Determination of peroxisomal fatty acyl-CoA oxidase activity using a lauroyl-CoA-based fluorometric assay

A simple, sensitive fluorometric method for the determination of peroxisomal fatty acyl-CoA oxidase (EC 1.3.99.3) activity has been developed. Studies of enzyme activity relative to subcellular distribution and to clofibrate induction indicate that this assay is specific for peroxisomal fatty acyl-CoA oxidase. The lauroyl-CoA-dependent production of H2O2 is quantitated by measuring the oxidation of 4-hydroxyphenyl-acetic acid to a fluorescent product in a horseradish peroxidase-coupled assay. Assays can be performed in either a fixed time or continuous mode. In either mode, H2O2 production is related to a change in fluorescence intensity through use of a standard curve generated with known amounts of H2O2. The use of lauroyl-CoA (12:0), rather than the more generally used substrate palmitoyl-CoA (16:0), provides significant advantages. Much of the substrate inhibition problem associated with palmitoyl-CoA has been avoided, and a greater than 4.5-fold higher specific activity has been achieved compared with a palmitoyl-CoA-based assay. In the fixed-time mode, linearity relative to time and to the amount of enzyme added has been established without resorting to the use of bovine serum albumin as a substrate binding medium. Sensitivity is estimated to be at least equal to that of the most sensitive methods reported, while reliability, versatility and range have been improved. Use of this method should greatly facilitate the study of peroxisomal beta-oxidation regulatory mechanisms in hepatocyte cell culture systems as well as in other circumstances where low activities or small samples must be assayed.     

A sensitive spectrophotometric assay for peroxisomal acyl-CoA oxidase.

A simple spectrophotometric assay was developed for peroxisomal fatty acyl-CoA oxidase activity. The assay, based on the H2O2-dependent oxidation of leuco-dichlorofluorescein catalysed by exogenous peroxidase, is more sensitive than methods previously described. By using mouse liver samples, cofactor requirements were assessed and a linear relationship was demonstrated between dye oxidation and enzyme concentration. By using this assay on subcellular fractions, palmitoyl-CoA oxidase activity was localized for the first time in microperoxisomes of rat intestine. The assay was also adapted to measure D-amino acid oxidase activity, demonstrating the versatility of this method for measuring activity of other H2O2-producing oxidases.     

Detection of peroxisomal fatty acyl-coenzyme A oxidase activity.

It has been postulated that the peroxisomal fatty acid-oxidizing system [Lazarow & de Duve (1976) Proc. Natl. Acad. Sci. U.S.A. 73, 2043--2046; Lazarow (1978) J. Biol. Chem. 253, 1522--1528] resembles that of mitochondria, except for the first oxidative reaction. In this step, O2 would be directly reduced to H2O2 by an oxidase. Two specific procedures developed to detect the activity of the characteristic enzyme fatty acyl-CoA oxidase are presented, namely polarographic detection of palmitoyl-CoA-dependent cyanide-insensitive O2 consumption and palmitoyl-CoA-dependent H2O2 generation coupled to the peroxidation of methanol in an antimycin A-insensitive reaction. Fatty acyl-CoA oxidase activity is stimulated by FAD, which supports the flavoprotein nature postulated for this enzyme. Its activity increases 7-fold per g wet wt. of liver in rats treated with nafenopin, a hypolipidaemic drug. Subcellular fractionation of livers from normal and nafenopin-treated animals provides evidence for its peroxisomal localization. The stoicheiometry for palmitoyl-CoA-dependent O2 consumption, H2O2 generation and NAD+ reduction is 1 : 1 : 1. This suggests that fatty acyl-CoA oxidase is the rate-limiting enzyme of the peroxisomal fatty acid-oxidizing system.     

Intrinsic enoyl-CoA isomerase activity of rat acyl-CoA oxidase I

Rat peroxisomal acyl-CoA oxidase I is a key enzyme for the beta-oxidation of fatty acids, and the deficiency of this enzyme in patient has been previously reported. It was found that rat acyl-CoA oxidase I has intrinsic enoyl-CoA isomerase activity, which was confirmed using incubation followed with HPLC analysis in this study. Various 3-enoyl-CoA substrates with cis or trans configuration were synthesized and used in the study of enzyme substrate specificity. The isomerase activity of the enzyme was characterized through studies of kinetics, pH dependence, and enzyme inhibition. Most k(cat)/K(M) values of rat peroxisomal acyl-CoA oxidase I for isomerization reaction are comparable with those of authentic rat liver peroxisomal Delta(3)-Delta(2)-enoyl-CoA isomerase and rat liver peroxisomal multifunctional enzyme 1 when hexenoyl-CoA and octenoyl-CoA with cis- or trans-configuration were used as substrate. Glu421 was found to be the catalytic residue for both oxidase and isomerase activities of the enzyme. The isomerase activity of rat peroxisomal acyl-CoA oxidase I is probably due to a spontaneous process driven by thermodynamic equilibrium with formation of a conjugated structure after deprotonation of substrate alpha-proton. The energy level of transition state may be lowered by a stable dienolate intermediate, which gain further stabilization via charge transfer with electron-deficient FAD cofactor of the enzyme.     

Acyl-CoA synthetase and the peroxisomal enzymes of beta-oxidation in human liver. Quantitative analysis of their subcellular localization.

The presence of acyl-CoA synthetase (EC 6.2.1.3) in peroxisomes and the subcellular distribution of beta-oxidation enzymes in human liver were investigated by using a single-step fractionation method of whole liver homogenates in metrizamide continuous density gradients and a novel procedure of computer analysis of results. Peroxisomes were found to contain 16% of the liver palmitoyl-CoA synthetase activity, and 21% and 60% of the enzyme activity was localized in mitochondria and microsomal fractions respectively. Fatty acyl-CoA oxidase was localized exclusively in peroxisomes, confirming previous results. Human liver peroxisomes were found to contribute 13%, 17% and 11% of the liver activities of crotonase, beta-hydroxyacyl-CoA dehydrogenase and thiolase respectively. The absolute activities found in peroxisomes for the enzymes investigated suggest that in human liver fatty acyl-CoA oxidase is the rate-limiting enzyme of the peroxisomal beta-oxidation pathway, when palmitic acid is the substrate.     

Significance of catalase in peroxisomal fatty acyl-CoA beta-oxidation

Catalase activity was inhibited by aminotriazole administration to rats in order to evaluate the influence of catalase on the peroxisomal fatty acyl-CoA beta-oxidation system. 2 h after the administration of aminotriazole, peroxisomes were prepared from rat liver, and the activities of catalase, the beta-oxidation system and individual enzymes of beta-oxidation (fatty acyl-CoA oxidase, crotonase, beta-hydroxybutyryl-CoA dehydrogenase and thiolase) were determined. Catalase activity was decreased to about 2% of the control. Among the individual enzymes of the beta-oxidation system, thiolase activity was decreased to 67%, but the activities of fatty acyl-CoA oxidase, crotonase and beta-hydroxybutyryl-CoA dehydrogenase were almost unchanged. The activity of the peroxisomal beta-oxidation system was assayed by measuring palmitoyl-CoA-dependent NADH formation, and the activity of the purified peroxisome preparation was found to be almost unaffected by the administration of aminotriazole. The activity of the system in the aminotriazole-treated preparation was, however, significantly decreased to 55% by addition of 0.1 mM H2O2 to the incubation mixture. Hydrogen peroxide (0.1 mM) reduced the thiolase activity of the aminotriazole-treated peroxisomes to approx. 40%, but did not affect the other activities of the system. Thiolase activity of the control preparation was decreased to 70% by addition of hydrogen peroxide (0.1 mM). The half-life of 0.1 mM H2O2 added to the thiolase assay mixture was 2.8 min in the case of aminotriazole-treated peroxisomes, and 4 s in control peroxisomes. The ultraviolet spectrum of acetoacetyl-CoA (substrate of thiolase) was clearly changed by addition of 0.1 mM H2O2 to the thiolase assay mixture without the enzyme preparation; the absorption bands at around 233 nm (possibly due to the thioester bond of acetoacetyl-CoA) and at around 303 nm (due to formation of the enolate ion) were both significantly decreased. These results suggest that H2O2 accumulated in peroxisomes after aminotriazole treatment may modify both thiolase and its substrate, and consequently suppress the fatty acyl-CoA beta-oxidation. Therefore, catalase may protect thiolase and its substrate, 3-ketoacyl-CoA, by removing H2O2, which is abundantly produced during peroxisomal enzyme reactions.     

Peroxisomal oxidases and catalase in liver and kidney homogenates of normal and di(ethylhexyl)phthalate-fed rats.

1. Activities of peroxisomal oxidases and catalase were assayed at neutral and alkaline pH in liver and kidney homogenates from male rats fed a diet with or without 2% di(2-ethylhexyl)phthalate (DEHP) for 12 days. 2. All enzyme activities were higher at alkaline than at neutral pH in both groups. 3. The effect of the DEHP-diet on the peroxisomal enzymes was different in kidney and liver. Acyl-CoA oxidase activity was raised three- and sixfold in kidney and liver homogenates, respectively. The activity of D-amino acid oxidase decrease in liver, but increased in kidney homogenates. In liver homogenates, urate oxidase activity was not affected by the DEHP diet. The catalase activity was twofold induced in liver, but not in kidney. 4. The differences suggest that the changes of peroxisomal enzyme activities by DEHP treatment are not directly related to peroxisome proliferation. 5. DEHP treatment caused a marked increase of total and peroxisomal fatty acid oxidation in rat liver homogenates. 6. In the control group the rate of peroxisomal fatty acid oxidation was higher at alkaline pH than at neutral pH. 7. This rate was equal at both pH values in the DEHP-fed group, in contrast to the acyl-CoA oxidase activity. These results indicate that after DEHP treatment other parameters than acyl-CoA oxidase activity become limiting for peroxisomal beta-oxidation.     

Fluorometric assay of peroxisomal oxidases

The present paper deals with the adaptation of the fluorometric measurement of H2O2 originally described by Guilbault et al. (1967, Anal. Chem. 39, 271) for the assay of the peroxisomal oxidation of D-amino acids, L-alpha-hydroxyacids, uric acid, and acyl-CoA esters. The present work essentially covers three facets: (i) the general kinetics of the assay of peroxisomal oxidases and the influence of each component of the assay medium on these kinetics; (ii) the measurement of peroxisomal oxidase activities in subcellular fractions and tissues from human and untreated and clofibrate-treated rodents; and (iii) the comparison between the oxidase activities measured by the fluorometric and spectrophotometric methods.     

Inhibition of peroxisomal fatty acyl-CoA oxidase by antimycin A.

Peroxisomal fatty acyl-CoA oxidase was inhibited by micromolar concentrations of antimycin A, an inhibitor of mitochondrial respiration. The inhibition was observed with all three substrates tested, i.e. palmitoyl-CoA, trihydroxycoprostanoyl-CoA and hexadecanedioyl-CoA. The peroxisomal D-amino acid oxidase was also inhibited by antimycin, but the peroxisomal L-alpha-hydroxyacid oxidase and uric acid oxidase and the mitochondrial monoamine oxidase were not. The degree of inhibition of acyl-CoA oxidase by antimycin was strongly dependent on the amount of cellular protein present in the assay mixture: at a fixed antimycin concentration, the inhibition was gradually lost with increasing protein concentrations. At a fixed cellular protein concentration in the assay mixtures, the mitochondrial oxidation of glutamate or palmitoylcarnitine was inhibited at antimycin concentrations that were much lower than those required for the inhibition of fatty acyl-CoA oxidase. Our results, nevertheless, demonstrate that antimycin A must be used with caution, when it is added to homogenates or subcellular fractions in order to distinguish between mitochondrial and peroxisomal fatty acid oxidation.     

Acyl-CoA oxidase activity and peroxisomal fatty acid oxidation in rat tissues

Acyl-CoA oxidase, the first enzyme of the peroxisomal β-oxidation, was proved to be rate-limiting for this process in homogenates of rat liver, kidney, adrenal gland, heart and skeletal muscle. Acyl-CoA oxidase activity, based on H₂O₂-dependent leuko-dichlorofluorescein oxidation in tissue extract, was compared with radiochemically assayed peroxisomal β-oxidation rates. Dichlorofluorescein production was a valid measure of peroxisomal fatty acid oxidation only in liver and kidney, but not in adrenal gland, heart or skeletal muscle. Production of ¹⁴C-labeled acid-soluble products from 1-¹⁴C-labeled fatty acids in the presence of antimycin-rotenone appears to be a more accurate and sensitive estimate of peroxisomal β-oxidation than the acyl-CoA oxidase activity on base of H₂O₂ production. Chain-length specificity of acyl-CoA oxidase changed with the acyl-CoA concentrations used. Below 80 μM, palmitoyl-CoA showed the highest activity of the measured substrates in rat liver extract. No indications were obtained for the presence in rat liver of more forms of acyl-CoA oxidase with different chain-length specificity.     

Glucagon and fasting do not activate peroxisomal fatty acid beta-oxidation in rat liver

In a study of the endocrine control of peroxisomes, the effects of acute glucagon treatment and fasting on hepatic peroxisomal beta-oxidation in rats have been investigated. The activity of the rate-limiting peroxisomal beta-oxidation enzyme, fatty acyl-CoA oxidase, was measured to determine whether activation of peroxisomal beta-oxidation could account for the increase in total hepatic fatty acid oxidation following acute glucagon exposure. Catalase, a peroxisomal enzyme not directly involved in beta-oxidation, was also measured as a control for total peroxisomal activity. No changes with acute glucagon treatment of intact animals were observed with either activity as measured in liver homogenates or partially purified peroxisomal fractions. These observations indicate the lack of acute control by glucagon of peroxisomal function at the level of total enzyme activity. Previous work on the effects of fasting on hepatic fatty acid beta-oxidation [H. Ishii, S. Horie, and T. Suga (1980) J. Biochem. 87, 1855-1858] suggested an enhanced role for the peroxisomal beta-oxidation pathway during starvation. It was found that the peroxisomal beta-oxidation system, as measured by fatty acyl-CoA oxidase activity, does increase with duration of fast when expressed on a per gram wet weight liver basis. However, when this activity is expressed as total liver capacity, a decline in activity with increasing duration of fast is observed. Furthermore, this decline in peroxisomal capacity parallels the decline in total liver capacity for citrate synthase, a mitochondrial matrix enzyme, and total liver protein. These data indicate that peroxisomal beta-oxidation activity is neither stimulated nor even preferentially spared from proteolysis during fasting.     

Chlorpromazine and carnitine-dependency of rat liver peroxisomal beta-oxidation of long-chain fatty acids.

The enzyme targets for chlorpromazine inhibition of rat liver peroxisomal and mitochondrial oxidations of fatty acids were studied. Effects of chlorpromazine on total fatty acyl-CoA synthetase activity, on both the first and the third steps of peroxisomal beta-oxidation, on the entry of fatty acyl-CoA esters into the peroxisome and on catalase activity, which allows breakdown of the H2O2 generated during the acyl-CoA oxidase step, were analysed. On all these metabolic processes, chlorpromazine was found to have no inhibitory action. Conversely, peroxisomal carnitine octanoyltransferase activity was depressed by 0.2-1 mM-chlorpromazine, which also inhibits mitochondrial carnitine palmitoyltransferase activity in all conditions in which these enzyme reactions are assayed. Different patterns of inhibition by the drug were, however, demonstrated for both these enzyme activities. Inhibitory effects of chlorpromazine on mitochondrial cytochrome c oxidase activity were also described. Inhibitions of both cytochrome c oxidase and carnitine palmitoyltransferase are proposed to explain the decreased mitochondrial fatty acid oxidation with 0.4-1.0 mM-chlorpromazine reported by Leighton, Persico & Necochea [(1984) Biochem. Biophys. Res. Commun. 120, 505-511], whereas depression by the drug of carnitine octanoyltransferase activity is presented as the factor responsible for the decreased peroxisomal beta-oxidizing activity described by the above workers.     

The presence of acyl-CoA hydrolase in rat brown-adipose-tissue peroxisomes.

The subcellular distribution of acyl-CoA hydrolase was studied in rat brown adipose tissue, with special emphasis on possible peroxisomal localization. Subcellular fractionation by sucrose-density-gradient centrifugation, followed by measurement of short-chain (propionyl-CoA) acyl-CoA hydrolase in the presence of NADH, resulted in two peaks of activity in the gradient: one peak corresponded to the distribution of cytochrome oxidase (mitochondrial marker enzyme), and another peak of activity coincided with the peroxisomal marker enzyme catalase. The distribution of the NADH-inhibited short-chain hydrolase activity fully resembled that of cytochrome oxidase. The substrate-specificity curve of the peroxisomal acyl-CoA hydrolase activity indicated the presence of a single enzyme exhibiting a broad substrate specificity, with maximal activity towards fatty acids with chain lengths of 3-12 carbon atoms. The mitochondrial acyl-CoA hydrolase substrate specificity, in contrast, indicated the presence of at least two acyl-CoA hydrolases (of short- and medium-chain-length specificity). The peroxisomal acyl-CoA hydrolase activity was inhibited by CoA at low (microM) concentrations and by ATP at high concentrations (greater than 0.8 mM). In contrast with the mitochondrial short-chain hydrolase, the peroxisomal acyl-CoA hydrolase activity was not inhibited by NADH.     

Properties of fatty acyl-CoA oxidase from rat liver, a peroxisomal flavoprotein

Fatty acyl-CoA oxidase from rat liver was partially purified and characterized as a peroxisomal flavoprotein oxidase. A sedimentation coefficient of 7.7 S was estimated from sucrose gradients and a Stokes radius of 42.3 Å was deduced from gel-exclusion chromatography. These data allow to estimate a molecular weight of 136,000 and a frictional ratio of 1.1. FAD, specifically required as a prosthetic group, is weakly bound. Still, FAD displays greater affinity for the free ap$\text{o}$ enzyme than for the putative apoenzyme-substrate complex formed with palmitoyl-CoA. In addition, it was established that the subcellular distribution of the fatty acyl-CoA oxidase, in complete liver homogenates fractionated in Metrizamide density gradients, parallels that of the peroxisomal marker catalase.     

Differential effects of freezing on total hepatic peroxisomal fatty acid oxidation and on the rate-limiting enzyme, acyl-CoA oxidase

Fatty acid oxidation defects can be acutely fatal, leading to the collection of tissues which are frozen for future analysis. Since peroxisomes can also oxidize long-chain fatty acids, differentiation of the contributions from the peroxisome as opposed to the mitochondria is important. We studied the effects of freezing and storage of rat livers on peroxisomal and mitochondrial beta-oxidation as measured by cyanide sensitivity of the oxidation of [1-14C]oleoyl-CoA to 14CO2 and acid-soluble labeled products. In addition, we examined the effects of freezing and storage on the rate-limiting enzyme for peroxisomal beta-oxidation, acyl-CoA oxidase, by the H2O2 generation method. Marked reduction in the oxidation of [1-14C]oleoyl-CoA was found for both peroxisomal and mitochondrial systems upon freezing at -18 or -70 degrees C for 2 days which declined further on storage at these temperatures for 12 weeks. Loss of activity after freezing was greater for the mitochondrial than the peroxisomal beta-oxidation system. By contrast, acyl-CoA oxidase activity was resistant to these changes, maintaining prefrozen activities despite storage for 12 weeks. The contribution of the peroxisomal system to beta-oxidation was 32% of the total rate of oxidation of [1-14C]oleoyl-CoA in the rat liver. These findings indicate that the contributions of the peroxisomal system to total fatty acid oxidation may be considerable, that freezing of the liver results in drastic reduction in enzyme activities of both peroxisomal as well as mitochondrial beta-oxidation, but that the rate-limiting enzyme of the peroxisomal system, acyl-CoA oxidase, retains full activity despite freezing and storage.     

The effect of di(ethylhexyl)phthalate on fatty acid oxidation and carnitine palmitoyltransferase in various rat tissues

Male rats were fed a diet with or without 2% di(2-ethylhexyl)phthalate (DEHP) for 12 days. Total and peroxisomal oxidation rates of palmitic and arachidonic acid were increased in homogenates of liver and kidney after DEHP administration. The relative peroxisomal contribution to the total oxidation was only higher in liver. The activities of acyl-CoA oxidase and carnitine palmitoyltransferase were also higher in both tissues. Immunoblots showed that the increase of fatty acid oxidation was associated with a higher concentration of enzymes of peroxisomal and mitochondrial beta-oxidation. DEHP did not change total and peroxisomal fatty acid oxidation and activity of carnitine palmitoyltransferase of homogenates of heart and skeletal muscle. The cause for the tissue-specific response is discussed.     

Expression and purification of his-tagged rat peroxisomal acyl-CoA oxidase I wild-type and E421 mutant proteins

Rat peroxisomal acyl-CoA oxidase I is a key enzyme for the beta-oxidation of fatty acids, and the deficiency of this enzyme in patients has been previously reported. We cloned the gene of rat peroxisomal acyl-CoA oxidase I into a bacterial expression vector pLM1 with six continuous histidine codons attached to the 5' end of the gene. The cloned gene was overexpressed in Escherichia coli and the soluble protein was purified with a nickel HiTrap chelating metal-affinity column in 90% yield to apparent homogeneity. The specific activity of the purified His-tagged rat peroxisomal acyl-CoA oxidase I was 1.5 micromol/min/mg. It has been proposed that Glu421 is a catalytic residue responsible for deprotonation of alpha-proton of acyl-CoA substrate. We constructed four mutant expression plasmids of the enzyme, pACO(E421D), pACO(E421A), pACO(E421Q), and pACO(E421G) using site-directed mutagenesis. Mutant proteins were overexpressed in E. coli and purified with a nickel metal-affinity column. Kinetic studies of wild-type and mutant proteins were carried out, and the result confirmed that Glu421 is a catalytic residue of rat peroxisomal acyl-CoA oxidase I. Our overexpression in E. coli and one-step purification of the highly active N-terminal His-tagged rat peroxisomal acyl-CoA oxidase I greatly facilitated our further investigation of this enzyme, and our result from site-directed mutagenesis increased our understanding of the mechanism for the reaction catalyzed by peroxisomal acyl-CoA oxidase I.     

The oxidation of dicarboxylic acid CoA esters via peroxisomal fatty acyl-CoA oxidase

Evidence supporting a common peroxisomal beta-oxidation pathway for the coenzyme A thioesters of medium-chain-length dicarboxylic acids (DCn-CoA) and monocarboxylic acids (MCn-CoA) has been obtained. Using the mono-CoA esters of dodecanedioic acid (DC12-CoA) and lauroyl-CoA (MC12-CoA) as substrates, parallel inductions of activities and parallel increases in specific activities during purification of peroxisomal fatty acyl-CoA oxidase (EC 1.3.99.3) from rat liver after di(2-ethylhexyl)phthalate treatment were seen. The purified enzyme was used for antiserum production in rabbits; antiserum specificity was verified by immunoblot analysis. Coincident losses of oxidase activities with MC12-CoA and DC12-CoA were found in immunotitration experiments with rat liver homogenates, supporting the hypothesis that peroxisomal fatty acyl-CoA oxidase is solely responsible for the oxidation of medium-chain length dicarboxylic acid substrates. Kinetic studies with purified enzyme using the mono-CoA esters of sebacic (DC10-CoA), suberic (DC8-CoA), and adipic (DC6-CoA) acids along with DC12-CoA revealed substrate inhibition. Although these substrates exhibited similar calculated Vmax values, with decreasing chain length, the combination of increasing Km values and decreasing substrate inhibition constant (Ki) caused the maximum obtainable velocity to decrease. These studies offer an explanation for the previously observed limit of the ability of peroxisomes to chain-shorten dicarboxylates and increased urinary excretion of adipic acid when peroxisomal oxidation of dicarboxylic acids is enhanced.     

Interactions between the omega- and beta-oxidations of fatty acids

Long-chain monocarboxylic, omega-hydroxymonocarboxylic and dicarboxylic acids were activated approximately at the same rate by rat liver homogenates into their CoA esters (2-3 U/g liver). These acyl-CoA were substrates for rat liver peroxisomal beta-oxidation. The distribution of the peroxisomal oxidation of these substrates was also studied in various tissues. Rat liver mitochondria were capable of oxidizing long-chain monocarboxyl- and omega-hydroxymonocarboxylyl-CoAs but not dicarboxylyl-CoAs. When the mitochondrial preparations were incubated in coupling conditions, the addition of either free decanoic acid or free 10-hydroxydecanoic acid resulted in an increase of the oxygen uptake conversely to the addition of decanedioic acid. The comparative study of the chain-length substrate specificity of peroxisomal fatty acyl-CoA oxidase and mitochondrial fatty acyl-CoA dehydrogenase activities revealed that, actually, both types of organelles, peroxisomes and mitochondria, contain "oxido-reductases" active on long-chain monocarboxylyl-CoAs, omega-hydroxymonocarboxylyl-CoAs and dicarboxylyl-CoAs.