Tension, microtubule rearrangements, and the proper distribution of chromosomes in mitosis

Chromosoma (Berl) (1989) 98:33-39 CHROMOSOMA Springer-Verlag 1989 Tension, microtubule rearrangements, and the proper distribution of chromosomes in mitosis Jeffrey G. Ault* and R. Bruce Nicklas Department of Zoology, Duke University, Durham, NC 27706, USA Abstract. The basis for stable versus unstable kinetochore orientation was investigated by a correlated living-cell/ultrastructural study of grasshopper spermatocytes. Mal-oriented bivalents having both kinetochores oriented to one spindle pole were induced by micromanipulation. Such malorientations are stable while the bivalent is subject to tension applied by micromanipulation but unstable after tension is released. Unstable bivalents always reorient with movement of one kinetochore toward the opposite pole. Microtubules associated with stably oriented bivalents, whether they are mal-oriented or in normal bipolar orientation, are arranged in orderly parallel bundles running from each kinetochore toward the pole. Similar orderly kinetochore microtubule arrangements characterize mal-oriented bivalents fixed just after release of tension. A significantly different microtubule arrangement is found only some time after tension release, when kinetochore movement is evident. The microtubules of a reorienting kinetochore always include a small number of microtubules running toward the pole toward which the kinetochore was moving at the time of fixation. All other microtubules associated with such a moving kinetochore appear to have lost their anchorage to the original pole and to be dragged passively as the kinetochore proceeds to the other pole. Thus, the stable anchorage of kinetochore microtubules to the spindle is associated with tension force and unstable anchorage with the absence of tension. The effect of tension is readily explained if force production and anchorage are both produced by mitotic motors, which link microtubules to the spindle as they generate tension forces. Introduction Bipolar orientation of the two kinetochores of each chromosome in a cell is required for proper chromosome distribution to daughter cells in mitosis. An important feature of the dynamics of the spindle is that mistakes in kinetochore orientation, mal-orientations, can be corrected. Malorientations often occur during prometaphase but are not stable, and one or both kinetochores usually reorient until a stable bipolar orientation is achieved. The tension produced in a bipolar orientation appears to be important for * Present address: Department of Genetics, University of Hawaii, Honolulu, HI 96822, USA Offprint requests to: 1.G. Ault stability. Tension has been shown to stabilize normally unstable orientations of bivalents (pairs of chromosomes in meiosis) (Nicklas and Koch 1969; Henderson and Koch 1970). For example, a bivalent in a living grasshopper spermatocyte can be micromanipulated to cause both kinetochores to orient to one pole. This mal-orientation persists, i.e., it is stable, only if the bivalent is subjected to tension by either pulling on it with a microneedle (Nicklas and Koch 1969) or by interlocking it with a bivalent similarly oriented to the opposite pole so that the two mal-oriented bivalents will pull against each other (Henderson and Koch 1970). If tension is released before anaphase, reorientation occurs, leading to a stable bipolar orientation. Microtubules might be expected to have a role in determining the degree of stability of different chromosome orientations. Spindle microtubules are certainly mechanically involved in chromosome behavior, if only because kinetochore microtubules are responsible for the physical attachment of chromosomes to the spindle (Nicklas et al. 1982) and, so influence the direction of chromosome movement (Church and Lin 1985; Nicklas and Kubai 1985). In this paper, we compare kinetochore microtubule arrangements of mal-oriented bivalents that are under tension, and hence have stable orientations, and those that are not under tension and so are likely to reorient. Materials and methods Spermatocytes of the grasshoppers M elanoplus differentia/is (Thomas) and Melanoplus sanguinipes (Fabricius) were used. The methods used in maintaining laboratory colonies were as previously described (Ault 1984). The procedures used for the culture, observation, micromanipulation, and electron microscopy of single cells as well as for constructing computer-generated three-dimensional stereo pairs of chromosomes and surrounding microtubules have been described elsewhere (Nicklas et al. 1979, 1982; Nicklas and Kubai 1985, and references therein). The process of inducing chromosome mal-orientation by micromanipulation is illustrated in Figure 1, and the experimental design was as follows. One of the kinetochores of a bivalent in normal bipolar orientation was detached from the spindle using a microneedle and then moved so that it faced the same pole as the other still attached kinetochore. That this repositioned kinetochore was attached to the pole it faced was verified by pulling on the bivalent with the microneedle. Once a mal-orientation was estab-

34 Fig. 1. Photomicrographs from a cinematographic record of an induced mal-oriented bivalent of Melanoplus differentia/is fixed while it was under tension. Arrows indicate the kinetochores of the bivalent: before detachment (0.0 min), after the upper kinetochore was detached from the spindle and moved so as to face the lower pole (3.0 min; the detached kinetochore is on the left), after mal-orientation to the lower pole and after the bivalent was pulled upward with a micromanipulation needle, producing tension (7.4 min), and at fixation (8.2 min). The bivalent was under tension for 3.0 min. Bar represents 10 1.1m lished, it was stabilized by maintaining the pull on the needle so as to put the entire bivalent under tension. The cell was then fixed either while the bivalent remained under tension or at different times after tension was released. The cell was processed for electron microscopy, and the arrangement of kinetochore microtubules of the bivalent was determined by reconstruction from serial thin sections and displayed using computer-generated three-dimensional stereo pair drawings. Four cells were fixed while the bivalent was under tension and six were fixed after tension was released. Of the latter six cells, two were fixed before the start of reorientation and four during reorientation, that is, while both kinetochores remained facing a single pole (two cells) or after one kinetochore began to move out of this relationship to the pole (four cells). The results are considered without regard to species because no differences in chromosome behavior ofbivalents or spindle fine structure were observed either in the present study or in previous investigations (Nicklas and Kubai 1985). For each cell illustrated the species is given in the figure legend. ' f Results Background Detachment of a kinetochore from the spindle in Melanoplus spermatocytes always occurs at or near the kinetochore, leaving at most one short microtubule associated with the kinetochore but usually none (Nicklas and Kubai 1985). This is an observation that assures us that kinetochore microtubules found later are not remnants of the previous spindle attachment of the kinetochore, but are newly derived, either by nucleation of new microtubules at the kinetochore or by capture of microtubules that pre-existed in the spindle as non-kinetochore microtubules. Note that clean detachment is known only for chromosomes in stable bipolar orientations. Chromosomes detached when in unstable orientations might well retain some kinetochore microtubules if those micro tubules are unanchored to the spindle. Microtubules associated with chromosomes under tension The cell shown in Figures 1 and 2 was fixed while the mal-oriented bivalent was under tension. Most of the kine- Fig. 2. Electron micrograph of the mal-oriented bivalent in Figure 1. Arrow points to the hole created by the micromanipulation needle. Bar represents 2 1.1m tochore microtubules associated with the mal-oriented chromosome run in parallel toward one pole, ending close to it (Fig. 3); none run toward the other pole, the one from which a kinetochore was detached in order to effect malorientation. Two of the microtubules associated with the previously detached kinetochore are unusual in that they run perpendicular to the spindle (Fig. 3). These may have resulted from detachment, perhaps by nucleating from small (less than 0.05 J.lm) unrecognizable microtubule fragments that remained associated with the kinetochore after detachment. Such fragments would not be recognizable (Nicklas and Kubai 1985) and conceivably would not remain in a parallel arrangement after detachment. Microtubules associated with chromosomes not under tension Two induced mal-oriented bivalents were released from tension and fixed before reorientation began. In one case, both

35 Fig. 3. Three-dimensional reconstruction (stereo pair) of the mal-oriented bivalent in Figure 1. Distal ends of kinetochore microtubules are boxed. Two kinetochore microtubules (arrowheads) of the previously detached kinetochore, on the left, are shown running perpendicular to the rest of the spindle fiber. Bar represents 211m kinetochores moved to within a few microns of the pole immediately after tension was released and then remained nearly stationary until fixation. This is the usual behavior of a bivalent in unipolar mal-orientation and is expected to be followed by reorientation of one kinetochore or the other. Both spindle fibers of this bivalent were similar to those of mal-oriented bivalents under tension, being composed primarily of parallel bundles of microtubules running toward a pole. Only one of the kinetochore microtubules associated with the previously detached kinetochore was unusual in that it made an angle of approximately 45 degrees with respect to the remaining microtubules of the fiber. In another case, the mal-oriented bivalent behaved differently (Fig. 4). Its two kinetochores were not at equal distances from the pole while tension was applied. After tension was released, the more distant, previously detached kinetochore (left in Fig. 4) moved toward the pole, passing the other kinetochore. The latter kinetochore moved only slightly poleward, about 1 jlm. In the course of this movement, the second kinetochore also turned slightly so as to face more toward the opposite pole (Fig. 4). The microtubule arrangements of the two kinetochores were different (Fig. 5). Neither kinetochore had microtubules running toward the opposite pole. The kinetochore microtubules of the previously detached kinetochore, the one that moved most, were skewed, not parallel to each other like those associated with bivalents under tension. In contrast, the kinetochore microtubules of the kinetochore that was not previously detached, that moved less and that turned slightly are more parallel with one another. When the spatial relationship of the microtubule bundles of the two kinetochores is compared using stereo pairs, it is evident that the microtubule bundle of each runs in a slightly different direction. Only the bundle associated with the previously detached kinetochore runs directly toward the pole. The microtubule bundle of the other kinetochore runs to a site slightly away from the pole, with the exception of a few curved microtubules (e.g., Fig. 5, arrowhead). It appears as if the spindle fiber resisted the poleward movement of the other kinetochore, which bowed a few kinetochore microtubules, but turned most away from the pole. Reorientation A kinetochore reorients by turning away from one pole and actively moving toward the other pole (Figs. 6--9). As was previously mentioned, turning of a kinetochore may sometimes occur passively as the result of its physical connection to a moving partner kinetochore (Figs. 4, 14.4 min print, and 8, 6.6 min print). Active poleward movement of reorienting kinetochores (Figs. 6, 6.0 and 10.7 min prints, and 8, 8.2 min print) is always correlated with the presence of at least a few kinetochore microtubules running in the direction of movement (Figs. 7 and 9). However, kinetochore microtubules are usually found running toward both poles- even a kinetochore that had turned completely from one pole to the other (Fig. 8) had one short microtubule running back toward the pole it originally faced (Fig. 9). One cell in particular illustrates the complexity involved in stability during reorientation. In this cell, the reorienting kinetochore was fixed midway in its turn toward the upper pole (Fig. 6). It is interesting that the kinetochore microtubule extending farthest toward the upper pole passes through the kinetochore and ends close to the lower pole (Fig. 7). All kinetochore microtubules running toward the lower pole, the pole to which both kinetochores were previously oriented, were tracked to their ends or until they were lost in the thicket of microtubules very close to the pole (Fig. 7). These microtubules of the reorienting kinetochore appeared to be remnants of the previous orientation: their lengths are approximately equal to the kinetochore-topole distance of the previous orientation (Fig. 6, 6.0 min print) and they fan out from the kinetochore as if, after losing their anchorage to the spindle near the lower pole, they were dragged along by the movement of the kinetochore toward the upper pole. It is noteworthy that the other kinetochore also displayed a disorderly arrangement of microtubules. Only two kinetochore microtubules ended near the pole (Fig. 7, arrowheads). Most of the rest were short microtubules pointing in various directions. The cinematographic record shows this kinetochore to be turning away from the lower pole (Fig. 6), as if the orientation was no longer stable. This movement may be a passive result of the reorienting movement of the other kinetochore of the

36 Fig. 4. Photomicrographs of an induced mal-oriented bivalent of Me/anop/us differentia/is fixed after it was released from tension. Arrows indicate the kinetochores of the bivalent: before detachment (0.0 min), just after the bivalent was released from tension (5. 7 min; the kinetochore that was detached from the spindle and induced to reorient is on the left), while the left kinetochore was moving toward the lower pole (10.6 min), and at fixation when the left kinetochore was closer to the lower pole and the right one passively swung upward (14.4 min) as the other kinetochore moved. The bivalent was under tension for 2.1 min. Bar represents 10 Jlm Fig. 5. Three-dimensional reconstruction (stereo pair) of the mal-oriented bivalent in Figure 4. The bivalent was fixed 8.8 min after tension was released. Distal ends of kinetochore microtubules are boxed. The kinetochore microtubules of the left kinetochore which was moving toward the pole are straight and, though this bundle of microtubules is splayed, all microtubules run in general toward the pole. A kinetochore microtubule (arrowhead) of the stationary kinetochore on the right bends toward the pole. Bar represents 2 Jlm bivalent or it may indicate that both kinetochores of the bivalent are actively reorienting. The turning kinetochore was associated with two short kinetochore microtubules running in the general direction of the upper pole. A curvilinear kinetochore microtubule (Fig. 7, double arrowhead) is particularly suggestive that this kinetochore was reorienting and so had lost its anchorage to the spindle. When the cell was fixed, the distal end of the curved microtubule was a long way from the lower pole (Fig. 7), but this microtubule is long enough (5.4 11m) to have extended all the way to the pole originally (e.g., at 6.4-8.6 min, Fig. 6). The observations just presented suggest that orderly arrangement of kinetochore micro tubules is not an immediate consequence of the establishment of bipolar orientation. A similar conclusion can be drawn from observations on another cell (Figs. 8 and 9). After the reorienting kinetochore in Figure 8 completed its turn from one pole to the other, its kinetochore microtubules were still disordered as were those of the non-reorienting kinetochore (Fig. 9). Each of the kinetochores was associated with a kinetochore microtubule running toward the pole opposite to the one with which it was oriented (Fig. 9). That is, as in the previous case, after tension was released, both kinetochores exhibited a disorderly arrangement of kinetochore microtubules, a circumstance that suggests the instability of the orientation of both kinetochores. It appears that order is only regained after the bipolar orientation has been maintained under tension for some time. Discussion An unstably oriented kinetochore will reorient until a stable orientation is achieved. Reorientation involves the loss of the unstable orientation and establishment of a new orientation to the opposite pole. How new orientations occur may in part be answered by the elegant work of Mitchison and Kirschner (1985). They have shown that in vitro kinetochores have the ability to capture pre-existing polar microtubules. Thus, a reorienting kinetochore may capture microtubules emanating from a pole it initially does not face if these microtubules happen to be close enough to the face of the kinetochore (Ault 1986). As was observed in this study, a kinetochore may be turned to face more directly toward the microtubules running from the more distant pole simply as a result of the movement of the second kinetochore of the bivalent. Thus, the structure of the bivalent can promote bipolar orientation (Ostergren 1951; Nicklas 1967) even when both of its kinetochores originally orient to the same pole. Evidently, during reorientation kinetochore microtu-

37 Fig. 6. Photomicrographs showing kinetochore reorientation of a large bivalent of Melanoplus diff'erentialis. Both kinetochores (arrows) of the bivalent were in focus before detachment (0.0 min) and while the bivalent was under tension ( 4.6 min; the previously detached kinetochore is on the right). Tension was maintained for 1.7 min and was released at 4.7 min. While the reorienting kinetochore was moving toward the upper pole (6.0 min and 10.7 min), the other kinetochore turned away from the lower pole (6.4 min, 8.6 min, 10.2 min, and 11.0 min). Fixation occurred at 11.0 min. Bar represents 10 Jlm Fig. 7. Three-dimensional reconstruction (stereo pair) of the reorienting bivalent in Figure 6. The bivalent was fixed 6.3 min after tension was released. Distal ends of the kinetochore microtubules are boxed. The kinetochore microtubule of the reorienting kinetochore that extended farthest toward the upper pole (arrows and boxes at both ends) ran through the kinetochore and ended apparently close to the lower pole but actually far above it, as seen in stereo view. Two kinetochore microtubules of the nonreorienting kinetochore that ended near the lower pole (centriole and pericentriolar material are outlined) are indicated by single arrowheads and a curvilinear one by a double arrowhead. Bar represents 2 Jlm buies are not all equally effective in kinetochore movement. Reorienting kinetochores usually have microtubules running toward both poles (Church and Lin 1985; Nicklas and Kubai 1985; Steffen 1986; this paper), but the new orientation which begins with very few microtubules (Nicklas and Kubai 1985) prevails, and the kinetochore can turn away from the pole to which it was originally unstably oriented. Apparently, the larger number of microtubules involved in the original unstable orientation cannot resist the movement of the kinetochore toward the other pole. Perhaps this is so because the microtubules of the unstable orientation are no longer anchored to the pole so that the motors associated with the few microtubules running toward the other pole are sufficient to move both the chromosome and the unanchored microtubules. The subset of kinetochore microtubules actively involved in kinetochore movement are thought to be those that run from the kinetochore toward the pole, parallel to the direction of the movement (Church and Lin 1985; Nicklas and Kubai 1985). In certain well-fixed materials,

38 Fig. 8. Kinetochore reorientation of an induced mal-oriented bivalent of Melanoplus sanguinipes after it was released from tension. Arrows indicate the kinetochores of the bivalent: before detachment (0.0 min), while the bivalent was under tension (5.0 min, the previously detached kinetochore is on the right), during reorientation (6.6 min, tension was released at 5.2 min), and at fixation (8.2 min). The bivalent was under tension for 3.0 min. Bar represents 10 ~m Fig. 9. Three-dimensional reconstruction (stereo pair) of the reorienting bivalent in Figure 8. The bivalent was fixed 3.0 min after tension was released. Distal ends of most kinetochore microtubules are boxed. A short kinetochore microtubule (single arrowhead) of the reorienting kinetochore extends downward, toward the pole to which the kinetochore was previously oriented (this microtubule has the proximal end boxed). A kinetochore microtubule (double arrowhead) of the non-reorienting kinetochore also extends in the opposite direction from its fellows ~ toward the upper pole. Bar represents 2 ~m these kinetochore microtubules often extend all the way to the pole (Church and Lin 1985; Nicklas and Kubai 1985). Short kinetochore microtubules and those that are skewed are thought to be excluded from this subset of active microtubules, even though skewed microtubules are more common with moving kinetochores (Steffen 1986). In Figure 5, the kinetochore microtubules of the moving kinetochore are more skewed than those of the stationary kinetochore. Skewed kinetochore microtubules appear to be microtubules which are not anchored to the pole and therefore can be passively displaced as the kinetochore moves. Indeed, only one kinetochore microtubule is needed for kinetochore movement (Church and Lin 1985; Nicklas and Kubai 1985) and those kinetochore microtubules not directly involved would be either passively pushed or pulled as movement proceeds. Kinetochore microtubules that are not anchored to a pole may be subject to depolymerization. The set of microtubules involved in unstable orientations eventually dissociates from reorienting kinetochores. By the time a reorienting kinetochore has completed its turn toward the opposite pole (Fig. 8), only short kinetochore microtubules running back toward the pole to which the kinetochore was unstably oriented are observed (Fig. 9). If kinetochores associate with microtubules by capturing microtubules originating from the poles, many of the observed short microtubules may be remnants of microtubules that have lost their anchorage to a pole and are in the process of depolymerizing. An orientation is stable and reorientation cannot occur as long as kinetochore microtubules remain anchored to a pole. Anchorage to a pole is evident during mitotic force production (Nicklas and Staehly 1967; Nicklas et al. 1982;

39 Nicklas 1983; Nicklas and K ubai 1985) and bivalents under tension obviously remain under tension only because both force production and anchorage to a pole persist. Tension may stabilize an orientation by keeping it engaged in force production. Chromosome behavior within the spindle would be easily explained if force production and mechanical linkage were one and the same (Nicklas and Koch 1969; Nicklas and K ubai 1985). An orientation that no longer engages in force production would lose its anchorage to a pole, and the kinetochore would be free to interact with and move toward the other pole. A kinetochore in such unstable orientation might reorient or it might capture yet other microtubules from the same pole and thus remain oriented to it. In either case the new orientation would be anchored to the pole as long as it was engaged in force production. Several consecutive unstable orientations to the same pole would appear as a stable one - an example of this could be univalents that move to a pole and remain there. Such consecutive unstable orientations to the same pole would be rare with bivalents because a change in the position of the kinetochore would occur repeatedly as a passive result of the movement of the other kinetochore and this would promote eventual reorientation. Normally, only bipolar orientations would be stable. Here the opposing forces persist and so does the associated anchorage. Thus, motors that simultaneously pull on kinetochore microtubules and anchor them to the spindle provide a straightforward explanation for the stabilizing effect of tension. Tension stabilization of the poleward anchorage is not so readily explained if the kinetochore itself is the site of force production, which may be true in anaphase (e.g., Gorbsky et a!. 1987). The site and nature of force producers in prometaphase remains an open question. The stable unipolar orientation of the sex chromosome in M. sanguinipes (Ault 1984, 1986) complies with the above interpretation. This orientation is unusual in that it is stable without tension. Orientation stability is attributed to the large number of non-kinetochore micro tubules that impinge on the sex chromosome. These microtubules apparently prevent reorientation by constraining the chromosome so that it cannot turn toward the opposite pole (Ault 1986). In this regard, these non-kinetochore microtubules are like kinetochore microtubules engaged in force production - both maintain the anchorage to the pole. Acknowledgements. We especially thank Angela Brown, Andrea McKibbins, and Suzanne Ward for their skilled technical assistance, and Dr. Donna Kubai for her advice on electron microscopy of single cells and for her other helpful suggestions. This investiga- tion was supported in part by a Charles W. Hargitt Research Fellowship in Cell Biology from Duke University and by grant GM 13745 from the Institute of General Medical Sciences, NIH. References Ault JG (1984) Unipolar orientation stability of the sex univalent in the grasshopper (Melanoplus sanguinipes). Chromosoma 89:201-205 Ault JG (1986) Stable versus unstable orientations of sex chromosomes in two grasshopper species. Chromosoma 93:298-304 Church K, Lin HP (1985) Kinetochore microtubules and chromosome movement during prometaphase in Drosophila metanagaster spermatocytes studied in life and with the electron microscope. Chromosoma 92:273-282 Gorbsky GJ, Sammak PJ, Borisy GG (1987) Chromosomes move poleward in anaphase along stationary microtubules that coordinately disassemble from their kinetochore ends. 1 Cell Bioi 104:9-18 Henderson SA, Koch CA (1970) Co-orientation stability by physical tension: a demonstration with experimentally interlocked bivalents. Chromosoma 29:207-216 Mitchison TJ, Kirschner MW (1985) Properties of the kinetochore in vitro. II. Microtubule capture and A TP-dependent translocation. J Cell Biol101 :766-777 Nicklas RB (1967) Chromosome micromanipulation. II. Induced reorientation and the experimental control of segregation in meiosis. Chromosoma 21:17-50 Nicklas RB (1983) Measurements of the force produced by the mitotic spindle in anaphase. 1 Cell Bioi 97:542-548 Nicklas RB, Koch CA (1969) Chromosome micromanipulation. III. Spindle fiber tension and the reorientation of mal-oriented chromosomes. J Cell Bioi 43:40-50 Nicklas RB, Kubai DF (1985) Microtubules, chromosome movement, and reorientation after chromosomes are detached from the spindle by micromanipulation. Chromosoma 92:313-324 Nicklas RB, Staehly CA (1967) Chromosome micromanipulation. I. The mechanics of chromosome attachment to the spindle. Chromosoma 21 : 1-16 Nicklas RB, Brinkley BR, Pepper DA, Kubai DF, Rickards GK ( 1979) Electron microscopy of spermatocytes previously studied in life: methods and some observations on micromanipulated chromosomes. J Cell Sci 35: 87-104 Nicklas RB, Kubai DF, Hays TS (1982) Spindle microtubules and their mechanical associations after micromanipulation in anaphase.j Cell Biol95:91-104 Ostergren G (1951) The mechanism of co-orientation in bivalents and multivalents. Hereditas 37: 85-156 Steffen W (1986) Relationship between the arrangement of microtubules and chromosome behaviour of syntelic autosomal univalents during prometaphase in crane fly spermatocytes. Chromosoma 94:412-418 Received November 10, 1988 Accepted by J.H. Taylor