Molecular characterization of p62, a mitotic apparatus protein required for mitotic progression.

A 62-kDa (p62) mitotic apparatus-associated protein is important for the proper progression of mitosis in sea urchin embryos (Dinsmore, J. H., and Sloboda, R. D. (1989) Cell 53, 769-780). We have isolated and characterized a full-length p62 cDNA of 3374 base pairs which encodes an extremely acidic polypeptide of 411 amino acids having a calculated Mr of 46,388 and a pI of 4.01; p62 is a unique protein with no significant identity to any known proteins. Southern and Northern blot analyses demonstrate that the gene for p62 is present once in the sea urchin genome and the corresponding mRNA is present in unfertilized eggs and in early embryos through and up to the gastrula stage. Sequence analysis suggests certain regions may participate in chromatin association and microtubule binding, an observation that is consistent with previous immunological data (Ye, X., and Sloboda, R. D. (1995) Cell Motil. Cytoskeleton 30, 310-323) as well as data reported herein. Confocal microscopy reveals that during interphase the protein binds to chromatin in the nuclei of sea urchin eggs. In the germinal vesicles of clam oocytes at prophase of meiosis I, p62 binds to the condensed chromosomes. Currently, truncated clones of p62 are being used to identify the tubulin and chromatin binding domains.

Controlled microtubule disassembly occurs during anaphase A, the stage of mitosis characterized by the movement of the chromosomes toward the poles, an event thought to be powered by molecular motors that reside at the kinetochores. The kinetochore microtubules selectively disassemble during anaphase A, and if the microtubules are prevented from disassembling, e.g. by treatment with the microtubule stabilizing, anti-tumor drug Taxol, then mitosis ceases. Microtubule depolymerization can be induced by calcium, and calcium has been shown to result in the shortening of the distance separating the poles and movement of the chromosomes toward the poles (1). Furthermore, microtubule disassembly is necessary for the reactivation of chromosome movement in lysed cells (2), and isolated mitotic chromosomes can be translocated in association with a single microtubule in vitro as the result of microtubule depolymerization (3,4). Thus, a critical step in anaphase is the controlled disassembly of the kinetochore microtubules.
We have previously demonstrated that a protein of 62 kDa (p62), a substrate of a calcium/calmodulin-dependent protein kinase, is associated with the mitotic apparatus (5,6). Both p62 and the kinase responsible for its phosphorylation copurify with mitotic apparatuses isolated from first cell cycle sea urchin embryos (5,7). Importantly, the phosphorylation of p62 has been shown to correlate directly with the disassembly of microtubules in isolated mitotic apparatuses (5). Furthermore, microinjection of sea urchin embryos with affinity-purified antibodies to p62 inhibits progression through the cell cycle, specifically at the metaphase to anaphase transition (6). Because p62 is present at constant levels through at least two cell cycles (8), p62 dephosphorylation must be due to the action of a specific phosphatase and not to the degradation of the protein, as occurs, for example, with the cyclins. In fact, the existence of protein phosphatase 1 in the mitotic apparatus isolated from sea urchin embryos that dephosphorylates p62 has been demonstrated by pharmacological studies with phosphatase inhibitors (9). Finally, the subcellular localization of p62 varies with the cell cycle. During mitosis, p62 is associated with the mitotic apparatus, particularly with the microtubules (6,8,10), while during interphase, p62 resides in the nucleus (8,10).
Based on this background, our working hypothesis is that phosphorylation and dephosphorylation of p62 play key roles in controlling cell division, particularly during anaphase A, by permitting kinetochore microtubule disassembly at the appropriate time during mitosis. As the logical next step in elucidating the mode of action of p62 in mitosis, we present here a characterization of the polypeptide based on its deduced amino acid sequence obtained from cloned cDNA. The p62 polypeptide is highly acidic, yet contains distinct subdomains of alternating clusters of basic and acidic residues. Furthermore, confocal microscopy reveals that during interphase the protein binds to chromatin in the nuclei of sea urchin eggs; in the germinal vesicles of clam oocytes at prophase of meiosis I, p62 binds to the condensed chromosomes.

Generation of Peptide Fragments of p62 and Protein Microsequencing
Electroelution was used to purify p62, as described previously (10). The purified, denatured protein was concentrated using a single Centricon-10 ultrafiltration device according to the instructions of the manufacturer (Amicon, Inc., Beverly, MA).
CNBr Cleavage-Purified p62 (100 l) was dialyzed against 0.5 M NH 4 HCO 3 at 4°C overnight, lyophilized, and resuspended in 150 l of 50 mg/ml CNBr (Eastman Kodak Co.) in 70% (v/v) formic acid and incubated overnight in the dark at room temperature. An equal volume of deionized water was then added, and the sample was shell-frozen in liquid N 2 and lyophilized. Thirty microliters of reducing buffer (31.25 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 5% ␤-mercaptoethanol, 0.004% pyronin Y) were added, and the sample was boiled for 4 min and resolved by electrophoresis on a 5-15% gradient polyacrylamide SDSgel (11). The fragments were blotted to polyvinylidene difluoride mem-brane in CAPS 1 transfer buffer (10 mM CAPS, 10% MeOH, pH 11.0) at 300 mA at 4°C for 75 min. The membrane was rinsed with deionized water, and stained for 1 min with Coomassie Blue R-250 in 40% MeOH, 1% acetic acid and destained in 50% MeOH for 1 min followed by three rinses with deionized water. Bands of interest were excised and loaded directly into the sequencer cartridge of an Applied Biosystems-476 A protein sequenator.
Trypsin Digestion-Trypsin digestion of p62 blotted to Trans-Blot membrane (Bio-Rad) was used in a parallel approach to obtain sequence data from internal regions of p62. After electrophoresis and electroblotting, the membrane was rinsed with Milli-Q water three times for 5 min each, then immersed for 1 min in a solution containing 0.1% Amido Black in 10% acetic acid and destained for 1 min in 5% acetic acid. The band corresponding to p62 was excised, rinsed with Milli-Q water, air-dried, and sent to the Wistar Protein Microchemistry Laboratory at the Wistar Institute (Philadelphia, PA) where in situ proteolytic digestion of p62 with trypsin was performed. Three peptides identified by HPLC from the tryptic digest were chosen, isolated, and sequenced in a gas phase sequenator.

Sea Urchin cDNA Expression Library and Screening
A sea urchin ZAP cDNA library provided courtesy of Dr. Jonathan Scholey (University of California, Davis) was screened (12) with affinity-purified p62 polyclonal antibodies (see Ye and Sloboda (10) for antibody characterization) that had been preabsorbed with an Escherichia coli/phage lysate. One positive clone, ZAP h1, was identified using this procedure. The pBluescript SK Ϫ containing the h1 clone was removed from the ZAP vector by in vivo excision, and the resulting clone was named SKh1. As an initial step in clone verification, the fusion protein produced by ZAP h1 was used to affinity purify antibodies from a crude antiserum obtained from a second rabbit, different than the one that produced the antibodies used for library screening.

RT-PCR and Library Screening
Two regions with minimal codon degeneracy of the longest tryptic peptide of p62, Trp-3, were chosen to design degenerate sense and antisense primers for use in PCR. The sequence of the Trp-3 peptide is NH 2 -DDDRNVIEVETINFDGDTVIQPLLSLR-COOH. The sense primer, referred to as 3A up (5Ј-GAYGAYGAYMGIAAYGTIAT-3Ј, where I ϭ inosine, M ϭ A ϩ C, and Y ϭ C ϩ T), was designed from the first seven residues of the Trp-3 peptide. The antisense primer, referred to as 3A down (5Ј-ARIGGYTGDATIACIGTRTC-3Ј, where D ϭ A ϩ G ϩ T and R ϭ A ϩ G), was designed from residues 17-23 of the Trp-3 peptide. Finally, a pool of 26-mer degenerate oligonucleotides, referred to as 3A (5ЈGARGTIGARACIATIAAYTTYGAYGG-3Ј), was designed from residues 8 -16 in the internal sequence of Trp-3. The 3A oligos were used to screen sequences generated by the PCR reaction using the 3A up and 3A down primers. cDNA was synthesized from poly(A)-rich RNA of unfertilized sea urchin eggs using an oligo(dT) 16 primer in a reverse transcription reaction and was then subjected to two rounds of PCR amplification using primer pair 3A up and 3A down . Analysis of the amplified PCR products showed a diffuse band in the expected size range of 68 base pairs. The 68-base pair PCR product was subcloned into the pCR TM vector (Invitrogen), and the resulting colonies were screened by hybridization for a specific PCR product using ␥-32 P-end-labeled 3A oligonucleotides; positive clones were confirmed by DNA sequencing. One such colony, named TA 8, was shown by sequence analysis to contain a specific DNA that could be translated into 23 amino acids identical in sequence to the Trp-3 peptide. Clone TA 8 was then used to screen the cDNA library by plaque hybridization.

DNA Sequencing and Analysis; in Vitro Transcription and Translation
Plasmid DNA was isolated from recombinant clones using the plasmid mini kit from Qiagen and sequenced via the dideoxy chain termination method on an Applied Biosystems model 373A automated se-quencer. Either vector-specific or custom oligonucleotide primers were used to sequence the cDNA clones in primer-directed sequencing using the Taq Dye Primer cycle sequencing kit (Applied Biosystems). All clones were sequenced at least twice on each strand. Sequence analysis was performed using the programs of the University of Wisconsin Genetics Computer Group (UW GCG) software package (13). The FASTA (14) and BLASTP (15) programs were used to search the data libraries of GenBank TM 79.0, SWISS-PROT 26.0, and PIR 38.0 through the NCBI E-mail Server. Sequence comparison was performed using BESTFIT of the GCG software. Coupled in vitro transcription and translation was performed with 1 g of cloned DNA in the reticulocyte lysate-based TNT system (Promega, Madison WI) in the presence of [ 35 S]methionine essentially as described by the manufacturer except as follows. The luciferase-directed reaction was driven by the T3 promoter at 30°C for 120 min. The SKh11-directed reactions were incubated with either T3 or T7 polymerase in reticulocyte lysate at 37°C for 30 min and then at 30°C for 90 min. The products were resolved by SDS-PAGE, the gel dried, and the translation products detected by PhosphorImaging (Molecular Dynamics, Sunnyvale, CA).

RNA Preparation and Northern Blot Analysis
Sea urchins of the genus Strongylocentrotus purpuratus were purchased from Marinus, Inc. (Westchester, CA). Spawning, fertilization, and embryo culture were carried out in 0.45 m Millipore-filtered sea water at 15°C. Embryos were collected at various stages, and total RNA was isolated using the RNeasy Total RNA kit according to the instructions of the manufacturer (Qiagen). About 15 g of total RNA from each sample were denatured at 65°C for 20 min and then subjected to electrophoresis on a formaldehyde/agarose gel. The resolved RNA was blotted onto nitrocellulose by capillary transfer and prehybridized at 42°C for 4 h in 2.5 ϫ prehybridization mix (1 ϫ mix is 25 mM PIPES, 2 M NaCl, 0.25% (w/v) each of Sarkosyl, polyvinylpyrrolidone-40, bovine serum albumin, and Ficoll 400), 50% formamide, and 0.2 mg/ml denatured salmon sperm DNA. RNA on the membrane was hybridized with probe at a specific activity of 2.0 ϫ 10 6 cpm/ml in the above-described buffer containing 10% dextran sulfate for 48 h at 42°C. The probe corresponded to the entire p62 cDNA that had been labeled with [␣-32 P]dCTP by random priming using Ready-To-Go TM DNA labeling beads from Pharmacia Biotech Inc. and dCTP. After hybridization, the membrane was washed once for 30 min in 2 ϫ SSC, 750 M sodium pyrophosphate, and 0.05% (w/v) Sarkosyl at room temperature and four times, for 30 min each, in 0.1 X SSC, 375 M sodium pyrophosphate, and 0.05% (w/v) Sarkosyl at 50°C, and then exposed to a PhosphorImager screen for 2 h at room temperature.

Genomic DNA Preparation and Southern Blot Analysis
Genomic DNA was isolated from unfertilized sea urchin eggs by the hexadecyltrimethylammonium bromide (CTAB) DNA extraction protocol provided by Dr. Kathy Foltz, University of California, Santa Barbara, as follows: 100 l of packed eggs were resuspended in a 1.5-ml microcentrifuge tube with 700 l of prewarmed (60°C) CTAB lysis buffer (100 mM Tris-HCl, pH 8.0, 1.4 M NaCl, 20 mM EDTA, 2% CTAB [w/v], 1% polyvinylpyrrolidone, M r 360,000 (polyvinylpyrrolidone-360 (w/v), and 0.2% 2-mercaptoethanol (the latter added just before use). The tube was incubated at 60°C for 30 to 60 min. The lysate was extracted with an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1), followed by chloroform:isoamyl alcohol (24:1). After extraction, the aqueous layer was transferred to a new tube and mixed with two-thirds volume of ice cold isopropanol to precipitate genomic DNA, which was sedimented by centrifugation for 5 min at room temperature. The pellet was washed in 500 l of wash buffer (76% ethanol, 10 mM ammonium acetate) at room temperature for 5 min. After a brief centrifugation, the DNA pellet was dried using a Speed Vac (Savant Instruments, Inc., Farmingdale, NY) and resuspended in 30 l of TE (10 mM Tris, 1 mM EDTA, pH 8.0).
Twenty micrograms of genomic DNA were digested to completion with EcoRI, KpnI, or HindIII and electrophoresed on an 0.8% agarose gel. The gel was treated with 0.25 N HCl for 15 min, denatured for 30 min in 1.5 M NaCl, 0.5 M NaOH, and neutralized for 30 min in 0.5 M Tris-HCl, pH 7.4. After rinsing the gel in 2 ϫ SSC, the DNA was transferred to nitrocellulose by capillary transfer in 20 ϫ SSC. The membrane was prehybridized at 65°C in prehybridization solution (see previous section), then hybridized with 32 P-labeled p62 cDNA in the same solution at 65°C for 12 h. After hybridization, the membrane was washed at room temperature in 2 ϫ SSC, 0.5% SDS, four times, for 30 min each, then exposed to a PhosphorImager screen at room temperature overnight.

Nucleus Immunolabeling Experiments
Sea urchin nuclei were isolated with membranes intact from Lytechinus pictus eggs (8). Germinal vesicles were isolated from oocytes of the surf clam Spisula solidissima as follows. Oocytes were washed three to four times in seawater. One milliliter of packed oocytes was collected by centrifugation in a clinical centrifuge at maximum speed, washed for 30 s in 1 M glycerol, then washed for 60 -90 s in 1 M glycerol, 10 mM NaH 2 PO 4 , pH 8.0. Oocytes were resuspended in 20 volumes of Nonidet P-40 buffer (0.25 M sucrose, 50 mM PIPES, pH 6.9, 0.75 mM MgSO 4 , 10 g/ml leupeptin, 1 mM dithiothreitol, 160 M Pefabloc, and 0.5% Nonidet P-40), applied over a 0.4 M sucrose in 50 mM PIPES, pH 6.9, cushion, and centrifuged for 10 min in a clinical centrifuge at threefourths speed. The germinal vesicles in the pellet were washed several times by resuspension and sedimentation through sucrose.
Isolated nuclei or germinal vesicles were fixed in 90% methanol, 50 mM EGTA at Ϫ20°C, and processed for indirect double label immunofluorescence (10) using a mixture of affinity purified antibodies to p62 and DNA antibodies, the latter kindly provided by Dr. Robert Goldman (Northwest University Medical School). Secondary antibodies were a mixture of fluorescein-labeled goat anti-human antibodies and rhodamine-labeled goat anti-rabbit antibodies. Slides were viewed using an LSM-410 inverted confocal laser scanning microscope from Carl Zeiss, Inc.

RESULTS
Protein Microsequencing-Purified p62 was cleaved with CNBr to yield two cleavage products, pL and pS, with estimated molecular masses of 36 and 20 kDa, respectively. These peptides were specific to CNBr cleavage as they were absent from control samples which revealed slight degradation of p62 due to formic acid alone. Digestion with CNBr also generated several other peptides that were detectable on Coomassie Bluestained SDS-gels, but these were not efficiently transferred to the polyvinylidene difluoride membrane. When a duplicate gel was transferred to nitrocellulose and probed with affinity purified p62 antibodies, both pL and pS as well as several smaller peptides were recognized by the antibodies. When phosphorylated p62 was cleaved with CNBr, peptides pL and pS were radioactive, indicating they contained one or more residues that could be phosphorylated under the conditions employed in the in vitro phosphorylation reaction of p62 (5). This observation also indicates that p62 is phosphorylated in vitro at more than one site, assuming that the CNBr cleavage reaction went to completion. Peptides pL and pS were sequenced, and five sequential residues from each peptide were obtained with confidence. These sequences were named CNBr-1 (NH 2 -AKEYF-COOH) and CNBr-2 (NH 2 -KGKGD-COOH).
Library Screening-Two approaches, an antibody screen and a PCR-based screen, were used simultaneously to identify p62 clones from a sea urchin cDNA library. Both approaches identified the same clone, SKh1, which encodes tryptic peptides Trp-1 and Trp-3, as well as the CNBr-derived peptides, CNBr-1 and CNBr-2, and contains an open reading frame of 225 residues (25 kDa). However, the 225-residue open reading frame extends to the 3Ј end of the clone, there is no in frame stop codon, poly(A) addition signal, or poly(A) tail. Thus, SKh1 is a partial cDNA clone of p62, and thus could be used to rescreen the cDNA library to identify a full-length clone. Using this partial cDNA as a probe, an additional 24 positive clones were obtained by hybridization screening. Clone SKh11, with the largest insert, was sequenced completely, and the sequence and deduced amino acid sequence are shown in Fig. 1. SKh11 is composed of 3374 bases, and contains a single open reading frame encoding 411 amino acid residues which is preceded by 173 bases of 5Ј-untranslated region. The open reading frame is followed by a 3Ј-untranslated region of 1968 bases that terminates in a short poly(A) sequence. The correct reading frame of p62 was specified both by this single, long, open translation unit and by the presence of all four microsequenced peptide sequences in this reading frame (Fig. 1). The first ATG in the open reading frame, at nucleotide position 174 of SKh11, is in the proper context for eukaryotic initiation. The ATG is preceded by an A at the Ϫ3 position, where a purine residue has been shown to be important for the initiation of translation (16), and a G at the ϩ4 position (17). Upstream of this ATG there exist five, in-frame stop codons (at nucleotide positions 6, 96, 111, 126, and 132). Although five additional in-frame ATGs (at nucleotide positions 702, 1038, 1102, 1227, and 1260) are also present, these ATGs either lack the consensus sequence for optimal translational initiation, a feature that becomes important in the analysis of eukaryotic translation, or encode small peptides that do not contain p62 tryptic and CNBr peptide sequences. Therefore, none of these codons represent translational initiation sites for intact p62. The coding region is terminated by a stop codon (TGA) at nucleotide position 1407; a series of TGAs follow throughout the entire extended 3Јuntranslated region. The 3Ј-untranslated region includes three putative poly(A) addition signals (ATTAAA) at nucleotide positions 2337, 2385, and 3336, and a short poly(A) tail of nine bases. The ATTAAA at nucleotide position 3336 is located 24-base pairs upstream of the poly(A) tail, and thus has a high potential to be the polyadenylation signal. The other two poly-adenylation signals are most likely not used, at least by the transcript that generated clone SKh11, as they are Ͼ35 bases upstream of the poly(A) tail. Furthermore, no correspondingly smaller species of mRNA were detected by Northern blot analysis (see Fig. 3). Thus, it is doubtful that those signals are functional in vivo, at least in early embryos. All of the above demonstrate that SKh11 represents a full-length cDNA clone of p62.
In Vitro Translation of p62-The protein encoded by clone SKh11 has a calculated M r of 46,388. On SDS-gels, p62 has been estimated to have a M r of 62,000 (5). Thus, the molecular mass predicted from the sequence of SKh11 is 16 kDa less than the apparent molecular mass estimated from SDS-gels. Although phosphorylation may contribute to the anomalous migration of a given protein on SDS-gels (18), this 16-kDa discrepancy in molecular mass of p62 most likely results from the regions in the primary sequence that are extraordinarily rich in acidic residues. Indeed, highly acidic amino acid sequences can cause large anomalous migrations of proteins on SDS-gels. This has been inferred intuitively, as such proteins bind SDS with relatively low efficiency, and demonstrated directly by fusion studies using cloned CENP-B cDNAs (19). To confirm directly that the acidic character of p62 is responsible for this migration anomaly, the insert from clone SKh11 was transcribed and translated. When the T7 promoter in SKh11 is used to drive a coupled transcription-translation reaction with reticulocyte lysate, a T7 dependent polypeptide is produced that migrates on SDS gels at M r 62,000 (Fig. 2, lane B). No similar product is produced when T3 polymerase, whose promoter lies downstream of the insert, is used as the negative control (Fig. 2, lane A). The positive translation control for this experiment is luciferase, having a M r of 61,000 (Fig. 2, lane C). Thus, these data show that the large molecular weight of p62 estimated from SDS-gels is most likely due to anomalous SDS binding, and support the conclusion that SKh11 is a full-length cDNA clone of p62.
Northern and Southern Blot Analyses-Northern blot analysis was performed to determine the size and expression of p62 mRNA during early development in the sea urchin embryo. High stringency hybridization between the insert of SKh11 and total RNA identified a 3.5-kilobase message in unfertilized eggs and in embryos at all stages of development investigated (Fig.  3). It is clear that the p62 message is present maternally in unfertilized eggs and throughout development up to the gastrula stage. Only one mRNA transcript was detected during early development in the sea urchin. The size of the message The RNAs were transferred to nitrocellulose and hybridized in 50% formamide at 42°C with a DNA probe obtained from the full-length cDNA clone SKh11, prepared by random primer labeling of the insert from SKh11. Lanes 1-9, RNA from developmental stages as indicated above each lane. corresponds to that of the insert of clone SKh11, indicating that the cloned cDNA corresponds to a full-length transcript.
To determine the copy number of the gene encoding p62 in the sea urchin genome, as well as to identify related genes, high molecular mass DNA from unfertilized sea urchin eggs was digested with EcoRI, KpnI, or HindIII, and probed with the insert from SKh11. The resultant autoradiogram is shown in Fig. 4. The hybridization pattern was the same under all stringency conditions tested suggesting that there are no other sequences in the sea urchin genome related to p62. Note that the recognition sites for EcoRI and KpnI are not found within clone SKh11, whereas HindIII is present once in SKh11 (nucleotides 831-836). These data show that the gene encoding p62 is presence once in the sea urchin genome.
The entire polypeptide is relatively acidic with a pI of 4.01. Two-dimensional gels and immunoblots of mitotic apparatus proteins confirm that p62 has a pI of about 4.0 (10). However, analysis of the data in Fig. 5 shows that the acidic and basic residues are clustered, thus producing distinct subdomain structure in p62 which may be functionally important. Highly acidic amino acid segments are characteristic of a number of nuclear and chromosomal proteins and are thought to play a role in chromatin binding (19) possibly through electrostatic interactions with basic histones. Moreover, in p62 subdomains m2 and m4 and domain C are basic. Several microtubuleassociated proteins (MAPs) have basic microtubule binding domains, suggesting that these basic regions in p62 may represent one or more microtubule-binding domains.
When hydropathy was analyzed by the Kyte-Doolittle algorithm (20), the results showed two characteristic regions in p62. Domain N is composed of a combination of both hydrophilic and hydrophobic amino acids. This suggests that domain N may adopt an internal position in the tertiary structure or, perhaps more likely, p62 interacts via hydrophobic interactions with one or more other polypeptides in vivo. By contrast, the majority of the remaining 321 residues, domains M and C, are charged and thus are highly hydrophilic. Finally, p62 does not contain any P-loop motifs (of the form GXXXXGK(T/S/G)) for nucleotide binding. Thus, p62 is not predicted to have microtubule-based motor activity.
The p62 primary sequence was also analyzed against the PROSITE data bank for conserved protein motifs and potential post-translational modification sites. Surprisingly, a search for known calcium/calmodulin-dependent protein kinase phosphorylation sites failed to reveal any such sites in p62. However, certain regions in p62 have similarity to a base consensus sequence for this kinase, and one or more of these may represent potential calcium/calmodulin-dependent protein kinase phosphorylation sites in p62. The literature (21,22) indicates that calcium/calmodulin-dependent protein kinase phosphorylation sites, while similar, vary greatly in actual sequence. One can distill the varying domains, however, to a minimal essential consensus composed of a basic residue followed by two to three residues and then the phosphorylated residue. When p62 is searched for this consensus, similar sequences are noted at six positions. These sites will be discussed in relation to the calcium/calmodulin-dependent phosphorylation of p62 in more detail in the discussion section. Multiple potential phosphorylation sites for other kinases were also identified in the deduced amino acid sequence generated by clone SKh11, including cAMPdependent protein kinase, protein kinase C, casein kinase II, and protein tyrosine kinase. The possible biological significance of these phosphorylation sites with respect to the function of p62 will be discussed later.
GenBank™ Search-Protein sequence similarity between p62 and known proteins was analyzed by searching the data libraries of GenBank TM 79.0, SWISS-PROT 26.0, and PIR 38.0 as well as the yeast genome data base. No significant homologies between p62 and any known proteins were noted. However, p62 showed some degree of similarity to a superfamily of nucleolar phosphoproteins which includes nucleolar protein NO38 from Xenopus (23), nucleophosmin from chicken (24), rat protein B23 (25), and nucleolar phosphoprotein B23 from human (26,27). These phosphoproteins are the major nucleolar proteins of growing eukaryotic cells. They have been found to associate with intranucleolar chromatin and preribosomal particles and bind to single-stranded nucleic acids (28). Although their function is unknown, these proteins are thought to play a role in pre-rRNA transcription and ribosome assembly (29). Overall, p62 and the rat protein B23 (25) are 61% similar. The similarity, which is due mainly to the extensive stretches of glutamate and aspartate residues that occur in both proteins, is not significant. However, significant sequence homology is confined to a stretch of 29 amino acids from residues 96 to 124 (Fig. 6) in the N domain of p62 and the NH 2 -terminal region of B23. A closer comparison of this region with other proteins identified exclusively 12 members from the nucleolar phosphoprotein superfamily, which are all 100% homologous in the region encompassing residues 105-115 of p62. However, no function has thus far been elucidated for this conserved region from rat B23 or other members of the family. p62 is clearly distinct from rat protein B23, as little sequence similarity is observed in the remainder of p62 and B23. When the p62 sequence was searched for consensus nuclear localization signals, none were found, an observation that is consistent with previous immunofluorescence data (10), which indicated that p62 interacts directly with chromatin prior to nuclear envelope reassembly.
Immunodetection of Phosphoserine, Phosphothreonine, and Phosphotyrosine in p62-As mentioned earlier, CNBr cleavage of in vitro phosphorylated p62 detected two major radioactive cleavage products via autoradiography, suggesting that, in vitro, p62 is phosphorylated on more than one residue by calcium/calmodulin-dependent protein kinase (5). To investigate this possibility, Western blot analysis of mitotic aparatus proteins not previously subjected to in vitro phosphorylation was performed using antibodies specific for phosphoserine, phosphothreonine, and phosphotyrosine residues. As shown in Fig.  7, all three antibodies react with p62, while other polypeptides on the blot react variably. This suggests that p62 contains at least one of each type of phosphoamino acid and is thus phos-phorylated at multiple sites as suggested by CNBr cleavage studies. However, this result does not indicate whether any or all of these in vivo phosphoepitopes result from calcium/ calmodulin-dependent protein kinase activity.
Nucleus Immunolabeling-As revealed by homology searches, p62 showed some similarity to certain nuclear proteins and DNA-binding proteins. To study the possible interaction between p62 and chromatin/chromosomes, laser confocal double label immunofluorescence microscopy of isolated nuclei was performed using affinity purified antibodies to p62 and antibodies to DNA. When nuclei were isolated from sea urchin eggs, processed for indirect immunofluorescence, and scanned by laser confocal microscopy, p62 was observed distributed throughout the nucleus, coincident with the decondensed chromatin characteristic of these cells, which have completed meiosis and are awaiting fertilization to activate the cell cycle ( Fig.  8, a-c). By comparison, a similar analysis with germinal vesicles isolated from surf clam oocytes, which are arrested at prophase of meiosis I and have the bulk of their chromatin condensed and bound to the inner nuclear envelope, demonstrates that p62 resides throughout the nucleus, in association not only with the condensed chromosomes but also with other filamentous networks within the nucleoplasm (Fig. 8, d-f ). These networks may represent decondensed chromatin, the transcription of which is required for oocyte maintenance. After stimulation of the completion of meiosis by KCl activation, these networks become less evident as the filamentous material progressively associates with the chromosomes during the final stages of chromatin condensation. Concurrently, p62 staining becomes restricted to the condensed chromatin as well (Fig. 8, g-i). These data provide direct evidence that p62 can be found in association with chromatin/chromosomes as suggested by the sequence analysis. DISCUSSION This study reports the sequence and characterization of p62, a mitotic apparatus-associated phosphoprotein important for mitotic progression in the sea urchin embryo. By SDS-PAGE FIG. 6. Amino acid sequence comparison of a segment of p62 with a segment of rat nucleolar protein B23. The deduced amino acid sequence of p62 (top) was aligned with the rat nucleolar protein B23 (bottom) using BESTFIT of the GCG package. Significant homology was detected in the portion of the sequence indicated here. Residues 96 -124 of p62 show a high degree of homology with residues 96 -124 of the rat nucleolar protein. The GCG software indicates identical residues with vertical lines between a given pair of residues, while conservative substitutions are indicated by two dots and similar residues are indicated by a single dot.

FIG. 7. Western blot analysis of p62 phosphorylated in vivo.
Mitotic apparatus proteins were separated by SDS-PAGE, transferred to nitrocellulose, and probed with monoclonal antibodies to phosphoserine (lane 1), phosphothreonine (lane 2), and phosphotyrosine (lane 3), respectively. Note that p62 (indicated by the arrow) reacts with all three phosphoamino acid antibodies, while other proteins in the mitotic apparatus preparation react variably, depending on the antibody employed.
p62 has been estimated to have a relative molecular mass of 62 kDa (5), yet the relative molecular mass predicted from the cDNA sequence (Fig. 1) is 46 kDa. The large apparent mass on SDS-PAGE is due to the acidic properties of the protein, and this was demonstrated directly by in vitro translation of the cloned insert (Fig. 2). Although phosphorylation can affect the mobility of proteins on SDS gels (18), the major contributor to the migration anomaly noted here is the acidic nature of p62 (see Figs. 1 and 2) (19). The detection of phosphorylated CNBrderived peptides by autoradiography, and the positive immunoreaction of p62 with anti-phosphoamino acid antibodies (Fig.  7), suggest that p62 is phosphorylated at multiple sites in vivo and in vitro. However, phosphopeptide mapping of p62 will be required to identify the number and position of these sites.
Previous studies (6,8,10) have shown that p62 associates with microtubules in the mitotic spindle. It is believed that microtubule-binding to associated proteins (MAPs) is mediated through electrostatic interactions between the acidic COOH terminus of ␤-tubulin and basic MAP domains (30,31). For example, the microtubule binding domain shared by MAP 2, tau, and MAP 4 is characterized by three to four imperfect 18-amino acid repeats containing the characteristic tetramer sequence PGGG (32)(33)(34)(35)(36). By contrast MAP 1A and MAP 1B have a repeated binding motif of KKE(EIV) (37,38). In addition, CLIP-170, a cytoplasmic linker protein that binds endocytic vesicles to microtubules, contains a conserved repeated motif, GKN(D/S)G, shared by rat DP-150, the Drosophila glued protein, and the yeast protein BIK1 (39). None of the above microtubule binding motifs was found in p62, and p62 shares no sequence homology, at the amino acid level, with other known MAPs. EMAP, a 77-kDa echinoderm MAP originally identified in eggs by cycles of pH-and temperature-dependent microtubule assembly and disassembly, has been reported in sea urchin, sand dollars, and starfish (40) to localize to interphase and mitotic microtubule arrays. This 77-kDa EMAP contains basic and slightly acidic regions in its deduced amino acid sequence but lacks the above-described characteristic microtubule-binding sequences (41). Basic and acidic charge distribution is typical of microtubule-binding proteins, and thus the basic region of EMAP may represent a novel microtubule binding domain. However, direct tests of this hypothesis, by, for example, constructing fusion proteins containing various EMAP domains, have not yet been reported. While a known microtubule-binding sequence in p62 similarly cannot be identified due to sequence homology alone, previous immunoelectron microscopy of isolated mitotic apparatuses (8) as well as immunofluorescence studies (6, 10) strongly suggest the association of p62 with microtubules in the mitotic spindle, perhaps either directly, through a basic, yet novel, microtubule-binding motif, or indirectly through other adapter proteins or MAPs.
Conversely, it has been reported that tubulin interacts with MAP-2 or cytoplasmic dynein at an EGEE sequence located in the acidic COOH terminus of ␣-tubulin (31, 42) (but see also Cleveland et al. (43)); an EGEE sequence (residues 147-150) is also present in the acidic subdomain, m1, of p62. A BLASTP search of the protein data base using this sequence detected many unrelated proteins. The significance of this sequence for non-tubulin proteins remains to be determined, but it may represent a motif for interaction of non-tubulin proteins with MAPs. In this regard, p62 could interact with certain MAPs through this element, and then the microtubule-binding ability of the MAP could mediate the association of p62 with microtubules.
p62 was originally identified as a substrate of a Ca 2ϩ / calmodulin-dependent protein kinase (Cam kinase) which copurified with p62 in the sea urchin mitotic apparatus. Three classes of Cam kinases have been described thus far. Cam kinase I and Cam kinase III are specific to brain and certain other tissues, and each has a very narrow substrate specificity. By contrast, Cam kinase II is a multifunctional protein kinase FIG. 8. Laser confocal immunofluorescence of L. pictus nuclei and S. solidissima germinal vesicles. Nuclei were isolated from eggs of the sea urchin L. pictus and germinal vesicles were isolated from oocytes of the surf clam S. solidissima and processed for double label immunofluorescence using anti-DNA antibodies and affinity-purified antibodies to p62. Shown here are the laser confocal images. The distribution of DNA is revealed in the fluorescein channel (green, in a, d, and g), while p62 is shown in the rhodamine channel (red, in b, e, and h). c, f, and i show the superimposed fluorescence images of p62 and DNA; the yellow color results when the two images are coincident. In the sea urchin interphase nuclei, chromatin (a) and p62 (b) appear to be distributed more or less evenly throughout the nucleus (c). In nonactivated clam germinal vesicles, which are arrested at prophase of meiosis I, the bulk of the DNA is condensed into chromosomes (d) and most, but not all, of the p62 (e) is associated with these condensed chromosomes, as indicated by superimposition of the images (f ). When germinal vesicles are activated to proceed through meiosis, complete condensation of the chromatin begins (g) and p62 (h) becomes progressively more associated with the chromosomes (i).
having broad tissue distribution and substrate specificity. p62 is likely phosphorylated by Cam kinase II based on the multifunctional property of the kinase and the fact that active Cam kinase II has been detected in isolated sea urchin mitotic spindles (7). One of the remarkable features of Cam kinase II phosphorylation site sequences is their variability. The enzyme recognizes the following motif, RXXS*/T* (asterisks indicate the phosphorylated residue) in both proteins and peptide substrates (22). Representative examples are synapsin I (44), tyrosine hydroxylase (45), and the autophosphorylation sites on the ␣ and ␤ subunits of Cam kinase II itself (46). However, different phosphorylation site sequences other than this consensus motif have been identified in many other proteins. For example, myelin basic protein has the phosphorylation site sequence RSKYLAS*AST (47), ATP-citrate lyase contains R/KTAS*FSESR (48), and acetyl-CoA carboxylase has FIIGSVS*EDN (49). In general, these sites, as well as the consensus motif, have a characteristic serine or threonine 2-6 residues following a basic residue. The primary sequence of p62 does not contain the consensus recognition sequence RXX(S*/ T*) for Cam kinase II. However, considering the variability of the phosphorylation domains in CaM kinase II substrates, this is not surprising. In the p62 sequence, there are several S or T residues following one or more basic residues, such as 159 KKGS, 209 RPAPS, 226 KDGT, 355 KKTYS, and 369 KSPS. Some of these sequences may serve as phosphorylation sites for Cam kinase II. Indeed, when p62 is phosphorylated in vitro, two CNBr peptides, referred to as pL and pS, are obtained, both of which are radiolabeled. Both pL and pS contain potential Cam kinase phosphorylation sites, 159 KKGS in pL, 209 RPAPS and 226 KDGT in pS, which likely account for this observation. However, it should be noted that results from in vitro phosphorylation experiments may produce different phosphorylated residues than occurs in vivo. This question is currently being addressed by identifying and comparing the phosphoamino acid residues in p62 produced by in vivo versus in vitro phosphorylation reactions.
In addition to Cam kinase II, analysis of the p62 sequence revealed the presence of several potential phosphorylation sites for cAMP-dependent protein kinase, protein kinase C, casein kinase II, and tyrosine kinase. It is interesting to note the potential of cAMP-dependent phosphorylation of p62. cAMPdependent phosphorylation of p62 in the isolated mitotic apparatus does not occur to any great extent (5), perhaps because the site is already phosphorylated when the mitotic apparatus is isolated. Of interest in this regard is a recent report (50) showing that cAMP-dependent protein kinase A is activated by maturation promoting factor and required for the transition from mitosis to interphase. The authors indicate that the cAMP-protein kinase A pathway is activated in the embryonic cell cycle, either directly or indirectly, by maturation promoting factor. These results suggest that the control of activationinactivation cycles of the cAMP-protein kinase A pathway via maturation promoting factor may play a critical role in regulating transitions through M phase of the cell cycle. Recall this is the time in the cell cycle in which p62 is required, as antibodies to p62 arrest the cell cycle at the metaphase-to-anaphase transition when they are injected into dividing sea urchin embryos (6). Furthermore, previous work has shown that microinjection of the protein inhibitor of cAMP dependent protein kinase blocks the first division in sea urchin embryos prior to or during spindle formation (51).
p62 is a nuclear protein (10), is chromatin/chromosome-associated (Fig. 8), and is highly acidic (Fig. 5). Other chromatinassociated proteins, such as HMG-1 (52), also contain acidic domains. Indeed, a large class of nuclear proteins called A Ϫ proteins contain an extended region (or regions) enriched in acidic residues (19,53). This class includes various proteins whose structures are not conserved and whose functions are not identical. It has been demonstrated that A Ϫ proteins exhibit significant binding to the core histones in vivo (54). This suggests that the acidic regions of p62 may serve to bind p62 to histones or to the basic domains in other chromosomal proteins. Considering the cell cycle-dependent nuclear localization of p62, this interaction is a likely candidate for the mechanism that targets p62 to the nucleus in late anaphase (10). It will be important to determine how p62 interacts with chromosomes/ chromatin and what biological function this interaction serves in vivo. To this end, cloned p62 sequences are currently being used in in vitro expression experiments to obtain purified p62, and truncated clones are being used to identify the tubulin and chromatin binding domains of this polypeptide.
How might calcium-and calmodulin-dependent phosphorylation of p62 be involved in mitosis? In vivo, microtubules are sensitive to intracellular calcium concentrations, an effect mediated by calmodulin. For example, microinjection of fibroblasts with calcium-saturated calmodulin induces a rapid disruption of microtubules at the injection site (55,56). Furthermore, microinjection of calcium chelators inhibits mitosis (57)(58)(59), again implicating calcium in this process. Finally, indirect immunofluorescence has localized calmodulin to the kinetochore microtubules of the mitotic apparatus (60,61), and, in a related approach, fluorescently labeled calmodulin microinjected into cells binds to the kinetochore microtubules (62,63). Taken together, these studies strongly implicate calcium and calmodulin in the controlled depolymerization of microtubules that is necessary for anaphase A. Consistent with this hypothesis is the observation that the calmodulin inhibitor, chlorpromazine, blocks mitosis in mammalian cells (64) and that antisense calmodulin RNAs under the control of the Zn 2ϩ -inducible metallothionein promoter cause a transient cell cycle arrest in mouse cells in the presence of Zn 2ϩ (65). Therefore, calcium may act in concert with calmodulin to destabilize the kinetochore microtubules at the metaphase to anaphase transition. This in turn may permit chromosome movement to the poles along the now depolymerizing kinetochore microtubules, via a motor localized to the kinetochore (66 -75). One obvious molecule that may be involved in the cascade of events that begins with calcium release and ends with the successful completion of mitosis is p62, which interacts with the microtubules of the mitotic apparatus (6,9,10), thus stabilizing them. When calcium is released to initiate anaphase, calcium-calmodulin dependent protein kinase II is stimulated to phosphorylated p62. We have previously shown that this kinase is a component of the isolated mitotic apparatus and that it can phosphorylate p62 in vitro (7). Mitotic apparatus microtubules with p62 phosphorylated are unstable (5) and therefore may be readily disassembled by the anaphase machinery that is undoubtedly localized to the kinetochores. This model provides the framework for the design of experiments to test a number of important hypotheses predicted by the model. Such experiments are currently in progress.