A Mutant of Tetrahymena Thermophila with a Partial Mirror-Image Duplication of Cell Surface Pattern

/. Embryol exp. Morph. Vol. 49, pp. 167-202, 1979 Printed in Great Britain © Company of Biologists Limited 1979

A mutant of Tetrahymena thermophila with a partial mirror-image duplication of cell surface pattern

I. Analysis of the phenotype

By MARIA JERKA-DZIADOSZ1 AND JOSEPH FRANKEL2 From the Department of Zoology, University of Iowa

SUMMARY Cells of a mutant clone, CU-127, of Tetrahymena thermophila (formerly T. pyriformis, syngen 1) manifest three anatomical abnormalities. First, the stable number of ciliary meridians is 21-25, above the usual number (17-21) in this species. Second, up to 30 % of the cells have two oral apparatuses (OAs), one normal and the other abnormal. Third, more than one-half of the cells possess two distinct sets of contractile vacuole pores (CVPs). In some living cells two contractile vacuoles are seen. These abnormalities have persisted un- changed during more than 500 generations of vegetative propagation, and are similarly expressed in subclones. The normal and abnormal OAs are topographically segregated, with normal OAs developing along the 'primary oral axis' and abnormal OAs developing along a 'secondary oral axis' that is situated 170° of the cell circumference to the cell's right of the primary oral axis. CVPs always appear within this 170° arc and never within the com- plementary 190° arc to the left of the primary oral axis. A unique feature of the CU-127 clone is the commonly expressed mirror image reversal of the structural pattern of OAs that develop along the secondary oral axis. The primordia of such OAs initially appear (as usual) to the cell's left of a ciliary meridian, but as membranelles develop they frequently come to be oriented in a mirror image of the normal pattern, and an undulating membrane sometimes develops on the wrong (left) side of the oral primordium. When two sets of CVPs are formed, their average positions are roughly equidistant with respect to the two oral axes, with the two sets located 50-60° to the right and left respectively of the primary and secondary oral axes. Such cells are thus bilaterally symmetrical about a plane defined by the central longitudinal axis and the halfway point between the two CVP sets (see Fig. 25). This plane bisects the cell into a normal and a 'reversed' half-cell. However, only oral asymmetry and large-scale CVP positioning are subject to such reversal; all ciliary meridians remain of normal asymmetry and all CVPs are situated on the left side of CVP meridians. The fact that major aspects of large-scale cellular organization can be reversed while the 'fine-positioning' associated with the ciliary meridians remains normal indicates that the two aspects of cell organization are distinct. 1 Author's address: Department of Cell Biology, M. Nencki Institute of Experimental Biology, Pasteura 3, Warsaw 02-093, Poland. 2 Author's address (for reprints): Department of Zoology, University of Iowa, Iowa City, Iowa 52242, U.S.A. 168 M. JERKA-DZIADOSZ AND J. FRANKEL

INTRODUCTION Positioning of cell surface organelle systems in ciliated protozoa involves interactions taking place over very small as well as relatively large intracellular distances. There has been considerable discussion in the literature over the degree to which the mechanisms governing long-range positioning differ from those controlling short-range positioning (Frankel, 1974, 1975; Jerka-Dziadosz, 1974; Lynn & Tucker, 1976; Lynn, 1977; Sonneborn, 1975). The paradigm for short-range positional interactions is the propagation of the ciliary meridian. Within this ensemble new structures such as basal bodies and accessory fibrillar and microtubular systems are positioned in definite spatial relations to nearby pre-existing structures. The proof of the determinative role of the pre-existing topographic arrangement within ciliary meridians was the demonstration that an 180° inversion of that arrangement is faithfully propagated. This was shown first in Paramecium (Beisson & Sonneborn, 1965) and later in Tetrahymena (Ng & Frankel, 1977; Ng & R. Williams, 1977). In Tetrahymena it was further ascertained that the geometry of the ciliary meridian also controls the fine- positioning of the contractile vacuole pore (Ng, 1977, 1978) as well as aspects of the positioning of cortically situated mitochondria (Aufderheide, 1978). Investigation of positioning of structures such as new oral apparatuses has been carried out mainly by microsurgical experiments on large ciliates, such as Stentor (Tartar, 1962; Uhlig, 1960) and urostylids (Jerka-Dziadosz, 1974, 1977). The results of these experiments suggest that positioning over long intracellular distances may involve 'gradient-fields' that may be analogous to those operating in multicellular development (Frankel, 1974, 1975; Jerka-Dziadosz, 1974). In Tetrahymena, morphometric analyses of the cell latitudes at which new oral apparatuses are formed (Lynn & Tucker, 1976; Lynn, 1977) and of the longitudes at which new contractile vacuole pores appear (Nanney, 1966 a) suggest that both structures are determined relationally as joint functions of reference points (or axes) and aspects of overall cell size. A decisive judgement of the relationship of long-range positioning systems to the better-understood short-range systems could be made if it were possible to rotate or reverse either system within the spatial domain of another. One way of doing this would be to obtain a propagated geometrical reversal of some aspect of large-scale cellular organization. A propagated reversal of asymmetry of feeding structures has already been obtained in mirror-image doublets of hypotrich ciliates (Faure-Fremiet, 1945; Tchang, Shi & Pang, 1964; Tchang & Pang, 1965; Dryl & Totwen-Nowakowska, 1972), but these cases have not been subjected to detailed cytological analysis. Recently, however, we have encountered a clone of Tetrahymena thermophila that indefinitely perpetuates the capacity to manifest a mirror-image reversal of oral structures at a well- defined cellular position and also maintains a corresponding reversal of contractile vacuole pore positions. Since in Tetrahymena (unlike the hypotrichs Mirror-image duplication in Tetrahymena 169 referred to above) both oral structures and contractile vacuole pores develop in the close neighborhood of ciliary meridians (reviewed in Frankel & Williams, 1973), this reversal achieves the desired superimposition of two positional systems. The cytological analysis to be presented here shows that these two systems are indeed separate and dissociable. The accompanying paper (Frankel & Jenkins, 1979) demonstrates that the unique reversal of asymmetry in this clone is under genie control, and provides further evidence for the separation of short-range and long-range positional controls.

MATERIALS AND METHODS The CU-127 clone of Tetrahymena thermophila (Nanney & McCoy, 1976) that is the subject of this investigation was obtained in December 1976 from the laboratory of Dr P. Bruns. It had previously been subjected to mutagenesis in 10 /Ag/ml of iV-methyl-A^'-nitrosoguanidine, followed by short-circuit genomic exclusion, a protocol designed to select homozygous cells rapidly (Bruns, Brussard & Kavka, 1976). CU-127 was one of a subset of morphologically abnormal clones among a larger number of clones that had been screened for temperature sensitivity following the above-mentioned mutagenesis and genomic exclusion protocol (Bruns & Sanford, 1978). All but one of the experiments were performed on a sample received directly from Dr Bruns' laboratory in December 1976. One experiment (cf. Table 6; Table 7, II) was carried out on another sample that had first been sent by Dr Bruns to Dr D. L. Nanney's laboratory for cryo- preservation and was received from there in February 1977; this is designated CU-127 (111.). The media used in experiments were all axenic, and consist of four different recipes: TGVS: 0-3% bacto-tryptone (Difco), 0-5% glucose, vitamins, and salts (Frankel, 1965); 1 % PPY: 1 % proteose peptone (Difco) plus 01 % yeast extract (Difco); 2 % PPY: 2 % proteose peptone plus 0-5 % yeast extract; Dryl's: Dryl's inorganic medium made up as described by Nelsen & DeBault (1978). Stocks were maintained at 28 °C in axenic tube cultures containing 5 ml of 1 % PPY or TGVS medium, with transfer daily or every second day [weekly for CU-127 (111.)]. Fernbach flasks (500 ml, Jena Glaswerk) containing 100 ml of medium were inoculated with one to five drops of 1-day-old tube culture. Flasks containing TGVS medium were inoculated with cells from TGVS tubes, while flasks with 1 % PPY or 2 % PPY medium were inoculated from 1 % PPY tubes. The flask cultures were incubated for 20-24 h at 28° ± 1 °C to yield mid-log phase cultures (cell density 15000-50000 per ml). In some cases such cultures were shifted to a bath set at 40° ±0-1 °C (39-5 °C inside flask) and maintained at that temperature for various durations (usually 4-5-5 h), following which, in certain experiments, the flask was returned to 28 °C for 1-5 h. Samples were fixed at various intervals. In a few experiments cultures were fixed after being allowed to enter stationary 170 M. JERKA-DZIADOSZ AND J. FRANKEL phase by 48 h continuous maintenance at 28 °C, or by 24 h following a shift of a mid-log phase culture from 28 to 40 °C. In still other experiments, cells grown in 1 % PPY at 28 °C were washed by centrifugation and transferred to Dryl's inorganic medium and maintained under various temperature regimens. In one experiment, the cells were maintained in 1 % PPY and TGVS tubes, with daily transfer, for 10 successive days at 39-5 °C. After the tenth such serial tube-culture had reached late-log phase, the contents of the entire tube were poured into a Fernbach flask containing 100 ml of medium, kept at 39-5 °C for 3 h, and fixed. These cultures had thus spent about 100 cell generations at 39-5 °C prior to fixation. In another experiment, a CU-127 culture growing in 1 % PPY was subjected to clonal expansion. Thirty single cells were removed by a micropipette from a log-phase flask culture growing in 1 % PPY at 28 °C, and were individually inoculated into depressions of three-spot depression slides, each containing 0-5 ml pen-strep 1 % PPY (Frankel, Jenkins, Doerder & Nelsen, 1976). They were kept in these depressions at 25° for 2 days. These 30 subcultures were transferred from the depressions to culture tubes each containing 5 ml of 1 % PPY medium, and maintained at 25° for 2 further days. The 30 cultures were then fixed for silver impregnation as they were entering stationary phase, 25-30 generations after the original single-cell isolation. All but two of the 30 subclones were terminated at the time of fixation. One subclone (CU-127-1) was retained and used in further experiments. Cells were prepared for counting as described earlier (Frankel, 1965) and counted in a model A Coulter Counter (Coulter Electronics). Food vacuole formation was assayed by uptake of carmine particles. Silver impregnation was performed according to Frankel & Heckmann (1968) with slight modifications (Nelsen & DeBault, 1978). Measurements of silver impregnated specimens were made with a filar micrometer eyepiece (American Optical). Protargol staining was done according to the procedure of Ng & Nelsen (1977). Cells were fixed for scanning electron microscopy in 1 % osmic acid for 3 min, and submitted to critical point drying following the procedure of Ruffolo (1974) with the omission of amyl acetate. Cells were observed in a Cambridge stereoscan S4 microscope.

RESULTS 1. General comparison of the phenotype of CU-127 and wild-type cells Cells of the CU-127 clone were 'originally scored as flat cells' (J. W. McCoy, personal communication). This flattening is generally not very pronounced, and CU-127 cells otherwise differ only slightly from wild-type cells in size and form. Like wild-type (WT) cells, they have a single macronucleus and generally a single micronucleus (Fig. 9 a). The rate of population growth and the density attained at stationary phase are similar for CU127 and WT. CU-127 grows well at high temperature (39-5 °C), and differs from WT only in tending to die after Mirror-image duplication in Tetrahymena 171

FIGURES 1 AND 2 Scanning electron micrographs of CU-127 cells. The bars in both micrographs represent 10 /.cm. Fig. 1. A polar view of a cell with two oral apparatuses (OAs), indicated by arrows. The OA on the viewer's right is the structurally normal primary OA, while the OA on the viewer's left is the secondary OA. The arrows point to the first membranelle of each of these two OAs. Note that the secondary OA has three membranelles, but no undulating membrane. Fig. 2. A lateral view of a dividing cell with one OA and two oral primordia. The normal oral axis is at the left (viewer's right) edge of the cell, where both OA and OP are visible. The secondary oral axis is at the right (viewer's left) edge of the cell, where only an OP is present. Note the single set of contractile vacuole pores (CVPs) neai the posterior end of the cell (arrow). prolonged maintenance at 39-5 °C without transfer, which is the probable basis on which CU-127 was originally selected as a 'temperature sensitive' putative mutant (Bruns & Sanford, 1978). CU-127 differs dramatically from WT in three major aspects of cortical pattern. First, the number of ciliary meridians is higher in CU-127 cells than in WT cells, with typically 21-25 in the former and 17-21 in the latter. The high meridian number in CU-127 cells is stable even when cells are propagated by frequent transfer, conditions under which WT populations that include cells with unusually high ciliary meridian numbers tend to return to a 'stability center' near 19 ciliary meridians (Nanney, 19666, 1968; Frankel, unpublished observations). Ciliary meridians of CU-127 are of normal polarity and asymmetry, and also of normal ultrastructural organization. A second striking cortical anomaly manifested by CU-127 cells is the presence of secondary abnormal oral areas at a well defined location nearly 180° opposite 172 M. JERKA-DZIADOSZ AND J. FRANKEL

OA-1

** / •• &* * M- * * / ' 4 / ••/ «• // £f jiff A / * ./*< Mirror-image duplication in Tetrahymena 173 from the normal OA (Figs. 1, 4). A certain proportion of dividing CU-127 cells produce a similarly abnormal secondary oral primordium (OP) nearly opposite from the normal OP (Fig. 2). On the side of the cell with the normal OA and OP, both the anatomy and mode of development of OA are indistinguishable from those of WT cells. The longitudinal strip of cell surface along which the normal OA is located anteriorly and the OP is developing subequatorially prior to cell division is here called the 'primary oral axis'. Correspondingly, on the opposite side of the cell the longitudinal strip of the cell where the abnormal OA is located and along which the abnormal OP develops is termed the 'secondary oral axis'. This is unambiguously distinguished by the abnormality of oral structures and their irregular occurrence (see section 2b). The secondary oral axis also lacks a cytoproct such as is normally observed at the posterior end of the primary axis. Finally, the majority of CU-127 cells possess two sets of contractile vacuole pores (CVPs). A cell is said to possess one CVP set if all CVPs are situated near adjacent meridians. All WT cells have one CVP set situated immediately to the left1 of the posterior ends of one or two adjacent meridians that are located approximately 22 % of the cell circumference to the right of meridian no. 1 (Nanney, 1966a; Nanney, Chow & Wozencraft, 1975). Some CU-127 cells possess a single CVP set situated at the position usual in normal cells (Figs. 2, 4, and the posterior cell in Fig. 5), but the majority have two CVP sets separated by one, two, three or four ciliary meridians (Fig. 3 and anterior cell in Fig. 5). The mean number of CVPs within each set in cells with two sets (1-57 and 1-63 for the left and right site respectively) approaches the mean number of CVPs in cells that have only one set (1-84). Each of the two CVP sets is at least transiently underlain by a separate contractile vacuole, as two such vacuoles are frequently 1 Throughout this report, right and left refer to the observer's right and left, assuming that he stands inside the animal so that his anterior-posterior axis coincides with that of the animal, and he keeps turning around his own longitudinal axis to face the surface of the animal.

Figs. 3-5. Silver impregnated CU-127 cells. The bar in each micrograph represents 10 /tm. Fig. 3. A cell with two CVP sets (arrows) separated by four ciliary meridians lacking CVPs. Fig. 4. A cell with two OAs and one CVP set (arrow). The apparent reversal of the orientation of oral structures in the primary OA (OA-1) is a focusing artifact. Fig. 5. A dividing cell with two CVP sets just anterior to the division furrow (small arrows), and one CVP set near the posterior end (large arrow). Note that the two CVP sets of the anterior fission product are positioned to the right and left of the single CVP set of the posterior fission product. Fig. 6. A living CU-127 cell photographed under bright field optics. Two contractile vacuoles, indicated by arrows, are evident just anterior to the division furrow, while one is present at the posterior end of the cell. The bar represents 10 /on. 12 KMB 49 174 M. JERKA-DZIADOSZ AND J. FRANKEL

Table 1. Number of OAs and CVP sets in CU-127 cells under various growth conditions

Phenotype Tpm. 0/ Il\/Al I till /o ' \J VCI 11 °/ perature in oral 1 OA 1 OA 2 OA 2OA ( Growth. regimen* devel- 1 CVP 2 CVP 1 CVP 2 CVP 2 CVP Medium phase (°Q opment set sets set sets 2 OAs sets TGVS Log 28 30 41 46 6 7 13 53 TGVS Stat 28 0 44 54 1 0 1 54 1 %PPY Log 28 24 41 44 6 9 15 53 1 %PPY Stat 28 0 40 60 0 0 0 60 2%PPY Log 28 27 17 55 7 21 28 76 2%PPY Stat 28 5 24 69 1 6 7 75 2%PPY Log 39-5 27 36 59 1 4 5 63 * The 28° cultures were fixed after growth of flask cultures to mid-log phase 1:20-24 h) or to early stationary phase (48 h). The 39-5° culture was sampled after continuous growth at that temperature for about 100 cell generations (see Methods).

observed at the same cell latitude in CU-127 cells (Fig. 6); this situation has not previously been reported for Tetrahymena. The two neighboring contractile vacuoles empty their contents asynchronously. The 'fine-positioning' (Ng, 1977) of CVPs is always normal in CU-127 cells: CVPs are observed close to the left side of ciliary meridians or sometimes directly on the axis of meridians, but not on the right side of the meridians (Figs. 2-4). Although they superficially resemble homopolar doublets, induced by division blockage (Faure-Fremiet, 1948) or by failure of separation of conjugating cells (Nanney et al. 1975) that are in the process of reverting to singlets (Nanney, 1966b; Nanney et al. 1975), the CU-127 cells are profoundly different from such doublets in two respects. First, the expression of the CU-127 phenotype has been stable for over 500 generations of propagation in our laboratory, with no substantial augmentation nor diminution of the degree of doubleness of CVP sets or OAs. Second, instead of possessing two similar normal oral axes, CU-127 cells possess a primary oral axis that is completely normal plus a secondary oral axis that is strikingly abnormal (see Results, sections 3 and 4).

2. Analysis of expression of CU-127 (a) Expression in non-dividing cells under different conditions The expression of secondary OAs and double sets of CVPs by CU-127 cells was analyzed in combinations of three media, two growth phases (log and stationary), and several temperature regimens. Typical results are shown in Table 1. Three basic findings emerge. First, expression is not temperature sensitive: cultivation at high temperature, either continuously (Table 1) or for Mirror-image duplication in Tetrahymena 175 1-OA/1-OP 1-OA/2-OP 2-OA/l-OP 2-OA/2-OP

To t;il

0 14 13 0 10 1 0 0 1 11 66

{7-8)

0 0 0 1 76

22 U4-5) (.8-5) Fig. 7. Tabulation of configurations of OAs, oral primordia, and CVPs in dividing CU-127 cells. Each of the 16 sketches schematically represents one of the 16 possible configurations of OAs and CVP sets in dividing cells. Each black dot indicates a CVP set. The boldface numbers indicate actual sums; the italicized numbers in parentheses give values expected on the basis of random associations of numbers of OAs and oral primordia assuming that the two are independently determined. For example, in series A the proportion of 1-OA is 52/66 or 0-787, while the proportion of 1-OP is 37/66 or 0-561; the proportion of joint occurrence of 1-OA and 1-OP assuming independence is 0-787x0-561, or 0-442. This leads us to expect that 0-442 of the total of 66 cells, or 29-2 cells, should have both 1-OA and 1-OP. (A) A culture grown in 1 % PPY to mid-log phase at 28 °C, then fixed. (B) A culture grown in 2 % PPY to mid-log phase at 28 °C followed by exposure to 39-5 °C for 5-5 h followed by return to 28 °C for 1-5 h, then fixed. (C) The aggregated results from tallies of various other samples, all in log phase at the time of fixation. short periods (not shown), does not in itself raise or lower the expression of doubleness of CVP sets. Second, expression of secondary OAs (but not of double CVP sets) is invariably very low in stationary phase. Third, expression of secondary OAs and of double CVP sets are mutually independent: the proportion of the four possible joint phenotypic classes (1OA/1 CVP, 1 OA/2 CVP, 2 OA/1 CVP, 2 OA/2 CVP) within every sample is close to what would be expected on the basis of random combinations of the frequency of expression of each separate trait. In general, except for the scarcity of secondary OAs in stationary phase, the similarity of expression of OA and CVP anomalies under the investigated conditions is more striking than any differences (a tendency for higher expression in 2 % PPY does not attain statistical significance). 176 M. JERKA-DZIADOSZ AND J. FRANKEL

Table 2. Percentage of CU-127 cells with two sets of organelles

Oral apparatus , Contractile vacuole pore (CVP) Dividing cells Non- Dividing cells Non- Anterior Posterior dividing , * > dividing Sample* (OA) (OP) Cellsf Anterior Posterior cellst A. (1 % PPY) 21 44 15 80 55 53 B. (2 % PPY) 42 53 31 86 70 74 C. Various 30 37 18 66 53 61 * See legend of Fig. 7. t Percentages apply to randomly chosen non-dividing cells from the same samples (A, B) or the same set of samples (C) used in the analysis of dividing cells.

(b) Expression in dividing cells Dividing CU-127 cells were analyzed in order to gain insight into the dynamics of propagation of the secondary oral structures and double CVP sets. The results of an analysis of two samples, plus the remaining collection of relevant data from numerous other samples, are tabulated in Fig. 7 according to membership in each of 16 possible combinations of OAs and CVP sets (it should be noted that a dot in the diagrams indicates a set of CVPs, not a single CVP). Fifteen of the 16 possible combinations have in fact been observed. OA and CVP patterns are largely independent of each other, and there are no invariant associations between the two division products for either oral or CVP patterns (cf. Figs. 2 and 5). Ascertainment of whether numbers of OAs and CVP sets are randomly associated in anterior and posterior division products demands a more searching analysis. First, the aggregate proportion of secondary OAs and double CVP sets in anterior and posterior division products separately was computed by summing across the appropriate columns in Fig. 7 and adjust- ing the totals to percentages; these percentages are shown in Table 2, where they are also compared to the corresponding percentages observed in non-dividing cells. It is evident that dividing cells manifest a higher expression of secondary oral primordia and generally also of double anterior CVP sets than do non- dividing cells. The relative excess of doubleness of newly formed structures (oral primordia, anterior CVP sets) suggest that newly formed structures may not always persist. The next step is to use the aggregate percentages of individual traits to compute the expected frequencies of random combinations, and then compare the expected to the observed proportions. This comparison is shown for the oral structures in Fig. 7 (observed values bold, expected numbers italicized and in parentheses). The association is clearly not random, as there is a highly significant excess of homogeneous combinations (1 OA/1 OP, 2 OA/2 OP) and a deficiency of heterogeneous combinations (1 OA/2 OP, Mirror-image duplication in Tetrahymena 177 2 OA/1 OP). The same is true, though in a less dramatic fashion, for CVP sets (not shown). This suggests some tendency to propagate the pre-existing state of expression. Finally, the striking asymmetry between the two heterogeneous combinations, with the 1 OA/2 OP class much more common than the 2 OA/1 OP class, is a reflexion of the fact that dividing CU-127 cells have a higher frequency of secondary oral primordia than of secondary OAs (Table 2). A similar excess of the 2 CVP(ant)/l CVP(post) class relative to the 1 CVP(ant)/ 2 CVP(post) class was observed. In a sample of 50 dividing cells of CU-127 (111.) fixed 16 months later all of the above findings, excepting only the excess of 2 CVP(ant)/! CVP(post) over 1 CVP(ant)/2 CVP(post), were confirmed. The observations summarized in Fig. 7 may thus be presumed to reflect constant characteristics of the CU-127 clone. In a different approach to the problem of continuity of states of expression, a single CU-127 culture was expanded into 30 subclones by selection of single celJs, with fixation and silver impregnation of the subclones 25 to 30 generations later. The expression of the OA and CVP abnormality was indistinguishable in these subclones.

(c) Expression during oral replacement Oral replacement is an alternative mode of oral development (Frankel & Williams, 1973), typically observed in T. thermophila during stationary phase (Kaczanowski, 1976) and during transformation to a 'rapid swimmer' phenotype (Nelsen, 1978). CU-127 cells also undergo oral replacement under both of these circumstances. The normal primary OA undergoes typical oral replacement with involvement of the undulating membrane (Frankel, 1969). In CU-127 cells undergoing the 'rapid swimmer' transformation, the abnormal secondary OA, which generally lacks an undulating membrane (see section 4) is replaced by an oral primordium that is formed adjacent to its right postoral meridian, generally some distance back from its anterior end. In some cases, transforming CU-127 cells may form an abortive division furrow on the side of the cell with the secondary oral axis while simultaneously undergoing typical oral replacement on the side with the primary oral axis. In contrast, CU-127 cells entering stationary phase generally undergo replacement only of the primary OA, while the secondary OA is simultaneously resorbed without being replaced. This is probably one reason for the low proportion of secondary OAs in stationary phase cells.

(d) Relation of secondary OAs and double CVP sets to cell size and number o ciliary meridians. CU-127 cells in predivision stages of oral development [stages 1-5, see Frankel & Williams (1973) and section 4 below] were scored for OA and OP number and cell length and width. The results, presented in Table 3, show that the expression of secondary oral structures is differently related to cell size and to 178 M. JERKA-DZIADOSZ AND J. FRANKEL

Table 3. Association between ciliary meridian number, cell size, and number of oral structures in CU-127 cells cultured in 2%PPY

Ciliary imeridian no. Cell length Cell width r configuration Mean ±S.D. n Mean ±S.D. n Mean ±S.D. n 1 OA-1 OP 22-56* ±0-71 32 47-25 ±2-51 32 3200 ±2-77 32 1 OA-2 OP 23-20* ±0-77 15 48-90 ±3-83 20 32-67 ±3-28 20 1 OA Total 47-88t ±3-16 52 32-26$ ±2-96 52 2 OA-1 OP 22-80* ±0-98 21 53-74 ±3-07 22 35-50 ±3-52 22 2 OA-2 OP 2305* ±0-97 19 53-05 ±3-86 22 34-56 ±4-59 22 2 OA Total 53-39| ±3-57 44 35-07$ ±4-07 44 * Ciliary meridian no., 1 OPv. 2 OP; t, = 2-44, 0-025> P > 001. t Cell length, 1 OA v. 2OA; / = 800, P < 0001 $ Cell width, 1 OA v. 2OA; / = 3-90, P < 0001 number of ciliary meridians. Cells with two OAs are significantly longer and wider than cells with one OA. Within each OA class there is no association between number of OPs and cell size. Increased size is therefore at most a consequence, and not a cause, of stomatogenic activity along the secondary oral axis. Cells with two OAs did not, in this experiment, have a significantly greater average number of ciliary meridians than cells with one OA. Instead, cells with two oral primordia had a number of ciliary meridians that was slightly but significantly greater than cells with one OP. Hence, development of a secondary OP may sometimes be associated with a gain of one or occasionally more ciliary meridians. Other data (Table 5) suggest that cells double for both OAs and CVP sets have (on the average) about one more ciliary meridian than cells that are single for both characteristics.

3. Cytogeometry of CU-127 (a) Location of the secondary OA The secondary OA, when present, appears to a first approximation to be located 180° opposite to the primary OA (Fig. 1). A more careful examination indicates that this is not strictly true. To quantitate the relative positions of the OAs, we will follow a convention whereby the primary oral axis is operationally defined by the right postoral ciliary meridian of the primary OA, and the secondary oral axis is similarly defined by the right postoral meridian of the secondary OA. If one assumes equal spacing of ciliary meridians, then cells in which the two OAs are located precisely opposite each other on the cell circumference should be separated by an equal number of ciliary meridians on both sides. For Mirror-image duplication in Tetrahymena 179

Table 4. Relative location of primary and secondary OAs in CU-127 cells with 2 OAs

Ciliary meridian number Number A of Total Right* Left* cells 20 10 10 1 21 10 11 2 22 10 12 6 11 11 7 23 10 13 2 11 12 38 24 11 13 35 12 12 24 25 11 14 2 12 13 12 26 13 13 1 The data in this table are based on those of the cells from experiments 1 and 2 (cf. Table 5) in which two OAs are present and the position of both can be ascertained. * 'Right' and 'left' refer to distances between the primary and secondary oral axes measured in intermeridional intervals to the right and left respectively of the primary oral axis. For conventions of measurement, see the text.

example, in a cell with 24 ciliary meridians, the secondary oral axis should be 12 intermeridional spaces distant from the primary axis in both directions. Such equality is observed in slightly fewer than 50 % of the cells with an even total number of ciliary meridians (Table 4). In the remainder, the distance in ciliary meridians between the primary and secondary oral axis is less when measured to the right of the primary axis than when measured to the left; for example, most cells with 24 ciliary meridians have 11 intermeridional intervals between the two oral axes measured to the right of the primary axis, and 13 intermeridional intervals between the two oral axes measured to the left of the primary oral axis (Table 4). In cells with an odd total number of ciliary meridians, equality of distance between oral axes on the two sides is inherently impossible, and in these cases all cells manifest a slightly shorter distance between oral axes to the right of the primary axis than to the left of the primary axis (Table 4). An opposite inequality (shorter distance to the left of the primary axis than to the right) was never observed. In the experimental series summarized in Table 4 most cells with 2 OAs had 11 or 12 intermeridional intervals on the side of the cell to the right of the primary oral axis (the side bearing the CVPs). In the much later examination of CU-127 (111.), when virtually all cells possessed 22 ciliary meridians (cf. Table 7), 90 % of the cells with 2 OAs manifested 10 intermeridional intervals on the side of the cell to the right of the primary oral axis, 12 on the left. It thus appears that some rather precise geometric instructions 180 M. JERKA-DZIADOSZ AND J. FRANKEL

Table 5. CVP cytogeometry of CU-127

Number of Position of CVP sets* Ciliary merid. Expt. Culture CVP Sample Line no. conditions OAs sets Mean +s.D. CVP1* CVP CVP2* size CU-127 la % PPY-log-28c 22-66 ±1-20 79-8" 24 CU-127 lb % PPY-log-t 22-50 ±0-62 84-5° 18 CU-127-1 2a % PPY-log-28c 23-23 ±0-65 76-7° 26 CU-127-1 2b -log-t 22-56 ±1-22 861° 25 CU-127 la %PPY- log-28° 23-52 ±1-13 84-8° 17 CU-127 lb %PPY-•log-t 22-40 ±1-13 76-1° 15 CU-127-1 2a % PPY •Iog-28C 24-14 ±0-37 85-8° 7 CU-127-1 2b %PPY- log-t 23-75 ±0-44 83-8a 16 CU-127 la % PPY-log-28° 2 22-96 ±1-13 62-9° (89-7")§ 116-5° 31 CU-127 lb % PPY-•log-t 2 22-82 ±0-86 52-7° (821°) 111-5° 28 3 CU-127-1 2a % PPY-•log-28° 2 23-76 ±0-60 60-9° (85-6 ) 110-4° 25 CU-127-1 2b % PPY- log-t 2 23-56 ±0-58 54-5° (83-3°) 112-1° 25 CU-127 la % PPY-log-28° 2 23-90 ±0-75 56-3° (79-93) 103-5° 22 CU-127 lb %PPY- log-t 2 23-17 ±0-86 52-2° (790°) 105-8° 28 CU-127-1 2a %PPY- log-28° 2 24-09 ±0-94 61-4° (86-6°) 111-8° 11 CU-127-1 2b X PPY- log-t 2 23-84 + 0-37 55-3° (82-3°) 109-3° 19 * For conventions of measurement, see Fig. 8 and text. t Grown to mid-log phase at 28 °C, followed by exposure to 39-5 °C for 5-5 h, followed by return to 28°Cfor 1-5 h, then fixed. t Grown to mid-log phase at 28 °C, followed by exposure to 39-5 °C for 4 h, then fixed. § Midpoint between two CVP sets indicated in parentheses. See Fig. 8.

must govern the positioning of the secondary OA relative to the primary OA such that the secondary OA is seldom placed much more or less than 170° to the right of the primary OA.

(b) Arrangement of CVP sets One of the unique features of the CU-127 clone is the location of its two CVP sets on the cell circumference. Positions of CVP sets analyzed in a set of four experiments are tabulated according to numbers of OAs and of CVP sets present in the cell (Table 5). The measurements are expressed in terms of the central angle convention of Nanney (1966a). Computation of this angle is made with reference to a plane extending from the central longitudinal axis of the cell to the right postoral meridian of the primary oral axis. The 'central angle' is measured at the intersection of this plane and a horizontal line extending from the central axis to the midpoint within the CVP set (for conventions used in ascer- taining this midpoint, see Nanney, 1966a). The central angle is thus a measure of the position of each CVP set on the cell circumference relative to an oral axis. Three further remarks must be made about this measure, (a) Measurement Mirror-image duplication in Tetrahymena 181

OA-2 OA-1

OA-1

CVP-?*-—1—•" CVP-1 OA-1

CVP-2 CVP-1 CVP (CVP) Fig. 8. Diagrammatic representation of various geometric possibilities of positions of OAs and CVPs in CU-127 cells. The dashed vertical line in the diagram of the cell represents the central longitudinal axis of rotation. A horizontal circle embracing the cell circumference is shown in the posterior region of the cell, and is projected outwards to diagrams of four configurations of OAs and CVPs. These four configurations, from left to right, correspond to the four classes from top to bottom (1 OA-1 CVP, 2 OA-1 CVP, 1 OA-2 CVP, 2 OA-2 CVP) in Table 5. The labelling of OAs and CVPs follows the same conventions as in Tables 5 and 7, and the angles a and /? are those tabulated in Table 7 (Group 1).

is to the right of the primary oral axis, (b) Measurement of surface position of CVP sets is made by counting ciliary meridians, not by measuring actual surface distances, so the unit of distance is thus intermeridional intervals, not micro- meters. As ciliary meridians appear to be uniformly spaced around the cell circumference, the former is proportional to the latter, (c) The 'central angle' convention is purely descriptive, and carries no mechanistic implications other than indicating that there is some general proportionality of positioning (see Nanney, 1966 a). All CVPs in CU-127 cells are situated in the space between the two oral axes that is to the right of the primary axis. As indicated earlier, either one or two sets of CVPs are present within this space. In CU-127 cells with only one set of CVP's, the average position of the midpoint of that set with respect to the plane defined by the primary oral axis is 82°, or 23 % of the entire cell circumference (Table 5, Fig. 8). This is very close to the 75-80° (20-22 %) characteristic of normal T. thermophila (Nanney, 1967b\ Nanney et al. 1975; Frankel, 1972). In CU-127 cells with two sets of CVPs, the average midpoint of the set closer to 182 M. JERKA-DZIADOSZ AND J. FRANKEL

Table 6. Position of newly established CVP sets

Position of CVP sets Number Sample Type of cell of OAs CVP1 CVP 2 size Anterior component of divider 1 51-4° 109-8° 28 (51-4°) (109-3°) (23)* 2 45-0° 100-9° 12 Non-dividing cell (control) 1 50-6° 100-6° 28 2 51-3° 98-4° 13 * The subset in which the posterior as well as the anterior component of the dividing cell has only one OA. The data in this table are based on a single sample of CU-127 (111.) that was grown at 28 °C in a flask culture of 1 % PPY and fixed in mid-log phase. Only cells with 2 CVP sets are included. the primary oral axis (CVP set 1) is situated at a position described by a central angle of about 55°, and that of the other (CVP set 2) at an angle of about 110°, both measured to the right of the primary oral axis (Table 5, Fig. 8). The mean halfway point between these two CVP sets (Table 5, values in parentheses) is about the same as the average position of the CVP midpoint in those cells with only one CVP set. It should be stressed that the average CVP positions tabulated in Table 5 and depicted in Fig. 8 do not faithfully represent the positions encountered in each individual CU-127 cell. The distribution of locations of CVP sets is quite broad. When there are two CVP sets they may be close together or far apart. When there is only one CVP set, it may be positioned about halfway between the two OAs or it may be positioned off-center, occasionally even at a location typical for either CVP set in those cells that have two sets. However, the distribution of positions of CVP sets in cells with one CVP set is broadly monomodal, with 76 % of the CVP sets situated between 65° and 95° to the right of the primary oral axis. In cells with two CVP sets, 74 % of these sets are located outside of these limits. Hence the geometrical construct of Fig. 8, though idealized, is not artifactual (see also Fig. 5). Both the means and distributions of CVP locations indicate that the positions of CVP sets are unrelated to whether or not a secondary OA is present (Table 5). However, the measurements described thus far were made on cells with established CVP sets, and it is possible that sets of CVPs initially formed in a cell with two OAs might subsequently come to reside in a cell with only one OA, either by resorption of the secondary OA or by original formation of CVPs in an anterior division product with two OAs followed by subsequent segregation of these CVPs at the next fission into a posterior division product of a cell that had formed only one oral primordium. For this reason, when CU-127 (111.) was examined, the CVP cytogeometry of dividing as well as of non-dividing cells was Mirror-image duplication in Tetrahymena 183

Table 7. Relational geometry of CVPs in CU-127 cells with 2 OAs and 2 CVP sets

CVP distances A Ciliary OA 1 to CVP 1 OA 2 to CVP 2 m p r i H — 111CI lii A \ 1 A ian Sample Mean Mean Group* no. size distancef S.D. at distancef S.D. fit a < p a = fi a> fi 1 21 1 2-75 4-25 — ] — — 22 4 3 04 ±0-58 49.70 3-50 ±0-95 57-2° 2 1 1 23 21 3-41 ±0-55 53-4° 417 ±0-76 65-3° 15 2 4 24 34 3-63 ±0-50 54-5° 4-30 + 0-96 64-5° 19 5 10 25 10 3-80 ±0-92 54-7° 4-65 ±0-96 69-8° 7 2 1 23-69 70 3-54 53-8° 4-26 64-7° 44 10 16 11 20 1 2-75 3-25 — 1 — — 22 32 2-81 ±0-63 46-0° 3-86 ±0-50 63-1° 28 1 3 24 1 4-75 — — 3-75 — — — — 1 22-00 34 2-87 46-9° 3-84 62-8° 29 1 4 * Group I includes those of the 2 OA-2 CVP set cells from expts. 1 and 2 (cf. Table 5) for which complete data on positions of both OAs and CVP sets are available. Group II includes the 2 OA-2 CVP set cells from the CU-127 (III.) sample (cf. Table 6). In Group 11 the anterior and posterior components of dividing cells are included together with the non-dividing cells. Group II was fixed about 16 months later than Group I. t CVP distance as measured in intermeridonal intervals (to obtain CVP positions in the sense of Nanney (1966a, 19676), add 1 to the CVP distances). t 'Central angles' defined with reference to the right postoral meridians of OA 1 (a) and OA 2 (fi). For clarification, see Fig. 8 and text.

assessed. The major question is whether, in those anterior moieties of dividing cells which possess two CVP sets, the position assumed by these sets is affected, by the presence or absence of oral structures along the secondary axis. The results, presented in Table 6, show that (a) the average position of the two CVP sets is similar in prospective anterior division products with newly formed CVPs to what it is in non-dividing cells with established CVPs and (b) the location of the two newly developed CVP sets is similar in cells with one and two sets of oral structures. Secondary oral structures thus exert a minimal direct influence, if any, on the positioning of the CVP sets to the left of the secondary oral axis. If one switches one's frame of reference from the primary to the secondary oral axis, one can construct a reference plane extending from the central cell axis to the right postoral ciliary meridian of the secondary OA, and compute the position of CVP sets to the left of that plane. When one examines the position of the single CVP set of cells with 2 OAs and 1 CVP set in this manner, the average distance of the CVP midpoint to the left of the secondary oral axis is only slightly greater than the distance measured to the right of the primary oral axis (data not shown). When a similar calculation is carried out for cells 184 M. JERKA-DZIADOSZ AND J. FRANKEL

9 a 9b Fig. 9 (a) A protargol-stained cell showing the primary OA and OP. M-l, M-2, and M-3 indicate membranelles 1, 2, and 3 respectively. Note the posterior tapering of M-l. UM = undulating membrane. The single macronucleus is seen out of focus in the center of the cell, the dividing micronucleus is on the viewer's right. The bar represents 10 /tm. (b) The arrangement of basal bodies in a normal, completed oral apparatus. Redrawn from McCoy (1974). The bar represents 10 fim. with 2 OAs and 2 CVP sets, it becomes apparent that CVP 2 is situated at a somewhat larger distance to the left of the secondary oral axis (angle /?, Fig. 8) than is CVP 1 to the right of the primary oral axis (angle a, Fig. 8) (Table 7). The data in Table 7 are consistent with the assumption that positioning of CVPs in CU-127 as in wild-type cells takes place through measurement of relative rather than absolute distances, so that the 'central angles' remain the same as number of ciliary meridians vary. This assumption received further support from a limited analysis of progeny clones homozygous for the 'Janus' allele that determines the unique geometrical confirmation described here but allows considerable variation in number of ciliary meridians (Frankel & Jenkins, 1979); janus cells with as few as 18 or as many as 31 ciliary meridians differ little if at all from each other or from CU-127 in the relative placement of OAs and CVPs. Mirror-image duplication in Tetrahymena 185

4. Secondary oral structures (a) Structure and function of secondary OAs As mentioned earlier (Results, section \b), while primary OAs of CU-127 cells are indistinguishable from OAs of WT cells (Fig. 9), secondary OAs are structurally abnormal. The abnormalities consist of varying combinations of three basic defects: (a) absence of oral components, typically the undulating membrane (see Fig. 9 b for normal components and their arrangement), (b) reduced size or apparent fragmentation of components, and (c) reversed asymmetry of components. A characteristic, though by no means universal, combination of defects is present in the secondary OA shown in Fig. 1, in which the undulating membrane is absent and the three membranelles manifest the reverse of the usual right-left asymmetry, i.e. they slant from the anterior-left to the posterior-right rather than the typical anterior-right to posterior-left (cf. Fig. 9). A few abnormal OAs manifest an undulating membrane at the left margin (Fig. 10) rather than the right margin as is normal. However, some secondary OAs are nearly normal, whereas others are very tiny, with only one or two membranellar fragments, suggestive of ongoing resorption. The generally reduced size and commonly encountered abnormal configuration of secondary OAs makes it relatively easy to distinguish primary and secondary OAs in living cells, even when observed under bright field optics. When fed with carmine powder in 1 % PPY medium, CU-127 cells with two OAs formed food vacuoles only with their primary OAs. The membranelles of secondary OAs appeared stiffly outstretched and did not beat properly. Secondary OAs sometimes do possess a buccal cavity (Fig. 1) but lack the deep fibers described by Nilsson & Williams (1966). It is not known whether they possess a food vacuole forming region.

(b) Development of secondary oral primordia A secondary OP initially appears as a field of basal bodies to the left of a ciliary meridian (Fig. 10) or midway between two ciliary meridians (Fig. 11). If an anterior secondary OA is also present, the ciliary meridian(s) in question can be identified as the right postoral meridian (Fig. 10) or the right and left postoral meridians for the above two situations respectively (the primary OP, like the OP of WT cells, appears to the left of the right postoral meridian). In no case did a secondary OP arise adjacent to the right side of a ciliary meridian, and in all CU-127 cells the ciliary meridians were of the usual polarity and asymmetry, with longitudinal microtubule bands to their right (Figs. 10, 11) and new basal bodies forming anterior to old ones (Figs. 10-12). When a secondary OP was present it appeared at the same cell latitude as did the primary OP. Except for their common tendency to appear midway between two ciliary meridians, there was nothing atypical revealed by light microscopy about the position or 186 M. JERKA-DZIADOSZ AND J. FRANKEL Mirror-image duplication in Tetrahymena 187 geometry of secondary oral primordia at early stages of their development. The secondary oral primordia were, however, invariably smaller than their primary counterparts. Oral primordia in late stages of differentiation often manifested reversed right-left asymmetry. This can be seen by comparing protargol stained stage-5 secondary oral primordia (Figs. 12-14) with a similarly stained primary OP fixed at an equivalent stage (Fig. 9 a). Reversed membranelles are seen in Fig. 12, and reversed undulating membranes in Figs. 13 and 14. The reversed undulating membrane in Fig. 14 bears a single row of cilia as does a normal undulating membrane (Nilsson & Williams, 1966), and the cilia of that row are short, as is characteristic of normal stage-5 oral primordia (Buhse, Stamler & Corliss, 1973; Ruffolo & Frankel, 1972). The subtler details of the reversed oral primordia, less clearly evident on the photographs, suggest that the secondary oral primordia, though defective, are true mirror-images of the complete primary oral primordia. The cilia of the undulating membrane shown in Fig. 14 appear on close examination to be attached to the outermost of the two staggered rows

FIGURES 10-14 Protargol-stained preparations of CU-127 cells, focused on the secondary oral primordia. All photographs are printed in the same orientation, with the cell anterior at the top of the page and viewer's right corresponding to cell's left. In each micrograph the bar represents 10 fim. Fig. 10. A cell with a secondary OA and an incipient (stage 1) oral primordium (OP). Only the undulating membrane (UM) of the anterior OA is in focus. The OP is situated to the cell's left of the right postoral ciliary meridian. Note that new unciliated basal bodies within ciliary meridians near tbe OP are situated immediately anterior to old ciliated ones (arrows). Longitudinal microtubule bands are faintly visible to the right (viewer's left) of the ciliary meridians. Fig. 11. A cell lacking a secondary OA, with a stage-1 secondary OP (thick arrow) situated midway between two ciliary meridians. Thin arrows point out longitudinal microtubule bands situated to the right of the two ciliary meridians on either side of the OP. Fig. 12. A cell with a differentiated (stage late-4) secondary OP. The two upper arrows show the tapering of the first membranelle, with three rows of basal bodies at the left-anterior end, two rows at the right-posterior end. The ciliary meridian to the cell's right of the OP is interrupted by the membranelles, while the meridian to the left of the OP remains continuous. The posterior arrow indicates a new unciliated basal body anterior to an old ciliated one. Fig. 13. A portion of a cell with a secondary OP. Although only tiny fragments of membranelles remain, the well differentiated state of the undulating membrane (TJM) and the equatorial discontinuities in the ciliary meridians indicate that this is an advanced (stage 5) oral primordium, possibly undergoing resorption. Note the diagonal couplets of basal bodies within the UM, and the posterior continuity of the UM with the ciliary meridian situated to the cell's left of the remainder of the OP. Fig. 14. A portion of a dividing cell, with a secondary OP. The undulating membrane (UM) bears a single row of short cilia. Ciliated membranelles (out of focus) are visible to the right of the UM. 188 M. JERKA-DZIADOSZ AND J. FRANKEL

e/ ^ • f- *•*•<&

y •••••• •••

15 Mirror-image duplication in Tetrahymena 189 of basal bodies, mirroring the analogous configuration of undulating membranes of normal oral primordia. The third, outermost row of basal bodies within the first membranelle terminates at a somewhat more anterior position than do the other two rows (Fig. 9 b). This feature is readily observed in favorable protargol preparations of stage-5 primary oral primordia (it is responsible for the 'taper' of the first membranelle of the OP in Fig. 9 a) and is present in mirror-image in those stage-5 secondary OPs that have well-developed first membranelles (Fig. 12). Examination of the geometrical configurations of basal bodies and cilia during intermediate stages of development gives further indications of mirror asymmetry of development of secondary oral primordia. During normal oral development basal bodies within the OP become associated into couplets (McCoy, 1974) that later organize into membranelles (initially double rows of basal bodies) beginning at the anterior-left margin of the OP (Fig. 15). In a secondary OP the couplets are formed but their association into membranelles tends to occur in a less regular fashion than in a normal OP; however, there is a definite tendency for differentiation of membranelles to begin at the anterior- right margin of the OP, with progression posteriorly and later to the left - the mirror image of the normal geometry of membranelle assembly (Fig. 16). Examination of ciliary outgrowth with scanning electron microscopy adds further evidence for a geometrical pattern of development of secondary oral primordia that often approximates a mirror image of the normal pattern. We have confirmed in primary oral primordia the observations of Buhse et al. (1973) that the three constituent basal body rows of membranelles sprout cilia in an ordered sequence. Cilia grow out first on the innermost row, then the middle row, and finally the outermost row of each membranelle. Figure 17(6) shows an early stage of this process (at a time when there are still only two rows of basal bodies within each membranelle) in a primary OP. The innermost, right row of cilia within each membranelle is partly grown out, while the outer, second row to its left consists only of short stubs. Figure 18(6) illustrates a secondary OP in a similar stage of development: again, one sees nearly longitudinal rows of partly grown cilia adjacent to rows of ciliary stubs, but now each row of stubs

Figs. 15 and 16. Photographs and drawings of protargol preparations of CU-127 cells focused on oral primordia in an early stage of formation of oral membranelles (stage 3). Photographs are shown above camera lucida drawings of the same preparations. Orientation is the same as in Figs. 10-14. In each illustration the bar represents 10 /tm. Fig. 15 (a) A photograph of a stage-3 primary OP. The arrow indicates the region of initiation of membranelle differentiation, (b) A camera lucida drawing of the same preparation. Fig. 16 (a) A photograph of a stage-3 secondary OP. The arrow indicates the region of probable initiation of membranelle differentiation, (b) A camera lucida drawing of the same preparation.

13 E M B 49 M. JERKA-DZIADOSZ AND J. FRANKEL Mirror-image duplication in Tetrahymena 191 is to the right of the adjacent row of cilia. This same mirror-image pattern of ciliary outgrowth in primary and secondary oral primordia is evident at a later stage, after the third row of basal bodies has formed within each membranelle (cf. Williams & Frankel, 1973) and the third, outermost row of cilia is growing out. In a primary OP this outermost row of stubby cilia is seen on the left- anterior margin of each developing membranelle (Fig. 19); in a secondary OP this short outermost row is on the right-anterior margin (Fig. 20). The pattern of ciliary outgrowth in secondary oral primordia is somewhat less clear than in primary primordia mainly because the membranelles are smaller and closer to one another in the secondary primordia, with the third membranelle often missing. One further aspect of this geometric reversal of oral differentiation in secondary oral primordia becomes evident when one examines the relationship of the developing oral primordia to the surrounding ciliary meridians. As mentioned earlier, secondary as well as primary oral primordia typically (though not invariably) begin their development to the left of a 'stomatogenic' ciliary meridian. As differentiation of membranelles proceeds, normal oral primordia tend to remain closely associated with that ciliary meridian (Fig. lid). At later stages the portion of that meridian which is adjacent to the oral primordium loses its cilia, merges with the primordium, and presumably participates in the differentiation of the undulating membrane (Figs. 19, 21 a). Simultaneously, the differentiating membranelles interrupt the continuity of the ciliary meridian immediately to the left of the OP (Fig. 21 a). The evolving relationship between secondary OPs and. the adjacent ciliary meridians is quite different. The differentiating secondary OP becomes dissociated from the 'stomatogenic' meridian to its right (Fig. 18 a). Later, if substantial membranelles are formed, these membranelles interrupt the continuity of the meridian to its right (Figs. 12, 21 b)

FIGURES 17 AND 18 Scanning electron micrographs of CU-127 cells, showing oral primordia in an early stage of membranelle formation (stage early-4). Whole cells are shown above, enlargements of their oral primordia below. Orientation of cells is as in the preceding figures. The bars in the upper micrographs represent 10/tm, those in the lower micrographs represent 1 /tm. Fig. 17 (a) A cell with a primary OA and a stage early-4 primary OP. Note the close association of the OP with the adjacent right postoral ciliary meridian, (b) Enlarge- ment of OP. Three membranelles are developing in the anterior-left portion of the OP. Each membranelle has one row of partly-grown cilia and one row of ciliary stubs (arrows) to its anterior-left. Fig. 18 (a) A cell with a stage early-4 secondary OP, and lacking a secondary OA. Note the absence of close contact of the OP with the ciliary meridian to its right. An associated CVP set is seen near the posterior end of the cell (arrow), (b) Enlargement of OP. Membranelles are developing on the right side of the OP. Each membranelle has one row of partly-grown cilia plus one row of ciliary stubs (arrows) to its right. Developing membranelles are more closely packed together than in a normal OP. 13-2 192 M. JERKA-DZIADOSZ AND J. FRANKEL Mirror-image duplication in Tetrahymena 193 in much the same way that the membranelles of the primary OP interrupt the continuity of the meridian to its left. The comparison of Figs. 21 (a) and (b), representing primary and secondary oral axes of the same cell, indicates this particularly clearly. The configuration on the other (left) side of the secondary OP is more variable: in a few cases the meridian to the left of a secondary OP abuts on a reversed undulating membrane (Fig. 13) much as the meridian to the right of the primary OP does on the normal undulating membrane; more typically, no well organized undulating membrane is produced by the secondary OP, and the ciliary meridian to the left of the OP tends to remain uninterrupted (Figs. 12,216). It should be emphasized that the above account has deliberately emphasized the cases in which development of the secondary OP is most complete and its reversed character best manifested. Not all secondary oral primordia are of this type. A tally of 36 silver impregnated cells with advanced (stage 5) primary oral primordia that also possessed secondary oral primordia revealed that 14 of the secondary oral primordia were right-left reversed much as described above, 4 were nearly normal (not reversed), 1 was ambiguous (with an undulating membrane on both sides), and 17 were highly incomplete or defective, suggesting either retardation in development or ongoing resorption. An example of a secondary oral primordium that would be scored as nearly normal is shown in Fig. 22, while Fig. 23 indicates an 'ambiguous' primordium with nearly transverse

Figs. 19-20. Enlarged scanning electron micrographs of CU-127 oral primordia at a late stage of membranelle formation (stage early-5). Orientation is as in preceding figures. Bars represent 1 /*m. Fig. 19. A primary OP in stage early-5. The three membranelles (M-l, M-2, and M-3) possess two rows of partly grown out cilia apiece, plus one row of ciliary stubs (open arrow) on the outermost (anterior-left) margin of each membranelle. The undulating membrane (UM) is undergoing alignment of its single ciliary row. Fig. 20. A secondary OP in stage late-4 or early-5. Only the first membranelle (M-l) is fully developed, and the second membranelle (M-2) is short. Ciliary stubs (open arrow) are situated on the anterior-right margins of these membranelles. The cilia on the left side of the OP (right of micrograph) are not organized in the configuration typical of a UM, and may instead represent an abnormal membranelle. Fig. 21. Two sides of the same CU-127 cell oriented as in preceding figures, (a) The side with the primary oral axis. A normal OA is present. The OP is in stage late-4. Membranelles are well differentiated and the UM is undergoing alignment of its cilia. The two arrows indicate the anterior and posterior ends of the portion of the left postoral ciliary meridian that is interrupted by the membranelles of the OP. (b) The same cell rotated to reveal the secondary oral axis. An OA is lacking. The OP is highly abnormal, with one irregular reversed membranelle situated between the arrows that delineate the portion of the ciliary meridian to the right of the OP that is interrupted. Note that the ciliary meridian to the left of the OP (extreme right of the photograph) bends around the OP without any interruption. Bars represent 10 /*m. 194 M. JERKA-DZIADOSZ AND J. FRANKEL

FIGURES 22 AND 23 Enlarged scanning electron micrographs of secondary CU-127 oral primordia at late stages of membranelle formation. Orientation is as in preceding figures. The bars represent 1 /im. Fig. 22. An OP with three small membranelles (M-l, M-2, and M-3) that are approximately transverse in orientation. The clump of short cilia on the right side of the OP might represent a partly organized undulating membrane (UM). Fig. 23. An OP with two membranelles (M-l, M 2), oriented somewhat more trans- versely than is normal. Remaining cilia are scattered on both sides of the posterior portion of the primordium, suggesting ambivalence in positioning of the undulating membrane. membranelles and indications of undulating membrane development on both sides. In many cases in which secondary oral primordia were small or development was apparently arrested, ciliary meridians to the right and left of the oral primordia both remained uninterrupted. Whereas only a small proportion of differentiating secondary oral primordia are geometrically nearly normal, at least one-half of the fully developed secondary OAs have membranelles with normal or near-normal orientation. This difference is probably due mainly to selective resorption of the most abnormal oral structures; however, some reorientation of membranelles may Mirror-image duplication in Tetrahymena 195 also take place during late stages of oral development. Advanced oral primordia with 'ambiguous' orientation, as in Fig. 23, may be undergoing such reorientation.

DISCUSSION (A) Demonstration of a reversed cortical field The CU-127 clone is characterized by unique abnormalities in the structure of oral apparatuses (OAs) and in the positioning of contractile vacuole pores (CVPs) juxtaposed on structurally normal ciliary meridians. We will argue that the characteristic abnormalities of this clone are manifestions of a propagated morphogenetic field of reversed asymmetry. Abnormalities of the pattern of oral development in Tetrahymena that have previously been observed (Frankel, 1964; Frankel, Nelsen & Jenkins, 1977 and unpublished; Kaczanowski, 1976; Nelsen, 1970) did not manifest any change of the fundamental asymmetry of the OA. In CU-127 cells, however, oral primordia (OP) frequently develop with a clear reversal of left-right asymmetry. Although such oral structures were not analyzed at the ultrastructural level, examination of features such as the relative lengths of oral membranelles and of ciliary rows within membranelles suggests that the reversal is a true situs inversus; i.e. the abnormal oral primordia cannot be rotated so as to superimpose them on the normal ones. Although the basis of this situs inversus is itself unknown, it is independent of the structural asymmetry of the ciliary meridians. The ciliary meridians adjacent to the reversed oral primordia are characterized by normal configurations of accessory microtubule bundles, and by normal polarity of basal body proliferation. A further symptom of the normal morphogenetic influence of these ciliary meridians is that oral primordia destined to become abnormal initially develop adjacent to the normal (left) side of a 'stomatogenic meridian', and only become reversed during the later development of membranelles. The reversed oral primordia are never complete, and there are intermediate cases that are suggestive of a conflict between two incompatible systems of pattern control. The resulting abnormal OAs are then readily resorbed. The resorption is probably a non- specific response to abnormality of oral structure (cf. Frankel, 1964; Nelsen, 1970). CU-127 cells produce both abnormal OAs and also normal ones that are in no known way different from OAs of wild type cells. The two types of OAs are spatially segregated from one another. The fully normal OAs are generated along one longitudinal axis, here called the 'primary axis'. Stomatogenesis never fails along this axis, and the OAs produced there are always normal. Abnormal OAs are generated only along a 'secondary axis' that is almost - but not quite - opposite to the primary axis. The secondary axis is visibly abnormal in two ways, first in the nature of the OAs produced there (never completely normal, frequently reversed), and second in the sporadic manifestation of stomatogenic activity. Oral structures may be completely absent along 196 M. JERKA-DZIADOSZ AND J. FRANKEL the secondary axis of a dividing cell, or there may be an OA but no OP, or an OP without an OA, or both an OA and an OP. The production of oral structures along the secondary axis appears to reflect a dynamic state, with a probability of stomatogenesis during division that is always less than one, and a probability of resorption of previously formed OAs that probably depends on the degree of abnormality, and perhaps also the age, of the OA. Although the expression of oral development along the secondary axis is variable, the location of that axis as indicated by the circumferential position of OAs that do develop is remarkably constant. This suggests the existence of a continuously propagated morphogenetic field which underlies the secondary structures. The mode of expression of double CVP sets provides further evidence for the constant existence of such a field, and the geometry of this expression as well as that of the secondary OAs indicates that this field is reversed in CU-127 cells. A clear understanding of this conclusion requires a comparison of CVP cytogeometry in CU-127 and wild-type cells. In wild-type T. thermophila (= T. pyriformis, syngen 1) the position of the CVP midpoint is at an angle of close to 80° to the right of the oral axis as defined by the right postoral ciliary meridian (Nanney, 19676; Frankel, 1972). When the position of the oral axis undergoes ' slippage' the location of the CVP set undergoes a corresponding shift (Nanney, 1967a). This geometrical association, as well as the fact that CVP positioning involves assessment of & proportionate rather than an absolute distance around the cell circumference, led to the notion that oral and CVP positioning are coordinately controlled by a morphogenetic field of some kind (Nanney, 1966a, 1967a). Homopolar doublets of genotypically wild type cells behave as two separate (Nanney, 1966a) though not entirely independent (Nanney et al. 1975) domains with respect to CVP positioning, with each component 'semicell' of the duplex positioning its CVP set as if it were an entire singlet cell. The angle between each of the two similar oral axes of a typical Tetrahymena doublet and the corresponding CVP midpoint is thus close to 40° of the approximately 180° width of the semicell, rather than 80° of the 360° circumference of the duplex (Fig. 24, top). The CVP cytogeometry of CU-127 cells with one CVP set is remarkably similar to that of wild-type cells, provided that one ignores the secondary OAs (Fig. 24, bottom left). However, CU-127 'doublets' (cells with two OAs and two CVP sets) manifest a remark- able difference from wild-type doublets: in CU-127 cells there is one CVP set associated with the primary OA as in a typical doublet, and a second CVP set that is associated with the secondary OA, but invariably to its left rather than its right (Fig. 24, bottom right). The reversed direction of CVP determination in CU-127 doublets suggests that these cells might have a morphogenetic field underlying the secondary oral axis that is of reversed, handedness. We are now in a position to demonstrate why the postulated morphogenetic field underlying the secondary oral axis must be continuously propagated in a manner that is independent of its expression. First, the expression of double Mirror-image duplication in Tetrahymena 197

Singlet. Doublet

CVP-2

OA OA-2 OA-1

CVP-1

CVP

OA-1 OA-1 (OA-2)

CVP-1 CVP-2 CVP Fig. 24. Comparison of CVP cytogeometry of wild type (WT) and CU-127 cells with one and two sets of CVPs. Illustrative conventions are as in Fig. 8. Configurations of WT cells are shown in the top half of the diagram, those of CU-127 cells in the bottom half; cells with single CVP sets on the left, those with double sets on the right. The angles indicated are averages for cells with a total of 24 ciliary meridians. Data for W.T. are those of Nanney et al. (1975) for 13 +13 row doublets extrapolated to 12 + 12 row doublets; data for CU-127 are from Table 7, with the angle between OA 2 and CVP 2 modified as indicated in Fig. 25 and accompanying text.

sets of CVPs and of secondary OAs are largely independent of each other (Results, section 2a). Second, the positions of CVP sets, whether one or two, are the same irrespective of whether or not secondary OAs are present (Results, section 3 b and Tables 5 and 6). These two observations rule out any direct influence of secondary OAs themselves on positioning of CVPs and strengthen the conclusion that the secondary OAs and the associated CVP positions are both separate manifestations of a reversed underlying morphogenetic field. A corollary of this conclusion is that the reference point(s) for the cell longitudes at which CVPs and probably also new oral structures are positioned are not the OAs themselves, though, spatially correlated with them when they are present. Is the mutual arrangement of these two morphogenetic fields completely symmetrical? The results summarized in Table 7 suggest a negative answer, as the average distance from the secondary oral axis to the midpoint of the second CVP set is greater than the average distance from the primary oral axis to the midpoint of the first CVP set. However, in the Results (section 3 b) the oral axes 198 M. JERKA-DZIADOSZ AND J. FRANKEL

—/OA-1

Fig. 25. The bilateral symmetry of CU-127 cells with two OAs and two CVP sets. (a) A schematized polar projection of a CU-127 cell with 23 ciliary meridians, 2 OAs, and 2 CVP sets viewed from above so that the anterior end is in the center. The primary oral axis is on the viewer's right side of the diagram, the secondary oral axis on the viewer's left. Oral primordia are shown undergoing membranelle development. Ciliary meridians are represented as rows of dots (basal bodies) radiating from the center (cell apex). The average positions of the midpoints of the two CVP sets on the cell circumference are indicated by blackened circles. The two oral reference axes are shown as dashed lines, the primary axis along meridian no. 1, the secondary axis between meridians 11 and 12 (see text), (b) A more highly schematized representation of the bilateral symmetry of the cell shown in (a). The two oral axes and the positions of the two CVP sets are shown. The plane of bilateral symmetry is indicated by the heavy vertical line. were arbitrarily defined by the positions of the right postoral meridians. If we instead consider the actual course of development in the two OAs, the right postoral ciliary meridian of the secondary oral axis is not equivalent to the corresponding ciliary meridian of the primary oral axis. Due to the right-left reversal of membranelle development, the membranelles of the secondary OP come to abut upon the right postoral meridian of the secondary axis much as the membranelles of the primary OP abut on the left postoral meridian of the primary axis (Figs. 21, 25 a). The longitudinal axis that has the same geometrical relation to the secondary oral structures as the right postoral meridian does to the primary oral structures is most likely situated approximately midway between the ciliary meridians to the right and left of the secondary OP. The positions of these two equivalent axes are shown by the heavy dashed lines in Fig. 25(a). Using this construct, the secondary oral axis is now approximately one-half of an intermeridional interval to the left of where it had earlier been Mirror-image duplication in Tetrahymena 199 placed. This, then, provides a justification for subtraction of 0-5 from the positions of CVP sets located to the left of the secondary axis, lowering the average distances from the 4-26 and 3-84 recorded in Table 7 to 3-76 and 3-34, both not much greater than the respective mean distances of 3-54 and 2-87 from the primary oral axis to its associated CVP set (this adjustment was already incorporated into Fig. 24). This correction allows us to schematize a CU-127 cell that has two OAs as being bisected by a mirror-plane separating two bilaterally symmetrical halves (Fig. 25b). These halves may be considered as two separate and mutually symmetrical domains under the control of the morphogenetic fields underlying the respective oral axes, one normally directed and the other reversed. The two halves are also capable of manifesting some physiological independence, as when one half attempts to divide while the other undergoes oral replacement following a shift to starvation conditions (Results, section 2 c). The geometrical idealization shown in Fig. 25 requires one important quali- fication with regard to position of CVP sets. Many CU-127 cells, even those with two OAs, have only one CVP set, at a position that is the same as in normal singlets. In those CU-127 cells that have two CVP sets, the average positions of these sets are further away from the oral axes than they are in typical WT doublets (Fig. 24, bottom right). However, these averages conceal a broad range of variation in CVP positions, with reference angles ranging from 35 to 75°. CU-127 cells with two CVP sets can therefore be viewed as expressing a con- tinuum of cytogeometric states with respect to positioning of CVPs, ranging between the limits of a 'doublet' condition with the reference angles typical of WT doublet cells (but with one of the angles reversed in direction) to a condition closely approaching a 'singlet' state (e.g. with only one ciliary meridian separating the two CVP sets). Then, the apparent 'singlet' state of CU-127 cells could be viewed as an actual doublet state in which the two oppositely directed reference angles reach a limit at which the two sites of CVP determination become superimposed, so that only one CVP set can be scored. In fact, such sets often display clear indications of actual doubleness, such as the appearance of CVPs along four adjacent ciliary meridians, a condition never found in wild-type cells yet one which has to be scored as a 'single' set according to our objective scoring rules. On this view, the true percentage of double CVP sets in CU-127 cells may really be 100 %; what varies is the type of expression of these two sets, de- pending on a varying degree of interaction between the two underlying morphogenetic fields. When there is no interaction, each system operates as if the other did not exist, and thus positions CVPs at an angle of about 85°, as in singlet cells (but with one angle reversed, hence the resulting superimposition). When there is maximal interaction, the cell is 'divided up' between two non-inter- penetrating morphogenetic fields as in Fig. 25(b), and the cells's CVP cytogeometry is the same as that of a wild-type homopolar doublet cell except for the reversed handedness of one of the fields. Somewhat surprisingly, the 200 M. JERKA-DZIADOSZ AND J. FRANKEL strength of the interaction between the two morphogenetic fields as manifested by the varying CVP reference angles is largely independent of the expression of oral development along the secondary axis.

(B) Duality of control of cortical asymmetry The reversal that generates (or possibly reflects) bilaterality of cellular organization in CU-127 cells affects asymmetry of OAs and the large-scale aspect of the positioning of the CVP sets which was earlier shown to be relationally determined (Nanney, 1966a). It is essential to note that the bilaterality is not all-encompassing, as it does not extend to the structure of the ciliary meridians, to the side of the meridian on which the OP first appears, and to the 'fine positioning' of CVPs relative to adjacent ciliary meridians (cf. Ng, 1977). Thus, one aspect of cell organization is reversed in one cell-half while another aspect remains normal. This dissociation demonstrates that the two aspects of cell organization are to a considerable degree independent of each other. Another consequence of this same independence is presented in the accompanying paper (Frankel & Jenkins, 1979), in which the kinetics of expression of cortical reversal and of change in number of ciliary meridians are shown to be mutually independent following acquisition or loss of homozygosity for the controlling janus gene. We therefore suggest that what remains normal in the CU-127 clone is a short-range positional system directly associated with ciliary meridians, whereas the aspect subject to reversal is a long-range system comparable to the 'gradient-fields' operating in multicellular development. We should finally note that the existence of a dual system of positional controls in ciliates is supported by substantial earlier evidence derived mainly from microsurgical studies of larger ciliates (e.g. see Jerka-Dziadosz, 1974; Frankel, 1974). The experimental result that most closely approximates the geometrical situation in thd CU-127 clone is an operation performed by Uhlig (1960) in which the development of reversed oral structures was brought about within a context of normal ciliary meridians (see interpretation in Frankel, 1974, pp. 466-467). In Stentor the reversal of oral asymmetry is demonstrably related to a right-left reversal of the 'zone of stripe contrast' (Tartar, 1956, 1960) that is thought to be a manifestation of an underlying circular gradient (Uhlig, 1960). Two greatly different methods of experimentation applied to Tetrahymena and Stentor respectively have thus supported remarkably similar conclusions.

The authors would like to thank Mr Richard J. Hollis for suggesting one of the experiments, and Drs J. Wynne McCoy and Stephen F. Ng for advice concerning analysis and interpretation of the data. We also wish to express our appreciation for helpful criticisms of various drafts of the manuscript by Drs Karl Aufderheide, Giinter Cleffmann, Anne W. K. Frankel, Krystina Golinska, Leslie M. Jenkins, Andrzej Kaczanowski, Denis Lynn, J. Wynne McCoy, David L. Nanney, E. Mario Nelsen, Stephen F. Ng, and Norman E. Williams. This research was supported by grant no. HD 08485 from the U.S. National Institutes of Health. Mirror-image duplication in Tetrahymena 201

REFERENCES AUFDERHEIDE, K. (1978). Specific associations of mitochondria with the cortex of Tetra- hymena thermophila. J. Protozool. 25, 7a (abstract). BEISSON, J. & SONNEBORN, T. M. (1965). Cytoplasmic inheritance of the organization of the cell cortex in Paramecium aurelia. Proc. natn. Acad. Sci. U.S.A. 53, 275-282. BRUNS, P. J., BRUSSARD, T. B. & KAVKA, A. B. (1976). Isolation of horaozygous mutants after induced self-fertilization in Tetrahymena. Proc. natn. Acad. Sci. U.S.A. 73, 3243- 3247. BRUNS, P. J. & SANFORD, Y. N. (1978). Mass isolation and fertility testing of temperature sensitive mutants in Tetrahymena. Proc. natn. Acad. Sci. U.S.A. 75, 3355-3358. BUHSE, H. E. JR., STAMLER, S. J. & CORLISS, J. O. (1973). Analysis of stomatogenesis by scanning electron microscopy in Tetrahymena pyriformis strain W during synchronous cell division. Trans. Am. microsc. Soc. 92, 95-105. DRYL, S. & TOTWEN-NOWAKOWSKA, I. (1972). Motile behaviour in two forms of double animals Stylonchia mytilus. Acta Protozool. 11, 301-305. FAURE-FREMIET, E. (1945). Symetrie et polarite chez les cilies bi-ou-multicomposites. Bull. biol. Fr. Belg. 79, 106-150. FAURE-FREMJET, E. (1948). Doublets homopolaires et regulation morphogenetique chez le Cilie Leucophrys patula. Archs Anat. microsc. Morph. exp. 37, 183-203. FRANKEL, J. (1964). The effects of high temperatures on the pattern of oral development in Tetrahymena pyriformis. J. exp. Zool. 155, 403-436. FRANKEL, J. (1965). The effect of nucleic acid antagonists on cell division and oral organelle development in Tetrahymena pyriformis. J. exp. Zool. 159, 113-148. FRANKEL, J. (1969). Participation of the undulating membrane in the formation of the oral replacement primordium of Tetrahymena pyriformis. J. Protozool. 16, 26-35. FRANKEL, J. (1972). The stability of cortical phenotypes in continuously growing cultures of Tetrahymena pyriformis. J. Protozool. 19, 648-652. FRANKEL, J. (1974). Positional information in unicellular organisms. /. theor. Bid. 47, 439-481. FRANKEL, J. (1975). Pattern formation in ciliary organelle systems of ciliated protozoa. In Cell Patterning, Ciba Foundation Symposium, New Series no. 29, pp. 25-49. Amsterdam: Elsevier. FRANKEL, J. & HECKMANN, K. (1968). A simplified Chatton-Lwoff silver impregnation procedure for use in experimental studies with ciliates. Trans. Am. microsc. Soc. 87, 317-321. FRANKEL, J. & JENKTNS, L. M. (1979). A mutant of Tetrahymena thermophila with a partial mirror-image duplication of cell surface pattern. II. Nature of genie control. /. Embryol. exp. Morph. 49, 203-227. FRANKEL, J., JENKINS, L. M., DOERDER, F. P. & NELSEN, E. M. (1976). Mutations affecting cell division in Tetrahymena pyriformis. I. Selection and genetic analysis. Genetics 83, 489-506. FRANKEL, J., NELSEN, E. M. & JENKINS, L. M. (1977). Mutations affecting cell division in Tetrahymena pyrifcrmis, syngen 1. II. Phenotypes of single and double homozygotes. Devi Biol. 58, 255-275. FRANKEL, J. & WILLTAMS, N. E. (1973). Cortical development in Tetrahymena. In The Biology of Tetrahymena (ed. A. M. Elliott), pp. 375-409. Dowden, Hutchinson and Ross. JERKA-DZIADOSZ, M. (1974). Cortical development in Urostyla. II. The role of positional information and preformed structures in formation of cortical pattern. Acta Protozool. 12, 239-274. JERKA-DZIADOSZ, M. (1977). Temporal coordination and spatial autonomy in regulation of ciliary pattern in double forms of hypotrich ciliate, Paraurostyla weissei. J. exp. Zool. 200, 23-31. KACZANOWSKI, A. (1976). An analysis of mp gene affected morphogenesis in Tetrahymena pyriformis, syngen 1. /. exp. Zool. 196, 215-230. 202 M. JERKA-DZIADOSZ AND J. FRANKEL LYNN, D. H. (1977). Proportional control of organelle position by a mechanism which similarly monitors cell size of wild type and conical form-mutant Tetrahymena. J. Embryol. exp. Morph. 42, 261-274. LYNN, D. H. & TUCKER, J. B. (1976). Cell size and proportional distance assessment during determination of organelle position in the cortex of the ciliate Tetrahymena. J. Cell Sci. 21, 35-46. MCCOY, J. W. (1974). New features of the tetrahymenid cortex revealed by protargol staining. Acta Protozool. 8, 155-159. NANNEY, D. L. (1966a). Cortical integration in Tetrahymena: an exercise in cytogeometry. /. exp. Zool. 161, 307-317. NANNEY, D. L. (19666). Corticotypes in Tetrahymena pyriformis. Amer. Nat. 100, 303-318. NANNEY, D. L. (1967 a). Cortical slippage in Tetrahymena. J. exp. Zool. 166, 163-169. NANNEY, D. L. (19676). Comparative corticotype analyses in Tetrahymena. J. Protozool. 14, 690-697. NANNEY, D. L. (1968). Patterns of cortical stability in Tetrahymena. J. Protozool. 15,109-113. NANNEY, D. L., CHOW, M. & WOZENCRAFT, B. (1975). Considerations of symmetry in the cortical integration of Tetrahymena doublets. /. exp. Zool. 193, 1-14. NANNEY, D. L. & MCCOY, J. W. (1976). Characterization of the species of the Tetrahymena pyriformis complex. Trans. Am. microsc. Soc. 95, 664—682. NELSEN, E. M. (1970). Division delays and abnormal oral development produced by col- chicine in Tetrahymena. J. exp. Zool. 175, 69-84. NELSEN, E. M. (1978). Transformation of Tetrahymena thermcphila: development of an inducible phenotype. Devi Biol. 66, 17-31. NELSEN, E. M. & DEBAULT, L. E. (1978). Transformation in Tetrahymena pyriformis - description of an inducible phenotype. /. Protozool. 25, 113-118. NG, S. F. (1977). Analysis of contractile vacuole pore morphogenesis in Tetrahymena pyriformis by 180°-rotation of ciliary meridians. /. Cell Sci. 25, 233-245. NG, S. F. (1978). The precise site of origin of the contractile vacuole pore in Tetrahymena and its morphogenetic implications. Acta Protozool. (In the Press.) NG, S. F. & FRANKEL, J. (1977). 180°-rotation of ciliary rows and its morphogenetic implications in Tetrahymena pyriformis. Proc. natn. Acad. Sci. U.S.A. 74, 1115-1119. NG, S. F. & NELSEN, E. M. (1977). The protargol staining technique: an improved version for Tetrahymena pyriformis. Trans. Am. microsc. Soc. 96, 369-376. NG, S. F. & WILLIAMS, R. J. (1977). An ultrastructural investigation of 180°-rotated ciliary meridians in Tetrahymena pyriformis. J. Protozool. 24, 257-263. NILSSON, J. R. & WILLIAMS, N. E. (1966). An electron microscope study of the oral apparatus of Tetrahymena pyriformis. C. r. Trav.Lab. Carlsberg 35, 119-139. RUFFOLO, J. J. (1974). Critical point drying of protozoan cells and other biological specimens for scanning electron microscopy. Trans. Am. microsc. Soc. 93, 124-130. RUFFOLO, J. J. & FRANKEL, J. (1972). Observations on stomatogenesis in Tetrahymena pyriformis GL-C using scanning electron microscopy. /. Protozool. 19 (suppl.), 21. SONNEBORN, T. M. (1975). Positional information and nearest neighbour interactions in relation to spatial patterns in ciliates. Annee Biol. 14, 565-583. TARTAR, V. (1956). Further experiments correlating primordium sites with cytoplasmic pattern in Stentor coeruleus. J. exp. Zool. 132, 269-298. TARTAR, V. (1960). Reconstitution of minced Stentor coeruleus. J. exp. Zool. 144, 187-207. TARTAR, V. (1962). Morphogenesis in Stentor. Adv. Morphogen. 2, 1-26. TCHANG TSO-RUN, N., SHI XIN-BAI & PANG YAN-BIN (1964). An induced monster ciliate transmitted through three hundred and more generations. Scientia Sinica 13, 850-853. TCHANG TSO-RUN, N. & PANG YAN-BIN (1965). Conditions for the artificial induction of monster jumelles of Stylonchia mytilus which are capable of reproduction. Scientia Sinica 14, 1332-1338. UHLIG, G. (1960). Entwicklungsphysiologische Untersuchungen zur Morphogenese von Stentor coeruleus Ehrbg. Arch Protistenk 105, 1-109. WILLIAMS, N. E. & FRANKEL, J. (1973). Regulation of microtubules in Tetrahymena. I. Electron microscopy of oral replacement. /. Cell Biol. 56, 441-457. {Received 18 July 1978, revised 26 September 1978)