Mechanisms of chromosomal aberration production II. Aberrations induced by 5-bromodeoxyuridine and visible light

We have allowed synchronized V79B Chinese hamster tissue culture cells to incorporate 5-bromodeoxyuridine (BUdR) during one DNA synthetic (S) period of the cell cycle and then determined chromosomal aberration yields induced by illumination of the cells with visible light during the succeeding pre- and post-DNA-synthetic (G₁and G₂) phases of the cell cycle. At the level used, BUdR by itself induces no aberrations. Illumination during the G₁ phase following incorporation induces aberrations of the chromatid type, but none of the chromosome type. All types of chromatid aberrations are induced, including isochromatid deletions and exchange types. In contrast, when cells are illuminated during the immediately following G₂ phase, large numbers of achromatic lesions and chromatic deletions are seen at the first post-illumination mitosis, but no isochromatid deletions and few exchange-type aberrations occur. When G₂-illuminated cells are examined in their second mitosis, however, chromatid aberrations of all types are again seen.

These results are interpreted within the “repair” model of chromosomal aberration production by UV light presented earlier³. The model assumes that the vertebrate chromosome is mononeme, consisting of but a single DNA double helix during the prereplication G₁ phase. The initial lesions induced by illumination of BUdR-containing DNA are believed to be single-chain breaks, and the observation that G₁ illumination produces only chromatid-type aberrations is taken as additional evidence for the mononeme chromosome. Conversion of single-chain breaks into double chain breaks through the action of a single-strand nuclease is postulated to account for the production of chromatid deletions at the first mitosis of G₂-illuminated cells. The action of this enzyme, plus a recombinational or post-replication repair mechanism, are postulated to account for the production of isochromatid deletions in G₁-illuminated cells. A rapid decline in achromatic lesion frequency with increasing time between G₂ illumination and fixation of the cells is considered evidence for rapid rejoining of most of the initial chain breaks.     

Mechanisms of chromosomal aberration production I. Aberration induction by ultraviolet light

We have studied chromosomal aberration production in V-79 Chinese hamster tissue culture cells by UV light administered during the post-DNA-synthetic G₂ phase of the cell cycle. The treatment produced achromatic lesions and some chromatid deletions in the first post-irradiation mitosis, but no isochromatid deletions or chromatid exchange aberrations. In contrast, when G₂ UV-irradiated cells were examined in their second post-irradiation mitosis, there were significant yields of chromatid-type aberrations of all types, including isochromatid deletions and chromatid exchanges.

We have earlier reported²¹ that UV-irradiation during the pre-DNA-synthetic G₁ phase of the cell cycle induces only chromatid aberrations and also that most chromosomal aberration production by UV in G₁ can be photoreactivated in cells possessing the photoreactivating enzyme. We present here a model for chromosomal aberration production by UV. In the model all aberration production is enzymatically mediated, a consequence of the functioning of known molecular repair mechanisms. The important elements in the model are the following:

1. (1) The vertebrate chromosome is mononeme; i.e., contains but a single DNA double helix during the prereplication G₁ phase of the cell cycle.

2. (2) The UV-induced DNA lesion leading to the production of most aberrations is the cyclobutane dimer between adjacent pyrimidines in one polynucleotide strand.

3. (3) Single chain breaks appear at metaphase as achromatic lesions.

4. (4) Dimer removal sometimes leaves unrepaired single chain gaps, possibly as a result of incomplete excision repair.

5. (5) The single-stranded DNA opposite a single chain gap can be cleaved by a single-strand DNAase.

6. (6) Gaps are left in newly synthesized DNA polynucleotide chains opposite defective template chains (i.e., opposite dimers and chain breaks).

7. (7) Double-strand breaks present following local DNA replication may “spread” to the other chromatid by a recombinational process between template and new polynucleotide chains, one from each of the homologous double helices.

The model predicts the occurrence of isoachromatic lesions and of chromatid deletions paired (isolocus) with achromatic lesions. Though often not reported, both do, in fact, occur. In addition, the model accounts for the phenomenon of sister-chromatid exchange as a manifestation of a recombinational, or post-replication, repair mechanism. Finally, the model offers a simple interpretation of chromosomal aberration production by a variety of chemical agents.     

Dynamcis of the thymidine triphosphate pool during the cell cycle of synchronized 3T3 mouse fibroblasts

To investigate whether resting cells of 3T3 mouse fibroblasts carry out de novo synthesis of deoxyribonucleoside triphosphates, we determined the turnover of the thymidine triphosphate pool of G₀ cells obtained by starvation of cultures for platelet-derived growth factor. These cells were contaminated by less than 1% S-phase cells. In the absence of deoxyribonucleosides in the medium one million G₀ cells contained 5 pmole of dTTP with a turnover of 0.09 pmole/min. S-phase cells in comparison contained a 20 times larger dTTP pool with a more than 200-fold faster turnover. Our results suggest that G₀ cells carry out a slow but finite de novo synthesis of deoxyribonucleoside triphosphates to satisfy the cells' requirement for DNA repair and mitochondrial DNA synthesis.     

DNA polymerase delta mediates increase in exchange production by X-radiation in human lymphocytes moving from G0 to G1

Earlier work of several laboratories established that the yields of radiation-induced ring and dicentric chromosomes are greater when human peripheral blood lymphocytes are irradiated in GH₁ some hours after phytohemagglutinin stimulation than if they are irradiated in G₀ before stimulation. Post-treatment of lymphocytes irradiated in G₀ with the DNA polymerase inhibitor aphidicolin, which is effective against both pol and pol δ, produces a similar increase in ring and dicentric yield. We found that aphidicolin post-treatment was much less effective in increasing ring and dicentric yield increases in cells irradiated in G₁ four to five hours after stimulation. Because we had earlier found specific inhibitors of DNA pol ineffective in producing increased yields in either G₀ or G₁ lymphocytes, we conclude that much of the G₀ to G₁ increase in yields is mediated by pol δ.     

Increased DNA polymerase and ATP-dependent deoxyribonuclease activities following DNA damage in Mycobacterium smegmatis

Summary Treatment of growing cultures of Mycobacterium smegmatis with alkylating agents (methyl methaneusulphonate, ethyl methanesulphonate, nitrogen mustard, or mitomycin C) or with ultraviolet light resulted in enhanced specific activities of a DNA polymerase and of an ATP-dependent deoxyribonuclease. Similar results had previously been obtained with hydroxyurea and with iron limitation. The three of these treatments which were tested (methyl methane-sulphonate, mitomycin C and hydroxyurea) produced strand breaks or alkali-labile regions in the DNA of this organism. The increased enzyme activities could be prevented by simultaneous treatment with inhibitors of protein synthesis.In contrast, treatment of the cultures with intercalating agents (ethidium bromide, acridine orange, or proflavine), 5-fluorouracil, caffeine, or nalidixic acid, inhibited DNA synthesis without increasing the enzyme activities. These treatments did not produce strand breaks in the DNA of this organism.The results support the hypothesis that, in M. smegmatis, damage to DNA induces increased synthesis of enzymes associated with DNA repair.     

Static Cytofluorometry and Fluorescence Morphology of Mitochondria and DNA in Proliferating Fibroblasts

The shape, distribution, and content of mitochondria in individual cells were examined during the cell cycle phases (G₀/G₁, S, G₂ mitosis) in living human fibroblasts by static cytofluorometry and fluorescence microscopy. The morphocytochemical evaluations were performed in cell cultures submitted to double supravital fluorochrome staining with Hoechst 33342 and DiOC₆ to label DNA and mitochondria, respectively. The staining modalities were based on the stability of mitochondrial labeling. The G₁ to early S phases were characterized by the presence of filamentous mitochondria, except during the early postmitotic period. During late S, G₂, and mitotic phases, mitochondrial mass reached its highest value and mitochondria became short and numerous. During the last stage of mitosis, mitochondria were distributed among daughter cells through a cytoplasmic bridge.     

Potentiating effects of organomercuries on clastogen-induced chromosome aberrations in cultured Chinese hamster cells

Mercury compounds are among the most serious environmental pollutants. In this communication, the potentiating effects of organic and inorganic mercuries on clastogen-induced chromosome aberrations were studied in Chinese hamster CHO K1 cells. Post-treatment with monoalkylated mercuries — methyl mercuric chloride (MeHgCl) and ethyl mercuric chloride (EtHgCl) - increased the number of breakage-and exchange-type aberrations induced by 4-nitroquinoline 1-oxide (4NQO) and methyl methanesulfonate. With the DNA crosslinking agents mitomycin C (MMC) and cisplatin, MeHgCl enhanced both types of aberrations while EtHgCl enhanced breakage-type aberrations only. Since these monoalkylated mercuries did not show clastogenic effects by themselves under the present experimental conditions, the increases in chromosome aberrations were not additive. Dialkylated mercuries (dimethyl mercury and diethyl mercury) and inorganic mercuries (HgCl and HgCl₂) did not show any potentiating effects.

When MMC- or 4NQO-treated cells were post-treated with MeHgCl during the G1 phase, both breakage- and exchange-type aberrations were enhanced. Treatment with EtHgCl during the G1 phase also enhanced both types of aberrations induced by 4NQO. With MMC, however, G1 treatment with EtHgCl did not show any potentiating effect. MeHgCl and EtHgCl treatments during the G2 phase enhanced breakage-type aberrations only.

Based on these results, the following possible mechanisms for potentiation of clastogenicity by monoalkylated mercuries were suggested; (1) they interfere with repair of base lesions induced by 4NQO and MMS during the pre-replicational stage, thereby increasing unrepaired DNA lesions which convert into DNA double-strand breaks in S phase, (2) MeHgCl (but not EtHgCl) also inhibits repair of crosslinking lesions during the pre-replicational stage, and (3) their G2 effects enhance breakage-type aberrations only.     

Targeted deletion of mouse Rad1 leads to deficient cellular DNA damage responses

The Rad1 gene is evolutionarily conserved from yeast to human. The fission yeast Schizosaccharomyces pombeRad1 ortholog promotes cell survival against DNA damage and is required for G₂/M checkpoint activation. In this study, mouse embryonic stem (ES) cells with a targeted deletion of Mrad1, the mouse ortholog of this gene, were created to evaluate its function in mammalian cells. Mrad1^-/- ES cells were highly sensitive to ultraviolet-light (UV light), hydroxyurea (HU) and gamma rays, and were defective in G₂/M as well as S/M checkpoints. These data indicated that Mrad1 is required for repairing DNA lesions induced by UV-light, HU and gamma rays, and for mediating G₂/M and S/M checkpoint controls. We further demonstrated that Mrad1 plays an important role in homologous recombination repair (HRR) in ES cells, but a minor HRR role in differentiated mouse cells.     

Circadian and Circannual Variations of Cell Cycle Distribution in the Mouse Bone Marrow

The cell cycle distribution of bone marrow cells from the femurs of female C3H mice has been investigated by flow cytometry according to the time of the day and month of the year. Both circadian and seasonal variations were found for the different cell cycle phases as well as the total cell numbers per femur. Both the mesor, the acrophase and the amplitude of the S, G₂ and (G₁ + G₀) phases varied significantly in some months, while in other months only insignificant rhythms were found. The relative cell cycle distribution only partly reflected variations in the total numbers of proliferating cells, since the total cell number per femur was also variable.

The total numbers of cells in DNA synthesis seem to be higher in the first part of the year, indicating increased cell proliferation during winter and spring. In this period the acrophases of DNA synthesis and G₂ were in the morning, while the second half of the year showed the peak later in the day.

In general, hemopoietic cell proliferation seems to constitute a labile equilibrium with rapidly changing activities.     

Inter- and intra-chromosomal distribution of chromatid breaks induced by X-rays during G2 in human lymphocytes

Cultures of blood from healthy adults were irradiated 48 h after stimulation with 240 R of X-rays and fixed after various time intervals (0–2 h, 2–4 h, 4–6 h). ³HTdR was added to several cultures after irradiation. Mitotic and labelling indices were used to distinguish between two cell samples inside the irradiated G₂ population: D − cells reaching mitosis without mitotic delay and a high frequency of chromatic breaks and D + cells with mitotic delay and which, during the delay, repair most of the damage produced. After R banding 450 chromatid deletions were located in each of the two cell samples. The D + cells showed a higher frequency of breaks than the D − cells with decreasing chromosome size, in the telomeric and centromeric region and in the junction between the R + and R − bands. These results can be interpreted as indicative of a non-random distribution of repair processes both between and within chromosomes.     

X-ray induction of 6-thioguanine-resistant mutants in division arrested, G0/G1 phase Chinese hamster ovary cells

The cytotoxic and mutagenic effect of X-irradiation was determined with Chinese hamster ovary cells arrested in the G₀/G₁ phase of the cell cycle through 9 days incubation in serum-free medium. In comparison with exponential phase cultures, the arrested cells showed increased cytotoxicity and mutation induction over the dose range of 50–800 rad. Exponential cultures showed a linear mutant frequency-survival relationship while the arrested cells showed a biphasic linear relationship. A post irradiation holding period of 24 h does not result in any change in the mutant frequency. The increased sensitivity of the arrested cells to the mutagenic effects of X-rays appears to be a cell-cycle phase phenomenon. Upon readdition of serum, the arrested cells re-enter the cell cycle in a synchronous manner, reaching S phase at 10–12 h. Cells irradiated at 5 h after serum addition, i.e. in G₁, show a similar does response for mutant frequency, while those irradiated at 10 h or later, i.e. in late G₁, S or G₂, show lower mutation induction. These observations are consistent with a chromosome interchange mechanism of mutation induction by X-rays, possibly through interactions between repairing regions of the DNA. Irradiation of cells in the G₀/G₁ phase allow more time for such interactions in the absence of semiconservative DNA replication.     

Nitric oxide blocks the cell cycle of mouse macrophage-like cells in the early G2+M phase

The effects of nitric oxide produced by macrophage-like cells (Mml) on the cell cycle were investigated. Mml cells lost proliferative activity in the presence of interleukin-6 (IL-6) and a subpopulation accumulated in the G₂+ M phase. This level increased in proportion to the incubation time. The DNA content of the cells was slightly lower than that of Mml cells treated with vinbrastine or demecolcine, drugs which block the cell cycle in the M phase. The peak of the early G₂+M phase was reduced by treatment with N^G-mono-methyl- -arginine. However, after treatment with exogenous nitric oxide or sodium nitroprusside, the G₀/G₁ phase increased, but the early-G₂+M and the S phase decreased. The flow cytometry pattern in IL-6-treated Mml was the same as that of cytochalasin B-treated Mml. These data suggest that endogenous nitric oxide affects the microfilament system of IL-6-treated Mml cells and blocks the cell cycle in the early G₂+M phase.     

Sensitivity to cell killing and the induction of cytogenetic damage following gamma irradiation in wild-type and thymidine kinase-deficient Friend mouse erythroleukaemia cells

Wild-type Friend erythroleukaemia (clone 707) cells and 2 thymidine kinase-deficient subclones, 707BUE and 707BUF, having thymidine kinase activities of 1.4% and 0.7% that of clone 707, were compared for sensitivity to killing and the induction of cytogenetic damage following irradiation. Three doses of gamma irradiation were used (150, 300 and 450 cGy), and cells were harvested for metaphase spreads after 4, 8, 12, 15, 29 and 43 h. G₂ delay was evident at 4 h following gamma irradiation in the 3 cell clones examined, and recovery of mitosis was observed to be dose-dependent. G₂ delay was found to be most prolonged in subclone 707BUE and most prompt in clone 707. Increased sensitivity to the induction of cytogenetic aberrations at all three doses was apparent in the 2 thymidine kinase-deficient subclones (as compared to wild-type cells) at 15, 29 and 43 h. Th thymidine kinase-deficient subclones also showed increased sensitivity to gamma radiation-induced cell killing. Furthermore, subclone 707BUE consistently exhibited greater to gamma irradiation than did the subclone with lower thymidine kinase activity, 707BUF. The importance of thymidine kinase levels and extended G₂ delay for DNA repair processes is discussed.     

Sensitization of human cells by inhibitors of DNA synthesis following the action of DNA-damaging agents

Inhibitors of DNA synthesis 1-β-arabinofuranosylcytosine (Ac) and hydroxyurea (Hu) taken together drastically sensitized human cells to the killing effect of DNA-damaging agents. For UV-irradiation this sensitization depended on the cells′ ability for excision repair.

By using viscoelastometric methods of measurement of double-strand breaks (DSB) in the genome, it was established that the first DSB were generated after incubation of the damaged cells in the mixture of inhibitors at about the same dose when sensitization appeared.

A scheme is proposed to described molecular events associated with the phenomenon studied.     

Molecular forms of acetylcholinesterase in developing Torpedo embryos

We have studied the evolution of acetylcholinesterase molecular forms during the embryonic development of Torpedo marmorata, in the electric organs and in the electric lobes of the central nervous system. In the early stages of development (35 mm embryos, ‘myogenic phase’ of electric organ development), globular forms of acetylcholinesterase (G₄ and G₂) are abundant in both tissues and the collagen-tailed form A₁₂ is already present. In the electric organs, this form accumulates rapidly after the 55–60 mm stage (‘electrogenic phase’), when synapse formation first commences. Although the molecular characteristics of the collagen-tailed forms, and particularly their aggregation properties, do not appear to change during development, their solubilization requires higher concentrations of MgCl₂, as the electrocytes mature, suggesting that they become more tightly integrated in a better organized basal lamina. The smaller collagen-tailed form A₈ shows a transient increase which coincides approximately with the maximal accumulation of A₁₂, suggesting that it is an intermediate in its synthesis. The accumulation of the hydrophobic G₂, which eventually becomes predominent in the adult electric organs, lags behind that of A₁₂. The functional significance of this important fraction of acetylcholinesterase is therefore not that of a pool of precursor for the synthesis of A₁₂. In the electric lobes, the tetrameric form (G₄) is abundant during development, as well as G₂ and G₁ at certain stages, but the A₁₂ form is predominant in the adult.     

Synergistic enhancement of the frequency of chromatid aberrations in cultured human lymphocytes by combinations of inhibitors of DNA repair

Whole-blood cultures of human lymphocytes were exposed in the G₂-phase to caffeine and to 4 inhibitors of DNA synthesis, hydroxyurea (HU), 2′-deoxyadenosine (dAdo), 1-β- -arabinofuranosylcytosine (araC) and aphidicolin (Aph), either individually or in pairs resulting in 10 different possible combinations. Since dAdo is rapidly deaminated in whole-blood cultures, all treatments involving dAdo were carried out in the presence of an inhibitor of the enzyme adenosine deaminase (ADA). The G₂-treatments were carried out on 3 different types of culture, (1) cultures that had not previously been exposed to any mutagenic treatment, (2) cultures that had been irradiated with 0.4 Gy of X-rays 3.5 h before harvesting, and (3) cultures that had been exposed for 2 h to 4 × 10⁻⁵ M thiotepa (TT) in G₀ immediately before stimulation with phytohaemagglutinin. The aim of the study was to find out in which combinations the inhibitors enhanced synergistically the frequencies of spontaneous and induced chromatid aberrations. In all 3 types of experiment, synergistic effects were observed with most of the 10 combinations, those involving HU being particularly effective. A very strong synergistic enhancement was also obtained when dAdo was combined with Aph.     

Circadian rhythms of DNA synthesis in nasopharyngeal carcinoma cells

Nasopharyngeal carcinoma (NPC) occurs frequently in southern China. The circadian rhythm of DNA synthesis of a poorly differentiated NPC human cell line (CNE2) was investigated as an experimental prerequisite for designing chrono-chemotherapy schedules for patients with this disease. Twenty-two nude mice with BALB/c background were synchronized alternatively in 12 h of light and 12 h of darkness (LD12:12) for at least 3 wk prior to the transplantation of a CNE2 tumor fragment into each flank (area of ∼2×2 mm²). Ten days later, a tumor sample (area of ∼5 mm²) was obtained at 3, 9, 15, and 21 h after light onset (HALO) alternatively from different sites in each mouse. Single-cell suspensions were prepared and stained with propidium iodide. Cellular DNA content was measured with flow cytometry. Data were analyzed by ANOVA and cosinor methods. The average proportion of tumor cells in G₁, S or G₂-M phase varied according to circadian time with statistical significance. The maximum occurred at 9 HALO for G1, 2 HALO for S and 21 HALO for G₂-M phase cells. The approximate average distribution patterns of G₁ and G₂-M phases of cosine curve was 24 h. This was not the case for S-phase cells, which displayed a bimodal temporal pattern. Inter-individual variability in peak time was large, possibly due to relatively sparse sampling time. Nevertheless, no more than 6% of the time series displayed a maximum at 3 HALO for G₁, 21 HALO for S and 15 HALO for G₂-M. The cell cycle distribution of this human NPC cell line displayed circadian regulation following implantation into nude mice. The mechanisms involved in this rhythm and its relevance to the chrono-chemotherapy of patients deserve further investigation.     

Differentiation is coupled to changes in the cell cycle regulatory apparatus of human embryonic stem cells

Mouse embryonic stem cells (mESC) exhibit cell cycle properties entirely distinct from those of somatic cells. Here we investigated the cell cycle characteristics of human embryonic stem cells (hESC). HESC could be sorted into populations based on the expression level of the cell surface stem cell marker GCTM-2. Compared to mESC, a significantly higher proportion of hESC (GCTM-2⁺ Oct-4⁺ cells) resided in G₁ and retained G₁-phase-specific hypophosphorylated retinoblastoma protein (pRb). We showed that suppression of traverse through G₁ is sufficient to promote hESC differentiation. Like mESC, hESC expressed cyclin E constitutively, were negative for D-type cyclins, and did not respond to CDK-4 inhibition. By contrast, cyclin A expression was periodic in hESC and coincided with S and G₂/M phase progression. FGF-2 acted solely to sustain hESC pluripotency rather than to promote cell cycle progression or inhibit apoptosis. Differentiation increased G₁-phase content, reinstated cyclin D activity, and restored the proliferative response to FGF-2. Treatment with CDK-2 inhibitor delayed hESC in G₁ and S phase, resulting in accumulation of cells with hypophosphorylated pRb, GCTM-2, and Oct-4 and, interestingly, a second pRb⁺ GCTM-2⁺ subpopulation lacking Oct-4. We discuss evidence for a G₁-specific, pRb-dependent restriction checkpoint in hESC closely associated with the regulation of pluripotency.     

The formation and repair of single-strand breaks in DNA of cultured mammalian cells treated with UV-light, methylating agents or 4-nitroquinoline-1-oxide.

The technique of sedimentation in alkaline sucrose was used to examine the formation and repair of single-strand (SS) breaks in cultured mammalian cells that were treated with methyl methanesulfonate (MMS), methyl nitrosourea (MNUA), 4-nitroquinoline-1-oxide (4NQO) or UV-light. The SS breaks induced by MMS and 4NQO were largely repaired by HeLa cells during a 5-h post-treatment incubation. The SS breaks induced by MNUA and UV-light were not repaired by HeLa cells. L-cells were not able to repair the SS breaks induced by any of the agents, which correlates with the deficiency of these cells for repair synthesis of DNA. The following conclusions are discussed. MNUA and UV-light produce modifications in DNA which are not repaired but are translated into SS breaks in alkali. MMS produces SS breaks intracellularly but these are not derived from a simple depurination of methylated purines. 4NQO produces a modification in DNA which is translated into an SS break in alkali but which can be removed by an intracellular process.     

Nitric oxide inhibits DNA-adduct excision in nucleotide excision repair

Sustained induction of nitric oxide (NO) in chronic inflammation may be mutagenic, through DNA damage induction and/or DNA repair inhibition. Although there is good evidence that NO can cause DNA damage, how NO is involved in DNA repair remains elusive. By using DNA synthesis inhibitors to accumulate DNA strand breaks in comet assay, we show that NO and peroxynitrite inhibit DNA-adduct excision in human fibroblasts damaged by UVC, 4-nitroquinoline 1-oxide, benzo[a]pyrene dihydrodiol epoxide, cisplatin, or mitomycin C, but not with methyl methane sulfonate. Treating cells with arsenite increased NO production and also inhibited the DNA-adduct excision induced by UVC, 4-nitroquinoline 1-oxide, benzo[a]pyrene dihydrodiol epoxide, cisplatin, and mitomycin C, but not by methyl methane sulfonate, H(2)O(2), sodium nitrosoprusside, or 3-morpholinosydnonimine. Arsenite inhibition of DNA-adduct excision was decreased by NO synthase inhibitors and NO scavengers. The nuclear extract prepared from fibroblasts pretreated with sodium nitrosoprusside, dipropylenetriamine NONOate, 3-morpholinosydnonimine, or arsenite also showed decreased activity in excising the DNA adducts induced by UVC and cisplatin but not by methyl methane sulfonate or H(2)O(2) plus Fe. These results are consistent with the notion that NO, peroxynitrite, and arsenite inhibit the DNA-adduct excision in nucleotide excision repair but not that in base excision repair.