Effects of substitution at serine 40 of tyrosine hydroxylase on catecholamine binding.

Phosphorylation of Ser40 in the regulatory domain of tyrosine hydroxylase activates the enzyme by increasing the rate of dissociation of inhibitory catecholamines [Ramsey, A. J., and Fitzpatrick, P. F. (1998) Biochemistry 37, 8980-8986]. To probe the structural basis for this effect and to ascertain the ability of other amino acids to functionally replace serine and serine phosphate, the effects of replacement of Ser40 with other amino acids were determined. Only minor changes in the Vmax value and the Km values for tyrosine and tetrahydropterin were seen upon replacement of Ser40 with alanine, valine, threonine, aspartate, or glutamate, in line with the minor effects of phosphorylation on steady-state kinetic parameters. More significant effects were seen on the binding of dopamine and dihydroxyphenylalanine. The affinity of the S40T enzyme for either catecholamine was very similar to that of the wild-type enzyme, while the S40E enzyme was similar to the phosphorylated enzyme. The S40D enzyme had an affinity for DOPA comparable to the phosphorylated enzyme but a higher affinity for dopamine than the latter. With both catecholamines, the S40V and S40A enzymes showed intermediate levels of activation. The results suggest that the serine hydroxyl contributes to the stabilization of the catecholamine-inhibited enzyme. In addition, the S40E enzyme will be useful in further studies of the effects of multiple phosphorylation on tyrosine hydroxylase, while the alanine enzyme does not provide an accurate mimic of the unphosphorylated enzyme.

Evidence for an essential arginine in the flavoprotein nitroalkane oxidase.

The flavoprotein nitroalkane oxidase from the fungus Fusarium oxysporum catalyzes the oxidative denitrification of primary or secondary nitroalkanes to yield the respective aldehydes or ketones, hydrogen peroxide and nitrite. The enzyme is inactivated in a time-dependent fashion upon treatment with the arginine-directed reagents phenylglyoxal, 2,3-butanedione, and cyclohexanedione. The inactivation shows first order kinetics with all reagents. Valerate, a competitive inhibitor of the enzyme, fully protects the enzyme from inactivation, indicating that modification is active site directed. The most rapid inactivation is seen with phenylglyoxal, with a k(inact) of 14.3 +/- 1.1 M(-1) min(-1) in phosphate buffer at pH 7.3 and 30 degrees C. The lack of increase in the enzymatic activity of the phenylglyoxal-inactivated enzyme after removing the unreacted reagent by gel filtration is consistent with inactivation being due to covalent modification of the enzyme. A possible role for an active site arginine in substrate binding is discussed.

Substrate dehydrogenation by flavoproteins.

Enzymes with tightly bound FMN or FAD as cofactor catalyze the oxidation of a wide range of substrates. The chemical versatility of the isoalloxazine ring provides these enzymes with a range of potential mechanisms. Recent progress in elucidating the mechanisms of oxidation of organic substrates by flavoenzymes is described, focusing on the oxidation of alcohols, amino and hydroxy acids, amines, and nitroalkanes. With each family of enzymes, an attempt is made to integrate mechanistic, structural, and biomimetic data into a common catalytic mechanism.

Probing the relative timing of hydrogen abstraction steps in the flavocytochrome b2 reaction with primary and solvent deuterium isotope effects and mutant enzymes.

Flavocytochrome b(2) catalyzes the oxidation of lactate to pyruvate. Primary deuterium and solvent kinetic isotope effects have been used to determine the relative timing of cleavage of the lactate O-H and C-H bonds by the wild-type enzyme, a mutant protein lacking the heme domain, and the D282N enzyme. The (D)V(max) and (D)(V/K(lactate)) values are both 3.0 with the wild-type enzyme at pH 7.5 and 25 degrees C, increasing to about 3.6 with the flavin domain and increasing further to about 4.5 with the D282N enzyme. Under these conditions, the (D20)V(max) values are 1.38, 1.18, and 0.98 for the wild-type enzyme, the flavin domain, and the D282N enzyme, respectively; the (D20)(V/K(lactate)) values are 0.9, 0.44, and 1.0, respectively. The (D)k(red) value is 5.4 for the wild-type enzyme and 3.5 for the flavin domain, whereas the solvent isotope effect on this kinetic parameter is 1.0 for both enzymes. The V(max) values for the wild-type enzyme and the flavin domain are 32 and 15% limited by viscosity, respectively. In contrast, the V/K(lactate) value for the flavin domain increases about 2-fold at moderate concentrations of glycerol. The data are consistent with a minimal chemical mechanism in which the lactate hydroxyl proton is not in flight in the transition state for C-H bond cleavage and there is an internal equilibrium involving the lactate-bound enzyme prior to C-H bond cleavage which is sensitive to solution conditions. Removal of the hydroxyl proton may occur in this pre-equilibrium.

Use of pH and kinetic isotope effects to dissect the effects of substrate size on binding and catalysis by nitroalkane oxidase.

The flavoprotein nitroalkane oxidase catalyzes the oxidation of a broad range of primary and secondary nitroalkanes to the respective aldehydes or ketones, with production of hydrogen peroxide and nitrite. The V/K values for primary nitroalkanes increase with increasing chain length, reaching a maximum with 1-nitrobutane [Gadda, G., and Fitzpatrick, P. F. (1999) Arch. Biochem. Biophys. 363, 309-313]. In the present report, pH and deuterium kinetic isotope effects with a series of primary nitroalkanes and phenylnitromethane as substrates have been used to dissect the effects of chain length on binding and catalysis. The apparent pKa value for a group that must be unprotonated for catalysis decreases from about 7 to 5.3 with increasing size of the substrate. The D(V/K) values for these substrates decrease from 7.5 with nitroethane to 1 with phenylnitromethane. These results show that increasing the size of the substrate results in an increased partitioning forward to catalysis. The D(V/K) and DVmax values at pH 5.5 have been used to calculate the effect of substrate size on the Kd values for primary nitroalkanes. The Kd values decrease with increasing length of the substrate, with a deltadeltaG(binding) of 1.7 kcal mol(-1) for each additional methylene group. Such a value is less than the value of 2.6 kcal mol(-1) previously determined for the effect of a methylene group on the V/K value [Gadda, G., and Fitzpatrick, P. F. (1999) Arch. Biochem. Biophys. 363, 309-313], suggesting that the total energy available per methylene group is used not only to enhance binding but also to increase the rate of catalysis.

Reversing the substrate specificities of phenylalanine and tyrosine hydroxylase: aspartate 425 of tyrosine hydroxylase is essential for L-DOPA formation.

The catalytic domains of the pterin-dependent enzymes phenylalanine hydroxylase and tyrosine hydroxylase are homologous, yet differ in their substrate specificities. To probe the structural basis for the differences in specificity, seven residues in the active site of phenylalanine hydroxylase whose side chains are dissimilar in the two enzymes were mutated to the corresponding residues in tyrosine hydroxylase. Analysis of the effects of the mutations on the isolated catalytic domain of phenylalanine hydroxylase identified three residues that contribute to the ability to hydroxylate tyrosine, His264, Tyr277, and Val379. These mutations were incorporated into full-length phenylalanine hydroxylase and the complementary mutations into tyrosine hydroxylase. The steady-state kinetic parameters of the mutated enzymes showed that the identity of the residue in tyrosine hydroxylase at the position corresponding to position 379 of phenylalanine hydroxylase is critical for dihydroxyphenylalanine formation. The relative specificity of tyrosine hydroxylase for phenylalanine versus tyrosine, as measured by the (V/K(phe))/(V/K(tyr)) value, increased by 80000-fold in the D425V enzyme. However, mutation of the corresponding valine 379 of phenylalanine hydroxylase to aspartate was not sufficient to allow phenylalanine hydroxylase to form dihydroxyphenylalanine at rates comparable to that of tyrosine hydroxylase. The double mutant V379D/H264Q PheH was the most active at tyrosine hydroxylation, showing a 3000-fold decrease in the (V/K(phe))/(V/K(tyr)) value.

Australian bat lyssavirus infection: a second human case, with a long incubation period.

In December 1998, a 37-year-old Queensland woman died from a rabies-like illness, 27 months after being bitten by a flying fox (fruit bat). Molecular techniques enabled diagnosis of infection with Australian bat lyssavirus (ABL), the second human case to be recognised and the first to be acquired from a flying fox. It must be assumed that any bat in Australia could transmit ABL; anyone bitten or scratched by a bat should immediately wash the wounds thoroughly with soap and water and promptly seek medical advice.

Identification of a cysteine residue in the active site of nitroalkane oxidase by modification with N-ethylmaleimide.

The flavoprotein nitroalkane oxidase catalyzes the oxidative denitrification of primary or secondary nitroalkanes to the corresponding aldehydes or ketones with production of hydrogen peroxide and nitrite. The enzyme is irreversibly inactivated by treatment with N-ethylmaleimide at pH 7. The inactivation is time-dependent and shows first-order kinetics for three half-lives. The second-order rate constant for inactivation is 3.4 +/- 0.06 m(-)(1) min(-)(1). The competitive inhibitor valerate protects the enzyme from inactivation, indicating an active site-directed modification. Comparison of tryptic maps of enzyme treated with N-[ethyl-1-(14)C]maleimide in the absence and presence of valerate shows a single radioactive peptide differentially labeled in the unprotected enzyme. The sequence of this peptide was determined to be LLNEVMCYPLFDGGNIGLR using Edman degradation and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. The cysteine residue was identified as the site of alkylation by ion trap mass spectrometry.

The aromatic amino acid hydroxylases.

The enzymes phenylalanine hydroxylase, tyrosine hydroxylase, and tryptophan hydroxylase constitute the family of pterin-dependent aromatic amino acid hydroxylases. Each enzyme catalyzes the hydroxylation of the aromatic side chain of its respective amino acid substrate using molecular oxygen and a tetrahydropterin as substrates. Recent advances have provided insights into the structures, mechanisms, and regulation of these enzymes. The eukaryotic enzymes are homotetramers comprised of homologous catalytic domains and discrete regulatory domains. The ligands to the active site iron atom as well as residues involved in substrate binding have been identified from a combination of structural studies and site-directed mutagenesis. Mechanistic studies with nonphysiological and isotopically substituted substrates have provided details of the mechanism of hydroxylation. While the complex regulatory properties of phenylalanine and tyrosine hydroxylase are still not fully understood, effects of regulation on key kinetic parameters have been identified. Phenylalanine hydroxylase is regulated by an interaction between phosphorylation and allosteric regulation by substrates. Tyrosine hydroxylase is regulated by phosphorylation and feedback inhibition by catecholamines.

Mutation of serine 395 of tyrosine hydroxylase decouples oxygen-oxygen bond cleavage and tyrosine hydroxylation.

Ser395 and Ser396 in the active site of rat tyrosine hydroxylase are conserved in all three members of the family of pterin-dependent hydroxylases, phenylalanine hydroxylase, tyrosine hydroxylase, and tryptophan hydroxylase. Ser395 is appropriately positioned to form a hydrogen bond to the imidazole nitrogen of His331, an axial ligand to the active site iron, while Ser396 is located on the wall of the active site cleft. Site-directed mutagenesis has been used to analyze the roles of these two residues in catalysis. The specific activities for formation of dihydroxyphenylalanine by the S395A, S395T, and S396A enzymes are 1.3, 26, and 69% of the wild-type values, respectively. Both the S395A and S396A enzymes bind a stoichiometric amount of iron and exhibit wild-type spectra when complexed with dopamine. The K(M) values for tyrosine, 6-methyltetrahydropterin, and tetrahydrobiopterin are unaffected by replacement of either residue with alanine. Although the V(max) value for tyrosine hydroxylation by the S395A enzyme is decreased by 2 orders of magnitude, the V(max) value for tetrahydropterin oxidation by either the S395A or the S396A enzyme is unchanged from the wild-type value. With both mutant enzymes, there is quantitative formation of 4a-hydroxypterin from 6-methyltetrahydropterin. These results establish that Ser395 is required for amino acid hydroxylation but not for cleavage of the oxygen-oxygen bond, while Ser396 is not essential. These results also establish that cleavage of the oxygen-oxygen bond occurs in a separate step from amino acid hydroxylation.

Identification of an essential tyrosine residue in nitroalkane oxidase by modification with tetranitromethane.

The flavoprotein nitroalkane oxidase from Fusarium oxysporum catalyzes the oxidation of nitroalkanes to the respective aldehydes or ketones with production of nitrite and hydrogen peroxide. The enzyme is irreversibly inactivated by incubation with tetranitromethane, a tyrosine-directed reagent, at pH 7.3. The inactivation is time-dependent and shows first-order kinetics for two half-lives of inactivation. Further inactivation can be achieved upon a second addition of tetranitromethane. A saturation kinetic pattern is observed when the rate of inactivation is determined versus the concentration of tetranitromethane, indicating that a reversible enzyme-inhibitor complex is formed before irreversible inactivation occurs. Values of 0.096 +/- 0.013 min(-1) and 12.9 +/- 3.8 mM were determined for the first-order rate constant for inactivation and the dissociation constant for the reversibly formed complex, respectively. The competitive inhibitor valerate protects the enzyme from inactivation by tetranitromethane, suggesting an active-site-directed inactivation. The UV-visible absorbance spectrum of the inactivated enzyme is perturbed with respect to that of the native enzyme, suggesting that treatment with tetranitromethane resulted in nitration of the enzyme. Comparison of tryptic maps of nitroalkane oxidase treated with tetranitromethane in the presence and absence of valerate shows a single peptide differentially labeled in the inactivated enzyme. The spectral properties of the modified peptide are consistent with nitration of a tyrosine residue. The amino acid sequence of the nitrated peptide is L-L-N-E-V-M-C-(NO(2)-Y)-P-L-F-D-G-G-N-I-G-L-R. The possible role of this tyrosine in substrate binding is discussed.

Iso-mechanism of nitroalkane oxidase: 1. Inhibition studies and activation by imidazole.

The flavoprotein nitroalkane oxidase catalyzes the oxidation of primary and secondary nitroalkanes to aldehydes and ketones, respectively, transferring electrons to oxygen to form hydrogen peroxide. The steady-state kinetic mechanism of the active flavin adenine dinucleotide-(FAD-) containing form of the enzyme has been determined with nitroethane at pH 7 to be bi-ter ping-pong, with oxygen reacting with the free reduced enzyme after release of the aldehyde product. The V(max) value is 5.5 +/- 0.3 s(-)(1) and the K(m) values for nitroethane and oxygen are 3.3 +/- 0.6 and 0.023 +/- 0.007 mM, respectively. The free reduced enzyme forms a dead-end complex with nitroethane, with a K(ai) value of 30 +/- 6 mM. Acetaldehyde and butyraldehyde are noncompetitive inhibitors versus nitroethane due to formation of a dead-end complex between the oxidized enzyme and the product. Acetaldehyde is an uncompetitive inhibitor versus oxygen, indicating that an irreversible isomerization of the free reduced enzyme occurs before the reaction with oxygen. Addition of unprotonated imidazole results in a 5-fold increase in the V(max) value, while the V/K values for nitroethane and oxygen are unaffected. A 5-fold increase in the K(ai) value for nitroethane and a 6.5-fold increase in the K(ii) value for butyraldehyde are observed in the presence of imidazole. These results are consistent with the isomerization of the free reduced enzyme being about 80% rate-limiting for catalysis and with a model in which unprotonated imidazole accelerates the rate of isomerization.

Mechanism of nitroalkane oxidase: 2. pH and kinetic isotope effects.

Nitroalkane oxidase catalyzes the oxidation of nitroalkanes to aldehydes or ketones with production of nitrite and hydrogen peroxide. pH and kinetic isotope effects with [1, 1-(2)H(2)]nitroethane have been used to study the mechanism of this enzyme. The V/K(ne) pH profile is bell-shaped. A group with a pK(a) value of about 7 must be unprotonated and one with a pK(a) value of 9.5 must be protonated for catalysis. The lower pK(a) value is seen also in the pK(is) profile for the competitive inhibitor valerate, indicating that nitroethane has no significant external commitments to catalysis. The (D)(V/K)(ne) value is pH-independent with a value of 7.5, whereas the (D)V(max) value increases from 1.4 at pH 8.2 to a limiting value of 7.4 below pH 5. The V(max) pH profile decreases at low and high pH, with pK(a) values of 6.6 and 9.5, respectively. Imidazole, which activates the enzyme, affects the V(max) but not the V/K(ne) pH profile. In the presence of imidazole at pH 7 the (D)V(max) value increases to a value close to the intrinsic value, consistent with cleavage of the carbon-hydrogen bond of the substrate being fully rate-limiting for catalysis in the presence of imidazole.

Effects of phosphorylation on binding of catecholamines to tyrosine hydroxylase: specificity and thermodynamics.

As the catalyst for the rate-limiting step in the biosynthesis of the catecholamine neurotransmitters, the activity of tyrosine hydroxylase is tightly regulated. A principle means of posttranslational regulation is reversible phosphorylation of serine residues in an N-terminal regulatory domain. Phosphorylation of serine 40 has been shown to have a large effect on the rate constant for dissociation of dopamine and a much smaller effect on that for DOPA [Ramsey, A. J., and Fitzpatrick, P. F. (1998) Biochemistry 37, 8980-8986]. To determine the structural basis for the differences in affinity and to further test the validity of the previously proposed model for regulation, the effects of phosphorylation of serine 40 on the affinities for a series of catechols have been determined. The affinities of the unphosphorylated enzyme vary by 3 orders of magnitude due to differences in the rates of dissociation. The highest affinities are found with catecholamines which lack a carboxylate. The affinities of the phosphorylated enzyme show a much smaller range. In the case of binding of dihydroxyphenylalanine, the decrease in affinity upon phosphorylation is due primarily to a decrease in the enthalpy of the interaction. Based upon these results, a structural model for the effect of phosphorylation is proposed.

Influence of steric bulk and electrostatics on the hydroxylation regiospecificity of tryptophan hydroxylase: characterization of methyltryptophans and azatryptophans as substrates.

Tryptophan hydroxylase is a pterin-dependent amino acid hydroxylase that catalyzes the incorporation of one atom of molecular oxygen into tryptophan to form 5-hydroxytryptophan. The substrate specificity and hydroxylation regiospecificity of tryptophan hydroxylase have been investigated using tryptophan analogues that have methyl substituents or nitrogens incorporated into the indole ring. The products of the reactions show that the regiospecificity of tryptophan hydroxylase is stringent. Hydroxylation does not occur at the 4 or 6 carbon in response to changes in substrate topology or atomic charge. 5-Hydroxymethyltryptophan and 5-hydroxy-4-methyltryptophan are the products from 5-methyltryptophan. These products establish that the hydroxylating intermediate is sufficiently potent to hydroxylate benzylic carbons and that the direction of the NIH shift in tryptophan hydroxylase is from carbon 5 to carbon 4. The effects on the V/K values for the amino acids indicate that the enzyme is most sensitive to changes at position 5 of the indole ring. The V(max) values for amino acid hydroxylation differ at most by a factor of 3 from that observed for tryptophan, while the efficiencies of hydroxylation with respect to tetrahydropterin consumption vary 6-fold, consistent with oxygen transfer to the amino acid being partially or fully rate limiting in productive catalysis.

Phenylalanine residues in the active site of tyrosine hydroxylase: mutagenesis of Phe300 and Phe309 to alanine and metal ion-catalyzed hydroxylation of Phe300.

Residues Phe300 and Phe309 of tyrosine hydroxylase are located in the active site in the recently described three-dimensional structure of the enzyme, where they have been proposed to play roles in substrate binding. Also based on the structure, Phe300 has been reported to be hydroxylated due to a naturally occurring posttranslational modification [Goodwill, K. E., Sabatier, C., and Stevens, R. C. (1998) Biochemistry 37, 13437-13445]. Mutants of tyrosine hydroxylase with alanine substituted for Phe300 or Phe309 have now been purified and characterized. The F309A protein possesses 40% less activity than wild-type tyrosine hydroxylase in the production of DOPA, but full activity in the production of dihydropterin. The F300A protein shows a 2.5-fold decrease in activity in the production of both DOPA and dihydropterin. The K(6-MPH4) value for F300A tyrosine hydroxylase is twice the wild-type value. These results are consistent with Phe309 having a role in maintaining the integrity of the active site, while Phe300 contributes less than 1 kcal/mol to binding tetrahydropterin. Characterization of Phe300 by MALDI-TOF mass spectrometry and amino acid sequencing showed that hydroxylation only occurs in the isolated catalytic domain after incubation with a large excess of 7, 8-dihydropterin, DTT, and Fe(2+). The modification is not observed in the untreated catalytic domain or in the full-length protein, even in the presence of excess iron. These results establish that hydroxylation of Phe300 is an artifact of the crystallography conditions and is not relevant to catalysis.

Characterization of 2-oxo-3-pentynoate as an active-site-directed inactivator of flavoprotein oxidases: identification of active-site peptides in tryptophan 2-monooxygenase.

2-oxo-3-pentynoate has been characterized as an active-site-directed inhibitor of selected flavoprotein oxidases. Tryptophan 2-monooxygenase is irreversibly inactivated in an active-site-directed fashion. The addition of FAD affords no protection from inactivation, whereas the competitive inhibitor indole-3-acetamide fully protects the enzyme from inactivation. The inactivation follows first-order kinetics for at least five half-lives. The rate of inactivation shows saturation kinetics, consistent with the formation of a reversible complex between the alkylating agent and the enzyme before inactivation occurs. Values of 0.017 +/- 0.0005 min-1 and 44 +/- 7 microM were determined for the limiting rate of inactivation and the apparent dissociation constant for 2-oxo-3-pentynoate, respectively. Tryptic maps of tryptophan 2-monooxygenase treated with 2-oxo-3-pentynoate show that two peptides are alkylated in the absence of indole-3-acetamide but not in its presence. The two peptides were identified by mass spectrometry as residues 333-349 and 503-536. Based upon sequence analysis, cysteine 511 and either cysteine 339 or histidine 338 are the likely sites of modification. In contrast, incubation of D-amino acid oxidase or nitroalkane oxidase with 2-oxo-3-pentynoate results in a loss of 55% or 100%, respectively, of the initial activity. In neither case does a competitive inhibitor affect the rate of inactivation, suggesting that the effect is not due to modification of active-site residues.

Site-directed mutants of charged residues in the active site of tyrosine hydroxylase.

The active site of tyrosine hydroxylase consists of a hydrophobic cleft with an iron atom near the bottom. Within the cleft are several charged residues which are conserved across the family of pterin-dependent hydroxylases. We have studied four of these residues, glutamates 326 and 332, aspartate 328, and arginine 316 in tyrosine hydroxylase, by site-directed substitution with alternate amino acid residues. Replacement of arginine 316 with lysine results in a protein with a Ktyr value that is at least 400-fold greater and a V/Ktyr value that is 4000-fold lower than those found in the wild-type enzyme; substitution with alanine, serine, or glutamine yields insoluble enzyme. Arginine 316 is therefore critical for the binding of tyrosine. Replacement of glutamate 326 with alanine has no effect on the KM value for tyrosine and results in a 2-fold increase in the KM value for tetrahydropterin. The Vmax for DOPA production is reduced 9-fold, and the Vmax for dihydropterin formation is reduced 4-fold. These data suggest that glutamate 326 is not directly involved in catalysis. Replacement of aspartate 328 with serine results in a 26-fold higher KM value for tyrosine, a 8-fold lower Vmax for dihydropterin formation, and a 13-fold lower Vmax for DOPA formation. These data suggest that aspartate 328 has a role in tyrosine binding. Replacement of glutamate 332 with alanine results in a 10-fold higher KM value for 6-methyltetrahydropterin with no change in the KM value for tyrosine, a 125-fold lower Vmax for DOPA formation, and an only 3.3-fold lower Vmax for tetrahydropterin oxidation. These data suggest that glutamate 332 is required for productive tetrahydropterin binding.

Substrate specificity of a nitroalkane-oxidizing enzyme.

The flavoprotein nitroalkane oxidase from Fusarium oxysporum catalyzes the oxidation of nitroalkanes to aldehydes with production of hydrogen peroxide and nitrite. The substrate specificity of the FAD-containing enzyme has been determined as a probe of the active site structure. Nitroalkane oxidase is active on primary and secondary nitroalkanes, with a marked preference for unbranched primary nitroalkanes. The V/K values for primary nitroalkanes increase with increasing length of the alkyl chain, reaching a maximum with 1-nitrobutane, suggesting a hydrophobic binding site sufficient to accommodate a four carbon chain. Each methylene group of the substrate contributes approximately 2.6 kcal mol-1 in binding energy. The V/K values for substrates containing a hydroxyl group are two orders of magnitude smaller than those of the corresponding nitroalkanes, also consistent with a hydrophobic binding site. 3-Nitro-1-propionate is a competitive inhibitor with a Kis value of 3.1 +/- 0.2 mM.