The conjusome: A novel structure in Tetrahymena found only during sexual reorganization

A diagnostic feature of the Ciliophora is the possession of two types of differentiated nuclei in the same cell: the germline micronucleus, and the somatic macronucleus. The polycopy macronucleus is transcriptionally active, and determines most of the phenotype of the cell, while the diploid micronucleus is transcriptionally silent and contributes little to cellular phenotype (Ng, 1986; Prescott, 1994). This distinction between micro- and macronuclei, sexual versus vegetative function, is analogous to the germ line/soma line differentiation found in metazoans. During conjugation, cells of complementary mating types pair and initiate the sexual process of nuclear reorganization (Bruns and Brussard, 1974). After micronuclear meiosis and subsequent reciprocal exchange and fusion of haploid gamete nuclei, mitotic progeny of the zygotic nucleus differentiate and give rise to the new macronuclei and micronuclei (Nanney, 1953; Orias, 1986). After the second post-zygotic mitosis in Tetrahymena thermophila, two undifferentiated nuclei migrate to the anterior pole and two to the posterior pole of the cell. The two nuclei that migrate anteriorly become determined to develop into macronuclei, while the two that migrate posteriorly will become micronuclei (Nanney, 1953). During this interval when new macronuclei (macronuclear anlagen) and new micronuclei are developing, the parental macronucleus begins its programmed nuclear death by a highly regulated process (Nanney, 1953). Thus, late in conjugation, three different nuclear fates occur in parallel within a common cytoplasm: apoptosis of the parental macronucleus (Davis et al., 1992; Mpoke and Wolfe, 1996; Madireddi et al., 1996), micronuclear differentiation, and macronuclear differentiation.

Large-scale rearrangements and changes of the genome take place within the macronuclear anlagen of ciliates (Coyne et al., 1996; Klobutcher and Jahn, 1991; Madireddi et al., 1995; Prescott, 1994). These events include transcriptional activation (Wenkert and Allis, 1984), DNA excision (Yao et al., 1984; Yao and Gorovsky, 1974), chromatin remodeling (Allis and Wiggins, 1984; Stargell et al., 1993), nucleolar biogenesis and gene amplification (Orias, 1986; Prescott, 1994; Yao, 1986; Yao and Gall, 1979), and chromosome breakage (Yao et al., 1990) followed by telomere addition (Greider and Blackburn, 1989). In Tetrahymena anlagen, specific micronuclear-limited DNA segments are removed (∼15% of the genome; Yao and Gorovsky, 1974). Pdd1p, a conjugation-specific polypeptide, localizes to electron-dense heterochromatic-like structures at the periphery of developing anlagen (Smothers et al., 1997a) and co-localizes with micronuclear-specific DNA sequence deletion elements in the anlagen (Madireddi et al., 1996). These heterochromatic structures, resembling mature nucleoli, have now been shown to be unique organelles in developing anlagen (Smothers et al., 1997a). The abundant nature of Pdd1p suggests that it plays a structural role, forming a complex with deletion-specific DNA-binding factors, thereby promoting a change in chromatin structure that brings appropriate boundaries of eliminated sequences into the correct conformation for excision (Coyne et al., 1996; Madireddi et al., 1994) and ligation. Recently, biochemical analyses have demonstrated a physical link between Pdd1p and germ-line restricted chromatin (Smothers et al., 1997b). At certain times in conjugation, transient levels of Pdd1p have also been found in the parental macronucleus (Madireddi et al., 1994). Programmed DNA degradation in the parental and in developing macronuclei occurs at approximately the same phase of conjugation (Austerberry et al., 1984; Davis et al., 1992; Mpoke and Wolfe, 1996; Yokoyama and Yao, 1984). No direct functional link has been found between Pdd1p and DNA degradation in the parental macronucleus. However, Pdd1p is present when the parental macronucleus begins to degenerate (Madireddi et al., 1994).

Observations, using phase contrast or Nomarski optics, of living, immobilized pairs of conjugating T. thermophila at specific phases of conjugation, revealed a hitherto unreported structure in the anterior of each conjugant partner. This structure did not label with DAPI or other DNA-specific stains. Staining with antibodies directed against Pdd1p revealed strong labeling. This paper is a description of this new structure, which we have named the conjusome.

Genetically marked strains of Tetrahymena thermophila Nanny & McCoy, 1976, CU 427 (Chx/Chx[cy-s]VI) and CU 438 (Pmr/Pmr[pm- s]IV) were grown independently to densities of ∼250,000-500,000 cells/ml in proteose peptone and yeast extract medium (PPY) as described by Ng and Frankel (1977), and starved in Dryl’s solution (Dryl, 1959) for 18-24 hours prior to mixing. The cell strains were provided by Peter Bruns (Cornell University, Ithaca, NY). In order to induce cells to be sexually reactive, equal numbers of cells of complementary mating type were mixed following starvation, according to established procedures (Martindale et al., 1982). Matings with greater than 80-90% pairing and kinetics similar to those previously described (Martindale et al., 1982) were observed in all experiments reported. All time points, reported in hours, refer to the time from mixing of complementary mating types.

Living, conjugating cells were picked individually at appropriate times by micropippeting and were placed into a rotocompressor (Aufderheide, 1986). Immobilized, rotocompressed conjugants were examined using high resolution Nomarski DIC or Zernike phase contrast optics on an Olympus BH2 microscope. No deleterious effects were observed when cells were placed back into growth medium and cultured.

Living cells were prepared for in vivo fluorescence by incubating conjugating cells in 0.3 mg/ml DAPI and 0.1 mg/ml Acridine Orange simultaneously for one hour (similar to the method of Mpoke and Wolfe, 1997). Cells were then resuspended and washed 3× in Dryl’s before observation on a UV fluorescent microscope. Observations were also made using only Acridine Orange (0.3 mg/ml) in living cells. All observations were made while cells were under rotocompression. Kodak Royal Gold 400 was used to record color images.

All cells were prepared for immunostaining as previously described (Wenkert and Allis, 1984), with the following modifications: after fixation and resuspension in methanol, cells were washed twice in PBS and then stored overnight in PBS with 2% BSA. Cells were incubated for one hour the following day at room temperature with immune or preimmune serum (appropriately diluted in PBS according to the titer of the immune serum with 2% BSA). The cells were then washed twice in PBS followed by a one hour incubation in Cy3-conjugated goat anti- rabbit serum (Sigma). For DNA staining, 0.1 mg/ml DAPI in 70% EtOH was added to the cell suspension after labeling and then washed out with PBS before mounting. This concentration of DAPI allowed excellent detection of all nuclei with a mercury lamp without interfering with the red fluorescence imparted by the Cy3- immunolabeled structures. The specificity of the primary antibody to Pdd1p has been demonstrated in previous reports (Madireddi et al., 1994). Background basal body staining is a phenomenon common to rabbit polyclonal sera (Turksen et al., 1982). Labeled cells were resuspended in 150 μl of PBS, mounted on slides and examined on an Olympus BH2 microscope using an epifluorescence illumination system mounted with standard fluorescent clusters. To record black and white images, Kodak Technical Pan Film was exposed using an Olympus C-35AD-2 camera.

Cytochemical localization of acid phosphatase was done according to Anderson’s adaptation of the Gomori technique (Anderson, 1982). Observations were made using brightfield optics at appropriate time points in conjugation.

Appropriately staged mating cells were fixed in 0.5% glutaraldehyde, 2.0% paraformaldehyde in 0.1 M cacodylate buffer, dehydrated and embedded in LR white and polymerized at 65°C for 24 hours. Ultrathin sections were mounted on colloidon coated copper grids and were stained for 5 minutes in aqueous uranyl acetate and 30 seconds in Reynolds’ lead citrate or subjected to immunodetection.

For immunodetection, grids were incubated for 1 hour in blocking solution (2% BSA and 0.1% sodium azide in PBS). Grids were then incubated for 1 hour in Pdd1p primary antiserum diluted 1:100 in PBS solution followed by 5× 5 minute washes in PBS. Samples were then incubated for 40 minutes with secondary antibody (goat anti-rabbit conjugated to 10 nm gold beads; Amersham). Secondary labeling was followed by 5× 5 minute washes in PBS and then two washes in dH₂O and then stained with uranyl acetate and Reynold’s lead citrate as described above. All examinations were done on a Zeiss 10C transmission electron microscope operating at 60 kV.

With the use of the rotocompressor (Aufderheide, 1986), the cells were held motionless, thus offering the observer a detailed view of living cells. Observations of conjugants showed that the time of appearance of the conjusome corresponded closely with the condensation of the parental macronucleus. The parental macronucleus became ‘grainy’ in appearance at 5.5 to 6.5 hours postmixing of complementary mating types. This conspicuous event correlated with the development of macronuclear anlagen and the appearance of the conjusome (Fig. 1A). Although the mass matings were fairly synchronous, the precise timing of mating events can differ some among conjugating pairs. Visual markers, such as the appearance or state of mitotic spindles, were necessary to identify the exact stage of conjugation in a given pair. We were thus able to correlate the conjusome with specific nuclear events during conjugation.

The appearance of the conjusome at approximately 6 hours corresponded very closely to the time when anlagen first appeared following the last post-zygotic mitosis (arrows, Fig. 2B). The conjusome was usually in close proximity, and anterior, to the two anlagen at this stage in development (Figs 1A, 2A,B). The parental macronucleus also started to condense at about this time. The series of Nomarski images in Fig. 1 show the conjusome within two different living, conjugating pairs at early and late stages of parental macronucleus condensation. The conjusome was present throughout the time that the parental macronucleus condensed (condensation takes approximately 0.5-1.5 hour). The parental macronuclei linger in their condensed state for a number of hours before undergoing pycnosis and complete degeneration and resorption. The early conjusome was small, less than 1.0 μm in diameter (Fig. 2A). It subsequently (see Fig. 2B) was capable of reaching dimensions rivaling those of early, developing anlagen (5.0-7.0 μm in diameter). The conjusome is normally spherical in form, although it occasionally has assumed a somewhat crescent and/or elongated shape. Once the old macronucleus condensed, the conjusome began to shrink and eventually disappeared.

Fig. 3 shows a time series of anti-Pdd1p staining in the three structures that display this protein at selected times during sexual reorganization: the parental macronucleus, the conjusome, and the developing macronuclei. The parental macronucleus showed antibody labeling for Pdd1p at roughly 4.0 hours post mixing. This labeling persisted for about 1.0-3.0 hours, and ceased at approximately the same time the conjusome appeared. The conjusome appeared and also stained positive for Pdd1p at approximately 6 hours, which is 0.5-1.0 hour before the developing anlagen began to stain with the Pdd1p antibody. During the time that the conjusome was present, it labeled very intensely, often much more so than either the parental or developing macronuclei for Pdd1p (see Fig. 2). At the light microscopic level, Pdd1p was distributed evenly within the structure. Developing anlagen began to stain positive for Pdd1p at about 7.0 hours into conjugation. The parental macronuclei subsequently showed the presence of Pdd1p for a second time, at 10.0-12.0 hours. Pdd1p staining of the parental macronucleus, which is often asynchronous in the two cells of a conjugating pair, lasted approximately 0.5-1.0 hour. Conjugation ends with the separation of cells at about 12 hours. The anlagen continue to develop and remain Pdd1p positive until about 15 hours from the time of mixing (Madireddi et al., 1994). Occasionally, small structures other than the conjusome stained for Pdd1p in fixed specimens. These structures were often seen near the parental macronucleus or conjusome. It is important to note that occasionally, the DNA stain DAPI, revealed that these smaller structures in fact contained DNA. This is in distinct contrast to the conjusome, which was always DAPI negative and contained no detectable DNA.

Acridine orange staining for DNA, RNA and lysosomal activity in the conjusome was negative in living cells (Fig. 4B). At the time the conjusome was visible, the structure also did not stain for acid phosphatase by the Gomori technique, a classic indicator of digestive vacuoles and lysosomes. As expected, there were other acid phosphatase positive bodies in these cells (data not shown). Furthermore, immunocytochemistry using antibodies directed against ubiquitin showed no detectable ubiquitin in the conjusome (data not shown). The conjusome thus did not show characteristics typical of a lysosome or of the ubiquitin/proteasome pathway.

We examined conjugant cells using transmission electron microscopy (TEM). TEM revealed that the conjusomes were not membrane bounded (Fig. 5). The conjusome was composed of a coarse reticulum (see arrows in Fig. 6B). This reticulation forms a contiguous three dimensional meshwork that appears to be the framework of the conjusome. Intermingled within this reticulation were ribosomes. The concentration of ribosomes within the structure did not appear different from that of the cytoplasm. However, electron-lucent bodies found throughout the cytoplasm were excluded from the conjusome. The presence of Pdd1p was substantiated by immungold labeling using the Pdd1p antibody. Interestingly, the labeling (arrowheads) colocalized almost exclusively with the reticulum within the conjusome (Fig. 6B).

To characterize the possible function(s) the conjusome has in the cell during conjugation, antibody staining was examined in a mutant line that does not develop new macronuclei. Nuclei in the cnj9 mutant fail to complete the second post-zygotic division (Cole and Soelter, 1997). This inability to undergo the last mitosis left each cell with two nuclei, which both appeared to remain micronucleus-like. They are somewhat larger than normal micronuclei, but smaller than developing anlagen. In conjugating pairs of cnj9 cells that show this characteristic phenotype, Pdd1p labeling was seen only in a very large conjusome at 6.0-11.0 hours postmixing (Fig. 7A). As in wild- type, the conjusome in this mutant lacked DNA (Fig. 7B). The conjusome in cnj9 cells persisted longer than in wild-type cells, although the duration of the persistance varied from pair to pair. It is important to note that the conjusomes persisted longer in this mutant and conjugation events appear to be delayed to some degree in cnj9 cells (Cole and Soelter, 1997). In the cnj9 mutant, the parental macronucleus stained positive for Pdd1p early on, as in the wild-type. The timing of events in labeled cnj9 cells and the conjusome was also substantiated by observations of living, rotocompressed, conjugants of this mutant type.

Documentation of conjugation and sexual reorganization in Tetrahymena thermophila has been done predominantly by examination of fixed specimens (Nanney, 1953; Ray, 1956; Martindale et al., 1982; Numata et al., 1985; Orias, 1986; Gaertig and Fleury, 1992; Madireddi et al., 1994; Nelson et al., 1994). With the use of the rotocompressor (Aufderheide, 1986), we have been able to immobilize and observe sexual reorganization in living cells. Rotocompression allows the experimenter to have an unprecedented view into the cell because it holds the specimen still and also gently flattens it. Observations of living, conjugant pairs with Nomarski and phase-contrast optics have allowed us to discover a number of important phenomena that occur during conjugation, most notably the appearance of a previously undocumented structure, the conjusome.

We have shown here that a transient structure appears during sexual reorganization in Tetrahymena. The conjusome appears in the anterior of conjugating cells during the time that early macronuclear anlagen development takes place. With the microscopic techniques we have used, the conjusome is visible for a period of 2-4 hours during conjugation. The ontogeny of the conjusome is not clear at this time. The conjusome ranges in diameter from approximately 1 μm to 7 μm. The conjusome is not membrane bounded and consists of a reticulum that forms its structural basis. This reticulum stains positive for the chromodomain protein Pdd1p. We hope to identify other proteins within the conjusome to see how they are temporally regulated.

We have a number of hypotheses suggesting what role the conjusome is playing during conjugation. The first is that the conjusome is acting as a ‘dump’ for Pdd1p and possibly other components from the parental macronucleus. In this fashion it may be degrading the components it collects. The second hypothesis is that the cells recycle Pdd1p, and possibly other components, from the parental macronucleus into the developing macronuclear anlagen. In this regard, the conjusome acts as a reservoir for components from the old macronucleus and stores or processes them for the newly developing macronuclei. Finally, the conjusome may be acting as a production or distribution center for Pdd1p. It is possible that combinations of any of the above mentioned three hypotheses may operate.

Thus far, we have been unable to document enzymatic activity in the conjusome that would indicate that the conjusome is acting as a site for degradation of any components from the parental macronucleus. Moreover, a site of degradation uncompartmentalized by a membrane is not consistent with usual eukaryotic lysosomal activity. Furthermore, degradation taking place by the usual ubiquitin/proteasome pathway seems unlikely since we do not find detectable amounts of ubiquitin in the conjusome. We do, however, have evidence that small Pdd1p-positive structures exist during the time the conjusome is present. Some small structures also contain DNA. It might be possible these smaller structures originated from the condensing parental macronucleus. These Pdd1p-positive structures could be precursors to the conjusome or it may be that these structures are part of some unknown transport mechanism that shuttles Pdd1p and possibly other components into and/or out of the conjusome.

If the Pdd1p protein found in the conjusome originates in the parental macronucleus, then it would seem plausible that Tetrahymena are using the conjusome as a reservoir to store or process this protein and possibly other components. The temporal coordination of the presence and absence of Pdd1p in the conjusome and the developing and parental macronuclei fits with this hypothesis. The close proximity of the conjusome to the developing macronuclei and the disappearance of the conjusome as Pdd1p staining in the developing macronuclei increases also is consistent with this view. In addition, the data obtained from the mutant cnj9 lend credence to the idea that the components of the conjusome are destined to be transferred to the developing anlagen. When anlagen are not present, the conjusome has nowhere to deliver the Pdd1p, and thus the amount of Pdd1p remains high and the conjusome appears large. These observations suggest that the conjusome serves as a reservoir, possibly storing the Pdd1p and other components to support development in the anlagen. The cnj9 data also do not lend support to the degradation hypothesis.

The presence of ribosomes within the conjusome could mean that Pdd1p is translated within the structure. The high concentration of Pdd1p in the structure and the significant translation rate at this period in development (Madireddi, 1994) makes it reasonable to believe that its production may take place within the conjusome as well. The translation of Pdd1p in the conjusome is very intriguing when considering that the fate of macronuclear anlagen is dependent on their localization in the anterior of the cell. It is possible, however, that little or no translation take place in the conjusome. The localization of Pdd1p may merely be the result of some unknown mechanism which shuttles cytoplasmically translated Pdd1p to the conjusome for further processing or distribution.

If Pdd1p originates from the parental macronucleus and is stored or processed in the conjusome for subsequent use in the developing anlagen, the conjusome and its associated structures might be involved in some aspects of epigenetic phenomena that have been observed in ciliates (Meyer and Duharcourt, 1996; Chalker and Yao, 1996). It has clearly been shown that the parental macronucleus communicates with, and can have a maternal effect on, the developing macronuclei (Preer, 1997). The mechanism of this type of communication is unknown at this time, but experiments (Sonneborn and Schneller, 1979; Epstein and Forney, 1984) show that some somatic mutations are transmitted in a maternal fashion and not inherited in a typical Mendelian style.

The conjusome is spatially associated with the developing anlagen in the anterior of the cell during the time of nuclear differentiation. It has been shown that the developmental fates of nuclei are related to their location within the cell (Kaczanowski et al., 1991; Gaertig and Fleury, 1992) at critical times. The presence of the conjusome in the cell anterior at the time of macronuclear determination may be merely coincidental. The conjusome might be acting independently of these events and is merely coordinate with nuclear events in time and not causally linked to their development. This association warrants further investigation. Interestingly, the conjusome has some structural resemblance to the granulofibrillar material (GFM) found in the mitochondrial cloud in Xenopus laevis (Heasman et al., 1984) and the germline (P) granules of Caenorhabditis elegans (Wolf et al., 1983). Similar polar granules have been found in Drosophila (Mahowald, 1971; Illmense and Mahowald, 1974). It has been suggested that there is a relationship between the GFM and these germinal granules (Heasman et al., 1984). The significance of this structural similarity to the function of the conjusome is not known.

The microscopy data we have presented demonstrate that the conjusome has a unique structural appearance. Immunofluoresence data suggest that its protein constituents are uniquely expressed during conjugation and specifically localized to the conjusome. The appearance of the conjusome is time-limited, and highly coordinated to nuclear events during conjugation. All of these data suggest to us a specific function for this structure. For these reasons, we believe the conjusome is worthy of recognition as an organelle in Tetrahymena. We would also be interested to learn whether homologous structures are found in other ciliophorans. We are investigating further the structures and functions of the conjusome in development during the events of sexual reorganization.

We are grateful to the Peter Bruns lab (Cornell University, Ithaca NY) for the strains of Tetrahymena used in this study. Part of this study was supported by funds from the Texas A&M University Interdisciplinary Intitiatives Grant IR194-05.