Human p14(Arf): an exquisite sensor of morphological changes and of short-lived perturbations in cell cycle and in nucleolar function

The human Ink4a/Arf tumor suppressor locus encodes two distinct products: p16(Ink4a) which prevents phosphorylation and inactivation of the retinoblastoma protein and, p14(Arf), a nucleolar protein which activates the function of the tumor suppressor p53 protein in the nucleoplasm in response to oncogenic stimulation through an as yet ill-defined mechanism. Here we show that the level of endogenous p14(Arf) and its balance between the nucleolus and the nucleoplasm in HeLa cells are exquisitely sensitive to changes in cell morphology and to short-lived perturbations in cell cycle and in nucleolar function such as those induced by the cyclin-dependent kinase inhibitor, roscovitine, and the casein kinase II and RNA synthesis inhibitor, DRB. Most remarkably, whereas p14(Arf) predominantly concentrates in the nucleolus of interphase cells and transiently disappears between metaphase and early G1 under normal growth conditions, it massively and reversibly accumulates in the nucleoplasm of postmitotic and S-phase cells upon short-term treatment with roscovitine and, at a lesser extent, DRB. In line with the fact that the nuclear level of p53 reaches a peak between mid-G1 and the G1/S border in p53-expressor cells which lack Arf expression, these results provide a clue that, in p53+/Arf+ cells, Arf proteins might serve both to speed and to amplify p53-mediated responses in conditions and cell cycle periods in which the mechanisms involved in p53 stabilization and activation are not fully operational. They further suggest that human endogenous p14(Arf) might activate p53 pathways in physiologic situations by acting inside the nucleoplasm, especially when normal cell cycle progression and nucleolar function are compromised.

Common and reversible regulation of wild-type p53 function and of ribosomal biogenesis by protein kinases in human cells

Two specific inhibitors of cyclin-dependent kinase 2 (Cdk2), roscovitine and olomoucine, have been shown recently to induce nuclear accumulation of wt p53 and nucleolar unravelling in interphase human untransformed IMR-90 and breast tumor-derived MCF-7 cells. Here, we show that the early response of MCF-7 cells to roscovitine is fully reversible since a rapid restoration of nucleolar organization followed by an induction of p21(WAF1/CIP1), a downregulation of nuclear wt p53 and normal cell cycle resumption occurs if the compound is removed after 4 h. Interestingly, similar reversible effects are also induced by the casein kinase II (CKII) inhibitor, 5,6-dichloro-1-beta-D-ribofuranosylbenzimidazole. Upon short-term treatment also, both compounds significantly, but reversibly, reduce the level of 45S precursor ribosomal RNA. Cells exposed to the two types of protein kinase inhibitors for longer times keep exhibiting altered nucleolar and wt p53 features, yet they strikingly differentiate in that most roscovitine-treated cells fail to ever accumulate high levels of p21(WAF1/CIP1) in contrast with DRB-treated ones. In both cases, however, the cells eventually fall into an irreversible state and die. Moreover, we found that constitutive overexpression of p21(WAF1/CIP1) alters the nucleolar unravelling process in the presence of DRB, but not of roscovitine, suggesting a role for this physiological Cdk inhibitor in the regulation of nucleolar function. Our data also support the notion that both roscovitine- and DRB-sensitive protein kinases, probably including Cdk2 and CKII, via their dual implication in the p53-Rb pathway and in ribosomal biogenesis, would participate in coupling cell growth with cell division.

Sustained accumulation of the mitotic cyclins and tyrosine-phosphorylated p34cdc2 in human G1-S-arrested cancer cells but not untransformed cells

Coupling mitosis to the completion of DNA replication in cycling embryonic extracts from Xenopus eggs appears to rely on blocking the activation of the tyrosine-phosphorylated p34cdc2/cyclin B, which continues to build up when S phase is inhibited by adding unreplicated DNA (Smythe, C., and Newport, J. W., Cell, 68: 787-797, 1992). We show here that a similar mechanism might be operative in human tumor-derived cells, which, during a thymidine-aphidicolin block, stop progressing through S phase and thereby fail to undergo mitosis. Under such conditions, indeed, cancer cells do continue to accumulate cyclin A, cyclin B1, and tyrosine-phosphorylated p34cdc2 to supranormal levels, a phenomenon that does not occur in untransformed, nonimmortalized human fibroblasts. Thus, in human cancer cells, the onset of active accumulation of cyclin A and cyclin B1 can be uncoupled from transit through the G1-S and S-G2 borders, respectively, and, as in simple embryonic cell cycles, the coupling of mitosis to the completion of S phase presumably relies, at least in part, on the prevention of premature activation of the tyrosine-phosphorylated p34cdc2/cyclin B1 complex.

Delaying the onset of M phase in NIH 3T3 cells blocked in early S phase occurs via accumulating cyclin B1 and tyrosine-phosphorylated p34cdc2 in the nucleus

An affinity-purified antibody (anti-Cdc2C) raised against the carboxy terminal sequence LDNQIKKM of p34cdc2 uncovered in NIH 3T3 cells a protein subpopulation, the location and the level of accumulation of which evolve during progression through the cell cycle: it first emerges inside the nucleus in late G1/early S phase and continues to build up principally in this location throughout S phase; a cytoplasmic expression then becomes apparent near the end of S phase, develops during G2 and sometimes prevails over the nuclear expression; it finally relocates to the nucleus in early prophase. We propose that a major part of this subpopulation would represent p34cdc2 molecules existing inside a complex with cyclin B1. NIH 3T3 cells arrested in early S phase with aphidicolin do not commit prematurely to mitosis which indicates that the regulatory pathway involved in preserving the temporal order of S and M phases is functioning in these conditions. Conjugated Western blot analysis and immunofluorescence microscopy showed that cyclin A, cyclin B1 and tyrosine-phosphorylated p34cdc2 continue to build up predominantly in the nucleus of the arrested cells. After release from the block, the cells rapidly reenter S and G2 phases and, concomitantly, cyclin B1 and tyrosine-phosphorylated p34cdc2 relocate to the cytoplasm before redistributing again in the nucleus in early prophase. These data would suggest that delaying the onset of M phase in NIH 3T3 cells in which the rate of DNA replication is reduced, is first ensured by a mechanism that prevents the cytoplasmic relocation of inactive p34cdc2/cyclin B1 complexes continually forming in the nucleus once the G1 period of mitotic cyclin instability is over.

Human cyclin B1 is targeted to the nucleus in G1 phase prior to its accumulation in the cytoplasm

The immunolocalisation patterns of cyclin B1 have been investigated in various tumour-derived and untransformed human cells, using one polyclonal (B7/B8) and two different monoclonal anti-human cyclin B1 antibodies, GNS1 and GNS11. In actively dividing cell populations, GNS11 reveals uniquely a cytoplasmic pool of cyclin B1 that rapidly increases after the onset of S phase; yet, B7/B8 and GNS1 detect, besides this cytoplasmic cyclin B1 population, a moderate but clear nuclear concentration of the protein in cells that have not yet entered S phase. As for confluent populations of untransformed and tumour cells, they become enriched in G1- and G2-arrested cells that characteristically display a discrete nuclear or an intense cytoplasmic GNS1 immunostaining respectively. Altogether, our immunofluorescence data conjugated with the results of a detailed biochemical analysis suggest that human cyclin B1 would exist in situ as two distinguishable molecular entities differentially susceptible to in vitro degradation and exhibiting different timing, kinetics and site of accumulation during the cell cycle: one form (recognized by the polyclonal and GNS1 antibodies but apparently not by GNS 11) starts to build up inside the nucleus prior to entry in S phase and the other (in which both the GNS11 and GNS1 epitopes are readily accessible) emerges and accumulates in the cytoplasm beyond the G1/S boundary.

Cell cycle-dependent regulation of nuclear p53 traffic occurs in one subclass of human tumor cells and in untransformed cells

We have analyzed the regulation of subcellular compartmentation of mutant and wild-type (WT) p53 proteins as a function of the cell cycle using immunofluorescence microscopy and referring to different markers of position in the cell cycle in different human cells expressing either mutated (KHOS-240, A 431, and T47-D cells) or WT (WI 38 and MCF-7 cells) p53. The mutant p53 proteins present in the KHOS-240, A 431, and T47-D tumor-derived cell lines enter very rapidly in the nucleus in early postmitotic cells before the chromosomes have fully decondensed; they continue accumulating in this location without any obvious cytoplasmic retention throughout the cell cycle until prophase. Such behavior is similar to that observed for the WT p53 associating with SV40 large T antigen in human WI 38 cells transformed by SV40, but it is in contrast to the behavior of the WT p53 protein present in both the untransformed WI 38 and the tumor-derived MCF-7 cells. In these latter systems, the highest nuclear concentrations of the WT protein are always found in G1 cells that still fail to exhibit a high rate of nuclear cyclin A; past the G1-S transition, the nuclear level of WT p53 tends to decrease, possibly to the benefit of cytoplasmic expression, whereas that of cyclin A concomitantly increases, suggesting that the nuclear accumulation of WT p53 becomes restricted during the phase of DNA replication. As for Saos-2 cells stably transfected with the temperature-sensitive p53Ala-143 mutant, they become arrested before the G1-S transition with a heavy pool of nuclear p53 at 32.5 degrees C, the temperature at which the transcriptional activity of p53Ala-143 is restored. All these data are compatible with the presently acknowledged primary role for WT p53, which would be to brake transit through the G1-S border possibly by directly transactivating the p21cip1 protein.

Highly specific antibody to Rous sarcoma virus src gene product recognizes nuclear and nucleolar antigens in human cells

An antiserum to the Rous sarcoma virus-transforming protein pp60v-src, raised in rabbits immunized with the bacterially produced protein alpha p60 serum (M. D. Resh and R. L. Erikson, J. Cell Biol. 100:409-417, 1985) previously reported to detect very specifically a novel population of pp60v-src and pp60c-src molecules associated with juxtareticular nuclear membranes in normal and Rous sarcoma virus-infected cells of avian and mammalian origin, was used here to investigate by immunofluorescence microscopy localization patterns of Src molecules in human cell lines, either normal or derived from spontaneous tumors. We found that the alpha p60 serum reveals nuclear and nucleolar concentrations of antigens in all the human cell lines tested and in two rat and mouse hepatoma cell lines derived from adult tumorous tissues but not in any established rat and mouse cell lines either untransformed or transformed by the src and ras oncogenes. Both the nuclear and nucleolar stainings can be totally extinguished by preincubation of the serum with highly purified chicken c-Src. We show also that the partitioning of the alpha p60-reactive proteins among the whole nucleus and the nucleolus depends mostly on two different parameters: the position in the cell cycle and the degree of cell confluency. Our observations raise the attractive possibility that, in differentiated cells, pp60c-src and related proteins might be involved not only in mediating the transduction of mitogenic signals at the plasma membrane level but also in controlling progression through the cell cycle and entry in mitosis by interacting with cell division cycle regulatory components at the nuclear level.

Slow time-dependent cellular transformation induced at restrictive temperature by ts-src mutants

Cell lines infected with a temperature-sensitive Rous sarcoma virus have been widely used to study the temporal dynamics of various transformation parameters following downshift from the non-permissive temperature to the permissive temperature as it is considered that, at the non-permissive temperature, the infected cells exhibit the morphological and growth characteristics of normal cell whereas, at the permissive temperature, they exhibit characteristics of the transformed state. We show here that the apparently normal state in which tsPA1-infected FR3T3 cells are directed at the restrictive temperature is not a stationary stable state, but rather a transient one which continuously drifts as the cells are grown and passaged at this temperature and which eventually ends up as a new transformed state (T2) with morphological and growth properties definitely different from those belonging to the transformed state (T1) at the permissive temperature. The establishment of the transformed T2 state at the restrictive temperature occurs concomitantly with a steady accumulation of an intracellular pool of pp60v-src in the vicinity of the nucleus, whose traffic towards the plasma membrane is released following downshift to the permissive temperature, leading to the reappearance of transformation parameters characteristic of the transformed T1 state. Our finding raises the possibility that the v-src protein encoded by the tsPA1 mutant of Rous sarcoma virus may induce cellular transformation via two different pathways, leading to two different transformation states, depending on at which temperature the infected cells are grown. Various possible mechanisms that could be involved in the time-dependent establishment of a transformed state by ts-src mutants at the restrictive temperature are discussed.

Immunolocalization of the cellular src protein in interphase and mitotic NIH c-src overexpresser cells

The mouse mAb, mAb 327, that recognizes specifically both pp60v-src and pp60c-src in a wide variety of cells, has been used to determine precisely the various locations of pp60c-src in NIH c-src overexpresser cells, using the technique of immunofluorescence microscopy. In interphase cells, the protein exhibits two main distributions: one that appears uniform and in association with the cell surface and the other that is patchy and juxtanuclear and coincides with the centrosomes. The juxtanuclear aggregation of pp60c-src-containing patches depends on microtubules and does not seem to occur within the Golgi apparatus and the rough ER. At the G2-to-M-phase transition, a drastic change in the localization patterns of pp60c-src takes place. We also report experiments in which the NIH c-src overexpresser cells were exposed to Con A for various times to induce a redistribution of the cell surface Con A receptors. We show that, at each stage of the Con A-mediated endocytotic process, the Con A-receptor complexes redistribute into structures to which pp60c-src appears also to be associated: at first, into patches that form at the cell surface level and then, into a cap that stands at the cell center in a juxtanuclear position and that coincides with the Golgi apparatus. During this capping process, pp60c-src-containing vesicles continue to accumulate in a centriolar spot, as in interphase, Con A-untreated cells, from which Con A is excluded. The significance of the intracellular locations of pp60c-src to the possible functions of the protein is discussed. Also, the distribution patterns of the cellular protein in the NIH c-src overexpresser cells are compared with those of pp60v-src in RSV-transformed cells. The differences observed are discussed in relation with the differences in transforming capacities of the two proteins. Finally, the possible physiological significance of the association between pp60c-src and the structures generated after the binding of Con A to its surface receptors is addressed.

Serine/threonine-specific protein kinase activity associated with viral pp60src protein

Three different types of experiments are presented in this paper, the results of which converge to indicate that the viral src protein associates with and modulates the activity and/or the specificity of a serine/threonine protein kinase. Firstly, a 60-kDa protein from extracts of FR3T3 rat fibroblasts transformed by wild-type Rous sarcoma virus (SRD-FR3T3) is shown to be immunoprecipitated with a monoclonal antibody (mAb) raised against bacterially produced pp60v-src, the mAb327 [Lipsich, L. A., Lewis, A. J. & Brugge, J. S. (1983) J. Virol. 48, 352-360] and to be phosphorylated in vitro at serine/threonine/tyrosine residues, in the ratio 25:53:22. Under the same experimental conditions, the pp60c-src protein immunoprecipitated with mAb327 from extracts of NIH c-src overexpresser cells is phosphorylated exclusively on tyrosine residues. Secondly, the results of immunoprecipitation experiments using a tumor-bearing rabbit (TBR) serum and reported in an earlier work [David-Pfeuty, T. & Hovanessian, A. (1984) Eur. J. Biochem. 140, 325-342], together with those reported here, suggest that the TBR-immunoprecipitated pp60v-src coprecipitates with a cellular protein related to the 60-kDa subunit of the Ca2+/calmodulin protein kinase II from brain. Finally, partially purified preparations of pp60v-src, but not of pp60c-src, are shown to contain a Ca2+/calmodulin-dependent protein kinase activity that phosphorylates a 52-kDa protein substrate.