/ . Embryo!, exp. Morph. Vol. 65, pp. 57-71, 1981 57Printed in Great Britain © Company of Biologists Limited 1981

Polymorphism and specificity of positioningof contractile vacuole pores in a ciliate,,

Chilodonella steini

By JANINA KACZANOWSKA1

From the Institute of Zoology, Warsaw University

SUMMARY

The unicellular ciliate Chilodonella steini has a well-defined flat and ciliated ventral field.During divisional morphogenesis two sets of new contractile vacuole pores (CVPs) areformed on this field. During final pattern formation some of these CVP primordia and theold parental set of CVPs are completely resorbed. Primary pattern of distribution of theCVP primoidia and final pattern of distribution of the matured CVPs manifest an intraclonalpolymorphism.

From analysis of this polymorphism some features of mechanism(s) of CVP patterndetermination are deduced:

1. There is a strict, short-distance negative control of appearance of CVP primordia atsites of oral morphogenesis and around the ventral field.

2. Certain indeterminacy of large-scale patterning of CVP primordia is observed over thearea competent to yield CVP formation. However, within this area three longitudinal sectorswith a high probability of occurrence of CVP primordia are alternated with sectors nearlydeprived of their occurrence.

3. Positive control of probability of occurrence and of specificity of location is found forcertain CVP primordia. An interaction of mechanism of positioning on cellular longitudesand latitudes is proposed to account for these facts.

4. The resorption of supernumerary CVP primordia does not alter the character of theglobal map of distribution of CVP primordia achieved during primary pattern formation.The primordia located at some latitudes persist, whereas others are resorbed at random.It is suggested that all CVP primordia which do not mature at the time of stabilization ofdivisional morphogenesis are resorbed. Thus the global map of CVPs distribution wouldresult from the sum of the individual determinations of the fates of each CVP primordium,superimposed on an initial spatially non-uniform distribution of CVP primordia.

INTRODUCTION

Ciliate are single-celled organisms with an ordered pattern of distribution oftheir organelles over the cell cortex. There is evidence that the mechanism oflarge-scale positioning of organelles is encoded in the ciliate genome (Heckmann& Frankel, 1968; Jerka-Dziadosz & Frankel, 1979) to ensure this species-specificorder. However, in some ciliates the polymorphism exists. This may result frommodification of the global map by effects of pre-existing organization of theciliate cortex (Beisson & Sonneborn, 1965; Grimes, 1976; Ng & Frankel, 1977;

1 Author's address: Institute of Zoology, Warsaw University, Warsaw 00-927, Poland.

58 J. KACZANOWSKA

Final patternResorption

Primary pattern

1 •1 •1 •

• !

|oiiiii

• ;

o|

Attenuation

: ° •• o

•

Random selection

Atten

i

: •

• !

• !• i

| O• 1

uation

«•

•

•

V0

••

Random selection

o

0

' O

o

Positive control

•9

0

0

•

0

Random selection

Fig. 1. For legend see opposite.

Ng, 1979), either from the gradual regulation of patterns in the progeny of anabnormally patterned initial cell (Nanney, 1968), or from a probabilistic modeof determination of the number and of the location of organelles. This latterpossibility is explored in this report on polymorphism of the disposition ofcontractile vacuole pores (CVPs) on the flat ventral surface of a ciliate,Chilodonella steini.

Any geometric description of the position of an organelle on the cell cortexrequires specification of a coordinate system. The coordinate system projectedover the cell surface allows one to measure absolute or relative distances fromchosen reference points. At least two parameters are indispensable for the exact

Contractile vacuole pores patterning in a ciliate 59

placement of the measured point on a two-dimensional surface. In a Cartesiensystem these parameters correspond to latitude and longitude (Frankel, 1979).In the polarized cortical layer of ciliates the evenly spaced meridional rows ofciliary basal bodies form the natural meridians for measuring longitudes of theorganelle, with one of them - the stomatogenic meridian - serving as a refer-ence. Nanney (1966a, b, 1967, 1968) discovered certain cytogeometric rules ofpositioning of the CVP on longitudes in Tetrahymena. He gave a formal explana-tion of positioning of the CVP in a given cell as a consequence of specificationof an inductive angle between the stomatogenic meridian and the area in whichthe CVP is found. The variability of positioning of the CVP between meridiansis described in terms of afield angle, which defines the sector of the cell surfacewhich may be competent to yield a CVP.

It is already established (Kaczanowska, 1974) that CVPs in Chilodonellacucullulus strain X form on three longitudinal sectors; the CVPs vary in numberfrom 5 to 11, and their location is variable. CVPs are positioned at intersectionsof the longitudinal sectors with specific radii measured from a reference point,a site of stomatogenesis. If cells are microsurgically miniaturized (Kaczanowska,1975), the proximal radius describes the exact placement of the anterior obli-gatory CVPs, while posterior CVPs are drastically reduced in number.

It is believed that an analysis of the polymorphism of CVP primordia and ofCVPs distributions over the ventral field of the related species Chilodonellasteini makes possible to delineate some general characters of a large-scalemechanism(s) operating in CVP-pattern determination. We begin by describingthe number and distribution of CVP primordia and of CVPs in different speci-mens of Ch. steini at the same morphogenetic stage of division, when old CVPs

Fig. 1. Theoretical models of CVP primordia and CVPs distribution within a zonecompetent to yield CVP formation in Chilodonella steini. Boxes represent an entirearea of the ventral field. Black circles represent CVP primordia (left boxes), ormatured CVPs (right boxes). White circles mark resorbed CVP primordia absent infinal patterns (right boxes). Model A. Negative control, or inhibition of CVPprimordia appearance in the extreme 'forbidden' zones. Sharp boundaries (solidlines) between CVP competent zone and CVP deprived zones. Primordia arerandomly distributed within CVP competent zone. This might lead, followingresorption of some CVP primordia, to either of two alternatives: Al -an attenuationof the CVP competent zone (indicated by transposition of solid lines, while dashedlines mark positioning of original boundaries), or A2 - maintenance of the originalboundaries with random selection of CVP primordia for resorption. Model B.Probabilistic model of CVP primordia distribution along a preferred meridian(heavy line), but with a high dispersion of placement of CVP primordia. Delineatedby dashed lines external zones mark zones of a very low probability of CVP primordiaoccurrence. Following resorption there is either Bl - an attenuation of the originaldispersion (internal dashed lines), or B2 - maintenance of the original dispersion.Model C. Positive control of placement of CVP primordia at the intersection oftwo coordinates (crossing heavy lines). This might be followed by either: Cl - main-tenance of CVP primordia at that intersection (dashed circle), or C2 - by a randomselection of CVPs,

60 J. KACZANOWSKA

and newly induced CVP primordia still coexist. Next we may investigate whetherthere are any similarities of the test patterns. The following issues are considered:

1. Area occupied with CVP structures in all tested specimens apparentlybelong to regions competent to yield CVPs. Remaining areas may be either lesscompetent to yield CVP formation, or are inhibited in CVP formation. If anarea occupied with CVP primordia (or CVPs) is very sharply marked off fromareas deprived of CVP primordia (or CVPs), this suggests that there are zones'forbidden' from yielding CVP structures, i.e. they are under some form ofnegative control. If however CVP-competent areas gradually fade or mergewith empty areas, this suggests that only positive controls are operating. Itmeans that the appearance of CVP structures is most likely along certainmeridians, with a gradual decrease in probability at more distant locations(Fig. 1, models A and B, left boxes).

2. Grimes & L'Hernault (1979) and Frankel (1979) have established adifferent character of positioning operating on the longitudes and latitudes of aciliate cell. Then the question arises whether or not these two supposed mech-anisms act independently in determination of the position of a given organelle.It might be ascertained whether, in Ch. steini, the probability of occurrence ofCVP structures and the degree of longitudinal dispersion remain constant withina given sector at all of its latitudes. In the event of some cooperation betweenthe two putative mechanisms involved in positioning of CVP primordia, someCVP primordia within the same sector would be spatially more precisely locatedthan others (Fig. 1, model C left box).

3. In Ch. steini two sets of new CVP primordia appear during divisionalmorphogenesis while the old set of parental CVPs still persists. The primarypatterns of the distribution of CVP primordia is then changed by the resorptionof certain supernumerary CVP primordia. During final pattern formation, allsupernumerary CVP primordia and the old set of parental CVPs completelydisappear. The comparison of the primary and final patterns is made to testwhether resorption of some CVP primordia may modify the general characterof their distribution. Does it expand the 'forbidden' areas by eliminating CVPprimordia at the boundaries or does it occur throughout the CVP areas merelydecreasing their number but not affecting the boundaries of the CVP zone?If it is not at random this proves that final pattern determination takes place intwo steps: first a broad outline of CVP areas, second some modification ofthese areas by the process of CVP primordia resorption. The alternative possi-bility, of random resorption of some of CVP primordia over the ventral field,would, if it occurred, still leave open the question of the reason of this random-ness. The two different models of final pattern regulation subsequent to each ofthe three types of initial pattern establishment are depicted schematically inFig. 1 (right boxes).

Data reported here are taken as evidence that:1. There is a very strict negative control of appearance of CVP primordia in

Contractile vacuole pores patterning in a ciliate 61

some 'forbidden' areas. These areas are confined to the sites of stomatogenesisand to the border of the ventral field.

2. In the remaining areas competent to yield CVP primordia, three sectorsof high probability of occurrence of CVP primordia manifest different widths.

3. Some CVP primordia are spatially much more precisely determined thanothers.

4. Resorption of the supernumerary CVP primordia during final patternformation does not alter the character of CVP distribution over the ventralfield. However, certain CVPs, which are invariant elements of pattern, are neverresorbed.

MATERIALS AND METHODS

A clone of Chilodonella steini (Ciliata, Kinetofragmophora; Radzikowski &Goiembiewska, 1977) line 237/10 that did not self in the immaturity period(Kaczanowski, Radzikowski, Malejczyk & Polakowski, 1980) was isolatedfrom one exconjugant. Six months later, one subclone was reinitiated. Allpreparations of this subclone were made during a one-month period. Othercells tested were derived from the same subclone about 10 months later, whenat least some cells entered into permanent selfing, with retention of old macro-nucleus (Kaczanowski et al. 1980).

The general characteristic of this species and the methods of culturing of thecells followed these of Radzikowski & Goiembiewska (1977). The mean genera-tion time of the cells varied from 12-19 h when they v/ere maintained in a normaldaily photoperiod and fed every second day.

Cells from 2-day mass clonal cultures were used for silver impregnation(method of Frankel & Heckmann, 1968). Well-silvered specimens in earlydivision were selected for mapping and counting of their CVPs and CVP prim-ordia if they were properly dorsoventrally embedded in gelatin and if all ciliarymeridians were clearly distinguished. In 41 dividers of the subclone 237/10immatured, protocols and maps were made of cortical parameters of 1842 CVPprimordia and of 587 parental CVPs. If any coordinate system was tested (Figs.6, 7, 8 and Table 1), all of the data were grouped and then statistically analysed(Sokal & Rohlf, 1969). In Fig. 8 three peaks were revealed in all three histograms.The significance of these results was tested for a given coordinate system bycomputing for every individual (n = 41) the difference in number of CVPprimordia between a sector selected a priori in this coordinate system as havinga high number of CVP primordia and neighbouring right and left sectors of thesame width selected as having a low number of CVP primordia. In Table 1 inall specified sectors mean numbers and sd values of CVP primordia and ofCVPs were calculated. To estimate the rate of decrease of total number ofCVPs as compared to the total number of CVP primordia in specified sectors,the ratio of the number of CVPs occurring in a given sector of CVP primordiain it was calculated for every respective parental to anterior daughter, and

3 EMB 65

62 J. KACZANOWSKA

Fig. 2. A ventral field of morphostatic Ch. steini. The CVPs are dispersed amongthe ciliary meridians (arrows).Fig. 3. A ventral field of early dividing Ch. steini. In addition to old parental CVPs,new CVP primordia (arrows) are seen at the left sides of some ciliary meridians.The R, M and L sectors are marked off by dashed lines.Fig. 4. A ventral field of dividing Ch. steini at a stage, somewhat more advanced thanthat in Fig. 3. Segment A-4 is labelled with an arrow.Fig. 5. A ventral field of an advanced divider of Ch. steini. Morphogenetic movementof the oral segments is seen (heavy arrow). The new CVP primordia are undergoingtransformation into the matured CVPs for daughter cells (arrows), while others aregradually discarded.

parental to posterior daughter patterns of the same specimen. Then pooled dataof the mean of all these ratios for every sector were compared with a Cochranand Cox test. The additional control cells fixed 10 months later involving onlymorphostatic cells were tested for reproducibility and the maintenance of thepolymorphism of cells.

Some statistical calculations (r correlation coefficients) have been made inthe Center of Statistical Calculations of Warsaw University by Msc I. Wozniak.

RESULTS

(1) Divisional morphogenesis o/Chilodonella steini

Divisional morphogenesis of Ch. steini conforms to the general schemedescribed for this genus (Radzikowski & Golembiewska, 1977).

Ch. steini is a flat asymmetric ciliate, with the ventral surface covered withciliary meridians and subapical oral apparatus encircled by an oral ciliature.CVPs are distributed only over the ventral surface, but they never appear nearthe oral apparatus or at the margin of the ventral surface. In silvered specimensCVPs appear as round black circles between ciliary meridians (Fig. 2, arrows).

The first signs of approaching division are an increase of body size, differ-entiation of oral ciliary segments for the prospective posterior daughter cell

Contractile vacuole pores patterning in a ciliate 63

(opisthe), and differentiation of two sets of new CVP primordia for the daughtercells as little spots or perpendicular slits near the left side of certain ciliarymeridians (Fig. 3, arrows). Oral ciliary segments and one somatic segment (theso-called A-4 segment, Radzikowski, 1966; Kaczanowska, 1971) for the pre-sumptive opisthe differentiate in the subequatorial region of the ventral field.Cells in this stage were selected for further study.

In the following stage of morphogenesis (Fig. 4), all oral segments of thepresumptive opisthe and the ciliary segment A-4 begin to migrate by rotatingaround the centre of the oral region. While the ciliary meridians are passivelytransmitted into successive anterior daughter cells (proters), in opisthes theenumeration of meridians of the postoral left part of the ventral field becomesaltered due to a gain of one meridian from the A-4 segment. A-4 inserted back-wards between a stomatogenic meridian (no. 1) and the meridian to its right(no. 2). When enumeration of meridians follows the rules for Tetrahymena(Elliott & Kennedy, 1973), the former meridian 1 now becomes meridian n, theformer n becomes n-\ etc. (Figs. 4, 5). This slippage compensates for the usualloss of the extreme left meridian, which is not represented in the equatorial zoneand is passed entirely to the proter. As a result of these migrations, the circumoralsegments turn about 120° while the preoral segment is reversed and positionedanteriorly to them (Fig. 5). Some new CVP primordia change into long perpen-dicular slits, while others remain with no transformation (Fig. 4).

At a later stage of morphogenesis the old, parental oral apparatus (pharyngealbasket) is resorbed (Fig. 5) and very quickly two new oral apparatuses for thenascent daughters are formed: in situ for the proter and in the centre of a regionof rotating ciliary segments for the opisthe. The CVP primordia gradually aretransformed into the final round orifices in the middle of the intermeridionalspace, while the parental set of CVPs and some supernumerary new CVPprimordia are resorbed.

In early dividers (Fig. 3) new CVP primordia are readily distinguished fromparental CVPs by their shape and fine positioning (slits or dots in the left sideof a ciliary meridian versus round bigger circles in the middle of the inter-meridional space). In this stage, the future fission line is marked by the ruptureof stomatogenic and left meridians. The right margin of the fission line mayalso be discerned as a small indentation of the dorsal surface. The distributionof CVP primordia may, therefore, be analysed separately in prospective protersand opisthes. In this stage the A-4 segment has not yet been added, and the oldpattern of CVPs is perfectly preserved.

(2) Variability of the cortical patterns and of the total number ofCVPs and CVP primordia

Dividing specimens of Ch. steini manifested an array of corticotypes (i.e.total number of ciliary meridians) from 24 to 30 with corticotypes 27 and 28in the majority of cells. In cells with corticotypes 26, 27 and 28 of the group

3-2

64 J. KACZANOWSKA

tested for CVPs and CVP primordia, the distribution of the total dimensionsof the ventral field were nearly identical. If the first (furthest right) stomatogenicmeridian is designated as no. 1, it divides the whole set of ciliary meridians intoright (nos. 2, 3, 4 etc.) and left ciliary meridians (nos. n, n-\, «-2, etc.). Cells ofthe same corticotypes often have different patterns of the total number of rightand left meridians.

The total number of parental CVPs varies from 8 to 25 with a mean numberof 14-7 ± 3-9 (n = 41). The newly formed sets of the CVP primordia includedrespectively 13 to 35 (mean 22-1 ± 5-0) for proters and 10 to 38 (mean 22-8 ± 6-2;n = 41) for opisthes. There is a significant (P = 0-01) difference between thetotal number of parental CVPs and the number of CVP primordia in descendants.It is deduced that some of the CVP primordia are discarded during formationof final pattern (about 34-5 % of the total number of CVP primordia).

There is no statistically significant correlation between the total number ofparental CVPs and the number of CVP primordia in proters (r = 0-14) or inopisthes (r = 0-18). There is a slight positive correlation of the total numberof the CVP primordia in proter and opisthe (r = 0-55 with P = 0-05).

The total number of the CVP primordia observed in particular intermeri-dional space is variable (from 0 to 10). Among 41 tested dividers no two cellshad identical patterns of distribution of CVPs or CVP primordia.

Hence in Ch. steini there is a polymorphism of corticotypes, of right/leftpattern of ciliary meridians, and of a number and disposition of CVP primordiaand CVPs.

(3) Localization of the CVP primordia in early dividers, primary patterns

The maps of CVP primordia show that in all specimens there are zones strictlydevoid of CVP primordia. These 'forbidden' zones include the margin of theventral field, the regions of the parental oral apparatus and of the preoral ciliarysegment, the area of the future oral primordium for the opisthe, and the regionof the forming fission line.

The remaining surface of the ventral field is more or less competent to yieldCVP primordia. However, these competent zones include longitudinal sectorsof relatively high frequency of occurrence of CVP primordia. The right (R) sectorfollows the curvature of the right margin of the ventral field, but at some distancefrom it. The median (M) sector appears in the right half of the ventral fieldparallel to the main longitudinal axis of cells. This sector is roughly composedof two separate groups of CVP primordia for every daughter cell: one justposterior to the oral areas, which corresponds to the sites of CVP-1 describedin the related species, Ch. cucullulus (Kaczanowska, 1974), and the second atthe rear end of the ventral surface of the prospective daughter cell, correspondingto the posterior CVP-4 site of Ch. cucullulus. Finally, the left (L) sector appearsin the middle portion of the left part of the ventral field. All of these sectors arediagramatically marked in Fig. 3.

Contractile vacuole pores patterning in a ciliate 65

4i

3-

OJ ~^r-

14 12 10 8 _2 - 4 - 6 - 8 -10 -12

Fig. 6. Diagram of the mean number of CVP primordia per given intermeridionalspace counted from the stomatogenic axis for prospective proters of three differentcorticotypes. Corticotype 28 (n = 8) - dashed line, corticotype 27 (n = 11) - heavyline, and corticotype 26 (n = 6) - dashed and dotted line.

0J

14 12 10 8 - 2 -4 -6 -8 -10 -12

Fig. 7. Diagram of the mean number of CVP primordia per given intermeridionalspace counted from the stomatogenic axis for prospective opisthes of three differentcorticotypes. Illustrative conventions and number of studied patterns as in Fig. 6.

Comparison of means of occurrence of CVP primordia in particular inter-meridional spaces in cells of the same corticotypes (Fig. 6 for proters and Fig. 7for opisthes) revealed that histograms for the proters and for the opisthes arevery similar. The R, M and L sectors of frequent occurrence of CVP primordiaalternate with four sectors of infrequent occurrence. The sectors at the left andright edges of the histograms belong to the 'forbidden' zones, while the sectorslocated between the R, M and L sectors, are characterized by a relativeabsence of the CVP primordia. The different corticotypes have a very similarvariability of the mean frequency of CVP primordia. The longitudes of the CVPprimary pattern might be computed for the whole ventral field by combiningdata for the prospective proters and opisthes for all 41 specimens. To testwhether the observed R, M and L sectors might be also specified with reference

66 J. KACZANOWSKA

1 3 5 7 9 11 13 15 17 19 21 23 25 27 2929 27 25 23 21 19 17 15 13 11 9 7 5 3 114 12 10 8 6 4 2 n -2 -4 -6 -8 - 1 0 - 1 2 -14

Fig. 8. Diagram of the mean number of CVP primordia per given intermeridionalspace occurring both in prospective proter and opisthe in 41 specimens using thefollowing reference coordinates: (a) the right extreme meridian, solid line; (b) theleft extreme meridian, dashed and dotted line; (c) the stomatogenic axis, dotted line.Note that three different conventions of counting of ciliary meridians were used forordering data about the CVP primordia deployment: convention (a), the mostright ciliary meridian is designated as a ciliary meridian 1 and followed by 2, 3, etc.;convention (b), the most left ciliary meridian is designated as a ciliary meridiannumber 1 and followed in reverse direction by 2, 3, etc.; convention (c), furthestright postoral meridian is designated as a ciliary meridian number 1, to the rightthis meridian is followed with ciliary meridians 2, 3, etc., to the left this meridian isfollowed with ciliary meridians n, n-l,n-2, etc. This convention corresponds to that inFigs. 6 and 7. Three conventions of enumerations indicated on three abscissae.Ordinate: mean number of CVP primordia per intermeridional space.

to the boundaries of the ventral field, the positioning of all CVP primordia wasthen assessed in all specimens in terms:

(1) of the number of meridians from the right margin of the ventral field,(2) or in terms of the number of meridians from the left margin of the ventral

field,(3) or in terms of the number of meridians from the stomatogenic axis.In three histograms (Fig. 8) the R, M and L peaks are visualized in all three

coordinate systems. Statistical tests were carried out to assess the significanceof the peaks appearing within the three coordinate systems with respect to theneighbouring valleys. Particular intermeridional spaces were selected on an apriori basis as having either a high or low probability of appearance of CVPprimordia. It was found that statistically significant results were obtained forthe R and L sectors only if these were constituted of three intermeridionalspaces, not one or two.

The three intermeridional spaces of the R peak had a significantly greateraverage number of CVP primordia than did the neighbouring valleys when these

Contractile vacuole pores patterning in a ciliate 67

spaces were counted from the right margin of the ventral field; similar resultswere obtained for the left L peak with reference to the left margin. When thepositions of the right and left peaks were enumerated with respect to the stomato-genic axis (Fig. 8, dotted line), these peaks were significant when compared tothe proximal valleys (i.e. those between the stomatogenic axis and the peak)but, surprisingly, not with respect to the distal zones situated near themargins.

The M peak differed from the two other peaks in that it was made up of onlya- single intermeridional space, and that it was significantly specified only withrespect to the stomatogenic axis. This can be clearly appreciated by noting inFig. 8 how much sharper this peak is in the spatial system keyed to the stomato-genic axis (dotted line) than in that keyed to the right (solid line) or left (dashedline) margins.

The simplest interpretation consistent with this analysis is that the M sectorconsists of only a single intermeridional space while the R and L sectors consistof three or, in the case of the L sector, probably more such spaces. The stomato-genic axis probably serves as the primary reference point for the M sector andmay participate in specifying the R and L sectors; however, the 'forbidden'sectors at the two margins are significantly specified only in relation to thenearby cell margins.

Although, as mentioned earlier, the distribution of CVP primordia is proba-bilistic, and no two cells have an identical distribution, certain specific CVPprimordia can be followed more reliably than others. We will here consider thetwo CVP primordia that may appear in the M sector, one of them (CVP-1)located a short distance posterior to the oral region, the other (CVP-4) situatednot far from the posterior end of the nascent cells (Figs. 3, 5), as well as CVP-5,the anterior preoral CVP primordium in the R sector (these CVP primordia arenumbered as in Ch. cucullulus stock X; Kaczanowska, 1974, 1975). The CVP-1primordium was found in all but one of the 82 proter and opisthe examinedpatterns, and was always placed in the same intermeridional space with referenceto the stomatogenic axis. Thus it occurred that CVP-1 primordium was a stableelement of primary pattern of Ch. steini.

The CVP-4 primordium, on the other hand, was absent in 17 of the 82 patterns,preferentially in opisthes (but see also Fig. 4), and its location with respect tothe stomatogenic axis was not absolutely fixed; in fact, all of the variation inthe M sector is accounted for by variation in the placement of the CVP-4primordium.

The preoral CVP-5 primordium failed to appear in only 2 out of the 82daughter patterns. Its position, though not completely invariant, is mainlyrestricted to the interior intermeridional space within three ones constitutingthe R peak.

It thus appears that the level of indeterminacy in both the occurrence and thepositioning of specific CVP primordia differs, both for primordia located at

68 J. KACZANOWSKA

Table 1

Mean number of CVP primordia (n = 82), of CVPs (n = 41), and ratios of thetotal number of CVPs to the total number of CVP primordia in respective parentaland daughter patterns in selected sectors specified in a coordinate system with thestomatogenic axis as a reference (explained in text).

Sector R R, M M M, L

Intermeridional spaces 9,8,7 6,5,4 2 1, n, n-\ n-5,n-6,n-7CVP primordia 6-90±3-40 1-23 ±1-41 1-94±108 0-58 ±0-75 4-85 ±2-22(mean number and sd)

CVPs 4-35±20O O-75±O-8O l-54±0-98 0-22±0-52 3-58±l-43(mean number and sd)

Ratio: 0-94±0-93 036±0-62 100±0-85 016±0-38 0-93±0-81number of CVPs

number of CVP primordia(mean number and sd)

Number of tested pairs 82 72 79 36 81

different latitudinal levels in the same sector (CVP-1 and CVP-4) and in differentsectors at a somewhat similar latitude (CVP-1 and CVP-5).

(4) Comparison of the primary and final patterns of CVP distribution

In Ch. steini, only about 65-5 % of the CVP primordia persist in the finalpatterns. The remainder are resorbed. The question arises whether the proba-bility of resorption is uniform over the whole competent area of occurrance ofCVPs, or whether it is specifically confined to certain sectors (Fig. 1, rightboxes).

Some decrease in the mean number of CVPs is observed in all of the specifiedsectors, both in the peaks of preferential occurrence of CVP primordia and inthe valleys of relative absence of these primordia (Table 1). Further, the ratiosof CVPs to CVP primordia did not differ significantly among sectors, as evalu-ated by the Cochran and Cox test (P = 0-05). This result strongly suggests theuniform resorption of CVP primordia over the whole CVP competent zone(corresponding to models of the alternatives nos. 2 among right boxes in Fig. 1).

But on the other hand, when specific CVPs (matured CVP-1, CVP-4 andCVP-5) were considered, CVP-1 was found to be absent at its expected site onlyin 1-6% of the specimens, CVP-5 was absent in 8 % of the specimens, whilethe posterior CVP-4 was absent in 52-8 % of the cells. Thus the CVP-1 primordiumtends to persist at a non-random fashion (Fig. 1 Cl model), while the CVP-4primordia are much more readily resorbed.

These different data are taken as evidence for a generally uniform resorptionof the total number of CVP primordia that is superimposed upon, and inde-pendent of, the spatially non-uniform probability of formation of CVP primordiaand of their persistence.

Contractile vacuole pores patterning in a ciliate 69

DISCUSSION

In Chilodonella steini, CVP primordia may occur near any ciliary meridiansexcept those in certain 'forbidden' areas. The borders of the ventral field andthe site of stomatogenesis were used here as a priori reference points for CVPpositioning on longitudes. Absolute absence of CVP primordia in these areasis consistent with a hypothesis of a short-distance inhibitory effect of the siteof stomatogenesis and of the boundaries of the ventral field on the competenceto yield CVP primordium formation.

Every meridian on the remainder of the ventral field is able to support CVPformation, although the probability of this event is much higher in three longi-tudinal sectors. From this result three conclusions may be drawn: (a) CVPorganellogenesis is not restricted to particular meridians but rather to certainterritories, (b) there is some indeterminacy in the large-scale mechanism ofspecification of these territories, as CVP primordia sometimes form outside ofthe territories and since even within the territories the disposition of CVPprimordia is variable, and (c) there is some periodicity of longitudes of highprobability of the occurrence of CVP primordia in a form of the R, M and Lsectors alternated with sectors of low probability of CVP primordia occurrence.The right and left sectors of a high probability of occurrence of CVP structurescover more than one intermeridional space. In terms of Nanney's formalism(19666) they represent a broad field angle. The median sector, however, islimited to one intermeridional space.

A virtual stability of occurrence and of localization of the CVP-1 primordiumwith respect to the stomatogenic meridians, which undergo cortical slippage inevery generation of opisthes, indicates that the stomatogenic axis is not inheritedby the structural identity of a particular meridian, but as a territory in whichstomatogenesis takes place. The position of the M sector, and especially theCVP-1 primordium is determined in relation to this territory.

At least CVP-1 primordium placement is determined much more specificallythan the placement of the other CVP primordia. This suggests that along a givensector of high probability of appearance of CVP primordia, there exists, at somelatitudes a spatial constraint on the CVP placement along longitudes (Fig. 1,model C). Thus a cytogeometric model of CVP distribution in Chilodonella(Kaczanowska, 1974) may result from some cooperation of mechanisms ofpositioning on longitudes and on latitudes, perhaps in a form of a mosaic ofnodes of increased specificity of CVP positioning.

Ultrastructural investigations (Kaczanowska & Moraczewski, in preparation)indicate that during late division stages some CVP primordia are very advancedin their differentiation, while other neighbouring ones are still in an early stageof development. This asynchronous development of individual CVP primordiaand next resorption of part of them, while others persist evidence a very localcharacter of completion of CVP organellogenesis, which is different from the

70 J. KACZANOWSKA

global character of assessment of CVP-competent territories observed at thecellular level of organization. This statement is consistent with a distinction madebetween large-scale and short-range mechanisms of patterning in ciliates(Frankel, 1979).

The final pattern of CVPs results from the spatially uniform resorption ofabout 35 % of the total number of CVP primordia. Resorption of the super-numerary CVP primordia does not modify the global map of CVP distributionover the ventral field. However, CVP primordia occurring at particular sites(e.g. CVP-1) are positively selected to persist. This juxtaposition of positiveselection of certain of the CVP primordia and a randomness of the global fatesof CVPs observed at the cellular level of organization might be understood byassuming an early structural maturation of the CVP primordia positioned atsites of increased specificity of CVP positioning. They might be sufficientlydeveloped at the critical period of divisional morphogenesis (Kaczanowska &Kiersnowska, 1976) of Chilodonella to escape resorption.

Thus the global map of CVPs distribution in Ch. steini would result from thesum of the individual determinations of the fates of each CVP primordium,superimposed on an initially spatially non-uniform distribution of CVPprimordia.

In terms of the set of theoretical models of CVPs distribution on the ventralsurface of Ch. steini (Fig. 1) presented here data are consistent with model Aapplied to dissect a CVP competent zone out of 'forbidden' zones at the sitesof stomatogenesis and around the border of the ventral field. On remainingzone there are three preferred sectors of CVP primordia occurrence with certaindispersion of their placement (Model B). However, along these sectors thepositive control of placement of certain CVP primordia is also established(perhaps consistent with Model C). Resorption, though globally random,involves a positive selection of at least CVP-1 primordium (Model Cl).

I am most grateful to Dr Joseph Frankel for extensive discussions, helpful suggestions andcriticisms in the development and final shaping of this manuscript. The author would liketo thank Dr Maria Jerka-Dziadosz, Dr Krystyna Golinska and Dr Andrzej Kaczanowskifor critical reading of the draft of this manuscript.

This work is partially supported by a research grant of the Polish Academy of SciencesP.A.N.-II. 1.3.7.

REFERENCES

BEISSON, J. & SONNEBORN, T. M. (1965). Cytoplasmic inheritance of the organization of thecell cortex in Paramecium aurelia. Proc. natn. Acad. Sci., U.S.A. 53, 275-282.

ELLIOT, A. M. & KENNEDY, J. R. (1973). Morphology of Tetrahymena. In Biology ofTetra-hymena (ed. A. M. Elliot), pp. 57-87. Stroudsbury, Pa: Dowden, Hutchinson and Ross Inc.

FRANKEL, J. (1979). An analysis of cell-surface patterning in Tetrahymena. In Determinantsof Spatial Organization (ed. S. Subtelny & I. R. Konigsberg), pp. 215-246. New York:Academic Press.

FRANKEL, J. & HECKMANN, K. (1968). A simplified Chatton-Lwoff silver impregnation pro-cedure for use in experimental studies with ciliates. Trans. Amer. microsc. Soc. 87, 317-321.

Contractile vacuole pores patterning in a ciliate 71GRIMES, G. W. (1976). Laser microbeam induction of incomplete doublets of Oxytricha

fallax. Genet. Res. Cambridge 27, 213-226.GRIMES, G. W. & L'HERNAULT, S. W. (1979). Cytogeometrical determination of ciliary

pattern formation in the hypotrich ciliate Stylonychia mytilus. Devi Biol. 70, 372-295.HECKMANN, K. & FRANKEL, J. (1968). Genie control of cortical pattern in Euplotes. J. exp.

Zool. 168, 11-38.JERKA-DZIADOSZ, M. & FRANKEL, J. (1979). A mutant of Tetrahymena thermophila with a

partial mirror image duplication of cell surface pattern. I. Analysis of phenotype. / .Embryol. exp. Morph. 49, 167-202.

KACZANOWSKA, J. (1971). Studies on topography of the cortical organelles of Chilodonellacucullulus. (O.F.M.). III. Morphogenetic movements, regional multiplication of kineto-somes and cytokinesis in normal dividers and after phenethyl alcohol treatment. ActaProtozool. 8, 84-103.

KACZANOWSKA, J. (1974). The pattern of morphogenetic control in Chilodonella cucullulus.J. exp. Zool. 187, 47-62.

KACZANOWSKA, J. (1975). Shape and pattern regulation in regenerants of Chilodonellacucullulus (O.F.M.). Acta Protozool. 13, 343-360.

KACZANOWSKA, J. & KIERSNOWSKA, M. (1976). Thermosensitivity of pattern in a ciliateChilodonella cucullulus. J. exp. Zool. 196, 135-148.

KACZANOWSKI, A., RADZIKOWSKI, S., MALEJCZYK, J. & POLAKOWSKI, I. (1980). Study onintraclonal pairing in Chilodonella steini. Evidence of abortive conjugation. / . exp. Zool.213, 262-269.

NANNEY, D. L. (1966tf). Cortical integration in Tetrahymena. An exercise in cytogeometry.J.exp. Zool. 161, 307-318.

NANNEY, D. L. (19666). Corcicotype transmission in Tetrahymena. Genetics 54, 955-968.NANNEY, D. L. (1967). Cortical slippage in Tetrahymena. J. exp. Zool. 166, 163-169.NANNEY, D. L. (1968). Cytogeometric integration in the ciliate cortex. Ann. N.Y. Acad. Sci.

193, 14-28.NG, S. F. (1979). The precise site of origin of the contractile vacuole pore in Tetrahymena

and its morphogenetic implications. Acta Protozool. 18, 305-312.NG, S. F. & FRANKEL, J. (1977). 180°-rotation of ciliary rows and its morphogenetical

implications in Tetrahymena pyriformis. Proc. natn. Acad. Sci., U.S.A. 74, 1115-1119.RADZIKOWSKI, S. (1966). Study on morphology, division and postconjugation morpho-

genesis in Chilodonella cucullulus (O.F. Mueller). Acta Protozool. 4, 90-96.RADZIKOWSKI, S. & GOLEMBIEWSKA, M. (1977). Chilodonella steini, morphology and culture

method. Protistologica 13, 381-389.SOKAL, R. R. & ROHLF, F. J. (1969). Biometry. San Francisco: Freeman, Cooper and Co.

{Received 6 May 1980, revised 10 April 1981)