Elevation of a threonine phospholipid in polyoma virus transformed hamster embryo fibroblasts.

Analysis of the lipids of normal hamster embryo fibroblasts and polyoma virus transformed fibroblasts shows a decrease in phosphatidylcholine and phosphatidylethanolamine and a marked increase in a threonine phospholipid after transformation. Transformed cells also react differently with fluorodinitrobenzene and trinitrobenzenesulfonate. phosphatidylethanolamine of transformed cells reacts to a greater extent with both probes. Phosphatidylserine and the threonine phospholipid of both cells do not react with trinitrobenzenesulfonate. The threonine phospholipid is provisionally identified as phosphatidylthreonine.     

FBJ osteosarcoma virus in tissue culture. III. Isolation and characterization of non-virus-producing FBJ-transformed cells.

Hamster and rat cell lines have been established that have been transformed by FBJ murine sarcoma virus (FBJ-MuSV) but that do not produce virus. The hamster cell line originated from an osteosarcoma that appeared in a hamster inoculated at birth with an extract of a CFNo1 mouse FBJ-osteosarcoma. The rat cell line was obtained by transferring the FBJ-MuSV genome to normal rat kidney cells in the absence of the FBJ type C virus (FBJ-MuLV), which, usually in high concentration, accompanies the FBJ-MuSV. Both transformed hamster and rat cell lines contain the FBJ-MuSV genome, which can be rescued by ecotropic and xenotropic murine type C viruses. This rescued genome produces characteristic FBJ-MuSV foci in tissue culture and, in appropriate animal hosts, induces osteosarcomas typical of those induced by FBJ-MuSV. FBJ-MuSV was isolated originally from a parosteal osteosarcoma that occurred naturally in a mouse. Since there was no previous history of passage of the agent through any other animal species, these non-virus-producing hamster and rat cells transformed by FBJ-MuSV should be very helpful in molecular studies examining the origin of spontaneous sarcoma genomes in mice.     

Translocation of phosphatidylthreonine and -serine to mitochondria diminishes exponentially with increasing molecular hydrophobicity

Some cultured cells contain significant amounts of a rarely recognized phospholipid, phosphatidylthreonine. Since phosphatidylthreonine is a structural analog of phosphatidylserine, the question rises whether it is transported to mitochondria and decarboxylated to phosphatidylisopropanolamine therein. We studied this issue with hamster kidney cell-line using a novel approach, i.e. electrospray mass-spectrometry and stable isotope-labeled precursors. Scanning for a neutral loss of 155, which is characteristic for phosphatidylisopropanolamine, indicated that this lipid is indeed present. The identity of phosphatidylisopropanolamine was supported by the following: (i) it co-chromatographed with phosphatidylethanolamine; (ii) its molecular species profile was similar to that of phosphatidylethanolamine; (iii) its head group was labeled from ¹³C-threonine; and (iv) its concentration increased in parallel with phosphatidylthreonine. Tests with solubilized decarboxylase and subcellular fractionation studies indicated that the low cellular content of phosphatidylisopropanolamine is due to inefficient decarboxylation, rather than poor translocation of phosphatidylthreonine to mitochondria. Importantly, the average hydrophobicity of phosphatidylisopropanolamine molecular species was significantly less than that of phosphatidylthreonine species, indicating that hydrophilic phosphatidylthreonine species translocate to mitochondria far more rapidly than hydrophobic ones. Parallel results were obtained for phosphatidylserine. These findings imply that efflux from the ER membrane could be the rate-limiting step in the phosphatidylthreonine and -serine translocation to mitochondria.     

Phorbol ester stimulates the hydrolysis of phosphatidylethanolamine in leukemic HL-60, NIH 3T3, and baby hamster kidney cells

Treatment of leukemic HL-60, NIH 3T3, and baby hamster kidney (BHK-21) cells, prelabeled with [2-14C]ethanolamine, with 12-O-tetradecanoylphorbol-13-acetate (TPA), a potent activator of protein kinase C, resulted in increased degradation of both 14C-labeled phosphatidylethanolamine and its alkenyl (plasmalogen) derivate. A half-maximal and a maximal (approximately 3.4-fold) stimulation of ethanolamine phospholipid degradation required 3 and 10-20 nM TPA, respectively. TPA had a similar concentration-dependent stimulatory effect on the hydrolysis of phosphatidylcholine in cells previously prelabeled with [methyl-14C]choline. Increased phospholipid degradation was not accompanied by the formation of lysophosphatidylethanolamine, indicating that a phospholipase A-type enzyme was not involved. About 80% of total water-soluble degradation products was ethanolamine, suggesting that phospholipid hydrolysis was catalyzed by a phospholipase D-type enzyme. Increased formation of ethanolamine with exposure of cells to TPA was observed only after a 10-min lag period. Mezerein, bryostatin, sn-1-oleoyl-2-acetylglycerol, and polymyxin B, all of which mimic the action of TPA on protein phosphorylation in vivo, also stimulated the hydrolysis of ethanolamine phospholipids in HL-60 cells, suggesting that the TPA effect was mediated by protein kinase C.     

An extra species of lysine transfer ribonucleic acid in polyoma virus-transformed cells in tissue culture

Isoaccepting tRNAs from various mouse cells were fractionated on columns of benzoylated DEAE cellulose. Lysine tRNA from mouse embryo, adult mouse liver and kidney, primary mouse embryo cells in tissue culture, and an established tissue culture line of mouse fibroblasts (3T3) has two peaks of isoaccepting tRNA; lysine tRNA from two established lines of polyoma virus-transformed cells contains an additional peak of lysine tRNA. The extra peak in transformed cells comprises about 25% of the acceptor capacity for lysine. It is stable to denaturation and renaturation and can be chromatographed, stripped of lysine, recharged, and rechromatographed. The extra peak is present in tRNA from transformed cells and absent in tRNA from normal cells regardless of whether the lysyl-tRNA ligase used for aminoacylation is from normal or transformed cells. Isoaccepting tRNAs for arginine, leucine, serine, valine, histidine, and tyrosine reveal similar profiles for the various tRNAs from normal and transformed cells.     

Density-dependent inhibition of cell growth is correlated with the activity of a cell surface phosphatidylinositol-specific phospholipase C.

We previously identified a novel phospatidylinositol-specific phospholipase C (PI-PLC) present at the surface of Swiss 3T3 cells using a fluorescent analog of PI and showed that this cell surface PI-PLC (csPI-PLC) activity increases with increasing cell density (J. Biol. Chem. 265, 5337-5340 (1990)). Here, we find that csPI-PLC activity also increased in cultures in which growth was inhibited due to serum deprivation. csPI-PLC was inhibited by agents that inhibit other mammalian PI-PLCs, but not by treatments which inhibit glycosyl PI-PLCs (GPI-PLCs). We also extended our studies using fluorescent PI to other cell types and found that csPI-PLC activity was present only in cell lines that exhibit growth inhibition upon reaching confluency (Swiss 3T3, 3T3-L1, BALB/c 3T3, and normal rat kidney (NRK) fibroblasts), but not in cell lines that are tumorigenic and/or do not exhibit growth inhibition in a density-dependent manner (Chinese hamster ovary (CHO), mouse L, SV-40 transformed BALB/c 3T3 (SV-T2), baby hamster kidney (BHK), and Chinese hamster lung (V79) fibroblasts). These results support the hypothesis that csPI-PLC plays a role in the density-dependent inhibition of cell growth.     

A procedure to introduce protein molecules into living mammalian cells

Although several methods are now available by which to introduce macromolecules into cultured living mammalian cells, each has limitations on its adoption as a general means, for a variety of purposes. We describe here a simple procedure to introduce protein molecules into various living mammalian cells. This procedure is based upon the finding that mammalian cells, after exposure to a low concentration of a phospholipid (L-alpha-lysophosphatidylcholine) in the presence of high (hypertonic) concentrations of glycerol became permeable to protein molecules and that a significant portion of the exposed cells regain their viability following incubation in the appropriate growth medium. We have demonstrated that diphtheria toxin (A fragment), horseradish peroxidase and antibodies against SV40 T-antigens are incorporated into living mouse erythroleukemia (Friend) cells, baby hamster kidney (BHK) cells and mouse fibroblasts (C3H), respectively. The volume introduced into a single cell (mouse Friend cells) is approx. 3 X 10(-15) liter, which is comparable to those with other systems. Parameters affecting permeability to protein molecules and viability of the treated cells were also investigated with these and other cell lines.     

Retention of phospholipid acyl groups is not characteristic of neoplastic cells in culture.

Phospholipid acyl group retention of normal and mouse sarcoma virus transformed cell lines has been examined. The cells were fed [1-14C]oleate for 24 h and then grown to non-radioactive medium for 4 days. Both the transformed rat and mouse cell lines continued to release 14C-labeled fatty acids and phospholipids into the culture medium similarly to the uninfected cells, and showed shifts in 14C between phospholipid subclasses. Thus, the phospholipid acyl group stability reported for L cells is not characteristic of oncogenic cells in culture.     

Endogenous xenobiotic enzyme levels in mammalian cells.

The response of mammalian cell lines to chemicals depends, in part, on the exogenous activation system used for the induction of a biological response. This could be attributed to differences in the expression of enzymes involved in xenobiotic metabolism. We have measured the activities of benzo[a]pyrene hydroxylase, dimethylaminoazobenzene N-demethylase, catalase, superoxide dismutase, peroxidase and glutathione-S-transferase in human lymphoblast TK6, mouse lymphoma L5178Y, Chinese hamster ovary (CHO) and lung (V79) and mouse C3H10T1/2 cell lines as well as in primary hepatocytes and S9 preparations of liver from male F344 rats. Nitroreductase was also measured in some of these preparations. Human lymphoblast TK6 and mouse C3H10T1/2 cells had the capacity to metabolize dimethylaminoazobenzene and the latter cell line also metabolized benzo[a]pyrene, indicating the presence of constitutive mono-oxygenase activity. Cytochrome P450 could not be detected spectrophotometrically in the cell lines. Western blot analysis indicated that P450 from the P450IIA family is expressed in C3H10T1/2 cells. Reactivity was also observed with an antibody to P450IA2; however, the identity of this protein remains uncertain. Superoxide dismutase, catalase and peroxidase, which protect cells against oxygen radical damage, were found in all the cell lines and in rat hepatocytes and S9. The human lymphoblast TK6 cell line, however, had the least of each of these three enzymes. Glutathione-S-transferase activity was detected at varying levels in all cell types. Nitroreductase activity was high in S9 and Chinese hamster ovary cells and lower in mouse lymphoma and Chinese hamster V79 cells.     

A new lymphocyte-specific gene which encodes a putative Ca2+-binding protein is not expressed in transformed T lymphocyte lines

The LSP1 gene is a new lymphocyte-specific gene which is expressed in normal mouse B and T lymphocytes and in transformed B cells but not (or in much smaller amounts) in nine T lymphoma lines tested. No LSP1 mRNA is found in myeloid cells or in liver, kidney, or heart tissue. Inspection of the predicted LSP1 protein sequence reveals the presence of two putative Ca2+-binding domains in the LSP1 protein. Southern blotting analysis of genomic DNA from mouse liver suggests that the LSP1 gene is present as one copy per haploid genome. Similar analysis of genomic DNA extracted from three transformed B cell lines and five transformed T cell lines shows that the absence of LSP1 mRNA in T cell lines is not due to deletion or gross rearrangements of the LSP1 locus. With the use of the mouse LSP1 cDNA as a probe we can detect a cross-hybridizing RNA species in four normal human functional T cell lines but not in three transformed human T cell lines. This suggests that at least part of the DNA sequence and the expression pattern of the LSP1 gene is conserved between mouse and man. These conserved features, together with the particular expression pattern and the protein sequence homologies, suggest that the LSP1 protein is involved in a Ca2+-dependent aspect of normal T cell growth.     

Biosynthetic labelling of membrane lipids of eukaryotic cells in tissue culture by a novel type of fluorescent fatty acids.

W-Anthryl labelled fatty acids with hydrocarbon chains of different lengths (C8, C11, C15) and different degrees of unsaturation have been incorporated into the membrane lipids of three different cell lines in tissue culture by addition of these 3H-labelled precursor fatty acids to the growth medium. The cell lines were baby hamster kidney cells (BHK 21), Chang liver cells and the RN6 cell line derived from a chemically induced Schwannoma tumor cell clone. Cell growth was normal. The quantitative analysis on the basis of radioactivity determinations demonstrated that the fluorescent-labelled fatty acids were introduced into the neutral lipid fraction (triglycerides, diglycerides, and cholesterol esters, all present in small amounts), but mainly into the phospholipid classes phosphatidylcholine, -ethanolamine and -serine, and to a lesser extent, as N-acyl component of sphingolipids (sphingomyelins, ceramides, mono- and diglycosylceramides). Cell fractionation studies indicated that the membranes of all subcellular particles were labelled with the fluorescent probes in their lipid moieties. These w-anthryl fatty acids are the first type of fluorescent lipid precursors which can be incorporated biosynthetically in vivo into membrane lipids of eukaryotic cells. The effective incorporation of the bulky fluorescent anthryl group in the terminal position of fatty acids of different chain lengths into the complex membrane lipids of the cell gives proff of 1) their uninhibited membrane transport, 2) their activation by the acyl-CoA synthetase and 3) their substrate properties for the O- acyl and N-acyl transferases in phospho- and sphingolipid biosynthesis.     

Intercellular communication and tissue growth

Summary Epithelial cells of normal rat (adult) liver and hamster embryo in tissue culture communicate through membrane junctions: the membrane regions of cell contact are highly ion-permeable. Cancerous counterparts of these cells, cells from Morris' and Reuber's liver tumors and from x-ray-transformed embryo cultures, do not communicate under the same experimental conditions. These cells also fail to communicate with contiguous normal cells. Cancerous fibroblastic cells from a variety of tissues, including cells transformed by virus, x-radiation and chemicals, communicate as well as their normal counterparts; this is so for long- and short-term cell cultures. Communication in some fibroblastic cells is sensitive to components of blood serum: normal and transformed hamster embryo fibroblasts, which communicate when cultured in medium containing fetal calf serum, appear to lose communication in medium containing calf serum; the converse holds for hamster (adult) fibroblasts and 3T3 cells.The preceding papers of this series appeared in the Journal of Cell Biology.Trainee of the National Institutes of Health, National Cancer Institute, Grant CA 05011.     

Monoclonal antibodies to murine CD3 epsilon define distinct epitopes, one of which may interact with CD4 during T cell activation

The TCR is comprised of two variable chains that confer specificity, called alpha:beta or gamma:delta, physically associated with five different molecules that comprise the complex known as CD3. Antibodies to this complex are very useful, as they react with all T lymphocytes. A rat mAb to mouse CD3 has been prepared. It reacts with 100% of T cells in all mouse strains tested but with no other cell type. It binds to the CD3 epsilon chain. This antibody activates cloned T cell lines and normal T cells, provided suitable accessory cells and signals are present. This antibody detects a determinant similar to but not identical with those detected by two previously reported hamster anti-CD3 epsilon antibodies. This antibody fixes C efficiently, and it is thus useful for depletion of T cells from bulk populations. Activation of T cells by one of the three different anti-CD3 epsilon antibodies was inhibited by the Fab fragment of anti-CD4, similar to the effects of anti-CD4 Fab on two previously reported anti-TCR V region antibodies that bind a CD3 epsilon-associated epitope. This further defines a site involving TCR V regions and CD3 epsilon with which CD4 appears to associate during T cell activation.     

Anandamide stimulates phospholipase D activity in PC12 cells but not in NIH 3T3 fibroblasts

The endogenous cannabinoid arachidonoylethanolamide was previously reported to have no effects on the phospholipase D activity in Chinese hamster ovary cells expressing the human brain-specific cannabinoid receptor, while in mouse peritoneal cells, delta9-tetrahydrocannabinol stimulated this enzyme. In this work, arachidonoylethanolamide (0.1-1 microM) was found to stimulate the phospholipase D-mediated phospholipid hydrolysis in rat adrenal pheochromocytoma PC12 cells, but not in mouse NIH 3T3 fibroblasts. The phospholipase D-activating effects of arachidonoylethanolamide were comparable to those elicited by phorbol ester and nerve growth factor, while arachidonic acid (1 microM) had no effects. The results show that, depending on the cell type, arachidonoylethanolamide can be an activator of the phospholipase D system.     

Erythrocyte Membrane-specific Antigens common to Several Species of Rodentia

THE rare existence of organ-specific antisera transcending species differences was pointed out by Coombs¹. Similar new organ, or “erythrocyte-specific”, antisera reacting selectively to the erythrocytes of several species of Myomorpha are described here. The antisera, “anti-A”, were prepared by the method of Adachi and Furusawa²: adult guinea-pigs were sensitized with the “antigen-A” (the protein portion of dd mouse erythrocyte membrane). The detailed procedure for preparing labelled antibody, the staining method and its specific reactivity to mouse erythrocytic cells have been reported before³. Circulating erythrocytes from various animals were stained with the fluorescein isothiocyanate-labelled “anti-A”. Smeared cell preparations were fixed with acetone at ? 20° C for 20 min and incubated with the labelled antisera, which had been thoroughly absorbed with the emulsion of mouse liver and kidney to remove the cross-reacting antibodies, at 37° C for 90 min. A light microscope with a darkfield condenser was used for the fluorescent observations using Orsen's interference filter system. Clear fluorescent reactions, though of varied intensities, were observed with those animals classified as Myomorpha, that is, mouse, rat, golden hamster and Chinese hamster (Table 1, Fig. 1). Mammalian erythrocytes other than those of the guinea-pig used as the sensitized animal showed faint fluorescences. Nucleated erythrocytes of the chick, quail, turtle, toad and goldfish were negative to the reaction. The erythrocyte-specific reactivity of the labelled antisera was examined with the liver, kidney and thymus cells of the animals whose erythrocytes displayed intense fluorescences; they showed, however, no fluorescence.     

Phosphorylation of C/EBPbeta at a consensus extracellular signal-regulated kinase/glycogen synthase kinase 3 site is required for the induction of adiponectin gene expression during the differentiation of mouse fibroblasts into adipocytes

Stimulation of adipogenesis in mouse preadipocytes requires C/EBPbeta as well as activation of the MEK/extracellular signal-regulated kinase (ERK) signaling pathway. In this study, we demonstrate that phosphorylation of C/EBPbeta at a consensus ERK/glycogen synthase kinase 3 (GSK3) site regulates adiponectin gene expression during the C/EBPbeta-facilitated differentiation of mouse fibroblasts into adipocytes. First, we show that exposure of 3T3-L1 preadipocytes to insulin, dexamethasone (DEX), and isobutylmethylxanthine (MIX) leads to the phosphorylation of C/EBPbeta at threonine 188. Pretreating the cells with a MEK1-specific inhibitor (U0126) significantly attenuates this activity. Similarly, these effectors activate the phosphorylation of T188 within an ectopic C/EBPbeta overexpressed in Swiss mouse fibroblasts, and this event involves both MEK1 and GSK3 activity. We further show that expression of C/EBPbeta (p34kD LAP isoform) in Swiss mouse fibroblasts exposed to DEX, MIX, and insulin induces expression of peroxisome proliferator-activated receptor gamma (PPARgamma) and some adiponectin but that it does not activate expression of FABP4/aP2. In fact, complete conversion of these fibroblasts into lipid-laden adipocytes, which includes activation of FABP4 and adiponectin expression, requires their exposure to a potent PPARgamma ligand such as troglitazone. Expression of a mutant C/EBPbeta in which threonine 188 has been modified to alanine (C/EBPbeta T188A) can induce PPARgamma production in the mouse fibroblasts, but it is incapable of stimulating adiponectin expression in the absence or presence of troglitazone. Interestingly, replacement of T188 with aspartic acid creates a C/EBPbeta molecule (C/EBPbeta T188D) that possesses adipogenic activity similar to that of the wild-type molecule. The absence of adiponectin expression correlates with a reduced amount of C/EBPalpha in the adipocytes expressing the T188A mutant suggesting that C/EBPalpha is required for expression of adiponectin. In fact, ectopic expression of PPARgamma in C/EBPalpha-deficient fibroblasts (NIH 3T3 cells) produces a modest amount of adiponectin, whereas expression of both PPARgamma and C/EBPalpha in NIH 3T3 cells facilitates production of abundant quantities of adiponectin. These data demonstrate that phosphorylation of C/EBPbeta at a consensus ERK/GSK3 site is required for both C/EBPalpha and adiponectin gene expression during the differentiation of mouse fibroblasts into adipocytes.     

Effects of density-dependent inhibition of growth on phospholipid metabolism in monolayer cultures of animal cells.

The onset of density-dependent inhibition of growth is correlated with a change in the synthesis of the major phospholipids of either BHK-21 (hamster) and 3T3 (mouse) cells. The change consists of inhibition of phosphatidylcholine synthesis while the synthesis of phosphatidylethanolamine remains constant or increases. The incorporation of radioactive precursors reflects turnover rates, since no significant differences were observed in phospholipid composition. No major differences in fatty acyl chains of these same phospholipids were observed as a function of growth rate despite considerable individuality of relative patterns of polar lipids in both cell lines. No other density-dependent differences in phospholipid metabolism were observed which were common to both cell lines. These results suggest that major aspects of growth regulation may be metabolic rather than compositional and that regulation of the enzymes of lipid metabolism may be useful probes of effector mechanisms in growth control.     

Characteristics of somatic cell hybrids (mouse x Chinese hamster) with a different ratio of chromosome sets from the parent species. II. Thymidine kinase activity]

The activity of thymidine kinase (TK) was studied in series of somatic cell hybrids between the mouse cell line 3T3-4E (TK-) and Chinese hamster cells M-15-1 (HGPRT-). Four groups of hybrid lines with different ratio of parental chromosome sets have been investigated: 1) three lines containing one hamster and one mouse chromosome set (1 hs+1 ms); 2) one line with 2 hs+1 ms; 3) one line containing 3 hs+1 ms and 4) one line containing 1 hs+2 ms. Mixtures of extracts from the parental cells were shown to possess the expected TK activity. The calculation of the activity per cell revealed that the 1 hs+1 ms and 2 hs+1 ms hybrid lines possessed about 50% of the initial hamster cell TK activity. The decreased TK activity in these hybrids might be due either to a loss of hamster chromosomes or to some inhibitory effect of mouse genome in cells with the studied ratio of parental sets. The enzyme activity in the 3 hs+1 ms hybrid was as expected, about three times greater than that of hamster cells.     

Molecular cloning of a murine glycerol-3-phosphate acyltransferase-like protein 1 (xGPAT1)

A novel murine glycerol-3-phosphate acyltransferase-like protein 1 (named xGPAT1) has been cloned. The mouse xGPAT1 gene is located on mouse Chromosome 2, spans >19 kb, and consists of at least 23 exons. The protein is 32% identical and 72% similar to mouse mitochondrial GPAT (mtGPAT) on the amino acid level. Sequencing analysis confirmed that xGPAT1 has a 2403-bp open reading frame (ORF) that encodes an 801-amino acid protein with an estimated molecular mass of 89.1 kDa. A hydropathy plot of the deduced xGPAT1 protein showed a high degree of similarity with that of the mtGPAT protein. Using 5′-rapid amplification of cDNA ends, two alternate, untranslated exon 1 (1a and b) isoforms were obtained, generating variants xGPAT1-v1 and xGPAT1–v2. xGPAT1-v1 is expressed in mouse heart, liver, spleen, kidney and murine inner medullary collecting duct 3 (mIMCD3) cells, while xGPAT1-v2 is expressed in mouse liver, spleen, kidney, white and brown adipose tissues and 3T3-L1 pre- and post-adipocytes. xGPAT1 was distributed in the membrane fraction and showed GPAT activity when epitope-tagged xGPAT1 was expressed in Chinese hamster ovary (CHO)-K1 cells.     

Expression and amplification of cloned rat liver tyrosine aminotransferase in nonhepatic cells

A full-length cDNA for the rat liver enzyme tyrosine aminotransferase has been used to construct mammalian expression vectors by recombinant DNA techniques. These vectors, which have employed either a simian virus 40 or a Rous sarcoma virus promoter, were transfected into a variety of nonhepatic mammalian cell lines in culture. Transient expression of tyrosine aminotransferase was readily observed after transfection into monkey COS cells and mouse L cells. Stable clones that express cloned tyrosine aminotransferase have been isolated from mouse L cells, hamster Wg1a fibroblasts, and Chinese hamster ovary (CHO) cells. A vector capable of expressing both tyrosine aminotransferase and dihydrofolate reductase was stimulated to undergo amplification by treatment with methotrexate in a CHO cell line deficient in the latter enzyme. Levels of tyrosine aminotransferase as much as 50-fold higher than typically seen in glucocorticoid-induced hepatoma cells were achieved in some CHO clones by this technique. The tyrosine aminotransferase produced at these highly amplified levels appeared structurally normal and had no major harmful effects on the cells.