Proportional control of organelle position by a mechanism which similarly monitors cell size of wild type and conical form-mutant Tetrahymena

Analysis of pattern formation in ciliates has been largely confined to dorsoventrally flattened ciliates such as the hypotrichs Euplotes and Urostyla and the cyrtophorine Chilodonella. In these ciliates, the field boundaries are considered coincident with the border between dorsal and ventral surfaces. The development of cirral primordia in Urostyla altered by microsurgery (Jerka-Dziadosz & Frankel, 1969; Jerka-Dziadosz, 1974, 1977), the spatial distribution of ciliary units on the dorsal surface of Euplotes (Frankel, 1975a), and positioning of contractile vacuole pores in Chilodonella (Kaczanowska, 1974, 1975) have been explained with reference to these boundaries. However, it is difficult to define field boundaries in ‘radially’ symmetrical ciliates such as Tetrahymena and Paramecium (see Nanney, 1968; Sonneborn, 1974; Frankel, 1974). The interaction of positional signals (e.g. diffusing morphogens) emanating from one or several reference points could be sufficient to specify the pattern of cortical structures in these ciliates.

It has been established that dividing Tetrahymena position the new oral primordium or mouthparts in a regulative manner, proportional to the body or cell length (Lynn & Tucker, 1976). This relationship was demonstrated with log dividers and first and second post-starvation dividers of Tetrahymena corlissi. Since these three types of dividers are similarly shaped, it is not possible to test whether body length or overall body size (e.g. surface area, volume, biomass) is the determining parameter nor what the reference points for oral primordium positioning are. Genetically related ciliates of different shape would be necessary for this purpose. Doerder, Frankel, Jenkins & De Bault (1975) have isolated a form-mutant of Tetrahymena thermophila. This form--mutant, named conical, results from the action of a single recessive gene which changes the overall cell shape from ovoid to conical and places the anterior mouthparts on average further from the anterior pole. The gene apparently does not affect cortical characteristics such as number of ciliary rows, number of ciliary units within these rows, and positions of contractile vacuole pores (Doerder et al. If it is assumed that this gene also does not affect the mechanism responsible for positioning the oral primordium, then it is possible to test the hypothesis that primordium position is determined with reference to overall body size rather than to some single linear parameter such as cell length.

This report will demonstrate that similar proportional positioning of the oral primordium of both wild-type (co+) and conical (co) cells occurs both as a function of cell length and, more importantly, as a function of cell size estimated by the product of cell length and cell width. The reference point for the measurement is apparently the anterior mouthparts rather than the anterior pole of the cell.

Tetrahymena thermophila (formerly Tetrahymena pyriformis syngen 1, see Nanney & McCoy, 1976) strain D was obtained from Dr D. L. Nanney and cultured axenically in 2-0 % proteose-peptone with 0·5 % yeast extract (PP) or in tryptone-dextrin-vitamin-salts (TS) medium (Frankel, 1965) at 28 °C. Cells in TS medium were sampled only once in mid-log phase, while cells in PP medium were sampled repeatedly from mid-log to early stationary phase. Tetrahymena corlissi strain WT, clone TC-2, was cultured axenically in 2·0 % proteose-peptone and 0·1 % yeast extract. The methods for obtaining log dividers and first and second post-starvation dividers of T. corlissi have been described (Lynn & Tucker, 1976).

Both T. thermophila and T. corlissi were stained by the Chatton-Lwoff wet-silver procedure, following the directions of Frankel & Heckmann (1968) for T. thermophila and those of Corliss (1953) for T. corlissi.

Measurement of silver-stained organisms was performed in two ways. Dr J. Frankel generously supplied the original data obtained during the description of the co mutant (Doerder et al. 1915). Measurements of cell proportions of one sample of each genotype, cultured in TS medium, were made with an ocular micrometer as described in the original paper.

Measurements of the PP cultured T. thermophila and T. corlissi were made using a Leitz filar ocular micrometer mounted on a Leitz Ortholux microscope. For T. thermophila, cell proportions were measured from three (co⁺) or four (co) slides representing cells fixed in middle to late log phase. The distances measured on ventrally oriented specimens whose long axis was approximately on the horizontal plane are as shown in Doerder et al. (1975; fig. 14, p. 247) or in Lynn & Tucker (1976; fig. 1, p. 37). Only specimens in division stages 1 – 2 have been included.

The oral primordium begins to develop several μ m posterior to the anterior mouthparts and adjacent to a ciliary row (kinety 1) which runs between the two sets of mouthparts in organisms at stages 1 – 2 of fission (Fig. 1). By the end of stage 2 the three membranelles are not yet apparent and kineties in the presumptive furrow region have not yet been interrupted.

The distance between mouthparts (d, Fig. 1) in dividing co⁺ and co organisms at fission stages 1-2 is proportionately related to body length (f, Fig. 1). The TS cultured co⁺ cells varied between 37 – 46 μ m in length by 12 – 24 μ m in width (iv, Fig. 1), and d ranged from 6 to 11 μ m; the co cells varied from 27-39 μ m in length by 20 – 27 μ m in width, and d ranged from 4 to 12 μ m (Table 1). Although there is a significant difference between cell lengths of co⁺ and co cells, d is not significantly different for the two cell types. Hence, the ratio of d to I differs in co⁺ and co cells (Table 1). If co⁺ and co cells are assessing cell size by the same mechanism (see Introduction), then some estimate of cell size should prove similar for both types. The product of cell length and cell width, l × w, for co⁺ and co cells is not significantly different (Table 1), although these cells are shaped very differently (Fig. 1). This suggests that the positional mechanism determines the location of the oral primordium as a function of cell size, rather than cell length, since the former character is similar and the latter character is different for the two genotypes while the distance between mouthparts remains the same.

The y-intercept is not significantly different from zero (t = 1·65; D.F. = 46). Thus, the distance between mouthparts is exactly proportional to this estimate of cell size.

The anterior mouthparts in co⁺ and co are not positioned at similar distances from the anterior end of the cell; in co⁺ cells the average preoral distance is 3·8 μ m while in co cells it is 6·2 μ m (Table 1). The oral primordium is positioned at a similar, though statistically significantly different, distance from the anterior end in co⁺ and co cells (Table 1), while it is an average of 14·4 μ m and 5·4 μ m respectively from the posterior pole of these two cell types. Although this suggests that the anterior pole could be an important reference point, the next section will demonstrate this to be unlikely.

A second experiment was made to determine if proportionality was demonstrable in co⁺ and co individually as well as jointly. Several different samples of ciliates were fixed and stained 3 years after the original isolation of the co mutant and 2 years after the TS experiment described above. In this second sample totalling 50 individuals of each genotype, co⁺ and co cells have diverged in cortical characteristics from the cells in the first isolation. For example, the modal number of kineties is now 20(17 – 22; n = 50) for co⁺ and 16(13 – 19; n = 50) for co cells where previously these were 19(16-21; n = 196) and 18(14 – 21; n = 253) respectively, while the number of postoral rows is 1-4 (1-2; n = 30) for co⁺ and 1·1 (1 – 2; n = 30) for co cells where previously these were considered quite similar (Tables 3, 5 in Doerder et al. 1975). This degree of variation in cortical features has been observed over an extended time within co⁺ strains of T. pyriformis (Frankel, 1972, personal communication). Furthermore, co+ and co cells which were cultured in the very rich PP medium are substantially larger than in the previous experiment in the less nutrient-rich TS medium. The co⁺ cells have become differentially wider so that their width is no longer significantly less than that of co cells (Table 2). However, the characteristic difference between the ovoid shape of co⁺ cells and the conical shape of co cells (Fig. 1) remains undiminished.

Since relative cell size has changed since the TS experiment, the comparison of l × w reveals a significant difference between co⁺ and co cells (Table 2). However, these changes in cortical features and shape have not greatly affected the spacing of the mouthparts in the two cell types. The average distance is quite similar though significantly different in the two cell types (Table 2).

Regression analysis of the PP experiment demonstrates that here also there is a proportional relationship between d and l in these genetically different strains of T. thermophila (Fig. 2). Comparison of the ratio d/l shows a distinct difference in proportionality of these two parameters for co⁺ and co cells (Table 2). Yet, the slopes of the regression lines (Fig. 2) are not significantly different. If d is regressed upon the estimate of cell size I x w, co⁺ and co cells fall along the same line which goes through the origin (Fig. 3). The two lines fitted to co⁺ and co separately (given as d = 0·00973 (l) (w) −1·31 for co⁺, and d = 0·00851 (l) (w) + 1·09 for cd) do not fit the scatter of points significantly better than the single line d = 0·00907 (l) (w) for co⁺ and co (where F = 2·43 < 2·68 at P = 0·05 for D.F. = 3/99).

Since l × w is a rather crude approximation of cell size, a better estimate was derived using calculus. Measurements of the profiles of 10 co⁺ and 20 co cells were used to derive equations for the cell shape. By integration, volumes and surface areas were calculated for each cell and d was then regressed upon these estimates. These further estimates of cell size were not able to explain any more of the variation than the l × w estimate.

As in the previous experiment, preoral and postprimordium distances in co⁺ and co are very different. However, the preprimordium distances are not significantly different (Table 2). Again, this suggests that the anterior pole could be an important reference point. If the preprimordium distance, p, is regressed upon l l × w, each genotype lies on a different line. In neither case does a single line suffice. For example, p = 0·0111 (l) (w) + 9·40 for co⁺ and p = 0·00941 (Z)(w) +14 – 17 for co are significantly better at accounting for the variation at P = 0·05 than the single line p = 0·00775 (l) (w) +15 – 29 for co⁺ and co jointly. In addition, none of these lines demonstrates exact proportionality as the y-intercepts are very significantly different from zero.

The number of basal bodies in kinety 1 between the mouthparts has been counted in silver-stained co⁺ and co cells (Table 2). Silver-stained basal body counts are a good estimate of the true basal body number between mouthparts (Lynn & Tucker, 1976; Lynn, unpublished observations). There is a statistically significant difference in this number in co+ and co cells although a similar distance separates the mouthparts (Table 2). As observed in T. pyriformis strain W (Lynn & Tucker, 1976), there is a great deal of variation in the number of basal bodies even within strains (Table 2).

To test that a similar relationship between d and cell size exists in another species of Tetrahymena, individuals of T. corlissi were measured. There is a significant regression when d is regressed upon l × w for these 3 types of dividers (Fig. 4). Although, as is apparent from Fig. 4, d is related to l × w in a proportional manner when all three cell types are included, the best fit line for each cell type does not coincide exactly with this regression line.

A mechanism which proportionately assesses cell length has been suggested to position the oral primordium in dividing T. corlissi (Lynn & Tucker, 1976). The results of the present study demonstrate that a similar mechanism which monitors at least cell length also determines the position of the oral primordium in dividing T. thermophila co⁺ and co cells.

The ratio d/l is different for each strain (Tables 1, 2). As co⁺ and co cells have been demonstrated to be quite similar for a number of cortical characteristics (Doerder et al. 1975), the differences in d/l ratios and in assessment of d as a function of l do not in themselves refute the assumption that the same mechanism positions the oral primordium in both cell types. Therefore, it is assumed that the co gene has not altered the fundamental positioning mechanism. In fact, the regression of d on l reveals that both co⁺ and co cells have the same proportion of cell length contributing to the estimation of d (Fig. 2) and thus, perhaps, share a similar underlying positional mechanism.

Cell length may not be the crucial parameter which is the reference for the distance assessment, since d/l ratios are different and the same proportion of I contributes to the estimation of d (Fig. 2). A simple estimate of cell size is the product of l × w which is likely to be an estimate of cell surface area rather than cell volume or biomass. Undoubtedly, it would be highly correlated with all three. In the TS experiment, co⁺ and co are on average not significantly different when d and I × w are compared (Table 1). In the PP experiment, although d and l × w are now different since the two genotypes have diverged morphologically, regression analysis clearly shows that the cells could be making an identical proportional assessment of I × w (Fig. 3). A relationship to cell size is exhibited by different sized dividers of T. corlissi (Fig. 4).

For several other reasons, cortical surface area is likely to be the component of a ciliate’s size which is used to determine the distance which the oral primordium is from the anterior mouthparts. Many ciliates are able to change shape rapidly and consequently volume varies considerably. The ciliate cortex and cell surface are much more stable, being resistant to deformation since they are composed of a complex array of basal bodies, microtubules, and microfilaments. If the mechanism of size assessment requires a stable cell parameter, surface area (or l × w?) is more likely than volume. Already the cell surface has been implicated in the control of cellular events. De Terra (1974, 1975) demonstrated that the cortex of Stentor can control nuclear division. Transplanted cortical components, especially the oral region of the cortex, have an inhibitory or inductive effect on the cell to which they are transplanted (Uhlig, 1960; Tartar, 1961). It is possible that changes in surface area (perhaps correlated with changes in number of ciliary rows?) might also be a factor in the distribution of basal bodies among ciliary rows in Tetrahymena. The researches of Nanney and co-workers (Nanney, 1971; Nanney & Chow, 1974) have demonstrated that as the number of ciliary rows increases, the number of basal bodies within a row decreases. However, the exact relationships between number of ciliary rows and number of basal bodies per row to cell length, cell width, and distance between mouthparts have yet to be analysed.

The old or anterior mouthparts of ciliates have been considered an important, and by some, an essential reference point for positional determination of developing cortical structures. Kaczanowska (1974, 1975) has concluded this from her studies of contractile vacuole pore positioning in Chilodonella. Sonneborn (1974) arrived at the same conclusion in his review of ciliate morphogenesis.

In co⁺ and co cells, the preprimordium and postprimordium distances are different for cells of the same average size (i.e. when l × w is the same, Table 1). Moreover, the regression of preprimordium distance on I or I × w does not demonstrate a similar relationship for both genotypes. If the anterior pole is the reference point, these two results suggest that a different assessment of cell size operates in each genotype. On the other hand, the interoral distance d is the same when I × w is the same (Table 1) and there is a similar relationship when d is regressed on I or I × w for both genotypes. Thus, if the anterior mouthparts are the reference point, these results suggest that the same exact assessment of cell size is operating in each genotype. Since both genotypes are similar in a number of other cortical characters (Doerder et al. 1975), this second alternative is preferred. The mo3 mutant of T. thermophila is arrested during division so that chains of cells are formed (Frankel, Jenkins & De Bault,1976) The cells within the chain which do not have an independent anterior pole still place an oral primordium at some distance from the old mouthparts. What this distance is related to is presently uncertain. It is likely that the mo3 mutant will provide a crucial test between the above alternative hypotheses.

However, this is not to say that the anterior pole is irrelevant in shape and pattern formation. The facts that the anterior mouthparts in co⁺ and co cells are on average and predictably at different distances from the anterior pole and that these cells have different cell shapes suggest that some mechanism, perhaps under control of the co gene, alters the overall cortical patterning. The above discussion has assumed that one mechanism shapes the cell after cytokinesis to yield the co⁺ and co phenotypes and another mechanism determines the position of the oral primordium.

The measurement for oral primordium position could proceed along kinety 1. However, the spacing of the mouthparts is unlikely to involve a count of the absolute number of cortical units or basal bodies from the anterior mouth parts. The variability in number of basal bodies between mouthparts of these strains of T. thermophila and in T. corlissi (Lynn & Tucker, 1976) seems to preclude this possibility. Frankel (personal communication) has isolated a mutant disl which has a highly disorganized kinety pattern, including kinety 1, and yet this strain manifests normal positioning of the oral primordium. This clearly indicates that kinety 1 is not essential for positioning the oralprimordium.

Jerka-Dziadosz (1974) has presented an analysis of modified ‘sand-hill’ models to account for the pattern of cortical development in Urostyla cells which have been morphologically altered by microsurgery. The pattern is explained in terms of gradients re-establishing themselves in an altered field. Frankel (1974, 1975b) has discriminated explicitly between the concept of a graded distribution of a property and gradients of diffusing morphogens. At this time, Frankel prefers to discuss the phenomenon of pattern formation employing the abstract concept of a graded property.

There are two major obstacles to the establishment of a morphogenetic diffusion gradient in ciliates, assuming that an appropriate source can be chosen. First, is there sufficient time in the cell cycle to establish a gradient by diffusion? Secondly, is the cytoplasm ever free from cyclotic movements which would prevent the establishment or reduce the equilibrium stability of such a gradient?

There is an obvious choice for a source in the differentiated ciliate. It is the anterior oral apparatus. There is strong evidence from the researches of Uhlig (1960) and Tartar (1961) that the oral apparatus has the properties of a morphogenetic ‘source’. Indirect evidence is provided by analysis of pattern formation in Chilodonella (Kaczanowska, 1974), T. corlissi (Lynn & Tucker, 1976), and T. thermophila (see Results) that the anterior oral apparatus is a primary reference point for the determination of the position of the new oral apparatus.

There is also an appropriate time in the cell cycle in some ciliates when cyclosis, the rapid movement of endoplasm, ceases. In Paramecium, cyclosis ceases during division at the time when the oral primordium is migrating and the macronucleus is dividing (Sikora & Kuznicki, 1976). This stability lasts for 5 – 15 min. Although this phenomenon has not been observed in Tetrahymena and would have to occur in the middle of the cell cycle when the oral primordium develops, a cessation of equal time would be more than adequate to establish a gradient by diffusion. For ciliates, this condition may not be necessary as there is no cyclosis in the epiplasm and cortical ectoplasm which are themselves very stable cytoplasmic regions (Sibley & Hanson, 1974). Diffusion of a morphogen might occur through this ectoplasmic region of the cytoplasm and thus be unaffected by endoplasmic cyclosis. Even if the properties of the ectoplasm are somewhat different from the general properties of cytoplasm assumed in Crick’s analyses (1970, 1971), there are at least two orders of magnitude more time available between divisions than the minimal time required by the model.

The anterior oral apparatus of Tetrahymena could be the source of a morphogen which diffuses through the ectoplasmic regions of the cell. There is sufficient time in the cell cycle for a diffusion gradient to reach an equilibrium stability. If a ‘circular gradient’ exists in Tetrahymena (Nanney, 1966, 1968; Nanney, Chow & Wozencraft, 1975) as it seemingly does in Stentor (Uhlig, 1960) and if it extends from the ‘stomatogenic kinety’ around the cell, the position of the oral primordium could be exactly specified by the end-boundary of the circular gradient and by the diffusing morphogen.

I would like to thank Dr Joseph Frankel for his enthusiastic encouragement, for his careful criticisms, and especially for his generous provision of original data and silver-stained specimens. I am also indebted to Dr John B. Tucker and Mr C. D. Sinclair for their advice and criticism.

This research has been supported by grants B/SR/88418 and B/SR/5894.5 from the Science Research Council (U.K.) and by grant No. 08485 from National Institutes of Health (U.S.) awarded to J. Frankel. The author was REFERENCES supported as a NATO Postdoctoral Fellow by the National Research Council of Canada.