The INCENPs are two polypeptides of 135 ×103 and 150×103 Mr that enter mitosis as tightly bound chromosomal proteins, but subsequently leave the chromosomes altogether and become associated with the central spindle and cell cortex at the contractile ring. In the experiments reported here we have used confocal microscopy and immunoelectron microscopy to provide a detailed picture of the intracellular location of these proteins during mitosis. The experiments have not only revealed a number of new details concerning the properties of the INCENPs in mitosis, but have revealed a number of novel aspects of the mitotic process itself. The first of these is the existence of a sequential pathway of structural changes in the chromosomes that occurs during metaphase. This pathway is revealed by the existence of four distinct INCENP staining patterns in mitotic cells. In ‘early’ and ‘early/mid’ metaphase, the INCENPs gradually become concentrated at the centromeres, forming a ring at the center of the metaphase plate. During *mid/late’ metaphase they exit from the chromosomes, so that by late metaphase they are found solely in streaks that traverse the plate parallel to the spindle axis. The streaks probably correspond to INCENPs closely associated with microtubule bundles, perhaps as part of the stem body material. Examination of transverse optical sections of the spindle interzone during early anaphase reveals an unexpectedly high degree of order. The INCENP antigens are localized on fibers that are organized into a hollow ring 8 μm in diameter and ∼4μm beneath the cell cortex. Measurement of cellular dimensions in the confocal microscope reveals that the maximum diameter of early anaphase cells lies across the spindle equator, so that when the cleavage furrow forms, it does so around the maximum circumference of the cell. During anaphase, a subpopulation of the INCENP antigen becomes localized to the cortex where the furrow will subsequently form. This occurs prior to any other evidence of furrowing. Thus, binding of the INCENPs to this region may represent an early step in furrow formation. Together, these results suggest that the INCENPs may represent a new class of ‘chromosomal passenger’ proteins that are carried to the spindle equator by the chromosomes and subsequently perform a cytoskeletal role following their release from the chromosomes at the metaphase:anaphase transition.

The events of mitosis involve a number of structural transitions. These include chromosome condensation, nuclear envelope breakdown, movement of chromosomes to the metaphase plate (congression), sister chromatid disjunction, movement of the sister chromatids to the spindle poles, and cytokinesis. These events have often been thought to be distinct, rather than inter-related processes. In particular, there has been a historical division between those whose studies have focused on the chromosomes and those who have preferred the various components of the cytoskeleton. This trend has recently begun to change, and there has been considerable interest from both camps in the structural makeup and functional characterization of the centromere/kinetochore, particularly since this interface between the chromosome and the spindle has been shown to be the probable location of an important mechanochemical motor component for anaphase chromosome movements (Gorbsky et al. 1988; Brinkley et al. 1988; Nicklas, 1989; Steuer et al. 1990; Pfarr et al. 1990).

The distinction between chromosomes and cytoskeleton in mitosis has recently been blurred through the identification of a number of components that may provide a link between chromosomal and cytoskeletal functions in mitosis. The first of these proteins to be identified were designated INCENPs (INner CENtromere Proteins – 135, 155 ×103 Mr) because they are located between sister chromatids in the inner centromere of chromosomes from colcemid-blocked cells (Cooke et al. 1987).

The INCENPs were initially described in experiments designed to identify novel polypeptides that might be important for mitotic chromosome structure. To do this, we isolated the insoluble chromosome scaffold fraction from purified mitotic chromosomes (Lewis and Laemmli, 1982) and used this fraction as the antigen for the isolation of monoclonal antibodies.

The INCENPs turned out to be of particular interest because of the pattern of changes in structural affiliation that they undergo during mitosis. During interphase, INCENPs are detected solely in the nucleus by both immunofluorescence and subcellular fractionation (Cooke et al. 1987). They enter mitosis as chromosomal proteins, and the INCENPs from colcemid-arrested cells are among the most tightly bound chromosomal proteins known (as judged from their near quantitative retention in the chromosome scaffold; Cooke et al. 1987). Nonetheless, we found that the INCENPs detach completely from the chromosomes at the onset of anaphase, becoming associated with the overlapping microtubules of the central spindle and with the cell cortex at the contractile ring. As mitosis progresses, the INCENPs eventually become concentrated in the intercellular bridge flanking the midbody (Cooke et al. 1987).

This pattern of movements undergone by the INCENP antigens is not unique. A concurrent study identified a human autoantigen of 38 × 103Mr that appears to undergo an identical series of movements during mitosis (Kingwell et al. 1987). In addition, three studies have subsequently identified components of the kinetochore domain that are chromosomal through metaphase, but become associated with the central spindle (and ultimately, the intercellular bridge) during subsequent mitosis (Pankov et al. 1990; Compton et al. 1991; Yen et al. 1991). Finally, a monoclonal antibody analysis of mitotic spindle proteins identified a pair of antigens, known as CHOI, that are associated with the spindle during anaphase/telophase in a pattern apparently identical to that seen with the INCENPs (Sellitto and Kuriyama, 1988). The CHOI antigens have been proposed to be markers for the stem body material that coats anti-parallel overlapping microtubules of the spindle interzone during anaphase and telophase (Buck and Tisdale, 1962; McIntosh and Landis, 1971; McDonald et al. 1977; Sellitto and Kuriyama, 1988). They have recently been shown to be essential for mitotic progression (Nislow et al. 1990).

The pattern of changes undergone by the INCENPs during mitosis suggested that these antigens could be involved in several important mitotic functions, including regulation of sister chromatid disjunction, sliding apart of the two half-spindles during anaphase, and establishment of the cleavage furrow (Cooke et al. 1987). The first two of these functions have also been suggested for the other antigens discussed above (Kingwell et al. 1987; Sellitto and Kuriyama, 1988; Pankov et al. 1990; Compton et al. 1991; Yen et al. 1991). As the first step towards determination of the function of the INCENP proteins, we have now undertaken a coordinated analysis of their location during the various stages of mitosis using laser scanning confocal microscopy and immunoelectron microscopy.

In addition to the specific information provided about the INCENPs themselves, our results reveal several fundamentally new facts about the mitotic process itself. In particular, we find that the chromosomes undergo a pathway of sequential structural changes during metaphase. We also find that the spindle interzone of anaphase cells has an unexpectedly high degree of order. Finally, we have found that when anaphase cells prepare to cleave, they assemble a contractile apparatus around their greatest circumference. Transport of the INCENP antigens to the cortex may be an early event in furrow formation.

Buffers

KB is 10 mm Tris-HCl, pH 7.7, 0.15 M NaCl, 0.1% BSA. PBS is 10ma sodium phosphate, pH 7.4, 0.15 M NaCl, Ima EGTA, 0.01 % NaN3.

Cells

Chicken hepatoma-derived line DU249 (Langlois et al. 1974) was obtained from R. Eisenman (Hutchinson Cancer Center, Seattle), and was cultured in RPMI 1640+5% fetal bovine serum (Hyclone).

Antibodies

Anti-ENCENP monoclonal antibody 3D3 has been described (Cooke et al. 1987). Rabbit antibodies recognizing the INCENPs were obtained by immunization with a cloned INCENP fusion protein. These clones and the reagents derived from them will be the subject of a future publication (A. Mackay & W.C.E., unpublished results). The rabbit antibodies yield staining patterns that are indistinguishable from those obtained with the monoclonal antibody.

Confocal microscopy

INCENP.-tubulin double labeling. DU249 cells in various stages of mitosis were obtained by subjecting flasks of exponentially growing cells to three sequential mitotic shakeoffs at 5- to 10–min intervals. The detached cells were harvested by centrifugation following the third shakeoff, washed once with warm serum-free medium, and permitted to attach to Adhesio-Slides (M+M Developments, Ottawa, Canada) for 5–10min at 37°C. Cells were then fixed with 4 % paraformaldehyde in PBS for 5 min at room temperature, washed three times (2 min each) with KB+0.02% Triton X-100, once with KB, and incubated with rabbit anti-LNCENP diluted 1:500 in KB for 30min at 37°C. Slides were then washed three times (2 min, 5 min, 3 min) with KB, and incubated with a mixture of adsorbed biotinylated antirabbit (Vector Labs) diluted 1:100 plus mouse anti-tubulin (Accurate) diluted 1:30 in KB for 30 min at 37 °C. (The biotinylated secondary antibody was adsorbed with a concentrated extract made from paraformaldehyde-fixed mitotic chicken cells in order to minimize non-specific binding.) Following washing as before, slides were incubated in a mixture of streptavidin:Texas Red (BRL) diluted 1:800 plus fluorescein-conjugated sheep anti-mouse (Cappel), diluted 1:40 in KB for 30 min at 37 °C. Final washing was as described above, preparatory to mounting in 25% glycerol in PBS plus 100 mg ml−1 of DABCO (l,4-diazabicyclo(2.2.2)-octane; Aldrich) as an anti-fade agent.

INCENP.DNA double labeling

This was performed essentially as described above, with the following modifications. Primary antibody incubation utilized monoclonal anti-INCENP ascites fluid, diluted 1:1000 in KB. Secondary antibody incubation was with biotinylated goat anti-mouse (Vector) diluted 1:100 in KB. Third incubation was in streptavidimfluorescein (BRL) 1:500 plus 50 μgml−1 RNase A. In the final washes, thé second (5 min) step was with KB containing 5 μg ml−1 of propidium iodide.

Slides were examined using a BioRad MRC-500 or MRC-600 laser scanning confocal microscope (White et al. 1987) mounted on a Nikon Optiphot microscope using either Olympus 60-X or 100-X S-plan apochromat lenses (both of NA 1.4). Images were obtained by photography of the monitor using Kodak Ektachrome-400 film

Immunoelectron microscopy

Chicken DU249 cells obtained by mitotic shake-off, were placed on adhesio slides, (MM Developments) and fixed in PLP (0.01 M sodium meta periodate, 0.075 M lysine and 2 % paraformaldehyde in 0.037 NaPO4 buffer, pH 7.4; McLean and Nakane, 1974), for 15 min. After fixation the cells were washed once in D-PBS (8.06 mM Na2HPO4,1.47 mw KH2PO4,137mMNaCl, 2.7 mM KC1, 0.68 mM CaCl2, 0.492 mM MgCl2) and permeabilized in D-PBS with 0.05% Triton X-100 for 5 min. The cells were then placed in D-PBS with 1 % BSA for 5 min, and subsequent antibody dilutions (mouse monoclonal antibody 1:250), were done in the same buffer. After 1 h in primary antibody, the slides were washed and placed in goat anti-mouse antibody conjugated to 1 run gold (Janssen) for 2-4 h. The cells were fixed with 2 % glutaraldehyde in D-PBS for 30 min and silver enhanced using the gum arabic/silver lactate method (Danscher, 1981). The slides were dehydrated through a graded ethanol series and embedded in Polybed/812 (Polysciences). Thin sections were cut and placed on copper grids and stained with uranyl acetate and lead citrate.

Procedure for synchronizing mitotic cells

Mitotic cells were released from monolayer cultures by six cycles of selective detachment at 5- to 10-min intervals (pre-shakes). Following the sixth selective detachment, the released cells were removed from the culture medium, which was then returned to the flasks for a further 8 min at 37 °C. A final round of selective detachment was carried out, yielding a synchronous population of mitotic cells that had all entered mitosis during the 8-min period. These cells were washed once in serum-free medium, plated on Adhesio slides, and incubated at 37°C. The centrifugations, resuspension and plating took a further 10 min. At 5-min intervals thereafter slides were transferred to 4% paraformaldehyde and processed for detection of INCENPs by indirect immunofluorescence. DNA was stained with propidium iodide. The elides were then examined in the confocal microscope, and metaphases were scored into four morphological classes based on their INCENP staining pattern. DNA morphology was noted (and was consistent with the proposed pathway), but was not used in the scoring.

Distribution of the INCENPs from interphase through prometaphase

When the distribution of INCENPs and DNA was examined in prophase cells by confocal microscopy, the two patterns were found to be largely superimposed, although the INCENP staining was considerably more diffuse (Fig. 1C,D). Detailed examination revealed that the distribution of the INCENPs along the chromosomal arms is not uniform. Certain regions of the chromosome stain more intensely with the anti-INCENP antibodies than others. Two regions of reduced staining are indicated by arrows in Fig. 1C,D. This irregularity of staining may account for the speckled nature of the INCENP fluorescence staining in interphase cells (Fig. 1A,B).

Fig. 1.

Distribution of INCENP antigens from mterphase through prometaphase as seen by confocal microscopy. (A,B) Interphase. (A) Tubulin staining. (B) INCENP staining. Bar, 10μm. (C,D) Prophase. (C) Propidium iodide staining of DNA. (D) INCENP staining. Two regions where the INCENP staining is relatively faint compared with the DNA distribution are shown by arrows. (E,F) Late prometaphase. (E) DNA staining. (F) INCENP staining. The arrows indicate the centromeres of two chromosomes still in the process of congressing to the metaphase plate. INCENP staining in all panels was with monoclonal antibody (mAB) 3D3. Bars: C–F, 5μm.

Fig. 1.

Distribution of INCENP antigens from mterphase through prometaphase as seen by confocal microscopy. (A,B) Interphase. (A) Tubulin staining. (B) INCENP staining. Bar, 10μm. (C,D) Prophase. (C) Propidium iodide staining of DNA. (D) INCENP staining. Two regions where the INCENP staining is relatively faint compared with the DNA distribution are shown by arrows. (E,F) Late prometaphase. (E) DNA staining. (F) INCENP staining. The arrows indicate the centromeres of two chromosomes still in the process of congressing to the metaphase plate. INCENP staining in all panels was with monoclonal antibody (mAB) 3D3. Bars: C–F, 5μm.

The INCENPs remain chromosomal throughout prometaphase; however, as the chromosomes condense further, the uneven distribution of antigen along the chromosome becomes even more visible than it was at prophase. The INCENP staining in prometaphase cells appears to be concentrated at both centromeres and telomeres, and to be less intensely distributed on the euchromatic chromosome arms. Fig. IE,F shows a late prometaphase where only a few chromosomes have not yet completed their congression to the plate. The concentration of INCENP antigen at two of the centromeres is indicated by arrows.

Distribution of INCENPs during metaphase

The distribution of the INCENPs changes substantially as metaphase progresses. Four distinct patterns of INCENP staining were revealed in cells that were in chromosomal metaphase (i.e. no separation of sister chromatids was visible). We have shown that these patterns correspond to a temporal series of changes as cells progress through metaphase. For ease of discussion, we will first describe the patterns using terms that reflect this interpretation, and then present the evidence supporting our conclusion that the various patterns form a temporal sequence.

The metaphase distribution of INCENPs is shown in longitudinal optical sections in Fig. 2, and transverse sections in Fig. 3. For the purpose of this discussion, we define longitudinal sections as sections that cut the metaphase cell parallel to the spindle axis. Such sections are perpendicular to the metaphase plate. We use the term transverse to describe sections that cut the metaphase cell perpendicular to the spindle axis. Such sections fall in the plane of the metaphase plate.

Fig. 2.

Longitudinal optical sections showing the distribution of ENCENPs in cells at various stages of metaphase. (See Fig. 4 for the evidence that these correspond to a temporal sequence of changes throughout metaphase.) (A,B) Early metaphase (INCENPs on centromeres and protruding chromosome arms). (C,D) Early/mid metaphase (INCENPs concentrated on centromeres). (E,F) Mid/ late metaphase (complex distribution of INCENPs with both centromere labeling and streaks between chromosomes). (G,H) Late metaphase (ENCENPs in streaks crossing the metaphase plate). (A,C,E,G) Propidium iodide staining of the DNA. (B,D,F,H) INCENP staining with mAB 3D3. Bar, 5μm.

Fig. 2.

Longitudinal optical sections showing the distribution of ENCENPs in cells at various stages of metaphase. (See Fig. 4 for the evidence that these correspond to a temporal sequence of changes throughout metaphase.) (A,B) Early metaphase (INCENPs on centromeres and protruding chromosome arms). (C,D) Early/mid metaphase (INCENPs concentrated on centromeres). (E,F) Mid/ late metaphase (complex distribution of INCENPs with both centromere labeling and streaks between chromosomes). (G,H) Late metaphase (ENCENPs in streaks crossing the metaphase plate). (A,C,E,G) Propidium iodide staining of the DNA. (B,D,F,H) INCENP staining with mAB 3D3. Bar, 5μm.

Fig. 3.

Transverse optical sections showing the distribution of INCENPs in the metaphase plate of early, early/mid and late metaphase cells. (A) Early metaphase; (B) early/mid metaphase; (C,D) late metaphase. (A–C) INCENP labeling with mAB 3D3. (D) DNA labeling. The location of three of the spots of INCENP staining that correspond to gaps in the DNA staining are indicated by arrows. Bar, 5μm.

Fig. 3.

Transverse optical sections showing the distribution of INCENPs in the metaphase plate of early, early/mid and late metaphase cells. (A) Early metaphase; (B) early/mid metaphase; (C,D) late metaphase. (A–C) INCENP labeling with mAB 3D3. (D) DNA labeling. The location of three of the spots of INCENP staining that correspond to gaps in the DNA staining are indicated by arrows. Bar, 5μm.

In cells that our kinetic analysis (see below) has subsequently shown to be in early metaphase, the chromosomes are organized in a loose radial array at the spindle equator (Fig. 3A). Protruding chromosome arms radiate outwards from the more compact chromatin at the center of the metaphase plate (Fig. 2A). (The thin plane of optical section in the confocal microscope makes it difficult to visualize more than one or two chromosomal arms in longitudinal sections.) The INCENP staining in these cells resembles that seen at prometaphase. The antigen is present on the chromosomes, typically appearing as ‘fuzzy’ balls coincident with the centromeres (Figs 2A,B and 3A), and also as foci coincident with the chromosomal telomeres.

In early/mid metaphase the chromosomal morphology is indistinguishable from that in early metaphase cells (Fig. 2C); however, the INCENP staining appears distinctly different. The staining is still co-localized with the centromeres, where it looks like fuzzy balls. However, the staining of the euchromatic arms and telomeres is much reduced (Figs 2C,D and 3B). The net result is that the INCENP antigen appears to become concentrated at the center of the metaphase plate. We do not know if this change involves actual movement of the INCENP antigen, or if it reflects a selective retention of antigen at centromeres and selective loss elsewhere.

By mid/late metaphase both the chromosome morphology and INCENP staining patterns have changed. The metaphase plate now consists of a single condensed mass of chromosomes: distinct chromosomal arms are no longer visible (Fig. 2E). In these cells, the INCENP antigen is concentrated at the center of the metaphase plate, much as in early/mid metaphases. However, careful examination of the fluorescence pattern reveals that the nature of the staining has changed. Some punctate foci of staining are still observed; however, these are intermingled with streaks that appear to traverse the plate parallel to the spindle axis (Fig. 2F).

The distribution of DNA in late metaphase is indistinguishable from that of mid/late metaphase: the chromosomes are highly concentrated on the metaphase plate. The INCENP antigen is now present solely as streaks traversing the plate parallel to the spindle axis (Fig. 2G,H). When the distribution of DNA is examined in parallel, the streaks of antigen are found to correspond to regions of reduced DNA staining, suggesting that the INCENPs may no longer be bound to the chromosomes. This is shown best by examination of transverse optical sections of late metaphase cells. In this view, the lines appear as spots surrounded by a background of more diffuse staining (Fig. 3C,D). If the INCENP antigen and DNA pattern are displayed simultaneously on the video monitor (as in Fig. 7A) it can be seen that the INCENPs are concentrated in regions of reduced DNA staining between the chromosomes. The locations of three of the INCENP spots determined in this way are indicated by arrows in Fig. 3C,D.

The late metaphase pattern of INCENP staining was missed in our earlier study of the INCENP antigens (Cooke et al. 1987), probably because it occurs in a relatively small percentage of the mitotic cells. When INCENP staining morphology was scored for a random mitotic population, 44 of 303 metaphases (15 %) had the ‘late’ morphology. We now routinely recognize this pattern in the standard fluorescence microscope, however.

The interpretation that the staining patterns described above correspond to a temporal sequence of changes in the distribution of INCENPs during metaphase was initially suggested by the changes in chromosomal morphology seen at each stage. Thus, an ‘early’ INCENP pattern is only seen in metaphase plates where the chromosomes are relatively decondensed, while the late pattern is only seen in cells with highly condensed metaphase plates. Chromosomes continue to condense throughout metaphase, with the metaphase plate gradually increasing in compactness.

We developed a protocol to test whether the different INCENP distributions seen in metaphase cells represent inherent population variability or reveal a sequential pathway of structural states (Materials and methods). Briefly, a synchronous population of mitotic cells that entered mitosis during an 8-min period was induced to progress through mitosis attached to slides. Samples fixed at 5-min intervals were processed for detection of the INCENPs. The slides were examined in the confocal microscope, and metaphases scored into the four morphological classes described above, on the basis of their INCENP staining pattern.

The data of Fig. 4 confirm that the early and early/mid patterns are succeeded by the mid/late and late distributions in turn, as cells spend increasing lengths of time in mitosis. Early and early/mid metaphases, which together make up 88 % of the total metaphases at 5 min, drop to 7 % by 15 min. In contrast, mid/late metaphases peak at 10 min, while late metaphases rise from 2% at 5 min to 84 % at 20 min. This experiment thus offers strong support for the interpretation that the various INCENP staining patterns correspond to a series of sequential structural states during metaphase.

Fig. 4.

Sequential progression of INCENP labeling patterns in cells passing through metaphase in synchrony. Cells entering mitosis within an 8 – min interval were plated on Adhesio slides in warm serum-free medium and incubated for the indicated time prior to fixation and staining using mAB 3D3 (Materials and methods). Metaphase cells from each time point were examined by standard and confocal microscopy, and sorted into the four morphological classes shown in Figs 2 and 3.

Fig. 4.

Sequential progression of INCENP labeling patterns in cells passing through metaphase in synchrony. Cells entering mitosis within an 8 – min interval were plated on Adhesio slides in warm serum-free medium and incubated for the indicated time prior to fixation and staining using mAB 3D3 (Materials and methods). Metaphase cells from each time point were examined by standard and confocal microscopy, and sorted into the four morphological classes shown in Figs 2 and 3.

The distribution of the INCENPs deduced from the confocal micrographs was confirmed by immunoelectron microscopy. In these experiments, the cells were exposed to specific antibody prior to embedding, as in the immunofluorescence experiments. The bound antibody was then localized using 1 run colloidal gold conjugated to an anti-mouse secondary antibody.

In early metaphases (presumed to be such because the chromosomes form a loose radial array), the gold grains were found to be concentrated throughout the centromeric regions of the chromosomes (Fig. 5A). In other cells, where the metaphase plate appears to be slightly more compacted, the INCENP staining is characterized by the simultaneous presence of both streaks and clusters of gold particles distributed across the centromeric chromatin (Fig. 5B). On the basis of the chromosomal morphology and INCENP staining pattern, these cells are likely to correspond to one or both of the mid-metaphase stages described above. The high background gold labeling of the cytoplasm between the chromosomes may be equivalent to the diffuse INCENP staining also observed by confocal microscopy (Fig. 2D,F), or it might simply reflect nonspecific labeling by the immunogold reagents.

Fig. 5.

Immunoelectron microscopy of the localization of INCENPs in early, mid and late metaphase cells. (A) Early metaphase (INCENPs largely concentrated on centromeres); (B) mid metaphase (complex distribution of INCENP antigen); (C,D) late metaphase (streaks of INCENP antigen traversing the metaphase plate are clearly seen not to be on the chromosomes). INCENPs were detected with mAB 3D3. Bar, 1 μm.

Fig. 5.

Immunoelectron microscopy of the localization of INCENPs in early, mid and late metaphase cells. (A) Early metaphase (INCENPs largely concentrated on centromeres); (B) mid metaphase (complex distribution of INCENP antigen); (C,D) late metaphase (streaks of INCENP antigen traversing the metaphase plate are clearly seen not to be on the chromosomes). INCENPs were detected with mAB 3D3. Bar, 1 μm.

We also observed cells where the chromosomes were highly condensed and compacted on the plate. From the chromosome morphology, we concluded that these cells were probably in late metaphase. In these cells, the gold particles were found to be concentrated in linear tracks that traverse the metaphase plate between the densely packed chromosomes (Fig. 5C,D). Thus, immunoelectron microscopy confirms that when the INCENPs are present as streaks traversing the metaphase plate, the antigen is no longer associated with the chromosomes.

The streaks of INCENP antigen that traverse the metaphase plate run parallel to the spindle axis. Such a localization is what would be expected if the antigen were associated with microtubules. (This is also suggested by the localization of the antigens in anaphase cells - see below.) Unfortunately, the INCENP antigenicity is extremely sensitive to fixatives, and we have not yet been able to find conditions suitable for simultaneous detection of INCENPs and microtubules by electron microscopy. However, two observations suggest that the INCENPs may become associated with microtubules following their detachment from the chromosomes during late metaphase.

First, when INCENPs and tubulin are co-localized by confocal microscopy, the INCENP staining is seen to be closely associated with microtubules that penetrate the metaphase plate (Fig. 6). Second, when cells are subjected to brief treatment with high concentrations of colcemid (exposure to 4 – 8 μM colcemid for 15 min) a substantial redistribution of the INCENPs takes place. Colcemid treatment apparently causes the antigens to become concentrated in intense foci at the inner centromeres between the sister chromatids (Fig. 7B). This staining is much more discrete than that observed in unblocked metaphase cells, apparently because the diffuse background staining is greatly reduced (compare with Figs 3C,D and 7A).

Fig. 6.

Colocalization of ENCENPs and tubulin in metaphase cells. (A,B) ENCENP (A) and tubulin (B) staining pattern in an early/mid metaphase cell. (C) Colocalization of ENCENPs with microtubules traversing the metaphase plate. This shows a superimposition of A and B. (D,E) Tubulin (D) and ENCENP (E) localization on a second cell in late metaphase. ENCENPs are present as streaks crossing the metaphase plate. Colocalization of one of the ENCENP streaks with microtubule staining is indicated by an arrow. Microtubules were detected with commercial monoclonal anti-tubulin; ENCENPs were detected with rabbit serum RaC. Bar, 5 μm.

Fig. 6.

Colocalization of ENCENPs and tubulin in metaphase cells. (A,B) ENCENP (A) and tubulin (B) staining pattern in an early/mid metaphase cell. (C) Colocalization of ENCENPs with microtubules traversing the metaphase plate. This shows a superimposition of A and B. (D,E) Tubulin (D) and ENCENP (E) localization on a second cell in late metaphase. ENCENPs are present as streaks crossing the metaphase plate. Colocalization of one of the ENCENP streaks with microtubule staining is indicated by an arrow. Microtubules were detected with commercial monoclonal anti-tubulin; ENCENPs were detected with rabbit serum RaC. Bar, 5 μm.

Fig. 7.

Colocalization of ENCENPs and DNA in metaphase cells in the absence and presence of colcemid. (A) Late metaphase from a normal cell. INCENPs, white; DNA, red. Note the ring-like distribution of th e INCENPs. The spots of ENCENP labeling lie in between the chromosomes. (B) Metaphase treated with colcemid (two adjacent optical sections shown). INCENPs, white; DNA, red. The INCENP label is hyperconcentrated in the inner centromeres. Bar, 5 μm.

Fig. 7.

Colocalization of ENCENPs and DNA in metaphase cells in the absence and presence of colcemid. (A) Late metaphase from a normal cell. INCENPs, white; DNA, red. Note the ring-like distribution of th e INCENPs. The spots of ENCENP labeling lie in between the chromosomes. (B) Metaphase treated with colcemid (two adjacent optical sections shown). INCENPs, white; DNA, red. The INCENP label is hyperconcentrated in the inner centromeres. Bar, 5 μm.

Changes in the distribution of INCENPs during early anaphase

The INCENP streaks remain behind at the position of the metaphase plate when sister chromatids separate at the onset of anaphase. Examination of cells in early anaphase reveals that an increase in the length of the INCENP streaks accompanies sister chromatid separation (compare Fig. 8A-D with Fig. 2H, noting the difference in magnification). These streaks are difficult to measure with the confocal microscope, since they are not necessarily parallel to the plane of optical section. Thus the following measurements underestimate the length of the streaks.

Fig. 8.

Colocalization of INCENPs and DNA in anaphase cells. (A-D) Very early anaphase; (E,F) mid anaphase; (G,H) late anaphase. The DNA staining is shown in A, C, E, G. INCENP staining (with mAB 3D3) is shown in B, D, F, H. The direction of chromatid separation is shown by inverted arrows in panels A, C, F. (F) The earliest anaphase cell (judged by the separation of the sister chromatids) in which labeling of the equatorial cortex was observed. Bar, 5 μm

Fig. 8.

Colocalization of INCENPs and DNA in anaphase cells. (A-D) Very early anaphase; (E,F) mid anaphase; (G,H) late anaphase. The DNA staining is shown in A, C, E, G. INCENP staining (with mAB 3D3) is shown in B, D, F, H. The direction of chromatid separation is shown by inverted arrows in panels A, C, F. (F) The earliest anaphase cell (judged by the separation of the sister chromatids) in which labeling of the equatorial cortex was observed. Bar, 5 μm

We have used the software provided with the confocal microscope to determine that in late metaphase the streaks are 2.09±0.36izm in length (28 separate late metaphases scored). This value doubles to 4.3±0.81 /an in early anaphases (defined as anaphases in which no INCENP labeling of the cell cortex can yet be detected - see below). These very early anaphases are rare (9 scored); however, the length difference is nonetheless clearly significant. (Anaphase in cultured chicken cells is extremely rapid, being complete in 1-2 min (Hughes and Swann, 1948).) The length of the INCENP streaks again decreases to 2.06±0.38 gm (64 measured) in mid anaphase cells.

Association of the INCENPs with the cell cortex during anaphase

In our initial description of the INCENP antigens, we noted that the antigens appeared to be located to both the central spindle and cell cortex above the spindle equator. In the present study we have been able to screen larger numbers of early anaphases, and have noted that INCENPs are not detected at the cell cortex until sister chromatid separation is well under way. We have measured the spacing between sister chromatids in early anaphase cells and correlated this with the presence or absence of INCENPs at the cortex. Representative data are presented in Fig. 8E-H, and the relevant measurements are summarized in Fig. 9. We have been unable to detect labeling of the cortex in any cells where the separation of sister chromatids was less than 4.2 μ m.

Fig. 9.

The relationship between separation of the sister chromatids and presence of detectable INCENP antigen at the equatorial cortex. The diagram shows the axis along which D was determined. The graph shows the distribution of early anaphase cells in which no INCENP labeling of the cortex was observed (•), and cells in which labeling was observed (×). Vertical position in the graph is arbitrary. The minimum separation at which surface labeling was observed was 4.2 μm. The maximum separation at which labeling of the cortex was not observed was 5.3 μm.

Fig. 9.

The relationship between separation of the sister chromatids and presence of detectable INCENP antigen at the equatorial cortex. The diagram shows the axis along which D was determined. The graph shows the distribution of early anaphase cells in which no INCENP labeling of the cortex was observed (•), and cells in which labeling was observed (×). Vertical position in the graph is arbitrary. The minimum separation at which surface labeling was observed was 4.2 μm. The maximum separation at which labeling of the cortex was not observed was 5.3 μm.

The close association of the INCENPs with the cell cortex in the cleavage furrow seen by confocal microscopy (Fig. 8) was readily confirmed by immunoelectron microscopy (Fig. 10). Gold label was found to be concentrated in a region immediately beneath the cell membrane (labeled m in Fig. 10). In both optical and electron microscope sections of anaphase cells, the staining is confined to the cortex in the cleavage furrow itself, and is absent from the cell surface above or behind the chromosome mass (labeled u in Fig. 10).

Fig. 10.

Immunoelectron microscopy of INCENPs in telophase cells showing the close association of the INCENPs with the cell membrane, m indicates regions where gold particles are seen in close association with the cell membrane in the cleavage furrow. u indicates regions of the membrane outside of the cleavage furrow that are unlabeled with gold particles, n, nucleus. INCENPs were detected with mAB 3D3. Bar,1 μm

Fig. 10.

Immunoelectron microscopy of INCENPs in telophase cells showing the close association of the INCENPs with the cell membrane, m indicates regions where gold particles are seen in close association with the cell membrane in the cleavage furrow. u indicates regions of the membrane outside of the cleavage furrow that are unlabeled with gold particles, n, nucleus. INCENPs were detected with mAB 3D3. Bar,1 μm

Localization of the INCENPs to the equatorial cell cortex (where the cleavage furrow will subsequently form) apparently occurs during the very early stages of cleavage furrow formation. INCENPs are detected at the cortex while the cells are still apparently quite round, with no visible evidence for formation of a cleavage furrow (Figs 8E,F and 13). This impression is confirmed by comparisons of the diameters of these cells across the cleavage furrow with the diameter of metaphase cells in a similar plane (across the metaphase plate).

The diameter of early anaphase cells across the spindle equator is, surprisingly, greater than that parallel to the spindle axis. Early and mid anaphase cells measured 15.1±0.8μm across the cleavage furrow perpendicular to the spindle axis (26 cells measured), and 13.8±0.9 μ m parallel to the spindle axis. The small standard deviation of these measurements (±5 %) suggests that we measured a uniform population of cells in which the cleavage furrow had not yet begun to contract. The diameter of the cell equator decreases dramatically following the onset of cleavage at mid-anaphase. As a control for the above measurements, the average diameter of 26 metaphase cells was found to be 16.0±0.8gm perpendicular to the spindle axis in the plane of the metaphase plate and 13.9±0.9 gm parallel to the spindle axis. These measurements are summarized in Fig. 17 (below).

Organization of INCENPs in the spindle interzone during anaphase and telophase

The cortical labeling with INCENP antibodies is most prominent only during mid/late anaphase. In contrast, the labeling of internal streaks in the spindle interzone is observed throughout anaphase and telophase.

Double staining for INCENPs and tubulin suggests strongly that the streaks of INCENP antigen are closely associated with microtubules (Fig-11A,B). In favorable optical sections, the region of INCENP staining appears to colocalize with regions of anti-parallel overlap of polar microtubules. This conclusion is also supported by the observation that the INCENPs become highly focussed in the intercellular bridge as cleavage progresses (Fig-11C). Electron microscopy of cells undergoing cytokinesis confirms that the INCENP antigens become concentrated in the intercellular bridge (Fig. 12). The electron-dense structure of the midbody (Bellairs and Bancroft, 1975) is unlabeled under these conditions. It should also be noted that INCENP antigen is undetectable in the nuclei of cells undergoing cytokinesis (Figs 11 and 12).

Fig. 11.

Simultaneous colocalization of INCENPs and either tubulin or DNA in late anaphase and telophase cells. (A,B) INCENP: tubulin double localization in late anaphase cells. INCENPs, orange; tubulin, blue. INCENPs are intimately associated with the microtubules in the region of overlap. (Note early/mid metaphase cell at lower left of A.) (C) INCENP: DNA localization during cytokinesis. INCENPs, white; DNA, orange. INCENPs are concentrated in the intercellular bridge flanking the midbody. INCENP staining in A, B was with rabbit serum RaC; that in C was with mAB 3D3. Bar, 5 μm.

Fig. 11.

Simultaneous colocalization of INCENPs and either tubulin or DNA in late anaphase and telophase cells. (A,B) INCENP: tubulin double localization in late anaphase cells. INCENPs, orange; tubulin, blue. INCENPs are intimately associated with the microtubules in the region of overlap. (Note early/mid metaphase cell at lower left of A.) (C) INCENP: DNA localization during cytokinesis. INCENPs, white; DNA, orange. INCENPs are concentrated in the intercellular bridge flanking the midbody. INCENP staining in A, B was with rabbit serum RaC; that in C was with mAB 3D3. Bar, 5 μm.

Fig. 12.

Concentration of the INCENPs in the intercellular bridge flanking the midbody in cells undergoing cytokinesis. This micrograph shows that ENCENP antigens are not detected in the midbody region or in the reforming G1 nuclei. INCENPs were detected with mAB 3D3. Bar, 1μm.

Fig. 12.

Concentration of the INCENPs in the intercellular bridge flanking the midbody in cells undergoing cytokinesis. This micrograph shows that ENCENP antigens are not detected in the midbody region or in the reforming G1 nuclei. INCENPs were detected with mAB 3D3. Bar, 1μm.

The total length of the INCENP staining was preserved in telophase intercellular bridges, which contained two mirror-image stained regions, each of 1.16±0.14 μm (24 measured). This yields a total stained length of 2.3 μm, not significantly different from the length of the stained regions in late metaphase and mid anaphase. The two stained regions in early intercellular bridges were separated by an unstained region about 0.5 μm across.

We have used optical sectioning of anaphase cells to examine the organization of the INCENP antigen in the spindle interzone in greater detail. When the optical sections (Fig. 13) are used to construct stereo pairs (Fig. 14B), it can be seen that an outer ring of labeled cortex surrounds an inner ring of labeled streaks. These are labeled c and r, respectively, in Fig. 13. The streaks are presumably bundles of microtubules with associated INCENP antigen (Fig. 11). These bundles are located several micrometers beneath the surface. Fig. 14A shows a stereo pair constructed from a series of oblique sections of an early anaphase cell. This shows the ring-like distribution of labeled bundles particularly clearly. Cortical labeling was not yet present in this cell.

Fig. 13.

Optical sections of the distribution of INCENPs in a mid anaphase cell. INCENP labeling of the cortex (c) and inner ring (r) is indicated. Section spacing, 1.5μm. INCENPs were detected with mAB 3D3. Bar, 10μm.

Fig. 13.

Optical sections of the distribution of INCENPs in a mid anaphase cell. INCENP labeling of the cortex (c) and inner ring (r) is indicated. Section spacing, 1.5μm. INCENPs were detected with mAB 3D3. Bar, 10μm.

Fig. 14.

Stereo pairs showing the distribution of INCENPS in anaphase cells. (A) Oblique view of the inner ring of microtubule bundles in an early anaphase cell. (B) Stereo view calculated from the optical sections of Fig. 13. Bar, 10μm.

Fig. 14.

Stereo pairs showing the distribution of INCENPS in anaphase cells. (A) Oblique view of the inner ring of microtubule bundles in an early anaphase cell. (B) Stereo view calculated from the optical sections of Fig. 13. Bar, 10μm.

The most informative views of the organization of the INCENP bundles in the spindle interzone were obtained when anaphase cells that had adhered to the Adhesio slide at their polar surface were viewed in transverse optical section. Several of these images, which are unlike anything we have observed previously, are shown in Fig. 15. The images show a ring of intensely labeled spots surrounded by a thin continuous ring of label. The latter corresponds to the labeling of the cell cortex at the cleavage furrow (labeled c in Fig. 13), while the ring of spots apparently corresponds to cross-sections of the labeled INCENP bundles observed in longitudinal sections of anaphase cells (labeled r in Fig. 13). Fig. 15F shows a rare end-on view of a telophase cell.

Fig. 15.

Transverse sections through the cleavage furrow of anaphase cells showing the distribution of the INCENPs. All panels are stained with mAB 3D3 and were simultaneously stained with propidium iodide to show the DNA (which was absent from the planes of section shown). The INCENPs are located in a ring of spots, which correspond to the streaks labeled r in Fig. 13. (A-E) Mid anaphase cells. Comparison of the INCENP and DNA staining of the cell shown in F suggests that this cell was in early telophase. INCENPs were detected with mAB 3D3. Bar, 5μm.

Fig. 15.

Transverse sections through the cleavage furrow of anaphase cells showing the distribution of the INCENPs. All panels are stained with mAB 3D3 and were simultaneously stained with propidium iodide to show the DNA (which was absent from the planes of section shown). The INCENPs are located in a ring of spots, which correspond to the streaks labeled r in Fig. 13. (A-E) Mid anaphase cells. Comparison of the INCENP and DNA staining of the cell shown in F suggests that this cell was in early telophase. INCENPs were detected with mAB 3D3. Bar, 5μm.

That these images are cross-sections through the spindle interzone is indicated by two factors. First, when the chromosomes of these cells are double stained for DNA with propidium iodide, no DNA stain is visible in these focal planes by confocal microscopy, although the separated sister chromatids are readily visualized in focal planes both above and below those in which the double ring of INCENP staining is seen. Second, the INCENP labeling of the cortex is only seen in the spindle interzone of these cells.

Double staining for both tubulin and INCENP antigens in this region suggests that the stained filaments are either microtubules or microtubule bundles. When a cell is double-labeled for INCENPs and tubulin, and examined at two different optical planes (INCENPs through the cleavage furrow; tubulin from closer to the aster), the microtubules projecting from the aster lead to the INCENP-containing filaments (Fig. 16).

Fig. 16.

Simultaneous staining of INCENPs and microtubules in superimposed transverse sections from different locations in the spindle interzone of an anaphase cell. (A) Microtubules extend outwards from the aster (taken several am to one side of the cleavage furrow). (B) The ring of INCENP-containing spots at the mid-plane of the spindle interzone. The arrows in the two panels show several of the spots that lie at the end of the microtubule bundles shown in A. The arrows were placed using a color superposition of the two staining patterns (see cover of this issue of the journal). INCENPs were detected with rabbit serum RaC. Bar, 5μm.

Fig. 16.

Simultaneous staining of INCENPs and microtubules in superimposed transverse sections from different locations in the spindle interzone of an anaphase cell. (A) Microtubules extend outwards from the aster (taken several am to one side of the cleavage furrow). (B) The ring of INCENP-containing spots at the mid-plane of the spindle interzone. The arrows in the two panels show several of the spots that lie at the end of the microtubule bundles shown in A. The arrows were placed using a color superposition of the two staining patterns (see cover of this issue of the journal). INCENPs were detected with rabbit serum RaC. Bar, 5μm.

Fig. 17.

Diagram summarizing the distribution of INCENPs in cells in late metaphase, early anaphase and mid anaphase. The cells and INCENP streaks are drawn to scale using the indicated dimensions, which were determined using the software provided with the confocal microscope.

Fig. 17.

Diagram summarizing the distribution of INCENPs in cells in late metaphase, early anaphase and mid anaphase. The cells and INCENP streaks are drawn to scale using the indicated dimensions, which were determined using the software provided with the confocal microscope.

The most striking aspect of these images is the extent to which the INCENP bundles are evenly distributed in a single circular ring located 3 – 5 μm beneath the plasma membrane. The regularity of the structure is surprising. The diameter of this ring, 8.2±0.7 μm (21 measured), is relatively constant, and few, if any, prominently labeled spots are observed in the interior. In many cases (Fig. 15B and D are clear examples) the most strongly labeled spots are distributed in a distinctly symmetrical fashion.

The structure of the INCENP ring may be initiated during metaphase. A similar stained ring of spots is also observed in transverse optical sections through the plane of the metaphase plate of many late metaphase cells where the INCENP staining is concentrated in spots. However, the INCENP staining at metaphase is composed of a larger number of more weakly stained units (Fig. 7 A) than those observed at anaphase, suggesting that the microtubules may be bundled to a lesser extent.

INCENP staining patterns reveal a pattern of progressive structural changes in the chromosomes that occurs throughout metaphase

Metaphase has primarily been thought to be a period of waiting prior to the onset of anaphase, during which a feedback mechanism or checkpoint (Hartwell and Weinert, 1989) ensures completion of the alignment of the chromosomes at the spindle equator (Zirkle, 1970; Rieder and Alexander, 1989). It has long been known that chromosomes continue to condense and the spindle gradually becomes shorter and more robust with increasing time spent in metaphase (Taylor, 1959). Most models explaining the metaphase:anaphase transition have been focussed primarily on the nature of the attachment of chromosomes to the spindle. Our new results indicate, however, that the structure of the chromosomes themselves is different at different times during metaphase.

The changes that we see involve alterations in the relative distribution of the INCENP antigens. In early metaphase these antigens are found both at centromeres and telomeres. By early/mid metaphase they are concentrated primarily at the centromeres. The most surprising aspect of our results is that during mid/late metaphase the INCENPs undergo a progressive detachment from the chromosomes, becoming concentrated in streaks that traverse the metaphase plate parallel to the spindle axis. By late metaphase, the INCENPs appear to be fully released from the chromosomes. This release from the chromosomes has been confirmed by immunoelectron microscopy.

The detachment of INCENPs from the chromosomes may be dependent on transfer to microtubules that penetrate the metaphase plate. This is suggested by a number of observations. First, the INCENPs colocalize with microtubules that traverse the metaphase plate (Fig. 6). Second, immunoelectron microscopy reveals that the INCENPs traverse the late metaphase plate as fine lines, as expected if they were associated with filamentous elements (Fig. 5). Third, the association between INCENPs and microtubules during anaphase and telophase is readily apparent (Fig. 11). Fourth, exposure of metaphase cells to colcemid causes a dramatic rearrangement of the INCENPs, which become highly concentrated in the inner centromere (Fig. 7B).

Why should exposure to colcemid alter the distribution of the INCENPs so markedly? Release of the INCENPs from the chromosomes might require physiological signals that accompany the initiation of anaphase, which normally requires a fully assembled mitotic spindle with all chromosomes aligned at the metaphase plate (Zirkle, 1970; Rieder and Alexander, 1989). Furthermore, by causing microtubule disassembly, colcemid treatment may remove the non-chromosomal binding site for the INCENPs, causing them to remain ‘trapped’ on chromosomes. ‘The subsequent concentration of the bound antigen into tight foci in the inner centromeres may result from the continued chromatin condensation that occurs during exposure to colcemid.

The zone occupied by the INCENP fibers during late metaphase is limited to the width of the metaphase plate (Fig. 2). Thus, if the INCENPs do transfer to microtubules at this time, binding only occurs within this limited region. It cannot simply be true that the INCENPs have an affinity for microtubules in areas where adjacent tubules have opposite polarity. The spindle contains a significant number of polar microtubules that penetrate the metaphase plate between chromosomes and extend for significant distances beyond (K. McDonald, personal communication). There must also be some additional interactions, presumably with the chromosomes, to limit the extent of lateral spread of the INCENPs.

‘Late’ metaphase, with the INCENPs detached from the chromosomes, may actually represent a very early step in the initiation of anaphase. The degradation of the A-type cyclins occurs during metaphase, as judged by a lack of visible separation of sister chromatids in certain cells in which cyclin A is no longer detected (Lehner and O’Farrell, 1990). Cyclin A degradation thus defines an early step in the activation of anaphase events. It will be therefore be important in future experiments to examine the temporal relationship between the detachment of the INCENPs from the chromosomes and cyclin A degradation. If a correlation is observed, then models in -which the INCENPs play a role in the regulation of sister chromatid pairing (Cooke et al. 1987) remain worthy of consideration.

Regardless of the function of the INCENPs, our results suggest that current thinking about the regulation of the metaphase:anaphase transition may need to be broadened. In addition to the current belief that anaphase onset is triggered by kinetochore:microtubule interactions as the chromosomes achieve a bipolar orientation, our observations raise the possibility that the timing of the metaphase:anaphase transition could in principle also be sensitive to a pathway of structural change that occurs within the chromosomes themselves.

The cleavage furrow forms around the greatest circumference of anaphase DU249 cells

Both metaphase and anaphase DU249 cells are anisotropic: the diameter measured perpendicular to the spindle axis is longer than that measured parallel to the spindle axis (Fig. 17). This is surprising for anaphase cells, since it means that the cleavage furrow must form around the largest circumference of the cells. The anisotropy further suggests that the spindle exerts some influence on the cell shape. One explanation of the observed results is that tension along the axis of the spindle is somehow transmitted to the cell cortex, resulting in the slight flattening of the cell at the spindle poles. A weak connection between the spindle poles and the plasma membrane was previously suggested by Carlson (1952).

A transient elongation of the INCENP streaks accompanies sister chromatid separation

The INCENP streaks that form during mid/late and late metaphase double in length in early anaphase, and then return to their initial length by mid anaphase. These length changes are highly reproducible. Although the reason for the doubling in length is unknown at present, two possibilities suggest themselves.

First, the INCENPs could be ‘elastic’ and could remain attached in some way to the chromosomes during early anaphase. Thus, as the chromatids begin their movements towards the spindle poles, the INCENP fibers might undergo transient stretch. Later on, as the chromatids become more widely spaced, the INCENPs may detach and ‘snap back’ to their original dimensions.

Alternatively, the transient elongation of the INCENP streaks might result from sliding of the microtubules with which they are associated. This suggests that the initiation of anaphase may involve (or be accompanied by) a transient burst of microtubule sliding at the metaphase plate. Such sliding is unlikely to explain the entire lengthening of the INCENP streaks, however. Although spindle elongation does occur in early anaphase in some cases (Hughes and Swann, 1948; Oppenheim et al. 1973), movements of the magnitude required (∼2μm) have not been observed.

The length of microtubules coated with INCENPs (∼2μm) appears to change little during the subsequent course of anaphase and telophase. In fact, the INCENPs appear to remain more or less fixed at the location of the metaphase plate as the daughter cells migrate apart during anaphase B. This raises the possibility that the INCENPs may be components of a static substructure through which microtubules slide, as in diatoms (Wordeman and Cande, 1987; Wordeman et al. 1989).

INCENP bundles are present in a highly ordered ring in the spindle interzone

The distribution of the INCENPs in anaphase cells is complex, and for the purposes of this discussion we will break it down into two categories: INCENPs associated with the central spindle, and INCENPs associated with the cell cortex at the cleavage furrow. These distributions are shown di agrammatical ly in Fig. 17. Since there are two INCENP polypeptides of 135 and 150×103Mr (Cooke et al. 1987), it is possible that one of them may be associated with microtubules while the other associates with the plasma membrane. We are currently attempting to develop antibody probes that are specific for each INCENP species,

The INCENPs are found in intimate association with the microtubules of the central spindle during early anaphase, as soon as the sister chromatids have separated sufficiently for this region to be resolved. The portion of the spindle interzone associated with the INCENPs exhibits a surprisingly ordered structure, with the microtubules clustered into a single ring of bundles located 4—5 gm beneath the cortex. These bundles are distributed with considerable radial symmetry, suggesting that this portion of the dividing cell may be more ordered than has generally been supposed. We do not yet know if the hollow ring of INCENP-decorated bundles corresponds to the entire spindle in the interzone of these cells, or if there are other microtubules within the ring that are not associated with INCENPs. Such internal microtubules are not detected when INCENPs and tubulin are colocalized simultaneously (data not shown); however, it is possible that smaller microtubule bundles may not be stable under the mild fixation conditions required for INCENP visualization.

The distribution of the INCENPs in anaphase cells closely resembles that of the stem body material, which has been previously identified as a dark amorphous matrix that coats antiparallel microtubules of the central spindle during anaphase (Buck and Tisdale, 1962; McIntosh and Landis, 1971; McDonald et al. 1977). The function of the stem body material is unknown, although both motor and structural roles have been suggested (McDonald et al. 1977; Kingwell et al. 1987; Sellitto and Kuriyama, 1988; Wordeman and Cande, 1987). Stem body components could be involved either in the anaphase B separation of spindle poles (Sellitto and Kuriyama, 1988; Wordeman and Cande, 1987), or in maintaining the structural integrity between the two half-spindles (McIntosh et al. 1979; Kingwell et al. 1987) (or both).

Paradoxically, when the function of one putative stem body component was disrupted in vivo mitosis was blocked, but at a stage prior to anaphase. Injection of monoclonal antibody to the CHOI antigens (Sellitto and Kuriyama, 1988) results in the arrest of most cells in a state similar to metaphase dr late prometaphase (Nislow et al. 1990). This microinjection experiment thus suggests that stem body components may be required for mitotic events prior to the onset of anaphase. This may make it difficult to determine their role (if any) in anaphase and telophase.

INCENP localization to the equatorial cortex may be an early event in cleavage furrow formation

The consequences of concentrating the INCENPs at the cell cortex during anaphase are unknown. Several observations from our work are consistent with the possibility that the INCENPs might be involved in early events in either the establishment or the function of the contractile ring at the cleavage furrow.

First, we consistently observe INCENP staining of the cortex before there is any other evidence for contractile ring formation. Thus the timing of the movement of the INCENPs to the surface resembles that observed with myosin (Fujiwara and Pollard, 1976). In other published studies, cleavage furrow components were not observed to become concentrated at the equatorial cortex until later in anaphase (Nunnally et al. 1980; Sanger et al. 1989). It will be important in future experiments to time the order of appearance of INCENPs and myosin at the equatorial cortex by double staining with specific antibodies.

Second, immunoelectron microscopy reveals that the association of the INCENPs with the cortex is an extremely close one, as would be expected for a component of the contractile ring (Schroeder, 1973).

Third, it is now widely accepted that the ‘cleavage stimulus’ that is responsible for specifying the location of the contractile ring is directed to the appropriate region of the cell surface by forces that are focussed midway between asters (Rappaport, 1986). We have recently found that the localization of the INCENPs to the equatorial cortex of anaphase cells apparently requires microtubules.

If anaphase cells are incubated in the presence of colcemid, the INCENP staining apparently becomes randomized across the cell surface (W.C.E., preliminary results).

Fourth, it is well known that the mechanics of contractile ring function are based on an actomyosin network (Schroeder, 1973; Fujiwara and Pollard, 1976; Nunnally et al. 1980; Sanger et al. 1989; Cao and Wang, 1990). The possibility that the INCENPs might be involved in some way with the functioning of this network is strongly supported by preliminary sequence analysis of the cloned INCENPs, which reveals significant amino acid sequence similarities to several proteins that are known to interact with actin and tropomyosin (A. Mackay and W.C.E., unpublished results). Our future molecular analysis of the structure and function of the INCENPs will be initially directed towards determination of the role (if any) played by these proteins in the cleavage furrow.

Do the chromosomes participate in establishment of the architecture of the dividing cell?

Our experiments show that among the proteins that move with the chromosomes to the metaphase plate are some that are released from the chromosomes at (or just before) the metaphase:anaphase transition. These released proteins then mark the position of the metaphase plate once chromosomes move away from it during anaphase. During anaphase, one or more of the proteins becomes focussed to the cortex above the spindle equator (midway between spindle poles) in a process that apparently depends on microtubules.

These observations raise the possibility that the INCENPs, while they enter mitosis as components of the chromosomes, might principally function during mitosis after the onset of anaphase when they are no longer associated with the chromosomes. We thus suggest that the INCENPs may define a new class of chromosomal passenger proteins, whose association with the mitotic chromosomes is important primarily as a mechanism for positioning them properly in order to fulfil their roles after anaphase onset. This hypothesis, which we explore in greater depth elsewhere (Earnshaw and Bemat, 1991), suggests an involvement of chromosomes in mitosis beyond that previously assumed. Chromosomes may be more than passive repositories of genetic information. In addition, they may make essential structural contributions to the organization of the central spindle and the dividing cell.

We are grateful to Dr G. Sluder for many helpfill discussions, and apologize for doing our part to keep the concept of metaphase alive. We thank Dr A. Mackay for permitting us to quote his unpublished results; and also thank Drs A. Mackay, R. Rappaport, J.B. Rattner, G. Sluder, J. Tomkiel and E. Wood for their comments on the manuscript, which doesn’t correspond to their views in some cases. This work was supported by NTH grant GM30985 to W.C.E.

Bellaibs
,
R
and
Bancroft
,
M.
(
1975
).
Midbodies and beaded threads
.
Am. J. Anal
.
143
,
393
398
.
Brinkley
,
B. R.
,
Zinkowski
,
R P.
,
Mollon
,
W. L.
,
Davis
,
F. M.
,
Piseona
,
M. A.
,
Pershouse
,
M.
and
Rao
,
P. N.
(
1988
).
Movement and segregation of kinetochores experimentally detached from mammalian chromosomes
.
Nature
336
,
251
254
.
Buck
,
R C.
and
Tisdale
,
J. M.
(
1962
).
The fine structure of the midbody of the rat erythroblast
.
J. Cell Biol
.
13
,
109
115
.
Cao
,
L.
and
Wang
,
Y.
(
1990
).
Mechanism of the formation of contractile ring in dividing cultured animal cells. I. Recruitment of preexisting actin filaments into the cleavage furrow
.
J. Cell Biol
.
110
,
1089
1095
.
Carlson
,
J. G.
(
1952
).
Microdissection studies of the dividing neuroblast of the grasshopper Chortophaga vindifasciata (De Geer)
.
Chromosoma
5
,
199
220
.
Compton
,
D. A.
,
Yin
,
T. J.
and
Cleveland
,
D. W.
(
1991
).
Identification of novel centromere/kinetochore associated proteins using monoclonal antibodies generated against human mitotic chromosome scaffolds
.
J. Cell Biol, (in press)
.
Cooke
,
C. A.
,
Heck
,
M. M. S.
and
Earnshaw
,
W. C.
(
1987
).
The INCENP antigens: movement from the inner centromere to the midbody during mitosis
.
J. Cell Biol
.
105
,
2053
2067
.
Danscher
,
G.
(
1981
).
Localization of gold in biological tissue: a photochemical method for light and electron microscopy
.
Histochemistry
71
,
81
88
.
Earnshaw
,
W. C.
and
Bernat
,
R L.
(
1991
).
Chromosomal passengers: Toward an integrated view of mitosis
.
Chromosoma (in press)
.
Fujiwara
,
K.
and
Pollard
,
T. D.
(
1976
).
Fluorescent antibody localization of myosin in the cytoplasm, cleavage furrow, and mitotic spindle of human cells
.
J. Cell Biol
.
71
,
848
875
.
Gorbsky
,
G. J.
,
Sammak
,
P. J.
and
Borisy
,
G. G.
(
1988
).
Microtubule dynamics and chromosome motion visualized in living anaphase cells
.
J. Cell Biol
.
106
,
1185
1192
.
Hartwell
,
L. H.
and
Weinert
,
T. A.
(
1989
).
Checkpoints: controls that ensure the order of cell cycle events
.
Science
246
,
629
634
.
Hughes
,
A. F.
and
Swann
,
M. M.
(
1948
).
Anaphase movements in the living cell. A study with phase contrast and polarized light on chick tissue cultures
.
J. exp. Biol
.
25
,
45
70
.
Kingwell
,
B.
,
Frttzler
,
M. J.
,
Decoteau
,
J.
and
Rattner
,
J. B.
(
1987
).
Identification and characterization of a protein associated with the stembody using autoimmune sera from patients with systemic sclerosis
.
Cell Motil. Cytoskel
.
8
,
360
367
.
Langlois
,
A. J.
,
Lapis
,
K.
,
Ishizaki
,
R
,
Beard
,
J. W.
and
Bolocnesi
,
D. P.
(
1974
).
Isolation of a transplantable cell line induced by the MC29 avian leukosis virus
.
Cancer Res
.
34
,
1457
1464
.
Lehner
,
C. F.
and
O’farrell
,
P. H.
(
1990
).
The roles of Drosophila cydins A and B in mitotic control
.
Cell
61
,
535
547
.
Lewis
,
C. D.
and
Laemmli
,
U. K.
(
1982
).
Higher order metaphase chromosome structure: evidence for metalloprotein interactions
.
Cell
29
,
171
181
.
McDonald
,
K.
,
Pickett-Hbaps
,
J. D.
,
McIntosh
,
J. R
and
Tippit
,
D.
(
1977
).
On the mechanism of anaphase spindle elongation in Diatoma vulgare
.
J. Cell Biol
.
74
,
377
388
.
McIntosh
,
J. R
and
Landis
,
S. C.
(
1971
).
The distribution of spindle microtubules during mitosis in cultured human cells
.
J. Cell Biol
.
49
,
468
497
.
McIntosh
,
J. R
,
McDonald
,
K. L.
,
Edwards
,
M. K.
and
Ross
,
B. M.
(
1979
).
Three-dimensional structure of the central mitotic spindle of Diatoma vulgare
.
J. Cell Biol
.
83
,
428
442
.
McLean
,
I. W.
and
Nakane
,
P. K.
(
1974
).
Periodate-lysine- paraformaldehyde fixative: a new fixative for immunoelectron microscopy
.
Histochem Cytochem
.
22
,
1077
1083
.
Nicklas
,
R B.
(
1989
).
The motor for poleward chromosome movement in anaphase is in or near the kinetochore
.
J. Cell Biol
.
109
,
2245
2255
.
Niblow
,
C.
,
Sellitto
,
C.
,
Kuriyama
,
R
and
McIntosh
,
J. R
(
1990
).
A monoclonal antibody to a mitotic microtubule-associated protein blocks mitotic progression
.
J. Cell Biol
.
Ill
,
511
-
522
.
Nunnally
,
M. H.
,
D’Angelo
,
J. M.
and
Craig
,
S. W.
(
1980
).
Filamin concentration in cleavage furrow and midbody region: frequency of occurrence compared with that of alpha-actinin and myosin
.
J. Cell Biol
.
87
,
219
226
.
Oppenheim
,
D. S.
,
Hauschka
,
B. T.
and
McIntosh
,
J. R
(
1973
).
Anaphase motions in dilute colchicine. Evidence of two phases in chromosome segregation
.
Expl Cell Res
.
79
,
95
105
.
Pankov
,
R
,
Lemieux
,
M.
and
Hancock
,
R
(
1990
).
An antigen located in the kinetochore region in metaphase and on polar microtubule ends in the midbody region in anaphase, characterized using a monoclonal antibody
.
Chromosoma
99
,
95
101
.
Ptahr
,
C. M.
,
Coue
,
M.
,
Gbissom
,
P. M.
,
Hays
,
T. S.
,
Porter
,
M. E.
and
McIntosh
,
J. R
(
1990
).
Cytoplasmic dynein is localized to kinetochores during mitosis
.
Nature
345
,
263
265
.
Rappaport
,
R
(
1986
).
Establishment of the mechanism of cytokinesis in animal cells
.
Int. Rev. Cytol
.
105
,
245
281
.
Rieder
,
C. L.
and
Alexander
,
S. P.
(
1989
).
The attachment of chromosomes to the mitotic spindle and the production of aneuploidy in newt lung cells
. In
Mechanisms of Chromosome Distribution and Aneuploidy
(ed.
M. A.
Resnick
and
B. K.
Vig
), pp.
185
-
194
. New York: Alan R Lias.
Sanger
,
J. M.
,
Mittal
,
B.
,
Dome
,
J. S.
and
Sanger
,
J. W.
(
1989
).
Analysis of cell division using fluorescently labeled actin and myosin in living PtK2 cells
.
Cell Motil. Cytoskel
.
14
,
201
219
.
Schroeder
,
T. E.
(
1973
).
Actin in dividing cells: Contractile ring filaments bind heavy meromyosin
.
Proc. natn. Acad. Sci. U.S.A
.
70
,
1688
1692
.
Sbllttto
,
C.
and
Kurjyama
,
R.
(
1988
).
Distribution of a matrix component of the midbody during the cell cycle of Chinese hamatar ovary cells
.
J. Cell Biol
.
106
,
431
439
.
Stiver
,
E. R.
,
Wordeman
,
L.
,
Schroer
,
T. A.
and
Sheetz
,
M. P.
(
1990
).
Localization of cytoplasmic dynein to mitotic spindles and kinetochores
.
Nature
345
,
266
268
.
Taylor
,
E. W.
(
1959
).
Dynamics of spindle formation and its inhibition by chemicals
.
J. biophys. biochem. Cytol
.
6
,
193
196
.
White
,
J. G.
,
Amos
,
W. B.
and
Fordham
,
M.
(
1987
).
An evaluation of confocal versus conventional imaging of biological structures by fluorescence light microscopy
.
J. Cell Biol
.
105
,
41
48
.
Wordeman
,
L.
and
Cande
,
W. Z.
(
1987
).
Reactivation of spindle elongation in vitro is correlated with the phosphorylation of a 205 kd spindle-associated protein
.
Cell
50
,
535
543
.
Wordeman
,
L.
,
Masuda
,
H.
and
Cande
,
W. Z.
(
1989
).
Distribution of a thiophosphorylated spindle midzone antigen during spindle reactivation in vitro
.
J. Cell Sci
.
93
,
279
285
.
Yin
,
T. J.
,
Compton
,
D. A.
,
Earnshaw
,
W. C.
and
Cleveland
,
D. W.
(
1991
).
CENP-E, a human centromere associated protein released from chromosomes at the onset of anaphase
.
EMBO J. (in press)
.
Zirkle
,
R. E.
(
1970
).
UV-microbeam irradiation of newt-cell cytoplasm: Spindle destruction, false anaphase, and delay of true anaphase
.
Rad. Res
.
41
,
516
537
.