Dissociation of Emerin from Barrier-to-autointegration Factor Is Regulated through Mitotic Phosphorylation of Emerin in a Xenopus Egg Cell-free System*

Emerin is the gene product of STA whose mutations cause Emery-Dreifuss muscular dystrophy. It is an inner nuclear membrane protein and phosphorylated in a cell cycle-dependent manner. However, the means of phosphorylation of emerin are poorly understood. We investigated the regulation mechanism for the binding of emerin to chromatin, focusing on its cell cycle-dependent phosphorylation in a Xenopus egg cell-free system. It was shown that emerin dissociates from chromatin depending on mitotic phosphorylation of the former, and this plays a critical role in the dissociation of emerin from barrier-to-autointegration factor (BAF). Then, we analyzed the mitotic phosphorylation sites of emerin. Emerin was strongly phosphorylated in an M-phase Xenopus egg cell-free system, and five phosphorylated sites, Ser49, Ser66, Thr67, Ser120, and Ser175, were identified on analysis of chymotryptic and tryptic emerin peptides using a phosphopeptide-concentrating system coupled with a Titansphere column, which specifically binds phosphopeptides, and tandem mass spectrometry sequencing. An in vitro binding assay involving an emerin S175A point mutant protein suggested that phosphorylation at Ser175 regulates the dissociation of emerin from BAF.

membrane targeting to chromatin at an early stage of nuclear assembly (1). The interactions between some of the inner nuclear membrane proteins and chromatin are regulated through phosphorylation of these inner nuclear membrane proteins. The phosphorylation mechanisms for LBR and LAP2␣ and 2␤ are well understood (2)(3)(4)(5)(6)(7). LBR directly binds to DNA in vitro and dissociates on phosphorylation by cdc2 kinase and other kinase(s) in a mitotic egg extract (3). LAP2␤ binds to lamin B1 and chromatin, and cell cycle-dependent phosphorylation of LAP2␤ cancels this binding (4). Phosphorylation of these inner nuclear proteins, therefore, is likely to be one of the key mechanisms that control the interactions between the inner nuclear proteins and components of the nuclear lamina as well as chromatin. In this study, we focused on the mitotic phosphorylation of emerin, one of the inner nuclear membrane proteins.
Human emerin is a serine-rich protein exhibiting an apparent mass of 34 kDa on SDS-PAGE (8) and is phosphorylated in a cell cycle-dependent manner (9). Emerin belongs to the LEM (LAP2␤, emerin, MAN1) protein family, whose members have approximately a 40-residue domain named the LEM (10). These proteins directly bind to barrierto-autointegration factor (BAF) (11)(12)(13). BAF is a DNA-bridging protein with a dimer mass of 20 kDa and is highly conserved in metazoans, and the BAF interactions with both DNA and LEM proteins are critical for nuclear membrane targeting to chromatin and chromatin decondensation during nuclear assembly (14). At the onset of mitosis, emerin disperses from the NE to the endoplasmic reticular (ER) network, and is re-localized to the surface of the central region of chromatin, called the "core" region, during telophase (15,16). An LEM domain deletion mutant of emerin cannot be re-localized to this region, suggesting that the binding of emerin to BAF through the LEM domain is essential for this recruitment (16). It is also known that emerin has many binding partners, including transcriptional repressors and intermediate filament proteins (17)(18)(19)(20)(21)(22)(23)(24)(25). In particular, binding to A-type lamin is essential for the retention of emerin in the NE in the interphase. Furthermore, a deletion mutant of emerin residues 95-99 (⌬95-99), which causes Emery-Dreifuss muscular dystrophy and cannot bind to lamin A, exhibits aberrant cell cycle-dependent phosphorylation forms (9). The study also suggested that the phosphorylation of emerin regulates the binding of emerin to lamin A (9). Thus, we were interested in the cell cycle-dependent regulation of the binding of emerin to chromatin and BAF.
We first examined the binding of emerin to chromatin by means of a binding assay involving a GST-fused N-terminal fragment of emerin and chromatin in a Xenopus egg cell-free system. We also analyzed the cell cycle-dependent phosphorylation states and sites of emerin. Phosphopeptides derived from emerin treated with a Xenopus egg mitotic cytosol were separated by means of a Titansphere column, and five phosphorylation sites were identified on mass spectrometry. Furthermore, an in vitro binding assay involving an emerin point mutant revealed that Ser 175 phosphorylation is responsible for the dissociation of emerin from BAF. 6 Tag BAF-Cloning of the nucleoplasmic region of human emerin (⌬TM, amino acid residues 1-213) was performed by PCR. PCR was carried out with a human testis cDNA library using the following primers: 5Ј-CGG-GATCCCCATGGACAACTAGCAGAT-3Ј and 5Ј-CGGGATCCA-GAGCACGGTTTTCAGG-3Ј. The PCR product was digested with BamHI and then inserted into the pBluescript II SK(Ϫ) or pGEX 3X vector (Novagen) at the BamHI site at the 3Ј end of GST. To generate a point mutant with the serine at position 175 replaced with alanine (S175A-⌬TM), a GeneTailor mutagenesis kit (Invitrogen) was used according to the manufacturer's procedure. PCR was carried out with pBluescript II SK(Ϫ)-emerin ⌬TM using the following primers: 5Ј-CT-GTTTCGCCTCCAGGGCCTCCCTGGACC-3Ј and 5Ј-CCTGGAG-GCTGAAACAGGGCGGTAGTCGT-3Ј, followed by verification by DNA sequence analysis. The pBluescript II SK(Ϫ)-S175A-⌬TM was digested with BamHI, and the resulting fragment was inserted into the pGEX 5X-3 vector (Novagen) at the BamHI site. To construct the pGEX 3X-emerin ⌬LT plasmid, pBluescript II SK(Ϫ)-emerin ⌬TM was digested with BglII and BamHI, and the resulting fragment was inserted into the pGEX 3X vector using the BamHI site in the vector. The cDNA clone of human BAF (accession no. BC005942) was purchased from Invitrogen. To obtain His tag BAF, the coding region of BAF was PCRamplified using primers 5Ј-CGGGATCCCGATGACAACCTCCCAA-AAGCA-3Ј and 5Ј-CGGAATTCATGCAAGAGCGAGAATCC-3Ј. The PCR product was digested with BamHI and EcoRI, and then inserted into the pET28c vector at the BamHI and EcoRI sites. The DNA sequences of the inserts in plasmid pGEX 3X-emerin ⌬TM and pET28c-BAF were confirmed using an ALF DNA sequencer (Amersham Biosciences).

Construction of GST-fused Emerin Fragment Proteins and His
Preparation of Xenopus Egg Cytosol Fractions-Xenopus eggs were collected, dejelled, and then lysed to prepare S-phase and M-phase cytosol fractions as described previously (26,27).
Chromatin Binding Assay-Using beads bearing GST-emerin ⌬TM or GST-emerin ⌬LT, the chromatin binding assay was carried out as previously described (3,28) except for the use of 20,000 Xenopus sperm chromatin per assay. When pretreatment of the beads bearing GSTemerin with an Escherichia coli extract containing BAF was necessary, it was carried out as follows. E. coli cells expressing His tag BAF were sonicated vigorously and then centrifuged at 12,000 ϫ g for 10 min. A 50-l aliquot of the supernatant was reacted with beads bearing GSTemerin ⌬TM in binding buffer (20 mM Tris-HCl (pH 7.6), 134 mM NaCl, and 0.1% Tween-20 either containing 125 g/ml DNase or not) at 4°C for 3 h. The beads thus treated were washed three times with binding buffer and then used for the chromatin binding assay.
Typically, Xenopus sperm chromatin, which was demembranated with lysolecithin and subsequently decondensed with heated Xenopus egg cytosol, in 20 l of extraction buffer (10,000 per l in 50 mM HEPES-KOH (pH 7.7), 250 mM sucrose, 50 mM KCl, 2.5 mM MgCl 2 , 2 mM 2-mercaptoethanol), was added to the beads, which were thus treated suspended in 10 l of extraction buffer. After incubation at 4°C for 10 min, the binding reaction was stopped by pipetting 15-l samples onto glass slides spotted with 12 l of a fixing solution (extraction buffer containing 3% formaldehyde and 6 g/ml Hoechst dye 33342). The fixed samples were observed by fluorescence microscopy. 100 -200 beads were observed for every sample, and "the percentage of beads with bound chromatin" was determined. This value was used as an index of the affinity of beads bearing emerin fragments and chromatin. The values in the figures are indicated after subtraction of a blank value. The blank value (Ͻ10%) was determined in each experiment using Sepharose beads with GST-bound instead of GST-emerin fragments. The significance of the pretreatment of beads bearing emerin fragments with various reagents in the chromatin binding assay was evaluated by means of Student's t test (n ϭ 3). In a previous study (28), we compared this method and an established in vitro binding method involving soluble proteins and chromatin, showing that this bead method gives the same results as the established method. Therefore, we used this method to determine the affinity of protein fragments to chromatin in this study.
In Vitro Binding Assay of Emerin and BAF-A supernatant containing His tag BAF was prepared as described above. Beads bearing ϳ10 g of ⌬TM or S175A-⌬TM preincubated with Xenopus cytosol or extraction buffer were washed twice with binding buffer (20 mM Tris-HCl (pH 7.6), 134 mM NaCl and 0.1% Tween 20), and then incubated with BAFexpressed E. coli extract at 4°C for 3 h. The beads were washed three times with binding buffer. The sample thus obtained was subjected to SDS-PAGE and then transferred to a nitrocellulose membrane. BAF bound to beads was detected with an anti-His tag antibody and chemical luminescence. Emerin phosphorylated with a Xenopus egg cytosol fraction was separated by 10% gel SDS-PAGE, and the emerin band was excised. Phosphorylation of the protein was detected with ProQ diamond stain (Molecular Probes) according to the manufacturer's instructions and a reference (29).
In Vitro Dissociation Assay of Emerin and BAF-Beads bearing ϳ10 g of ⌬TM were pretreated with a supernatant containing His tag BAF as described above. After washing three times, the beads thus treated were again treated with buffer or cell cycle-dependent Xenopus egg cytosol fraction at 23°C for 1 h. The beads were washed, subjected to SDS-PAGE, and transferred to a nitrocellulose membrane. BAF bound to the beads was detected as described above. Phosphorylation Assay of Emerin Fragments with a Xenopus Egg Cytosol Fraction-Approximately 3 g of GST-fused emerin bound to glutathione-Sepharose beads was incubated with S-phase and M-phase egg cytosol fractions containing 1 Ci of [␥-32 P]ATP at 23°C for 1 h. After washing twice with phosphate-buffered saline containing 0.05% Tween 20 (PBS-T), the beads were supplemented with 2 mM ATP and then incubated at 4°C for 10 min. The proteins thus treated were separated by SDS-PAGE and visualized by CBB R-250 staining. After drying the gel, phosphorylation was detected with Fuji x-ray film. The emerin ⌬TM bands were excised from the gel and the phosphorylated emerin ⌬TM was quantified by scintillation counting.
Phosphopeptide Mapping-Approximately 30 g of emerin ⌬TM or ⌬LT was phosphorylated as described above, separated by SDS-PAGE, and then transferred to a nitrocellulose sheet. The full-length emerin band was excised, soaked in 0.5% poly(vinyl pyrrolidone) K-30 in 100 mM acetic acid for 30 min at 37°C, and then washed extensively with water. The protein was digested with trypsin or chymotrypsin in 50 mM NH 4 HCO 3 for 16 h at 37°C. The released peptides were dried, dissolved in water, and then loaded onto a cellulose TLC plate (Funacell, Funakoshi Co., Tokyo). Electrophoresis, in the first dimension, was performed at pH 8.9 (1% ammonium carbonate) for 20 min at 1000 V, and ascending chromatography, in the second dimension, was performed with a solvent system of 37.5% 1-butanol, 25% pyridine, and 7.5% acetic acid in water (v/v). The dried plate was exposed to Fuji x-ray film.
Separation of Phosphopeptides Derived from Emerin ⌬TM Treated with Mitotic Cytosol Using a Titansphere Column-This experiment was carried out according to the method of Kuroda et al. (30) with some modification. Beads bearing ϳ100 g of GST-fused emerin ⌬TM were treated with a Xenopus egg M-phase cytosol fraction as described above except that [␥-32 P]ATP was omitted. The beads thus treated were separated by 10% gel SDS-PAGE and visualized by CBB R-250 staining. Emerin ⌬TM bands were excised from the gel and then in-gel digested with chymotrypsin or trypsin in 50 mM NH 4 HCO 3 at 37°C for 16 h. The peptides thus obtained were dried, dissolved in solvent A (Milli Q grade water containing 0.1% (v/v) trifluoroacetic acid), and then applied to a Titansphere column (4.0-mm inner diameter ϫ 10-mm column, GL Science Co., Japan) equilibrated with solvent A at the flow rate of 0.1 ml/min for 30 min. Phosphopeptides trapped on the Titansphere column were eluted with solvent C (0.5 M H 3 PO 4 ⅐NaOH (pH 8.0)) at the flow rate of 0.5 ml/min for 15 min. The thus eluted phosphopeptides were directly applied to a reversed-phase silica-base CAPCELL PACK C8 column (4.6-mm inner diameter ϫ 150-mm column, Shiseido, Japan), briefly washed with solvent A, and then eluted with a 90-min linear gradient, 0 -45%, of solvent B (acetonitrile containing 0.0886% (v/v) trifluoroacetic acid) at a flow rate of 1 ml/min. The phosphopeptides thus isolated were analyzed with an AXIMA-CFR MALDI-TOF MS (Shimadzu Co., Japan) using CHCA as a matrix.
Dephosphorylation of Phosphopeptides-Dephosphorylation of phosphopeptides derived from M-phase cytosol-treated emerin ⌬TM was carried out according to the method of Kuyama et al. (31). The phosphopeptides were dried, dissolved in 40 l of 46% hydrofluoric acid (HF, Wako, Japan), and then incubated at room temperature for 1.5 h. The peptides thus treated were dried, dissolved in 1 l of 40% acetonitrile containing 0.1% (v/v) trifluoroacetic acid, and then analyzed with an AXIMA-CFR MALDI-TOF MS.
MS/MS Sequencing-Phosphopeptides, which were separated with the Titansphere and C8 columns, were applied to an Inertsil ODS column (0.2-mm inner diameter ϫ 50 mm, GL Science Co., Japan) equilibrated with solvent D (2% (v/v) acetonitrile containing 0.1% (v/v) formic acid) and then eluted with a 20-min linear gradient, from 5 to 55% of solvent E (98% (v/v) acetonitrile containing 0.1% (v/v) formic acid) at the flow rate of 1.5 l/min using a MAGIC 2002 system (Michrome BioResources, Inc.). The phosphopeptides thus eluted were directly introduced to ESI-IT MS, LCQ Deca XP (Thermo Electron, San Jose, CA), equipped with a nanospray interface (AMR Inc., Japan) and a metal nanosprayer (GL Science Co.). To obtain sequence information on the eluted phosphopeptides, the mass spectrometer was operated in the ion select mode, where the MS scan was followed by the MS/MS scans of the calculated mass of the phosphopeptide as parent mass.

RESULTS
The Binding of Emerin ⌬TM to Chromatin-Two kinds of emerin fragments, i.e. emerin ⌬TM (residues 1-213) and emerin ⌬LT (residues 37-213), fused to GST were expressed in E. coli and used in this study (Fig. 1A). Emerin ⌬TM lacks the transmembrane domain, and emerin ⌬LT lacks both the transmembrane domain and most of LEM domain. GST-fused emerin ⌬TM (⌬TM) and GST-fused emerin ⌬LT (⌬LT) were purified from E. coli extract using a GSH-Sepharose bead (Fig. 1B). Western blotting with anti-GST antibody of ⌬TM and ⌬LT preparations, which were purified by GSH-Sepharose showed that smaller protein bands observed in Fig. 1B (lanes 1 and 2) indicated by asterisk were GST-containing degradation products of ⌬TM and ⌬LT (data not shown). All experiments in this study were done using beads bearing GST-emerin ⌬TM or ⌬LT. To determine whether ⌬TM interacts with chromatin and its interaction is regulated in a cell cycle-dependent manner, like for some other inner nuclear membrane proteins, i.e. LBR and LAP2␤, or not, we performed an in vitro chromatin binding assay. We previously developed this assay method to analyze the binding of inner nuclear membrane proteins to chromatin (28). When beads bearing ⌬TM were preincubated with buffer in the absence of the Xenopus egg cytosol fraction, they bound to chromatin slightly (column 1 in Fig.  2). However, when they were preincubated with a synthetic phase cytosol fraction (SC), the binding of chromatin to beads was stimulated (compare columns 1 and 2 in Fig. 2). Preincubation with a mitotic phase cytosol fraction (MC) did not stimulate the binding (column 3 in Fig. 2). Moreover, the once-stimulated chromatin binding activity of SC-treated beads was suppressed on subsequent incubation with MC (compare columns 2 and 4 in Fig. 2). On the other hand, the oncesuppressed chromatin binding activity of MC-treated beads was activated on subsequent incubation with SC (compare columns 3 and 5 in Fig. 2). These results demonstrated that the emerin fragment thus expressed can bind to chromatin, and that the chromatin binding assay method can be used to analyze the cell cycle-dependent binding of emerin to chromatin in vitro. The stimulation of the binding of ⌬TM to chromatin on SC treatment seemed to be independent of phosphorylation of ⌬TM, because the stimulation was not suppressed on pretreatment of SC with apyrase for ATP depletion or a wide-spectrum protein kinase inhibitor, i.e. staurosporine (compare columns 2, 6, and 7 in Fig.  2). Furthermore, the stimulation of the binding did not occur on SC treatment of the beads bearing GST-emerin ⌬LT (compare columns 12 and 13 in Fig. 2). These results indicated that the stimulation might be caused by the binding of the BAF in SC to ⌬TM, which is known to mediate the binding of emerin to chromatin, because (i) the stimulation of the binding was not suppressed by kinase inhibitors, (ii) the stimula-  [11][12][13][17][18][19][20][21][22][23][24][25]. Slashed boxes indicate the hydrophobic amino acid-rich region, which was demonstrated by Wolff et al. (31). Two GST-fused emerin fragments were constructed and used in this study (middle and lower). B, SDS-PAGE of GST fusion proteins. GST-emerin ⌬TM (⌬TM) and GST-emerin ⌬LT (⌬LT) were expressed in E. coli and then purified with glutathione-Sepharose beads. Beads bearing ⌬TM (1), ⌬LT (2), or marker proteins (M) were analyzed by SDS-PAGE on a 10% gel and then stained with CBB. The bands with the added asterisk at the right are GSTcontaining degradation products of ⌬TM and ⌬LT. The values at the left are the relative molecular masses of the marker proteins.
tion was not observed for ⌬LT, which lacks the BAF binding domain (mentioned below in more detail), and (iii) Xenopus egg cytosol contains 12 M BAF (14). On the other hand, suppression of the binding of SC-treated ⌬TM to chromatin on subsequent treatment with MC should be caused by the phosphorylation of emerin, because the suppression was prevented by apyrase or a kinase inhibitor, i.e. staurosporine (columns 8 -11 in Fig. 2).
Participation of BAF in the Binding of Emerin ⌬TM to Chromatin-To clarify the stimulation mechanism for the binding of ⌬TM to chromatin on SC treatment, we examined whether the binding of emerin to chromatin is mediated or not by BAF in our assay system. Beads bearing ⌬TM were treated with an E. coli extract containing expressed His tag BAF. The beads thus treated were used for the chromatin-binding assay (Fig. 3). ⌬TM treated with the E. coli extract containing His tag BAF (Fig. 3A, lane 2) bound to chromatin, although beads treated with buffer or the blank E. coli extract (Fig. 3A, lane 1 or 3, respectively) could not bind to chromatin. The binding of His tag BAF to the beads bearing ⌬TM in this assay system was confirmed by Western blotting with anti-His tag antibody (Fig. 3B). These results clearly show that the binding of beads bearing ⌬TM to chromatin is mediated by BAF and also support our idea that the stimulation of the binding of beads bearing ⌬TM to chromatin on pretreatment with SC is mediated by the binding of BAF to emerin.
M-phase-specific Phosphorylation of Emerin ⌬TM Suppressed the Binding to BAF-We examined whether the binding of emerin ⌬TM beads to BAF is cell cycle-dependent or not. Beads bearing ⌬TM were pretreated with E. coli extract containing His tag BAF to bind His tag BAF to ⌬TM. The beads thus treated were further treated with a buffer or cell cycle-dependent Xenopus egg extract (Fig. 4A). In the cases of treatments with the buffer and SC, His tag BAF bound to ⌬TM remained on beads (Fig. 4A, buffer and SC). However, His tag BAF bound to ⌬TM disappeared by MC treatment (Fig. 4A, MC). These results showed that MC, but not SC, treatment of ⌬TM-BAF complex causes dissociation of BAF from ⌬TM. On the other hand, beads bearing ⌬TM pretreated with buffer, SC, or MC were treated with the E. coli extract containing His tag BAF, then proteins bound to the beads were separated by SDS-PAGE and transferred to a nitrocellulose membrane, and BAF was detected with an anti-His tag antibody (Fig. 4B). The binding of BAF was suppressed by the pretreatment of ⌬TM beads with MC but not that with buffer or SC (Fig. 4B, buffer, SC, and MC). Furthermore, SC treatment followed by MC treatment (SC-MC) suppressed the binding of BAF to beads bearing ⌬TM (Fig. 4B, SC-MC). These results were consistent with the earlier results. The phosphorylation levels of emerin shown in Fig. 4C were demonstrated by ProQ diamond staining, which is known as a means of phosphoprotein staining (compare Fig. 4C, ProQ stain with CBB stain) (29). The weak "Pro Q staining" observed for the band of ⌬TM treated with buffer (Fig. 4C,  buffer) is nonspecific staining. Therefore, staining over the background can be considered to be phosphoprotein-specific staining. The bandshift on SDS-PAGE of emerin on treatment with MC (Fig. 4C, MC and SC-MC) was consistent with in vivo phosphorylated emerin in mitoticphase lymphoblastoid cells (9). However, the four bands depending on the phosphorylation states reported for lymphoblastoid cell were not clear in this system. ⌬TM preparations treated with MC (MC and SC-MC) were strongly phosphorylated and the binding of His-BAF was strongly suppressed. On the other hand, pretreatment of MC with apyrase or staurosporine to prevent the mitotic phosphorylation of emerin canceled this suppression activity (Fig. 4, B and C, MCϩApy. and MCϩSta.). A part of phosphorylation of ⌬TM by MC seems to be SCϩApy. and SCϩSta., respectively) or MC pretreated with 8 milliunits of apyrase or 5 M staurosporine (columns 8 -11; MCϩApy., MCϩSta., SC-MCϩApy., and SC-MCϩSta., respectively) at 23°C for 20 min. The beads thus treated were incubated with 20,000 decondensed sperm chromatin at 4°C for 10 min, and then observed by fluorescence microscopy after staining of DNA with Hoechst 33342. The "percentage of beads with bound chromatin" values were determined as described under "Materials and Methods" after subtraction of the value for blank GST beads. Beads bearing ⌬LT (ϳ4 g) treated with buffer (column 12) or SC (column 13) were reacted with chromatin in the same way as for the beads bearing ⌬TM. The results are the means Ϯ S.D. for three independent experiments. *, a significant difference from the respective control (p Ͻ 0.05).  25,000 decondensed sperm chromatin were added to the beads thus treated, followed by incubation for 10 min. Then, the "percentage of beads with bound chromatin" values were determined as described in Fig. 2. The results are the means Ϯ S.D. for three independent experiments. *, a significant difference from the respective control (p Ͻ 0.05). B, confirmation of the binding of emerin to BAF. Proteins bound to beads treated as in A were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and then incubated with an anti-His antibody. Bound antibodies were detected as enhanced chemiluminescence. A bar at the right indicates the position of 17-kDa marker protein electrophoresed.
caused by staurosporine-insensitive kinase, because 5 M staurosporine could not cause complete inhibition (Fig. 4, ProQ stain and MCϩSta.). Therefore, we concluded that the binding of emerin to BAF is suppressed by M-phase-specific direct or indirect phosphorylation. Thereafter, we focused on analysis of the cell cycle-dependent phosphorylation of emerin, especially mitotic phosphorylation sites of emerin.
Cell Cycle-dependent Phosphorylation States of Emerin-We next examined the cell cycle-dependent phosphorylation of emerin in a Xenopus egg cell-free system (Fig. 5). ⌬TM on beads was treated with egg cytosol fractions containing [␥-32 P]ATP, washed with binding buffer, and then analyzed by SDS-PAGE. The gel was stained with CBB and then subjected to autoradiography. The ⌬TM was strongly phosphorylated upon treatment with MC, although the phosphorylation by SC was very low (Fig. 5A, Autoradiography) (9). A mitotic phase-specific band-shift was also observed for ⌬LT treated with MC (Fig. 5A, ⌬LT). These results show that one or more phosphorylation sites that are outside of the LEM domain caused the main band-shift, because ⌬LT lacking almost all of the LEM domain showed a clear band-shift on treatment with MC (Fig. 5A, ⌬LT lanes). To compare the amounts of incorporated phosphate groups, ⌬TM treated with SC or MC in the presence of [␥-32 P]ATP was electrophoresed, and ⌬TM bands were excised from the gel and counted (Fig. 5B). The ⌬TM phosphorylated in the M-phase was 6.6 times as strong as that in the S-phase. These results demonstrated that the phosphorylation of emerin in the S-phase is very low.
To determine the mitotic phosphorylation state of emerin, and the differences in the phosphorylation site(s) between ⌬TM and ⌬LT, we performed phosphopeptide mapping (Fig. 6). Beads bearing ⌬TM and ⌬LT were treated with SC or MC in the presence of [␥-32 P]ATP. After washing, the beads thus treated bearing ⌬TM and ⌬LT were subjected to SDS-PAGE, and proteins were transferred to a nitrocellulose mem-brane. Then, the full-length emerin bands were excised from the membrane and digested with trypsin (Fig. 6A) or chymotrypsin (Fig. 6B). The peptides generated were subjected to two-dimensional separation on a cellulose plate. Although many spots of phosphorylated peptides were observed for the preparations derived from both ⌬TM and ⌬LT treated with MC, the major spots indicated with arrows in Fig. 6 were not observed for ⌬LT. When 32 P-labeled ⌬TM and ⌬LT were digested with chymotrypsin, well focused patterns were obtained (Fig. 6B). In these patterns, major spots 1 and 2 and three weak spots, 8, 10, and 11, were absent in ⌬LT map (compare Fig. 6B, ⌬TM and ⌬LT). These results suggested the following three possibilities: 1) at least one major phosphorylation site exists in the LEM domain, 2) deletion of the LEM domain causes obstruction of a major phosphorylation site present in other than the LEM domain, and 3) BAF binds an emerin kinase and Beads bearing ⌬TM (ϳ10 g) were pretreated with E. coli extract containing His tag BAF at 4°C for 3 h. After washing, these beads were treated with extraction buffer (buffer) or a cell cycle-dependent Xenopus egg cytosol fraction (SC or MC) at 23°C for 1 h. Then, proteins on beads thus treated were subjected to 12% gel SDS-PAGE, transferred to a nitrocellulose membrane, and then incubated with an anti-His tag antibody. Ten percent of loading amount of His-BAF in the reaction mixture was also reacted with an anti-His tag antibody (Load). Bound antibodies were detected by enhanced chemiluminescence. B, the binding of BAF to emerin was suppressed by mitotic phosphorylation of emerin. Beads bearing ⌬TM (ϳ10 g) were pretreated with SC in the case of (SC-MC), and then subsequently treated with buffer (Buffer), SC (SC), MC (MC and SC-MC), MC pretreated with apyrase (MCϩApy.), or MC pretreated with staurosporine (MCϩSta.), respectively. BAF bound to beads were detected by anti-His tag antibody as in A. C, the phosphorylation states of emerin. Beads bearing ⌬TM treated as in B were subjected to 10% gel SDS-PAGE, and then gel was stained with ProQ diamond to confirm the phosphorylation states of emerin (ProQ stain). Total emerin protein was stained by CBB (CBB stain). Bars at the right indicate positions of 17-, 43-, and 66-kDa marker proteins electrophoresed.  recruits it to ⌬TM. These three possibilities are explained under "Discussion." The phosphorylation levels of emerin fragments treated with SC were much lower than for those treated with MC (Fig. 6A), although the phosphorylation patterns of SC-treated samples were found to be very similar to those of MC-treated ones when the autoradiography films were superposed (data not shown). Therefore, we only examined the mitotic phosphorylation sites in the following experiments.
Identification of the Mitotic Phosphorylation Sites of Emerin-To identify the mitotic phosphorylation sites of emerin by means of mass spectrometry sequencing, we used a phosphopeptide separation system comprising a combination of a Titansphere column, which specifically binds phosphopeptides, and a reverse phase C8 column (see "Materials and Methods"). For this purpose, ⌬TM was purified by glutathione-Sepharose and phosphorylated by MC and then digested with proteinases. We mainly used a chymotryptic digest of emerin, because a chymotryptic digest gave a clearer phosphopeptide map than a tryptic digest, as can be seen in Fig. 6. The chymotryptic or tryptic phosphopeptides derived from phosphorylated ⌬TM were bound to the Titansphere column and then eluted with 0.5 M phosphate buffer (pH 8.0). The phosphopeptide fraction thus obtained was directly introduced onto the reverse phase C8 column, eluted with a linear gradient, and then fractionated (Fig. 7A). The molecular masses of the peptides thus obtained were determined by MALDI-TOF MS. Some of the determined masses were consistent with the calculated phosphopeptide masses (Fig. 7B, HF(Ϫ), and TABLE ONE). To confirm that these peptides contained phosphate groups, samples were pretreated with HF to hydrolyze phosphate groups and then the mass shift was analyzed by MALDI-TOF MS (Fig. 7B, HF(ϩ)). The mass shifts of 80 or 160 Da indicate that the obtained peptides were phosphopeptides and that one or two sites in their sequences were phosphorylated ( Fig. 7B and TABLE ONE). As can be seen in TABLE ONE, every phosphopeptide had more than two possible phosphorylation sites. Therefore, we performed MS/MS sequencing by means of ESI-IT MS to determine which residues are phosphorylated. The mass spectrum of the product ions of m/z 1367.7 and 685.1 ((Mϩ1H) ϩ and (Mϩ2H) 2ϩ of the 1367.7-Da peptide, respectively) allowed localization of the phosphorylation site to residue Ser 175 by b and y ion series. This was so because we could detect the b 7 ion (RPVSASR) without a phosphate group on product ion m/z 1367.7 sequencing and the b 8 (RPVSASRS) ion, which includes a phosphate group, and the y 4 ion (SLDL) without a phosphate group on product ion m/z 685.1 sequencing (Fig. 7C). In contrast to the case of 1367.7-Da peptide, we could not determine which site, Ser 175 or Ser 176 , was phosphorylated in 2080.0-Da peptide. However, we focused on only Ser 175 because Ser 175 , but not 176 Ser, was identified as a phosphorylation site by the analysis indicated above. Using the same approach for determination of the phosphorylated residue, the phosphorylation sites of other phosphopeptides (Ser 49 , Ser 66 , Thr 67 , and Ser 120 ) were also determined (TABLE ONE). However, we could not completely exclude the possi-  Ser 175 Phosphorylation of Emerin by MC Responsible for the Emerin-BAF Dissociation-Although five mitotic phosphorylation sites of emerin were identified and one possible phosphorylation site, i.e. Ser 175 , was found on MS/MS sequencing, we could not detect any phosphorylation site in the LEM domain. When these identified phosphorylation sites and phosphopeptide maps of ⌬TM and ⌬LT were compared, it was suggested that the phosphopeptide spots absent for the ⌬LT digest (major spots 1 and 2 and minor spots 10 and 11 in Fig. 6B) might not be due to the LEM domain and that phosphorylation at these sites is regulated through an unknown mechanism by the LEM domain. Then, we applied the phosphopeptide separation system to ⌬TM and ⌬LT phosphorylated with a mitotic extract to determine the phosphorylation site(s) corresponding to the absent spots in the phosphopeptide maps in Fig. 6B. Surprisingly, a peak corresponding to a peptide containing phosphorylated Ser 175 (arrow in Fig. 8A, ⌬TM, corresponding to peak 1 in Fig. 7A) had completely disappeared for the ⌬LT digest (Fig. 8A, ⌬LT), although there was no clear change in other peaks. We could not examine whether the peak corresponding to peak 2 in Fig. 7A, a peak corresponding to a fragment of the peak 1 material, disappeared or not, because the peak in Fig. 8A was too small. These results indicated that phosphorylation at Ser 175 , which is located outside of the LEM domain, might be affected through some unknown mechanism by the LEM domain and suggested that the phosphorylation participates in emerin-BAF dissociation. Then, we generated a point mutant at Ser 175 of ⌬TM (S175A-⌬TM) replaced with an alanine. Using this mutant, we could confirm that the missing peak from the elution pattern of phosphopeptides derived from ⌬LT is that of the Ser 175 -containing peptide (Fig. 8A,  S175A). Furthermore, we performed phosphopeptide mapping to confirm that the Ser 175 phosphorylation was affected by deletion of the LEM domain. As can be seen in Fig. 8B, spots 1 and 2, which disappeared from the ⌬LT pattern, also completely disappeared from the S175A-⌬TM pattern. Spots 12-14 appeared irregularly (compare ⌬TM patterns in Figs. 6 and 8). Spot 4 in Fig. 8B S175A shifted up when compared with ⌬TM and ⌬LT in Fig. 8B. However, the shift seems to be within a fluctuation of the spot, because the spot shifts slightly from one gel to another (compare Fig. 6B, ⌬TM to Fig. 8B, ⌬TM and Fig. 6B, ⌬LT to Fig.  8B, ⌬LT). These results suggested that the LEM domain is necessary for the Ser 175 phosphorylation by MC. Then, we carried out an in vitro FIGURE 8. Ser 175 phosphorylation by MC participates in emerin-BAF dissociation. A, the separation pattern of phosphopeptides derived from ⌬LT and S175A-⌬TM. Approximately 100 g of ⌬TM, ⌬LT, or S175A-⌬TM treated with MC as in Fig. 7 was chymotrypsinized. The peptides thus obtained were applied to a phosphopeptide separation system as in Fig. 7. The peak indicated by the arrow coincides with peak 1 in Fig. 7. B, beads bearing ϳ30 g of ⌬TM, ⌬LT, or S175A-⌬TM were incubated with 20 l of SC or MC containing 10 Ci of [␥-32 P]ATP at 23°C for 1 h. The proteins thus treated were separated by SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and then chymotrypsinized. The phosphopeptides thus obtained were separated by electrophoresis at pH 8.9 (horizontal direction; cathode to the left) and by ascending chromatography, as in Fig. 6. The points of sample application can be seen as dots near the bottom-left corners. C, S175A-⌬TM prevented the mitotic dissociation of BAF. Beads bearing ⌬TM or S175A-⌬TM (ϳ10 g) were treated with MC at 23°C for 1 h. The beads thus treated were incubated with an E. coli extract expressing His tag BAF, separated by 12% gel SDS-PAGE, and then transferred to a polyvinylidene difluoride membrane. BAF that bound to the beads was detected with an anti-His antibody as shown in Fig. 4. binding assay using S175A-⌬TM and BAF to determine whether the Ser 175 phosphorylation regulates the dissociation of emerin and BAF or not (Fig. 8C). The S175A point mutant treated with MC retained binding activity toward BAF, although ⌬TM treated with MC dissociated from BAF. This result showed that the Ser 175 phosphorylation in the M-phase participates in the dissociation of emerin and BAF.