Mitotic Arrest and Enhanced Nuclear Protein Phosphorylation in Human Leukemia K562 Cells by Okadaic Acid, a Potent Protein Phosphatase Inhibitor and Tumor Promoter*

We investigated the effects of the non-phorbol tumor promoter okadaic acid on human leukemia K562 cells. It was found that okadaic acid potently and reversibly inhibited cell growth, with a nearly complete inhibition of thymidine uptake seen at about 10 nM. The cytotoxicity of okadaic acid was characterized by a marked mitotic arrest of the cells exhibiting scattered chro-mosomes and abnormal anaphase-like structures, a phenomenon distinct from the typical metaphase arrest caused by colchicine. Okadaic acid (10-1,000 nM) greatly stimulated phosphorylation of a number of nuclear proteins in K562 cells. Phosphorylation of many of the same proteins was also stimulated by 12-0-tetradecanoylphorbol-l3-O-acetate, a protein kinase C activator. The present findings, consistent with recent reports that okadaic acid is a potent inhibitor of protein phosphatases 1 and 2A (PP1 and PPBA) shown to be essential for normal mitosis, provided evidence for the first time that okadaic acid inhibition of PPl/ PPZA resulted in enhanced nuclear protein phosphorylation and subsequent mitotic arrest. in cell-free com- IO') homogenized in 1 homogenization centrifuged yield and post- nuclear fractions. Aliquots (0.25 ml) of the incubated (in 0.5 for at 30 with 0.5 mM CaCI, (in excess of EGTA), 10 pg/ml of phosphatidylserine, 10 mM MgCI,, and 10 p~ [y-:"P] ATP (ahout 2 X 10' cpm), with or without TPA or okadaic acid (Fig. 4R), essentially as

We investigated the effects of the non-phorbol tumor promoter okadaic acid on human leukemia K 5 6 2 cells. It was found that okadaic acid potently and reversibly inhibited cell growth, with a nearly complete inhibition of thymidine uptake seen at about 10 nM. The cytotoxicity of okadaic acid was characterized by a marked mitotic arrest of the cells exhibiting scattered chromosomes and abnormal anaphase-like structures, a phenomenon distinct from the typical metaphase arrest caused by colchicine. Okadaic acid (10-1,000 nM) greatly stimulated phosphorylation of a number of nuclear proteins in K 5 6 2 cells. Phosphorylation of many of the same proteins was also stimulated by 12-0tetradecanoylphorbol-l3-O-acetate, a protein kinase C activator. The present findings, consistent with recent reports that okadaic acid is a potent inhibitor of protein phosphatases 1 and 2 A (PP1 and PPBA) shown to be essential for normal mitosis, provided evidence for the first time that okadaic acid inhibition of P P l / PPZA resulted in enhanced nuclear protein phosphorylation and subsequent mitotic arrest.
Okadaic acid, a polyether fatty acid and a diarrhetic shellfish poisoning factor isolated from marine sponges (I), is an inhibitor of protein phosphatases PP1 and PP2A' (2) and a potent non-phorbol tumor promoter (3). The molecular basis for the tumor-promoting activity of okadaic acid is thought to be due to its inhibition of PP1 and PP2A, two of four major protein phosphatases of mammalian cells (4), thus enabling maintenance of a high phosphorylation level of cellular proteins, an effect that is similarly achieved by TPA, a potent tumor promoting phorbol ester and protein kinase C activator ( 5 ) . In this regard, the PPl/PP2A family of protein phosphatases might function as tumor suppressors by dephosphorylating cellular proteins. Several aspects of okadaic acid on cellular regulation have been reviewed recently (6,7). Other 1 and 2A, respectively; TPA, 12-O-tetradecanoyl-12-O-acetate; ALP, alkyllysophospholipid (1-O-octadecyl-2-0-methyl-glycero-3-phosphocholine); Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; EGTA, (ethylenebis(oxyethylenenitrilo)]tetraacetic acid. biological effects of the toxin include relaxation (8,9) and contraction (9) of isolated aorta, contraction of isolated tenia coli (IO), increased Caz+ current in isolated cardiac myocytes ( l l ) , increased phosphorylation of cellular proteins in hepatocytes (12) and adipocytes (121, increased generation of Ca'+independent form of Ca"/calmodulin-dependent protein kinase I1 in cerebellar granule cells due to phosphorylation of the enzyme (13), stimulation of glucose transporter (12) and its phosphorylation in adipocytes (14), stimulation of Na/K/ 2C1 cotransporter in erythrocytes ( X ) , and activation of protein kinase activity in adipocytes (16).
Recent evidence indicates that the cell cycle is regulated by protein phosphorylation/dephosphorylation reactions. A serine/threonine protein kinase ~34'~"', the catalytic subunit of maturation promoting factor or growth-associated histone H1 kinase (17), and expression of the cdc2 family of genes have been shown to be essential for eukaryotic cells to progress from G1 to M phase (for example, . Phosphorylation of some nuclear proteins (such as histone H1, lamin B, and nucleolin) by ~3 4 "~" has been suggested to be responsible for the mitotic events which include disassembly of the nucleus, chromosome condensation, and generation of mitotic spindle (21). PP1 and PP2A have also been shown to be essential for mitosis. It appears, however, that they are likely involved in mitotic progression (22, 23), a role apparently different from that of ~34'~'' for mitotic initiation or G1 to M phase progression (18-20) as mentioned above. Dephosphorylation by PP1 and PP2A of ~3 4 '~' ' substrates and other unidentified proteins to their interphase levels of phosphorylation would be an important event in mitotic progression after metaphase or for cells to escape from mitosis.
In the present studies, we report that the PPI and PP2A inhibitor okadaic acid potently inhibited proliferation of human leukemia K562 cells, which was characterized by enhanced nuclear protein phosphorylation and a pronounced mitotic arrest.

EXPERIMENTAL PROCEDURES
Materials-Okadaic acid was a kind gift of Dr. Thomas R. Soderling or purchased from Moana BioProducts (Honolulu, HI); TPA was from LC Service (Woburn, MA); ALP was from Calbiochem; ["HI thymidine was from ICN Radiochemicals (Irvine, CA); human leukemia cell line K562 was from American Type Culture Collection (Rockville, MD); media and supplies for cell culture were from GIBCO; and colchicine and other compounds were from Sigma. Cellular Studies"K562 cells were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated bovine calf serum, in a humidified incubator a t 37 "C in 95% air, 5% CO,. Cells at the midlog phase were used in all experiments reported herein. ["HIThymidine uptake was determined essentially as described by Bastian et al.
(24). Briefly, cells (0.5-1 X lo5 ml/well) were incubated with [:'HI thymidine (1 pCi) in the culture medium without serum, with or without the agents to be studied, for 1 h a t 37 "C, and the uptake was stopped by addition of 1 ml of ice-cold 10% trichloroacetic acid. Cells were collected by centrifugation, washed twice with 1 ml each of 5% trichloroacetic acid, and finally dissolved in 0.25 ml of 0.5% deoxycholic acid in 0.1 N NaOH for radioactivity determination. About 5-15% of the total ["Hlthymidine was incorporated into cells under the experimental conditions. Cell viability was determined by exclusion of 0.2% trypan blue dye. Cells were stained with Wright-Giemsa stain for microscopic examination and estimation of cell population caught a t mitotic phase.

Okadaic Acid and Mitotic Arrest
Na, pH 7.4) were treated with 0.003% saponin for 5 min a t 37 "C, pelleted hy centrifugation, and washed once with 10 ml of the medium. Permeabilized cells were suspended in 1 ml of the medium and aliquots (0.25 ml) of the cell suspension were incubated (in 0.5 ml) lor 20 min a t 30 "C in the presence of 10 p M [y-?>P]ATP (containing about 0.5-1 X 10' cpm) and 100 p~ CaCI,, with or without 10-1,000 nM okadaic acid, 100 nM TPA, and 30 p~ ALP (Fig. 3). Cells were then recovered by centrifugation, suspended in 0.5 ml of ice-cold homogenization buffer (10 mM Tris/HCI, pH 7.4, 240 mM sucrose, 2 mM ECTA, 2.5 mM MgCI,, 10 mM mercaptoethanol, and 1 mM phenylmethanesulfonyl fluoride), and homogenized (20 up-and-down strokes) with a tight-fitted Teflon-glass homogenizer. Aliquots (0.4 ml) of the homogenate were centrifuged a t 1,000 X g for 5 min to yield the post-nuclear (supernatant) and crude nuclear (pellet) fractions, and the nuclear fraction was taken up in 0.4 ml of the homogenization buffer. To the fractions were added 0.25 volume of gel sample buffer (25), and aliquots (80 pl, corresponding to 6.4 X lo5 cells) of the samples were suhjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and autoradiograph as described elsewhere (25). The cell permeabilization procedure described above was predetermined to be optimum because of high cell viability and [y-'"PIATP uptake and protein phosphorylation. Alternatively, cellular protein phosphorylation was carried out using the native (saponinuntreated) K562 cells prelabeled with ?'P, (Fig. 4A), according to the procedures we described earlier for HL60 cells (25). Finally, protein phosphorylation was also examined in the cell-free system for comparison. Briefly, K562 cells (2 X IO' ) were homogenized in 1 ml of homogenization buffer and centrifuged to yield the nuclear and postnuclear fractions. Aliquots (0.25 ml) of the fractions were incubated (in 0.5 ml) for 20 min a t 30 "c with 0.5 mM CaCI, (in excess of EGTA), 10 pg/ml of phosphatidylserine, 10 mM MgCI,, and 10 p~ [y-:"P] ATP (ahout 2 X 10' cpm), with or without TPA or okadaic acid (Fig. 4R), essentially as described (26).

RESULTS
Effects of okadaic acid on cell growth, indicated by incorporation of ["Hlthymidine into DNA, and viability of K562 cells were examined. The toxin was found to markedly inhibit cell growth a t 10 and 15 nM, whereas it had no effect a t 1 and 5 nM (Fig. lA). Inhibition of thymidine uptake was accompanied by a progressive loss of cell viability (Fig. 1B). It was noted that at days 1 and 2 when thymidine uptake was inhibited by okadaic acid 10 and 15 nM, cell viability was essentially unaffected; cell death began only after day 3, indicating that its action was cytostatic, not acutely cytocidal.
In light of okadaic acid inhibition of PPl/PPZA (2,6,7), enzymes reported to be essential for mitosis ( examined microscopically the cells treated with the toxin under the conditions predetermined to be inhibitory to cell growth without causing cell death. In the control experiment ( Fig. 2 A ) , mitotic cells such as the one caught at the prophase/ metaphase were observed only occasionally. When cells were treated with 10 nM colchicine for 24 and 48 h (Fig. 2, R and C ) , high fractions of the cell population were found to be arrested a t metaphase characterized by chromosome overcondensation, a result consistent with the well known effect of the microtubule-depolymerizingalkaloid. In contrast, treatment with 10 nM okadaic acid for 24, 48, and 72 h (Fig. 2, D nM okadaic acid with 5 or 10 nM colchicine did not produce a greater effect than that produced by either alone. It has been reported previously that mitotic arrest by colchicine is either reversible or irreversible depending upon the treatment conditions (27). In the present study we observed that K562 cells which had been preincubated with 10 nM okadaic acid for up to 24 h restored the normal rate of thymidine uptake when the cells were subsequently cultured for an additional 4 days in the toxin-free medium ( Table I).
The findings strongly suggested that the cytotoxicity of okadaic acid was reversible under the experimental conditions.
Protein phosphorylation experiments using saponin-permeabilized K562 cells indicated that okadaic acid and TPA stimulated phosphorylation of many common proteins (Fig.  3). Okadaic acid was highly effective a t 10 nM (lane 3 ) , the lowest concentration tested, and the phosphorylation was further stimulated by 100 and 1,000 nM of the toxin (lunes 4 and 5 ) . It was noted that 100 nM okadaic acid (lane 4 ) appeared to be a t least as effective as 100 nM TPA (lune 2). and a combination of the two agents (lune 11 )  achieved by 1,000 nM okadaic acid (fane 5). ALP (30 PM), a specific protein kinase C inhibitor (26), seemed to preferentially inhibit phosphorylation stimulated by TPA compared with that by okadaic acid. Further analysis of phosphoproteins revealed that the majority of the '"P-labeled roteins, whose phosphorylation was stimulated by okadaic acid or TPA, was found in the nuclear fraction but not in the post-nuclear fraction (data not shown).
TPA and okadaic acid increased phosphorylation of similar proteins in the nuclear fraction of the native (not saponinpermeabilized) K562 cells (Fig. 4A). It was further noted that both agents markedly increased phosphorylation of many similar proteins in the post-nuclear fractions (Fig. 4A), in contrast to low phosphorylation of post-nuclear proteins in the permeabilized cells mentions earlier. Again, TPA and okadaic acid increase phosphorylation of similar proteins when the nuclear fraction derived from the native K.562 cells e " " " was subjected to phosphorylation (Fig. 4R). Phosphorylation of proteins in the isolated post-nuclear fraction was stimulated by okadaic acid hut not by TI'A ( Fig. 4H). Distinct differences in phosphorylation patterns were noted under various conditions. For example, the 24-and 23-kDa species in the permeabilized intact cells (Fig. 3), the 38-and 36-kDa species in the native intact cells (Fig. 4A), and the 8 5 . 80-. and 53-kDa species in the isolated subcellular fraction (Fig.  4R) were specific major phosphoproteins detected in the presence of TPA or okadaic acid, clearly underscoring the notion that cellular protein phosphorylation in situ (or in uiuo) and in vitro could he qualitatively or quantitatively different.

DISCUSSION
The present findings indicated for the first time that okadaic acid inhibited growth (thymidine uptake), caused mitotic arrest, and stimulated phosphorylation of a number of nuclear proteins. All of the above cellular effects occurred at a toxin concentration as low as 10 nM, strongly suggesting that inhibition of PPl/PPZA ( K , of okadaic acid = 0.05-10 nM) might be a causal factor. Because higher concentrations (100 and 1,000 nM) of the toxin produced greater effects in cells and homogenates, its possible actions on other protein phosphatases or unidentified targets could not be completely excluded.
Two other major features concerning our protein phosphorylation studies merited comments. First, okadaic acid and TPA stimulated phosphorylation of many of the same cellular proteins. These findings showed that essentially all of the substrate proteins for protein kinase C in K562 cells were also substrates for PPl/PP2A, consistent with the earlier suggestion that the phosphatases are likely to be the chief enzymes that reverse the action of protein kinase C (6). In this respect, it was puzzling that TPA, which stimulated phosphorylation of an apparently similar set of proteins as did okadaic acid, failed to arrest K562 cells at mitosis (data not shown). Second, many proteins whose phosphorylation was stimulated by okadaic acid and/or TPA were localized in the nuclear fraction. The observations were of particular significance because ~34'"' protein kinase and PI'I/PP2A, all essential for mitosis, are acting primarily on nuclear proteins. The identities of these nuclear phosphoproteins remained unclear, however.
We found that 10 nM okadaic acid caused mitotic arrest of K562 cells, which was characterized by chromosome overcondensation and chromatin degeneration (Fig. 2). These observations were surprisingly similar to those of Axton P t 01. (22) on ganglions from drosophila which has a mutated gene encloding PP1, indicating that inhibition of PP1 (and/or PP2A) by the toxin and absence of the enzyme itself could lead to a similar end result. The findings appeared to support the contention that PPl/PP2A might be required in inactivating p34'""' or reversing the effects of the kinase after metaphase (22). It was not totally unexpected, therefore, that okadaic acid inhibition of PPl/PI"2A could interfer with normal progression of mitosis from metaphase to anaphase and beyond. Clearly, phosphorylation and dephosphorylation of certain proteins, occurring in an orderly manner, are essential for normal mitotic process. I t has been reported that reorganization of interphase microtubules as a mitotic spindle in starfish embryos requires both p34""' and okadaic acidsensitive PPl/PPZA activities (28). In this regard, it is of interest to investigate whether the toxin-induced mitotic arrest of K562 cells would exhibit abnormal spindle morphology similar to that seen in the drosophila mutant lacking P P I (22).