This report introduces a new system in the study of programming of genomic function during development of the somatic nucleus of Paramecium tetraurelia. Previous works have established a definite, but replaceable, role of the germ nuclei (micronuclei) in oral development in the asexual cycle; their removal from the cell generates viable amicronucleate cell lines, which characteristically suffer a transient period of growth depression marked by abnormal oral development. Such cell lines gradually recover, showing that a compensatory mechanism is activated in the absence of the germ nuclei to bring the cell back to near-normal. To test the notion that the somatic nucleus (macronucleus) is involved in this compensation, cells possessing micronuclei were treated with 5-azacytidine during sexual reproduction when new somatic nuclei develop. These cells were then propagated asexually for a number of fissions in the absence of the drug, and thereafter micronuclei were removed from them. The amicronucleate cell lines generated in this manner clearly did not suffer a depression as severe as the untreated controls did in terms of growth rate and oral development, and they recovered much sooner. This supports the notion that the somatic nucleus is the physical basis of the compensatory mechanism. This study suggests that the stomato-genic sequences in question normally become repressed in the somatic nucleus developing in sexual reproduction, and that 5-azacytidine administered to the cells at this time could alter this programme which then persists during subsequent asexual propagation. The possibility that the somatic nucleus is programmed by methylation of cytosine at the 5’ position is discussed.

The somatic nucleus of ciliated protozoa offers a favourable system for studying the programming of genomic expression, which occurs shortly after formation of the zygotic nucleus in sexual reproduction. After two mitoses, the zygotic nucleus generates two somatic nuclei (macronuclei) and two germ nuclei (micronuclei). At this time, the new somatic nuclei (macro-nuclear anlagen) are programmed to express one of alternative allelic states; such established states are stably inherited in subsequent asexual clonal propagation.

The paradigm for such programming of the somatic nucleus is the determination of alternative mating types in the Paramecium aurelia species complex (Sonneborn, 1977). Other systems include aberrant expression of the trichocyst discharge phenotype in stock dll3 (Sonneborn & Schneller, 1979), the expression of surface antigens in mutant d48 (Preer, 1986) and, in Tetrahymena thermophila, the expression of the SerH cell surface antigen genes (Doerder & Berkowitz, 1987). The underlying mechanism of programming of the somatic nucleus in these systems (and also in others - see Sonneborn, 1977 for summary) is not understood, apart from the involvement of genomic processing in d48. External factors, notably temperature, are known to affect determination of the developing somatic nucleus, in favouring the expression of particular allelic states.

The present report introduces a new system, in Paramecium tetraurelia, for the study of programming of genomic expression of the developing macronucleus, and suggests DNA methylation as an underlying mechanism. This concerns the function of the somatic nucleus in the development of the oral apparatus (Figs 1 and 2). During binary fission, the ciliate produces a new oral apparatus to be passed on to the posterior daughter cell. Stomatogenesis in asexual reproduction, however, is not solely dictated by the somatic nucleus but also involves the micronucleus. Amicronucleate cell lines suffer, shortly after their generation by elimination of the micronuclei, a transient depression period of about 30 fissions, characterized by slow fission rate and abnormal stomatogenesis (1986) have demonstrated that the micronucleus plays a definitive, but replaceable, role in stomatogenesis during asexual reproduction. The point of interest presently concerns the mechanism of replacement of this stomatogenic function of the micronucleus in its absence.

Figs 1-9

Figs 1 and 2 show the ventral view of a silverimpregnated postautogamous micronucleate cell with a normal oral apparatus, in two focal levels. The dorsal (DP) and ventral peniculi (VP) are each made up of four closelypacked basal body rows. At their dorsal side lies the quadrulus (Q) comprising four basal body rows that are spread out in the anterior half. At level S, the ventral peniculus terminates while the dorsal peniculus and the quadrulus make a left turn to form an S-shape spiral along he wall of the cytopharynx (C). The vestibule, the left side of which (LV) is in focus, leads to the buccal cavity. PF, postoral microtubules; MF, fragments of the preautogamous macronucleus. The horizontal lines define the anteroposterior span of the quadrulus for assessing the length of the oral apparatus. X1800. Figs 3-9 show silver-impregnated amicronucleate cells with abnormal oral apparatuses, all in ventral view and X1800. Figs 3-8 are vegetative cells showing reduction in the length of the oral apparatus. In addition, in Fig. 3 the quadrular basal body rows in the anterior part are misaligned (arrow). In Fig. 4, the quadrulus appears to be subdivided into an anterior portion (aQ) and a posterior portion (pQ) along a horizontal line. The three rightmost basal body rows of the two parts have shifted laterally with respect to each other. The two rightmost basal body rows of the anterior part are slightly out of focus. FV, food vacuole. In Fig. 5, the quadrulus and dorsal peniculus are split horizontally into anterior and posterior portions. The anterior dorsal peniculus (aDP) is connected with the posterior quadrulus (pQ). The basal bodies of the anterior quadrulus (aQ) also exhibit a chaotic pattern. In Fig. 6, all three membranelles are split horizontally into two anterior and posterior portions and are laterally displaced with respect to each other. The anterior quadrulus (aQ) is connected with the posterior dorsal peniculus (pDP); the anterior dorsal peniculus (aDP) is connected with the posterior ventral peniculus (pVP) to form a single smooth band of four closely packed basal body rows. aVP, anterior portion of ventral peniculus; pQ, posterior portion of quadrulus. In Fig. 7, two basal body rows of the dorsal peniculus are spread out in the anterior part (arrow), resembling the quadrulus. Fig. 8 shows a cell lacking a buccal cavity. The oral membranelles and the cytopharyngeal membrane (CM) all become exposed on the ventral surface. This cell does not feed. Fig. 9 shows an amicronucleate cell after the sexual cycle (autogamy). No oral membranelles or buccal cavity are developed and the pre-existing oral apparatus is resorbed, so that the cell is left with only a few dozen basal bodies in a shallow oral depression (arrow). More examples of abnormal oral features of amicronucleates can be found in previous publications (Ng & Mikami, 1981; Ng, 1988).

Figs 1-9

Figs 1 and 2 show the ventral view of a silverimpregnated postautogamous micronucleate cell with a normal oral apparatus, in two focal levels. The dorsal (DP) and ventral peniculi (VP) are each made up of four closelypacked basal body rows. At their dorsal side lies the quadrulus (Q) comprising four basal body rows that are spread out in the anterior half. At level S, the ventral peniculus terminates while the dorsal peniculus and the quadrulus make a left turn to form an S-shape spiral along he wall of the cytopharynx (C). The vestibule, the left side of which (LV) is in focus, leads to the buccal cavity. PF, postoral microtubules; MF, fragments of the preautogamous macronucleus. The horizontal lines define the anteroposterior span of the quadrulus for assessing the length of the oral apparatus. X1800. Figs 3-9 show silver-impregnated amicronucleate cells with abnormal oral apparatuses, all in ventral view and X1800. Figs 3-8 are vegetative cells showing reduction in the length of the oral apparatus. In addition, in Fig. 3 the quadrular basal body rows in the anterior part are misaligned (arrow). In Fig. 4, the quadrulus appears to be subdivided into an anterior portion (aQ) and a posterior portion (pQ) along a horizontal line. The three rightmost basal body rows of the two parts have shifted laterally with respect to each other. The two rightmost basal body rows of the anterior part are slightly out of focus. FV, food vacuole. In Fig. 5, the quadrulus and dorsal peniculus are split horizontally into anterior and posterior portions. The anterior dorsal peniculus (aDP) is connected with the posterior quadrulus (pQ). The basal bodies of the anterior quadrulus (aQ) also exhibit a chaotic pattern. In Fig. 6, all three membranelles are split horizontally into two anterior and posterior portions and are laterally displaced with respect to each other. The anterior quadrulus (aQ) is connected with the posterior dorsal peniculus (pDP); the anterior dorsal peniculus (aDP) is connected with the posterior ventral peniculus (pVP) to form a single smooth band of four closely packed basal body rows. aVP, anterior portion of ventral peniculus; pQ, posterior portion of quadrulus. In Fig. 7, two basal body rows of the dorsal peniculus are spread out in the anterior part (arrow), resembling the quadrulus. Fig. 8 shows a cell lacking a buccal cavity. The oral membranelles and the cytopharyngeal membrane (CM) all become exposed on the ventral surface. This cell does not feed. Fig. 9 shows an amicronucleate cell after the sexual cycle (autogamy). No oral membranelles or buccal cavity are developed and the pre-existing oral apparatus is resorbed, so that the cell is left with only a few dozen basal bodies in a shallow oral depression (arrow). More examples of abnormal oral features of amicronucleates can be found in previous publications (Ng & Mikami, 1981; Ng, 1988).

Two previous studies have implicated the involvement of the somatic nucleus in replacing micronuclear function (in Stylonychia, Atnmermann, 1970; in Paramecium bursaria, Fokin, 1983; discussed in Ng, 1986). Directly relevant evidence is obtained by manipulating, with cytidine analogues, DNA synthesis in the somatic nucleus of amicronucleate P. tetraurelia (Ng, 1988) (see also Discussion). Our working hypothesis is that normally in the somatic nucleus some cytosine bases associated with the stomatogenic DNA sequences in question are methylated, and the methylated state is maintained in the presence of the micronucleus. Following removal of the micronucleus, these sequences gradually become activated and their activation is promoted by demethylation. This hypothesis is in line with the notion of gene activation by demethylation revealed in other systems (reviewed by Adams & Burdon, 1982; Doerfler, 1983; Bird, 1984; Razin & Cedar, 1984; Taylor, 1984). It is reasonable to assume, furthermore, that methylation is established during development of the somatic nucleus at the beginning of the clonal cycle, and thereafter this programme remains stable in subsequent asexual propagation unless the micronuclei are removed. The present study tests this notion by attempting to interfere with the programming of somatic nucleus, by administering 5-azacytidine to micronucleate cells in sexual reproduction. The drug-treated cells were then propagated asexually for a certain number of fissions before removing their micronuclei. We here report that amicronucleate cell lines derived from such drug-treated cells indeed suffered a distinctively less severe depression, and they recovered much sooner than usual.

Cells and culture

Paramecium tetraurelia stock 51 mating type VII was employed. The cells were cultured in cerophyl medium (2·5gl−1, phosphate-buffered at pH around 7), bacterized with Enterobacter aerogenes and supplemented with stigmasterol (5 mg I−1). The general methods followed those of Sonneborn (1950, 1970). All experiments were performed at 27 °C and back cultures were stored at 15–17 °C.

Drug treatment

5-azacytidine (Sigma Co., St Louis, USA) was used in three experiments. In each, the drug solution, freshly prepared, was administered to the cells in plastic depression slides, three times on three successive days covering the period from about 5 fissions before autogamy (uniparental sexual reproduction) up to the first postautogamous cell fission (detailed in Table 1). During this period, DNA duplicated about 9 times in the presence of the drug before the stage of development of the macronuclear anlagen (about 5 times in the vegetative micronucleus before autogamy; ‘4 times during autogamy comprising one premeiotic, one pregametic and two postzygotic duplications). During maturation of the macronucleus, additional rounds of DNA duplication also occurred in the presence of the drug.

Table 1

The condition of treatment of cells with 5-azacytidine in three experiments

The condition of treatment of cells with 5-azacytidine in three experiments
The condition of treatment of cells with 5-azacytidine in three experiments

After the last drug treatment, postautogamous cells were isolated from each experiment. These were then expanded into clones by allowing exponential growth in glass depression slide cultures. These clones were then ready for emicronucleation.

Emicronucleation

The micromanipulation set-up for removing the two micronuclei from the cell was as described by Ng (1981), except that one injection needle was used. The two micronuclei were removed simultaneously from the cell, or in others removal of the first one was followed later by propagation of the cell line and removal of the second one. Several fissions after the operation, some cells were stained to evaluate the absence of micronuclei. Altogether 16 amicronucleate cell lines from 12 postautogamous clones were generated. Twelve of the cell lines originated between 4 and 20 fissions after autogamy; four between 30 and 45 fissions. Their clonal origin and the age when they were enucleated are detailed in Fig. 11. Each amicronucleate cell line is designated according to experiment (I, II, or III), and also its clonal origin (1, 2, 3 or 4); a and b denote two amicronucleate cell lines obtained independently by two operations on the same clone. The control set contains amicronucleate cell lines derived, at similar clonal ages (see Fig. 11), from cells not treated with the drug.

Sampling and cytology

For nuclear staining, aceto-orcein without osmium fixation was used (Beale & Jurand, 1966). For monitoring the recovery of amicronucleate cell lines, they were sampled periodically at different clonal ages for silver impregnation (Chatton & Lwoff, 1936; Corliss, 1953). Their oral structures were examined and the length of the oral apparatus, defined by the anteroposterior span of the longest oral membranelie (quadrulus), was measured with an ocular micrometer under X1000 phase-contrast optics (see Figs 1 and 2).

Determination of cellular fission rate

The best-looking cells of the amicronucleate cell lines were selected for propagation. Nine cells were transferred, three into each depression slide well. The number of cells was scored on the next day. This was repeated on consecutive days for several days and the scores were pooled to give the mean fission rate during a certain period.

Statistics

For comparing two percentages, 2×2 G-test (applying William’s correction when n<200) was performed. For comparing two mean values (fission rate or length of the oral apparatus) t-test was performed. For comparing two ratios involving small numbers exact probability of the binomial was calculated (Sokal & Rohlf, 1981).

As in previous observations (Ng & Mikami, 1981; Ng & Tam, 1987; Chau & Ng, 1988), three amicronucleate cell lines derived from cells not treated with the drug in this study exhibited growth depression, marked by slow growth rate and abnormal oral development (Figs 3-8), but they gradually recover to near-normal. Compared to the three controls, 13 amicronucleate cell lines derived from drug-treated cells in three experiments exhibited a distinctively less severe depression. The amicronucleate control and experimental cell lines are compared, at defined periods after their origin, in terms of fission rate, length of oral apparatus, normality of oral membranelles, and presence of buccal cavity, as follows.

Fission rate

Micronucleates normally divide about 4 times a day at 27°C. Amicronucleates have a lower fission rate, especially during the depression period (Table 2). Compared to controls, the experimental cell lines propagated faster, especially during the first 20 fissions after enucleation. Indeed, during the next period (21–40 fissions) of propagation, the experimental cell lines had attained a fission rate equal to, or surpassing, that of the controls after 41 fissions. Hence, the experimental cell lines regained vigour much sooner than the controls did. Amicronucleate cell lines of experiment III, for unknown reasons, grew more slowly in the third period (after 41 fissions); a higher frequency of oral abnormality was also seen in this period (see below).

Table 2

Mean fission rate of amicronucleate cell lines in daily reisolation cultures at various periods (fissions) after enucleationa,b

Mean fission rate of amicronucleate cell lines in daily reisolation cultures at various periods (fissions) after enucleationa,b
Mean fission rate of amicronucleate cell lines in daily reisolation cultures at various periods (fissions) after enucleationa,b

Oral length

During the depression period, the mean length of the oral apparatus of amicronucleate cell lines declined. The oral length was reduced to a much lesser extent in the experimentáis, and they also recovered much sooner (Fig. 10A-D). This conclusion is evident when experimental and control cell lines are compared (Table 3; except for cell lines of experiment III at 70-109 fissions). Moreover, experimental cell lines I-2b and 1-3 did not exhibit the characteristic depression ‘trough’; instead, their mean oral length never fell below 24 μm, a value equivalent to those having recovered.

Fig. 10

The mean oral length of amicronucleate cell lines during asexual propagation after emicronucleation. Vertical bar above or below points on curves show + or − standard error of mean. Arrows represent the time of emicronucleation of the cell lines. In Experiment HI, the curves are not connected to their origins because of the uncertainty in the early part of the curves due to the lack of sampling before 10 fissions since origination of the amicTonucleate cell lines. It is clear that if a depression exists very shortly after removal of the micronuclei, as is the case with the cell lines of Experiments I and II, the cell lines of Experiment IH have recovered within twenty fissions.

Fig. 10

The mean oral length of amicronucleate cell lines during asexual propagation after emicronucleation. Vertical bar above or below points on curves show + or − standard error of mean. Arrows represent the time of emicronucleation of the cell lines. In Experiment HI, the curves are not connected to their origins because of the uncertainty in the early part of the curves due to the lack of sampling before 10 fissions since origination of the amicTonucleate cell lines. It is clear that if a depression exists very shortly after removal of the micronuclei, as is the case with the cell lines of Experiments I and II, the cell lines of Experiment IH have recovered within twenty fissions.

Table 3

The highest and lowest values of mean oral length (in μm) of amicronucleate cell lines at various periods (fissions) after enucleationa,b

The highest and lowest values of mean oral length (in μm) of amicronucleate cell lines at various periods (fissions) after enucleationa,b
The highest and lowest values of mean oral length (in μm) of amicronucleate cell lines at various periods (fissions) after enucleationa,b

The response to the drug appears to depend on dosage and probably also number of rounds of DNA duplication in the presence of the drug. Cell lines of experiment I showed the greatest response compared to the other two (Table 3). In experiment III, the drug dosage was halved. In experiment II, the postautogamous cells were starved, and hence the developing macronuclear anlagen had probably undergone fewer rounds of DNA duplication in the drug during the 3rd treatment (see Table 1). This is because DNA accumulation in the cell cycle is sensitive to nutrient level (Ching & Berger, 1986a,b); in starved postautogamous cells, the macronuclear anlagen are slow to develop intense staining with aceto-orcein indicating a slow rate of DNA accumulation.

The effect of the drug persisted for a large number of cell generations. Three experimental cell lines (I–3, II-4 and III-3), derived after 30–45 fissions since drug treatment of progenitor autogamous cells, also recovered much sooner than the control cell line (2b) did (Table 3). Hence, the effect of the drug lasted for at least 30 –45 rounds of DNA replication during asexual propagation. Additionally, some data also indicated that five amicronucleate cell lines generated even later, at 85–100 fissions, from drug-treated cells of experiments I and II, also recovered sooner than the controls did (S. F. Ng & S. C. Loo, unpublished).

Pattern of oral membranelles

The percentages of abnormal oral membranelles at various periods of propagation of the cell lines are given in Fig. 11. Table 4 compares the experimental cell lines with each other, and also with the controls. The conclusion of this analysis parallels that on mean oral length, in that the experimental cell lines recovered sooner; the difference between experiments I and II is, however, not as obvious.

Fig. 11

Sampling of amicronucleate cell lines in the clonal life, and the percentage (sample size) of abnormal oral membranelles in each sample.

Fig. 11

Sampling of amicronucleate cell lines in the clonal life, and the percentage (sample size) of abnormal oral membranelles in each sample.

Table 4

The highest and lowest frequencies of abnormal oral membranelles in amicronucleate cell lines at various periods (fissions) after enucleationa,b

The highest and lowest frequencies of abnormal oral membranelles in amicronucleate cell lines at various periods (fissions) after enucleationa,b
The highest and lowest frequencies of abnormal oral membranelles in amicronucleate cell lines at various periods (fissions) after enucleationa,b

Amicronucleate cell lines do not fully recover (Ng & Mikami, 1981; Ng & Tam, 1987). Even after prolonged propagation (before clonal aging), up to 10% of their oral apparatuses are abnormal. Inspection of Fig. 11 suggested that, in later periods of propagation when the cell lines had largely recovered, the experimental ones were less likely to produce abnormal oral membranelles. Because of the low percentages involved, this difference is not obvious from the 2×2 G-test in Table 4, but becomes apparent from an alternative analysis, as follows.

The design is to test whether most samples of the recovered experimental cell lines had produced a lower percentage of abnormal membranelles than that of the controls, by considering only those samples with ⩽ 10% abnormal membranelles, which represented the situation after recovery. Taking the lowest control value (6·8% abnormal, see Fig. 11) as reference, it is found that most samples of the recovered experimental cell lines had a smaller percentage of cells with abnormal oral membranelles (Exp. I, 16 samples less than 6-8% abnormal vs 1 sample more than 6·8% abnormal; II, 15 vs 2; III, 10 vs 4; against the null hypothesis of 1:1 the expected exact probabilities are 0·001, 0·001 and 0·06, respectively). Similarly, compared to the highest control value after recovery (8-2%, Fig. 11), the corresponding ratios of the experimental group are: I, 17:0; n, 16:1; 12:2 (P< 0-0001, P = 0-0001 and P = 0·006, respectively).

Presence of buccal cavity

One of the stomatogenic abnormalities exhibited by amicronucleates is reduction of the size of buccal cavity. In the extreme case where no buccal cavity forms, the oral membranelles become exposed on the cell surface (Fig. 8). Such abnormality arises sporadically (<5%) in exponential amicronucleate cultures. Cells lacking buccal cavity become starved and enter autogamy. Amicronucleate autogamous cells resorb the pre-existing one as usual, but fail to produce a new one, and become astomatous after autogamy (Fig. 9) (Ng & Mikami, 1981; Ng & Newman, 1984; Ng & Tam, 1987). There-fore, the appearance of postautogamous astomatous cells in exponential cultures, indicating loss of buccal cavity in these cells, provides an additional parameter for assessing the extent of stomatogenic abnormality, as follows.

From exponentially growing control cell lines, 7 samples yielded postautogamous astomatous cells; 5 did not. This ratio, 7:5, differs from those of the three experiments (I, 1:17; II, 5:20; III, 6:12; exact probabilities, 0·000001, 0·003 and 0·05, respectively). Thus, amicronucleates derived from drug-treated cells were more capable of developing a functional oral apparatus. Again, the drug had produced the greatest response in experiment I.

The present study shows that the drug 5-azacytidine applied to Paramecium during autogamy can alter the programming of the developing somatic nucleus. Amicronucleate cell lines derived from such cells a dozen or more fissions later suffered a much less severe depression, in terms of growth rate and oral development, than those derived from untreated controls. The simplest interpretation of this finding is that, following fertilization, the normal programming of the new somatic nucleus involves repression of some of its stomatogenic DNA sequences. After removal of the micronuclei, such sequences are gradually activated to replace the stomatogenic function of the micronuclei. The normal repression of such sequences during development of the somatic nucleus can be prevented by treatment with 5-azacytidine and the altered programme is maintained during subsequent asexual propagation. The macronuclear sequences having been activated are thus ready to take over the function of the micronuclei, in initiating recovery of the cell line, once the micronuclei are removed.

Implicit in the above interpretation is the assumption of a role of the micronucleus in maintaining the macronuclear programme established in sexual reproduction. Since all amicronucleate cell lines (from this and previous studies) invariably enter depression soon after removal of the micronuclei, this indicates that the macronuclear stomatogenic sequences in question are all along repressed owing to the presence of the micronucleus. It appears, furthermore, that this role of the micronucleus is only permissive, and not instructive, in that the micronucleus does not actively create the state of repression of those sequences, nor can it reverse their state of activation once established by drug treatment. This is simply because after drug treatment the activated state of the macronuclear sequences coexists with the micronuclei during asexual propagation.

The above interpretation centres around the notion of repression of macronuclear stomatogenic DNA sequences as a normal programme during macronuclear development, and their activation under unusual circumstances. Can the present observations be explained otherwise? 5-azacytidine is known to be weakly mutagenic in prokaryotes and to cause neoplastic transformation of mouse cells; but there is no unequivocal support for a mutational basis for the biological effect of the drug on eukaryotes (Doerfler, 1983; Taylor, 1984). A mutational explanation for our observations, in view of the specific response to the drug, would require that the macronuclear stomatogenic sequences in question be mutational hotspots. The drug is also known to inhibit methylation of t-RNA (Taylor et al. 1984). The heritability of the effect of the drug over a considerable number of cell generations, however, calls for an explanation at the level of the genome.

What is the mechanism underlying the effect of 5-azacytidine in the present system? The analogue is commonly held to be responsible for demethylation of 5-methylcytosine in connection with gene activation (for review see Doerfler, 1983; Bird, 1984; Razin & Cedar, 1984; Taylor, 1984; Taylor et al. 1984). The possibility also exists that the analogue incorporated into DNA affects the affinity between genes and their regulatory proteins (Christman et al. 1985). The notion of genomic programming by methylation of cytosine is worth exploring in ciliates, for to date 5-methylcytosine has not been detected biochemically in Paramecium (Cummings et al. 1974), Tetrahymena (Gorovsky et al. 1973; Hattman et al. 1978; Pratt & Hattman, 1981), Oxytricha (Rae & Spear, 1978) and Stylonychia (Ammermann et al. 1981). Instead, N6-methyadenine is present. Ciliates thus resemble some insects, notably Drosophila, as exceptions among eukaryotes in not containing detectable levels of 5-methylcytosine in the genome (Cedar et al. 1979; Rae & Steele, 1979; Eastman et al. 1980). It has been speculated that the Drosophila genome may contain ‘an exceedingly low amount of 5-methylcytosine in highly strategic positions’ (Doerfler, 1983). If so, methylation may serve as a specific mechanism, in ciliates and insects, for the exclusive control of expression of a few genes.

There are three reasons for hypothesizing that cytosine methylation is a key feature in programming the macronuclear stomatogenic genes. First, the limit of sensitivity of biochemical detection of 5-methylcytosine, in several ciliates, is at best 0·01 molepercent of cytosine (see Pratt & Hattman, 1981). Taking a genome size of about 60 000 kb and a GC content of 26% (as in the sibling species P. primaurelia, see Freiburg, 1988), the presence of fewer than 800 5-methylcytosine bases would not be detected. Since gene activation may involve demethylation of only a few 5-methylcytosine bases (e.g. Toniolo et al. 1984) there is ample room for postulating programming by methylation in Paramecium.

Second, in a complementary study, a number of cytidine analogues have been administered to amicronucleate Paramecium cell lines that were still in the growth depression period to see if these could promote recovery (Ng, 1988). Their recovery from depression was indeed speeded up by 5-azacytidine, 5-aza-deoxycytidine, and also 5-fluoro-2’-deoxycytidine. On the other hand, cytidine and other analogues unaltered at the 5’ position, including 6-azacytidine, 2’-fluoro-2’-deoxycytidine and cytosine-β-D-arabinofuranoside did not promote recovery. Such response to 5’-specific analogues suggests that these analogues affect the 5’-methylation of cytosine, and that demethylation underlies the basis of recovery of amicronucleates.

Third, cells treated with 5-azacytidine in this and the previous study (Ng, 1988) propagated normally. Their tolerance to the relatively high dosage of the drug speaks against a widespread occurrence of 5-methylcytosine in the genome, since if this were the case then demethylation would affect not only the stomatogenic sequences in question, but also other genomic sites as well. It is reasonable to assume that normally only a few genomic sites of the somatic nucleus of Paramecium are methylated, and these are mostly connected with expression of the stomatogenic sequences in question. The tolerance may also be rationalized on the basis of other mechanisms of action of 5-azacytidine, such as the interaction between DNA and regulatory proteins (Christman et al. 1985), which will have to specifically affect only the stomatogenic sequences in question.

In Tetrahymena, methylation of adenine at the N6 position occurs during development of the somatic nucleus in sexual reproduction, at about the same time as, and hence thought to be involved in, DNA rearrangement and transcriptional activation of the macronuclear genome (Blackburn et al. 1983; Harrison & Karrer, 1985; Harrison et al. 1986). The link between methylation of adenine and gene activation, however, has not yet been established. In view of the present finding in Paramecium, the notion that methylation of adenine could instead be involved in the silencing of some sequences in the somatic nucleus merits further investigation.

We thank Ms M. F. Chau for performing some of the enucleation manipulations.

Adams
,
R. L. P.
&
Burdon
,
R. H.
(
1982
).
DNA methylation in eukaryotes
.
CRC Crit. Rev. Biochem
.
13
,
349
384
.
Ammermann
,
D.
(
1970
).
The micronucleus of the ciliate Stylonychia mytilus; its nucleic acid synthesis and its function
.
Expl Cell Res
.
61
,
6
12
.
Ammermann
,
D.
,
Steinbruck
,
G.
,
Baur
,
R.
&
Wohlert
,
H.
(
1981
).
Methylated bases in the DNA of the ciliate Stylonychia mytilus
.
Eur. J, Cell Biol.
24
,
154
156
.
Beale
,
G. H.
&
Jurand
,
A.
(
1966
).
Three different types of matekiller (mu) particles in Paramecium aurelia (syngen 1)
.
J. Cell Sci
.
1
,
31
34
.
Bird
,
A. P.
(
1984
).
DNA methylation versus gene expression
.
J. Embryol. exp. Morph
.
83
Supplement
,
31
40
.
Blackburn
,
E. H.
,
Pan
,
W. C.
&
Johnson
,
C. C.
(
1983
).
Methylation of ribosomal RNA genes in the macronucleus of Tetrahymena thermophila
.
Nucleic Acid Res
.
11
,
5131
5145
.
Cedar
,
H.
,
Solage
,
A.
,
Glaser
,
G.
&
Razin
,
A.
(
1979
).
Direct detection of methylated cytosine in DNA by use of the restriction enzyme Mspl
.
Nucl. Acids Res
.
6
,
2125
2132
.
Chatton
,
E.
&
Lwoff
,
A.
(
1936
).
Techniques pour l’étude des Protozoaires, spécialement de leurs structures superficielles (cinétome et argyrome)
.
Bull. Soc. Microsc. Fr
.
5
,
25
39
.
Chau
,
M. F.
&
Ng
,
S. F.
(
1988
).
Stomatogenic function of the micronucleus in Paramecium jenningsi
.
Eur. J. Protist. (in press)
.
Ching
,
A. S. L.
&
Berger
,
J. D.
(
1986a
).
The timing of initiation of macronuclear DNA synthesis is set during the preceding cell cycle in Paramecium tetraurelia. Analysis of the effects of abrupt changes in nutrient level
.
Expl Cell Res
.
167
,
177
190
.
Ching
,
A. S. L.
&
Berger
,
J. D.
(
1986b
).
Control of cell division in Paramecium tetraurelia. Effects of abrupt changes in nutrient level on accumulation of macronuclear DNA and cell mass
.
Expl Cell Res
.
167
,
191
202
.
Christman
,
J. K.
,
Schneiderman
,
N.
&
Acs
,
G.
(
1985
).
Interaction of DNA methyltransferase and other non-histone proteins isolated from Friend erythroleukemia cell nuclei with 5-azacytosine residues in DNA
.
In Progress in Clinical and Biological Research. Biochemistry and Biology of DNA Methylation
, pp.
105
118
.
New York
:
Alan R. Liss, Inc
.
Corliss
,
J. O.
(
1953
).
Silver impregnation of ciliated protozoa by the Chatton-Lwoff Technic
.
Stain Technol
.
28
,
97
100
.
Cummings
,
D. J.
,
Tait
,
A.
&
Goddard
,
J. M.
(
1974
).
Methylated bases in DNA from Paramecium aurelia
.
Biochim. biophys. Acta
374
,
1
11
.
Doerder
,
F. P.
&
Berkowitz
,
M. S.
(
1987
).
Nucleo-cytoplasmic interaction during macronuclear differentiation in ciliate protists: Genomic basis for cytoplasmic control of SerH expression during macronuclear development in Tetrahymena thermophila
.
Genetics
117
,
13
23
.
Doerfler
,
W.
(
1983
).
DNA methylation and gene activity
.
A. Rev. Biochem
.
52
,
93
124
.
Eastman
,
E. M.
,
Goodman
,
R. M.
,
Erlanger
,
B. F.
&
Miller
,
O. J.
(
1980
).
5-methylcytosine in the DNA of the polytene chromosomes of the Diptera Sciara coprophila, Drosophila melanogaster and D. persimilis
.
Chromosoma
79
,
225
239
.
Fokin
,
S. I.
(
1983
).
Morphology and viability of cells in the amicronucleate clones of Paramecium caudatum
.
Tsitologiya
25
,
811
817
(in Russian with English summary).
Freiburg
,
M.
(
1988
).
Organization and expression of the nuclear genome
.
In Paramecium
(ed.
H. D.
Gdrtz
), pp.
141
154
.
Berlin
:
Springer-Verlag
.
Gorovsky
,
M. A.
,
Hattman
,
S.
&
Pleger
,
G. L.
(
1973
).
(N6)methyl adenine in the nuclear DNA of a eukaryote, Tetrahymena pyriformis
.
J. Cell Biol
.
56
,
697
701
.
Harrison
,
G. S.
,
Findley
,
R. C.
&
Karrer
,
K. M.
(
1986
).
Site specific methylation of adenine in the nuclear genome of a eukaryote, Tetrahymena
.
Molec. cell. Biol.
6
,
2364
2370
.
Harrison
,
G. S.
&
Karrer
,
K. M.
(
1985
).
DNA synthesis, methylation and degradation during conjugation in Tetrahymena thermophila
.
Nucl. Acids Res
.
13
,
73
87
.
Hattman
,
S.
,
Kenny
,
C.
,
Berger
,
L.
&
Pratt
,
K.
(
1978
).
Comparative study of DNA methylation in three unicellular eukaryotes
.
J. Bacterial
.
135
,
1156
1157
.
Ng
,
S. F.
(
1981
).
Transplantation of the germ nucleus in Paramecium tetraurelia and its genetic consequences
.
Protistologica
17
,
489
496
.
Ng
,
S. F.
(
1986
).
The somatic function of the micronucleus of ciliated protozoa
.
Progr. Protistol.
1
,
215
286
.
Ng
,
S. F.
(
1988
).
Cytidine analogues and stomatogenic recovery in amicronucleate Paramecium tetraurelia and Paramecium jenningsi
.
J. Protozoal, (in press)
.
Ng
,
S. F.
&
Mikami
,
K.
(
1981
).
Morphogenetic role of the germ nucleus in Paramecium tetraurelia
.
Protistologica
17
,
497
509
.
Ng
,
S. F.
&
Newman
,
A.
(
1984
).
The role of the micronucleus in stomatogenesis in sexual reproduction of Paramecium tetraurelia: conjugation of amicronucleates
.
Protistologica
20
,
517
523
.
Ng
,
S. F.
&
Tam
,
L. W.
(
1987
).
Stomatogenesis in Paramecium tetraurelia. The roles of the micronucleus and the pre-existing oral apparatus
.
Eur. J. Protistol
.
23
,
141
151
.
Pratt
,
K.
&
Hattman
,
S.
(
1981
).
Deoxyribonucleic acid methylation and chromatin organization in Tetrahymena thermophila
.
Molec. cell. Biol
.
1
,
600
608
.
Preer
,
J. R.
, IR
(
1986
).
Surface antigens of Paramecium
.
In The Molecular Biology of Ciliated Protozoa
(ed.
J. G.
Gall
), pp.
301
339
.
OrLondo
:
Academic Press
.
Rae
,
P. M. M.
&
Spear
,
B. B.
(
1978
).
Macronuclear DNA of the hypotrichous ciliate Oxytricha fallax
.
Proc. natn. Acad. Sci. U.S.A
.
75
,
4992
4996
.
Rae
,
P. M. M.
&
Steele
,
R. E.
(
1979
).
Absence of cytosine methylation at C-C-G-G and G-C-G-C sites in the rDNA coding regions and intervening sequences of Drosophila and the rDNA of other higher insects
.
Nucl. Acids Res
.
6
,
2987
2995
.
Razin
,
A.
&
Cedar
,
H.
(
1984
).
DNA methylation in eukaryotic cells
.
Int. Rev. Cytol
.
92
,
159
185
.
Sokal
,
R. R.
&
Rohlf
,
F. J.
(
1981
).
Biometry
. 2nd. ed.
San Francisco
:
Freeman and Co
.
Sonneborn
,
T. M.
(
1950
).
Methods in the general biology and genetics of Paramecium aurelia
.
J. exp. Zool
.
113
,
87
148
.
Sonneborn
,
T. M.
(
1970
).
Methods in Paramecium research
.
In Methods of Cell Physiology
, vol.
4
(ed.
D. M.
Prescott
), pp.
241
339
.
New York
:
Academic Press
.
Sonneborn
,
T. M.
(
1977
).
Genetics of cellular differentiation: stable nuclear differentiation in eucaryotic unicells
.
A. Rev. Genet
.
11
,
349
367
.
Sonneborn
,
T. M.
&
Schneller
,
M. V.
(
1979
).
A genetic system for alternative stable characteristic in genomically identical homozygous clones
.
Devel. Genet
.
1
,
21
46
.
Taylor
,
J. H.
(
1984
).
DNA Methylation and Cellular Differentiation
.
Cell Biology Monographs
, vol.
11
.
New York
:
Springer-Verlag
.
Taylor
,
S. M.
,
Constantinides
,
P. A.
&
Iones
,
P. A.
(
1984
).
5-azacytidine, DNA methylation, and differentiation
.
Curr. Top. Microbiol. Immunol
.
108
,
115
127
.
Toniolo
,
D.
,
D’Urso
,
M.
,
Martini
,
G.
,
Pérsico
,
M.
,
Tufano
,
V.
,
Battistuzzi
,
G.
&
Luzzatto
,
L.
(
1984
).
Specific methylation pattern at the 3’ end of the human housekeeping gene for glucose 6-phosphate dehydrogenase
.
EMBO J
.
3
,
1987
1995
.