Abnormal microtubule deployment during defective macronuclear division in a Paramecium mutant

Cell division requires precise spatial and temporal adjustment between nuclear and cytoplasmic division. Asymmetric divisions yielding cell products of unequal size, such as occur during budding in yeast or during ‘oblique’ spindle construction and associated spiral cleavage in various invertebrate blastomeres (Grant, 1978), provide striking evidence for specific control of the positioning of dividing nuclei within cells. However, the mechanisms by which such control is effected remain obscure. This problem can be approached with Parameciumby studying mutants in which cytoplasmic division proceeds normally although nuclei are aberrantly positioned and an unequal partitioning of nuclear material between daughter cells results. Most of these Parameciummutants belong to a particular class that consistently display 2 main phenotypic abnormalities: both nuclear and trichocyst positioning are abnormal (Beisson & Rossignol, 1975; Ruiz, Adoutte, Rossignol & Beisson, 1976). Trichocysts are secretory vesicles that normally attach to certain plasma membrane ‘docking sites’ where they remain until discharge is stimulated. In these mutants the trichocysts never attach to the plasma membrane and the dividing macronucleus never reaches the normal dorsal subcortical position described in the accompanying paper (Tucker, Beisson, Roche & Cohen, 1980).

This paper analyses defective macronuclear division during binary fission in the mutant tam 8(a representative of the class of mutants discussed above) in terms of spatio-temporal correlations between abnormal positioning, defective elongation, abnormal and irregular shaping, and abnormal nucleoplasmic microtubule deployment, for the dividing macronucleus. In addition, the arrangement of microtubules in the tips of abnormally positioned trichocysts and the separation spindles of dividing micronuclei (that become unequally distributed between daughter organisms) are described. These examinations appear to rule out the possibility that the tam 8mutation exerts its influence by interfering with microtubule assembly in a marked fashion. It is argued that the pleiotropic effects of the tam 8mutation indicate that considerable spatial control of macronuclear division is exercised by the cell cortex.

The mutant tam 8(Beisson & Rossignol, 1975) was isolated after nitrosoguanidine treatment of stock <14-2of P. tetraureliaand first screened on the basis of its lack of trichocyst discharge. The trichocysts are normal in shape but fail to attach to the plasma membrane and do not exhibit the saltatory motion (Aufderheide, 1978) that is performed by trichocysts prior to plasma membrane attachment in wild-type organisms.

Paramecia were cultured using procedures described by Sonnebom (1970) at 27 °C in Scotch Grass infusion or cerophyl infusion supplemented with β-sitosterol (0·4 μg/ml). Infusions were inoculated with Klebsiella pneumoniae24 h before inoculation with Paramecium.

Organisms were isolated from log-phase cultures, transferred to microscope slides, fixed by adding one drop of Dippell’s (1955) stain and then examined using bright-field microscopy. These preparations were not covered with coverslips, so that organisms could be gently rolled over in the drops by blowing a jet of air at them to ascertain the location of the macronucleus (central or subcortical). Cell lengths, macronuclear lengths, and the lengths of the portions of macronuclei situated in proters (in cases where a cleavage furrow was visible) were measured using an ocular micrometer. Micronuclei and micronuclear spindles were examined using phase-contrast microscopy in organisms stained with Azure A (Dalamater, 1951).

The procedures used for electron microscopy were those described by Tucker et al. (1980).

A detailed comparison of macronuclear division in tam 8with that which occurs in stock dj-2of the wild-type organism (see Tucker et al. 1980) from which this mutant is derived has been undertaken. The division stages referred to below have been distinguished, using the criteria described by Tucker et al. (1980), on the basis of changes in the lengths and shapes of organisms and the progress of cleavage furrowing. These cortex-associated changes all appear to proceed normally in tam 8. The fine structure of 4 tam 8organisms has been examined (2 at stage 4 and 1 each at stages 5 and 6) to ascertain the spatial organization of microtubules in the macronucleus and monitor changes in its cross-sectional area during the period of most marked nuclear elongation.

In an interfission tam 8organism the macronucleus is more or less spherical (Fig. 1 D) and its intracellular position varies from one individual to another. It can apparently be located almost anywhere inside the cell. The macronucleus of an interfission wild-type organism is usually somewhat elongate and spheroidal and is located alongside the oral apparatus (Fig. 1 A).

Both tam 8and wild-type organisms increase in length during binary fission. There is little time for cell length increase during the rapid (Tucker et al. 1980) elongation of wild-type macronuclei. In terms of its temporal correlation with cell elongation, the elongation of tam 8macronuclei takes place more slowly than that of wild-type nuclei, although the time taken for cell division is similar in tam 8and wild-type organisms (about 20 min at 27 °C). In addition, elongation is less extensive and final lengths are more variable (Fig. 2). In some stage 6 tam 8organisms macronuclei had undertaken almost no elongation. The elongation factor (ratio of mean length at late division stages to mean length at early division stages is 3 ·6 for wild-type organisms and 2·3 for tam 8organisms − 64% of the wild-type value).

During stage 2 the macronucleus fails to migrate towards the anterior dorsal cortical region. Examination of stained organisms while they are rolled over (see Materials and methods) reveals that during stage 3 macronuclei remain more or less centrally positioned close to an organism’s longitudinal axis and that they do not adopt a flattened elliptical shape against the dorsal cortex as is the case in stage 3 wild-type organisms. Nor does a tam 8macronucleus belatedly take up a dorsal position in close contact with the cortex during stage 4 (Fig. 4) or subsequently (Fig. 1E). Although dorsoventral flattening does not occur at stage 3, such flattening was found along a portion of a stage 4 nucleus (Fig. 5). Microtubules were slightly concentrated at each extremity of the flattened cross-sectional profiles of this portion of the nucleus, but not at the sides of portions exhibiting a more or less circular cross-sectional profile (Fig. 4). These microtubules were less numerous and less closely packed together than those in the marginal bands of elliptical stage 3 wild-type macronuclei.

Cross-sectional areas and the shapes of their cross-sectional profiles vary considerably along the lengths of elongating stage 4 tam 8macronuclei (Figs. 4, 5). Such marked variations do not occur in wild-type organisms. Furthermore, very pronounced longitudinally oriented folds in the nuclear envelope were present along some portions of both of the stage 4 nuclei for which sequences of cross-sections were examined (Fig. 6); folding on this scale was never observed for wild-type nuclei.

Considerable numbers of microtubules have assembled in the nucleoplasm of elongating nuclei by stage 4. As in wild-type organisms, most of them (79–92%, Table 1) are longitudinally oriented but their distribution is somewhat different. The marked peripheral concentration of microtubules at stage 4 in wild-type macronuclei is not apparent in tam 8nuclei (Fig. 7). At stage 4 there are 43–67% internal microtubules (microtubules situated more than 1 μm from the nearest portion of the nuclear envelope) (Table 1) compared with 15–48% in wild-type organisms. As for wild-type macronuclei, some of the peripheral microtubules are coated with dense granules. Whether tubules are also attached to nucleoli was not ascertained; extensive sequences of longitudinal thin sections of elongating tam 8macronuclei were not prepared.

Correlated with the slower, less extensive, and more variable elongation of tam 8macronuclei compared with those of wild-type organisms are the greater and more variable cross-sectional areas of tam 8macronuclei, and the substantially greater numbers of microtubule profiles per nuclear cross-section (N), during the period (stages 4–6) of most marked macronuclear elongation (Table 1, Figs. 7, 8). For example, at stage 4 the cross-sectional areas of tam 8nuclei lie in the range 65–122 μm²with a range of 779–1134 for N, while in wild-type organisms these values are 25–58 μm²and 347–565, respectively. The value of Ndecreases as macronuclei elongate in both tam 8and wild-type organisms. Decreases during stage 5 and 6 (after most of the elongation of wild-type nuclei is completed) may be largely a consequence of microtubule breakdown, since macronuclei contain few microtubules as fission is completed (Fig. 8). One of the stage 4 tam 8macronuclei had a much higher number of microtubules per cross-section than the maximum value recorded for wild-type nuclei (Fig. 8). It is not clear whether this is an indication that a greater than normal number of microtubules assemble in tam 8macronuclei or if Nrises briefly to such high values in wild-type organisms at a stage (between 2 and 3) that has not been examined ultrastructurally.

During stages 5 and 6 the tam 8macronucleus usually fails to narrow and then divide at its mid-point; it is pinched in 2 by the advancing cleavage furrow. The constricted portion forms a ‘nuclear bridge’ (Fig. IF). Occasionally, however, a macronucleus may undergo a more or less central thinning out, similar to that which occurs in wild-type organisms (Fig. 1c) before the cleavage furrow is sufficiently advanced to exert a pinching action. Furthermore, the macronucleus is often not positioned medially with respect to the poles of an organism, so that the cleavage furrow plane does not coincide with the mid-point of the nucleus. As a result segregation of macronuclear material to daughter organisms can be very unequal (Fig. 9). In extreme cases of such ‘slippage’ one of the daughter organisms does not receive a portion of a macronucleus and an ‘amac’ cell results. Considerable numbers of amac cells are always present in log-phase cultures of tam 8.

Since the tam 8lesion interferes with microtubule organization in the defectively dividing macronucleus, does it also perturb microtubule organization and activity in the cell’s other microtubular systems?

Ciliary beating and ultrastructure appear to be normal. Although trichocysts do not exhibit saltatory movement (Aufderheide, 1978) and their tips do not attach to the plasma membrane in tam 8, microtubules seem to be arranged (Fig. 3) in a manner identical with that described for trichocyst tips in P. caudatum(Bannister, 1972). Each micronucleus elongates extensively during binary fission in a wild-type organism as it produces a long (up to 80 μm) intranuclear microtubular separation spindle (Jurand & Selman, 1970; Stevenson & Lloyd, 1971). This separation spindle is very unusual; it consists mainly of tubules with diameters of up to 32 nm in thin-sectioned material (Tucker, 1979). Microtubules in the macronucleus and cytoplasm have diameters of about 24 nm. This peculiarity is also manifested by the separation spindles of tam 8micronuclei (Fig. 10) which, unlike tam 8macronuclei, do not differ in any marked way from those of wild-type organisms (in terms of cross-sectional area or microtubule number, spacing, and arrangement). Nor did examination of organisms stained with Azure A reveal any delay or reduction in the elongation of micronuclear separation spindles. However, segregation of daughter micronuclei is substantially impaired in tam 8organisms; 56% of the interfission organisms examined contained more, or less, than the normal pair of micronuclei (Fig. 11).

The dividing macronuclei of tarn 8organisms elongate less extensively and more slowly than those of wild-type organisms. Correlated with this, the intranuclear microtubules are not peripherally concentrated in tam 8, and the numbers of microtubules per nuclear cross-section persisting at relatively late division stages (4 and 5) are about the same as those at an earlier stage (3) in wild-type organisms prior to the main rapid phase of elongation. These findings are compatible with other evidence (Tucker et al. 1980) that elongation is promoted by a peripherally situated microtubule-sliding mechanism. Reduced elongation in tam 8may also be partly due to delay in elongation. Some tam 8nuclei stop elongating although putative daughter nuclei have not separated. Disassembly of microtubules at stage 6 may intervene to reduce still further the potential of the microtubules for involvement in nuclear elongation. The failure of tam 8macronuclei to adopt an elliptical shape during stage 3 is another indication that intranuclear microtubules are abnormally deployed in these nuclei, since such shaping is spatially correlated with a marginal microtubule band in wild-type organisms (Tucker et al. 1980).

In tam 8organisms trichocyst motility and positioning are abnormal, micronuclear segregation is irregular, and the shaping and positioning of macronuclei are affected especially during binary fission. Why does the tam 8lesion produce this range of defects?

Examination of dividing macronuclei indicates that nucleation of microtubule assembly inside nuclei and subsequent elongation of these tubules are not impaired to any marked degree. The apparently normal organization of microtubules in cilia, around trichocyst tips, and in micronuclear separation spindles also suggests that the mutation does not have a direct effect on the assembly or basic structure of the organism’s microtubules. Possibly the genetic lesion interferes with the action of microtubule-associated contractile components. Intracellular motility in general involves a variety of ATPases such as dynein and actomyosin complexes, as well as a range of proteins that regulate the force-generating activities of these molecules (Goldman, Pollard & Rosenbaum, 1976; Dustin, 1978; Stebbings & Hyarns, 1979; Roberts & Hyams, 1979). Different forms of motility within a single cell may exploit a different selection of these components. It may be that the action of only one such component is directly affected by the tam 8lesion, so that micronuclear elongation and ciliary action proceed normally while macronuclear elongation and shaping, and the saltation and transport of trichocysts to the cell surface are impaired. The irregular segregation of micronuclei may be due to malfunction of a procedure that is not directly associated with microtubules (see below).

There is an alternative way of accounting for the defects inflicted by the tam 8mutation. The lesion may interfere with cell-surface-mediated inter-actions.

In tam 8organisms the dividing macronucleus does not take up a dorsal subcortical position yet the longitudinal orientation of elongating macronuclei and their abnormally distributed nucleoplasmic microtubules still occurs. Hence, such orientation is not dependent on subcortical positioning, although it is possible that adoption of an elliptical shape (which does not take place in tam 8)and formation of an elliptical marginal band of microtubules requires the close nucleo-cortical association that is achieved only in wild-type organisms. Interfission tam 8macronuclei also have abnormal shapes and positions. The positioning of other organelles is abnormal. Trichocysts do not exhibit saltatory motion (Aufderheide, 1978) and cortical attachment. In wild-type organisms pairs of daughter micronuclei become localized at opposite poles after breakdown of separation spindles and before the completion of cleavage (Tucker et al. 1980). This localization may ensure the even segregation of daughter micronuclei in wild-type organisms that is lacking in tam 8. Since micro-nuclear elongation proceeds normally in tam 8it is perhaps subcortical localization that is defective. Hence the mutation is perturbing events at the cortex, in the general cytoplasm, and within the macronucleus. Is the primary defect in some cytoplasmic or cortical component? Are all the nuclear abnormalities a consequence of defective nucleo-cortical interactions rather than a direct intranuclear effect of the tam 8lesion? This is a distinct possibility bearing in mind the pronounced nucleo-cortical interactions reported for other ciliates. Macronuclear shaping, division, and DNA synthesis are sensitive to the state of the cortex in Stentor(De Terra, 1978). Analysis of Tetrahymenamutants reveals that cortical development and macronuclear division are morphogenetically related, as are construction of the cortical oral organelles and micronuclear division (Frankel, Jenkins & De Bault, 1976). Furthermore, micro-nuclear spindles make close contact with the cortex in dividing Tetrahymena(Jaeckel-Williams, 1978).

If the tam 8mutation exerts its effect only viamacronucleo-cortical interactions then these influence the spatial arrangement of the intranuclear microtubules. What sort of signals might be involved and how could they cross the nuclear envelope? A variety of cytoskeletal networks extend from the vicinity of nuclei to the cell surface in a range of cell types during interfission (see Weber, Pollack & Bibring, 1975). It has been suggested that intermediate filaments are involved in the control of nuclear positioning; they still form an extensive network during cell division (Blose, 1979). There is experimental evidence for cytoskeletal interactions between nuclei and plasma membranes (Wang, Gunther & Edelman, 1975; Berke & Fishelson, 1979). A mutation that affects cytoskeletal components connecting trichocysts and nuclei to plasma membranes might disturb control of macronuclear division, trichocyst/plasma membrane attachment, and polar localization of micronuclei. No such connexions between macronucleus and cortex have been detected in wild-type organisms (Tucker et al. 1980) but they may be of a type that is not preserved during preparation for electron microscopy, such as discrete regions of subsurface cytoplasmic gelation.

It is especially interesting that the samepositional defects (for both macro- and micro-nuclei) as those described above for tam 8are caused by a number of other mutations at different chromosomal loci which allprevent attachment of trichocysts to the plasma membrane (Cohen & Beisson, submitted for publication). This is further evidence for positional interactions between the cell surface, nuclei and trichocysts in Paramecium.

We thank Mr D. L. J. Roche for skilful assistance with electron microscopy. A training fellowship from the Ligue Nationale Français contre le Cancer to J.C., support from the Délégation Generale à la Recherche Scientifique et Technique (grant no. 77-70267) to J.B., and from the Science Research Council (U.K.) to J.B. T. are gratefully acknowledged.