Solvent Perturbation Studies Research Articles

Kinetic and spectroscopic probes of motions and catalysis in the cytochrome P450 reductase family of enzymes

There is a mounting body of evidence to suggest that enzyme motions are linked to function, although the design of informative experiments aiming to evaluate how this motion facilitates reaction chemistry is challenging. For the family of diflavin reductase enzymes, typified by cytochrome P450 reductase, accumulating evidence suggests that electron transfer is somehow coupled to large-scale conformational change and that protein motions gate the electron transfer chemistry. These ideas have emerged from a variety of experimental approaches, including structural biology methods (i.e. X-ray crystallography, electron paramagnetic/NMR spectroscopies and solution X-ray scattering) and advanced spectroscopic techniques that have employed the use of variable pressure kinetic methodologies, together with solvent perturbation studies (i.e. ionic strength, deuteration and viscosity). Here, we offer a personal perspective on the importance of motions to electron transfer in the cytochrome P450 reductase family of enzymes, drawing on the detailed insight that can be obtained by combining these multiple structural and biophysical approaches.

Nature of the Energy Landscape for Gated Electron Transfer in a Dynamic Redox Protein

Conformational control limits most electron transfer (ET) reactions in biology, but we lack general insight into the extent of conformational space explored, and specifically the properties of the associated energy landscape. Here we unite electron-electron double resonance (ELDOR) studies of the diradical (disemiquinoid) form of human cytochrome P450 reductase (CPR), a nicotinamide adenine phosphate dinucleotide (NADPH)-linked diflavin oxidoreductase required for P450 enzyme reduction, with functional studies of internal ET to gain new insight into the extent and properties of the energy landscape for conformationally controlled ET. We have identified multiple conformations of disemiquinoid CPR, which point to a rugged energy landscape for conformational sampling consistent with functional analysis of ET using high-pressure stopped-flow, solvent, and temperature perturbation studies. Crystal structures of CPR have identified discrete "closed" and "open" states, but we emphasize the importance of a continuum of conformational states across the energy landscape. Within the landscape more closed states that favor internal ET are formed by nucleotide binding. Open states that enable P450 enzymes to gain access to electrons located in the FMN-domain are favored in the absence of bound coenzyme. The extent and nature of energy landscapes are therefore accessible through the integration of ELDOR spectroscopy with functional studies. We suggest this is a general approach that can be used to gain new insight into energy landscapes for biological ET mediated by conformational sampling mechanisms.

Free Energy Contributions to Direct Readout of a DNA Sequence

The energetic contributions of individual DNA-contacting side chains to specific DNA recognition in the human papillomavirus 16 E2C-DNA complex is small (less than 1.0 kcal mol(-1)), independent of the physical and chemical nature of the interaction, and is strictly additive. The sum of the individual contributions differs 1.0 kcal mol(-1) from the binding energy of the wild-type protein. This difference corresponds to the contribution from the deformability of the DNA, known as "indirect readout." Thus, we can dissect the energetic contribution to DNA binding into 90% direct and 10% indirect readout components. The lack of high energy interactions indicates the absence of "hot spots," such as those found in protein-protein interfaces. These results are compatible with a highly dynamic and "wet" protein-DNA interface, yet highly specific and tight, where individual interactions are constantly being formed and broken.

Differential conformational environment of tryptophan in epsilon native prototoxin and active toxin from Clostridium perfringens type D.

The tryptophan content of Clostridium perfringens type D epsilon protoxin and toxin was found to be one residue per molecule of protein. N-bromosuccinimide in the presence of urea cleaves the tryptophan with total loss of lethality in both toxin and prototoxin. Fluorescence spectroscopy, circular dichroism (CD) and 10% ethylene glycol solvent perturbation studies showed that the tryptophan in epsilon toxin and that in prototoxin have different conformational environments. The tryptophan is more on the surface in the prototoxin than in the toxin molecule. NBS causes total loss of lethality of the toxin with its ellipticity coming to almost zero in the near UV region of the CD.

Chemical mechanism of β-xylosidase from Trichoderma reesei QM 9414: pH-dependence of kinetic parameters

The variation of kinetic parameters of β-xylosidase from Trichoderma reesei QM 9414 with pH was used to elucidate the chemical mechanism of the p-nitrophenyl β-D-xylopyranoside hydrolysis. The pH-dependence of V and V/K m showed that a group on the enzyme with a pK value of 3.20 must be unprotonated and a group with a pK value of 5.20 must be protonated for activity and both are involved in catalysis. Solvent-perturbation studies indicated that these groups are neutral acid type. Temperature dependence of kinetic parameters suggested the stickiness of the substrate at lower temperatures than the optimum and the calculated ionization enthalpies pointed to carboxyl groups as responsible for both pKs. Chemical modification with triethyloxonium tetrafluoroborate and protection with the substrate studies demonstrated essential carboxyl groups on the enzyme. Profiles of pK i for D-gluconic acid lactone indicated that a group with a pK value of 3.45 must be protonated for binding and it has been assigned to the carboxyl group of D-gluconic acid formed by lactone ring breakdown in solution.

Electrostatic environment surrounding the activation loop phosphotyrosine in the oncoprotein v-Fps.

Autophosphorylation of Tyr-1073 in the activation loop of the oncoprotein v-Fps enhances the phosphoryl transfer reaction without influencing substrate, ATP, or metal ion binding affinities [Saylor, P., et al. (1998) Biochemistry 37, 17875-17881]. A structural model of v-Fps, generated from the insulin receptor, indicates that pTyr-1073 chelates two arginines. Mutation of these residues to alanine (R1042A and R1066A) results in weakly phosphorylated enzymes, indicating that one electropositive center is insufficient for attaining maximum loop phosphorylation and concomitant high catalytic activity. While the turnover rate for R1066A is similar to that for a mutant lacking a phosphorylatable residue in the activation loop, the rate for R1042A is 50-fold slower. While solvent perturbation studies suggest that the former is due to a slow phosphoryl transfer step, the latter effect results from a slow conformational change in the mutant, potentially linked to motions in the catalytic loop. Binding of a stoichiometric quantity of Mg(2+) is essential for ATP binding and catalysis, while binding of an additional Mg(2+) ion activates further the wild-type enzyme. The affinity of the R1066A enzyme for the second Mg(2+) ion is 23-fold higher than that of the phosphorylated or unphosphorylated form of wild-type v-Fps, with substrate binding unaffected. Conversely, the affinity of R1066A for a substrate mimic lacking a phosphorylation site is 12-fold higher than that for the phosphorylated or unphosphorylated form of wild-type v-Fps, with binding of the second Mg(2+) ion unaffected. A comparison of these enzyme-independent parameters indicates that Arg-1042 and Arg-1066 induce strain in the active site in the repressed form of the enzyme. While this strain is not relieved in the phosphorylated form, the improvements in catalysis in activated v-Fps compensate for reduced metal and substrate binding affinities.

PH studies to elucidate the chemical mechanism of penicillin acylase from Kluyvera citrophila.

The variation with pH of the kinetic parameters of penicillin acylase from Kluyvera citrophila has been used to gain information about the chemical mechanism of the reaction catalysed by the enzyme. The pH-dependence of log (V/Km) for penicillin G showed that a group with a pK value over 4.7 must be deprotonated and that a group with a pK value over 9.7 must be protonated in the free enzyme for activity. The solvent perturbation and temperature studies indicated that these groups are respectively of cationic and neutral acid type with ionization enthalpies of 29.7 and 111 kJ/mol. It was proved that penicillin G sulphoxide is a reversible linear competitive inhibitor with respect to the hydrolysis of penicillin G. The similarity of the pH profile and the magnitude of the pK values derived from the dissociation constant, Ki, suggest that both groups are concerned with the binding of penicillin G and its analogues to the enzyme. It is proposed that binding of substrate involves the formation of hydrogen bonds between the substrate and the essential ionizable groups in the enzyme which lie within the hydrophobic environment of the active site of penicillin acylase. This suggestion is supported by the finding that the profile of V (Vmax.) is similar to the V/Km profile, except that the low and high pK values are respectively shifted downward and upward due to the entry of substrate. Moreover, the bell shape of the V profile indicated that they are also essential in the catalytic steps subsequent to binding.

Chemical mechanism of β-glucosidase from Trichoderma reesei QM 9414. pH-dependence of kinetic parameters

The variation of kinetic parameters of beta-glucosidase from Trichoderma reesei QM 9414 with pH was used to gain information about the chemical mechanism of the reaction catalysed by this enzyme. The pH-dependence of Vmax. and Vmax./Km for p-nitrophenyl beta-D-glucopyranoside showed that a group with a pK value of 4.3 must be unprotonated and a group with a pK value of 5.9 must be protonated for activity. Temperature and solvent-perturbation studies indicated that these groups are a histidine residue and a carboxy group respectively. Profiles of pKi for maltose as competitive inhibitor showed that binding is prevented when a group on the enzyme with a pK value of 4.5 becomes protonated.

Studies on the assembly and stability of the metal-thiolate clusters of metallothionein in dimethyl sulfoxide.

Solvent perturbation studies in 100% dimethyl sulfoxide (Me2SO) have been carried out on rabbit liver metallothionein (MT) in an effort to learn more about the factors stabilizing the three-dimensional structure and the mechanism of cluster formation. As indicated by the electronic absorption spectra of Co7-metallothionein, the reconstituted protein preserves its structural integrity in this solvent. Minor spectral differences between water and Me2SO were fully reversible. The titration of apoMT with cobalt(II) in Me2SO, followed by UV-visible-near-infrared electronic absorption, circular dichroism, magnetic circular dichroism, and EPR spectroscopy, indicate that the protein can refold in this solvent. A comparison with the previous titration data in water reveals that the first four titration steps in both solvents are identical, indicating a thermodynamically controlled folding process. However, the reversed order of the cluster completion between Me2SO and water may suggest the involvement of a kinetically controlled folding process in the last three titration steps. A new cluster form developed with approximately nine Co(II) equivalents.

Plastocyanin cytochrome f interaction.

Spinach plastocyanin and turnip cytochrome f have been covalently linked by using a water-soluble carbodiimide to yield an adduct of the two proteins. The redox potential of cytochrome f in the adduct was shifted by -20 mV relative to that of free cytochrome f, while the redox potential of plastocyanin in the adduct was the same as that of free plastocyanin. Solvent perturbation studies showed the degree of heme exposure in the adduct to be less than in free cytochrome f, indicating that plastocyanin was linked in such a way as to bury the exposed heme edge. Small changes were also observed when the resonance Raman spectrum of the adduct was compared to that of free cytochrome f. The adduct was incapable of interacting with or donating electrons to photosystem I. Peptide mapping and sequencing studies revealed two sites of linkage between the two proteins. In one site of linkage, Asp-44 of plastocyanin is covalently linked to Lys-187 of cytochrome f. This represents the first identification of a group on cytochrome f that is involved in the interaction with plastocyanin. The other site of linkage involves Glu-59 and/or Glu-60 of plastocyanin to as yet unidentified amino groups on cytochrome f. Euglena cytochrome c-552 could also be covalently linked to turnip cytochrome f, although with a lower efficiency than spinach plastocyanin. In contrast, a variety of cyanobacterial cytochrome c-553's and a cyanobacterial plastocyanin could not be covalently linked to turnip cytochrome f.

States of tryptophan residues in castor bean hemagglutinin: the perturbing effects of solvents on tryptophans in hemagglutinin and its complexes as analyzed by UV-difference spectroscopy.

The states of tryptophan residues in castor bean hemagglutinin (CBH) were analyzed by solvent perturbation studies employing ultraviolet difference spectroscopy. Eight out of 22 tryptophan residues in CBH were exposed to ethylene glycol and glycerol, suggesting that the remaining 14 tryptophan residues are buried in the interior of the CBH molecule. The fraction of tryptophan residues accessible to the perturbant decreased with increase in the molecular size of the perturbant, and only 2 tryptophan residues were exposed to polyethylene glycol 600. Upon binding with raffinose, 2 tryptophan residues were shielded from the perturbing effect of the solvent, and binding of lactose reduced the number of tryptophan residues accessible to the perturbant by 1 mol per mol of protein. Binding of galactose, however, did not change the accessibility of tryptophan to the perturbant. On the other hand, the accessibility of tyrosine to the perturbant remained unchanged after binding with raffinose and lactose, suggesting that tyrosine is not directly involved in the saccharide binding of CBH. Based on these results, it is proposed that one tryptophan residue at the saccharide-binding site on each B-chain of CBH lies on the surface of the protein molecule and is located at a subsite which is accessible to a glucopyranoside moiety in the lactose molecule or a glycopyranosyl-fructofuranosyl moiety in the raffinose molecule, whereas such a residue is not present at the galactopyranoside-recognition site.

Intramolecular interactions, mesomerism and dynamics in actinomycin D studied by 15N NMR spectroscopy.

We present a detailed conformational study of 15N-labelled actinomycin D in different organic solvents using 1H, 15N and two-dimensional (2D) NMR techniques at 30.4 MHz and 50.6 MHz. The assignment of the threonine and valine 15N resonances to the individual residues on the alpha- or beta-lactone rings was achieved via heteronuclear shift-correlated 2D NMR experiments. The solvent perturbation studies allow an estimation of the solvent accessibility of the nitrogens and carbonyl groups. Evidence is presented that the pentapeptide rings of actinomycin D have different conformations in polar and in apolar solvents. The chromophoric N10 is efficiently solvent-protected, the solvent-dependence of its 15N resonance resulting from solvent interactions at other positions of the molecule and from solvent-dependent changes in the twisting of the chromophoric system. The chromophoric 2-amino nitrogen is shown to exhibit a strong sp2 character due to the formation of a conjugated system with the carbonyl group at C1. Such a conjugation requires a non-planar chromophoric ring system. Additionally, a hydrogen bond connecting the 2-amino and the 1-carbonyl group was detected. In some solvents, two resonances appear for the 2-amino nitrogen implying the presence of the 2-amino group in two different conformations. The possible implications of the non-planarity of the chromophore for the intercalation process and for the biological activity of the drug are discussed.

Chemical mechanism of saccharopine dehydrogenase (NAD +, l-lysine-forming) as deduced from initial rate pH studies

The variation of kinetic parameters with pH has been determined so as to gain insight into the chemical mechanism of the saccharopine dehydrogenase (NAD +, l-lysine-forming)-catalyzed reaction. In the direction of reductive condensation of lysine and α-ketoglutarate (reverse reaction), the V K profile for lysine shows a group with a p K of 6.3 must be unprotonated and a group with a p K of 8.0 must be protonated for activity. Similar p K's are obtained in the p K i profile for ornithine, which acts as a linear competitive inhibitor with respect to lysine. Temperature and solvent perturbation studies show that these groups are probably histidines. The V K profile for α-ketoglutarate reveals a single group with p K = 8.4 (probably lysine) that must be protonated. It is proposed that one of the histidines is involved in the binding of the ϵ-amino group of the substrate lysine and the positively charged lysine residue hydrogen bonds to the carbonyl oxygen of α-ketoglutarate. In the direction of saccharopine cleavage, the V K profile for saccharopine shows that two groups with p K values of 6.0 and 7.1, possibly a histidine and lysine, must be unprotonated for its reaction with the enzyme · NAD + complex. The log V-pH plots for the forward and reverse reactions both show sigmoidal curves. At low pH, the activity is lower for the forward reaction, and is higher for the reverse reaction. The ionization of a single group appears to be responsible for the change in activity. A tentative scheme for the chemical reaction is presented.

Examination of the solvent perturbation technique as a method to identify enzyme catalytic groups.

The present study was undertaken for the purpose of evaluating the solvent perturbation technique as a method to identify enzyme catalytic residues. For establishment of expected directions and sizes of pKa perturbations for different types of acids in different classes of solvents, a study of the pKa of a series of acids in mixed solvent systems was carried out. Consistent with previous findings, the presence of organic solvents (25% v/v) increased the pKa values of neutral acids while it decreased or did not change the pKa values of cationic acids. The size of the perturbation observed was dependent on the nature of the organic solvent and on the polarity of the neutral form of the acid. The solvent perturbation studies were then extended to the catalytic aspartate residue of yeast hexokinase. The pKa of this residue was determined from the MgATP V/K profile measured in the presence and absence of organic solvents (25% v/v). While dimethylformamide and methanol induced small but perhaps significant increases in the observed pKa, dimethyl sulfoxide and propylene glycol did not. The pKa values, from the MgATP V/K profiles measured in the presence of fully saturating glucose, were not significantly increased by the organic solvents. The pKi vs. pH profile for the competitive inhibitor lyxose was also measured in the presence and absence of organic solvents. While methanol (25% v/v), dimethylformamide (25% v/v), and dioxane (17.5% v/v) induced a large increase in the pKa, propylene glycol and dimethyl sulfoxide (25% v/v) did not. The results from this investigation indicate that the solvent perturbation technique should not be relied upon indiscriminately.

Fluorescence spectroscopic studies of pyrene-actin adducts.

Reaction kinetic studies of the sulfhydryl-directed fluorescent probes N-(1-pyrene)maleimide (PM) and N-(1-pyrenyl)iodoacetamide with actin from rabbit skeletal muscle showed that there were three accessible sulfhydryl groups in actin. Fluorescence spectral studies showed energy transfer from aromatic amino acid residues to fluorophore reacted at Cys-373, as well as weak excimer fluorescence probably due to doubly labeled molecules at Cys-10 and Cys-373. These results provide further evidence that trytophan and tyrosine residues are located near the probe attached to Cys-373 or Cys-10 and the latter two thiols are in close proximity. In age PM-labeled F-actin, the succinimido ring of PM underwent intramolecular aminolysis, resulting in large emission spectral changes and increased excimer fluorescence. Solvent perturbation studies indicate that the probes were located in a hydrophobic environment; their quantum yield and spectrum properties were very sensitive to changes in the microenvironment. Nanosecond-pulse fluorimetry studies revealed complex fluorescence emission decays with three intrinsic lifetimes in adducts with low molecular weight thiols as well as in labeled proteins. Fluorescence lifetimes were 17, 48 and 111 ns for the pyrenemaleimide adduct of actin, and 3, 14 and 60 ns for the pyrenyliodoacetamide adduct. Supporting evidence is given for the argument that multiple fluorescence lifetimes are an intrinsic property of the pyrene derivatives and are not due to the presence of impurity or heterogeneity in the protein reaction sites. Because of their high sensitivity and long lifetimes, pyrene derivatives are extremely useful.

Studies on carbohydrate binding to a lectin purified from Streptomyces sp.

The anti-B specific lectin produced by Streptomyces sp. was shown to have two carbohydrate-binding sites with binding constants of 8.3.10 3 M- 1 (15°C) and 2.2· 3 M- 1 (4°C) for l-rhamnose and d-galactose, respectively, calculated according to Scatchard plots. The binding of specific sugars to the lectin not only induced a peculiar ultraviolet difference spectrum showing a blue shift of tryptophan absorption, but also canued crystallization of the lectin at a concentration of 1 mg per ml or more. The solvent-perturbation studies on the lectin showed that the number of solvent-exposed tryptophan (or average extent of exposure) was two in the absence of l-rhamnose, and three in the presence of the sugar. This suggests that one tryptophan residue appears outside as the result of sugar-binding to the lectin, which is reflected by the difference spectra. Oxidation of two tryptophan residues with N-bromosuccinimide led to complete loss of carbohydrate-binding activity of the lectin, indicating that these residues are important for retaining the activity.

22] The use of pH studies to determine chemical mechanisms of enzyme-catalyzed reactions

This chapter discusses the use of pH studies to determine chemical mechanisms of enzyme-catalyzed reactions. The pH profiles that are of the most value are of pKi for competitive inhibitors or for metal-ion activators, log (V/K) for one or slower, nonsticky substrates, and isotope effects on V/K for these same substrates. Profiles of pKi for competitive inhibitors will show the required state of protonation of the inhibitor or of groups on the enzyme for binding. The log (V/K) versus pH profile for a nonsticky substrate shows the correct pK values of groups necessary for both binding and catalysis. Those pK values not present in the pKi profiles are groups that act as acid-base catalysts during the reaction, or whose protonation state is important for the chemical reaction, but not for binding. Profiles of V/K for nonsticky substrates can also be used in temperature-variation and solvent-perturbation studies to determine the nature of the groups involved.

Conformational differences between high clotting human alpha-thrombin and nonclotting gamma-thrombin.

The conformations of human alpha-thrombin and gamma-thrombin have been compared by circular dichroism, solvent perturbation different spectroscopy, and chemical modification. Circular dichroism studies indicate that proteolytic conversion of alpha-thrombin to gamma-thrombin is accompanied by considerable conformational changes which include a decrease in alpha-helical content from 5-7% to 0-1%. Solvent perturbation at pH 6.0 obtained with 20% ethylene glycol, 20% glycerol, and 20% dimethyl sulfoxide indicates an apparent exposure of 3.5 +2- 0.2 tryptophan and 7.8 +/- 0.1 tyrosine residues in alpha-thrombin and 4.6 +/- 0.2 tryptophan and 9.2 +/0 0.3 tyrosine residues in gamma-thrombin. This increased exposure is substantiated by the greater reactivity of tryptophan residues in gamma-thrombin toward dimethyl (2-hydroxy-5-nitrobenzyl) sulfonium bromide. It suggests that gamma-thrombin is a less compact molecule than the parent alpha-thrombin. Solvent perturbation studies of alpha-thrombin and gamma=thrombin inhibited by phenyl-methanesulfonyl fluoride showed that 0.3 +/- tryptophan and 0.9 +/- 0.3 tyrosine residues in alpha-thrombin and 0.6 +/- 0.3 tryptophan and 1.3 +/- 0.4 tyrosine residues in gamma-thrombin were blocked by the inhibitor. These subtle differences in the extent of blocking of tyrosine and tryptophan suggest a tighter conformation in the catalytic site of gamma-thrombin compared to that of alpha-thrombin.

Chemical mechanism of the reaction catalyzed by dihydrofolate reductase from Streptococcus faecium: pH studies and chemical modification.

The variation with pH of the kinetic parameters associated with dihydrofolate reductase from Streptococcus faecium has been used to gain information about the chemical mechanism of the reaction catalyzed by the enzyme. The pH dependence of log V/K for dihydrofolate showed that a group with a pK value of 4.7 must be ionized and that a group with a pK value of 6.6 must be protonated for activity. Temperature and solvent perturbation studies indicate that these groups are probably the carboxyls of the glutamate moiety of dihydrofolate and of an aspartate residue on the enzyme, respectively. The similarity of the pH profile and the magnitude of the pK value for the linear competitive inhibitor 2,4-diaminopteridine suggest that the carboxyl group is concerned with the binding of dihydrofolate and its analogues to the enzyme. This conclusion is confirmed by the result that a group with a pK value of 6.7 must be protonated for the binding of methotrexate. It is proposed that the binding involves the formation with N-5 of dihydrofolate or N-1 of methotrexate of a hydrogen bond which has considerable ionic character and which lies within a hydrophobic environment. Further, it is suggested that the same hydrogen acts as an auxiliary catalyst which facilitates hydride transfer from NADPH to dihydrofolate for its conversion to tetrahydrofolate. Evidence to support this suggestion comes from the finding that the V profile is similar to the V/K profile except that the pK of the group which must be protonated for maximum enzyme activity is shifted upward by about 2 pH units. Such an increase in a pK value is consistent with the formation of a hydrogen ionic bond in the ternary enzyme-NADPH-dihydrofolate complex. The results of inactivation experiments with trinitrobenzenesulfonate appear to indicate that a lysine residue is necessary to maintain the enzyme in its active conformation.

Investigation of the catalytic mechanism of yeast inorganic pyrophosphatase.

P1,P2-Bidentate Co(NH3)4PP was found to be a competitive inhibitor of pyrophosphatase vs. MgPP (Kis = 8.7 mM, pH 7) and, in the presence of Mg2+, an active substrate as well. P1,P2-Bidentate Cr(III) complexes of pyrophosphate, imidodiphosphate, and methylenediphosphonate were also competitive inhibitors vs. MgPP (pH 5.9; Kis = 0.2, 0.2, and 0.4 mM, respectively). In the presence of Mg2+, P1,P2-bidentate Cr(H2O)4PP was found to have a Km 10-fold greater and a turnover number 36-fold smaller than MgPP at pH 5.9. Mg2+, Mn2+, Co2+, Zn2+, Cd2+, Ni2+, and Fe2+ activate the CrPP--pyrophosphatase reaction, while Ca2+ and Ba2+ are not activators but serve as competitive inhibitors vs. Mg2+ (Kis = 0.35 and 2.3 mM). At levels above 0.1 mM, Mn2+, Co2+, and Zn2+ show activator inhibition. Kinetic studies with CrPP and Mg2+ suggest that the kinetic mechanism is rapid equilibrium ordered, with CrPP adding before Mg2+. pH studies of the MgPP/Mg2+ reaction and the CrPP/Mg2+ reaction suggest that the active form of the substrate is (MgPP)2- and that the uncomplexed metal ion cofactor interacts with at least two active-site residues, one possibly via H bonding and the other by direct coordination. The former group (pKa = 5.6) appears on the basis of temperature and solvent perturbation studies to be a carboxylic acid. The MgPP reaction also requires that an active-site residue (pKa = 7.5) be protonated. Temperature and solvent perturbation studies suggest that this residue is an amine. A mechanism accounting for these observations is presented.