Autoantibodies to a novel cell cycle-regulated protein that accumulates in the nuclear matrix during S phase and is localized in the kinetochores and spindle midzone during mitosis

The centromere region of chromosomes appears to be the central orchestrator of many events that occur during mito-sis. Its primary function is to direct the movement of chro-mosomes through the activity of microtubule-based motor enzymes located in or near its kinetochore domain (Gorb-sky et al., 1987; Nicklas, 1989; McIntosh and Pfarr, 1991; Sawin and Scholey, 1991). Compelling evidence for this can be found in the recent detection in the kinetochore of microtubule-based motor activities with opposite polarities, plus-end- and minus-end-directed (Hyman and Mitchison, 1991); the immunolocalization within this region of mem-bers of the dynein and kinesin families of motor proteins (Pfarr et al., 1990; Steuer et al., 1990); and the isolation of a yeast centromere-associated protein complex with motor activities (Hyman et al., 1992). A second possible function of the centromere is to act as a transport structure for the correct positioning at the metaphase plate of specific pro-teins involved in late mitosis. This notion is supported by the recent identification of a new class of cell cycle-regu-lated proteins, termed ‘chromosomal passengers’, which are concentrated in the centromeres during early mitosis and are then relocated in the spindle midzone, the spindle poles or the midbody after the metaphase/anaphase transition (Earnshaw and Bernat, 1991). Although the functions of these proteins are unknown, their subcellular localizations are consistent with roles in chromosome movement and cytokinesis.

Essential for a detailed understanding of the structure and function of the centromere is the identification of its pro-tein components. Some of the information available on cen-tromere-associated proteins is derived from studies involv-ing anti-centromere autoantibodies from patients with the CREST subset of the systemic autoimmune disease sclero-derma (Moroi et al., 1981; Brenner et al., 1981). The avail-ability of these autoantibodies has been instrumental in the identification of the centromere-associated autoantigens CENP-A, B, C and D (Earnshaw and Rothfield, 1985; King-well and Rattner, 1987), and the chromatid-linking proteins (CLiPs) (Rattner et al., 1988). Other centromere-associated antigens, such as the inner centromere proteins (INCENPs) (Cooke et al., 1987) and the 312 kDa kinesin-like protein CENP-E (Yen et al., 1991, 1992), have been identified with murine monoclonal antibodies raised against chromosomal scaffolds. Centromere-associated proteins can be classified into two major categories: those that are present in cen-tromeres throughout the cell cycle, such as CENP A, B, C and D; and those that bind in a cell cycle-dependent fash-ion, like the INCENP and CLiP proteins, and CENP-E (reviewed by Brinkley et al., 1992).

In the present study, using human autoantibodies as probes, we report the identification and initial characteri-zation of a novel cell cycle-regulated nuclear protein, pro-visionally designated p330^d, which accumulates in the nuclear matrix during S phase and is redistributed to kine-tochores and spindle midzone during mitosis. We propose that p330^d is a novel member of the growing class of chro-mosomal passenger proteins.

HeLa, U-87 MG, WI-38, HEp-2, PtK₂, NRK, 3T3 and Indian muntjac cells (American Type Culture Collection) were grown in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and 10 μg/ml gentamicin. MOLT-4 (human T cell leukemia) cells were grown in RPMI 1640 medium supplemented with 10% FBS, 2 mM L-glutamine, 50 μg/ml gentamicin solution (Sigma, St Louis, MO), 10 mM sodium pyruvate (Irvine Scientific, Santa Ana, CA), and non-essential amino acids (Irvine Scientific). Cell cultures were maintained in a humidified 37°C incubator with a 5% CO₂ atmos-phere. Peripheral blood lymphocytes from a healthy blood donor were prepared by Ficoll-Hypaque (Histopaque 1077) (Sigma) frac-tionation and stimulated to proliferate with 10 μg/ml phyto-hemagglutinin (PHA) (Sigma) for 72 hours.

Human autoimmune sera were from the collection of the W. M. Keck Autoimmune Disease Center Laboratory serum bank. Human autoimmune reference serum to NuMA (Ta) was kindly supplied by Dr David Pettijohn (University of Colorado, Denver) and was used to identify our own anti-NuMA serum (BL), which was subsequently used for detection of NuMA. The mouse mon-oclonal antibody 1D12 to DNA has been characterized previously (Kotzin et al., 1984) and was a kind gift from Dr Robert Rubin (The Scripps Research Institute, La Jolla). The rabbit anti-PCNA antibody 734 was a kind gift from Dr Yoshinao Muro (The Scripps Research Institute, La Jolla), and its characterization will be described elsewhere. For indirect immunofluorescence (IIF) studies, cells grown on coverslips in this laboratory and com-mercially available HEp-2 slides (Bion, Park Ridge, IL or Kallestad, Chaska, MN) were employed. Cells grown on cover-slips were washed with phosphate buffered saline (PBS), fixed for 15 minutes at room temperature in 2% formaldehyde buffered with PBS, and permeabilized with 0.5% Triton X-100 for 2 minutes at room temperature. Alternative fixation procedures included 100% methanol for 10 minutes at −20°C, followed by 100% acetone for 5 minutes at −20°C; 100% methanol for 10 minutes at −20°C, fol-lowed by 0.5% NP-40 (in PBS) on ice for 5 minutes; and 70% ethanol for 10 minutes at −20°C. Coverslips or commercial slides were incubated with human sera at a 1:200 dilution in PBS for 30 minutes at room temperature. For double-label IIF, the human sera were mixed with murine monoclonal anti-DNA antibody or rabbit anti-PCNA antibody diluted at 1:50. FITC-conjugated goat anti-human or anti-rabbit, and rhodamine-conjugated goat antimouse or anti-human IgG antibodies (Caltag, So. San Francisco, CA) were used as secondary detecting reagents. After the IIF procedure, coverslips/commercial slides were examined using either an Olympus BH-2 fluorescence microscope or a Bio-Rad MRC-600 argon/krypton laser confocal microscope. Confocal images were collected simultaneously from rhodamine and fluorescein channels, and merged using a Bio-Rad CoMos software.

For IIF studies on chromosomal spreads, actively growing Indian muntjac cells were arrested in metaphase by treatment with 0.1 μg/ml colcemid for 16 hours and then harvested by mild trypsinization, followed by several washes with DMEM and PBS. Chromosome spreads were prepared by the method of Merry et al. (1985) and then processed for IIF as described above. Chro-mosomes labeled with antibodies were counterstained with ethid-ium bromide (1 μg/ml in PBS) for 5 minutes and rinsed in water for 5 minutes.

WI-38 human fetal lung fibroblasts were seeded onto Lux Per-manox dishes (Electron Microscopy Sciences, Fort Washington, PA) and processed as monolayers after 2 days in culture. Cells were rinsed briefly with Hanks’ balanced salt solution (HBSS) and then fixed for 30 minutes at room temperature with 3% formalde-hyde buffered with HBSS. Following several rinses with HBSS, cells were permeabilized for 5 minutes at room temperature with 0.5% Triton X-100 in HBSS, rinsed with HBSS, and blocked by incubation with 1% normal goat serum (NGS) in HBSS for 30 minutes at room temperature. Fixed and permeabilized cells were then incubated overnight at 4°C with a 1:100 dilution of human autoimmune serum made in 1% NGS/HBSS. Cells were then rinsed with HBSS, incubated for 2 hours at 37°C with a 1:50 dilu-tion of affinity-purified goat anti-human IgG coupled to horse-radish peroxidase (Cappel, Durham, NC) in 1% NGS/HBSS, rinsed with HBSS, rinsed with 50 mM Tris-HCl (pH 7.6), and then incubated for 5 minutes with 1 mg/ml diaminobenzidine tetrahydrochloride (Polysciences, Warrington, PA) in 50 mM Tris-HCl (pH 7.6) containing 0.03% H₂O₂. The peroxidase reaction was terminated by rinsing with distilled water and then intensi-fied by incubation in 1% OsO₄ for 30 minutes. Following osmi-fication, cells were dehydrated in ethanol and embedded in Polybed 812 (Polysciences). Embedded cells were sectioned as monolayers and examined in the electron microscope unstained.

For immunoblotting, cells grown in culture dishes were solubi-lized directly in Laemmli’s sample buffer (Laemmli, 1970) con-taining the following protease inhibitors: leupeptin (5 μg/ml), pep-statin (2 μg/ml), phenylmethylsulfonyl fluoride (PMSF) (175 μg/ml), aprotinin (5 μg/ml), p-chloromercuriphenylsulfonic acid (CMSA; 0.1 mM), N-ethylmaleimide (1 mM), and α₂-microglob-ulin (1 μg/ml). The extract was passed several times through a 27-gauge needle, heated at 100°C for 5 minutes, and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using a 5% gel. After electrophoresis, proteins were transferred at 4°C to polyvinyldifluoride (PVDF) (Millipore, Bedford, MA) membranes for 5 hours at 500 mA, using a Bio-Rad Trans-Blot Electrophoretic Transfer Cell equipped with a cooling system. The transfer buffer contained 10% methanol. Membranes were cut into strips and blocked overnight at 4°C with PBS containing 0.1% casein and 0.1% gelatin. Individual strips were incubated with human serum diluted 1:100 in blocking solution. Bound antibod-ies were detected with [¹²⁵I]Protein A (ICN Biochemicals, Irvine, CA) followed by autoradiography.

For affinity purification of antibodies from western blots, PVDF strips containing the bands of interest were blocked overnight with PBS/casein/gelatin as described above, incubated for 2 hours with autoimmune sera at 1:50 dilution, and washed extensively with PBS/Tween-20. Antibodies were then eluted from the membranes by incubation for 2 minutes with 200 mM KPO ₄, pH 2.5, 150 mM NaCl, 0.1% BSA, followed by neutralization with 1 M Tris (pH 8.7). Finally, antibodies were concentrated using Centricon-30 microconcentrators (Amicon, Beverly, MA) and stored at −20°C for subsequent immunofluorescence and immunoblotting studies.

Peripheral blood lymphocytes (both quiescent and PHA-stimulated) and continuously cycling MOLT-4 cells were used in the flow cytometry studies. 1×10⁶ cells were fixed and permeabilized with 100% methanol for 10 minutes at −20°C. Following three washes with PBS, the cells were incubated in suspension with human sera diluted at 1:200 for 30 minutes at room temperature. After PBS washes, the cells were incubated with a FITC-conju-gated goat anti-human antibody for 30 minutes. Cells were again washed with PBS, resuspended in PBS containing propidium iodide (10 μg/ml) and RNase A (20 μg/ml), and kept dark on ice until used in a flow cytometry analysis, which was performed in a FACScan flow cytometer (Becton Dickinson Immunocytometry Systems, San Jose, CA). Five to ten thousand cells were collected and stored in list mode for each sample. Debris and damaged cells were excluded by gating on a forward and side scatter dot plot and on the DNA histogram. Electronic compensation was used to compensate for overlapping fluorescence between different chan-nels. The data obtained were evaluated with FACScan software (Becton Dickinson).

The method of Staufenbiel and Deppert (1984) for the in situ preparation of nuclear matrices devoid of DNA, RNA and solu-ble nuclear proteins was employed with some minor modifica-tions. Briefly, HeLa cells grown on coverslips were washed three times for 5 minutes each at 4°C with Kern-Matrix (KM) buffer (10 mM MES, pH 6.2, 10 mM NaCl, 1.5 mM MgCl₂, 0.5 μM PMSF, and 10% (v/v) glycerol) to affix the cells to coverslips. Cells were then washed for 30 minutes with KM buffer containing 1% (v/v) NP-40 at 4°C, with gentle shaking. Cells were washed with KM buffer as before to remove detergent and then treated successively with DNase I (50 μg/ml) for 15 minutes at 37°C, 2 M NaCl for 15 minutes at 4°C, and a mixture of DNase I/RNase A (50 μg/ml) for 15 minutes at 37°C. Each treatment was followed by washes with KM buffer. Cells were fixed with methanol/acetone and processed for immunofluorescence as described above.

In an analysis of several hundred human autoimmune sera, we identified 18 sera from patients with miscellaneous dis-ease conditions that shared an unusual cell cycle-related staining pattern by IIF (Fig. 1). The IIF analysis was per-formed using commercial HEp-2 slides, which are prepared from continuously cycling HEp-2 cells and contain cells representing all the stages of the cell cycle from G₁ to M. The staining pattern in interphase cells consisted of very fine dots distributed throughout the entire nucleus but spar-ing the nucleoli. Variations in the staining intensity of inter-phase nuclei could be observed. Metaphase cells exhibited discrete dots associated with chromosomes. Staining of the spindle midzone and midbody region was also observed in mitotic cells. The cell cycle-related distribution of the target antigen did not depend on the fixation methods employed, suggesting that it was not an artifact of the fixation proce-dures and that it reflects the in vivo distribution of the target antigen. In addition, titration of the serum antibodies to their end-point dilution (i.e. negative IIF, in most cases >1:1000) did not lead to any change in the IIF pattern. The antigen appears to be conserved in mammals, since it was detected by IIF on a variety of cell lines, including human HEp-2, HeLa, U-87 MG and WI-38; rat NRK; rat kangaroo PtK₂; Indian muntjac; and mouse 3T3.

To detect the target antigen recognized by the autoim-mune sera, whole HeLa extracts were solubilized with Laemmli SDS-sample buffer, resolved in 5% polyacry-lamide gels, and subjected to immunoblotting analysis. Fig. 2 shows an immunoblot with representative sera. The sera reacted with a doublet in the 330 kDa region. No other common reactivities were detected, either in 5% gels or in 10% gels (data not shown). The common recognition of a 330 kDa doublet by the autoimmune sera strongly suggested that this protein doublet, which will be referred to as p330^d hereafter, is the target antigen identified by IIF. To establish this association, antibodies were affinity puri-fied from PVDF strips containing p330^d. In order to have an internal control in these experiments, separate strips con-taining p330^d and the nuclear mitotic apparatus antigen (NuMA, 240 kDa) were reacted with a mixture containing the relatively monospecific autoimmune sera JG (anti-p330^d, Fig. 2, lane 3) and BL (anti-NuMA, Fig. 2, lane 2). Both sera were used at 1:50 dilution. After incubation for 2 hours and following extensive washes with PBS/Tween-20, antibodies were eluted from the PVDF membranes and analyzed by IIF and western blotting. Staining of HEp-2 cells with the mixture of sera JG and BL is shown in Fig. 3A. Note that the mixture gave a combination IIF pattern that consists of homogeneous staining of all nuclei and spin-dle poles produced by the NuMA antibodies, and punctate staining of metaphase chromosomes produced by the p330^d antibodies. Antibodies eluted from the region of the blot containing the NuMA protein gave the typical NuMA stain-ing with no punctate staining of metaphase chromosomes (Fig. 3B). On the other hand, antibodies eluted from the 330 kDa region displayed a staining pattern indistinguish-able from that of the unfractionated anti-p330^d sera. As shown in Fig. 3C,D, the characteristic patterns seen during interphase and mitosis (metaphase chromosomes and spin-dle midzone) were observed. Midbody staining was also present (not shown). These affinity-purified antibodies showed no NuMA staining and their specificity was con-firmed by immunoblotting (data not shown).

Double-labeling experiments using unfractionated anti-p330^d serum JG and a mouse anti-DNA monoclonal anti-body were conducted to follow the intracellular distribution of p330^d during the cell cycle. These experiments were per-formed using commercial HEp-2 slides, which gave IIF images identical to those observed in HEp-2, HeLa or WI-38 cells grown on coverslips. Chromosome condensation and positioning as viewed by antibody staining were used to identify the different stages of the cell cycle. Fig. 4 shows the distribution of p330^d throughout the cell cycle as visualized by laser confocal microscopy. As shown above, nuclear staining of interphase cells by the p330^d sera dis-played a remarkable variability (Fig. 4A). No significant cytoplasmic staining was observed in these cells. As cells entered prophase, most of p330^d appeared as discrete dots (Fig. 4B), some of which could be seen in pairs, indicative of possible staining of the centromere region. During metaphase, the p330^d dots were localized to the region of aligned chromosomes, presumably associated with the cen-tromeres (Fig. 4C). In addition, a relatively intense stain-ing in the area surrounding the chromosomes could be observed during this phase. Since this staining was observed with the 18 sera and with affinity-purified antibodies from the 330 kDa doublet (Fig. 3C), we presume that it was prob-ably due to the release of a population of non-chromoso-mal-associated antigen upon nuclear envelope breakdown combined with the typical autofluorescence of metaphase cells. During early anaphase, p330^d was still detected in association with chromosomes (Fig. 4D), although it was evident that most of the staining at this phase was associ-ated with the spindle midzone. By early telophase, p330^d was no longer detected in the chromosomes but was seen predominantly associated with the cleavage furrow, form-ing a belt-like structure (Fig. 4E). As telophase proceeded, p330^d was concentrated in the intercellular bridge regions flanking the midbody (Fig. 4F) and disappeared gradually as the daughter cells separated.

To determine whether p330^d is associated with cen-tromeres during mitosis, we stained cycling Indian muntjac and PtK₂ cells, as well as chromosomal spreads derived from colcemid-treated Indian muntjac cells. The autoanti-bodies produced the double-dot pattern typical of cen-tromere staining (Fig. 5A-C). Analysis of chromosomal spreads indicated that p330^d was located in the external sur-face of centromeres, presumably within the kinetochore domain (Fig. 5B). Pairs of centromeres stained with anti-p330^d serum JG in prometaphase PtK₂ cells appeared to be more widely separated (Fig. 5C) than pairs stained with autoimmune serum KS (Fig. 5D), which is relatively mono-specific for the autoantigen CENP-B (Ochs and Press, 1992). This difference in the resolution of the centromere pairs between the two sera can be attributed to the known central location of CENP-B within the centromere (Pluta et al., 1990) and the likelihood of a more external kinetochore location for p330^d.

We employed immunoelectron microscopy to explore further the intracellular distribution of p330^d during the cell cycle of WI-38 cells stained with anti-p330^d serum JG. The distribution was consistent with that observed by IIF and confocal microscopy in HEp-2 cells. In interphase cells (Fig. 6A), p330^d appeared as fine dots distributed through-out the nucleoplasm but sparing the nucleoli. In prophase (Fig. 6B), the staining was more punctate and was confined to the regions between the condensing chromosomes, pre-sumably at the centromeres. At metaphase (Fig. 6C), p330^d staining was clearly seen concentrated at centromere regions on the chromosomes of the metaphase plate. When chromosomes began to separate during early anaphase (Fig. 6D), p330^d staining could still be observed at the cen-tromeres although it was predominantly confined to the spindle midzone. The redistribution of p330^d was seen more dramatically during late anaphase (Fig. 6E), when it could not be detected in the separated chromosomes but was localized in the spindle midzone. By telophase, p330^d was completely excluded from the decondensing chromosomes (Fig. 6F).

The cell cycle-related pattern of p330^d immunoreactivity suggested that the expression of this protein was associated with proliferation. To explore this aspect further, normal and PHA-stimulated human peripheral blood lymphocytes were stained with either anti-p330^d serum JG or normal human serum (NHS) and processed for flow cytometry analysis. Fig. 7A shows overlay histograms of normal lym-phocytes stained with JG and NHS sera. These lympho-cytes, which are known to represent nonproliferative G₀ cells, displayed no p330^d positivity, as judged by the over-lapping of the p330^d and NHS signals. However, stimula-tion of lymphocytes with PHA for 72 hours, which allowed the cells to go through one cell cycle prior to analysis, showed that p330^d staining increased markedly relative to the NHS staining (Fig. 7B), indicative of increased levels of p330 ^d in these proliferating cells.

The expression of p330^d during the cell cycle was also analyzed by flow cytometry in continuously cycling MOLT-4 cells stained simultaneously with p330^d autoanti-bodies and propidium iodide. Fig. 7C,D shows the fluores-cence distribution of cells stained with NHS and anti-p330^d serum JG, respectively. p330^d levels started to increase steadily at the G₁/S boundary and reached a maximum in G₂/M. Fig. 7E shows the fluorescence distribution of cells blocked in mitosis with nocodazole for 5 hours. Overlay histograms comparing p330^d levels in exponentially grow-ing cells (continuous line) and in nocodazole-blocked cells (dotted line) are shown in Fig. 7F. As expected, high levels of p330^d were observed in mitosis-arrested cells. Note that the peak in the G₁ population detected in exponentially growing cells disappeared after treatment with nocodazole, suggesting that it could have been composed of either very late G₁ cells and/or newly divided daughter cells that still retained residual levels of p330^dstaining. Various fixation methods were used in these experiments with no detectable differences in the profiles.

The above experiments indicate that the expression of p330^d at the protein level, as monitored by flow cytome-try, begins around the G₁/S transition. To determine whether this expression is concomitant with the onset of DNA replication, we simultaneously stained commercial HEp-2 slides with anti-p330^d serum JG and a rabbit anti-body to PCNA. The pattern of PCNA immunoreactivity in mammalian cells, which resembles the topographical pat-tern of DNA synthesis, has been described in detail before (Madsen and Celis, 1985) and was used to recognize dif-ferent stages of interphase. As shown in Fig. 8, the first evi-dence of p330^d staining was observed during early S phase. Cells in the G₁/S boundary showing PCNA staining were negative for p330^d. The intensity of nuclear staining with p330^d antibodies increased through S phase and reached maximum levels during G₂, a phase at which PCNA stain-ing could no longer be detected. These results indicate that the expression of p330^d begins during early S phase, shortly after the onset of DNA replication.

To obtain some insight into the biochemical properties of p330^d, HeLa cells were subjected to the extraction proce-dures described by Staufenbiel and Deppert (1984) for the preparation of nuclear matrices in situ. Cells grown on cov-erslips were first extracted with 1% NP40, then digested with DNase I, re-extracted with 2 M NaCl, and further digested with DNase I and RNase A. This procedure yields a ‘nuclear matrix’ preparation that is devoid of 90% of the total cellular protein and approximately 99% of the total DNA and RNA. After treatment, cells were fixed and stained with autoantibodies to p330^d. In addition, autoanti-bodies to NuMA, which is an established nuclear matrix protein (Yang et al., 1992), and to the non-nuclear matrix-associated antigens, PCNA and DNA/histones were employed as controls in these experiments. Fig. 9 shows the IIF staining with the autoantibodies after the final DNase/RNase digestion step. As expected, the staining with NuMA antibodies was not affected by the extraction procedure (Fig. 9A), whereas staining with PCNA and DNA/histone antibodies was completely abolished (data not shown). As shown in Fig. 9B, staining of interphase cells with p330^d antibodies was not affected after the treatment. Mitotic cells exhibited a weaker diffuse staining with no visible centromere staining (data not shown). These results suggest that p330^d is a component of the nuclear matrix as defined by the procedure of Staufenbiel and Deppert (1984).

Autoantibodies have been used extensively to identify and characterize intracellular autoantigens and to delineate their roles in important biological processes such as DNA repli-cation and repair, RNA processing and cell division (Tan, 1989, 1991). In this study we have employed autoimmune sera to characterize a novel cell cycle-regulated nuclear pro-tein, p330^d, which accumulates in the nuclear matrix during S phase but is relocated to centromeres during early mito-sis and to the spindle midzone and midbody after the metaphase/anaphase transition. The visualization of p330^d at the external surface of centromeres is indicative of a pos-sible location at the kinetochore domain. Electron microscopy studies using immunogold labeling procedures should allow a more precise localization within this domain.

The autoimmune sera recognized a 330 kDa doublet in western blots. An association between this doublet and the distinct IIF pattern exhibited by the sera was confirmed by the exact reproduction of both the IIF pattern and immunoblotting reactivity with affinity-purified antibodies from the 330 kDa region. The possibility that antibodies bound non-specifically to the PVDF membranes were eluted in these experiments was excluded by the specific reactivities displayed by the affinity-purified p330^d and NuMA antibodies. In future studies, it would be particu-larly important to determine whether the appearance of the doublet in western blots might reflect differences in the sub-cellular localization or solubility of p330^d during the cell cycle, post-translational modifications, or proteolytic break-down.

Our flow cytometry data indicate that the expression of p330^d begins around the G ₁/S transition and reaches a peak during G₂/M. Double-labeling IIF experiments using anti-p330^d and anti-PCNA antibodies showed more definitively, however, that p330^d expression begins during early S phase, shortly after the onset of DNA replication. This was con-firmed by IIF experiments with synchronized HeLa popu-lations obtained at various times after release from a double thymidine block (data not shown). p330^d expression was also found to be associated with proliferation, since its levels were negative in resting peripheral blood lympho-cytes but increased upon mitogen stimulation. This associ-ation with proliferation was confirmed by flow cytometry studies with differentiating HL60 cells showing that p330^d levels decreased markedly as differentiation progressed (A. Baez, C. Casiano, M. Cortes and K. Torres, unpublished results). The cell cycle-dependent expression of p330^d resembles that of two other proliferation-associated nuclear antigens, Ki-S1 (160 kDa) and Ki67 (345/395 kDa dou-blet), which are increasingly expressed from late G₁ to G₂/M (Landberg et al., 1990; Kreipe et al., 1993). How-ever, p330^d clearly differs from these antigens in its intra-cellular distribution. p330^d thus joins a growing group of proliferation-associated nuclear antigens and could poten-tially be employed as a marker to assess cell proliferation under normal and pathological circumstances.

The accumulation of p330^d in the nuclear matrix during S phase suggests that, in addition to its role in mitosis, this protein might participate in events in which the nuclear matrix is known to be actively involved during interphase, including DNA replication, higher-order chromatin organ-ization, and gene regulation (Cook, 1988; Verheijen et al., 1988). p330^d may also be required for events leading to the G₂/M transition. It is well established that this transition is characterized by a profound reorganization of cellular archi-tecture that includes the relocation of proteins such as NuMA (Yang et al., 1992), NMP125 (Marugg, 1992), mitotin (Todorov et al., 1988) and H1B2 (Nickerson et al., 1992) from the nuclear matrix to the mitotic apparatus.

The simultaneous presence of p330^d in the centromeres and in the area surrounding the chromosomes during metaphase suggests the possible existence of two popula-tions of this protein: one bound to centromeres and likely to be involved in kinetochore-directed chromosome move-ment; and the other, matrix-associated and perhaps playing a supportive role. It is tempting to speculate that this matrix-bound population could be responsible for most of the flu-orescence associated with the spindle midzone during late mitosis, since early anaphase cells showed strong staining of the central spindle while centromere staining was still evident. An alternative possibility is that the sera might con-tain two different antibody populations reacting with dif-ferent cellular antigens. In future experiments it should be possible to distinguish between these alternatives by using experimentally induced antibodies to p330^d.

p330^d shares some similarities with other members of the recently proposed class of chromosomal passenger proteins. This class include the INCENPs (135-155 kDa doublet; Cooke et al., 1987; Earnshaw and Cooke, 1991), the 37A5 antigen (140-150 kDa doublet; Pankov et al., 1990), the TD60 autoantigen (60 kDa; Andreasen et al., 1991), a 38 kDa autoantigen (Kingwell et al., 1987), and the kinesin-like protein CENP-E (312 kDa; Yen et al., 1991, 1992). Like all these proteins, p330^d is found in association with centromeres during early mitosis, with the spindle midzone during anaphase, and with the midbody during telophase. This unusual subcellular redistribution is indicative of pos-sible roles in centromere/kinetochore maturation and assembly, chromosome movement, regulation of the metaphase/anaphase transition, stabilization of the plus ends of polar microtubules located in the spindle midzone, and determination of the location of the cleavage furrow (reviewed by Andreasen et al., 1991; Cooke et al., 1987; Earnshaw and Bernat, 1991; Rattner, 1992). Sequence information on these proteins and microinjection studies should provide valuable insights into their cellular func-tions. It would also be of interest to determine whether their redistributions and activities are regulated by phosphoryla-tion/dephosphorylation events and/or by interactions with other cellular structures.

An issue that deserves further investigation is the rela-tionship between p330^d and other chromosomal passenger proteins. A recent report described the identification with autoantibodies of a novel 400 kDa doublet protein, desig-nated CENP-F, whose subcelullar localization resembles that of p330^d (Rattner et al., 1993). We compared, by IIF and immunoblotting, the reactivity of an autoimmune serum (VD) employed by these investigators with the reactivity of our autoimmune sera and found that serum VD recognized a 330 kDa doublet and gave an IIF pattern similar to that observed with our autoimmune sera. These results strongly suggest that p330^d and CENP-F are the same protein. Another protein with a high molecular mass and a mitotic distribution similar to that of p330^d is CENP-E. One clear difference, however, between these two proteins, as detected by IIF, is their subcellular localization during inter-phase and early mitosis. CENP-E staining is absent from the nucleus during interphase and prophase, consistent with the absence of a consensus nuclear targeting signal within its primary structure (Yen et al., 1992). It accumulates in the cytoplasm during late S and G₂ and first associates with centromeres during prometaphase (Yen et al., 1992). After metaphase, it follows a distribution identical to that of p330^d. On the other hand, p330^d begins accumulating in the nucleus during early S phase, it is not detected in the cytoplasm by IIF, and first becomes associated with cen-tromeres during early prophase. These differences in IIF staining would favor the argument that CENP-E and p330^dare different proteins located in separate subcellular com-partments prior to their redistribution to centromeres during early mitosis. Alternatively, the antibodies could be recog-nizing different epitopes of the same protein that become accessible only during specific stages of the cell cycle. If this is the case, however, it would be difficult to explain the failure to detect CENP-E on immunoblots of nuclear or cytoplasmic extracts prior to G₂ (Yen et al., 1991, 1992). A third possibility, as mentioned previously, is that the p330^d serum may contain various antibody populations or crossreacting antibodies that react simultaneously with dif-ferent cellular antigens, one of which could be CENP-E. Establishing a definite relationship between p330^d and other chromosomal passenger proteins will require a comparison of their amino acid sequences. As a step forward in this direction we have initiated the cDNA cloning of p330^d. Thus far, partial sequences obtained from four cDNA clones encoding portions of p330^d show no significant homology to either CENP-E or any other known protein (unpublished data).

In conclusion, we have described a novel cell cycle-reg-ulated nuclear protein whose behavior during mitosis makes it a likely candidate to join the new class of chromosomal passenger proteins. Detailed information on the structure and function of p330^d and related proteins should provide valuable insights into the mechanisms underlying mitosis. Further experiments on this protein will be aimed at inves-tigating its molecular structure and function, the mecha-nisms of its cell cycle-regulation, its physical interactions with other nuclear and centromere-associated proteins, and its potential use as a marker of cell proliferation.

We are very grateful to George Klier for assistance at the con-focal microscope facility of the Department of Cell Biology; Kevin Sullivan, Göran Roos, Robert L. Rubin, Haruhiko Imai and K. Michael Pollard for discussion and helpful suggestions; Thomas Stein for technical assistance in immunoelectron microscopy; Carol Peebles for help in searching for anti-p330^d sera; and J. B. Rattner and Marvin Fritzler for sharing unpublished data and for providing human anti-CENP-F serum VD. This work was supported by grant AR32063 from NIH. G. L. was a post-doctoral fellow supported by grant 3223-B92-02R from the Swedish Cancer Society. This is publication number 8018-MEM of The Scripps Research Institute.