Mitotic HeLa Cells Research Articles

Protein Phosphatase-1α, γ1, and δ: Changes in Phosphorylation and Activity in Mitotic HeLa Cells and in Cells Released from the Mitotic Block

Protein phosphatase-1 is phosphorylated “in vitro” by cdc2-cyclin B (E. Villa-Moruzzi,FEBS Lett.304, 211–215, 1992). In the present study the phosphatase-1 isoforms α, γ1, and δ were analyzed in mitotic (nocodazole-blocked) HeLa cells. Phosphorylation on threonine increased in γ1 and δ at mitosis. α was phosphorylated only in mitotic cells and mainly on serine. Exposure of permeabilized mitotic cells to a peptide that inhibits cdc2 decreased the phosphorylation of the isoforms. Cell fractionation indicated that phosphatase-1 was over 90% inactivated and phosphorylated in the soluble, but not in the chromosomal fraction of mitotic cells. Immunoprecipitation from the mitotic soluble fraction indicated that only γ1 and δ, but not α, were inactivated. Altogether the data pointed to a correlation between phosphatase-1 inactivation and phosphorylation in mitotic cells. cdc2-cyclin B might be the kinase (or one of the kinases) that phosphorylates phosphatase-1. In cells released from the mitotic block, the phosphatase-1 activity in the soluble, but not in the nuclear fraction, increased progressively, reaching control values by 16 h. Immunoprecipitation indicated that the increase in activity was due to α and δ only. On the other hand, the activity of γ1 remained low, and this was also the only isoform that remained phosphorylated, though less than in mitotic cells. Also in the case of the cells released from mitosis, a correlation may exist between phosphorylation and inactivation of phosphatase-1. However, the regulation of phosphatase-1 is complex and may involve also regulatory subunits that are still unknown. Altogether, the results indicated the differential regulation of the phosphatase-1 isoforms both at mitosis and in G1 cells.

Cell-cycle related regulation of poly(A) polymerase by phosphorylation.

The poly(A) tail found on almost all eukaryotic messenger RNAs is important in enhancing translation initiation and determining mRNA stability. Control of poly(A)-tail synthesis thus has the potential to be a key regulatory step in gene expression and is indeed known to be important during early development in many organisms. To study a possible basis for such regulation, we examined phosphorylation of poly(A) polymerase (PAP) by p34(cdc2)/cyclin B (maturation/mitosis-promoting factor, MPF). We show here that PAP can be phosphorylated in vivo and in vitro by MPF. Consistent with this, PAP becomes hyperphosphorylated both during meiotic maturation of Xenopus laevis oocytes and in HeLa cells arrested at M phase, times in the cell-cycle when MPF is known to be active. We show further that hyperphosphorylation by MPF dramatically reduces the activity of purified PAP, and that PAP isolated from mitotic HeLa cells is similarly inhibited by hyperphosphorylation. This repression probably contributes to the well established reductions in poly(A)+ RNA and/or protein synthesis known to occur in M-phase cells.

Role of cdc25-C phosphatase in the immediate G2 delay induced by the exogenous factors epidermal growth factor and phorbolester.

Studies on the link between cellular signalling and cell cycle control at the G2 checkpoint have shown that, in HeLa cells, epidermal growth factor (EGF) and the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) rapidly inhibit the G2-M transition by preventing the key component of mitosis-promoting factor (MPF), p34cdc2, from expressing protein kinase activity. The kinase activity of active MPF is not inhibited; rather, the conversion of pre-MPF to MPF, i.e., the activating dephosphorylation of p34cdc2, at tyrosine is rapidly blocked (Barth and Kinzel, 1994, Exp. Cell Res. 212:383-388; 1995, J. Cell. Physiol., 162:44-51). The phosphatase responsible, cdc25-C, is activated by phosphorylation in mitotic cells starting at the G2-M transition in an autocatalytic loop with MPF (Hoffmann et al., 1993, EMBO J. 12:53-63). We now show that, concomitant with the prevention of MPF activation, EGF and TPA induced a reduction of the activity of cdc25-C in synchronized cultures. Furthermore, treatment of mitotic HeLa cells with TPA did not influence the kinase activity of MPF but caused a rapid decrease of the specific enzyme activity of cdc25-C, probably due to dephosphorylation of the enzyme, as indicated by reduced binding of monoclonal MPM-2 antibody specific for phosphoepitopes in M phase. Because of its inability to induce signalling during division, EGF failed to influence the activity of cdc25-C in mitotic cells. The scenario in cells late in G2 that are committed to enter mitosis may be as follows: In those cells where the signalling pathways responding to EGF as well as those responding to TPA are still open, cdc25-C is prevented by dephosphorylation from exceeding the threshold level of activity required to initiate the activation of and the autocatalytic feedback loop with p34cdc2 and to enter mitosis. Therefore, cdc25-C appears to represent part of an interface between cellular signalling and cell cycle control in G2 phase.

Role of cdc25‐C phosphatase in the immediate G2 delay induced by the exogenous factors epidermal growth factor and phorbolester

Studies on the link between cellular signalling and cell cycle control at the G2 checkpoint have shown that, in HeLa cells, epidermal growth factor (EGF) and the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) rapidly inhibit the G2-M transition by preventing the key component of mitosis-promoting factor (MPF), p34cdc2, from expressing protein kinase activity. The kinase activity of active MPF is not inhibited; rather, the conversion of pre-MPF to MPF, i.e., the activating dephosphorylation of p34cdc2, at tyrosine is rapidly blocked (Barth and Kinzel, 1994, Exp. Cell Res. 212:383–388; 1995, J. Cell. Physiol., 162:44–51). The phosphatase responsible, cdc25-C, is activated by phosphorylation in mitotic cells starting at the G2-M transition in an autocatalytic loop with MPF (Hoffmann et al., 1993, EMBO J. 12:53–63). We now show that, concomitant with the prevention of MPF activation, EGF and TPA induced a reduction of the activity of cdc25-C in synchronized cultures. Furthermore, treatment of mitotic HeLa cells with TPA did not influence the kinase activity of MPF but caused a rapid decrease of the specific enzyme activity of cdc25-C, probably due to dephosphorylation of the enzyme, as indicated by reduced binding of monoclonal MPM-2 antibody specific for phosphoepitopes in M phase. Because of its inability to induce signalling during division, EGF failed to influence the activity of cdc25-C in mitotic cells. The scenario in cells late in G2 that are committed to enter mitosis may be as follows: In those cells where the signalling pathways responding to EGF as well as those responding to TPA are still open, cdc25-C is prevented by dephosphorylation from exceeding the threshold level of activity required to initiate the activation of and the autocatalytic feedback loop with p34cdc2 and to enter mitosis. Therefore, cdc25-C appears to represent part of an interface between cellular signalling and cell cycle control in G2 phase. © 1996 Wiley-Liss, Inc.

Evidence that the endogenous histone H1 phosphatase in HeLa mitotic chromosomes is protein phosphatase 1, not protein phosphatase 2A.

Histone H1 is highly phosphorylated in mitotic HeLa cells, but is quickly dephosphorylated in vivo at the end of mitosis and in vitro following cell lysis. We show here that okadaic acid and microcystin-LR block the in vitro dephosphorylation of H1 and that they do so directly by inhibiting the histone H1 phosphatase rather than by some indirect mechanism. The concentrations of microcystin and okadaic acid required for inhibition strongly suggest that the histone H1 phosphatase is either PP1 or an unknown protein phosphatase with okadaic acid-sensitivity similar to PP1. The histone H1 phosphatase is predominantly located in chromosomes with at most one copy for every 86 nucleosomes. This tends to support its identification as PP1, since localization in mitotic chromosomes is a characteristic of PP1 but not of the other known okadaic acid-sensitive protein phosphatases. We also show that treatment of metaphase-arrested HeLa cells with staurosporine and olomoucine, inhibitors of p34cdc2 and other protein kinases, rapidly induces reassembly of interphase nuclei and dephosphorylation of histone H1 without chromosome segregation. This result indicates that protein kinase activity must remain elevated to maintain a mitotic block. Using this as a model system for the M- to G1-phase transition, we present evidence from inhibitor studies suggesting that the in vivo histone H1 phosphatase may be either PP1 or another phosphatase with similar okadaic acid-sensitivity, but not PP2A.

Mimosine Arrests the Cell Cycle after Cells Enter S-Phase

L-Mimosine (β-N-[3-hydroxy-4-pyridone]-α-aminopropionic acid)—a rare amino acid derived fromMimosaandLeucaenaplants—arrests cells reversibly late during G1 phase or at the beginning of S-phase. If mimosine were to arrest cells immediately before S-phase, it would provide a superb tool for the investigation of the initiation of DNA synthesis. Therefore, we reexamined the point of action of mimosine. Mitotic HeLa cells were released into 200 μMmimosine and grown for ∼10 h to block them, before the cells were permeabilized and the amino acid removed by washing them thoroughly. On addition of the appropriate triphosphates, DNA synthesis—measured by the incorporation of [32P]dTTP—began immediately; as it is known that such permeabilized cells cannot initiate DNA synthesis but can only resume elongating previously initiated chains, mimosine must arrest after DNA synthesis has begun. Moreover, cells grown in mimosine assembled functional replication factories—detected by immunolabeling after incorporation of biotin–dUTP—that were typical of those found early during S-phase. Disappointingly, it seems that mimosine—like aphidocolin—blocks only after cells enter S-phase.

Dynamic changes of NuMA during the cell cycle and possible appearance of a truncated form of NuMA during apoptosis.

We have demonstrated that dynamic redistribution of nuclear-mitotic apparatus (NuMA) protein in the cell cycle is correlated temporally and spatially with its biochemical modifications. In interphase, NuMA behaves solely as a 220 kDa nuclear matrix-associated protein. After initiation of DNA condensation during mitosis, NuMA is phosphorylated by Cdc2 kinase into a 240 kDa form which is transported quickly to the centrosomal region. Once cells have passed the metaphase-anaphase transition, the 240 kDa form of NuMA either becomes a 180 kDa truncated form which is fated to be degraded completely before mitotic exit, or returns to the 220 kDa form that relocates to the daughter nuclei and remains throughout interphase. Apparently, a proteolytic enzyme is activated during the late stages of mitosis. After induction of a 180 kDa form of NuMA in interphase HeLa cells by 5,6-dichloro-1-beta-D-ribofuranosylbenzimidazole, nuclear apoptotic phenomena including chromatin condensation, DNA fragmentation, and micronucleation were observed. However, the same treatment did not induce apoptosis in mitotic phase-arrested HeLa cells. The 180 kDa form of NuMA was demonstrated to be a truncated product, at least lacking the tail domain. When HL60 cells were stimulated by diverse apoptosis inducers such as camptothecin, staurosporine, cycloheximide, and A23187, the extent of NuMA cleavage to produce a 180 kDa product was comparable with the degree of oligonucleosomal laddering. NuMA cleavage is likely to be a consequence of the onset of apoptosis. The intact 220 kDa NuMA functions in interphase cells to retain the nuclear structural integrity. Additionally, NuMA appears to act as a nuclear structural target for a death protease during apoptosis.

Arachidonic acid mobilization is suppressed during mitosis: role of cytosolic phospholipase A2 activation

In interphase HeLa cells, incubation with histamine or thapsigargin led to the rapid release of arachidonic acid. The release was absolutely dependent on Ca2+, consistent with the activation of an 85 kDa cytosolic phospholipase A2 (cPLA2). In metaphase-arrested HeLa cells, by contrast, the stimulation of arachidonate release by these agents was inhibited by more than 90%. The lack of arachidonic acid release by mitotic cells was at least partly expected, since histamine- or thapsigargin-induced Ca2+ influx and elevations of cytosolic free Ca2+ are known to be strongly inhibited during mitosis [Preston, Sha'afi and Berlin (1991) Cell Regul. 2, 915-925]. Indeed, incubation of interphase cells with the Ca2+ ionophore A23187 alone induced a high level of arachidonate release. However, even A23187 failed to elicit release from mitotic cells. Since the Ca(2+)-dependent release of arachidonate by many cell types is promoted by preincubation with ligands that activate receptors of the tyrosine kinase class, and tumour promoters that lead to the phosphorylation of cPLA2, we determined if the responses of mitotic HeLa cells could be modified by this 'priming' process. We first established that epidermal growth factor and phorbol 12-myristate 13-acetate were effective priming agents in interphase cells: cells preincubated with the hormone or tumour promoter showed a 2-fold stimulation of thapsigargin- or A23187-induced arachidonic acid release. However, none of the priming agents reversed the lack of mitotic cell response. This refractoriness was not caused by destruction of cPLA2 during mitosis: by Western blotting, cPLA2 of interphase and mitotic cells was shown to be present in comparable amounts. Moreover, cPLA2 activities measured in extracts of interphase and mitotic cells were also comparable. Surprisingly, mitotic cPLA2 appeared to be constitutively phosphorylated in non-hormone-treated (control) cells. The results indicate a novel mechanism of regulation by cPLA2 activity in mitotic cells.

Mitotic HeLa cells contain a CENP-E-associated minus end-directed microtubule motor.

A minus end-directed microtubule motor activity from extracts of HeLa cells blocked at prometaphase/metaphase of mitosis with vinblastine has been partially purified and characterized. The motor activity was eliminated by immunodepletion of Centromere binding protein E (CENP-E). The CENP-E-associated motor activity, which was not detectable in interphase cells, moved microtubules at mean rates of 0.46 micron/s at 37 degrees C and 0.24 micron/s at 25 degrees C. The motor activity co-purified with CENP-E through several purification procedures. Motor activity was clearly not due to dynein or to kinesin. The microtubule gliding rates of the CENP-E-associated motor were different from those of dynein and kinesin. In addition, the pattern of nucleotide substrate utilization by the CENP-E-associated motor and the sensitivity to inhibitors were different from those of dynein and kinesin. The CENP-E-associated motor had an apparent native molecular weight of 874,000 Da and estimated dimensions of 2 nm x 80 nm. This is the first demonstration of motor activity associated with CENP-E, strongly supporting the hypothesis that CENP-E may act as a minus end-directed microtubule motor during mitosis.

Golgi membrane vesicles in HeLa mitotic cells are identified with monoclonal antibody made against Golgi cisternal membrane protein p138.

A monoclonal antibody (mAbG3A5) recognizing p138 antigen was used to identify the Golgi cisternal membrane and determine behavior of Golgi fragments during mitosis in HeLa cells. At the start of mitosis, Golgi stacks identified with the mAbG3A5 antibody were fragmented into fine membrane vesicles which were distributed throughout the cytoplasm leaving only the region of the chromosome cluster unoccupied. On Western immunoblotting analysis, p138 was found associated with the membrane fraction prepared from mitotic HeLa cells having a buoyant density the same as that of interphase Golgi membranes. In addition to the fine membrane vesicles, clusters labeled with mAbG3A5 antibody were frequently observed in mitotic cells. They numbered 11 on average per mitotic cell and consisted of fine membrane vesicles of which membrane region was labeled with the mAbG3A5 antibody. This fact indicates that the membrane vesicles in mitotic Golgi clusters were also part of the fragments of Golgi cisternae. The number of mitotic Golgi clusters per mitotic cell was constant from prophase to anaphase, increasing twofold at telophase, although the average size of mitotic Golgi cluster remained unchanged throughout mitosis. The increase in number of mitotic Golgi clusters at telophase was accompanied by decrease in immunofluorescence of fine membrane vesicles. Treatment with nocodazole caused the disappearance of the mitotic Golgi clusters from prophase cells; however upon removal of it, they were reformed. These results suggest that during mitosis the Golgi apparatus were fragmented to fine membrane vesicles leaving only a part as mitotic Golgi clusters and were reassembled through tentative clustering of the fine membrane vesicles at the end of mitosis.

Phosphorylation of Rap1GAP during the Cell Cycle

Rap1GAP (for Rap1 GTPase Activating Protein) is an 89 kD protein that highly stimulates the intrinsic GTPase activity of the small GTP binding protein Rap1. It has been shown that Rap1GAP is phosphorylated in vitro by purified p34cdc2 kinase, which regulates the G2/M transition of the cell cycle. In this work, we have studied the phosphorylation of Rap1GAP during the cell cycle and showed that Rap1GAP is phosphorylated in vivo in interphasic and mitotic Hela cells; the electrophoretic mobility of Rap1GAP from mitotic cells is reduced compared with that from interphasic cells, suggesting that the mitotic form of the protein is hyperphosphorylated. As the cdc2 kinase is specifically active during mitosis, we sought to investigate whether it actually phosphorylates Rap1GAP during this phase of the cell cycle. We show that p34cdc2 coimmunoprecipitated from mitotic Hela cell lysates with an anti human cyclin B1 antibody, but not from interphasic cell lysates, is able to phophorylate efficiently wild-type Rap1GAP, but not a mutant in which the putative consensus site for phosphorylation by the cdc2 kinase (serine 484) has been altered. Moreover, depletion of p34cdc2 from mitotic extracts abolishes the phosphorylation of Rap1GAP by such lysates. These results therefore strongly suggest that Rap1GAP is indeed a substrate of the cdc2 kinase during mitosis. This phosphorylation does not affect the stimulation of the GTPase activity of Rap1 by Rap1GAP but may play a role in regulating the interaction of Rap1GAP with other proteins involved in the cellular functions regulated by Rap1 and Rap1GAP.

Activation of the cdc25C Phosphatase in Mitotic HeLa Cells

The cdc2-activator cdc25C was immunoprecipitated from HeLa cell extracts and assayed as tyrosine phosphatase (PTP) using tyrosine-phosphorylated myelin basic protein. The PTP activity was 12-fold higher in immunocomplexes from mitotic (nocodazole-arrested) than from asynchronous cells. This difference is due to enzyme activation, since the same amount of cdc25C was immunodetected in both conditions. However, mitotic cdc25C had Mr 59,000, while a 56,000-59,000 doublet was detected in immunocomplexes from asynchronous cells. The PTP activity of mitotic cdc25C was decreased by treatment with Phosphatase-2A catalytic subunit (but not with Phosphatase-1), with re-appearance of the 56,000 polypeptide. cdc25C was also found associated with cdc2-p13-Sepharose complex and its PTP activity was 7-fold higher in samples from mitotic than from asynchronous cells. cdc25C and cdc2 co-migrated during gel filtration and the higher activity of mitotic cdc25C was retained through gel filtration.

Determination of the radiation sensitivity of the stromal cells in the murine long-term bone marrow culture by measuring the induction and rejoining of interphase chromosome breaks

Purpose : To determine the radiosensitivity of bone marrow stromal cells, the rate of interphase chromosome breakage and rejoining of stromal cells in the murine long term bone marrow culture and of human skin fibroblasts were compared. Methods and Materials : The cells were irradiated with doses up to 6 Gy and repair times up to 6 hr were investigated. After induction of premature chromosome condensation by fusing the cells with mitotic HeLa cells, the number of interphase chromosome fragments was counted. Results : The number of radiation induced breaks was found to be not significantly different for both cell types with 6.16 ± 0.26 breaks per Gray for the fibroblasts and 5.96 ± 0.20 breaks per Gray for the stromal cells. A significant difference was observed in the repair rate. The fibroblasts rejoined 39.6% of the breaks induced initially during the first hour after irradiation and 5.6 ± 1.84 breaks remained unrejoined after 6 hr, while the stromal cells were able to rejoin 63.2% in 1 hr and had 2.05 ± 0.07 breaks unrejoined after 6 hr. Conclusion : If the well substantiated assumption is made, that the capacity to repair DNA double strand breaks or interphase chromosome breaks is correlated with the cellular radiosensitivity, this findings indicate, that murine bone marrow stromal cells are more radioresistant than human skin fibroblasts.

Mechanism of chromosome exchange formation in human fibroblasts: insights from "chromosome painting".

We have used the techniques of premature chromosome condensation (PCC) and fluorescence in situ hybridization (FISH) with a library for human chromosome 4 to analyze the rate of rejoining of chromosome breaks and development of exchange aberrations in AG1522 human fibroblasts. AG1522 cells were irradiated in plateau phase with 10 Gy and fused with mitotic HeLa cells either immediately after irradiation or at intervals up to eight days later. The slides were then hybridized with the chromosome 4 library and unrejoined breaks and exchange events (visualized as bicolor chromosomes) scored in these cells. At the earliest time point after irradiation, the number of exchange events in the irradiated cells was low, but increased with kinetics similar to that of the joining of the breaks. Furthermore, when we analyzed those cells which had exchange events for their distribution, almost all of the cells initially contained one exchange event (1 bicolor chromosome). As time progressed, the number of cells containing exchanges with two exchange events per cell increased as the number with one exchange event per cell decreased. Extrapolation of the number of exchange events to zero time (with an estimate of 20 min for the fusion and condensation times) gave a value consistent with zero exchanges at zero time after irradiation. In a separate experiment, we also scored AG 1522 cells at the first metaphase after a dose of 6 Gy and were able to show that as many as 50% of the complete exchanges were non-reciprocal in nature, that is, the two broken ends of a single break in chromosome 4 joined to two different chromosomes. These data support the classical breakage-and-reunion model rather than the Revell Exchange Theory of exchange formation.

Sphingolipid transport in mitotic HeLa cells.

Mitotic and interphase HeLa cells were labeled with [3H]serine. Ceramide and its derivatives, lactosylceramide and sphingomyelin, were biosynthetically labeled under both conditions. Only in the absence of nocodazole, as the cells entered telophase, was an additional glycosphingolipid synthesized which was identified as GA2 (GalNAc(beta 1,4)Gal(beta 1,4)Glc(beta 1,1)Cer). Ceramide, the basic sphingolipid precursor, is synthesized in the endoplasmic reticulum, whereas its immediate derivatives are synthesized in early Golgi compartments. Transport of newly synthesized proteins from the endoplasmic reticulum to the Golgi is inhibited in mitotic cells while ceramide acquires early Golgi modifications under the same conditions, suggesting that ceramide can be delivered to the Golgi by a different route. Since GA2 is synthesized in late Golgi, its absence in mitotic cells strongly argues for an in vivo inhibition of intra-Golgi transport, an observation with important implications for the mechanism of Golgi division.

A Radiation-Induced Inhibitor of Chromosome Condensation and Nuclear Envelope Breakdown in HeLa Cells

An in vitro microscopic assay for mitosis-inducing activity in mitotic HeLa cells was developed and used to demonstrate that cells irradiated and arrested in G2 phase of the cell cycle contain an inhibitor of mitosis. This assay system has a number of advantages over other assays including the use of autologous components (HeLa nuclei and mitotic cell extracts) in contrast to the microinjection method with Xenopus oocytes and without the requirements for microinjection expertise and Xenopus oocytes. The radiation-inducible inhibitor was detected at the lowest radiation dose tested (2 Gy) with maximal activity achieved within 30 min after radiation. Inhibitor activity decayed with time after radiation (2 Gy) with no activity detected at 6 h even though the cells remained in G2 phase, suggesting that either synthesis or activation of additional components is necessary for recovery from G2 arrest. The inhibitor activity was not detected in irradiated cells treated with caffeine to induce premature recovery from G2 arrest.

Effects of okadaic acid on mitotic HeLa cells.

Mitotic HeLa cells were treated with different concentrations of okadaic acid (OA), known to inhibit phosphatase 1 and 2A activities. The cytological effects on the course of mitosis were studied at the light microscopic, immunoflourescence and electron microscopic levels. At the lowest concentration used (1 nM), OA did not show any effect on mitosis, but at higher concentrations it showed pronounced effects. The mitotic chromosomes became scattered, the mitotic spindle became deranged and the cells failed to enter anaphase. At the electron microscopic level formation of isolated microtubules and regular trilaminar kinetochores were observed. An extensive growth of the endoplasmic reticulum could be noted in these cells. Decondensation of chromatin and nuclear envelope re-formation could be seen only after withdrawal of OA. A high frequency of multinucleate cells could be found after 24 h of recovery. Cells treated with 100 nM OA for 3 hours showed diplochromosomes in over 50% of mitotic cells after 24 h recovery. These were presumably formed due to the failure of sister chromatid separation in the earlier mitosis in the presence of OA. At the electron microscopic level the diplochromosomes showed a quadruplet structure. The role of phosphatase 1 in controlling some late mitotic events, i.e. sister chromatid separation, MPF-inactivation and nuclear envelope re-formation etc., is discussed.

Okadaic acid inhibits sister chromatid separation in mammalian cells

Mitotic HeLa cells were treated with different concentrations of okadaic acid inhibiting phosphatase 2A activity alone or in addition to phosphatase 1 activity. Phosphatase 2A inhibition alone had no visible effect on mitosis, but inhibition of both phosphatase 1 and 2A produced mitotic abnormalities, including inhibition of anaphase mimicking the effect of colchicine. Recovery experiments in okadaic acid-free medium showed formation of diplochromosomes, indicating a failure of sister chromatid separation in the treated mitotic cells. The universality of the phosphatase 1 requirement in sister chromatid separation is discussed.

Partial displacement of histone H1 from chromatin is required before it can be phosphorylated by mitotic H1 kinase in vitro.

The massive nonselective and reversible phosphorylation of histone H1 during mitosis is a universal phenomenon among eukaryotes. The growth-associated kinase responsible for this phosphorylation is identical to the maturation promoting factor, a key regulator of the cell cycle. Here we showed that growth-associated kinase, isolated from mitotic HeLa cells which were capable of phosphorylating HeLa H1 in vitro with high activity and mostly at the same sites phosphorylated during mitosis in vivo (assayed by two-dimensional analysis of tryptic phosphopeptides), did not significantly phosphorylate chromatin-bound or nuclear H1 in vitro. Its inability to phosphorylate chromatin-bound H1 did not change when the amount of kinase was increased or the incubation was prolonged. The resistance of chromatin-bound H1 to phosphorylation did not result from chromatin aggregation. Rapid phosphorylation of H1 in vitro, as well as in a nuclear system, was restored when NaCl concentrations were raised above 200 mM where H1:DNA interactions are weakened. At 300 mM NaCl, chromatin-bound H1 was phosphorylated in a subset of the sites observed for free H1 phosphorylated in vitro. These results suggest that active displacement of H1 from chromatin DNA may take place before H1 can be fully phosphorylated during mitosis.

Three-dimensional immunogold labeling of nuclear matrix proteins in permeabilized cells

A preembedment labeling procedure is described for the three-dimensional (3D) labeling of nuclear matrix proteins in permeabilized cells. The procedure is based on the use of ultra-small (1 nm) gold particles as a marker system. This marker penetrates the nucleus more efficiently than the conventionally used 5-10 nm colloidal gold probes. Dehydration is performed by freeze-substitution to preserve the ultrastructure of the cell as optimally as possible. During freeze-substitution the samples are stained by uranyl ions to stain the cellular material throughout the resin section. The 3D gold-labeled and uranyl-stained specimen is embedded in Epon resin and semi-thin (0.2-0.5 microns) sections are made for stereo electron microscopy. The applicability of this method is illustrated by the localization of nuclear matrix-associated nuclear bodies in permeabilized interphase and mitotic HeLa cells.