Quantification of protein thiols and dithiols in the picomolar range using sodium borohydride and 4,4'-dithiodipyridine

Experimental determination of the number of thiols in a protein requires methodology that combines high sensitivity and reproducibility with low intrinsic thiol oxidation disposition. In detection of disulfide bonds, it is also necessary to efficiently reduce disulfides and to quantify the liberated thiols. Ellman's reagent (5,5'-dithiobis-[2-nitrobenzoic acid], DTNB) is the most widely used reagent for quantification of protein thiols, whereas dithiothreitol (DTT) is commonly used for disulfide reduction. DTNB suffers from a relatively low sensitivity, whereas DTT reduction is inconvenient because the reagent must be removed before thiol quantification. Furthermore, both reagents require a reaction pH > 7.0 where oxidation by ambient molecular oxygen is significant. Here we describe a quick and highly sensitive assay for protein thiol and dithiol quantification using the reducing agent sodium borohydride and the thiol reagent 4,4'-dithiodipyridine (4-DPS). Because borohydride is efficiently destroyed by the addition of acid, the complete reduction and quantification can be performed conveniently in one tube without desalting steps. Furthermore, the use of reverse-phase high-performance liquid chromatography for the thiol quantification by 4-DPS reduces the detection limit to the picomolar range (equivalent to 1 microg of a 50-kDa protein containing 1 thiol) while at the same time maintaining low pH throughout the procedure.     

Copper(II) (3,5-diisopropylsalicylate)2 oxidizes thiols to symmetrical disulfides and oxidatively converts mixtures of 5-thio-2-nitrobenzoic acid and nonsymmetrical 5-thio-2-nitrobenzoic acid disulfides to symmetrical disulfides

L-cysteine, D-penicillamine, and L-glutathione were oxidized to symmetrical disulfides in the presence of Cu(II)(3,5-DIPS)2 and air-oxygen at physiologic pH, 7.3. Air-oxygen caused the oxidation of thiol reduced copper, Cu(I), to Cu(II), as evidenced by expected spectrophotometric changes in these reaction mixtures. L-cysteine, D-penicillamine, and L-glutathione formed mixed disulfides and TNB with the addition of DTNB to solutions of these thiols. The observed order of reactivity for these thiols with DTNB was: L-cysteine greater than D-penicillamine greater than L-glutathione. Surprisingly, Cu(II)(3,5-DIPS)2 converted these mixed disulfides to their symmetrical disulfides and DTNB, and although the initial conversion rate was rapid, complete conversion required more than two hours. These observations suggest caution with regard to the spectrophotometric determination of thiols immediately after the addition of Ellman's reagent. These results also clarify an earlier report concerning the oxidation of thiols by Cu(II)(o-phenanthroline)2 and offer caution with regard to the determination of thiols using DTNB in the presence of copper complexes. Spectrophotometric data are provided in support of the suggestion that analysis of plasma or cellular samples for thiols be done in the absence of copper(II) complexes to avoid false negative results.     

2,4-Dinitrophenyl [14C]cysteinyl disulfide allows selective radioactive labeling of protein thiols under spectrophotometric control

2,4-Dinitrophenyl [1-14C]cysteinyl disulfide readily introduces by disulfide exchange [14C]cysteine as a label into proteins with exposed thiols. The release of an equivalent amount of colored 2,4-dinitrothiophenolate allows the labeling reaction to be followed spectrophotometrically. In reaction with two cysteine residues of rabbit skeletal muscle actin, the thiol selectivity of the reagent corresponded to that of 5,5'-dithiobis(2-nitrobenzoic acid) (Ellman's reagent) and was superior to that of N-[14C]ethylmaleimide. Labeling of single SH groups of actin and papain proceeded faster than titration with Ellman's reagent under the same conditions. The [14C]cysteine label could be removed under mild conditions, e.g., with dithiothreitol, but proved to be stable during cyanogen bromide degradation of the protein and peptide purification. 2,4-Dinitrophenyl cysteinyl disulfide can be easily prepared within a few hours.     

Two continuous spectrophotometric assays for methionine aminopeptidase

Two spectrophotometric assays have been developed for methionine aminopeptidases (MetAPs). The first method employs a thioester substrate which, upon enzymatic removal of the N-terminal methionine, generates a free thiol group. The released thiol is quantitated using Ellman's reagent. The MetAP reaction is conveniently monitored on a UV-VIS spectrophotometer in a continuous fashion, with the addition of an excess of Ellman's reagent into the assay reaction. Two tripeptide analogues were synthesized and found to be excellent substrates of both Escherichia coli MetAP and human MetAP2 (k(cat)/K(M) = 2.8 x 10(5) M(-1) s(-1) for the most reactive substrate). In the second assay method, the MetAP reaction is coupled to a prolyl aminopeptidase reaction using Met-Pro-p-nitroanilide as substrate. MetAP-catalyzed cleavage of the N-terminal methionine produces prolyl-p-nitroanilide, which is rapidly hydrolyzed by the prolyl aminopeptidase from Bacillus coagulans to release a chromogenic product, p-nitroaniline. This allows the MetAP reaction to be continuously monitored at 405 nm on a UV-VIS spectrophotometer. The assays have been applied to determine the pH optima and kinetic constants for the E. coli and human MetAPs as well as to screen MetAP inhibitors. These results demonstrate that the current assays are convenient, rapid, and sensitive methods for kinetic studies of MetAPs and effective tools for screening MetAP inhibitors.     

Thyroliberin stimulates rapid hydrolysis of phosphatidylinositol 4,5-bisphosphate by a phosphodiesterase in rat mammotropic pituitary cells. Evidence for an early Ca2+-independent action.

Benzyl p-guanidinothiobenzoate hydrochloride was synthesized and demonstrated to be useful for active-site titration of bovine trypsin, bovine thrombin, human lung tryptase, bovine activated protein C, human Factor XIIa fragment and bovine Factor Xa beta. The titration is based on rapid formation of a stable acyl-enzyme with a stoichiometric release of benzyl thiol. Thiol production is measured quantitatively by including 4,4'-dithiodipyridine in the reaction mixture and measuring the increase in absorbance at 324 nm. Ellman's reagent has also been successfully employed, allowing measurement at 410 nm. Unlike p-nitrophenyl p'-guanidinobenzoate, the thioester titrant reacts slowly with chymotrypsin A alpha thus eliminating interference by this enzyme in most titrations. Advantages of this reagent as a titrant include: flexibility in detection of the released thiol, selectivity between trypsin and chymotrypsin-like enzymes, minimal pH-dependence of the epsilon of the absorbing species, relative stability of the reagent under titration conditions, and high epsilon at pH 7.2 with either 4,4'-dithiodipyridine or Ellman's reagent. The reagent should prove useful as an alternative to p-nitrophenyl p'-guanidinobenzoate hydrochloride for the determination of active-site concentrations of the enzymes employed, as well as of other related enzymes.     

Thiols as classical and slow-binding inhibitors of IMP-1 and other binuclear metallo-beta-lactamases

The inhibitory effect of a variety of thiol compounds on the function of binuclear metallo-beta-lactamases, with a particular focus on IMP-1 from Pseudomonas aeruginosa, has been investigated. Thiol inhibitors, depending on their structural features, fall into two categories, one in which inhibition at neutral pH was instantaneous and the other in which inhibition was time-dependent. While mercaptans with anionic substituents in the vicinity of their SH groups exhibited the former type of inhibition, neutral thiols appear to induce a slow, time-dependent isomerization of the initially formed EI complex to a tighter EI complex. Kinetic parameters describing the latter process were obtained by fitting progress curves of substrate hydrolysis using standard and numerical procedures. The failure of charged thiols to exhibit slow binding is suggested to be due to a rapid isomerization of the initial EI complex. Slow binding in the case of neutral thiols was observed only below pH 8. Studies on the pH dependence of catalysis by IMP-1 revealed that (i) enzyme inactivation at low pH is a slow process with presumably two groups with a pK(a) of approximately 5.2 in the protein being responsible for the loss of activity, (ii) inhibition by thiols is independent of pH between pH 5 and 9, and (iii) an apparent enhancement of the catalytic activity of IMP-1 by thiols occurs at pH <5. The last mentioned phenomenon is explained by a model in which mercaptans retard the proton-dependent isomerization of the enzyme. Studies on the thiol-mediated inhibition of the binuclear forms of Bacteroides fragilis (CcrA) and Bacillus cereus (BcII strain 5/B/6) metallo-beta-lactamase have revealed that while CcrA was instantaneously albeit moderately inhibited by mercaptans, BcII mimicked IMP-1 in its interaction with thiols. These differences are proposed to be due partly to the structural divergence of these proteins in the vicinity of Zn2.     

A direct spectrophotometric assay for peptide deformylase.

A direct UV-VIS spectrophotometric assay has been developed for peptide deformylase. This assay employs a novel class of peptide mimetics as deformylase substrates which, upon enzymatic removal of the N-terminal formyl group, rapidly release free thiols. The released thiols are quantitated using Ellman's reagent. A variety of peptide analogues that contain beta-thiaphenylalanine or beta-thiamethionine as the N-terminal residue were synthesized and found to be excellent substrates of the peptide deformylase from Escherichia coli (k(cat)/K(M) = 6.9 x 10(5) M(-1) s(-1) for the most reactive substrate). The deformylase reaction is conveniently monitored on a UV-VIS spectrophotometer in a continuous fashion. The versatility of the assay has been demonstrated by its application to kinetic characterization of the deformylase, pH profile studies, and enzyme inhibition assays. The assay can also be performed in an end-point fashion. The results demonstrate that this assay is a simple, highly sensitive, and rapid method to study kinetic properties of deformylases without the use of any coupling enzymes.     

The Sulfonylurea-Inhibited NADH Oxidase Activity of HeLa Cell Plasma Membranes has Properties of a Protein Disulfide–Thiol Oxidoreductase with Protein Disulfide–Thiol Interchange Activity

Plasma membrane vesicles of HeLa cells are characterized by a drug-responsive oxidation of NADH. The NADH oxidation takes place in an argon or nitrogen atmosphere and in samples purged of oxygen. Direct assay of protein thiols by reaction with 5,5-dithiobis-(2-nitrobenzoic acid) (DTNB; Ellman's reagent), suggests that protein disulfides may be the natural electron acceptors for NADH oxidation by the plasma membrane vesicles. In the presence of NADH, protein disulfides of the membranes were reduced with a concomitant stoichiometric increase in protein thiols. The increase in protein thiols was inhibited in parallel to the inhibition of NADH oxidation by the antitumor sulfonylurea LY181984 with an EC₅₀ of ca. 30 nM. LY181984, with an EC₅₀ of 30 nM, also inhibited a protein disulfide–thiol interchange activity based on the restoration of activity to inactive (scrambled) RNase and thiol oxidation. The findings suggest that thiol oxidation, NADH-dependent disulfide reduction (NADH oxidation), and protein disulfide–thiol interchange in the absence of NADH all may be manifestations of the same sulfonylurea binding protein of the HeLa plasma membrane. A surface location of the thiols involved was demonstrated using detergents and the impermeant thiol reagent p-chloromercuriphenylsulfonic acid (PCMPS). The surface location precludes a physiological role of the protein in NADH oxidation. Rather, it may carry out some other role more closely related to a function in growth, such as protein disulfide–thiol interchange coupled to cell enlargement.     

Reactivity of biologically important thiol compounds with superoxide and hydrogen peroxide.

The reactivities of glutathione, cysteine, cysteamine, penicillamine, N-acetylcysteine, dithiothreitol and captopril with superoxide generated from xanthine oxidase and hypoxanthine, and with reagent hydrogen peroxide, have been investigated. Rates of thiol loss on adding hydrogen peroxide, and superoxide-dependent thiol loss and oxygen uptake were measured. The relative reactivities of the different thiols with both oxidants were inversely related to the pK of the thiol group, such that at pH 7.4, penicillamine was the most reactive. N-acetylcysteine weakly reactive and no reaction was seen with captopril. For hydrogen peroxide, the calculated rate constants for the reaction with the thiolate anion all fell within the range 18-26 M(-1) s(-1). With superoxide, our results are consistent with each thiol reacting via a short chain that consumes oxygen and regenerates superoxide. Only with some of the thiols, was the consumed oxygen recovered as hydrogen peroxide. Reported values for the rate constant for the reaction of thiols with superoxide vary over four orders of magnitude, with the highest being > 10(5) M(-1) s(-1). Due to the complexity of the chain reaction, no study so far has been able to obtain accurate values and we consider the best estimates to be in the 30 to 1000 M(-1) s(-1) range.     

An amperometric inhibitor biosensor for the determination of reduced glutathione (GSH) without any derivatization in some plants

In this study, an amperometric biosensor based on cucumber tissue homogenate was developed for the determination of glutathione. Cucumber (Cucumis sativus L.) tissue homogenate was used as the biological material. The cucumber tissue homogenate was cross-linked with gelatine using glutaraldehyde and fixed on a pretreated teflon membrane. The principle of the measurements was based on the determination of the decrease in the differentiation of oxygen level which had been caused by the inhibition of ascorbate oxidase in the biological material by glutathione. Determinations were carried out by standard curves which were obtained by the measurement of the decrease in the consumed oxygen level related to glutathione concentration. Optimization and characterization studies of the biosensor were carried out and a linearity in the gamma-L-glutamyl-L-cysteinyl-glycine (GSH) concentration range 0.1-2 microM was obtained when 600 microM ascorbic acid was used as a substrate. The repeatability experiments (n = 7) revealed that for 1.5 microM GSH, the average value (x), standard deviation (S.D.) and variation coefficient (C.V.) were 1.517 microM, 4.72 x 10(-5) 3.11%, respectively. The biosensor useful lifetime was at least 2 months. The results of some plant samples analyzed with the presented biosensor agreed well with the spectrophotometric method (Ellman's reagent) used as a reference.     

Fluorescence-based detection of thiols in vitro and in vivo using dithiol probes

Thiols play a central role in maintaining biological homeostasis. Their levels can change dramatically in response to oxidative stress associated with toxic insults, bacterial infection, and disease. Therefore, a reagent that can monitor thiol levels both in vitro and in vivo would be useful for assays and as a biomarker. Such a reagent should (i) be selective for thiols, (ii) be able to penetrate cell walls, and (iii) have a low reduction potential so as not to create oxidative stress in a cell. We have developed such a fluorescent reagent (DSSA) based on a dithiol linker: (i) the use of a dithiol linker makes it selective for thiols; (ii) the use of fluorophores that populate neutral states at physiological pH improves cell wall penetration; and (iii) because of the reagent's low reduction potential (-0.60 V), it will not stress cells oxidatively. For example, 5 microM of reagent is responsive to changes in glutathione levels in the physiologically relevant range of 1 to 10mM, yet this would oxidize less than 1% of cellular glutathione. In Escherichia coli, decreased thiol levels were detected in cells deficient in glutathione synthesis. In zebrafish embryos, the DSSA reagent permitted detection of unusually high thiol levels in the zebrafish chorion.     

Reactivity of the essential thiol of Klebsiella aerogenes urease. Effect of pH and ligands on thiol modification

The kinetics of Klebsiella aerogenes urease inactivation by disulfide and alkylating agents was examined and found to follow pseudo-first-order kinetics. Reactivity of the essential thiol is affected by the presence of substrate and competitive inhibitors, consistent with a cysteine located proximal to the active site. In contrast to the results observed with other reagents, the rate of activity loss in the presence of 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB) saturated at high reagent concentrations, indicating that DTNB must first bind to urease before inactivation can occur. The pH dependence for the rate of urease inactivation by both disulfide and alkylating agents was consistent with an interaction between the thiol and a second ionizing group. The resulting macroscopic pKa values for the 2 residues are less than 5 and 12. Spectrophotometric studies at pH 7.75 demonstrated that 2,2'-dithiodipyridine (DTDP) modified 8.5 +/- 0.2 mol of thiol/mol of enzyme or 4.2 mol of thiol/mol of catalytic unit. With the slow tight binding competitive inhibitor phenyl-phosphorodiamidate (PPD) bound to urease, 1.1 +/- 0.1 mol of thiol/mol of catalytic unit were protected from modification. PPD-bound DTDP-modified urease could be reactivated by dialysis, consistent with the presence of one thiol per active site. Analogous studies at pH 6.1, using the competitive inhibitor phosphate, confirmed the presence of one protected thiol per catalytic unit. Under denaturing conditions, 25.5 +/- 0.3 mol of thiol/mol of enzyme (Mr = 211, 800) were modified by DTDP.     

Characterization of normal, glutathione-deficient and arginase-deficient sheep erythrocytes by 1H-NMR spectroscopy

Normal sheep erythrocytes as well as glutathione- (GSH-) deficient and arginase-deficient sheep erythrocytes have been characterized by 1H nuclear magnetic resonance spectroscopy. The GSH deficiency is a result of defective amino acid transport (lesion 1), diminished gamma-glutamylcysteine synthetase activity (lesion 2), or both (lesions (1 + 2)). 1H-NMR spectra of normal sheep erythrocytes are similar to those for human erythrocytes, and consist of resonances from a number of small intracellular molecules, including GSH. In contrast, the resonances for GSH in the GSH-deficient erythrocytes are much weaker, and strong resonances are observed for lysine, threonine and ornithine or arginine, depending on the arginase activity, in erythrocytes with lesion 1 and lesions (1 + 2). A comparison of the intensity of GSH resonances in spectra for normal and GSH-deficient erythrocytes with GSH levels determined spectrophotometrically following reaction with the nonspecific thiol reagent 5,5'-dithiobis(2-nitrobenzoate) (DTNB) indicates that either not all of the GSH determined with Ellman's reagent is free and observable by 1H-NMR or that not all of the thiol determined by Ellman's reagent is GSH. If the latter is the case, the GSH levels determined with Ellman's reagent for erythrocytes with lesions (1 + 2) are most affected, which might account for their high susceptibility to oxidative stress.     

Protein determination using bicinchoninic acid in the presence of sulfhydryl reagents

The bicinchoninic acid (BCA) copper reagent, developed for quantification of proteins, was found to react with thiol reagents in a linear and reproducible manner. The reactivity with thiols closely matched the extinction coefficient determined for the Cu(I)-BCA complex [6.6 X 10(3) liters (mol Cu.cm)-1], suggesting that the reaction is quantitative. This reaction interferes with the accurate determination of protein concentrations. A method was developed for determining protein concentrations in the presence of thiol reagents using the BCA protein reagent. The procedure involves preincubation of the protein solution with iodoacetamide prior to addition of the BCA protein reagent. Iodoacetamide does not react with the BCA reagent by itself. In the presence of a 10-fold molar excess of iodoacetamide over thiol equivalents, the reaction of the thiol with the BCA reagent is prevented. The method is simple and allows the assay of solutions of proteins which have been stabilized by the addition of thiol reagents.     

Essential and nonessential thiols of yeast hexokinase. Reactions with iodoacetate and iodoacetamide.

The reaction of yeast hexokinase with iodoacetate or iodoacetamide has been investigated in detail, using pure hexodinase B. Of the four thiols in each subunit of the molecule, two (the "apparently essential thiols") are alkylated rapidly at 35 degrees, and the enzymic activity is lost in parallel with their reaction. The other two thiols react subsequently to completion, but at a very much slower rate. In the conditions use, no other uptake of the reagent occurs elsewhere during these thiol alkylations. Electrophoretically homogeneous kialkylated and tetraalkylated protein species are formed, in the two stages of the reaction. The inactivating reaction at 35 degrees with the apparently essential thiols is second order. The rate constant increases with increasing pH, in the range pH 7.0-8.5, in a manner consistent with control of the reaction by a group with pKa of approximately 10. The absolute (pH independent) rate constant is of the same order as that for a normal thiol in model compounds. The availability of the apparently essential thiols appears to be associated with some conformational change in the molecule in the monomer form: it declines at high ionic strengths, is maximal at intermediate values where the dimer first dissociates, but is lowered in the dimer at very low ionic strengths. The reaction also shows a sharp temperature dependence: the dimer at 30 degrees (in constrast to 35 degrees) shows no availability of the apparently essential thiols. A similar transition to a state permitting fast inactivation is found with pH, above pH 8.5. The reaction of the two apparently essential thiols is strongly inhibited by glucose. ATP and ADP, and their Mg complexes, protect significantly, but less effectively than does glucose. The affinities of these substrates at the active site of the enzyme are measured in this protection system. These various reactions appear to be of value for identifying the cysteine-containing regions that are involved in the active center or in its maintenance in the structure.     

Effect of bridging the two essential thiols of myosin on its spectral and actin-binding properties.

The circular dichroic and fluorescent spectral properties of the myosin head (subfragment I (SFI)) modified by covalently bridging the two essential thiol groups have been examined. CD spectra of SFI with the two thiols linked through reaction with a bifunctional reagent, N, N'- p-phenylenedimaleimide, show enhancement of the 282-nm minimum similar to that observed for the long-lived kinetic intermediate (Mg**MgADP-Pi) formed during the ATP cleavage reaction. No significant perturbation of the CD band at 282 nm is seen on blocking both thiol groups with the monofunctional reagent N-ethylmaleimide. The fluorescence emission maximum also shifts to lower wavelengths following covalent bridging (from 343 to 340 nm), but no change in fluorescent intensity has been detected. Formation of the covalent bridge completely inhibits interaction of the modified protein with F-actin. These results suggest that the local conformational state of the polypeptide chain formed on bridging the two thiol groups exhibits certain similarities with the state produced following binding of MgATP to native myosin.     

The reaction of papain with Ellman''s reagent [5,5′-dithiobis-(2-nitrobenzoate) dianion]

1. The report by Robyt et al. (1971) that the 2-nitro-5-mercaptobenzoate dianion (Nbs(2-)) produced by reaction of papain with the 5,5'-dithiobis-(2-nitrobenzoate) dianion (Nbs(2) (2-); Ellman's reagent) cleaves the three disulphide bonds in papain is shown to be incorrect. 2. When partially active papain containing approx. 0.4 mol of thiol/mol of protein is incubated with excess of Nbs(2) (2-) at pH8, Nbs(2) (2-) reacts with the protein in an amount stoicheiometric with the cysteinyl thiol group of papain to produce Nbs(2-) in an amount stoicheiometric with the original papain cysteinyl thiol group, and the catalytically inactive mixed disulphide, papain-Nbs(-). 3. Papain catalyses the hydrolysis of Nbs(2) (2-) at pH10.5 probably by nucleophilic catalysis involving the enzyme's thiol group. 4. These results cast very serious doubts on the claim by Robyt et al. (1971) to have established a new general method for the determination of cystinyl disulphide residues in proteins.     

Aromatic thiol pKa effects on the folding rate of a disulfide containing protein

The production of proteins via recombinant DNA technology often requires the in vitro folding of inclusion bodies, which are protein aggregates. To create a more efficient redox buffer for the in vitro folding of disulfide containing proteins, aromatic thiols were investigated for their ability to increase the folding rate of scrambled RNase A. Scrambled RNase A is fully oxidized RNase A with a relatively random distribution of disulfide bonds. The importance of the thiol pK(a) value was investigated by the analysis of five para-substituted aromatic thiols with pK(a) values ranging from 5.2 to 6.6. Folding was measured at pH 6.0 where the pK(a) value of the thiols would be higher, lower, or equal to the solution pH. Thus, relative concentrations of thiol and thiolate would vary across the series. At pH 6.0, the aromatic thiols increased the folding rate of RNase A by a factor of 10-23 over that observed for glutathione, the standard additive. Under optimal conditions, the apparent rate constant increased as the thiol pK(a) value decreased. Optimal conditions occurred when the concentration of protonated thiol in solution was approximately 2 mM, although the total thiol concentration varied considerably. The importance of the concentration of protonated thiol in solution can be understood based on equilibrium effects. Kinetic studies suggest that the redox buffer participates as the nucleophile and/or the center thiol in the key rate determining thiol disulfide interchange reactions that occur during protein folding. Aromatic thiols proved to be kinetically faster and more versatile than classical aliphatic thiol redox buffers.     

Microanalysis of the metabolic intermediates of lactose synthesis in human milk and plasma using bioluminescent methods

Sensitive bioluminescent methods were developed to measure the metabolites glucose, glucose 6-phosphate (G6P), glucose 1-phosphate (G1P), UDP-glucose, and UDP-galactose in human milk and lactose and galactose in human plasma. The bioluminescent methods measured NADH produced by coupled enzymatic assays derived from equivalent spectrophotometric methods. We found that the long chain fatty acids in human milk (C10-C16) inhibited the bioluminescent reactions. This inhibition was overcome by adding defatted bovine serum albumin to the reaction mixture containing the bioluminescent enzymes. It also was necessary to modify methods of deproteinizing milk and blood plasma to accommodate small sample volumes. In the development of these assays emphasis was given to simplicity of reagent preparation, minimizing cost, and ease of use. The detection limit for the bioluminescent method for NADH was 0.28 nM for a 20-microliters sample. For the assays of the metabolites, recoveries ranged from 91 to 107%. For sample sizes of 2 to 5 microliters of protein free sample, the detection limits for milk were G1P, 0.09 microM; G6P, 0.05 microM; UDPhexose, 0.07 microM; UDP-Glc, 0.03 microM; glucose, 9 microM; and for plasma, lactose, 0.76 microM, galactose, 0.31 microM. The bioluminescent methods gave equivalent results to spectrophotometric methods for the measurement of blood lactose and milk glucose.