Proteasome activator 28γ (PA28γ) allosterically activates trypsin-like proteolysis by binding to the α-ring of the 20S proteasome

Proteasome activator 28γ (PA28γ/REGγ) is a member of the 11S family of proteasomal regulators that is constitutively expressed in the nucleus and implicated in various diseases, including certain cancers and systemic lupus erythematosus. Despite years of investigation, how PA28γ functions to stimulate proteasomal protein degradation remains unclear. Alternative hypotheses have been proposed for the molecular mechanism of PA28γ, including the following: (1) substrate selection, (2) allosteric upregulation of the trypsin-like (T-L) site, (3) allosteric inhibition of the chymotrypsin-like (CT-L) and caspase-like (C-L) sites, (4) conversion of the CT-L or C-L sites to new T-L sites, and (5) gate opening alone or in combination with a previous hypothesis. Here, by mechanistically decoupling gating effects from active site effects, we unambiguously demonstrate that WT PA28γ allosterically activates the T-L site. We show PA28γ binding increases the Kcat/Km by 13-fold for T-L peptide substrates while having little-to-no effect on hydrolysis kinetics for CT-L or C-L substrates. Furthermore, mutagenesis and domain swaps of PA28γ reveal that it does not select for T-L peptide substrates through either the substrate entry pore or the distal intrinsically disordered region. We also show that a previously reported point mutation can functionally switch PA28γ from a T-L activating to a gate-opening activator in a mutually exclusive fashion. Finally, using cryogenic electron microscopy, we visualized the PA28γ-proteasome complex at 4.3 Å and confirmed its expected quaternary structure. The results of this study provide unambiguous evidence that PA28γ can function by binding the 20S proteasome to allosterically activate the T-L proteolytic site.     

Multicolor BiFC analysis of competition among G protein β and γ subunit interactions

We have applied multicolor BiFC to study the association preferences of G protein β and γ subunits in living cells. Cells co-express multiple isoforms of β and γ subunits, most of which can form complexes. Although many βγ complexes exhibit similar properties when assayed in reconstituted systems, knockout experiments in vivo suggest that individual isoforms have unique functions. BiFC makes it possible to correlate βγ complex formation with functionality in intact cells by comparing the amounts of fluorescent βγ complexes with their abilities to modulate effector proteins. The relative predominance of specific βγ complexes in vivo is not known. To address this issue, multicolor BiFC can determine the association preferences of β and γ subunits by simultaneously visualizing the two fluorescent complexes formed when β or γ subunits fused to amino terminal fragments of yellow fluorescent protein (YFP-N) and cyan fluorescent protein (CFP-N) compete to interact with limiting amounts of a common γ or β subunit, respectively, fused to a carboxyl terminal fragment of CFP (CFP-C). Multicolor BiFC also makes it possible to determine the roles of interacting proteins in the subcellular targeting of complexes, study the formation of protein complexes that are unstable under isolation conditions, determine the roles of co-expressed proteins in regulating the association preferences of interacting proteins, and visualize dynamic events affecting multiple protein complexes. These approaches can be applied to studying the assembly and functions of a wide variety of protein complexes in the context of a living cell.     

Combinatorial expression of bacterial whole mevalonate pathway for the production of β-carotene in E. coli

The increased synthesis of building blocks of IPP (isopentenyl diphosphate) and DMAPP (dimethylallyl diphosphate) through metabolic engineering is a way to enhance the production of carotenoids. Using E. coli as a host, IPP and DMAPP supply can be increased significantly through the introduction of foreign MVA (mevalonate) pathway into it. The MVA pathway is split into two parts with the top and bottom portions supplying mevalonate from acetyl-CoA, and IPP and DMAPP from mevalonate, respectively. The bottom portions of MVA pathway from Streptococcus pneumonia, Enterococcus faecalis, Staphylococcus aureus, Streptococcus pyogenes and Saccharomyces cerevisiae were compared with exogenous mevalonate supplementation for β-carotene production in recombinant Escherichia coli harboring β-carotene synthesis genes. The E. coli harboring the bottom MVA pathway of S. pneumoniae produced the highest amount of β-carotene. The top portions of MVA pathway were also compared and the top MVA pathway of E. faecalis was found out to be the most efficient for mevalonate production in E. coli. The whole MVA pathway was constructed by combining the bottom and top portions of MVA pathway of S. pneumoniae and E. faecalis, respectively. The recombinant E. coli harboring the whole MVA pathway and β-carotene synthesis genes produced high amount of β-carotene even without exogenous mevalonate supplementation. When comparing various E. coli strains – MG1655, DH5α, S17-1, XL1-Blue and BL21 – the DH5α was found to be the best β-carotene producer. Using glycerol as the carbon source for β-carotene production was found to be superior to glucose, galactose, xylose and maltose. The recombinant E. coli DH5α harboring the whole MVA pathway and β-carotene synthesis genes produced β-carotene of 465 mg/L at glycerol concentration of 2% (w/v).     

PA28α overexpressing female mice maintain exploratory behavior and capacity to prevent protein aggregation in hippocampus as they age

With age, protein damage accumulates and increases the risk of age‐related diseases. The proteasome activator PA28αβ is involved in protein damage clearance during early embryogenesis and has demonstrated protective effects against proteinopathy. We have recently discovered that adult female mice overexpressing PA28α (PA28αOE) have enhanced learning and memory, and protein extracts from their hippocampi prevent aggregation more efficiently than wild type. In this study, we investigated the effect of overexpressing PA28α on aging using C57BL/6N×BALB/c F2 hybrid mice. We found that the hippocampal anti‐aggregation effect was maintained in young adult (7 months) to middle‐aged (15 months) and old (22 months) PA28αOE females. While the PA28αOE influence on learning and memory gradually decreased with aging, old PA28αOE females did not display the typical drop in explorative behavior—a behavioral hallmark of aging—but were as explorative as young mice. PA28αOE lowered PA28‐dependent proteasome capacity in both heart and hippocampus, and there was no indication of lower protein damage load in PA28αOE. The life span of PA28αOE was also similar to wild type. In both wild type and PA28αOE, PA28‐dependent proteasome capacity increased with aging in the heart, while 26S and 20S proteasome capacities were unchanged in the timepoints analyzed. Thus, PA28αOE females exhibit improved hippocampal ability to prevent aggregation throughout life and enhanced cognitive capabilities with different behavioral outcomes dependent on age; improved memory at early age and a youth‐like exploration at old age. The cognitive effects of PA28αβ combined with its anti‐aggregation molecular effect highlight the therapeutical potential of PA28αβ in combating proteinopathies.     

Molecular determinants of desensitization and assembly of the chimeric GABAA receptor subunits (α1/γ2) and (γ2/α1) in combinations with β2 and γ2

Two γ-aminobutyric acid_A (GABA_A) receptor chimeras were designed in order to elucidate the structural requirements for GABA_A receptor desensitization and assembly. The (α1/γ2) and (γ2/α1) chimeric subunits representing the extracellular N-terminal domain of α1 or γ2 and the remainder of the γ2 or α1 subunits, respectively, were expressed with β2 and β2γ2 in Spodoptera frugiperda (Sf-9) cells using the baculovirus expression system. The (α1/γ2)β2 and (α1/γ2)β2γ2 but not the (γ2/α1)β2 and (γ2/α1)β2γ2 subunit combinations formed functional receptor complexes as shown by whole-cell patch–clamp recordings and [³H]muscimol and [³H]flunitrazepam binding. Moreover, the surface immunofluorescence staining of Sf-9 cells expressing the (α1/γ2)-containing receptors was pronounced, as opposed to the staining of the (γ2/α1)-containing receptors, which was only slightly higher than background. To explain this, the (α1/γ2) and (γ2/α1) chimeras may act like α1 and γ2 subunits, respectively, indicating that the extracellular N-terminal segment is important for assembly. However, the (α1/γ2) chimeric subunit had characteristics different from the α1 subunit, since the (α1/γ2) chimera gave rise to no desensitization after GABA stimulation in whole-cell patch–clamp recordings, which was independent of whether the chimera was expressed in combination with β2 or β2γ2. Surprisingly, the (α1/γ2)(γ2/α1)β2 subunit combination did desensitize, indicating that the C-terminal segment of the α1 subunit may be important for desensitization. Moreover, desensitization was observed for the (α1/γ2)β2γ2 receptor with respect to the direct activation by pentobarbital. This suggests differences in the mechanism of channel activation for pentobarbital and GABA.     

Large-scale production of the carbohydrate portion of the sialyl–Tn epitope, α-Neup5Ac-(2→6)- -GalpNAc, through bacterial coupling

α-Neup5Ac-(2→6)- -GalpNAc, the carbohydrate portion of sialyl–Tn epitope of the tumor-associated carbohydrate antigen, was prepared by a whole-cell reaction through the combination of recombinant Escherichia coli strains and Corynebacterium ammoniagenes. Two recombinant E. coli strains overexpressed the CMP-Neup5Ac biosynthetic genes and the α-(2→6)-sialyltransferase gene of Photobacterium damsela. C. ammoniagenes contributed to the production of UTP from orotic acid. α-Neup5Ac-(2→6)- -GalpNAc was accumulated at 87 mM (45 g/L) after a 25-h reaction starting from orotic acid, N-acetylneuraminic acid, and 2-acetamide-2-deoxy- -galactose.     

Assembly and Intracellular Targeting of the βγ Subunits of Heterotrimeric G Proteins

The assembly in living cells of heterotrimeric guanine nucleotide binding proteins from their constituent α, β, and γ subunits is a complex process, compounded by the multiplicity of the genes that encode them, and the diversity of receptors and effectors with which they interact. Monoclonal anti-β antibodies (ARC5 and ARC9), raised against immunoaffinity purified βγ complexes, recognize β subunits when not associated with γ and can thus be used to monitor assembly of βγ complexes. Complex formation starts immediately after synthesis and is complete within 30 min. Assembly occurs predominantly in the cytosol, and association of βγ complexes with the plasma membrane fraction starts between 15–30 min of chase. Three pools of β subunits can be distinguished based on their association with γ subunits, their localization, and their detergent solubility. Association of β and α subunits with detergent-insoluble domains occurs within 1 min of chase, and increases to reach a plateau of near complete detergent resistance within 30 min of chase. Brefeldin A treatment does not interfere with delivery of βγ subunits to detergent-insoluble domains, suggesting that assembly of G protein subunits with their receptors occurs distally from the BFA-imposed block of intracellular membrane trafficking and may occur directly at the plasma membrane.     

Biosynthesis of 20-hydroxyecdysone in Ajuga hairy roots: fate of 6α- and 6β-hydrogens of lathosterol

The fate of 6α- and 6β-hydrogens of lathosterol during the transformation into 20-hydroxyecdysone was chased by feeding [3α,6β-²H₂]- and [3α,6α-²H₂]-lathosterols to hairy roots of Ajuga reptans var. atropurpurea. The behavior of 6β-hydrogen, which mostly migrated to the C-5 position of 20-hydroxyecdysone, was in agreement with that of C-6 hydrogen of cholesterol. The results strongly supported the view that cholesterol and lathosterol are first metabolized into 7-dehydrocholesterol, which is then converted into 20-hydroxyecdysone via 7-dehydrocholesterol 5α,6α-epoxide in the hairy roots.     

Large-scale production of the carbohydrate portion of the sialyl–Tn epitope, α-Neup5Ac-(2→6)-d-GalpNAc, through bacterial coupling

α-Neup5Ac-(2→6)-d-GalpNAc, the carbohydrate portion of sialyl–Tn epitope of the tumor-associated carbohydrate antigen, was prepared by a whole-cell reaction through the combination of recombinant Escherichia coli strains and Corynebacterium ammoniagenes. Two recombinant E. coli strains overexpressed the CMP-Neup5Ac biosynthetic genes and the α-(2→6)-sialyltransferase gene of Photobacterium damsela. C. ammoniagenes contributed to the production of UTP from orotic acid. α-Neup5Ac-(2→6)-d-GalpNAc was accumulated at 87 mM (45 g/L) after a 25-h reaction starting from orotic acid, N-acetylneuraminic acid, and 2-acetamide-2-deoxy-d-galactose.     

A novel role for PA28gamma-proteasome in nuclear speckle organization and SR protein trafficking

In eukaryotic cells, proteasomes play an essential role in intracellular proteolysis and are involved in the control of most biological processes through regulated degradation of key proteins. Analysis of 20S proteasome localization in human cell lines, using ectopic expression of its CFP-tagged α7 subunit, revealed the presence in nuclear foci of a specific and proteolytically active complex made by association of the 20S proteasome with its PA28γ regulator. Identification of these foci as the nuclear speckles (NS), which are dynamic subnuclear structures enriched in splicing factors (including the SR protein family), prompted us to analyze the role(s) of proteasome-PA28γ complexes in the NS. Here, we show that knockdown of these complexes by small interfering RNAs directed against PA28γ strongly impacts the organization of the NS. Further analysis of PA28γ-depleted cells demonstrated an alteration of intranuclear trafficking of SR proteins. Thus, our data identify proteasome-PA28γ complexes as a novel regulator of NS organization and function, acting most likely through selective proteolysis. These results constitute the first demonstration of a role of a specific proteasome complex in a defined subnuclear compartment and suggest that proteolysis plays important functions in the precise control of splicing factors trafficking within the nucleus.     

15β-hydroxysteroids (Part IV). Steroids of the human perinatal period: The synthesis of 3α,15β,17α-trihydroxy-5α-pregnan-20-one and its A/B-ring configurational isomers

In recent years several 15β-hydroxysteroids have emerged pathognomonic of adrenal disorders in human neonates of which 3α,15β,17α-trihydroxy-5β-pregnan-20-one (2) was the first to be identified in the urine of newborn infants affected with congenital adrenal hyperplasia. In this investigation we report the synthesis of the three remaining 3ξ,5ξ-isomers, namely 3α,15β,17α-trihydroxy-5α-pregnan-20-one (3), 3β,15β,17α-trihydroxy-5α-pregnan-20-one (7) and 3β,15β,17α-trihydroxy-5β-pregnan-20-one (8) for their definitive identification in pathological conditions in human neonates. 3β,15β-Diacetoxy-17α-hydroxy-5-pregnen-20-one (11), a product of chemical synthesis was converted to the isomeric 3 and 7, while conversion of 15β,17α-dihydroxy-4-pregnen-3,20-dione (4), a product of microbiological transformation, resulted in the preparation of 8. In brief, selective acetate hydrolysis of 11 gave 15β-acetoxy-3β,17α-dihydroxy-5-pregnen-20-one (12) which on catalytic hydrogenation gave 15β-acetoxy-3β,17α-dihydroxy-5α-pregnan-20-one (13) a common intermediate for the synthesis of the 3β(and α),5α-isomers. Hydrolysis of the 15β-acetate gave 7, whereas oxidation with pyridinium chlorochromate gave 15β-acetoxy-17α-hydroxy-5α-pregnan-3,20-dione (14) which on reduction with -Selectride and hydrolysis of the 15β-acetate gave 3. Finally, hydrogenation of 4 gave 15β,17α-dihydroxy-5β-pregnan-3,20-dione (10) which on reduction with -Selectride gave 8.     

Cloning of cDNAs coding for the Gαil and Gαi2 G-proteins from chick brain

We have cloned and characterized several cDNAs coding for G-protein inhibitory α subunits (Gai) from a chick brain cDNA library. Based on homology to Ga subunits from other eukaryotes, these clones were designated chick Gαil and Gαi2. On the deduced amino-acid level, Gαi1 and Gαi2 were found to be 98 and 95% identical to rat Goal and Gαi2, respectively. Using RNase protection analysis, the Gαi1 and Gαi2 mRNAs were found to be expressed in chick atria, ventricle, lung, liver, brain and kidney.     

Identification of a BRAF/PA28γ/MEK1 signaling axis and its role in epithelial-mesenchymal transition in oral submucous fibrosis

Oral submucous fibrosis (OSF) is a chronic and insidious oral potentially malignant disorder associated with a 4–17% risk of oral squamous cell carcinoma (OSCC). Our previous study found that proteasomal activator 28 gamma (PA28γ) is frequently overexpressed in oral squamous cell carcinoma and negatively correlated with poor patient prognosis. However, the role of PA28γ in the occurrence and development of OSF remains unclear. Here, we screened PA28γ-related genes and investigated their function in OSF. We demonstrated that the expression of PA28γ was positively associated with MEK1 and gradually elevated from normal to progressive stages of OSF tissue. Arecoline, a pathogenic component of OSF, could upregulate the protein levels of PA28γ and phosphorylated MEK1 and contribute to epithelial to mesenchymal transition (EMT) in epithelial cells. Notably, PA28γ could interact with MEK1 and upregulate its phosphorylation level. Furthermore, arecoline upregulated BRAF, which can interact with PA28γ and upregulate its protein level. Additionally, BRAF, PA28γ, and MEK1 could form protein complexes and then enhance the MEK1/ERK signaling pathways. The concrete mechanism of the protein stability of PA28γ is that BRAF mediates its degradation by inhibiting its ubiquitination. These findings underscore the instrumental role of PA28γ in the BRAF/MEK1 pathway and enhanced EMT through MEK1/ERK activation in OSF.Subject terms: Mechanisms of disease, Oral cancer     

Characterization of the 5′-to-5′linked adult α- and β-globin genes from three sciaenid fish species (Pseudosciaena crocea, Sciaenops ocellatus, Nibea miichthioides)

Recently, we cloned the adult α-globin genes from large yellow croaker Pseudosciaena crocea, cuneate drum Nibea miichthioides and red drum Sciaenops ocellatus. All these α-globins have a unique Gly insertion at the 47th residue. In this paper, the three sciaenid globin complexes were identified and compared in detail. Linkage analysis indicated that the sciaenid α- and β-globin genes were oriented head-to-head relative to each other. The sciaenid intergenic regions between the linked α- and β-globin genes were the smallest in reported fish globin gene complexes to date. Classical promoter elements were condensed and the CCAAT box unstable duplication was found in these regions. The promoter function of the intergenic region from large yellow croaker was tested by transient expression of EGFP in Vero cells. We also described a method for studying luciferase reporter gene transient expression in primary fish erythrocytes. We used the method to assess the promoter strength of the three intergenic regions between the sciaenid α- and β-globin genes.     

Enantioselective ester hydrolase from Sphingobacterium sp. 238C5 useful for chiral resolution of β-phenylalanine and for its β-peptide synthesis

A novel enzyme, β-phenylalanine ester hydrolase, useful for chiral resolution of β-phenylalanine and for its β-peptide synthesis was characterized. The enzyme purified from the cell free-extract of Sphingobacterium sp. 238C5 well hydrolyzed β-phenylalanine esters (S)-stereospecifically. Besides β-phenylalanine esters, the enzyme catalyzed the hydrolysis of several α-amino acid esters with l-stereospecificity, while the deduced 369 amino acid sequence of the enzyme exhibited homology to alkaline d-stereospecific peptide hydrolases from Bacillus strains. Escherichia coli transformant expressing the β-phenylalanine ester hydrolase gene exhibited an about 8-fold increase in specific (S)-β-phenylalanine ethyl ester hydrolysis as compared with that of Sphingobacterium sp. 238C5. The E. coli transformant showed (S)-enantiomer specific esterase activity in the reaction with a low concentration (30 mM) of β-phenylalanine ethyl ester, while it showed both esterase and transpeptidase activity in the reaction with a high concentration (170 mM) of β-phenylalanine ethyl ester and produced β-phenylalanyl-β-phenylalanine ethyl ester. This transpeptidase activity was useful for β-phenylalanine β-peptide synthesis.     

3β-hydroxysteroid dehydrogenase activity in tissues of the human fetus determined with 5α-androstane-3β,17β-diol and dehydroepiandrosterone as substrates

3β-Hydroxysteroid dehydrogenase (3β-HSD)/Δ^5→4-isomerase activity in steroidogenic tissues is required for the synthesis of biologically active steroids. Previously, by use of dehydroepiandrosterone (3β-hydroxy-5-androsten-17-one, DHEA) as substrate, it was established that in addition to steroidogenic tissues 3β-HSD/Δ^5→4-isomerase activity also is expressed in extraglandular tissues of the human fetus. In the present study, we attempted to determine whether the C-5,C-6-double bond of DHEA serves to influence 3β-HSD activity. For this purpose, we compared the efficiencies of a 3β-hydroxy-5-ene steroid (DHEA) and a 3β-hydroxy-5α-reduced steroid (5α-androstane-3β,17β-diol, 5α-A-diol) as substrates for the enzyme. The apparent Michaelis constant (K_m) for 5α-A-diol in midtrimester placenta, fetal liver, and fetal skin tissues was at least one order of magnitude higher than that for DHEA, viz the apparent K_m of placental 3β-HSD for 5α-A-diol was in the range of 18 to 40 μmol/l (n = 3) vs 0.45 to 4 μmol/l for DHEA (n = 3); for the liver enzyme, 17 μmol/l for 5α-A-diol and 0.60 μmol/l for DHEA, and for the skin enzyme 14 and 0.18 μmol/l, respectively. Moreover, in 13 human fetal tissues evaluated the maximal velocities obtained with 5α-A-diol as substrate were higher than those obtained with DHEA. A similar finding in regard to K_ms and rates of product formation was obtained by use of purified placental 3β-HSD with DHEA, pregnenolone, and 3β-hydroxy-5α-androstan-17-one (epiandrosterone) as substrates: the K_m of 3β-HSD for DHEA was 2.8 μmol/l, for pregnenolone 1.9 μmol/l, and for epiandrosterone 25 μmol/l. The specific activity of the purified enzyme with pregnenolone as substrate was 27 nmol/mg protein·min and, with epiandrosterone, 127 nmol/mg protein·min. With placental homogenate as the source of 3β-HSD, DHEA at a constant level of 5 μmol/l behaved as a competitive inhibitor when the radiolabeled substrate, [³H]5α-A-diol, was present in concentrations of 20 to 60 μmol/l, but a lower substrate concentrations the inhibition was of the mixed type; similar results were obtained with [³H]DHEA as the substrate at variable concentrations in the presence of a fixed concentration of 5α-A-diol (40 μmol/l). These findings are indicative that both steroids bind to a common site on the enzyme, however, the binding affinity for these steroids appear to differ markedly as suggested by the respective K_ms. Studies of inactivation of purified placental 3β-HSD/Δ^5→4-isomerase by an irreversible inhibitor, viz 5,10-secoestr-4-yne-3,10,17-trione, were suggestive that the placental protein adopts different conformations depending on whether the steroidal substrate has a 5α-configuration, e.g. epiandrosterone, or a C-5,C-6-double bond e.g. DHEA or pregnenolone. The lower rates of product formation obtained with placenta and fetal tissues by use of 3β-hydroxy-5-ene steroids as substrates when compared with those obtained with 3β-hydroxy-5α-reduced steroids may be explained by a combination of factors, including: (i) inhibition of 3β-HSD activity by end products of metabolism of 3β-hydroxy-5-ene steroids, e.g. 4-androstene-3,17-dione formed with DHEA as substrate; (ii) higher binding affinity of the enzyme for 3β-hydroxy-5-ene steroids—and possibly for their 3-oxo-5-ene metabolites; (iii) lack of a requirement for the isomerization step with 5α-reduced steroids as substrates, and (iv) the possible presence in fetal tissues of an enzyme with 3β-HSD activity only (i.e. no Δ^5→4-isomerase).     

Posttranslational Modifications and β/γ Chain Associations of Human Laminin α1 and Laminin α5 Chains: Purification of Laminin-3 from Placenta

Laminins assemble into trimers composed of α, β, and γ chains which posttranslationally are glycosylated and sometimes proteolytically cleaved. In the current paper we set out to characterize posttranslational modifications and the laminin isoforms formed by laminin α1 and α5 chains. Comparative pulse–chase experiments and deglycosylation studies in JAR cells established that the M_r 360,000 laminin α1 chain is glycosylated into a mature M_r 400,000 band while the M_r 370,000 laminin α5 chain is glycosylated into a M_r 390,000 form that upon secretion is further processed into a M_r 380,000 form. Hence, despite the shorter peptide length of α1 chain in comparison with the α5 chain, secreted α1 assumes a larger size in SDS–PAGE due to a higher degree of N-linked glycosylation and due to the lack of proteolytic processing. Immunoprecipitations and Western blotting of JAR laminins identified laminin α1 and laminin α5 chains in laminin-1 and laminin-10. In placenta laminin α1 chain (M_r 400,000) and laminin α5 chain (M_r 380,000/370,000 doublet) were found in laminin-1/-3 and laminin-10/-11. Immunohistochemically we could establish that the laminin α1 chain in placenta is deposited in the developing villous and trophoblast basement membrane, also found to contain laminin β2 chains. Surprisingly, a fraction of the laminin α1 chain from JAR cells and placenta could not be precipitated by antibodies to laminin β1–β3 chains, possibly pointing to an unexpected complexity in the chain composition of α1-containing laminin isoforms.