Mn(II)-sulfoximine-ATP complex is seen by the addition of ADP. This experimental result supports the experimental findings of Ronzio and Meister [Proc. Nat. Acad. Sci. USA59, 164 (1968)], who established that ATP phosphorylates methionine sulfoximine, thereby producing an inactive enzyme. The allosteric effectors, AMP and Trp, have little effect on the epr spectrum of the complex formed from Mn(II), enzyme, sulfoximine, and ADP, suggesting the absence of direct coordination of AMP or Trp to the bound Mn(II). The prr and epr results reported herein with glutamine synthetase from S. typhimurium when compared to those seen for the enzyme from E. coli [Villafranca et al., Biochemistry15, 544 (1976)] demonstrate some similarities but also many substantial differences between the enzymes from these two bacterial sources.