A perichromosomal region contains proteins phosphorylated during mitosis in Xenopus Laevis cells

Mitosis involves large-scale structural reorganisation in all eukaryotic cells. These controlled and precise alterations are necessary to ensure stable division of the nucleus. The first phenotypic event that signifies a cell’s entry into mitosis is an orderly condensation of the chromatin into identifiable chromosomes, an event that is essential for ensuring the protection and separation of the genetic material. This condensation probably involves a higher-order packing of nucleosomes through histone Hl interactions (Thoma et al. 1979), linked at intervals to a proteinaceous ‘scaffold’ located in the centre of the chromosome and containing topoisomerase II (for review, see Gasser and Laemmli, 1987). However, little is known of the non-histone proteins occurring outside this core, or of the events involved in initiating and maintaining the condensed state.

Recent work has implicated a number of protein kinases in the control of mitosis. The cytoplasm of unactivated eggs of amphibians contains an activity termed maturation promoting factor, or MPF (Adlakha and Rao, 1986), that induces nuclear envelope breakdown and chromosome condensation when introduced into mature oocytes. An essential component of this activity has been identified as the homologue of the cdc2 gene product of Schizosac-charomyces pombe, p34cdc2 (Dunphy et al. 1988; Gautier et al. 1988; Lohka, 1989). A protein kinase activity is associated with this protein and a number of experiments indicate that this property is responsible for initiating mitotic events (e.g. see Riabowol et al. 1989; Lamb et al. 1990). However, the biochemical substrates of this kinase activity have not yet been defined, nor have the affects that these phosphorylations might exert on their activity. Histone Hl is readily phosphorylated in vitro by p34cdc2, and such phosphorylation, together with histone H3 phosphorylation, is known to be correlated with chromosome condensation (e.g. see Bradbury et al. 1974; Gurley et al. 1978; Ajiro et al. 1983; Matsumoto et al. 1980; Mueller et al. 1985). In addition, the src gene product, pp60c-src, has also been identified as a possible substrate for MPF (Shenoy et al. 1989; Morgan et al. 1989), suggesting that a series of protein kinases may interact during mitosis initiation. A number of protein phosphorylations have been observed to occur in a mitosis-specific manner (e.g. see Davis et al. 1983; Sahasrabuddhe et al. 1984; Lohka et al. 1987; Karsenti et al. 1987), not all of which necessarily occur by direct action of p34cdc2. Therefore, the molecular mechanisms by which p34cdc2 initiates mitosis, and the processes that follow from this primary event remain obscure. It is clear, though, that phosphorylation is an important component of mitotic initiation, and so mitosisspecific phosphorylations may be important mediators of this control.

This paper describes the identification of a polypeptide in somatic cells of Xenopus laevis that is phosphorylated mainly during mitosis. This polypeptide is distinguished from other mitotic phosphorylations in that it is located exclusively around, but not within, the condensed chromosomes, and has an extremely large apparent relative molecular mass on SDS-polyacrylamide gels. The perichromosomal location persists throughout all stages of mitosis and extends for a distance from the chromosome but not into the surrounding cytoplasm. The location and phosphorylation pattern suggests that the distribution is established by phosphorylation and serves as a storage mechanism for nuclear proteins during mitosis.

The Xenopus laevis adult kidney epithelial cell line A6 was grown in Dulbecco’s Modified Eagle’s Medium, diluted to 75% strength, and supplemented with 10% foetal calf serum. Monoclonal antibodies directed against nucleoplasmm were isolated as previously described (Dingwall et al. 1987; Dilworth et al. 1987).

Nucleoplasmin was purified from unactivated eggs of X. laevis by affinity chromatography using monoclonal antibody PA3C5 as previously described (Richardson et al. 1988).

Purified nucleoplasmin was dephosphorylated in vitro by treatment with calf intestinal alkaline phosphatase (Sigma Type VII-N). A 50 μg sample of nucleoplasmin was incubated with 20i.u. (20 μg) of alkaline phosphatase for 4h at 30°C in 7mM Tris-HCl, pH8.5, 3MM ZnCl₂, 3HIM MgCl₂, and 0.2t.i.u.ml⁻¹ aprotinin (Sigma). Controls were incubated similarly but without the addition of alkaline phosphatase.

Nucleoplasmin was immunoprecipitated with monoclonal antibody containing tissue culture fluid (TCF) and protein A-bearing Staphylococcus aureus Cowan I (SAC) (Kessler, 1975) as previously described (Dingwall et al. 1987; Dilworth et al. 1987).

All SDS-PAGE was run as described by Laemmli (1970). Proteins were stained when necessary with 0.5% PAGE Blue 83 (BDH) in 40% methanol, 10% acetic acid.

Cells grown in flasks were washed with phosphate-buffered saline (PBS) then dissolved in situ with Laemmli sample buffer (2% SDS). The samples were then boiled and loaded onto 5% SDS–PAGE and electrophoresed. After separation the proteins were transferred onto nitrocellulose (Amersham International) by the method of Towbin et al. (1979). Nitrocellulose was always batch-tested before use, aB a wide variation in binding ability for phosphorylated polypeptides has been observed. Strips of the nitrocellulose were cut and the unoccupied binding sites saturated with foetal calf serum. Antibody binding proteins were then detected by incubation with antibody-containing TCF, biotinylated anti-mouse Ig and ^S-labelled streptavidin (Amersham International) as described in the Amersham protocol book. Dephosphorylation, when necessary, was performed before saturation of unreacted sites. Nitrocellulose strips were incubated at 30°C for 4h in 75 mw Tris-HCl, pH8.8, 3mM ZnClz, 3 HIM MgCl_z±20pg alkaline phosphatase, or 50 mM sodium citrate, pH 5.0, 3mM MgCl₂±10i.u. acid phosphatase from white potato (Sigma TypeVII), washed twice and then processed as above.

X. laevis tissue culture cells were seeded onto glass coverslips and allowed to grow for 48 h before use. After appropriate incubation with or without colchicine (Sigma), the cells were washed with PBS and then fixed with 3.7% formaldehyde solution in PBS for 15 min at room temperature. The cells were then washed with PBS, permeabilised with 1% Nonidet P40 for 5 min, washed again, and incubated with monoclonal antibody-containing TCF for 1 h at room temperature. After washing with PBS, the bound antibody was detected with a 1:50 dilution of Texas Red-conjugated sheep anti-mouse IgG (Amersham International). After washing with PBS the coverslips were mounted in Aquamount (BDH) and viewed on a Nikon Optiphot microscope by epifluorescent illumination. Photographs were taken onto Ilford XP-1 film.

I2A3 is a monoclonal antibody of IgGl subclass that was raised against nucleoplasmin purified from unactivated eggs of X. laevis. It reacts well with nucleoplasmin from eggs (Fig. 1, lane e), probably binding within the proteaseresistant ‘core’ region at the N terminus of the molecule (Dingwall et al. 1987), but shows a variable reaction with nucleoplasmin isolated from oocytes. The major difference between egg and oocyte nucleoplasmin is the extent of phosphorylation (Cotten et al. 1987; Dingwall, Mills and Laskey, personal communication); egg nucleoplasmin contains a significantly greater number of phosphate groups. This additional phosphorylation retards the migration of egg nucleoplasmin on SDS-PAGE (Fig. 1, lanes a and c). To determine if this phosphorylation is affecting the binding of I2A3 antibody, immunoprecipitation was performed before and after treatment of egg nucleoplasmin with high levels of calf intestinal alkaline phosphatase (Fig. 1). Dephosphorylation of egg nucleo-plasmin resulted in an anticipated increase in gel mobility, and also reduced the interaction with I2A3 antibody below detectable limits (Fig. 1, lane g). A control antibody directed against the ‘tail’ sequence of nucleoplasmin showed no loss of reaction (Fig. 1, lane f), suggesting that treatment was not inhibiting antibody binding non-specifically. Similar results have been obtained after dephosphorylation with alkaline phosphatase from Escherichia coli, and acid phosphatase from sweet potato (data not shown). It is likely, therefore, that I2A3 recognizes phosphorylated nucleoplasmin only.

To determine whether a similar phosphorylation-specific binding of I2A3 antibody occurs in somatic cells, immunofluorescence analysis was performed on X. laevis tissue culture cells. Fig. 2 shows the result of such an experiment at low-power magnification. Most actively growing cells exhibited a low level of fluorescence within the nucleus, but a few cells showed a much stronger signal (Fig. 2A). Closer inspection of these brightly fluorescing cells suggested that they were in mitosis when fixed. To determine whether the high level of staining was related to mitotic events, cells were arrested in mitosis by the addition of colchicine for 12 h prior to fixation, then immunofluorescence was performed with 12A3. Fig. 2B shows that this dramatically increased the proportion of brightly fluorescing cells, exactly matching the increase in mitotic cell numbers found in these cultures. These antibody binding cells were therefore probably in mitosis when fixed.

The fluorescent staining pattern observed could represent the presence of nucleoplasmin (or a somatic variant; Schmidt-Zachmann et al. 1987; Cotten and Chalkley, 1987) or a cross-reaction of I2A3 with another polypeptide species. To investigate these possibilities, Western blot analyses of Xenopus tissue culture cell lysates were performed. Fig. 3 shows the results of such an experiment. Although a weak reaction was observed in the 30·10³M_r molecular weight region of the gel (the migration position of nucleoplasmin, not showm), the major reacting polypeptides migrated considerably more slowly. Two species were easily detected, one migrating with an apparent molecular weight of approximately 130×10³M_r (pl30), and another migrating much more slowly than a 200×10³M_r marker (myosin), and therefore probably with a molecular weight in excess of 400×10³M_r (p400+; Fig. 3, lane b). Neither polypeptide was seen as a major species on a similar gel of the cell lysate stained with Coomassie Blue prior to transfer. Treatment of the cells with colchicine before lysis resulted in a large increase in the binding of antibody to p400+ only (Fig. 3, lane b′)-pl30, and a polypeptide reacting with a control antibody, showed no alteration. It is likely, therefore, that pl30 represents the low level of staining seen in interphase nuclei, whereas p400+ is responsible for the strong staining of mitotic cells seen above.

As I2A3 binds phosphorylated nucleoplasmin specifically, alterations in the binding of pl30 and p400+ after dephosphorylation were investigated. Fig. 4 shows that incubation of the polypeptides bound to nitrocellulose with either alkaline or acid phosphatase prior to reaction with I2A3 considerably reduces antibody binding to both pl30 and p400+. A control antibody reaction shows no alteration in binding, suggesting that non-specific inhibition is not occurring. The inability to abolish binding completely in this case probably reflects the difficulty in removing all phosphate from a polypeptide when bound to nitrocellulose. Therefore, it is likely that I2A3 only binds to the phosphorylated forms of both pl30 and p400+, and that phosphorylation of the latter occurs at a much higher level during mitosis.

The subcellular location of p400+ was investigated by immunofluorescence and high-power examination of mitotic cells. Fig. 5A shows that when reacted with I2A3 the mitotic condensed chromosomes are clearly visible as an unstained region against a high level of antibody binding around the chromosome. Despite the absence of a nuclear envelope during mitosis, staining does not spread throughout the cytoplasm but is confined to the chromosome mass. To ensure that this absence of binding was not a consequence of the antibody being unable to penetrate the chromosome, a control monoclonal antibody reacting with histone Hl was used. Fig. 5B shows that this antibody binds to the whole chromosome, with no discernible extrachromosomal staining.

The distribution of p400+ was farther investigated by examining cells at various stages of mitosis. Fig. 6 shows that the ‘perichromosomal’ staining occurs at all stages of mitosis. Phosphorylated p400+ fluorescence can be seen before the chromosomes have fully condensed and only decreases once decondensation has begun. Therefore, throughout mitosis the condensed chromosomes are surrounded by p400+ in its highly phosphorylated form.

The fundamental processes and controlling mechanisms involved in entry into and progress through mitosis are beginning to be understood. The realization that a homologue of the cdc2 gene found in the yeast S. pombe is also present and playing a key role in mitosis in higher eukaryotes (see Lohka, 1989, for review) has provided an essential reference point for these studies. The intrinsic protein kinase activity of the p34cdc2 protein and the demonstration that this is essential for mitotic control, together with a wealth of earlier circumstantial evidence, suggests strongly that many of the events are controlled by protein phosphorylation-déphosphorylation. This has intensified the search for proteins that are phosphorylated specifically during mitosis. A number of protein phosphorylations have been observed to be mitosis-specific CDavis et al. 1983; Sahasrabuddhe et al. 1984; Lohka et al. 1987; Karsenti et al. 1987) but little is understood of their activities. This paper extends these observations to another set of proteins and reports the isolation of a monoclonal antibody that recognizes these proteins only in their phosphorylated state.

I2A3 is a monoclonal antibody raised against X. laevis nucleoplasmin that also interacts with two polypeptides found in somatic cells. In each case the interaction is abolished if the proteins are dephosphorylated. This phosphorylation dependence suggests that the phosphate group is either part of the antigenic site or induces a conformational change necessary for generating this site. The binding of I2A3 to SDS-denatured polypeptides in a Western blot suggests that the former is more likely. This raises the interesting possibility that the antibody binding region is similar to the sequences recognized by a protein kinase that phosphorylates all three protein species. The cell cycle-dependent phosphorylation of nucleoplasmin and p400+ suggest that this kinase is regulated and may be part of the mitosis-controlling mechanism. The lack of an obvious cell cycle dependence on the phosphorylation of pl30 is intriguing and may reflect a change in substrate specificity of the kinase during the cell cycle. The specificity of the I2A3 antibody will enable this possibility to be investigated further.

I2A3 reacts with a high M_r polypeptide in Xenopus somatic cells. This polypeptide (p400+) shows a dramatic increase in phosphorylation during mitosis. The rapid increase in immunoreactive material on entry into mitosis, with very little evidence for intermediate levels of fluorescence, strongly suggests that the polypeptide is present in an unmodified form during other stages of the cell cycle and is phosphorylated on entry into mitosis and dephosphorylated on exit, though rapid synthesis and degradation cannot be eliminated completely. This situation can be investigated more fully once an antibody against the unmodified polypeptide becomes available. The likely dramatic increase in phosphorylation of p400+ during mitosis and its unusual properties raise intriguing questions concerning its activity. The distribution around condensed chromosomes suggests that it may have a role in either establishing or maintaining the condensed state, or of protecting the chromosomes during their separation at mitosis. Previously, Hl, and possibly H3, phosphorylation has been connected with chromosome condensation states (Bradbury et al. 1974; Gurley et al. 1978; Matsumoto et al. 1980; Ajiro et al. 1983; Mueller et al. 1985) although this is not a complete correlation (see, e.g., Tanphaichitr et al. 1976; Fischer and Laemmli, 1980; Kyrstal and Poccia, 1981; Allis and Gorovsky, 1981). Hl is also known to be a substrate for the p34cdc2 kinase in vitro. It will be interesting to determine whether p400+ and nucleoplasmin are also direct substrates of p34cdc2 or whether other protein kinases in a cascade reaction intervene. However, a more unusual role for p400+ is also suggested by its location and phosphorylation state. There are other reports of ‘perichromosomally’ located proteins (Ochs et al. 1983; McKeon et al. 1984; Chaly et al. 1984; Hugle et al. 1985a; Hugle et al. 19855; Spector and Smith, 1986; Schmidt-Zachmann et al. 1987; Luji et al. 1987; Wataya et al. 1987). Interestingly, one of these is related to nucleoplasmin (Schmidt-Zachmann et al. 1987) and has been shown to be phosphorylated by p34cdc2 (Peter et al. 1990). Many of these proteins show a redistribution from the nucleolus or the nuclear lamina to the perichromosomal region during mitosis, suggesting that this region may be involved in storage of proteins during mitosis. The presence of many structural or specialised nuclear proteins that may be harmful when free in the cytoplasm, together with a protein related to a known storage protein, nucleoplasmin (see Dilworth and Dingwall, 1988, for a review of other roles for nucleoplasmin), suggests that proteins may be located in a storage complex bound to the chromosomes during mitosis. This would also ensure an equal distribution of essential nuclear components between the resulting daughter nuclei after mitosis. Phosphorylation is therefore likely to be involved in specifying the redistribution of these proteins to this location and establishing the perichromosomal structure. This model suggests at least three possible roles for p400+: (1) as a structural component of the perichromosomal complex; (2) as a storage protein providing a ‘molecular chaperone’ activity for the proteins organized at this site; or (3) as a nuclear protein being stored here. The perichromosomal region has been examined by electron microscopy (Capeo and Penman, 1983) and shown to contain a filamentous structure that could not be identified. It is feasible that the unique nature of p400+ (very few cellular proteins migrate as slowly on SDS-PAGE) could have precluded its previous identification as a structural component of the nucleus.

The large size of p400+ suggests that it may be formed from a number of linked smaller subunits. These bonds withstand boiling in SDS under conditions known to dissociate the pentameric structure of nucleoplasmin, an exceptionally stable non-covalent interaction that can resist some SDS treatments. Therefore, it is probable that any bonds between subunits are covalent. In addition, the polypeptide does not react with other anti-nucleoplasmin monoclonal antibodies and migrates as a discrete species on SDS-PAGE (the few immunoreactive species seen below the major band on long exposures probably represent proteolysis of the intact molecule during isolation). It is therefore likely that the polypeptide is not a multimer of nucleoplasmin but a discrete species, and so very unusual in its size. Determining the role of such a large protein within the cell will be very interesting. It may prove possible to investigate the function of this protein, and particularly the phosphorylation identified here, by using I2A3 as a tool. The binding of I2A3 antibody to the phosphorylated polypeptide may prevent access by a phosphatase during exit from mitosis. Therefore, the effects on chromosome condensation, and other cell cycle processes, of I2A3 antibody microinjection into cells may provide important insights into the role of this phosphorylation in vivo.

I particularly thank Professor R. A. Laskey and all other members of the CRC Molecular Embryology Research Group for their stimulating discussions and help throughout the project. I also thank Mrs B. Rodbard for typing the manuscript. This work has been generously supported by the Cancer Research Campaign, UK.