Immunoassay of serum polypeptide hormones by using 125I-labelled anti(-immunoglobulin G) antibodies.

1. A technique for indirectly labelling antibodies to polypeptide hormones, by combining them with radioactively labelled anti-(immunoglobulin G) is described. (a) 125I-labelled anti-(rabbit immunoglobulin G) and anti-(guinea-pig immunoglobulin G) antibodies with high specific radioactivity were prepared after purification of the antibodies on immunoadsorbents containing the respective antigens. (b) Rabbit immunoglobulin G antibodies to human growth hormone, porcine glucagon and guinea-pig immunoglobulin G antibodies to bovine insulin and bovine parathyroid hormone were combined with immunoadsorbents containing the respective polypeptide hormone antigen. (c) The immunoglobulin G antibodies to the polypeptide hormones were reacted with 125-I-labelled anti-(immunoglobulin G) antibodies directed against the appropriate species of immunoglobulin G,and the anti-hormone antibodies were combined with the hormone-containing immunoadsorbent. (d) 125I-labelled anti-(immunoglobulin G) antibodies and anti-hormone antibodies were simultaneously eluted from the hormone-containing immunoadsorbent by dilute HCl, pH 2.0. After elution the anti-(immunoglobulin G) antibodies and antihormone antibodies were allowed to recombine at pH 8.0 and 4 degrees C. 2. The resultant immunoglobulin G-anti-immunoglobulin G complex was used in immunoradiometric (labelled antibody) and two-site assays of the respective polypeptide hormone. 3. By using these immunoassays, concentrations down to 90pg of human growth hormone/ml, 100 pg of bovine insulin/ml, 80 pg of bovine parathyroid hormone/ml and 150 pg of glucagon/ml were readily detected. Assays of human plasma for growth hormone and insulin by these methods showed good agreement with results obtained by using a directly 125I-labelled anti-hormone antibody in an immunoradiometric assay of human growth hormone or by radioimmunoassay of human insulin. 4. The method described allows immunoradiometric or two-site assays to be performed starting with as little as 450 ng of polypeptide hormone-antibody protein. An additional advantage of the method is that a single iodination of the readily available antibodies to immunoglobulin G allows the establishemnt of several polypeptide hormone assays     

Mice immunized to insulin develop antibody to the insulin receptor

We immunized mice with insulin and found that those strains that develop insulin antibodies subsequently produce insulin-like activity in amount equivalent to 300–400 ng insulin per ml serum. The activity was due exclusively to IgG2 antibodies. Bioactivity could be blocked efficiently by insulin antibodies from guinea pigs and from mice. The active IgG2 also displaced labeled insulin from fat cells. Preliminary in vivo studies have indicated that the appearance of insulin-like antibodies in the mouse resulted in abnormal glucose homeostasis and “down regulation” of insulin receptors. These results indicate that immunization to insulin can initiate an idiotype-anti-idiotype network resulting in antibodies to the hormone receptor.     

Effects of insulin on inositol phosphate production in cultured rat hepatocytes

Addition of vasopressin (100 nM) to rat hepatocytes prelabelled with [3H]inositol stimulated the production of inositol phosphates in the presence of 20 mM Li+. Preincubation of hepatocytes with insulin (50 nM) or glucagon (10 nM) had no significant effect alone but enhanced the effects of vasopressin after a lag period of at least 1 min. The effects of insulin and glucagon appeared additive in this respect. Insulin also enhanced the norepinephrine-mediated stimulation of inositol phosphate accumulation. The enhancement by insulin of the effects of vasopressin required at least 0.5-5 nM insulin and did not involve changes in [3H]inositol lipid labelling or IP3 phosphatase activity. The effect of insulin appeared insensitive to prior treatment of hepatocytes with pertussis toxin (200 ng/ml for 18-24 h) or cholera toxin (100 ng/ml for 3-4 h). The glucagon enhancement of the effects of vasopressin was not affected by pertussis toxin but was mimicked by cholera toxin. The response of hepatocytes to vasopressin in the absence of Li+ was smaller and more transient. Under these conditions a 5 min prior incubation with insulin inhibited the stimulation by vasopressin of inositol phosphate accumulation. A similar inhibitory effect of prior insulin exposure on the transient activation by vasopressin of exogenous phosphatidylinositol 4,5-bisphosphate breakdown by hepatocyte homogenates was also seen. These data indicate that insulin, although having no effect on basal inositol phosphate accumulation, can either enhance or antagonise the effects of vasopressin in primary rat liver hepatocyte cultures depending on the experimental conditions.     

Expression, purification, and characterization of C-terminal amidated glucagon in Streptomyces lividans

Glucagon, a peptide hormone produced by alphacells of Langerhans islets, is a physiological antagonist of insulin and stimulator of its secretion. In order to improve its bioactivity, we modified its structure at the C-terminus by amidation catalyzed by a recombinant amidase in bacterial cells. The human gene coding for glucagon-gly was PCR amplified using three overlapping primers and cloned together with a rat alpha-amidase gene in plasmid pMGA. Both genes were expressed under control of the strong constitutive promoter of aph and secretion signal melC1 in Streptomyces lividans. With Phenyl-Sepharose 6 FF, Q-Sepharose FF, SP-Sepharose FF chromatographies and HPLC, the peptide was purified to about 93.4% purity. The molecular mass of the peptide is 3.494 kDa as analyzed by MALDI TOF, which agrees with the theoretical mass value of the C-terminal amidated glucagon. The N-terminal sequence of the peptide was also determined, confirming its identity with human glucagon at the Nterminal part. ELISA showed that the purified peptide amide is bioactive in reacting with glucagon antibodies.     

Complement-mediated unspecific binding of immunoglobulins to some endocrine cells.

Unspecific binding of immunoglobulins to gastrin G cells, glucagon A cells and somatostatin D cells of the gastric mucosa or pancreas, as well as to the calcitonin-somatostatin cells of rabbit thyroid has been found to occur through a non antigen-antibody mechanism mediated at least in part by the C1q fraction of complement. The phenomenon represents a major drawback in hormone immunohistochemistry, which can be prevented by incubating the specific anti-hormone sera with anti-C1q antibodies or with complement-fixing immunocomplexes.     

Insulin stimulates tyrosine phosphorylation of its receptor beta-subunit in intact rat hepatocytes.

We studied the phosphorylation of the beta subunit of the insulin receptor in intact freshly isolated rat hepatocytes, labelled with [32P]Pi. Insulin receptors partially purified by wheat-germ agglutinin chromatography were immunoprecipitated with either antibodies to insulin receptor or antibodies to phosphotyrosine. Receptors derived from cells incubated in the absence of insulin contained only phosphoserine. Addition of insulin to hepatocytes led to a dose-dependent increase in receptor beta-subunit phosphorylation, with half-maximal stimulation being observed at 2 nM-insulin. Incubation of cells with 100 nM-insulin showed that, within 1 min of exposure to the hormone, maximal receptor phosphorylation occurred, which was followed by a slight decrease and then a plateau. This insulin-induced stimulation of its receptor phosphorylation was largely accounted for by phosphorylation on tyrosine residues. Sequential immunoprecipitation of receptor with anti-phosphotyrosine antibodies and with anti-receptor antibodies, and phosphoamino acid analysis of the immunoprecipitated receptors, revealed that receptors that failed to undergo tyrosine phosphorylation were phosphorylated on serine residues. The demonstration of a functional hormone-sensitive insulin-receptor kinase in normal cells strongly supports a role for this receptor enzymic activity in mediating biological effects of insulin.     

Lack of vasopressin receptors in liver, but not in kidney, of ob/ob mice.

The activity of phosphorylase a was measured in isolated hepatocytes from fed lean and ob/ob mice after addition of vasopressin, angiotensin, phenylephrine and glucagon. The binding of these hormones to purified liver plasma membranes was also determined. In hepatocytes of ob/ob mice, no increase in phosphorylase a was measured after addition of vasopressin, whereas the other hormones promoted an increase in the activity of the enzyme. No specific vasopressin receptors could be measured on purified liver plasma membrane of ob/ob mice. A decrease in the number of receptors for angiotensin and glucagon, without modification of the affinity, was also observed. No restoration of the number of vasopressin receptors was observed in liver of ob/ob mice starved for 3 days or in younger (5-6 weeks) animals. Vasopressin receptors and vasopressin-stimulated adenylate cyclase, measured on purified kidney medulla membranes, were similar in both lean and ob/ob mice. The data indicate a selective lack of vasopressin receptors and metabolic response in liver of the ob/ob mouse.     

Complement-mediated unspecific binding of immunoglobulins to some endocrine cells

Summary Unspecific binding of immunoglobulins to gastrin G cells, glucagon A cells and somatostatin D cells of the gastric mucosa or pancreas, as well as to the calcitonin-somatostatin cells of rabbit thyroid has been found to occur through a non antigen-antibody mechanism mediated at least in part by the C_1q fraction of complement. The phenomenon represents a major drawback in hormone immunohistochemistry, which can be prevented by incubating the specific anti-hormone sera with anti-C_1q antibodies or with complement-fixing immunocomplexes.     

Pancreatic hormones are expressed on the surfaces of human and rat islet cells through exocytotic sites

Human and rat insulin cells show insulin immunoreactivity, and glucagon cells show glucagon immunoreactivity on their membrane surfaces, respectively. The reaction occurs in the form of small dots on the islet cell surface and colocalizes with the chromogranin family of secretory granule markers. Electron microscopy reveals the labeling to occur at sites of exocytotic granule release, involving the surfaces of extruded granule cores. The surfaces of islet cells were labeled both by polyclonal and monoclonal antibodies, excluding that receptor-interacting, anti-idiotypic hormone antibodies were responsible for the staining. Human insulin cells were surface-labeled by monoclonal antibodies recognizing the mature secretory products, insulin and C-peptide but not with monoclonal antibodies specific for proinsulin. Thus, routing of unprocessed preproinsulin to the cell surface may not account for these results. It is concluded that the staining reflects interactions between the appropriate antibodies and exocytotic sites of hormone release.     

Interaction of insulin analogs, glucagon, growth hormone, vasopressin, oxytocin, and scrambled forms of ribonuclease and lysozyme with glytathione-insulin transhydrogenase (thiol: protein-disulfide oxidoreductase): dependence upon conformation.

Interactions of several proteins with glutathione-insulin transhydrogenase (GIT) have been investigated by determining their ability to inhibit degradation of 125I-labeled insulin catalyzed by GIT. The inhibition by every insulin analog (des-Asn-des-Ala-pork insulin, desoctapeptide-pork insulin, des-Ala-pork insulin, pork insulin, proinsulin, and guinea pig insulin) was competitive vs. competitive vs. insulin indicating that they function as alternate substrates. The insulin analogs with the least hormonal activity showed the highest potency as inhigitors of insulin degradation. Whereas native ribonuclease and lysozyme showed little or no inhibition, their scrambled forms (i.e. reduced and randomly reoxidized) showed competitive inhibition with a potency greater than that of insulin. These results suggest that the conformation of the substrate or inhibitor is probably the major factor in determining the specificity for (or binding to) the enzyme. Studies withother peptide hormones showed competitive inhibition with vasopressin and oxytocin and noncompetitive inhibition with glycagon. The inhibition with growth hormone could be either competitive or noncompetitive. The inhibition by glucagon and growth hormone (physiologic antagonists of insulin) could serve as a control mechanism to modulate the activity of enzyme. The following showed very little or no inhibition; the native and scrambled form of pepsinogen, trypsin inhibitor of beef pancreas and of lima bean, C-peptide of pork proinsulin, and heptapeptide (B23-B29) of insulin.     

Mitochondrial function after acute alteration of the endogenous insulin-to-glucagon ratio

Mannoheptulose (2g/kg i.p.) increases serum glucagon and decreases serum insulin via its effect on pancreatic islet cells. These changes in endogenous hormone status had effects on rat liver mitochondria that were comparable to the effects of injecting porcine glucagon (0.5 mg/kg i.p.). Mitochondrial adenine nucleotide content was increased 38 or 39% by mannoheptulose or glucagon respectively, citrulline synthesis by 165 or 193%, pyruvate carboxylation by 113 or 135%, coupled respiration by 34 or 42%, and uncoupled respiration by 40 or 54%. We conclude that the reciprocal changes in endogenous insulin and glucagon brought about by mannoheptulose offer a useful and interesting alternative to glucagon injection for studying the effects of these pancreatic hormones on liver mitochondria.     

Immunoprecipitation of the lutropin receptor. Loss of receptor molecules during down-regulation.

Specific anti-(lutropin receptor) antibodies were produced by immunizing rabbits with lutropin receptor purified from pseudopregnant rat ovary. The anti-receptor serum at 1:100 dilution together with anti-(rabbit gamma-globulin) serum immunoprecipitated 70% of 3H-labelled, purified lutropin receptor and 42% of 125I-chorio-gonadotropin-receptor complex. The antiserum inhibited hormone binding to rat ovarian particles. Pseudopregnant rat ovarian particles were labelled with periodate/NaB3H4 and solubilized with Triton X-100. The Triton X-100 extract was subjected to immunoprecipitation using the anti-receptor serum. When the immunoprecipitate was dissolved and analysed by polyacrylamide-gel electrophoresis in sodium dodecyl sulphate under reducing conditions followed by fluorography, a receptor polypeptide with an apparent Mr 95000 was detected. A receptor down-regulating dose of choriogonadotropin was injected into pseudopregnant rats and their ovaries were removed and homogenized 4 days later, and analysed for immunoprecipitable receptors as above. No receptor molecules were found. Accordingly, the lutropin receptor molecules actually disappear rather than merely become masked from hormone during homologous down-regulation.     

Circadian oscillations of thyroid hormones, insulin and glucagon in the blood of laboratory rats in the course of the year

The concentration of thyroid hormones and insulin in the serum that of glucagon in the plasma and glucose in the blood was determined at 3-hour intervals, in the course of one day, at different times of the year, in adult male rats (Wistar strain) of a conventional breed kept under standard conditions with a 12:12 h light:dark regimen. The lowest thyroid hormone and glucagon concentrations and the highest insulin and blood glucose levels were found in the winter. In various seasons, circadian oscillations of the thyroid hormones culminated in the dark part of the day and that of glucagon in the light part, with the exception of the autumn. Circadian oscillation of insulin levels culminated at different times of day during the year. The pronounced changes found in the examined hormones in the laboratory rat at various times of the year are evidently the outcome of adaptation to changes in external environmental conditions during phylogenesis. In the polarity of changes in these indicators between the winter and the summer or the spring, the laboratory rat bears the closest resemblance to wild mammals.     

Circadian Dependent Effect of Epidermal Growth Factor,Insulin and Glucagon on Hepatic Pyruvate Kinase and Malic Enzyme of Mice

The influence of epidermal growth factor (EGF), 0.75 μg g^?1; insulin, 1.5 μg g^?1; glucagon, 1.25 ygg^?1 and their combinations on the activities of hepatic pyruvate kinase (PK) and malic enzymes (ME) was monitored. Male CD2F₁ mice were treated toward the end of the light or dark periods, 9 or 23 /tours after /ights on (9 or 23 HALO), and subgroups of six mice were killed at 4,8 or 12 hr post-treatment. PK and ME activities from control mice were well characterized by cosine curves. The PK activity was maximal when ME activity was minimal at the transition from light to dark (9 HALO plus 4 hr) and PK was at a minimum when ME was highest (23 HALO plus 4 hr). Both enzymes were influenced by at least one peptide hormone, and the effects were strongly circadian -stage dependent. The only effect attributed to EGF was an increase of PK activity (23%) 12 hr after injection at 23 HALO. PK activity was increased by insulin (23%) at 23 HALO (4 hr after injection), but not at 9 HALO, and decreased (17%) by glucagon 12 hr after injection at 9 HALO. Several reductions in PK activity in response to various combinations of peptides were observed, and appeared to be caused by glucagon but influenced by insulin. The activity of ME was decreased (33%) in response to insulin 4 hr after injection at 23 HALO but not at 9 HALO and increased (60-70%) by glucagon alone or in combinations with insulin or EGF, or both, at 4 hr after injection at 9 HALO but not at 23 HALO. In general, when ME activity was altered by either insulin or glucagon, PK activity was also altered in the opposite direction, and the effects of glucagon were opposed by insulin.     

Metabolic actions of vasopressin,glucagon and adrenalin in the intact rat

Metabolic effects of vasopressin, glucagon and adrenalin were compared, in intact rats, especially in regard to time courses of effects.Hyperglycaemia was transient in response to vasopressin, prolonged following adrenalin, and, surprisingly, was not discernible after glucagon, except in response to a very large dose. Vasopressin decreased and adrenalin increased, the plasma free fatty acid concentration; both hormones decreased the triacylglycerol level. Muscle glycogen concentrations, measured in heart, diaphragm and skeletal muscle, exhibited small changes, with complex time courses, following hormone administration. Vasopressin brought about a rapid but transient activation of hepatic glycogen phosphorylase which resembled that due to adrenalin. The activation by glucagon of phosphorylase was greater and more prolonged, despite the absence of hyperglycaemia. In response to vasopressin, there was an increase in plasma insulin. Incorporation of ¹⁴C from [¹⁴C] glucose into glycogen or fatty acids was not influenced by vasopressin. Taken together, these results may be explained by rapid metabolic action of vasopressin on hepatic glycogenolysis, whereas adrenalin has multiple prolonged actions.     

Circadian Dependent Effect of Epidermal Growth Factor, Insulin and Glucagon on Hepatic Pyruvate Kinase and Malic Enzyme of Mice

The influence of epidermal growth factor (EGF), 0.75 μg g^-1; insulin, 1.5 μg g^-1; glucagon, 1.25 ygg^-1 and their combinations on the activities of hepatic pyruvate kinase (PK) and malic enzymes (ME) was monitored. Male CD2F₁ mice were treated toward the end of the light or dark periods, 9 or 23 /tours after /ights on (9 or 23 HALO), and subgroups of six mice were killed at 4,8 or 12 hr post-treatment. PK and ME activities from control mice were well characterized by cosine curves. The PK activity was maximal when ME activity was minimal at the transition from light to dark (9 HALO plus 4 hr) and PK was at a minimum when ME was highest (23 HALO plus 4 hr). Both enzymes were influenced by at least one peptide hormone, and the effects were strongly circadian -stage dependent. The only effect attributed to EGF was an increase of PK activity (23%) 12 hr after injection at 23 HALO. PK activity was increased by insulin (23%) at 23 HALO (4 hr after injection), but not at 9 HALO, and decreased (17%) by glucagon 12 hr after injection at 9 HALO. Several reductions in PK activity in response to various combinations of peptides were observed, and appeared to be caused by glucagon but influenced by insulin. The activity of ME was decreased (33%) in response to insulin 4 hr after injection at 23 HALO but not at 9 HALO and increased (60-70%) by glucagon alone or in combinations with insulin or EGF, or both, at 4 hr after injection at 9 HALO but not at 23 HALO. In general, when ME activity was altered by either insulin or glucagon, PK activity was also altered in the opposite direction, and the effects of glucagon were opposed by insulin.     

Insulin and glucagon stimulation of (Na+-K+)-ATPase transport activity in isolated rat hepatocytes

The effects of insulin and glucagon on the (Na+-K+)-ATPase transport activity in freshly isolated rat hepatocytes were investigated by measuring the ouabain-sensitive, active uptake of 86Rb+. The active uptake of 86Rb+ was increased by 18% (p less than 0.05) in the presence of 100 nM insulin, and by 28% (p less than 0.005) in the presence of nM glucagon. These effects were detected as early as 2 min after hepatocyte exposure to either hormone. Half-maximal stimulation was observed with about 0.5 nm insulin and 0.3 nM glucagon. The stimulation of 86Rb+ uptake by insulin occurred in direct proportion to the steady state occupancy of a high affinity receptor by the hormone (the predominant insulin-binding species in hepatocytes at 37 degrees C. For glucagon, half-maximal response was obtained with about 5% of the total receptors occupied by the hormone. Amiloride (a specific inhibitor of Na+ influx) abolished the insulin stimulation of 86Rb+ uptake while inhibiting that of glucagon only partially. Accordingly, insulin was found to rapidly enhance the initial rate of 22Na+ uptake, whereas glucagon had no detectable effect on 22Na+ influx. These results indicate that monovalent cation transport is influenced by insulin and glucagon in isolated rat hepatocytes. In contrast to glucagon, which appears to enhance 86Rb+ influx through the (Na+-K+)-ATPase without affecting Na+ influx, insulin stimulates Na+ entry which in turn may increase the pump activity by increasing the availability of Na+ ions to internal Na+ transport sites of the (Na+-K+)-ATPase.     

Synergistic stimulation of the Ca2+ influx in rat hepatocytes by glucagon and the Ca2+-linked hormones vasopressin and angiotensin II

Glucagon was added to isolated rat hepatocytes, either alone or together with vasopressin or angiotensin II, and the effects on the initial 45Ca2+ uptake rate were investigated. Addition of glucagon alone which increased cyclic AMP content of the cells slightly increased the initial 45Ca2+ uptake rate. When glucagon was added together with vasopressin or angiotensin II--both of which when added separately increase the initial 45Ca2+ uptake rate but did not affect the cellular content of cyclic AMP--the measured initial 45Ca2+ uptake rate was larger than the sum of that seen with each hormone alone. This indicates that glucagon and Ca2+-linked hormones synergistically enhanced the Ca2+ influx in rat hepatocytes. These effects of glucagon can be mimicked by dibutyryl cyclic AMP or forskolin, suggesting that cyclic AMP augments both the resting Ca2+ and the vasopressin- or angiotensin II-stimulated influx. Measurement of the initial 45Ca2+ uptake rate as a function of the extracellular Ca2+ concentration indicated that the increase in the Ca2+ influx resulting from single or combined glucagon and vasopressin administration occurred through a homogeneous population of Ca2+ gates. These hormones were found to raise both the apparent Km for external Ca2+ and the apparent Vmax of the Ca2+ influx. The maximal increase in these two parameters was observed when the two hormones were added together. This suggests that glucagon and vasopressin synergistically stimulate the same Ca2+ gating mechanism. The dose-response curves for the action of glucagon or vasopressin applied in the presence of increasing concentrations of vasopressin or glucagon, respectively, showed that each hormone increases the maximal response to the other without affecting its ED50. It is proposed that glucagon and the Ca2+-linked hormones control the cellular concentration of two intermediates which are both necessary to allow Ca2+ entry into the cells.     

Metabolic actions of vasopressin, glucagon and adrenalin in the intact rat.

Metabolic effects of vasopressin, glucagan and adrenalin were compared, in intact rats, especially in regard to time courses of effects. Hyperglycaemia was transient in response to vasopressin, prolonged following adrenalin, and, suprisingly, was not discernible after glucagon, except in response to a very large dose. Vasopressin decreased and adrenalin increased, the plasma free fatty acid concentration; both hormones decreased the triacylglycerol level. Muscle glycogen concentrations, measured in heart, diaphragm and skeletal muscle, exhibited small changes, with complex time courses, following hormone administration. Vasopressin brought about a rapid but transient activation of heaptic glycogen phosphorylase which resembled that due to adrenalin. The activation by glucagon of phosphorylase was greater and more prolonged, despite the absence of hyperglycaemia. In response to vasopressin, there was in increase in plasma insulin. Incorporation of 14C from [14C]glucose into glycogen or fatty acids was not influenced by vasopressin. Taken together, these results may be explained by rapid metabolic action of vasopressin on hepatic glycogenolysis, whereas adrenalin has multiple prolonged actions.     

Comparison of insulin and glucagon pulsatile secretion between the rat and dog pancreas in-vitro

Sustained pulses of insulin and glucagon were obtained from the isolated perfused in vitro rat pancreas. The respective periodicity of hormone release (peak to peak interval) was calculated by the Pulsar computer algorithm as insulin 5.8 +/- 0.3 min and glucagon 6.5 +/- 0.25 min. Because pulsatile insulin secretion is absent in type II diabetics, pulsatile islet hormone secretion could theoretically be regulated directly by intra-islet hormone interactions or indirectly by hormone sensitive nerve feedback, possibly from a venous hormone sensitive receptor system within the pancreas. To test the possible contributions of these systems in pulse regulation, the direction of perfusion was reversed in both rat and dog pancreata to prevent hormone contact with putative venous hormone receptors. The periodicity of hormone secretion was unchanged by reversed perfusion in both species. As vascular perfusion of islet cells is normally B to A to D, these results suggest that neither intra-islet hormone interactions nor intra-pancreatic insulin or glucagon sensitive nerve feedback systems are responsible, on an acute basis, for the regulation of pulsatile insular secretion from the normal pancreas. Insulin regulates net glucagon secretion but does not acutely influence glucagon pulses. The presence of pulses during retrograde perfusion may be the result of the entrainment of the pacemaker-islet system. These observations are consistent with the presence of an independent pacemaker and neural coordinating system within the dog and rat pancreas which may influence both the A- and B-cell.