Purification of D-myo-inositol 1,4,5-trisphosphate 3-kinase from rat brain

The ATP-dependent, calmodulin-sensitive 3-kinase responsible for the conversion of D-myo-inositol 1,4,5-trisphosphate to D-myo-inositol 1,3,4,5-tetrakisphosphate has been purified 2,700-fold from rat brain to a specific activity of 2.3 mumol/min/mg protein. A method of purification is described involving chromatography on phosphocellulose, Orange A dye ligand, calmodulin agarose, and hydroxylapatite columns. Neither the highly purified enzyme nor enzyme eluting from the phosphocellulose column were activated by Ca2+. However, enzyme in the 100,000 x g supernatant from rat brain was activated by Ca2+ over the range from 10(-7) to 10(-6) M and Ca2+ sensitivity of the purified enzyme was restored by the addition of calmodulin. The enzyme has a catalytic subunit Mr of 53,000 by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Size exclusion chromatography of the purified enzyme on a Superose 12 column gave a Mr value of 70,000, indicating that the purified enzyme was present as a monomer. In contrast, the 100,000 x g supernatant and the purified enzyme after addition of calmodulin and 10(-6) M Ca2+ chromatographed on size exclusion chromatography with a Mr of 150,000-160,000. These results imply that the native enzyme is a dimeric structure of two catalytic subunits plus calmodulin. The purified enzyme showed a Km of 0.21 +/- 0.08 microM for D-myo-inositol 1,4,5-trisphosphate and had a pH optimum of 8.5. Addition of calmodulin increased both the Km and the Vmax of the purified enzyme about 2-fold. The high affinity of the 3-kinase for D-myo-inositol 1,4,5-trisphosphate together with its activation by Ca2+/calmodulin suggests that this enzyme may exert an important regulatory role in inositol phosphate signaling by promoting the formation of additional inositol polyphosphate isomers.     

Partial purification and some properties of rat brain inositol 1,4,5-trisphosphate 3-kinase.

An enzyme which catalyses the ATP-dependent phosphorylation of inositol 1,4,5-trisphosphate [Ins(1,4,5)P3] was purified approx. 180-fold from rat brain cytosol by (NH4)2SO4 precipitation, chromatography through hydroxyapatite, anion-exchange fast protein liquid chromatography and gel-filtration chromatography. Gel filtration on Sepharose 4B CL gives an Mr of 200 x 10(3) for the native enzyme. The inositol tetrakisphosphate (InsP4) produced by the enzyme has the chromatographic, chemical and metabolic properties of Ins(1,3,4,5)P4. Ins(1,4,5)P3 3-kinase displays simple Michaelis-Menten kinetics for both its substrates, having Km values of 460 microM and 0.44 microM for ATP and Ins(1,4,5)P3 respectively. When many of the inositol phosphates known to occur in cells were tested, only Ins(1,4,5)P3 was a substrate for the enzyme; the 2,4,5-trisphosphate was not phosphorylated. Inositol 4,5-bisphosphate and glycerophosphoinositol 4,5-bisphosphate were phosphorylated much more slowly than Ins(1,4,5)P3. CTP, GTP and adenosine 5'-[gamma-thio]triphosphate were unable to substitute for ATP. When assayed under conditions of first-order kinetics, Ins(1,4,5)P3 kinase activity decreased by about 40% as the [Ca2+] was increased over the physiologically relevant range. This effect was insensitive to the presence of calmodulin and appeared to be the result of an increase in the Km of the enzyme for Ins(1,4,5)P3. Preincubation with ATP and the purified catalytic subunit of cyclic AMP-dependent protein kinase did not affect the rate of phosphorylation of Ins(1,4,5)P3 when the enzyme was assayed at saturating concentrations of Ins(1,4,5)P3 or at concentrations close to its Km for this substrate.     

Characterization of D-myo-inositol 1,4,5-trisphosphate phosphatase in rat brain

Rat brain homogenates contain significant amounts of inositol 1,4,5-trisphosphate phosphatase in both 180,000xg (60 min) particulate and supernatant fractions. As other membrane-bound enzymes (e.g. guanylate cyclase), particulate inositol 1,4,5-trisphosphate phosphatase activity is highly sensitive to low concentrations of Triton X-100 (0.03%). Higher concentrations of detergent (1%) partially solubilized the enzyme. Thiol blocking agents (e.g. p-hydroxymercuribenzoate) inactivate inositol 1,4,5-trisphosphate phosphatase activity (an effect reversed with 2-mercaptoethanol). It is thus suggested that enzymatic activity requires the presence of -SH groups.     

Regulation of D-myo-inositol 1,4,5-trisphosphate 3-kinase by cAMP-dependent protein kinase and protein kinase C

The Ca2(+)-mobilizing second messenger D-myo-inositol 1,4,5-trisphosphate (Ins(1,4,5)P3) is converted to the putative messenger D-myo-inositol 1,3,4,5-tetrakisphosphate by Ins(1,4,5)P3 3-kinase. We found that cAMP-dependent protein kinase and protein kinase C phosphorylate, and thereby modulate, the activity of Ins(1,4,5)P3 3-kinase. cAMP-dependent kinase introduced a stoichiometric amount of phosphate at serine 109 of the 53-kDa polypeptide and caused a 1.8-fold increase in Vmax, whereas the protein kinase C-dependent phosphorylation reduced the Vmax to one-fourth of that of the unphosphorylated enzyme. Upon prolonged incubation, protein kinase C introduced phosphate at multiple sites in Ins(1,4,5)P3 3-kinase, and the resulting inactivation of the enzyme appeared to be well-correlated with the simultaneous phosphorylation of two major sites, serine 109 and serine 175. The Km for Ins(1,4,5)P3 was not affected significantly after phosphorylation by either protein kinase. We propose, therefore, that the phosphorylation of Ins(1,4,5)P3 3-kinase by cAMP-dependent kinase and protein kinase C constitutes mechanisms of cross-talk between cellular signaling pathways that use various second messengers such as inositol phosphates, diacylglycerol, Ca2+, and cAMP.     

Purification and characterization of inositol 1,4,5-trisphosphate 3-kinase from pig aortic smooth muscle.

Inositol 1,4,5-trisphosphate (InsP3) 3-kinase, which phosphorylates InsP3 to form inositol 1,3,4,5-tetrakisphosphate, was purified to apparent homogeneity by (NH4)2SO4 fractionation and sequential chromatographic steps on DEAE-sepharose, calmodulin-Affi-Gel and DEAE-5PW h.p.l.c. The purified enzyme had a specific activity of 24.4 nmol of inositol tetrakisphosphate formed/min per mg of protein, which represented a purification of approx. 195-fold with a 0.29% recovery, compared with the cytosol fraction of the muscle. SDS/polyacrylamide-gel electrophoresis showed a single protein-staining band of Mr 93,000. Moreover, the major protein peak, of Mr 84,000, was detected by TSK gel G3000SW gel-permeation chromatography of the purified sample. As this value was approximately consistent with the Mr determined by SDS/polyacrylamide-gel-electrophoretic analysis, the InsP3 3-kinase might be a monomeric enzyme. The purified enzyme had a Km for InsP3 of 0.4 microM, with an optimum pH range of 5.8-7.7. The enzyme was maximally activated by calmodulin, with a stoichiometry of 1:1.     

Syntheses of D-myo-inositol 1,4,5-trisphosphate affinity ligands.

A mixture of 2,3,6-tri-O-benzoyl-4,5-di-O-benzyl-D-myo-inositol and 1,3,6-tri-O-benzoyl-4,5-di-O-benzyl-D-myo-inositol, obtained during our synthesis of D-myo-inositol 1,4,5-trisphosphate [C.E. Ballou and W. Tegge, Proc. Natl. Acad. Sci. U.S.A., 86 (1989) 94-98], was separated after tetrahydropyranylation of the free hydroxyl group in each. 2,3,6-Tri-O-benzoyl-4,5-di-O-benzyl-1-O- (tetrahydro-2-pyranyl)-D-myo-inositol was debenzylated and the two free hydroxyl groups were phosphorylated by a dibenzyl phosphoramidite procedure. The tetrahydropyranyl group was then removed, and phosphorylation at position 1 with benzyl 3-(benzyloxycarbonylamino)propyl di-N-isopropylphosphoramidite, followed by oxidation and deprotection, provided 1-[3-aminopropoxy(hydroxy)phosphinyl]-D-myo-inositol 4,5-bisphosphate. This compound was coupled to activated agarose to prepare an affinity matrix for the isolation of D-myo-inositol 1,4,5-trisphosphate-binding proteins, and it was coupled to 4-azido-2-hydroxybenzoic acid to give a product that was labeled with 125I to prepare a photoactivable derivatizing reagent. The new derivatives retain significant biological activity as assessed by their ability to stimulate the release of stored Ca2+ from the endoplasmic reticulum of permeabilized rat basophilic leukemia cells.     

D-myo-inositol 1,4,5-trisphosphate phosphatase in skeletal muscle.

The presence and subcellular distribution of D-myo-inositol 1,4,5-trisphosphate phosphatase (InsP3ase) in rabbit fast-twitch skeletal muscle were investigated. A specific InsP3ase was found in both sarcotubular-membrane and soluble fractions. Membrane-bound InsP3ase accounted for 60-65% of total activity. The InsP3ase was detected both on the surface membranes and on the InsP3-sensitive intracellular Ca2+ store, i.e. the sarcoplasmic reticulum. The Km for inositol 1,4,5-trisphosphate (InsP3) ranged between 15 and 18 microM, and the highest Vmax. (19.6 nmol of InsP3 hydrolysed/min per mg of protein) was measured in a membrane fraction enriched in transverse tubules. Several known inhibitors of InsP3ase, e.g. 2,3-bisphosphoglycerate, CdCl2 and EDTA, were active on skeletal-muscle InsP3ase. Total InsP3ase activity of both rabbit and frog skeletal muscle was comparable with that of rabbit brain, liver and main pulmonary artery (smooth muscle). The present results are consistent with the hypothesis that InsP3 plays a role in excitation-contraction coupling in skeletal muscle [Volpe, Salviati, Di Virgilio & Pozzan (1985) Nature (London) 316, 347-349].     

C-glycoside based mimics of D-myo-inositol 1,4,5-trisphosphate

Epimeric C-glycoside based polyphosphates, alpha- and beta-D-glucopyranosylmethanol 3,4,1'-trisphosphates (8 and 9) were prepared from D-glucose. The key intermediate, allyl 2,6-di-O-benzyl-alpha-D-glucopyranoside, was prepared in five steps (67% yield) from allyl alpha-D-glucopyranoside without the need for chromatography. Compounds 8 and 9 were shown to be full agonists at the Ins(1,4,5)P3 receptors of permeabilised hepatocytes, but with markedly different potencies. Such C-glycoside analogues are worthy of further development as Ins(1,4,5)P, receptor ligands.     

Ca2+/calmodulin-sensitive inositol 1,4,5-trisphosphate 3-kinase in rat and bovine brain tissues

Inositol 1,4,5-trisphosphate (Ins P3) 3-kinase catalyzes the ATP-dependent phosphorylation of Ins P3 to Inositol 1,3,4,5-tetrakisphosphate (Ins P4). Ca2+/calmodulin (CaM)-sensitivity of Ins P3 3-kinase was measured in the crude soluble fraction from rat brain and different anatomic regions of bovine brain. Kinase activity was inhibited in the presence of EGTA (free Ca2+ below 1 nM) as compared to Ca2+ (10 microM free Ca2+) or Ca2+ (10 microM free Ca2+) and CaM (1 microM). Ca2+-sensitivity was also seen for the cAMP phosphodiesterase measured under the same assay conditions, but was not for the Ins P3 5-phosphatase. DEAE-cellulose chromatography of the soluble fraction of rat brain or bovine cerebellum resolved a Ca2+/CaM-sensitive Ins P3 3-kinase (maximal stimulation at 1 microM Ins P3 substrate level was 2.0-3.0 fold).     

Characterization of D-myo-inositol 1,4,5-trisphosphate phosphatase in rat liver plasma membranes

D-myo-Inositol 1,4,5-trisphosphate has been previously demonstrated to act as a second messenger for the hormonal mobilization of intracellular calcium in rat liver. In this study, the breakdown of D-myo-inositol 1,4,5-trisphosphate by a phosphatase activity was characterized. Using partially purified subcellular fractions, it was found that D-myo-inositol 1,4,5-trisphosphate phosphatase (I-P3ase) specific activity was highest in the plasma membrane fraction, while D-myo-inositol 1,4-bisphosphate phosphatase specific activity was highest in the cytosolic and microsomal fractions. The plasma membrane I-P3ase was Mg2+-dependent with optimal activity observed at 0.5-1.5 mM free Mg2+. The enzyme had a neutral pH optimum, suggesting that it was neither an acid nor alkaline phosphatase. Neither LiCl nor NaF inhibited the I-P3ase activity. However, both L-cysteine and dithiothreitol stimulated the activity 2-fold. Spermine (2.0 mM) inhibited the I-P3ase activity by 50%, while putrescine and spermidine had little or no effect.     

Interaction of synthetic D-6-deoxy-myo-inositol 1,4,5-trisphosphate with the Ca2(+)-releasing D-myo-inositol 1,4,5-trisphosphate receptor, and the metabolic enzymes 5-phosphatase and 3-kinase.

The ability of D-6-deoxy-myo-inositol 1,4,5-trisphosphate [6-deoxy-Ins(1,4,5)P3], a synthetic analogue of the second messenger D-myo-inositol 1,4,5-trisphosphate [Ins(1,4,5)P3], to mobilise intracellular Ca2+ stores in permeabilised SH-SY5Y neuroblastoma cells was investigated. 6-Deoxy-Ins(1,4,5)P3 was a full agonist (EC50 = 6.4 microM), but was some 70-fold less potent than Ins (1,4,5)P3 (EC50 = 0.09 microM), indicating that the 6-hydroxyl group of Ins(1,4,5)P3 is important for receptor binding and stimulation of Ca2+ release, but is not an essential structural feature. 6-Deoxy-Ins(1,4,5)P3 was not a substrate for Ins (1,4,5)P3 5-phosphatase, but inhibited both the hydrolysis of 5-[32P]+ Ins (1,4,5)P3 (Ki 76 microM) and the phosphorylation of [3H]Ins(1,4,5)P3 (apparent Ki 5.7 microM). 6-Deoxy-Ins (1,4,5)P3 mobilized Ca2+ with different kinetics to Ins(1,4,5)P3, indicating that it is probably a substrate for Ins (1,4,5)P3 3-kinase.     

Cell membrane permeable esters of D-myo-inositol 1,4,5-trisphosphate

d-myo-inositol 1,4,5-trisphosphate (Ins(1,4,5)P3, or IP3) is a ubiquitous second messenger that regulates cytosolic Ca2+ activities ([Ca2+]i). To study this signaling branch in intact cells, we have synthesized a caged and cell permeable derivative of IP3, ci-IP3/PM, from myo-inositol in 9 steps. Ci-IP3/PM is a homologue of cm-IP3/PM, a caged and cell permeable IP3 ester developed earlier. In ci-IP3/PM, 2- and 3-hydroxyl groups of myo-inositiol are protected by an isopropylidene group; whereas in cm-IP3/PM, a methoxymethylene is used. Ci-IP3/PM can be loaded into cells non-invasively to high concentrations without activating IP3 receptors (IP3Rs). UV uncaging of loaded ci-IP3 released i-IP3, a potent agonist of IP3Rs, and evoked Ca2+ release from internal stores. Interestingly, elevations of [Ca2+]i by i-IP3 lasted longer than [Ca2+]i transients by m-IP3, the uncaging product of cm-IP3. To understand this difference, we measured the metabolic stability of i-IP3 and m-IP3. Like natural IP3 which is known to be rapidly metabolized in cells, m-IP3 could only be detected within several seconds after uncaging cm-IP3. In contrast, i-IP3 was metabolized at a much slower rate. By exploiting different metabolic rates of m-IP3 and i-IP3, we developed two procedures for activating IP3Rs in cells without UV uncaging. The first method involves photolyzing ci-IP3/PM in vitro to generate i-IP3/PM. Successive additions of low micromolar i-IP3/PM to NIH 3T3 cells caused graded Ca2+ releases, confirming that "quantal Ca2+ release" occurs in fully intact cells with normal ATP supplies and undisrupted endoplasmic reticulum. The second technique utilizes two photon uncaging. After locally illuminating cells loaded with cm-IP3 with femtosecond-pulsed near-infrared light (730 nm), we observed a burst of Ca2+ activity in the uncaging area. This local Ca2+ rise rapidly propagated across cells and could be repeated many times in different sub-cellular locations to produce artificial Ca2+ oscillations of defined amplitudes and frequencies. The complementary advantages of these IP3 prodrugs should provide new approaches for studying IP3-Ca2+ signaling in intact cell populations with high spatiotemporal resolutions.     

Purification of bovine brain inositol-1,4,5-trisphosphate 5-phosphatase.

In bovine brain, two soluble inositol-1,4,5-trisphosphate (InsP3) 5-phosphatases, which catalyse the dephosphorylation of InsP3 to inositol 1,4-bisphosphate, have been separated by DEAE-Sephacel. Type I, i.e. the first eluted enzyme, is the main soluble form and is reminiscent of the membrane-bound enzyme by multiple criteria. Type I was purified to apparent homogeneity by a method involving chromatography on DEAE-Sephacel, Blue-Sepharose, Sephacryl S-200, phosphocellulose, and C18 HPLC. A single protein band of 42-43 kDa was identified by SDS/PAGE, corresponding to the peak of maximal activity. InsP3 5-phosphatase was purified to apparent homogeneity to a final yield of 45-50 micrograms protein. The minimal estimate value of the Vmax for InsP3 5-phosphatase was in the range 20-35 mumol.min-1.mg protein-1.     

Mode of activation of bovine brain inositol 1,4,5-trisphosphate 3-kinase by calmodulin and calcium.

The effect of Ca2+ and calmodulin (CaM) on the activation of purified bovine brain Ins(1,4,5)P3 kinase was quantified and interpreted according to the model of sequential equilibria generally used for other calmodulin-stimulated systems. Two main conclusions can be drawn. (i) CaM.Ca3 and CaM.Ca4 together are the biologically active species in vitro, as is the case for the great majority of other calmodulin targets. (ii) These species bind in a non-co-operative way to the enzyme with an affinity constant of 8.23 x 10(9) M-1, i.e. approx 10-fold higher than for most calmodulin-activated target enzymes. The dose-response curve of the activation of Ins(1,4,5)P3 kinase by calmodulin is not significantly impaired by melittin and trifluoperazine, whereas under very similar assay conditions the half-maximal activation of bovine brain cyclic AMP phosphodiesterase requires over 30-50-fold higher concentrations of CaM when 1 microM melittin or 20 microM-trifluoperazine is present in the assay medium. Similarly, 1 microM of the anti-calmodulin peptides seminalplasmin and gramicidin S, as well as 20 microM of N-(6-aminohexyl)-5-chloro-1-naphthalene-sulphonamide (W7), do not inhibit the activation process. These data suggest that binding and activation of Ins(1,4,5)P3 kinase require surface sites of calmodulin which are different from those involved in the binding of most other target enzymes or of model peptides.     

Synthesis of glucopyranoside-based ligands for D-myo-inositol 1,4,5-trisphosphate receptors

Adenophostins A and B are naturally occurring glyconucleotides that interact potently with receptors for D-myo-inositol 1,4,5-trisphosphate, an important second messenger molecule in most cell types. Here we describe the design and synthesis of glucopyranoside-based analogues of adenophostin A lacking the adenine component. The key synthetic strategy involves glycosylation of selectively protected alcohols, derived from methyl beta-D-ribofuranoside or 1,4-anhydroerythritol, using glycosyl donors synthesised from 2,6-di-O-benzyl-D-glucopyranose derivatives. Further elaboration and deprotection of the coupled products gave two trisphosphate analogues; methyl 3-O-alpha-D-glucopyranosyl-beta-D-ribofuranoside 2,3',4'-trisphosphate ("ribophostin") and (3'S,4'R)-3'-hydroxytetrahydrofuran-4'-yl alpha-D-glucopyranoside 3,4,3'-trisphosphosphate ("furanophostin"). The route to furanophostin was further modified to give (3'S,4'R)-3'-hydroxytetrahydrofuran-4'-yl alpha-D-glucopyranoside 3'-phosphate 3,4-bisphosphorothioate, the first phosphorothioate-containing adenophostin analogue.     

Purification of bovine brain inositol 1,4,5-trisphosphate 3-kinase. Identification of the enzyme by sodium dodecyl sulphate/polyacrylamide-gel electrophoresis.

Treatment of human platelets with concentrations of benzyl alcohol up to 50 mM augmented adenylate cyclase activity when it was assayed in the basal state and when stimulated by prostaglandin E1 (PGE1), isoprenaline or NaF. Benzyl alcohol antagonized the stimulatory effect exerted on the catalytic unit of adenylate cyclase by the diterpene forskolin. Benzyl alcohol did not modify the magnitude of the inhibitory response when the catalytic unit of adenylate cyclase was inhibited by using either low concentrations of guanosine 5'-[beta gamma-imido]triphosphate, which acts selectively on the inhibitory guanine nucleotide-regulatory protein Gi, or during alpha 2-adrenoceptor occupancy, by using adrenaline (+ propranolol). Some 34% of the potent inhibitory action of adrenaline on PGE1-stimulated adenylate cyclase was obliterated in a dose-dependent fashion (concn. giving 50% inhibition = 12.5 mM) by benzyl alcohol, with the residual inhibitory action being apparently resistant to the action of benzyl alcohol at concentrations up to 50 mM. Treatment of membranes with benzyl alcohol did not lead to the release of either the alpha-subunit of Gi or G-protein subunits. The alpha 2-adrenoceptor-mediated inhibition of adenylate cyclase was abolished when assays were performed in the presence of Mn2+ rather than Mg2+ and, under such conditions, dose-effect curves for the action of benzyl alcohol on PGE1-stimulated adenylate cyclase activity were similar whether or not adrenaline (+propranolol) was present. We suggest that (i) alpha 2-adrenoceptor- and Gi-mediated inhibition of PGE1-stimulated adenylate cyclase may have two components, one of which is sensitive to inhibition by benzyl alcohol, and (ii) the Gi-mediated inhibition of forskolin-stimulated adenylate cyclase exhibits predominantly the benzyl alcohol-insensitive component.     

D-myo-inositol 1,4,5-trisphosphate 3-kinase A is activated by receptor activation through a calcium:calmodulin-dependent protein kinase II phosphorylation mechanism.

D-myo-inositol 1,4,5-trisphosphate [Ins(1,4,5)P3] 3-kinase, the enzyme responsible for production of D-myo-inositol 1,3,4,5-tetrakisphosphate, was activated 3- to 5-fold in homogenates of rat brain cortical slices after incubation with carbachol. The effect was reproduced in response to UTP in Chinese hamster ovary (CHO) cells overexpressing Ins(1,4,5)P3 3-kinase A, the major isoform present in rat and human neuronal cells. In ortho-32P-labelled cells, the phosphorylated 53 kDa enzyme could be identified after receptor activation by immunoprecipitation. The time course of phosphorylation was very similar to that observed for carbachol (or UTP)-induced enzyme activation. Enzyme phosphorylation was prevented in the presence of okadaic acid. Calmodulin (CaM) kinase II inhibitors (i.e. KN-93 and KN-62) prevented phosphorylation of Ins(1,4,5)P3 3-kinase. Identification of the phosphorylation site in transfected CHO cells indicated that the phosphorylated residue was Thr311. This residue of the human brain sequence lies in an active site peptide segment corresponding to a CaM kinase II-mediated phosphorylation consensus site, i.e. Arg-Ala-Val-Thr. The same residue in Ins(1,4,5)P3 3-kinase A was also phosphorylated in vitro by CaM kinase II. Phosphorylation resulted in 8- to 10-fold enzyme activation and a 25-fold increase in sensitivity to the Ca2+:CaM complex. In this study, direct evidence is provided for a novel regulation mechanism for Ins(1,4,5)P3 3-kinase (isoform A) in vitro and in intact cells.     

A purification strategy for inositol 1,4,5-trisphosphate 3-kinase from rat liver based upon heparin interaction chromatography.

Rat liver inositol 1,4,5-trisphosphate [Ins (1,4,5)P3] 3-kinase was purified in high yield by a three-step procedure reliant upon chromatography on heparin and calmodulin agarose. Purified enzyme was stable in the presence of the detergent 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulphonate (CHAPS) (0.1-0.5%) and the sulphydryl reducing reagent dithiothreitol (DTT). The purified enzyme was activated 2-3-fold by Ca2+ (1 microM) in the presence of calmodulin. Pyrophosphate and heparin were identified as inhibitors of the enzyme.