Perchloric acid enhances sensitivity and reproducibility of the ferric-xylenol orange peroxide assay

The ferrous oxidation in xylenol orange (FOX) assay for hydroperoxides suffers from very narrow pH optimum in the range 1.7-1.8. Most published protocols recommend 25 mM sulfuric acid as the solvent, but this in practice does not ensure the maintenance of correct pH in the presence of materials such as samples of biological origin. Substitution of perchloric for the sulfuric acid resulted in a lowering of the optimum pH of the assay to 1.1, a decreased dependence of the absorbance of the ferric-xylenol orange complex on acid concentration and decreased sensitivity to added compounds. Molar absorption coefficients of hydrogen peroxide, cumene, and butyl hydroperoxides and of hydroperoxide groups generated in oxidized protein and lipids were determined and found to be higher than in sulfuric acid. The optimum concentration of perchloric acid proved to be 110 mM. The new assay was designated as PCA-FOX, to distinguish it from the FOX methods based on sulfuric acid.     

Determination of peroxides and hydroperoxides in irradiated solutions of nucleic acid constituents and DNA

Two methods are described for determining micromolar concentrations of H₂O₂ and organic hydroperoxides produced in irradiated solutions of nucleic acid derivatives. In the first technique, ferrous ions are oxidized and the ferric ions so formed produce a colored complex with the indicator xylenol orange; the yield of peroxide is measured spectrophotometrically. The second procedure involves modifications of the standard iodine technique.     

Ferrous ion oxidation in the presence of xylenol orange for detection of lipid hydroperoxide in low density lipoprotein.

A simple and sensitive method for the direct measurement of lipid peroxides in lipoprotein and liposomes is described. The method is based on the principle of the rapid peroxide-mediated oxidation of Fe2+ to Fe3+ under acidic conditions. The latter, in the presence of xylenol orange, forms a Fe(3+)-xylenol orange complex which can be measured spectrophotometrically at 560 nm. Calibration with standard peroxides, such as hydrogen peroxide, linoleic hydroperoxide, t-butyl hydroperoxide, and cumene hydroperoxide gives a mean apparent extinction coefficient of 4.52 x 10(4) M-1 cm-1 consistent with a chain length of approximately 3 for ferrous ion oxidation by hydroperoxides. Endoperoxides are less reactive or unreactive in the assay. The assay has been validated in the study of lipid peroxidation of low density lipoprotein and phosphatidyl choline liposomes. By pretreatment with enzymes known to metabolize peroxides, we have shown that the assay measures lipid hydroperoxides specifically. Other methods for measuring peroxidation, such as the assessment of conjugated diene, thiobarbituric acid reactive substances and an iodometric assay have been compared with the ferrous oxidation-xylenol orange assay.     

Determination of hydroperoxides by the ferric-xylenol orange method

Abstract

Organic hydroperoxides are some of the first semi-stable products of the interaction between free radicals (and other reactive oxygen species) with biological systems, so that they are potential indicators of the formation and effects of these reactive molecules. Many assays have been developed for the detection and measurement of hydroperoxides, but all have significant drawbacks. For example, while the acid iodometric technique is the only one giving indisputably quantitative results,¹ it requires anaerobic conditions, making it difficult to use in routine assays. A promising alternative method, with similar sensitivity but unaffected by oxygen, was developed for a wide range of hydroperoxides. It is based on the reaction between the hydroperoxide and ferrous iron in acid medium and the measurement of the ferric iron produced by formation of a complex with xylenol orange. The assay was initially used to measure H₂O₂, but later also in the measurement of a wide range of organic hydroperoxides, including some present in biological fluids.² In attempting to apply the assay to solutions of oxidized proteins, we found that accurate hydroperoxide concentrations could not be obtained by the recommended protocols. Therefore, we examined the variables affecting the assay and validated a new protocol able to provide more reliable results.     

Measurement of Plasma Hydroperoxide Concentrations by the Ferrous Oxidation-Xylenol Orange Assay in Conjunction with Triphenylphosphine

We describe the application of the FOX2 (ferrous oxidation in xylenol orange, version 2) method to the measurement of hydroperoxides in plasma. Authentic plasma hydroperoxides can be determined by a strategy in which the hydroperoxide reductant, triphenylphosphine, is used to discriminate between the background signal generated by ferric ions present in plasma and that generated by hydroperoxide in plasma. The approach was validated by extraction of total lipids from plasma using ethyl acetate prior to assay with the FOX2 reagent. Plasma from 23 normal individuals contained hydroperoxide in the range of 0.22 to 7.8 μM with a mean of 3.02 μM and a population standard deviation of 1.85 μM. After partitioning with ethyl acetate, plasma hydroperoxide levels ranged from 0.22 to 6.22 μM, with a mean value of 2.52 μM and a population standard deviation of 1.65 μM.     

Separation, detection, and quantification of hydroperoxides formed at side-chain and backbone sites on amino acids, peptides, and proteins

Hydroperoxides are major reaction products of radicals and singlet oxygen with amino acids, peptides, and proteins. However, there are few data on the distribution of hydroperoxides in biological samples and their sites of formation on peptides and proteins. In this study we show that normal-or reversed-phase gradient HPLC can be employed to separate hydroperoxides present in complex systems, with detection by postcolumn oxidation of ferrous xylenol orange to the ferric species and optical detection at 560 nm. The limit of detection (10-25 pmol) is comparable to chemiluminescence detection. This method has been used to separate and detect hydroperoxides, generated by hydroxyl radicals and singlet oxygen, on amino acids, peptides, proteins, plasma, and intact and lysed cells. In conjunction with EPR spin trapping and LC/MS/MS, we have obtained data on the sites of hydroperoxide formation. A unique fingerprint of hydroperoxides formed at alpha-carbon (backbone) positions has been identified; such backbone hydroperoxides are formed in significant yields only when the amino acid is part of a peptide or protein. Only side-chain hydroperoxides are detected with free amino acids. These data indicate that free amino acids are poor models of protein damage induced by radicals or other oxidants.     

Specificity of the ferrous oxidation of xylenol orange assay: analysis of autoxidation products of cholesteryl arachidonate

Autoxidation of polyunsaturated fatty acids and esters leads to a complex mixture containing hydroperoxides and cyclic peroxides. The oxidation mixture of cholesteryl arachidonate, which has been characterized by a variety of mass spectrometry techniques, was subject to analysis by conventional thiobarbituric acid-reactive substance (TBARS) and ferrous oxidation in xylenol orange (FOX) assays. Our results indicate that the FOX assay is not specific for hydroperoxides. Cyclic peroxides, such as monocyclic peroxides and serial-cyclic peroxides, give a positive FOX response even after triphenylphosphine reduction. We suggest that bicyclic endoperoxides are the major TBARS active compounds present in cholesteryl arachidonate oxidation mixtures. These compounds give a positive FOX assay before reaction with triphenylphosphine but negative TBARS and FOX assays after this reaction. Caution should be exercised when the FOX assay is used to analyze highly oxidized lipids, especially arachidonyl-containing lipids.     

Influence of ionizing radiation, application of iron ions and their chelate complexes on the oxidative status of blood serum of rats

Influence of ionizing radiation, ions of iron and their chelate complexes on the oxidative status of blood serum of rats has been investigated. Animals were irradiated by gamma-rays 60Co at a dose of 4 Gy. Ions of iron and iron chelates with nitrilotriacetic acid and citric acid were introduced into animals intra-abdominally at a doze of 10 mg of iron on 1 kg of body weight. The oxidative status of blood serum was determined according to the estimated content of oxidizing peroxide equivalents which oxidize ferrous iron in ferric iron with the subsequent estimation of ferric iron by means of xylenol orange. We also estimated the total content of iron in blood serum using ferrozine as an indicator. The oxidative status was defined 24 and 96 hours after irradiation and 2 hours after introduction of iron ions and their chelates. The research conducted has shown that the concentration of oxidizing peroxide equivalents in serum and the total iron concentration increase 1.47 times and 1.63 times correspondingly 24 hours after irradiation. The increase in the content of oxidizing peroxide equivalents and iron owing to Fenton's reaction can lead to the appearance of OH* radical and raise the level of damage of nuclear and membrane structures in irradiated cells. 2 hours after introduction of iron ions and their chelates, the content of oxidizing peroxide equivalents increased in the blood serum of irradiated and non-irradiated rats, and the maximum effect was observed when introducing ferrous iron and its chelate with citric acid.     

Promotion of lipid oxidation by selenate and selenite and indicators of lipid peroxidation in the rat

The AIN-93 reformulation of the AIN-76A rodent diet includes a change in selenium supplement from sodium selenite to sodium selenate to reduce dietary lipid peroxidation. A change to selenate as the standard form of Se in rat diets would render results from previous work using selenite less relevant for comparison with studies using the AIN-93 formulation. To critically examine the rationale for the AIN-93 recommendation, we prepared Torula yeast basal diets patterned as closely as possible after the AIN-93 formulation and supplemented with 0, 0.15 (adequate), or 2.0 (high) mg selenium/kg diet as sodium selenite or sodium selenate. Livers isolated from male Sprague-Dawley rats fed these diets for 15 wk showed no differences in thiobarbituric acid-reactive substances or lipid hydroperoxides measured with the ferrous oxidation in xylenol orange method. Lipids isolated from samples of high-selenate and high-selenite diets showed no differences in conjugated dienes. The addition of selenate or selenite to soybean oil did not result in an altered Oil Stability Index. These results demonstrate that selenate is not less likely than selenite to cause oxidation of other dietary components. Benefits of selenate over selenite in the diets of rodents remain to be demonstrated. Results included in this paper were presented at the meeting of Experimental Biology 98, San Francisco, CA, April 18–22, 1998, and published in abstract form (Moak, M. A., Johnson, B. L., & Christensen, M. J. [1998] On the AIN-93G recommendation for selenium. FASEB J. 12, A824).     

Interactions of Nitric Oxide with Lipid Peroxidation Products under Aerobic Conditions: Inhibitory Effects on the Formation of Malondialdehyde and Related Thiobarbituric Acid-Reactive Substances

Under aerobic conditions, exposure of peroxidized lipids to nitric oxide (NO) was found to result in a rapid decrease in the levels of thiobarbituric acid-reactive substances (TBARS). Addition of 10–100 μM NO to rat brain homogenates preincubated for 2 h at 37°C caused up to a 20% decrease in the levels of TBARS compared to controls. A similar inhibitory effect was observed on TBARS produced by Fe²⁺-induced decomposition of 15-hydroperoxyeicosatetraenoic acid (15-HPETE), due apparently to NO-induced decomposition of the hydroperoxide (ferrous oxidation/xylenol orange assay). Prostaglandin G₂ (PGG₂, 35 μM), as a model bicyclic endoperoxide, and malondialdehyde (MDA, 20 μM), the main component of TBARS, proved also susceptible to degradation by NO or NO donors (diethylamine NONOate, DEA/NO) at concentrations of 100 μM or higher in 0.05 M phosphate buffer, pH 7.4, and at 37°C, as indicated by the reduced response to the TBA assay. No significant effect on TBARS determination was caused by nitrite ions. These and other data indicate that NO can inhibit TBARS formation by decomposing primary lipid peroxidation products, chiefly 15-HPETE and related hydroperoxides, and, to a lesser extent, later stage TBARS precursors, including bicyclic endoperoxides and MDA, via nitrosation and other oxidative routes, without however affecting chromogenic reactions during the assay.     

Indices of oxidative stress. 2. Lipid peroxides

Two methods of the determination of lipid peroxidation products have been compared which are based on Fe(II) oxidation by them at acid pH values in the presence of xylenol orange which binds Fe(III) have been compared. The first method uses cumene hydropeoxide as an internal standard. In the second one, lipid peroxides are previously reduced by triphenylphosphine and these substances content is measured as a difference of the production of complexes with xylenol orange and iron ions in the control (with reduction) and experimental sample (without reduction). The optimization of measurement conditions is described. The levels of lipid peroxides in goldfish tissues assayed simultaneously by two methods were similar. The method with cumene hydroperoxide needs less amounts of biological material; moreover, there is no necessity in a calibration curve. Effects of hyperoxia on lipid peroxide levels in goldfish tissues were studied with the cumene method. Within the first hours of hyperoxia this index increased 13-times in the liver and 2-times in the brain and muscle. The further exposure rebounded this parameter to the initial level. Levels of lipid peroxides positively correlated with levels of end products of lipid peroxidation (thiobarbiturate acid reactive substances) in the goldfish tissues. The method of quantification of lipid peroxides with cumene is recommended for wide using in biological investigations.     

Peroxidase activity of cytochrome c in its compact state depends on dynamics of the heme region

Cytochrome c (cyt c) is a small globular hemoprotein with the main function as an electron carrier in mitochondrial respiratory chain. Cyt c possesses also peroxidase-like activity in the native state despite its six-coordinated heme iron. In this work, we studied the effect of increasing urea concentration in the range from 0 M to 6 M at pH 7 (pH value of the bulk solvent) and pH 5 (pH value close to negatively charged membrane) on peroxidase-like activity of cyt c. We show that peroxidase-like activity, measured by guaiacol oxidation and the ferrous oxidation in xylenol orange methods, correlates with the accessibility of the heme iron, which was assessed from the association rate constant of cyanide binding to cyt c. Cyt c peroxidase-like activity linearly increases in the pre-denaturational urea concentrations (0–4 M) at both studied pHs without an apparent formation of penta-coordinated state of the heme iron. Our results suggest that dynamic equilibrium among the denaturant-induced non-native coordination states of cyt c, very likely due to reversible unfolding of the least stable foldons, is pre-requisite for enhanced peroxidase-like activity of cyt c in its compact state. Dynamic replacement of the native sixth coordination bond of methionine-80 by lysines (72, 73, and 79) and partially also by histidines (26 and 33) provides an efficient way how to increase peroxidase-like activity of cyt c without significant conformational change at physiological conditions.     

Effect of dialysis on oxidative stress in uraemia

Abstract

It has been postulated that dialysis of patients with chronic renal failure (CRF) is associated with increased lipid peroxidation which may contribute to vascular and other complications of the syndrome. In the present study, a specific and precise technique [ferrous oxidation in xylenol orange (FOX) assay] was used to measure plasma lipid hydroperoxides (ROOHs) in three groups of uraemic patients. Patients were either studied before starting dialysis (n= 12) or on continuous ambulatory peritoneal dialysis (CAPD, n= 12) or haemodialysis (HD, n= 36) and compared to healthy controls (n=20). Plasma ROOHs were markedly elevated in HD patients compared with the controls (7.01±2.9 µM versus 4.25±2.05 µM; P < 0.005, Mann-Whitney test). Plasma ROOH concentrations in the CAPD patients were increased but not significantly higher than controls (5.36±3.56 µM versus 4.25±2.05 µM). By contrast, no differences in ROOH levels were found between controls and predialysis patients. There was no difference in plasma thiobarbituric acid reactive substances (TBARS)between control and the three CRF groups. Absolute and cholesterol standardised plasma α-tocopherol levels were lower in the patients (whether they were on dialysis or not) than in the controls (18.62±6.88 µM versus 22.73±5.33 µM; P < 0.01 and 1.99±1.88 µM/mM versus 5.25±1.0 µM/mM; P < 0.0005, respectively). This study provides direct evidence that enhanced oxidative stress in CRF patients is related to the dialysis treatment rather than the disease itself. Further studies will be necessary to establish the relationships between plasma measures of oxidative stress and cardiovascular complications in CRF patients under dialysis and whether treatment with antioxidants may reduce oxidative stress or reverse adverse effects associated with dialysis.     

Method for the simultaneous determination of free/protein malondialdehyde and lipid/protein hydroperoxides

A simple and sensitive method is presented for the simultaneous quantification (spectrophotometric and spectrofluorimetric) of the main lipid and protein peroxidation products after their initial fractionation: free malondialdehyde (FrMDA), protein-bound malondialdehyde (PrMDA), total hydroperoxides (LOOH), and protein hydroperoxides (PrOOH). FrMDA and PrMDA (released from proteins by alkaline hydrolysis) are measured after the reaction of MDA with thiobarbituric acid (TBA) under acidic conditions, by the specific fluorimetric quantification of the resulting MDA–(TBA)₂ adduct chromophore. The measurement of LOOH and PrOOH is based on the reaction of Fe³⁺ (resulting from the reaction of LOOH and PrOOH with Fe²⁺) with xylenol orange (XO) and the photometric quantification of the resulting XO–Fe complex. The sensitivity of the assays for FrMDA/PrMDA and LOOH/PrOOH is 20 and 100 pmol, respectively. The method was applied successfully on human plasma and can be used for the evaluation of oxidative stress in both basic and clinical research.     

In the presence of ferritin, visible light induces lipid peroxidation of the porcine photoreceptor outer segment

We studied the synergistic effect of visible light and ferritin on the lipid peroxidation on a fraction of porcine photoreceptor outer segment (POS). Reaction mixtures containing the POS fraction and horse spleen ferritin were irradiated under white fluorescent light mainly at 17,000 lx or incubated under dark conditions at 37°C. The lipid peroxidation was evaluated by both the thiobarbituric acid method and the ferrous oxidation/xylenol orange method. The irradiation-induced lipid peroxidation was affected by some experimental factors such as the irradiation dose and acidity of the material. When the irradiation was stopped, the lipid peroxidation was also stopped; thereafter, the re-irradiation induced lipid peroxidation. Moreover, this lipid peroxidation was inhibited by desferrioxamine, an iron chelator, or by dimethylthiourea, a hydroxyl radical scavenger, suggesting that the lipid peroxidation involves hydroxyl radicals generated via the Fenton reaction by iron ion released from ferritin. The lipid peroxidation did not take place under dark conditions or in the absence of ferritin. This study suggested the possibility that the visible light-induced lipid peroxidation of the POS fraction in the presence of ferritin may participate in the etiology of human retinal degenerative diseases as the human retina is exposed to light for life.     

H2O2-Mediated Damage to Lysosomal Membranes of J-774 Cells

The effects of hydrogen peroxide on cell viability and, in particular, on lysosomal integrity were investigated in a model system of cultured, established, macrophage-like J-774 cells. The cells were found to rapidly degrade added hydrogen peroxide, withstanding concentrations 250μM without cell death; however, all tested concentrations (100-500/μM) substantially decreased cellular ATP to approximately the same degree. Concentrations of hydrogen peroxide 500/μM resulted in a pronounced and rapid decrease in cell viability preceded by the loss of lysosomal integrity, as judged by the relocalization of acridine orange, a lysosomotropic weak base, in pre-labelled cells. Hydrogen peroxide-induced relocalization of acridine orange and cell death were either enhanced or much prevented, according to if the cells were initially allowed to endocytose ferric iron or the specific iron-chelator deferoxamine, respectively. Depletion of ATP, however, was not associated with the loss of lysosomal integrity and viability regardless of iron or deferoxamine pretreatment. Pre-exposure to E-64, an inhibitor of lysosomal thiol proteases, resulted in the reduction of both lysosomal membrane damage and cell death. The results are interpreted as indicating (i) generation of hydroxyl radicals within the secondary lysosomal compartment due to the occurrence of reactive ferrous iron, leading to (ii) peroxidative alterations of the lysosomal membrane resulting in (iii) loss of lysosomal membrane integrity with dissipation of the proton gradient and leakage of lysosomal contents, including hydrolytic enzymes, into the cell sap. The partial protection by E-64 may result from hydroxyl radical scavening by accumulated non-degraded autophagocytosed lysosomal material, and/or decreased availability of reactive redox-cycling iron due to decreased enzymatic digestion of autophagocytosed iron-containing metalloproteins. Moreover, our results show that the normal lysosomal content of iron, capable of redox cycling, of the cell line under study is enough to induce oxidative damage leading to loss of lysosomal integrity. It is suggested that lysosomal damage may be an important cause of cell degeneration under conditions of increased intra- or extracellular hydrogen peroxide-formation.     

In the presence of ferritin,visible light induces lipid peroxidation of the porcine photoreceptor outer segment

We studied the synergistic effect of visible light and ferritin on the lipid peroxidation on a fraction of porcine photoreceptor outer segment (POS). Reaction mixtures containing the POS fraction and horse spleen ferritin were irradiated under white fluorescent light mainly at 17,000 lx or incubated under dark conditions at 37°C. The lipid peroxidation was evaluated by both the thiobarbituric acid method and the ferrous oxidation/xylenol orange method. The irradiation-induced lipid peroxidation was affected by some experimental factors such as the irradiation dose and acidity of the material. When the irradiation was stopped, the lipid peroxidation was also stopped; thereafter, the re-irradiation induced lipid peroxidation. Moreover, this lipid peroxidation was inhibited by desferrioxamine, an iron chelator, or by dimethylthiourea, a hydroxyl radical scavenger, suggesting that the lipid peroxidation involves hydroxyl radicals generated via the Fenton reaction by iron ion released from ferritin. The lipid peroxidation did not take place under dark conditions or in the absence of ferritin. This study suggested the possibility that the visible light-induced lipid peroxidation of the POS fraction in the presence of ferritin may participate in the etiology of human retinal degenerative diseases as the human retina is exposed to light for life.     

Electrons initiate efficient formation of hydroperoxides from cysteine

Amino acid and protein hydroperoxides can constitute a significant hazard if formed in vivo. It has been suggested that cysteine can form hydroperoxides after intramolecular hydrogen transfer to the commonly produced cysteine sulfur-centered radical. The resultant cysteine-derived carbon-centered radicals can react with oxygen at almost diffusion-controlled rate, forming peroxyl radicals which can oxidize other molecules and be reduced to hydroperoxides in the process. No cysteine hydroperoxides have been found so far. In this study, dilute air-saturated cysteine solutions were exposed to radicals generated by ionizing radiation and the hydroperoxides measured by an iodide assay. Of the three primary radicals present, the hydroxyl, hydrogen atoms and hydrated electrons, the first two were ineffective. However, electrons did initiate the generation of hydroperoxides by removing the –SH group and forming cysteine-derived carbon radicals. Under optimal conditions, 100% of the electrons reacting with cysteine produced the hydroperoxides with a 1:1 stoichiometry. Maximum hydroperoxide yields were at pH 5.5, with fairly rapid decline under more acid or alkaline conditions. The hydroperoxides were stable between pH 3 and 7.5, and decomposed in alkaline solutions. The results suggest that formation of cysteine hydroperoxides initiated by electrons is an unlikely event under physiological conditions.     

Toward Arabidopsis thaliana hydrophilic metabolome: assessment of extraction methods and quantitative 1H NMR

Our goal was to establish the hydrophilic metabolome of heterotrophic Arabidopsis thaliana cells grown in suspension, a cellular model of plant sink tissues. Water‐soluble metabolites were extracted using four protocols: perchloric acid, boiling ethanol, methanol and methanol/chloroform (M/Chl). They were detected and quantified using ¹H nuclear magnetic resonance (NMR) spectroscopy at 400 MHz. Extraction yields and reproducibility of the extraction methods were investigated. The effects of cell harvest protocol, cell grinding and lyophilization and storage conditions on the measured metabolic profiles were also studied. These quantitative studies demonstrated for the first time that the four extraction protocols commonly used do lead to quite similar molecular compositions as analyzed by ¹H NMR. The M/Chl method proved effective and reliable to prepare series of physiologically significant extracts from plant cells for ¹H NMR analysis. Reproducibility of the detected metabolome was assessed over long periods of time by analyzing a large number of separate extracts prepared from independent cultures. Larger variations in the NMR metabolite profiles could be correlated to changes in physiological parameters of the culture medium. Quantitative resolved ¹H NMR of cell extracts proved to be robust and reliable for routine metabolite profiling of plant cell cultures.     

Measurement of choline and choline metabolite concentrations using high-pressure liquid chromatography and gas chromatography-mass spectrometry

We have developed a reproducible and sensitive procedure for the isolation and measurement of choline, phosphocholine, glycerophosphocholine, phosphatidylcholine, lysophosphatidylcholine and acetylcholine in a single 100-mg sample of biological tissue. Tissues were spiked with 14C-methyl- and 2H-methyl- or 15N-choline labeled internal standards for each compound. They were extracted with chloroform/methanol/water and the aqueous and organic phases were dried. The organic phase was resuspended in chloroform/methanol (1/1, v/v) and an aliquot was applied to a silica-gel thin-layer chromatography plate. The plate was developed in chloroform/methanol/water (65/30/4, v/v). Segments which cochromatographed with external standards of phosphatidylcholine and lysophosphatidylcholine were stained, scraped, and hydrolyzed in 6 M methanolic-HCl at 80 degrees C for 60 min, liberating free choline. The aqueous phase was resuspended in methanol/water and injected onto a silica HPLC column. Choline and its metabolites were eluted using a binary nonlinear gradient of acetonitrile/ethanol/acetic acid/1 M ammonium acetate/water/0.1 M sodium phosphate (800/68/2/3/127/10, v/v changing to 400/68/44/88/400/10, v/v). Peaks were detected with an on-line radiometric detector, collected, and dried under vacuum. Each choline ester was digested in 6 M HCl at 80 degrees C to form choline. Choline was then converted to the propionyl ester and demethylated with sodium benzenethiolate. This volatile derivative was then isolated using gas chromatography and measured with a mass selective detector. Deuterated internal standards were used to correct for variations in recovery. Choline, glycerophosphocholine, phosphocholine, phosphatidylcholine, lysophosphatidylcholine, and acetylcholine were measured in rat liver, heart, muscle, kidney, plasma, red blood cells, and brain and in human plasma.(ABSTRACT TRUNCATED AT 250 WORDS)