Evidence for the association of rna with the ciliary basal bodies of Tetrahymena

Basal bodies, and the morphologically similar centrioles, are small (0·2 × 0·5 μm) cylindrical organelles composed in part of microtubules and are found in all Metazoa, some eukaryotic protists, and lower plants. Centrioles are found at the poles of the mitotic apparatus and at some other locations in the cell from which microtubules radiate. The term basal body has been reserved for a centriole serving as the base of a cilium or flagellum. The functions of basal bodies and centrioles are unknown, but due to their locations, it seems possible that they serve as foci for generation of other structures containing microtubules. Interest in centrioles today centres about their possible roles in the control of cell division and morphology. Several reviews of centrioles are available; the reader is referred to recent articles of Fulton (1971) and Wolfe (1972).

Some classical microscopists maintained that centrioles were self-replicating, since nascent centrioles or basal bodies appeared first near mature structures (Wilson, 1925;

Lwoff, 1950). Gall (1961) confirmed these observations with electron microscopy. In certain organisms, however, centrioles appear with no apparent parental structure (Dirksen, 1961; Dingle & Fulton, 1966) or else near structures other than mature centrioles (see Wolfe (1972), p. 175 for review). Fulton (1971) and Pickett-Heaps (1971) reviewed the ultrastructure literature concerning centriole formation and concluded that there is insufficient evidence for the older conclusion that whole centrioles, or smaller entities within centrioles, were self-reproducing.

The problems of centriole formation might be clarified by a determination of the chemical composition of isolated centrioles. Unfortunately, there is no satisfactory method of centriole isolation. Cells having free centrioles generally have fewer than 5 per cell. Ciliated cells, which may each contain thousands of basal bodies, have their centrioles integrated within larger structures which have been difficult to dissociate without destroying the structure of the centriole itself (Seaman, 1960; Argetsinger, 1965). Direct studies of chemical composition, therefore, have been restricted mostly to impure preparations of basal bodies contained in ciliate pellicles which have been relatively easy to isolate by methods based on the ethanol-digitonin procedure of Child & Mazia (1956).

Due to early notions of self-replication of centrioles, chemical analysis of ciliate pellicles, or subfractions, has centred around attempts to isolate specific basal body DNA. The results of this analysis have been mixed (Seaman, 1960; Argetsinger, 1965; Hoffman, 1965; Hufnagel, 1969) with later results indicating no detectable specific DNA (Flavell & Jones, 1971). These studies have generally suffered from uncertain contamination with nuclear or mitochondrial DNA, and were not sensitive enough to detect DNA in concentrations lower than 0-05 % of pellicle weight. There has been no reported effort made to isolate a specific RNA from pellicles or basal bodies (Fulton, 1971).

Randall & Disbrey (1965) applied acridine orange staining plus fluorescence microscopy to search for nucleic acids in fixed Tetrahymena pellicles. This fluores cence technique can detect less than 1 × 10⁸ Daltons of RNA or DNA at a small site (Mayor & Diwan, 1961; Mayor, 1963). Basal bodies fluoresced bright yellow-green, a colour characteristic of stained double-stranded nucleic acids (method reviewed by Kasten, 1967). Fluorescence was found only in synchronized cells shortly before cell division, and was eliminated by treatment of fixed pellicles with DNase, but not RNase, before staining. Smith-Sonneborn & Plaut (1967) repeated this experiment with Paramecium pellicles and obtained similar and more convincing results. Again fluorescence was detected only in pellicles isolated from certain synchronized cells, but was abolished by RNase as well as DNase. The fluorescence changed from yellow-green to the red colour characteristic of single-stranded nucleic acid after treatment with acid and formaldehyde which supposedly denatured DNA. These fluorescence studies, although not wholly consistent, provide the strongest evidence to date for the presence of a nucleic acid, DNA, in basal bodies. Complementary studies with tritiated thymidine labelling and autoradiography in the work cited above have given less definitive results due to lower sensitivity and resolution.

Three problems remain with the fluorescence studies and the conclusions derived from them. (1) Isolated pellicles or related structures may have been contaminated with DNA from nuclear chromatin or mitochondria near the basal bodies (Flavell & Jones, 1971; Younger et al. 1972). (2) Cells giving fluorescing pellicles came from only certain stages of the ciliate growth cycle. This finding may be interpreted as indicating either the transient presence of nucleic acid in basal bodies or else cell cycle-dependent blocking of fluorescence by other molecules. (3) The low resolution of light microscopy has not allowed localization of the fluorescing molecules within basal bodies.

This report extends studies of acridine orange fluorescence in Tetrahymena pellicles. We have developed a new method of pellicle isolation giving cleaner and more stable preparations than those available previously. We now find that all basal bodies can be made to fluoresce regardless of the phase of the cell cycle. Further, fluorescence appears to arise from RNA, not DNA. Implications of these findings are discussed. A preliminary version of part of this work has been reported in abstract form (Hartman, Moss & Gurney, 1972).

Tetrahymena pyriformis, strain W, was obtained from Dr W. Balamuth of this University. Two types of synthetic growth medium were used. Tryptone-yeast extract medium contained, in one litre of water, 10 g Bacto-Tryptone (Difco), 5 g yeast extract (Difco), and 5 g NaCl. A modification of Frankel’s medium (personal communication from Dr J. Frankel, University of Iowa) contained, in one litre of water, 3 g Bacto-Tryptone (Difco), 5 g D-glucose, 0·3 g MgSO₄,1·8 mg nicotinic acid, 1·5 mg calcium pantothenate, 1·0mg thiamine hydrochloride, 0·9 mg riboflavine monophosphate, 0·2 mg pyridoxine hydrochloride, 40 μg DL-6-thioctic acid, 40 μg folic acid, and 4 μg biotin. Cells were cultured in one litre of medium in a 4-1. Ehrlenmeyer flask at 25 °C without aeration until the cell count reached 2 × 10⁵ organisms per ml (late log phase). In some experiments cells were synchronized by temperature changes according to the procedures of Zeuthen as described by Randall & Disbrey (1965).

One litre of cells (2 × 10⁸ organisms) was sedimented (1000 g, 5 min, 4 °C) and resuspended at 25 °C in 25 ml 0·1 M sodium ethylenediaminetetra-acetate (EDTA) pH 7·0 plus 1 % (v/v) Tween 80 (Baker Chemical Co.). Concentrated cells could remain motile in this homogenization medium for about 30 min. After resuspension cells were homogenized at 4000 rev/min at 25 °C using a ‘Virtis 45’ rotary homogenizer with a single blade 4 cm in diameter for 5 to 40 min. In some experiments 6 mM diethylpyrocarbonate (DEP) was added to concentrated cells immediately before homogenization. The homogenate was monitored with phase-contrast microscopy during homogenization. Whole cells were progressively converted to whole pellicles. Pellicles were then washed 3 times by sedimentation (2000 g, 1 min, 25 °C) and resuspension in fresh homogenization medium.

Washed pellicles were air-dried on microscope slides and fixed in modified Carnoy’s solution as described by Randall & Disbrey (1965). Fixed pellicles on slides were stained for 30 min at 25 °C in 0·025 mg/ml acridine orange as described by Smith-Sonneborn & Plaut (1967). Fixed and stained specimens were then examined with water between slide and coverslip. We used phase-contrast and fluorescence microscope equipment similar to that described by Smith Sonneborn & Plaut (1967). Colour photographs were taken with Kodak High Speed Ektachrome film, Type B, and developed at a film speed of ASA 320.

Fixed pellicles on slides were treated before staining with one of the enzyme or protein solutions described below. Ribonuclease was RNase A, Worthington code RAF, dissolved at 1 mg/ml in RNase buffer, which contained 10 mM Tris-HCl pH 8·0, 1 mM sodium EDTA. RNase was heated in solution at 80 °C for 10 min to ensure no contamination with deoxyribo-nuclease activity. Deoxyribonuclease was the electrophoretically purified beef pancreas enzyme, Worthington code DPFF, dissolved at 1 mg/ml in DNase buffer which contained 10 mM Tris-HCl pH 7·5, 100 mM NaCl, 5 mM MgCl₂. Pronase (Calbiochem grade B) was dissolved in RNase buffer at 1 mg/ml, preincubated at 37 °C for 1 h to destroy contaminating nuclease activities, and then diluted to 10 μg/ml in RNase buffer. Rabbit histone (gift of Dr R. D. Cole) and egg white lysozyme (Sigma) were dissolved in RNase buffer at 1 mg/ml. Slides were also treated in either buffer alone. The standard exposure of fixed pellicles to enzymes or proteins was 3 h at 37 °C, except for pronase which was 1 h. After treatment, enzymes and proteins were removed by washing slides in 0·1N acetic acid and then in distilled water.

Fixed pellicles on slides were treated for 1 h at 25 °C with 50 mM 2-mercaptoethanol in RNase buffer before staining.

Washed pellicles were fixed in 10 mM sodium phosphate, pH 7·5, plus 1·6% glutaraldehyde for 30 min at 2 °C, centrifuged and stained in the pellet with 1% OsO₄ for 30 min at 25 °C, washed in 70% ethanol, dehydrated, embedded and polymerized in Shell Epon 812. Thin sections were stained with lead citrate and uranyl acetate and then examined in a Siemens electron microscope at magnifications of 4000, 8ooo, 20000 and 40000.

Stable RNA associated with pellicles was labelled by incubating cells for 48 h in Frankel’s medium with 0·5 μCi/ml [8-³H]guanosine (Schwartz-Mann 6’6 Ci/mmol) or 0·5 μCi/ml [5-³H]uridine (New England Nuclear 42 Ci/mmol). The cells were then removed from radio active medium by sedimentation, resuspended in fresh non-radioactive medium, and incubated 24 h further before harvesting and pellicle isolation as described above.

Washed pellicles at 10⁷ per ml were extracted with a detergent-containing buffer (0·01 M Tris-HCl pH 8·0, 1·0 mM EDTA, 0·5% (w/v) sodium dodecyl sulphate (SDS)) for 30 min at 25 °C by vigorous pipetting or shaking. The mixture was then centrifuged once (1000g, 5 min, 25 °C). The supernatant was saved and extracted twice at 25 °C with an equal volume of distilled phenol which had been shaken with the SDS-containing buffer described above. RNA was precipitated from the final aqueous phase by adding NaCl to 0-3 M and then adding 2 volumes of 95% ethanol. After chilling for 18 h at —20 °C, RNA was recovered by sedi mentation (19000 g 45 min, o °C) and dissolved in hybridization buffer (0·06 M Na₂HPO₄, 0·06 M NaH₂PO₄, 1 mM EDTA, pH 6·8). Some contaminating glycogen was removed from the final solution by sedimentation (150000 g, 30 min, 0 °C). Recovery of acid-precipitable radio-activity from washed pellicles was > 80%.

Ribosomes were isolated from cultures labelled as described above and from unlabelled cultures according to procedures described by Kumar (1969). Ribosomes were further purified by sedimentation through solutions containing 1·35 M and 2·0 M sucrose according to the procedure of Blobel (1971). Ribosomal RNA was then extracted from purified ribosomes according to Kumar’s (1969) procedures and dialysed extensively against 0·12 M sodium phosphate, 1 mM EDTA, pH 6·8. RNA concentrations were determined using = 200. Typically A₂₆₀ /A₂₃₀A₂₆₀ =1·9–2·1 and /A₂₃₀ =2·2–2·4

Unlabelled cells were sedimented and resuspended at 2 × 10⁷/ml in ice-cold DNA extraction buffer (0·01 M Tris-HCl pH 8·0, 0·5 M NaCl, 0·01 M EDTA), and then lysed by adding Sarkosyl NL-97 (Geigy) to 1 % (w/v). Immediately after lysis, the suspension was mixed with an equal volume of distilled phenol which had been liquified by shaking with DNA extraction buffer, and then emulsified by vigorous stirring at 4 °C for 10 min. The emulsion was broken by centri-fugation and the aqueous phase was extracted again with fresh phenol. Nucleic acids were precipitated from the final aqueous phase with 2 volumes of 95% ethanol and recovered by sedimentation (19000 g, 20 min, 4 °C). The precipitate was dissolved in a low-salt buffer (10 mM Tris-HCl pH 8·0, 1 mM EDTA) and RNA was digested by incubation at 37 °C for 30 min with 10 μg/ml pancreatic RNase (which had been heated at 80 °C 15 min) and 25 units/ml T₁ RNase. After ribonuclease digestion, self-digested pronase was added to 25 μg/ml together with SDS to 0·5 % and incubation was continued at 37 °C for 30 min. After pronase treatment, NaCl was added to 0·3 M and the solution was extracted twice more with phenol and precipitated with ethanol as described above. The precipitate was dissolved and then extensively dialysed in 0·12 M sodium phosphate, 1 mM EDTA, pH 6·8. Finally, some contaminating glycogen was removed by sedimentation as described for RNA above. The con centration was adjusted to 200 μg/ml in the same phosphate EDTA solution assuming an extinction coefficient = 220 and sonicated under a nitrogen atmosphere at 0°C at full power in a Raytheon 10⁴-Hz sonic oscillator. In our hands this sonication sheared DNA to a rather uniform size of approximately 300000 Daltons in double-stranded molecular weight ⁠. The ratios of absorbance A₂₆₀/A₂₈₀ =1·8–1·9 and A₂₆₀/A₂₈₀ =2·1–2·3

Samples of ³H-RNA in 0·3 ml or less were applied to 12 ml 15–30% (w/v) linear sucrose gradients prepared in 10 mM Tris-HCl pH 8·0, 0·1 M NaCl, 1 mM EDTA, and sedimented at 25000 rev/min for i8hat2O°CinaBeckman Spinco SW36 rotor. After centrifugation fractions of 0·5 ml were collected from the top by displacing the contents of the centrifuge tube with 50 % sucrose. Some samples were collected directly into vials for scintillation counting and mixed with 0·5 ml H₂O and 10 ml Aquasol (New England Nuclear). To measure the amount of acid-precipitable radioactivity in sucrose gradients, fractions were collected into test tubes and samples of the fractions were precipitated for 30 min at 0°C in 10% trichloroacetic acid (TCA) in the presence of 100 μg/ml of purified chick embryo DNA. The precipitates were collected on Whatman GF/C filters, washed 5 times with 5 ml 10% TCA at o °C and once with ethanol. Dried filters were treated with 0 3 ml NCS solubilizer (Amersham-Searle) for 30 min at 56 °C, and then the filters were counted in a toluene-based scintillation fluid. To measure the amount of acid-precipitable radioactivity which was also hydrolysable in alkali, portions of sucrose gradient fractions, which were measured for acid prccipitability, were hydrolysed for 1 h at 80 °C plus 15 min at 100 °C in 1 N NaOH, then chilled, neutralized with HC1, precipitated with TCA, and counted as above.

Hybridization was carried out in sealed polypropylene tubes which contained 0-06 M Na₂HPO₄, 0·06 M NaH₂PO₄, 1 mM EDTA, pH 6·8, plus 11–17 μg purified sonicated TetraJiymena DNA and 9–10 μg of the purified Tetrahymena³H-RNA to be tested. The tritiated RNA was extrac ted from either ribosomes or pellicles. To test for homology between tritiated RNA and ribosomal RNA, unlabelled ribosomal RNA was also included in amounts varying between o and 300 μg per tube. To denature the DNA, the tubes were first heated to 100 °C for 10 min, and then to anneal homologous RNA and DNA, the tube was incubated at 60 °C for 18 h. This annealing corresponded to C_ot = 2–3 mol l.⁻¹ s⁻¹ in terms of DNA phosphate concentration (Britten & Kohne, 1968). At the end of the 18-h incubation, the tubes were chilled at 0 °C and then unhybridized RNA was destroyed by digestion with pancreatic RNase (10 μg/ml) and T₁ RNase (25 units/ml) for 1 h at 37 °C. The remaining hybridized radioactive RNA was then precipitated with TCA and counted as described above. In the absence of unlabelled competitor ribosomal RNA, 5% of the 8000–13000 input cpm of rRNA were recovered as hybridized. Under similar conditions, 2–3% of the 8000–13000 input cpm of pellicle RNA were recovered as hybridized.

Over 80 % of the cells could be converted to whole pellicles in 10 min of homogenization. Longer homogenization converted the remainder, but broke pellicles also. The properties of the pellicles depended on the growth medium; cells grown in tryptone yeast extract were harder to homogenize, yet yielded more shear-resistant pellicles than cells grown in Frankel’s medium. Addition of 0·1% DEP to the homogenization medium appeared to toughen cells and pellicles, hence longer homogenization was required. By growing cells in Frankel’s medium and homogenizing 40 min with added DEP, we could convert over 50% of the cells to whole pellicles, and this was the procedure we chose for most of the work reported here. Fewer than 0·5 % whole cells remained in the pellicle preparation. Light micrographs of isolated pellicles are shown in Figs. 3 and 4. Extended homogenization without DEP often resulted in disappearance or irregular spacing of the usually regularly spaced dark basal bodies seen in phase-contrast microscopy. Cells synchronized by heat shocks (Materials and methods) and isolated without DEP just before division also gave stable pellicles.

Thin-section electron micrographs of isolated pellicles showed inner membranes, basal bodies, and little else (Figs. 5–8). By comparison with the figures of Nozawa & Thompson (1971) it was evident that our pellicles lacked mucocysts and some outer membranes. Basal bodies appeared to be intact except for the dense central core (Dippell, 1968) which was missing when pellicles were isolated without DEP (Fig. 7). This core was present in pellicles isolated with DEP (Fig. 8). No mitochondrial structures were detected in electron micrographs. Whole mitochondria from Tetrahymena are large (ca. 1 μm in diameter; see Nozawa & Thompson, 1971) and should have been readily apparent if present.

Fluorescence microscopy gave a regular distribution of yellow-green fluorescence in or near basal bodies (Figs. 9, 10). Irregular red fluorescence was occasionally seen also (Fig. 9) but not near basal bodies. Double exposure of phase-contrast and fluorescence photographs demonstrated that the dark basal bodies seen in phase-contrast were coincident with the yellow-green fluorescent spots. Fluorescence was especially intense in the mouth region where basal bodies were closely spaced. Pellicles isolated with DEP or from synchronized cells without DEP gave stronger fluorescence than pellicles isolated without DEP and from unsynchronized cells. Every pellicle which showed regularly spaced basal bodies in phase contrast, regardless of variations in cell culture or pellicle isolation, showed some yellow-green fluorescence similar to that shown in Fig. 9.

Pellicles homogenized over 30 min without DEP lost dark basal bodies seen in phase-contrast. These pellicles also lost the characteristic regular yellow-green fluorescence. Further, all pellicles isolated in the absence of DEP usually lost the ability to be stained for fluorescence during RNase or histone treatment. Again the loss of regular fluorescence was accompanied by the loss of a regular distribution of dark spots seen in phase-contrast, even though the fixed pellicle structure remained otherwise intact. The use of DEP during homogenization prevented loss of fluorescence and visible basal bodies in phase-contrast. The mechanism of action of DEP could be a further fixing of pellicle structure or inhibition of nuclease activity during pellicle isolation (Rosén & Fedorcsák, 1966). Because of these observations, we felt that it was important to discriminate between loss of fluorescence due to the loss of the whole basal body structure from the pellicle and loss of fluorescence due to loss of stainable molecules within the structure.

Treatment of DEP-isolated pellicles with RNase did not eliminate basal bodies seen in phase-contrast (Fig. 9) but did reduce fluorescence considerably (Fig. 12). Residual yellow-green fluorescence was often detected, but was subject to irreversible fading within 10 s of the exciting illumination, and was always too dim to photograph (Fig. 12). Smith-Sonneborn & Plaut (1967) also reported that RNase removed yellow-green fluorescence but chose to interpret the finding as a blocking of accessibility of DNA to the dye. In our hands RNase was not noticeably effective in reducing fluorescence after 1 h of incubation, but was reproducibly effective after incubation of 3 h or longer.

Treatment of DEP-isolated pellicles with DNase did not alter pellicle fluorescence when used at a concentration of 1 μg/ml for 15 h at 37 °C. The enzyme preparation was tested for activity at 5 μg/ml in DNase buffer on a solution of 1 /Jg/ml of tritiated DNA from SV40 virus (kindly provided by Mr R. Macdonald). The enzyme rendered 95 % of the radioactive DNA acid-soluble within 10 min. The activity of the DNase preparation used was also tested on whole Tetrahymena fixed on slides. Cells receiving no enzyme treatment showed bright yellow fluorescence from the nucleus and a paler yellow fluorescence from the pellicle. DNase-treated cells showed only pellicle fluorescence. Hence it seems that the yellow-green fluorescence of basal bodies was not susceptible to active DNase.

Treatment of DEP-isolated pellicles with either buffer alone or with histone or lysozyme also had no effect on pellicle fluorescence. Of the proteins tested, only RNase abolished fluorescence.

Both mercaptoethanol and pronase changed the colour of pellicle fluorescence from yellow-green to orange. In Fig. 13, orange basal bodies, in pellicles treated with mercaptoethanol, are seen in a mixed field of pellicles and whole cells made by partial homogenization. Nuclei fluoresced bright yellow while basal bodies fluoresced orange. The orange fluorescence faded within 1 min leaving a very pale yellow-green residual fluorescence at the basal bodies. The effect of pronase was harder to control than mercaptoethanol because pronase often removed even DEP-isolated pellicles from the slide. To create a gradient of pronase treatment, we applied the pronase solution to the slide in droplets of about 50 fi\ rather than covering the whole slide with solution. The enzyme usually removed pellicles at the centre of the droplet but we observed orange-fluorescing basal bodies at the edge of the droplet. Again the bright orange fluorescence faded quickly, leaving a dim yellow-green fluorescence at the basal bodies. We think it likely that the change in colour from yellow-green to orange was due to changes in protein structure near the basal bodies and not due to direct action of the reagents on nucleic acids.

To search for further evidence for RNA in Tetrahymena pellicles, we labelled growing cells for 48 h with either [³H]guanosine or [³H]uridine and then incubated cells further for 24 h in fresh non-radioactive medium. The purpose of this protocol was to confine labelling to stable RNA species. Under these conditions, approximately 1 % of the radioactivity incorporated into acid-precipitable material of the whole cell was incorporated into isolated pellicles. When cells were labelled with [³H]thymidine under similar conditions, approximately 0·1% of acid-precipitable radioactivity was recovered in pellicles.

The species of RNA receiving radioactive label were analysed by sucrose gradient sedimentation. Fig. 1 A shows this analysis applied to the supernatant fraction of a whole cell homogenate prepared during pellicle isolation, after removing pellicles by sedimentation. The 3 major species of stable RNA found were the 2 ribosomal species of 25 s and 17 s, present in the approximate mass ratios of 2 to 1 (Prescott, Bostock, Gamow & Lauth, 1971), and the 4 s transfer RNA. RNA prepared from purified pellicles (Fig. 1 B) revealed 2 similar larger species, now in approximately equal amounts, and very little smaller RNA. The presence of the 2 larger species indicated probable contamination of pellicles with ribosomes, although the 2 species were not as sharply resolved as in analysis of whole cell homogenates, nor were the ratios of the species the same as expected in whole ribosomes. Hence, we felt that there was a good possibility of the presence of stable RNA other than ribosomal RNA in pellicles.

The radioactive RNA in pellicles was analysed further by competitive DNA–RNA hybridization. We used conditions of approximately equal overall DNA and RNA concentrations in this work and annealed to a C_ot value of 2–3 (Britten & Kohne, 1968) based on DNA concentration. According to the data of Allen & Gibson (1972), the for Tetrahymena DNA was 200–300, with a small amount of redundant sequences. We estimate from these data that our hybridization could have involved no more than the most redundant 15 % of the Tetrahymena genome. If the labelled RNA were confined to a few stable species, however, we could have hybridized under conditions of excess RNA. The results of this analysis (Fig. 2) showed that purified labelled ribosomal RNA could be competed to about 5 % of the value of uncompeted hybridization by adding to the hybridization mixture a 30-fold excess of RNA in the form of non-radioactive rRNA. This value of 5 % was close to that expected (3 %), under theoretical conditions of RNA excess during hybridization. The values of 33 % at an excess ratio of 5, and of 17 % at a ratio of 10 deviated from the expected values (20% and 10%, respectively), indicating that the assumption of excess RNA was not valid at lower ratios. When labelled pellicle RNA was used with unlabelled ribosomal RNA as competitor, 60 % of radioactivity hybridized without competitor could be competed away by adding 30 times the amount of unlabelled ribosomal RNA, indicating that about 65 % (correcting for the uncompeted 5 % residue observed above at the same competitor RNA concentration) of the extracted pellicle RNA was ribosomal. The remaining 35 % of pellicle RNA was not ribosomal, however, nor was it transfer RNA, judging from the lack of 4 s RNA in Fig. 1B.

Our observations on the fluorescence of pellicles differ in 2 major ways from the reports of Randall & Disbrey (1965) and of Smith-Sonneborn & Plaut (1967). We think the disagreements can be explained by differences in methods of pellicle isolation and in treatments of the isolated pellicles.

First, we found that all basal bodies in a given preparation of pellicles made from cells in any phase of the growth cycle could be made to stain with acridine orange and fluoresce with a yellow-green colour. By contrast, in the previous work only a fraction of the population gave fluorescent basal bodies. Randall & Disbrey (1965) found it necessary to synchronize cells to obtain any appreciable fluorescence. We have found also that the method of pellicle isolation influenced the stability of stain able material, and that for unknown reasons synchronized cells gave more stable pellicles than log-phase cells. Because of this last observation we think that the mechanical stability of pellicles or of basal bodies within pellicles isolated by the older ethanol methods was not sufficient to show fluorescence except in stronger pellicles from cells which appeared at certain times in the growth cycle. The reagent diethyl-pyrocarbonate (DEP) could be used to strengthen pellicles from cells in late log phase of the growth cycle and to produce fluorescent basal bodies in all pellicles. Two known actions of DEP are covalent cross-linking of protein and inactivation of nucleases (Rosén & Fedorcsák, 1966). Both of these activities may have been important in the preservation of the otherwise labile pellicle structure for fluorescence microscopy. DEP was not required to produce fluorescence, however, since tryptone-yeast extract grown cells in late log phase gave fluorescent pellicles without the use of DEP. This particular point of disagreement with the older work is important because if fluorescence is accepted as evidence for the association of nucleic acid with basal bodies, then the nucleic acid was present continuously.

Another known activity of DEP is a reaction with RNA resulting in an opening of the ring structures of bases and mono- or dicarbethoxylation of adenylate, cytidylate, and guanylate residues (Leonard, McDonald & Reichmann, 1970; Oberg, 1971; Henderson, Kirkegaard & Leonard, 1973). Phosphodiester bonds were not broken although the infectivity of viral RNA was lost (Oberg, 1970; Oxenfelt & Årnstrand, 1970). The modifications of RNA have also resulted in increased resistance to nucleases (Solymosy et al. 1971; Henderson et al. 1973). We think that reaction of RNA with DEP should not have reduced binding of acridine orange, however, since the phosphodiester bonds were not broken. It should be noted that we used a lower concentration of DEP, 6 mM, than the saturating concentrations, 40–60 raM, used in the work cited above.

The second point of disagreement with the older work concerns the results of enzyme treatment between fixing the pellicles and staining them. We found that only RNase reduced fluorescence; Randall & Disbrey (1965) found that only DNase abolished fluorescence; and Smith-Sonneborn & Plaut (1967) found that RNase, DNase, histone and protamine abolished or reduced fluorescence. We are unable to explain the lack of effect of RNase noted by Randall & Disbrey (1965) except to suggest that their treatment (10 mg/ml of RNase in water, 1 h at 37 °C) may have been insufficient despite the high concentration used. The effects of proteins other than RNase might be explained by residual binding of the proteins to basal bodies thus blocking acridine orange staining, as suggested by Smith-Sonneborn & Plaut (1967). We have been able to remove histone and lysozyme from treated pellicles by washing the slides with 0·1N acetic acid and afterwards observed bright fluorescence. The pre-caution of acid washing was not reported in previous work.

The yellow-green colour of the fluorescence observed has been accepted in earlier work as evidence for the presence of double-stranded nucleic acid in basal bodies. The colour of the fluorescence, however, does not always depend directly on the strandedness of nucleic acid; rather, colour probably depends generally on the rigidity of the dye-binding structure (Kasten, 1967) which can be determined by other molecular interactions as well. Single-stranded nucleic acids in solution are generally flexible and allow the basic acridine orange molecules bound to nucleotide phosphate groups to stack upon each other. By contrast, the rigid double-stranded nucleic acids do not allow bound dye molecules to interact with each other. It is generally accepted (Kasten, 1967, p. 148) that the interaction between dye molecules determines the colour of fluorescence. Stacked acridine orange gives the ‘metachromatic’ red and unstacked acridine orange gives the ‘orthochromatic’ yellow-green. A relevant case is the observation of Mayor & Diwan (1961) who found that tobacco mosaic virus, which contains single-stranded RNA, fluoresced green when stained with acridine orange in the unfixed state and red when stained after methanol treatment. We think that the fluorescence of our fixed pellicles was similar to that observed in unfixed TMV. Our yellow-green fluorescence could be changed to orange, however, by treating fixed pellicles with mercaptoethanol before staining. We think that our mercaptoethanol treatment did not denature double-stranded nucleic acid because the fluorescent colour of nuclei in the same preparation remained yellow. Smith-Sonneborn & Plaut (1967) also observed the same change in fluorescent colour of basal bodies after treating pellicles with acid and formaldehyde, and chose to interpret the change as due to denaturation of double-stranded DNA. We feel that an alternative interpretation is more consistent with all results available to date, namely, that the yellow-green fluorescence in or near basal bodies resulted from single-stranded RNA held in rigid configuration by protein.

Our cytological evidence for RNA therefore consists of 3 findings. (1) RNase reduced pellicle fluorescence; (2) DNase active enough to destroy nuclear fluorescence did not change pellicle fluorescence; (3) Mercaptoethanol changed pellicle fluorescence from yellow-green to red while leaving nuclear fluorescence yellow.

Another possible interpretation of these results, which we think to be less plausible, is that basal bodies were associated with acidic macromolecules other than nucleic acid. Acid mucopolysaccharide, phospholipid, or phosphoprotein are possible candidates, but it is hard to explain why RNase should have reduced fluorescence if molecules other than RNA were binding the dye.

In obtaining cytological data we have relied on the many clear fields obtainable with the new procedure of pellicle isolation. Biochemical data by contrast, have been more difficult to obtain because we needed nearly perfectly clean pellicles to give meaningful evidence of one or more RNA species which could be associated with basal bodies. In the work reported here, 1 % of stable RNA was confined to the pellicle fraction, as deduced from our radioisotope labelling experiments. The pellicle fraction contained 0·1–0·5 % unbroken cells, however, and further, most pellicle preparations contained a small amount of irregular red-fluorescing material, probably containing RNA, as shown in Fig. 9. With our present development of pellicle isolation, therefore, it was not surprising that bulk RNA isolated from a pellicle preparation was contaminated with ribosomal RNA of the cell. Contamination with rRNA was suggested by the sucrose gradient data in Fig. 1B and was confirmed by hybridization data in Fig. 2. The hybridization data did reveal, however, that approximately 35 % of the pellicle RNA which could hybridize to DNA under our conditions did not compete with ribosomal RNA. Under the same conditions, hybridization competition of ³H-rRNA by unlabelled rRNA was 95 %, thus demonstrating that the hybridization conditions could have shown homology to rRNA. We believe that the remaining uncompeted 35% could not have been messenger RNA or its precursors because the conditions of incubation with radioisotopes would have labelled only more stable species and also because the conditions of hybridization would have hybridized only RNA complementary to more redundant DNA. Further, from the sucrose gradient data of Fig. 1 B, it seems impossible that a fraction as large as 35 % could have been confined to 4 S transfer RNA. Finally, since we have observed no mitochondria in the electron micrographs, we think it is unlikely that a significant amount of mitochondrial RNA could have been contained in the pellicle preparation.

Our biochemical evidence for RNA associated with basal bodies is therefore the finding of stable non-ribosomal RNA in the pellicle fraction, and the elimination of the other known stable RNA’s by other considerations. Further work is needed to purify basal body RNA and thus to confirm its existence directly, but this preliminary study should serve to guide the purification. Basal body RNA should be complementary to a DNA present in at least 50 copies in the Tetrahymena genome, and should have a larger sedimentation coefficient than 4 Svedbergs.

The case for the association of RNA with centrioles and basal bodies is strengthened by 2 sets of observations made with the electron microscope. Stubblefield & Brinkley (1967) found that a dark staining ‘foot’ attached to the A microtubule of each triplet in Chinese hamster cell centrioles disappeared after RNase treatment. Dippell (1968) noted that the dense core material in Paramecium basal bodies was sensitive to RNase. In our work, it may be significant that pellicles prepared with DEP contained dense cores and fluoresced more brightly than pellicles prepared without DEP and which lack cores.

Two recent observations indicate that RNA synthesis may be required for basal body or centriole formation. Younger et al. (1972) found that the regeneration of the 15000–20000 basal bodies in the oral membranellar bands of Stentor was blocked by actinomycin D, which prevented DNA-dependent RNA synthesis, but not by several inhibitors of DNA synthesis. Similarly Stubblefield & DeFoor (1972) found that actinomycin D prevented procentriole formation in synchronized hamster cells. Inhibitors of DNA synthesis did not prevent procentriole formation.

The finding of RNA in basal bodies and the apparent requirement for DNA-dependent RNA synthesis in their formation suggest an outline of a mechanism of formation. We speculate that basal bodies are nucleated by specific RNA molecules which are synthesized in the nucleus at certain stages of the cell cycle, then migrate to the cytoplasm and bind at well defined positions on or near mature basal bodies. The hypothetical’ morphic’ RNA has specific binding properties which result in recognition of the mature structure and a structural template character for nucleation of the new one. Also, release of this RNA in the absence of mature structures may trigger de novo information of the basal bodies in Naegleria (Dingle & Fulton, 1966). Finally, the example of RNA in the composition and development of one complex cellular structure, the centriole, illustrates a possibly important principle of controlled nucleation in the organization of other cellular structures such as microtubule organizing centres (Pickett-Heaps, 1971).

We are grateful to Dr Daniel Branton for the use of a fluorescence microscope, and ro Mr Paul Moss and Dr Thelmn Dunnebacke Dixon for electron microscopy. This work was supported by Public Health Service Research Grant CA 12407 from the National Cancer Institute. J. P. Puma was a predoctoral trainee of the Public Health Service Grant GM 01389 from the National Institute of General Medical Sciences.