(Received for publication, July 2, 1996, and in revised form, November 7, 1996)
From the Department of Biological Sciences, Dartmouth College, Hanover, New Hampshire 03755 and the Marine Biological Laboratory, Woods Hole, Massachusetts 02543
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 CleavagePurified p62 (100 µl) was dialyzed against
0.5 M NH4HCO3 at 4 °C overnight,
lyophilized, and resuspended in 150 µl 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 N2 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 SDS-gel (11). The fragments were blotted to
polyvinylidene difluoride membrane in CAPS1
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 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
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 NH2-DDDRNVIEVETINFDGDTVIQPLLSLR-COOH. The
sense primer, referred to as 3Aup
(5-GAYGAYGAYMGIAAYGTIAT-3
, where I = inosine, M = A + C,
and Y = C + T), was designed from the first seven residues of the
Trp-3 peptide. The antisense primer, referred to as 3Adown (5
-ARIGGYTGDATIACIGTRTC-3
, where D = A + G + T and R = A + G), was designed from residues 17-23 of the Trp-3 peptide. Finally, a
pool of 26-mer degenerate oligonucleotides, referred to as 3A (5
GARGTIGARACIATIAAYTTYGAYGG-3
), 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 3Aup and
3Adown 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 3Aup and
3Adown. 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 pCRTM vector
(Invitrogen), and the resulting colonies were screened by hybridization
for a specific PCR product using
-32P-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 sequencer. 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 GenBankTM 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 µg of cloned DNA in the reticulocyte lysate-based TNT system (Promega, Madison WI) in the presence of [35S]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 µm 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 µg 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 × prehybridization mix (1 × 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 × 106 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
[-32P]dCTP by random priming using
Ready-To-GoTM DNA labeling beads from Pharmacia Biotech
Inc. and dCTP. After hybridization, the membrane was washed once for 30 min in 2 × SSC, 750 µM sodium pyrophosphate, and
0.05% (w/v) Sarkosyl at room temperature and four times, for 30 min
each, in 0.1 X SSC, 375 µM 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 µl of packed eggs were resuspended in a 1.5-ml microcentrifuge tube with 700 µl 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, Mr 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 µl 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 µl 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 × SSC, the DNA was transferred to nitrocellulose by capillary transfer in 20 × SSC. The membrane was prehybridized at 65 °C in prehybridization solution (see previous section), then hybridized with 32P-labeled p62 cDNA in the same solution at 65 °C for 12 h. After hybridization, the membrane was washed at room temperature in 2 × 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 NaH2PO4, 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 MgSO4, 10 µg/ml leupeptin, 1 mM dithiothreitol, 160 µM 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 three-fourths 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 20 °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.
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 Blue-stained 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 (NH2-AKEYF-COOH) and CNBr-2 (NH2-KGKGD-COOH).
Tryptic peptides were generated and purified by HPLC. Three peptides were selected, sequenced, and the resulting peptide sequences were called Trp-1 (NH2-LGLNESTNLDIGLQPPVT-COOH), Trp-2 (NH2-NVIEVETINFDGETVIQPLLSLR-COOH), and Trp-3 (NH2-DDDRNVIEVETINFDGETVIQPLLSLR-COOH). Note that Trp-2 is a fragment of Trp-3 formed by cleavage after an Arg residue found at position number four of Trp-3.
Library ScreeningTwo 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 3 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
5-untranslated region. The open reading frame is followed by a
3
-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
3 position, where a purine residue has been
shown to be important for the initiation of translation (16), and a G
at the +4 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 3
-untranslated region. The 3
-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
polyadenylation signals are most likely not used, at least by the
transcript that generated clone SKh11, as they are >35 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.
Nucleotide and deduced amino acid sequence of p62 cDNA clone SKh11. The cDNA clone SKh11 was identified from a cDNA library using antibodies and RT-PCR. Shown here is the nucleotide sequence of SKh11 and the corresponding primary sequence. Tryptic peptides Trp-1 and Trp-3 are shown by underlines beneath the deduced amino acid sequence, and the two CNBr peptides are shown in bold type and by underlines beneath the deduced amino acid sequence. Asterisks denote the first in frame stop codon. Three polyadenylation sequences (ATTAAA) are underlined; the putative poly(A) addition signal is the one at nucleotides 3336-3341. GenBankTM accession no. U76750[GenBank].
In Vitro Translation of p62
The protein encoded by clone
SKh11 has a calculated Mr of 46,388. On
SDS-gels, p62 has been estimated to have a Mr 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 Mr 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 Mr
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 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.
Sequence Analysis of p62
Analysis of the amino acid
composition reveals that the p62 polypeptide is rich in glutamyl
(19.7%), aspartyl (13.9%), and lysyl (12.9%) residues. No cysteine
is present in p62, indicating that p62 does not fold by intrachain
disulfide bonds or form interchain disulfide bonds with other proteins.
Based on amino acid charge distribution and hydrophobicity, the p62
polypeptide can be divided into three distinct domains (Fig.
5): an NH2-terminal acidic domain (residues
1-120, Mr = 13,457; isoelectric point (pI) = 4.21), called domain N, a middle, also acidic domain (residues
121-320, Mr = 22,684; pI = 3.79), called
domain M, and the COOH-terminal basic domain (residues 321-411,
Mr = 10,281; pI = 10.16), named domain C. Domain M can be further divided into five subdomains, named m1 to m5.
There are three acidic (m1, m3, and m5) and two basic subdomains (m2
and m4) which alternate with each other and are either extremely rich
in aspartyl and glutamyl residues or in lysyl residues. Subdomain m1
(pI = 3.25) consists of a stretch of 36 amino acids (residues
121-156) 67% of which are acidic (17 Glu, 7 Asp). Subdomain m2
(pI = 12.06) contains 25 amino acids (residues 157-181), of which
40% are basic (8 Lys, 2 Arg). Acidic residues are absent from m2.
Subdomain m3 (pI = 3.04) is composed of 19 amino acids (residues
182-200), 95% of which are acidic (12 Glu, 6 Asp). Subdomain m4
(pI = 10.58) spans 45 amino acids (residues 201-245) and contains
29% basic residues (10 Lys, 2 Arg, 1 His). Subdomain m5 (pI = 2.77) is composed of 75 amino acids (residues 246-320), of which 81%
are acidic (36 Glu, 25 Asp).
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 microtubule-associated 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 cAMP-dependent 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.
GenBankTM SearchProtein sequence similarity between p62 and
known proteins was analyzed by searching the data libraries of
GenBankTM 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
NH2-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
phosphorylated 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.
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 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 CNBr-derived 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 (32-36). 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
Ca2+/ calmodulin-dependent protein kinase (Cam
kinase) which co-purified 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 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 159KKGS, 209RPAPS, 226KDGT,
355KKTYS, and 369KSPS. 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,
159KKGS in pL, 209RPAPS and 226KDGT
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. cAMP-dependent 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 activation-inactivation 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 chromatin-associated proteins,
such as HMG-1 (52), also contain acidic domains. Indeed, a large class
of nuclear proteins called A 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
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 (57-59), 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 Zn2+-inducible metallothionein promoter cause a transient cell cycle arrest in mouse cells in the presence of Zn2+ (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.
We thank Dr. Robert E. Palazzo for his encouragement and for his critical reading of the manuscript. Generous colleagues provided us with essential research materials and protocols, as indicated under "Materials and Methods." We thank in this regard Drs. Kathy Foltz, Robert Goldman, and Jonathan Scholey.