The pegs on the decorated tubules of the contractile vacuole complex of Paramecium are proton pumps

Though one of the first organelles to be visualized within Paramecium (see Patterson, 1980) or any cell following the development of the light microscope, the precise function of contractile vacuole complexes (CVCs) has yet to be resolved. It is generally assumed that this organelle system is involved in osmoregulation and that it accumulates and expels a hypotonic fluid. However, the ionic composition of this fluid and how fluid segregation is achieved are not known.

Recent studies have shown that the membranes of the CVC of Dictyostelium discoideum are enriched in vacuolar-type proton pumps (V-ATPase) (Heuser et al., 1993; Fok et al., 1993; Nolta et al., 1993). The V-ATPase is one of three major ATP-hydrolyzing proton pump classes known in nature (see Harvey, 1992; Nelson, 1992), possibly the primordial class (Nelson, 1994). These pumps have been visualized in replicas of quick-freeze freeze-dried preparations of Dictyostelium as 15 nm ‘pegs’ scattered liberally over the cytosolic surface of the membranes of the CVC. This has led to the suggestion that these V-type proton pumps may produce a proton gradient to drive other ions such as sodium, bicarbonate or ammonium ions into the CVC (Heuser et al., 1993). The presence of such pegs on the CVC membranes of Paramecium (McKanna, 1976) and Vorticella (McKanna, 1974) has been known for many years, based on thin-section electronmicroscopic images. Pegs were also reported in the CVC of the fresh water sponge Spongilla lacustris (Brauer and McKanna, 1978). We have frequently observed indications of these pegs in our own studies of Paramecium both in thin-section and freezefracture electron microscopy (Hausmann and Allen, 1977; Allen et al., 1990). They decorate the bundles of membrane tubules that are collectively known as the decorated spongiome. These bundles are connected to a network of smooth tubules (collectively called the smooth spongiome) along the collecting canals. All of these structures are components of the 5 to 10 radial arms of each CVC.

Recently, we showed that in this ciliate these bundles of decorated membrane tubules form the locus of fluid segregation (Ishida et al., 1993). In the present study and by using two antibodies known to be specific for the V₁ complexes in other cells, we show that these pegs on the decorated tubules of Paramecium are, in fact, the V₁ complexes of the V-ATPases. We also show: (1) that the ATPase inhibitor, concanamycin B, inhibits the CVC’s fluid output but does not affect the integrity of the decorated tubules; and (2) that cold treatment, known to cause the dissociation of the isolated V₁V₀ complexes, causes the in vivo dissociation of the decorated tubules so that the antigens become dispersed throughout the cytosol.

Cells of Paramecium multimicronucleatum, syngen 2, cultured in an axenic medium (Fok and Allen, 1979), were harvested at mid-log phase of growth. Concanamycin B, a potent and specific inhibitor of ATP-dependent acidification of macrophage lysosomes and endosomes (Woo et al., 1992), was a gift from Dr K. Miwa (Central Research Laboratories, Ajinomoto Co., Japan). Secondary antibodies were purchased from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA). All other chemicals were obtained from Sigma Chemical Co. (St Louis, MO).

Monoclonal antibody (mAb) DS-1, whose specificity has been described (Allen et al., 1990), was used to identify the decorated tubules of the decorated spongiome. For simplicity, the antigen to which DS-1 binds will be called A4. Monoclonal antibody SS-1 (Clone C38.2.2) was obtained according to standard procedures for hybridoma production (Kohler and Milstein, 1975). This mAb labeled the smooth spongiome at the light microscopic level and will be called the G4 antigen. In contrast to the DS-1-specific A4 antigen, the SS-1-specific G4 antigen was not reactive after glutaraldehyde fixation. Also, neither the A4 nor the G4 antigens under either reduced or nonreduced conditions were detectable on immunoblots.

A polyclonal antibody that was raised against the V-ATPase subunits of Dictyostelium discoideum, and affinity purified using purified V-ATPase (Dictyostelium proton pump antibody; Nolta et al., 1993) was used to identify the general location of the V-ATPase in Paramecium. A monospecific polyclonal antibody against the 57 kDa B-subunit of the V-ATPase of chromaffin granules (Nelson, 1992), referred to in this paper as anti-57 kDa antibody, was a gift from Dr N. Nelson (Roche Institute of Molecular Biology, Nutley, NJ). Antigens for these two polyclonal antibodies were determined using SDS-PAGE and immunoblotting as described (Fok et al., 1988).

Living cells were concentrated, placed on copper mounts and slamfrozen onto polished copper blocks at −195°C. Fractures, etching and platinum-carbon replicas were made as described (Allen et al., 1989). Replicas were viewed in a Zeiss 10A transmission electron microscope (TEM).

For indirect immunofluorescence experiments, Paramecium was fixed in 3% formaldehyde in 50 mM phosphate buffer (pH 7.2), and acetone permeabilized at −20°C for 20 minutes. These cells were then incubated sequentially with a primary antibody followed by a secondary antibody, as previously reported (Allen et al., 1990). For double labeling, cells were incubated first with a mixture of two primary antibodies, followed by their respective secondary antibodies. Cells were examined in a Zeiss light microscope (Thornwood, NY) equipped with epifluorescence illumination.

For immunoelectron microscopy, cells were fixed in 0.25% gultaraldehyde in 50 mM phosphate buffer (pH 7.2) at room temperature for 15 minutes. These cells were pelleted and embedded in 4% gelatin and processed for cryosectioning and immunogold staining as described (Allen et al., 1990). To reduce the nonspecific staining, all sections were preincubated with 5% normal goat serum and/or bovine serum albumin prior to incubation with a primary antibody.

The expulsion frequency and the diameter of the contractile vacuole were measured in cells whose motility was restricted. Fluid output of each contractile vacuole was obtained by multiplying its expulsion frequency by its maximum volume at late diastole (Ishida et al., 1993).

A stock solution of concanamycin B (10 mM) in dimethylsulfoxide was diluted immediately before use in axenic culture medium which lacked phosphatidylethanolamine. Preliminary experiments had shown that the potency of this inhibitor was much lower when phosphatidylethanolamine was present. This could be caused by the binding of concanamycin B to the phospholipid vesicles thereby lowering its effective concentration. To determine the dose effect of concanamycin B on the total contractile-vacuole fluid output, cells were incubated in varying concentrations of concanamycin B while keeping the dimethylsulfoxide concentration at 0.01% or less.

For cold treatment, cells in axenic culture medium were quickly cooled to 12°C or 1°C for 30 minutes. Cells were then fixed for immunofluorescence and cryoimmunogold labeling as above. For recovery experiments, cells were returned to room temperature and the fluid output of their contractile vacuoles was measured.

Decorated tubules of the decorated spongiome of the CVCs of Paramecium are often arranged as bundles of 50 nm diameter, membrane-bound cylinders. These cylinders are ‘decorated’ by rows of ∼15 nm globular pegs helically wound around their cytosolic surfaces when viewed in both conventional thin sections (Fig. 1A) as well as in cryo-thin sections (Fig. 1B). However, they can be visualized in greatest detail using the quick-freeze deep-etch (QF-DE) technique. By QF-DE, it can be seen that ridges formed from a two-start helical pattern of 30 nm wide compact protrusions are separated by grooves of 10 nm in width (arrow, Fig. 1C). This is reminiscent of the F₁F₀ subunits on the mitochondrial cristae of the same cell, although a one-start helix with a longer pitch is found on the cristae (Allen et al., 1989). Each ridge is composed of a double row of pegs spaced at 15 nm (center-to-center). Thus, using a formula for the length of a helix (Allen et al., 1989), the number of pegs along a decorated tubule can be calculated to be ∼467/μm. For each micrometer of a decorated tubule, the membrane surface area and the enclosed volume (inset b, Fig. 1C) are about 0.16 μm² and 7×10⁻⁴μm³, respectively. Thus, the ratios of membrane surface to luminal volume and of pegs to luminal volume are very high along the length of these tubules.

At high magnification in our best QF-DE replicas, each peg appears to be composed of several subunits, and many resemble tetrads (arrow, Fig. 1C) or three rows of diads (insert a, Fig, 1C). In favorable fractures, apparent luminal extensions of the protein complexes can be seen in a region where the tubule has been fractured obliquely (arrowhead, Fig. 1C, and small arrow in inset a). These luminally exposed globular units are 11 nm in diameter.

To estimate the cell’s total load of pegs on its two CVCs, we obtained a rough approximation of the total length of decorated tubules in 16 randomly selected cross sections of CVC radial arms. The average total length of such tubules was determined to be roughly 16 μm per 70 nm thick cryosection. Earlier measurements had shown the average length of the decorated spongiome obtained from measuring 505 radial arms to be 37.4 μm in interphase cells (Allen et al., 1990). From these numbers we could roughly estimate that there were 8,600 μm of decorated tubules per radial arm. At 467 pegs per μm of decorated tubules, an average radial arm would contain around 4 million pegs. Since there are 5 to 10 radial arms per CVC, and two CVCs per cell, a rough estimate of the total load of pegs per cell would be between 40 and 80 million.

Previous work in our laboratory has shown that mAb DS-1 to the A4 antigen specifically labels the decorated tubules of the CVC (Allen et al., 1990). Unfortunately, A4 was shown to be unreactive on blots of either whole cell homogenates or isolated membrane fractions known to be enriched in the decorated tubules. Thus the identity of this antigen could not be determined. To determine if the pegs were part of the V-ATPase, we used antibodies specific for the V-type proton pumps of other organisms to see if they would cross-react with the decorated tubules of Paramecium. The Dictyostelium proton pump antibody reacted with a 75 kDa polypeptide in blots of Paramecium (Fig. 2, lane 1) and strongly labeled the decorated tubules (Figs 3 and 4). Double labeling showed that both the A4 (Fig. 3B) and the 75 kDa antigens (Fig. 3A) were located on this decorated spongiome, while double exposure showed the colocalization most dramatically (Fig. 3C). Similar results were obtained using cryo-thin sections where antigens were located on the decorated tubules singly (Fig. 4A,B) or in double labeling experiments (Fig. 4C).

To confirm that the decorated tubules have authentic V-type proton pump polypeptides, we obtained a monospecific polyclonal antibody to the 57 kDa B-subunit of the V₁ complex of chromaffin granules. Immunoblotting showed that this antibody cross-reacted with a somewhat larger 66 kDa polypeptide in Paramecium (Fig. 2, lane 2). This anti-57 kDa antibody gave essentially the same labeling pattern as mAb DS-1 (compared Fig. 5A with Fig. 3B) or as the Dictyostelium proton pump antibody by both immunofluorescence (compare Fig. 5A with Fig. 3A) and immunogold labeling (Fig. 6D).

With all three antibodies, the label was mostly on the cytosolic sides of the tubules of the decorated spongiome where the 15 nm pegs are located (Fig. 6A-D). None of the membranes of the smooth spongiome, the collecting canals, or the contractile vacuole itself was labeled, and nonspecific background labeling was minimal. These results showed that the pegs on the decorated tubules of Paramecium are, beyond doubt, components of a V-ATPase.

Concanamycin B is a macro-cyclic lactone structurally related to the V-ATPase inhibitor bafilomycin A. The latter drug was the first specific V-ATPase inhibitor to be discovered and was found to be effective in fungi, plants, and in animal cells (Bowman et al., 1988), but it had no noticeable effect on the function of the CVC in Paramecium (unpublished observation). Concanamycin B, which was found to potently and specifically inhibit the V-ATPase activity of endosomes and lysosomes in J774 macrophages (Woo et al., 1992), was shown to reversibly block the CVC function in Dictyostelium discoideum (Heuser et al., 1993). In the present study, we found that it also inhibited the fluid output of the contractile vacuoles of P. multimicronucleatum in a dose-dependent manner after cells were exposed to this drug for 30 minutes (Fig. 7A). Half maximal inhibition was achieved at approximately 15 nM of the drug. In one experiment where cells were exposed to 30 nM concanamycin B for 30 minutes and then transferred to a hypotonic solution containing this same concentration of drug, the cells became swollen and lysed (data not shown). Thus, these cells were not able to protect themselves from a hypoosmotic shock when the V-ATPase was inhibited.

The inhibitory effect of concanamycin B on fluid output in Paramecium was reversible. After a 30 minute incubation in 30 nM concanamycin B, cells transferred to a fresh axenic culture medium began to recover their CVC’s activity soon after washout. Recovery was essentially complete within 3 hours. This dramatic and reversible inhibitory effect produced by concanamycin B indicated that an ATP-dependent acidification mechanism, the putative V-ATPase, is intimately involved in the water segregation of the CVC of Paramecium. To test for the effects of concanamycin B on the structure of the CVCs, cells were treated with this drug (30-1,000 nM) for 30 minutes, and then fixed and processed for immunolabeling with mAb DS-1. Interestingly, concanamycin B had almost no effect on the structure of the decorated spongiome (Fig. 5B).

It is known that cold causes the dissociation of some isolated proton pump complexes into their subunit polypeptides. This was first shown for the F₁F₀-ATPase of mitochondria (Pullman et al., 1960) and eubacteria. More recently, it has also been shown that cold together with ATP and MgCl₂ cause the dissociation of the V₁ from the V₀ subunit of the V-ATPase of chromaffin granule membrane vesicles (Moriyama and Nelson, 1989). To see what effect cold treatment would have on the decorated tubules in vivo, Paramecium was incubated at 12°C or 1°C for 30 minutes, fixed and then processed for immunofluorescence using mAbs DS-1 and SS-1. Functionally, when the cells were at 1 °C for 30 minutes, all fluid output stopped. However, upon warming, fluid output rapidly resumed, returning to normal in a few hours.

The morphology of the decorated spongiome in these cells could be divided into three categories. The first category consisted of cells with normal morphology (Fig. 8A). The second category consisted of cells with remnants of the decorated spongiome along the radial arms but the decorated spongiome appeared to be in the process of dissociating or disintegrating (Fig. 8C). The last group consisted of cells that had no discernible decorated spongiome and the A4 antigen was dispersed throughout the cytosol (Fig. 8E). That the smooth spongiome was still intact in all cases is shown in Fig. 8B,D and F. The distribution of cells in the three structural categories is shown in Fig. 9. At ambient temperature, less than 10% of the cells were in category 2. Some of these cells might have been in the process of cell division, as it is known that the CVCs undergo partial dissociation (Allen et al., 1990) during this stage. After 30 minutes at 12°C, almost half of the cells had a decorated spongiome that belonged to category 2. When cells were exposed to 1°C, almost 90% of the cells had a decorated spongiome belonging to either category 2 or 3. Fig. 10 shows that the decorated spongiome is entirely absent in a cell exposed to 1°C for 30 minutes that was in category 3. These data verified that cold temperature had little effect on structures of the CVC other than the decorated spongiome which, for the most part, was dissociated from the smooth spongiome (Fig. 10). However, whether the V₁ subunits themselves were dissociated from their basepieces in the dispersed decorated spongiome could not be determined in this study.

Identical results were obtained when these cells were labeled with either Dictyostelium proton pump antibody or the anti-57 kDa antibody (data not shown).

It has been known for some time that the decorated tubules of the CVCs in Paramecium are decorated with pegs (McKanna, 1976). Using QF-DE we confirm here that the decorated tubules support two helically wound arrays of pegs (Fig. 1C). These pegs are similar to those in the CVC of Dictyostelium in both appearance and in size. The pegs in Dictyostelium were identified as the V-ATPase by showing their cross-reactivity with the antibody to the B subunit of the V-ATPase of chromaffin granules (Heuser et al., 1993). However, while it might be reasonable to speculate that the pegs on the decorated tubules of the CVCs in Paramecium are also proton pumps, definitive evidence for this conclusion was lacking.

In this report, we provide four lines of evidence to show that the pegs on the decorated tubules of Paramecium are parts of these proton pumps. Firstly, these pegs have a topography and structural complexity that closely resembles the F₁F₀ proton pumps on the cristae of the mitochondria of Paramecium (Allen et al., 1989). Both the pegs on the CVC and the proton pumps in mitochondria are complex globular units of ∼15 nm and 10 nm, respectively, which are located on the cytosolic or matrix sides of the membrane. At high magnification in QF-DE replicas, the pegs appear similar to the F₁ complex in the mitochondria and each peg appears to be composed of several subunits. Pegs in both the CVC and in the mitochondria exist as two closely spaced rows that wind around the supporting membrane tubules in a helical pattern, although, in the CVC the pegs are arrayed in a two-start helical pattern. Secondly, an affinity-purified Dictyostelium proton pump antibody and an antibody to the B subunit of the chromaffin granule V-ATPase both cross-react with these pegs. Immunological labeling of cryosections shows that the antigens are located predominantly at the surface of the membrane tubules where the pegs of the decorated spongiome are located (Fig. 6A-D). These results show that the pegs are probably part of the V₁V₀ proton pumps. Thirdly, concanamycin B, a specific V-ATPase inhibitor, completely and reversibly inhibited the fluid output of the CVCs. Lastly, subjecting cells to a cold temperature results in the dissociation of the decorated spongiome from the smooth spongiome so that the A4 antigen, as well as the 75 and 66 kDa polypeptides of the proton pump are found dispersed in the cytosol instead of being specifically associated with the intact smooth spongiome along the collecting canals.

In this study, we determined that Paramecium, like Dictyostelium, has a large number of pegs/proton pumps associated with its CVC membranes. However, unlike Dictyostelium where most of the membrane compartments of the CVC bear the pumps, the pumps in Paramecium are found only on the decorated spongiome, the locus of water segregation in this cell. The smooth spongiome, collecting canals, ampullae and the contractile vacuole, presumably have other functions such as ion exchange, fluid transport, fluid storage, or membrane storage, as they are normally devoid of proton pumps. Thus, it appears there may be even more ‘division of labor’ in the CVC of Paramecium than has been described in the CVC of Dictyostelium (Nolta and Steck, 1994).

The decorated tubules in Paramecium have an extremely high proton pump-to-luminal volume ratio. We estimate there are 6.6×10⁵ pumps per μm³ of lumen or 467 pumps per μm of tubule length. This comes out to 1 proton pump for every 1,500 nm³, which is roughly the volume of one V₁ complex. One Paramecium cell of average size has roughly 1.3×10⁵μm of total decorated tubule length or a total of 40 to 80 million proton pumps in its two CVC. Such a large number of pumps around a total volume of 85.7 μm³ has the potential of moving a great number of protons into a very small volume within a very short time. Yet the pH of the culture medium in which the Paramecium cells are multiplying exponentially and which contains the fluid expelled by the CVs not only does not become acid but actually becomes slightly alkaline with time (A. K. Fok, unpublished observation). Thus, any protons generated by these pumps must be quickly removed from solution or neutralized by anions present in the CVC. It seems reasonable to begin to look for ion exchangers in the CVC which could be driven by the potential energy developed by this very large number of proton pumps.

Our results from the experiments using concanamycin B, a specific V-ATPase inhibitor, show that this ATPase activity is necessary for fluid segregation to continue. This drug strongly inhibits fluid segregation, but has minimal effect, if any, on the integrity of the decorated tubules. On the other hand, cold treatment results in the loss of the decorated spongiome from around the smooth spongiome without affecting the smooth spongiome itself. In preliminary TEM experiments we have seen evidence of vesiculated decorated tubules which apparently have lost their V₁ complexes as they are no longer labeled with mAb DS-1. These experiments as well as the effects of hyperosmotic treatment on the CVC are currently under study. In summary, while our earlier paper showed that the decorated tubules are the sites of fluid segregation, the results of the present report demonstrate that the pegs on the decorated tubules are proton pumps whose integrity and whose ATPase activity are essential for ongoing fluid segregation.

We thank Dr N. Nelson, Roche Institute of Molecular Biology, Nutley, NJ, for the gift of the anti-57 kDa antibody and Dr K. Miwa, Ajinomoto Co., Kanagawa, Japan, for the gift of concanamycin B. We also thank Dr Y. Naitoh, Tsukuba, Japan, for stimulating and helpful discussions. This work was supported in part by National Science Foundation grants MCB 90 17455 (to A.K.F.) and MCB 92 06097 (to R.D.A.). The Biological Electron Microscope Facility is partially supported by an NIH RCMI grant RR-03061 and, in the past, through NSF instrumentation grants.