Contact-Mediated or Soluble Signal Molecule?

After starvation, cells of complementary mating types can be mixed thereby provoking the mating reaction. This involves a series of developmental events that remodel the anterior cell membrane and its associated cytoskeleton, and results in the formation of a remarkable and dynamic system of cell-cell junctions. This period of obligatory cell-cell interaction lasts between one and two hours.^15,21,23 There has been some controversy as to whether or not costimulation involves soluble signal molecules, or strictly physical cell-cell contact. In reviewing the literature, both appear to play a legitimate role, though cell-cell contact seems the principle trigger.

The first suggestion that there might be a soluble factor associated with mating in Tetrahymena came from Takahashi.²⁴ More extensive evidence came from the Wolfe laboratory.^25,26 These early studies demonstrated that both costimulation and pair formation were profoundly delayed when cells were washed into fresh starvation medium. This delay was eliminated if, after washing, cells were returned to a conditioned starvation medium. The factor that was present in conditioned medium was a macromolecule (retained by dialysis) that was heat-stable (30 minutes in boiling water) and secreted constitutively by cells in either nutrient or starvation medium from sexually mature or immature cells. It had no mating-type specificity, and its synthesis was blocked by cycloheximide treatment suggesting that it was an unusually heat-stable, secreted protein. Wolfe named this compound FAC (factor active in conjugation).²⁷ He went on to demonstrate that FAC acts like neither a hormone (as seen in yeast mating systems and the ciliate Blepharisma) nor an agglutinin (as seen in Chlamydomonas and Paramecium). It does not trigger conspicuous changes in the recipient cell, nor does it cause immediate aggregation of cells. Rather, it seems to somehow facilitate those cell-cell contacts that bring about both costimulation and subsequent cell adhesion. Adair et al²⁵ suggested we think of this factor as a type of extra-cellular matrix molecule, a provocative idea in that it would suggest a role for ECM proteins in cells that do not form tissues, yet need to stabilize cell-cell interactions for this one event in their life histories. Consistent with this notion is the demonstration that both initiation and costimulation require trace amounts of calcium ion.²²

Recent studies lend further support to the idea of a soluble mating factor, and raise new questions. Fujishima et al²⁸ characterized a very early morphological response in cells undergoing costimulation. Starved Tetrahymena of complementary mating types exhibit a subtle, yet reproducible and consistent “rounding” of their shape within ten minutes of mixing. Stirring a mixed culture (which can prevent costimulation and subsequent pairing) also prevented cell rounding. When starved cells were washed and mixed into fresh starvation medium, there was a 30-minute delay in both rounding and pairing that was eliminated when cells were washed into conditioned medium. Medium from cells exhibiting a complementary mating type could not bring about rounding in starved cells until both mating types were present. In short, rounding was induced when starved cells of complementary mating types were able to undergo collisions in the presence of some secreted factor (possibly Wolfe's FAC).

Finally, in a very recent study Anafi et al (personal communication) demonstrate that ten minutes after one mixes starved cells of complementary mating types, there is a wave of tyrosine phosphorylation on cytoplasmic proteins as revealed by immuno-fluorescence and western blotting using a monoclonal antibody directed against phospho-tyrosine residues. Furthermore, starved cells of one mating type show elevated tyrosine phosphorylation in response to cell-free, conditioned medium from another mating type. Significantly, conditioned media only stimulates phosphorylation in cells of a different mating type from the cells that were used to condition the medium. This result, if valid, may force us to reopen the question of whether or not there are mating-type specific signal molecules secreted by Tetrahymena in addition to soluble factors that promote costimulation and pair formation.

Activation and Maturation

Costimulation can be divided into two stages: activation (first half hour following mixing) and maturation (30 to 120 minutes following mixing,²⁹ see Fig. 3, Fig. 4). During the initial “activation stage,” costimulation is sensitive to cycloheximide (10 μg/mL), actinomycin D (20 μg/mL), physical agitation and nutritional enhancement of the medium. After this initial period of activation, loose pairs begin to form that can still be disrupted by either feeding, or mechanical agitation. Curiously, there is a brief period in which loose, “homotypic” pairing can occur.³⁰ That is, cells of the same mating type form brief associations. These are intrinsically unstable, however, and give way to heterotypic pairing that is reinforced during the subsequent maturation step.

By one hour, mixed cells become “fully stimulated”.¹⁰ This condition is revealed when cells are maintained in physical contact with one another for a full hour while weak-pair associations are disrupted through periodic bouts of mechanical agitation. Such cells will form pairs immediately following cessation of the last bout of agitation, but only if they have accumulated one hour of nonshaken contact with one another.

Ninety minutes after mixing, pairs that are violently disrupted (as opposed to gentle disruption) produce cell isolates that can complete the full 12-hour developmental program involving meiosis, mitosis, self-fertilization and differentiation of the macronuclear anlagen.^31,32 This strange scenario suggests that after only one and a half hours of interaction, the conjugation program can be initiated and maintained even in the absence of a mating partner. Two hours after mixing, (1.5 hrs after initial cell pairs are observed), pairs have fully matured. They are now resistant to feeding, cycloheximide, and agitation. This period may be associated with the formation of actual pores connecting the cytoplasm of the mating partners²² (see below).

Membrane Events

In response to costimulation, Tetrahymena remodel their anterior tips in a process called tip transformation.³³ Specifically, physical collisions induce a shape change in the anterior region of the cell that begins approximately 30 minutes after mixing. The anterior pole changes from a pointed cap that forms from the convergence of numerous longitudinal ciliary rows and intervening ciliary ridges (Fig. 5A), to a region of smooth, naked membrane (Fig. 5D, asterisk). In costimulated cells, the intervening ciliary ridges converge in a seam, or chevron just anterior to the Oral Apparatus. This seam appears to be the site of novel membrane synthesis. In nonmating cells, the endo-membrane system is dispersed predominantly in the form of dictyosomes, or simplified membrane lamellae associated with ciliary basal bodies and the “deep face” of cortical mitochondria.³⁴ The only time a well-organized, multi-lamellate golgi apparatus has been reported in Tetrahymena is during mating when it forms adjacent to the developing conjugation junction (Fig. 6).³⁵ As tip transformation proceeds, the ridge-seam broadens, producing a blunt, naked, protruding region of membrane at the anterior tip that extends down the ventral face just anterior to the Oral Apparatus (Fig. 5B-D). When a sufficient number of costimulated cells complete tip transformation, loose adhesions form between mating partners (Fig. 7A). As mentioned earlier, transient adhesions can and do form between homotypic pairs (same mating types), but these fail to stabilize and are replaced by heterotypic pairings.³⁰ As the region of adhesion broadens it forms the “fusion plate”⁹ (Fig. 7BB). This becomes perforated with hundreds of 0.1-0.2 μm pores that join the cytoplasm of the two mating partners (Fig. 8). Rows of intra-membrane particles form at the perimeter of the fusion plate, as revealed by freeze fracture electron microscopy.⁹

Figure 5

Scanning electron micrographs showing “tip transformation” (reprinted with permission of the publishers from Wolfe and Grimes). OA indicates the oral apparatus at the anterior end of the cell. A) This shows the anterior end of a de-ciliated (more...)

Figure 6

Transmission electron micrographs showing a fully formed conjugation junction. A) This image was captured just prior to nuclear exchange (courtesy of Judy Orias). The transfer pronucleus is pressed up against the conjugation junction that has already (more...)

Figure 7

Scanning electron micrographs of mating pairs (reprinted with permission from the publishers from Wolfe and Grimes, and Wolfe et al). A) This figure shows loosely paired cells. B) Mating pairs are shown with a mature conjugation junction.

Figure 8

This figure depicts both (A) a freeze fracture image of the conjugation junction and (B) a transmission EM image of a microtome section that grazes the conjugation junction. A) This freeze fracture image shows the surface membrane of one mating partner, (more...)

Membrane Proteins and Cell Adhesion: Concanavalin a Receptors

In 1976, Ofer et al³⁶ demonstrated that the plant lectin concanavalin A (25 micrograms per mL) could inhibit pairing in costimulated Tetrahymena, and cause loosely bound pairs to dissociate. ConA is known to bind to mannose-containing glycoproteins. When conA was conjugated to Fluorescein (FITC-conA) and delivered to a mixed, mating culture, it was found to bind strongly to the developing conjugation junction in Tetrahymena.³⁷ The pattern of conA receptors on Tetrahymena during costimulation and pairing has been carefully documented using both FITC-conjugated conA with fluorescence microscopy,^37-41 and ferritin-conjugated conA with transmission electron microscopy.^37,42 FITC-conA decorates the oral region in nonpairing cells, and is also incorporated into food vacuoles. As cells undergo costimulation, labeling of food vacuoles diminishes (ceasing by 20 minutes),³⁷ suggesting that mating cells cease to feed. Within 15 minutes of mixing, conA-receptors appear in “small clusters near the anterior end, but posterior to the actual tip” (Fig. 9A). Subsequently, punctate conA labeling appears within the region of membrane remodeling located at the anterior tip of the cell (the future conjugation junction), and consolidates into a vertical line (Fig. 9B). Morphologically, this region appears as a region of smooth membrane devoid of the cilia, or membrane ridges that characterize the anterior surface of costimulated, nonmating cells.^9,33 y one hour after mixing, conA labeling is dense and uniform within the, now broadened and transformed tip (Fig. 9C). Pairing follows shortly, and fluorescence becomes restricted to a heart-shaped ring around the perimeter of the conjugation junction³⁹ (Fig. 9D, Fig. 10, Fig. 11).

Figure 9

This figure, (modified from Wolfe and Feng, 1988), depicts the pattern of appearance of concanavalin A receptors at the anterior tip of a costimulated cell. In stage A) receptors appear at the oral apparatus and in patches posterior to the cell's un-modified (more...)

Figure 10

Fluorescence images of conA receptors during tip transformation and pair formation. A) This cell has been costimulated for 1.5 hours. A rhodamine (TRITC)-conjugated conA probe has been employed (red), and a DAPI nuclear counter stain reveals the nucleus (more...)

Figure 11

This image depicts isolated conjugation junctions labeled with FITC-conjugated conA. This preparation was made from sonicated mating pairs produced by Johanna Savage. A) 400X micrograph showing heart-shaped conjugation junctions. B) 600X oil immersion (more...)

It has been pointed out that conA molecules exhibit cross-linking activity, and this could drive the observed “tipping” phenomenon in which conA receptors form aggregates. This has been ruled out by first fixing cells with a variety of agents, and then labeling with conA.^37-39 The study by Watanabe's group is of particular interest in that they demonstrated that when conA labeling is performed prior to observation or fixation, surface labeling is underestimated. Only the most intensely labeled regions appear. When cells are fixed (3 % glutaraldehyde) and then labeled with FITC-conA, a more extensive labeling appears in the anterior end of the cell, though the details of conA tipping and ring formation remain unchanged.

Several studies set out to search for newly synthesized conA-binding glycoproteins as cells underwent costimulation and pairing. Van Bell⁴³ performed extensive SDS-PAGE analysis on mating and costimulating cells and probed gels with ¹²⁵I-conA. Although a number of conspicuous conA-binding proteins were detected, none showed changes in their level of expression during prepairing or conjugation. Wolfe and Feng,⁴¹ isolated cytoskeletal frameworks (triton-extracted cortical residues) and analyzed conA-binding proteins associated with these by Western Blots using conA as a probe. Again, there were no changes associated with pair formation. Other studies identified a number of conA labeling surface proteins (Dentler, 1992). The most recent study by Driscoll and Hufnagel⁴⁴ discovered a conA-labeling 28kD protein associated with ciliary membranes that appears to be down regulated in response to starvation, but again, no conA receptors seemed to be synthesized during the period of costimulation. These studies suggest that costimulation induced conA receptor tipping is brought about not by novel synthesis of the conA receptors themselves, but by reorganization of existing surface glycoproteins resulting in recruitment and aggregation of these membrane proteins at the anterior, conjugation junction. Curiously, general membrane protein synthesis is essential for conA receptor tipping to occur. Pair formation can be blocked by general inhibitors of protein synthesis such as cycloheximide.²² More specifically, tunicamycin, a reagent that blocks protein glycosylation, blocks pair formation as well.⁴⁵ These studies suggest that though conA receptors may already exist diffusely distributed on the surface membrane prior to costimulation, their redistribution and aggregation at the future conjugation junction may depend upon synthesis of other surface membrane glycoproteins. In 1987, Pagliaro and Wolfe⁴⁶ demonstrated that binding of conA to membrane receptors causes certain conA binding proteins to become tightly associated with the underlying cytoskeletal frameworks. Specifically a 23 kD and a 25 kD protein become triton-insoluble when cells are first exposed to conA prior to extraction. This becomes intriguing when coupled with the later discovery that two calcium-binding proteins (TCBP23 and TCBP25) sharing the same molecular weights can be isolated from cortical fractions and appear to decorate the conjugation junction (see below).

One can envision a model in which costimulation induces rapid synthesis of a spatially localized “anchor protein” in the newly formed membrane of the conjugation junction. ConA receptors that were diffusing freely in the plasma membrane would then aggregate at this site and help to mediate cell adhesion while triggering cytoskeletal association of cytoplasmic proteins. In short, conA receptor tipping is one of the first biochemical changes associated with pairing, and it appears to be essential for mediating early cell adhesion events that precede construction of the conjugation junction.

Membrane Events

At the time of pore formation, the junction becomes stabilized and begins to permit cytoplasmic exchange.⁵⁹ The initial pores show fairly consistent dimensions (0.1-0.2 μm in diameter). From casual examination of a number of TEM images, it appears that more pores form over regions of membrane that are adjacent to a nucleus. Over the next 2 hours, pores broaden (possibly fusing with neighboring pores) creating ever larger apertures separated by ever thinner regions of double membrane reticulum. As the transfer pronucleus associates with the membrane of the conjugation junction, that is where we see the most pronounced broadening and fusing of adjacent membrane pores. The largest pores documented by Wolfe,⁹ had diameters of nearly 0.45 μm, though more typically 0.32 μm (ten times the area of the smaller, earlier pores). This transformation (depicted in Fig. 16) reaches an extreme when the membrane reticulum separating the apertures is reduced to a network of branching inter-connected tubules with circular cross section of uniform, 90 μm diameter (Fig. 13). This transformation has been beautifully captured by Judy Orias employing both conventional, and high voltage TEM.^48,49 As junction membranes are transformed into this reticulum of branching spaghetti, the electron dense scaffold proteins disappear, and the membranes appear naked (Fig. 13B,D). It is likely that the curtain-organization of the membrane at this point represents an energetically favored state, and does not require protein scaffolding to maintain it (E. Orias, personal communication).

Figure 16

This illustration provides a model for the membrane remodeling that accompanies formation of the conjugation junction (artwork courtesy of Eduardo Orias). In part (A) 0.1-0.2 μm pores have formed between the two cell membranes. In part (B) junctional (more...)

Figure 13

TEM images of the conjugation junction during pronuclear exchange (reprinted with permission of the publishers from Orias). A) The moment of pronuclear exchange has been captured in this TEM photograph. Black and white arrows lie over the transfer pronculei (more...)

Nuclear-Cytoskeletal Events

As events remodel the membranes of the conjugation junction, the micronucleus of each mating partner undergoes a series of three divisions, meiosis I, meiosis II and a third gametogenic nuclear division that is mitotic in nature.^60-65 During meiosis, there is some form of nuclear assessment going on.⁶⁶ When a nucleus is found “defective” by some, as yet ill-understood criterion (aneuploidy, loss of chromosomes, broken chromosomes), meiosis terminates, and all four meiotic products are destroyed. When a nucleus “passes” assessment, one of the four haploid products becomes tethered to the conjugation junction via microtubules⁶⁷ and/or 14 nm filaments,^68,69 where it is shielded from signals that trigger destruction of the other three meiotic products. The surviving, nucleus initiates a third, gametogenic nuclear division. This tethering has been named “nuclear selection”.⁶⁷ The third nuclear division produces what may be accurately thought of as two gametic pronuclei. The resident pronucleus remains in the cytoplasm of its parent cell, while the transfer pronucleus becomes pressed against the conjugation junction. At this point, both resident and transfer pronuclei of a single mating partner show strands of microtubules between them⁶⁷ and they (and the conjugation junction) are decorated with the proteins fenestrin, and TCBP-25 (see below). The transfer pronucleus bears a striking cytological asymmetry in the electron microscope. On the cytoplasmic side, the pronucleus has acquired a thatched cap of short microtubules cris-crossing to form a type of net (Fig. 13C). Some microtubules even appear to form “T”-junctions with one another (Fig. 14). On the junction side, the pronucleus is nearly devoid of microtubules, and closely pressed against the, now-perforated plasma membrane. It is widely assumed that some form of microtubule-based motility is responsible for the impending nuclear exchange event. Vinblastine and nocodazole (drugs that destabilize microtubules) are effective at preventing nuclear exchange.^70-73 Both fluorescence microscopy using antisera raised against tubulin,67 and electron microscopy⁴⁹ reveal microtubule involvement. Some type of mechanical pressure appears to be generated in that (A) the transfer pronuclei lose their spherical shape and become flattened or lens-shaped as they press up against the conjugation junction, and (B) the junction itself bows out, deforming as the microtubule driven pronuclei press into it from opposing sides (Fig. 13A). The transfer pronuclei pass, side by side at the same anterior-posterior level within the cell⁷⁴ migrate towards the partner's resident pronuclei, and undergo karyogamy (pronuclear fusion). After pronuclear transfer, the conjugation junction is severely disrupted, and the cap of short microtubules and fenestrin that had previously decorated the nucleus remains behind as a “patch” in the gap within the conjugation junction (Fig.15). How the two plasma membranes are reestablished prior to pair separation has never been studied, though membrane pores persist long after nuclear exchange. A model depicting membrane changes during nuclear exchange appears in Figure 16.

Figure 14

This TEM image (courtesy of Judy Orias) reveals a remarkable series of microtubule “T”-junctions associated with the pronuclear basket.

Figure 15

This figure, affectionately named “the morning after”, shows the conjugation junction just after migratory pronuclei (M) have been exchanged (photo courtesy of Judy Orias). Black and white arrow indicates the direction of migration of (more...)

The Conjusome: A p-Granule Nuage Assemblage?

In 1999, Janetopolous et al,⁷⁵ discovered a novel, nonmembrane bound, electron-dense structure located adjacent to the nuclear exchange junction. It appeared late in conjugal development at the time that the developing macronucleus was being remodeled. Antibodies to the chromodomain protein pdd1 (implicated in chromatin remodeling during development)^76-78 labeled this structure.⁷⁵ This structure, the conjusome, begs further inquiry. It resembles p-granules seen in the embryos of Caenorhabditis and Drosophila, or the so-called nuage of higher metazoan embryos. These structures have been implicated in distinguishing germ cells from somatic cell lineages early in the development of numerous metazoan embryos. The molecular machinery that is involved in germ-cell determination of Drosophila has been found in Tetrahymena, where it is involved in the developmental remodeling of the somatic nucleus.^79-81 It is tantalizing to contemplate that the conjusome may represent a ciliate, unicellular version of the nuage assemblage that drives this fundamental developmental decision that distinguishes nuclear fates. It is also interesting to note that the conjusome assembles in the anterior cytoplasm that is associated with the conjugation junction, very near where the golgi assembles early in development.

Fenestrin

This 64kD protein was first identified from a screening of monoclonal antibodies generated against cortical residues of Tetrahymena pyriformis.⁷⁴ The antibody 3A7 cross-reacted with proteins from both T. pyriformis and T. thermophila. In vegetative cells, the 3A7 antibody binds to all the “windows” in the cell: the oral apparatus, contractile vacuole pores, the cytoproct and the fission ring. In mating cells this antibody decorates the conjugation junction on the cytoplasmic face of the membrane in much the same pattern as that seen by FITC-conA on the extracellular face of the membrane, though it remains in the central disk and is not excluded to the perimeter as seen with conA labeling. 3A7 also decorates both the migratory and stationary pronuclei, though not before the third gametogenic nuclear division. Fenestrin appears to adhere to the cytoplasmic face of the sub-membrane scaffolding (ss) first recognized by Wolfe.⁹ It also appears to decorate filaments that extend out from the conjugation junction and appear to anchor the pronuclei. Recently, a group claims to have sequenced the fenestrin gene by mass spectrometry of the protein fragments.⁸⁹

Calmodulin Homologs

Takemasa et al^90,91 cloned the genes for two Tetrahymena calmodulin homologs: TCBP-25 and TCBP-23 respectively. These genes encode EF – hand, Ca²⁺ binding proteins. Polyclonal antisera were generated directly against these proteins that had been expressed in E. coli.^92,93 Both proteins appear to associate with the epiplasm. Furthermore, the antisera to TCBP-25 labels migratory and stationary pronuclei in mating cells, and faintly decorates the exchange junction as well.⁹⁴ This pattern is quite similar to the fenestrin labeling of the conjugation junction, though other “windows” into the cell were not labeled. A similar protein was found in Paramecium, (PCBP 25α)where it was shown to be a target for phosphorylation.⁹⁵ These findings are significant in that they may implicate calcium ions in the dynamic processes associated with nuclear exchange during conjugation.

49 kD Filament-Forming Protein

This curious protein was isolated from Tetrahymena thermophila by in vitro assembly/disassembly.⁹⁶ It appears to form 14 nm filaments in the presence of calcium and ATP. A polyclonal antiserum was generated against it^97,98 and its gene was later cloned.⁹⁹ The sequence revealed it to have homology to citrate synthase and it has been touted as a dual-function protein.¹⁰⁰ Immuno-fluorescence revealed that this protein forms fine cables that radiate from the conjugation junction at the time of nuclear selection, and appear to associate with the gametic pronuclei.^68,69 This protein is of interest for two reasons. First, this protein may be involved in capturing one of the meiotic nuclei, shielding it from programmed nuclear elimination, and tethering it to the conjugation junction for a subsequent gametogenic mitosis. Second, if its assembly is mediated by calcium, then we may see one reason for calmodulin family proteins to also be assembled at the conjugation junction.