Effects of oxysterols on cell viability, inflammatory cytokines, VEGF, and reactive oxygen species production on human retinal cells: cytoprotective effects and prevention of VEGF secretion by resveratrol.

BACKGROUND AND AIMS: Oxysterols are assumed to play important roles in age-related macular degeneration, a major cause of blindness. So we characterized the cytotoxic, oxidative, inflammatory, and angiogenic activities of oxysterols (7ß-hydroxycholesterol (7ß-OH), 7-ketocholesterol (7KC), 25-hydroxycholesterol (25-OH)) in human retinal ARPE-19 cells, and evaluated the protective effects of resveratrol (Rsv: 1 µM), a polyphenol from red wine. METHODS: ARPE-19 cells were treated with 7ß-OH, 7KC, or 25-OH (5-40 µg/mL; 24-48 h) without or with Rsv. Cell viability was determined using trypan blue and the MTT assay. Cell death was characterized by electron microscopy and in situ detection of activated caspases with fluorochrome-labeled inhibitors of caspases. Reactive oxygen species (ROS) production was measured with hydroethidine. ELISA methods and a cytometric bead assay were used to quantify cytokines involved in inflammation (IL-8, IL-1ß, IL-6, IL-10, IL-12p70, TNF-α, MCP-1) and VEGF. RESULTS: 7ß-OH and 7KC triggered a caspase-independent cell death process associated with the presence of multilamellar cytoplasmic structures evocating phospholipidosis, increased ROS production, and IL-8 secretion. 7ß-OH enhanced VEGF secretion. No cytotoxic effects were identified with 25-OH, which highly stimulated ROS production, MCP-1, and VEGF secretion. With oxysterols, no IL-10, TNF-α, and IL-12p70 secretion were detected. 25-OH induced IL-8 secretion through the MEK/ERK½ signaling pathway, and Rsv showed cytoprotective activities and inhibited VEGF secretion. CONCLUSION: 7ß-OH, 7KC, and 25-OH have cytotoxic, oxidative, inflammatory, and/or angiogenic activities on ARPE-19 cells. As Rsv has some protective effects against oxysterol-induced cell death and VEGF secretion it could be valuable in ARMD treatment.

Evidence for the presence of one carboxyl group in the catalytic center of D-beta-hydroxybutyrate dehydrogenase. Inactivation and binding studies with carbodiimide reagents.

D-beta-Hydroxybutyrate dehydrogenase D-3-hydroxybutyrate: NAD+ oxidoreductase, EC 1.1.1.30), a phosphatidylcholine-requiring enzyme, was irreversibly inactivated by a water-soluble carbodiimide, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDAC) or a hydrophobic carbodiimide, N,N'-dicyclohexylcarbodiimide (DCCD). The inactivation is pseudo-first-order with a kinetic stoichiometry of about 1. Phospholipid-free apoenzyme was more sensitive towards these reagents than reconstituted phospholipid-enzyme or membrane-bound enzyme forms. Reduced coenzyme (NADH) protected the enzyme against the inactivation, while oxidized coenzyme (NAD+) in presence of 2-methylmalonate (a pseudo-substrate) gave a better protection. It was found that the phospholipid-free apoenzyme bound about 1 mol [14C]DCCD. This incorporation was prevented by EDAC, indicating that both reagents react at the same site. [14C]Glycine ethyl ester, a nucleophilic compound which reacts specifically with the carboxylcarbodiimide derivative was incorporated to the enzyme (1 mol [14C]glycine ethyl ester per polypeptide chain), whatever its form, in the presence of DCCD or EDAC. These results indicate the presence of one carboxyl group probably located at or near the coenzyme-binding site and near the interacting domain of the enzyme with phospholipid.

Metabolic control of the expression of mitochondrial D-beta-hydroxybutyrate dehydrogenase, a ketone body converting enzyme.

The effects of various metabolic conditions inducing an overproduction of ketone bodies in the rat were studied at different levels of D-beta-hydroxybutyrate dehydrogenase expression, i.e., enzymatic activity and protein content in purified mitochondria, and translational activity of isolated free cytosolic polysomes. The strongest variations were obtained in diabetes mellitus where the D-beta-hydroxybutyrate dehydrogenase expression is largely decreased. Insulin can reverse this strong effect. Modulation of liver enzyme activity and of enzyme content was observed under the other conditions tested, i.e., a decrease and an increase in starvation and hyperlipidemic conditions, respectively. A comparative study was carried out on the enzyme of extrahepatic tissues, i.e., heart, kidney and brain. The results indicate that the D-beta-hydroxybutyrate dehydrogenase function appears to be controlled, at least at the translational, post-translational and catalytic levels.

Properties of peroxisomes from jerboa (Jaculus orientalis).

This report presents new data on mammalian peroxisomes by studying an unusual rodent: the jerboa (Jaculus orientalis). This animal exhibits some unique peroxisomal properties compared to the rat, such as higher cyanide-insensitive palmitoyl-CoA oxidase specific activity, pattern differences in SDS-PAGE peroxisomal proteins as well as in acyl-CoA oxidase immunoblotting. There is also a peculiar response to a peroxisome proliferator, ciprofibrate. With 250 ppm of ciprofibrate in the diet for 2 weeks, we observed a limited liver peroxisome proliferation as well as a palmitoyl-CoA oxidase activity, enzyme content and mRNA increase. However, there was no increase in catalase activity, nor hepatomegaly which are prominent features of peroxisome proliferation in rats treated under the same conditions. The palmitoyl-CoA oxidase activity increase was weak in the kidney and not observed in the heart. Other subcellular organelle marker enzyme activities did not significantly change, especially the mitochondrial D-3-hydroxybutyrate and succinate dehydrogenases, lysosomal acid phosphatase, cytosolic lactate dehydrogenase and the endoplasmic reticulum NADPH-cytochrome c reductase. However, the activity of the liver membrane endoplasmic reticulum linked omega-lauryl hydroxylase (cytochrome P450 IV A1) increases after ciprofibrate treatment. Jerboa also behaves differently compared to the guinea pig after ciprofibrate treatment since the guinea pig has a weak response towards peroxisome proliferators. In conclusion, this first peroxisome study utilizing a different type of rodent as a laboratory animal, reveals that the jerboa shows unique peroxisome properties and responds in a moderate manner to a peroxisome proliferator, ciprofibrate, without leading to any increase in liver mass. This supports the fact that fibrate molecules may have different targets depending upon the species.

The analysis of modified peroxisome proliferator responsive elements of the peroxisomal bifunctional enzyme in transfected HepG2 cells reveals two regulatory motifs.

Peroxisome proliferators (PPs) are non-genotoxic carcinogens in rodents. They can induce the expression of numerous genes via the heterodimerization of two members of the steroid hormone receptor superfamily, called the peroxisome proliferator-activated receptor (PPAR) and the 9-cis retinoic acid receptor (RXR). Many of the PP responsive genes possess a peroxisome proliferator response element (PPRE) formed by two TGACCT-related motifs. The bifunctional enzyme (HD) PPRE contains 3 such motifs, creating DR1 and DR2 sequences. PPAR and RXR regulate transcription via the DR1 element while DR2 modulates the expression of the gene via auxiliary factors in HepG2 cells.

Cloning and tissue expression of two cDNAs encoding the peroxisomal 2-enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase in the guinea pig liver.

The 2-enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase (HD) is the second enzyme of the peroxisomal beta-oxidation pathway. In human and rat, only one HD mRNA has been so far detected in the liver. This paper reports for the first time in a mammal species, the guinea pig, the cloning and sequencing of two cDNAs encoding an HD. The 3,274 nucleotide-cDNA is a strictly identical but longer copy of the 2,494 nucleotide-form. A 2,178 bp-open reading frame encodes a protein of 726 amino acids (M(r) 79.3 kDa) with the peroxisomal-targeting signal (tripeptide SKL) at the carboxyterminus. Northern blot analysis of HD mRNA identified three mRNAs of respective sizes 3.5, 2.6 and 1.6 kb in the guinea pig liver and kidneys.

Photolabelling of D-beta-hydroxybutyrate apodehydrogenase with azidoaryl phospholipids.

Two synthetic photoactive azidoarylphosphatidylcholines were used to investigate the level of interaction between D-beta-hydroxybutyrate apodehydrogenase (apoBDH), an amphipathic membrane protein, with the hydrophobic domain of phospholipids. The two synthetic lecithins, PL I (1-myristoyl-2-12-N-(4-azido-2-nitrophenyl) aminododecanoyl-sn-glycero-3-phosphocholine) and PL II (1-myristoyl-2-(2-azido-4-nitrobenzoyl)-sn-glycero-3-phosphocholine), are able to reactivate the non-active purified apoBDH as well as the non-photoactive homologs, indicating that the photoreactive chemical groups are without effect on the cofactor properties of phosphatidylcholine. Photoirradiation of reconstituted complexes between phospholipid containing azidoaryllecithin and apoBDH leads to a covalent binding of some synthetic lecithin molecules on the protein. The labelling, about 3 times higher with PL II than with PL I, suggests that the area of interacting domain of BDH with the hydrophobic moiety of phospholipid is more important at or near the surface of the lipid bilayer than in the inner part. This approach is further demonstration that BDH is an integral protein.

Peroxisome proliferators and peroxisome proliferator activated receptors (PPARs) as regulators of lipid metabolism.

Peroxisome proliferation (PP) in mammalian cells, first described 30 years ago, represents a fascinating field of modern research. Major improvements made in its understanding were obtained through basic advances that have opened up new areas in cell biology, biochemistry and genetics. A decade after the first report on PP, a new metabolic pathway (peroxisomal beta-oxidation) and its inducibility by peroxisome proliferators were discovered. More recently, a new type of nuclear receptor, the peroxisome proliferator-activated receptor (PPAR), has been described. The first PPAR was discovered in 1990. Since then, many other PPARs have been characterized. This original class of nuclear receptors belongs to the superfamily of steroid receptors. With activation of cell signal transduction pathways, the occurrence of PPARs provides, for the first time, a coherent explanation of mechanisms by which PP is triggered. Nevertheless, although many compounds or metabolites are capable of activating PPARs, the natural direct ligands of these receptors have not been, up to now, clearly identified, with, however, the exception of 15-deoxy-12,14-prostaglandin J2 which is the ligand of PPAR gamma 2 while leukotrien LTB4 binds PPAR alpha. At this stage, the hypothesis of some orphan PPARs (ie receptors without known ligand) can not be ruled out. Despite these relatively restrictive aspects, the mechanisms by which activation of PPARs leads to PP become clear; also, coherent hypotheses among which a scenario involving receptor phosphorylation or a heat shock protein (ie HSP 72) can be proposed to explain how PPARs would be activated. The aim of this note is to review recent developments on PPARs, to present members up to now recognized to belong to the PPAR family, their characterization, functions, regulation and mechanisms of activation as well as their involvement in lipid metabolism regulation such as control of beta-oxidation, ketogenesis, fatty acid synthesis and lipoprotein metabolism. As an introducing section, a brief review of the major events between the first report of PP in mammals and the discovery of the first PPAR is given. Another section is devoted to current hypotheses on mechanisms responsible for PPAR activation and PP induction. Rather than an exhaustive presentation of cellular alterations accompanying PP induction, a dynamic overview of the lipid metabolism is provided. By assessing the biological significance of this organellar proliferative process, the reader will be led to conclude that the discovery of PPARs and related gene activation through peroxisome proliferator responsive element (PPRE) makes PP induction one of the most illustrative examples of control that occurs in lipid metabolism.

Biochemical properties of liver peroxisomes from rat, guinea pig and human species and the influence of hormonal status on rat liver acyl-CoA oxidase mRNA content.

Liver peroxisomes from three different species, rat, guinea pig and man, have been purified by ultracentrifugation on a discontinuous Nycodenz gradient. Several biochemical parameters were tested in order to compare the basic peroxisomal properties of liver from rat, a species strongly responsive to peroxisome proliferators, and guinea pig and man, two weakly responsive species. Polypeptide patterns were compared and the bands in guinea pig and man comigrating with the two major bands in rat, catalase at 66 kDa and urate oxidase at 35 kDa, appeared in low amounts. However, other polypeptides are similar throughout these species especially in guinea pig as revealed by cross-immunoreactivity using an anti-rat peroxisomal protein rabbit immune serum. Specific activities of peroxisome acyl-CoA oxidase and microsome omega-lauryl hydroxylase have comparable rates in rat and guinea pig liver, but in human liver the activities are much lower. There is a cross-hybridization between acyl-CoA oxidase mRNA probed by rat liver acyl-CoA oxidase cDNA among the three species at a medium stringency. But interestingly, acyl-CoA oxidase mRNA from guinea pig and man appear to be larger in size. On the other hand, the hormonal status does not seem to have a significant effect on the rat liver acyl-CoA oxidase mRNA level suggesting at most that insulin, corticosterone and estradiol have no direct effect on acyl-CoA oxidase gene expression, which contrasts with the well-known effect of peroxisome proliferators.

Possible occurrence for histidyl and cysteyl residues in the catalytic center of rat liver mitochondrial D (-)-beta-hydroxybutyrate dehydrogenase.

1. Rat liver mitochondrial D(-)-beta-hydroxybutyrate dehydrogenase (submitochondrial particles and partially purified preparation) is inhibited by some dicarboxylates, especially by malonate and succinate. The inhibition is reversible and competitive with beta-hydroxybutyrate while uncompetitive with acetoacetate, NAD and NADH: the inhibition is maximal at pH 6 and decrease with increasing pH. 2. Diethylpyrocarbonate (which reacts preferentially with histidyl residues at pH 6.6) inactivates the dehydrogenase at pH 6.1, beta-hydroxybutyrate protects against inactivation, this inactivation being almost completely released by hydroxylamine. The diethylpyrocarbonate-treated enzyme shows an absorbance increase at 242 nm which is characterisitic of reaction between diethylpyrocarbonate and histidyl residue. 3. The optimum pH of the enzyme for beta-hydroxybutyrate oxidation is around 8.2, while for acetoacetate reduction, the optimum pH is around 7. 4. All these results favour the existence of a histidyl residue in the catalytic center and taking into account previous results concerning the effect of thiol reagents on the same enzyme and especially, the protective effect of NAD+ and NADH against these reagents [11] we discuss the possible occurrence of, at least, one histidyl and one cysteyl residue on the catalytic center.

Alkylation at the active site of the D-3-hydroxybutyrate dehydrogenase (BDH), a membrane phospholipid-dependent enzyme, by 3-chloroacetyl pyridine adenine dinucleotide (3-CAPAD).

The structure of the rat liver's D-3-hydroxybutyrate dehydrogenase (BDH) active site has been investigated using an affinity alkylating reagent, the 3-chloroacetyl pyridine adenine dinucleotide (3-CAPAD). This NAD+ analogue reagent strongly inactivates the enzyme following a concentration- and time-dependent process with a stoichiometry of approximately 1. The reagent reacts at the coenzyme binding site as revealed by the efficient protection by NADH. The effect of 3-CAPAD is stronger with the enzyme into its natural membrane environment than with the lipid-free purified apoBDH or with the reconstituted apoBDH-mitochondrial phospholipid complex. The pH-dependent effect on the inactivation process is in agreement with the participation of protons in the catalytic mechanism of BDH. Furthermore, this study exhibits the phospholipid activating role in BDH catalytic activation.

Variations of specific mRNA and polypeptide contents of rat liver D-beta-hydroxybutyrate dehydrogenase during an experimental diabetes mellitus.

The expression of the rat liver D-beta-hydroxybutyrate dehydrogenase (BDH) gene was investigated at different levels: the level of its specific mRNA, the protein content and the enzymatic activity. By using a cDNA probe, we found that the BDH mRNA was about 2 kb and we report here that the decrease of BDH activity in diabetic rats is due to a reduction in the content of the enzyme, which is proportional to a diminution in the amount of the BDH mRNA. We also show that insulin is able to reverse this diabetes effect by restoring the level of BDH mRNA, the BDH content and thus its activity. This result indicates that in vivo the control of the expression of the BDH gene by insulin is mainly transcriptional and/or post-transcriptional (mRNA stability).

The essential cationic charge of phospholipid polar head in the reactivation of D-beta-hydroxybutyrate apodehydrogenase revealed by cationic surfactants.

Attempts to reactivate purified D-beta-hydroxybutyrate apodehydrogenase, a lecithin-requiring enzyme, have been carried out using neutral, anionic, cationic and zwitterionic surfactants. Cationic and zwitterionic compounds exclusively are able to partially replace phosphatidylcholine, the reactivating phospholipid. The extent of reactivation depends on the steric hindrance of the polar head and on the hydrophobic tail length. A molecule bearing a positive charge and an aliphatic chain is the sole structure absolutely required for activity. However the presence of a negative charge is important for enzyme binding to amphiphilic structures and for the efficiency of reactivation.

Lipid-protein interactions in biomembranes studied through the phospholipid specificity of D-beta-hydroxybutyrate dehydrogenase.

Since the biological membranes are fundamental units in the living cells, the studies of lipid-protein interactions are crucial for the understanding of their structure, functions and properties. Beside hydrophobic interactions between fatty acids chain of phospholipids and intrinsic membrane proteins, the interactions between charged groups of the protein with the polar heads of phospholipids generally confer the specificity which may be absolute or preferential. This paper reports essential results obtained these last few years with D-beta-hydroxybutyrate dehydrogenase (BDH) from inner mitochondrial membrane, one of the most interesting and best documented examples of a lipid-requiring enzyme. This is a review of the molecular basis--knowledge and strategy of study--of the lipid specificity for membrane protein functions.

Ciprofibrate stimulates protein kinase C-dependent phosphorylation of an 85 kDa protein in rat Fao hepatic derived cells.

The effect of ciprofibrate on early events of signal transduction was previously studied in Fao cells. Protein kinase C (PKC) assays performed on permeabilized cells showed a more than two-fold increase in PKC activity in cells treated for 24 h with 500 microM ciprofibrate. To show the subsequent effect of this increase on protein phosphorylation, the in vitro phosphorylation on particulate fractions obtained from Fao cells was studied. Among several modifications, the phosphorylation of protein(s) with an apparent molecular mass of 85 kDa was investigated. This modification appeared in the first 24 h of treatment with 500 microM ciprofibrate. It was shown to occur on Ser/Thr residue(s). It was calcium but not calmodulin-dependent. The phosphorylation level of this/these protein(s) was reduced with kinase inhibitors and especially with 300 nM GF-109203X, a specific inhibitor of PKC. All these results suggest that the phosphorylation of the 85 kDa protein(s) is due to a PKC or to another Ser/Thr kinase activated via a PKC pathway. A possible biochemical candidate for 85 kDa protein seems to be the beta isoform of phosphatidylinositol 3-kinase regulatory subunit.

Variations of D (-)-beta-hydroxybutyrate dehydrogenase activity of rat liver mitochondria during post-natal development.

1. D(-)-beta-hydroxybutyrate dehydrogenase specific activity of rat liver mitochondria changes during ontogenesis: at birth, the activity is low, then increases to a maximum at 12 days, decreases until 50 days to keep constant thereafter. At the same time, mitochondrial protein amount increases regularly while succinatecytochrome c reductase specific activity slightly increases after birth to keep constant afterwards. 2. The observed changes in activity of D(-)-beta-hydroxybutyrate dehydrogenase are not related to possible interactions between the enzyme and phospholids since addition of lecithin to mitochondria does not change the activity. 3. Electrophoresis of mitochondrial proteins isolated from rats at different development stages demonstrates the presence of a protein band characterized by the same electrophoretic mobility as beta-hydroxybutyrate dehydrogenase and by significative changes of its proportion during maturation: the relative amount of this protein increases from the new-born to the 10-12 days old rat, to decrease afterwards. 4. These findings may signify that the increased activity of the enzyme with a maximum at 10-12 days followed by a decrease is related to the rate of the enzymes biosynthesis.

Preparation of a dansylated fibrate, a new fluorescent tool to study peroxisome proliferation. Effect on hepatic-derived cell lines.

The synthesis of a dansylated fibrate (DNS-X) has been performed in order to identify the cellular affinity sites of peroxisome proliferators and to establish the subcellular localization of such molecules. DNS-X has been obtained by coupling the dansy1 chloride with the amine resulting from the bezafibrate alkaline hydrolysis. The purified DNS-X has been further characterized by spectrum analysis (UV-Vis, fluorescence, [1H]/[13C]-NMR and mass). At 250 microM and incubated for 48 h with the rat hepatic derived cells (Fao cells), DNS-X stimulates 12-fold the palmitoyl-CoA oxidase, a peroxisome proliferation marker enzyme. This increase is comparable to the one obtained with well known peroxisome proliferators such as bezafibrate or ciprofibrate. The stimulation by DNS-X is specific for the overall molecule since neither the dansyl chloride, the amine, nor the precursors of DNS-X are active. The increase of palmitoyl-CoA oxidase activity is correlated with the increase of the enzyme amount as shown by immunoblotting. In agreement with the species-specificity of the fibrate neither DNS-X, bezafibrate nor ciprofibrate significantly increase palmitoyl-CoA oxidase activity and the enzyme amount in human hepatic-derived cells, HepG2. This work shows that the dansylated fibrate is a new fluorescent tool to study the subcellular localization and identification of high affinity binding sites, then further on, to elucidate the peroxisome proliferation mechanism and the action of hypolipidaemic agents of the fibrate family.

Loss of response of carnitine palmitoyltransferase I to okadaic acid in transformed hepatic cells.

The specific activity of carnitine palmitoyltransferase I (CPT-I) was similar in mitochondria isolated from rat Fao and human HepG2 hepatoma cells and from rat hepatocytes, but almost twofold higher in permeabilized hepatoma cells than in permeabilized hepatocytes. Short-term exposure to okadaic acid induced a ca. 80% stimulation of CPT-I in hepatocytes, whereas no significant response of the enzyme from hepatoma cells was evident. Thus, the high CPT-I activity displayed by hepatoma cells may be reached by hepatocytes upon challenge to okadaic acid. Reconstitution experiments with purified mitochondrial and cytoskeletal fractions showed that the cytoskeleton of hepatocytes produced a more remarkable inhibition of CPT-I than the cytoskeleton of Fao cells. The present data may be explained by a disruption of interactions between CPT-I and cytoskeletal components in tumor cells that may be involved in the okadaic acid-induced activation of hepatic CPT-I as previously suggested.

Transcriptional and post-transcriptional analysis of peroxisomal protein encoding genes from rat treated with an hypolipemic agent, ciprofibrate. Effect of an intermittent treatment and influence of obesity.

The treatment of rats with ciprofibrate, a potent peroxisome proliferator, led to increased levels of the peroxisomal acyl-CoA oxidase (ACO) mRNA. How ciprofibrate functions to elevate ACO mRNA is not known. To help determine the mechanism of ciprofibrate action, in vitro transcription assays were performed. It was determined that ciprofibrate was responsible for a 3.5-fold stimulation of the rate of ACO transcription within 24 hr of ingestion. It was also observed that the transcription rate stimulation following a 2-week ciprofibrate treatment of Wistar rats was maintained following 4 weeks of ciprofibrate withdrawal. Re-introduction of the drug after the 4-week pause resulted in greater stimulation than was initially observed. The results demonstrate that the effect of ciprofibrate is rapid and persists at least twice as long as the initial treatment period. In Zucker rats, both lean and obese, ACO mRNA levels were examined following 2 weeks of ciprofibrate treatment (1 or 3 mg/kg body weight/day). The presence of increased blood levels of triglycerides did not increase ciprofibrate action on transcription, although basal levels of transcription of peroxisomal enzymes were higher in obese rats. The increase in the ACO mRNA level was greater than the transcription rate stimulation suggesting a post-transcriptional regulation.

Delayed effects of ciprofibrate on rat liver peroxisomal properties and proto-oncogene expression.

Peroxisome proliferators (PPs) are non-genotoxic carcinogens in rodents. Their reversible effects on rat liver have been studied with ciprofibrate and fenofibrate. We found that with the hypolipemic drug fenofibrate a pause of 28 days is sufficient for a return to normal status, whereas with the highly potent PP ciprofibrate, the stimulation of ACO mRNA levels remains after its withdrawal. We investigated the effects of the renewal of the treatment with PPs on other peroxisomal parameters and proto-oncogene expression using Wistar rats. Interestingly, c-myc expression was enhanced even upon drug withdrawal, and was more stimulated by the second exposure to ciprofibrate, while c-fos expression was unaltered. However, only slight differences in c-Ha-ras expression were observed. Therefore, the effects of PPs in the Wistar rats are not totally reversible within 28 days following withdrawal, depending on the drug used. These delayed effects of ciprofibrate could be a key to our understanding the hepatocarcinogenic effect of PPs in rodents.