Alteration of cell cycle-dependent histone phosphorylations by okadaic acid. Induction of mitosis-specific H3 phosphorylation and chromatin condensation in mammalian interphase cells.

Effects of okadaic acid (OA), a protein phosphatase inhibitor, on chromatin structure and phosphorylation of histones were examined using HeLa and N18 cells. The chromatin condensation in HeLa cells was mild and resemble prometaphase nuclei, while the condensation in N18 cells was extensive and chromatin became a compact body. H2A in HeLa cells was extensively and consistently phosphorylated at the same site throughout the cell cycle, and H3 was demonstrated to be phosphorylated at the mitosis-specific site Ser10. In contrast, H1 phosphorylation was rapidly decreased in most sites within 3 h. The reduction of H1 phosphorylation was accompanied by a quantitative change in the set of H1 phosphopeptides. During the early phase of the OA treatment, H1 phosphorylation was transiently elevated in tandem, whereas H3 phosphorylation reached a maximum somewhat later. The results suggest that mitosis-specific events (cdc2/H1 kinase activation, H1 superphosphorylation, mitosis-specific H3 phosphorylation and chromatin condensation) induced by OA are sequentially associated. The changes appear to reflect a molecular mechanism similar to that operating in normal mitosis.

Effects of okadaic acid (OA), a protein phosphatase inhibitor, on chromatin structure and phosphorylation of histones were examined using HeLa and N18 cells. The chromatin condensation in HeLa cells was mild and resemble prometaphase nuclei, while the condensation in N18 cells was extensive and chromatin became a compact body. H2A in HeLa cells was extensively and consistently phosphorylated at the same site throughout the cell cycle, and H3 was demonstrated to be phosphorylated at the mitosis-specific site Ser 10 . In contrast, H1 phosphorylation was rapidly decreased in most sites within 3 h. The reduction of H1 phosphorylation was accompanied by a quantitative change in the set of H1 phosphopeptides. During the early phase of the OA treatment, H1 phosphorylation was transiently elevated in tandem, whereas H3 phosphorylation reached a maximum somewhat later. The results suggest that mitosisspecific events (cdc2/H1 kinase activation, H1 superphosphorylation, mitosis-specific H3 phosphorylation and chromatin condensation) induced by OA are sequentially associated. The changes appear to reflect a molecular mechanism similar to that operating in normal mitosis.
The control of cell cycle progression is one of the most basic biological systems for which involvement of protein phosphorylation and dephosphorylation has been shown (1,2). These processes are controlled by protein kinase and phosphatase activities. To understand the mechanisms of chromosome kinetics in mitosis, it is thus crucial to analyze the role of nuclear protein phosphorylations at the initiation of chromatin condensation.
Okadaic acid (OA), a specific inhibitor of serine/threonine protein phosphatase (18 -20) and a potential tumor promoter (21), is known to affect mammalian cells and induce mitosislike processes. The chemical causes rounding of cells, PCC, spindle deconstruction and enhancement of H1 kinase activity (22)(23)(24)(25). The chemical also induces apoptotic cells (26,27). These phenomena accompany an elevation of overall cellular protein phosphorylations (28,29). The inhibition of phosphatase activity by OA is associated with increased phosphorylation in which pp1 or pp2A phosphatases are involved. Thus it is possible to investigate the relationship between chromatin condensation and histone phosphorylation during cell proliferation using this agent (30).
In the present investigation, we examined critically the effects of okadaic acid on chromatin condensation and histone phosphorylations in both HeLa and N18 cells and found that OA induces H3 phosphorylation in a mitosis-specific site concomitantly with chromatin condensation in interphase cells.

MATERIALS AND METHODS
Chemicals-OA was provided by Dr. S. Yamada of Nagoya University or purchased from Wako Chemical Co. 6-Dimethylaminopurine (6-DMAP) was purchased from Sigma Stock solution of OA was 1 mM dissolved in dimethyl sulfoxide.
Morphological Observations-HeLa (S-3) or N18 (mouse neuroblastoma) cells were cultured in dishes (9-cm diameter, Falcon) for 2 days. The cells were treated with or without chemicals for 30 min to 2 h. For detailed observation of cell nuclei, cells were plated on coverslips and treated with 0.1-1.0 M of OA or 0.5-1.0 mM 6-DMAP. After washing two times with phosphate-buffered saline solution, the cells were fixed with Carnoy's solution (methanol:acetic acid, 3:1) for 30 min and stained with 50 g/ml Hoechst 33258 for 30 min, then washed twice with 90% ethanol and observed by immunofluorescence microscopy (Nikon, Fluophot).
Preparation of 32 P-Labeled Histones-Cells were grown in plastic bottles (Coster 3275). For the collection of mitotic cells, logarithmically growing cells were treated with 0.02 g/ml colcemid for 16 h. Then, loosely attached cells were shaken off and collected. For 32 P labeling, cells (approximately 3.5 ϫ 10 7 ) grown either in dishes or in suspension culture were resuspended in a phosphate-free medium (8) containing [ 32 P]orthophosphate (40 Ci/ml) with or without chemicals. The cultures were incubated for 3 h or 30 min for pulse labeling at 37°C. Cells were partially lysed with a solution containing 80 mM NaCl, 20 mM EDTA, and 1% Triton X-100 (pH 7.6). Total histones were extracted twice from chromatin with 0.4 N H 2 SO 4 and the proteins precipitated with 4 volumes of ethanol as described earlier (14).
Acid-Urea Triton X-100 Polyacrylamide Gel Electrophoresis-Samples containing 30 g of histones were resolved by electrophoresis on 30-cm-long 12% polyacrylamide gel containing 7.5 M urea, 5% acetic acid, and 6 mM Triton X-100 (31). The proteins were run at 120 V for 48 h, and the gels were stained with 0.2% Amido Black and autoradiographed as described previously (14). For the quantitation of protein * This work was supported in part by The Sagawa Cancer Research Fund, Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
content, gels were subjected to scanning with a densitometer at 565 nm with a Digital Densitrol (type DMU-33C, Toyo Kogaku Co.). Each protein band in the gel was cut out, and the radioactivity of 32 P in the protein was estimated by Cerenkov counting.
Resolution of Phosphopeptides on Cellulose TLC Plates-The procedures used for trypsin digestion and tryptic phosphopeptide mapping of phosphorylated histones have been described previously (32). Phosphorylated histones were resolved by SDS-polyacrylamide gel electrophoresis (33). Individual histones was recovered electrophoretically from gels. After trichloroacetic acid precipitation, proteins were dissolved in 0.05 M sodium bicarbonate and digested with 2% trypsin for 4 h at room temperature. Resolution of the phosphopeptides in the first dimension was by electrophoresis in buffer containing buthanol, acetic acid, water, pyridine, 50:25:900:25 (pH 4.7) at 550 V for 50 min on cellulose TLC plates (Merck, Art 5716) under Isoper (Esso Co.). After the plates were dried, resolution in the second dimension was by ascending chromatography in a solution containing n-butanol, acetic acid, water, pyridine (36.6:11.4:45.3:56.7).
Estimation of H1 Kinase Activity in OA-treated Cells-Cells (4 ϫ 10 6 ) were lysed partially with buffer A (10 mM Tris⅐HCl (pH 7.4), 3 mM MgCl 2 , 0.5 mM phenylmethylsulfonyl fluoride) containing 0.4% Nonidet P-40. Nuclear pellet was washed once with buffer A without Nonidet P-40. From the nuclear pellet, 0.4 M NaCl soluble proteins were extracted. The extract was centrifuged at 100,000 ϫ g for 30 min. The clarified extract was diluted with beads buffer (34) and was reacted with p13 suc1 -Sepharose beads (Oncogene Sci, Inc.) for 90 min. The beads were then washed three times with beads buffer A. H1 kinase activity associated with the beads was estimated in a reaction mixture (50 l) containing 20 mM Tris⅐HCl (pH 7.5), 1 mM EGTA, 1 mM dithiothreitol, 10 mM MgCl 2 , 10 g of HeLa histone H1, 1 M API (an inhibitor peptide for protein kinase A, Sigma) and 20 Ci of [␥-32 P]ATP (DuPont NEN) with incubation at 30°C for 15 min. The reaction was stopped by adding 20% trichloroacetic acid on ice. H1 was precipitated and washed once with acetone containing 0.3% HCl and twice with acetone. Then, the proteins were run on SDS gels. Bands were detected by autoradiography of dried gels. For quantitation, labeled bands were excised, and radioactivity was determined by Cerenkov counting.

RESULTS
Chromatin Condensation in OA-treated Cells-To investigate the fine structure of HeLa nuclei treated with OA, cells on coverslips were fixed and stained with DNA binding dyes ( Fig.  1). Most of the HeLa cells treated with OA at 0.1-1.0 M detached from the substratum. In the cells remaining attached, weak chromosome condensation was observed in approximately 80% of nuclei (Fig. 1B), whereas untreated cells had round nuclei with a few mitoses (Fig. 1A). The condensation in HeLa cells was mild and resembled prometaphase nuclei. The extent of condensation varied, suggesting that it depends on the cell stage at the time of administration. Most mitotic nuclei were overcondensed with reduction in size and rounding (Fig.  1B, indicated by an arrowhead). In nuclei pretreated with colcemid for 16 h prior to the addition of OA for 3 h, the number of prometaphase-like or compact nuclei was remarkably increased (data not shown). However, most nuclei of N18 cells were extensively condensed and showed round body with no chromatin fibers (data not shown).
Analysis of Acid-soluble Nuclear Proteins and Their Phosphorylations in the Presence of OA or DMAP-Since protein phosphorylations are associated with chromosome condensation, histone phosphorylations were examined. HeLa cells were labeled with 32 P in the presence of OA or 6-DMAP, and the histones were extracted and analyzed by acid-urea-Triton X-100 gel electrophoresis (Fig. 2). Electrophoretic patterns of the proteins were the same for both control and inhibitortreated cells, except for additional H2A bands caused by phosphorylation ( Fig. 2A, lanes 2 and 3, H2A.1 and H2A.2). H2A and H1 are extensively phosphorylated as observed in most growing cells (Fig. 2A, lane 7). In the OA-treated cells, H2A phosphorylation was remarkably increased ( Fig. 2A, lanes 8 and 9) with two subtypes, H2A.1 and H2A.2, incorporating 2 and 2.5 times more of 32 P, respectively ( Fig. 2A and Table I). It was unexpected that the level of H1 phosphorylation was less than half of that in control cells. H3 was rapidly phosphorylated in all three subtypes (H3.1, H3.2/H3.3) and further increased with an increase in the OA concentration, whereas H2B and H4 remained unphosphorylated. With the N18 cells (Fig. 2B), similar results to OA were obtained, although the phosphorylation rate was lower than HeLa (Fig. 2B, lanes 8  and 9). The data therefore indicated that OA affected histone phosphorylation in a histone-specific manner in the two mammalian interphase cells. Thus, 1) H2A phosphorylation was increased, 2) H1 phosphorylation was extensively decreased, 3) H3 phosphorylation was newly induced in interphase cells, and 4) H2B and H4 remained in an unphosphorylated state as in the control cells, 6-DMAP inhibited both H2A and H1 histone phosphorylations of HeLa cells ( Fig. 2A, lanes 11 and 12). The chemical also suppressed the mitotic H3 phosphorylation induced by colcemid ( Fig. 2A, lane 10). The HeLa cells with colcemid contained approximately 48% mitotic cells. The decrease of these phosphorylations corresponded to the treatment dose (0.5 mM in lane 11 and 1.0 mM in lane 12) and also to the extent of chromatin decondensation (data not shown). A similar influence with histone dephosphorylation by 6-DMAP was observed (Fig. 2B, lanes 11 and 12) in the N18 cells. Quantitative data for histone phosphorylations in both HeLa and N18 cells in the presence of these two chemicals are summarized in Table  I.
Critical examination of the effective dose of OA for induction of H3 phosphorylation and decreased H1 phosphorylation revealed clear differences between N18 and HeLa cells (Fig. 3). The minimum concentration of OA for induction of H3 phosphorylation in the HeLa case was 0.2 M (Fig. 3, lane 8), whereas the induction in N18 cells at 0.2 M was very low (Fig.  3, lane 3). The extent of H3 phosphorylation in N18 cells rapidly increased with low concentrations of OA (0.1-0.5 M) (Fig.  3, lanes 2-4), to a greater extent than with colcemid (data not shown). H3 phosphorylation reached maximum levels at 0.5 M in HeLa cells (Fig. 3, lane 9), against 1 M in N18 cells (Fig. 3,  lane 10). In contrast, H1 phosphorylation decreased with an increasing dose of OA in both cell types.
Comparison of Histone Phosphopeptides between OA-treated Cells and M Cells-To analyze whether the changes of histone phosphorylations by OA reflect any alteration in the number of phosphorylation sites, [ 32 P]phosphopeptide maps of the three kinds of histones were prepared and compared with those of control cells (Fig. 4). The phosphopeptide of H2A in the OAtreated cells was single (Fig. 4b) and migrated to the same position as with H2A in untreated cells (Fig. 4a). The H2A spot in OA-treated cells was more extensively labeled than that in the control H2A. The positions of individual H1 phosphopeptides with OA (Fig. 4d) were also similar to those in untreated cells (Fig. 4c), but the intensity of the individual phosphopeptide differed considerably. The 32 P radioactivity in most spots was reduced and one spot demonstrated intensification (Fig.   4d, indicated by an arrow). The two major H3 phosphopeptides with OA (Fig. 4f) were the same as the two mitotic H3 phosphopeptides (3-1 and 3-2 in Fig. 4e). The amino acid sequences of the phosphopeptides, 3-1 and 3-2 were shown previously to contain the same residue Ser(P) 10 ; the peptide 3-2 is a product of incomplete proteolysis (17). The results indicate that OA does not induce an extra phosphorylation site in both H2A and H1, but does induce H3 phosphorylation in interphase cells exactly at the same site as in mitotic cells and some minor sites.

Relationship between H1 Kinase Activity and H1 and H3 Phosphorylations in HeLa Cells in the Presence of OA-Since
OA-induced chromatin condensation in interphase cells, H1 superphosphorylation and mitosis-specific H3 phosphorylation were investigated simultaneously (Fig. 5). The H1 and H3 phosphorylations in OA-treated cells were examined by pulse labeling with 32 P (Fig. 5a). The H1 kinase activity in the 0.4 M NaCl-soluble nuclear fraction was examined in vitro using HeLa histone H1 as a substrate (Fig. 5a). At the same time, H1 and H3 phosphorylations in aliquots of OA-treated cells were examined by pulse labeling with 32 P (Fig. 5b). During the first 30 min after the OA treatment, the incorporation of 32 P into H1 did not essentially differ from that of untreated cells (Fig. 5, a  and b), and H3 phosphorylation was negligible. H1 phosphorylation was elevated and peaked after 1 h, and H3 phosphorylation reached a maximum level at around 1.5 h. Thereafter, H1 phosphorylation returned to the normal level by 2 h, whereas H3 phosphorylation was maintained at a relatively high level. The data indicate that the kinetics of H3 phosphorylation differed from that of H1 phosphorylation. Quantitative   data on H1 kinase activity, H1 phosphorylation, and H3 phosphorylation are shown in Fig. 6.

DISCUSSION
The present study confirmed that OA causes interphase cells to demonstrate mitosis-like characteristics in terms of both morphological and biochemical parameters. Thus, 1) most OAtreated interphase cells became round in shape with condensed chromosomes, and 2) mitosis-specific histone H3 phosphorylation was induced, although the other cell cycle-dependent histone phosphorylations (H1 and H2A) were also influenced. Our data imply that OA inhibition of phosphatase results in accumulation of phosphorylated H3 proteins, so condensation of chromatin was induced without the completion of DNA synthesis. 6-DMAP, in contrast, decondenses chromatin and concomitantly dephosphorylates H3 even in cells pretreated with colcemid. The chemical is known to inhibit both MPF kinase activity and protein phosphorylations during the maturation of oocytes (35,36). Therefore, 6-DMAP inactivates both H1 and H3 kinases and thereby dephosphorylates H1 and H3 phosphorylations, which are accumulated in the presence of colcemid, leading to chromatin decondensation. This result is consistent with that of Th'ng et al. (37), who used staurosporine, another protein kinase inhibitor, and found it induced both H1 and H3 dephosphorylations and chromatin decondensation in FM3A cells. This evidence suggests that chromatin condensation must be regulated by a protein kinase/protein phosphatase related mechanism.
OA-induced chromatin condensation of HeLa cells was relatively weak as compared with the case with N18 or other cells (22). Steinmann et al. (38) reported that chromatin in human and mouse cells is hard to condense by OA. The different extent of chromatin condensation among cell types may derive from different rates of H1 kinase induction. We found that induction of H1 kinase by OA was only two times enhanced in HeLa, against 10 times in baby hamster kidney cells (22). This might be due to a lower content of factors supporting H1 kinase activity, like cyclin (38).
It is known that the Ser/Thr-protein phosphatase inhibitor,

FIG. 4. Comparison of histone phosphopeptides in OA treated cells and mitotic cells.
HeLa cells were treated with either OA (1 M) for 3 h or colcemid (0.04%) for 16 h. For the final 3 h, cells were labeled with 32 P in the presence of the chemical. H1 was extracted with 5% perchloric acid directly from nuclear pellet. H2A and H3 were obtained from total histones separated on SDS-polyacrylamide gels and electrophoretical elution. Fifty g of H1 or 15 g of H2A or H3 was digested with 2% trypsin. The minimum counts of radioactivity loaded on TLC plates were approximately 2,500 cpm for H1 and 1,500 cpm for H2A and H3. Resolution of the phosphopeptides on the TLC plates in the first dimension was carried out by electrophoresis, then by chromatography in the second dimension as described under "Materials and Methods." a, H2A in untreated cells; b, H2A in OA-treated; c, H1 in untreated; d, H1 in OA-treated; e, H3 in mitotic cells; f, H3 in OA-treated. The arrow in d indicates a spot with increased relative intensity of 32 P. 3-1 and 3-2 in e were identified as peptides containing Ser 10 residue in an earlier study (17). Small black dots indicate the origin of the phosphopeptides. The cells were lysed, 0.4 M NaCl soluble nuclear proteins were prepared, and H1 kinase activity was examined as described under "Materials and Methods." Remaining cells were pulse-labeled for a 30-min interval with 32 P, and histones were prepared. The 32 P-labeled histones were run on acid-urea-triton gels. a, H1 kinase activity in the nuclear extract; b, the incorporation of 32 P into histones H1 and H3.

FIG. 6. Relationship between H1 kinase activity and H1 and H3
Phosphorylations. The three items, H1 kinase activity in the nuclear extract (छ) and the incorporation of 32 P into histones H1 (E) and H3 (q) in the presence of OA as shown in the Fig. 5, were quantitated (cpm/g) and depicted together.
OA, specifically inactivates both protein phosphatases, pp1 and pp2A. The pp2A is inhibited completely with 1-2 nM OA, whereas pp1 requires 1 M (18,20). At the low concentration of OA (2 nM) giving pp2A inhibition, no appreciable effect was observed in terms of H2A and H3 phosphorylations or chromatin structure. Therefore, the data indicate that H2A or H3 dephosphorylations during the cell cycle involve pp1 rather than pp2A.
The mechanism by which OA induces such a burst increase of H3 phosphorylation remains unclear, and the question of whether H3 can be phosphorylated in its natural state or whether the phosphorylation requires some structural change in chromatin still awaits an answer. The H3 phosphorylation itself was probably accelerated by OA inhibition of the relevant phosphatase rather than by H3 kinase activation, since Ser 10 in nucleosomes associated with H1 is hardly phosphorylated in vitro (17,39).
In the presence of OA, the H1 phosphorylation level decreased after 2-3 h in the present case. Paulson et al. (40) similarly found that mitotic H1 phosphorylation of HeLa cells was inhibited by OA using retardation of phosphorylated proteins in gel electrophoresis. The present data indicate that 1) dephosphorylation of H1 occurs not only in M cells but also in interphase cells and 2) the dephosphorylation occurs only in H1 but not in H2A or H3. One possible explanation for this might be that H1 kinase is inactivated by OA with longer treatment as reported by Schonthal and Feramisco (41). The possibility is supported by the evidence that H2A and H3 phosphorylations are not inhibited by protein kinases other than cdc2/H1 kinase. Alternatively, the protein phosphatase responsible for H1 phosphorylation in vivo is neither pp1, pp2A, or the other new protein phosphatase, as discussed by Paulson et al. (40). Although it has been reported that pp2A (42) or pp1 (43) is involved, the phosphatase responsible in mammalian cells has not been determined yet. The remaining phosphorylated H1 demonstrated that the dephosphorylation of a specific peptide was very slow or uneffected by OA. This indicates that the efficiency of dephosphorylation by a protein phosphatase may depend on the primary structure of the phosphorylation site (44).
The present study revealed that H3 phosphorylation associated with OA occurs at the mitosis-specific site (Ser 10 ). H3 phosphorylation is the most closely associated with chromatin condensation among the three phosphorylatable histones during the cell cycle. Extensive H3 phosphorylation has been observed thus far with various kinds of chromatin condensation; 1) normal mitosis from prophase to anaphase (4), 2) PCC in a hybrid cells between S and M phase cells (13), 3) PCC in tsBN2 at the nonpermissive temperature (14,15), and 4) PCC in the presence of protein phosphatase inhibitor, OA (present study). In most cases, there is an accompanying high level or transient elevation of H1 phosphorylation. It was earlier proposed that a high level of mitotic H1 phosphorylation may facilitate mitotic H3 phosphorylation (17,39). It is possible that this site of H3 could become exposed by the transient elevation of H1 phosphorylation and thereby become accessible to H3 kinase. Subsequent to this step, phosphorylated H1 may not be directly involved in chromatin condensation until H1 phosphatase removes the phosphate. This notion may explain the recent controversy regarding the role of H1 in chromatin condensation in that it is not essential in the amphibian embryonic system in vitro (45,46), whereas it is in growing organisms (47). On the other hand, it is known that OA phosphorylated on 80-kDa proteins binding at AT-rich and on nuclear matrix-associated 70-kDa proteins (48). The effect of these phosphorylated nonhistone proteins is not rule out for the chromatin condensation. The H1 kinase activated by OA is known to be Cdc2-type protein kinase associated with cyclin B, which is same as in mitotic cells (22). The present evidence that transient H1 and mitotic H3 phosphorylations occur in the presence of OA indicates that chromatin condensation with OA in interphase cells is induced by a biochemical mechanism similar to that operating in normal mitosis. The causal relationship between chromatin condensation and H3 phosphorylation now requires elucidation, as well as determination of H3 kinase(s) actions.