Nucleophosmin Is Required for DNA Integrity and p19^Arf Protein Stability^†

Generation of NPM-deficient mice.

NPM-deficient mice were generated by Lexicon Genetics Inc. (The Woodlands, TX). An embryonic stem (ES) cell clone with a single gene trap event in the NPM locus was used to generate the NPM-deficient mouse line (genetic background: C57BL/6 × 129/Ola) (43). The vector used to create this particular ES cell clone was VICTR37 (8.1 kb) containing a splicing acceptor site, an internal ribosome entry site-BGeo cassette for positive selection, and the splice donor site of the first exon of the BTK gene. The insertion was determined by genomic PCR and occurred 488 bp downstream of exon I (AACCTACAGGTGGGGTCTTTCA*GTCCGAGCAGGCGCCCTGAGG; the retroviral long terminal repeat sequence is underlined, and the asterisk indicates the insertion site).

PCR genotyping.

A PCR-based strategy was developed to distinguish wild-type and mutant NPM alleles. The following primers were used: 5′-TAGAAGCAGCGGCCTGAACACTA; 5′GATGCTGACTCCGCCTACCAGA; and 5′-GCCTTGCAAAATGGCGTTACTT (see Fig. 1A). A 336-bp fragment indicates the presence of the mutated allele, whereas a 641-bp fragment is amplified from the wild-type allele (see Fig. 1B). Trp53 genotyping was determined as previously described (19)

Histology.

For sectioning, embryos were fixed in 4% paraformaldehyde 1-phosphate-buffered saline overnight, embedded in paraffin, sectioned, and stained in hematoxylin and eosin.

Immunohistochemistry.

For bromodeoxyuridine (BrdU) experiments, pregnant females were injected intraperitoneally with 150 μg of BrdU/g of body weight and sacrificed 2 h later. The following antisera were used: anti-BrdU (1/200; Boehringer Mannheim), anti-cleaved caspase 3 (1/200; Cell Signaling), anti-NPM (NPMc) (10), anti-H2AX (1:100; Upstate), and anti-p53 (monoclonal antibody raised against mouse p53 protein, clone AI25, kindly provided by Kristian Helin) An appropriate biotinylated secondary antibody was then applied to the slides. Avidin-conjugated peroxidase (ABC and DAB kits from Vector Laboratories) was used for immunostaining.

TaqMan real-time quantitative reverse transcription-PCR assays.

Total RNA (1 μg) was reverse transcribed in a final volume of 20 ul using a commercial kit (Gibco-BRL). The quantitative PCR assays were performed following the manufacturer's specifications (PE Applied Biosystems). Primer pairs and TaqMan probes were designed by Applied Biosystems. Each sample was analyzed in triplicate.

Immunoblotting and immunoprecipitation.

Western blot experiments were performed as described (9). The primary antibodies used were: sheep polyclonal anti-p53 (1:1000; Ab7; Oncogene); rabbit polyclonal anti-p53 phospho-Ser15 (1/1,000; Cell Signaling), monoclonal anti-p21 (1/300; F5; Santa Cruz); polyclonal anti-p19^Arf (1:1,000; Abcam); monoclonal anti-NPM (NPMc; NPMa) (10); polyclonal anti-β-tubulin (1:1,000, H-235, Santa Cruz); monoclonal antivinculin (1:1,000; hVIN-1; Sigma), monoclonal anti-Flag (1:1,000; Sigma), polyclonal anti-NPM (B19) (9), and anti-H2AX (1:1,000; Upstate)

Immunofluorescence.

Immunofluorescence analysis was performed as previously described (9). The antibodies used were polyclonal anti-p19^Arf (1:200; Abcam); monoclonal anti-NPM (NPMc), monoclonal antinucleolin (1:500; MS-3; Santa Cruz); polyclonal antifibrillarin (1:300; A-16; Santa Cruz), anti-phosphorylated ATM (1:500; Rockland), and anti-γH2AX (1:300; Upstate).

Cell culture, transfection, and infection.

Mouse embryo fibroblasts (MEFs), Phoenix, and 293T cells were cultured at 37°C and 5% CO₂ in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Transient transfections were performed using the standard calcium phosphate precipitate method for Phoenix and 293T cells and FuGENE 6 (Roche) for MEFs. For infection experiments, we used a Pinco-based vector expressing either only the green fluorescent protein (GFP) or GFP-tagged wild-type NPM and mutant protein. For expression of RasV12, c-myc, and p19^Arf we used the pBabe retroviral vector carrying puromycin resistance. For interference experiments we used a lentivirus-based vector called Lentilox3.7 (31). Empty or recombinant retroviral or lentivirus vectors were transfected into the Phoenix and 293T packaging cell lines, respectively, and after 48 h the supernatants were used to infect target cells.

Cell cycle analysis.

Replicative DNA synthesis and DNA content were analyzed using bivariate flow cytometry. Experiments have been carried out as described previously (9). Briefly, cells were pulsed with BrdU 1 μM, fixed and treated with 2 M HCl for 20 min followed by 2 volumes of 0.1 M sodium borate (pH 8.5). After washing cells were incubated with anti-BrdU antibody for 1 hour, washed and incubated with fluorescein isothiocyanate-conjugated secondary antibody. Cells were counterstaining overnight with 5 μg/ml propidium iodide.

Plasmids.

For short interfering RNA we used a lentivirus-based vector (Lentilox 3.7), in which we cloned (HpaI-XhoI cloning sites) the following double-stranded oligonucleotides. For NPM interference we used two different double-standard oligonucleotides that targeted two different regions of NPM mRNA (for region 1, FW 5′TggctgacaaagactatcacTTCAAGAGAgtgatagtctttgtcagccTTTTTTC and REV 5′TCGAGAAAAAAggctgacaaagactatcacTCTCTTGAAgtgatagtctttgtcagccA; for region 2, FW 5′TgagcaccagttgtcattaaTTCCAAGAGAttaatgacaactggtgcAcTTTTTTC and REV 5′TCGAGAAAAAAgagcaccagttgtcattaaTCTCTTGGAAttaatgacaactggtgcAcA). Lowercase letters represent the region that matches with the sequence of the targeted gene, and uppercase letters represent the common part of the oligonucleotides required for cloning and hairpin formation.

Analysis of newly synthesized rRNA.

Experiments have been performed as previously described (2). Briefly, cells were labeled with 2.5 uCi of [³H]uridine per ml in complete medium for 30 min, washed twice, and then incubated in complete medium without the radioactive precursor for 2 h. Total RNA was isolated with Trizol (Invitrogen), quantitated, and loaded into 1% agarose gels containing formaldehyde. After electrophoretic separation the RNA was transferred to Hybond N⁺ membranes (Amersham), and dried membranes were treated with En³Hance (Perkin Elmer) and then subjected to autoradiography.

Yolk sac cell preparation.

The yolk sac was dissected from 10.5-day embryos, rinsed to remove maternal blood, and incubated with 0,1% collagenase (Sigma) in 20% fetal bovine serum for 45 min at 37°C. Dispersed yolk sac was drawn through a 70-um strainer and washed with Iscove's minimal medium plus 10% fetal bovine serum. The cell suspension was counted and seeded in semisolid medium 3236 (StemCell Technologies) supplemented with 15% fetal bovine serum, 100 ng/ml stem cell factor, 5 ng/ml interleukin-3 and interleukin-6. After 10 days colonies can be counted and collected cells are replated. After the first replating the cell population have been cultured both in semisolid or liquid (Dulbecco's modified Eagle's medium) medium.

Mid-stage embryonic lethality of NPM^−/− embryos.

An ES cell clone with a specific gene trap event in the NPM locus was microinjected into host blastocysts to produce an NPM gene trap mouse line (43). The insertion gene trap vector, based on Moloney murine leukemia virus, integrated between exon 1 and exon 2 of the NPM locus (Fig. 1A). Mice heterozygous for this mutation were viable and fertile. However, matings between heterozygous animals produced no viable homozygous mutant offspring (F₂ in Table 1), indicating a recessive lethal phenotype (see Fig. 1B for representative genotypes and Western blots of wild-type and mutated NPM tissues).

Closer inspection revealed that all homozygous embryos (NPM^−/−) died in utero between 9.5 and 10.5 days postcoitum (dpc) (Table 1) and exhibited marked growth deficiency and gross developmental alterations (Fig. 1C). Histological examination of various embryonic tissues from NPM^−/− mutants did not reveal developmental features characteristic of their embryonic stage, suggesting that the growth defect is due to a developmental arrest (Fig. 1C, representative sections of neuroepithelium from wild-type and null 10.5 dpc embryos; see also Fig. S1 in the supplemental material). No viable mutants were observed beyond embryonic day E10.5 (Table 1). Consistent with an overall impact of the mutation throughout the whole embryo at 9.5 to 10.5 dpc, NPM was highly and ubiquitously expressed at this developmental stage, as demonstrated by immunohistochemistry (Fig. 1C). The placental structures of NPM^−/− embryos were smaller but normally developed (not shown).

p53-dependent apoptosis in the NPM^−/− embryos.

To investigate the cellular basis of the embryonic lethality associated with the NPM mutation, we investigated the effect of NPM expression on apoptosis by measuring in situ activation of caspase-3 by immunohistochemistry. Caspase-3-positive cells were only occasionally observed in wild-type embryos, as expected. In the mutant embryos, instead, caspase-3-positive cells were present at a high proportion in all the tissues examined (Fig. 1D for representative results). To investigate the molecular basis for the increased apoptosis in the NPM-null embryos, we measured levels of p53 expression by immunohistochemistry. Anti-p53 staining was low or undetectable in the wild-type embryos, while a large number of p53-positive cells were found in all NPM^−/− embryonic tissues (Fig. 1E for representative results). Quantitative PCR and Western blotting analysis of a number of p53 target genes (p21, Mdm2, pig8, and Bax) confirmed the transcriptional competence of stabilized p53 in the knockout embryo (Fig. 1F and G). It appears, therefore, that p53 is constitutively activated in NPM^−/− embryonic cells.

To investigate whether p53 activation is responsible for the observed mid-stage embryonic lethality, widespread embryo apoptosis, and in vitro MEF senescence, the NPM mutation was transferred into a p53-null background (19). Matings between NPM^+/− and p53^+/− or p53^−/− mice produced no viable doubly homozygous mutant offspring, indicating that the absence of p53 does not rescue embryonic lethality of the NPM mutation (Table 2). Analysis of embryo development revealed the presence of double-knockout embryos up to 12.5 dpc, a stage when all NPM^−/− embryos have been readsorbed (Table 3). Moreover, the NPM^−/− p53^−/− (double-knockout) embryos were bigger and more developed than NPM^−/− embryos at the same developmental stage and more similar to their p53^−/− controls (compare Fig. 2A with Fig. 1C). Analysis of anti-caspase-3-positive cells revealed a marked reduction of apoptotic cells in the double-knockout mutant compared to NPM^−/− embryos (Fig. 2B). It appears, therefore, that loss of p53 has a moderate effect on the mid-stage embryonic lethality associated with mutation of NPM, while it rescues the apoptosis observed in vivo.

NPM^−/− cells can be grown in vitro only if p53 is lost (double-knockout MEFS) or functionally inhibited (yolk sac cells).

We then tried to propagate in vitro fibroblasts from NPM^−/− and wild-type embryos (MEFs). Upon seeding, the fibroblasts adhered to the culture dishes, did not proliferate, and acquired a senescence-like phenotype (not shown). This experiment was performed several times using MEF preparations from NPM^−/− embryos at different stages (9, 10, and 10.5 dpc), and always gave the same result.

To investigate whether the growth arrest observed in NPM^−/− MEFs is the direct consequence of NPM deficiency, we down-regulated expression of NPM in growing wild-type MEFs using lentivirus vectors expressing NPM-specific short hairpin RNAs (siNPM). Expression of siNPM led to the almost complete disappearance of the NPM protein in about 70% of the infected cells, as evaluated by immunofluorescence analysis (Fig. 2C). Western blot analysis of the same cell population revealed a consistent decrease in cell proliferation of fibroblast infected with the siNPM lentivirus (Fig. 2C). Analysis of BrdU incorporation performed in the same experiment (day 3) showed a dramatic decrease of BrdU-positive cells in the MEFs infected with the siNPM lentivirus compared to control cultures (from 27 to7% in the experiment shown in Fig. 2C). Comparable results were obtained using lentivirus vectors that expressed different NPM-specific short hairpin RNAs (not shown; see Materials and Methods). It appears, therefore, that depletion of NPM expression in wild-type MEFs induces cell cycle arrest, suggesting that the growth arrest observed with NPM^−/− MEFs is a cell-autonomous effect due to NPM deficiency.

Notably, wild-type MEFs infected with siNPM lentiviruses expressed moderately higher protein levels of the p53 target p21 compared to control cultures (Fig. 2C), suggesting that the cell cycle arrest observed after NPM depletion is due to p53 activation. This hypothesis was confirmed by the finding that MEFs derived from double-knockout embryos did not undergo growth arrest in culture and showed unlimited growth potential (as evaluated by the 3T3 assay; not shown). The growth properties of double-knockout MEFs are described below (Fig. 4B).

Activation of p53 is antagonized by cytokines, as shown by the finding that hematopoietic precursors grown in vitro in the presence of cytokines are resistant to p53-induced apoptosis (30). Therefore, we investigated whether NPM^−/− hematopoietic cells can be propagated in vitro. Single-cell suspensions from E9.5 to E10.5 yolk sacs were plated in methylcellulose in the presence of cytokines and macroscopic colonies (CFC) were scored after 14 days. Thereafter, single-cell suspensions from pooled primary CFC colonies were replated into secondary CFC cultures or grown in liquid cultures.

As we had hypothesized, NPM^−/− yolk sac cells were able to grow under both culture conditions (Fig. 2D, left panel, and data not shown). On the contrary, 24 h of cytokine deprivation provoked a marked reduction of the ability of NPM^−/− cells to form CFCs, while had no effect on wild-type cells (Fig. 2D, right panel). Reexpression of NPM in cytokine-deprived NPM^−/− cells restored the normal levels of CFC colonies (Fig. 2D). It appears, therefore, that NPM^−/− yolk sac cells can be propagated in vitro only if grown in the presence of cytokines.

Analysis of the number of CFC colonies, however, showed that the efficiency of colony formation and growth rate of NPM^−/− yolk sac cells (in the presence of cytokines) was slightly, yet consistently, lower than that of control wild-type cells (Fig. 2D, left panel, and data not shown); this lower proliferation rate correlated with a moderate degree of p53 activation in the same cells. In fact, analysis of the status of p53 activity revealed a slight increase of p53 phosphorylation (Ser 18) and p21 expression in the NPM^−/− cultures (Fig. 2D, middle panel). Notably, reexpression of NPM into NPM^−/− hematopoietic precursors (by retroviral gene transfer) reduced levels of p53 phosphorylation and p21 expression and restored the efficiency of colony formation (Fig. 2D). Together, these findings indicate that NPM^−/− cells can be grown in culture only if p53 is lost, as in the p53^−/− NPM^−/− MEFs, or functionally inhibited, as occurs with NPM^−/− yolk sac cells when grown in the presence of cytokines.

p53 activation in NPM^−/− cells is part of a DNA damage response.

We next investigated the mechanisms responsible for p53 activation in NPM^−/− cells. It has been proposed that NPM inhibits p53-dependent Arf activity by targeting it to nucleoli, thus segregating Arf from p53 (20). To test whether activation of p53 in NPM^−/− cells is due to Arf activation, we down-regulated expression of NPM in Arf^−/− MEFs, using lentivirus vectors expressing NPM-specific short hairpin RNAs. Down-regulation of NPM expression in Arf^−/− cells (Fig. 3A) led to increased p53 stabilization (Fig. 3A), accumulation of p21 (Fig. 3A), and decreased cellular growth (Fig. 3B), demonstrating that Arf is not critical for activation of p53 following NPM depletion.

Activation of p53 is part of a checkpoint response to DNA damage, which involves activation of ATM/ATR kinases, phosphorylation of a set of downstream targets, including histone H2AX, and their accumulation at sites of DNA double-strand breaks or stalled replication forks (14). Western blotting analysis of H2AX phosphorylation (γ-H2AX) in total embryo lysates revealed higher level of γ-H2AX in the NPM^−/− samples compared to wild-type samples (Fig. 3C). immunohistochemistry analysis of various embryo tissues confirmed markedly increased γ-H2AX focus formation in the NPM^−/− samples (Fig. 3D for representative sections). Likewise, Western blotting (Fig. 3C) and immunofluorescence (Fig. 3E) experiments revealed increased levels of γ-H2AX and γ-H2AX focus formation in cultured NPM^−/− yolk cells compared to the wild-type samples. Staining of yolk sac cells with antibodies against phosphorylated ATM revealed the presence of phosphorylated ATM foci in the NPM^−/− but not wild-type cells, which largely colocalized with the γ-H2AX foci (Fig. 3F). Notably, reexpression of NPM in NPM^−/− yolk cells restored levels of γ-H2AX and phosphorylated ATM foci to the background levels of wild-type cells (Fig. 3F and 3C). It appears, therefore, that loss of NPM leads to the constitutive activation of a DNA damage response, suggesting that p53 activation in NPM^−/− cells is secondary to the loss of a checkpoint function of NPM on DNA integrity.

Increased cell proliferation in NPM^−/− embryos and double-knockout MEFs.

We then investigated the effects of NPM expression on cellular proliferation by measuring BrdU incorporation in wild-type and NPM^−/− embryos and comparing the growth properties of double-knockout and p53^−/− MEFs. To analyze BrdU incorporation in embryos, 10.5 dpc pregnant female mice were intraperitoneally injected with BrdU and sacrificed 45 min later. The recovered embryos were analyzed for BrdU incorporation in situ by immunohistochemistry.

Microscopic examination of embryo sections from different tissues revealed a random dispersion of BrdU-positive cells in the mutant tissues, at variance with their normal counterparts, where cycling and resting cells typically cluster within separate areas (Fig. 4A for representative sections from neuroepithelium and dorsal root ganglia). Analysis of the frequency of BrdU-positive cells revealed a slightly yet consistently higher number of cycling cells in the NPM^−/− samples than in the wild-type counterparts (Fig. 4A). To obtain a quantitative estimate of BrdU positivity, we performed a FACS analysis of cell suspensions obtained from wild-type and NPM^−/− embryos of different stage (E9.5 and 10.5 dpc). The results showed an approximately 45% increase of BrdU-positive cells in NPM^−/− embryos (see Fig. 4A for representative results obtained from 10.5 dpc embryos). Together, these data demonstrate that NPM expression is not critical for cell cycle progression, as previously hypothesized (15), and that, on the contrary, it negatively regulates progression through the cell cycle.

Comparison of the growth properties of double-knockout and p53^−/− MEFs revealed that double-knockout fibroblasts grow markedly faster than p53^−/− controls (see Fig. 4B for one representative curve). BrdU incorporation analysis performed on logarithmically growing cell population revealed an approximately 35% increase of BrdU-positive cells in the double-knockout cells (Fig. 4B). It appears, therefore, that loss of NPM expression correlates with increased proliferation both in vivo and in cultured cells. In the latter, however, increased proliferation can only be observed if p53 is lost, since the loss of NPM is also responsible for the activation of a p53-dependent checkpoint. Whether, in the NPM^−/− fibroblasts, the increased proliferation is responsible for the accumulation of DNA replication alterations, which then activate a (p53-dependent) intra-S-phase checkpoint (1) remains, however, to be determined.

NPM is critical for the nucleolar localization and stability of Arf protein.

Arf inhibits cell proliferation through both p53-dependent and -independent mechanisms (34). Since NPM physically interacts with Arf and protects it from degradation (20), the inhibitory effect of NPM on cellular proliferation might be mediated by Arf itself. Therefore, we investigated the effects of NPM on Arf expression. Immunofluorescence analysis of p53^−/− and p53^−/− NPM^−/− (double-knockout) MEFs revealed that Arf, in the absence of NPM, is nuclear-diffuse and completely excluded from nucleoli, which remain as intact organelles, as shown by staining with other nucleolar markers, such as nucleolin (Fig. 5A) and fibrillarin (not shown). Notably, Arf levels in the double-knockout MEFs were markedly lower than in the control p53^−/− MEFs (Fig. 5B, left panel). Reexpression of NPM in double-knockout cells restored normal levels (Fig. 5B), compared to p53^−/− cells, and nucleolar localization of Arf (Fig. 5B).

Since levels of arf mRNA were comparable in the p53^−/− and double-knockout MEFs (Fig. 5C, left panel), we investigated whether NPM regulates Arf stability. To this end, p53^−/− and double-knockout cells were exposed to 100 μM cycloheximide to inhibit de novo protein synthesis, harvested at serial time points following treatment, and assessed for Arf levels by Western blotting. The half-life of Arf in p53^−/− MEFs was approximately 12 h (Fig. 5C, right panel). In the double-knockout cells the stability of the Arf protein was markedly reduced, with a half-life of about 2 to 4 h (Fig. 5C, right panel). Therefore, NPM is indispensable for the nucleolar localization of Arf and critical for its stability. These data are consistent with previous reports showing destabilization and partial delocalization of Arf following down-regulation of NPM expression (20).

NPM is not a downstream target of Arf.

Upon activation by abnormally high thresholds of mitogenic signals, such as those imposed by oncogenic Myc or Ras, Arf activates p53 via Mdm2, enabling cells to stop proliferating (35). Arf can inhibit ribogenesis (36) and, though less efficiently, suppress proliferation of cells that lack both p53 and Mdm2, implying interaction with other regulators (39). Due to its reported properties of stimulating the cell cycle and ribogenesis, NPM has been proposed as a critical downstream target of Arf (18). Upon activation, Arf would bind NPM directly, thus inducing its degradation or blocking its nucleus-cytoplasm shuttling activity (7). Therefore, we evaluated the effects of NPM on the ability of Arf to inhibit cellular growth and ribosome biogenesis in p53^−/− cells.

p53^−/− and double-knockout MEFs infected with control or Arf-expressing retroviruses were analyzed for proliferation capacity or labeled with [³H]uridine in order to evaluate newly synthesized rRNA. Exogenously expressed Arf localized predominantly within nucleoli in the p53^−/− MEFs, as expected (not shown), while it was nuclear diffuse and mainly excluded from nucleoli in the double-knockout cells (Fig. 5D). Consistent with the observed effect of NPM on Arf stability, levels of exogenous Arf were lower in double-knockout than in p53^−/− cells (Fig. 4D). The ability of overexpressed Arf to reduce cell proliferation was comparable in p53^−/− and double-knockout MEFs (Fig. 5D). Notably, the proliferation rate of the various MEF populations (controls and p19^Arf infected) nicely correlated with the levels of p19^Arf (Fig. 4D). As an example, p19^Arf-infected double-knockout cells and p53^−/− control cells expressed comparable levels of p19^Arf and showed similar growth rates (Fig. 5D). Finally, the ability of overexpressed Arf to reduce maturation of the 28S rRNA was comparable in p53^−/− and double-knockout MEFs (Fig. 5E). The effect of NPM expression on rRNA processing was also negligible under basal conditions (Fig. 5E). It appears, therefore, that expression of NPM is not rate limiting for rRNA processing and that the ability of ectopically overexpressed Arf to inhibit ribogenesis and cellular proliferation (in the absence of p53) is independent of NPM and its nucleolar localization.

NPM regulates the biological activities of Arf.

Mitogenic stimuli, as imposed by enforced expression of c-Myc or the RasV12 mutant, induce expression of Arf, which leads to p53 stabilization. The absence of p53 and/or Arf expression suppresses this phenotype and increases the chances of cellular transformation (12). Moreover, Arf inhibits Myc through direct binding, independently of its function on p53 (29). Therefore, to investigate the functions of Arf that are p53 independent, we ectopically expressed c-Myc in p53^−/− and double-knockout MEFs. Expression of Myc in p53^−/− MEFs increased the rate of cell proliferation (Fig. 6A) and induced their oncogenic conversion, as measured by the ability of Myc-expressing cells to form colonies in semisolid medium (Fig. 6A). Comparable levels of overexpressed Myc (Fig. 6A) provoked, in the double-knockout MEFs, a higher proliferation rate and increased frequency of transformation (Fig. 6A). Notably, the size of methylcellulose colonies formed by the double-knockout cells was considerably larger than that of the controls (Fig. 6A). Comparable results, in terms of transforming ability and colony growth, have been obtained by infecting the same cells with RasV12 (Fig. 5B). In the cases of both RasV12 and Myc, oncogene-induced Arf stabilization was modest in the double-knockout MEFs, and p19^Arf levels remain markedly lower than in the p53^−/− MEFs.

Taken together, these results indicate that, in the absence of NPM, the p53-independent effects of Arf on cellular proliferation are less potent, either under basal conditions or after oncogenic stimuli, likely due to reduced stability of the Arf protein.