EXPERIMENTAl.
CELL
RESEARCH
195,
145-153
(1991)
Monoclonal Antibody CC-3 Recognizes Phosphoproteins in Interphase and Mitotic Cells ALAINTHIBODEAUANDMICHEL OntogmPsc
ct Ghnhtiyw
Mokkulaire,
Ccntre
de Recherche
du
INTRODUCTION Understanding the mechanism and regulation of mitosis will require the identification and characterization of proteins or activities specific to different phases of the cell cycle. Recent progress in immunological techniques and in the preparation of antigenic material has permitted the production of several antibody probes recognizing diverse molecules specific to the mitotic phase (M phase) or that differentially react during the interphase to M phase transition. Using an antiserum raised against a nonhistone nuclear protein preparation, Lyrequests Centre (QuChec),
should he addressed de recherche du GlV 4G2, Canada.
at: Ontoge&se CHUL, 2705
VINCENT’ and
D4partement
de Mhdecine,
UniuersitP
Laclal
dersen and Pettijohn [ 11reported the first evidence that a predominantly interphase nuclear protein (NuMA) becomes part of the mitotic apparatus during M phase. Microtubule-associated proteins isolated from different species have also been extensively used as immunizing material giving rise to a panoply of mAbs directed against a variety of interphasejmitotic structures [2-S]. Other related mAbs have been produced by immunizing animals with isolated mitotic spindles [9, lo], chromatin-containing nuclear extracts [ 111, DNA-binding proteins from chick embryo extracts [la], and mitotic cell extracts [13, 141. In most of these studies, the recognized antigens behave like cell division-related molecules in the sense that they are present in interphase (nuclear or cytoskeletal) structures and become part of mitotic components (mitotic spindle, centrosomes, kinetochores, midbody) during mitosis. The transition from int.erphase to M phase is marked by an increase in the phosphorylation of several intracellular proteins. For instance, histone Hl becomes highly phosphorylated [ 15,161 and depolymerization of the nuclear envelope lamina is accompanied by lamin phosphorylation [17,181. Similarly, structural modifications of the cytoskeleton at the onset of mitosis correlate with alterations in the state of phosphorylation of vimentin [ 19, 201, keratin [21], and desmin [22]. Cell cycle-modulated phosphorylation has also been reported on proteins which are part of specialized mitotic structures including the mitotic spindle [23,24], centrosomes [25], microtubule organizing centers [26], and proteins that undergo intracellular distribution during the interphase to M phase transition [7, 271. The search for new molecules involved in embryonic development led us to obtain monoclonal antibodies (mAbs) that reacted in a cell cycle-dependent manner [28]. In the present study, we show t.hat one of them, CC-3 (previously designated Z4E9 [as]), recognizes a phylogenetically well-conserved 255-kDa phosphoprotein that is sequestered within the nucleus during interphase. At mitosis, CC-3 reveals a complete new set of phosphoproteins that are associated with the mitotic apparatus. In both cases, the reactivity of CC-3 depends on the phosphorylation of the antigens since it is abolished by the removal of phosphate groups.
Among a library of monoclonal antibodies (mAbs) recognizing developmental markers in the chick embryo, mAb CC-3 was selected because of its differential immunostaining of mitotic cells. The intracellular distribution of the CC-3 antigen (CC-3a) throughout the cell cycle was visualized by immunolocalization. In interphase cells CC-3a resided in the nucleus and was arranged in distinct extranucleolar clusters. At prophase, the nuclear reactivity of CC-3a considerably increased and subsequently extended to the cytoplasm at metaphase. From metaphase through anaphase, most of the reactivity was associated with the mitotic apparatus. During cytokinesis CC-3a was detected in the midbody and also in discrete speckles dispersed throughout the cytoplasm. The initial interphase pattern was then restored in the two daughter nuclei. Immunoblot analysis demonstrated that a 255-kDa phosphoprotein was present only in the interphase nucleus and that a complete new set of phosphoproteins accounted for the mitotic cell reactivity. The binding of CC-3 was dependent on the phosphorylation of its antigens. CC-3a is an evolutionary conserved molecule; it is present in such phylogenetically distant species as Drosophila and humans. Furthermore, the unique behavior of CC-3 on sections of normal, embryonic, and regenerative tissue and in cell culture immunostaining make it a reliable tool to identify mitotic foci. ( 1991 Academic Press. Inc.
’ 7’0 whom reprint Gknktique Molkculaire, Laurier. Sainte-Fov
CHIJL
et houl
145
0014.4827/91
$3.00
146
THIBODEAIJ
MATERIALS
AND
METHODS
Fertilized White leghorn eggs were obtained from a local hatchery and maintained in a humidified air-forced incubator at 38°C. The embryos were staged according to the number of hours or days of incubation as indicated. Antibodirs. MAb CC-3 (IgG) was obtained after immunization of a Balb/C mouse with isolated pharyngeal regions from 72-h chick embryos (281. The CC-3 hybridoma was grown in Iscove’s modified Dulbecco’s medium supplemented with 10% fetal calf serum and the culture supernatant was used for immunofluorescence. CC-3 ascites fluid (1:lOOO dilution) was used for immunoblots. CC-3 was revealed using either a fluorescein isothiocyanate (FITC)-conjugated goat IgG anti-(mouse immunoglobulins) (FITCGAM) in immunofluorescence or a iz51-labeled GAM (““I-GAM) (10 pCi/pg) in immunoblot studies. Immunofluorescent
staining
of
cell
cultures
and
tissue
swtion.s.
Primary fibroblast cell cultures from ‘i-day-old chick embryos (CEF) were prepared and maintained as described [29]. The Drosophila cell line (KcIII-10) [30] was grown in D-22 supplemented with 10% decomplemented fetal calf serum (FCS) at 23°C [31]. Chinese hamster ovary (CHO) cells were grown in Iscove’s modified Dulbecco’s medium (IMDM) supplemented with 10% FCS. Mouse cells (NIH 3T3), potoroo kidney cells (PtK,), human fibroblasts (primary culture), and human amniocytes (primary culture) were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% FCS in a humidified 5% CO, atmosphere at 37°C. For immunost,aining, cells were washed with PBS (8 mMNa,HPO, * 7 H,O, 1.5 mM KH,PO,, 140 mM NaCI, :1 mM KCl, pH ‘i.4), fixed in methanol for 20 min at -20°C and finally soaked in PBS. Paraffin embedding and tissue sectioning were carried out as previously described [32]. Fixed cells and rehydrated tissue sections were incubated with CC-3 for 45 min at room temperature, washed in PBS, and incubated with FITCGAM in the same conditions. Immunofluorescence observation was performed on a Zeiss Axiophot microscope equipped with epifluorescence optics. Cell .s~~nchronization. CHO cell cultures were allowed to reach an exponentially growing phase in 75cm2 culture flasks (Costar). The monolayer was then gently squirted with the overlying medium to get rid of loosely attached cells and the medium was replaced with 1.5 ml of complete medium (IMDM/lO% FCS) containingcolcemid (0.1 fig/ ml). After a 12-h incubation, the nonadherent cells were removed by selective mechanical detachment [33], pelleted, resuspended in IMDM containing colcemid, and kept on ice. This first collection was referred to as the mitotic fraction. The remaining adherent cell layer was lightly trypsinized (0.0125% trypsin, 90 s) to remove the persistent adherent mitotic cells and this fraction was discarded. The cells still attached to the bottom of the Hask were collected by complete trypsinization (0.05% trypsin, 5 min) of the layer, washed with complete medium, resuspended in IMDM, and kept on ice. This second collection was referred to as the nonmitotic fraction. The cells of both fractions were finally pelleted and resuspended in electrophoresis sample buffer containing 2 mM sodium orthovanadate as a phosphatase inhibitor [34]. To estimate the mitotic index, an aliquot of cells was swollen in 0.8% sodium citrate buffer (5-20 min) at room temperature, dried on a microscope slide, and fixed in methanohacetic acid (3:l). The slides were stained with propidium iodide (0.17 eg/ml; 30 s) and mitotic cells were scored as those showing individual condensed chromosomes in fluorescence microscopy. The mitotic index of the mitotic fraction was 85% while the nonmitotic fraction was over 98% pure. Preparation of cell and tissue extracts. Whole cell and chick embryo brain extracts were obtained by directly homogenizing the cells or the tissue in electrophoresis sample buffer. To obtain CEF nuclear and cytoplasmic fractions, cells were washed with PBS, resuspended in HMK buffer (15 mM Hepes, 5 mM MgCl,, 2 mM KCl, 1 mM PMSF, 1 pg/ml aprotinin, pH 7.2), and allowed to swell 10 min on ice. One volume of HMK buffer containing 0.5 A4 sucrose was then added
AND
VINCENT
and the cell membranes were disrupted using a Dounce homogenizer. The percentage of disrupted cells (>99% after 100 strokes) was rou tinely monitored by phase contrast microscopy. The cytoplasmic and nuclear fractions were separated by centrifugation at 1OOOg. The former was concentrated by ethanol precipitation and recovered by sedimentation at 15OOg. The nuclear fraction (1OOOg pellet) and the cvtoplasmic fraction (15OOg pellet) were dissolved in the electrophoresis sample buffer. E:lpctrophoresisandimmunoblottin~. Polypeptidesinelectrophoresis sample buffer (63 mM TrissHCl, 10% glycerol, 2.3% SDS 5% rj-mercaptoethanol, 0.005’;> bromophenol blue, pH 6.8) were heated at 95°C for 5 min and separated on 7.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Protein standards used in gel migration were: actin (42 kDa) bovine serum albumine (68 kDa), phosphorylase o (94 kDa). spectrin /j (220 kDa), and spectrin (Y (240 kDa). Protein separation was verified by Coomassie blue staining. For immunoblot analysis, proteins were electrophoretically transferred to nitrocellulose membranes (Gelman Sciences, Ann Arbor) by the method of Towbin et al. [35]. The transfer efficiency was confirmed by Ponceau S staining. Prior to antibody contact, the ni trocellulose membranes were saturated with 5% blotto (PBS contain ing 5% skim milk powder) for 1 h at 37°C. The membranes were then reacted with U-3 (45 min at 37”C), washed in PBS containing 0.05!;, Tween-20 with several solution changes, incubated with ““I-GAM in the same conditions, and finally washed as above. Immunoblots were exposed to Kodak XAR-5 film. AIkalinc~ phosphatasr twatmfvzt. Nitrocellulose strips to which electrophoretically separated polypeptides had been transferred were rinsed in Tris/NaCl (50 mM Tris. 150 mM NaCl. pH 7.5), blocked with 5% blotto in Tris/NaCl, and rinsed again in Tris/NaCl. The strips were then deposited in Tris/MgCl, (100 mMTris, 1 mM MgCl,, pH 8.0) containing a cocktail of protease inhibitors (5 pg/ml antipain, 5 &ml leupeptin, 5 pg/ml pepstatin, 1 pg/ml aprotinin, and 1 mM PMSF) and alkaline phosphatase (molecular biology grade, BoehringerMannheim) was added at 15 units/ml and the strips were incubated 30 min at :37”(‘ followed by 90 min at 25°C. After digestion, the strips were subsequently washed with PBS, PBS containing 0.05”C Tween-20 and then resaturated with PBS containing 5% blotto. To test the effects of phosphatase treatment on immunofiuorescence, PtK, cells were briefly rinsed with Tris/NaCl, fixed in methanol, washed again with Tris/NaCI, and soaked in Tris/MgCI, containing the protease inhibitor cocktail. Alkaline phosphatase was then added at 5OLlOO units/ml and the cells were incubated 69 min at 37°C followed by 60 min at 25°C. After digestion the cells were washed with PBS. The nitrocellulose strips and the cells were finally reacted with the primary and secondary antibodies. Control experiments were carried out by incubating nitrocellulose strips or fixed cells in the same digestion conditions without adding enzyme or in the presence of enzyme with 50 mM sodium phosphate (pH 8.0) as an alkaline phosphatase inhibitor [as].
RESULTS Reactivity Sections
Pattern of CC-3 on Chick and Cell Cultures
Embryo
Tissue
Our hybridoma supernatant screening procedure [28] allowed direct visualization of the pattern of CC-3 reactivity on 72-h chick embryo tissue sections. As shown in Fig. lA, CC-3 reacted with scattered cells demonstrating no obvious tissue, cell lineage, or region specificity, except for the internal periphery of the neural tube where the labeled cells seemed to accumulate. A higher magnification of the neural tube (Fig. 1B) shows that several positive cells are aligned in the germinal zone
CELL
DIVISION-RELATED
it was gathered (Fig. 2G). Intracellular
FIG. 1. Histological staining of chick embryo tissue sections with (‘C-3. Immunofluorescence is observed (A, B) on transverse sections at the pharyngeal level of a 72-h embryo, (C) at the mesencephalic level of a 40-h embryo and (D) on a longitudinal section of the limb bud of a 96 h-stage embryo. CC-3 differentially identify mitotic cells. nt. neural tube. (AI 35~. (B) 70x, (C) 1:30x, (D) 45~.
near the ventricular surface where a high mitotic index is present. The immunoreactivity was not restricted to this particular stage of development. For example, CC-3 also stained individual cells in the neural tube on a 40-h embryo and in the wing bud of a 96-h embryo (Figs. 1C and D, respectively). In order to identify the cellular and subcellular structures targeted by mAb CC-3, the immunofluorescence staining pattern of CEF was investigated. As was the case with histological sections, CC-3 brightly stained a limited number of cells (Fig. 2A), whereas the others exhibited a weaker nuclear fluorescence. A closer examination of the strongly reacting cells demonstrated that these cells were undoubtedly undergoing mitosis. Indeed, metaphase cells displayed a strong staining of mitotic structures, namely, the spindle and the centrosomes (Fig. 2C). A diffuse staining was also observed throughout the cytoplasm. The chromosomes aligned at the metaphase plate were not immunoreactive. In agreement with this result, CC-3 staining of chromosomal spreads, in which microtubular structures have been depolymerized, showed that the immunoreaction was only present in the cytoplasmic portion and in a sheath wrapping each chromosome but that the chromatin, as such, was not labeled (data not shown). After the duplicated chromosomes segregated reactivity was found in the centrosomal areas (depolymerized kinetochore microtubules) and on the polar microtubules (Fig. 2E). Furthermore, bright speckles can be observed over the cytoplasmic staining at this stage. The reactivity at interphase was entirely restricted to the nucleus where
147
PHOSPHOPROTEINS
into
distinct
Compartmentalization
extranucleolar
clusters
of the CC-3 Antigen
Since the immunofluorescence was strictly confined to the nucleus in fixed interphase cells, we tried to confirm this by cell fractionation. Total extract and nuclear and cytoplasmic fractions prepared from a primary culture of CEF were immunoblotted with CC-3. Figure 3 shows that the CC-3 antigen (CC-3a) was present on a molecule of 255 kDa whose immunoreactivity was entirely recovered in the nuclear fraction. In these unsynchronized cultures, in which mitotic cells represented not more than 3% of the total cells (assessed by CC-3 and DNA staining), it can be assumed that the 255-kDa band accounted for the nuclear immunoreactivity of the interphase cells. Furthermore, the inability to immunodetect a-tubulin in the nuclear fraction (data not shown) suggests that no or at most trace residual cytoskeleton elements contaminated this fraction. Other minor CC-3-immunoreactive bands were occasionally observed but their relative intensities were low and fluctuating compared with that of the 255-kDa band (see for instance, Figs. 5 and 6). Intracellular M Phase
Distribution
of the CC-3 Antigens
during
Since the immunofluorescence pattern of CC-3 (nuclear clusters, mitotic apparatus) was very well conserved in all species examined (not shown, see also Fig. 6), immunofluorescence studies were performed on PtK, cells in order to visualize the stepwise transition in the pattern of reactivity from interphase to M phase. These cells were found to be particularly well suited for this purpose since, unlike other cell types, they do not round up when entering mitosis. During mitotis, a change in CC-3a reactivity was first seen at prophase, when the nucleus became heavily stained (compared to the interphase nucleus) and the positioned centrosomes were recognized (Fig. 4A). In metaphase cells, the immunoreactivity was diffusely distributed throughout the cytoplasm, being preferentially associated with the mitotic apparatus (spindle and centrosomes) (Fig. 4B). During the separation of the duplicated chromatids, as the kinetochore microtubules depolymerized, the cytoplasmic reactivity faded but a signal was preserved at the mitotic poles and in the polar microtubule area. Some fluorescence was also visible at the periphery of the chromosomes (Fig. 4C). At early telophase, when the cleavage furrow was initiated, the centrosomes were still reactive while the median region of the midbody corresponding to the overlapping of the polar microtubules fluoresced more strongly than the rest of the cytoplasm (Fig. 4D). This region of the midbody was strongly labeled throughout cytokinesis and the nuclear fluorescence slowly lighted up as the chromatin decon-
148
FIG. 2. Cytological staining of CEFs with CC-S. Immunofluorescence is ohserved in (A) a random cell population, (C) in a cell at metaphase, (E) during anaphase-telophase transition, and (G) at interphase. The micrographs on the right, (B, D, F, and H) represent the corresponding phase contrast image. The micrograph in C has been underexposed to emphasize the fluorescence of the mitotic apparatus over that of the cytoplasm. Interphase nuclei (arrows) and mitotic cells (arrowheads) are indicated. CC-S stains the interphasic nucleus and the mitotic apparatus. ce, centrosomes; ch, chromosomes; ms, mitotic spindle, no, nucleolus; nu, nucleus. (A, B) 160X, (C-H) 500X.
densed (Figs. 4E-4G). In the earliest stage of cytokinesis, discrete cytoplasmic speckles were also detected (Fig. 4E). As the daughter cells reachedcomplete separa-
tion, the labeling of the midbody progressively diminished and the interphase nuclear reactivity pattern was completely restored (Fig. 4H).
CELL
A
255
B
DIVISION-RELATED
c
-
FIG. 3. Intracellular compartmentalization of the CC-3 antigen. Immunoblot anal>-sis of (A) total extracts, (B) nuclear fraction, and (C) cytoplasmic fraction of CEFs. Nuclear and cytoplasmic fractions were isolated from the same number of cells. The relative molecular mass of CC-3 antigen is shown in kilodaltons.
CC-3 Antigen
Is Phosphorylated
The differential reactivity of the CC-3 throughout the cell division cycle led us to investigate possible posttranslational modifications of its antigen. Until now, several antibodies that have been described with similar cyclic activity have been associated with phosphorylated epitopes (see discussion). To verify whether CC-3 binding was phosphate-dependent, nitrocellulose strips to which chick embryo brain homogenates had been transferred were probed with CC-3 after an alkaline phosphatase treatment. Phosphatase treatment completely abolished the immunoreactivity of the 255-kDa
PHOSPHOPROTEINS
149
band (Fig. 5 lane C). To determine whether the loss of reactivity was due to the phosphatase action or to proteolysis and/or protein loss, strips were incubated in the same buffer devoid of enzyme (Fig. 5 lane B) or in the presence of enzyme with the addition of 50 mM sodium phosphate (Fig. 5 lane D). In both cases, the immunoreactivity was fully preserved. Ponceau S staining profiles of the nitrocellulose strips before and after phosphatase treatment showed that this treatment did not affect the adsorption of proteins on the nitrocellulose membrane. Since the CC-3a was detected by immunofluorescence in a variety of species, the same alkaline phosphatase treatment was carried out on electroblotted whole extracts of cells from different organisms. The phosphatase treatment of these nitrocellulose strips completely abolished the capacity of CC-3 to recognize its antigen in chick fibroblasts, Drosophila KcIII-10, marsupial PtK,, human fibroblasts, human amniocytes, and murine NIH 3T3 (Fig. 6 lanes B, D, F, H, J, and L, respectively). This immunoblot also demonstrated that CC3a was associated with a molecule of the same size (255 kDa) in all of the species tested (Fig. 6, lanes A, E, G, I, and K), except for Drosophila (230 kDa) (Fig. 6 lane C). To confirm the phosphate dependence of CC-3 reactivity in immunofluorescence, alkaline phosphatase treatments were carried out on cultured cells. PtK, cells were incubated with the enzyme and then reacted with CC-3. Phosphatase treatment almost completely removed the nuclear immunofluorescence signal from the treated cells (compare Fig. 7B and Fig. 7A). However, even with more extensive digestion (Fig. 7C) some clustered fluorescence was still detected. Cells incubated in phosphatase buffer without the enzyme immunoreacted
FIG. 4. Topological distribution of the CC-3 antigen during M phase in PtK, cells. Immunofluorescence prophase, (B) at metaphase, (C) late anaphase, (D) during anaphase-telophase transition, and (E-H) at various micrograph in E has been overexposed to allow visualization of the cytoplasmic speckles. Interphase nuclei centrosomes: ch, chromosomes; mb, midbody; ms, mitotic spindle; no, nucleolus. 550~.
is observed in cells (A) at stages of cytokinesis. The (arrows) are indicated. ce,
THIBODEAU
150
FIG. 5. Phosphorylation of the CC-3 antigen. Nitrocellulose strips to which electrophoresed tot,al hrain proteins extracted from a 11-day chick embryo had been transferred were reacted with CC-3 after being treated as follows: A, no treatment; B, incubated in alkaline phosphatase buffer without enzyme; C, incubated with alkaline phosphatase (15 U/ml); D, incubated as in C in the presence of50 mM NaPO,. The dephosphorylation of the CC-3 antigen completely abolished its immunoreactivity.
at the same intensity shown). CC-3 Reveals
Mitosis-Specific
as untreated
cells (data not
AND
VINCENT
interphase or M phase. In interphase cells, the staining was restricted to distinct intranuclear clusters. According to the immunoblot experiments, this activity was associated with a 255-kDa phosphoprotein which was entirely recovered in the nuclear fraction of cells. During mitosis, the immunofluorescence first increased in the prophase nucleus and dispersed in the cytoplasm upon nuclear membrane breakdown. The staining was present in the mitotic structures throughout cell division. When analyzed on immunoblots, the M phasespecific activity appeared as a complex set of phosphoproteins ranging from 50 to >300 kDa. The mechanisms by which the immunoreactivity appears in the cytoplasm and becomes associated to mitotic structures remain to be elucidated. Although it seems to coincide with the breakdown of the nuclear envelope, it would be premature at present to associate this distribution with a migration of antigens from the nuclear region to the surrounding cytoplasm. However, a relationship between the interphase nuclear 255.kDa phosphoprot.ein and the M phase-specific phosphoproteins cannot be ruled out. Different mechanisms could account for the sudden cell division-related transition in the immunoreactivity pattern. The M phase-related antigens could be present in the interphase cytoplasm but masked by other cellular components that would be removed (modified or degraded) at the onset of mitosis. Such a covering of antigens during a phase of the cell cycle has been observed [8, 371. Another possibility would be that the M phase-specific antigens bearing phosphate-dependent epitope structures exist in an unphosphorylated form during interphase and become
Phosphoproteins
Since CC-3 recognized a 255-kDa phosphoprotein in unsynchronized cultures consisting of mostly interphase cells, we decided to perform immunoblot analyses of the more intensively immunoreactive mitotic cells. As expected, CC-3 recognized the 255-kDa band in a whole cell extract of an unsynchronized CHO cell culture (Fig. 8A). However, extracts of CHO cells arrested at metaphase displayed a complex immunoblot profile showing reactive proteins with molecular masses ranging from 50 to >300 kDa (Fig. 8B). The same profile has been obtained with mitotic cells isolated by a selective detachment procedure in which no drug was added (not shown). On the other hand, only the 25%kDa band immunoreacted in whole cell extracts of nonmitotic cells from the same culture (Fig. BC). Binding of CC-3 to the mitosis-specific bands was also found to be phosphatedependent (not shown). DISCUSSION
Monoclonal antibody CC-3 displays a different immunoreactive pattern depending on whether the cells are in
ABCDEFGHI
J
K
L
255 -
FIG. 6. CC-3 antigen phosphorylation in cells of various species. Immunohlot analysis of (A, B) total extracts of CEFs; (C, D) drosophila-KcIII-10 cells; (E, F) potoroo-PtK, cells; (G, H) human fibroblasts; (I, J) human amniocytes, and (K, L) mouse-NIH 3T3 cells without treatment (A, C, E, G, I, and K) or after an alkaline phosphatase treatment (15 1J/ml) (B, D, F. H, J, and I,).
151
FIG. 7. Effect of dephosphorylation after alkaline phorphatase treatments :100x.
on CC-3 immunofluorescence reactivity. Immunofluorescence using (B) 50 LJ/ml or (C) 100 (J/ml. Exposure and development
phosphorylated at mitosis thus permitting CC-3 binding. A dual action of phosphatases and kinases during the cell cycle has been described for the proto-oncogene c-Abl where an increase in the activity ~34”~~’ kinase above that of phosphatases during mitosis leads to the phosphorylation of the mitotic-specific sites [38]. Kinases specific Sor M phase and associated with the mitotic apparatus have been reported [24, 25, 39, 40, for review see 41,421. A de nouo synthesis and phosphorylation of the CC-3 antigens at the very beginning of mitosis cannot be excluded either. Compared to other antibodies showing cell cycle-related immunoreactivities, it is obvious that CC-3 belongs to a small family of closely related antibody probes recognizing epitopes common to both nuclear and mitotic structures. Anti-NuMA protein antibody
A
255
B
-
c
-
240 220
-
94
I.
-68 *I))
-
42
FIG. 8. Immunoreactivity pattern of CC-3 on mitotic cell extracts. lmmunoblot analysis of total extract of CHO cells from (A) an unsynchronized culture; (B) of a mitotic fraction; and (C) of the nonmitotic fraction of the same population of cells as in (B). An aliquot equivalent to 150,000 cells was loaded in each well. CC-3 reveals a complete new set of M phase-specific phosphoproteins.
is observed in PtK, cells (A) before or times were identical for A, B, and C.
recognizes a nuclear phosphoprotein [ 1, 271, with a molecular size ranging from 220 to 280 kDa, that becomes associated with spindle poles during mitosis in a limited number of species. Among a group of antibody probes raised against microtubule-associated proteins are: a mAb recognizing a 200-kDa polypeptide that undergoes a nuclear/mitotic apparatus relocalization only in human cells [2]; a mAb detecting a 280-kDa protein present in the nucleus and in some mitotic related structures of a variety of cell types [6]; and an immune serum directed against a 220-kDa phosphoprotein which is associated with the nuclear matrix and the mitotic spindle of nonneural cells [7]. Using mitotic cell extracts as antigen, Rao and collaborators [13] produced mAbs (MPM1 and MPM-2) that recognized numerous conserved phosphoproteins ranging from 40 to >200 kDa during mitosis, but showing a weak reaction when analyzed on immunoblotted extracts derived from cells in interphase. Further studies showed that MPM-2 binds to a heterogeneous set of multiple phosphoproteins on immunoblots from a variety of interphase cell extracts [43, 441. Recently, mAb C9 was shown to detect a 250-kDa protein sequestered in the nucleus and to react with M phase-specific polypeptides. However, binding of C9 to its antigens was not phosphate-dependent [8]. A similar redistribution during the interphase to M phase transition has been also described from immunolocalization experiments of proteins associated with small nuclear ribonucleoprotein particles (snRNPs) [45, 461, SV40 Tantigen [47] as well as for the proto-oncogene products of c-fos [48] and c-myc [49]. The distinguishing feat,ures of CC-3 as a new member of the interphase to M phase transition-related antibody family are its recognition of a well-conserved 255. kDa phosphoprotein in the interphase nucleus and of a complete new set of phosphoproteins associated with the mitotic apparatus and present in the cytoplasm during mitosis. In both cases, the phosphorylation is obligatory for the binding of CC-3 to its antigens. A similar phosphodependence has been reported for MPM-2 antibody in mitotic cell extracts [13] but, unlike CC-3, MPM-2 showed no or weak reaction in interphase cell
152
THIBODEAIJ
extracts. This, however, does not imply that the epitope is the precise phosphorous accepting site since phosphorylation in the vicinity or elsewhere on the protein could cause a conformational change thus creating or making available the epitope structure. Immunoblot analysis did not reveal a 255-kDa band in mitotic extracts. The interphase nuclear antigen could be selectively dephosphorylated and even degraded or its relative electrophoretic mobility may be altered because of a change in its phosphorylation state as was shown for other proteins [50]. Another related class of antibody probes consists of mAbs and autoimmune sera that recognized molecules (PCNA) that are expressed only in proliferating/cycling cells [51-551. Among these, mAb Ki-67 [56,57] has been extensively used in clinical pathology to monitor the growth fraction in human tumor cells and is thus considered to be of prognostic importance [58-601. In parallel, CC-3 proved to be an excellent probe to localize mitotic foci and evaluate the mitotic index. In our hands, CC-3 differentially stained mitotic cells in normal embryonic tissue and regenerative newt limb bud (not shown) as well as in a variety of cultured cells including human, fixed either in alcohol or aldehyde. Experiments are now in progress to determine whether there are structural analogies between the interphase nuclear antigen and those recognized during M phase and to investigate the mechanisms by which M phase-specific forms are phosphorylated. The results of these experiments should shed light on the structural and functional implications of the CC-3 antigens during the interphase to M phase transition and in the early events of nuclear assembly.
AND
VINCENT Diaz-Nido,
8. 9. 10. 11.
13
Davis, F. M., Tsao. ‘I’. Y., Fowler, S. Ii., Proc. Natl. ,4md. S’ci. ITSA 80, 2926-2930.
14
Rao, I’. N., Zhao, J-T., (‘641 SC,;. 93, 6X-69.
15.
Bradburg, (1973)
16. 17. 18.
3.
Biol.
4.
5.
Bradbury,
E. M., 2.57
247,
(ierace,
J. R. (1983)
J. Cell
I,.,
Proc.
Natl.
J. S., Klausner, Acad.
Sci.
USA
R. D., and Sandoval, 82, 1146-1150.
Inglis.
K. (1987) and Rao,
R. K., and Ashorn,
R. .J., Matthews. 33,
Inglis,
&xp.
(‘r,Il Kes.
P. N. (1983) C. L. (1989)
J.
H. R., and Sarner,
N.
13L139.
R. J.. and Matthews,
H. R. (1974)
Nn-
XII.
and
Ottaviano, Evans.
20.
(:hou, (1990)
21.
Celis,
Hlohel,
(19801 Cell 19, %Y;-28’i. L. (19851 J. Hiol. (‘hem.
(i.
Y., and Gerace,
260,
22.
L. M. (1982)
1’(,11 29,
X-52.
Y.-H., Bischotf, ,J. H., Beach, (‘P/Z 62, 106:1&1071.
R. M., and Fink,
D., and
Goldman,
J. IX., Larsen, 14291434.
Hiol
624
(‘hou.
Y.-H.,
Kosevear,
Acad.
23. 24.
Wordeman, Dinsmore,
‘5.
Koriyama,
26.
Vandre,
“7.
I’. M.,
Fey, S. ,J., and Celis,
R. D.
A. (1983)
J. (‘cll
97,
Natl.
Sri.
I!.SA
86,
E., and Goldman. 1885 1889.
1~. D. (19891
f’roc.
I,., and Cande, W. Z. (1987) (‘c/I 50, 535 :j4:L .J. H., and Sloboda. R. D. (1988) (‘ell 53, 7699780. R. ( 1989j
(‘ytoskrlc~ton
(‘c/i Motif.
12, 90- 103.
F. M.. Kao, P. N., and Borisy, (;. G. (1984) I’rrr. Nat1 Acnd. Sci. ITSA 81, 4439-4443. I’rice, (‘. M., and Pettijohn, D. E. (1986) I&J. (‘PI/ Krs. 166, 295-3 I I
D. D., Davis,
‘I‘hibodeau, (1989)
A.,
Histochem.
29.
Vincent, 192.
30.
Echalier, G.. and 268, 1771-1774.
3I
Echalier,
32.
Vincent,
M.,
Duchaine, J.. Simard, J. 2 1, 348-X56.
and I,ahaie,
C. (1988)
Ohanessian,
G., and Ohanessian, M., and Duchaine,
J. I,., and
Hioch
CELL
RESEARCH
195,
145-153
(1991)
Monoclonal Antibody CC-3 Recognizes Phosphoproteins in Interphase and Mitotic Cells ALAINTHIBODEAUANDMICHEL OntogmPsc
ct Ghnhtiyw
Mokkulaire,
Ccntre
de Recherche
du
INTRODUCTION Understanding the mechanism and regulation of mitosis will require the identification and characterization of proteins or activities specific to different phases of the cell cycle. Recent progress in immunological techniques and in the preparation of antigenic material has permitted the production of several antibody probes recognizing diverse molecules specific to the mitotic phase (M phase) or that differentially react during the interphase to M phase transition. Using an antiserum raised against a nonhistone nuclear protein preparation, Lyrequests Centre (QuChec),
should he addressed de recherche du GlV 4G2, Canada.
at: Ontoge&se CHUL, 2705
VINCENT’ and
D4partement
de Mhdecine,
UniuersitP
Laclal
dersen and Pettijohn [ 11reported the first evidence that a predominantly interphase nuclear protein (NuMA) becomes part of the mitotic apparatus during M phase. Microtubule-associated proteins isolated from different species have also been extensively used as immunizing material giving rise to a panoply of mAbs directed against a variety of interphasejmitotic structures [2-S]. Other related mAbs have been produced by immunizing animals with isolated mitotic spindles [9, lo], chromatin-containing nuclear extracts [ 111, DNA-binding proteins from chick embryo extracts [la], and mitotic cell extracts [13, 141. In most of these studies, the recognized antigens behave like cell division-related molecules in the sense that they are present in interphase (nuclear or cytoskeletal) structures and become part of mitotic components (mitotic spindle, centrosomes, kinetochores, midbody) during mitosis. The transition from int.erphase to M phase is marked by an increase in the phosphorylation of several intracellular proteins. For instance, histone Hl becomes highly phosphorylated [ 15,161 and depolymerization of the nuclear envelope lamina is accompanied by lamin phosphorylation [17,181. Similarly, structural modifications of the cytoskeleton at the onset of mitosis correlate with alterations in the state of phosphorylation of vimentin [ 19, 201, keratin [21], and desmin [22]. Cell cycle-modulated phosphorylation has also been reported on proteins which are part of specialized mitotic structures including the mitotic spindle [23,24], centrosomes [25], microtubule organizing centers [26], and proteins that undergo intracellular distribution during the interphase to M phase transition [7, 271. The search for new molecules involved in embryonic development led us to obtain monoclonal antibodies (mAbs) that reacted in a cell cycle-dependent manner [28]. In the present study, we show t.hat one of them, CC-3 (previously designated Z4E9 [as]), recognizes a phylogenetically well-conserved 255-kDa phosphoprotein that is sequestered within the nucleus during interphase. At mitosis, CC-3 reveals a complete new set of phosphoproteins that are associated with the mitotic apparatus. In both cases, the reactivity of CC-3 depends on the phosphorylation of the antigens since it is abolished by the removal of phosphate groups.
Among a library of monoclonal antibodies (mAbs) recognizing developmental markers in the chick embryo, mAb CC-3 was selected because of its differential immunostaining of mitotic cells. The intracellular distribution of the CC-3 antigen (CC-3a) throughout the cell cycle was visualized by immunolocalization. In interphase cells CC-3a resided in the nucleus and was arranged in distinct extranucleolar clusters. At prophase, the nuclear reactivity of CC-3a considerably increased and subsequently extended to the cytoplasm at metaphase. From metaphase through anaphase, most of the reactivity was associated with the mitotic apparatus. During cytokinesis CC-3a was detected in the midbody and also in discrete speckles dispersed throughout the cytoplasm. The initial interphase pattern was then restored in the two daughter nuclei. Immunoblot analysis demonstrated that a 255-kDa phosphoprotein was present only in the interphase nucleus and that a complete new set of phosphoproteins accounted for the mitotic cell reactivity. The binding of CC-3 was dependent on the phosphorylation of its antigens. CC-3a is an evolutionary conserved molecule; it is present in such phylogenetically distant species as Drosophila and humans. Furthermore, the unique behavior of CC-3 on sections of normal, embryonic, and regenerative tissue and in cell culture immunostaining make it a reliable tool to identify mitotic foci. ( 1991 Academic Press. Inc.
’ 7’0 whom reprint Gknktique Molkculaire, Laurier. Sainte-Fov
CHIJL
et houl
145
0014.4827/91
$3.00
146
THIBODEAIJ
MATERIALS
AND
METHODS
Fertilized White leghorn eggs were obtained from a local hatchery and maintained in a humidified air-forced incubator at 38°C. The embryos were staged according to the number of hours or days of incubation as indicated. Antibodirs. MAb CC-3 (IgG) was obtained after immunization of a Balb/C mouse with isolated pharyngeal regions from 72-h chick embryos (281. The CC-3 hybridoma was grown in Iscove’s modified Dulbecco’s medium supplemented with 10% fetal calf serum and the culture supernatant was used for immunofluorescence. CC-3 ascites fluid (1:lOOO dilution) was used for immunoblots. CC-3 was revealed using either a fluorescein isothiocyanate (FITC)-conjugated goat IgG anti-(mouse immunoglobulins) (FITCGAM) in immunofluorescence or a iz51-labeled GAM (““I-GAM) (10 pCi/pg) in immunoblot studies. Immunofluorescent
staining
of
cell
cultures
and
tissue
swtion.s.
Primary fibroblast cell cultures from ‘i-day-old chick embryos (CEF) were prepared and maintained as described [29]. The Drosophila cell line (KcIII-10) [30] was grown in D-22 supplemented with 10% decomplemented fetal calf serum (FCS) at 23°C [31]. Chinese hamster ovary (CHO) cells were grown in Iscove’s modified Dulbecco’s medium (IMDM) supplemented with 10% FCS. Mouse cells (NIH 3T3), potoroo kidney cells (PtK,), human fibroblasts (primary culture), and human amniocytes (primary culture) were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% FCS in a humidified 5% CO, atmosphere at 37°C. For immunost,aining, cells were washed with PBS (8 mMNa,HPO, * 7 H,O, 1.5 mM KH,PO,, 140 mM NaCI, :1 mM KCl, pH ‘i.4), fixed in methanol for 20 min at -20°C and finally soaked in PBS. Paraffin embedding and tissue sectioning were carried out as previously described [32]. Fixed cells and rehydrated tissue sections were incubated with CC-3 for 45 min at room temperature, washed in PBS, and incubated with FITCGAM in the same conditions. Immunofluorescence observation was performed on a Zeiss Axiophot microscope equipped with epifluorescence optics. Cell .s~~nchronization. CHO cell cultures were allowed to reach an exponentially growing phase in 75cm2 culture flasks (Costar). The monolayer was then gently squirted with the overlying medium to get rid of loosely attached cells and the medium was replaced with 1.5 ml of complete medium (IMDM/lO% FCS) containingcolcemid (0.1 fig/ ml). After a 12-h incubation, the nonadherent cells were removed by selective mechanical detachment [33], pelleted, resuspended in IMDM containing colcemid, and kept on ice. This first collection was referred to as the mitotic fraction. The remaining adherent cell layer was lightly trypsinized (0.0125% trypsin, 90 s) to remove the persistent adherent mitotic cells and this fraction was discarded. The cells still attached to the bottom of the Hask were collected by complete trypsinization (0.05% trypsin, 5 min) of the layer, washed with complete medium, resuspended in IMDM, and kept on ice. This second collection was referred to as the nonmitotic fraction. The cells of both fractions were finally pelleted and resuspended in electrophoresis sample buffer containing 2 mM sodium orthovanadate as a phosphatase inhibitor [34]. To estimate the mitotic index, an aliquot of cells was swollen in 0.8% sodium citrate buffer (5-20 min) at room temperature, dried on a microscope slide, and fixed in methanohacetic acid (3:l). The slides were stained with propidium iodide (0.17 eg/ml; 30 s) and mitotic cells were scored as those showing individual condensed chromosomes in fluorescence microscopy. The mitotic index of the mitotic fraction was 85% while the nonmitotic fraction was over 98% pure. Preparation of cell and tissue extracts. Whole cell and chick embryo brain extracts were obtained by directly homogenizing the cells or the tissue in electrophoresis sample buffer. To obtain CEF nuclear and cytoplasmic fractions, cells were washed with PBS, resuspended in HMK buffer (15 mM Hepes, 5 mM MgCl,, 2 mM KCl, 1 mM PMSF, 1 pg/ml aprotinin, pH 7.2), and allowed to swell 10 min on ice. One volume of HMK buffer containing 0.5 A4 sucrose was then added
AND
VINCENT
and the cell membranes were disrupted using a Dounce homogenizer. The percentage of disrupted cells (>99% after 100 strokes) was rou tinely monitored by phase contrast microscopy. The cytoplasmic and nuclear fractions were separated by centrifugation at 1OOOg. The former was concentrated by ethanol precipitation and recovered by sedimentation at 15OOg. The nuclear fraction (1OOOg pellet) and the cvtoplasmic fraction (15OOg pellet) were dissolved in the electrophoresis sample buffer. E:lpctrophoresisandimmunoblottin~. Polypeptidesinelectrophoresis sample buffer (63 mM TrissHCl, 10% glycerol, 2.3% SDS 5% rj-mercaptoethanol, 0.005’;> bromophenol blue, pH 6.8) were heated at 95°C for 5 min and separated on 7.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Protein standards used in gel migration were: actin (42 kDa) bovine serum albumine (68 kDa), phosphorylase o (94 kDa). spectrin /j (220 kDa), and spectrin (Y (240 kDa). Protein separation was verified by Coomassie blue staining. For immunoblot analysis, proteins were electrophoretically transferred to nitrocellulose membranes (Gelman Sciences, Ann Arbor) by the method of Towbin et al. [35]. The transfer efficiency was confirmed by Ponceau S staining. Prior to antibody contact, the ni trocellulose membranes were saturated with 5% blotto (PBS contain ing 5% skim milk powder) for 1 h at 37°C. The membranes were then reacted with U-3 (45 min at 37”C), washed in PBS containing 0.05!;, Tween-20 with several solution changes, incubated with ““I-GAM in the same conditions, and finally washed as above. Immunoblots were exposed to Kodak XAR-5 film. AIkalinc~ phosphatasr twatmfvzt. Nitrocellulose strips to which electrophoretically separated polypeptides had been transferred were rinsed in Tris/NaCl (50 mM Tris. 150 mM NaCl. pH 7.5), blocked with 5% blotto in Tris/NaCl, and rinsed again in Tris/NaCl. The strips were then deposited in Tris/MgCl, (100 mMTris, 1 mM MgCl,, pH 8.0) containing a cocktail of protease inhibitors (5 pg/ml antipain, 5 &ml leupeptin, 5 pg/ml pepstatin, 1 pg/ml aprotinin, and 1 mM PMSF) and alkaline phosphatase (molecular biology grade, BoehringerMannheim) was added at 15 units/ml and the strips were incubated 30 min at :37”(‘ followed by 90 min at 25°C. After digestion, the strips were subsequently washed with PBS, PBS containing 0.05”C Tween-20 and then resaturated with PBS containing 5% blotto. To test the effects of phosphatase treatment on immunofiuorescence, PtK, cells were briefly rinsed with Tris/NaCl, fixed in methanol, washed again with Tris/NaCI, and soaked in Tris/MgCI, containing the protease inhibitor cocktail. Alkaline phosphatase was then added at 5OLlOO units/ml and the cells were incubated 69 min at 37°C followed by 60 min at 25°C. After digestion the cells were washed with PBS. The nitrocellulose strips and the cells were finally reacted with the primary and secondary antibodies. Control experiments were carried out by incubating nitrocellulose strips or fixed cells in the same digestion conditions without adding enzyme or in the presence of enzyme with 50 mM sodium phosphate (pH 8.0) as an alkaline phosphatase inhibitor [as].
RESULTS Reactivity Sections
Pattern of CC-3 on Chick and Cell Cultures
Embryo
Tissue
Our hybridoma supernatant screening procedure [28] allowed direct visualization of the pattern of CC-3 reactivity on 72-h chick embryo tissue sections. As shown in Fig. lA, CC-3 reacted with scattered cells demonstrating no obvious tissue, cell lineage, or region specificity, except for the internal periphery of the neural tube where the labeled cells seemed to accumulate. A higher magnification of the neural tube (Fig. 1B) shows that several positive cells are aligned in the germinal zone
CELL
DIVISION-RELATED
it was gathered (Fig. 2G). Intracellular
FIG. 1. Histological staining of chick embryo tissue sections with (‘C-3. Immunofluorescence is observed (A, B) on transverse sections at the pharyngeal level of a 72-h embryo, (C) at the mesencephalic level of a 40-h embryo and (D) on a longitudinal section of the limb bud of a 96 h-stage embryo. CC-3 differentially identify mitotic cells. nt. neural tube. (AI 35~. (B) 70x, (C) 1:30x, (D) 45~.
near the ventricular surface where a high mitotic index is present. The immunoreactivity was not restricted to this particular stage of development. For example, CC-3 also stained individual cells in the neural tube on a 40-h embryo and in the wing bud of a 96-h embryo (Figs. 1C and D, respectively). In order to identify the cellular and subcellular structures targeted by mAb CC-3, the immunofluorescence staining pattern of CEF was investigated. As was the case with histological sections, CC-3 brightly stained a limited number of cells (Fig. 2A), whereas the others exhibited a weaker nuclear fluorescence. A closer examination of the strongly reacting cells demonstrated that these cells were undoubtedly undergoing mitosis. Indeed, metaphase cells displayed a strong staining of mitotic structures, namely, the spindle and the centrosomes (Fig. 2C). A diffuse staining was also observed throughout the cytoplasm. The chromosomes aligned at the metaphase plate were not immunoreactive. In agreement with this result, CC-3 staining of chromosomal spreads, in which microtubular structures have been depolymerized, showed that the immunoreaction was only present in the cytoplasmic portion and in a sheath wrapping each chromosome but that the chromatin, as such, was not labeled (data not shown). After the duplicated chromosomes segregated reactivity was found in the centrosomal areas (depolymerized kinetochore microtubules) and on the polar microtubules (Fig. 2E). Furthermore, bright speckles can be observed over the cytoplasmic staining at this stage. The reactivity at interphase was entirely restricted to the nucleus where
147
PHOSPHOPROTEINS
into
distinct
Compartmentalization
extranucleolar
clusters
of the CC-3 Antigen
Since the immunofluorescence was strictly confined to the nucleus in fixed interphase cells, we tried to confirm this by cell fractionation. Total extract and nuclear and cytoplasmic fractions prepared from a primary culture of CEF were immunoblotted with CC-3. Figure 3 shows that the CC-3 antigen (CC-3a) was present on a molecule of 255 kDa whose immunoreactivity was entirely recovered in the nuclear fraction. In these unsynchronized cultures, in which mitotic cells represented not more than 3% of the total cells (assessed by CC-3 and DNA staining), it can be assumed that the 255-kDa band accounted for the nuclear immunoreactivity of the interphase cells. Furthermore, the inability to immunodetect a-tubulin in the nuclear fraction (data not shown) suggests that no or at most trace residual cytoskeleton elements contaminated this fraction. Other minor CC-3-immunoreactive bands were occasionally observed but their relative intensities were low and fluctuating compared with that of the 255-kDa band (see for instance, Figs. 5 and 6). Intracellular M Phase
Distribution
of the CC-3 Antigens
during
Since the immunofluorescence pattern of CC-3 (nuclear clusters, mitotic apparatus) was very well conserved in all species examined (not shown, see also Fig. 6), immunofluorescence studies were performed on PtK, cells in order to visualize the stepwise transition in the pattern of reactivity from interphase to M phase. These cells were found to be particularly well suited for this purpose since, unlike other cell types, they do not round up when entering mitosis. During mitotis, a change in CC-3a reactivity was first seen at prophase, when the nucleus became heavily stained (compared to the interphase nucleus) and the positioned centrosomes were recognized (Fig. 4A). In metaphase cells, the immunoreactivity was diffusely distributed throughout the cytoplasm, being preferentially associated with the mitotic apparatus (spindle and centrosomes) (Fig. 4B). During the separation of the duplicated chromatids, as the kinetochore microtubules depolymerized, the cytoplasmic reactivity faded but a signal was preserved at the mitotic poles and in the polar microtubule area. Some fluorescence was also visible at the periphery of the chromosomes (Fig. 4C). At early telophase, when the cleavage furrow was initiated, the centrosomes were still reactive while the median region of the midbody corresponding to the overlapping of the polar microtubules fluoresced more strongly than the rest of the cytoplasm (Fig. 4D). This region of the midbody was strongly labeled throughout cytokinesis and the nuclear fluorescence slowly lighted up as the chromatin decon-
148
FIG. 2. Cytological staining of CEFs with CC-S. Immunofluorescence is ohserved in (A) a random cell population, (C) in a cell at metaphase, (E) during anaphase-telophase transition, and (G) at interphase. The micrographs on the right, (B, D, F, and H) represent the corresponding phase contrast image. The micrograph in C has been underexposed to emphasize the fluorescence of the mitotic apparatus over that of the cytoplasm. Interphase nuclei (arrows) and mitotic cells (arrowheads) are indicated. CC-S stains the interphasic nucleus and the mitotic apparatus. ce, centrosomes; ch, chromosomes; ms, mitotic spindle, no, nucleolus; nu, nucleus. (A, B) 160X, (C-H) 500X.
densed (Figs. 4E-4G). In the earliest stage of cytokinesis, discrete cytoplasmic speckles were also detected (Fig. 4E). As the daughter cells reachedcomplete separa-
tion, the labeling of the midbody progressively diminished and the interphase nuclear reactivity pattern was completely restored (Fig. 4H).
CELL
A
255
B
DIVISION-RELATED
c
-
FIG. 3. Intracellular compartmentalization of the CC-3 antigen. Immunoblot anal>-sis of (A) total extracts, (B) nuclear fraction, and (C) cytoplasmic fraction of CEFs. Nuclear and cytoplasmic fractions were isolated from the same number of cells. The relative molecular mass of CC-3 antigen is shown in kilodaltons.
CC-3 Antigen
Is Phosphorylated
The differential reactivity of the CC-3 throughout the cell division cycle led us to investigate possible posttranslational modifications of its antigen. Until now, several antibodies that have been described with similar cyclic activity have been associated with phosphorylated epitopes (see discussion). To verify whether CC-3 binding was phosphate-dependent, nitrocellulose strips to which chick embryo brain homogenates had been transferred were probed with CC-3 after an alkaline phosphatase treatment. Phosphatase treatment completely abolished the immunoreactivity of the 255-kDa
PHOSPHOPROTEINS
149
band (Fig. 5 lane C). To determine whether the loss of reactivity was due to the phosphatase action or to proteolysis and/or protein loss, strips were incubated in the same buffer devoid of enzyme (Fig. 5 lane B) or in the presence of enzyme with the addition of 50 mM sodium phosphate (Fig. 5 lane D). In both cases, the immunoreactivity was fully preserved. Ponceau S staining profiles of the nitrocellulose strips before and after phosphatase treatment showed that this treatment did not affect the adsorption of proteins on the nitrocellulose membrane. Since the CC-3a was detected by immunofluorescence in a variety of species, the same alkaline phosphatase treatment was carried out on electroblotted whole extracts of cells from different organisms. The phosphatase treatment of these nitrocellulose strips completely abolished the capacity of CC-3 to recognize its antigen in chick fibroblasts, Drosophila KcIII-10, marsupial PtK,, human fibroblasts, human amniocytes, and murine NIH 3T3 (Fig. 6 lanes B, D, F, H, J, and L, respectively). This immunoblot also demonstrated that CC3a was associated with a molecule of the same size (255 kDa) in all of the species tested (Fig. 6, lanes A, E, G, I, and K), except for Drosophila (230 kDa) (Fig. 6 lane C). To confirm the phosphate dependence of CC-3 reactivity in immunofluorescence, alkaline phosphatase treatments were carried out on cultured cells. PtK, cells were incubated with the enzyme and then reacted with CC-3. Phosphatase treatment almost completely removed the nuclear immunofluorescence signal from the treated cells (compare Fig. 7B and Fig. 7A). However, even with more extensive digestion (Fig. 7C) some clustered fluorescence was still detected. Cells incubated in phosphatase buffer without the enzyme immunoreacted
FIG. 4. Topological distribution of the CC-3 antigen during M phase in PtK, cells. Immunofluorescence prophase, (B) at metaphase, (C) late anaphase, (D) during anaphase-telophase transition, and (E-H) at various micrograph in E has been overexposed to allow visualization of the cytoplasmic speckles. Interphase nuclei centrosomes: ch, chromosomes; mb, midbody; ms, mitotic spindle; no, nucleolus. 550~.
is observed in cells (A) at stages of cytokinesis. The (arrows) are indicated. ce,
THIBODEAU
150
FIG. 5. Phosphorylation of the CC-3 antigen. Nitrocellulose strips to which electrophoresed tot,al hrain proteins extracted from a 11-day chick embryo had been transferred were reacted with CC-3 after being treated as follows: A, no treatment; B, incubated in alkaline phosphatase buffer without enzyme; C, incubated with alkaline phosphatase (15 U/ml); D, incubated as in C in the presence of50 mM NaPO,. The dephosphorylation of the CC-3 antigen completely abolished its immunoreactivity.
at the same intensity shown). CC-3 Reveals
Mitosis-Specific
as untreated
cells (data not
AND
VINCENT
interphase or M phase. In interphase cells, the staining was restricted to distinct intranuclear clusters. According to the immunoblot experiments, this activity was associated with a 255-kDa phosphoprotein which was entirely recovered in the nuclear fraction of cells. During mitosis, the immunofluorescence first increased in the prophase nucleus and dispersed in the cytoplasm upon nuclear membrane breakdown. The staining was present in the mitotic structures throughout cell division. When analyzed on immunoblots, the M phasespecific activity appeared as a complex set of phosphoproteins ranging from 50 to >300 kDa. The mechanisms by which the immunoreactivity appears in the cytoplasm and becomes associated to mitotic structures remain to be elucidated. Although it seems to coincide with the breakdown of the nuclear envelope, it would be premature at present to associate this distribution with a migration of antigens from the nuclear region to the surrounding cytoplasm. However, a relationship between the interphase nuclear 255.kDa phosphoprot.ein and the M phase-specific phosphoproteins cannot be ruled out. Different mechanisms could account for the sudden cell division-related transition in the immunoreactivity pattern. The M phase-related antigens could be present in the interphase cytoplasm but masked by other cellular components that would be removed (modified or degraded) at the onset of mitosis. Such a covering of antigens during a phase of the cell cycle has been observed [8, 371. Another possibility would be that the M phase-specific antigens bearing phosphate-dependent epitope structures exist in an unphosphorylated form during interphase and become
Phosphoproteins
Since CC-3 recognized a 255-kDa phosphoprotein in unsynchronized cultures consisting of mostly interphase cells, we decided to perform immunoblot analyses of the more intensively immunoreactive mitotic cells. As expected, CC-3 recognized the 255-kDa band in a whole cell extract of an unsynchronized CHO cell culture (Fig. 8A). However, extracts of CHO cells arrested at metaphase displayed a complex immunoblot profile showing reactive proteins with molecular masses ranging from 50 to >300 kDa (Fig. 8B). The same profile has been obtained with mitotic cells isolated by a selective detachment procedure in which no drug was added (not shown). On the other hand, only the 25%kDa band immunoreacted in whole cell extracts of nonmitotic cells from the same culture (Fig. BC). Binding of CC-3 to the mitosis-specific bands was also found to be phosphatedependent (not shown). DISCUSSION
Monoclonal antibody CC-3 displays a different immunoreactive pattern depending on whether the cells are in
ABCDEFGHI
J
K
L
255 -
FIG. 6. CC-3 antigen phosphorylation in cells of various species. Immunohlot analysis of (A, B) total extracts of CEFs; (C, D) drosophila-KcIII-10 cells; (E, F) potoroo-PtK, cells; (G, H) human fibroblasts; (I, J) human amniocytes, and (K, L) mouse-NIH 3T3 cells without treatment (A, C, E, G, I, and K) or after an alkaline phosphatase treatment (15 1J/ml) (B, D, F. H, J, and I,).
151
FIG. 7. Effect of dephosphorylation after alkaline phorphatase treatments :100x.
on CC-3 immunofluorescence reactivity. Immunofluorescence using (B) 50 LJ/ml or (C) 100 (J/ml. Exposure and development
phosphorylated at mitosis thus permitting CC-3 binding. A dual action of phosphatases and kinases during the cell cycle has been described for the proto-oncogene c-Abl where an increase in the activity ~34”~~’ kinase above that of phosphatases during mitosis leads to the phosphorylation of the mitotic-specific sites [38]. Kinases specific Sor M phase and associated with the mitotic apparatus have been reported [24, 25, 39, 40, for review see 41,421. A de nouo synthesis and phosphorylation of the CC-3 antigens at the very beginning of mitosis cannot be excluded either. Compared to other antibodies showing cell cycle-related immunoreactivities, it is obvious that CC-3 belongs to a small family of closely related antibody probes recognizing epitopes common to both nuclear and mitotic structures. Anti-NuMA protein antibody
A
255
B
-
c
-
240 220
-
94
I.
-68 *I))
-
42
FIG. 8. Immunoreactivity pattern of CC-3 on mitotic cell extracts. lmmunoblot analysis of total extract of CHO cells from (A) an unsynchronized culture; (B) of a mitotic fraction; and (C) of the nonmitotic fraction of the same population of cells as in (B). An aliquot equivalent to 150,000 cells was loaded in each well. CC-3 reveals a complete new set of M phase-specific phosphoproteins.
is observed in PtK, cells (A) before or times were identical for A, B, and C.
recognizes a nuclear phosphoprotein [ 1, 271, with a molecular size ranging from 220 to 280 kDa, that becomes associated with spindle poles during mitosis in a limited number of species. Among a group of antibody probes raised against microtubule-associated proteins are: a mAb recognizing a 200-kDa polypeptide that undergoes a nuclear/mitotic apparatus relocalization only in human cells [2]; a mAb detecting a 280-kDa protein present in the nucleus and in some mitotic related structures of a variety of cell types [6]; and an immune serum directed against a 220-kDa phosphoprotein which is associated with the nuclear matrix and the mitotic spindle of nonneural cells [7]. Using mitotic cell extracts as antigen, Rao and collaborators [13] produced mAbs (MPM1 and MPM-2) that recognized numerous conserved phosphoproteins ranging from 40 to >200 kDa during mitosis, but showing a weak reaction when analyzed on immunoblotted extracts derived from cells in interphase. Further studies showed that MPM-2 binds to a heterogeneous set of multiple phosphoproteins on immunoblots from a variety of interphase cell extracts [43, 441. Recently, mAb C9 was shown to detect a 250-kDa protein sequestered in the nucleus and to react with M phase-specific polypeptides. However, binding of C9 to its antigens was not phosphate-dependent [8]. A similar redistribution during the interphase to M phase transition has been also described from immunolocalization experiments of proteins associated with small nuclear ribonucleoprotein particles (snRNPs) [45, 461, SV40 Tantigen [47] as well as for the proto-oncogene products of c-fos [48] and c-myc [49]. The distinguishing feat,ures of CC-3 as a new member of the interphase to M phase transition-related antibody family are its recognition of a well-conserved 255. kDa phosphoprotein in the interphase nucleus and of a complete new set of phosphoproteins associated with the mitotic apparatus and present in the cytoplasm during mitosis. In both cases, the phosphorylation is obligatory for the binding of CC-3 to its antigens. A similar phosphodependence has been reported for MPM-2 antibody in mitotic cell extracts [13] but, unlike CC-3, MPM-2 showed no or weak reaction in interphase cell
152
THIBODEAIJ
extracts. This, however, does not imply that the epitope is the precise phosphorous accepting site since phosphorylation in the vicinity or elsewhere on the protein could cause a conformational change thus creating or making available the epitope structure. Immunoblot analysis did not reveal a 255-kDa band in mitotic extracts. The interphase nuclear antigen could be selectively dephosphorylated and even degraded or its relative electrophoretic mobility may be altered because of a change in its phosphorylation state as was shown for other proteins [50]. Another related class of antibody probes consists of mAbs and autoimmune sera that recognized molecules (PCNA) that are expressed only in proliferating/cycling cells [51-551. Among these, mAb Ki-67 [56,57] has been extensively used in clinical pathology to monitor the growth fraction in human tumor cells and is thus considered to be of prognostic importance [58-601. In parallel, CC-3 proved to be an excellent probe to localize mitotic foci and evaluate the mitotic index. In our hands, CC-3 differentially stained mitotic cells in normal embryonic tissue and regenerative newt limb bud (not shown) as well as in a variety of cultured cells including human, fixed either in alcohol or aldehyde. Experiments are now in progress to determine whether there are structural analogies between the interphase nuclear antigen and those recognized during M phase and to investigate the mechanisms by which M phase-specific forms are phosphorylated. The results of these experiments should shed light on the structural and functional implications of the CC-3 antigens during the interphase to M phase transition and in the early events of nuclear assembly.
AND
VINCENT Diaz-Nido,
8. 9. 10. 11.
13
Davis, F. M., Tsao. ‘I’. Y., Fowler, S. Ii., Proc. Natl. ,4md. S’ci. ITSA 80, 2926-2930.
14
Rao, I’. N., Zhao, J-T., (‘641 SC,;. 93, 6X-69.
15.
Bradburg, (1973)
16. 17. 18.
3.
Biol.
4.
5.
Bradbury,
E. M., 2.57
247,
(ierace,
J. R. (1983)
J. Cell
I,.,
Proc.
Natl.
J. S., Klausner, Acad.
Sci.
USA
R. D., and Sandoval, 82, 1146-1150.
Inglis.
K. (1987) and Rao,
R. K., and Ashorn,
R. .J., Matthews. 33,
Inglis,
&xp.
(‘r,Il Kes.
P. N. (1983) C. L. (1989)
J.
H. R., and Sarner,
N.
13L139.
R. J.. and Matthews,
H. R. (1974)
Nn-
XII.
and
Ottaviano, Evans.
20.
(:hou, (1990)
21.
Celis,
Hlohel,
(19801 Cell 19, %Y;-28’i. L. (19851 J. Hiol. (‘hem.
(i.
Y., and Gerace,
260,
22.
L. M. (1982)
1’(,11 29,
X-52.
Y.-H., Bischotf, ,J. H., Beach, (‘P/Z 62, 106:1&1071.
R. M., and Fink,
D., and
Goldman,
J. IX., Larsen, 14291434.
Hiol
624
(‘hou.
Y.-H.,
Kosevear,
Acad.
23. 24.
Wordeman, Dinsmore,
‘5.
Koriyama,
26.
Vandre,
“7.
I’. M.,
Fey, S. ,J., and Celis,
R. D.
A. (1983)
J. (‘cll
97,
Natl.
Sri.
I!.SA
86,
E., and Goldman. 1885 1889.
1~. D. (19891
f’roc.
I,., and Cande, W. Z. (1987) (‘c/I 50, 535 :j4:L .J. H., and Sloboda. R. D. (1988) (‘ell 53, 7699780. R. ( 1989j
(‘ytoskrlc~ton
(‘c/i Motif.
12, 90- 103.
F. M.. Kao, P. N., and Borisy, (;. G. (1984) I’rrr. Nat1 Acnd. Sci. ITSA 81, 4439-4443. I’rice, (‘. M., and Pettijohn, D. E. (1986) I&J. (‘PI/ Krs. 166, 295-3 I I
D. D., Davis,
‘I‘hibodeau, (1989)
A.,
Histochem.
29.
Vincent, 192.
30.
Echalier, G.. and 268, 1771-1774.
3I
Echalier,
32.
Vincent,
M.,
Duchaine, J.. Simard, J. 2 1, 348-X56.
and I,ahaie,
C. (1988)
Ohanessian,
G., and Ohanessian, M., and Duchaine,
J. I,., and
Hioch
