Identification of Protein Kinase C (PKC) Phosphorylation Sites on Human Lamin B POTENTIAL ROLE OF PKC IN NUCLEAR LAMINA STRUCTURAL DYNAMICS*

Protein kinase C (PKC) is activated at the nuclear membrane in response to a variety of mitogenic stimuli. In human leukemic cells, the ,911 PKC isotype is selec- tively translocated and activated at the nucleus. We recently identified the nuclear envelope component lamin B1 as a major substrate for nuclear PKC both in whole cells and in vitro. Using highly purified human BII PKC and isolated nuclear envelopes from the human promyelocytic (HL60) leukemia cell line, we have now determined the major sites for BII PKC-mediated lamin B phosphorylation. Using a combination of cyanogen bromide cleavage, direct microsequencing, tryptic phosphopeptide, and phosphate release analyses, two major sites of PKC-mediated phosphorylation, Ser3@' and Ser406, have been identified. These sites lie within the carboxyl-terminal domain of lamin B immediately adjacent to the central a-helical rod domain. Function- ally, &I PKC-mediated phosphorylation of these sites leads to the time-dependent solubilization of lamin B indicative of mitotic nuclear envelope breakdown in vitro. PKC-mediated lamin B phosphorylation is inhibited by the active hibitor and Two observations indicate that PKC-mediated nitrocellulose Autoradiography and immunoblot analy- sis was performed using a specific anti-lamin antibody as previously described (12). Lamin B phosphorylation and content was evaluated by laser densitometric scanning.


Identification of Protein Kinase C (PKC) Phosphorylation Sites on Human Lamin B
POTENTIAL ROLE OF PKC IN NUCLEAR LAMINA STRUCTURAL DYNAMICS* (Received for publication, September 22, 1992) Barbara A. HocevarS, David J. Burns  Protein kinase C (PKC) is activated at the nuclear membrane in response to a variety of mitogenic stimuli. In human leukemic cells, the ,911 PKC isotype is selectively translocated and activated at the nucleus. We recently identified the nuclear envelope component lamin B1 as a major substrate for nuclear PKC both in whole cells and in vitro. Using highly purified human BII PKC and isolated nuclear envelopes from the human promyelocytic (HL60) leukemia cell line, we have now determined the major sites for BII PKC-mediated lamin B phosphorylation. Using a combination of cyanogen bromide cleavage, direct microsequencing, tryptic phosphopeptide, and phosphate release analyses, two major sites of PKC-mediated phosphorylation, Ser3@' and Ser406, have been identified. These sites lie within the carboxyl-terminal domain of lamin B immediately adjacent to the central a-helical rod domain. Functionally, &I PKC-mediated phosphorylation of these sites leads to the time-dependent solubilization of lamin B indicative of mitotic nuclear envelope breakdown in vitro.
PKC-mediated lamin B phosphorylation is inhibited by 1) a monoclonal antibody directed against the active site of PKC, 2) a PKC pseudosubstrate inhibitor peptide, and 3) a PKC peptide substrate. Two observations indicate that PKC-mediated lamin B phosphorylation and solubilization is due to direct phosphorylation of lamin B by PKC rather than indirect activation of a cdc2 kinase. Neither immunodepletion with p13suc1 Sepharose beads nor the presence of a ~3 4 "~"~ kinase peptide substrate had any effect on PKC-mediated lamin B phosphorylation. Therefore, we conclude that & PKC represents a physiologically relevant lamin kinase that can directly modulate nuclear lamina structure in vitro. Nuclear BII PKC, like ~3 4 " "~ kinase, may function to regulate nuclear lamina structural stability during cell cycle.
The nuclear lamina is a proteinaceous fibrillar network underlying the inner nuclear membrane that gives the nucleus structural integrity, segregates the nuclear interior from the surrounding cytoplasm and provides attachment sites for * This work was supported by grants from the Mathers Charitable chromatin (1)(2)(3). The major constituents of this structure, the nuclear lamins, are members of the intermediate filament protein family (1). The nuclear lamina is a dynamic structure which undergoes growth during cell cycle progression and is reversibly depolymerized at the Gz/M phase transition of cell cycle (2). Many of the structural dynamics of the lamina are thought to be regulated by the reversible phosphorylation of the nuclear lamins. Mitotic depolymerization is dependent upon the stoichiometric hyperphosphorylation of the lamins, an event which contributes to the increased solubility of mitotic lamins (4).
Following mitosis, the lamins become increasingly insoluble and are progressively dephosphorylated as they reassociate to form the polymeric lamina (2,4).
Recent attention has focused on ~34'~'' kinase, the active component of M-phase promoting factor, since it is required for G2/M phase progression, and it can elicit a mitotic-like nuclear envelope breakdown (NEBD) in uitro. However, biochemical evidence suggests the involvement of multiple lamin kinases in NEBD in uiuo ( 7 , 10).
Recent work from this laboratory demonstrated that treatment of human leukemia cell lines with the mitogenic PKC activator bryostatin 1 leads to the rapid translocation and activation of PKC at the nucleus where it induces timeand dose-dependent phosphorylation of lamin B, the sole lamin species expressed in these cells (14,15). From these data we suggested that nuclear ~I I PKC activation and subsequent lamin B phosphorylation was involved in mitogenic stimuli, including the proliferative effects of bryostatin (14,15). Although PKC appears to be a physiologically relevant lamin kinase, its physiologic role remains to be elucidated.
In the present study, we report the identification of the major sites of phosphorylation on human lamin B mediated by 011 PKC. These sites are restricted to the carboxyl-terminal tail region immediately adjacent to the central a-helical coil domain. This region of the lamins has been found to be phosphorylated during mitosis and is thought to be involved in the regulation of lamin assembly and disassembly. In addition, we show that PKC-mediated phosphorylation can cause mitotic-like depolymerization of the nuclear lamina in uitro. These results indicate that PKC, like ~3 4 '~" ' kinase, may play a role in regulating nuclear lamina structural dynamics during the cell cycle.
Expression, Isolation, and Chnracterization of Human @II PKC Produced in the Baculovirus-Insect Cell Expression System-Two independent clones for human ,911 PKC were isolated from a human temporal cortex library (Stratagene). Clone 1 contained a deletion between base pairs 1189 and 1261 based on the previously published human sequence (17). Clone 2 was a partial clone with sequence beginning at approximately base pair 400 of the previously published human (311 sequence. The two clones were spliced together to generate a full-length human BIl PKC construct as follows: a 1.1-kilobase BamHI/SacI fragment from clone 1 (representing the 5' end of @II PKC) was ligated to a 0.93-kilobase SacI/BglII fragment from clone 2 (3' end of ,&I). The resulting 2.0-kilobase BarnHI/BglII fragment (full-length human PI1 PKC) was gel isolated and then ligated into the BarnHI site of the baculovirus transfer vector, pVL941.
The generation of recombinant On PKC baculovirus was accomplished as previously described (18). Recombinant human PI[ PKC was produced in Spodoptera frugiperda cells (Sf91 infected with @]I PKC baculovirus. Sf9 cells were routinely harvested and characterized 60-72 h after infection. PKC was partially purified from Sf9 cells expressing recombinant human @I1 PKC as previously described (18). Protein kinase C activity was determined using the protein kinase C vesicle assay of Ogita et al. (19) except that the reaction buffer was slightly modified. Briefly, a reaction mixture containing the following components (final concentrations): 30 pg of phosphatidylserine, 0.5 pg of diacylglycerol, 100 p~ CaC12, 47.5 p M EGTA, 10 mM MgC12, 20 mM HEPES, pH 7.5, 200 pg/ml histone type H1, 30 p M [32P]ATP (NEN Research) was combined with partially purified enzyme in a final volume of 250 pl. This mixture was incubated for 10 min at 30 "C and filtered onto Whatman GF/C filters. The filters were washed with 10 ml of 10% trichloroacetic acid and subjected to liquid scintillation counting. Phosphatidylserine and diacylglycerol were omitted from the reaction mixture for no lipid control reactions. We routinely obtained 1-2 pg of PKC/1 X lo6 infected cells. Specific activities of the recombinant enzyme were typically 500-1,000 nmol/ min/mg. Isotype purity of the recombinant enzyme was confirmed by immunoblot analysis using isotype-specific synthetic peptide-derived rabbit antibodies as previously described (14).
Phosphorylation of Lamin B and Synthetic Peptides in Vitro-For isolation of phosphorylated lamin B for cyanogen bromide cleavage, trypsin digestion, and microsequence analysis, purified HL60 nuclear envelopes were combined with human PII PKC in PKC phosphorylation buffer (50 mM Tris-HC1, pH 7.5, 10 mM MgS04, 1 mM dithiothreitol, 100 p~ CaCI2, 1 mM ATP containing 50 pCi/ml [32PlATP, and 100 nM bryostatin 1). Reactions were carried out for 30 min at 37 'C. The reaction mixture was solubilized and subjected to twodimensional isoelectric focusing/SDS-PAGE as previously described (11). Resolved proteins were transferred to nitrocellulose and subjected to autoradiography. The identity of phosphorylated lamin B was confirmed by immunoblot analysis with a rabbit anti-lamin antibody as previously described (12), with the exception that antibody detection was achieved by enhanced chemiluminescence (Amersham Corp.) as described by the manufacturer.
Peptides corresponding to the two identified tryptic fragments containing putative PKC phosphorylation sites were synthesized by the Peptide Synthesis Laboratory at the University of Kentucky. Synthetic peptides were incubated with PKC in PKC phosphorylation buffer as described above. Reactions were terminated by centrifugation in Millipore Ultrafree-MC filter units (10,000 molecular weight cut-off) to separate the PII PKC from the peptide. The resulting filtrate was subjected to tryptic phosphopeptide mapping as described below.
Cyanogen Bromide Cleavage and Microsequencing Analysis of Phosphorylated Lamin B-Lamin B, phosphorylated either in whole cells or in vitro with PKC, was resolved by two-dimensional isoelectric focusing/SDS-PAGE and transferred to nitrocellulose sheets as described above. Immobilized lamin B was incubated with 50 mg/ml cyanogen bromide (CNBr, Kodak) in 70% (v/v) formic acid for 18 h. The cleavage supernatants were separated from the nitrocellulose, diluted 1:l with dHsO, and taken to dryness in a speed-vac (Savant). Cleavage efficiency was assessed by Cerenkov counting. Typically, >95% of input counts were recovered in the supernatant fraction. The dried residue was washed twice with dHzO and taken to dryness. The washed CNBr fragments were solubilized in Laemmli sample buffer, boiled, and subjected to SDS-PAGE in 10-20% acrylamide gradient gels. The gels were washed twice in pH 11.0 CAPS transfer buffer (10 mM 3-cyclohexylamino-1-propane sulfonic acid, 10% methanol, v/v, 0.5 mM dithiothreitol) and transferred to polyvinylidene fluoride (PVDF, Schleicher and Schuell) membrane. The PVDF membrane was stained with Coomassie Blue to visualize CNBr fragments, destained, and subjected to autoradiography.
For microsequencing analysis, the resulting phosphorylated CNBr fragment was excised from the PVDF membrane and submitted to automated microsequencing on an Applied Biosystems 477A Protein Sequencer at the Case Western Reserve Core Molecular Biology Facility.
Tryptic Phosphopeptide Analysis of Phosphorylated Lamin B-In vivo or in vitro phosphorylated lamin B, immobilized on nitrocellulose as described above, were incubated in a solution of 0.5% (w/v) polyvinylpyrrolidone (PVP-40) in 100 mM acetic acid at 37 "c for 30 min to block nonspecific adsorption of trypsin. Trypsin digestion and two-dimensional phosphopeptide mapping was carried as described by Boyle et al. (21) on cellulose thin layer plates (Kodak). Recovery of counts was monitored by Cerenkov counting of the digests and was typically 90% or greater. Electrophoresis was carried out in pH 1.9 buffer (formic acid/glacial acetic acid/deionized water, 50156:1794, v/v) for 20 min at 1000 V. Chromatography was performed for 3 h in either phosphochromatography buffer (n-butanol/pyridine/glacial acetic acid/deionized water, 75:50:15:60, v/v) or isobutyric acid buffer (isobutyric acidln-butanol/pyridine/glacial acetic acid/deionized water, 65:2:5:3:29, v/v). Tryptic phosphopeptides were visualized by autoradiography at -70 "C.
Phosphate Release Assays of Tryptic Phosphopeptides-Tryptic phosphopeptides, separated by two-dimensional thin layer chromatography as described above, were scraped and eluted from the cellulose plates into pH 1.9 buffer and subjected to 10 cycles of manual Edman degradation as described (21). Phosphate release was monitored by electrophoresis on TLC plates at pH 1.9 using free ["-PI orthophosphate as a marker. Radioactivity was monitored by Cerenkov counting. Total recovery of input counts was greater than 90% after correction for decay.
Nuclear Envelope Breakdown Assays-Human &I PKC was incubated with nuclear envelopes from 1 X lo7 HL60 cells in reaction buffer (50 mM Tris-HC1, pH 7.5, 50 mM NaCI, 10 mM MgSO4, 1 mM dithiothreitol, 100 p~ CaCl2, 20 mM @-glycerophosphate, 20 mM NaF, and 1 mM ATP) for the indicated times at 37 "c. In some experiments, 1 pCi of [3zP]ATP was included to monitor incorporation into lamin B. Reactions were terminated by the addition of Nonidet P-40 to a PKC Phosphorylation Sites on Human Lamin B 7547 final concentration of 1% (w/v) followed by a 10-min incubation on ice. The reaction mix was microfuged a t 12,000 X g for 10 min at 4 "C, and supernatants and pellets were subjected to SDS-PAGE analysis in 8% acrylamide gels followed by transfer to nitrocellulose as previously described (1 1). Autoradiography and immunoblot analysis was performed using a specific anti-lamin antibody as previously described (12). Lamin B phosphorylation and content was evaluated by laser densitometric scanning.

RESULTS
Expression and Characterization of Human PKC in the Baculovirus-Insect Cell System-Our previous studies in human leukemic cells demonstrated the translocation and activation of a PKC activity at the nucleus following bryostatin treatment where it phosphorylated nuclear lamin B (11,14,15). Moreover, we recently identified this nuclear-activated PKC as the isotype (14, 15). We therefore wished to use purified human ,811 PKC to determine the sites of phosphorylation on human lamin B. To obtain a source of human PI^ PKC that was not cross-contaminated with other PKC isotypes, we utilized the baculovirus-insect cell system to express recombinant human PI1 PKC. This system has been used successfully for the production of bovine a, rat PII, and y-PKC (22). Since uninfected Sf9 insect cells contain very little detectable PKC activity and no immunoreactive PKC, infection with recombinant baculovirus yields a convenient source for isotype pure PI1 PKC (22). The isotype purity of this preparation was confirmed by immunoblot analysis using our previously characterized isotype-specific antibodies recognizing the a, PI, PII, and y isotypes of PKC (14). Recombhant PII PKC was immunoreactive with the PII-specific antibody but not the other isotype-specific antibodies indicating the absence of multiple isotypes in the preparation (Fig. 1B).
To assess that the recombinant PI1 PKC behaved like native PKC, the Ca2+ and phospholipid dependencies of the enzyme were determined as described under "Experimental Procedures" (Fig. 1, C and D ) . Recombinant PII PKC exhibited Ca2+-phosphatidylserine-dependent histone kinase activity with maximal activity observed at -150 PM Ca2+ and 10 mol 9% phosphatidylserine, similar to values reported previously for PII PKC from other sources (22). Therefore, our recombinant human PI1 PKC preparation is isotype pure and exhibits enzymatic properties consistent with those of the native enzyme.
Human Protein Kinase C-mediated Phosphorylation Is Confined to the Carboxyl-terminal Domain of Lamin B-Nuclear lamin B belongs to the family of intermediate filament proteins and as such possesses several structural motifs common to all family members. The most prominent of these is the presence of a large, highly conserved a-helical domain which makes up the central portion of the molecule. This region is thought to be responsible for the formation of a highly stable coiled-coil dimer between two lamin molecules (3). Flanking the central a-helical region are non-helical NH2and COOH-terminal domains (3). Previous analysis has identified the NHz-and COOH-terminal domains immediately adjacent to the a-helical coil domain as a region of multiplesite phosphorylation (5, 8, 10, 23). It appears clear that phosphorylation of these sites is involved in the regulation of vitro phosphorylated lamin B revealed that the major phosphorylation sites mediated by PI1 PKC reside in a fragment which migrates as a doublet of -10-12 kDa consistent with phosphorylation of the COOH-terminal domain adjacent to the central a-helical domain (Fig. 2). To confirm the identity of this fragment, the lower band of the doublet was excised and subjected to direct microsequencing analysis (Fig. 3). The sequence obtained identified this fragment as corresponding to the predicted CNBr cleavage product G~u~~~ to Met4%. This fragment corresponds to the carboxyl end of the central ahelical domain and the flanking region of the carboxyl-terminal domain. Sequence analysis of the upper band in the doublet was inconclusive; however, two-dimensional tryptic phosphopeptide maps of the two bands were identical to each other and to those generated by tryptic digestion of intact lamin B suggesting identity of the actual phosphorylation sites within these fragments (data not shown). The most likely explanations of the doublet are that the bands are related fragments generated by incomplete CNBr cleavage, or alternatively that the upper band may represent a more highly phosphorylated form of the lower fragment. In this regard, phosphorylation of intact lamin B causes the protein to migrate more slowly in SDS-PAGE gels (4; see also Fig. 6).

FIG. 2. CNBr digestion analysis of lamin B phosphorylated by j?n PKC in vivo and in vitro.
Panel A, HL60 cells were equilibrated with 100 pCi/ml of [32P]orthophosphoric acid 60 min prior to 30 min of treatment with 100 nM bryostatin 1 to stimulate lamin B phosphorylation (11). Nuclear envelopes were then isolated, and samples were subjected to SDS-PAGE analysis, transferred to nitrocellulose, and digested with CNBr as described under "Experimental Procedures." The resulting CNBr digest was resolved by SDS-PAGE on a 10-20% acrylamide gradient gel, transferred to PVDF membrane, and subjected to autoradiography. The 10-12-kDa CNBr doublet corresponding to the COOH-terminal domain adjacent to the a-helical region is designated by arrows. The 24-kDa CNBr fragment corresponding to the NHP-terminal domain is designated by an arrowhead. € 3 , HL60 nuclear envelopes phosphorylated in vitro by PJI PKC as described under "Experimental Procedures" were subjected to CNBr digestion analysis as described in panel A .
Interestingly, phosphorylated lamin B isolated from whole cells contains a phosphorylated fragment of -24 kDa in addition to the 10-12-kDa doublet (Fig. 2 A ) . Based upon its molecular mass, this 24-kDa band may correspond to the predicted NH2-terminal domain CNBr fragment. However, this fragment is not highly phosphorylated in vivo and is not detected after in vitro phosphorylation (Fig. 2B), where phosphorylation can be directly attributed to 0 1 1 PKC. These results indicate that the major phosphorylation sites mediated by PII PKC are contained within the carboxyl-terminal domain of lamin B immediately adjacent to the central a-helical domain.

P I I PKC Phosphorylates Lamin B Predominantly at Se?95
and Ser405-Although microsequencing of the phosphorylated CNBr fragment localized PII PKC-mediated phosphorylation sites to the COOH-terminal domain of lamin B, specifically between Glu"' and Met4%, the large size of the fragment precluded direct sequencing through potential PKC phosphorylation sites. Previous phosphoamino acid analysis of human lamin B phosphorylated either in whole HL60 cells or in vitro with PKC indicated that phosphorylation was confined to serine residues within the molecule (11). Furthermore, analysis of the deduced amino acid sequence of human lamin B revealed the presence of only three potential PKC phosphorylation sites, Ser395, Ser405, and Ser4O8, within the identified CNBr fragment (no potential Thr sites exist in this fragment). All three sites conform to the determined consensus PKC phosphorylation motif of S*/T'-X-K/R. As indicated in Fig. 3, each of these potential sites localize to separate tryptic fragments. Therefore, tryptic digestion of intact lamin B employed to generate small peptides that would contain only one potential PKC phosphorylation site within this region.
Tryptic digests of in vivo and in vitro phosphorylated lamin B were subjected to two-dimensional thin layer chromatography and autoradiography. Lamin B phosphorylated in vitro by PKC generates two predominant "P-labeled tryptic fragments labeled spots 1 and 2 (Fig. 4, right panel). In vivo labeled lamin B (Fig. 4, left panel) generates these two major spots and an additional spot not seen in in vitro labeled lamin B (spot 3 ) .
Attempts to directly microsequence in vitro phosphorylated tryptic fragments after HPLC or two-dimensional TLC separation proved unsuccessful, so alternative methods were employed to identify the phosphorylated residues, including phosphate release analysis and comparative chromatographic migration analysis. [32P]Phosphate released after each cycle of manual Edman degradation was employed to detect a phosphorylated residue at that cycle. For this analysis, tryptic residues were separated by two-dimensional TLC, eluted from the cellulose plate, and subjected to manual Edman degradation as described (21). Analysis of an aliquot removed after each cycle revealed [32P]phosphate release after cycle 3 in peptide 1 and after cycle 6 in peptide 2, with approximately In vivo and BIIPKC in uitro phosphorylated lamin B was subjected to tryptic digestion as described (21). The resulting tryptic digest was subjected to two-dimensional TLC analysis as described first dimension electrophoresis was performed in pH 1.9 buffer (formic acid/acetic acid/dH20 50156:1794) at 1000 V for 20 min followed by second dimension ascending chromatography in phosphochromatography buffer (n-butanol/pyridine/ acetic acid/dH20 75:5015:60) for 3 h. Phosphopeptides were visualized by autoradiography. from lamin B phosphorylated with BII PKC in vitro (Fig. 4) were eluted from cellulose TLC plates in pH 1.9 buffer and subjected to manual Edman degradation as described (21). Aliquots from each cycle were evaluated for phosphate release as described (21). Total recovery of initial input counts was greater than 90% as monitored by Cerenkov counting. Results are displayed graphically as counts released/cycle as a percentage of total recovered counts. 90% of the input counts recovered in the phosphorylated cycle, indicating a single phosphorylation site in each peptide (Fig. 5). Therefore, peptide 1 was tentatively identified as A403SSSR, with phosphorylation a t Ser405. Peptide 2 was tentatively identified as L3Y3PSPSSR, with phosphorylation at Phosphate release analysis also discounted Ser40R as a possible site of phosphorylation.

V C I E E I D V D G K F I R L K N T S E Q D Q P M~~
To further confirm the identity of the tryptic fragments, comparative chromatographic migration analysis was performed in two different solvent systems. A migration or Rf value was calculated from the known amino acid composition of each peptide for each solvent system as described (21) and compared to the migration of the tryptic fragments following two-dimensional TLC separation. Table I compares the actual migration of peptides 1 and 2 in the two chromatographic solvents along with predicted relative mobilities for each peptide. As shown, the experimentally determined mobilities correlate well with predicted values, again corroborating the proposed assignment of peptides 1 and 2. In addition, synthetic peptides corresponding to the sequences of peptides 1 and 2 show nearly identical migration with the respective peptide following one-dimensional chromatography (R, values in bold, Table I). Based on the above results, the major sites of PKC-mediated phosphorylation of lamin B are identified as Ser395 and Ser405.

PII PKC Phosphorylation Leads to Solubilization of Nuclear
Envelope-associated Lamin B-A hallmark of the G2/M phase transition in eukaryotes is NEBD, a process requiring hyperphosphorylation of the lamins (4). We therefore sought to determine if PKC could elicit NEBD in uitro utilizing a standard lamin solubilization assay as previously described (25). This assay takes advantage of the fact that after lamina disassembly, presumably by hyperphosphorylation of the lamins, the lamins become soluble in the reaction mixture. For lamin B, which remains membrane bound in the cell during mitosis, non-ionic detergent is added to obtain freely soluble lamin B in the reaction supernatant.
As can be seen in Fig. 6A, incubation of nuclear envelopes with PKC leads to solubilization of lamin B as evidenced by loss of lamin B from the nuclear envelope pellet fraction ( P ) and recovery of the protein in the Nonidet P-40 supernatant (S). This solubilization was accompanied by PKCmediated phosphorylation of lamin B. Lamin B solubilization was rapid and time-dependent, being apparent by 5 min and becoming maximal by 120 min (Fig. 6B). This time course is similar to that observed for lamin solubilization using extracts from mitotic cells (26). These data demonstrate that PII PKC can elicit nuclear lamina disassembly in uitro.
Since the p34cdc2 kinase has also been shown to elicit nuclear lamina disassembly in uitro, we assessed whether a cdc2-like activity was present in our PKC or nuclear envelope preparations. This did not seem likely since our nuclear envelopes contain no intrinsic kinase activity (11,14), and TABLE I Mobility of tryptic phosphopeptides during chromatography Two-dimensional-TLC analysis was performed on phosphopeptides generated from tryptic digestion of lamin B. Electrophoresis was carried out in pH 1.9 buffer (formic acid, 88%/acetic acid/dH20, 50156:1794) followed by chromatography in the indicated buffers.  the recombinant PKC preparation exhibited only Ca+* and phospho1ipid:dependent kinase activity (Fig. 1). Likewise, the identified phosphorylation sites, Ser395 and Ser405, do not conform to the CDC2 kinase consensus motif (27). However, further evidence that the observed NEBD was elicited directly by PKC comes from the observation that PII PKC-mediated lamin B phosphorylation is inhibited by: 1) a monoclonal antibody against the PKC active site which specifically inhibits PKC activity, 2) a PKC substrate peptide, and 3) a PKC pseudosubstrate inhibitor peptide. In contrast, preincubation of the reaction mix with ~1 3~"~' beads, which specifically bind cdc2 kinase, failed to remove the lamin kinase activity (Fig.  7A). Human cdc2 kinase prepared from mitotically arrested HL60 cells, as described under "Experimental Procedures" also phosphorylates lamin B (Fig. 7B). Lamin B phosphorylation mediated by cdc2 kinase was inhibited by a cdc2 substrate peptide, but not by the PKC pseudosubstrate inhibitor peptide (Fig. 7B). Histone phosphorylation is included to demonstrate isolation of the active mitotic form of cdc2 kinase (20). cdc2 phosphorylation of lamin B also was not inhibited by the PKC monoclonal antibody nor PKC substrate peptide (data not shown). These observations indicate that although both kinases can phosphorylate lamin B, the lamin solubilization observed in the NEBD assays shown in Fig. 6A can be attributed to direct phosphorylation by PI1 PKC.

DISCUSSION
Existence of Multiple Lamin Kinases-The intermediate filament nuclear lamins are integral components of the nuclear envelope which provide structural integrity to the nucleus during interphase (1)(2)(3). Disassembly of the lamin network is thought to be required for nuclear envelope breakdown at the time of mitosis (3). Previously it was shown that phosphorylation of the lamins precedes disassembly while dephosphorylation accompanies reassembly of the lamin network following chromosome segregation, leading to the hypothesis that phosphorylation is the key trigger for lamina disassembly (2,4). Thus, the search for "mitosis-specific" lamin kinases was initiated.
Previously, we have shown in HL60 and K562 cells that lamin B is a major substrate for PKC both in vivo and in vitro (11,14,15). In the present report we have identified the major sites of PKC-mediated lamin B phosphorylation as Ser395 and Ser405. In addition, phosphorylation of the lamins by PKC has been demonstrated by other groups both in vivo (16) and in vitro (10,23). Clearly PKC has been identified as a physiologically relevant lamin kinase but the functional consequence of PKC-mediated lamin phosphorylation remained to be elucidated.
Recently, several other kinases also have been identified as physiologically relevant lamin kinases, most notably, ~34'~'' kinase (5)(6)(7)(8), and S6 kinase I1 (9,10). Interestingly, in a detailed study of phosphorylation sites on human lamin C by Ward and Kirschner it was found that sequences surrounding the identified phosphorylation sites did not match the consensus phosphorylation site motifs of any single kinase (10). Therefore, they postulated that several lamin kinases act on the lamins to modulate lamina structure during the cell cycle. Specifically, they proposed that phosphorylation of sites within the carboxyl-terminal domain by one kinase may facilitate subsequent phosphorylation of adjacent sites by other lamin kinases. Indeed, the existence of multiple lamin kinases is corroborated by our study, since treatment of intact HL60 cells with bryostatin generates an additional lamin B phosphorylation site which cannot be directly attributed to phosphorylation by PKC (Fig. 2 A ) , suggesting the possible involvement of another lamin kinase. Such cooperative or hierarchal phosphorylation involving multiple lamin kinases could explain the phosphorylation of sites on mitotic lamins which are consensus sites for different kinases. In this regard, we have found that Pa PKC phosphorylates human lamin B in vitro a t more than 50 times the rate of mitotic cdc2 kinase suggesting that PKC is a better intrinsic lamin kinase than is cdc2 kinase? It is intriguing to speculate on the existence of a hierarchal phosphorylation scheme involving PKC and cdc2 kinase. If such a scheme exists, our data would suggest that PKC first phosphorylates lamin B at Ser395 and Ser405 thereby facilitating subsequent cdc2 kinase-mediated phosphorylation at the adjacent site, Ser3". This possibility remains to be tested.
Role of Lamin Phosphorylation in Regulating Nuclear Lamina Structure-Although it has been demonstrated that phosphorylation of the lamins is important in regulating lamin disassembly preceding mitosis, conflicting data exists in the literature regarding mitotic phosphorylation sites as well as sites which are key for disassembly of the lamins. Fig. 8 summarizes the identified phosphorylation sites on the lamins attributed to the known lamin kinases. Peter et al. (5) demonstrated that phosphorylation of Ser" on chicken lamin BZ by cdc2 kinase resulted in lamina disassembly in vitro. In B. A. Hocevar   further studies, this group demonstrated that phosphorylation at this site was sufficient to cause head-to-tail depolymerization of in vitro synthesized chicken lamin B2 polymers (8).
Although this site conforms to the consensus phosphorylation sequence requirements of cdc2 kinase, this is not a consensus phosphorylation site for PKC and indeed is not detected after in vitro phosphorylation by PKC in our study (Fig. 2B). In contrast, Ward and Kirschner (10) identified Ser392 as the only mitosis-specific phosphorylation site on human lamin C, while phosphorylation at Se132 (the human lamin C homologue of Ser" on chicken lamin B2) and Ser4" increased in phosphorylation over interphase levels. It is important to note that while Ser4M has been identified as a S6 kinase I1 phosphorylation site on lamin C, sequence differences renders the analogous site (Ser405) on human lamin B1 a consensus PKC phosphorylation site (see Fig. 8). Indeed in this report, we identify Ser405 as a major site of PKC-mediated phosphorylation. Interestingly, in a study by Heald and McKeon (24) it was found that a double mutation of Ser22 and Ser392 on human lamin A was required to partially block lamina disassembly in vivo, while single mutations of these sites were insufficient to cause this blockade indicating the importance of carboxyl-terminal phosphorylation in lamina disassembly in vivo. In the present study, we have demonstrated that & PKC-mediated phosphorylation of Ser395 and Ser405 on human lamin B1 is sufficient to cause nuclear lamina disassembly in vitro.
The common theme that emerges from these in vivo and in vitro phosphorylation studies is that phosphorylation of both the NH2-and COOH-terminal domains immediately adjacent to the a-helical coil region is important in lamina disassembly in vivo. The lamins are known to dimerize, presumably through interaction of the a-helical regions, and phosphorylation of residues close to these regions could cause disruption of the hydrophobic interactions holding the network together. In addition, phosphorylation of these sites may be involved in other processes proposed to be mediated by the lamins, such as chromatin attachment (28) and DNA replication (29). sients have been observed to precede nuclear envelope breakdown in amphibian eggs as well as mammalian Swiss 3T3 cells (30,31). In fact, microinjection of the calcium chelator BAPTA as well as severe Ca2+ deprivation by treatment with ionomycin in EGTA-containing media blocks nuclear envelope breakdown, suggesting the involvement of a Ca2+-sensitive kinase (31). Moreover, rat diploid fibroblasts treated with staurosporine analogs, selective inhibitors of PKC, fail to enter mitosis and instead undergo DNA re-replication (32). Finally, a functional homologue of PKC identified in Saccharomyces cerevisiae, termed PKCl, was shown to be essential for cell viability. Cells depleted of the PKCl gene product arrested at a point in cell cycle following S phase but prior to mitosis, while deletion of the gene resulted in recessive lethality (33). Although PKC has been implicated in the process of NEBD, no aspect of NEBD has previously been shown to be directly modulated by PKC. In this report, we identify the major sites of phosphorylation mediated by PII PKC as Ser395 and Ser405 in the COOH-terminal domain of human lamin B immediately adjacent to the a-helical region. Furthermore, phosphorylation of these sites is sufficient to cause NEBD in vitro.
Since these residues have been demonstrated to be involved in modulating lamina structure during mitosis (lo), we propose that PI, PKC may also serve as a mitotic lamin kinase.
Identification of the relative roles of individual phosphorylation sites and their proposed kinases in modulating nuclear lamina structural dynamics in vivo remain to be determined.