Identification of a Novel Phosphorylation Site on Histone H3 Coupled with Mitotic Chromosome Condensation*

Histone H3 (H3) phosphorylation at Ser10 occurs during mitosis in eukaryotes and was recently shown to play an important role in chromosome condensation inTetrahymena. When producing monoclonal antibodies that recognize glial fibrillary acidic protein phosphorylation at Thr7, we obtained some monoclonal antibodies that cross-reacted with early mitotic chromosomes. They reacted with 15-kDa phosphoprotein specifically in mitotic cell lysate. With microsequencing, this phosphoprotein was proved to be H3. Mutational analysis revealed that they recognized H3 Ser28phosphorylation. Then we produced a monoclonal antibody, HTA28, using a phosphopeptide corresponding to phosphorylated H3 Ser28. This antibody specifically recognized the phosphorylation of H3 Ser28 but not that of glial fibrillary acidic protein Thr7. Immunocytochemical studies with HTA28 revealed that Ser28 phosphorylation occurred in chromosomes predominantly during early mitosis and coincided with the initiation of mitotic chromosome condensation. Biochemical analyses using32P-labeled mitotic cells also confirmed that H3 is phosphorylated at Ser28 during early mitosis. In addition, we found that H3 is phosphorylated at Ser28 as well as Ser10 when premature chromosome condensation was induced in tsBN2 cells. These observations suggest that H3 phosphorylation at Ser28, together with Ser10, is a conserved event and is likely to be involved in mitotic chromosome condensation.

Histone H3 (H3) phosphorylation at Ser 10 occurs during mitosis in eukaryotes and was recently shown to play an important role in chromosome condensation in Tetrahymena. When producing monoclonal antibodies that recognize glial fibrillary acidic protein phosphorylation at Thr 7 , we obtained some monoclonal antibodies that cross-reacted with early mitotic chromosomes. They reacted with 15-kDa phosphoprotein specifically in mitotic cell lysate. With microsequencing, this phosphoprotein was proved to be H3. Mutational analysis revealed that they recognized H3 Ser 28 phosphorylation. Then we produced a monoclonal antibody, HTA28, using a phosphopeptide corresponding to phosphorylated H3 Ser 28 . This antibody specifically recognized the phosphorylation of H3 Ser 28 but not that of glial fibrillary acidic protein Thr 7 . Immunocytochemical studies with HTA28 revealed that Ser 28 phosphorylation occurred in chromosomes predominantly during early mitosis and coincided with the initiation of mitotic chromosome condensation. Biochemical analyses using 32 P-labeled mitotic cells also confirmed that H3 is phosphorylated at Ser 28 during early mitosis. In addition, we found that H3 is phosphorylated at Ser 28 as well as Ser 10 when premature chromosome condensation was induced in tsBN2 cells. These observations suggest that H3 phosphorylation at Ser 28 , together with Ser 10 , is a conserved event and is likely to be involved in mitotic chromosome condensation.
During G 2 interphase to M phase, the relaxed interphase chromatin is converted into mitotic condensed chromosomes, a process considered to be essential for the following nuclear division to correctly separate parental genetic information into two daughter cells. However, little is known of mechanisms regulating the packing of DNA into mitotic condensed chromosomes (for a review, see Ref. 1).
Histones are major protein constituents of chromatin in eu-karyotic cells and are divided into two main groups. Core histones are wrapped by DNA as octamers, consisting of two H2A-H2B dimers and a tetramer of (H3⅐H4) 2 (2)(3)(4). The linker histone H1 (H1) binds to each nucleosome near the site where the DNA helix enters and leaves the core histones (for reviews, see Refs. [5][6][7]. Earlier studies focused on the relationship between H1 hyperphosphorylation and chromosome condensation; H1 hyperphosphorylation is temporally associated with entry into mitosis and depends on Cdc2 kinase activity (Refs. 8 -13; for reviews, see Refs. 14 and 15). However, recent studies revealed that chromatin condensation can occur without H1 hyperphosphorylation (16,17) or H1 itself (18 -20). Therefore, the biological significance of mitotic H1 hyperphosphorylation remains unknown. In contrast to H1 phosphorylation, H3 phosphorylation level is negligible during interphase and reached maximum during mitosis in mammalian cells (9). In Tetrahymena, H3 phosphorylation occurred only in the mitotic micronucleus, which divides mitotically, but not in macronucleus, which divides amitotically without obvious chromosome condensation (21). When mammalian interphase cells were fused with mitotic cells, premature chromosome condensation (PCC) 1 was accompanied by a significant increase in the level of H3 phosphorylation (22,23). PCC induced by the phosphatase inhibitor was associated with enhanced H3 phosphorylation (16,17). H3 phosphorylation also occurred together with PCC induced at a nonpermissive temperature in tsBN2 baby hamster kidney cells (24). These observations indicate that H3 phosphorylation may coincide with chromosome condensation.
Mitotic H3 phosphorylation is known to occur at Ser 10 in the amino-terminal tail (Refs. 9 and 25; for reviews, see Refs. 14 and 26). Using a site-and phosphorylation state-specific antibody (for a review, see Ref. 27) for this site, it was demonstrated that there is a tight correlation between H3 Ser 10 phosphorylation and mitotic chromosome condensation (28,29). This Ser 10 phosphorylation is conserved through eukaryotes (29). Recently, Wei and co-workers (30) produced strains Tetrahymena, which carry the H3 mutant gene (S10A) as the only gene encoding the major H3 protein. By using these strains, they demonstrated that H3 Ser 10 phosphorylation is required for proper chromosome condensation and segregation. However, the S10A mutation in H3 did not completely disrupt chromosome condensation. Thus, they proposed that other factor(s) or other H3 posttranslational modification(s), including phosphorylation at other site(s), may also play important roles in chromosome condensation (30).
In the present study, we have identified Ser 28 as the mitotic H3 phosphorylation site, using immunological and biochemical approaches. H3 Ser 28 phosphorylation coincided with mitotic chromosome condensation in several types of cultured cells. In addition, this phosphorylation was also observed when PCC was induced in early S phase-synchronized tsBN2 baby hamster kidney cells.
Cell Culture-U251 human glioma cells, HeLa human cervical carcinoma cells, NIH 3T3 mouse fibroblastic cells, Madin-Darby bovine kidney cells, baby hamster kidney (BHK) cells, and tsBN2 cells (a temperature-sensitive mutant of BHK cells; Ref. 33) were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum in a humidified atmosphere of 5% CO 2 in air.
Immunofluorescence Microscopy-Cells growing on glass coverslips were fixed with 3.7% formaldehyde in ice-cold PBS for 10 min and then treated with methanol at Ϫ20°C for 10 min. Incubation with primary antibodies diluted in PBS containing 1% sucrose and 1% bovine serum albumin was for 2 h at 37°C. After three washes with PBS, cells were incubated for 1 h with appropriate secondary antibodies diluted 1:100 and subsequently washed with PBS. Then DNAs were stained with 0.5 g/ml propidium iodide (Sigma) or 0.5 g/ml 4Ј,6-diamidino-2-phenylindole (Roche Molecular Biochemicals) for 10 min at room temperature.
Fluorescently labeled cells were examined, using an Olympus BH2-RFCA microscope or an Olympus LSM-GB200 confocal microscope.
Preparation of Interphase or Early Mitotic Cells-Just before cells reached confluence, the cells were arrested in early mitosis by adding 50 ng/ml nocodazole for 12 h. Early mitotic cells were collected by mechanical shake off, and the adherent cells served as interphase cells. Cell lysates for SDS-PAGE were prepared as follows. Interphase or mitotic cells were rinsed twice with PBS, treated with 10% trichloroacetic acid for 1 h at 4°C, and then collected. After centrifugation, cell pellets were dissolved in Laemmli's sample buffer, with brief sonication.
Production of Wild Type or Mutant Histone H3.3 Protein-A fulllength human wild type (WT) histone H3.3 cDNA (pT7T3D-Pac) was obtained from American Type Culture Collection (Manassas, VA). Mutagenesis of Ser 10 to Ala (S10A), Ser 28 to Ala (S28A), or Thr 118 to Ala (T118A) in histone H3.3 was done using the QuikChange site-directed mutagenesis kit (Stratagene) and mutagenic primers. The open reading frame of each cDNA was amplified by polymerase chain reaction with a 5Ј-primer (5Ј-AAGGATCCGGCTCGTACAAAGCAGACTGCCCGC-3Ј) and a 3Ј-primer (5Ј-AGGTCGACTTAAGCACGTTCTCCACGTATGCG-G-3Ј), which produce a BamHI site at a 5Ј-end of and a SalI site at a 3Ј-end of each cDNA. Each amplified cDNA was inserted into the BamHI-SalI site of plasmid pQE-31 (Qiagen). The sequence of each cDNA was confirmed by sequencing, using the dideoxy termination method and a DNA sequencer (Applied Biosystems). For bacterial expression, Escherichia coli strain BL21(DE3)pLysS (Stratagene) was transformed with each plasmid. WT or each mutant (S10A, S28A, or T118A) histone H3.3 was expressed in E. coli as His 6 -tagged protein and purified on nickel-nitrilotriacetic acid-agarose resin (Qiagen) under 8 M urea denaturing conditions according to the manufacturer's protocol. Before the phosphorylation assay, each His 6 -tagged protein was dialyzed with the renaturing buffer (2 mM EGTA, 20 mM 2-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, 20 mM Tris-Cl (pH 8.0)) for at least 12 h at 4°C.
Phosphorylation Assay-The catalytic subunit of cAMP-dependent protein kinase (A kinase) was prepared from bovine heart by the method of Beavo et al. (34). Cdc2 kinase was purified from mouse mammary tumor (FM3A) cells, as described (35). The phosphorylation reaction was performed for 30 min at 37°C in 20 l of 25 mM Tris-Cl (pH 7.5), 1 mM MgCl 2 , 100 M ATP, 0.1 M calyculin A, and 110 g/ml wild type or mutant (S10A, S28A, or T118A) His 6 -tagged histone H3.3 in the presence of either 5 g/ml A kinase or 0.5 g/ml Cdc2 kinase, with or without [␥-32 P]ATP. Reaction mixtures were boiled in Laemmli's sample buffer and subjected to SDS-PAGE.
Extraction of Histones from Mitotic HeLa Cells-Mitotic HeLa (S-3) cells were prepared as described above. Aliquots of cells were labeled with [ 32 P]orthophosphate at 40 Ci/ml for an additional 3 h (17). Cells were treated with 10% trichloroacetic acid and then collected. Cell pellets were partially lysed with the solution (80 mM NaCl, 20 mM EDTA, and 1% Triton X-100) in the presence or absence of 100 nM okadaic acid. Total histones were extracted three times with 0.4 N sulfaric acid for 10 min and precipitated with 4 volumes of ethanol as described (24).

One Type of Rat Monoclonal Antibodies Raised against a
Phosphopeptide Corresponding to GFAP Phosphorylation at Thr 7 Cross-reacts with Early Mitotic Chromosomes-We earlier reported that GFAP is phosphorylated at Thr 7 , Ser 13 , and Ser 34 by Rho-associated kinase (Rho-kinase) at cleavage furrow during cytokinesis (37)(38)(39). We developed a site-and phosphorylation state-specific monoclonal antibody for Thr 7 on GFAP, using synthetic phosphopeptide PG7 (CERRRVpT 7 SAARR; also see Fig. 3A) as an antigen for immunization. Then we obtained two types of rat monoclonal antibodies that recognize GFAP phosphorylated at Thr 7 but neither unphosphorylated GFAP nor GFAP phosphorylated at other sites including Ser 8 , Ser 13 , and Ser 34 (Fig. 1, A and B). One type (referred to as TMG7) specifically reacted with GFAP at the cleavage furrow during cytokinesis (late mitosis) in U251 glioma cells (Fig. 1C). Surprisingly, the immunoreactivity of the other type (referred to as TA22) was observed not only at the cleavage furrow during late mitosis but also at chromosomes during early mitosis in glioma cells (Fig. 1C). Since GFAP is an intermediate filament protein expressed specifically in the cytoplasm of astroglial cells (40 -42), TA22 may recognize the antigen(s) other than GFAP during early mitosis.
TA22 Recognizes the Phosphorylation of Histone H3 in the Early Mitotic Cell Lysate but Not in the Interphase Cell Lysate-To identify the TA22-reacted protein in the chromosome during early mitotic phase, Western blot analysis of U251 cell lysates was carried out. As shown in Fig. 2A, TA22 immunoreacted with about 15-kDa protein in the early mitotic cell lysate but not in the interphase cell lysate. This TA22 immunoreactivity for the mitotic 15-kDa protein disappeared after treatment of the transferred membrane with the protein phospha-tase (-PPase; a dual specificity phosphatase), suggesting that TA22 recognized the phosphoepitope in the 15-kDa protein (Fig. 2B). To clarify the molecular identity, we determined three peptide sequences derived from this immunoreacted protein: STELLIR, RVIT, and DIQLARRIRGER (Fig. 2C). They are found within the sequence of the human histone H3.1 or H3.3 (for a review, see Ref. 43). Since histone H3 is one of core histones wrapped by DNA as octamers in eukaryotic nuclei (2)(3)(4), TA22 may recognize H3 phosphorylation in the chromosomes during early mitosis.
TA22 Recognizes H3 Phosphorylation at Ser 28 but Not at Ser 10 -We next examined whether TA22 recognized recombinant His 6 -tagged histone H3.3 phosphorylated by A kinase or Cdc2 kinase in vitro. As shown in Fig. 2D, TA22 reacted with H3.3 phosphorylated by A kinase but not by Cdc2 kinase. H3 was reported to be phosphorylated at Ser 10 , Ser 28 , and Thr 118 by A kinase in vitro (44,45). Ser 10 was also identified as an in vivo H3 phosphorylation site (9, 25). Then we produced several mutant proteins of histone H3.3 at each of the phosphorylation sites (Ser 10 , Ser 28 , or Thr 118 3 Ala (S10A, S28A, or T118A,  200). B, after early mitotic U251 cell lysates were resolved by SDS-PAGE, the gel was transferred onto a PVDF membrane as described above. The membrane was preincubated with 1ϫ protein phosphatase buffer (50 mM Tris-Cl (pH 7.5), 0.1 mM Na 2 EDTA, 5 mM dithiothreitol, 0.01% Brij 35, 2 mM MnCl 2 ) in the absence (ϩ buffer) or the presence (ϩ -PPase) of 100 g/ml protein phosphatase (New England Biolabs Inc.) for 1 h at 30°C. Each treated membrane was immunoblotted with the antibody TA22 (dilution 1:200) and then stained with Coomassie Brilliant Blue. The arrowheads indicate the position of the TA22-immunoreactive band. C, the TA22-immunoreactive band at 15 kDa in the mitotic U251 cell lysate was cut from transferred membrane and digested with lysyl-endopeptidase (57). The resulting peptides were fractionated by C18 column chromatography and subjected to amino acid sequence analysis. Three peptide sequences derived from this 15-kDa protein were determined (square). H3.1 and H3.3 represent the sequence of the human histone H3.1 and H3.3, respectively. D and E, recombinant His 6 -tagged histone H3.3 (wild type; WT) was phosphorylated by A kinase or Cdc2 kinase with or without [␥-32 P]ATP in vitro as described under "Experimental Procedures." Each mutant histone H3.3 in which Ser 10 , Ser 28 , or Thr 118 was changed to Ala (S10A, S28A, or T118A) was phosphorylated by A kinase as described above. Radiolabeled bands were visualized using autoradiography ( 32 P). After SDS-PAGE, nonradioactive samples were stained with Coomassie Brilliant Blue or transferred onto a PVDF membrane. The membrane was immunoblotted with the antibody TA22 (dilution 1:200).

FIG. 1. Immunoreactivities of rat monoclonal anti-PG7 antibodies.
A, reactivity of TM-G7 or TA22 to unphosphorylated GFAP (Control) and GFAP phosphorylated by Rho-associated kinase (Rhokinase). Recombinant GFAP and GFAP phosphorylated by Rho-kinase was prepared as described (38). After SDS-PAGE (100 ng in each lane), samples were transferred onto a polyvinylidene difluoride filter (PVDF) membrane. The membrane was immunoblotted with the antibody TM-G7 or TA22 (dilution 1: 200) and then stained with Coomassie Brilliant Blue (CBB). B, specificity of TM-G7 or TA22, determined by a competition assay. GFAP phosphorylated by Rho-kinase was immunoblotted with TM-G7 or TA22 (dilution 1:200) after preincubation with synthetic peptides (50 g/ml G7, PG7, PG8, PG13, or PG34). As a control, the phosphorylated GFAP was immunoblotted with TM-G7 or TA22 (dilution 1:200) after preincubation with TBS-T. The arrowheads indicate the position of GFAP phosphorylated by Rho-kinase. C, confocal micrographs of U251 glioma cells stained with TM-G7 or TA22 (green). DNAs were stained with propidium iodide (red). Images represent projections of Z-series scans. Bar, 10 m. respectively)). TA22 reacted with wild type (WT), S10A and T118A phosphorylated by A kinase but not with S28A phosphorylated by A kinase (Fig. 2E). These results suggested that TA22, a monoclonal antibody originally raised against phospho-Thr 7 on GFAP, recognized H3 phosphorylation at Ser 28 but not at the other sites including Ser 10 , which is a well known mitotic H3 phosphorylation site.
Production and Characterization of a Site-and Phosphorylation State-specific Antibody HTA28 for H3 Ser 28 -The rat monoclonal antibody TA22 recognized the phosphorylation of at least two proteins including H3 Ser 28 ( Fig. 1 and 2), much like monoclonal antibody MPM-2, which can recognize phosphorylated forms of several proteins during mitosis (46). Therefore, with this antibody it is difficult to obtain clear data concerning H3 Ser 28 phosphorylation in cells, because there is the possibility that it might recognize several phosphoproteins, including H3, in mitotic chromosomes.
To overcome this difficulty, we produced a site-and phosphorylation state-specific antibody HTA28 for H3 Ser 28 , using a synthetic peptide PH28 (CKAARKpS 28 APATGGV; Fig. 3A), corresponding to H3 phosphorylation at Ser 28 as an antigen for immunization. This rat monoclonal antibody HTA28 specifically recognized the phosphorylation of H3 Ser 28 (Fig. 3B) but not that of GFAP-Thr7 (Fig. 3C). To compare H3 Ser 28 phosphorylation with H3 Ser 10 phosphorylation, we also prepared the rabbit polyclonal antibody ␣PH10, which can specifically recognize H3 phosphorylation at Ser 10 (Fig. 3B).
H3 Phosphorylation Occurs Not Only at Ser 10 but Also at Ser 28 during Mitosis- Fig. 4A shows various types of cells doubly labeled with ␣PH10 and HTA28. ␣PH10 stained with mitotic nuclei (arrowheads) and some interphase nuclei (arrows), probably late S and/or G 2 nuclei (28,47). In contrast, HTA28 specifically immunoreacted with mitotic nuclei (arrowheads) but not with interphase nuclei. Western blot analyses revealed that HTA28 and ␣PH10 immunoreacted with the band at 15 kDa corresponding to the position of H3 in the mitotic cell lysate (Fig. 4B). Although ␣PH10 also immunoreacted weakly with the band at 15 kDa in interphase cell lysate, no HTA28 immunoreactivity was observed in interphase cell lysate (Fig. 4B). In HeLa and U251 cells, HTA28 immunoreactivity appeared at the onset of mitosis (prophase), was maintained until metaphase, decreased at the beginning of an-aphase, and disappeared during late anaphase (Fig. 5). Similar results were obtained, using other types of cultured cells including Madin-Darby bovine kidney, NIH3T3, and BHK cells (data not shown; also see Fig. 4). This antibody has no reactivity with cytoplasmic proteins, including GFAP in glioma cells (Fig. 5A). These results suggested that H3 may be phosphorylated not only at Ser 10 but also at Ser 28 during mitosis. This temporal distribution of H3 phosphorylation at Ser 28 was similar to that at Ser 10 , except that H3 Ser 10 phosphorylation may occur in some interphase nuclei (Fig. 4).
Next, we examined mitotic H3 phosphorylation sites, using total histones extracted from mitotic HeLa cells. Total histones were extracted with or without okadaic acid (one of the serine/ threonine phosphatase inhibitors) from mitotic HeLa cells. As shown in Fig. 6A, ␣PH10 immunoreacted with the band at 15 kDa corresponding to the position of H3 in mitotic cell lysate (lane d) and in total histones (lanes e and f). In contrast, although HTA28 immunoreacted with the band at 15 kDa in mitotic cell lysate (lane g), no signal or only a faint signal at 15 kDa was detected in total histones extracted without okadaic acid (lane h). However, the HTA28-immunoreactive band at 15 kDa was observed in total histones extracted with okadaic acid (lane i). Fig. 6B shows tryptic phosphopeptide mapping of H3 histones purified from 32 P-labeled total histones by HPLC. Ser 10 -phosphorylated peptides (arrowheads) were detected, regardless of whether total histones were extracted with (b) or without (a) okadaic acid from 32 P-labeled mitotic HeLa cells. (Fig. 6B). However, Ser 28 -phosphorylated peptide (arrow) was detected only when the extraction procedure was performed in the presence of okadaic acid (Fig. 6B). These results indicated that H3 may be dephosphorylated at Ser 28 (but not excessively at Ser 10 ) on the extraction procedure of total histones, and our immunological observations that H3 phosphorylation occurs not only at Ser 10 but also at Ser 28 during mitosis were thus confirmed.
H3 Ser 28 Phosphorylation Coincided with Chromosome Condensation-H3 was phosphorylated at Ser 28 specifically during early mitosis. This phosphorylation appeared to correlate spatiotemporally with the initiation of the chromosome condensation (Figs. 4 and 5). To further confirm this possibility, we used tsBN2 cells, a temperature-sensitive mutant of BHK cells: these cells exhibit PCC at nonpermissive temperature (33). PCC under this condition mimics normal chromosome condensation, as evidenced by premature activation of Cdc2 kinase and the association of mitosis-specific antigen recognized by a monoclonal antibody MPM-2 with PCC chromosome (48). To determine if H3 phosphorylation occurs at Ser 10 and Ser 28 accompanied with PCC, tsBN2 and BHK parental cells were synchronized in early S phase by adding aphidicolin at 32.5°C (permissive temperature). Upon temperature shift up to 41°C (nonpermissive temperature) for 4 h, PCC was induced in about 10% of tsBN2 cells, but not in any BHK cells (Fig. 7). Upon incubating at 32.5°C (permissive temperature) for an additional 4 h, no PCC was observed in each cell line (Fig. 7). H3 phosphorylation at Ser 10 and Ser 28 was observed only in tsBN2 cells where PCC was induced at 41°C (Fig. 7). Together with data shown in Fig. 4 and 5, these results suggested that H3 phosphorylation at Ser 28 , as well as Ser 10 , correlated closely with chromosome condensation. DISCUSSION A major new finding in this study is that histone H3 is phosphorylated not only at Ser 10 but also at Ser 28 during mitosis. This H3 phosphorylation at Ser 28 , together with that at Ser 10 , correlates with mitotic chromosome condensation in various types of cultured cells and with PCC induced in tsBN2 cells. FIG. 3. Specificity of antibodies ␣PH10 and HTA28 analyzed by immunoblotting. A, amino acid sequences of the synthetic phosphopeptides PG7 and PH28 corresponding to the phosphorylation of GFAP at Thr 7 and H3 at Ser 28 , respectively. A phosphoresidue is indicated by an amino acid residue and P within a circle. B, reactivity of ␣PH10 and HTA28 to wild type (WT) or each mutant histone H3.3 (S10A, S28A, or T118A), phosphorylated by A kinase. The membrane was immunoblotted with the antibody ␣PH10 (dilution 1:10,000) or HTA28 (dilution 1:200). C, reactivity of TA22 and HTA28 to unphosphorylated GFAP (Control) and GFAP phosphorylated by Rho-kinase (Rho-kinase). The membrane was immunoblotted with the antibody TA22 or HTA28 (dilution 1:200) and then stained with Coomassie Brilliant Blue.
H3 was considered to be phosphorylated only at Ser 10 during mitosis (9,25,28,29). Why had only Ser 10 been identified as the mitosis-specific H3 phosphorylation site? It may be due to the H3 purification procedure from mitotic cells. Most studies determining the in vivo H3 phosphorylation site used the total histones extracted from the isolated chromatin in 32 P-labeled mitotic cells. By using the method similar to the reported one (9,25), we could also detect only Ser 10 as the mitosis-specific H3 phosphorylation site (Fig. 6). However, both Ser 10 and Ser 28 were detectable when the extraction of total histones from mitotic cells was done in the presence of phosphatase 1 and 2A inhibitor, okadaic acid (Fig. 6). Thus, H3 may be dephosphorylated at Ser 28 (but not excessively at Ser 10 ) upon extracting total histones from mitotic cells. Therefore, Ser 28 may be first identified as one of the mitosis-specific phosphorylation sites by using the site-and phosphorylation state-specific antibody.
Identification of a novel H3 phosphorylation site raised the question about the ratio of phosphorylated H3 Ser 28 to phosphorylated H3 Ser 10 or total H3 histones in early mitotic cells. Biological analyses revealed that the phosphorylation level of H3 Ser 28 was lower than that of H3 Ser 10 in early mitotic cells (Fig. 6B). However, H3 Ser 28 was more sensitive to phosphatase(s) than H3 Ser 10 and dephosphorylated to some extent during the purification even in the presence of the phosphatase inhibitor (Fig. 6A, lanes g and i). Thus, we consider that Ser 28 phosphorylation also occurs to some extent, although the phosphorylation level of H3 Ser 28 may be slightly lower than that of H3 Ser 10 .
H3 Ser 10 phosphorylation correlated closely with chromosome condensation and was required for proper chromosome condensation and segregation (see Introduction). Recently, this phosphorylation was reported to be required for the initiation, but not maintenance, of mammalian chromosome condensation (47). H3 Ser 28 phosphorylation occurred specifically during early mitosis, at least in mammalian cells and correlated closely with mitotic chromosome condensation and PCC induced in tsBN2 cells (Figs. 4 -7). While the biological significance of H3 Ser 28 phosphorylation remains unknown, these observations raised the possibility that H3 phosphorylation at Ser 28 , as well as Ser 10 , may play an important role in chromosome condensation in mammalian cells.
SMC (structural maintenance of chromosomes) proteins are also revealed to play important roles in chromosome condensation; these proteins are probably directly involved in chromosome condensation (for reviews, see Refs. 1 and 49 -52). However, the relationship between H3 phosphorylation and SMC proteins remains unknown. Recently, mitotic H3 phosphorylation was reported to promote the disassociation between the H3 amino-terminal tail and DNA (53). Since the H3 amino-terminal tail emerged from the nucleosome at the entry and exist points of DNA (4), the phosphorylation of the H3 amino-terminal tail at Ser 10 and Ser 28 might reduce their affinity for DNA  ; lanes a, d, and g) and total histones (2 g of protein/lane) were loaded on lanes and resolved by SDS-PAGE. The gel was stained with Coomassie Brilliant Blue (CBB) or transferred onto a PVDF membrane. The membrane was immunoblotted with the antibody ␣PH10 (dilution 1:10,000) and HTA28 (dilution 1:200). B, total histones were extracted with (b) or without (a) okadaic acid from 32 P-labeled mitotic HeLa cells. Then H3 histones were separated from total histones by reverse-phase HPLC and subjected to tryptic phosphopeptide mapping, as described under "Experimental Procedures" (horizontal dimension, electrophoresis with butanol/acetic acid/water/pyridine (50:25:900:25), pH 4.7) at 550 V for 50 min (vertical dimension, ascending chromatography in butanol/acetic acid/water/pyridine (48.8:15.2:60.4:75.6)). The bottom left corner shows the spot origin. The arrowheads and an arrow indicate the 32 P spot identified as peptides containing Ser 10 and Ser 28 residue, respectively, in an earlier study (45). and facilitate the targeting of condensing factors including SMC proteins.
H3 Ser 10 phosphorylation also occurred in some interphase nuclei where H3 Ser 28 phosphorylation was not observed (Fig.  4, arrows). Other groups reported similar observations that H3 Ser 10 phosphorylation initiated during late S phase or G 2 phase in mammalian cells (28,47). In contrast, H3 Ser 28 phosphorylation appeared to be initiated at the onset of mitosis, prophase ( Fig. 4 and 5). These observations raise the question about mechanisms regulating the phosphorylation at these two sites on H3. We consider two possibilities. 1) One possibility is the existence of at least two H3 kinase activities in a cell cycle-dependent manner. Protein kinase(s) activated during S or G 2 interphase might phosphorylate H3 at Ser 10 but not at Ser 28 . Since Ser 10 and Ser 28 on H3 were phosphorylated in tsBN2 cells where Cdc2 kinase was prematurely activated (Fig. 7), other H3 kinase(s) might phosphorylate H3 at both Ser 10 and Ser 28 after the activation of Cdc2 kinase. This mitotic H3 kinase might not be Cdc2 kinase, because H3 Ser 28 was not phosphorylated by Cdc2 kinase in vitro (Fig. 2D). 2) The second possibility is a different site specificity of H3 phosphatase(s). Since H3 was dephosphorylated exclusively at Ser 28 during the extraction procedure of total histones (Fig. 6), Ser 28 might be more sensitive to phosphatase(s) than Ser 10 even in cells. Thus, during S or G 2 interphase, the Ser 28 phosphorylation level might be too low to detect, although both Ser 10 and Ser 28 could be phosphorylated by H3 kinase activity. Since phosphatase 1 was reported to be phosphorylated and inactivated by Cdc2 kinase during mitosis (54 -56), the phosphorylation level at Ser 28 as well as Ser 10 might be elevated by inactivation of phosphatase(s) during mitosis.
In conclusion, we first found that H3 Ser 28 phosphorylation occurred specifically during early mitosis and coincided with chromosome condensation. Further analyses of protein kinase(s) and phosphatase(s) regulating H3 phosphorylation level at Ser 10 and Ser 28 will help to elucidate the molecular mechanism of some mitotic events including chromosome condensation.